Bringing Finance to Pakistan s Poor:

Transcription

1 Puli Dislosure Authorized Puli Dislosure Authorized Puli Dislosure Authorized Puli Dislosure Authorized Bringing Finne to Pkistn s Poor: A Study on Aess to Finne for the Underserved nd Smll Enterprises My Ttin Nenov Ceile Thioro Ning with Anjum Ahmd

2 Contents Arevitions Aknowledgements Exeutive summry Reommendtions summry vi viii ix xiii Chpter 1 Overview, Finnil Mrket Struture, Regultions, nd Poliies Expnding Aess to Finne, Links to Growth, nd Poverty Redution 1.2. Ojetives of the Report 1.3. Pkistn rief mrket overview 1.4. Bnking Setor 1.5. Mirofinne 1.6. Islmi finne 1.7. SME finne 1.8. Remittnes 1.9. Finnil inlusion nd the new finnil regultory nd tehnologil infrstruture Chpter 2 Aess to Finne: Evidene from the Demnd Side Vriility of finnil ess ross rurl/urn res, provines, inome, gender, nd other popultion hrteristis 2.2. High interest in finnil servies nd low finnil litery 2.3. Low nk ess ntionwide, due to lk of inome / informtion 2.4. Informl finnil servies 2.5. Svings nd investment 2.6. Borrowing 2.7. Insurne 2.8. Pyment nd hek servies 2.9. Suggested strtegies for expnding ess to finne for the underserved Chpter 3 Aess to Finne for the Underserved Providers of finnil servies 3.2. Produts 3.3. Mirofinne lients nd redit growth trends 3.4. Soures of finning 3.5. MFI performne 3.6. Mirofinne nd Moile Telephone Tehnology in Pkistn 3.7. Informl finne 3.8. Regultion 3.9. Growing the mirofinne setor Chpter 4 Improving Finnil Aess for SMEs Aess to finne for SMEs: Supply-side evidene 4.2. Aess to finne for SMEs: Demnd-side evidene 4.3. Constrints for improving SMEs ess to finne 4.4. Helping nks serve SMEs Chpter 5 Hrnessing Remittnes for Aess to Finne Overview of the remittne mrkets 5.2. Mrket plyers 5.3. Distriution of remittnes y region nd inome level

3 5.4. Determinnts of remittnes nd historil trends 5.5. Usge ptterns nd servie fees 5.6. Outreh nd sope of remittne tehnologies 5.7. Poliy nd regultory frmework 5.8. Suggested venues for tion Chpter 6 Expnding Aess to the Underserved: An Ation Pln The Role of the Privte Setor 6.2. The Role of the Puli Setor 6.3. The Role of Puli-Privte Prtnerships Annex to Chpter Annex to Chpter Annex to Chpter Annex 4.1 Smple desription of KfW SME Survey Annex 4.2 Frmework for movle ollterl in Pkistn Annex to Chpter Appendix A Dt Methodology nd Clirtion 140 Appendix B A2F Questionnire 152 Vrile Desription 162 Biliogrphy 164 List of Boxes Box 1.1. Finne nd Link to Growth Box 1.2. Household Survey for Pkistn (Demnd Side Dt) Box 1.3. Other Progrms in Pkistn for Finnil Inlusion Box 2.1. Fts out Segmenting Pkistni Femle Clients Box 3.1. The story of Mirofinne in Pkistn Box 3.2. Lessons from Suessful Postl Finnil Systems Box 3.3. Cse study: Tehnology Innovtions hve Improved Bk-End Proessing nd Expnded Delivery Chnnels in the Mirofinne Vlue Chin Box 3.4. Bsi Bnking in Indi, Mexio, nd South Afri Box 4.1. SME Bnk Box 4.2. Ntionl Bnk of Pkistn SME Lending Box 4.3. How Big is the Potentil Mrket for Providing Credit to SMEs? Box 4.4. Chin SME Lending Projet Box 5.1. A Closer Look t HBL Box 5.2. Innovtive UBL Produts Box 5.3. Zro Exhnge Brnhing Out, Inluding Online Box 5.4. Indin NGO Helping Domesti Remittnes Box 5.5. Moile Bnking: G-Csh (Philippines) nd M-Pes (Keny) Box 5.6. Kiosks to Help Rurl Outreh List of Tles Tle 1.1 Bsi Finnil Inditors, Interntionl Comprison Tle 1.2 Brekdown of Lons (Domesti Opertions) y Setor Tle 1.3 Comprtive Position of Numer of Brnhes in Pkistn (De 2007) Tle 1.4 Rurl Credit Mrket Tle 1.5 Frmer Aess to Finne Tle 1.6 Miro Finne Outreh in Pkistn Tle 1.7 Islmi Bnking Plyers

4 Tle 1.8 Credit to Privte Setor, nd Growth in Lending Tle 2.1 Aess of to Finning, y Servie Tle 2.2 Finnil Litery Tle 2.3 Litery Mesures in Rurl nd Urn Ares Tle 2.4 Forml nd Informl Borrowing Tle 2.5 Purpose of the Lon (perent of ll orrowers) Tle 2.6 Ftors Considered efore Tking Personl Long Term Lon Tle 2.7 Rejetion Rtes for Existing Borrowers (% of ll forml orrowers) Tle 3.1 Mirofinne Penetrtion ross Asi Tle 3.2 Mrket Plyers Tle 3.3 Distriution of Miro-lients Tle 3.4 Key Performne Inditors of MFBs (%) Tle 3.5 Moile Telephone Susriers Tle 3.6 Interest Rtes Chrged in the Informl Mrket in Pkistn, 2005 Tle 4.1 Distriution of SME Finne mong Bnks Tle 4.2 Brekdown of Lending to SME y Type of Fility (Rs illion) Tle 4.3 Credit to Privte Setor: A Profile (Rs., Billion) Tle 4.4 Ese of Otining Credit Tle 4.5 Business Development in the Pst 24 Months Tle 4.6 Demnded Lon Produt Fetures Tle 4.7 SME Applition Proess for Bnk Lons Tle 5.1 Snpshot of the Min Chnnels of Interntionl Forml Money Trnsfer Tle 5.2 Perentge Post Offies y Provine nd Rurl / Urn res Tle 5.3 Proportion of Rurl vs / Urn Bnk Brnhes Tle 5.4 Regionl Distriution of Remittnes Tle 5.5 Shre of Remittnes in Household Inome for the Poorest Quintile (%) Tle 5.6 Remittne Growth Tle 5.7 Cost of Sending $300 to Pkistn through Vrious Chnnels List of Figures Figure 1.1 Deposits, Lons nd Assets of the Bnking System Figure 1.2 Credit y Setor Figure 1.3 Trends in Sving Rtes Figure 1.4 Bnking Setor Penetrtion Figure 1.5 Bnking Setor Performne Figure 2.1 Pkistn Aess Strnd Figure 2.2 Shre of Bnked, y provine Figure 2.3 Women nd Mirofinne Figure 2.4 Produt Penetrtion Figure 2.5 Csh-sed Inome ross Setors Figure 2.6 Produt Penetrtion (s % of nked) Figure 2.7 Cost of Bnk Aess Figure 2.8 Resons for Prtiipting in Committee Figure 2.9 Government / Privte Svings Aounts Figure 2.10 Requirements Before Reeiving Lon Figure 2.11 Soures of Funding of Emergeny Pyment Figure 2.12 Resons for not Hving Insurne (% of not insured) Figure 2.13 Mens of Reeiving Inome Figure 3.1 Numer of Borrowers nd Lon Portfolio y Lending Methodology Figure 3.2 Soures of Finning for MFIs Figure 3.3 Averge Lon Size Figure 3.4 Cellulr Penetrtion nd Projeted Moile Telephone Trends y Provine

5 Figure 4.1 Aess to Finne for SMEs in Pkistn, Indi nd Bngldesh Figure 4.2 Externl Soures of Working Cpitl Figure 4.3 Externl Soures of Fixed Cpitl Figure 4.4 Lon Applition nd Disursement y Firm Size Figure 4.5 Bnk Aess Inreses with Wider Rnge of Aeptle Assets to Seure Lon Figure 5.1 Interntionl Forml Remittnes Figure 5.2 Remittnes Interntionl Comprison Figure 5.3 Top Domesti Remittne Chnnels Figure 5.4 Remittnes y Country Figure 5.5 Overses Workers By Ctegory Figure 5.6 Moile Consumer Bse y Provine Figure 5.7 Moile Penetrtion y Provine List of Annex Tles Annex Tle A1.1 Indies of Finnil Aess (men vlues y region nd y ountry) Annex Tle A1.2 Weighted Averge Lending nd Deposit Rtes, My 2008 Annex Tle A1.3 Trends in Infltion Annex Tle A2.1 Determinnts of Aess to Forml nd Informl Finnil Servies, Proits Annex Tle A2.2 Determinnts of Aess to Forml nd Informl Finnil Servies, Multinomil Logits Annex Tle A2.3 F-sttisti omprisons of likelihood for finnil ess Annex Tle A2.4 Determinnts of Interest in Finnil Mtters Annex Tle A2.5 Determinnts of Willingness to Enter the Finnil Setor Annex Tle A2.6 Determinnts of Aess to Bnking Annex Tle A2.7 Determinnts of Aess to Forml nd Informl Borrowing Annex Tle A2.8 Determinnts of Aess to Forml nd Informl Svings Annex Tle A2.9 Determinnts of Aess to Speifi Forml nd Informl Finnil Produts Annex Tle A2.10 Religious Considertions s Determinnt of Aess to Finne, for Rnge of Produts Annex Tle A2.11 Determinnts of Aess to Finne of the Self Employed Annex Tle A2.12 Resons for Sving Annex Tle A2.13 Resons for Borrowing Annex Tle A4.1: Prinipl Setor of Respondent s Business Annex Tle A4.2: Definition of Firm Size Annex Tle A4.3: Legl Sttus of Smpled Enterprises Annex Tle A4.4: Lotion of Business of Smpled Enterprises Annex Tle A4.5: Averge Yers of Shooling of Owners of Smpled Enterprises Annex Tle A5.1 Workers Remittnes y Setor (Millions US$) Annex Tle A5.2 Inlnd Money Orders (Figures in Millions) Annex Tle A5.3 Postl Orders (Figures in Millions) Annex Tle A5.4 Rurl Vs Urn Split of Bnk Brnhes in Provines Annex Tle A5.5 Workers Remittnes to Pkistn y FY nd Country (Millions US$) Annex Tle A5.6 Perentge Shre of Remittnes in Household Inome (PIHS ) Annex Tle A5.7 Remittnes, Contriution to Household Inome nd Poverty Rtio Annex Tle A5.8 Determinnts of Remittnes in Pkistn Annex Tle A5.9 Orgniztions tht re Helping nd Funding the Glol Remittnes Mrket

6 List of Arevitions A2F Aess to Finne (survey) ABL Allied Bnk Limited ADB Asin Development Bnk AJK Azd Jmmu nd Kshmir AML Anti-Money Lundering ATM Automted Teller Mhine CIB Credit Investment Bureu DFI Development Finne Institutions DFID Deprtment for Interntionl Development EC Exhnge Compny FANA Federlly Administered Northern Are FATA Federlly Administered Tril Are FBS Federl Bureu of Sttistis FIP Finnil Inlusion Progrm FY Fisl Yer GDP Gross Domesti Produt GOP Government of Pkistn HBL Hi Bnk Limited HIES Household Inome Integrted Survey ICA Investment Climte Assessment MCB Muslim Commeril Bnk MFB Mirofinne Bnk MFI Mirofinne Institution MTO Money Trnsfer Orgniztion NBP Ntionl Bnk of Pkistn NGO Nongovernmentl Orgniztion NIFT Ntionl Internk Finnil Teleommunition NRSP Ntionl Rurl Support Progrm NSS Ntionl Svings Sheme NWFP North West Frontier Provine OECD Orgniztion for Eonomi Co-opertion nd Development PAR Portfolio t Risk PLS Aount Profit nd Loss Shring Aount PMN Pkistn Mirofinne Network POS Point of Sle PPAF Pkistn Poverty Allevition Fund PPSB Pkistn Post Svings Bnk PRISM Pkistni Reltime Internk Settlement Mehnism PSLM Pkistn Soil nd Living Stndrds Mesurement Survey PSU Primry Smpling Unit ROA Returns on Assets ROE Returns on Equity ROSCA Rotting Credit nd Svings Assoition RSP Rurl Support Progrm Pge vi

7 SBP SECP SME SMS SSU SWIFT UAE UBL USA Stte Bnk of Pkistn Seurities nd Exhnge Commission of Pkistn Smll nd Medium Enterprises Short Messge Servie Seondry Smpling Unit Soiety for Worldwide Internk Finnil Teleommunition United Ar Emirtes United Bnk Limited United Sttes of Ameri Curreny onversion: Exhnge rte t the time of the study (De 2008) ws Rs 1=$ All dollr mounts re in U.S. dollrs. Pge vii

8 Aknowledgements This report owes muh to our ollegues t Stte Bnk of Pkistn, espeilly Dr. Seed Ahmed, Diretor of the Finnil Inlusion Progrm. The dt olletion would hve not een possile without the guidne, onsite ssistne, dvie, dt provision, nd expert sttistil ontriution of FBS. The oopertion, ssistne, nd joint work with DFID, FinMrk Trust, SDC, nd PMN hve een invlule. The oordintion nd ledership of the survey work y Ftimh Afzl of PMN hs ridged the oundries of the impossile in pioneering this mjor new dt-gthering exerise in the ountry. The A. C. Nielsen (the survey house) stff, nd espeilly Fiz Jmil, hve put in long hours of dedited effort nd diligene. Finlly, the tem thnks ll mjor Pkistni Bnks nd other finnil institutions, who took time to disuss nd rinstorm on pertinent issues, s well s vrious stkeholder groups nd espeilly the Pkistni men nd women who ptiently went long with our enumertors during the pilot lunh of the survey. The effort hs enefited from le dvisers inluding Hroon Shrif (DFID), Zoi Andrew (DFID), Sturt Andrew (DFID), Srh Hennell (DFID), Srh Zk, Gregory Chen (Shorenk), Lio Solh (KfW), Mihel Kortenush (BFC), Khlid Sirj, Ijz Ni, Fisl Bri, Nmoos Zheer, Isfndyr Zmn Khn, Anjli Kumr, Alsi Demirgu-Kunt, Consolte Rusgr, P.S. Shrinivs, Yoko Doi, Shn Khwr, Kspr Rihter, Zhid Hsnin, Nouo Yoshid, Tomoyuki Sho, Pul Wde, Bo Cull, Bill Zi, Ann Rennie, nd Mhesh Uttmhndni nd Rhel (Rh) Shhid-Sless (who together ontriuted the expert nnex on seured trnstions). Eri Mnes ws instrumentl with dvie nd dt for Chpter 4. Kirn Afzl provided outstnding reserh ssistne. The tem owes prtiulr ppreition to Shmsuddin Ahmd nd Simon Bell for overll guidne, review, nd dvie. The exellent support from Az Rshid, Skm Adul Hye, Imtiz Ahmd Sheikh, Ruin Geizl Quner, Mrjorie Espiritu, nd Mrgret Murry hve ove ll mde this work possile. The hpter uthors re Anjum Ahmd nd Ttin Nenov (Chpter 1), Ceile Thioro Ning nd Ttin Nenov (Chpter 2), Stephen Rsmussen (Chpter 3), Auror Ferrri nd Mukt Joshi (Chpter 4), nd Mrtin Aslop (Chpter 5). Shnthi Divkrn ontriuted to reserh for Chpter 1. Ceile Thioro Ning further ontriuted the tehnology setions to ll hpters. Ceile Thioro Ning, Mukt Joshi, nd Anjum Ahmd tirelessly helped with the overll reserh, report drfting, nd review. Ttin Nenov oordinted, edited, nd led the survey nd report work. Pge viii

9 Exeutive Summry Aess to finning is now widely knowledged s pth to meningful eonomi inlusion nd redution in poverty. Poliy efforts to inrese ess to finne in Pkistn hve tken time to er fruit, ut now ess is indeed expnding quikly in ertin finnil setors (mirofinne, remittnes) leit from very low se. Nevertheless, poliy mesures nnot single-hndedly inrese finnil ess; finnil institution willingness to expnd ess in Pkistn hs een stinted y slow tehnologi dvnes, wek legl foundtions, nd unsuitle finnil proesses nd produts. Poor soioeonomi onditions, gender is, nd low levels of si edution nd finnil litery remin rriers, ut perhps the single strongest driver of low demnd for finnil ess hs een inome. The verge Pkistni household remins outside the forml finnil system, sving t home nd orrowing from fmily or friends in ses of dire need. Fourteen perent of Pkistnis re using finnil produt or servie of forml finnil institution (inluding svings, redit, insurne, pyments, nd remittne servies). When informl finnil ess is tken into ount, 50.5 perent of Pkistnis hve ess to finne. Informl ess n our through the orgnized setor (though ommittees, shopkeepers, moneylenders, hwl/hundi money trnsfers, nd so forth), or informlly through friends nd fmily. In omprison, 32 perent of the popultion hs ess to the forml finnil system in Bngldesh, nd this figure mounts to 48 perent in Indi nd 59 perent in Sri Lnk (World Bnk, 2008). Of the nerly 50 perent of Pkistnis who do not engge in either the forml or informl finnil system, we estimte out 19 perent hve voluntrily exluded themselves through lk of understnding, wreness, or need, due to poverty, or for religious resons. Finnil exlusion preludes people from reduing risk, mnging flututions in inome, nd investing in miroenterprises or in helth nd edution Shre of the popultion with forml finnil ess Bngldesh Chin Indi Indonesi Kore Mlysi Nepl Pkistn Philippines Singpore Sri Lnk Thilnd Finnilly Served Voluntrily exluded Formlly Inluded Informlly Served (orgnized setor) Inf. Served (unorg. setor) Finnilly Exluded 0% 14.3% 45.5% 50.5% 81.4% 100% Mjor onstrints to finnil ess, in spite of poliy reforms, rise from the high levels of poverty, omined with low wreness of nd informtion out ville finnil servies, s well s gender is. Currently, finnil institutions limit their servies expnsion to individuls nd enterprises with high nd preditle inome. And yet there is nothing inherently un-serviele out low-inome, informlly employed, rurl, or femle lients. In ft, the onsiderle gender is is ompletely eliminted mong formlly-servied individuls. Differenes etween urn nd rurl finnil inlusion re ompletely eliminted one the effet of inome nd other individul hrteristis is tken out. Tehnology n e hrnessed to help expnd geogrphil outreh, s well s overome low litery levels. Physil ess n e stepped up using two-pronged strtegy, in view of limited finnil infrstruture nd penetrtion vi existing genies with higher penetrtion suh s the Pkistn Post Offie, s well s vi new tehnology solutions suh s rnhless nking nd moile nking. Simplified finnil proesses nd proedures, lient segmenttion, nd produt diversifition n help lower osts nd mnge risks etter. New pprohes suitle for smller enterprises, suh s Pge ix

10 nk downsling, re workle tools to hieve sustinle smll nd medium enterprise (SME) lending produts. Further integrtion of finnil servies for the underserved mirofinne, remittnes, smll enterprise finne would strengthen finnil provider sustinility, vi improved ompetition, effiieny, nd exposure of finnil institutions to mrket disipline. A mjor ostle tht hs rendered produt- nd lient- expnsion diffiult nd unneessrily risky hs een the limited informtion on mrkets, segments, nd instruments. This report nd the ttendnt pioneering ntionwide ess to finne survey, prt of onsiderle donor-supported Stte Bnk of Pkistn (SBP) finnil inlusion progrm, is ment to fill in the gp. Despite reforms ess to finnil produts remins limited. The Stte Bnk of Pkistn (SBP) hs emrked on n ggressive pth of expnding finnil mrket overge vi enling, if strit, regultion, yet outreh hs lgged ehind the ountry s growth nd development needs. Reforms in the pst dede hve resulted in strong nks with stedy performne. As result, Pkistn s finnil system hs grown signifintly in the lst few yers, nd ess nd penetrtion of finnil produts hs een expnding, though gin, from very smll se: Over hlf of the popultion sves, ut only 8 perent entrust their money to forml finnil institutions. One-third of the popultion orrows, ut only 3 perent use forml finnil institutions to do so. Mirofinne hs grown t 40 perent per yer sine yet mirofinne ess extends to only 1.7 million out of n dult popultion of out 80 million. Interntionl remittnes hve grown t 29 perent sine yet only 2.3 perent of Pkistnis send or reeive remittnes, while hlf of remittnes, inluding domesti flows, re trnsmitted informlly. Agriulturl disursement grew y 44 perent in yet rurl redit demnd remins unmet the finnil system rehes only 15 perent of frmers. Life insurne is the most used insurne produt, nd demnd is high for rop insurne produts - yet only 1.9 perent re insured. 9.0% 8.0% 7.0% 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% Forml svers 8.2% Forml orrowers 2.5% 2.3% 2.2% 1.9% 1.0% 1.0% 0.2% Remittne users Mirofinne users Insurne users Islmi finne users Housing finne users SMEs using nks Finnil ess is low mong the poorer, women, smll nd miroenterprises nd in rurl res, though mrket studies suggest they re vile ustomers. Forml mrkets ould lern from nd ooperte with informl rrngements, to inrese outreh. In testimony to the ommeril viility of these lient segments, informl finne seems to e ple of profitly serving suh groups t resonly low premium to forml servies. Most forml finnil produts remin high-end, limited to urn rih eduted mles employed in the forml setor. The forml setor ould lern lot from nd prtner with informl providers their servies re pereived s eing more geogrphilly essile, less omplex, with fewer requirements, nd esier to understnd. For one, forml finne ould differentite their produts more, ttuning them to the speifi needs of vrious popultion segments, suh s women. Requirements ould e less onerous using tehnology nd flexiility - strit doumentry, gurntor, nd ollterl requirements for eoming ustomer, nd inome nd informtion/wreness onstrints hve ostruted more ggressive growth in forml finnil ess. Disprte s forml ess might e mong rurl nd urn res, men nd women, inome levels, s well s long edution nd employment setors, interest in finnil servies is virtully identil. The hllenge is to trnslte finnil interest into forml ess nd usge. Poverty nd lk of informtion on finnil servies predite lk of interest the pereption of irrelevne of the finnil system to the everydy lives of onsiderle shre of the Pkistni popultion. Low geogrphi outreh, omplex proedures, nd produts poorly suited to lient needs strengthen the pereption of forml setor irrelevne for the vst mjority of the popultion. An nlysis of pereptions of finnil servies links the populrity of informl finne to its minimum ess requirements in diret ontrst Pge x

11 to forml finne doumenttion, reditworthiness requirements, nd ssoited fees tht overurden popultion groups suh s women, low-inome or rurl popultions. Mirofinne Mirofinne in Pkistn represents low 0.2 perent of totl finnil ssets, though forml mrkets growth is seond fstest in South Asi (fter Afghnistn). The forml mirofinne setor rehes less thn 2 perent of the poor, s opposed to over qurter in Bngldesh, Indi, nd Sri Lnk. The informl setor n e ompetitive, nd hs good lessons to offer to its forml ounterprt. There is still onsiderle room for growth of mirofinne in Pkistn the estimted potentil mrket size is in the rnge of million tive orrowers, nd some estimtes ple the numer s high s 35 million. Women re poorly explored lientele with tremendous potentil. While mirofinne poliy nd servies hve foused on redit, there is onsiderle potentil for other produts, espeilly svings. A key hllenge to mirofinne institutions (MFIs) in Pkistn is rising onsiderle funding to grow, ttining sustinility, nd etter integrting with finnil mrkets. MFIs rely onsiderly on nonommeril funding ommeril liilities to gross lending portfolio re rely 21 perent. Profitility nd performne in the mirofinne setor is low ut improving. In 2000, mirofinne ws elevted to ore spet of the government s poverty redution progrm. In spite of SBP enourgement, ommeril nks hve shown little ppetite to servie mirofinne lients. The SBP strtegy of offering nk liense to stronger MFIs hs proven more suessful, though the hoped-for deposit moiliztion hs not mterilized with the pe expeted, nd the outreh of mirofinne nks (MFBs) remins limited. MFBs ount for 31 perent of the mirofinne lending portfolio, nd 85 perent of its growth. The Pkistn mirofinne mrket hs muh potentil for rpid outreh expnsion, nd fes onsiderle unstisfied demnd, espeilly for svings produts. Moile tehnology n help expnd ess onsiderly, espeilly in the informl setor. The finnil setor hs not yet tken up SBP enourgement to tht effet, nd will unlikely hnge ourse given the reent finnil risis fllout. Yet, it is importnt to persevere in this gend, whih diretly links into poverty redution. Promising strtegies inlude finnil wreness mpigns, strengthening of MFI viility nd ommeril sustinility, inlusion of women nd lient segmenttion, nd development of svings produts. Smller size of produts, nd ulk servie might etter ttrt lower-inome groups. The inresing use of tehnology will mke this pproh vile usiness proposition for nks s well s ffordle for lients. Two pprohes hve een used interntionlly to ddress high trnstion osts due to low popultion density, smll verge lons, nd low household svings the Grmeen nd Br low-teh, low-ost, high-volume models of mirofinne, nd the high-teh, low-ost, high-volume pprohes developed in Keny or the Philippines. With lose to 90 perent overge nd 59 perent reh (nd no gender divide), moile nking holds muh promise to inrese ess. A potentilly mjor plyer in ess to finne for the underserved is the Pkistn Post Offie, with its over 13,000 offies, nd urrent efforts to upgrde tehnology. Under government poliy enourgement, some MFBs hve experimented with linking up with Pkistn Post in id to expnd outreh (suh s in Brzil nd Chin). Smll Enterprise Finne Smll nd miro enterprises hve seen worsening of ess to finne, while medium-size enterprises hve seen improvements. SMEs re the growth engines of the eonomy due to their ility to rete jos, foster entrepreneurship, nd provide depth to the industril se of the eonomy. Yet SMEs get disproportiontely smll shre of redit reltive to their eonomi importne. There re 3.2 million SMEs in Pkistn, whih onstitute over 90 perent of ll privte enterprises in the industril setor, employ nerly 78 perent of the nongriulture lor fore, nd ontriute over 30 perent to gross domesti produt (GDP). Yet SME lending ounts for 16 perent of totl lending volume nd only 4 perent of totl ustomers. Similrly, 3.6 perent of firms use lons for investment (s ompred with 12.7 perent for South Asi), nd only 13.9 perent use them for expenses (34.5 perent in South Asi). Pge xi

12 Smll nd miro firms finne internlly 90 perent of working pitl nd 81 perent of new investment. Studies estimte smll enterprise redit demnd gp of Rs 277 illion (ompred with totl urrent SME redit t Rs 400 illion). Enterprises do not seem to e exluded from finnil mrkets due to poor performne. Insted, n inomplete legl nd regultory frmework nd non-sme-friendly produts nd proedures hmper inresed SME lending. Indiret osts legl fees, ollterl registrtion, nd doumenttion mke nk lending expensive for SMEs. A typil smll usiness lon requires up to 27 steps for the nk nd nine meeting with lients. An enling role hs een plyed y the expnsion of the Credit Investment Bureu s (CIB s) sope in 2006, the SME Poliy 2007, whih emphsized SME ess to finne, nd ove ll the new SBP Prudentil Regultions for SMEs. However, nks ontinue to find it diffiult to serve SMEs profitly for severl resons. First, the legl frmework (nmely, the seured trnstions regime nd, to lesser extent, the redit informtion infrstruture) limits the pool of potentil pplints. Further regultory hllenges inlude slow ourt, the stlled SME At of 2006, nd prolemti tx system. Seond, nk produts re not tilored to SMEs, resemling insted orporte lending prties. Finlly, nks do not hve orgniztionl strutures nd monitoring tools onduive to hieving high effiieny. SME demnd-side ftors further onstrin the mrket, inluding limited SME ounting, udgeting, nd plnning pity nd poor entrepreneuril skills. Continued promotion of n enling environment for SME lending nd lrge-sle downsling effort involving oth the puli nd privte setors n forge rpid growth in SME lending. Inresing ess to finne for SMEs lso requires reting seured trnstions lw tht llows ll SMEs to use movle ollterl, ttrting n institutionl investor with trk reord in SME lending, nd ssisting other nks to go down mrket. While progress hs lso een sustntil on redit ureus, more ould e done to filitte the retion of redit histories y SMEs. Remittnes Remittnes to Pkistn re estimted t round $16 illion nd growing fst, ut forml flows do not reh the poor, women, nd rurl res, where servie is mostly informl. Interntionl flows (through oth forml nd informl hnnels) totl $9 illion, with domesti flows t pproximtely $6.95 illion. These remittne flows n ply vlule role in providing foreign exhnge, ut more importntly lso offer signifint potentil to support inomes of poor nd vulnerle groups. In Pkistn, however, forml remittnes hve not een mjor prt of inome for poorer households, minly euse of limited ess. Trnsmission networks work well in urn res, though outreh to rurl nd remote lotions re diffiult, nd servies re not suffiiently ustomized to lient needs (suh s women who might need doorstep delivery, nd the poor who rrely hve the requisite doumenttion nd ounts). Pkistn Post hs lrge rurl network nd is the most ommon hnnel for domesti remittnes, ut servies remin reltively ineffiient. Home onsumption onstitutes the lrgest use of remittnes. SBP hs tken vrious mesures tht hve signifintly inresed remittnes through forml hnnels, though lrge shre of domesti remittnes remins informlly trnsferred. SBP hs een enourging the privte setor towrd providing moile nking solutions. The moile overge (t out 90 perent of the popultion of Pkistn), nd the suess of moile money trnsfer solutions in other ountries, suggests tht moile phone nking offers signifint potentil to sle up ess to finnil servies in Pkistn. To stimulte outreh to remote lotions, SBP hs een enourging the privte setor to provide moile nking solutions. Other solutions from interntionl experiene inlude Indin innovtions suh s the nongovernmentl orgniztion (NGO) Adhikr, whih developed n effiient domesti ustomized trnsfer servie, nd ICICI Bnk, whih extended its outreh to remote villge enters vi omputer kiosks. New prtnerships mong remittne mrket plyers nd other finnil entities oth within Pkistn nd rod hold muh promise. Further dvnes in formlizing money trnsfer flows will ring new lientele nd motivtion for effiieny gins nd ustomiztion of servies to lient needs. Pge xii

13 Reommendtions Summry The Wy Ahed: Poliy Options A mjor drive to enhne finnil inlusion would involve joint effort of SBP, the ntionl government, privte setor, the ommunity, nd donor efforts. The est formul for rpid sling-up of ess is to rely on tehnology, litery gins, finnil re-engineering of proesses nd produts, nd n enling legl nd institutionl frmework.* The Role of the Privte Setor Aess to ll Finnil Servies Diversifying the produt rnge nd segmenting lients to inrese outreh, simplify proedures, lower osts, nd mnge risks, to etter ter to lient needs. Speifi suggestions inlude: Use of lterntive forms of ollterl, suh s soil ollterl, ompulsory svings, personl gurntees, rops or mhinery purhsed, household ssets; Use of trditionl sving rrngements nd Rotting Svings nd Credit Assoitions; Smller size of produts, nd ulk servie, to etter ttrt lower-inome groups; Lower lon size nd deposit size to etter mth women s needs, given their lower inomes; Frequent repyments so tht instllments re smller nd orrespond to women s inome yles; Litery should not e requirement to ess finnil servie; Innovtive wys to reh ustomers suh s deentrlized opertions, use of moile units, operting units loted ner women lients, trnstions t lients doorsteps, use of femle stff. Fous on the promising mrket nihe of finnil ess for heds of households, espeilly of interest in rurl res). Rehing out to the femle lient: Women s ilities to etter mnge det, their stronger svings ptterns nd lient loylty present n untpped profitle lient se for the finnil nd mirofinne setors in Pkistn. Understnding women s needs more preisely, nd refleting those in the finnil produts nd the provider s poliies nd proedures, would ensure n inrese in women s ess to finne in spite of ulturl norms, gender segregtion, low litery, nd inomes. Glol experiene suggests offering women redit tht is untied to speifi use, insted llowing the orrower to suggest the tivity. Lerning from the informl finnil setor s minimum requirements, flexiility, nd ost/time effiieny: Developing linkges with the informl setor would help formlize the setor, inresing the forml setor s lient se. Rehing out to the rurl lient y leverging tehnology: Tehnology n lower osts, enlrge geogrphil outreh, inrese produt qulity, help enhne redit The Role of the Puli Setor Creting wreness of the enefits of ess to finnil servies: Further gins in finnil litery re ritil, though more ritil is inresing wreness of finnil servies to promote trust in the setor, s well s informtion out servies nd produts ville. A ntionl wreness mpign n support finnil inlusion, espeilly for women, s well s enourge people to open nk nd svings ounts. Awreness-retion nd trust-uilding ould e forged through soil moiliztion nd mss medi. Strengthening institutions: Growth in ess to finne will e elerted y n integrted finnil system nd strong regultory frmework. Stronger institutions (inluding Seurities nd Exhnge Commission of Pkistn nd CIB) re mjor prt of rpid inrese in finnil inlusion. Upgrding the existing redit ureu mnged y SBP nd onsolidting the hievements in inresing its overge to the whole finne servie setor inluding NGO MFIs n ple mny more potentil orrowers within reh of some ess to finne. SBP ould lso filitte the retion of redit history for SMEs y mndting the redit ureu to ollet informtion from utility nd teleom ompnies. To filitte SME lending monitoring, SBP should mend SME portfolio reporting requirements to inlude volumes / numer of lons in four su-rkets (Rs <2M, Rs 2-6M, Rs 6-25M nd Rs 25-75M). A more effiient Pkistn Post is must (following the suessful exmples of Brzil nd Chin), to pitlize on its lrge network nd outreh in rurl res. Government should explore wys to improve remittne nd other servies nd speed up the utomtion of postl rnhes. Given the signifint developmentl potentil of Pkistn Post to enhne finnil ess, it needs moderniztion in Pge xiii

14 informtion, nd provide innovtive pplitions for servie delivery. Despite some regultory nd operting hllenges, tehnology solutions hve enled poor people to ess finnil servies, for exmple: Innovtive nk devies n enhne outreh/orrespondent nking (suh s Indi s ICICI Bnk omputer kiosks in remote villge enters). Brnhless nking vi moile phones nd other devies hs the potentil to derese operting osts y s muh s 12 perent nd n help shift some of the finnil flows from informl to forml hnnels, in prtiulr if omined with other orrespondent nking hnnels. Chllenges inlude regultory, seurity, nd supervision diffiulties, nd limits on the rnge of servies (s in the Philippines G-Csh nd Keny M-Pes models). The Grmeen nd Br low-teh, low-ost, high-volume models present suessful experiene for mirofinne. Prtnerships mong finnil institutions (ommeril nks nd MFBs, linkges with Pkistn Post nd NGOs) n unlesh remittnes nd other finnil servies for the rurl poor. A useful exmple is tht of the Indin NGO Adhikr developed n effiient domesti ustomized trnsfer servie. Bsi nking hs hd some suess (Indi, Mexio, South Afri), in prtiulr when nks hve voluntrily offered ommeril si nking. Home-grown solutions inlude United Bnk Limited s servies for ill pyments, money trnsfers; Bnk Al-Flh s moile nking with Wrid Teleom; nd the remittne servie of Etislt s (UAE)-Smrt (Philippines). Aess to Mirofinne Improving MFI sustinility nd ility to muster ommeril funding/svings deposits, nd their further integrtion into the finnil system: MFIs need to improve effiieny, risk mngement, nd profitility to inrese reline on ommeril funding. opertions nd regultions. Following most suess ses in Est Asi, s well s mny other ontinents, PPSB should e pled under the supervision of SBP. Creting n enling environment for expnding ess to the underserved: Regultions should keep up with the needs of the setor nd tehnologil developments, to enle expnsion. Simultneously, n enling environment should go hnd-in-hnd with refully hosen government presene. Indisriminte susidies, espeilly foused on interest-rtes, n e detrimentl to the expnsion of the setor, s they not only distort pries ut rowd out effiient institutions nd produts. The government should resist populist pereptions tht low interest rte funding n serve developmentl purpose. Even more detrimentl re stte-owned institutions reted to promote finnil ess. Evidene in the se of Pkistn tht suh institutions (for exmple, SME Bnk) tully improve ess is wek; rther, these efforts wste vlule puli resoures tht ould more usefully e deployed elsewhere, nd eliminte the level plying field for mrket prtiipnts. Enourging positive puli pereptions on essiility nd sfety would help. A further strtegy is refousing on mirosvings nd deposits olletion, given the lrge untpped demnd for suh produts nd is supported y interntionl experiene: svings methods tht hve worked for mirofinne inlude doorstep olletion shemes nd periodi ontriution or ommitment progrms. Aess to SME Lending Refousing on mirosvings: Interntionl experiene points to regultory methods of promoting svings, suh s mthing shemes nd tx dvntged shemes. Creting omplete nd well-funtioning seured trnstions regime: Seurity interests over movle ssets should e esy nd llowed on most ssets nd y every entity (oth physil nd juridil persons). Priority rnkings should lso e lerly defined mong those who might hve lims on property offered s ollterl. The new seured trnstions regime should lso inlude ple (suh s Pge xiv

15 Crrying out thorough nk downsling progrm nd modernizing SME nking. Key fetures of the downsling progrms tht hve worked inlude: Long-term tehnil ssistne to implement the neessry sustntil hnges; Creful seletion of nk dvisors nd ontent of the tehnil ssistne; Prtiiption of mix of ommitted nks to rete ompetition; Adequte performne greements for prtiipting nks. Aess to Remittnes Reduing informlity. Informlity will derese upon the introdution of effiient, low-ost, esy-ess remittne servies without prohiitive identifition requirements. Inresing nk ounts n lso help inrese remittnes through forml hnnels. New tehnologies n redue osts nd mke it possile to servie res where trditionl nk rnh models re not vile. registry) for mking priority interests pulily known, nd enforement of seurity interests for ll ssets should e fst nd low-ost. Promoting inititives proving demonstrtion effet of nk downsling. Introduing n institutionl investor with trk reord in SME lending, idelly vi selling the SME nk or giving ontrolling rights on its ord to n institutionl investor. Supporting long-term tehnil ssistne progrms for seleted nks ( good exmple is the Chin SME lending progrm). Formlizing informl remittnes, prtiulrly domesti: Puli poliies imed t supporting tehnil nd finnil litery, omined with edution on the enefits of forml systems would lso help. Supporting remittne servies of Pkistni nks rod, to oost interntionl remittnes nd forge llines with interntionl nks. One strtegy would e to set the reimursement rte through prtner nks higher thn remittnes through Pkistni nk networks. Pkistni missions rod ould disseminte informtion to immigrnts on lowest ost nd est soures of money trnsfer (s is done y the Mexin mission to the United Sttes). Expnding Pkistni nk presene nd remittne servie provision rod, prtiulrly in the mjor remittne soure ountries. These inlude prtnerships, or innovtive wys to reh ustomers (for exmple, Hi Bnk Limited s model for ultivting lients nd fostering regulr remittne hits in the Middle Est). Pperwork should e kept to minimum, reognizing low litery levels, nd tehnology interfes should not e exessively omplex. Doorstep delivery models would help remittne reipients, who re generlly women or elder groups. Promoting the struturing of interntionl flows into investments: Interntionl good prties inlude (1) pkging remittnes with pyment servies (suh s Bnsefi in Mexio); nd (2) orgnizing nd trgeting Dispor networks rther thn tul remittnes flows. At the mro level, Dispor onds issues (for exmple, in Isrel, Indi, Ghn) or seuritiztion of future remittnes flows (in Brzil, Slvdor) hve een used, lthough seuritiztion is ostly. At the miro level, governments hve filitted trgeted Dispor funds. The Role of Puli-Privte Prtnerships A onrete wy to mke progress in expnding ess to the underserved is to form puli-privte working groups on mirofinne, smll enterprise finne, nd remittnes, to strt dilogue to tkle key hllenges in the setor. Priority themes should inlude legl nd regultory issues, mrket trnspreny, ompetition nd ost, nd reserh nd dt issues. Puli-privte disussions on rnhless nking ould fous on rriers to industry response to reent SBP regultory inentives, s well s review the Pyment nd Eletroni Fund Trnsfer At, dt privy or seurity regultion tht ould filitte e- or m-pyments, nd the rnhless nking regultions providing for nk-nonnk prtnerships nd use of gents in money trnsfer servies (uilding on the U.K. Puli-Privte Working Group on Remittnes nd the Remittnes Tsk Fore, nd the World Bnk remittne inititive). Pge xv

16 1. Overview, Finnil Mrket Struture, Regultions, nd Poliies 1.1. Expnding Aess to Finne, Links to Growth, nd Poverty Redution Finnil ess is now widely knowledged s pth to meningful inlusion nd redution in poverty. Inresed ess to finnil servies hs signifint impt on poverty (Clessens nd Tzioumis, 2006) (Box 1.1). Aess to finne in Pkistn hs n importnt potentil for signifint improvements. Credit to the privte setor mounts to 29 perent of gross domesti produt (GDP), s individuls nd SMEs prefer to rely on retined ernings to finne their working pitl, investment, housing finning, nd other finnil needs. Of the totl popultion, 14 perent hve ess to forml finne, nd out 40 perent hve no finnil ess to forml or informl finnil systems ltogether. In omprison, 32 perent of the popultion in Bngldesh hs ess to the forml finnil system, s do 48 perent in Indi nd 59 perent in Sri Lnk (World Bnk, 2008). Poliy efforts to inrese ess to finne in Pkistn hve tken time to er fruit, ut ess now is expnding quikly in ertin finnil setors, leit from very low se. To understnd the diffiulties in poliy mesures to single-hndedly rise finnil ess, one hs to reognize the importne of soioeonomi onditions, si edution, nd finnil litery on ess. Lk of informtion on ville finnil servies, omined with high levels of poverty, low litery rtes, nd gender is, result in low levels of finnil inlusion. Aess to svings ount, to redit, to insurne, to miro- nd smll nd medium enterprise (SME) finne, nd to remittnes redues risk nd vulnerility y llowing households to etter mnge flututions in inome, nd it enles the poor to invest in miroenterprises or in essentil servies like helth nd edution. Lk of ess to the forml finnil setor perpetutes the poverty trp. Lk of well-funtioning finnil mrkets hs disproportiontely dverse onsequenes for the poor, who hve redit requirements ut few ssets tht n serve s ollterl. They re thus shut out of forml finne mrkets. Poorer households depend minly on expensive informl or noninstitutionl soures. BOX 1.1: FINANCE AND LINK TO GROWTH The importne of rod finnil servies outreh n e justified in severl wys. The first rgument uilds on the theoretil nd empiril finne nd growth literture, s surveyed y Levine (2005) nd the importne of well-developed finnil system for eonomi development nd poverty llevition (Bek, Demirgu-Kunt, nd Levine 2004 nd Honohn 2004). Finnil mrket imperfetions suh s informtionl symmetries, trnstion osts nd ontrt enforement osts re prtiulrly inding on poor or smll entrepreneurs who lk ollterl, redit histories, nd onnetions. Without rod ess, suh redit onstrints mke it diffiult for poor households or smll entrepreneurs to finne high return investment projets, reduing the effiieny of resoure llotion nd hving dverse implitions for growth nd poverty llevition (Glor nd Zeir 1993). Seond, one of the hnnels through whih finnil development fosters eonomi growth is the entry of new firms (Klpper, Leven, nd Rjn, 2004) nd the Shumpeterin proess of retive destrution. This implies tht tlented newomers hve ess to the neessry finnil servies, inluding externl finne. Aess to finne for lrge prts of the popultion is thus seen s importnt to expnd opportunities nd ssure thriving privte setor with effiient distriution of resoures (Rjn nd Zingles 2003). Soure: Bek, Demirgu-Kunt, nd Peri (2005) This report exmines oth the forml (regulted) finnil setor, nd the informl setor (moneylenders, redit from supplier of goods, hwl/hundi money trnsfers, svings nd lending vi ommittees nd friends/fmily). Informl finne is disussed in setions 2.4 (demnd), nd 3.7 (supply). In the Pkistni ontext, the finl finnil setor inludes the nking setor regulted y the Stte Bnk of Pkistn (SBP), the nonnking finnil institutions regulted y the Seurities nd Exhnge Commission of Pkistn, Pkistn Post Offie, Diretorte of Ntionl Svings, mirofinne institutions (MFIs), remittnes through liensed exhnge ompnies, nd the insurne setor. The report suggests strtegies for stronger ross-pollintion etween the two setors, formlizing or dopting prties of the informl setor, nd putting the setor on more sustinle footing; it lso points out lessons out Pge 1

17 lient knowledge nd orienttion, nd (sometimes) effiieny, in the informl setor, from whih the forml setor ould enefit or emulte. Reent dvnes in ess to finne tehnology mke it possile to reh roder groups of people t lower ost nd risk. An estimted 1 illion people re urrently onneted to pyment systems in developing ountries. Yet finnil institutions in mny ountries hve een relutnt to expnd into tht mrket, inluding in Pkistn, due to poor informtion, low puli wreness, inpproprite tehnology, nd unsuitle finnil proesses nd produts. Eqully widely knowledged re the dt gps tht persist when ttempting to understnd rriers to ess. Greter vilility of dt in Pkistn n provide preise mesures on finnil ess to the poor, to rurl res, s well s informtion on the hrteristis of the finnilly underserved nd informl finne Ojetives of the Report The primry purpose of this study is to mesure nd desrie the stte of finnil servie provision to underserved segments of the mrket in Pkistn, prtiulrly those with low inomes nd smll enterprises, nd to identify wys improve investment nd rete inlusive mrkets tht meet the needs of underserved people nd enterprises. The new dt nd study re of vlue to ommeril providers who re le to design produt strtegies round the segmenttion nd trends highlighted y the dt, to poliymkers who my e onsidering new legisltion imed t improving the funtioning of finnil mrkets, nd to donors who my e mking investment nd funding deisions to inrese ess to ertin regions or popultion groups. Further, the dt n e used to ondut ross-ountry omprisons nd onstrut seline for future referene tht will help guide poliy development on finnil inlusion. The survey hs lredy een onduted in severl ountries, inluding Brzil, Indi, Mexio, nd South Afri. An in-depth piture of finnil ess ws otined y ntionl Aess to Finne (A2F) household survey fousing on the demnd side of finnil mrkets. A speifi effort ws mde to pture the elements of ess to finne tht lie outside of the forml finnil systems suh s informl finning from shops, fmily-sed usiness run out of the house, soilly sed rotting svings/lending shemes, nd others tht might not even e reognized s finne y survey respondents nd therefore require speifi tehniques to eliit the dt nd fts. The ntionl household survey lso provides insights into regionl issues, inluding ess to finne onerns speifi to urn nd rurl res. The survey dt on household ess to finne is first of its kind nd is not urrently ville for the ountry. Comining demnd nd supply spets, the findings n e of onsiderle enefit to those urrently underserved y the finnil setor. The report fouses on the underserved popultion groups nd enterprises, those with informl ess to finne or with no suh ess ltogether. It does not over lrge orporte lending, onsumer lending, nd more omplex finnil instruments suh s seurities, money mrket instruments, nd re-insurne. SBP hs identified severl priority res for finnil development, with hevy emphsis on the underserved, inluding SMEs nd mirofinne. The report piks up on some of those issues, s well s on the mrket for remittnes, the rurl nd urn poor, gender nd inome dimensions of ess, nd informl finne. The dt gthered do not fous on speifi set of institutions or finnil instruments; insted, the pproh is very muh demnd driven: vi household surveys, the dt permits us to eliit the ext set of institutions nd instruments in demnd nd those in use, y different portions of the popultion (y inome, region, rurl/urn res, gender, ge, nd so forth). In order to provide omplete piture, the report lso reviews the supply dt, to id poliy formultion nd ssist with the understnding of the Pkistn-speifi finnil system fetures. 1 1 A word of ution: while the household survey ollets representtive smple, inluding upper-inome households who use more sophistited onsumer finne instruments, the report does not expliitly fous on the Pge 2

18 BOX 1.2: HOUSEHOLD A2F SURVEY FOR PAKISTAN (DEMAND-SIDE DATA) The A2F household survey is omprehensive ntionl household survey of ll the min finnil servies (trnstion nking, svings, redit nd insurne), needs, nd usge mong onsumers, in oth the forml nd informl setors. The survey design is sed on joint methodology developed y FinMrk Trust (South Afri) nd the World Bnk, nd rooted in eonomi fundmentls. The design hs enefited from ontinuous improvements s the survey hs een used in severl lrge ountries round the world, s well s from reful ustomiztion to Pkistn onditions vi fous-group disussions nd piloting. The im of this demnd-side study is to estlish redile enhmrks nd highlight opportunities for innovtion in produt nd delivery, s well s suggest promising venues for deepening nd rodening ess to finne. The dt olleted permit us to: Mesure nd trk the lndspe of ess to finnil servies ross ll the min produt tegories trnstion nking, svings, redit, nd insurne, in oth the forml nd informl setors nd institutionl tegories ommeril nks, other regulted institutions, semiforml nonregulted institutions, inluding memership-sed ones, nd informl or villge-sed institutions. Understnd hrteristis of those who re finnilly exluded; Segment the mrket; Identify opportunities for expnsion of finnil servies to the un- nd underserved segments of the mrket; Understnd the sope of the popultion of vulnerle poor whose needed finnil trnstions re too smll for ny finnil institution to provide profitly. The dt lso permit us to trk the diverse ptterns of ess to finnil servies ross hrteristis suh s ge, gender, ethni group, nd re of residene. Further, the dt extend over spetrum of res of finnil usge nd interest, from exmining qulity of life nd poverty, to ttitudes to nd the use of tehnology, s well s levels of finnil litery. The survey overs 10,305 households in ll regions of Pkistn exluding the tril res. At n initil stge, the stndrdized questionnire ws ustomized to Pkistni onditions, to ensure high-qulity dt. Fous interviews were held in urn nd rurl res of Sindh, Punj, North West Frontier Provine (NWFP), Blohistn, nd Azd Jmmu nd Kshmir (AJK), for purposes of lirtion nd gthering of supplementry qulittive informtion. At seond stge, households were surveyed on the following topis in detil: - Bsi household demogrphis; finnil litery; - Soioeonomi hrteristis; psyhogrphis/ttitudinl; - Household inome; - Aess to finnil servies; nks; provider differentitions; - Svings; - Lons/redit; - Insurne; - Money trnsfer/remittnes; - Pyment nd reeipts. Detiled desriptions of the survey design, methodology, smple, nd dt spets, s well s the questionnire itself, re presented in Appendix A. The A2F household survey overge is ntionlly representtive nd reflets oth rurl nd urn ess to finne issues (Box 1.2). In the se of rurl res, the report does not get into the tehnilities of speilized griulturl finne instruments nd their supply side, s this is full topi in its own right. Frmer finning nd insurne issues re signifintly different from finning issues for other mrket prtiipnts, nd household survey would not e well suited to ring out the relevnt prolems nd mrkets nd ess issues for suh instruments, suh s mortgge finne, utomoile finne, onsumer durles lending, redit rds, nd sophistited heking nd sving ounts. On the other hnd, it does ttempt to onvey est-prtie nd poliy options s onerns simple instruments relevnt for the underserved, suh s deit rds, nd very si (limited) ounts suh s smrt rds nd simplified trnstion ounts. Pge 3

19 issues (onsider, for exmple, the issue of wether-sed insurne). The dt do over rurl issues nd differenes, nd the report eliits issues of geogrphi nd demogrphi ess, s well s types of instruments used in rurl forml nd informl finning. Finlly, the A2F survey dt were olleted etween Otoer 2007 nd Mrh 2008, nd the nlysis in this report refers to this time frme. In prtiulr, the nlysis ignores susequent events ssoited with the glol finnil risis nd its fll-out within Pkistn Pkistn A Brief Mrket Overview Pkistn s finnil system grew signifintly in the pst few yers. At present, totl finnil ssets hve rehed $215 illion (150 perent of GDP). While this rtio ompres fvorly with regionl enhmrks (Tle 1.1), it is modest ompred with ountries like Chin (543 perent), Indi (298 perent of GDP), Mlysi (384 perent), nd Thilnd (211 perent). Nevertheless, Pkistn s nking setor hs gined dynmism, profitility, nd strength, with deposit se rehing $62 illion nd gross dvnes $47 illion ntionlly in August Supported y growing finnil intermedition proess, nks ggregte profitility rose to $1.8 illion in The nking system onstituted 44.5 perent of the totl ssets of finnil institutions on Mrh 31, Tle 1.1: Bsi Finnil Inditors, Interntionl Comprison Stoks Credit to Privte Setor (% of GDP) Equity Mrket Cpitliztion (% of GDP) Vlue Trded (% GDP) Privte, Bonds (% GDP) Puli Bonds (% GDP) Totl Finnil Assets/GDP GNI Per Cpit, ($) Bngldesh 37.7% 10.0% 7.1% % 470 Brzil 49.8% 104.3% 44.5% 2.9% 3.6% 205.1% 5910 Chin 114.5% 189.8% 237.5% 0.2% 0.4% 542.5% 2360 Indi 47.4% 155.4% 94.6% 0.7% 0.2% 298.3% 950 Indonesi 25.4% 48.9% 26.1% 2.1% 1.1% 103.6% 1650 Mlysi 108.8% 180.2% 83.0% 4.4% 7.1% 383.5% 6540 Pkistn 29.4% 48.9% 70.0% 0.6% 1.3% 150.2% 870 Philippines 23.8% 71.6% 20.3% 1.7% 11.3% 128.7% 1620 Sri Lnk 34.0% 23.3% 2.9% 0.3% 0.2% 60.8% 1540 Thilnd 84.2% 79.8% 44.0% 1.4% 1.2% 210.6% 3400 Soure: World Development Inditors 2008, dt for Note: GNI=gross ntionl inome. Totl finnil ssets re defined in simplified mnner, sed on ggregte of ssets desried in this tle. Pkistn hs suessfully implemented signifint finnil setor reforms over the pst out 15 yers, strting with grnt of lienses to numer of new privte nks in the erly 1990s, moderniztion of the governne nd regultory frmework of the nking setor in the lte 1990s, nd the privtiztion of mjor puli setor nks sine the erly to mid-2000s. The uthorities hve tken steps to phse out or reorgnize most of the government-owned development finne institutions, hve put in ple severl inititives to promote the growth of the mirofinne setor, nd hve llowed more freedom to insurne ompnies. The tive pe of reform reted resonly well-developed, diverse, nd sophistited finnil mrket, given Pkistn s inome level. In line with these reforms, the privte setor redit touhed the figure of Rs 2,523 illion in My 2008, s ompred with Rs illion yer erlier. SME redit hs inresed from Rs 18 illion in fisl 2000 to Rs 403 illion on Mrh 31, 2008, though the inrese is entirely ounted for y medium, not smll, enterprises (SBP 2008d). Consumer redit ounted for 14 perent of totl outstnding dvnes t the end of Mrh Agriulture redit rose from less thn Rs 40 illion in fisl 2000 to Rs 200 illion in Pge 4

20 fisl The ggregte numer of orrowers hs risen, from 2.7 million in 2003 to out 5.5 million y Deemer House uilding lons stood t Rs illion in My 2008 wheres the totl housing finne mrket of Pkistn stood t Rs 126 illion on Deemer 31, 2007 (SBP 2008e), douling its size from Mirofinne lons (miroredit, mirosvings, nd miroinsurne) worth Rs 22.6 illion were disursed in 2007 through extension of 1.8 million mirolons. Presently, the tive lients of miroredit re round 1.7 million. In spite of reent hievements, ess to finnil servies remins quite limited in Pkistn. The predominnt shre of the finnil system, the nking setor, is mostly foused on lrge enterprise lending, with n inresing interest in onsumer finning (though still on very smll sle), to the reltive neglet of SMEs, rurl res, mirofinne, nd the poor. There is little understnding of the min rriers to wider provision of finnil servies, or the opportunities tht exist for finnil ompnies in underserved mrket segments. One of the resons for the lk of improvements in ess provision is the limited vilility of dt on ptterns of ess to nd usge of finnil servies mong different popultion groups Bnking Setor Totl deposits (exluding internk) stood t Rs 3,202 illion t the end of 2006 nd Rs 3,779 illion in August The puli setor nks ounted for 21 perent of deposits, domesti privte nks for 74 perent, nd foreign nks for 5 perent. The lon portfolio of the nking system is lso growing nd rehed Rs 2,336 illion in It stood t Rs 2,853 illion in August The mjor hunk of growth in lons portfolio ould e ttriuted to lol ommeril privte nks. Figure 1.1: Deposits, Lons, nd Assets of the Bnking System Soure: SBP (2008f) Note: 2008 dt re for Mrh 31, Deposits Advnes Assets Tle 1.2: Brekdown of Lons (Domesti Opertions) y Setor Figure 1.2: Credit y Setor Lons Outstnding Numer of Borrowers Amount Shre (%) Numer Shre (%) (Rs Bn) (Rs 000) Corporte setor % % SMEs % % Agriulture prodution % % Consumer finne % % Commodity opertions % % Stff lons % % Other % % Totl % % Soure: SBP figures for Mrh 2008 Note: SBP regultions define SME lons s lons to enterprises or sole proprietors up to Rs 75 million. The nking system ontinues to grow with its lne sheet size pssing the level of Rs 5.2 trillion in Mrh The shre of lol ommeril privte nks in totl ssets inresed to 74.8 perent. In 2 Provisionl figure (SBP 2008g: Tle 5.3) Pge 5

21 ontrst, the shres of puli setor ommeril nks, speilized nks, nd foreign nks stood t 19.3 perent, 2.4 perent, nd 3.6 perent, respetively. Tle 1.2 shows the limited ess to nk finne of enterprises nd individuls. Fousing only on the onsumer side, ording to ltest (unpulished) SBP figures, there re only 16 million personl (nonusiness) nk ounts nd 5.5 million personl lons. Moreover, only 25 perent of the totl nk deposits nd 17 perent of the totl orrowers re from rurl res. In vlue terms their shres re even smller, 10 perent nd 7 perent of the totl vlue of deposits nd dvnes, respetively (Akhtr 2008). The Pkistn Post Offie mnges 3.6 million svings ounts, through 12,343 rnhes, with 70 perent of suh ounts holding svings elow Rs 10,000 ($128). Most of the nk lending is onentrted in few lrge mnufturing ompnies (Figure 1.2). Aggregte dt for ll redit y orrower size shows skewed distriution: 0.4 perent of nk orrowers ount for 65 perent of ll nk redit nd more thn 5 million orrowers ount for the remining one-third of lons. At the very top there is even more onentrtion; the lrgest 50 orrowers ount for 37 perent of ll redit outstnding. Limited ess to servies is ptured y the low level of rnh penetrtion, espeilly in rurl res, whih hs held k the growth of svings nd impted redit distriution systems (Figure 1.3). Figure 1.3: Trends in Sving Rtes Consumer lending hs een the fstest-growing segment in the pst severl yers, y wide mrgin, t 320 perent in terms of numer of orrowers nd slightly less thn 500 perent in lon mount, with 2.9 million lients urrently etween 2003 nd Mortgge lending grew from Rs 4 illion in 2003 to Rs 67.4 illion in In tndem with growth, the setor hs improved its effiieny drmtilly. Applition fees for strt-up personl lons in Pkistn re found to e round 7 perent, one of the lowest in rod spetrum of ountries (World Bnk 2008). Similrly, Pkistn rnks well mong the ountries with very low omplexity of personl lon pplitions, nd rnks high mong the ountries with mny free usge fetures for nk ounts. Soure: SBP Suh positive effets of ompetition nd effiieny improvements hve not ppered yet in the SME lending setor, though the setor experiened douling of SME redit in Interntionlly, Pkistn ompres unfvorly on usiness lon proessing times nd ollterl to lon rtio, ut does ompete suessfully on diret proessing ost (Annex Tle A1.1). Similrly, 3.6 perent of firms use lons for investment (s ompred with 12.7 perent for South Asi), nd only 13.9 perent use them for expenses (versus 34.5 perent in South Asi). Overll, 37.6 perent of firms onsider ess to finne onstrint, s ompred with 27.8 perent in South Asi (Annex Tle A1.1). Agriulturl disursement hs inresed sustntilly (y 44 perent etween 2003 nd 2007, to 1.35 million orrowers) euse of greter ess, nd holds onsiderle promise for medium-term development ommeril nks hve ggressively lent, to the result of meeting SBP s trget of Rs 200 illion set for fisl Outstnding griulture dvnes ount for 6 perent of totl dvnes nd servie 2 million lients. However, urrent redit meets only 45 perent of the griulture requirements. Pge 6

22 Figure 1.4: Bnking Setor Penetrtion DEMOGRAPHIC BRANCH PENETRATION (NUMBER OF BRANCHES PER 100,000 PEOPLE) GEOGRAPHIC BRANCH PENETRATION (NUMBER OF BRANCHES (ATMS) PER 1,000 SQUARE KM) Repuli of Kore Singpore Philippines Sri Lnk Thilnd Indi Pkistn Bngldesh Chin Singpore Repuli of Kore Bngldesh Indi Philippines Sri Lnk Thilnd Pkistn Chin DEMOGRAPHIC ATM PENETRATION (NUMBER OF ATMS PER 100,000 PEOPLE) GEOGRAPHIC ATM PENETRATION (NUMBER OF ATMS PER 1,000 SQKM) Repuli of Kore Singpore Thilnd Philippines Sri Lnk Chin Pkistn Bngldesh Singpore Repuli of Kore Bngldesh Indi Philippines Sri Lnk Thilnd Pkistn Chin DEPOSIT ACCOUNTS PER CAPITA (NUMBER OF DEPOSITS PER 1,000 PEOPLE) LOAN ACCOUNTS PER CAPITA (NUMBER OF LOANS PER 1,000 PEOPLE) Singpore 1671 Singpore 513 Thilnd 1423 Philippines 302 Thilnd 248 Bngldesh 229 Bngldesh 55 Pkistn 192 Keny 70 Pkistn PERCENT OF ADULTS WITH ACCESS TO AN ACCOUNT WITH A FINANCIAL INTERMEDIARY Singpore Repuli of Kore Sri Lnk Thilnd Indi Chin Bgldesh Vietnm Philippines Pkistn Keny Soure: WB Development Inditors ATM=Automted Teller Mhine. Pge 7

23 Bnking Setor Penetrtion Inditors of physil ess to finne in Pkistn suggest need to further improve the nking infrstruture (Figure 1.4). Speifilly, the numer of utomted teller mhines (ATMs) is t less thn 0.53 ATMs per 100,000 people. Compre this with the demogrphi ATM penetrtion figure for Sri Lnk, whih is four ATMs per 100,000 people, nd to the world medin figure of 10. In terms of demogrphi penetrtion of nk rnhes, Pkistn is lose to the world medin levels t 5 rnhes per 100,000 people. This is still onsiderly lower thn the figures of 6 for Indi nd 7 for Sri Lnk. The ountry reords out 22 lon ounts per 1,000 people, ompred with Bngldesh s 55 nd the world medin of 81. The use of deposit servies is eqully low the ountry hs 192 deposit ounts per 1,000 people, s ompred with 229 in Bngldesh nd 529 for the world medin. A new fous is needed on strtegy to remove rriers to finnil servies ess nd tively enourge wider penetrtion y nks nd other ommeril providers. For this, etter informtion is required on urrent levels of ess to finnil servies s well s on impediments to wider ess. Doing Business 2008 inditors show tht only 46 perent of dults in Pkistn re overed y the puli registry (ompred with 86 perent in ountries of the Orgniztion for Eonomi Co-opertion nd Development [OECD] nd 49 perent in Chin), while the privte ureus over less thn 1.4 perent (ompred with 59 perent in OECD ountries) Bnking Setor Performne Pkistn s nking setor hs gined dynmism, profitility, nd strength with deposit se rehing $62 illion nd dvnes of $47 illion ntionlly. Supported y growing finnil intermedition proess, nks ggregte profitility rose to $1.8 illion in Minimum pitl requirements for nks were rised to Rs 1.5 illion y Deemer 2004 nd Rs 2 illion y Deemer SBP hs lso required nks to inrese their pid-up pitl y Rs 1 illion yer, to reh Rs 6 illion y the end of Deemer As result, onsolidtion of the nking mrket hs tken ple, nd stronger nks hve emerged. Bnks pitliztion nd qulity of ssets hve helped rise the risk weighted pitl dequy rtio to 13.2 perent. All nks mintin pitl dequy rtios well ove the minimum interntionl stndrd of 8 perent, exept for stte nks. Figure 1.5: Bnking Setor Performne Gross NPLs to Gross Advnes ROA fter tx Intermedition ost Interest rte spred Soure: Pkistn Finnil Setor Assessment 2005, SBP (2005) with SBP dt updte s of Mrh 31, Note: Interest rte spred defined s return on lons nd dvnes minus ost of deposits. Pge 8

24 Growth of nk ssets hs piked up onsiderly sine the lte 1990s (Figure 1.5), up to 18.8 perent nd perent most reently (Deemer 2007 nd Mrh 2008, respetively). Nonperforming lons of the nking system hve een delining overll, though they re still high for the speilized nks (Figure 1.5). The pitl dequy rtios of the nking system show n inresing trend. They delined for ll types of nks (exept for speilized ones) over the period of ut hve shown n inresing trend sine then. For puli setor ommeril nks, the pitl dequy rtio rehed 15.2 in 2006 from -1.3 in Speilized nks, however, hve een reporting negtive pitl dequy rtios. In ddition, the strengthening profitility of the nking setor n e ssessed through inresing returns on ssets (ROAs). As group, ll of the ommeril nking groups exept foreign nks hve witnessed inrese in ROA. Intermedition osts hve styed stle, while interest rte spreds hve een on the inrese. Although the ROAs hve een positive nd growing for most of the period under onsidertion, return on equity of the nking system experiened slight deline over due to the inrese in pitl to meet minimum pitl requirement nd the higher pproprition of profits y the nks. Stte-owned nks, however, hve performed very poorly in terms of the profitility Overview of Mrket Plyers in the Bnking Setor The numer of sheduled nks in Pkistn ws 40 t end Deemer 2007 (Tle 1.3). The numer of ommeril nks hs delined from 41 to 26 through mergers, quisitions nd losures. In the nking setor, speilized government-owned nks hve mde only limited ontriution in their nihe mrkets. As of Deemer 2007, there were 18 nks involved in Islmi nking with network of 288 rnhes in the ountry. Of these, six re full-fledged Islmi nks with 185 rnhes nd 12 of the existing ommeril nks hve 103 rnhes tht re offering Islmi nking servies (SBP 2008i:2). Tle 1.3: Comprtive Position of Numer of Brnhes in Pkistn (June 2008) Rurl Urn Five ig nks 2,247 3,315 Privte nks 143 1,830 Speilized nks Islmi nks Foreign nks - 51 Totl 2,580 5,763 Soure: SBP, Septemer Pkistn Post Offie Another potentilly mjor plyer in ess to finne for the underserved is the Pkistn Post Offie. It hs network of 13,419 rnhes throughout the ountry nd is signifint provider of finnil servies, inluding svings, insurne, nd remittnes, through 7,276 nk rnhes. The Pkistn Post Svings Bnk (PPSB) serves s n gent of the Ministry of Finne for rnge of finnil servies inluding svings moiliztion, life insurne, postl giro ounts, nd money trnsfers. It is present throughout the ountry nd is the only nking servie ville in some remote res. Pkistn Post hs een le to upgrde the tehnology for its opertions with help from the Islmi Development Bnk. Internet servies were lunhed in the lte 1990s nd re widely ville t post offies, ttrting the interest of smll entrepreneurs nd providing limited finnil servies. The PPSB offers severl svings shemes, nd in 2006 it hd 3.6 million svings ount holders. In ddition, Pkistn Post ts s n gent to sell government-ked svings instruments. The Postl Life Insurne offers 10 insurne options nd hd 252,810 tive poliies in Pkistn Post provides severl options for ntionl nd Pge 9

25 interntionl remittnes in ddition to hosting Western Union in some of its rnhes, nd ting s n gent to Western Union for interntionl remittnes. Government poliy hs een to enourge finnil institutions to use the extensive network of post offie rnhes to extend outreh to more people. In reent yers two MFBs, Khushhli Bnk nd First MiroFinneBnk, hve linked up with the Pkistn Post to offer servies t their rnh offies. To dte, though, the volume of this usiness hs remined smll. Given the signifint developmentl potentil of Pkistn Post to enhne finnil ess, espeilly geogrphil penetrtion, it is importnt to py speifi ttention to pity uilding nd strengthening of its performne nd moderniztion, s well s stndrds. Sine Pkistn Post dtes k to the yer 1873 when the Government Svings Bnk At, 1873 ws promulgted, its finnil servies re not under SBP regultion, nd thus hve limited potentil for tehnil supervision nd performne improvement. A notle lesson from the Est Asin Experiene is the onsiderle dvnement of postl finnil opertions one they move under entrl nk supervision nd regultion. Box 3.2 presents suessful experienes from Brzil nd Chin nd lessons lerned from interntionl se studies on postl finnil systems Unmet Rurl Credit Demnd Only 25 perent of the totl nk deposits nd 17 perent of the totl orrowers re from rurl res. In vlue terms, their shres re even smller, 10 perent nd 7 perent of the totl vlue of deposits nd dvnes, respetively (Akhtr 2008). Only 15 perent of frmers re rehed y the finnil system s whole (inluding the ommeril nks, griulturl nks, nd other finnil institutions). Totl redit to rurl Tle 1.4: Rurl Credit Mrket (Rs) Totl rurl popultion of Pkistn 97.0 million Totl rurl households 15.0 million Frm 9.2 million Nonfrm 5.8 million Estimted redit demnd illion Totl griulturl redit supplied (Rs) illion Frm redit 51.5 illion Nonfrm redit 58.2 illion Totl miroredit 28.4 illion Totl miroredit (rurl) 22.0 illion Totl miroredit (urn) 6.4 illion res is Rs illion, while rurl Soure: redit demnd is estimted s twie this volume (Rs 250 illion). The ulk of rurl finne volumes re in griulturl finne (Rs illion), with lned shre of frm nd nonfrm redit. While miroredit volumes re skewed towrd rurl res, rurl miroredit urrently represents merely 17 perent of totl rurl redit. The ess to finne gp is even strker in the se of frmers. The few upper-inome lrge frmers enjoy lmost 10 times higher ess to forml finne (38.7 perent) thn the frmers in the poorest quintile (6.5 perent). Lrge frmers get even more informl redit thn their less-fortunte ounterprts 82.0 perent vs perent. Suh skewed distriution of frm redit negtively ffets the poverty level of griulturl households, nd prevents fster development in rurl res. Tle 1.5: Frmer Aess to Finne Finnilly Inluded Formlly Inluded Informlly Inluded Finnilly Exluded Poorest 70.7% 6.5% 70.2% 29.3% 2 nd quintile 81.8% 11.5% 80.6% 18.2% 3 rd quintile 83.5% 13.2% 82.9% 16.5% 4 th quintile 82.0% 24.9% 81.7% 10.0% Rihest 89.6% 38.7% 82.0% 10.4% Soure: Pkistn Aess to Finne Dtse, Pge 10

26 1.5. Mirofinne (Chpter 3) Outside of the nking system, little finning is ville for the underserved (Tle 1.6), though this is promising re for long-term growth. Though ess to finne hs stedily improved over the yers, it remins too low to mke rel differene in terms of the poverty impt of growth. Vrious estimtes of potentil mrket demnd for mirofinne ple potentil lient figures in the tens of millions, s ompred with tul lient figures of 1.7 million urrently (out of totl popultion of over 160 million). Mirofinne penetrtion in the region is higher, t 35 perent in Bngldesh, 25 perent in Indi, nd 29 perent in Sri Lnk. Afghnistn hs reltively new mirofinne industry with out 437,000 tive lients s of Mrh Tle 1.6: Mirofinne Outreh in Pkistn 2007 No. of rnhes 1,343 Totl no. of orrowers 1,471,295 Averge lon size (Rs) 13,520 Averge lon lne (Rs) 10,286 Gross lon portfolio (Rs in million) 15,134 Currently, the network of mirofinne providers is 1,343 rnhes with out Rs 15 illion portfolio. Among mirofinne providers, Khushli Bnk lone provides overge in 86 distrits. hene ounting for 47 perent of this network. The three mirofinne entities, nmely, Ntionl Rurl Support Progrm (NRSP), Khushli Bnk, nd Kshf Foundtion ounted for pproximtely 70 Soure: MiroWth. Issue 6, Annul (2007). perent of the setor s tive lients. In the middle of 2008, totl outreh mesured y tive orrowers ws out 1.7 million in mrket where it is estimted tht there re 7 million poor households. In ft, sed on experiene in more mture mirofinne mrkets like Bngldesh, the size of the potentil mrket in Pkistn ould e s high s million sine the vulnerle poor, those living ove the poverty line ut without ess to most forml finnil servies, lso use mirofinne servies. Some estimtes of potentil demnd y women ple the potentilly vile svings lients t 35 million, nd vile orrowers t 22 million. Muh remins to e done to pture potentil mrket of suh proportions, inluding in terms of litery improvement, wreness-rising, nd ost redution of ess vi tehnology improvements, s well s reline on tried nd tested low-teh solutions suh s the Grmeen model. A seond hllenge for mirofinne is for servie providers to eome profitle so tht servie provision to poor people n grow on sustinle sis. A study of South Asin MFIs done in 2005 showed tht only 42 perent of miroredit orrowers in Pkistn reeived servies from profitle MFIs, the lowest perentge in South Asi (Mirofinne Informtion Exhnge, Consulttive Group to Assist the Poor, nd the World Bnk (2006). Most mirofinne providers rely on omintion of donor/nonommeril funding, nd nnot lim ommeril viility, with the exeption of the reently formed Mirofinne Bnks (MFBs). Retil providers suh s the six reently liensed MFBs operte on sustinle footing nd re expeted to solidify their position further y inresing reline on lient deposits. The remining mirofinne providers re lrgely not expeted to eome sustinle s lss. Those inlude (1) the ntionl nd provinil rurl support progrms nd hritle or speilized nongovernmentl orgniztions (NGOs); (2) the wholesle pex institution, the prinipl eing the Pkistn Poverty Allevition Fund (PPAF), whih relies on donor nd government finning support to refinne ommunity orgniztion whose numer hs rehed 60,000; nd (3) the NGOs nd the oopertives, whih re now mostly losed exept for one prinipl opertor, the Punj Coopertive Bnk, whih hs mnged to e protive ut relies on SBP redit lines (quite n unsustinle pproh) nd provinil government support nd hs to e eventully restrutured to e more finnilly vile. A third hllenge is tht mirofinne is lmost exlusively foused on lons, while other finnil servies, svings, trnsfers/remittnes, nd insurne re often more in demnd y poorer households. Indeed, this report finds tht demnd for svings, not redit, predomintes, nd is lrgely untpped. Unlike most other ountries, the mirofinne setor hs historilly foused more on men thn women, s produt delivery methods hve een poorly suited to women nd estlished ulturl mores. Pge 11

27 1.6. Islmi Finne In 2002, the first Islmi nking liense ws issued to Meezn Bnk. Currently there re six liensed, full-fledged Islmi Bnks nd 12 onventionl nks with stnd-lone Islmi nking rnhes (Tle 1.7). As of Deemer 31, 2007, the numer of Islmi nking rnhes stood t 288 in 47 ities in ll four provines of the ountry. These lso inlude interntionl nks suh s ABN Amro nd Stndrd Chrtered. Applitions for few more plyers re under onsidertion. The most populr Islmi produts urrently inlude murh, ijrh, nd diminishing mushrk finning. 3 The totl ssets portfolio in the Islmi Bnking setor stnds t Rs 200 illion s of Mrh 31, 2008; deposits were t Rs 142 illion nd finning nd investment t Rs 114 illion. Totl numer of orrowers/finnes is out 25,000. Tle 1.7: Islmi Bnking Plyers De-04 De-05 De-06 De-07 No. of Islmi nks No. of rnhes N.A. 185 No. of onventionl nks offering Islmi nking No. of onventionl nks rnhes offering Islmi nking N.A. 103 Totl numer of rnhes offering Islmi nking Soure: SBP Bnking System Review for the yer ending De 2005 Despite their rpid growth over the pst few yers, the Islmi nks re fing numer of prolems, inluding the lk of liquidity mngement instruments, 4 speifi Islmi-nk prudentil regultions, depositor risk for Islmi depositors, nd ertin lk of lrity with more omplex finnil instruments, suh s foreign urreny ounts, letters of redit, nd ssets to e generted in the future SME Finne (Chpter 4) There re 3.2 million SMEs in Pkistn. SMEs ount for over 90 perent of ll privte enterprises, ontriute to 30 perent of GDP, nd ount for 25 perent of exports of mnuftured goods. In spite of their importne in the eonomy, lending to SME still ounts for only 16 perent of totl lending in volumes nd only 4 perent of totl ustomers. While interntionl omprisons of SME lending re diffiult euse of inonsistent SME definitions, redit to Pkistn smll enterprises hs muh potentil to expnd, reltive to the interntionl experiene. Lending to SMEs (s defined y the Prudentil Regultions of SBP 5 ), despite the importne of SMEs in the Pkistni eonomy, hs shown slightly downwrd trend over the period Although the development of the SME setor rests ruilly on the vilility nd ess to finne t n ffordle 3 Murh finning is n investment prtnership, s of 2007 t out 40 perent of the totl finning y the Islmi nking institutions. Ijrh finning is n Islmi lese greement, ounting for out 30 perent of the totl finning. Diminishing mushrk finning is profit nd loss shring, wherey the lient purhses finnier s shre grdully, until ll of finnier s shre is purhsed y the lient, mking him sole owner, t out 17 perent of the totl finning y the Islmi nking institutions. 4 The issune of two sukuks (sset-ked Islmi onds), oth y the Wter nd Power Development Authority (WAPDA), hve somewht esed the sitution for some of the Islmi nks, ut there re no proper institutionl rrngements ville. 5 Prudentil Regultions for Smll nd Medium Enterprises Finning: Bnking Poliy nd Regultions Deprtment, SBP ( The SBP dt on SMEs provided in this setion re sed on this definition. Pge 12

28 ost, espeilly through the forml institutionl setup, privte setor redit to SMEs ounted only for 16.2 perent of totl redit to the privte setor in Deemer 2007 (Tle 1.8). SME lending grew muh more slowly thn other lending mrket segments, suh s onsumer lons, espeilly redit to smll, s opposed to medium, enterprises. One reson for suh disppointing expnsion of redit to SMEs is tht the risk-ssessment proess for SMEs is more omplex, espeilly in dynmi usiness environment with little relile informtion. Another reson is tht nks onsider SMEs s prt of their usiness (orporte) finne tivities, wheres SMEs should e prt of their retil usiness, euse SME lending will only work if less ttention is pid to risk-ssessment proesses nd more to effiieny nd speed of delivery to produe sle of opertions. In ft, mny nks view SMEs s mini-orportions nd use more or less the sme tools to ssess risks. This proves to e too expensive s well s ineffetive. Tle 1.8: Credit to Privte Setor, nd Growth in Lending Credit to Privte Setor (De 2007, Rs, illion) Growth in % during De CY03- De CY07 Numer of Numer of Setors of the Eonomy Amount Borrowers Borrowers Lons Corporte setor 1, n.. n.. SMEs Agriulture Consumer finne Totl 2, Soure: SBP The relutne of nks to pereive SMEs s profitle usiness venture is ompounded y the ft tht mny potentil SME lients do not hve immovle ssets, whih re the most-used form of ollterl. Weknesses in the existing ollterl regime, espeilly for movle property, omplite mtters further. Moreover, s the mjority of the SMEs re sole proprietors, nks nnot seure their lons with movle ssets either. In ft, on the sis of the existing lw, only limited liility ompnies n register lien over movle sset in the ompny registry. Other resons for the relutne of nks to lend to SMEs re the generl lk of relile forml redit informtion, given tht redit ureus re still in n erly stge of development, nd the ft tht most SMEs do not mintin proper ooks of ounts, mking it diffiult for trditionl nkers to ssess the viility of going onern Remittnes (Chpter 5) Interntionl remittnes hve grown t impressive rtes in Pkistn, in prt euse of rod formliztion effort y SBP, whih hs rought in over 70 perent of interntionl flows into the forml net. Not so the domesti remittne mrket, whih hs not enjoyed suh poliy ttention, nd on whih offiil dt re limited. Rough unoffiil evidene exists tht domesti flows re of onsiderle mgnitude, nd, s with interntionl flows, most of the trnsfer hppens in the informl setor. Figures from the SBP indite tht interntionl forml remittnes into Pkistn for were $6.45 illion, up from $2 illion in Reent studies suggest tht, in most ses, the mroeonomi enefits of remittnes in developing ountry re likely to e positive, nd there re lso onsiderle household-level enefits tht hve eonomi growth nd poverty redution effets. However, in Pkistn, remittnes do not pper to mke signifint portion of household inome for the poorest quintile. Interestingly, domesti remittnes onstitute lrger shre of inome thn interntionl flows for poorer households, while the trend reverses mong riher households. In prt, this is euse most migrnt workers do not pertin to the lowest inome groups. However, muh more importntly, limited ess to Pge 13

29 remittnes y the poorest popultion is mostly to lme for the wekened poverty effet of remittnes, in spite of the setor s spetulr reent growth nd performne. Remittnes to Pkistn hve een growing t n impressive rte in reent yers (29 perent etween fisl 2001 nd fisl 2008). The lrgest mount remitted is from the United Sttes (27.3 perent) followed y Sudi Ari (19.4 perent). Inresed remittnes hve supplemented foreign exhnge ernings nd hve helped mintin the rupee exhnge rte (though higher exhnge rtes lso tend to suppress export-led growth). The steps tken y SBP to formlize the remittne mrket through formtion of exhnge ompnies (ECs), nd redution in the ur nd internk exhnge rte, hve helped inrese forml remittne flows. Over 80 perent of forml remittnes ome through the forml nks, Hi Bnk Limited nd United Bnk Limited eing the two min plyers, while 17 perent ome through ECs nd out 2.5 perent through the Pkistn Post Offie. Western Union is very tive in the mrket nd hs set up prtnerships with vrious nks, ECs, nd Pkistn Post. Fees for sending remittnes from rod vry from free trnsfer to over $50, while the time tken for the trnsfer is from under 10 minutes to more thn two months. The iggest hllenge is to inrese rurl outreh of remittnes. Interntionl suess stories in this respet inlude moile phone nking strtegies sed on the G-Csh model of the Philippines or the M- Pes model of Keny. Another strtegy is to prtner more losely with plyers who hve etter rurl networks. While the urrent regultory set-up does not llow teleom opertors to provide finnil servies, lessons n e lerned from these teleom-led models nd pplied through prtnership rrngements etween nks nd nonnks. Other steps tht n e tken inlude forming prtnerships with nks rod, s well s expnding the outreh of Pkistni nks rod, informing, inresing wreness, nd filitting remittnes y Pkistni workers ord Finnil Inlusion nd the New Finnil Regultory nd Tehnologil Infrstruture Ntionl Poliies to Enhne Finnil Inlusion SBP hs dopted oordinting role with the Government of Pkistn (GOP) nd privte setor to implement finnil inlusion strtegy, fousing on mirofinne, SMEs, remittnes, Islmi nking, griulture finne, nd insurne, mong other res. Mny new SBP regultions hve een issued to enourge finnil development nd inlusion, for exmple eletroni nking, rnhless nking, the SME poliy frmework, nd others 6. The newly reted Development Finne Group with SBP in Septemer 2006 steers nd helps poliy implementtion, in the res of the legl nd regultory frmework for SMEs, griulture nd rurl redit, housing nd infrstruture finne, nd Islmi finne. The Finnil Inlusion Progrm (FIP) 7, developed with onsiderle donor king (from the U.K. Deprtment for Interntionl Development, or DFID) nd dopted in 2008, will onstitute n importnt nd promising inititive in this regrd, over the long term. 6 BPRD Cirulr No. 02 of 2008, Brnhless Bnking Regultions for Finnil Institutions desirous to undertke rnhless nking ( Prudentil Regultions for Smll nd Medium Enterprises Finning: Bnking Poliy nd Regultions Deprtment, SBP ( 7 Pge 14

30 SBP hs further introdued the Annul Brnh Liensing Poliy 8 with rod mp for 2008 nd eyond. This poliy requires ommeril nks with 100 rnhes or more to open t lest 20 perent of their rnhes outside ig ities nd set up rnhes in tehsil hedqurters, where no rnh of ny nk exists. The five ig nks hve, s result, more thn 40 perent of their rnhes in rurl res. However, due to this poliy, smller nks re keeping the numer of rnhes t less thn 100 only fter they mke onsious deision to go into rurl res do they exeed 100 rnhes. In ddition, SBP hs llowed nks to estlish su-rnhes, ooth nd servie enters of ommeril nks in inner regions where it is ostly to mintin full-fledged rnh. These su-rnhes n e mnged y skeleton stff nd n t s n extended rm of the ner-y rnh for performing limited nking funtions, suh s deposits, withdrwls, issune of demnd drfts, nd telegrphi trnsfers, nd lso filitte pyment of home remittnes to their enefiiries. Other utionry lessons should e tken into ount while pplying this poliy. The reent World Bnk study on Bnking the Poor suggests tht mndtory regionl requirements for new rnhes my redue the totl numer of rnhes (inrementl osts might trnslte into inrese in the ost of providing the dditionl servies, thus mking ess prohiitively expensive even when the servies re ville), nd while enourging su-rnhes or ooths, my fll short of enourging geny reltionships or orrespondents, where present mndtes pper unler. Sine Novemer 2005, ll ommeril nks operting in Pkistn re required to offer Bsi Bnking Aounts 9 to improve finning ess for the low-inome Pkistnis. Suh n ount n e opened, for exmple, with mximum deposit of Rs 1,000 rrying no fee, no minimum lne, nd full ATM ess. As of Mrh 2007, out 120,000 si ounts hd een opened. Those re still ssoited with high trnstion osts for nks, nd low demnd from ount holders who tend to sty with existing instruments in spite of low interest rtes. Competition is expeted to weken these disdvntges in the longer term. Suh regultory mesures should e treted with ution, however. The Bnking the Poor study suggests tht regultion in this re on its own is not enough nd indeed there is limited ssoition with etter ess. However, in mny ples where nks hve voluntrily offered ommeril produts of si nking hrter, there is positive ssoition with ess. The entrl nk with the support of its two prinipl trining institutions, Ntionl Institute of Bnking nd Finne (NIBAF) nd the Institute of Bnkers Pkistn (IBP), is implementing n inititive for finnil litery nd pity uilding for finnil institutions nd the puli, to leverge ville finnil servies while uilding momentum for the ommunity to ppreite the ville lending opportunities. This gend is importnt, nd this report lso finds low ess to finne mong individuls with lower levels of litery. Cuslity hs een deted y prtitioners nd poliymkers, however. While finnil litery rguments re quite ompelling, empiril evidene from Indonesi (Cole, Zi nd Smpson 2008) suggests tht finnil inentives (whih remove or redue ffordility rriers of opening nk ount), not finnil litery, hve n impt on the use nd uptke of finnil servies mong households. Nonetheless, this report ontends tht wreness of finnil options nd servie providers, s well s development of trust in suh providers, rther thn litery per se, hold the promise of inresing finnil ess ntionwide. The evidene on SMEs lso suggests tht there my e n importnt role for usiness trining results of usiness development ssistne hve een very promising, per interntionl experiene. In 2001, the Mirofinne Institutions Ordinne nd Supportive Prudentil Regultions ws dopted, llowing for the retion of ntionl nd regionl MFBs. SBP pulished new guidelines in 2006 for ommeril nk provision of mirofinne servies 10. The mirofinne ordinne ws mended in Annul Brnh Liensing Poliy, Bnking Poliy nd Regultions Deprtment, SBP ( 9 BPD Cirulr numer 30 of 2005 ( 10 SMED Cirulr numer 11 of 2006 ( Pge 15

31 (Finne Bill ) to ensure tht minimum lon size is rised to Rs 150,000 to llow providers to tilor their servies to mirousiness needs. Further, mirofinne providers re now llowed to engge in the remittne setor. With fous on expnding mirofinne to 3 million orrowers y 2010, strtegy for Expnding Mirofinne Outreh hs een developed y the SBP nd ws pproved y the GOP in Ferury The strtegy stresses ommeriliztion of the setor s key to finnil nd soil sustinility, nd enefits from extensive support y donors (DFID), vi the overll FIP inititive, in terms of redit s well s pity-uilding nd tehnology inititives support. While the long-term impt of the progrm, dopted in 2008, remins to e ssessed, it will undoutedly enounter short-term diffiulties ssoited with the glol finnil turmoil nd its fll-out in Pkistn. SBP introdued Prudentil Regultions for Agriulturl Finning nd Guidelines for Livestok nd Fisheries Finning, s well s simplified nd stndrdized lon douments 11. It hs further foused on enourging revolving redit sheme for three yers, under whih frmers n orrow for one yer nd ontinue to orrow without providing doumenttion eh yer; guidelines hve een issued on lending for livestok, fisheries, nd hortiulture susetors, nd progrms re eing lunhed for the diry setor. In ddition, smll frmer finning sheme hs een promoted sed on group lending nd rop lon insurne sheme hs een strutured tht is now eing offered y some insurne ompnies. SBP hs further llowed nks to finne ginst mrket/relizle/fored sle vlue of the griulture lnd, urn properties, gold/silver ornments, nd so forth, nd lso ginst two personl sureties (up to Rs 500,000, or $8,333), in ddition to the pssook to the lnd, nd so forth. These inititives hve ontriuted signifintly to enhning forml institutionl griulturl redit disursement for smll orrowers. SBP hs lso lid out n elorte prudentil regultory nd supervisory frmework tht onforms to the Interntionl Islmi Finnil Servie Bord s (IFSB) frmework, y wy of enling Islmi finne, whih SBP expets will onstitute lmost 12 perent of the Pkistn finnil system in the next five yers. In 2004, SBP issued detiled riteri for (1) setting up of sheduled Islmi ommeril nks sed on the priniples of Shrih (Islmi lw) in the privte setor, nd (2) for setting up of Islmi nking susidiries y existing ommeril nks. In the sme yer, SBP lso issued guidelines for opening of stnd-lone Islmi nking rnhes y existing nks. Islmi finne is still in its infny worldwide nd in Pkistn. Bsi regultions re outstnding, s re estimtes of potentil demnd mong vrious popultion groups, nd models for sustinle produt delivery. The setor hs muh growth potentil. SBP hs pulished new prudentil guidelines for SME finning, wherein nks n issue unseured lons to SMEs up to Rs 3 million 12. A new SME Poliy 2007 ws lunhed in fisl 2008, ttempting to define uniformly SMEs in the mnufturing, trde, nd servies setors for ll the stkeholders. The poliy doument gives rod frmework for promotion of SMEs y improving the regultory, fisl, nd usiness environment, s well s institutionliztion of the support struture. The doument further outlines strtegy for SME-led privte setor growth, poverty redution, nd jo retion. Setor Development Compnies hve lredy een set up for vrious industries, inluding gems nd jewelry, mrle nd grnite, furniture, diry produts, nd the hunting nd sporting rms setor. These ompnies work on puli-privte prtnership sis with the privte setor hving leding role in deision mking. In prllel, there re severl inititives to extend finnil servies to smll orrowers. These 11 Prudentil Regultions for Agriulture Finning, Bnking Poliy nd Regultions Deprtment, SBP ( Guidelines for Fisheries Finning, Agriulturl Credit Deprtment, SBP ( Guidelines for Livestok Finning, Agriulturl Credit Deprtment, SBP ( 12 Prudentil Regultions for Smll nd Medium Enterprises Finning: Bnking Poliy nd Regultions Deprtment, SBP ( Pge 16

32 inlude: SBP s SME finning strtegy tht dvotes promotion of redit gurntee shemes nd venture pitl funds, estlishment of SME trining enters nd pity uilding of nks SME lon offiers, SME redit-risk soring nd ompetitiveness enhmrking, s well s triennil survey on SMEs to estimte their finne demnd 13. Work is underwy to develop n pproprite redit enhnement mehnism tht will filitte nk lending to the setor. SME redit poliies hve hd mixed suess, espeilly s onerns the smll enterprises mrket, nd more work is needed in the ollterl nd redit informtion spets of SME lending. BOX 1.3: OTHER PROGRAMS IN PAKISTAN FOR FINANCIAL INCLUSION Ntionl Rurl Support Progrm: Estlished in 1991, NRSP is the lrgest rurl support progrm in the ountry in terms of outreh, stff nd development tivities. It is not-for-profit orgniztion registered under Setion 42 of Compnies Ordinne NRSP s mndte is to llevite poverty y hrnessing people s potentil nd undertke development tivities in Pkistn. It hs presene in 46 distrits in ll the four provines inluding Azd Jmmu nd Kshmir. NRSP is urrently working with more thn hlf million poor households orgnized into network of more thn 55,366 ommunity orgniztions. With sustined inrementl growth, it is emerging s Pkistn s leding engine for poverty redution nd rurl development. NRSP mnges one of the lrgest miroredit portfolios in Pkistn, with 282,421 tive lons s of Mrh 2007, nd holds 25 perent of the mirofinne mrket. NRSP provides vrious finnil servies to the ommunity orgniztion memers in rurl res, to help them implement their miroinvestment plns, inluding miroredit, miroinsurne, nd svings produts. Rozgr Sheme: The Ntionl Bnk of Pkistn, the lrgest ntionwide nk, is offering smll lons to the unemployed nd the poor to finne purhses of uto rikshws, set up utility stores under frnhise of the Utility Corportion, set up puli ll offies, nd so forth. Interest on these lons is fixed t 12 perent: hlf of the interest hrges re to e orne y the GOP nd the remining hlf y the orrowers. The government piks up first 10 perent of ll losses nd shres in the redit risk. Externl nd internl verifition of orrowers, referenes, nd gurntees re hndled y ICIL, representtive firm of Dun nd Brdstreet. This sheme is ntiipted to deliver out Rs 105 illion ($1.34 illion in lending) over five-yer period. Sine its lunh this yer, the sheme hs delivered Rs 2 illion to out 22,136 orrowers. Pkistn Poverty Allevition Fund: The PPAF ws reted in 2008 s privte setor pex fund to provide pity uilding nd lon funds to MFIs, primrily NGOs. By the end of 2005, it ws supporting out 35 orgniztions, lmost ll NGOs tht provide mirofinne servies. Sponsored y the GOP nd funded y the World Bnk nd leding donors, the PPAF follows model of puli-privte prtnership through lending of wholesle funds to ivil soiety orgniztions ommitted to ommunity outreh progrms for enhning inome nd eonomi welfre of the disdvntged. Speifilly, the PPAF provides funding for projets tht help generte inome, nd improve the physil nd soil infrstruture nd the skills of the vulnerle. On Ferury 28, 2007, the PPAF hd resoure se of Rs 49,560.2 million ($ million). Soure: Akhtr (2007) While these inititives re step in the right diretion, SBP needs to develop strong nd effetive monitoring nd evlution system to monitor the impt of these inititives nd to ring mid-ourse orretions where needed to ensure tht the ultimte ojetives for whih the inititives hve een tken re fully hieved. Other progrms for finnil inlusion in Pkistn re listed in Box Regultory Frmework for Brnhless Bnking 14 SBP hs een working with the Consulttive Group to Assist the Poor on using informtion nd ommunitions tehnology nd new rnhless nking models to reh mssive numers of presently 13 SME Poliy 2007 ( 14 BPRD Cirulr No. 02 of 2008, Brnhless Bnking Regultions for Finnil Institutions desirous to undertke rnhless nking ( Pge 17

33 unserved poor people. 15 Brnhless nking is defined s the delivery of finnil servies outside onventionl nk rnhes using informtion nd ommunitions tehnologies nd nonnk retil gents. Industry plyers, prtiulrly moile network opertors, see the enormous potentil of rnhless nking s profitle vlue-dded servie. Brnhless nking n help ddress the low point of sle infrstruture in Pkistn, in prtiulr the very low ATM penetrtion. There re, however, onsiderle hllenges to rnhless nking posed y urrent regultion in Pkistn tht need to e ddressed, while pying due regrd to the new nd enhned risks tht rnhless nking n rry. Some potentilly vile pths inlude nk- nd telom-led models or omintion of models. Regultory hnges would e required, long the lines of some of the following: (1) the use of gents outside nk rnhes, therey inresing the numer of servie points; (2) esier ount opening (oth onsite nd remotely) while mintining dequte know-your-ustomer stndrds; nd (3) permission for rnge of plyers to provide pyment servies nd issue e-money, therey enling innovtion from multiple soures. These hnges re onsiderle. For one, the legl frmework restrits the use of gents for sh-in/shout tivities nd ount opening, mjor impediment to orrespondent nking. It is unler whether it is leglly possile for ustomers to withdrw money outside of nk rnhes, though n exeption is provided for ATMs nd withdrwls t Authorized Merhnt Estlishments t Vrious Points of Sle sujet to sh limit of Rs 10,000. However, MFBs re permitted to ppoint gents nd open ooths for olletion of sh or for mking spot pyments, yet those powers re restrited y regultion to speifi irumstnes. Further, SBP hs grnted permission for moile nking (tht is, the use of moile vns nd smll servie enters). ECs, whih re lso regulted y SBP, n mke use of gents in remote res; however, these gents urrently re permitted to provide only one servie: pyment (in Pkistni rupees only) of remittnes from foreign ountries. The urrent nti-money lundering/omting the finning of terrorism regime, whih overs nks only, presents further ostles to the potentil introdution of rnhless tivities 16. Exmples inlude using the ntionl identity rd (pproximtely 20 perent of Pkistnis do not hve one), s well s requirements for new ustomers to e introdued y existing ones, whih is diffiult in res with low overge. Other regultions on enhned ustomer due diligene use diffiulties with remote ount opening nd other non-fe-to-fe trnstions. The rules on estlishing identity re more relxed for MFBs, where in fr-flung nd remote res where people, prtiulrly women, do not hve identity rds, the MFB my extend miro-redit y estlishing identity through other pproprite mens. Aess to finne in Pkistn hs lgged ehind the ountry s growth nd development needs, in spite of tive poliy support on the prt of SBP. Prt of this disppointing outome n e explined y very high growth rtes in ess in the pst five yers, leit from very low lient se. A seond pertinent onsidertion suggests tht finnil poliies lone re not suffiient to expnd ess, nd holisti pproh linked to si poverty redution nd finnil wreness-uilding would e more suessful. In ddition to finnil ess expnsion, the tsk of strengthening finnil provider sustinility, espeilly in the mirofinne setor, is of prmount importne. Competition, effiieny improvements, nd exposure of finnil institutions to mrket disipline hold the promise of ommeril viility nd reliility for the mirofinne setor. Competition nd effiieny improvements re lso needed if SME lending is to eome profitle. New pprohes suitle for smller enterprises, suh s nk downsling, re workle tools to hieve sustinle SME lending produts. In view of limited finnil infrstruture nd penetrtion, physil ess n e stepped up using two-pronged strtegy, vi existing genies with higher penetrtion suh s Pkistn Post, s well s vi new tehnology solutions suh s rnhless nking nd moile nking. 15 This setion is losely sed on Consulttive Group to Assist the Poor (2007). 16 BPD Cirulr Numer 20 of 2004 ( Anti-money Lundering Ordinne 2007 ( Pge 18

34 2. Aess to Finne: Evidene from the Demnd Side 17 This hpter overviews the detiled findings of the ntionl household survey on individul nd household interest, knowledge of, nd usge of finnil produts, inluding svings, redit, insurne, pyments, nd remittne servies, s well s informl finnil produts. The verge Pkistni household remins outside the forml finnil system, sving t home, nd orrowing from fmily or friends in ses of dire need. While only 14 perent is formlly inluded, out 40 perent of the popultion is ompletely exluded from forml or informl finnil systems. Disprte s forml ess might e mong rurl nd urn res, men nd women, inome levels, s well s long edution nd employment lines, interest in finnil servies is virtully identil. The hllenge is to trnslte finnil interest into ess. Inome nd finnil litery pper s key rriers to finnil exlusion when ontrolling for ll other hrteristis. Other drivers of finnil inlusion re employment (forml employment determining forml inlusion), ge, edution levels, or eing the min household inome erner. Moile phone use nd ess redue the hnes of eing finnilly exluded. The nked in ddition enjoy ess to ollterl while distne is only mrginl deterrent to eing nked. An nlysis of pereptions of finnil servies links the populrity of informl finne to its minimum ess requirements in diret ontrst to forml finne doumenttion, reditworthiness requirements, nd ssoited fees tht overurden popultion groups suh s women, nd low-inome or rurl popultions. The verge Pkistni household sves t home, orrows from fmily or friends in ses of dire need, nd does not use the forml finnil system. Aout 14 perent hve ess to forml finnil servies, 18 s ompred with 32 perent in Bngldesh, 48 perent in Indi, nd 59 perent in Sri Lnk (World Bnk, 2008). Adding informl finne, the proportion of finnilly servied popultion rises lose to 60 perent (Figure 2.1). Aout 40 perent of the popultion is ompletely exluded from oth the forml nd informl finnil system. Not ll of these neessrily wnt to e served, in ft, 18.6 perent do not see the usefulness of eing finnilly served. Of the remining 22.3 perent, 6.3 perent re employed nd ould e profitly served, ut hve remined outside the finnil system. Figure 2.1: Pkistn Aess Strnd Finnilly Served Voluntrily exluded Formlly Inluded Informlly Served (orgnized setor) Inf. Served (unorg. setor) Finnilly Exluded 0% 14.3% 45.5% 50.5% 81.4% 100% Soure: Pkistn Aess to Finne (A2F) Household Survey 2008 The survey dt permits us to look t speifi ess nd ppel issues long the lines of severl si finnil produts, in oth the forml nd informl setor: nking, svings nd investment, orrowing, informl finne nd ommittees, insurne, nd pyments. The findings show low ess to forml produts overll. Islmi finne overs less thn 1 perent of individuls, mostly high-inome urn mles. There re 2.5 perent forml orrowers s per the Aess to Finne (A2F) household survey, nd the orresponding Stte Bnk of Pkistn (SBP) figure for the numer of ounts (not lients) per dult 17 Wht this hpter identifies s demnd is the potentil demnd rising from hrteristis ptured through the ntionl household survey nd nlyzed though n eonometri nlysis model. It should not e understood s effetive demnd. 18 Seventeen perent per SBP figure (Akhtr 2007). Pge 19

35 popultion is 6.3 perent (4.9 million).19 Aout 1.7 million of the dult popultion re servied y the mirofinne setor (SBP dt) this is out 1 perent of the popultion nd out 2 perent of dults. Aout 2.3 perent of people send or reeive remittnes, nd 1.9 perent re insured. Less thn 1 perent of the popultion enefited from mortgges (ut 3.5 perent used lon for those purposes). Pkistn is sh-sed eonomy even in the forml setor, inome is primrily reeived in sh. Few households re formlly served, espeilly rurl nd lower-inome groups nd women. Within the informl finne setor, 35 perent of individuls sve t home, 21 perent in livestok, nd 14 perent use ommittees20 to ple their svings. A third of individuls hold lons or redits, mostly in the informl setor (33.7 perent, whih inludes 19 perent who orrow from fmily nd friends). Overll, little over hlf of the popultion hve forml or informl svings/investments, nd 35 perent hve redit (Tle 2.1). Close to 13 perent of individuls (10.4 perent rurl nd 17.2 perent urn) hve nk ounts. Bnking with Pkistn Post, however, is lmost eqully used mong rurl nd urn households. Another notle exeption is in essing svings for investment in produtive tivities, stronger in rurl thn in urn res. Tle 2.1: Aess of to Finning, y Servie Servie % Popultion Rurl Urn Using Svings/Investment 55.4% 55.7% 54.8% Lon nd redit 35.0% 38.7% 28.6% Bnked 12.9% 10.4% 17.2% Money trnsfers 2.3% 1.9% 3.2% Insurne S 1.9% 1.7% 2.3% Soure: Pkistn A2F Household Survey 2008 Severl ptterns of finnil ess in Pkistn ome out of the ntionl household survey nd re presented methodilly in the susequent setions elow: - Rurl res re underserved nd less finnilly literte ut disply n equl or higher level of interest in finnil servies; - Mles hve etter ess to finnil servies thn femles, though deision mking is joint nd willingness to lern out finnil mtters is omprle; - There re wide provinil differenes in finnil servies, litery, nd ess; - Lower-inome groups disply signifintly lower wreness of nd ess to finnil produts, though only mildly lower interest in finnil servies. Disprte s forml ess might e mong rurl nd urn res, men nd women, lower- nd upperinome people, s well s long edution nd employment lines, interest in finnil servies is virtully identil! The hllenge is to trnslte finnil interest into ess. The nlysis goes on to suggest speifi venues for pitlizing on the onsiderle promise for finnil expnsion. Strong growth rtes of finnil ess mesures lso point to the ountry s potentil, though they lso return the nlyst to the urrent sttus-quo with very limited levels of ess. The hpter reviews (1) finnil ess ptterns y inome, provines, gender, nd other individul hrteristis; (2) finnil litery nd interest issues; (3) informl finne nd ommittees; (4) nking; (5) svings nd investment; (6) orrowing; (7) insurne; nd (8) pyment nd hek servies. 19 SBP Sttistil Bulletin June The differenes etween SBP sttistis nd the Pkistn A2F survey sttistis re disussed in detil in Appendix A (Clirtion). Some differenes rise due to lustering of finnil servies in speifi res, whih ntionwide household survey of this size (10,305 households) nnot dequtely ddress. This is the se for mirofinne lients nd ommittees (prtiulrly the usiness ommittees in zrs). Other differenes re used y multiple ounts mintined y lients nd some doule ounting of lients mong finnil institutions (this is the se for nk lending), s well s some stle or intive ounts (this is the se for Pkistn Post). Differenes n further e explined y shortomings on the prt of litery nd understnding on prt of respondents. 20 Committees, n informl mehnism for group sving nd lending, re explined in setion 2.4 elow. Pge 20

36 2.1. Vriility of finnil ess is onsiderle ross rurl/urn res, provines, inome, gender, nd other popultion hrteristis For the medin individul, 32-yer old with edution up to 4th grde of elementry shool, living in rurl res, nd employed, joining the government or eoming frmer redues his or her hnes of remining finnilly exluded y whopping 10 nd 9 perentge points, respetively, while joining the orporte setor improves those hnes y only 1 perentge point (Annex Tle A2.2). Being womn, on the other hnd, redues one s hne of eing informlly served y 3 perentge points, nd respetively inreses the hnes y s muh to e finnilly exluded, though gender hs no mrginl effet on forml inlusion of ny eonomi mgnitude ( perentge points). Every 10 yers of ge dd one perentge point to the hnes of leving the rnks of the unservied nd joining the formlly inluded, though there is no signifint effet of ge on informl inlusion. Completing primry shool improves one s hnes of joining the forml finnil mrket y 0.6 perentge points. Moving into urn res improves one's hnes to join the informlly served y 8 perentge points ut does not ffet the hnes for forml inlusion. Moile phone use nd ess redue the hnes of eing finnilly exluded y 5 nd 10 perentge points, respetively. Wht re the ritil ftors for improving ess to finne? We segment our findings y nlyzing rurl-urn, stte, gender, nd other individul differenes. Lower ess to finne in rurl res might e determined y n indiret inome nd litery effet, rther thn ftors suh s time/ost of ess, though differenes in essiility re lerly onsiderle, etween urn nd rurl res (Annex Tles A2.1, A2.6 A2.10). Interestingly, lotion in rurl res lone would not hurt or improve one s hnes for finning, nd n urn ddress would not improve them either. While forml soures re muh more uiquitous in urn res, ontrolling for the effets of inome, edution, nd employment, oth lotions re eqully well served y forml nd informl finning. This suggests tht there is nothing inherently unserviele out rurl res. In ft, informl nd some forml hnnels re providing ess there quite suessfully, nd n e esily imitted y vrious Mirofinne Institutions (MFIs) to emulte their suess in ess provision. Post offie ess (nd interestingly, insurne) is signifintly higher for the rurl popultion. Wht is perhps inherently hrder with rurl res is ess to ertin instruments suh s redit/deit rds, nd, uriously, Islmi produts nd ommittees. Figure 2.2: Shre of Bnked, y provine Soure: Pkistn A2F Household Survey 2008 Wide provinil differenes emerge s to the likelihood of eing finnilly served (formlly nd informlly), s well s nked. Overll, Bluhistn nd the North West Frontier Provine (NWFP) show lower interest, wreness, nd ess thn other regions. Punj nd Azd Jmmu nd Kshmir (AJK) led other provines (Figure 2.2). Eonometri nlysis shows tht, while provinil effets re Pge 21

37 oserved, inome is the min driver of finnil inlusion oserved in Pkistn sttes. Hlf of the popultion in NWFP nd Bluhistn re mong lowest two quintiles, while Punj hs the highest popultion (48.9 perent) mong ll provines in the highest two quintiles. However, provinil differenes persist fter ontrolling for inome. Bnk essiility in Bluhistn might e one reson for low ess, though times nd ost to getting to nk is high in AJK s well. There re signifint differenes mong provines regrding requirements prior to reeiving forml lon. While 79 perent in Bluhistn nd 21 perent in NWFP hve to provide o-gurntor, this requirement is quoted y round 50 perent in Punj, Sindh, nd AJK. Differenes in finnil nd nking ess re signifint (Annex Tle A2.3, Pnel A). Women hve drmtilly less ess to the finnil setor, with the exeption of forml orrowing. There re fewer women with ess to nking servies (5.5 perent vs perent men), money trnsfers (1.4 perent vs. 3.3 perent men) nd insurne (0.6 perent vs. 3.3 perent men). 21 The sme pttern holds for other finnil produts generlly, women Figure 2.3: Women & Mirofinne re disdvntged in ess, with the onsistent exeptions of Pkistn Post nd ommittee servies (Annex Tles A2.9 A2.10). Women remin signifintly less likely thn men to hve ess to the finnil setor overll, fter ounting for the effet of inome nd other soio-eonomi differenes (Annex Tles A2.1 nd A2.6). This overll disdvntge elies omplex piture, however. Women enjoy signifintly less ess to orrowing (24.7 perent vs perent men), exept to forml orrowing. In ontrst, women hve lmost equl ess to svings (52.6 vs perent men), though they re signifintly more likely to sve informlly (Annex Tles A2.7 nd A2.8). 22 Soure: Competitiveness Support Fund (2008) In the South Asi region, in ontrst to Pkistn, mirofinne providers hve tended to fous on femle lients (Figure 2.3). Evidene hs umulted of the link etween mirofinne for nd empowerment of women. Further, finnil institutions lso tend to find women to e etter redit risks. Segmenting femle lients y offering them tilor-mde produts is promising strtegy to th up with regionl neighors in terms of femle ess to finne (Box 2.1). Out of roughly 47 million femles over the ge of 16, it is estimted tht potentilly 35 million re vile svings lients, nd 22 million re vile orrowers.. 21 Supply-side dt show similr piture. Penetrtion of the mrket for redit to women is estimted t 4 perent. The penetrtion in svings is slightly higher t 9 perent. Response from nks suggests tht round 4 perent of the Smll nd Medium Enterprise (SME) lients nd 5 perent of onsumer redit lients re women. Approximtely 19 perent of long-term interest erning deposit ounts nd 1 perent of totl urrent ounts re owned y women. The most reent ville finnil setor outreh estimte suggests there re urrently 4 million women essing forml finnil servies (redit nd/or svings). Of these, out 3 million women re nking with the trditionl finnil institutions nd out 1 million ess servies from MFIs. 22 Speifilly note the positive nd signifint oeffiients in Tle A2.6 in the regressions of forml versus informl nd nonorrowers, nd forml versus informl orrowers only. Note lso the positive nd signifint oeffiients in Tle A2.7 in the regressions of forml versus informl nd nonsvers, nd forml versus informl svers only. Pge 22

38 Box 2.1: Fts out Segmenting Pkistni Femle Clients Women round the gloe hve een oserved to etter mnge their det responsiilities. There is no evidene tht women offer less-ttrtive usiness proposition thn men from the sme inome tegory. Women tend to e net svers, nd lthough the verge size of womn s deposit tends to e less thn tht of mn, women tend to mintin their ount with speifi nk for longer while men my shift their ptronge for mrginl gins. Thus, y ignoring women s potentil nd seprte mrket segment with unique demnd dynmis, finnil institutions my e forgoing potentil revenue nd profits. Women tend to hve preferenes tht re distint from men s. Women s ttitudes towrd svings re onsistent with their desire to lern: women re more likely to follow udget thn men. They re lso more relutnt to tke dets nd re more risk-verse thn mle respondents. In ddition, women tend to prefer onveniene ove ll. Inditive of this my e the ft tht the nks think tht the shre of women mong their totl orrowing lients inresed fter the introdution of redit rds. Pkistni women lso tend to void going to nk rnhes: thus mny women nk through their lose reltives. Given the nks riteri for lending redit only to those (men nd women) who re either slried or hve usiness inome, mny women inluding those who my hve seure soure of rent inome re preluded from the nks finning inluding redit rds. Women might lso e fing more diffiult lon requirements more women thn men (59.8 perent vs 45.3 perent, respetively) ite the o-grntor/osignor s ondition to reeiving forml redit. Being less-experiened nking lients thn men, women tend to e more nxious out nking proedures; thus simplified proedures n e n effetive selling proposition to women. Finlly, women re rrely the min inome erners (only 3.6 perent of Pkistni women re), nd re lrgely more exposed thn men to reeiving irregulr inome. Over 85 perent of women, eing housewives, only reeive poket money s inome. It is therefore unsurprising tht the poorest quintile in terms of personl (not household) inome is three-qurters omposed of women nd only qurter y men. As result, women s trnstions will in generl e smller thn men s. Women s equl ess to moile ommunitions mke moile phones promising hnnel to reh women in terms of moile nking produts ut lso wreness mpigns on the enefits of ess to finne. Bnk Al-Flh is urrently mrketing women s redit rd promoted s A Credit Crd Exlusively for Women. Though ludle effort to ttrt dormnt mrket segment, this rd does not differ from regulr redit rd offered y the nk itself or ny other nk in terms of either the fetures or the eligiility riteri. Soure: Bnk Af Flh wesite (2008). Meezn Bnk, n Islmi nk, hs introdued Ldies Bnking. dedited nking setion only for women. This new servie provides n exlusive nking re s per the Shrih (Islmi lw) priniples for women ustomers in more omfortle environment. Under this inititive, dedited ustomer servies offier will provide onsultny for full rnge of the nk s servies inluding, 9 m to 5 pm online nking, term ertifites, urrent nd sving ounts, onsumer finning, nd 24/7 nking servies. Soure: (2008). Aess inreses with ge, n employed sttus, nd edution. 23 Household heds re found to hve etter ess to finnil servies The regression nlysis on whih the disussion is sed is in Annex Tle A2.2 nd more generlly in Annex Tles 1-6, whih re lso disussed in further detil in the following setions of this hpter. The proit nd multinomil logit regression frmeworks follow the speifition of Kumr et l. (2005). Pge 23

39 The setor of employment lso ffets ess to forml nd informl finne, to the extent tht government nd orporte setor employment ugment the proility of finnil ess, nd lorer sttus diminishes it. Frmers hve worse ess to the forml setor, ut unlike lorers, hve etter hnes of essing the informl setor thn remining unserved. Government employment is negtively ssoited with Islmi produt use. Surprisingly, lorers nd frmers hve weker ess to ommittees thn other (etter-employed) groups. Belonging to the forml setor in generl improves ess to finning, espeilly forml finning. If forml finne speifilly is wht one is fter, dd government jo, or employment in the forml/orporte setor, nd higher inome. A more surprising finding is tht the positive effet of ge on finnil ess disppers in the se of Islmi produts nd ommittees. Personl nd household inome is strongly ssoited with ess to finne, though they hve roughly omprle levels of interest in finnil mtters (45.2 perent of the poorest quintile re interested vs perent of the popultion on verge). The shre of the nked popultion inreses sixfold with inome levels, from very low levels (3.8 perent) mong the lowest inome quintile to 25.5 perent for the welthiest quintile (Figure 2.4). Yet mirofinne nd informl finne re y no mens privilege of lower-inome groups. Mirofinne is used four times s muh y the rihest quintile thn the poorest quintile. The lowest inome group fes on verge four times longer dely in rehing nk thn the highest inome quintile (though medins differ less). The lowest inome quintile fes higher solute osts thn the highest (medin of Rs 36 vs. Rs 26). Figure 2.4: Produt Penetrtion The upper-inome, espeilly mle, popultion hs muh higher ess to vrious forml finnil produts inluding produts suh s si nking ounts. Less thn 1 perent of the poorest nd seond poorest hve si nking ount, s ompred with 4.3 perent mong the highest inome quintile. Credit rds, deit rds, or Islmi finne re reserved for the riher, s lmost no one in the two poorest quintiles hd ess to these produts. Soure: Pkistn A2F Household Survey 2008 The top three inome quintiles hve signifintly higher ess to finne thn the lower quintiles. Interestingly, though the top three quintiles do not differ signifintly mong themselves in the se of ess to forml nd informl finne, they fre quite distintly in terms of nking ess, whih eh higher quintile otining mrkedly etter ess. The ottom two quintiles re never signifintly different from eh other (Annex Tle A2.3, Pnel B). Avilility of ollterl is positively ssoited with the proility of eing finnilly served, s well s nked. Interestingly, it lso hs signifint positive effet on ess to svings. There might e sustitution effet etween ess to ollterl, nd ess to Islmi produts nd insurne, s ollterl ess is signifintly negtively ssoited with ess to oth Islmi produts nd to insurne. 24 Exept in the se of informl finning (see, for exmple, the signifintly negtive ssoition with informl orrowing in Annex Tle A2.3 nd the signifintly negtive ssoition of ge with informl svings in Annex Tle A2.4). Pge 24

40 2.2. Interest in finnil servies is strong mong ll popultion segments, though finnil litery is lower mong rurl, lower-inome, nd femle groups Rurl res re more likely to express interest in finnil mtters thn urn res, ontrolling for inome. The sme is true for those with more edution, with employed sttus, heding the household, in the forml setor, nd regulrly using ell phone. Age nd gender do not impt interest in finnil mtters, though they re signifint determinnts of the willingness to ttin ess to finne, long with edution. The vilility of ollterl nd government jo lso signifintly positively ffet motivtion to join the finnilly served. These findings 25 hve importnt poliy implitions for finnil wreness mpigns, in terms of the seletion of relevnt udiene given limited resoures. In ft, 51.2 perent of the popultion express interest or strong interest in finnil mtters (51.7 perent in rurl res; Tle 2.2). Nerly two-thirds (60.6) of households follow wht is written or sid out finnil mtters, eqully in urn nd rurl res. Pkistni households get their finnil informtion predominntly from fmily (41.1 perent) nd the medi (24.5 perent). Rurl res rely muh less on medi thn urn (18.3 perent s ginst 35.1 perent), ut muh more on fmily (42.1 perent vs perent). This is true lso for lower-inome groups. Lower-inome groups hve muh less ess to nd wreness of finnil produts, though they hve roughly omprle levels of interest in finnil mtters (45.2 perent of the poorest quintile re interested vs perent of the popultion on verge). The most interest in further lerning, ross inome groups nd rurl/urn res, fouses on: preprtion of household/personl udget for 33.2 perent of the popultion, opening of nk ount (26.9 perent), nd svings methods (18.3 perent). Interestingly, insurne is not fous re of interest. Tle 2.2: Finnil Litery Men Rurl Urn Interested or strongly interested in finnil mtters (%) Follows wht is written/sid on finnil mtters (%) Soure: Pkistn A2F Household Survey 2008 The ulk of Pkistnis mke finnil deisions with fmily or spouse, not lone (78.1 perent). The effet is even more pronouned in rurl nd lower-inome groups. While women represent group tht is onsistently less finnilly literte nd more relint upon informl finnil soures thn men, they express strong desire to lern out sving nd udget plnning. Close to 44 perent of women show n interest in finnil servies, leit less thn men (59.4 perent). The onepts of profit nd lons, s well s ommittees nd moneylenders, seem loser to femles thn other topis. Mles enjoy etter ess to ll soures of informtion. Women get their news primrily from fmily memers, while men rely twie s muh s women on the medi. In ddition, women express stronger desire thn men to e eduted in res tht inlude prepring personl udget effetively nd lerning out wys to sve money. Women follow udget nd re more likely to sve regulrly, egging the question of their reltive exlusion s svings lients s ompred to men. They re lso more relutnt to tke on det (59.8 perent) thn men (48.2 perent) nd re more risk-verse (63.2 perent) thn mles (51.3 perent). 25 See Annex Tle A2.4 for eonometri results supporting this evidene. Pge 25

41 Tle 2.3: Litery Mesures in Rurl nd Urn Ares Herd nd do understnd Men Rurl Urn (perent) Lons Committees Profit Interest Bnk Moneylender Collterl Post offie sving ounts Mirofinne PLS/svings ount Insurne premium Credit rd Islmi nking Automted Teller Mhine (ATM) Moile nking Swift trnsfer Moile phone nking Household understnding of speifi produts is fir, with ommon finnil terms eing fmilir to over hlf to 90 perent of the popultion (Tle 2.3). Rurl res hve lower understnding, onsistently, with the exeption of the onepts of profit nd lons. Only 17.7 perent in rurl res understnd mirofinne (versus 26.9 perent in urn res). Rurl wreness of post offie svings ounts follows similr pttern. The lrgest differentils etween rurl nd urn res re in the understnding of more omplex produts suh s profit loss shring (PLS) /svings ounts, insurne premium, redit rd, utomted teller mhine (ATMs), Islmi Soure: Pkistn A2F Household Survey The low nk ess ntionwide is driven y inome onstrints nd the lk of informtion nd finnil edution. Of ll dults in Pkistn, only 12.9 perent re nked. This rnks low in the region, for exmple ompred with 66 perent in Indi nd 51 perent in Nepl. More thn 87 perent re not nked, yet 42 perent of those would like to e. Older, more eduted, employed persons, nd min inome erners, hve higher hne of eing nked. 26 Positive effets on nk ess re lso t work for individuls who hve ollterl, re employed in the forml/orporte setor, nd enjoy regulr ell phone use. Similrly, the three highest household inome quintiles re more likely to e nked thn poorer households. Lotion in rurl res does not hve signifint effet on the proility of eing nked. In ontrst, provines differ signifintly in the proility of their popultion eing finnilly served, even when informl finning is inluded. Bluhistn, due to inome, geogrphy, nd ulturl resons, lgs ehind in terms of offering its inhitnts finnil ess. Inome onstrints used y poor opportunities for produtive employment/entrepreneuril tivities re indeed the min reson for not eing nked, followed y lk of informtion/litery/wreness. Other resons (soio-ulturl, religious, moility, high fees, unfriendly stff, produt omplexity, or delys) re muh less ited thn personl resons. Lk of informtion is more ute mong women nd the poor non-nked. While svings re well understood (y 86.3 perent of the popultion), 83.8 perent do not onsider nk ounts essentil. 26 Demnd side regression nlysis introduing numer of explntory vriles to the proility of eing nked shows positive signifint effet of ge, edution levels, eing employed, nd eing the min inome erner on the proility of eing nked (Annex Tle A2.6). The regressions follow the speifition of Kumr (2005). Pge 26

42 Demnd-side evidene points to three min nks y mrket shre: Ntionl Bnk of Pkistn followed y Hi Bnk Limited nd Muslim Commeril Bnk re the most populr mong people with nk ounts, ll generting higher usge in urn res nd mong men. This suggests onsiderle potentil for rpid expnsion of ess to nking Csh-sed eonomy with high willingness of the non-nked to gin ess to nk servies Aout 41.6 perent of the non-nked popultion expressed high desire to own individul nk ounts, growing with inome. It is lso stronger mong men (51.4 perent) nd in urn res. Sving money ppers to e the min reson for the non-nked to desire nk ount, eqully for men nd women (ited y 26.7 perent of the popultion), nd ited twie s often in urn res (9.4 perent). The seurity of eing le to withdrw money s needed is stronger onern for women thn men (11.4 perent vs. 9.7 perent). Gurding ginst theft is ited y 3.9 perent on verge (twie s muh in urn res). Aessing lon is still three times less ited thn sving s motive to ess the nking setor, nd is mostly of interest to men (11.4 perent vs. 4.8 perent women). Mking deposits from inome is the third most populr reson, gin muh more populr in urn res. Figure 2.5: Csh-sed Inome Aross Setors Soure: Pkistn A2F Household Survey 2008 Csh domintes the inome reeived in forml nd informl setors. Ten perent of the popultion reeives inome in kind. Nerly 16 perent put ll or most of their inome into their nk ounts. Almost third don t put ny, in prtiulr women (lmost 39.2 perent) nd rurl res (32.5 perent). The poorest re three times more likely thn the rihest to not sve ny inome t the nk. New delivery infrstruture represented y eletroni trnsfers represents non-eligile shre of government nd slry pyments. Currently, 10.8 perent of government slry pyments nd 3 perent of pyments for pensions re mde through eletroni trnsfers. Colleting pensions or slry pyments re strong drivers for opening n ount. Independently of inome nd other hrteristis, eing government worker inrese one s hnes to e formlly inluded y 7.5 perent. More thn hlf of working Pkistnis re self-employed, in prtiulr mong the poor. 27 Adding other informlly engged groups (suh s housewives), we find tht onsiderle mjority of the popultion reeives irregulr nd /or informl ernings. This is signifint rrier to eing nked, given urrent nk servie proedures nd pproh. Credit informtion vilility ould permit the use of redit soring tehniques, whih would eliminte suh rriers to ess. Outside of orrowing, suh rriers rise due to the smll size of the trnstions nd the high per-ount ost to the nking institution. Yet tehnology dvnes hve gone long wy towrds resolving this issue Mirofinne Bnks (MFBs) interntionlly report up to 12 perentge points of redution in intermedition osts ttriutle to tehnology improvements. 27 The self-employed re predominntly mle (see Annex Tle A2.10), with lower edution thn the verge individul. Those informlly employed lso tend to e younger, nd live in urn res. The formlly self-employed tend to regulrly use ell phones, not so the informlly self-employed. Pge 27

43 Usge nd penetrtion of finnil produts is limited in rnge, nd not well suited to different mrket segments The min produts used y the mjority of nked popultion re si svings nd heking ounts (Figure 2.6; svings nd orrowing produts re disussed in more detil in setion 2.4 nd 2.5, respetively, lter in the hpter). Yet different popultion segments, suh s women, might require produts more ttuned to their needs, s illustrted in Box 2.1. Sving ounts re held y 8.2 perent of the popultion nd urrent/heking ounts y 6.5 perent. Crd penetrtion is low (2 perent), with 10-fold higher penetrtion of oth redit nd ATM rds in urn households. These findings re losely ssoited with the undeveloped ATM infrstruture highlighted in setion of Chpter 1. Crds nd svings ount re lso muh more frequent mong higher-inome mles thn lower-inome groups or women, with the exeption of Pkistn Post svings ounts nd ommittees. Figure 2.6: Produt Penetrtion (s % of nked) Soure: Pkistn A2F Household Survey 2008 A loser look t the determinnts of ess to speifi produts suh s redit/deit rds, heks, svings ounts, nd urrent ounts, further informs the nlysis. The proility of holding vrying nking instruments inreses with ge, edution levels, nd other hrteristis tht inlude eing the min inome erner nd eing employed. Bnk produt users re unlikely to e loted in rurl res. A position with government, ut not with the forml orporte setor, inreses the proility of eing nked. There is positive ssoition etween speifi produt usge nd the upper inome quintiles. Just like for ess to nking servies, women generlly hve less ess to speifi nking produts (Annex Tle A2.9). The ntionl household survey dt results suggest tht Islmi finne is high-end produt: the mle, riher, urn, household hed, more eduted, nd pertining to the forml setor popultion is signifintly more likely to use Islmi produts (Annex Tle A2.10, regression V). Interestingly, Islmi finne is inversely ssoited with ge, employment, nd vilility of ollterl, s well s with employment in the orporte setor. Only fifth of the popultion re willing to py more for Islmi produts thn onventionl nking produt. Women re less willing thn men to do so (14.7 perent vs perent). Aess to Pkistn Post nd insurne servies is determined y muh the sme hrteristis s stndrd nking produts ge, edution, employment, heding household, eing employed y government, nd eing prt of the forml setor. However, lotion in rurl re is positively ssoited with oth insurne nd post offie ess, s opposed to stndrd nking (non-postl) produt ess. Informl finne vi ommittees lso differs signifintly from stndrd nking ess in terms of determining ftors. The positive effet of employment is out the only ommon ftor we oserve. The ommittees re oriented towrd women, who re signifintly more likely to gin ess to finne vi Pge 28

44 rotting sving nd redit ssoition (ROSCA)-type sheme. They re further not foused on older groups, nd do not ter to other underserved groups suh s frmers nd lorers. These effets re rooted in the ulture nd trdition of mintining ommittees mong mrried women s soil undertking Chief ostles to ess inlude hevy doumenttion, nd lk of inome nd informtion The most frequently ited pereptions of forml finnil servie providers pertin to requirements for eoming ustomer (ssoited with 53.8 perent of the popultion). 28 Commeril nks require n identity doument, permnent ddress, redit referenes, nd gurntor. The urrent requirements for two identity douments on verge to open n ount in ommeril nk in Pkistn, 29 while not ppering to e serious hurdle for the popultion in generl, my relistilly represent signifint ostle to eing nked for ertin popultion groups, suh s women or low-inome groups. This ples Pkistn s numer one in South Asi, etter thn the South Asi verge (2.65 douments required), though elow top performers suh s Thilnd. 30 Aording to the World Bnk (2006), moreompetitive nking setors re less ureurti, s nks tht re vigorously ompeting ginst eh other tend to simplify proedures. The seond most importnt luster of issues defining pereptions of forml finne re personl resons, ited y 50.7 perent of the popultion, inluding wreness limittions nd self-exlusion. 31 Bluhistn highlights the lk of produt informtion s its mjor ostle. Lk of informtion is lso more ited y non-nked women nd is muh higher mong the poor. In ddition, 5 perent more women thn men express no need for nk ount, mking it the iggest reson mong women for not nking. This pereption dereses with inome, nd might lso e relted to the wreness of the enefits of finnil produts, espeilly mong women ( If I don t work, why would I need finnil instruments? ). Pereptions on essiility issues, while minor ostle overll (4.9 perent), 32 re more ute in rurl res. While n verge of 23.2 perent of the popultion onsider tht ommeril nks re too fr wy, one-third in rurl res pereive this issue. Committees nd informl moneylenders on the other end do not prompt essiility issues. Very few (6.5 perent) of urn households trvel more thn n hour to the nerest nk, s opposed to 34.5 perent in rurl res who trvel one hour or more (up to two dys). 33 Cost to reh nk is lso more ited y rurl res (2.8 perent of non-nked vs. 2.5 perent of non-nked in urn res). While lose to hlf of the urn popultion spend Rs 20 ($0.26) or less on 28 Speifilly, respondents felt tht they did not hve enough ollterl, too muh doumenttion ws involved, referenes were hrd to otin, they did not hve n identity doument, or did not hve enough money, regulr inome, or suffiient minimum lne to open n ount. 29 World Bnk (2008). While identifition doumenttion penetrtion is high (80.2 perent on verge, though lower for femle nd poorer inome groups), other douments re extremely hrd to ome y. Eletriity ills nd household ownership douments re distnt seond nd third in penetrtion (12.1 perent nd 8.3 perent on verge, respetively) nd re more frequently possessed y higher-inome households (with lrge gender divide in fvor of men in oth instnes). 30 With its top three nks mking up less thn 45 perent of the nking system ssets nd, on verge, only one doument required to open n ount, Thilnd is n exmple of virnt nking setor. 31 Personl resons inlude lk of informtion out nk produts, preferene to del in sh, no pereived need for eing nked, lk of self-onfidene, unesiness in the nk environment, no time to go to the nk, nd no knowledge of how to open n ount. 32 Aessiility issues inlude the ost nd time to reh nk. 33 Wlking is the most ommon mode of trnsporttion to reh finnil institution in urn res (38.9 perent) vs. puli trnsport in rurl res (48.5 perent rurl respondents ommute to their finnil institution y trin, rikshw, tong, us, or vn). Aout 43.9 perent of respondents omine other tivities while going to the nk; this is true ross the inome spetrum, though with women little less, t 36.3 perent vs perent men. Pge 29

45 one-wy trip to the nk, over hlf in rurl res (50.6 perent) hve to spend Rs ($0.26 to $1) nd 14.9 perent inur more thn $1 eh wy to reh nk (Figure 2.7). Figure 2.7: Cost of Bnk Aess Estimted Time to the Bnk (one wy) Estimted Cost to the Bnk (one wy) Soure: Pkistn A2F Household Survey 2008 Servie-relted resons for not eing nked seem less pronouned (5.1 perent). 34 Fees nd lengthy proedures, ited on verge y 1.9 perent of the nonnked popultion, do not pper s strong lok to eing nked. Only out 3.5 perent of the nonnked point to low returns s reson for not hving nk ount, while out the sme proportion of the nked (3.6 perent) list erning inome s reson for eing nked. This suggests wek ssoition etween hving n ount nd erning interest on it, onsistent with surveys in other prts of the world. 35 The slow servie nd somewht onerous doumenttion requirements of Pkistn Post nks re lso frequently ited. Pkistn rnks ner the top for the world in terms of highest opening fees (pplition fees nd other types of fees). While it tkes minutes to open n ount in mny developing ountries, Pkistn is mong the ountries where it tkes longer, on verge 2.8 dys, muh eyond the South Asi verge of 1.21 dys (World Bnk 2008). 36 While developing-ountry nks re more likely to hrge fees, tht tends to result in lower finnil ess, even fter djusting for inome. Soio-ulturl resons suh s trust, religion, nd ulturl rriers to ess re lso not importnt overll (8.5 perent of the popultion) ut re very strongly skewed towrd women nd Bluhistn. Suh resons for not nking re ited y 11.9 perent of the non-nked women (ompred with mrginl 1.3 perent mong non-nked men) nd over 9 perent of the non-nked popultion of Bluhistn nd Sindh. This reson does not pper strongly orrelted with inome. Religious resons suh s issues with pying or reeiving interest, nk mistrust, nd issues with orruption re ited y only 2.5 perent of the non-nked. 34 Servie-relted issues explored inlude: hrges nd fees re too high; returns re low, hours of opertion re inonvenient; long lines t the nk; stff re snoish nd not helpful, spek in omplited terms, do not understnd lient needs, or give importne to the lient. 35 A survey overing 2,000 urn individuls in Brzil lso highlighted wek ssoition etween rtes of returns nd mking deposits. See Aess to Finnil Servies in Brzil, p World Bnk (2008). Bnks in South Asi re lest likely to hrge monthly mintenne fee for nk ount nd hrge nnul deit rd fees, whih enourges more ess. Eighty-six perent of nks surveyed opened the ount within few minutes of the pplition or on the sme dy (World Bnk 2006). Pge 30

46 2.4. Informl finnil servies re pereived s more geogrphilly essile nd less omplex, nd they re the only finnil instruments tht the mjority of Pkistnis will ever know nd use. Informl finnil mrkets re disussed from supply stndpoint in setion 3.7. Informl servies hve fewer requirements nd re esier to understnd. As noted, informl svings re pervsive (63.6 perent of ll svers), e it t home (in sh, gold, or jewelry) or in ssets (livestok, lnd). Ninety-seven perent of orrowers lso orrow informlly, e it from shopkeepers (77.9 perent of ll orrowers) or from friends nd fmily (53.5 perent of ll orrowers). Groups tht re less trgeted y the forml setor rurl res, women, nd the poor hve higher shre of informl svings in totl svings. Setion 3.7 presents detiled disussion of supply issues for informl finne, while this setion fouses on distinguished hrteristis or the informl setor reltive to the informl setor nd on nother populr informl vehile, the ommittee. 37 Wht n e lerned from informl finnil systems? Using ess to redit s enhmrk, the pereption of orrowers is tht the informl setor opertes with muh lower servie requirements thn the forml setor, in prtiulr in terms of o-gurntor (ited y 4.8 perent vs 46.5 perent), immovle ollterl (1 perent vs 27.5 perent) nd minimum lnes (0.3 perent vs 10.3 perent) onditions (Figure 2.10). Over qurter (27.5 perent) of the popultion ite immovle ollterl s ondition for forml redit. In rurl res 36 perent ite immovle ollterl requirements. Setion 3.7 suggests tht informl soures might hve some dvntges over forml soures. In generl, informl soures of redit re more expensive thn forml soures, nd there is muh evidene to suggest tht they exploit poor people, ut they hve the dvntges of eing ville ll the time, not requiring doumenttion, operting outside the purview of forml uthorities, nd not requiring ollterl. Informl mrkets del effetively with the higher risk they fe from lients, sustituting soil ontrts nd lient informtion for immovle ollterl tht is often not ville to the poor, lending only for produtive uses (sh lending is frequently done t perent premium), nd on verge hrging only smll premium over the forml setor (23 perent vs 19 perent, Qdir 2005). Pereptions of servie providers give striking perspetive of gps etween forml nd informl institutions. Positive pereptions suh s sene of legl formlities, of servie hrges, nd of doumenttion rnk re frequently ssoited with ommittees or informl moneylenders. Conversely, negtive pereptions nd onstrints re the most frequently ited out forml finnil institutions, in prtiulr the need for redit referenes, doumenttion, nd o-gurntor. Fous on ommittees. Although the onept of ommittees is widely understood y 87.9 perent of the popultion, ommittees rnk only s the third svings hnnel, fr ehind home sving nd sving in ssets. They re lso only the third most populr orrowing hnnel, ehind shopkeeper lons nd fmily/friends. This is surprising, espeilly ompred with previous studies of limited smple sizes tht suggest the mjority of Pkistnis re involved in suh group shemes. Bllot ommittees re muh more frequently used thn other ommittees (nmed y 87 perent of ommittee users) mong the ommittee users. Committees re not used signifintly more y women (14.0 perent) thn men (14.5 perent). They do not ter to other underserved groups suh s frmers nd lorers. Finlly, ommittees re used more 37 Committees, lso lled ROSCA, re groups through whih prtiipnts join ontriutions to lend to single memer. Memers tke turns to ess ommittee lons, through n order determined y list, lottery, or ution. By ggregting individul funds nd hnneling them in pre-determined order to individul memers, ommittees ply n importnt intermedition funtion, sed on revolving funds. Committees do not require physil pitl s ollterl. Insted, repyment is sed on reiproity nd soil pressure. Pge 31

47 thn twie s muh in urn res thn in rurl res, prtiulrly in Punj nd Sindh. Their use lso inreses with inome. Women nd urn res re signifintly more likely to prtiipte in ommittees (lst regression in Annex Tle A2.9). The inidene of ommittee prtiiption is lso found to inrese with edution levels nd derese with ge. While eing employed in the forml or informl setor hs positive inidene on using ommittees, the professionl profile effet on the likelihood of ommittee prtiiption reinfores their urn nture. Committee prtiipnts re more likely to e forml orporte workers. Committees re mostly used to sve s well purhse household goods. Few prtiipte in ommittees to purhse fixed ssets (7.3 perent) or invest into usiness (12.1 perent) (Figure 2.8), lthough slightly lrger shre thn overll forml or informl svers (10.76 perent) list usiness investment s svings motive. Figure 2.8: Resons for Prtiipting in Committee Most users prtiipte in only one ommittee (84.1 perent). Contriutions to ommittees re mde monthly. Chnge in inome hs the highest impt on hnges in pyments to ommittees, with the exeption of Bluhistn, where fmily resons hve the highest impt on mounts pid to ommittees. While pereptions expressed on the servie of ommittees re very positive, nd involve trust, good servie, nd ese of use, women hve n even etter pereption of ommittees thn the verge individul. Soure: Pkistn A2F Household Survey Svings nd investment re the most populr finnil servie, nd mostly onentrted in the informl setor Among finnil produts, svings nd investments, y fr the most populr, re used y two-thirds of the popultion followed y lons nd redits, used y one-third. Among those who sve, svings re mostly done informlly (only 14.9 perent of svers go to forml finnil institutions). Aout 63.6 perent of svers sve t home. Committees re the seond most populr informl svings hnnel (25.7 perent) for the svers, s muh mong women s men (26.6 perent nd 24.8 perent, respetively). 38 Sving with fmily nd friends (6.7 perent) is done eqully mong rurl nd urn res, ross inome groups. Three-qurters of forml svers were lso sving informlly. This topi will e piked up on in more detil elow nd in the following two setions, whih disuss the intertion of the forml nd informl setors in generl nd in terms of orrowing speifilly. Demnd-side evidene links higher inome levels to higher proility of sving, formlly or informlly (Annex Tle A2.8). Rurl res re more likely to hold svings ounts overll, nd in the informl setor, while urn res hold more forml svings ounts. While women s ttitudes towrd finnil servies suggest higher svings motives thn men, women re exluded from ess to most svings hnnels; ll hrteristis held onstnt, women re found less likely thn men to e svings ount 38 The figure inludes group svings shemes. Pge 32

48 holders. Women re found to e more likely thn men to use informl svings hnnels thn not sve t ll. The other hrteristis ssoited with n inresed likelihood of sving inlude n employed sttus, edution, nd eing the min household inome erner. While ge is importnt for forml svings, it plys if nything negtive role in the proility for informl svings. Government workers re positively ssoited with svings (nd minly forml svings), while frmers re lso positively ssoited with svings ounts, ut upon disggregting, the effet minly omes from the informl setor ounts. Finlly, lorers hve lower proility to ess svings instruments. Regulr ell phone users re lso more likely to e svings ount users. Figure 2.9: Government / Privte Svings Aounts Only smll shre of the sving volume is ptured in the ommeril nking setor. The most populr svings produt, the PLS/svings ount, is used y only 11.2 perent of svers (Figure 2.9). This proportion inreses with inome levels (2.1 perent for lowest quintile vs perent for the highest), men (18.6 perent vs. 3.8 perent for women), nd urn res (16.2 perent vs. 8.4 perent for rurl). The penetrtion of post offie svings is ppreily less thn tht of post offie remittne produts (disussed in detil in Chpter 5). Despite its presene throughout Pkistn nd its existing remittne ustomer se, the Pkistn Post Bnk my not e Soure: Pkistn A2F Household Survey 2008 leverging its reltive proximity in some provines to diversify further into nking nd sving produts. Post offies nd Stte Life Insurne re the mostused nonnk forml finnil institutions for sving, used y 0.7 perent nd 0.6 perent, respetively, of those who sve, lthough well ehind the nks. Mirofinne svings (0.2 perent) nd Islmi svings (0.1 perent) represent minor svings hnnels. Within the forml setor, government svings (provident funds, pension funds, prize onds, nd government svings ertifites) nd privte svings (PLS/sving ount, post offie svings ount, mirofinne, Islmi svings, nd shres) produts re oth populr mong Pkistn households who sve. Among the privte nd government sving produts, PLS/svings re the most populr with over 11 perent people using them. While the penetrtion of pensions in nnuity grows s inome levels inrese, the produt is most populr in AJK nd mong men. Svings is for smoothing onsumption, not investment, though rurl res invest more. Resons for sving seem to revolve round immedite needs suh s food, household goods, nd medil nd shooling osts, in rurl nd urn res like. Mrrige nd deth/disility rnk high mong sving motives. Longer-term prospets suh s retirement nd future uying prospets suh s lnd re mrginlly ited, in prtiulr mong the poorest. Sving to invest in or strt usiness, on the other hnd, lthough overll low, is more ommon in rurl res (12.5 perent vs. 7.8 perent), irrespetive of inome. The ulk of rurl svers use money to invest in ttle (54.4 perent) or frm lnd (19 perent). Close to 1.3 perent of svers invest in working pitl suh s mhinery or uying inputs for smll usinesses. Retirement, pension fund, nd life insurne investments re mrginl, t 1.2 perent of svers on verge Regression results onfirm nd enrih these findings (Annex Tle A2.12). While oth forml nd informl svers re motivted y investment motives, edution, nd soio-religious resons, only informl svers re further driven Pge 33

49 2.6. While orrowing is mostly informl, strong version to ny type of det is oserved A third of the popultion (35 perent) hve lon or redit, ut only one in 14 reeived it from the forml setor. Almost hlf of the Pkistnis (44.9 perent) hve never ontrted lon (55.5 perent of women nd 33 perent of men). A strong version to det nd ssoited is ginst orrowing is oserved nd shred ross gender, rurl res, nd inome levels (61.3 perent). Religious ojetions to orrowing, while expressed y third of the popultion, do not represent suh signifint effet on orrowing s the version is, nd ome out insignifint in regressions (Annex Tle A2.10). To gin n understnding of the extent of orrowing version mong the popultion, onsider tht the most widespred form of orrowing (77.9 perent of ll orrowers) is getting goods on redit interest-free from shopkeeper, whih is more ulturlly determined prtie thn finnil trnstion. Borrowing interest free from friends nd fmily (53.5 perent) omes in lose seond. The use of shopkeepers, s well s fmily friends, s n informl soure of orrowing is sustntil mong rih nd poor, urn nd rurl res like. All other types of orrowing re rre (inluding moneylenders, interestingly, t 2.9 perent on verge). Personl lons from nk re ited y 1.8 perent of the orrowers nd griulturl lons re ited y only 2.1 perent of rurl orrowers. Credit rds usge is reserved to the highest inome quintile. Islmi nd mirofinne lons re sprsely used 0.03 perent of orrowers held n Islmi lon t the time of the survey. 40 Aout one-fifth of the popultion re willing to py more for Islmi thn trditionl produts, less so women thn men (14.7 perent vs perent). Hjj/Umrh shemes re used only y 0.2 perent of Pkistnis. Furthermore, 3.5 perent of the popultion ought, uilt, or renovted their houses with mortgge or forml or informl lon. World Bnk (2002) notes the onsiderle literture tht exmines the intertion of the informl nd forml redit setors. Some orrowers re found to e simultneously tive in oth mrkets, either euse they re unle to get ll the redit they need from one setor lone, or euse forml lenders might ondition their lending on the orrower s otining o-finning from other (presumly informl) soures. This is the se in Pkistn, oth in terms of svings (see setion 2.4), nd orrowing (Tle 2.4). Borrowers use their dul ess more in rurl res (one-third of urn forml orrowers lso ess the informl setor, vs. more thn hlf of forml orrowers in rurl res), onfirming the trend oserved in setion suggesting more redit-onstrined rurl mrket. Tle 2.4: Forml nd Informl Borrowing Urn Rurl Totl Forml redit holders Informl redit holders Forml nd informl redit holders Soure: Pkistn A2F Household Survey 2008 Aess to redit inreses with ge, nd employed sttus, nd edution (Annex Tle A2.7). The three highest inome quintiles re likely to orrow more thn poorer groups. Interestingly, nd onsistent with previous findings, rurl lotion lone does not exert negtive effet on ess to lons. Women re less likely to ontrt n informl lon, though more likely to ontrt forml lon, presumly euse y household needs nd medil, ident, emergeny, nd old-ge provisions. Presumly, forml svers hve overed these motives y other finnil produts. Femle, rurl, younger individuls, household memers other thn the household hed, frmers, nd less-eduted individuls re more likely to sve informlly. In ontrst, government employees nd regulr ell phone users re more likely to sve formlly. 40 We do not quote mirofinne lon figures due to suspeted survey is s disussed in Appendix A (Clirtion). Insted we go with offiil SBP figures, whih put mirofinne lients t 1.7 million. Pge 34

50 women s responsile money mngement hits re tken into onsidertion in the forml finnil setor. Overll, women seem to e less likely to orrow thn men, minly s self-expressed preferene. Heding household lso improved one s hnes to ess finnil servies. Positive effets on ess re lso t work for groups tht hve ollterl, nd re employed in the government or s frmer. Lorers fe signifintly lower ess to lons. Those who enjoy regulr ell use or hve ess to ell phone re signifintly ssoited with orrowing. It is worth noting tht informl orrowers re likely to hve ess to s moile phone ut not regulrly use it, s opposed to forml orrowers, who tend to e signifintly ssoited with regulr ell phone usge. Tle 2.5: Purpose of the Lon (% of ll orrowers) Forml Informl Urn Rurl Totl Urn Rurl Totl Personl lons Business/frm lons Soure: Pkistn A2F Household Survey 2008 Notes: Personl lons inlude: lons to uy food or goods, meet expenses of dowry, wedding, Hjj/Umrh, unforeseen emergeny osts (hospitl, hildirth, funerl), uy or uild house, trvel, uy luxury items/durles, nd py utility ills. Business lons inlude lons to uy goods for griulture or usiness, produtive ssets, meet expenses of usiness. Like sving, orrowing is used in mjority of ses to ttend to personl needs (deresing s inome inreses) rther thn to invest in usiness. Rurl res orrow muh more thn urn, nd tend to use the lon muh more for produtive purposes thn for onsumption (Tle 2.5). 41 While forml orrowers tke lons for vrious investment purposes (griulturl, nonfrm, nd housing investment), informl lients orrow exlusively for household needs. Borrowing for personl needs dereses s inome inreses, while orrowing to invest is positively orrelted to inome. Reltive to their orrowing portfolio, men orrow more for produtive tivities thn women. On verge, one-fifth of those who orrowed (19.1 perent) ever missed pyment, the poorest more thn twie s frequently s the rihest (25.9 perent vs perent). Figure 2.10: Requirements Before Reeiving Lon In ddition to redit requirements (see setion 2.7), hief ostles for orrowing inlude generl version to det, religious is, nd signifint likelihood of getting rejeted. Almost third of Pkistnis onsider tht interest rtes go ginst religious priniples (twie s mny s those who think tht interest rtes re not prolem). These views, lthough shred throughout inome groups, re strongest mong the highest inome groups nd men (43.8 perent versus 25.2 perent for women), suggesting tht they re not the mjor ostle Soure: Pkistn A2F Household Survey 2008 where finnil ess enhnement is relevnt mong the poor nd women. In ddition, religious ttitudes re only one prt of the negtive pereptions towrd redit, dwrfed y the generl version to eing in det, whih is pervsive ross gender (58.1 perent for men nd 64.2 perent for women), lotion, nd inome divides. 41 Regression nlysis demonstrtes similr ptterns (Annex Tle A2.13). Pge 35

51 Onerous nd omplex redit requirements re onsiderle ostle to forml redit. The most frequently ited forml lon requirements re gurntor (46.5 perent), movle ollterl (27.5 perent), nd utomti pyroll dedution (20.3 perent). Informl requirements re miniml to nonexistent (Figure 2.10). Over qurter (27.5 perent) of the popultion ite immovle ollterl s ondition for forml redit. In rurl res, 36 perent ite immovle ollterl requirements. Collterl onditions re muh less ited thn for informl lons. Tle 2.6: Ftors Considered efore Tking Personl Long-term Lon (% of forml, respetively informl, orrowers) Forml Lon Informl Lon Likelihood tht the lon will e grnted in the mount needed Getting the lon pproved s soon s possile The ffordility of the monthly instllment Their interest rte hrges Soure: Pkistn A2F Household Survey 2008 The likelihood of eing pproved is n importnt ftor onsidered efore tking personl long-term forml lon. Diverse s forml nd informl lons re, orrowers re preoupied with the sme issues: suessful outome, speed, ffordility, nd rtes, in tht order (Tle 2.6). Three-qurters of forml orrowers pplied within the pst 12 months (76.6 perent). Rejetion rtes were t out 10 perent, nd higher for rurl res (11.4 perent) nd for women (17 perent) (Tle 2.7). Aout 41.4 perent of the existing forml orrowers whose pplitions were rejeted re onfused out the reson for their unsuessful pplitions, suggesting servie issues on the prt of the nks. Other resons ited for unsuessful pplitions inlude existing over indetedness, sent redit history, insuffiient ssets, nd lk of existing ounts. Tle 2.7: Rejetion Rtes for Existing Borrowers (% of ll forml orrowers) Popultion Urn Rurl Men Women Existing orrowers with lon pplition in pst 12 months Shre of lon pplitions in pst 12 months whih were rejeted Rejetion rtes Soure: Pkistn A2F Household Survey 2008 Emergeny pyments re feture of the lives of urn nd rurl, rih nd poor; informl orrowing is the usul response to suh exigenies. Emergeny pyments not overed y regulr inome ffet rurl nd urn res eqully. The poorest households (15.4 perent) s well s the rihest ones (12.1 perent) hd to mke n unntiipted emergeny pyment in 12-month period prior to the survey. Emergeny pyments re more thn twie s often linked to involuntry thn voluntry pyments (70.1 perent vs. 35 perent of emergeny pyments). Involuntry pyments linked to helth issues dominte (48.9 perent of the popultion hd emergeny medil expenses in the pst yer). Religious nd soil purposes (wedding elertions were finned y 22.4 perent of Pkistnis in the pst yer), sore the highest mong voluntry pyments. Pge 36

52 Households tht mde n emergeny Figure 2.11: Soures of Funding of Emergeny Pyment pyment used informl orrowing most often, suh s lons from friends/reltives for more thn hlf of the ffeted households (67.1 perent), lons from shopkeepers (10.3 perent) or sh from home (13.4 perent) for n emergeny pyment. Forml orrowing soures were little mentioned y these households. Svings re lso wy to mitigte emergeny expenses nd re rnked seond in soures of funding (17.1 perent) fter lons from Soure: Pkistn A2F Household Survey 2008 friends/reltives Insurne is little used, though demnd is high for life nd vile rop insurne produts. Insurne is used y only 1.9 perent of the popultion. Its penetrtion is higher in urn (2.3 perent vs 1.7 perent in rurl) res nd mong the upper inome groups (4.6 perent vs. 0.3 perent for the lowest inome quintile). Forml insurne hnnels re more used thn informl ones. Life insurne is virtully the only produt (used y 90.8 perent of insurne enefiiries). Resons for not hving insurne point more towrd litery issues out onept/enefits rther thn out using insurne (Figure 2.12). Stte life Insurne is the most populr entity mong insurne ompnies (serving 71.4 perent of lients), followed y employer insurne shemes (11 perent), Admjee Insurne (5.2 perent), EFU Generl Insurne (3.4 perent) nd New Juilee Insurne (1.4 perent). Insurne is reserved for higher inome groups. Age, edution, employed sttus, government jo, nd elonging to the two highest inome quintiles ll individully inrese the inidene of eing n insurne holder. Interestingly, rurl res re signifintly positively ssoited with eing insured, though of ourse they lso profess the highest insurne needs (for rop/life insurne). Figure 2.12: Resons for not hving insurne (% of uninsured) Soure: Pkistn A2F Household Survey 2008 Pge 37

53 2.8. Little use of hek servies eyond government pyments. More thn twie s mny people in urn res know how to write hek s people in rurl res, nd three times s mny men s women. Bluhistn ppers to e the lest-profiient provine, orroorting erlier finnil litery results. Figure 2.13: Mens of Reeiving Inome Aout third of the hek-profiient popultion (33 perent) hd shed hek in the month prior to eing interviewed. Rurl/urn nd gender divides re onsistent through profiieny nd usge. Eight-four perent of hek users hve shed heks themselves t nk or Pkistn Post nk, while 17.2 perent hve shed heks through someone else. Possile resons for the wek uptke of heks s pyment infrstruture inlude widespred fees: third of those who shed hek in the month preeding the survey pid fees for this servie. Wide provine differenes Soure: Pkistn A2F Household Survey 2008 exist from 20 perent in NWFP to 38.1 perent in Sindh py fees for hek-relted servies. Inome nd edution, employed sttus, ge, vilility of ollterl, heding the household, nd working in government re signifint determinnts of using heks. Chek users lso tend to e regulr moile phone users. Women re less likely to use heks. The little uptke of heks n e linked to the sh-sed nture of the Pkistni eonomy. The literture finds tht other ftors underlying little hek usge interntionlly inlude high fees, delys in hek lerne, physil distne from nk rnhes, nd reliility of hek s pyment instrument given the inherent disinentive to write d heks, s well s trust in the system. Existing mesures set y SBP to enourge hek usge inlude provisions in the Negotile Instruments At to tegorize ouned heks s riminl offene, s well s tx mesures involving the withdrwl txes on sh. Reommended next steps would inlude tht SBP lunhes further wreness mpigns on these new provisions. 2.9 Suggested strtegies for expnding ess to finne for the underserved A mjor drive to enhne finnil inlusion would involve joint effort of SBP, the ntionl government, privte setor, the ommunity, nd donor efforts. The est et for rpid sling-up of ess is to rely on tehnology, litery gins, finnil re-engineering of proesses nd produts, nd n enling legl nd institutionl frmework. Wy Forwrd for the Privte Setor Diversifying the produt rnge to inrese outreh: The rod hed ertinly lso lies in produt diversifition, with more servies nd fewer requirements tering to the mss popultion of lowerinome pity. Smller size of produts, nd ulk servie might etter ttrt lower-inome groups. Mrket segmenttion nd produt diversifition n lso lower osts nd help nks risk mngement. Poliymkers nd nks would note the importne of the mrket for finnil produts trgeting heds of households, given the finding tht eing household hed hs signifint impt on finnil ess. Pge 38

54 This potentil is prtiulrly of interest in rurl res s well s lusters of urnized rurl migrnts into igger ities. Rehing out to the femle lient: Women present n untpped potentil finnil lient se for Pkistn. Better mngers of their det responsiilities nd more loyl lients, women re lso more ssiduous svers. Yet produts nd proedures re not ustomized to the speifiity of women s needs. Fousing on women hs een profitle proposition in interntionl mirofinne experiene. The privte setor hs strted to segment femle lients in Pkistn y offering them tilor-mde produts. This is promising usiness proposition nd strtegy for inresing ess to finne. Lerning from the informl finnil setor s suess: The gp etween the populrity of informl finne nd limited forml finnil setor outreh, omined with the pereption gp etween hevy proedure-driven forml setor nd informl setor with minimum requirements, lls for lerning from informl finnil systems nd developing linkges with the informl setor. Strengthening finnil infrstruture, y expnding redit informtion ureus nd other pyment systems, would llow nks to lower inding requirements suh s gurntors nd immovle ollterl, therefore ttrting new ustomers mong women, the poor, nd residents of rurl res. Rehing out to the rurl lient: Household survey evidene suggests tht there is nothing unserviele out rurl res. In ft, the lrger use of svings for produtive use in rurl res nd positive ssoition with insurne nd post offie finnil produts represent n untpped opportunity for ommeril servies. Leverging tehnology: Tehnology n e used s n enler of ost redutions within finnil institutions s well s for innovtive pplitions for servie delivery. Further ost redutions thnks to the use of tehnology within nks would llow expnsion of the ustomer se y deresing fees nd inresing the qulity of produt nd servies. The positive ssoition etween ess to forml finning nd regulr moile phone use rries potentil for new servies mong n existing ustomer se. The signifint ssoition etween the informlly served nd moile phone ess unleshes opportunities to tp into new forml ustomer se lredy operting in the informl setor. Wy Forwrd for the Puli Setor Creting wreness of the enefits of ess to finnil servies: Further gins in finnil litery re ritil. Awreness-retion nd trust-uilding ould e forged through soil moiliztion nd mss medi. Stronger institutions (inluding SECP nd Credit Informtion Bureu) re mjor prt of rpid inrese in finnil inlusion. More effiient post offie nking is must (following the suessful exmples of Brzil nd Chin). Given the signifint developmentl potentil of Pkistn Post to enhne finnil ess, it needs moderniztion in opertions nd regultions. Following most suess ses in Est Asi, s well s other ontinents, PPSB should e pled under the supervision of SBP. This is the se in Chin, Philippines, Thilnd, nd Sri Lnk, 42 s well s out hlf of the ountries with seprte postl nk in Afri, lmost ll ountries in Europe nd Centrl Asi, nd onsiderle numer of ountries in Ltin Ameri, s well, inluding Argentin nd Brzil. Creting n enling environment for expnding ess to the underserved: Regultions should keep up with the needs of the setor nd tehnologil developments, to enle expnsion. Simultneously, n enling environment should go hnd-in-hnd with refully hosen government presene. 42 Not in Indonesi, where hlf of postl servies re under privte setor provision. Pge 39

55 Indisriminte susidies, espeilly of the interest-rte vriety, n e detrimentl to expnsion of the setor, s they not only distort pries ut rowd out effiient institutions nd produts. The government should resist populist pereptions tht low interest-rte funding n serve developmentl purpose. Even more detrimentl re stte-owned institutions reted to promote finnil ess. In ddition to wek evidene in the se of Pkistn tht suh institutions tully improve ess (for exmple, SME Bnk), these efforts wste vlule puli resoures tht ould more usefully e deployed elsewhere, nd eliminte the level plying field for mrket prtiipnts. In sum, while finnil ess is low, nd mostly limited to informl finnil servies, interest in finne outstrips effetive ess rtes onsiderly. Poverty nd lk of informtion on finnil servies predite lk of interest the pereption of irrelevne of the finnil system to the everydy lives of onsiderle shre of the Pkistni popultion. Severl groups seem prtiulrly exluded, though mrket studies suggest they ould e vile ustomers if the forml finnil setor dopts pproprite servies nd instruments to reh these groups residents of rurl res nd women in prtiulr. In testimony to the ommeril viility of these lient segments, informl finne seems to e ple of profitly serving suh groups t resonly lower premium thn forml servies. Inome, employment, edution, nd eing the hed of the household hve signifintly positive impt on finnil ess, finding with signifint poliy implitions. The following hpters will fous on three suh servies, nd ttempt to point to potentil res of development nd ess expnsion, putting informtion on demnd nd supply together mirofinne, smll enterprise finne, nd remittne flows. Pge 40

56 Bringing Finne to Pkistn s Poor Chpter 3 3. Aess to Finne for the Underserved Mirofinne servies in Pkistn re offered oth formlly nd informlly. Informl mrkets re generlly hrterized y high interest rtes, widespred rtioning, segmenttion, nd sizele gp etween lending nd deposit rtes. Still, the informl setor n e ompetitive for lients with lterntive options, nd it hs good lessons to offer to its forml ounterprt. Forml mrkets re growing fst (t 40 perent), ut from negligile se. Mirofinne Bnks (MFBs) ounted for 31 perent of the mirofinne lending portfolio, nd 85 perent of its growth. Profitility nd performne in the mirofinne setor is low ut improving. A key hllenge is the sustinility of Mirofinne Institutions (MFIs), whih still rely onsiderly on nonommeril funding (ommeril liilities re rely 21 perent of totl lending portfolio). In the forml setor, 1.7 million mirofinne lients re reported, in popultion of over 160 million. The totl lending portfolio of ll MFIs in Pkistn stood t $340 million in 2007 (0.2 perent of the finnil system ssets). There is still onsiderle room for growth of mirofinne in Pkistn the estimted potentil mrket size is t lest million tive orrowers, nd some estimtes ple the numer s high s 35 million. Women re poorly explored lientele with tremendous potentil, s noted in Chpter 2. While mirofinne poliy nd servies hve foused on redit, there is onsiderle potentil for other produts suh s insurne, pyments, nd, ove ll, svings. There re severl venues tht n deliver expnded outreh nd ddress the demnd gp for mirofinne: enourging strong sustinle MFIs, improving litery nd puli wreness, employing tehnology (smrt rds, utomted teller mhines (ATMs), point of sle (POS) devies, rnhless nking, nd moile telephone networks). Two pprohes hve een used interntionlly to ddress high trnstion osts due to low popultion density, smll verge lons, nd low household svings the Grmeen nd BRAC low-teh, low-ost, high-volume models of mirofinne, nd the Philippines/Keny high-teh, low-ost, high volume pproh. Cell phones, used y over hlf of the popultion, hold muh promise to inrese ess. Informl servies ould e o-opted into the forml finnil system, lerning for their ompetitive spets. Produt diversifition towrd more svings produts is promising strtegy s well, s is fousing on women lients nd tiloring produts to their speifi needs. In 2000, mirofinne ws elevted to ore spet of the government s poverty redution progrm. In spite of Stte Bnk of Pkistn (SBP) enourgement, ommeril nks hve shown little desire to servie mirofinne lients. The SBP strtegy of offering nk liense to stronger MFIs hs proven more suessful, though the hoped-for deposit moiliztion hs not mterilized with the speed expeted, nd MFB outreh remins smll. The mirofinne setor hs emerged over the pst 20 yers (Box 3.1), nd hs foused hopes for expnded servie provision to underserved popultion. While there is long wy still to go to mth the outreh of informl servies, or even wht is eing provided y Tle 3.1 Mirofinne Penetrtion ross Asi Country Mirofinne Coverge of Penetrtion Poor Fmilies Bngldesh 17% 35% Sri Lnk 7% 29% Indi -- 25% 43 Vietnm 7% 25% Cmodi 4% 12% Indonesi 3% 11% Nepl 3% 8% Philippines 2% 6% Pkistn 1% 2% the other forml finnil servie providers, the generl view is tht the mirofinne setor hs the potentil to eventully reh most poor people with rnge of finnil servies through sustinle institutions. In Pkistn, 1.7 million mirofinne lients re reported, s ginst popultion of over 160 million. Mirofinne penetrtion in the South Asi region is higher, t 35 perent in Bngldesh, 25 perent in Indi, nd 29 perent in Sri Lnk (see Tle 3.1). The totl lending portfolio of ll MFIs stood t $340 million in 2007 (0.2 perent of the finnil system ssets). 43 Inludes indigenous self-help groups. Forml MFI estimtes for Indi mount to 3%. Pge 41

57 Bringing Finne to Pkistn s Poor Chpter 3 Soure: World Bnk (forthoming) 44 Chpter 3 looks t (1) mirofinne providers, (2) produts, (3) lients nd growth trends, (4) soures of funding, (5) MFI performne, (6) regultory issues, (7) the informl finnil setor, nd (8) the role of tehnology nd the regultory frmework for further growth of the setor Providers of Finnil Servies The mirofinne setor is mde up of severl types of key plyers. Mirofinne nks re liensed nd regulted y SBP under the MFI Ordinne Lienses n e grnted for distrit, regionl, provinil, nd ntionllevel nks. Lon sizes re urrently restrited to mximum of Rs 150,000, svings n e moilized from ny individul or orgniztion, nd in-ountry remittne servies n e provided. By erly 2008 there were eight MFBs (with Ntionl Rurl Support Progrm (NRSP) Bnk nd KASHF most reently) nd nother 2 4 prties plnning to estlish nk inluding BRAC (Tle 3.2). Speilized nongovernmentl orgniztions (NGO) MFIs provide only mirofinne servies. Like ll NGOs, they n provide lons ut nnot moilize deposits. By erly 2008, there were five suh orgniztions. Box 3.1: The Story of Mirofinne in Pkistn Development of the mirofinne setor in Pkistn ws initited y not-forprofit development orgniztions in the erly 1980s. The most influentil of these NGO experienes egn in 1982 when the Ag Khn Rurl Support Progrm ws estlished nd susequently spwned the rurl support progrm (RSP) movement. The Ag Khn Rurl Support Progrm ws lso the first NGO to trnsform its mirofinne progrm into MFB. Although two other RSPs were formed in 1989 nd 1992 nd other plyers lso egn to enter the mrket in the 1990s, the mirofinne movement remined smll nd not very visile. Network Lesing Compny ws estlished s listed lesing ompny in 1994 nd with some donor support it egn mirolesing progrm. The first ommeril nk to egin seprte mirofinne division, in 1995, ws the Bnk of Khyer, owned y provinil government. At out the sme time, the government estlished the First Women s Bnk. In ddition to supporting the needs of women entrepreneurs, the nk tried to estlish mirofinne progrm, lthough it hs not een prtiulrly suessful. More NGOs egn to offer miroredit servies during the 1990s, the most signifint eing the Kshf Foundtion, whih ws the first NGO estlished exlusively to provide mirofinne servies nd the first to e mnged y women nd hve only women lients. A mjor shift egn in the lte 1990s with the estlishment the PPAF, wholesle funding nd pity-uilding pex tht eme the min soure of funding for NGOs engged in mirofinne. In 2000, the government estlished Khushhli Bnk, MFB set up through speil ordinne. This ws followed in lte 2001 y the MFI Ordinne tht llowed for the retion of MFBs under SBP supervision. First MiroFinneBnk ws set up y the Ag Khn Development Network in erly 2002, nd over the next few yers, four more MFBs egn opertions. In 2008 the two lrgest NGO mirofinne orgniztions mnged y NRSP nd Kshf Foundtion pplied for lienses to estlish MFBs. These nks re expnding rpidly nd now ount for lmost hlf of the outreh of the mirofinne setor. RSPs re prtiulr type of not-for-profit rurl development orgniztion tht together hve lrge presene throughout the ountry. Of these, five lso provide mirofinne servies. Other NGO MFIs, possily 50 or more, provide some miroredit servies s one prt of their multidimensionl development progrms. Twelve of these orgniztions, representing lmost 100 perent of the totl outreh for this type of orgniztion, re inluded in the nlysis provided in the next setion. 44 These estimtes re sed on ountry reports y lol uthorittive orgniztions: Miroinvestment Support Fility for Afghnistn for Afghnistn, Credit nd Development Forum for Bngldesh, Aess nd MCril for Indi, Center for Mirofinne for Nepl, Pkistn Mirofinne Network (PMN) for Pkistn, nd Ntionl Development Trust Fund for Sri Lnk. Mirofinne rehes people living elow the poverty line s well s the vulnerle poor with slightly higher inomes. All of these people hve generlly een exluded from the forml finnil system. Pge 42

58 Bringing Finne to Pkistn s Poor Chpter 3 There re some ommeril finnil institutions with seprte mirofinne deprtments. Two of these, Bnk of Khyer nd Orix Lesing Compny, re inluded in the nlysis elow. But the reltive ontriution of these institutions hs een delining, nd they now represent very smll prt of the mirofinne setor. Tle 3.2: Mrket Plyers Soure: Mirowth, Issue 9 Qurter 3, Another potentilly mjor plyer in ess to finne for the underserved is the Pkistn Post Offie. It hs network of 13,419 rnhes throughout the ountry nd is signifint provider of finnil servies, inluding svings, insurne, nd remittnes, through 7,276 nk rnhes. The Pkistn Post Svings Bnk (PPSB) serves s n gent of the Ministry of Finne for rnge of finnil servies inluding svings moiliztion, life insurne, postl giro ounts, nd money trnsfers. It is present throughout the ountry nd is the only nking servie ville in some remote res. Pkistn Post hs een le to upgrde the tehnology for its opertions with help from the Islmi Development Bnk. Internet servies were lunhed in the lte 1990s nd re widely ville t post offies, ttrting the interest of smll entrepreneurs nd providing limited finnil servies. The PPSB offers severl svings shemes, nd in 2006 it hd 3.6 million svings ount holders. In ddition, Pkistn Post ts s n gent to sell government-ked svings instruments. The Postl Life Insurne offers 10 insurne options nd hd 252,810 tive poliies in Pkistn Post provides severl options for ntionl nd interntionl remittnes in ddition to hosting Western Union in some of its rnhes, nd ting s n gent to Western Union for interntionl remittnes. Government poliy hs een to enourge finnil institutions to use the extensive network of post offie rnhes to extend outreh to more people. In reent yers two MFBs, Khushhli Bnk nd First MiroFinneBnk, hve linked up with the Pkistn Post to offer servies t their rnh offies. To dte, though, the volume of this usiness hs remined smll, minly due to poor mngement, inility to trk funds trnsprently, limited pity to del with omplex servies, s well s little inititive nd vision in rodening the produt rnge in oopertion with privte finnil outfits. Box 3.2 presents suessful experienes from Brzil nd Chin nd lessons lerned from interntionl se studies on improving the effiieny nd viility of postl finnil systems. There is diverse group of other plyers in the Pkistn mirofinne mrket, inluding remittne firms, whih re the fous of Chpter While ommeril nd Islmi nks do not generlly fous on serving underserved popultions, the SBP does require them to provide some servies long these lines, inluding lowvlue svings ounts nd griulture redit. Four lesing ompnies hve provided mirolesing servies in 45 There re s mny s 53,000 primry oopertive soieties, nd until the lte 1990s there ws Federl Bnk for Coopertives s well s provinil oopertive nks tht provided finning to the oopertive soieties. But euse of some oopertive sndls in whih mny poor shreholders lost their money, the generlly poor qulity of finnil mngement within oopertives, nd lx oversight y the uthorities, the oopertive movement went into deline. While there re still mny oopertives ll over the ountry, little is known out the extent of their tive memership or their finnil performne. Pge 43

59 Bringing Finne to Pkistn s Poor Chpter 3 the pst dede. The volume of servies hs lwys remined smll ut t lest one lesing ompny still hs n tive progrm mnged under seprte deprtment. Over the pst 10 yers, there hs een growing interest in providing miroinsurne servies. Four ommeril insurne ompnies, one of whih is stte owned, provide miroinsurne servies nd the numer of lients served is growing. Most of these servies re provided in ollortion with mirofinne institutions. Box 3.2. Lessons from Suessful Postl Finnil Systems Brzil presents one of the most interesting ses in postl finnil systems. With no historil legy in postl finnil servies, Correios (Brzil s Post Offie) enefited from strong government poliy gend tht ws seeking to improve ess to finne nd tht used diret nd indiret inentives to ttrt new opertors. The government modified the regultory frmework to llow Correios to eome orrespondent of nk. Correios seleted, through trnsprent idding proess, strtegi prtner (Brdeso, the most importnt retil nk in Brzil) nd lunhed Bno Postl in Its rnge of servies spns from trditionl giro nd svings ounts to pyments nd remittnes, nd, sine 2003, mirofinne. Building effetive nd strong prtnerships etween stteowned postl opertors nd privte finnil institutions hs proven to e omplex nd umersome proess. In the se of Brzil, it took seven yers etween the first fesiility study nd the tul lunh of postl finnil servies. Tody, Bno Postl is leding plyer in retil nking in Brzil. Its suess rises from ler strtegi vision, new regultory frmework to reflet the government poliy, lned ontrt with the strtegi prtner, nd sustntil investments in the network (in prtiulr in informtion systems) to offer qulity servies. Chin Post offers nother suess story from whih Pkistn n drw vlule lessons. Chin Post hs operted the Postl Svings Bureu sine April 1984, key soure of revenue for the postl servie. Postl svings re provided t nerly 40,000 post offies pproximtely the sme size s Chin s entire nk rnh network (37,000 rnhes). Postl svings hve grown rpidly sine their opertionl lunh in 1986 nd in 2002 ounted for 189 million ounts nd more thn $65 illion in deposits, representing mrket shre of 8 perent nd the fifth lrgest deposit tker in Chin. The Postl Svings Bureu opertes n ATM network nd deit rd linked ounts ("green rds"). It lso provides 90 perent of privte remittnes in Chin (postl money orders). The volume of trnsfer opertions hs fetured fst growth mounting to 210 million trnstions in 2002, pproximtely 15 perent of the totl volume of shless pyments trnstions in Chin. A few lessons n e tken from interntionl experiene: Sustined politil support nd poliy reform drive re key suess ftor (Brzil, Romni); when this is lking, reform is diffiult to deliver (Ugnd). Historil institutionl usge of postl svings, to fund puli setor investment progrms, n onstitute mjor rodlok to signifint reform iming t ommerilizing the Postnk (Vietnm, Sri Lnk). Assessing mrket gps etween supply nd demnd n e powerful nlytil tool to onvine poliymkers of the potentil role tht post offies n ply in improving ess to finne (Brzil). Not splitting postl nd finnil servies slows down the potentil performne of the postl finnil servies (Nmii, Kzkhstn). Privte setor prtiiption, through ommeril prtnership (Brzil), pity-uilding investments, or equity prtiiption (Romni) n elerte the reform implementtion. Inluding postl opertors in pyment reforms nd enourging investment in pyment rds nd systems (Tnzni, Tunisi, South Afri) exploits est the omprtive dvntges of the postl system network. Linking mirofinne nd postl networks is ruil for onsiderle expnsion of finnil ess nd outreh. Different shemes n e envisged: the postl retil network distriutes on ehlf of lol MFI through servie-level greement, or the post offie ts s wholesler to MFIs. Soure: World Bnk (2006d) Prominent mongst the finnil institutions owned y federl nd provinil governments nd serving lower-inome lients re First Women s Bnk, Bnk of Khyer, Bnk of Punj, SME Bnk, nd the Zri Trqiti Bnk Ltd (formerly the griulture nk). The Centrl Diretorte of Ntionl Svings (CNDS, under the Ministry of Finne) orrows from seven Ntionl Svings Shemes (NSS). In 2006 the CNDS hd more thn 4 million svings ounts. It is estimted tht more thn two-thirds of ll smll-vlue svings ounts re held y these institutions, nd the perentge might e s high s 80 perent (Duflos, Ltortue, Mommrtz, Pge 44

60 Bringing Finne to Pkistn s Poor Chpter 3 Perrett, nd Stshen 2007). At the sme time, these institutions hve reeived lrge mounts of government nd donor susidies over the yers, nd most re still not prtiulrly well mnged Produts There is little vriety in the ville hoie of finnil produts for underserved popultions (Tle 3.3), though demnd for ll types of servies is not lking. Tle 3.3 Distriution of Mirolients Soure: Mirowth, Issue 9 Qurter 3, Mirofinne poliies nd tivity of nks nd other finnil institutions hve emphsized redit servies, though demnd for svings instruments is more roust. The vst mjority (over 90 perent) of lons in Pkistn follow the group lending model (Figure 3.1). Individul lons, usully lrger nd often tied to n identified miro enterprise, re mostly llotted to men, t lon mturities vrying from 3 to 18 months. Inresingly, MFIs re diversifying their redit produt offerings, with housing improvement lons nd miroleses on hirepurhse sis. Burki nd Shh (2007) present detiled desription of ommon redit produts in Pkistn. 46 Muh more n e done to expnd mirosvings, judging from existing demnd. 47 Produt ustomiztion to diverse lient needs is muh needed, espeilly for femle lients where onsiderle mrket potentil for svings exists. Offering svings produts n lso lower the ost of funds for MFBs. Other miro produts, suh s insurne nd pyments, re few nd fr etween, though they hve gret growth potentil. 48 Over the pst 5 yers, for exmple, nks in Brzil hve used tehnology to expnd outreh of finnil servies to millions of people tht were not previously served y the forml finnil setor. More thn 95% of the volume of these trnstions is relted to pyments. Figure 3.1: Numer of Borrowers nd Lon Portfolio y Lending Methodology 46 The MFI Ordinne defines mirofinne s very smll lons, presently of less thn Rs 150, The lrgest mirosvings providers re NSS nd PPSB, with $12 nd $0.85 illion in svings deposits, respetively. 48 In ddition to group life insurne, whih enjoys the highest populrity mount miro-insurne produts, MFIs hve strted experimenting with miro helth insurne. RSPs is the lrgest helth insurne sheme, overing out 200,000 of their group memers for hospitliztion enefits. Pge 45

61 Bringing Finne to Pkistn s Poor Chpter 3 Soure: Mirowth, Issue 9 Qurter 3, Mirofinne Clients nd Credit Growth Investments in the mirofinne industry hve resulted in setor outreh (tive lon lients) growing y 22 times, from round 60,000 tive orrowers in 1999 to 1.7 million tive orrowers y mid This trnsltes into n verge nnul growth rte of 40 perent, the seond highest in the region fter Afghnistn. The mjor shre of mirofinne lending goes to trde, followed y griulture nd livestok. Rurl res reeive the mjority of mirolending (56 perent), in ontrst to 44 perent going to urn res (SBP Mirowth). Women represented 47% of tive orrowers, 33% of svers, nd 57% of poliyholders, in the third qurter of In other South Asi ountries, even Afghnistn, etween two-thirds nd 95 perent of ll tive orrowers re women. Of the MFIs with signifint outreh, only Kshf Foundtion is designed to trget women speifilly. There re other MFIs fousing exlusively on women lientele, ut given their urrent pity, they re less likely to influene the trjetory of finnil servies ess to women. BRAC, whih hs reently strted opertions in Pkistn, is lso exlusively foused round women Soures of Finning The totl lending portfolio of ll MFIs ould reh over $1 illion y 2010, provided suh growth n e funded, nd ontinued demnd exists. The most promising funding soure whih ould keep up with the rpid growth of the setor is miro-svings, estimted t $1.2 illion y The distriution of noninstitutionl deposits held in forml finnil institutions shows only 0.04 perent of svings mounts nd 0.31 perent of the numer of svings deposits held t MFBs, s ompred with ommeril nks (74.4 perent nd 77.4 perent, respetively), Pkistn Post (1.5 perent nd 11.3 perent), nd NSS (24 perent nd 11 perent). 50 Another promising strtegy is for MFIs to deploy lrger perentge of their ssets into their lon portfolios. At the end of 2007 this stood t 56 perent, muh elow regionl stndrds (for exmple, more thn 80 perent in Bngldesh, 2006). This lso redues the overll sset yield of the setor sine funds tht re deployed in things other thn lon portfolios (investments, sh) yield lower returns. 51 The mirofinne setor reeived $463 million in donor funding, s well s indiret IFI investments. 52 Generlly, non-mrket-sed funding id hs not resulted in roust setor development. 49 Burki nd Mohmmed (2008) desries demnd for svings servies mong the urn poor. 50 June 2007 dt (PMN 2008). 51 Averge yield on investments is 4 10 perent depending upon the tenor nd risk of investments, wheres yield on portfolio verged 21.5 perent for the setor s whole in This inludes the Asin Development Bnk projet under whih Khushhli Bnk ws estlished ($150 million), grnts from ilterl donors ($31 million), nd two PPAF projets funded y the World Bnk ($215 million). The Interntionl Fund for Eonomi Development egn $32 million mirofinne projet with PPAF in 2007 nd nother projet with PPAF ($35 million) egn in erly The lrgest of the indiret investments is the Asin Development Bnk (ADB) projet ($322 million), s well s the U.K. Deprtment for Interntionl Development (DFID) support ($100 million) towrd strengthening SBP s Finnil Inlusion Progrm to inrese finnil ess to the poor nd smll enterprises. Pge 46

62 Bringing Finne to Pkistn s Poor Chpter 3 Ptterns in soures of funds vry y ountry sed lrgely on wht regultions llow, the extent to whih government nd donor soures of funding exist, nd MFI profitility. For exmple, mong NGO MFIs in Bngldesh, 29 perent of funding me from svings nd 37 perent from equity. 53 By omprison, Indin MFIs hve lmost no svings for regultory resons, nd re highly leverged with 7 perent equity nd 62 perent ommeril orrowing. The soures of finning for the gross lon portfolio of ll MFIs in Pkistn show some ler trends (Figure 3.2). Svings s soure of funds inresed, nd equity got inresingly leverged. The perentge shre of Figure 3.2: Soures of Finning for MFIs donor-funded PPAF lons 54 fell s MFBs 2004 GLP $120 million 3.5. MFI Performne 2007 GLP $340 million 2010 GLP $1,000 million The mirofinne setor in Pkistn is indeed rehing reltively poorer people. The verge mirofinne lon size in Pkistn is out 15 perent of GDP per pit inome (Figure 3.3), whih ompred very well with the muh lrger lon sizes prevlent in Estern nd Centrl Europe (95 perent of GDP per pit), Ltin Ameri (out 55 perent), nd the Middle Est nd North Afri (out 20 perent). Asi does even etter thn Pkistn, y lowering the verge mirolon size to td over 10 perent of GDP per pit. MFI portfolio qulity is dequte, nd eme ineligile for onessionry funding. The min hllenge in the future will e to inrese the mount in other liilities so s to mintin this t out hlf of the totl. This will require signifintly inresed use of ommeril funding y MFIs, mostly from domesti pitl mrkets. Current efforts y the industry nd the SBP to fill the projeted funding gp re foused on inresing ess y MFIs to ommeril soures of funding, prtly sed on the ft tht ommeril nks re quite liquid nd there is mrket ppetite for ond issues. 55 Figure 3.3 Averge Lon Size figures. This nlysis exludes Grmeen Bnk, the only MFB in Bngldesh. All the other MFIs re NGOs. 54 PPAF onsists of mix of lons nd grnts to MFIs. 55 The disussion, like the entire report, does not tke into ount the onoming finnil risis. Given the risis, the proess of inresing MFI ommeril funding eomes more diffiult, nd more urgent. Likely onsequenes of the mroeonomi risis inlude: (1) rising delinqueny nd slowing of growth s poor lients (or potentil lients) suffer severe hrdship, (2) svings flight from MFIs into sfer options, nd (3) less funding eing ville for MFIs. Pge 47

63 Bringing Finne to Pkistn s Poor Chpter 3 profitility is low ut improving. 56 Soure: SBP The finnil helth of the MFIs setor in Pkistn hs een onern (Tle 3.4). While sset qulity hs een improving (s evidened y delining nonperforming lons), return on ssets (ROA) hs seen deline, s hve operting suffiieny nd finnil sustinility. SBP hs enourged the etter institutions to otin nking liense, to improve funding Tle 3.4 Key Performne Inditors of MFBs (%) efforts nd to sujet them to * tighter prudentil regultion. Nonperforming lons to dvnes % 3.3% Overll, however, MFI Operting expenditure to gross inome % 135.5% finnil helth nd ROA % -5.3% mngement prties require Operting suffiieny** % 65.1% urgent ttention to improve Finnil sustinility*** % 50.6% sustinility, outreh, nd effiieny of the setor. Soure: SBP, Septemer *Dt for Mrh ** Operting Suffiieny= (Totl Inome less Inome from Dontions, et.) / (Totl Operting Expense plus Atul Cost of Funds). *** Finnil Self-Sustinility = (Totl Inome less Inome from Dontions, et.) / (Totl Operting Expenses plus Atul Cost of Funds plus Impliit Cost of Susidized Funds plus Impliit Cost of Equity; omputed t SBP disount rte) Mirofinne nd Moile Telephone Tehnology in Pkistn Reently, there hs een rpid inrese in the use of new tehnologies to inrese ess to finnil servies in mny ountries. This hs inluded the widespred use of smrt rds nd of moile telephone networks. In Pkistn, moile phone penetrtion hs inresed t rpid pe reently (see Tle 3.5). Hlf or more of ll Pkistnis hve ess to ell phone, inluding mong women nd rurl res (two-thirds in urn res). While more thn 86 perent of men hve their own ell phone, 40 perent of women do. These figures suggest muh higher ess to ell phones (ville in the household or within the extended fmily) thn the figure for regulr usge. Both ess to nd usge of ell phones inrese with inome. Bluhistn is the lest served (still, third of the popultion hve ess to moile phones). Per the survey, ell phone usge is 35.4 perent in Sindh, out 44 perent in North West Frontier Provine (NWFP) nd Azd Jmmu nd Kshmir (AJK), nd highest in Punj, t 51.5 perent (the figures in Tle 3.5, s estimted y moile phone providers, re somewht more onservtive). However ess to ell phone is muh higher: over hlf of the popultion in ll provines exept Bluhistn hve ess. The pervsive use of moile telephony retes redy mrket for finnil servie delivery over moile phones. This effet is somewht limited y the finding tht fe-to-fe intertions re highly vlued, eqully y men nd women (76 perent). Tle 3.5: Moile Telephone Susriers June 2005 June 2006 June 2007 Mrh 2008 Totl Susriers 12,771,203 34,506,557 63,159,857 82,514,536 Mrket overge* Punj Sindh NWFP Blohistn 3.4% 4.5% 1.7% 1.5% 8.5% 11.7% 3.6% 3.6% 23.9% 27.3% 8.1% 8.1% 35.1% 39.8% 23.0% 16.6% Soure: Pkistn Teleommunition Authority (2007) * These figures re provided y the moile telephone ompnies. Mrket overge is the perentge of the estimted totl numer of susriers who hve n tive SIM rd. 56 Portfolio t risk t 30 dys is elow 5%. The effets of the risis re yet to e ssessed t the time of printing of this report. Pge 48

64 Bringing Finne to Pkistn s Poor Chpter 3 Use of ell phones is still mostly onfined to voie servies (for 78.5 perent of ell phone users). Bsi dt servies, suh s Short Messge Servie (SMS), re used y 40.2 perent ell phone users, with higher usge in urn res (45 perent) thn in rurl (36.7 perent). SMS usge is lso more frequent mong men nd in AJK nd inreses s inome rises. Internet usge remins low nd is onentrted in urn res nd mong men. Internet nd e-mil re more essed /used t home. Close to two-thirds of Pkistnis (60.8 perent) hve their own prepid moile phones. Moilink is the most-used network, ross rurl nd urn res, exept in AJK nd NWFP, where Telenor is most often ited. There is lrge gender divide on ell phone pyments. While mjority of men py for themselves, the mjority (84.5 perent) of women s ell phone expenses re pid y their spouse or fmily. This limits somewht the usefulness of ell phone redit reords s soure of redit informtion. Nevertheless, World Bnk (2008) onludes, on the sis of ross-ountry nlysis, tht moile tehnology does expnd finnil ess. The potentil for using moile phone tehnologies to reh mny more people is evident from Figure 3.4. By July 2008, there were 89.3 million moile telephone onnetions nd the networks overge potentilly extends over 90 perent of the popultion. Teledensity in the ountry stnds t 56% s of Jnury Growth hs een very rpid over the pst few yers nd will ontinue to e in the future. It is lso ler tht even poor people re using moile telephones, sine it is esy nd inexpensive to get prepid ount. By omprison, during the pst 60 yers, the nks hve only een le to umulte 25 million deposit ounts, nd rriers to entry (doumenttion, minimum lne requirements, et.) hve kept most of the poor out of the nking system. Figure 3.4: Cellulr Penetrtion nd Projeted Moile Telephone Trends y Provine Soure: Qurterly Report of PTA (Jn - Mr 2008); Pkistn Teleommunition Authority, Annul Report 2007 With five moile opertors ked y strong interntionl shreholders, Pkistn s tive nd ompetitive moile phone mrket, omined with the presene of regultory guidelines for rnhless nking, provides n enling environment for prtnerships etween MFIs nd teleommunitions opertors. Suh prtnerships hve lredy mterilized in Pkistn, with the quisition in Novemer 2008 of Tmeer Mirofinne Bnk, pioneering new ommeril wys to inrese outreh to the underserved Informl Finne The summry in this setion is lrgely drwn from Qdir (2005). Pge 49

65 Bringing Finne to Pkistn s Poor Chpter 3 Informl finne mrkets hve long history pre-dting forml mrkets nd strong presene in most of rurl nd urn Pkistn. These mrkets nnot e stritly lssified. They re generlly stnd-lone, operting without the links tht hrterize well-integrted finnil mrkets. The multipliity of informl finne mrkets is refleted in the oserved diversity of trnstions in these mrkets, suh s ommittees or rotting sving nd redit ssoition (ROSCAs), moneylending, interlinked finning, nd suppliers redit, mong others. However, informl finne mrkets fe onstrints in getting ess to institutionl redit. This lk of wellfuntioning finnil mrkets hs n dditionl dverse effet on poverty, s the poor who hve redit requirements ut few ssets tht n serve s ollterl nnot resort to forml finne mrkets, ut re left insted t the mery of urrent onditions in informl mrkets. Qdir (2005) quotes 1996 estimte of informl orrowing in Pkistn t 78 perent of totl orrowing in the ountry. A omprison of the shre of informl rurl redit in Asin ountries shows tht this shre is high in ll ountries, espeilly Pkistn. The shre long with the reporting yer is: Bngldesh, 63 perent (1974); Indi, 70 perent (1972); Indonesi, 52 perent (1985); Mlysi, 62 perent (1986); Pkistn, 73 perent (1985); Philippines, 71 perent (1978); South Kore, 50 perent (1981); Thilnd, 52 perent (1985) (World Bnk, 1996). Informl Borrowing. It is very ommon for poor people in prtiulr to orrow from reltives, friends, nd other people with whom they hve lose reltionships, espeilly for onsumption smoothing. Reognized moneylenders operte ll over the ountry nd re often soure of lrger mounts of finne or re pprohed s lst resort. A populr soure of finnil servies for ll inome levels, espeilly women, re ROSCAs, ommonly known s ommittees. These tke different forms, ut ll involve prtiiption in group where everyone sves money y ontriuting to pool nd eh one in turn is le to ess nd use the pool mount. Urn finnil mrkets re different from rurl ones in ertin importnt respets. Mny urn mrkets ter to trders, espeilly wholeslers, nd re quite well developed in terms of the mounts of funds intermedited, the speed nd effiieny of the intermedition, nd the sophistition of prtiipnts nd of the mrket s whole. Suppliers' redit is ommon feture: in old mrkets with estlished plyers, s muh s 90 perent of trnstions re rried out on suppliers' redit resting on good fith. A hit (prhi, or informl reeipt) is the norm for mking usiness trnstions nd is not dishonored; it represents onvenient nd flexile method tht llows usiness to e onduted t rms-length nd does not require doumenttion or entil tx liilities. Credit in the rurl res is mostly supplied y rtis (ommission gents) nd other middlemen t high interest rtes through interlinked trnstions. The ute shortge of pitl t ffordle rtes severely onstrins the growth of the rurl eonomy nd prevents effiient resoure moiliztion nd risk mngement. Informl svings. Poor people sve in oth finnil nd nonfinnil forms ut often lk the ility to mke use of smll svings flows when lrger lump sums re needed. Even the ommittee system does not lwys meet this need. Given the derth of institutionl hnnels to tp into rurl svings, people in rurl res sve through trditionl hnnels. Sving in livestok, whih n e ought nd sold when needed, is good livelihood diversifition strtegy for low-inome households. Livestok is lso kept through shre lesing, whih is sving rrngement etween lndlords nd professionl strt (kmmis). Other trditionl sving rrngements in Punj re lled vrtn hnji nd wnghr, mening sking others for help on voluntry nd reiprol sis. Hording gold nd silver is lso ommon in rurl nd semi-urn res. Trditionlly, suh jewelry is supposed to e kept for lifetime nd trnsferred to the hildren of the fmily, to e utilized only s lst resort through sle or mortgge. Informl trnsfers involve trnsfers or exhnges etween households of sh, food, lothing, informl lons, nd other informl ssistne. While informl trnsfers help the poor in risk mngement, they re not dequte sustitutes for puli tion in soil protetion. A populr informl system of trnsferring money round the world is the hwl system, mrked y low ommissions, fst trnstions, little doumenttion, nd round-the- Pge 50

66 Bringing Finne to Pkistn s Poor Chpter 3 lok opertions. The system works through individul rokers or opertors olleting funds t one end of the pyment hin nd others distriuting the funds t the other end. This is n espeilly populr mens for migrnt workers to send money home. Tle 3.6: Interest Rtes Chrged in the Informl Mrket in Pkistn, 2005 Effetive Annul Interest Rte Auto-rikshw informl redit mrket (Lytton Rod, Lhore) 26.4% Fisld yrn mrket 17.31% to 46.15%, verge 28.39% Moneylenders in rurl res, for household nd onsumption 120 to 150% Moneylenders in rurl res, for usinessmen nd griulturists 12 to 50% Moneylenders verge, ll lons 55.2% Commission gents, for frm inputs to 15.38% Frm input delers 2.2% to 74.2 %, verge 25.81% Self-reported y frmers, for lending in kind 29.81% Self-reported y frmers, for lending of sh 85% Soure: Qdir (2005) Informl mrkets re generlly hrterized y high interest rtes, widespred rtioning, segmenttion, nd sizele gp etween lending nd deposit rtes. There is extreme vriility in the interest rte hrged y lenders for similr lon trnstions. Informl finne mrkets re generlly mrked y low levels of defult due to soil sntion, group sinerity, pst history, nd repet trnstions. Informl redit mrkets re mrked y widespred rtioning; tht is, there re upper limits on how muh orrower reeives from lender. Segmenttion is nother feture of informl redit mrkets. Typilly, moneylender serves fixed lientele, whose memers he lends to on repeted sis, nd is extremely relutnt to lend to outsiders. Qdir (2005) estimtes ost of informl orrowing of 23 perent on verge, s opposed to n verge of 19 perent in the forml mrket. The totl trnstion ost of informl lenders on verge onstitutes only 5 perent of the totl volume of lending. Interestingly, sh redit is onsiderly more expensive t 85 perent per yer (Tle 3.6). Soil sntion nd mrket limittions re the most ommon instruments for enforement of ontrts s well s for reovery of lons. Resorting to the legl system of the ountry is firly unommon. Moneylenders usully tke vrious preutionry mesures efore tking on new lient. These inlude the prtie of deling with the potentil lient in other mrkets, extensive srutiny of new lients, nd smll testing lons. In n environment of wek ontrtul enforement, those engged in usiness, espeilly rms-length trnstions, hve to e very disreet nd often rely on individul goodwill nd soil pressure in the sene of seurity (ollterl). In ses of defult, mrket plyers normlly medite nd deide out reeivles nd pyles nd, in extreme irumstnes, dispose of ssets. It is often sid tht informl finne mrkets exist euse forml mrkets re poorly developed nd mny people, espeilly the poor, do not hve ess to forml soures. But informl finnil mrkets still thrive even fter ll the efforts of the pst few dedes to expnd forml institutionl mrkets. This suggests tht informl soures might hve some dvntges over forml soures. In generl, informl soures of redit re more expensive thn forml soures nd there is muh evidene to suggest tht they exploit poor people, ut they hve the dvntges of eing ville ll the time, not requiring doumenttion, operting outside the purview of forml uthorities, nd not requiring ollterl. Moreover, mny upper-inome groups use informl mrket s sustitute for or omplement to forml ones, whih indites tht informl mrkets n e ompetitive. One response to reognition of the informtionl dvntges of informl finnil mrkets n e to try nd enourge them rther thn reple them y expnding forml finne to eonomi gents who re likely to use these funds in informl mrkets. Another response is to tully design nd help expnd MFIs tht will tke dvntge of lol-level informtion. The modern mirofinne movement hs mde use of some of the hrteristis to uild up its usiness, not lest sustituting soil ontrts for other forms of ollterl tht re often not ville to the poor. But the persistent use of informl finne in the fe of onerted efforts to Pge 51

67 Bringing Finne to Pkistn s Poor Chpter 3 improve ess to forml soures suggests tht muh more ould e done to lern from informl finnil systems s well s to develop linkges etween these systems, or even find wys to inorporte informl institutions suh s the hwl system into the forml setor Regultion In 2000, mirofinne ws elevted to ore spet of the government s poverty redution progrm. The introdution of the Mirofinne Institutions Ordinne 2001, under whih MFBs re liensed, ws driven y poverty-redution gend. The ordinne is not onerned with NGO mirofinne progrms nd reltes only to MFIs tht wish to moilize deposits. The MFBs re liensed nd regulted y the SBP under speilly designed prudentil regultions. There re four tegories of liense, eh with different pid-up pitl requirement: ntionl, provine, region, nd distrit. At first, the regultion did not led to rpid expnsion of servies, ut the regultor dopted n open pproh to the setor, inluding the retion of mirofinne onsulttive group hired y the SBP nd omposed of representtives from the setor, nd hnges in the ordinne nd regultions grdully mde this more ttrtive option. By 2008, there were six MFBs inluding the lrgest suh nk, Khushhli Bnk, tht ws initilly estlished in 2000 under speil ordinne ut reliensed under the MFI Ordinne in erly To dte, First MiroFinneBnk is the only se in whih n NGO progrm ws trnsformed into MFB, ut the two lrgest NGO progrms oth intend to pply for MFB lienses in It is expeted tht y the end of 2008 the shre of totl tive lients served y MFBs will e more thn the NGO totl for the first time. Considering tht demnd still fr outstrips supply nd tht it will tke huge mounts of funding to stisfy tht demnd over the next dede or more, MFBs tht moilize deposits nd n more esily ess ommeril funding hve the potentil to tke lrge shre of the mrket if they hve good usiness models. An importnt hnge in mirofinne investor mindset will id the ommeriliztion proess for MFBs. When the government enourged the ommeril nks to invest Rs 1.7 illion ($285 million) to pitlize Khushhli Bnk, some ommeril nk exeutives invested out of sense of orporte soil responsiility. This is hnging s more ommerilly minded investors enter the mrket. The Pkistn regultion is now onsidered to e the most onduive regultion in the region with respet to promoting finnil inlusion for poor people. Amendments to the MFI Ordinne nd its regultions mde in hve improved the sitution even further, nd the SBP is open to mking further djustments in future. The ordinne hs the potentil to signifintly improve the delivery of forml finnil servies to exluded popultions. In ft, the SBP hs tken on ledership role to promote nd filitte the development of the mirofinne setor. In ddition to mking the MFI Ordinne nd its regultions inresingly friendly, the SBP hs issued series of guidelines for ommeril nd Islmi nks to engge in the mirofinne setor Growing the Mirofinne Setor The min fous of government poliy iming to inrese ess to finne to lrger groups of poor nd underserved people is the development of the mirofinne setor. Reent poliy interventions inlude SBP regultions on rnhless nking, newly reted lss of MFBs, si nking, litery nd wreness efforts, nd vrious progrms to support mirofinne. These progrms hve spurred fst growth in the mirofinne setor, ut hve hd mixed results on improving the setor s sustinility nd rpidly expnding its outreh. More enourgingly, eletroni trnsfers lredy represent 10 perent of pyments for government workers nd 3 perent of pyments for pensions (SBP figures). Bsi nking is urrently used y 2.5 perent 58 For more detils out these mendments nd their implitions for MFBs, see Ahmed nd Shh (2007). Pge 52

68 Bringing Finne to Pkistn s Poor Chpter 3 on verge, espeilly mong men nd in rurl res (SBP figures). Muh more remins to e done, nd some promising strtegies re outlined elow. Privte Setor Ations Improving MFI sustinility nd ility to muster ommeril funding / svings deposits, nd their further integrtion into the finnil system: The min requirement for mirofinne outreh growth is the presene of strong, profitle MFIs. More ould e done to improve effiieny nd inrese profitility. One of the min onditions for sustinility is inresed reline on ommeril funding nd the ility to ttrt deposits, in order to fund n expnsion of outreh to more lients. Effiieny improvements will lso e required, in order to deliver finnil servies in heper mnner nd mnge risks etter. The ultimte gol for MFIs should e to firmly integrte into the finnil system. While the mirofinne setor in Pkistn is urrently too underdeveloped to e dequtely interlinked with the rest of the finnil setor, integrtion would gurntee tht MFIs re sujet to mrket signls nd re sustinle euse they would respond to the need to rise resoures from the mrket nd thus improve their finnil performne. The setor n refous from miroredit to mirosvings, given the lrge untpped demnd for suh produts. Besides eing the servie tht poor people wnt more thn ny other, svings moiliztion y MFBs will lso provide longer-term stle soure of funds to grow redit outreh. MFBs need to py more ttention to developing their svings servies, espeilly through strengthening their systems nd developing pproprite produts. The ommeril nks nd other finnil institutions tht provide svings servies n lso do more to refous further on poor people. World Bnk (2008) lso notes tht mny rih ountries nd some developing ountries re lso experimenting with new wys to promote svings, suh s mthing shemes nd tx dvntged shemes. Svings methods tht hve worked for mirofinne doorstep olletion shemes nd periodi ontriution or ommitment progrms re lso eing offered y some nks in developing ountries. Enlrging the produt rnge nd lient segmenttion: Client segmenttion llows finnil institutions to etter tilor produts to tstes nd lient needs, s well s redue osts nd mnge risks more effiiently. A prtiulrly relevnt exmple is gender segregtion. Understnding women s needs more preisely, nd refleting those in the finnil produts nd the provider poliies nd proedures, would ensure n inrese in women s ess to finne in spite of ulturl norms, gender segregtion, nd low litery nd inomes, s detiled in Box 2.1 nd the reommendtions setion in Chpter 2. Produts (svings, insurne, nd redit) for old ge, hildren s edution, pregnny nd medil expenses, nd livestok re few exmples of those tht tke ount of women s needs for life-yle events. Sving produts, whih re expeted to e espeilly populr, n e uilt upon trditionl sving rrngements nd ROSCAs tht women use. Lower lon size nd deposit size would permit etter mthing to women s needs, given their lower inomes. Repyments should e frequent so tht instllments re smller nd orrespond to women s inome yles. More simplified proedures, s well s doumenttion tht does not depend on the men in the household, will ddress women s relutne to vil themselves of finnil produts. Glol experiene suggests offering women redit tht is un-tied to speifi use, insted llowing the orrower to suggest the tivity. Aess would lso improve with the use of lterntive forms of ollterl, suh s soil ollterl, ompulsory svings, personl gurntees, rops or mhinery to e purhsed, or household ssets. Aess to finnil servies should e rodened eyond the hed of household to inlude more women from the sme household, espeilly through sving progrms. Home-sed usinesses should e given onsidertion. Litery should not e requirement to ess finnil servie. Simple poliies nd proedures tht speed the trnstion, lower trnstion ost for women, nd do not prelude uneduted women tend to mximize outreh to women lients. Deentrlized opertions, operting units loted ner women lients, use of moile units, nd trnstions t lients doorsteps tend to mke nking onvenient for women. Femle stff will improve Pge 53

69 Bringing Finne to Pkistn s Poor Chpter 3 pprohility for lients, nd llevite ulturl onerns. Some suessful exmples of home-grown solutions to women s nking is Bnk A;-Flh women s redit rd, nd Meezn nk s Ldies Bnking. Puli Setor Ations Expnding finnil wreness: Further gins in finnil litery re ritil, though even more ritil is the popultion s wreness nd etter understnding of finnil servies nd produts, whih will promote trust in the setor. A ntionl wreness mpign is needed to support finnil inlusion, espeilly for women, s well s to enourge people to open svings ounts. A lrge numer of people still do not hve ny ount t ll. While finnil litery rguments re quite ompelling, empiril evidene from Indonesi (Cole nd Zi 2008) suggests tht finnil litery hs very little if ny impt on the use/uptke of finnil servies mong households. To the ontrry, it is finnil inentives (whih remove/redue ffordility rriers of opening nk ount) tht hve n impt. Strengthening institutions: Aess to finne growth will e elerted y n integrted finnil system, nd strong regultory frmework. SBP is working towrds reting n enling frmework for ess, nd is set upon n mitious Finnil Inlusion Progrm to mke further inrods into the mtter. Other regultors nd institutions would need to rise to the osion s well, inluding the SECP. Among the importnt fetures of omplete finnil system is well-funtioning ntionl-level redit ureu for redit referening. All ommeril nks, development finnil institutions, lesing ompnies nd MFBs re urrently reporting to CIB on ll orrowers irrespetive of the size of lon. A privte setor inititive long these lines is eing prepred y the PMN memers with Dthek, onsumer finne redit ureu, to rete mirofinne redit ureu in the ompetitive Lhore mrket, whih is expeted to er initil fruit in By estlishing redit history nd thus potentil ollterl sustitute, redit ureu n e instrumentl in ess to finne for groups who my not hve sh or sset ollterl required to ess lon ut hve stellr redit history to present to the nk. Upgrding the existing redit ureu mnged y SBP to more thn just lk list nd inresing its overge to the whole finne servie setor inluding NGO-MFIs n ple mny more potentil orrowers within reh of some ess to finne. Puli - Privte Prtnerships Tehnology n help ddress effiieny hllenges within the mirofinne vlue hin, while opening the wy for innovtive pplitions designed to inrese outreh. Enhning outreh vi tehnology solutions n involve nks dopting omintion of devies s well s rnhless nking through ell phone nd moile devies. These re different diretions, nd eh hs shortomings nd diffiulties. In Pkistn, the nking ess infrstruture is prtiulrly wek, s noted in Chpter 1 there re few nk rnhes nd ATMs reltive to the popultion size nd geogrphi expnse of the ountry. In ddition, s demonstrted y the Brzil orrespondent nking experiene, ostles to expnding outreh n lso rise from ontrtul s well s regultory nd prudentil ftors regrding geny rrngements (Kumr et l. 2006). In ontrst, while moile nking presents regultory hllenges (s in the se of the Philippines G-Csh nd Keny M-Pes models), it n e promising hnnel to help shift some of the finnil flows from informl to forml hnnels, in prtiulr if omined with other orrespondent nking hnnels. Demnd-side results do show us very high moile penetrtion nd our eonomi nlysis highlights signifint positive linkges etween finnil inlusion nd regulr moile use, s well s informl inlusion nd ess to ell phone. Expnding outreh vi moile telephony, smrt rds, nd POS devies: Given the wide, reltively equitle, nd rpidly growing ess in Pkistn to moile phones, tehnology hs mjor potentil to eome onduit for ess to finne in the ountry. In ddition to ess expnsion, moile phones, smrt rds, POS devies, nd other tehnology improvements n lower trnstion osts, s well s help enhne redit informtion on muh wider popultion segment. The simpliity nd low ost of these servies hve enled poor people to use them esily nd suessfully in spite of their novelty nd reent penetrtion. This stnds in Pge 54

70 Bringing Finne to Pkistn s Poor Chpter 3 shrp ontrst to the omplexity nd lk of user-friendliness of trditionl nk produts, nd their reltive filure to penetrte wider popultion rnge. Box 3.3 illustrtes some of the glol good prties leverging tehnology to improve k-end proessing nd expnd delivery hnnels, inresingly through prtnerships etween nks nd non-nk institutions. These tehnologies hve een very suessful in promoting pyments servies worldwide. In Pkistn, given popultion preferenes nd needs, it is importnt to find wys to extend ess to svings servies s well, vi tehnology gins. Interntionl experiene points to (1) regultory methods of promoting svings, suh s mthing shemes nd tx dvntged shemes, s well s (2) svings methods whih hve worked for mirofinne doorstep olletion shemes nd periodi ontriution or ommitment progrms. But if the full potentil of this new pproh is to e relized, it will need to go well eyond the mirofinne setor, where there is lredy strong interest in using tehnology to reh more people nd lower osts, nd inlude the nks. So fr, ommeril nks hve not shown muh interest, though this might hnge s time goes on. Experimenttion with using moile telephone networks is just eginning. On the other hnd, rds hve een in use for mny yers, though not generlly y poor nd underserved popultions. As of Deemer 2007 the totl numer of tive rds in the Pkistn nking system stood t 6.7 million, s ompred to totl of 16 million personl nk ounts. Of these, 1.7 million were redit rds (25.4 perent) nd 4.8 million were deit rds Pge 55

71 Bringing Finne to Pkistn s Poor Chpter 3 Box 3.3 Cse study: Tehnology Innovtions hve Improved Bk-End Proessing nd Expnded Delivery Chnnels in the Mirofinne Vlue Chin Shred Mirofinne Servies Hus Mke Tehnology-led Effiieny Gins Aessile While the enefits of utomting ore nking systems to ut operting osts, stremline lending proesses, sle up nd integrte with the rest of the finnil setor, re widely reognized, these solutions often remin out of reh for individul mirofinne institutions. Chllenges inlude high upfront osts, onnetivity requirements, unvilility of tehnil support nd high mintenne osts. Outsouring mngement informtion systems to pplition servie providers (ASPs) who mnge entrlized mirofinne proessing hus hs emerged s solution to ddress some of these hllenges. The utomtion of MFIs opens the door to new produts nd servies. FINO (Finnil Informtion Network & Opertions Ltd), n ASP in Indi inuted within ICICI Bnk, now provides entrlized tehnology pltform vi iometri smrt rd to lients of MFIs, nks, insurne providers nd the Government. FINO lims to hve enrolled 5 million ustomers who perform lne enquiry, deposit nd withdrwl of sh nd sttement nd reeipt printing. Similrly, IBM Glol Servies is in the erly stges of plnning for Ltin Ameri wide strtegy for entrlized ICT pltform for mirofinne, loted in Mexio nd serving ll MFIs in Spnish-speking Ltin Ameri through we-sed solution. The pltform envisioned inludes model of servie delivery, hosting, nd tilored funtionlity. Innovtive Tehnology Applitions, in the Hnds of Mirofinne Agents or Clients, Inrese Outreh The Andhr Prdesh Government in Indi, fter suessful pilot rehing hlf million villgers, is rolling out the delivery of Government pensions nd unemployment enefits under the rurl employment gurntee progrmme. A network of nk usiness orrespondents opertes with Ner Field Communition-enled moile phone, synhronizing with nk s server over wireless onnetion nd fingerprint snner to enroll the villgers. All deposit/withdrwl trnstions get reported to the prtner nk on dily or hourly sis while the nk ensures tht the usiness orrespondent hs enough funds to tke re of dily withdrwls. During the pilot, the Government gve nks 2 per ent of the mounts eing trnsferred s ommission, to e shred with their usiness orrespondents. In Pkistn, Tmeer Bnk hs prtnered with lol retil outlets inluding phrmies, teleom outlets nd post offies to instll point of sle devies loser to the ustomers. Bnk ustomers (individul nd smll usinesses) n visit these registered outlets to withdrw nd mke deposits, repy lons, py utility ills nd remit money. Costs of using orrespondent gents s lterntives to setting up nk rnhes re reportedly 30 times lower set-up osts nd 100 times lower running osts. Similrly in Keny. Equity Bnk is using the Nkumtt hin of retil stores s rnhless nking gents while WIZZIT, South Afrin ell-phone nking fility with 16 million registered lients hs prtnered with Dunn, hin of 400 lothing stores ting s WIZZIT ount opening lotions. The Rurl nkers Assoition of the Philippines, with support from USAID, hs prtnered with Gloe Teleom s G-Csh to llow its retil lients in 375 rnhes to send sles reeipts nd nk deposits vi their moile phones, without eing sujet to minimum lne nd fees hrged y the ommeril nks. Mirofinne institutions under this progrm eome sh in nd sh out outlets tht re redited to onvert tul money into eletroni money nd vie vers. Mexio s Bnmex Bnk offers two produts, pre-pid msterrd redit rd nd svings rd, to n verge of 25 million lients in 240 Sorin stores, the seond lrgest Mexin retiler. The two ompnies pln to o-invest to set up 150 retiler-run nk units in Soure: CGAP, World Bnk (2008) Pge 56

72 Bringing Finne to Pkistn s Poor Chpter 3 (71.6 per ent). ATM-only rds were million or 2.8 per ent. Ninety-nine perent of the time, these rds were used for withdrwls while only one perent of the trnstions were deposits (envelope sed). Eh ATM hd n verge of 70 trnstions per dy, of n verge size of Rs 6,127. Until Deemer 2007, there were 2,618 ATM mhines (s ompred with over 8,000 nk rnhes), nd 52,474 POS terminls ross the ountry. Almost two-thirds (61 perent) of the nk rnh network onsisted of Rel Time Online Brnhes. One qurter of the totl trnstions in the system were eletroni-sed. Although the eletroni nking system is growing, it is still in its infny nd its expnsion in oming yers is si premise for the inlusion of the mjority of the popultion (Lindh de Montoy nd Hq 2008). Box 3.4 Bsi Bnking in Indi, Mexio, nd South Afri In Deemer 2005, Indi introdued new type of nk ount designed for the poor. The Zero Blne ount provided for zero minimum lne; simplified pplition forms, low mintenne hrges, trnspreny in dislosure of free trnstions limits, nd simplified identifition douments. It lso inluded smll overdrft fility. Bnks were urged to give wide puliity to the new no-frills ounts. More reent is Mexio s si nking regultion of July For ounts with less thn minimum lne, it provides list of miniml servies tht nks should offer free of fees to ll persons who omply with norml pplition requirements. There is no minimum opening mount, though nks n set ongoing minimum lne levels. There is lso n ongoing mximum lne of 165 times the dily minimum wge. After it is exeeded, nks n hrge ommissions for lient trnstions. The minimum servies to e provided inlude: opening nd mintining n ount, providing deit rd, llowing free deposits s well s free withdrwls nd ount inquiries from the ATMs of the providing nk. Mexin regultions hve lso provided for employer-sponsored si ounts. If employers hve n rrngement with nk for opening pyroll deposits for their employees, there is no ongoing minimum lne. If the employee leves, ommissions my e hrged on his or her ount. South Afri offers n exmple of ountry where there is voluntry ommitment to provide si nking, with the lunhing of its Mznsi Aounts in Otoer The government enourged nks to provide suh ounts, following the doption of its Finnil Chrter of 2003, whih lled for ess to nking for ll lients. The ount, offered y four mjor South Afrin nks, hs no minimum lne, no monthly mintenne fee, nd limited numer of free monthly deposits nd withdrwls. Mini-sttements y ell phone re ville. Wht hs the impt of si nking een? While si nking hs grown in populrity, informtion out its impt is still limited. One suh ssessment is done in Indi s Gulrg distrit, where 400,000 new no frills ounts opened etween August 2006 nd June 2007, under government-supported drive (Rmji 2007). The study found tht most respondents (75 perent) hd opened ounts for reeiving government funds under the Ntionl Rurl Employment Gurntee Progrm. Very few persons opened ounts for svings (4 perent), or trnstions purposes. It onluded tht there ws no lrge-sle impt on nk ess. The evidene from South Afri is more positive. The voluntry ode led to the opening of numer of new ounts million in the first yer lone, mounting to n dditionl 8.5 perent of totl ounts nd 4 perent of the popultion. Over 91 perent of new ount holders were previously unnked. Reent estimtes of the use of the Mznsi ount suggest 3.5 to 4 million users, of whom 60 perent re new to the nking system (Teshler nd Shneider 2008). But grdution from the Mznsi ount to regulr nking is diffiult. Bnks omplin tht the ount is unprofitle, with reltively low use nd high dormny nd ndonment rtes. Pge 57

73 Bringing Finne to Pkistn s Poor Chpter 3 Expnding outreh vi si nking mixed experiene: Another venue for hrnessing tehnology to enhne ess to finne is si nking (see Box 3.4 for interntionl experiene). 59 Bsi nking implies the provision of some pkge of free or low-ost servies to lients, with some restritions on the menu of servies nd on the mounts held in the ounts. For exmple, suh ounts my hve free trnstions up to some limit provided they re undertken y ATM rds nd do not use teller servies. Fees nd hrges for some trnstions my e wived up to free limit in terms of numers of trnstions. The World Bnk (2008) onludes tht regultion in this re, on its own, is indequte, nd there is positive ut wek ssoition etween si nking nd shre of the nked. However, in mny ples where nks hve voluntrily offered ommeril produts of si nking hrter, there is positive ssoition with ess. The experiene of Pkistn hs een mixed, wherey si ount regultion ws introdued in Novemer 2005, 60 without muh uptke from either nks or lients. In sum, the Pkistn mirofinne mrket hs muh potentil nd fes onsiderle unstisfied demnd, reting potentil for rpid outreh expnsion. The finnil setor hs not yet tken up SBP enourgement to tht effet, nd will unlikely hnge ourse given the reent finnil risis fllout. Yet, it is importnt to persevere in this gend, whih diretly links into poverty redution. Promising strtegies inlude finnil wreness mpigns, strengthening of MFI viility nd ommeril sustinility, inlusion of women nd lient segmenttion, nd fous on svings produts development. Smller size of produts, nd ulk servie (for exmple, miroinsurne, mirolesing, miro-home improvement lons, miropyment servies, griulture redit, remittnes servies, nd mirosvings), might etter ttrt lower-inome groups. The inresing use of tehnology (ATMs, POS, moile nking, si nking, rnhless nking, smrt rds) will mke this pproh vile usiness proposition for nks s well s ffordle for lients. With lose to 90 perent overge nd 59 perent reh (nd no gender divide), moile nking is n up-nd-oming opportunity to deliver remittnes. Inresing penetrtion will hve to rely on suh outfits s the Pkistn Post Offie, whose over 13 thousnd rnhes ntionwide present n ttrtive point of lunh for wider ess to finne, ut whose pity nd effiieny would enefit from n overhul, so s to permit effetive oopertion with the privte setor in rodening servie provision. 59 Countries introduing si nking inlude Mlysi, Mexio, Vietnm, nd Brzil, mong mny more. 60 It inludes minimum initil deposit of Rs 1,000 (round $17) for trnstions ount. There is no mintenne fee nd no minimum ongoing lne. It permits two free deposits nd two free hek withdrwls per month; nd unlimited free ATM withdrwls from the nks own ATMs. It inludes n nnul sttement of ount. Should the ongoing lne of the ount e zero for six months, the ount will e losed. Pge 58

74 Bringing Finne to Pkistn s Poor Chpter 4 4. Improving Finnil Aess for Smll nd Medium Enterprises (SMEs) There re 3 million SMEs in Pkistn; they onstitute more thn 90 perent of ll privte enterprises in the industril setor, employ nerly 78 perent of the nongriulture lor fore, nd ontriute over 30 perent to GDP. Smll nd miro enterprises hve seen worsening of ess to finne; they internlly finne 90 perent of working pitl nd 81 perent of new investment. In ontrst, medium-size enterprises nd those with redit history hve seen n improved ess to finne. Studies estimte n SME redit demnd gp of Rs 277 illion (ompred with urrent SME redit of Rs 400 illion). However, enterprises do not seem to e exluded from finnil mrkets euse of poor performne. Insted, n inomplete legl nd regultory frmework nd non-sme-friendly produts nd proedures hmper inresed SME lending. Indiret osts legl fees, ollterl registrtion, nd doumenttion mke nk lending expensive for SMEs. A typil smll usiness lon requires up to 27 steps for the nk nd nine meeting with lients. An enling role hs een plyed y the expnsion of Credit Investment Bureu s (CIB) sope in 2006; the SME Poliy 2007, whih emphsizes SME ess to finne; nd, ove ll, the new SBP Prudentil Regultions for SMEs. However, nks ontinue to find it diffiult to serve SMEs profitly for severl resons. First, the legl frmework (nmely, the seured trnstions regions nd, to lesser extent, the redit informtion infrstruture) limits the pool of potentil pplints. Seond, nk produts re not tilored to SMEs, resemling insted orporte lending prties. Finlly, nks do not hve orgniztionl strutures nd monitoring tools onduive to hieving high effiieny. SME demnd-side ftors, inluding limited SME ounting, udgeting, nd plnning pity further onstrin the mrket. Continued promotion of n enling environment for SME lending nd lrge-sle downsling effort involving oth the puli nd privte setors n forge rpid growth in SME lending. Inresing ess to finne for SMEs ould lso e filitted y ttrting n institutionl investor with trk reord in SME lending nd ssisting other nks to go down mrket Aess to Finne for SMEs: Supply-side Evidene SMEs ount for sustntil prt of the eonomy, yet smll enterprise lending remins limited. There re out 3.2 million enterprises in Pkistn, of whih out 3 million (93 perent) re SMEs. SMEs spred ross the eonomy with vrying density: Most re in wholesle nd retil trde nd resturnts nd hotels (53 perent), followed y other servies (27 perent) nd the mnufturing setor (20 perent). 61 The SMEs ensus shows tht SMEs ontriute over 30 perent to the GDP nd 25 perent to the ountry s totl export ernings, nd they employ lose to 70 perent of the lor fore in the mnufturing industry, servies, nd trde. Their shre in the mnufturing vlue ddition is estimted to e round 35 perent. Despite the importne of SMEs in the eonomy, s of Deemer 2007 fewer thn 200,000 orrow from the nking setor nd SME lending volumes (tht is, lons of up to Rs 75 million) ount only for 16 perent of totl redit. Although no disggregted supply-side dt re ville, demnd-side dt nd interviews with nks hve shown tht there is prtiulrly ute finning gp for lon sizes etween Rs 100,000, the mximum lon size tht Mirofinne Institutions (MFIs) n offer, nd Rs 5 million, the lon size rnge required minly y smll usinesses. Privte nks re the leding lenders in the SME finne mrket. Four of the five nks with the highest numer of individul shres in SME finne portfolios re privte ommeril nks. Among the puli setor nks, Ntionl Bnk of Pkistn (NBP) hs slightly more thn 8 perent of the totl SME finning extended y the nking setor. The puli setor nks (Bnk of Punj, Bnk of Khyer, nd First Women s Bnk Limited) hve the third lrgest shre of SME finning, fter the ig-five nks nd privte nks. Islmi nks re lso mking signifint improvements in pturing the SME finne mrket. Among these, Meezn Bnk hs the highest outstnding portfolio, pproximtely Rs 3.7 illion, 61 Eonomi Survey Pge 59

75 Bringing Finne to Pkistn s Poor Chpter 4 followed y Dui Islmi nd Dwood Islmi nks. Speilized nks shre in SME finning extended y the nking setor is low, 2.19 perent; SME nk hs the lrgest portfolio, Rs 8 illion. Foreign nks ply negligile role in SME finne (Tle 4.1). Most lons to SMEs go to finne working pitl. SBP dt show tht mjor portion (71 perent) of SME orrowing is spent on working pitl, followed y trde finning nd longterm / fixed investment (Deemer 2007). The rekdown of lending to SMEs, y type of fility, is presented in Tle 4.2 Tle 4.1: Distriution of SME Finne mong Bnks % Shre in Bnks SME Finne Portfolio Big-five nks Hi Bnk Limited Ntionl Bnk of Pkistn 7.86 Allied Bnk Limited 7.46 Muslim Commeril Bnk Limited 6.72 United Bnk Limited 5.81 Privte nks (exluding ig 5) Puli setor nks (exluding NBP) 4.82 Foreign nks 0.42 Islmi nks (exluding Al-Brk Bnk) 4.82 Speilized nks 2.19 Soure: SBP, Note: SBP figures for 1 st qurter 2008, resled to dd up to 100. Tle 4.2: Brekdown of Lending to SME y Type of Fility (Rs illion) Type of Fility De 04 De 05 De 06 De 07 Mr 08 Fixed investment Working pitl Trde finne Totl * Soure: SBP * Provisionl figure Government Poliies to Improve Finnil Aess to SMEs] The Government of Pkistn (GOP) nd SBP hve introdued poliies to improve the legl frmework nd rete demonstrtion effet. The GOP nd SBP hve lunhed numer of inititives to support the growth of SME lending. These n e rodly grouped in two res: inititives iming t filitting SMEs provision of ollterl nd redit history nd tivities imed t reting demonstrtion effet for other finnil institutions, tht is, finning tehnil ssistne for seleted finnil institutions to inrese SME lending, thus showing to other mrket prtiipnts tht the SME segment is indeed profitle one nd ultimtely reting ompetition in the segment. Finlly, to monitor SME lending, the SBP hs introdued more detiled SME definition on pilot sis nd requires nks to report on their SME portfolio on qurterly sis. SBP hs mde two unsuessful ttempts to inrese lending to SMEs; firstly through the retion of n effetive seured trnstions system nd seondly through the relxtion of ollterl requirements. The trditionl orporte lending produts of Pkistni nks emphsize seurity, espeilly immovle ollterl, udited finnil sttements, nd usiness plns, rther thn doing firsthnd reserh on sh flows nd usiness performne. This mkes it more diffiult for finnil institutions to extend lons to SMEs, whose ollterl is limited nd movle, nd whih do not find finnil sttements nd usiness plns esy to produe. The generl resistne of nks to ept movle ollterl is ompounded in Pkistn y seured trnstions system tht mkes it unsfe to lend to sole proprietor y ollterlizing the lon with movle sset. Currently, only limited liility ompnies n register, nd hene notify to the puli t lrge, lien over movle ssets. They do so in the ompny registry, whih is not ville for unlimited liility ompnies, s these re not registered (more detils on this re provided in setion 4.3). To remove this rrier for sole-proprietor SMEs, the SBP hs ttempted to introdue registry for liens over movle ssets. However, enountering delys in Pge 60

76 Bringing Finne to Pkistn s Poor Chpter 4 setting the new system in ple, the SBP introdued regultion in 2004 llowing unollterlized lending for lons up to Rs 3 million nd without finnil sttements for lons up to Rs 10 million. Despite the new leglity of this lending, the lending prtie is proving too risky for onservtive Pkistni nks. To filitte the retion of redit history for SMEs, SBP hs promoted the estlishment of redit ureus, oth puli nd privte. However, despite their improved performne, their overge is limited, espeilly for smller orrowers. Not unlike other ountries, Pkistni SMEs hve no or little forml redit history; this mkes forml finnil institutions more hesitnt to lend to the segment. To promote the retion of forml redit histories y smll usinesses, SBP founded puli CIB in The CIB reords oth positive nd negtive informtion out ompnies, s well s ll the ssets they use s ollterl for the lons. CIB performne ws enhned in 2006 with the introdution of n e-cib system, whih improved speed, reliility, nd seurity of dt. In tht sme yer, CIB overge ws expnded to ll lon sizes. In ddition to the stte-owned CIB, two privtely owned redit ureus now operte in the ountry. Despite their improved performne, redit ureus over only perent of the orrowers nd very few SMEs requiring lons smller thn Rs 6 million. This is prtly euse the nks lend to this segment nd prtly euse the ureus do not ollet informtion from the teleom nd utility ompnies, whih ould sustntilly ontriute to uilding redit history for SMEs. To rete demonstrtion effet for other mrket prtiipnts, the GOP hs supported the retion of speilized SME Bnk, whih, in three yers sine its ineption, hs served mere 1 perent of the mrket. To hllenge the mrket nd demonstrte tht the SME segment is not only vile ut very profitle, the GOP reted nk dedited to serving only SMEs (Box 4.1). SME Bnk ws reted y merging two filed stte-owned nks. In 2007, two yers fter strting opertions, SME Bnk hd only 2,200 lients (1.2 perent of totl SME orrowers) nd offered less thn 2 perent of totl SME finne extended y the nking setor. Box 4.1: SME Bnk To t s tlyst for the SME mrket, in 2002 the GOP reted SME Bnk Ltd through the merger of two filed Development Finnil Institutions, the Regionl Development Finne Corportion (RDFC) nd Smll Business Finne Corportion (SBFC). SME Bnk reeived ommeril nking liense in 2004 nd strted opertions in mid Its urrent shreholders re the Federl Government (92.68 perent) nd six nks (7.32 perent). The nk ws poised to e privtized shortly fter eoming opertionl. However, the Supreme Court deision to reverse the Pkistn Still Mills privtiztion, the ft tht the SME nk ws smll-tiket item, in the privtiztion list nd the inresingly deteriorting seurity sitution in the ountry hve put the privtiztion on hold. SME Bnk opertes totl of 27 rnhes: 13 ommeril nking rnhes nd 14 reovery rnhes. In Deemer 2007, its totl ssets were Rs 6.6 illion, nd its net outstnding lons mounted to Rs 2.1 illion. As of 2007, the nk hd only 2,200 lients or 1.2 perent of totl SME orrowers in the ountry for n verge lon size of Rs 840,000 over the three yers. In theory, the nk provides lons of terms up to seven yers. In prtie, however, working pitl lons of up to Rs 500,000 re given up to term of three yers, nd sset finne up to four yers; in exeptionl ses, the term n e up to five yers. Despite this growth, SME Bnk hs not een le to meet the redit needs of the SME setor; this stems from its use of trditionl lending tehnology tht does not distinguish etween lrge orportions nd smll usinesses, inpproprite produt design, nd its lengthy pplition proedures (it tkes on verge 40 dys to reeive lon). Unlike t other SME nks, SME Bnk lon offiers do not hve onus system sed on the performne nd size of the lon portfolio tht they generte nd mnge. A omprison with suessful SME nks in the region nd eyond is presented in the tle elow. Pkistn s SME Bnk hs y fr the worst portfolio performne with 72 perent of the gross lon portfolio in volume lssified s loss. The limited ttention pid to lending is lso refleted in the nk humn resoure poliy of employing only smll numer of lon offiers, less thn one per rnh. Soure: Author s interview with SME Bnk mngement Pge 61

77 Bringing Finne to Pkistn s Poor Chpter 4 Results for Seleted Smll Business Bnks (Deemer 2007) Pro Credit Ukrine BnoSol Bolivi BRAC Bnk Bngldesh SME Bnk Pkistn Totl ssets (million $) Yer of estlishment Numer of outlets or rnhes Brnhes: SME unit offies: 392 Numer of lon offiers 246* 297* 1, Averge time to disurse lon (dys) Gross lon portfolio (million in $) Averge lon size outstnding (in $) 7,358 2,360 4,067 13,703** Portfolio t risk (PAR) >30 dys 0.98% 1% 18% 72%*** Totl ssets for holding/fund (in $) 6 illion million 2.4 illion Not Applile Soure: Pro Credit Ukrine; BnoSol, Bolivi; BRAC Bnk, Bngldesh; SME Bnk, Pkistn. * Dt for **Averge lon size disursed ($). ***Perentge of the portfolio lssified s loss s per SBP regultions. Lk of shreholders with the required know-how is the primry reson for the filure of SME Bnk to tlyze the mrket. A numer of ftors, ut most notly two, hve ontriuted to the filure of the SME Bnk experiment. First, euse the intended privtiztion of the nk never took ple, the nk does not hve institutionl investors with the know-how required to support the nk (for exmple, the ord did not give ler nd mitious quntittive trgets to nk mngement). Seondly, eing reted from wht remined of two filed institutions ment tht SME Bnk mngement hd to expend energy on restruturing rther thn expnding nd modifying its operting model to serve new mrket segment. The GOP hs funded tehnil ssistne for NBP to enter the SME segment, with some suess. Strting in 2005, the GOP filitted the delivery of tehnil ssistne to the NBP to sustntilly expnd SME lending. Lending on sh-flow, not ollterl, sis ws introdued, nd lon offiers were put on performne-sed ontrts. NBP hs hieved some promising results fter three yers (Box 4.2). Nevertheless, while the progrm hs ertinly inresed NBP s lending to SMEs, the size of the SME lon portfolio remins quite smll. By omprison, similr progrms in Chin over the sme time period expnded to lons for n verge of $10,000 with portfolio t risk (PAR) >30 dys of 1 perent, totling $380 million in 15 regions in the ountry (see Box 4.4 in setion 4.3). The Pkistn progrm s limited suess is most likely ttriutle to the lk of ompetition. While downsling progrms with stte-owned nks n work, usully they re omined with downsling progrms for fster-moving privte nks, whih implement the required hnges t fster pe. Box 4.2: Ntionl Bnk of Pkistn s SME Lending NBP is stte-owned nk with the lrgest rnh network (1,243 rnhes) in the ountry. Its totl ssets exeeded $12 illion in In 2005, NBP signed tehnil ssistne greement with ShoreBnk Interntionl under U.S. Ageny of Interntionl Development (USAID)-finnil setor projet to sustntilly inrese smll usiness lending (lons typilly under Rs 2 million). As prt of the greement, Shore Bnk Interntionl introdued onus system for its lon offiers sed on the numer of lons issued nd qulity of the performne of the lon they mnged. Lon offiers were trined in the new lending methodology nd onsite support ws provided to tive rnhes. NBP nd ShoreBnk Interntionl re ontinuing to work together on new ilterl greement pid for y NBP tht runs through Initilly, the tsk of onvining lon offiers nd mngers to fous on smll enterprise lons proved to e diffiult. The rnhes lked performne-sed ulture, nd there were mny mngement lyers. Most redit deisions were sed on the sis of immovle ollterl. The hllenge for the ShoreBnk Interntionl tem ws to shift Pge 62

78 Bringing Finne to Pkistn s Poor Chpter 4 the fous from ollterl to underwriting on sh flows nd usiness performne. After extensive fieldwork, forms nd formts were designed to ssess repyment pities sed on projeted sh flows nd finnil rtio nlysis. In ddition, tool ws developed to ssist lon offiers in onverting informl sh ook reeipts into stndrd finnil douments. But more importnt thn the underwriting pproh ws the need to effetively motivte lon offiers to leve their rnh offies to mrket nd underwrite new lons t the usiness premises. The strt ws slow, ut the key step in the proess ws introduing regulr monitoring report tht ws shred t ll levels of NBP, providing trnsprent mesure of progress. This report, irulted t lest monthly, served to stimulte interest nd keep senior mngement nd rnh-level tems foused on mesurele ojetives. NBP s lons to SMEs inresed strting in June Over the first 18 months of the greement, 658 lons were issued for totl lon portfolio of Rs 1.4 illion. Dt on performne of outstnding lon portfolio re presented in the tle elow. Portfolio Qulity of the SME Lending Progrm (Novemer 30, 2007) Numer of lons 510 Amount of outstnding Rs 813,089,934 Amount of outstnding lon portfolio $13,439,503 Averge lon size $26,352 Numer of overdue lons >90 dys 8 % of overdue lons (volume) >90dys 2% Soure: ShoreBnk Interntionl stff In ddition, SBP hs introdued more detiled SME definition on pilot sis nd requested nks to report qurterly; lthough this is step in the right diretion, the rkets re too mny whih mkes nks reporting too omplex. To monitor SME lending, SBP reted n SME deprtment nd now requires nks to report ll lons smller thn Rs 75 million on qurterly sis. Thus, impliitly, SBP hs defined SME lons s those smller thn Rs 75 million; this definition is intended to filitte nk reporting. While this ould llow trking of lending to the segment to egin, the rekdowns re too mny nd might tully disourge nks from reporting urte informtion. Tle 4.3: Credit to the Privte Setor: A Profile (Rs illion) Deemer 2004 Deemer 2005 Deemer 2006 Deemer 2007 Setor of the Shre Shre Shre Shre Eonomy Amount (%) Amount (%) Amount (%) Amount (%) Corporte , SMEs Agriulture Consumer finne Commodity opertions Stff lons Other Totl 1, , , Soure: SBP Government efforts to inrese SME ess to finning hve hd modest results, with lons to SMEs delining s perentge of totl lending. Aginst the kdrop of growing nking setor, SME lending hs witnessed slight downwrd trend over the period SMEs ounted for 16.2 perent of totl redit to the privte setor in 2007, down from 17.5 perent in In ontrst, oth orporte nd onsumer finne grew, from 54 perent to 56 perent nd from 9 perent to 14 perent, respetively, in the sme period (Tle 4.3). The verge SME lon hs lso slightly inresed in size, rehing Rs 2.4 million. Pge 63

79 Bringing Finne to Pkistn s Poor Chpter Aess to Finne for SMEs: Demnd-side Evidene To omplement the rod piture of SME finnil ess presented from supply inditors, this setion mesures ess from the demnd side. The nlysis presented here is minly sed on the 2005 urn demnd survey, onduted y KfW. The survey overed 510 SME usinesses operting in mnufturing, trde, nd other servie setors in the Punji ities of Fisld, Gujrnwl, Lhore, Srgodh, nd Silkot. The survey defined SMEs s those usinesses employing not more thn 100 permnent employees. (See Annex 4.1 for detiled desription of the survey uilding loks, methodology, nd smpling). The nlysis here lso drws on the 2005 Europen Commission Survey nd the 2002 nd 2007 Investment Climte Assessment (ICA) surveys (World Bnk 2003 nd 2009). The Europen Commission Survey overed 100 mnufturing SMEs in the golden tringle, 62 while the 2002 nd 2007 ICA surveyed 965 nd 1,184 enterprises, respetively, in mnufturing nd servies in urn Pkistn. 63 It should e stressed tht, lthough the definitions of SMEs used y the Europen Commission nd ICA surveys vry slightly from those used in the KfW survey, the differene is negligile nd their key findings on finnil ess (ess, soures, ostles) oinide. All the dt presented here, unless speifilly mentioned otherwise, refer to the 2005 KfW dtset. Although onsiderle numer of Pkistni SMEs hve nk ount, few orrow from nks. The SMEs hve lso more limited ess to finne thn their ounterprts in Indi nd Bngldesh. Thirty-six perent of Pkistni SMEs hve nk ount, while only 7 perent hd t lest one lon outstnding in the previous three yers ( ). In omprison, in 2006, 43 perent of SMEs in Bngldesh, nd 95 perent in Indi, hd nk ount, 64 nd 32 perent nd 33 perent, respetively, hd lon from nk. Moreover, it should e highlighted tht Bngldeshi nd Indin SMEs lso hve ess to finne from other finnil institutions, minly MFIs in Bngldesh nd finne ompnies in Indi. Figure 4.1: Aess to Finne for SMEs in Pkistn, Indi nd Bngldesh Bngldesh 43 Bngldesh 32 Indi 95 Indi 33 Pkistn 36 Pkistn % SMEs with nk ount % SMEs with lon from nks Soure: Indi ICA 2006 Mnufturing Enterprise Survey, Bngldesh 2006 Rurl MSME Finne Servies Survey, Pkistn KfW Demnd Survey An SME study ws onduted y the Europen Commission in 2005 under its Pkistn Finnil Servies Setor Reform Progrm. The SME definition used ws: smll (medium) enterprises tht employ up to 50 persons ( persons), ssets exluding lnd nd uilding not more thn Rs 30 million (Rs million), nd sles up to Rs 100 million (up to Rs 300 million). 63 The 2002 ICA survey defines smll enterprises s those employing 1 19 employees nd medium enterprises s those employing employees. The sme definition of firm size is used for the nlysis of the 2007 ICA in this urrent study. 64 Dt for Bngldesh is sed on the 2006 Rurl MSME Finne Servies Survey, whih inludes 226 SMEs in ities outside of Dhk nd Chittgong. Dt for Indi refers to ICA 2006 mnufturing enterprise survey, whih inludes 2460 firms in 16 sttes ross Indi; only dt for SMEs (enterprises with less thn 100 employees) ws used in this nlysis. Pge 64

80 Bringing Finne to Pkistn s Poor Chpter 4 Over the period, SME ess to finne inresed, though this is ttriutle to inresed ess for medium enterprises rther thn for the whole segment. The numer of enterprises tht reported hving t lest one outstnding lon or overdrft fility inresed sustntilly from 2002 to 2005 (ICA 2002 nd 2007). However, when disggregting the dt, ess to finnil servies hs inresed only mrginlly for smll enterprises (from 10 perent to 11 perent) nd hs inresed sustntilly for medium-size enterprises (from 27 perent to 43 perent). This finding is in line with the result of interviews with nks tht identified finning gp t the ottom of the mrket (Rs 100,000 5 million). During the sme period, enterprises tht hd redit history found it esier to ess finne. Firms overed in the KfW survey tht hd pplied for lons in the period were sked whether ess to nk lons hd eome esier over the period, nd 53 perent limed tht is ws very/firly esy in 2004 versus 33 perent in Improvement in ess to finnil servies minly for medium-size enterprises nd for enterprises tht lredy hve redit history is proly short-term outome of the privtiztion. Immeditely post privtiztion, s nks redue nonperforming lons, reorgnize, nd strt expnding their portfolios, they tend to pply striter risk-mngement rules nd to fous on more-estlished lients. This, omined with redution of mndted lending to smller lients nd government-finned produts, often uses short-term redution in lending to the lower end of the mrket. Only fter ompetition in the more-estlished segment hs intensified do nks strt fousing on lower segment. The SME mrket gp is estimted t Rs 277 illion. The KfW survey sked respondents to estimte the mount of externl finning required to finne working pitl nd fixed investments in Bsed on these estimtes, the potentil mrket size for serving smll usinesses is pproximtely Rs 300 illion (see Box 4.3). More thn 90 perent of this demnd is still unserved, mking smll usiness finne highly ttrtive venture for profit- nd growth-oriented nks in Pkistn. Huge redit demnd oupled with miniml supply promises high profit mrgins nd strong growth for nks going down mrket. Thnks to the lrge mrket size, nks will e le to sustin growth rtes in SME lon portfolios for yers to ome if they ould understnd well this prtiulr mrket segment nd hnge their nking prties to effetively ter to the needs of smll usinesses. Pge 65

81 Bringing Finne to Pkistn s Poor Chpter 4 Box 4.3: How Big is the Potentil Mrket for Providing Credit to SMEs? Although it is impossile to preisely quntify the overll demnd for redit y smll usinesses in Pkistn, n pproximtion n e ttempted y extrpolting from the results of the 2005 KfW urn demnd survey. Aording to Smll nd Medium Enterprise Development Assistne estimtes, there re lose to 3.2 million miro, smll, nd medium-size enterprises in Pkistn. Close to hlf of ll usinesses surveyed required externl finning of Rs 100,000 5 million. (Eighty-two perent of respondents needed lons for working pitl nd 61 perent of this demnd fell into the Rs 100,000 5 million rnge; 40 perent of the respondents needed lons for fixed pitl, of whih 74 perent fell into the Rs 100,000 5 million rnge.) Fifty perent (61 perent of 82 perent) of SMEs require finning in the Rs.100,000 5 million rnge, nd there re round 1.6 million suh enterprises. However, not ll of these will qulify for nk lons. Assuming tht 70 perent of firms tht wnt to orrow re reditworthy nd tht 70 perent of these re orrowing t the sme time, there re round 8 million eligile SMEs in Pkistn. All of these lients wnt to orrow working pitl, nd round 500,000 wnt to orrow to finne fixed investments. For the ske of simpliity, it is ssumed tht fixed investment orrowers re suset of the lrger group needing working pitl nd only 20 perent of ll potentil orrowers will get prllel lon. Survey results indited tht, in the se of working pitl, 85 perent of lients require lons in the 100,000 1 million rnge, while 15 perent require lons of up to Rs 5 million. In the se of fixed investments, lon requirements re nerly evenly split. Forty-nine perent of lients wish to orrow in the lower rnge, while 51 perent wnt to tke out lon of etween Rs 1 nd 5 million. Amounts nd Purposes of Lons Required y Smll Businesses Lon Purpose Lower Rnge Upper Rnge (Rs 100,000 1 million) (Rs 1 5 million) Totl Borrowers Working pitl 680, , ,000 Fixed investments 78,400 81, ,000 The solute minimum totl lon volume required y smll usinesses n e derived y multiplying the numer of orrowers in eh tegory y the minimum lon mount. In other words, ll lower-rnge orrowers re ssumed to require only Rs 100,000. Upper-rnge lients re ssumed to require extly one lon of Rs 1,000,000 eh. Minimum Credit Needs of Smll Businesses in Pkistn Lon Purpose Lower Rnge Upper Rnge (Rs 100,000 1million) (Rs 1 5 million) Totl Credit Needs Working pitl 68,000,000, ,000,000, ,000,000,000 Fixed investments 7,840,000,000 81,600,000,000 89,000,000,000 All lons 277,000,000,000 While lulting the totl redit demnd, only miniml lon mounts were tken into ount for working nd fixed pitl needs. Hene, the totl redit demnd n e sfely rounded up to lose to Rs 300 illion nd perhps even higher. SMEs use limited nk funding for oth working pitl nd fixed investments; forml funding is even more limited for smll thn for medium enterprises. Approximtely 89 perent of working pitl nd 75 perent of fixed investments were finned from retined ernings in nd the perentges re even lrger for miro nd smll enterprises: 90 perent for working pitl nd 81 perent for new investment (Figures 4.2 nd 4.3). Miroenterprises lso do not sustitute nks with MFIs, ut rther orrow more from fmily nd friends. This is explinle y the more ute gp tht there is in the mrket for lon sizes etween Rs 100,000 nd Rs 5 million (equivlent to etween $1,280 nd $64,000). This segment is lso not served y nks for the resons mentioned ove, or y the mirofinne setor, whih is leglly ound to lend up to Rs 100,000 nd trditionlly foused on the lower end of the mrket. Pge 66

82 Bringing Finne to Pkistn s Poor Chpter 4 Figure 4.2: Externl Soures of Working Cpitl Trde redit from Trde redit from suppliers, 0.60% ustomers, 0.50% Other, 0.30% Figure 4.3: Externl Soures of Fixed Cpitl Equity (new usiness prtner), 1.20% Bnk lons, 3.70% Fmily nd friends, 5.50% Bnk lons, 10.20% Fmily nd friends, 13.50% Soure: KfW Demnd Survey 2005 Not only do miroenterprises pply less frequently for Tle 4.4: Ese of Otining Credit lons; when they do, they re muh more likely to e rejeted thn medium ones. Aess to redit eomes progressively more onstrined s the size of the enterprise eomes smller: only 10.7 perent of Very esy Firly esy 2002 (%) (%) miroenterprises pplied for redit ompred with 25 Firly diffiult perent of medium enterprises, nd nerly 75 perent of Very diffiult them were refused lon wheres over two-thirds of Totl medium enterprises mnged to seure the lon for whih they pplied (Figure 4.4) Soure: KfW Demnd Survey 2005 Figure 4.4: Lon Applition nd Disursement y Firm Size g e t n 40 re e P 20 0 Miro Smll A (10 19) Smll B (20 49) Medium (50 100) Applying for Credit Applitios Approved Enterprises do not seem to e exluded from finnil mrkets euse of poor performne. The Europen Commission survey reports tht 60 perent of the SMEs hve profit mrgins in the rnge of 5-10 perent, while 12 perent hve higher profit mrgins of perent. This is orroorted y SME owners fvorle pereptions of their own usiness performne in the preeding two yers ( ). Over 60 perent think tht usiness hs improved, while only 12 perent think tht usiness hs Soure: KfW Demnd Survey 2005 delined (Tle 4.5). It would seem tht mny SMEs re exluded from finnil mrkets not euse of d performne ut euse of gp in the mrket. Pge 67

83 Bringing Finne to Pkistn s Poor Chpter 4 Tle 4.5: Business Development in the Pst 24 Months Miro Smll A Smll B Medium Totl (10-19) (20-49) (50-100) Delined signifintly Delined somewht Remined the sme Improved somewht Improved signifintly Totl Soure: KfW Demnd Survey 2005 When sked, SMEs stte tht nks re only the third preferred soure of orrowing, minly euse of high diret nd indiret osts nd omplex produts. Of the 85 perent of the smple tht did not pply for redit in , the mjority (54 perent) ited vilility of other soures of finning s the reson. When sked to rnk their preferred soures, SMEs rnk nks only third, fter nongovernmentl orgniztion MFIs nd fmily/friends. Other soures re preferred euse of lower diret nd indiret osts (50 perent), euse nks do not offer Islmi nking produts (24), nd euse nks require lot of ollterl (10 perent). Diret osts inlude interest nd fees, indiret osts the ost of produing the required douments, time the owner needs to spend with nk offiils, nd length of the proess, whih trnsltes in missing erning opportunities) Bnks require immovle ollterl lmost exlusively. Firms interviewed in the 2007 ICA survey reported tht nks ept lmost exlusively immovle ssets nd personl ssets of the owner (lso minly immovle ssets) s ollterl; 92 perent of the interviewed SMEs hd to provide immovle ssets only or immovle nd movle ssets, with the ltter in seondry mount. Tle 4.6: Demnded Lon Produt Fetures Men Low diret nd indiret ost 4.37 Convenient repyment period 4.39 Asene of requirement for immovle property s ollterl 3.94 Convenient lotion of finnil institution 3.81 Qulity of servie of finnil institution s stff 3.75 Avilility of other finnil servies from sme provider 3.63 Soure: KfW Demnd Survey 2005 These findings re orroorted y the fetures tht SMEs look for in lon produts: low diret nd indiret osts, onvenient repyment periods, nd no requirements for immovle ollterl, in very lose order. Respondents ssessed the importne of eh spet of orrowing on the sle of 1 to 5, 1 eing lest importnt nd 5 eing very importnt (Tle 4.6). For the respondents, low diret nd indiret ost nd onvenient repyment period re the most importnt fetures of lon produt nd must e emphsized in the design of lon produt. The onstrints to improving SME ess to finne re noted in setion Constrints to Improving SME Aess to Finne Pkistni nks tret SMEs s lrge orporte enterprises, when in ft SME lending hs more similrities with onsumer finne. In reent yers, SME lending in reltive terms hs deresed, while lending to orporte nd onsumer finne hve grown. This is minly euse nks mke little or no distintion etween lrger orportions nd SMEs in terms of produt nd proedures. Similrities etween onsumer nd SME finne inlude low-vlue trnstions nd sensitivity to lon delivery time. However, unlike onsumer finne, SME lending requires more individulized ssessment of sh flows, euse SMEs often hve limited nd sometimes inurte written sttements. SMEs re hrterized y limited formlity nd re sensitive to lon delivery time nd omplex proedures. SMEs re usully smll, fmily-owned usinesses, run nd mnged y one or two people tking full responsiility for ll spets of the usiness. These individuls typilly know their usiness well ut lk the skills to produe written usiness nd finnil plns nd re disourged y the extensive doumenttion requirements of nks. Moreover, when suh enterprises require finning, they Pge 68

84 Bringing Finne to Pkistn s Poor Chpter 4 usully need it immeditely, euse their mngement rrely plns opertions fr in dvne. It lso is ommon tht these enterprises hve limited or no forml redit history, lk forml finnil sttements, nd possess few immovle ssets. Business nd household finnes re strongly intertwined, nd re minly sh. It should e noted tht the nking needs of SMEs re very different from those of lrge orportions, nd even mong the SMEs, produt preferenes differ signifintly. Lrge volumes, effiieny, nd qulity re key to profitle smll usiness lending. To profitly serve SMEs, nks need to minimize trnstion osts nd generte lrge numer of high-qulity lons. As in similr tivities with smll profit mrgins, nks need to inrese revenue y mking mny lons while lowering expenses for exmple, y mking lon offiers more produtive nd voiding d lons. An inomplete legl nd regultory frmework nd lk of SME-friendly produts nd proedures hmper inresed SME lending. Bnks in Pkistn find it diffiult to serve SMEs profitly for severl resons. First, the legl nd regultory frmework (nmely, the seured trnstions regions nd, to lesser extent, the redit informtion infrstruture) limits the pool of potentil pplints. Seond, lending produts nd nk proedures re more suited to lrge orportions thn smller enterprises. Finlly, nks do not hve orgniztionl strutures nd monitoring tools onduive to hieving high effiieny. In spite of sustntil improvements of the legl nd regultory frmework, work remins to e done in the re of seured trnstions. 65 Progress hs lso een sustntil on reting redit ureus, though more ould e done to filitte the retion of redit histories y SMEs. In spite of the reform tht the GOP nd SBP initited five yers go, there is still no oherent seured trnstions regime nd no registry where unlimited liility ompnies n register lien on movle ssets. SBP hs tried to ompenste for this y wiving ollterl requirements for very smll lons (SME prudentil regultion). However, nks hve proven to e too risk verse to tke full dvntge of the regultion nd still prefer lons ollterlized with immovle ssets. Both well-funtioning seured trnstions regime nd more omplete redit informtion system ould help gp the ridge etween demnd nd supply of SME lending. An effetive seured trnstions regime plys vitl role in ountry s finnil system. 66 By expnding the pool of ssets tht n e offered s ollterl, seured trnstions system n help expnd finnil ess for those who were previously exluded or hd only limited finnil ess. Furthermore, reformed seured trnstion regime n promote ess to redit nd llow redit to e extended on etter terms, y improving trnspreny in determining the detor s reditworthiness nd y inresing seurity in estlishing limnts to ollterl. Empiril evidene from numer of developing ountries suggests tht n inrese in the index of eptle ollterl is ssoited with etter finnil ess to nk lons for firms, inluding for SMEs (Figure 4.5). 65 SBP, in oopertion with ADB, hve reently ompleted detiled study on improving the seured trnstions regime nd the puli registry offie, inorporting issues of the retion of interest over movle ssets nd hrge for SMEs nd rurl finne. 66 Seured trnstion is defined s ny trnstion, regrdless of its form, tht is intended to rete seurity interest in personl property or fixtures, inluding goods, douments, nd other intngiles. A seurity interest is right given to one prty in the sset of nother prty to seure pyment or performne of n oligtion. Collterl onstitutes the ssets over whih seurity is grnted Pge 69

85 Bringing Finne to Pkistn s Poor Chpter 4 Figure 4.5: Bnk Aess Inreses with Wider Rnge of Aeptle Assets to Seure Lon % of working pitl from nk lons o eff: 33.4, t: 1.78 All Firms Muritius Thilnd Vietnm Ethiopi Sri Lnk Keny El Slvdor Burundi Senegl Cmeroon Zmi Niger Mlwi Mli South Afri Burkin Fso Cmodi P hilippines Gmi DRC Swzilnd Tnzni Angol Botswn Index Rnge o f Aeptle Co llterl Smll nd Miro Firms (<50 employees) Thilnd Keny M uritius El Slvdor Ethiopi Burundi Vietnm Niger Mlwi Sri Lnk Senegl So uth Afri Burkin Fso Zm i Mli Cmeroon Philip p ines Gm i DRC Cm o d i Tnzni Swzilnd Angol Botswn o eff: 28.17, t: 1.9 Index Rnge o f Aeptle Co llterl Soure: World Bnk (2008) Note : A ross-ountry regression of the % of firms with working pitl funded y nk lons on the index of eptle ollterl shows tht 10% inrese in the index is ssoited with 3.3% inrese in firms using nk lons to fund working pitl (t-stt: 1.78.) Note : A ross-ountry regression of the % of smll firms with working pitl funded y nk lons on the index of eptle ollterl shows tht 10% inrese in the index is ssoited with 2.8% inrese in smll firms using nk lons to fund working pitl (t-stt: 1.78.) The existing seured trnstions system hs mjor defiienies tht limit, in prtiulr, SME use of movle ollterl. Key omponents of n effetive seured trnstions regime re: 1. Cretion: estlishing lim to property to seure pyment of redit 2. Priority: determining rnking of lims over ollterl 3. Puliity: mking priority interests pulily known 4. Enforement: repossessing ollterl nd selling it for stisfying lims Annex 4.2 presents detiled nlysis of the seured trnstions regime in Pkistn, nd undersores severl shortomings of the system inluding ertin legl provisions tht slow down enforement, the lk of speilized ourts, nd the sene of omprehensive, single registry for ll movele hrges in respet of ll types of detors (orporte or otherwise). The system s min limittions inlude: Multiple lws nd systems: In Pkistn, there is no simple, unified sttute pertining to the retion of seurity over movle property. The multiple lws nd systems rete unertinty nd diffiulties in using movle property s ollterl. Another importnt limittion is tht the detor must possess ny ollterl eing pledged. In ddition, the legl frmework for enforement ginst not-yet-existent ollterl is wek. This ffets the use of griulturl produts (suh s future rops), inventory, or fter-quired property s ollterl, nd would mke it diffiult for mny smll usinesses tht do not hve signifint mount of other lsses of ssets eyond fter-quired property to otin seured redit. Avilility of seured trnstions over movle property only to registered ompnies: Another importnt limittion is tht the system of seured trnstions over movle property is ville primrily to registered ompnies only. This is euse the registrtion of seurity interests is generlly only ville for registered ompnies (unless the ssets eing used re themselves sujet to registrtion), nd not for ny other orgniztions or usiness entities flling outside of the Compnies Ordinne, This strongly disrimintes ginst smll usinesses, whih re not registered ompnies. Pge 70

86 Bringing Finne to Pkistn s Poor Chpter 4 High trnstion osts: There is no unified registry for ll types of seurity interests grnted y ll types of entities on ll types of movle property. Pkistn hs mixed system of Asset-Bsed Registries nd n Entity-Bsed Registry. Not ll registries operte on notie filing sis nd entire greements need to e filled. In ddition, presently, only the ompnies registry is ville online nd is networked ross the ountry. With pper-sed system, verifition of ollterl eomes time-onsuming proess tht does not llow potentil lenders rel-time serhes of existing liens. This sustntilly inreses the trnstion osts of filing seurity interests. Slow enforement proess: Enforement of det reovery is slow in Pkistn. A seurity holder hs to file suit in ourt to reover his lim. However, for finnil institutions, expedited proedures re ville under the Finnil Institutions (Reovery of Finnes) Ordinne, But even in these ses, should the detor dispute the enforement y the reditor (whih is usully the se), firly lengthy ourt proedure will ensue. In prtie, if finnil institution is enforing its seurity right under the speil summry proedures, it would tke one nd hlf to two yers for the enforement to e ompleted. In other ses, it my tke yers or more. By this time, most movle property would hve lost its vlue, thus disourging finnil institutions to ept suh ollterl. Hving forml redit history is prtiulrly importnt for SMEs, whih re often less known entities to the nking ommunity. Credit registry, through whih lenders shre informtion out their lients repyment reords, is n estlished wy of enhning the ility of orrowers to signl good redit reord. Credit registry is even more importnt for SMEs, mny of whih hve hd limited ontt with the nking ommunity (64 perent of the Pkistni SMEs do not hve nk ount), less forml ounts, nd more limited ility to produe written usiness pln. While progress on the redit infrstruture hs een sustntil, work remins to e done. In prtiulr, the inlusion of informtion from utilities nd teleom providers in the registry ould gretly filitte the retion of redit history for SMEs. This informtion ould then e used s the pltform through whih SMEs n in the future show trk reord of repyment to nks. Bnk proedures nd poliies used for SME lending re too omplex, mking them time onsuming nd ostly for oth prties. Bnks pply the sme rules nd poliies to smll enterprises s they pply to lrge orportions; s result, typil smll usiness lon requires up to 27 steps for the nk (or 20 hours) nd nine meeting with lients (or 10 hours of the potentil orrower). Totl time tken to otin lon n vry from 30 to 45 dys on verge. Suh long nd expensive proedures limit the numer of lons tht n e mde month per lon offier. Furthermore, it trnsltes into missed usiness opportunities for SMEs, euse the mnger/employee must spend time wy from the usiness to omplete proedures to otin lon. In ddition to these high trnstion osts, indiret osts in terms of legl fees, ollterl registrtion, nd doumenttion mke nk lending expensive for SMEs. Bnks hve limited speiliztion in SME Tle 4.7: SME Applition Proess for Bnk Lons lending nd no mesures of SME usiness Inditor Averge Numer performne. Bnks further shy wy from SMEs Steps 27 due to lower vlue per trnstion nd distintive Meetings with orrower 9 fetures of SMEs, whih re onsidered n Stff involved 8 dditionl redit risk. Bnk SME lending is thus minly driven y risk onerns nd no ttention is Borrower time (hours) Soure: Authors lultions. 10 relly pid to effiieny, whih is the key to serve this prtiulr mrket segment profitly. Risk onerns must refully e lned ginst onsidertions of ustomer orienttion nd effiieny when designing produts nd proedures for smll lients. To hieve higher effiieny, speiliztion of stff, trining, nd proedures re essentil. In Pkistn, nks do not engge suffiiently in suh prties. Pge 71

87 Bringing Finne to Pkistn s Poor Chpter 4 Box 4.4: Chin SME Lending Projet Sine 2005, Chin Development Bnk hs sponsored tehnil ssistne progrms with lines of redit to enourge nks to inrese SME lending. The progrm, whih is funded y KfW nd the World Bnk, inludes oth privtely nd stte-owned nks. The key omponents of the progrm re represented grphilly elow. As illustrted, eh prtiipting nk signs inding performne greement (inditing, for exmple, how mny lon offiers will e devoted to the new produt, how mny lons will e disursed per month, nd the PAR to whih the nks ommit). Projet Design Interntionl Finnil Institutions USD 100 Mill lon inl. up to USD 10 Mill for TA TA funds Onlending funds EUR 50 Mill lon plus EUR 3 Mill grnt for TA TA funds Apex Institution / Projet Mngement Unit CHINA DEVELOPMENT BANK On-lending funds Tehnil Assistne Provider Cpity uilding Institution uilding Trining Prtner Bnks PB PB PB PB PB PB PB Trget Group Miro nd Smll Enterprises Aout 12 nks in Chin hve prtiipted in the progrm, nd six of them hve lredy suessfully grduted. All nks hve some sort of lol government or government enterprise ownership, exept for one nk, whih is fully privtely owned. The progrm hs hieved very impressive results: s of June 2008, three yers sine the strt of its opertion, $380 million hd een disursed with n verge lon size of less thn $10,000 nd PAR for less thn 30 dys t less thn 1 perent. The progrm hd expnded to 15 regions nd 42 perent of the lons hd een disursed in rurl res. Produt profitility ssessments show tht most prtiipting nks roke even within two yers of strting the miro, smll, nd medium enterprise (MSME) lon usiness. The most profitle nk hd return on ssets of more thn 4 perent fter three yers of opertion. There re vrious inentives for nks to join the progrm. First, MSME lending is profitle nd the onsultnt working with the nk helps the nk to uild model tht n trk tht lending. Seond, lending to MSMEs llows the prtiipting nks to diversify their portfolio nd hene their risk. Third, the tehnil ssistne provided under the progrm fores nk mngement to think of their risk mngement prties for ll lending produts (lthough hnges re introdued only for MSMEs). Lstly, prtiipting nks mitigte finnil risks y diversifying their soures of funds with su-lons from the Chin Development Bnk. Key fetures ehind the suess of the produts inlude: (1) n effiient deision-mking proess, thnks to whih ustomer usully gets smll lon within three dys; (2) equl lon repyments required to e mde y ustomers every month, llowing the nk to losely monitor the risk espeilly in the erly dys; nd (3) flexile ollterl requirements mke miro lons more essile to smller lients. In some ses, ollterl requirements re wived ltogether. Soure: Disussion with Interntionle Projekt Consult (IPC) GmH, World Bnk stff, nd Chin Development Bnk. Pge 72

88 Bringing Finne to Pkistn s Poor Chpter 4 Most Pkistni nks do not hve speilized lon offiers, trining, nd proedures for SME lending. Additionl effiieny gins ould e hieved y mesuring this type of lending nd ligning stff rewrds to the qulity of the portfolio they mnge. Even though Pkistni nks hve sophistited mngement informtion systems, they do not urrently use them for monitoring the effiieny of individul stff memers nd lending deprtments, nd stff rewrds re usully unrelted to their performne. An exmple of suessful downsling progrm for nks is presented in Box Helping Bnks Serve SMEs To ridge the gp etween supply nd demnd, the GOP nd SBP ould ontinue to promote n enling environment tht filittes SME lending. This inludes, first nd foremost, reting omplete nd well-funtioning seured trnstions regime. Seurity interests over movle ssets should e esy nd llowed on most ssets nd y every entity (oth physil nd juridil persons). Priority rnkings should lso e lerly defined mong those who might hve lims on property offered s ollterl. The new seured trnstions regime should lso inlude ple (suh s registry) for mking priority interests pulily known, nd enforement of seurity interests for ll ssets should e fst nd hep. Seondly, to filitte the retion of redit history for SMEs, SBP should lso ensure tht the redit ureu ollets informtion from utility nd teleom ompnies. Thirdly, to filitte SME lending monitoring, SBP should minstrem the piloted reporting requirements nd simplify nk reporting for SME portfolios to inlude volumes nd numer of lons only in four su-rkets (Rs. <2M, Rs. 2-6M, Rs. 6-25M nd Rs M). The GOP should lso ontinue to promote inititives imed t proving demonstrtion effet, sine n enling environment is neessry ut not suffiient ondition for SME lending expnsion. These inlude ttrtion to the mrket of n institutionl investor tht hs trk reord in SME lending. This should idelly e hieved y selling the SME nk or giving ontrolling rights on its ord to n institutionl investor. To stimulte ompetition in the mrket from the very eginning, the GOP ould lso support long-term tehnil ssistne progrms for seleted nks. As illustrted y the Chin MSME lending progrm (see Box 4.4), for this progrm to e suessful, the GOP needs to ensure orret mrket inentives re in ple. Key fetures of progrms tht hve worked inlude: 1. Long-term tehnil ssistne is required to ensure tht the neessry sustntil hnges tke ple. Lending to smll usinesses requires profound hnges in the wy ommeril nks operte. With its high trnstion osts, smll usiness lending is only profitle if done in high numers with exellent portfolio qulity. It is therefore importnt to provide nks long-term support in this hllenging proess. Tehnil ssistne to improve lending tehnology should fous on: reduing trnstion osts for oth the nk nd the lient, inresing lon offier produtivity (in terms of numer of lons disursed), nd mintining high portfolio qulity. 2. Seletion of nk dvisors nd ontent of the tehnil ssistne re key to suess. The tehnil ssistne pkge should e omprehensive, euse the hnges tht re required re sustntil. In ddition to its sope, the suess of the tehnil ssistne progrm will depend upon three elements: seletion of ommitted nks, seletion of onsultny firms with strong reord in nking, nd lose monitoring y fully dedited nd experiened stff. The tehnil ssistne should inlude redesigning nk produts to meet lient needs, roust mngement informtion system, nd use of stff inentives linked to their performne. 3. A mix of ommitted nks should prtiipte in the progrm to rete ompetition mong them. Tehnil ssistne should e provided only to those nks tht re fully ommitted to SME lending. Serving suh enterprises requires hnge in the entire orporte ulture nd in the wy nks operte. Thus, only nks whose investors re willing to engge in suh sustntive trnsformtions should e Pge 73

89 Bringing Finne to Pkistn s Poor Chpter 4 offered tehnil ssistne. Often, nks with lrge nking network nd fous on retil lending hve omprtive dvntge in entering this mrket segment. Smller nks ould lso e trgeted, however, to t s tlysts. Mny ountries hve developed speilized SME progrms in stte-owned nks with lrge networks. However, suess usully tkes few yers, euse, s ws just noted, the hnges required re sustntil nd nks tke time in implementing them. It is therefore essentil tht tehnil ssistne e lso given to smll, fster-moving privte nks, whih trnsform t fster speed ut my reh stgntion due to their smller networks. 4. Performne greements for nks prtiipting in the progrm re key. Furthermore, performne greements should lso inorporte terms on the numer nd volume of lons disursed nd outstnding y given dte, s well s portfolio qulity inditors suh s keeping PAR over 30 dys to less thn 3 perent. The tehnil ssistne should lst t lest two yers nd performne greements should e monitored on monthly sis so tht timely remedil mesures ould e undertken when trgets re not met. In sum, while poliy efforts hve hd some suess, more work is needed, espeilly on nk downsling produts. Few SMEs in Pkistn orrow from nks, s ompred with SMEs in other ountries in the South Asi region inluding Indi nd Bngldesh. In spite of sustntil improvements of the legl nd regultory frmework, work remins to e done in the re of seured trnstions. Progress hs lso een sustntil on redit ureus, though more ould e done to filitte the retion of redit histories y SMEs. Aess to finne for smll nd miro enterprise is prtiulrly stlled. Yet those re the entities most likely to ontriute to employment nd poverty redution. Together with miro-lients, disussed in the previous hpter, nd remittne lients, the ojet of the following hpter, the underserved miro nd smll enterprises require urgent poliy mking to expnd ess to finne. Pge 74

90 Bringing Finne to Pkistn s Poor Chpter 5 5. Hrnessing Remittnes for Aess to Finne Remittnes to Pkistn re estimted t round $16 illion nd hve experiened onsiderle growth in reent yers. Interntionl flows (through oth forml nd informl hnnels) totl $9 illion, with domesti flows t pproximtely $6.95 illion. These remittne flows ply vlule role in supporting the eonomy y providing foreign exhnge nd improving the lne of pyments nd externl det position. They lso offer signifint potentil to support inomes of poor nd vulnerle groups. The lion s shre of remittnes is trnsferred y nks (80 perent), with further 17 perent ounted for y Exhnge Compnies (ECs) nd smll shre y the Pkistn Post Offie. Bnks operte fully utomted system tht runs seven dys week, hs exellent urn nd some good rurl overge, nd usully n deliver within 24 hours. The mjority of remittnes re reeived in less thn one week, lthough servie tends to e slower in rurl res. Pkistn Post hs lrge rurl network nd is the most ommon hnnel for domesti remittnes, ut servies remin reltively ineffiient. Home onsumption onstitutes the lrgest use of remittnes. SBP hs tken vrious mesures tht hve signifintly inresed remittnes through forml hnnels, though lrge shre of domesti remittnes ontinues to e trnsferred informlly. SBP hs een enourging the privte setor to provide moile nking solutions. The moile overge (out 90 perent of the popultion of Pkistn), nd the suess of moile money trnsfer solutions in other ountries, suggests tht moile phone nking offers signifint potentil to sle up ess to finnil servies in Pkistn. There is lredy lot of tivity in this setor nd oth nks nd moile phone ompnies re working to struture vile models Overview of the Remittne Mrkets 67 Reorded remittnes hve inresed fivefold sine This impressive growth hs een driven to some extent y the inresing efforts of SBP to ring remittne flows into the forml net. Ten yers go, out 15 perent of interntionl remittne flows me through forml hnnels, ompred with over 70 perent urrently. This hievement is impressive. Yet onsiderle shre of remittnes remins in the informl mrkets. There re no offiil dt of the overll size of the mrket. Vrious estimtes of interntionl informl remittnes put hwl flows t round the $2.5 illion mrk, in n overll interntionl flow of $9 illion. In ddition, household onsumption survey results indite evidene tht the domesti remittne mrket is, if nything, of omprle size to the interntionl one, ut might involve mny more trnstions of smller vlue. Unoffiil estimtes put the overll remittne mrket t over $16 million. To gin omprtive perspetive of the size of remittnes reltive to the eonomy, onsider foreign exhnge reserves, whih stood t $5.7 illion s of Septemer 2008, totl redit to the privte setor t $34.8 illion, nd stok mrket pitliztion t $44.7 illion. Remittne flows re potent lever for the Pkistn eonomy, nd signifint soure of foreign exhnge, s well s onsiderle help in improving Pkistn s externl det sitution, uilding the foreign exhnge reserves, strengthening the lne of pyments, nd preserving the vlue of the rupee. Moreover, remittnes n potentilly form lrge prt of inome for poor nd vulnerle groups (not yet the se in Pkistn), helping lift this popultion out of poverty. This hpter ttempt to understnd the struture, plyers, nd lients of the remittne mrket, inluding domestilly nd within the informl setor, in order to help o-opt more of the flows into the forml net, mke them essile to the poor nd underserved, nd hieve the ove-noted desirle gols. In solute terms, reorded interntionl remittnes surpssed $6.45 illion during fisl 2008, n verge nnul growth of 29 perent from 2001 (Figure 5.1). Worldwide, interntionl forml remittne flows to 67 The text does not tke into ount reent effets of the glol finnil risis, insted nlyzing the dt s of Reent events hve used remittnes to drop drstilly. Pge 75

91 Bringing Finne to Pkistn s Poor Chpter 5 developing ountries were $265 illion in 2007, nd unreorded flows were estimted t nother $80 to $185 illion (World Bnk 2008). Compred with other ountries in the region, Pkistn is t pr with Indonesi nd Bngldesh in solute terms, while Indi is fr hed with over $25 illion reeived in 2006, lthough the figures re omprle given the size of the eonomies (Figure 5.2). Remittnes in other South Asin ountries re even lrger in reltive terms. In Nepl, remittnes exeed export reeipts nd tourism, while in Sri Lnk, remittnes of $2.7 illion re greter thn te exports. In Bngldesh, remittnes re five times the level of foreign id. 68 Figure 5.1 : Interntionl Forml Remittnes Figure 5.2 : Remittnes : Interntionl Comprison Soures: SBP World Bnk Migrtion & Remittnes Ftook Mrket Plyers The min plyers in the forml remittne mrket re nks, ECs, nd Pkistn Post, while informl hnnels onsist of hwl or money rried y hnd. Exept for ECs, whih y lw nnot operte domesti money trnsfers, ll plyers provide domesti nd interntionl remittne servies. Tle 5.1: Snpshot of the Min Chnnels of Interntionl Forml Money Trnsfer Remittnes (Billion $) Mrket Shre (%) Brnhes Cost Bnks (Overll) Hi Bnk Free United Bnk Free EC's (Overll) > 1000 Other Servies Provided Vrious Bnking Produts Vrious Bnking Produts Zro Exhnge $5-$20 Foreign Exhnge Prtnership greements Al-Rjhi Bnk, Trvelex Western Union, Severl Others Western Union, Ri, Vigo nd others Wll Street Exhnge $5-$20 Foreign Exhnge Moneygrm, Instnt Csh nd others Pkistn Post $5-$20 Mil, Svings, Insurne nd others Western Union Soure: SBP 2007 The nks ount for 80 perent of interntionl remittnes oming through forml hnnels into Pkistn, given their long history in the setor (Tle 5.1). The formtion of exhnge ompnies in 2002 hs signifintly inresed flows s hundreds of smll moneyhngers hve een rought into the regulted setor nd now ount for 17 perent of interntionl forml remittnes. Both nks nd ECs hve multiple 68 Soure: Dwn Eonomi nd Business Review, Mrh 31 st, Krhi, Pkistn. Pge 76

92 Bringing Finne to Pkistn s Poor Chpter 5 orrespondents nd reltionships rod. The rekdown of interntionl forml remittnes y setor (detiled time-series) is ville in Annex Tle A5.1. Pkistn Post ounts for lmost 2.5 perent of the interntionl remittnes reeived, exlusively vi n greement with Western Union. Money trnsfer orgniztions (MTOs) (minly Western Union), hve n estimted shre of out perent in intermediting interntionl trnsfers. Setions disuss trnsfer servie providers in the forml interntionl setor. Domesti remittnes re dominted to lrger extent y informl flows, reltive to interntionl trnsfers. We estimte domesti remittne flows t $6.95 illion, sed on household survey estimtes of the shre of foreign nd domesti remittnes in monthly household inome s per the ltest Household Inome Integrted Survey ( ). The min forml hnnels for domesti remittnes re the Post Offie nd internk trnsfers. Pkistn Post is the min plyer in trnsferring domesti remittnes to rurl res, though the five ig nks lso hve presene in rurl res. Mirofinne entities still ply only minor role in remittnes. Setion overviews the limited dt ville on (mostly forml) domesti trnsfer servies. Disussions with vrious stkeholders suggest tht the estimted size of the hwl mrket is $2 to $4 illion. The informl mrket is disussed in more detil in setion Bnks The min plyers in the nking setor re the five ig nks of Pkistn: Hi Bnk Limited (HBL), United Bnk Limited (UBL), Ntionl Bnk of Pkistn (NBP), Muslim Commeril Bnk, nd Allied Bnk Limited, whih lso hve n interntionl presene. They hve een given speil inentives y the SBP for fst nd seure delivery of home remittnes to reipients in Pkistn. These nks hve speilized home remittne ells t their respetive hed offies nd ollorte to speed up remittne trnsfers mong their rnhes. Furthermore, they do not hrge interntionl ustomers for remitting money through their network, s SBP reimurses them. 69 This step y SBP helped to signifintly oost remittnes oming through nks. Initilly, in 2002, when reforms were eing tken to regulte the remittne mrket, these nks were given the inentive of getting R 1 for eh dollr tht ws remitted through their network. However, this poliy ws dropped nd repled with fee reimursement. Over the lst yer, interntionl forml remittnes oming through nks inresed y 13.4 perent. An nlysis of ntionl household survey responses shows positive pereption of nk remittne servies y domesti nd interntionl remittne users. Bnk rnhes using one s own ount re found the most essile remittne hnnel ntionwide (y 29.5 perent of the popultion), mong forml nd informl hnnels like. Bnk remittne servies using one s own ount re lso onsidered the lest-risky hnnel (y 38.7 perent of the popultion), while nk remittnes vi nother person s nk ount, re pereived s the riskiest money-trnsfer method (y 47.9 perent of the popultion). Interestingly, rurl res onsider nks, ut not the Pkistn Post, risky, nd urn res exhiit extly the opposite pereptions. Hi Bnk Limited The lrgest plyer in the nking setor is HBL. It ounts for $1.2 illion in interntionl forml remittnes for fisl 2007 or mrket shre of 22 perent, overll. It hs 40 interntionl rnhes with opertions in 25 ountries nd is lso the lrgest privte setor nk in Pkistn with over 1,400 domesti rnhes (Box 5.1). HBL lso hs lrge network in rurl res, with lmost 50 perent of its rnhes loted there. Hene, HBL s outreh in the rurl res is impressive nd llows it to reh reipients in fr- 69 SBP hs given ll nks the inentive to not hrge interntionl ustomers for wiring money, nd it reimurses them for eh trnstion. Pge 77

93 Bringing Finne to Pkistn s Poor Chpter 5 flung res quikly. Aout 30 perent of foreign remittnes oming through HBL go to rurl res. The medin trnstion mounts for remittnes reeived through HBL re $500 to $700. Box 5.1: A Closer Look t HBL HBL hs hieved its mrket-leder position through vigorous mrketing mpign in the Middle Est. Energeti junior stff sttioned in these ountries promote HBL servies t smll exhnge ompnies, nks, nd worker mps. HBL hs tried to mke the proess s onvenient nd simple s possile for Pkistni workers in the UAE, so tht they use forml hnnels insted of hwl. Stff in the Middle Est visit worker mps on pydy nd hold speil remittne mpigns during the Eid (fest) festivl. For regulr remitters, HBL hs reted speil ounts with options to send money to five different people in Pkistn. One the ount is set, the ddress nd nk detils of eh reipient is entered into the system. The ustomer only hs to ll nd give his ount numer nd the nme of the person he wnts to send money to, nd the nk proesses the trnstion. HBL hs ssoitions with nks nd ECs in mny ountries to filitte remittnes. One of its mjor prtners is Al-Rjhi Bnk in Sudi Ari, through whih remittnes, free of ost, n e hnneled to Pkistn. HBL lso hs n greement with Trvelex, n MTO sed in the United Kingdom. Overses pyments re reeived y HBL Pkistn through SWIFT. One in Pkistn, HBL hs fully utomted system tht llows the nk to redit reipient ounts every hour or issue drft if the reipient hs n ount in nother nk. HBL s unit deling with remittnes opertes seven dys week nd thus remittnes n reh reipients in most ples in Pkistn within 24 hours. HBL lls this servie Fst Trnsfer nd relies on tehnology ought from Misys Interntionl. Other HBL produts inlude fx money orders nd express hek remittnes. On the reipient side, HBL offers speil remittne Munf Plus Certifites ville in one-, three-, nd five-yer terms, whih hve returns rnging from 7 to 11 perent per yer. Soure: DFID United Bnk Limited The seond lrgest plyer in the mrket is UBL, whih rought in pproximtely $550 million in interntionl forml remittnes for the yer 2006, giving them n overll mrket shre of 10 perent. They hve network of over 1,000 domesti rnhes nd 15 overses rnhes, with strong presene in the Gulf region. UBL lso hs mny orrespondent exhnge ompnies nd nks overses nd is onstntly striving to form reltionships with new ones. Until July 2007, UBL hd ontrt with Moneygrm, wherey money sent through Moneygrm frnhises overses ould e piked up t UBL rnhes. However, sine August 2007, UBL hs een operting del with Western Union, nd within the first few months, lims to e hndling volumes s high s those they hndled with Moneygrm t pek times. UBL estimtes tht out 40 to 50 perent of remittnes reeived in their network re for rurl res (Box 5.2). Box 5.2 Innovtive UBL Produts UBL hs severl remittne produts other thn the simple wire trnsfers from interntionl rnhes. In 2002, it lunhed UBL Tezrftr ( servie similr to HBL s Fst Trnsfer), whih offers free remittne servie within the ountry or from rod to the doorstep of the reipient. This servie treled UBL s remittne usiness volume s it ensures free delivery of money within 24 hours for most urn res. It is open to ll, even those who re not UBL ount holders, nd the remittne is delivered t the doorstep of the reipients y n uthorized ourier or redited to their ounts. At the sme time, in the UAE, UBL lunhed sheme, whih, long with hrging no fee for the money trnsfer, lso gve exptrites onus of 10 points redeemle for 10 dirhms on their next trnsfer, if they remitted t lest Rs 7,000 ($114). UBL lso hs online money trnsfers from the United Sttes in ollortion with Pyquik. With this servie remitters n simply send money from redit rd or Internet hek to nyone in Pkistn. The money n e redited to n ount in ny nk in Pkistn or delivered to the reipient t his or her home. The fee for sending money through n Internet hek is only $5; with the redit rd, flt fee of $10 is hrged, plus 2.89 perent of the trnstion mount. The verge trnstion mount of this online money trnsfer lled lik n remit is $750. The reson for the high fee on the redit rd is tht Vis hrges UBL this mount to proess the trnstion. Soure: DFID Pge 78

94 Bringing Finne to Pkistn s Poor Chpter Exhnge Compnies The EC frmework ws enfored in 2002 to ring exhnge nd remittne usinesses under proper finnil disipline. With the trnsformtion of uthorized moneyhngers into forml ECs, lrge numer of undoumented nd sh interntionl trnstions me within the forml net. These trnstions re required to hve proper doumenttion, reord keeping, nd dherene to interntionlly epted Know Your Customer norms. All EC trnstions, exlusively interntionl money trnsfers, re fully utomted nd hve rel-time online network tht n e essed y ny rnh to trk nd mke pyments. Mny ECs hve set up ll enters where the frnhisees n ll nd hek trnstion sttus. The EC frmework lso enled SBP to sustntilly nrrow the gp etween internk nd ur mrket exhnge rtes, whih hs plyed n importnt role in improving the inflow of interntionl remittnes during the pst two to three yers. The lne of pyments for fisl 2005 for the first time inluded the reeipts nd pyments mde through the ECs nd onveys more omprehensive nd relisti piture of the trnstions on the externl ount of the ountry. The trnsformtion of uthorized moneyhngers into forml ECs, however, hs posed onsiderle hllenges for SBP in ensuring effetive regultory nd supervisory oversight, inulting orporte ulture, good governne, trnspreny, nd proper doumenttion nd reord keeping struture. The network of ECs is omprehensive, overing 100 perent of the urn mrket nd slowly mking inrods into rurl res. 70 Interntionl remittnes oming through ECs grew y 53.4 perent from fisl 2006 to fisl Aording to ECs, 90 perent of remittne trnstions they hndle re etween Rs 10,000 nd Rs 100,000. Zro Exhnge Zro Exhnge is the top EC in the remittnes mrket nd hs prtnered with Western Union (Box 5.3). Aout $260 million were hnneled through it lst yer. Zro Exhnge hs reeived numerous wrds in the lst ouple of yers, suh s Role Model EC from SBP nd the Federtion of Pkistn Chmer of Commere nd Industry, wrded y the Prime Minister of Pkistn, nd the Best Computer Sienes Corportion Mngement wrd from Western Union. Box 5.3 Zro Exhnge: Brnhing Out, Inluding Online Zro Exhnge hs lrge network of over 660 pyment lotions. It hs n ggressive pproh nd ssoitions with 15 interntionl MTOs suh s Ri nd Vigo, in ddition to Western Union. It hs lso opened its own Zro Exhnge rnhes overses in res densely populted with Pkistnis. In Pkistn, it hs prtnered with mny orgniztions to ensure prompt delivery of remittnes. Its prtners inlude Bnk of Punj, whih hs pproximtely 300 rnhes; Utility Stores, of whih 100 re involved in pying out remittnes; nd Punj Coopertive Bnk. In ddition, Zro Exhnge hs its own rnhes nd pyment ooths ll over Pkistn. It does not hve pyment ooths in rurl res, like most MTOs, s its systems re utomted nd online nd so the minimum requirement of setting up rnh or pyment ooth is phone nd Internet onnetion. However, it hs strtegilly set up rnhes in towns ner villges tht require remittne servies. For instne, it hs seven rnhes in Jhelum ( smll ity) s mny people from the surrounding villges hve migrted to Itly for work. Other thn the money-in-minutes servie offered in ollortion with Western Union, Zro Exhnge lso offers n online trnsfer servie lik n send. This servie tkes 2-3 dys for pyment proessing, nd money n e pid through redit rds or Internet heks. The trnstion n e trked nytime from the We nd n e pid out s sh, nk drft, nk deposit, or home delivery. The fee for pying with n Internet hek is $3 per $300, while with redit rds it is flt hrge of $8 plus 2.9 perent of the trnstion mount. Soure: DFID 70 Informtion on individul rnhes n e found on the wesite of eh EC. Pge 79

95 Bringing Finne to Pkistn s Poor Chpter 5 Wll Street Exhnge The min prtner of Moneygrm Interntionl in Pkistn is Wll Street Exhnge, sed in Krhi. Wll Street hs 7 8 other interntionl orrespondents, suh s U.K.-sed Instnt Csh. Wll Street Exhnge hs lose to 90 rnhes nd frnhise offies in 50 ities ross Pkistn, nd it plns to open more pyment lotions in the oming yer. In 2007, interntionl remittnes of $217 million me through the Wll Street Exhnge network, giving them mrket shre of 23 perent in the EC segment. The min produt offered y Wll Street Exhnge is the money in minutes servie for oth inwrd nd outwrd remittnes. It is fst nd onvenient servie, linked through the Internet nd ll enter with toll-free numer through whih reipients n trk their pyments. Similr to Zro Exhnge, the si requirements of Wll Street for opening pyment ooth re Internet nd phone onnetion. However, if Internet infrstruture is not ville, dil-up onnetions on wireless phones is used. With the wireless network, Wll Street plns to improve its outreh in smll ities nd rurl res Pkistn Post Pkistn Post hs network of 13,419 post offies ross Pkistn. Reently it strted utomtion of its servies nd hs lunhed n online Express mil trk nd tre system. In the future, Pkistn Post plns to offer n Instnt Money Trnsfer servie, olletion of utility ills from home, premier ourier servie, nd Tele Centers up to the tehsil hedqurter level. Aprt from eing mjor hnnel through whih domesti remittnes re sent to rurl res, it lso hs n greement with Western Union to disurse foreign remittnes through 1,700 lotions. For fisl 2007, interntionl remittnes through Pkistn Post were $135 million, or mrket shre of 2.5 perent. Growth in the preeding yer ws remrkle 58.9 perent, up from $85 million interntionl remittnes for fisl Out of the 1,700 lotions used to disurse remittnes from Western Union, out third re fully utomted with rel-time online system. Aout third hve fx filities, where one the remittne notifition is reeived, fx is sent to notify disursl. Aout one-third of lotions telephone into ll enter to hek the remittne sttus nd rry out the disursl. Aout 85 perent of the post offie rnhes re in rurl res. Tle 5.2 shows the distriution of post offies y provine nd urn/rurl sttus (Pkistn Post 2005). Only the 80 mjor generl post offies hve een utomted. All the trnstions routed to the other 13,000 post offies go through the generl post offies, nd so their reords re omputerized even though the post offies themselves re not utomted. Tle 5.2 Perentge Post Offies y Provine nd Rurl/Urn Ares Soure: Pkistn Post Annul Report The remittne produts offered re money orders, postl orders, postl drft, urgent money order, nd fx money order. The min produt used for domesti remittnes to rurl res is the money order, whih is Pge 80

96 Bringing Finne to Pkistn s Poor Chpter 5 issued up to mximum vlue of Rs 10,000. For higher denomintions, up to Rs 200,000, postl drfts re issued. Urgent nd fx money orders trnsfer money the sme dy, ut these servies re only ville in limited numer of ities (204 ities re served y urgent money orders nd 49 ities y fx money orders). The hrges for these servies re nominl. A 5 perent ommission is hrged for the regulr money orders; urgent nd fx trnsfers hve n dditionl flt hrge of Rs 30 pplied on top of the 5 perent ommission. Post offie servies re pereived s the lest-expensive forml method of remitting funds, ording to household survey results. Informl methods remin less expensive, with 25.6 perent of the popultion iting money sent vi fmily or friends s the lest ostly, 21.5 perent iting hek with fmily or friends, nd 15.2 perent iting hek tht is hnd-rried. This ontrsts with pereptions of the high ost of other forml remittne hnnels. The ECs re onsidered the most expensive (y 40.7 perent ntionwide). Interestingly, over hlf of rurl respondents omplined out the ost of post offie trnsfers. The pperwork for sending the money orders is strightforwrd nd simply requires the ddresses of the sender nd the reipient. In ities, the money order is delivered in 1 2 dys; however, in rurl res it tkes 3 5 dys offiilly. In prtie the money orders might tke longer, nd nedotl evidene suggests tht in very smll villges, postmen might hold on to the money for personl use nd deliver it when they feel like it. Offiilly, the sender is supposed to reeive signed reeipt from the reipient one the money order is delivered, ut in relity it is generlly never delivered. Evidene from the ntionl household survey lso points to the reltively slow servie of post offie money orders ompred with other domesti nd interntionl remittnes servies. Informl remittne hnnels sore just s well s forml ones on speed, with 23.6 perent of the popultion iting money sent with fmily or friends s the fstest trnsfer, while 21.4 perent nd 20.3 perent point to ECs (Western Union) nd eletroni nk trnsfers, respetively. The post offie is onsidered the slowest delivery hnnel y more thn hlf of respondents (51.8 perent) Money Trnsfer Orgniztions The worldwide mrket shre of Western Union in the intermedition of remittnes is 15 perent, nd in Pkistn it proly hs shre of perent. Its sh-to-sh nywhere in the world model is very onvenient, s is rel-time servie. It hs prtnerships with severl orgniztions in Pkistn, inluding Pkistn Post, UBL, NBP, Fysl Bnk, nd Zro Exhnge, whih in turn hs its own prtnerships with Bnk of Punj, Utility Stores, nd Punj Coopertive Bnk. These ssoitions trnslte into more thn 5,000 pyment lotions for Western Union ross Pkistn. Western Union hs n offie in Islmd with full tem tht hndles opertions in oth Pkistn nd Afghnistn. The ompny does not offer online remittne servies to Pkistn from the United Sttes (USA) like it does to Mexio nd Brzil. The worldwide mrket shre of Moneygrm is 7 8 perent, though it does not hve signifint presene in Pkistn. It hs prtnered with few ompnies, of whih only two re well known Wll Street Exhnge nd Bnk Al-Flh. Moneygrm hndles usiness in Pkistn vi its Dui offie. This explins the lk of mrketing nd n ggressive expnsion strtegy in Pkistn. Overll, sending money through MTOs is very onvenient s they re well loted interntionlly, nd pik up of funds within Pkistn is esy. Furthermore, s seen elow in setion 5.5, the servie offered y MTOs is very resonly pried. Pge 81

97 Bringing Finne to Pkistn s Poor Chpter Domesti Remittnes Dt on domesti remittnes re muh less redily ville thn omprle figures on interntionl trnsfers. This shortoming preludes poliy tion towrds enourging n effiient nd essile setor, in the mnner in whih interntionl remittne servies hve developed thnks to SBP efforts. We estimte domesti trnsfer volume t $6.95 million, roughly 90 perent of interntionl remittne flows in Pkistn. While domesti remittnes used to represent higher shre thn foreign remittnes in monthly household inome in Pkistn, ntionl household inome nd onsumption expenditure dt for fisl 2006 showed reversed trend, most proly due to the fster formliztion of foreign remittne flows. Rurl res ontinue to rely slightly more on domesti remittnes thn on interntionl trnsfers (Federl Bureu of Sttistis [FBS] 2006: Tle 11). However, ording to the ntionl household survey onduted for this report tht orroortes Pkistn Soil nd Living Stndrds Mesurement Survey (PSLM) dt, domesti remittnes re more frequently used y Pkistnis. In ft, there re 2.5 times more domesti remittne users mong Pkistni households thn foreign remittnes users. The regionl distriution of remittnes is disussed in setion 5.3. Figure 5.3 presents the most frequently used domesti remittne hnnels, sed on results from the Pkistn Aess to Finne (A2F) household survey. The populrity of the post offie for domesti remittnes is onfirmed y ntionl household survey results, whih ple the post offie s the most ommonly ited hnnel for domesti remittnes, espeilly in rurl res, more thn money or in kind through friends of fmily, or nk servies (vi rnh or eletronilly). Figure 5.3: Top Domesti Remittne Chnnels 30.0% 20.0% 10.0% 0.0% At Post Offie Money w ith friend/fmily memer At nk rnh using ow n ount Eletroni nk trnsfer Soure: A2F Survey The min forml hnnels for domesti remittnes re internk trnsfers nd the post offie. Informl hnnels inlude hnd-rrying the sh, or sking reltive or friend to do so. Pkistn Post trnsferred $297 million through money orders nd nother $0.32 million through postl drfts in fisl Aout $56 million were lso trnsferred through urgent nd fx money orders ut they were purely urn trnsfers euse these servies re not offered in rurl res. Annex Tles A5.2 nd A5.3 present rekdown of money order nd postl drft trnsfers y provine. Tle 5.3: Proportion of Rurl vs Urn Bnk Brnhes The nks with rurl presene in Pkistn re the five ig nks: HBL, NBP, UBL, MCB, nd Allied Bnk Limited. MCB nd HBL hve the est rurl rnh networks. ZTBL is nother nk with signifint rurl presene sine its im is to provide finnil servies to rurl lients nd in griulturl development. The 71 PSLM is ntionlly representtive household survey rried out y the FBS. Pge 82

98 Bringing Finne to Pkistn s Poor Chpter 5 proportion of rurl versus urn rnhes, ording to SBP, is s given in Tle 5.3. The provinil rurl split of nks is given in Annex Tle A5.4. Within the nking industry, the five ig nks re primrily responsile for rurl-urn trnsfers. These nks re moving from pper- to e-nking, nd this trend hs onsiderly shortened the time it tkes to trnsfer money to rel-time trnsfers. The totl proportion of rnhes tht re online ws 61 perent y the seond qurter of fisl Some nks offer free online trnsfers, while others hrge up to Rs 200 per trnstion. SBP hs llowed Mirofinne Bnks (MFBs) to engge in remittne servies s well; however, First Mirofinne Bnk Limited is the only MFB urrently offering this servie. It lunhed this remittne servie in 2003 nd, in the 12 months ending My 2005, it inresed the monthly volume of trnsfers from $3.46 million to $7.5 million. The numer of trnstions inresed from round 3,000 in My 2004 to round 6,500 in My The nk s typil trnsfer ustomers re low-inome trders, ut they get wlkin ustomers s well. The fee hrged is etween nd 0.2 perent, depending on the trnsfer mount (DFID [U.K. Deprtment for Interntionl Development] nd USAID [U.S. Ageny for Interntionl Development] 2005). The medin trnstion mount for this servie is $1, Informl Chnnels In Pkistn, s ross the world, money hs een trnsferred through n unoffiil system known s hwl, whih is sed on trust. No reeipts re given, ut the trnsfer is hep, fst, nd seure. The reson is tht hwldrs (people who do hwl) hve very low overheds, they n offer etter exhnge rte thn tht offered y nks, nd there is no red tpe. The servie is generlly relile, onvenient, nd offers nonymity. World Bnk (2008) notes tht informl remittne osts tend to e muh lower thn forml hrges (out three times lower in Bngldesh, for exmple). Hwl fees re only perent of the trnstion mount, while in forml hnnels they my e s high s 20 perent (Psss, 2004). Furthermore, the fees re quite flexile nd depend on the reltions etween the hwldr nd the lient, the ultimte destintion, nd the frequeny nd regulrity of pyments. Servie to remote res is esier nd fster nd is ommonly delivered to the doorstep, whih is very importnt in Pkistn, where women in rurl res re not very moile due to ulturl nd soil ftors. The min reson hwl is populr in Pkistn is euse of the lk of nking filities in rurl res, s well s due to the doorstep delivery, so tht the women or older people do not hve to trvel long distnes to reeive pyments. The mrket size estimtes of the informl remittne setor from our disussions re round $2 to $4 illion, inluding hwl nd self-rried money. Besides hwl-type systems, informl hnnels n tke vriety of forms: ethni stores, trvel genies, moneyhngers, ounter servies, or hnd delivery. Another informl remittne hnnel is in-kind remittnes, whih is the sending or rrying of remittnes in the form of goods for personl use or for resle in the informl mrket. Kzi (1989) notes tht this prtie is quite widespred he estimtes in-kind remittnes t out 16 perent for urn migrnts nd 11 perent for those returning to rurl res. In-kind remittnes hve eome more signifint sine introdution of lierl import poliies llowing duty-free imports. However, goods rought in under personl ggge re not reorded s remittnes or imports. The estimted shre of in-kind remittnes is round perent of overll remittnes (Kzi 1989). Pge 83

99 Bringing Finne to Pkistn s Poor Chpter Distriution of Remittnes y Region nd Inome Level Interntionl Figure 5.4 Remittnes y Country Figure 5.4 displys the distriution of interntionl remittnes y soure ountry. The distriution of remittnes from ountries rod is given in Figure 5.4. The lrgest soure ountry for remittnes to Pkistn is the USA, with Sudi Ari in seond ple, followed y the United Ar Emirtes (UAE) nd other Gulf ountries, nd the United Kingdom in fifth ple. From the UAE, the lrgest ontriutor is Dui, nd from the Gulf ountries it is Kuwit. From the Europen Union the min ontriutor is Germny. The ountries tht hve experiened the gretest growth in remittnes through the Pkistn Soure: SBP Annul Report orridor sine 2001 re the USA, Cnd, nd Itly, proly euse of n inresed effort y these ountries to rk down on the terrorist network, enourging senders to go through forml hnnels (see detiled rekdown y ountry of origin in Annex Tle A5.5). Historilly, the ulk of remittnes hve ome from the Gulf region, s it is loser nd esier to migrte to. Only fter the events of Septemer 11, 2001, did the USA eme mjor remittne-sending ountry. However, if informl flows re tken into ount, the remittnes oming from Gulf ountries would fr exeed those oming from the USA Ntionl The two provines with signifint shre of remittnes re North West Frontier Provine (NWFP) nd Punj. Together they ount for over 85 perent of oth domesti nd foreign remittnes. Furthermore, the rurl res in these provines reeive the ulk of the flows. Tle 5.4 shows the overll distriution in the provines nd gives the split y rurl/urn re. Tle 5.4: Regionl Distriution of Remittnes After Punj nd NWFP, Azd Jmmu nd Kshmir (AJK) reeives 6.2 perent of foreign remittnes, with 4.5 perent going to rurl res. With over 60 perent shre of the foreign remittnes for rurl res of Punj nd NWFP, it is ler tht remittnes re not only going to mjor ities ut re lso ville in villges, where trditionlly forml ess hs een limited. Sindh nd Bluhistn reeive only trivil mount of remittnes. It is interesting to note tht only 17 perent of the popultion of Pkistn resides in NWFP, ut it reeives the lrgest proportion of foreign remittnes, 47.5 perent. Divergent qulity nd essiility of servies ntionlly re refleted in differenes in provinil pereptions on vrious hnnels of remittne trnsfers. In NWFP, money in person is risky remittne method (pereived s suh y 57.5 perent of the NWFP dult popultion); in AJK, money trnsfers vi friends is risky (38 perent); in Bluhistn, using nk trnsfers is pereived s unsfe (38 perent), nd in Sindh nd Punj, using someone else s nk ount is onsidered riskiest y 25 perent nd 50 perent, respetively. Pge 84

100 Bringing Finne to Pkistn s Poor Chpter 5 The pttern of regionl remittne distriution follows the migrtion pttern s noted in Ar (2003). Using the 1998 ensus dt, she writes tht Punj nd Sindh were the min destintions for domesti migrtion. For Punj, muh of the urn immigrtion ws from rurl or other urn res of Punj. In ontrst, urn immigrnts in Sindh (predominntly in Krhi) were mostly from other provines of Pkistn. There ws very little emigrtion from Sindh to other provines. NWFP, on the other hnd, ws mjor soure of migrnts to ll provines, prtiulrly to Sindh. Gzdr (2003) notes tht the pttern of migrtion in Pkistn from lor-undnt res of NWFP nd Punj to urn enters of Punj nd Sindh onforms to si poverty-migrtion linkge. The regions of emigrtion re, indeed, those prts of the ountry where griulturl inomes re low nd voltile. And the eonomi gp etween res of emigrtion nd immigrtion is ler enough. Wges of sul dily lorers n rnge from round Rs 40 in southern Punj to round Rs 150 in Krhi Poverty nd Remittnes It is well reognized tht remittnes ply enefiil role in reduing poverty. Remittnes help fmilies inrese spending on si needs, suh s food nd lothing, s well s helth nd edution. Certin studies suggest tht out 10 to 20 perent is lso put in svings nd investment (IFAD 2006). 72 Adms nd Pge (2005) find tht 10 perent inrese in remittnes leds to 3.5 perent deline in the shre of people living in poverty. Yng nd Mrtinez (2005) determine tht remittnes not only led to derese in poverty of migrnt fmilies, ut lso hve spillover effets on nonmigrnt fmilies. Totl remittne inflows to Pkistn from fisl 2002 through fisl 2006 hve mounted to over $19 illion. Aout one-third of this signifint inflow of money is towrds rurl res, helping to relx udget onstrints nd llowing inresed expenditure on helth, edution, nd onsumer durles, ontriuting to poverty llevition. However, remittnes do not pper to mke signifint shre of household inome for the poorest quintile (Tle 5.6). Only for NWFP do remittnes form signifint shre of inome over 12 perent in oth rurl nd urn res. For Sindh nd Bluhistn, the shre of remittnes in household inome is less thn 3 perent. In Punj, for urn poor the shre is just 3.6 perent nd for rurl poor it is 8.8 perent. NWFP reeived signifint remittne flows fter the Otoer 2005 erthquke. Tle 5.5: Shre of Remittnes in Household Inome for the Poorest Quintile (%) Soure: Household Integrted Eonomi Survey 2005, Federl Bureu of Sttistis, Government of Pkistn Interestingly, we find tht remittnes form lrger shre of household inome for the rihest two quintiles (Annex Tle A5.6). For the rihest quintile, on verge, foreign remittnes re 5.5 perent of rurl household inome while for the poorest quintile, foreign remittnes re only 1.9 perent of rurl household inome. However, ross ll quintiles, domesti remittnes re the sme proportion of household inome, round 3.5 perent. In onlusion, foreign remittnes re muh more signifint for the well-off households. 72 World Bnk (2008) quotes estimtes in Ghn, where 70 perent of remittnes re spent on urrent expenditures nd less thn 30 perent on investments suh s lnd, ttle, or onstrution. In Mli, 80 to 90 perent is spent on onsumption. In the Philippines, 68 perent is spent on det repyment nd urrent needs, 13 perent on onsumer goods, 1 perent on edution, 5 perent on usiness pitl, 3 perent on lnd nd uildings, nd 1 perent on personl svings. Pge 85

101 Bringing Finne to Pkistn s Poor Chpter 5 In ontrst with the four provines in Tle 5.6, in AJK, 25 perent of the households re found to reeive remittnes, whih re their primry soure of inome (Suleri nd Svge 2006). Hwl is the most ommon mode of money trnsfer tht is used. The study lso found tht households whose livelihood inluded interntionl remittnes ppered less vulnerle to the erthquke s effets nd lso showed more resiliene fter remittnes were re-estlished. A study undertken y Od (2007) in the Chkwl distrit of Punj finds tht externl migrtion helps in poverty llevition when ompred with internl migrtion, s high rtio of poor were found in the group of internl migrnts, nd low inidene of poverty mong externl migrnts. 73 Furthermore, remittnes from externl migrnts were twie s high s those from internl migrnts in solute terms. The results of this study re orroorted y study in Ghn y Adms (2006) Determinnts of Remittnes nd Historil Trends There re roughly 3.7 to 4 million Pkistni workers overses, nd over 90 perent of them re prodution workers, like lorers, or servie workers suh s ooks nd opertors (Figure 5.5). Pkistn hs higher perentge of low-skilled migrnt lor when ompred with ountries like Indi nd Philippines, whih hve perent professionl migrnts, ut similr to the levels in Bngldesh or Afrin ountries. Figure 5.5 Overses Workers By Ctegory The min determining ftors of remittnes inlude eonomi tivity in the sender nd reipient ountries. Swmy (1981) finds tht 70 Soure: Bureu of Emigrtion & Overses Employment perent of vrition in remittnes is explined y eonomi tivity in the sending ountry. Until reently, strong growth in Pkistni equity mrkets, s well s signifint eonomi growth t 7 perent yer, hd ttrted signifint investment from overses. Remittnes re lso determined y strong lol needs suh s nturl dissters. Western Union s remittnes to AJK inresed y 400 perent fter the Otoer 2005 erthquke (Shh et l., unpulished). Tle 5.7 displys the nnul trend of workers remittnes from 1997 to Workers remittnes fell to $983 million in fisl 2000 from muh higher levels in the preeding dede. The min reson for the deline in remittnes through the nking setor ws tht people strted to prefer the hwl hnnel for its effiieny, low ost, nd outreh. Furthermore, following the nuler testing in My 1998 the foreign urreny ounts of residents nd exptrites were frozen, whih reted unertinty nd severe loss of onfidene in the government. Remittnes fell y lmost 29 perent in fisl 1999 nd further delined 7.2 perent the following yer. Tle 5.6 Remittne Growth Soure: SBP (2007); 73 See Annex Tle A5.7 for the ontriution of remittnes to household inome nd poverty rtio in the smple from the study. Pge 86

102 Bringing Finne to Pkistn s Poor Chpter 5 In the 2000s, SBP hs een trying to ur money lundering nd informl mens of money trnsfer, whih my hve used reversion to forml nking hnnels nd MTOs. The min resons for remittnes oming through the forml hnnels inlude the onversion of the moneyhngers to ECs, whih were in the forml net, the redued exhnge rte spred etween the nk nd ur rtes, nd inentives given y the Government of Pkistn (GOP) nd SBP to nks nd exptrites. All these efforts supported 400 perent inrese within six yers. Remittnes ontinued their signifint growth in fisl 2002 nd 2003, y nd 77.3 perent, respetively. In fisl 2004, the remittnes deresed y 8.6 perent, the deline oming minly from the UAE for the two following resons: (1) the Hjj Sponsorship Sheme 74 ws nelled, whih explined hlf the derese, nd (2) investment in rel estte ws opened up to foreigners in Dui, giving Pkistni exptrites n lternte ttrtive investment option. In fisl , remittnes inresed y 7.7, 10.3, 19.4, nd 14.9 perent, respetively, mostly euse of inresed eonomi tivity in the Gulf ountries ( result of higher oil pries) nd in the USA. The erthquke disster prompted mny Pkistnis overses to send inresed finnil support to reltives k home. The growth is further ttriutle to inresed investment in the Pkistni sset mrkets, prtiulrly the equity mrket in fisl The inrese from the UAE for fisl 2007 ws $146.5 million; this n e ttriuted to rising onstrution tivity during the rel estte oom nd inresed eonomi tivity s result of the oil oom (SBP 2007). Eonometri nlysis suggests tht edution, employed sttus, nd regulr ell phone use re individullevel demnd determinnts for sending nd reeiving remittnes (Annex Tle A5.8). Interestingly, women, rurl res, griulturl workers, household memers other thn the hed, nd prtiipnts in the informl setor re not less likely to e remittne lients. Higher inome levels lso hve higher hne of eing ssoited with ess to remittnes thn the poorest quintile, onfirming the poverty impt nlysis in setion This ould reflet the ft tht the poorest households my hve less opportunity to send fmily memer rod nd reeive foreign remittnes Usge Ptterns nd Servie Fees The ntionl household survey highlights reltively short delivery times for remittnes in Pkistn: it tkes week or less for the mjority of respondents to get sh remittnes within nd outside Pkistn. Only fifth of reipients of interntionl remittnes report delivery delys of more thn week. This is more pronouned in rurl res, where qurter respondents note delys of over week. The typil frequeny for reeiving domesti or interntionl remittnes is monthly or more frequently. The ntionl survey lso revels tht home onsumption onstitutes the lrgest use of remittnes ross for ll inome groups, rurl/urn res, nd gender. Other top uses of remittnes fter home onsumption inlude helth nd hild re, edution, nd usiness/lnd purhse relted-purposes. The limited presene nd outreh of Pkistni nks rod is the min prolem fing Pkistni nks in inresing remittnes through their networks. For senders, the most esily essile nks re their lol nks, however, sending money through them tkes time nd is expensive. Tle 5.8 gives n overview of hrges from four ities tht hve signifint shre in remittnes to Pkistn. 75 The lol nk hrges vry onsiderly depending on the speed of trnsfer (urgent vs norml), ultimte destintion, nd the mount sent. These hrges re omprle with those for Indi; however, the fees hrged for other ountries suh s Brzil nd Mexio re higher from the UAE nd Sudi Ari nd lower from the USA nd United Kingdom. The rtes for Western Union nd Moneygrm re quite similr ut the network of the former in 74 This sheme ws introdued to filitte the Islmi Pilgrimge (Hjj) to Me. In the 1990s, there ws shortge of foreign exhnge nd people who hd reltives rod ould sk them to sponsor the Hjj for them. 75 For Western Union nd Moneygrm, the hrges re sed on sending sh. Pge 87

103 Bringing Finne to Pkistn s Poor Chpter 5 Pkistn is fr more extensive. The remittne hrges for Internet heks re $3 10 while for redit rds flt fee of $8 20 is hrged plus perent of the mount remitted. A ompetitive online trnsfer servie in rel time t fee of $2 is offered y Doh Bnk in Qtr, whih hs prtnered with HBL nd Askri Bnk in Pkistn. Tle 5.7: Cost of Sending $300 to Pkistn through Vrious Chnnels U.S. (New York) U.K. (London) U.A.E. Sudi Ari (Riydh) (Dui) Western Union $10 $18.30 $6.80 $6.70 Moneygrm $10 $20.30 $5.50 $5.30 Foreign nks $35 to $60 $40 to $80 $6.8 to $27 $16 to $24 Pkistni nks Free Free Free Free Soure: Author s own reserh For MTOs, fees hrged t the sending end re divided etween three prties; the sending gent keeps perent, the MTO keeps perent, nd the reeiving gent keeps perent (Psss 2004). The reeiving gent will give shre to the frnhise for hndling the tul disursement. The ommission pid to frnhises in Pkistn y ECs is 5 perent of the mount disursed Outreh nd Sope of Remittne Tehnologies Roughly 64 perent of the popultion of Pkistn resides in rurl res; however, only 33 perent of the nking network is in these res. Limited ess to nking in rurl res requires innovtive solutions to sle up outreh (Box 5.4 presents one suh solution for Indi). Interntionlly, severl new remittne tehnologies hve een in use, suh s trnsferring money on prepid rds or moile phones; however, in Pkistn only online trnsfers re ville. One signifint dvntge of finnil servies vi moile phones is their ility to onveniently ring nking to women, who for ulturl nd soil resons do not go out frequently. Box 5.4: Indin nongovernmentl orgniztion (NGO) Helping Domesti Remittnes In Indi, the NGO Adhikr is piloting domesti money trnsfer servie for the lrge numer of migrnts who trvel from the estern stte of Oriss to work in the western stte of Gujrt. Adhikr s omprtive dvntges in money trnsfer servies inlude knowledge of ustomer needs nd preferenes, s well s the ility to servie lients in remote lotions. When designing its trnsfer system, Adhikr deided to leve the tul trnsmission of funds to Corportion Bnk, whih hs rnhes in oth Oriss nd Gujrt nd the infrstruture to mke timely, seure trnsfers. Although nk trnsfers tke ple regulrly etween Oriss nd Gujrt, most migrnts do not hve nk ounts nd find it ostly to visit nk rnh. Adhikr entrlizes the olletion nd dissemintion of these smll trnsfers nd routes them through one ount t the nk. This proess spreds trnstion osts over lrger numer of trnsfers, ringing down the per-trnstion fee. Adhikr is now looking to leverge the system y involving NGOs tht serve other distrits in Oriss s distriution gents, lowering per-trnstion fees even further while enling the NGOs to ern new soure of revenue to support their work. Soure: Consulttive Group to Assist the Poor (2005) Through the issune of the Guidelines for Moile Bnking Opertions of MFBs/Institutions, SBP hs een enourging the privte setor towrds providing moile nking solutions. This shows strong potentil s oth the nking nd teleom setors re mong the most virnt in the ountry. In ddition, Mirofinne Institutions (MFIs) re growing t roust rte nd n support inresed outreh s 60 perent of their network is in rurl res, lthough only MFBs re llowed to remit money domestilly. Pge 88

104 Bringing Finne to Pkistn s Poor Chpter 5 The network of ECs in Pkistn is rod nd overs ll high-volume res. The ECs nlyze where remittnes re flowing to nd open frnhises/pyment ooths ordingly, in nery towns or smll ities. However, s the systems of the ECs re utomted nd the minimum requirement for them to set up rnh is to hve phone onnetion nd Internet, they hve tehnilly not opened rnhes in rurl res, ut in ities ner villges tht reeive remittnes. The est strtegy for inresing finnil ess of the poor is to introdue moile phone pltforms. The outreh of the teleom setor is onsiderle the moile network in priniple overs out 90 perent of the popultion of Pkistn. The totl onsumer se rehed 89.3 million y July ; the provinil rekdown of the moile onsumer se is given in Figure 5.6 (Pkistn Teleommunitions Authority 2007). Figure 5.6: Moile Consumer Bse y Provine Figure 5.7: Moile Penetrtion y Provine Soure: Pkistn Teleommunition Authority (2007) The penetrtion of moile servies in eh provine is round 30 perent (Figure 5.7), ut this mounts to 80 perent of the totl popultion. However, AJK nd Northern Ares re still ehind nd the popultion overed is 20 perent, or out 1 million people. This moile overge, nd the suess of moile money trnsfer solutions in the Philippines nd Keny (Box 5.5), suggests tht moile nking offers signifint potentil to sle up ess to finnil servies. Speifilly, in the Philippines, the G-Csh model hs hieved exellent outreh nd n e used s model in Pkistn to link moile phones with finnil servies. Moile-enled usiness models hve low operting osts, nd if onsumers n eome svvy with the nking tehnology loded onto moiles, it offers the est wy forwrd for improving outreh of ll finnil servies, inluding remittnes. Aording to the Consulttive Group to Assist the Poor, using moile phone n redue the ost of typil nk trnstion y 80 perent in the Philippines, while Tmeer Bnk estimtes tht, in the Orngi slum of Krhi, the setup ost of nk rnh would e 30 times more thn the setup ost of n gent. Moile money trnsfers hve the potentil to help lower interntionl remittne osts y up to 75 perent nd domesti remittne osts y up to 90 perent, providing diret svings to poor households. 76 The ellulr density s of July 31, 2008, ws 55.6 perent. Pge 89

105 Bringing Finne to Pkistn s Poor Chpter 5 Box 5.5: Moile Bnking: G-Csh (Philippines) nd M-Pes (Keny) The G-Csh produt is fility linking the user s phone to sh ount or wllet. The ustomer n deposit or withdrw sh from this wllet s needed, while redit n e trnsferred etween moile users s well. It opertes on redit sis, where the user initilly deposits sh t designted deposit lotion suh s Gloe offies or redited retilers. This servie llows moile users shless purhsing t wide rnge of shops, diret redit from employer pyroll, ill pyment, nd inwrd remittne from overses Filipino workers. Initilly, the user registers through sending short messge servie (SMS) nd deposits money; there re no hrges for registrtion. Eh trnstion updte is sent y Gloe in the form of n SMS. Just s esily s sh is deposited t deposit lotion, it n esily e withdrwn; however, n eptle identity doument is required. Retil purhses re possile t severl thousnd prtiipting outlets. This involves request from the retiler originting the trnstion through his own ell phone terminl nd ustomer uthoriztion from their own phone. Airtime trnsfers, prepid top up, nd ill pyments n e rried out with similr ese. For eh trnstion, the ost is the stndrd SMS fee ($0.02). For sh deposits nd withdrwls, the fee hrged is 10 perent of the trnstion vlue with minimum of $0.19. Gloe mintins nd opertes leringhouse fility tht reords ll trnstions nd rrnges settlement etween the retilers nd the G-Csh ustomers. This results in Gloe hving responsiility for the usul nking issues of frud mngement nd money lundering prevention. However, the sh flot from the ustomer lnes is held y the ompny s usul usiness nk, lthough tht nk hs no knowledge of the individul user tivity or ount lnes. The user hs to uthorize eh trnstion y PIN nd the mximum lne tht n e mintined y ustomer is $189. The M-PESA moile money trnsfer pltform in Keny is very similr to G-Csh nd it speifilly trgets the unnked popultion. These people, who do not hve ess to onvenient or hep sh trnsfer servies or nk rnhes, simply register with n M-Pes gent y providing their Sfriom moile numer nd identify rd. One they hve done this nd deposited money with lol gent, they n send money to other moile phone users through n SMS even if they re not Sfriom users. This sh n esily e withdrwn t n M-Pes gent. Bnks re now eginning to prtner with Sfriom to provide further servies. For exmple, two Equity Bnk ount holders with Sfriom lines n now redit nd trnsfer sh into eh other s ount using their moile phones in rel time, sujet to the vilility of funds. Equity nk lients will lso e le to py their utility ills, purhse irtime, request their nk sttements nd perform other ruil nking enquiries. Soure: DFID SBP hs issued guidelines for rnhless nking, overing three different models: In the first, nks n work with prtiulr teleom provider to offer servies to the urrent ustomer se of tht ompny. In the seond, nk n offer moile phone nking servies to ustomers using moile onnetion of ny teleom. The third model is group of nks nd teleoms tht n join hnds to provide servie. 77 As teleom providers re not themselves prudentilly regulted, they re prevented from tking deposits nd so nnot themselves provide svings produts ( primry need for mny poor onsumers). This would lso hmper their ility to e le to provide lons. Hene, lthough pure teleom provider models n urrently e effetive for providing moile pyments, there is need for nking reltionship, or pproprite regultion, to llow ess to other finnil servies eyond remittnes through these models. To further inrese outreh of finnil servies, SBP hs provided more options for rnhless nking suh s to the Post Offie, fuel distriution ompnies nd other hin stores. However, exept for the Post Offie, other plyers re still not involved in providing finnil servies. There is lot of tivity in this setor nd oth nks nd moile phone ompnies re working to develop vile models. UBL hs lunhed produt sed on the seond model lled Orion, essile only to UBL ount holders. The servies ville through Orion re pying ills, trnsferring money mong other 77 Guidelines ville t Pge 90

106 Bringing Finne to Pkistn s Poor Chpter 5 Orion users, nd uying phone rds. In time, UBL plns to inrese the kind of trnstions tht n e done with Orion, long with dding remittne trnsfer filities. Bnk Al-Flh is urrently testing moile nking produt with Wrid Teleom. Another plyer tht will enter the moile nking mrket soon is Etislt of the UAE, the mjor shreholder of Pkistn Teleommunitions Compny nd Ufone. They hve signed ontrt with Smrt of Philippines in the UAE to offer remittne servies. The remittne servie will llow susriers of Etislt to send funds to Pkistn nd other ountries nytime, nywhere, using their moile phones. It will lso llow trking of trnstions. The only infrstruture required will e sh in/out merhnts nd perhps these n ride on the network of moile phone rd providers, whih re ll over Pkistn. In the USA, dul utomted teller mhine (ATM) rds re very populr with Mexin immigrnts. In Pkistn, none of the stkeholders ontted hd ny knowledge of dul ATM rds. However, vrious Pkistni nks in the Gulf region do provide for dul ATM rds, nd they re quite onvenient s reipients k home n withdrw money nytime. It is possile tht Pkistni exptrites might e giving their fmilies ATM nd redit/deit rds to use in Pkistn, given the lrge mounts of money proessed through ATMs nd point of sle (POS). 78 The Remittne Crd Model from Indi or Vis Giro from Brzil n e dopted (Box 5.6), s eletroni nking trnstions re on the rise nd this model my prove populr for urn res. Citink UAE hs reently strted Twin Aount Fility for exptrite Pkistnis who hve nking reltionship in the UAE ut lso need to send money to Pkistn. Aount holders n trnsfer money round the lok y phone, online, y ATM, or t rnhes. Customers in oth the UAE nd Pkistn will e given hekooks nd deit rds to settle pyments worldwide. This servie ws lunhed in August Ikoo offers prepid vis rd remittnes, whih re ville online nd provide onveniene nd flexiility, nd n e used t ny ATM or ssoited mhines. However, this requires esy ess to ATMs nd tht the reipients re ATM svvy. Although this servie is offered for Pkistn, the extent of its urrent use is unknown. In Brzil, remittnes through Vis re doing well, given the lrge popultion tht uses redit rds nd good infrstruture to support suh produt. As these exmples show, remittnes n e used s key entry point to ringing the unnked into the forml setor nd inresing finnil inlusion. Tht is, money trnsferred through finnil institutions pves the wy for other finnil produts nd servies to e provided. Remittnes re often the first, nd sometimes the only, finnil servie tht low-inome households use. Providing remittne servies llows nks to get to know nd reh out to unnked reipients, uild up informtion on the size nd stility of their sh flows, nd then offer other finnil servies suh s svings or redit. As remittnes re typilly lumpy, it mkes sense tht remittne reeiver my need seure wy to store tht money. Inresed svings will in turn enle n inrese in lonle funds, ggregte redit nd wider finnil intermedition. Eonometri evidene provides strong support for the notion tht remittnes promote roder finnil development in developing ountries, vi inreses in the ggregte level of deposits nd redit intermedited y the lol nking setor. This is importnt onsidering the extensive literture tht hs doumented the growth-enhning nd poverty-reduing effets of finnil development (Aggrwl et l. 2006). 78 Aording to the SBP Pyment Systems Deprtment, from Otoer 2006 to Mrh 2007, pproximtely $67 million worth of trnstions were proessed through ATMs, nd similrly $64 million worth of trnstions were proessed through POS (SBC Pyment Systems Deprtment 2007). Pge 91

107 Bringing Finne to Pkistn s Poor Chpter Poliy nd Regultory Frmework Sine the erly 1980s, SBP hs een tively working to support virnt remittnes mrket. In the 1980s, it enourged Pkistni nks to estlish rnhes in Middle Estern ountries to where most Pkistnis hd migrted. The foreign exhnge regime ws lierlized, nd the option of opening foreign urreny ounts in Pkistn ws given to exptrites, llowing the free inflow nd outflow of foreign urreny. After Septemer 11, 2001, SBP took further tions to formlize remittnes, suh s setting up entrlized home remittne ells. Bnks sumitted their remittne trgets to SBP nd n SBP tskfore ensured implementtion. A omplint nd monitoring ell ws set up to ddress puli omplints. The GOP gve inentives t this time to further oost remittnes, suh s exemption of ustom duties on personl ggge nd speil immigrtion lerne ounters. On governmentl level, SBP tems visited nd studied severl ountries to lern from Anti-Money Lundering (AML) lws nd suessful remittne prties. MFBs were given permission to offer remittne filities, nd SBP is working with ll finnil institutions to enhne effiieny nd trnspreny. Future government plns inlude dding new produts with etter rte of return to the Ntionl Svings Sheme for remittne enefiiries, to disourge onsumer spending nd hnnel svings into produtive investments. On the donor side, the World Bnk hs supported mny inititives, suh s estlishing new frmework to improve the qulity nd overge of dt, or formulting Generl Priniples for Interntionl Remittne Servies to help improve regultion, foster ompetition nd inrese trnspreny. The World Bnk hs lso ooperted with other interntionl finnil institutions to work on remittnes, nd AML/omting the finning of terrorism mesures. It hs lunhed prie nd ost dtse on remittnes to enhne trnspreny nd is working on improving pyment systems so tht they re quik, hep, sfe, effiient, ontestle, nd trnsprent. The Interntionl Monetry Fund is helping ountries implement Finnil Ation Tsk Fore (FATF) reommendtions for remittne systems. It is undertking omprehensive ssessment of the AML/omting the finning of terrorism mesures long with other Interntionl Finnil Institutions nd is developing n opertionl guide, s well s offering tehnil ssistne on their implementtion. The Interntionl Monetry Fund is lso working to improve sttistis nd how they re ptured in the lne of pyments. Annex Tle A5.9 gives snpshot of donor nd Interntionl Finnil Institution tivities to improve the remittne system Legl/Regultory Frmework The legl nd regultory frmework governing inwrd remittne trnstions is provided for under the Foreign Exhnge Regultion At, SBP implemented the At through generl or speil diretives issued s osionl irulrs nd notifitions. The primry instrutions regrding foreign exhnge trnstions re ontined in the Foreign Exhnge Mnul issued nd updted y SBP. The Foreign Exhnge Mnul provides tht there re no restritions on inoming remittnes. However, finnil institutions, nks, nd other uthorized delers (inluding ECs) hve to otin lienses to rry on foreign exhnge trnstions. All trnstions hve to e reported to SBP, identifying the mount nd the soure. Both nks nd ECs report on dily, weekly, nd monthly sis. Furthermore, the nks hve ess to Rel Time Gross Settlement System for low-volume, high-vlue trnstions. Pge 92

108 Bringing Finne to Pkistn s Poor Chpter 5 SBP hs een moving stedily towrd utomting the system nd providing rel-time settlements to keep up with interntionl est prties nd ntionl nd interntionl requirements. To this end, in 2007, the Prliment ented Pyment nd Eletroni Fund Trnsfer At, whih lid down solid foundtions for development nd further strengthening of pyment systems in the ountry, implemented y SBP In ollortion with Ntionl Internk Finnil Teleommunitions (NIFT), SBP hs een implementing the rel-time gross settlement system lled PRISM (Pkistn Rel Time Internk Settlement Mehnism). An Automted Clering House hs een set up, sed on MICR Reder nd Sorter equipment with ode written in house to filitte lol requirements nd hnging needs. The utomted lering house n e diretly essed y eh rnh nd plys role of entrl regulting uthority. NIFT ollets the outwrd lering diretly from eh nk, rrnges ll proessing, nd provides the net position to SBP. It provides servies to ll 42 ommeril nks nd their rnhes. NIFT enters hve een estlished in over 80 ities for 4,000 nk rnhes to provide NIFT Automted Clering House Servies. SBP is lso working on providing ll lol nks with utomted low-ost Soiety for Worldwide Internk Finnil Teleommunition (SWIFT) onnetivity nd lol U.S. dollr loop hs een set up with NIFT to expedite lering of dollr trnstions. The new system hs redued the lering time for U.S. dollr heks from three weeks to only four dys, long with ost svings to the ount holders. At the sme time, SBP is lso rpidly developing the e-nking infrstruture. All nks nd ECs re sujet to oth off- nd onsite inspetions nd ny violtion mkes them lile to penlties. Eh trnstion y the institutions hs to e reported on weekly nd monthly sis to SBP nd hs to e filtered through n in-house AML system. Suspiious trnstions something out of hrter or inonsistent with the history nd pttern of the individul ounts hve to e reported. SBP does the monitoring vi supervision units t 16 field offies. A omprehensive inspetion mnul overing ll res of opertions hs een developed to help with monitoring. The opening of eh new frnhise offie or pyment ooth of n EC is refully srutinized y SBP, inluding kground hek on the person who is opening the offie nd visit to the lotion to verify the dequy of the filities. SBP further onduts mystery shopping y sending underover stff to sertin the qulity nd type of servie reeived. Know Your Customer regultions re lso in fore. ECs do not mke pyments unless the reipient hs ntionl identity rd. Furthermore, SBP mndtes eh nk nd exhnge ompny to hve ompline deprtment tht monitors ounts nd trnstions nd updtes ustomer informtion on regulr sis. Proper doumenttion nd reord keeping hs to e mintined, while orrespondent nks hve to e refully hosen fter exmining their due diligene proedures Ostles In the rih urn res of Pkistn, very few ostles to effiient remittne flows remin. The network of nks nd ECs is widespred, the ost is resonle, nd the speed of trnsfer is prompt. However, similr filities re not ville to low-inome nd rurl res. SBP hs stimulted outreh y its reent rnh poliy, wherey nks with 100 rnhes or more re required to hve 20 perent of their rnhes in rurl res. EC expnsion in rurl res is severely restrited, euse their online systems require phone nd Internet onnetion. However, s moile nd wireless penetrtion eomes more ommon, this will improves. A omputerized-kiosk system with low-ost ATMs hs permitted onsiderle expnsion of remittne servies overge in rurl Indi (Box 5.7). Pge 93

109 Bringing Finne to Pkistn s Poor Chpter 5 Box 5.7: Kiosks to Help Rurl Outreh ICICI Bnk, privte Indin ommeril nk tht evolved out of ntionl development nk, offers wide rnge of finnil servies, inluding money trnsfers sent to Indi y nonresident Indins. In 2004, the nk s Money2Indi servie hd over 670 gent lotions nd reently extended its outreh to remote villge enters vi omputer kiosks. The omputer kiosk system work s follows: sender remits money trnsfer to the reipient s ICICI ount, either through n ICICI rnh offie or Money2Indi gent. As soon s the trnstion hs tken ple, the Money2Indi gent informs the kiosk opertor, who in turn informs the reipient. The reipient n then either ollet the remittne t ICICI or the kiosk, whih is equipped with low-ost ATM. ICICI estimtes tht kiosks n e profitly pled in villges s smll s 2,000 residents. This option is very ttrtive for rurl reipients, euse it elimintes trnstion osts involved in trveling to lrger town to visit nk rnh. Kiosks used y ICICI nk offer omintion of telephone, finnil, edutionl, nd other servies. Kiosk opertors re independent usiness people, remunerted through ommissions pid y servie providers nd user fees pid y ustomers. They py for set-up osts themselves, for whih they typilly otin prtil lon from ICICI Bnk. Sine ICICI Bnk does not inur ny fixed osts, the system hs proven ost-effetive wy for the nk to extend its outreh to rurl res. In mid-2004, pproximtely 150 kiosk opertors offered ICICI servies nd the nk hoped to inrese tht numer to over 2,000 in pproximtely 12 months. From n opertor s perspetive, the usiness model is only vile if multiple servies re routed through single kiosk. However, experiene indites tht the kiosks n eome profitle even without the money trnsfer servie, whih n esily e dded t lter dte. ICICI Bnk ws le to offer its Money2Indi servie due to onfluene of irumstnes: (1) the inventors of the kiosk system were seeking pproprite usiness pplitions for it, (2) the tehnology suited the needs of ICICI, nd (3) other nonfinnil servie providers, suh s ompnies tht offer edutionl nd helth informtion/dignosti servies, lso opted to use the kiosks, reting multiservie usinesses tht ensured opertor profits. These irumstnes, nd therefore the trnsmission system for Money2Indi, my e somewht diffiult to replite in other ountries. However, the exmple of omputer kiosks with low-ost ATMs my represent ost-effetive wy for finnil servie providers to expnd the outreh of money trnsfers, s well s other servies. Soure: Consulttive Group to Assist the Poor (2005) Pkistn Post hs the est network in rurl res, though there is spe for muh effiieny improvement. HBL hd n greement with Pkistn Post to deliver remittnes to rurl res where they do not hve rnhes, ut delys in the proessing nd delivery of money orders led HBL to nel the greement. A further issue remins in formlizing informl remittnes, espeilly domesti flows. A lrge perentge of people do not onsider Pkistn Post or forml nking setor s options euse they feel these re too omplited nd ineffiient. An ostle speifi to the Gulf re is tht most of the immigrnts from Pkistn re illiterte nd re shy of going to nks to remit money. These people find it muh esier to send money through hwldrs whom they know nd trust. The onveniene nd simpliity of hwl nnot e mthed y nks, s remitting money through nks requires proess of first going to the nk, then witing for the reipient s nk to e redited, nd finlly the reipient trveling to the nk to withdrw the pyment. With hwl, the money is delivered to the reipient s doorstep. An importnt rrier for moilizing remittnes through forml hnnels from the USA is tht the people who generlly send money ern in sh nd do not hve nk ounts. They re undoumented nd unnked, nd it is not possile for them to go to nk or n MTO. Furthermore, the worker lss is lso unfmilir with these orgniztions nd prefers the trust-sed mehnism of hwl. The key drivers of the formliztion of remittne flows inlude ost, ese of ess, identity requirements, trust, nd finnil (nd tehnologil) litery. If forml servies n e mde heper, esier to ess, user friendly without prohiitive identity requirements, nd with pproprite edution to lert people to the Pge 94

110 Bringing Finne to Pkistn s Poor Chpter 5 enefits of forml systems nd how to use them, then this will signifintly help ring informl flows into the forml setor. New tehnologies, suh s rnhless nking, offer signifint potentil to redue trnstions osts nd so n llow new usiness models for previously unrehed res to eome vile (s noted in setion 5.6 ove). They lso re onvenient nd ess rurl or other hrd-to-reh res where trditionl nk rnh models re unvile, nd so offer gret potentil towrds the key drivers of formliztion Suggested Avenues for Ation There re vitl roles for oth the government nd the privte setor to ply in hrnessing remittnes for ess to finne. The privte setor should e fundmentlly responsile for providing finnil servies, nd the government needs to ensure to the extent possile tht the usiness nd regultory environment is onduive to the privte setor provision of sfe, effiient nd equitle finnil servies. The most omprehensive step to inrese ess to remittnes nd potentilly to hrness ess to other finnil servies would e moile nking solutions. Over hlf of the popultion lredy regulrly uses moile phone nd eonometri nlysis hs highlighted tht remittne users re likely to e regulr moile phone users. Moile phones re ffordle, nd the network hs expnded immensely in reent yers, nd moile pyment pltforms offer signifint potentil for lowering ost nd inresing outreh for the rurl poor. One onrete wy to mke progress in the setor would e to form puli-privte working group on remittnes. This would enle dilogue etween the government nd privte ompnies (nks, ECs, nd MTOs, s well s teleoms where relevnt) to tkle key hllenges in the setor. Disussion ould fous on, mong other things, legl nd regultory issues, mrket trnspreny, ompetition nd ost, s well s reserh nd dt issues. The UK hs hd suessful experiene in this regrd with Puli-Privte Working Group on Remittnes, whih led to the retion of privte setor-led U.K. Remittnes Tsk Fore. The World Bnk hs lso onvened similr group t the interntionl level ringing together key tors from the puli nd privte setors. Pkistn n drw on this onsiderle experiene Privte Setor Ations Reduing informlity: Informlity will derese upon the introdution of effiient, low-ost, esy-ess, user-friendly, nd trusted remittne servies, without prohiitive identifition requirements. Developing new tehnology-sed models for finnil servies: Severl new remittne tehnologies, suh s trnsferring money on prepid rds or moile phones, re lredy in use in vrious ountries; however, only online trnsfers re ville in Pkistn. SBP hs issued guidelines for three models of rnhless nking: first, where nks work with prtiulr teleom provider to offer servies to the urrent ustomer se of tht provider; seond, where nks offers moile phone nking servies to their own ustomers using the moile onnetion of ny teleom; nd third, where group of nks nd teleoms jointly provide servie. 79 New tehnologies, suh s rnhless nking, offer signifint potentil to redue trnstions osts nd so n llow new usiness models for previously unrehed res where trditionl nk rnh models re unvile. There is lot of tivity in this setor nd oth nks nd moile phone ompnies re working to ome up with vile models. UBL hs lunhed produt tht provides servies for pying ills, trnsferring money etween ustomers, nd uying phone rds. Bnk Al-Flh is urrently testing moile nking 79 Guidelines ville t Pge 95

111 Bringing Finne to Pkistn s Poor Chpter 5 produt with Wrid Teleom, nd Etislt of the UAE; it hs signed ontrt with Smrt (of the Philippines) to offer remittne servies in the UAE. Bnks nd teleoms ompnies should work to develop further money trnsfer nd moile nking servies in Pkistn. Expnding nk presene nd servie provision rod: The limited presene nd outreh of Pkistni nks rod is the min prolem they fe in inresing remittnes through their networks. Foreign nks re more esily essile to remittne senders rod, ut these my provide slower or more expensive remittne servie to Pkistn, nd my not hve signifint outreh in rurl res. New mesures should e tken to inrese outreh rod, prtiulrly in the mjor remittne soure ountries. These ould e new prtnerships, or new innovtive wys to reh ustomers. HBL mnged to gin mrket shre y rehing out to the workers in the Middle Est. A similr strtegy ould e followed in ll ountries where sizele Pkistni worker ommunity exists. One these workers, who re the min senders, re prt of the proess nd system is set up to fit their needs, the volume of remittnes will grow further. Moreover, if they re given inentives or onuses sed on the mount they remit, they will eome more diligent with sending money. One vet to keep in mind is tht the litery level is low nd so pperwork should e kept to minimum nd tehnology interfes should not e exessively omplex. In ddition, the reipients of remittnes re generlly women or elder groups, for whom onveniene is most importnt. For this group, models tht re sed on doorstep delivery would e most suitle. Forming new prtnerships within Pkistn: Within the Pkistni finnil mrket, prtnerships nd linkges etween different institutions should e developed to filitte ess to remittnes nd other finnil servies y the rurl poor. For exmple, ommeril nks should explore working with MFBs to expnd their networks in rurl res Government Ations Creting n enling environment for the privte setor: SBP hs tken signifint steps over reent yers to improve the remittne mrket in Pkistn. It hs in some ses provided more options for rnhless nking suh s to the Post Offie, fuel distriution ompnies, nd other hin stores, or issued guidelines for rnhless nking, ut the privte setor hs yet to fully respond to these res. There my e other rriers or lk of dequte inentives preventing the intended response, nd hene the government should tlk to the privte setor to understnd these rriers nd work to remove them where possile. Telo-sed models suh s G-sh nd M-Pes hve shown some limittions, most importntly the lk of prudentil regultion with the ttendnt seurity nd supervision shortomings, s well s the nrrow rnge of servies tht ould e provided y suh models. Regultory reforms tht my e required inlude mendments to the Pyment nd Eletroni Fund Trnsfer At nd Brnhless Bnking regultions providing for nk-nonnk prtnership nd use of gents in money trnsfer servies. A dt privy or seurity lw n filitte e- or m- pyments. Continuing to ring informl remittnes into the forml system, prtiulrly on the domesti front: An enling environment for the privte setor, enourgement for opening of nk ounts, s well s improved postl offie effiieny will hve n dditionl effet to ontinue ttrting remittne flows to the forml setor. However, until effiient, hep, esy-to-ess, user friendly without prohiitive identity requirements nd trusted methods penetrte throughout the ountry, inluding in rurl res, demnd for informl remittnes, prtiulrly on the domesti front, will e importnt. Advnes in tehnil nd finnil litery will mtter s well, with pproprite edution to lert people to the enefits of forml systems nd how to use them. Supporting remittne servies of Pkistni nks rod: Further enourgement for Pkistni nks might e required to ontinue to oost interntionl remittnes nd forge llines with interntionl nks. One potentil strtegy would e to set the reimursement rte through prtner nks higher thn those Pge 96

112 Bringing Finne to Pkistn s Poor Chpter 5 oming solely through Pkistni nk networks. Alterntively, smll fee ould e hrged to senders when they use prtner nks in interntionl ountries. Pkistni missions rod ould do survey of options ville in their respetive mrkets nd disseminte the informtion to immigrnts on the hepest nd est soures of money trnsfer. For exmple, the Mexin mission in the USA pulishes suh informtion. A further suessful poliy rried out y the Mexin emssy in the USA is tht it hs issued type of identifition rd (Mtriul Consulr) for their immigrnts, nd with this rd Mexin immigrnts into the USA n open nk ounts. Perhps the GOP n disuss something similr with the U.S. nd other uthorities. Enourging the opening of nk ounts: Another step tht SBP should work towrd is enourging people to open nk ounts, s supported y positive pereptions on essiility nd sfety of nk rnhes s remittne hnnel. The high shre of nks in the interntionl remittne mrket in Pkistn lso indites tht inresing nk ounts should help inrese remittnes through forml hnnels. Currently, the minimum ount lne of Rs 10,000 is prohiitive for lrge segment of the soiety 80. To enourge people to open nk ounts, unomplited pperwork nd less-stringent lne requirements re needed. Bnks urrently hndling remittnes n redit ounts within 24 hours; however, delivering money or sending money orders y the Post Offie retes unneessry delys, nd the nking setor loses ustomers to MTOs nd informl mens of trnsfer. Seeking to improve the effiieny of the postl remittne servie: Pkistn Post hs signifint network nd strong outreh in rurl res. However, servies n e ineffiient. Government should explore wys to improve this servie nd speed up the utomtion of postl rnhes. Promoting the struturing of interntionl flows into investments: Chnneling remittne flows diretly into investment in the trdle setors of the eonomy hs een pursued suessfully in the interntionl experiene, to ountert loss of interntionl ompetitiveness ssoited with elevted remittne flows. Interntionl good prties inlude pkging remittnes with pyment servies (suh s Bnsefi in Mexio). Suessful poliies to struture interntionl flows into investments hve lso foused on orgnizing nd trgeting Dispor networks rther thn tul remittnes flows. At the mro level, Dispor onds issues (for exmple, in Ghn, Indi, nd Isrel) or seuritiztion of future remittnes flows (in Brzil nd El Slvdor) hve een used, lthough seuritiztion hs proven ostly. At the miro level, governments hve filitted trgeted Dispor funds. Remittnes to Pkistn re of onsiderle eonomi importne, inluding domesti flows, nd hve grown quikly in reent yers, providing foreign exhnge, improving the lne of pyments nd externl det position, nd supporting inomes of poor nd vulnerle groups. Trnsmission networks work well in urn res, though outreh to rurl nd remote lotions re diffiult, nd servies re not suffiiently ustomized to lient needs (suh s women who might need doorstep delivery). The lrge network of Pkistn Post offers promise for the future, ut onsiderle gins in effiieny re required. SBP hs tken vrious mesures tht hve signifintly inresed remittnes through forml hnnels, though lrge shre of domesti remittnes remins informlly trnsferred. To stimulte outreh to remote lotions, SBP hs een enourging the privte setor to provide moile nking solutions. Other solutions from interntionl experiene inlude Indin innovtions suh s the NGO Adhikr, whih developed n effiient domesti ustomized trnsfer servie, nd ICICI Bnk, whih extended its outreh to remote villge enters vi omputer kiosks. New prtnerships mong remittne mrket plyers nd other finnil entities oth within Pkistn nd rod hold muh promise. Further dvnes in formlizing money trnsfer flows will ring new lientele nd motivtion for effiieny gins nd ustomiztion of servies to lient needs. 80 In n effort to enourge the opening of nk ounts, SBP introdued Bsi Bnking Aounts (BBA) in 2005 s new produt, with no minimum lne requirements nd minimum opening lne of only Rs Pge 97

113 Bringing Finne to Pkistn s Poor Chpter 6 6. Expnding Aess to the Underserved: An Ation Pln A mjor drive to enhne finnil inlusion would involve joint effort of the Stte Bnk of Pkistn (SBP), ntionl government, privte setor, the ommunity, nd donors. The est et of rpid sling-up of ess is to rely on tehnology, finnil wreness improvements, finnil reengineering of proesses nd produts, nd n enling legl nd institutionl frmework The Role of the Privte Setor Aess to All Finnil Servies Diversifying the produt rnge to inrese outreh, lower osts, nd mnge risks: The rod hed ertinly lies in produt diversifition, with more servies nd less requirements tering to the mss popultion of lower-inome pity. Produts (svings, insurne, nd redit) for old ge, hildren s edution, pregnny nd medil expenses, nd livestok re few exmples of those tht tke ount of women s needs for life-yle events. Home-sed usinesses should e given onsidertion. Aess would lso improve with the use of lterntive forms of ollterl, suh s soil ollterl, ompulsory svings, personl gurntees, rops or mhinery to e purhsed or household ssets. Sving produts, whih re expeted to e espeilly populr, n e uilt upon trditionl sving rrngements nd rotting sving nd redit ssoitions tht women use. Smller size of produts, nd ulk servie, might etter ttrt lower-inome groups. Litery should not e requirement to ess finnil servie. Simple poliies nd proedures tht speed the trnstion, lower trnstion osts for women, nd do not prelude uneduted women tend to mximize outreh to women lients. Innovtive wys to reh ustomers suh s deentrlized opertions, operting units loted ner women lients, use of moile units, nd trnstions t lients doorsteps tend to mke nking onvenient for women. Femle stff will improve pprohility for lients nd llevite ulturl onerns. Poliymkers nd nks would note the importne of the mrket for finnil produts trgeting heds of households, given the finding tht eing household hed hs signifint impt on finnil ess. This potentil is prtiulrly of interest in rurl res s well s lusters of urnized rurl migrnts into igger ities. Rehing out to the femle lient, lient segmenttion: Women s ilities to etter mnge det, their stronger svings ptterns nd lient loylty present n untpped profitle lient se for the finnil nd mirofinne setors in Pkistn. Understnding women s needs more preisely, nd refleting those needs in the finnil produts nd the provider s poliies nd proedures, would ensure n inrese in women s ess to finne in spite of ulturl norms, gender segregtion, nd low litery nd inomes. Lower lon size nd deposit size would permit etter mthing to women s needs, given their lower inomes. Repyments should e frequent so tht instllments re smller nd should orrespond to women s inome yles. Glol experiene suggests offering women redit tht is untied to speifi use, insted llowing the orrower to elet the tivity. Lerning from the informl finnil setor s suess: The gp etween the populrity of informl finne nd limited forml finnil setor outreh, omined with the pereption gp etween hevy proedure-driven forml setor nd informl setor with minimum requirements, lls for lerning from informl finnil systems nd developing linkges with the informl setor. Strengthening finnil infrstruture, y expnding redit informtion ureus nd other pyment systems, would llow nks to lower inding requirements suh s gurntors nd immovle ollterl, therefore ttrting new ustomers mong women, the poor nd rurl res. Pge 98

114 Bringing Finne to Pkistn s Poor Chpter 6 Rehing out to the rurl lient, leverging tehnology nd prtnerships: within the Pkistni finnil mrket, prtnerships nd linkges etween different institutions should e developed to filitte ess to remittnes nd other finnil servies y the rurl poor. For exmple, ommeril nks should explore working with mirofinne nks (MFBs) to expnd their networks in rurl res. Enhning outreh vi tehnology solutions n involve omintion of devies to e dopted y nks to enhne outreh s well s rnhless nking through ell phone nd moile devies. These re very different diretions, nd eh hs shortomings nd diffiulties. In Pkistn, the nking ess infrstruture is prtiulrly wek, s noted in Chpter 1 nk rnhes nd ATMs re low s mesured ginst oth the popultion nd re of the ountry. As demonstrted y the Brzil orrespondent nking experiene, diffiulties in expnsion of outreh n lso rise due to ontrtul s well s regultory nd prudentil ftors regrding geny rrngements (Chidzero et l., 2006). Moile nking, on the other hnd, presents regultory hllenges (s in the se of the Philippines G-Csh nd Keny M-Pes models), ut n e promising hnnel to help shift some of the finnil flows from informl to forml hnnels, in prtiulr if omined with other orrespondent nking hnnels. Demnd-side results do show us very high moile penetrtion nd our eonomi nlysis highlights signifint positive linkges etween finnil inlusion nd regulr moile use, s well s informl inlusion nd ess to ell phone. Expnding outreh vi moile telephony, smrt rds, nd point of sle (POS) devies: Given the wide, reltively equitle, nd rpidly growing ess in Pkistn to moile phones, tehnology hs mjor potentil to eome onduit for ess to finne in the ountry. In ddition to ess expnsion, moile phones, smrt rds, POS devies, nd other tehnology improvements n lower trnstion osts, s well s help enhne redit informtion on muh wider popultion segment. The simpliity nd low ost of these servies hve enled poor people to use them esily nd suessfully in spite of their novelty nd reent penetrtion. This stnds in shrp ontrst to the omplexity nd lk of userfriendliness of trditionl nk produts, nd their reltive filure to penetrte wider popultion rnge. These tehnologies hve een very suessful in promoting pyments servies worldwide. In Pkistn, given popultion preferenes nd needs, it is importnt to find wys to extend ess to svings servies s well, vi tehnology gins. Interntionl experiene points to (1) regultory methods of promoting svings, suh s mthing shemes nd tx-dvntged shemes, s well s (2) svings methods tht hve worked for mirofinne doorstep olletion shemes nd periodi ontriution or ommitment progrms. But if the full potentil of this new pproh is to e relized, it will need to go well eyond the mirofinne setor, where there is lredy strong interest in using tehnology to reh more people nd lower osts, nd inlude the nks. So fr, ommeril nks hve not shown muh interest. Experimenttion with using moile telephone networks is just eginning. On the other hnd, rds hve een in use for mny yers, though not generlly y poor nd underserved popultions. As of Deemer 2007 the totl numer of tive rds in the Pkistn nking system stood t 6.7 million, s ompred to totl of 16 million personl nk ounts. Of these, 1.7 million were redit rds (25.4 perent) nd 4.8 million were deit rds (71.6 perent). ATM-only rds were million or 2.8 perent. Ninety-nine perent of the time, these rds were used for withdrwls while only one perent of the trnstions were deposits (envelope sed). Eh ATM hd n verge of 70 trnstions per dy, of n verge size of Rs 6,127. Until Deemer 2007, there were 2,618 ATM mhines (s ompred to over 8,000 nk rnhes), nd 52,474 POS terminls ross the ountry. Almost two-thirds (61 perent) of the nk rnh network onsisted of Rel Time Online Brnhes. One qurter of the totl trnstions in the system were eletroni-sed. Although the eletroni nking system is growing, it is still in its infny nd its expnsion in oming yers is si premise for the inlusion of the mjority of the popultion (Lindh de Montoy nd Hq 2008). Pge 99

115 Bringing Finne to Pkistn s Poor Chpter Aess to Mirofinne Improving Mirofinne Institutions (MFIs) sustinility nd ility to muster ommeril funding/svings deposits, nd their further integrtion into the finnil system: To rete strong, profitle MFIs, the setor needs to improve effiieny, risk mngement, nd profitility, s well s inrese reline on ommeril funding nd ttrt deposits. The ultimte gol for MFIs should e to firmly integrte into the finnil system. Limiting nonommeril funding for mirofinne nks (MFBs) would rete ompetition for funds nd drive for improvements. A further strtegy for the setor would e to refous from miroredit to mirosvings, given the lrge untpped demnd for suh produts. Besides eing the servie tht poor people wnt more thn ny other, svings moiliztion y MFBs will lso provide longer-term stle soure of funds to grow redit outreh. MFBs need to py more ttention to developing their svings servies, espeilly through strengthening their systems nd developing pproprite produts Aess to Smll nd Medium Enterprise (SME) Lending Crrying out thorough nk downsling progrm nd modernizing SME nking: Key fetures of the downsling progrms tht hve worked inlude: 1. Long-term tehnil ssistne is required to ensure tht the neessry sustntil hnges tke ple. Lending to smll usinesses requires profound hnges in the wy ommeril nks operte. Due to its high trnstion osts, smll usiness lending is only profitle if done in high numers with exellent portfolio qulity. Bnks ould enefit from long-term support in this hllenging proess. Tehnil ssistne to improve lending tehnology should fous on: reduing trnstion osts for oth the nk nd the lient; inresing lon offier produtivity (in terms of numer of lons disursed), nd mintining high portfolio qulity. 2. Seletion of nk dvisors nd ontent of the tehnil ssistne re key to suess. The tehnil ssistne pkge should e omprehensive, euse the hnges tht re required re quite sustntil. In ddition to its sope, the suess of the tehnil ssistne progrm will depend upon three elements: seletion of ommitted nks, seletion of onsultny firms with strong trk reord, nd lose monitoring y fully dedited nd experiened stff. The tehnil ssistne should inlude redesigning nk produts to meet lient needs, roust mngement informtion system, nd use of stff inentives linked to their performne. 3. A mix of ommitted nks should prtiipte in the progrm to rete ompetition mong them. Tehnil ssistne should e provided only to those nks tht re fully ommitted to SME lending. Serving suh enterprises requires hnge in the entire orporte ulture nd in the wy nks operte. Thus, only nks whose investors re willing to engge in suh sustntive trnsformtions should e offered tehnil ssistne. Often, nks with lrge nking network nd fous on retil lending hve omprtive dvntge in entering this mrket segment. Smller nks ould lso e trgeted, however, to t s tlysts. Mny ountries hve developed speilized SME progrms in stte-owned nks with lrge networks. However, suess usully tkes few yers, euse the hnges required re quite sustntil nd nks tke time in implementing them. It is therefore essentil tht tehnil ssistne e lso given to smll, fster-moving privte nks tht trnsform t fster speed ut my reh stgntion due to their smller networks. 4. Performne greements for nks prtiipting in the progrm re key. Furthermore, performne greements should lso inorporte terms on the numer nd volume of lons disursed nd outstnding y given dte, s well s portfolio qulity inditors suh s keeping portfolio t risk over 30 dys to Pge 100

116 Bringing Finne to Pkistn s Poor Chpter 6 less thn 3 perent. The tehnil ssistne should lst t lest two yers nd performne greements should e monitored on monthly sis so tht timely remedil mesures n e undertken when trgets re not met Aess to Remittnes Reduing informlity: Informlity will derese upon the introdution of effiient, low-ost, esy-ess, user-friendly, nd trusted remittne servies, without prohiitive identifition requirements. Developing new tehnology-sed remittne models: New tehnologies, suh s rnhless nking, offer signifint potentil to redue trnstions osts nd therefore llow new usiness models for previously unrehed res where trditionl nk rnh models re unvile. Bnks nd moile phone ompnies should ontinue working to implement vile models, following on existing efforts suh s United Bnk Limited s servies for pying ills, trnsferring money etween ustomers, nd uying phone rds; Bnk Al-Flh s moile nking produt with Wrid Teleom, nd the remittne venture of Etislt s (UAE)-Smrt (Philippines). Expnding Pkistni nk presene nd remittne servie provision rod: The limited presene nd outreh of Pkistni nks rod is the min prolem they fe in inresing remittnes through their networks. Foreign nks re more esily essile to remittne senders rod, ut these my provide slower or more expensive remittne servie to Pkistn nd my not hve signifint outreh in rurl res. New mesures should e tken to inrese outreh rod, prtiulrly in the mjor remittne soure ountries. These ould e new prtnerships, or new innovtive wys to reh ustomers. Hi Bnk Limited mnged to gin mrket shre y rehing out to the workers in the Middle Est. A similr strtegy ould e followed in ll ountries where sizele Pkistni worker ommunity exists. One these workers, who re the min senders, re prt of the proess nd system is set up to fit their needs, the volume of remittnes will grow further. Moreover, if they re given inentives or onuses sed on the mount they remit, they will eome more diligent with sending money. One vet to keep in mind is tht the litery level is low nd so pperwork should e kept to minimum nd tehnology interfes should not e exessively omplex. In ddition, the reipients of remittnes re generlly women or older groups, for whom onveniene is most importnt. For this group, models tht re sed on doorstep delivery would e most suitle The Role of the Puli Setor Aess to All Finnil Servies Creting wreness of the enefits of ess to finnil servies: Further gins in finnil litery re ritil, through more ritil is the presene of wreness of finnil servies to promote trust in the setor, s well s informtion out servies nd produts ville. A ntionl wreness mpign n support finnil inlusion, espeilly for women, s well s enourge people to open svings ounts. There is still lrge numer of people who do not hve ny ount t ll. Awreness-retion nd trustuilding ould e forged through soil moiliztion nd mss medi. Strengthening institutions (SECP, Credit Investment Bureu (CIB), Pkistn Post): Aess to finne growth will e elerted y n integrted finnil system, nd strong regultory frmework. Stronger institutions (inluding SECP nd CIB) re mjor prt of rpid inrese in finnil inlusion. Among the importnt fetures of omplete finnil system is well-funtioning ntionl-level redit ureu for redit referening. By estlishing redit history nd thus potentil ollterl sustitute, redit ureu n e instrumentl in providing ess to finne for groups tht my not hve sh or Pge 101

117 Bringing Finne to Pkistn s Poor Chpter 6 sset ollterl required to ess lon ut hve stellr redit history to present to the nk. All ommeril nks, development finnil institutions, lesing ompnies nd MFBs re urrently reporting to CIB on ll orrowers irrespetive of the size of lon. For the NGO-MFIs, speil CIB hs reently een developed nd SBP llows MFBs to shre redit history of their orrowers to MF-CIB. Consolidting these hievements nd ssuring the full usge nd funtionl vilility of the umulted dt, n ple mny more potentil orrowers within reh of some ess to finne. To filitte the retion of redit history for smll nd medium enterprises (SMEs), SBP should lso ensure tht the redit ureu ollets informtion from utility nd teleom ompnies. Finlly, to filitte SME lending monitoring, SBP should mend reporting requirements for SME portfolios to inlude volumes nd numer of lons in four su-rkets (Rs <2M, Rs 2-6M, Rs 6-25M nd Rs 25-75M). A more effiient Pkistn Post is must (following the suessful exmples of Brzil nd Chin), to pitlize on its signifint network nd strong outreh in rurl res. The Government of Pkistn (GOP) should explore wys to improve remittne nd other servies nd speed up the utomtion of postl rnhes. Given the signifint developmentl potentil of Pkistn Post to enhne finnil ess, it needs moderniztion in opertions nd regultions. Following most suess ses in Est Asi, s well s mny other ontinents, PPSB should e pled under the supervision of SBP. Creting n enling environment for expnding ess to the underserved: Regultions should keep up with the needs of the setor nd tehnologil developments, to enle expnsion. Simultneously, n enling environment should go hnd in-hnd with refully hosen government presene. Indisriminte susidies, espeilly foused on interest rtes, n e detrimentl to the expnsion of the setor, s they not only distort pries ut rowd out effiient institutions nd produts. The GOP should resist populist pereptions tht low interest-rte funding n serve developmentl purpose. Even more detrimentl re stte-owned institutions reted to promote finnil ess. In ddition to wek evidene in the se of Pkistn tht suh institutions tully improve ess (for exmple, SME Bnk), these efforts wste vlule puli resoures tht ould more usefully e deployed elsewhere, nd eliminte the level plying field for mrket prtiipnts Aess to Mirofinne Refousing on mirosvings: Interntionl experiene shows tht government svings promotion models n e onsidered, inluding wreness-rising, mthing shemes, nd tx dvntged shemes. Enourging positive puli pereptions on essiility nd sfety would help Aess to SME Lending Creting omplete nd well-funtioning seured trnstions regime: Seurity interests over movle ssets should e esy nd llowed on most ssets nd y every entity (oth physil nd juridil persons). Priority rnkings should lso e lerly defined mong those who might hve lims on property offered s ollterl. The new seured trnstions regime should lso inlude ple (suh s registry) for mking priority interests pulily known nd enforement of seurity interests for ll ssets should e fst nd hep. Continued promotion of inititives imed t proving demonstrtion effet of nk downsling. These inlude ttrtion to the mrket of n institutionl investor tht hs trk reord in SME lending. This should idelly e hieved y selling SME Bnk or giving ontrolling rights on its ord to n institutionl investor. To stimulte ompetition in the mrket from the very eginning, the GOP ould lso support long-term tehnil ssistne progrms for seleted nks. As illustrted y the Chin SME Pge 102

118 Bringing Finne to Pkistn s Poor Chpter 6 lending progrm, for this progrm to e suessful the government needs to ensure orret mrket inentives re in ple Aess to Remittnes Bringing informl remittnes into the forml system, prtiulrly on the domesti front: The high shre of nks in the interntionl remittne mrket in Pkistn indites tht inresing nk ounts should help inrese remittnes through forml hnnels. Currently, the minimum ount lne of Rs 10,000 (with the exeption of si nking ounts) is prohiitive for lrge segment of the soiety. To enourge people to open nk ounts, unomplited pperwork nd less stringent lne requirements re needed. Bnks urrently hndling remittnes n redit ounts within 24 hours; however, delivering money or sending money orders y Pkistn Post retes unneessry delys, nd the nking setor loses ustomers to Money Trnsfer Orgniztions nd informl mens of trnsfer. An enling environment for the privte setor, enourgement for opening of nk ounts, s well s improved postl offie effiieny will hve n dditionl effet to ontinue ttrting remittne flows to the forml setor. Advnes in tehnil nd finnil litery will mtter, s well, with pproprite edution to lert people to the enefits of forml systems nd how to use them. Supporting remittne servies of Pkistni nks rod: Further enourgement for Pkistni nks might e required to ontinue to oost interntionl remittnes nd forge llines with interntionl nks. One strtegy would e to set the reimursement rte through prtner nks higher thn remittnes oming solely through Pkistni nk networks. Pkistni missions rod ould do survey of options ville in their respetive mrkets nd disseminte the informtion to immigrnts on the hepest nd est soures of money trnsfer. For exmple, the Mexin mission to the United Sttes (USA) pulishes suh informtion. A further suessful poliy rried out y the Mexin emssy in the USA is tht it hs issued type of identifition rd (Mtriul Consulr) for their immigrnts, nd with this rd Mexin immigrnts in the USA n open nk ounts. Perhps the GOP n disuss something similr with the U.S. nd other uthorities. Promoting the struturing of interntionl flows into investments: Chnneling remittne flows diretly into investment in the trdle setors of the eonomy hs een pursued suessfully in the interntionl experiene, to ountert loss of interntionl ompetitiveness ssoited with elevted remittne flows. Interntionl good prties inlude pkging remittnes with pyment servies (suh s Bnsefi in Mexio). Suessful poliies to struture interntionl flows into investments hve lso foused on orgnizing nd trgeting Dispor networks rther thn tul remittnes flows. At the mro level, Dispor onds issues (for exmple, in Isrel, Indi, nd Ghn) or seuritiztion of future remittnes flows (in Brzil, Slvdor) hve een used, lthough seuritiztion hs proven ostly. At the miro level, governments hve filitted trgeted Dispor funds The Role of Puli-Privte Prtnerships A onrete wy to mke progress in expnding ess to the underserved is to form puli-privte working groups on mirofinne, smll enterprise finne, nd remittnes. This would enle dilogue etween the government nd privte ompnies (nks, MFIs, Exhnge Compnies, Money Trnsfer Orgniztions, s well s teleoms nd other mrket prtiipnts where relevnt) to tkle key hllenges in the setor. Priority themes should inlude, mong others, legl nd regultory issues, mrket trnspreny, ompetition nd ost, nd reserh nd dt issues. On rnhless nking, the privte setor hs yet to fully respond to reent SBP regultions, nd puli-privte disussions ould fous on Pge 103

119 Bringing Finne to Pkistn s Poor Chpter 6 rriers to reent SBP regultory inentives, s well s review the Pyment nd Eletroni Fund Trnsfer At, possile dt privy or seurity lw to filitte e- or m-pyments, nd the rnhless nking regultions providing for nk-nonnk prtnerships nd use of gents in money trnsfer servies. The United Kingdom hs hd suessful experiene in this regrd with Puli-Privte Working Group on Remittnes, whih led to the retion of privte setor-led U.K. Remittnes Tsk Fore. The World Bnk hs lso onvened similr group t the interntionl level, ringing together key tors from the puli nd privte setors. Pkistn n drw on this onsiderle experiene. Pge 104

120 Bringing Finne to Pkistn s Poor Annex Chpter 1 Tle A1.1: Indies of Finnil Aess: Men Vlues y Region nd Country Pge 105

121 Bringing Finne to Pkistn s Poor Annex Chpter 1 Soure: World Bnk (2008) Pge 106

122 Bringing Finne to Pkistn s Poor Annex Chpter 1 Tle A1.2. Weighted Averge Lending nd Deposit Rtes, My 2008 Items Gross Disursements (% per Annum) Outstnding Lons (% per Annum) Fresh Deposits (% per Annum) Outstnding Deposits (% per Annum) Inluding Zero Mrkup Exluding Zero Mrkup Inluding Zero Mrkup Exluding Zero Mrkup Inluding Zero Mrkup Exluding Zero Mrkup Inluding Zero Mrkup Exluding Zero Mrkup Puli Privte Foreign Speilised All Bnks Soure: Stte Bnk of Pkistn Weighted Averge Lending nd Deposit Rtes, Eonomi Dt setion. Tle A1.3. Trends in Infltion Consumer Prie Index Whole Sle Prie Index 2007 Generl Food Generl Food Mrh April My June July August Septemer Otoer Novemer Deemer Jnury Ferury Mrh Soure: Hussin, B., Ashfq, M. et l How to hieve food seurity. DAWN Eonomi & Business Review, August 25 31, P II. Pulished August 25, 2008 Pge 107

123 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.1 Determinnts of Aess to Forml nd Informl Finnil Servies, Proits Finnilly Served vs. Finnilly Exluded Formlly Served vs. All Others Formlly Served vs. Informlly Served Informlly Served vs. Finnilly Exluded Model I Model II Model III Model IV Model V Model I Model II Model III Model I Model II Model III Model I Model II Model III Femle (0.092) (0.069) (0.069) (0.069) (0.068) (0.108) (0.090) (0.090) (0.124) (0.110) (0.109) (0.101) (0.078) (0.077) Age (0.002) (0.001) (0.001) (0.001) (0.001) (0.002) (0.002) (0.002) (0.003) (0.003) (0.002) (0.002) (0.001) (0.001) Edution (0.020) (0.014) (0.014) (0.013) (0.013) (0.020) (0.019) (0.019) (0.019) (0.020) (0.021) (0.017) (0.014) (0.013) Rurl (0.068) (0.061) (0.061) (0.061) (0.060) (0.090) (0.081) (0.082) (0.102) (0.090) (0.093) (0.076) (0.068) (0.067) Employed (0.083) (0.063) (0.079) (0.063) (0.063) (0.116) (0.091) (0.094) (0.127) (0.113) (0.112) (0.090) (0.090) (0.073) Household hed (0.058) (0.046) (0.046) (0.047) (0.048) (0.080) (0.060) (0.060) (0.092) (0.064) (0.064) (0.064) (0.045) (0.047) Collterl (0.068) (0.061) (0.062) (0.062) (0.062) (0.103) (0.086) (0.084) (0.141) (0.094) (0.093) (0.091) (0.064) (0.065) Corporte (0.098) (0.085) (0.090) (0.084) (0.083) (0.134) (0.128) (0.116) (0.156) (0.137) (0.122) (0.109) (0.093) (0.087) Government (0.129) (0.131) (0.146) (0.129) (0.128) (0.180) (0.161) (0.143) (0.194) (0.175) (0.154) (0.182) (0.185) (0.172) Frmer (0.103) (0.075) (0.081) (0.076) (0.076) (0.121) (0.104) (0.092) (0.126) (0.112) (0.101) (0.097) (0.084) (0.078) Lorer (0.125) (0.097) (0.116) (0.097) (0.094) (0.131) (0.124) (0.128) (0.144) (0.129) (0.138) (0.123) (0.118) (0.094) Type of house (0.060) (0.044) (0.044) (0.043) (0.043) (0.096) (0.059) (0.060) (0.097) (0.065) (0.067) (0.062) (0.046) (0.046) Type of ltrine (0.071) (0.072) (0.072) (0.072) (0.072) (0.096) (0.085) (0.085) (0.097) (0.092) (0.093) (0.071) (0.075) (0.075) Household inome Q (0.075) (0.137) (0.165) (0.082) Household inome Q (0.073) (0.145) (0.169) (0.084) Pge 108

124 Bringing Finne to Pkistn s Poor Annex Chpter 2 Household inome Q (0.085) (0.134) (0.150) (0.096) Household inome Q (0.077) (0.160) (0.177) (0.083) Personl inome Q (0.079) (0.079) (0.080) (0.080) (0.091) (0.094) (0.097) (0.098) (0.085) (0.085) Personl inome Q (0.087) (0.087) (0.088) (0.089) (0.082) (0.087) (0.087) (0.090) (0.087) (0.089) Personl inome Q (0.117) (0.116) (0.119) (0.119) (0.110) (0.111) (0.107) (0.111) (0.115) (0.118) Personl inome Q (0.100) (0.100) (0.102) (0.103) (0.108) (0.107) (0.120) (0.119) (0.108) (0.110) Forml Setor (0.097) (0.113) (0.121) (0.103) Cell use (0.043) (0.045) (0.053) (0.064) (0.047) Cell ess (0.060) (0.076) (0.086) (0.062) Punj (0.161) (0.120) (0.121) (0.126) (0.130) (0.279) (0.178) (0.185) (0.305) (0.201) (0.208) (0.161) (0.122) (0.129) Sind (0.164) (0.127) (0.128) (0.130) (0.132) (0.251) (0.175) (0.183) (0.279) (0.197) (0.209) (0.159) (0.126) (0.130) NWFP (0.186) (0.158) (0.159) (0.161) (0.166) (0.298) (0.191) (0.195) (0.346) (0.220) (0.226) (0.193) (0.165) (0.173) Bluhistn (0.171) (0.145) (0.145) (0.146) (0.150) (0.308) (0.213) (0.216) (0.336) (0.232) (0.233) (0.168) (0.138) (0.143) AJK (0.182) (0.163) (0.164) (0.164) (0.169) (0.323) (0.243) (0.246) (0.346) (0.261) (0.267) (0.182) (0.165) (0.171) Numer of oservtions F(k,d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 109

125 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.2 Determinnts of Aess to Forml nd Informl Finnil Servies, Multinomil Logits Model I Model II Model III Model IV Model V Mrginl Effets (Finnilly Exluded) Mrginl Effets (Informlly Served) Mrginl Effets (Formlly Inluded) Mrginl Effets (Finnilly Exluded) Mrginl Effets (Informlly Served) Mrginl Effets (Formlly Inluded) Mrginl Effets (Finnilly Exluded) Mrginl Effets (Informlly Served) Mrginl Effets (Formlly Inluded) Mrginl Effets (Finnilly Exluded) Mrginl Effets (Informlly Served) Mrginl Effets (Formlly Inluded) Mrginl Effets (Finnilly Exluded) Mrginl Effets (Informlly Served) Mrginl Effets (Formlly Inluded) Femle (0.037) (0.038) (0.003) (0.031) (0.031) (0.001) (0.031) (0.031) (0.001) (0.031) (0.031) (0.001) (0.030) (0.031) (0.001) Age (0.001) (0.001) (0.000) (0.001) (0.001) (0.000) (0.001) (0.001) (0.000) (0.001) (0.001) (0.000) (0.001) (0.001) (0.000) Edution Rurl Employed (0.006) (0.006) (0.002) (0.006) (0.006) (0.001) (0.006) (0.006) (0.001) (0.005) (0.005) (0.001) (0.005) (0.005) (0.001) (0.029) (0.029) (0.002) (0.026) (0.026) (0.001) (0.026) (0.026) (0.001) (0.026) (0.026) (0.001) (0.025) (0.025) (0.001) (0.035) (0.035) (0.004) (0.028) (0.029) (0.003) (0.035) (0.035) (0.002) (0.029) (0.029) (0.002) (0.029) (0.029) (0.002) Household hed (0.023) (0.024) (0.003) (0.019) (0.019) (0.001) (0.019) (0.019) (0.001) (0.019) (0.019) (0.001) (0.019) (0.019) (0.001) Collterl (0.034) (0.035) (0.005) (0.026) (0.025) (0.001) (0.025) (0.025) (0.001) (0.025) (0.025) (0.001) (0.025) (0.025) (0.001) Corporte (0.040) (0.039) (0.005) (0.036) (0.035) (0.002) (0.038) (0.038) (0.002) (0.035) (0.035) (0.001) (0.035) (0.035) (0.001) Government Frmer (0.054) (0.066) (0.046) (0.067) (0.066) (0.016) (0.072) (0.072) (0.014) (0.067) (0.067) (0.012) (0.067) (0.067) (0.013) (0.033) (0.032) (0.004) (0.032) (0.032) (0.001) (0.033) (0.033) (0.002) (0.032) (0.032) (0.001) (0.032) (0.032) (0.001) Lorer (0.046) (0.045) (0.005) (0.039) (0.039) (0.003) (0.048) (0.048) (0.003) (0.039) (0.039) (0.002) (0.037) (0.037) (0.002) Type of House Type of Ltrine Household inome Q Household inome Q (0.023) (0.025) (0.007) (0.018) (0.018) (0.001) (0.018) (0.018) (0.001) (0.018) (0.018) (0.001) (0.018) (0.018) (0.001) (0.026) (0.025) (0.003) (0.030) (0.030) (0.001) (0.030) (0.030) (0.001) (0.030) (0.030) (0.001) (0.030) (0.030) (0.001) (0.027) (0.028) (0.005) (0.028) (0.030) (0.007) Household inome Q (0.032) (0.035) (0.008) Household inome Q (0.028) (0.033) (0.013) Personl inome Q (0.035) (0.034) (0.002) (0.034) (0.034) (0.002) (0.035) (0.035) (0.001) (0.034) (0.034) (0.001) Personl inome Q Personl inome Q (0.036) (0.035) (0.002) (0.035) (0.035) (0.002) (0.036) (0.036) (0.001) (0.036) (0.036) (0.001) Pge 110

126 Bringing Finne to Pkistn s Poor Annex Chpter 2 (0.048) (0.046) (0.004) (0.047) (0.046) (0.004) (0.048) (0.047) (0.002) (0.048) (0.047) (0.003) Personl inome Q (0.044) (0.044) (0.003) (0.044) (0.044) (0.003) (0.045) (0.045) (0.002) (0.045) (0.044) (0.002) Forml Setor (0.042) (0.042) (0.003) Cell use (0.017) (0.017) (0.002) (0.018) (0.019) (0.002) Cell ess Punj (0.025) (0.025) (0.001) (0.027) (0.042) (0.039) (0.040) (0.051) (0.069) (0.039) (0.051) (0.069) (0.038) (0.049) (0.062) (0.037) (0.050) (0.061) Sind (0.054) (0.054) (0.005) (0.047) (0.048) (0.002) (0.047) (0.048) (0.002) (0.047) (0.048) (0.001) (0.046) (0.046) (0.001) NWFP Bluhistn AJK (0.078) (0.078) (0.005) (0.048) (0.048) (0.002) (0.047) (0.047) (0.002) (0.046) (0.046) (0.001) (0.043) (0.043) (0.001) (0.062) (0.063) (0.005) (0.030) (0.030) (0.002) (0.029) (0.029) (0.002) (0.029) (0.029) (0.001) (0.027) (0.027) (0.001) (0.053) (0.053) (0.005) (0.065) (0.066) (0.002) (0.066) (0.066) (0.002) (0.065) (0.065) (0.001) (0.064) (0.064) (0.001) Numer of oservtions F(k,d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 111

127 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.3 F-Sttisti Comprisons Of Likelihood For Finnil Aess Pnel A F-sttisti omprisons of likelihood for finnil ess mong provines Bnked Finnilly served Pnel B F-sttisti omprisons of likelihood for finnil ess mong inome quintiles Bnked Finnilly served Pge 112

128 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.4 Determinnts of Interest in Finnil Mtters Interested in Finnil Mtters vs Others Model I Model II Femle (0.096) (0.073) Age (0.002) (0.001) Edution (0.014) (0.013) Rurl (0.060) (0.044) Employed (0.083) (0.070) Household hed (0.061) (0.052) Collterl (0.091) (0.065) Corporte (0.109) (0.096) Government (0.177) (0.123) Frmer (0.108) (0.081) Lorer (0.115) (0.084) Forml Setor (0.082) (0.069) Cell use (0.055) (0.042) Household inome nd house / ltrine effets Yes Personl inome nd house / ltrine effets Yes Provine effets Yes Yes Numer of oservtions F(k,d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 113

129 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.5 Determinnts of Willingness to Enter the Finnil Setor Those wnting to hve Their Own Bnk Aount vs. Others Model I Model II Mle (0.118) (0.100) Age (0.002) (0.002) Edution (0.017) (0.012) Rurl (0.068) (0.052) Employed (0.094) (0.079) Household hed (0.083) (0.066) Collterl (0.102) (0.072) Corporte (0.121) (0.104) Government (0.200) (0.166) Frmer (0.121) (0.095) Lorer (0.096) (0.096) House (0.068) (0.054) Ltrine (0.069) (0.061) Household inome (0.028) Personl inome (0.022) Forml Setor (0.097) (0.090) Cell phone use (0.056) (0.049) Mle * Household inome (0.015) Men *Personl inome (0.002) Provine effets Yes Yes Numer of oservtions F(k,d) Note: This regression is run on suset of the smple who do not presently hve nk ount Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 114

130 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.6 Determinnts of Aess to Bnking Bnked vs. Unnked Model I Model II Model III Model IV Model V Femle (0.099) (0.080) (0.081) (0.079) (0.080) Age (0.003) (0.002) (0.002) (0.002) (0.002) Edution (0.021) (0.019) (0.019) (0.019) (0.019) Rurl (0.090) (0.078) (0.077) (0.078) (0.078) Employed (0.110) (0.095) (0.102) (0.095) (0.095) Household hed (0.081) (0.068) (0.067) (0.067) (0.067) Collterl (0.105) (0.070) (0.072) (0.074) (0.073) Corporte (0.136) (0.122) (0.135) (0.120) (0.120) Government (0.176) (0.147) (0.161) (0.142) (0.142) Frmer (0.108) (0.083) (0.100) (0.084) (0.083) Lorer (0.166) (0.144) (0.139) (0.149) (0.150) Type of house (0.081) (0.056) (0.057) (0.056) (0.056) Type of ltrine (0.111) (0.099) (0.098) (0.098) (0.099) Household inome Q (0.139) Household inome Q (0.141) Household inome Q (0.145) Household inome Q (0.169) Personl inome Q (0.100) (0.099) (0.100) (0.102) Personl inome Q (0.088) (0.087) (0.090) (0.092) Personl inome Q (0.117) (0.116) (0.118) (0.119) Personl inome Q (0.112) (0.110) (0.110) (0.111) Forml Setor (0.120) Cell use (0.052) (0.051) Cell ess (0.079) Punj (0.306) (0.214) (0.214) (0.215) (0.222) Sind (0.275) (0.203) (0.202) (0.204) (0.210) NWFP (0.320) (0.219) (0.219) (0.218) (0.223) Bluhistn (0.325) (0.239) (0.240) (0.239) (0.244) AJK (0.347) (0.263) (0.263) (0.261) (0.267) Numer of oservtions F(k, d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 115

131 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.7 Determinnts of Aess to Forml nd Informl Borrowing Femle Borrowers vs. Non-orrowers Forml Borrowers vs. Others Forml Borrowers vs. Informl Borrowers Informl Borrowers vs. Nonorrowers Model I Model II Model III Model I Model II Model III Model I Model II Model III Model I Model II Model III (0.085) (0.066) (0.066) (0.165) (0.146) (0.145) (0.221) (0.196) (0.189) (0.086) (0.067) (0.067) Age Edution (0.002) (0.001) (0.001) (0.005) (0.004) (0.004) (0.005) (0.004) (0.004) (0.002) (0.001) (0.001) (0.013) (0.013) (0.013) (0.031) (0.028) (0.025) (0.033) (0.030) (0.029) (0.015) (0.013) (0.013) Rurl Employed Household hed Collterl (0.071) (0.061) (0.060) (0.088) (0.084) (0.087) (0.132) (0.104) (0.107) (0.077) (0.064) (0.063) (0.085) (0.088) (0.066) (0.242) (0.251) (0.213) (0.302) (0.299) (0.263) (0.088) (0.090) (0.067) (0.068) (0.051) (0.052) (0.129) (0.111) (0.110) (0.150) (0.125) (0.125) (0.070) (0.051) (0.052) (0.065) (0.066) (0.065) (0.182) (0.135) (0.133) (0.245) (0.163) (0.160) (0.065) (0.064) (0.064) Corporte Government Frmer (0.124) (0.091) (0.093) (0.197) (0.166) (0.151) (0.209) (0.186) (0.165) (0.123) (0.090) (0.092) (0.092) (0.094) (0.086) (0.140) (0.143) (0.135) (0.170) (0.171) (0.162) (0.100) (0.102) (0.094) (0.060) (0.064) (0.053) (0.145) (0.157) (0.157) (0.161) (0.173) (0.176) (0.058) (0.067) (0.055) Lorer (0.095) (0.090) (0.074) (0.181) (0.177) (0.167) (0.226) (0.224) (0.201) (0.100) (0.097) (0.077) Forml Setor Cell use Cell ess (0.087) (0.202) (0.221) (0.088) (0.047) (0.112) (0.118) (0.044) (0.058) (0.131) (0.147) (0.060) Household inome nd house / ltrine effets Yes Yes Yes Yes Personl inome nd house / ltrine effets Yes Yes Yes Yes Yes Yes Yes Yes Provine effets Yes Yes Yes Yes Yes Yes Yes Yes Yes Numer of oservtions F(k, d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 116

132 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.8 Determinnts of Aess to Forml nd Informl Svings Svers vs. Non-svers Forml Svers vs. Others Forml Svers vs. Informl Svers Informl Svers vs. Non-Svers Model I Model II Model III Model I Model II Model III Model I Model II Model III Model I Model II Model III Femle (0.092) (0.073) (0.073) (0.144) (0.098) (0.097) (0.165) (0.119) (0.117) (0.100) (0.076) (0.076) Age (0.002) (0.001) (0.001) (0.003) (0.003) (0.002) (0.003) (0.003) (0.003) (0.002) (0.001) (0.001) Edution (0.019) (0.013) (0.013) (0.020) (0.019) (0.019) (0.019) (0.018) (0.019) (0.018) (0.012) (0.012) Rurl (0.074) (0.060) (0.058) (0.086) (0.068) (0.069) (0.098) (0.072) (0.073) (0.081) (0.063) (0.061) Employed (0.068) (0.073) (0.067) (0.117) (0.122) (0.116) (0.133) (0.163) (0.150) (0.072) (0.081) (0.074) Household hed (0.056) (0.051) (0.052) (0.106) (0.086) (0.084) (0.101) (0.080) (0.079) (0.052) (0.048) (0.048) Collterl (0.092) (0.068) (0.068) (0.085) (0.107) (0.109) (0.101) (0.110) (0.113) (0.096) (0.067) (0.066) Corporte (0.085) (0.092) (0.076) (0.184) (0.176) (0.177) (0.220) (0.224) (0.226) (0.107) (0.121) (0.106) Government (0.106) (0.115) (0.098) (0.160) (0.142) (0.129) (0.180) (0.167) (0.156) (0.137) (0.135) (0.122) Frmer (0.101) (0.079) (0.072) (0.131) (0.105) (0.098) (0.140) (0.114) (0.108) (0.104) (0.081) (0.074) Lorer (0.099) (0.095) (0.079) (0.183) (0.181) (0.156) (0.205) (0.206) (0.188) (0.100) (0.096) (0.082) Forml Setor (0.086) (0.130) (0.151) (0.089) Cell use (0.058) (0.060) (0.071) (0.056) Cell ess (0.075) (0.107) (0.119) (0.077) Household inome nd house / ltrine effets Yes Yes Yes Yes Personl inome nd house / ltrine effets Yes Yes Yes Yes Yes Yes Yes Yes Provine effets Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Numer of oservtions F(k, d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 117

133 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.9 Determinnts of Aess to Speifi Forml nd Informl Finnil Produts Credit/Deit Crd Users vs. Non-users Chek Users vs. Nonusers Svings Aount Users vs. Non-users Current Aount Users vs. Non-users Islmi Clients vs. Others Post Offie Clients vs. Others Insurne Users vs. Nonusers Committee Users vs. Nonusers Model I Model II Model I Model II Model I Model II Model I Model II Model I Model II Model I Model II Model I Model II Model I Model II Femle (.242) (.146) (.169) (.109) (.147) (.100) (.121) (.097) (.376) (.156) (.185) (.180) (.150) (.111) (.089) Age Edution.218 Rurl Employed (.006) (.006) (.003) (.003) (.003) (.003) (.003) (.003) (.008) (.007) (.004) (.003) (.004) (.003) (.003) (.002) (.055) (.052) (.026) (.022) (.020) (.019) (.024) (.024) (.048) (.075) (.036) (.031) (.024) (.024) (.019) (.018) (.145) (.125) (.119) (.095) (.087) (.069) (.103) (.086) (.392) (.460) (.095) (.074) (.084) (.094) (.075) (.058) (.177) (.255) (.124) (.140) (.119) (.125) (.131) (.150) (.449) (.185) (.207) (.154) (.126) (.105) (.109) Household hed (.180) (.145) (.087) (.062) (.108) (.088) (.111) (.082) (.318) (.257) (.115) (.145) (.206) (.173) (.087) (.075) Collterl (.169) (.175) (.147) (.104) (.085) (.107) (.172) (.101) (.143) (.118) (.169) (.148) (.160) (.117) (.085) (.064) Corporte (.199) (.102) (.179) (.171) (.187) (.179) (.094) (.117) (.719) (.655) (.225) (.220) (.170) (.176) (.101) (.081) Government (.187) (.132) (.205) (.183) (.162) (.144) (.201) (.182) (.601) (.563) (.175) (.170) (.148) (.134) (.144) (.118) Frmer (.177) (.141) (.153) (.146) (.133) (.108) (.112) (.096) (.605) (.405) (.122) (.139) (.208) (.194) (.092) (.073) Lorer (.128) (.248) (.208) (.170) (.187) (.187) (.110) (.141) (.265) (.244) (.173) (.178) (.143) (.120) Forml Setor (.176) (.127) (.134) (.116) (.564) (.125) (.112) (.081) Household inome nd house / ltrine effets Yes Yes Yes Yes Yes Yes Yes Yes Personl inome nd house / ltrine effets Yes Yes Yes Yes Yes Yes Yes Yes Provine effets Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Numer of oservtions F(k, d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 118

134 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.10 Religious Considertions s Determinnt of Aess to Finne, for Rnge of Produts Forml Borrowers Informl Credit/De Svings Current Borrowers Forml vs. Borrowers it Crd Chek Aount Aount Islmi Post Offie vs. Nonorrowers Borrowers Informl vs. Non- Users vs. Users vs. Users vs. Users vs. Clients vs. Clients vs. vs. Others Borrowers orrowers Non-users Non-users Non-users Non-users Others Others Femle Insurne Users vs. Non-users Committee Users vs. Non-users (0.114) (0.196) (0.268) (0.118) (0.233) (0.152) (0.134) (0.124) (0.232) (0.179) (0.126) Age (0.002) (0.004) (0.004) (0.002) (0.008) (0.004) (0.004) (0.004) (0.007) (0.003) (0.005) (0.003) Edution (0.020) (0.036) (0.034) (0.021) (0.059) (0.028) (0.033) (0.027) (0.073) (0.037) (0.041) (0.024) Rurl (0.087) (0.130) (0.138) (0.090) (0.168) (0.105) (0.087) (0.116) (0.463) (0.132) (0.153) (0.084) Employed (0.118) (0.247) (0.324) (0.120) (0.262) (0.150) (0.162) (0.144) (0.220) (0.200) (0.160) (0.152) Household hed (0.073) (0.152) (0.157) (0.073) (0.091) (0.094) (0.121) (0.091) (0.347) (0.183) (0.192) (0.111) Collterl (0.111) (0.136) (0.158) (0.112) (0.163) (0.175) (0.184) (0.135) (0.172) (0.231) (0.202) (0.112) Corporte (0.128) (0.212) (0.223) (0.123) (0.133) (0.203) (0.205) (0.113) (0.713) (0.251) (0.193) (0.143) Government (0.132) (0.138) (0.175) (0.153) (0.140) (0.205) (0.184) (0.202) (0.759) (0.217) (0.163) (0.189) Frmer (0.096) (0.186) (0.195) (0.097) (0.191) (0.170) (0.144) (0.097) (0.399) (0.181) (0.203) (0.122) Lorer (0.129) (0.241) (0.256) (0.124) (0.236) (0.222) (0.198) (0.156) (0.335) (0.232) (0.134) Anti-Islmi (0.070) (0.120) (0.139) (0.073) (0.172) (0.099) (0.100) (0.114) (0.183) (0.127) (0.141) (0.077) Personl inome nd house / ltrine effets Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Provine effets Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Numer of oservtions F(k, d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 119

135 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.11 Determinnts of Aess to Finne of the Self-Employed Self-employed in Self-employed in Forml Setor vs. Informl Setor vs. Others Others Model I Model II Model I Model II Mle (0.113) (0.086) (0.136) (0.120) Age (0.002) (0.002) (0.003) (0.003) Edution (0.017) (0.017) (0.026) (0.020) Household hed (0.077) (0.069) (0.088) (0.089) Cell use (0.053) (0.047) (0.065) (0.048) Cell ess (0.101) (0.083) (0.109) (0.100) Urn (0.161) (0.115) (0.178) (0.140) Urn*mle (0.183) (0.124) (0.200) (0.148) Household inome nd house / ltrine effets Yes Yes Personl inome nd house / ltrine effets Yes Yes Provine effets Yes Yes Yes Yes Numer of oservtions F(k,d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 120

136 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.12 Resons for Sving Forml Svers vs. Others Informl Svers vs. Non-svers Model I Model II Model I Model II Femle (0.167) (0.119) (0.224) (0.159) Age (0.003) (0.003) (0.007) (0.006) Edution (0.016) (0.018) (0.045) (0.033) Rurl (0.087) (0.070) (0.093) (0.112) Employed (0.176) (0.162) (0.262) (0.203) Household hed (0.118) (0.083) (0.175) (0.179) Collterl (0.095) (0.111) (0.235) (0.152) Corporte (0.181) (0.208) (0.354) (0.297) Government (0.145) (0.139) (0.285) (0.245) Frmer (0.138) (0.112) (0.214) (0.184) Lorer (0.271) (0.207) (0.397) (0.344) Forml Setor (0.220) (0.158) (0.181) (0.233) Cell use (0.074) (0.063) (0.141) (0.110) Food nd household needs (0.058) (0.065) (0.146) (0.109) Medil needs (0.085) (0.072) (0.172) (0.134) Edution needs (0.085) (0.058) (0.279) (0.187) Investment needs (0.110) (0.099) (0.135) (0.258) Provision for ident nd deth (0.065) (0.068) (0.177) (0.150) Soil nd religious needs (0.087) (0.068) (0.236) (0.158) Old ge needs (0.109) (0.100) (0.352) (0.259) Household inome nd house / ltrine effets Yes Yes Personl inome nd house / ltrine effets Yes Yes Provine effets Yes Yes Yes Yes Numer of oservtions F(k,d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 121

137 Bringing Finne to Pkistn s Poor Annex Chpter 2 Tle A2.13 Resons for Borrowing Forml Borrowers vs. Others Informl Borrowers vs. Non-orrowers Model I Model II Model I Model II Femle (0.211) (0.208) (0.342) (0.252) Age (0.005) (0.004) (0.010) (0.006) Edution (0.024) (0.023) (0.094) (0.121) Rurl (0.139) (0.132) (0.615) (0.309) Employed (0.327) (0.349) (0.306) (0.229) Household hed (0.179) (0.152) (0.217) (0.142) Collterl (0.244) (0.134) (0.231) (0.159) Corporte (0.233) (0.213) (0.615) (0.329) Government (0.204) (0.219) (0.605) (0.670) Frmer (0.163) (0.176) (0.268) (0.252) Lorer (0.242) (0.242) (0.155) (0.231) Forml Setor (0.222) (0.287) (0.427) (0.365) Cell use (0.113) (0.103) (0.187) (0.188) Food nd household needs (0.148) (0.110) (0.198) (0.270) Medil, ident, funerl needs (0.235) (0.226) (0.328) (0.466) Agriulturl investment needs (0.140) (0.134) Non-frming investment needs (0.207) (0.152) (0.438) (0.412) Soil nd religious needs (0.171) (0.140) (0.609) (0.793) Home investment nd improvement (0.257) (0.235) Household inome nd house / ltrine effets Yes Yes Personl inome nd house / ltrine effets Yes Yes Provine effets Yes Yes Yes Yes Numer of oservtions F(k,d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 122

138 Bringing Finne to Pkistn s Poor Annex Chpter Smple Desription of KfW SME Survey Urn Demnd Survey In Mrh April 2005, KfW undertook survey of 510 SMEs in Pkistn in towns/ities (Lhore, Fisld, Gujrnwl, Srgodh, Silkot) of Punj distrit to understnd their finnil ess nd demnd for finnil servies. Survey Instrument The survey instrument overed the following res: Finnil performne (profits, sles, exports, investments) Reltions etween enterprises nd finnil providers (ese of otining externl finne in ) Struture of demnd for externl finning (intent to orrow, mounts, onditions demnded, reditworthiness) Pereption of ostles to doing usiness Choie of Lotions The survey ws onduted in five ities/towns of Punj. The ities/towns were seleted sed on the size of their popultion, to llow verifition of the differene etween demnd nd ess to finne sed on this riterion. Definition of Popultion Owners of miro, smll, nd medium usinesses operting in the eonomi setors of trde, prodution, nd servies in one of five seleted urn res is onsidered s the popultion of urn demnd survey. Seletion Method First segmenttion riterion ws developed sed on (town/ity, size of enterprise, eonomi setor). Then smpling res were enumerted with onentrtion of trget enterprises within eh urn lotion, nd speifi res were lloted to the interviewers. At lest five interviews were onduted in eh re to hve representtion of the whole ity/town. Seleted respondents were owners/hief mngers, hd employees up to 100, nd were t lest 18 yers of ge. Totl smple size for the urn demnd survey ws 510. Bsi Smple Dt The overwhelming mjority of the respondents were mle (98 perent) with verge ge of 35 yers. The tles elow report the smple omposition y setor, size, legl sttus, lotion of usiness, nd verge yers of shooling of owners/mngers. Pge 123

139 Bringing Finne to Pkistn s Poor Annex Chpter 4 Tle A4.1: Prinipl Setor of Respondent s Business Setor Frequeny Perentge Trde (wholesle nd retil) Mnufturing Servies Totl Soure: KfW Demnd Survey, 2005 Tle A4.2: Definition of Firm Size Firm Size Numer of Permnent Smple Size (%) Averge Numer of Employees Employees Miro (52.9) 4 Smll (36.9) 20 Medium (10.2) 75 Totl (100.0) 17 Soure: KfW Demnd, 2005 Tle A4.3: Legl Sttus of Smpled Enterprises Sttus Numer of Enterprises Shre of Totl (%) Not registered with uthorities Pulily listed ompny Privtely held limited ompny Sole proprietorship Prtnership Totl Soure: KfW Demnd Survey, 2005 Tle A4.4: Lotion of Business of Smpled Enterprises Lotion Numer of Enterprises Shre of Totl (%) Business offie My house Mrketple/shop Ftory/industril uilding Other Totl Soure: KfW Demnd Survey, 2005 Tle A4.5: Averge Yers of Shooling of Owners of Smpled Enterprises Lotion Numer of Enterprises Shre of Totl (%) Illiterte Red/write only Primry (Clss 5) Middle shool (Clss 8) Mtriulte (Clss 10) Intermedite (Clss 12) Grdute Postgrdute Totl Soure: KfW Demnd Survey, 2005 Pge 124

140 Bringing Finne to Pkistn s Poor Annex Chpter 4 4.2: Frmework for Movle Collterl in Pkistn Three ritil shortomings exist with respet to the frmework for movele ollterl in Pkistn: 1. The requirement for ourt orders to e otined efore proeeding to enforement, even for most types of seured reditors, signifintly erodes the utility of hrges over movele ollterl; 2. The lk of speilized ourts to del with enforement nd insolveny mtters onsiderly slows enforement; nd 3. The sene of omprehensive, single registry for ll movele hrges in regrd to ll types of detors (orporte or otherwise) limits the sope of movele seurity. For these nd other resons, lthough the legl mehnisms for reting seurity over movele ollterl exist nd re widely used, they ple in omprison to the use of rel property, prtiulrly for long-term usiness dets. A. Cretion of Seurity Interests over Movle Property Proess of retion: To the extent possile, the proess of reting seurity interest should e esy nd ost-effetive. One of the most effetive wys to do this is through hving simple, unified sttute deling with seurity interests. A simpler nd heper system of retion ould trnslte into lower trnstion osts, nd susequently, lower interest rtes for potentil detors. In Pkistn, however, there is no simple, unified sttute pertining to the retion of seurity over movle property. Aording to some users of the system, the multiple lws nd systems rete unertinty nd diffiulties in using movle property s ollterl. In generl, these lws inlude the Contrt At, 1872, Trnsfer of Property At, 1882, Compnies Ordinne, 1984, Civil Proedure Code, 1908, Bnking Compnies Ordinne, 1962, Finnil Institutions (Reovery of Finne) Ordinne, 2001, the Prudentil Regultions issued y the Stte Bnk of Pkistn, Non-Bnking Finne Compnies (Estlishment nd Regultion) Rules 2003, Sle of Goods At, Prtnership At 1932, The Cpe Town Convention nd Airrft Protool (Implementtion) Rules, 2003, nd ll other nking regultions issued from time to time. Types of seurity interests over movle property: Hypothetion: A hypothetion greement or letter of hypothetion opertes s hrge over movle property inluding ook dets nd reeivles. The essentil seurity onferred upon lender is the right to seize nd sell the hypotheted property in reliztion of the seurity. Stmp duty on n instrument reting hypothetion is unless exempted or remitted very high. Registrtion is required for ompnies reting suh seurity under the Compnies Ordinne, 1984 ut not for other forms of usiness orgniztions or individuls. Floting hrge: In ddition to hypothetion, nd to ensure tht the orrower is free to del with the property in the ordinry ourse of usiness, nd to ensure tht ll the orrower's ssets nd undertking re overed y the terms of the hrge, the lenders usully require the orrower to rete floting hrge y seprte deed. The hrge only rystllizes upon the hppening of one or more stipulted events of defult whereupon the lender is entitled to seize the property nd sell it to reover his dues. No interest in the property rues to the lender until rystlliztion nd the desription of the property is given only in the most generl terms. This hrge is stmped t nominl rte. Registrtion is required for ompnies reting suh seurity under the Compnies Ordinne, 1984 ut not for individuls or other entities. Generlly speking, suh hrge n e reted through two-step proess, requiring two douments: exeution of the underlying greement etween orrower nd detor nd registrtion of tht greement in the Compnies Registry. Pge 125

141 Bringing Finne to Pkistn s Poor Annex Chpter 4 Leses: Leses my e operting leses or finne leses. Nonnking finnil institutions tht provide lesing filities re regulted y lw nd ontrolled y the Seurities nd Exhnge Commission of Pkistn (SECP). At times, the Islmi mode of lesing, whih is strutured s n Ijr, is used. Here, ownership does not pss to the user. The min fetures of this form of lesing re tht written ontrt is mndtory; the sset should exist t the time of the ontrt; the sset must hve rel nd lwful use nd should not e used for n tivity prohiited y Islmi lw. Totl loss is the owner's responsiility, however; the user is responsile for dmge through misuse or negligene. If one prty dies, the ontrt eomes void. The purest form of Ijr is lmost the sme s n operting lese. An Ijr ontrt my lso e strutured s finne lese with ertin refinements nd ltertions. Restritions on who n eome prty to seurity greement: To the extent possile, nd tking into ount the ountry s speifi legl ulture, the lw should not limit the types of reditors nd detors tht n eome prty to seurity greement. Expnding the lsses of entities tht n serve s reditors ould trnslte into more ville redit in the mrket. On the other hnd, llowing more types of entities s detors ould men esier ess to redit for mny non-inorported entities, prtiulrly those tht re smll or medium-sized. Creditor: Under the Bnking Compnies Ordinne, 1962, reditor hs een defined s persons from whom deposits hve een reeived on the sis of prtiiption in profit nd loss nd nking ompny or finnil institution from whih finnil ommodtion or fility hs een reeived on the sis of prtiiption in profit nd loss, mrk-up in prie, hire-purhse, lese, or otherwise. There re no serious restritions on who my eome seured reditor under Pkistni lw. The Finnil Institutions (Reovery of Finnes) Ordinne, 2001 defines finnil institution s ny ompny whether inorported within or outside Pkistn whih trnsts the usiness of nking or ssoited or nillry usiness in the ountry through its rnhes within or outside Pkistn nd inludes government svings nk, ut exludes the Stte Bnk of Pkistn. The definition lso inludes other entities suh s modr,81 venture pitl, finning or lesing ompny, n investment nk, unit trust or mutul fund of ny kind, redit or investment institution, orportion, or ompny, or ny other ompny uthorized y the Federl Government to rry on similr usiness. Detor: While there is no restrition on who my grnt seurity right, there is restrition on the types of detors over whose ssets seurity interests n nd must e registered so they re enforele ginst third prties. A hrge my e registered under the Compnies Ordinne, 1984 with the Registrr of Compnies y ompny formed nd registered under the Compnies Ordinne. Prolems ssoited with the retion of seurity interest over movle property: It ppers tht one of the most importnt limittions on the use of movle property s ollterl in Pkistn is tht the detor must possess ny ollterl eing pledged. In ddition, the legl frmework for enforement ginst not-yet-existent ollterl is wek. This ffets the use of griulturl produts (suh s future rops), inventory, or fter-quired property s ollterl, nd would mke it very diffiult for mny smll usinesses tht do not hve signifint mount of other lsses of ssets eyond fterquired property to otin seured redit. Another importnt limittion is tht the system of seured trnstions over movle property is ville primrily to registered ompnies only. This is euse n effiient system of registrtion of seurity interests is generlly only ville for registered ompnies (unless the ssets eing used re themselves sujet to registrtion), nd not for ny other orgniztions or usiness entities flling outside of the Compnies Ordinne, For definition of modrs, plese see (p. 15) Pge 126

142 Bringing Finne to Pkistn s Poor Annex Chpter 4 Best Prtie Pkistni Prtie Reommendtion 1. The mnner of the retion of seurity interest* should e simple, fst, nd hep. 2. Any legl entity should e le to grnt nd tke seurity interest. 3. There should e no limit on the type of oligtion tht n e seured. 4. A seurity interest should e llowed to e grnted over ll types of movle property to mximize the detor s potentil to otin redit. 5. Assets used s ollterl should e llowed to e defined generlly in the seurity greement so tht revolving lsses of ssets (inluding ounts reeivles, inventory, et.) n e used s ollterl. As explined ove, the system of retion of seurity interest in Pkistn is miguous euse of the ft tht there re multiple lws tht govern seured trnstions in the ountry nd multiple types of seurity interests eh with their own system. Although ny entity is le to eome prty to seurity greement, prtilly, noninorported entities find it diffiult to grnt seurity interests over their movle ssets euse there re no effiient registrtion systems in ple for seurity interests over the movle ssets of suh entities. It ppers tht there re no limits on the types of oligtions tht n e seured. Future oligtions, foreign urreny oligtions, nd hnging pools of det n ll e seured, lthough foreign urreny oligtions re sujet to dditionl legisltion. However, these dditionl regultions do not signifintly ffet the ility to seure foreign urreny oligtions. In the text of the lw, it ppers tht there re no limits on the types of movle ssets tht n e used s ollterl. In prtie, however, there my e some diffiulties. For exmple, for the purposes of n ssignment of ook det, if the stmp duties re too high, there will e disinentive to use this type of ssets s ollterl. At lest in the ontext of floting hrge, the ssets do not need to e defined speifilly. For the purposes of registrtion of seurity interests over the ssets of noninorported entities, however, the ssets need to e defined speifilly. A single lw should e reted tht would govern seured trnstions over movle property in Pkistn regrdless of the type of sset or the type of seurity interest eing reted. The ountry s legl system should filitte the retion of seurity greement y mking n effetive seured trnstions system ville to ll types of legl entities. No dditionl reommendtions mde. Prtil ostles tht hinder the effiient pplition of the lw should e removed. The system of seurity interests over movle property tht llows for generl desription of ssets should e mde ville to ll types of usiness entities, inluding noninorported entities. Pge 127

143 Bringing Finne to Pkistn s Poor Annex Chpter 4 6. Future ssets should e llowed to e used s ollterl. 7. Multiple seurity interests should e llowed to e mde on the sme ollterl, to mximize the potentil of the detor to otin redit sed on the vlue of prtiulr sset. As ws explined erlier, one of the mjor limittions in Pkistn is tht the detor must own the ollterl t the moment of grnting seurity interest over the sset. This will limit the use of mny future ssets s ollterl. In prtie, nks, when tking floting hrge over the ssets of ompny, inlude in the greement luse tht prohiits the detor from using the ssets s ollterl for ny susequent lons from other reditors without the onsent of the first reditor. This is done euse, under the Pkistni system, often fixed hrges tke priority over floting hrges. So, if detor grnts fixed hrge over its ssets to ny susequent reditors, tht susequent reditor will hve priority over previous reditor with floting hrge. The limittion on the use of future ssets s ollterl should e removed. The lw of Pkistn ould e reformed to ommodte first-registered firstpriority system, whih would generlly not distinguish etween the types of interests, nd whih would ditte priority sed on the time of registrtion (with some resonle exeptions) nd not the type of interest. This would e system similr to wht hs or is eing dvned in ommon lw ountries suh s the United Sttes, Cnd, New Zelnd, nd Austrli. This system ould disourge the use of prohiitive luses y nks, whih would effetively disllow detor to use its ssets s ollterl with susequent reditors. However, it should e kept in mind tht, despite the enefits of this system, this would onstitute omplete overhul of the system, whih my not neessrily e desirle t this time. * Seurity interest/hrge in this setion refers to non-possessory seurity interest over movle property. B. Priority of Seurity Interests The rules onerning priority should e lerly lid out in the lw. In the sene of ler rules, reditors, not eing le to evlute their hnes of reovery upon defult nd ssess their risk, will offer redit t higher prie to ompenste for this unknown vrile. Seprte from the question of lrity is the question of rnking of reditors. This is n issue tht enompsses rod set of legl nd ulturl vlues. For exmple, wht type of priority should e given to employee wge lims? Wht rights should the stte hve to reover tx rrers from the detor? In generl, however, s trditionl ommeril lenders in mny eonomies tend to lend sed on the vlue tht the underlying seurity offers them (ffeted y suh ftors s the rnking of the seured reditor), the importne of ording reltively high priority to seured reditors is generlly viewed s hving positive impt on oth the vilility nd ost of redit. The issue of solute priority of seured reditors is often frmed s inry one (solute priority is good nd sene of it is d). This distintion, however, is flse one s virtully no dvned eonomy hs true solute priority for seured reditors. Rther, poliymkers must e mde wre of the need to lne ompeting interests nd the onsequenes of preferring one group of stkeholders over nother. Pge 128

144 Bringing Finne to Pkistn s Poor Annex Chpter 4 There is si sttutory priority order set out in lw. For non-possessory seurity interests grnted y ompnies, lerly the first to file system (with respet to filings in the Compnies Registry) pplies. With regrd to unseured lims, the following sheme of priorities generlly pplies in liquidtion: Government revenue nd txes; Wges to mximum of Rs 2,000 or 12 months, whihever is less; Holidy pyments; Employer ontriutions to pensions; nd Workmen s ompenstion. In prtie, however, seured reditors enfore outside of this liquidtion, leving very little if nything for these priority lims. Moreover, given the multipliity of lws deling with seured trnstions, it is diffiult to onretely determine priority euse of the prevlene of more esoteri forms of seurity tht might tke preedene (onstrution liens, mehnil liens, et.). This undersores the need for ringing ll types of seured trnstions over movles under one lw. Best Prtie 8. To ensure resonle ess to redit, seured reditors should hve reltively high rnking in priority vis-à-vis other interests one their interest is registered. 9. The lw should set out lerly the priorities of the different Pkistni Prtie The lws governing priorities re spred throughout the vrious lws tht govern seured trnstions. As suh, further investigtion is required to ensure the vlidity of ny oservtions mde in this re. However, it ppers tht numer of interests rnk ove registered seurity interest. These inlude reditors dding vlue to the ollterl (equipment repirers/mehni s lien), nd reditors storing the ollterl. Furthermore, it ppers tht reditors seured y mens of title finning hve priority over ll other seured reditors regrdless of the order of registrtion. Wht is interesting, however, is tht it ppers tht other interests suh s purhse money seurity interests, on fide purhsers without notie, or those who provide finning for rehilittion fter the ommenement of insolveny proeedings whih, in mny dvned systems do hve priority over seured reditors, do not enjoy suh priority over seured reditors in Pkistn. Finlly, it ppers tht, in the ourse of nkrupty proedure, the ollterl of the seured reditor is exempt, nd, s suh, the seured reditor enjoys priority. However, should the seured reditor wish to prtiipte in the proedures, txes nd wges will gin priority over the seured reditor. Although si legisltive sheme for priorities exists, it ppers tht, the ft tht mny seurity interests suh s stte lims, title finne greements, judgments, purhse money seurity interests, nd finnil lesing greements re not registered retes lot of onfusion t the moment of enforement of rights, sine determining priority eomes very Reommendtion Further reserh my prove to e enefiil to onfirm the findings in this setion, nd to develop priority system tht would est suit Pkistn. A system of registrtion tht would meet Pkistn s needs, nd tht would inrese preditility for seured reditors y Pge 129

145 Bringing Finne to Pkistn s Poor Annex Chpter 4 interests. diffiult. requiring the registrtion of vrious types of interests should e explored nd promoted. C. Puliity of Seurity Interests A registrtion system tht is not simple nd hep, nd registry tht is not effetive ould diminish the onfidene of seured reditors in the system. One side-effet of this ould e the vilility of redit t higher interest rtes. In the lterntive, seured reditors my eome more relutnt to mke redit ville t ll. In Pkistn, the system of registrtion of seurity interests over movle property is s follows: Registrtion of seurity interests: () Registrr of Compnies under the Compnies Ordinne, 1984: Registrtion pplies only to ompnies, tht is, odies orporte inorported under the Compnies Ordinne, 1984 of Pkistn, nd not to sttutory orportions, or prtnerships or firms or other uninorported forms of usiness orgniztion. Registrtion in the se of ompnies hs to e done t the Registrr of Compnies where the registered offie is situted, nd in the se of rnhes of ompnies, t the Compnies Registry where the prinipl offie in Pkistn is situted. The Compnies Ordinne, 1984 requires reditor to rete nd register hrge within 21 dys from the dte of exeution of the seurity greement. The entire proess from the sumission of the Registrtion Form to the issune of ertifite of registrtion must e ompleted within the sid time. () Vehiles Registry () Pkistn Airrfts Register under the Rules, 1994: In the se of mortgge over n irrft (exluding militry irrft), nottion must e mde in the Remrks Column of the Airrft Register mintined y the Civil Avition Authority of Pkistn (we hve een dvised tht, despite this prtie, there is no speifi provision providing for the registrtion of hrge over n irrft. This issue ould e investigted further). (d) Interntionl Registry under the Airrft Rules. We do not hve further informtion regrding the sope of this registry. It is importnt to highlight tht, while the Pkistn Airrfts Register nd the Interntionl Registry over seurity interests pertining to irrfts lone, the Registrr under the Compnies Ordinne registers hrge over ll types of movele ssets s long s the reditor is ompny inorported under the Compnies Ordinne. The ost for registering floting hrge is Rs 5,000 nd the ost for onduting serh is Rs 200. The list for presried fees is ontined in the Sixth Shedule of the Compnies Ordinne Serhes nnot e onduted online. The proess for the serh is firly strightforwrd. Any person n file n pplition for inspetion of files long with the presried fee of Rs 200. This involves physilly visiting the SECP nd inspeting Pge 130

146 Bringing Finne to Pkistn s Poor Annex Chpter 4 the relevnt register. It is theoretilly possile to do this in one dy, provided the fee is deposited erly in the dy. However, this my well streth to two dys. As regrding the proess for registrtion of the floting hrge, the pplition in the presried form (Form 10) long with relevnt douments, nmely, the doument reting the hrge nd the presried fee re sumitted physilly t the offie of the SECP. Normlly, if ll the douments re in order nd the SECP mkes no ojetion to the douments, it tkes few dys to register the hrge. Best Prtie Pkistni Prtie Reommendtion 10. All types of legl entities (reditors/detors) should e le to register seurity interests. 11. Registrtion should e mndtory for the purposes of mking the seurity interest enforele ginst third prties, in order to filitte the ese of determining priority y reditors. 12. With very limited exeptions, the first to file should hve first priority. 13. The seurity interest should extend to replements nd proeeds of the underlying sset in order to mke the use of revolving pools of ssets s ollterl prtil. As mentioned erlier, the registrtion of seurity interests n only e mde in n effiient mnner over the movle ssets of inorported entities. Other types of entities do not hve ess to suh n effiient system unless they use ssets tht themselves re registrle s ollterl. In tht se, seurity interests over those ssets ould e registered in suh sset-sed registries. As for seurity interests over other ssets of noninorported entities, the registrtion system is pper nd regionlly sed. It ppers tht the registrtion of seurity interests over movle ssets of non-inorported entities is not mndtory for the purposes of perfetion. Although not neessrily n issue, s mentioned in the previous setion, some seurity interests hve priority over the seured reditor, regrdless of the time of registrtion y the seured reditor. It ppers tht seurity interests extend to replements nd proeeds of the underlying sset in Pkistn. It would e enefiil to extend n effiient system of registrtion of seurity interests to other types of entities so tht priorities of seurity interests reted on the ssets of those entities n lso e more esily determined, therey mking it more ttrtive for reditors to tke the movle ssets of these entities s ollterl. Registrtion should not only e mde ville ut should e mde mndtory for seurity interests over the movle ssets of ll entities for the purposes of perfetion. The ountry s soil, eonomi, nd ulturl vlues need to e tken into ount to develop omprehensive system of priority tht works for Pkistn while, t the sme time, grnting the neessry protetion nd lrity to seured reditors. No further reommendtions mde. Pge 131

147 Bringing Finne to Pkistn s Poor Annex Chpter There should e notiefiling registry to redue the urden on the registering prty nd the registrtion system. It ppers tht registries do not funtion on notie-filing system. Rther, the entire greement is wht needs to e filed. An effiient system of notie-filing should e implemented for ll registries. 15. The registry should e geogrphilly linked ross the ountry so tht potentil seured reditor n sertin its priority over the sset more urtely. 16. The registry should e omputerized so tht reords n e registered nd essed fster. 17. There should e unified registry for ll types of seurity interests, grnted y ll types of entities on the most importnt types of movle property so tht potentil reditors n refer to one entrl lotion to sertin whether there re other seurity interests on prtiulr sset. 18. Registrtion (or mendments to the registrtion) should e inputted into the registrtion system in timely nd hep mnner in order to mke the system more relile nd essile. D. Enforement of Seurity Interests The Compnies Registry is online nd networked ross the ountry. However, this is not the se for the registries for seurity interests over movle ssets of non-inorported entities. An online nd regionlly networked system of registrtion should e mde ville for ll entities. Plese see the previous response. Plese refer to the previous reommendtion. There is no unified registry for ll types of seurity interests grnted y ll types of entities on ll types of movle property. Pkistn hs mixed system of Asset-Bsed Registries (Airrft Registry) nd Entity-Bsed Registry (Compnies Registry). Some lwyers hve omplined out the ft tht registrtions nd mendments re not inputted in timely nd urte mnner, espeilly in registries other thn the Compnies Registry. Preferly, unified registry tht would pply to ll types of seurity interests over ll types of movle property y ll types of entities should e reted. The system of registrtion should e timely nd hep in ll registries nd for ll types of seurity interests, entities nd ssets. Without n effetive enforement system, seured reditors will hve diminished ility to reover their interest. Creditors will nturlly inlude this usiness risk in their lon priing. Although it is tehnilly possile to otin pre-judgement lien over disputed ssets, in prtie this is rrely grnted in ses other thn those deling with rel property. This ould men higher interest rtes for detors trying to use movele ollterl s seurity. Pkistn s enforement system is s follows: Enforement mehnisms: Enforement (y finnil institutions) without the involvement of the ourt: The lws relting to nks nd finnil institutions provide for speil summry proedures for determining the judgment nd enforing the seurity. Pge 132

148 Bringing Finne to Pkistn s Poor Annex Chpter 4 Under setion 15 of the Finnil Institutions (Reovery of Finne) Ordinne of 2001, seured reditor my sell the pledged property (under possessory pledge) with miniml involvement of the ourt. The seured reditor is required to send two noties suessively to the defulting detor ompny demnding pyment of the outstnding sum within 14 dys from servie of eh notie. Therefter, finl notie my e served. If pyment is not mde within 30 dys fter tht notie, the finnil institution my, without the intervention of the ourt, sell the mortgged property or ny prt thereof y puli ution nd pproprite the proeeds thereof towrd totl or prtil stisftion of the outstnding mortgge money. As prerequisite, efore exerising its powers, the seured reditor is required to pulish notie giving t lest 30 dys for the sumission of offers with respet to the mortgged property y dvertising it in one reputle English nd Urdu dily newspper with wide irultion in the provine in whih the mortgged property is situted. The finnil institution shll lso send suh noties to ll persons who, to the knowledge of the finnil institution, hve n interest in the mortgged property s mortggees. However, ording to some of the users of the system, even though the greement etween the prties my provide for the seured reditor to enfore its right over the seurity diretly, in prtie, detor ompny is likely to otin sty order from the ourt nd initite legl proeedings ginst the seured reditor on ny point of lw or ft to restrin the liention of the seured sset for s long s possile. Enforement with the involvement of the ourt, where out-of-ourt enforement is not ville: The seured reditor my file suit for the reovery of the sum due from the detor ompny. The detor ompny is required to file n pplition for leve to defend within 30 dys from the dte of notie. Leve is only grnted if sustntil questions of lw or ft re rised nd my e ompnied y requirement for the deposit of sh or furnishing of seurity. Upon pronounement of judgment nd deree, the suit utomtilly onverts into exeution proeedings without the need for the seured reditor to file seprte pplition for the exeution of the judgment. Enforement under the Civil Proedure Code: A suit for the reovery of the sum seured or repossession of the seured property my e filed under the Civil Proedure Code. Agin, the ourt is extensively involved nd the proedure is lengthy euse of () the possiility of severl ppels from interloutory orders in suh proeedings; () the need to ommene seprte exeution proeedings fter otining finl deree tht is no longer ple of ppel; nd () the onsumption of dditionl time in filing exeution proeedings. In prtie, we hve een dvised y lol lwyers tht if it is finnil institution enforing its seurity right under the speil summry proedures, it would tke out 1.5 to 2 yers for the enforement to e ompleted. In other ses, it my tke pproximtely 10 to 15 yers or more. Best Prtie Pkistni Prtie Reommendtion 19. Prties should e le to gree to out-of-ourt enforement t the time of the retion of their seurity greement (sujet to some judiil protetions, seured reditors should e le to seize nd sell the ollterl upon defult without judiil As explined, seurity holder will hve to file suit in Court to reover his lim. However, for finnil institutions, expedited proedures re ville under the Finnil Institutions (Reovery of The ountry s eonomi needs nd legl struture must e studied to determine whether mking system of out-of-ourt enforement ville to ll types of entities s seured reditors would suit Pkistn or not. At present, some mehnisms for outof-ourt enforement exist for ertin types of seurity grnted to finnil institutions. These mehnisms re eing hllenged efore the Supreme Court for onstitutionl vlidity nd, in ny event, do not pper to pply to non-posessory seurity interests. Pge 133

149 Bringing Finne to Pkistn s Poor Annex Chpter 4 involvement). Alterntively, there should e fst-trk enforement proedure ville to seured reditors. 20. The detor should hve limited defensive mesures to stop the seizure nd sle of the sset y the seured reditor. Finnes) Ordinne, But even in these ses, should the detor dispute the enforement y the reditor (whih, s it ppers, is usully the se), firly lengthy ourt proedure will ensue. The detor, s it ppers, hs rod powers to hllenge the enforement y seured reditor (even finnil institution tht hs right to enfore out of ourt). Bsed on question of ft or lw, the detor n stop the seizure nd sle proedures, t lest temporrily. The lsses of ppel ville to detors need to e studied further to determine whether the imposition of ny limittions on the right to ppel is worth onsidering or not. For exmple, the right of the detor to ppel n e limited to very speifi irumstnes (for exmple, proving tht pyment hs lredy een mde). Outside of these restritive ourses of ppel, the detor should hve the right to hllenge the reditor in ourt fter enforement y the seured reditor hs tken ple. At present, most suh mtters fll under the generl jurisdition of the High Court. It my e worthwhile to explore the ide of hving speilized ourt of well-trined judges to del with ommeril disputes on n expedited sis. The onept of out-of-ourt enforement for seured reditors, while generlly ludle, needs to e pprohed with some ution in the sene of the neessry regultory frmework to prevent widespred mlfesne y reeivers, liquidtors nd trustees. This regultory frmework does not yet exist in Pkistn. Pge 134

150 Bringing Finne to Pkistn s Poor Annex Chpter 5 Tle A5.1: Workers Remittnes y Setor (Millions US$) Period Bnks Exhnge Compnies Postl Enshment & premium of FEBC/FCBC in Pk Rupees. Totl Jn, Fe, Mr, Apr, My, June, July, Aug, Sep, Ot, Nov, De, Jn, Fe, Mr, Apr, My, June, July, Aug, Sep, Ot, Nov, De, Jn, Fe, Mr, Apr, My, June, July, Aug, Sep, Soure: SBP 2007i Pge 135

151 Bringing Finne to Pkistn s Poor Annex Chpter 5 Tle A5.2: Inlnd Money Orders (millions) Soure: Pkistn Post, Annul Report 2005 Note: FC, NA & AJK: Federl Cpitl, Northern Ares, nd Azd Jmmu nd Kshmir Tle A5.3: Postl Orders (millions) Soure: Pkistn Post, Annul Report 2005 Note: FC, NA & AJK: Federl Cpitl, Northern Ares, nd Azd Jmmu nd Kshmir Tle A5.4: Rurl vs Urn Split of Bnk Brnhes in Provines Provine Urn / Rurl Brnhes Perent Pun j Sindh N.W.F.P Bluhistn AJK FATA Urn Urn Urn Urn Urn Urn 2,745 1, Rurl Rurl Rurl Rurl Rurl Rurl 1, Totl Brnhes 7, Soure: Pkistn Post, Annul Report 2005 Pge 136

152 Bringing Finne to Pkistn s Poor Annex Chpter 5 Tle A5: Workers Remittnes to Pkistn y FY nd Country (Millions US$) Item FY97 FY98 FY99 FY00 FY01 FY02 FY03 FY04 FY05 FY06 FY07 I. Csh 1, , , , , , , , , USA , , , , , U.K Sudi Ari , UAE Dui Au Dhi Shrjh Other Other GCC Countries Bhrin Kuwit Qtr Omn EU Countries Germny Frne Netherlnd Spin Itly Greee Sweden Denmrk Irelnd Belgium Norwy Switzerlnd Austrli Cnd Jpn Other Countries II. Enshment nd Profit in Pk.Rs of Foreign Exhnge Berer Certifites (FEBCs) & Foreign Curreny Berer Certifites (FCBCs) TOTAL (I+II) 1, , , , , , , , , Soure: SBP , Note: GCC: Gulf Coopertion Counil, EU: Europen Union Pge 137

153 Bringing Finne to Pkistn s Poor Annex Chpter 5 Tle A5.6: Perentge Shre of Remittnes in Household Inome, Soure: Pkistn Household Integrted Eonomi Survey (HIES) (The first quintile is omposed of the poorest 20 perent, the seond quintile of the next etter off 20 perent, nd the fifth quintile of the rihest 20 perent.) Annex Tle A5.7 Remittnes, Contriution to Household Inome, nd Poverty Rtio Soure: Od (2007) Pge 138

154 Bringing Finne to Pkistn s Poor Annex Chpter 5 Tle A5.8: Determinnts of Remittnes in Pkistn Remittne Users vs. Non-users Forml Remittne Users vs. Others Informl Remittne Users vs. Nonusers Model I Model II Model III Model I Model II Model III Model I Model II Model III Femle (0.142) (0.175) (0.163) (0.150) (0.189) (0.174) (0.243) (0.185) (0.183) Age (0.004) (0.003) (0.003) (0.005) (0.004) (0.004) (0.006) (0.004) (0.004) Edution (0.032) (0.030) (0.031) (0.039) (0.036) (0.037) (0.045) (0.035) (0.036) Rurl (0.110) (0.092) (0.091) (0.117) (0.098) (0.095) (0.193) (0.130) (0.137) Employed (0.144) (0.180) (0.142) (0.121) (0.204) (0.159) (0.211) (0.217) (0.193) Household hed (0.139) (0.134) (0.139) (0.153) (0.171) (0.177) (0.189) (0.161) (0.161) Collterl (0.113) (0.137) (0.137) (0.137) (0.120) (0.123) (0.152) (0.204) (0.204) Corporte (0.191) (0.180) (0.167) (0.203) (0.210) (0.201) (0.241) (0.202) (0.184) Government (0.211) (0.190) (0.182) (0.213) (0.204) (0.204) (0.351) (0.331) (0.310) Frmer (0.153) (0.123) (0.114) (0.141) (0.161) (0.147) (0.369) (0.291) (0.256) Lorer (0.226) (0.161) (0.200) (0.349) (0.308) (0.327) (0.239) (0.220) (0.219) Forml setor (0.158) (0.167) (0.226) Cell use (0.091) (0.132) (0.109) Cell ess (0.150) (0.185) (0.188) Household inome nd house / ltrine effets Yes Yes Yes Personl inome nd house / ltrine effets Yes Yes Yes Yes Yes Yes Provine effets Yes Yes Yes Yes Yes Yes Yes Yes Yes Numer of oservtions F(k,d) Stndrd errors in prenthesis (=Signifint t 1% level; =Signifint t 5% level; =Signifint t 10% level) Pge 139

155 Bringing Finne to Pkistn s Poor Annex Chpter 5 Tle A5.9: Orgniztions tht re Helping nd Funding the Glol Remittnes Mrket Are Primry tors TA prtners Donors Inditors None None None Tehnology Firms CGAP, infodev Limited: CGAP-Gtes fund Better usiness models Firms IDB, CGAP IFAD fund Inresed ess Firms, lol uthorities IFIs, CGAP, ilterl genies, IFIs, ilterls, World Bnk, World Bnk IFAD Regultion nd oversight Centrl Bnks World Bnk, BIS, some other FIRST, World Bnk, IMF IFIs. IMF Overll ntionl pyment system development Centrl Bnks World Bnk, BIS, IMF FIRST, World Bnk, IMF, IFIs Stndrds nd open networks Some firms World Bnk, BIS None Mrket studies Prie informtion Donor oordintion Glol prie info Glol depository of reserh nd informtion Glol olletion of wht works on the ground Convening uthorities Comprehensive ssistne in remittnes Consultnies, World Bnk AML Consultnies, Consumer protetion genies World Bnk, IFIs, ilterl genies Consultnies World Bnk AML, IFIs, ilterls DFID, ilterls World Bnk, urrently through the Generl Priniples Coordintion Group None Formerly DFID remittnes newsletter, now AessFinne newsletter nd some on World Bnk wesites (Pyment Systems, AML, nd DEC) None, exept Gtes/CGAP fund on tehnology World Bnk None, ut World Bnk hs the potentil Soure: Cirsino nd Andressen (2007) Note: IFAD: Interntionl Fund for Eonomi Development, BIS: Bnk for interntionl Settlements, DEC: Development Eonomi Deprtment (World Bnk), IDB: Inter-Amerin Development Bnk, FIRST: Finnil Setor Reform nd Strengthening Inititive. Pge 140

156 Bringing Finne to Pkistn s Poor Appendix A: Aess to Finne Survey: Dt Methodology nd Clirtion The Pkistn Aess to Finne (A2F) survey is omprehensive ntionl household survey of ll the min finnil servies (trnstion nking, svings, redit, nd insurne), needs, nd usge mong onsumers, in oth the forml nd informl setors. The survey ws onduted etween Otoer 2007 nd Mrh The Survey Instrument The A2F survey followed joint methodology developed y FinMrk Trust (South Afri) nd the World Bnk. The survey overed 10,305 households in ll regions of Pkistn exluding the federlly nd provinilly dministered tril nd northern res (FATA, FANA, nd PATA). At n initil stge, the stndrdized questionnire ws ustomized to Pkistni onditions, to ensure high-qulity dt. Fous interviews were held in urn nd rurl res of Sindh, Punj, North West Frontier Provine (NWFP), Blohistn, nd Azd Jmmu nd Kshmir (AJK), for purposes of lirtion nd gthering of supplementry qulittive informtion. At seond stge, households were surveyed on the following topis in detil: - Bsi household demogrphis; finnil litery; - Soioeonomi hrteristis; psyhogrphis/ttitudinl; - Household inome; - Aess to finnil servies; nks; provider differentitions; - Svings; - Lons/redit; - Insurne; - Money trnsfer/remittnes; - Pyment nd reeipts; - Attitudes to risk nd oping mehnisms. Eh of the ove setions explores in detil individuls hits on nking, sving, lons nd redit, money trnsfers, nd insurne. To route the respondents to these setions, there is kone setion within the questionnire tht first estlishes the rtio of people involved in the five res mentioned ove. This setion is sed on n exhustive list of produts nd servies pertining to the five themes nd sks the respondents whether they hve never used, used in the pst, or re urrently using these produts. Only those respondents who mention tht they re urrently using these produts re further routed to the respetive setion of the questionnire. Choie of Household versus Individul Level of Anlysis Given the highly individul nture of finnil dt, s opposed to most stndrd dt tht re ville t the household level, the questionnire ttempts to mesure oth spets. As suh, some questions re initilly sked of the household hed, nd permission is susequently sought from tht person to interview household memer seleted t rndom, from mong the dult household memers (defined s ove 18 yers of ge). This pproh, worked out s ompromise, permits the identifition of finnil ttitudes, litery, nd preferenes, s well s needs, t the individul level, s those re highly speifi to ge, employment sttus, nd gender, mong other hrteristis (within the household, inome levels n differ signifintly, espeilly long gender lines). At the sme time, suh vriles s inome would e Pge 141

157 Bringing Finne to Pkistn s Poor more preisely defined t the household level, given resoure shring. Yet, we hve oserved, household memers re not lwys wre of one nother s finnil trnstions, even etween husnd nd wife. Smple Design The survey design ws rried out y the Pkistn Federl Bureu of Sttistis (FBS). The survey ws designed to over ll four provines in Pkistn (Bluhistn, Singh, Punj, nd AJK), exept the tril nd northern res (FATA nd FANA), the ltter for seurity resons. The survey used the popultion frme of the FBS, whih divides the entire ountry into rurl nd urn enumertion loks. The ltest ensus on whih these enumertion loks were sed ws the Popultion Census 1998, where the enumertion loks in rurl res were fixed in 1998, those in urn res in Following the identifition of the rndomly seleted (strtified) list of enumertion loks (see 1 st stge of smpling proess, elow), A.C. Nielsen, the survey firm ontrted to rry out the survey, visited eh seleted enumertion lok nd onduted detiled listing of the households in the lok, nd their ext lotion. Given the lists, rndom seletion proedure ws used to selet 15 households from eh enumertion lok. The following setions provide detils on the smpling proedure. A two-stge strtified smple design hs een dopted for this survey. Seletion of primry smpling units (PSUs): Enumertion loks in the urn domin nd mouzs/dehs (villges) in rurl domin were used s primry smpling units (PSUs). Smple PSUs from eh ultimte strtum/sustrtum were seleted y proility proportionl to size (PPS) method of the smpling sheme. In this survey, popultion of rurl res nd households for urn res were dopted s mesure of size for seleting PSUs from the strt/ sustrt formed in urn nd rurl su universes of the survey. Seletion of seondry smpling units (SSUs): Households within eh smple PSU were onsidered s seondry smpling units (SSUs). Fifteen households were seleted from eh smple villge nd enumertion lok y rndom systemti smpling sheme with rndom strt. Smple Frme FBS uses different smpling frmes for urn nd rurl res. The definition of res s urn nd rurl is sed on the lssifition provided y the Popultion Census Orgniztion nd y the lol government nd rurl development deprtment of eh provine. With regrd to the rurl res, the lists of mouzs/dehs (villges) ording to the Popultion Census 1998 re used s smpling frmes. For urn res, the FBS hs developed its own urn res frme, dopting the quik-ount reord survey tehnique. All urn res omprising ities/towns hve een divided into smll ompt res known s enumertion loks identifile through mps. On verge, eh enumertion lok omprises households. Eh enumertion lok hs een further divided into low, middle, nd high groups sed on sujetive omprison with other households in the re. As per ltest survey pulitions onduted y FBS, the urn re smpling frme hs een updted in Pge 142

158 Bringing Finne to Pkistn s Poor Smple Size The totl smple size for this survey ws 10,500 interviews to e onduted throughout Pkistn. To ondut these interviews in totl 700 enumertion loks were seleted y FBS nd demrtion nd listing tivity ws onduted y the survey firm. The rekdown of these 700 enumertion loks is s shown in Tles A1 nd A2: Tle A1: Initil Distriution of Enumertion Bloks/ Villges (PSU) Provine/Are Urn Rurl Totl Punj Sindh NWFP Blohistn AJK Totl Tle A2: Initil Distriution of Households (SSU) Provine/Are Urn Rurl Totl Punj 1,950 3,000 4,950 Sindh 1,200 1,200 2,400 NWFP ,560 Blohistn ,140 AJK Totl 4,350 6,150 10,500 Strtifition The strtifition sheme is sed on geogrphil nd (in the se of urn res) lso inome onsidertions. The detils of the strtifition proedure re elow: Urn res: The urn res: Bhwlpur, Fisld, Gujrnwl, Hyderd, Islmd, Krhi, Lhore, Multn, Peshwr, Quett, Rwlpindi, Srgodh, Silkot, nd Sukkur hve een treted s independent strt. Eh of these ities hs further een sustrtified ording to low-, middle-, highinome groups sed on the informtion olleted in respet to eh enumertion lok. After exluding the popultion of lrge ities, the remining urn popultion in eh dministrtive division in ll provines hs een grouped nd treted s n independent strtum. The entire re of AJK ws onsidered seprte independent strtum. Rurl res: In the rurl res, the popultion of eh distrit in Punj, Sindh, nd NWFP provines were grouped together to onstitute strtum. For Bluhistn, eh distint dministrtive division ws tken s strtum. Agin, the AJK provine hs een onsidered s n independent strtum in rurl res. Demrtion of Enumertor Bloks Provision of smple res: The FBS hed offie provided list of the enumertor loks for the urn res nd list of the villges for rurl res with ll possile identifiers (for exmple, ptwr irle, qnoongo irle, tehsil, nd distrit nme) where the fieldwork for the study will e onduted. Demrtion of smple res: FBS regionl offies ssisted the regionl offies of the survey firm in physilly demrking urn enumertor res. The firm s regionl offies provided suffiient notie to the Pge 143

159 Bringing Finne to Pkistn s Poor FSB regionl offies, so they ould rrnge for their stff to ompny the firm s regionl stff to physilly identify the oundries of the enumertor res. Eh of the firm s nine regionl offies were responsile for oordinting this with ll FSB regionl offies (out 35) tht fll within its overge re. Filitting demrtion: To filitte timely demrtion, the FBS hed offie provided list of ll its regionl offies inluding their ddresses, phone numers, nd min ontt person nd deputy in hrge in the fol person s sene. The FBS hed offie lso ommunited to ll of its regionl offies out the ssistne in demrtion tht would e required from them y the survey firm. The regionl stff of the survey firm ws thus put in touh with the FBS regionl offies nd ooperted on the demrtion nd listing exerise. Similrly, the survey firm riefed its regionl offies out the study nd the protool for oordinting nd working with the FBS regionl stff: ll in dvne of visiting the offie, set up timetle for demrtion greele for oth prties, honor the ommitments mde, nd so forth. Trining of survey firm stff y FBS: FBS stff trined the survey firm stff on listing protools tht were to e undertken in eh of the seleted enumertion loks. These trinings were onduted in three min ities of the ountry to keep the group size per session mngele nd the trinings effetive. The min res of trining were: - How to selet strting point to demrk the entire re; - Wht to e inluded in listing with lok; - How to fill in the listing form; - Differene in listing methods etween urn nd rurl res. Listing of Households In eh seleted PSU, listing of households ws rried out to estimte the updted popultion sttistis nd develop the systemti sheme of seletion of household within the PSU. This ws undertken s simple ensus tivity in whih, one the FBS identified the PSU, the survey firm ounted eh house nd noted its household numer nd nme. Seletion Proedure The households within eh PSU were seleted y systemti smpling tehnique with rndom strt. In eh PSU, 15 households were seleted nd interviewed. This ws done from list of households, whih ws produed y listing rried out in eh seleted PSU s desried ove. Seletion of respondent: If household hd more thn one potentil vlid respondent, then the Kish Grid ws used to selet the tul respondent. This enled us to selet the respondent in rndomized mnner. If the seleted respondent ws not ville, three ll ks were mde to lote tht respondent. If fter three ll ks the respondent still ws not ville or refused to prtiipte, sustitute from the sme viinity ws seleted nd ontted. Sustitution: Sustitution of seleted respondent within household ws done only if the originl respondent ws found to e of unsound mind. In other ses, the entire household ws sustituted. Sustitution Methodology nd Listing in the Sustitute Ares FBS eleted to void oversmpling, nd therefore gve out sustitution enumertion loks in very limited ses, where seurity risk or nturl disster (floods) ws involved, or where there ws n insuffiient numer of households for vlid seletion euse the government hd rzed houses for lnd Pge 144

160 Bringing Finne to Pkistn s Poor development. If the survey firm found tht sustitution ws needed, it hd to otin letter from the relevnt uthority (for exmple, the lol tehsil dministrtion) verifying tht the re is not suitle for work euse of lw nd order or ny other reson. FBS first did its own onfirmtion of the sitution nd then provided sustitute enumertion re for the listing exerise. In the end, there were 12 sustituted loks; their distriution is provided in Tle A3: Tle A3: Distriution of Sustituted Bloks/Villges (PSU) Provine/Are Urn Rurl Totl Punj 4-4 Sindh NWFP 1-1 Blohistn AJK Totl Bloks dropped: In ddition to the sustitutions, 13 res hd to e dropped ltogether for seurity resons nd no sustitutions were possile. The rekdown of these loks is provided in Tle A4: Tle A4: Distriution of Bloks/Villges Dropped Provine/Are Urn Rurl Totl Punj Sindh NWFP Blohistn 1 1 AJK Totl The finl distriution of enumertion loks nd households is provided in Tle A5: Tle A5: Finl Distriution of Bloks/Villges Enumertion Bloks/Villges (PSU) Households (SSU) Provine/Are Urn Rurl Totl Provine/Are Urn Rurl Totl Punj Punj 1,950 3, Sindh Sindh 1,170 1, NWFP NWFP Bluhistn Bluhistn AJK AJK Totl Totl Weights Using the initil nd finl smple distriution, s well s tking into ount the strtifition y geogrphy nd inome, FBS omputed nd provided to the World Bnk set of household weights for eh of the 10,305 interviewed households. The survey firm then used the household weights to ompute the weight of eh respondent s funtion in their household omposition (the numer of household memers). Household weights re used for vriles tht re defined t the household level, suh s household inome. Individul weights re used for metris foused on personl spets, suh s nking reltionships, preferenes, nd ttitudes to risk. Pge 145

161 Bringing Finne to Pkistn s Poor Clirtion of Survey Results The A2F household survey ws lirted to pulily ville results from FBS, Stte Bnk of Pkistn (SBP), nd other supply-side soures, to determine possile ltent ises nd guge the rediility of the dt olleted. This ws prtiulrly importnt for the A2F dt set, sine it ws the first of its kind in Pkistn, nd used methodology tht hd not efore een used in Pkistn. Clirtion ws done ording to: Popultion dt Inome dt Bsi finnil ess dt Popultion Clirtion Bsed on the household survey dt, estimtes of totl popultion re derived using the following methodology. The totl popultion is estimted s follows: P = (P i ) = (A i u * S w i u) + (A i r * S w i r) AS w i u AS w i r = {SH i u * (IW i u ) } * S w i u) + {SH i r * (IW i r ) } * S w i r) AS w i u AS w i r where P is totl popultion, P i (u, r) is the urn (resp. rurl) provine popultion; A i (u, r) is the totl urn (resp. rurl) numer of dults per provine; S w i (u, r) is the urn (resp. rurl) weighted verge household size per provine; AS w i (u, r) is the urn (resp. rurl) weighted verge dult household size per provine; nd SH i (u, r) re the urn (resp. rurl) smpled households per provine; IW i (u, r) re the urn (resp. rurl) individul weights per provine; nd i (Punj, Sindh, Bluhistn, NWFP, AJK). The totl households per provine re estimted s: H i = SH i u * (HW i u ) + SH i r * (HW i r ), where H i re totl the urn (resp. rurl) households per provine, SH i (u, r) re the urn (resp. rurl) smpled households per provine; HW i (u, r) re the urn (resp. rurl) household weights per provine; nd i (Punj, Sindh, Bluhistn, NWFP, AJK). Given the ove lultions, the resulting popultion figures re presented elow (Tle A6). Tle A6: Popultion Counts per A2FS Survey Results Totl Households (HHs) Totl 18 Yers Popultion Totl Popultion Provine Urn Rurl Totl Urn Rurl Totl Urn Rurl Totl Punj 3,970,977 7,587,990 11,558,967 15,047,579 27,949,635 42,997,214 26,275,084 52,187,201 78,462,284 Sindh 2,514,365 1,991,098 4,505,463 10,710,184 9,625,282 20,335,466 17,913,027 17,581,638 35,494,665 NWFP 449,841 1,763,546 2,213,387 1,939,424 7,385,777 9,325,201 3,577,497 14,667,675 18,245,172 Bluhistn 214, ,793 1,027, ,233 3,685,858 4,595,091 1,601,401 6,557,489 8,158,890 Totl 7,150,147 12,155,427 19,305,574 28,606,420 48,646,552 77,252,972 49,367,009 90,994, ,361,011 Pge 146

162 Bringing Finne to Pkistn s Poor For the household survey, enumertion loks in urn res re sed on 2003 estimtes nd in villges in the rurl domin on 1998 estimtes, per the ltest popultion ensus. During eh ensus, FSB fixes enumertion loks t the size of out 250 people. The survey used the old enumertion loks (1998/2003), ut ws sed on fresh listing done in Given this onstrint, the lirtion exerise egs the question of relevnt popultion figures for omprison. The nturl option is to ompre the impliit totl popultion from the household survey to the mix of rurl popultion of 1998 nd urn popultion of Another possile omprtor is the 2007 popultion figure, s the survey ws onduted in Otoer 2007 Mrh The hoie would depend on the pttern of popultion growth we elieve we hve oserved during the pst dede. If we elieve growth mostly ourred in the numer of enumertion loks, ut not within loks, the relevnt omprtor is the 1998/2003 mix. If, however, we elieve tht growth mostly ourred within loks (through irth rte), then the 2007 popultion is the more urte omprtor figure. Nturlly, there will e some of oth effets t work, so we would expet the A2F popultion to fll somewhere in etween (s is the se, see elow). We elieve the first option is more resonle ssumption, ut we present oth omprisons. The hnge in popultion over the pst 10 yers n e ttriuted to migrtion (growth in the numer of enumertion loks) nd irth rtes (growth within enumertion loks). The numer of urn households in ll four provines hs shown positive inrese from 1998 to 2008, while the totl rurl households hve delined over the sme period, whih lerly suggests migrtion. The strong urniztion trend ehoes this onlusion. On the other hnd, the popultion inrese ttriutle to irth rtes for 10-yer period would ffet totl popultion strongly, ut dult popultion figures less so (muh of the inrese will e in the popultion under 18 yers of ge). There re no ftul popultion (or household ount) figures sine the ltest ensus to rrive t more reent popultion figures, puli soures use extrpolted growth rtes. Therefore, the omprtor tles elow present only totl popultion omprisons. For those resons, we elieve lirtion should e done in omprison with mix of 1998/2003 figures s desried ove, insted of ompring with the extrpolted 2007 popultion figure. Tking this limittion into ount, one would expet tht the household survey, whih ws sed on old enumertion loks devised during the ltest ensus in 1998/2003, would ffet the dt in the following wy: while some growth hs ourred within the 1998/2003 enumertion loks up to 2007/2008, most growth would e expeted in the form of new enumertion loks in peri-urn res, whih the household survey perfore does not pture. As suh, the relevnt omprisons would e with the popultion of 1998 (rurl) nd 2003 (urn), s elow (see Tle 2). The ises stemming from this limittion of FBS dt re disussed throughout the text; however, one ovious exmple is tht new lusters of rurl migrnts to ities, the people who re most likely to engge in some miroentrepreneuril tivity, re poorly ptured y the household survey, ising downwrds the mirofinne ess estimtes. In the report, we ution the reder in eh instne where the ove is is relevnt. This ompres losely to FBS figures for popultion (Tle A7). Thus, the relevnt totl popultion for omprison is the ddition of urn popultion in 2003 nd rurl popultion in 1998, whih is million. Bsed on Tle A1, the totl popultion from the household survey is estimted t million, or within 1.9 perent of the relevnt omprtor figure for totl popultion. For ompleteness, we lso present the omprison with 2007 popultion estimtes. Estimted popultion from the household survey (142.2 million) is n underestimtion if we onsider the urrent popultion estimtes from vrious soures: 159 million (World Resoures Institute) for 2006, 164 million (World Resoures Institute) for 2007, nd 164 million (Popultion Census Orgniztion) for Popultion estimtes sed on the household survey hve to e interpreted with vet tht FATA, FANA, nd PATA were not inluded in the smpling frme (estimted popultion for these two res is Pge 147

163 Bringing Finne to Pkistn s Poor pproximtely 3-4 million, not ounting some 2 million registered Afghni refugees). Tht still leves popultion differene of over 15 million. Some of tht will e explined y hildren orn sine 1998, nd will not ffet the dult popultion figure. However, some of this inrese is genuine is in the dt due to the usge of dted enumertion loks per FSB reords. The ove omprison shows the extent of possile downwrd is in overge, euse of growth within enumertion loks (s opposed to growth in the numer of loks). This is is less worrisome s ompred with is used y not overing new migrtion lusters reted post 2003, for the following reson. Although the estimted popultion figures from the household survey re understted, nd some of the inrese in popultion my e ttriuted to migrtion, most of the inrese in popultion is ounted for in n inrese in the numer of underge hildren, whih re not onsidered y the survey for purposes of ess/usge of finnil servies. Thus, the results sed on the survey still remin vlid nd representtive of the overll popultion of Pkistn, with the provision tht overge of newly reted enumertion loks sine 1998/2003 hs not een ensured. Tle A7: Popultion Counts per Census, using Urn/Rurl Mix per Enumertion Blok Dtes Provine Urn 1998 Growth rte* Urn 2003 Rurl 1998 Totl Punj 23,548,205 19% 28,082,663 50,878,320 78,960,983 Sindh 14,839,862 14% 16,977,598 15,600,031 32,577,629 NWFP 2,994,084 13% 3,369,045 14,749,561 18,118,606 Bluhistn 1,568,780 22% 1,910,626 4,997,105 6,907,731 AJK 80,271 17% 93,912 2,892,729 2,986,641 Totl Pkistn 43,031,202 17% 50,339,933 89,117, ,457,679 Soure: AJK totl popultion dt from; *The AJK urn growth rte estimte is not ville; verge urn growth rte for the ountry ws used insted. Given the tiny urn popultion in AJK, the impt of this pproximtion is negligile. Inome Clirtion As further roustness hek, we hve lirted the A2F survey results to the Ntionl Household Survey of FBS, in terms of household inome/onsumption. This lirtion is sed on potentil onerns tht respondents understte their inomes reltive to the Pkistn Soil nd Living Stndrds Mesurement Survey (PSLM) 2005/06 onsumption figures (perhps euse of underreporting or/nd not ounting some inome s suh euse of lk of knowledge of ll household memers inome soures, or lk of litery to estimte inome orretly). Missttement of inome onsidertions (inome is prtiulrly hzrdous vrile in self-reported surveys) led us to estimte inome sed on ojetive hrteristis (type of house, type of si menities, nd so forth), s explined elow. The enhmrk dt used re the monthly household onsumption dt of the totl expenditure figures from PSLM 2005/06, sed on monthly onsumer prie index dt from FBS. Consumption dt re used s proxy for inome dt, s monthly household inome dt is not ville in Pkistn ntionl household surveys. The PSLM onsumption dt ws djusted for 22.9 perent infltion rte etween the olletion periods of PSLM (Jnury 2006) nd the A2F survey (Jnury 2008). The tles elow present the lirtion of the A2F nd PSLM models t two levels: Frequenies of households within monthly household expenditures/inome thresholds in Figure A1 (nd ssoited umultive frequenies in Figure A2) Pge 148

164 Bringing Finne to Pkistn s Poor Monthly household expenditures/inome thresholds orresponding to frequeny distriution of households in Figure A3. Figure A1: Perentge Popultion within Monthly Household Expenditures/Inome Thresholds 20% 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% Figure A2: Cumultive Perentge Popultion within Monthly Household Expenditures/Inome Thresholds 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Cum. Frequenies Aess to Finne 2008 Monthly HH Inome Aess to Finne 2008 Monthly HH Inome Cum. Frequenies PSLM 2005/06 Monthly HH Expenditure (djusted) PSLM 2005/06 Monthly HH Expenditure (djusted) Figure A3: Upper Expenditures/Inome Thresholds per Popultion Perentile Rs300,000 Rs250,000 Rs200,000 Rs150,000 Rs100,000 Rs50,000 Rs0 1% 8% 26% 51% 64% 81% 89% 92% 94% 98% 99% 100% PSLM 2005/06 Consumption Upper Brket (djusted) Aess to Finne 2008 Inome Upper Brket Clirtion results show tht the distriution of the A2F survey inome results is wider thn the PSLM 2005/06 expenditure dt. The lowest inome quintile of the A2F survey reported higher inomes reltive to onsumption expenditure, while the rihest quintile reported lower inome thn onsumption. Suh devition of inome from onsumption is typil in the dt, the differene eing ounted for y svings. While the svings of low-inome groups re negligile, upper-inome groups tend to sve onsiderle prt of their inomes. Exept for these differenes t the upper nd lower mrgins of the inome dt, the A2F inome urve n e onsidered to overll fit the ntionl expenditure urve. While the overll inome results re deemed lirted to PSLM expenditure dt, the rnge of ftors tht ould explin the oserved differenes t the mrgins re disussed elow. Household Survey Mesurement Bis Pge 149

165 Bringing Finne to Pkistn s Poor 1) Respondent is: s frequently oserved in household surveys, euse of lk of knowledge of ll household memers inome soures or lk of litery to estimte inome orretly, poor respondents of the A2F survey my hve understted their inomes reltive to onsumption expenditure, while rih respondents my hve overstted their inome. 2) Questionnire is introduing differenes linked to onsumption from home prodution ftors: A2F inome survey responses on monthly household inome my not hve ptured home onsumption from the poor in griulture, while this dditionl inome flow would e reported in the onsumption survey. This ntiipted is ws, however, mitigted y n dditionl question in the A2F survey to pture onsumption from things the fmily grew s inome. Consumption nd Current Inome Bis 3) Differenes etween inome nd onsumption levels mong poor nd rih households: The differenes in fit t the lower- nd upper-inome/onsumption ends ould e explined y the oservtion tht onsumption my e higher thn inome mong poor households nd lower mong rih households, more likely to sve or invest exess inome over onsumption. 4) Current versus permnent inome is: While the stndrd hypothesis is tht onsumption ptures permnent inome, the lirtion is onduted etween onsumption nd the urrent inome levels mesured y the A2F household survey. The stndrd devition of urrent inome dt would e lrger thn of the permnent inome, hene higher perentges of households within ottom/top inome thresholds thn onsumption threshold equivlents. Corretive Tehniques for Bises Identified Aove To orret for potentil mesurement is, we use n lterntive methodology to mesure urrent inome, pioneered for Pkistn y Dr. Mrk Shreiner. The index developed y Dr. Shreiner lirtes inome y the menities used y households, suh s type of house nd other household ssets. This methodology hs een proven to e n dequte proxy to mesure inome levels of the Pkistni popultion. We use oth vriles in tests, for roustness, nd otin very similr results. Clirtion of Demnd-side Finnil Aess Vriles to SBP Supply-side Mesures Numer of Bnk Aounts The totl numer of nk ounts, s quoted y SBP (most reent figures re for Deemer 2006) is tht in Pkistn there re 26.6 million nk ounts. These inlude oth usiness nd individul ounts. We do not hve figures for the rekdown etween those two tegories, ut we hve tht rekdown in the se of nk lons speifilly. In tht se, out of 4.9 million totl lon ounts outstnding, 1.8 re privte usiness lons, nd 3.2 million re individul lons. If one would ssume tht similr proportion holds for ll nk ounts (dmittedly, lep-of-fith ssumption), we end up with n estimte of 17.4 million nk ounts. The A2F survey does not mesure the numer of ounts, ut wht we n oserve is the numer of dults who hve t lest one (nd possily more) nk ounts. We oserve 12.2 million people who report to urrently hve nk ounts, nd 18.9 million people who hve now or hve hd in the pst nk ount (weighted popultion figures, dults only). These figures re quite omptile if one tkes into ount tht people, espeilly household heds, would mke use of more thn one ount, for the use of spouse nd grown hildren, nd noting tht those with nk ounts re mostly the urn highinome groups. On Islmi lons, we hve 11.5 thousnd people who hold suh lons, nd 61 thousnd hve now or hve hd them in the pst. The SBP quotes 22.2 thousnd lons, whih gives out two lons per person. Note Pge 150

166 Bringing Finne to Pkistn s Poor gin tht Islmi lons re mostly held y riher urn mles. We find 127 thousnd lients for Islmi produts, nd 349 thousnd hve now or hve hd suh produt in the pst. Mirofinne On mirofinne, we hve not een le to reonile the dt. The A2F survey points to 231 thousnd people who hve t lest one mirofinne produt, nd 601 thousnd who hve it or hve hd it in the pst. The respetive SBP figure is muh higher, t 1.7 million mirofinne produts s reported y MFIs to SBP. There re severl resons to explin the divergene in estimtes. First, mirofinne lients my hve multiple ounts or work with more thn one MFI. Seond, the A2F survey mirofinne figures do not isolte mirofinne insurne produts (ll insurne produts re olleted in single mesure, whih preludes ny distintion of mirofinne insurne from trditionl produts). Third, nd perhps most signifintly, mirofinne lients tend to luster in urn res, nd my e re under-represented in the A2F survey. This is is pointed out in relevnt setions of the report, nd SBP figures re used. Figure A4: Distriution of Ative Mirofinne Borrowers Soure: Mirowth, Issue 7 Qurter 1, To understnd in detil the lustering is, onsider the following thoughts on popultion growth ptterns. If the lirtion ove orretly identifies the is of underoverge of newer (urnized) enumertion loks, whih ppered fter the ltest ensus nd re therefore systemtilly not overed y the survey, tht will result in severe underoverge of younger villge migrnts. There is onsiderle evidene tht mirofinne produts re predominntly used in lusters in urnized res, whih lends even further redene to this onjeture. Figure 4, presenting mp of tive mirofinne orrowers, exhiits signifint lustering of orrowers round urn res, nd further onfirms the ove is. Pkistn Post Offie Finlly, respondents report 1.7 million people with Pkistn Post svings ount, nd 5.7 million hve ever hd suh n ount. The SBP-reported figure is 3.6 million post offie svings ounts. Wheres it is unresonle to ssume multiple ounts in most of the ses (these re low-inome svings, for the most prt), some of the resons for this differene lie most likely in the numer of intive ounts tht Pge 151