Cost-Benefit Analysis for the Introduction of a Bus Rapid Transit System in Yangon, Myanmar

Transcription

1 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu Cost-Beefit Aalysis for the Itroductio of a Bus Rapid Trasit System i Yago, Myamar Hiroori Kato, Akihiro Iagi 2, Nozomi Saito 3, ad Phyo Thet Thet Htu 4 Departmet of Civil Egieerig, Uiversity of Tokyo 7-3-, Hogo, Bukyo-ku, Tokyo , Japa Phoe: ; Fax: kato@civil.t.u-tokyo.ac.jp 2 Departmet of Civil Egieerig, Uiversity of Tokyo 7-3-, Hogo, Bukyo-ku, Tokyo , Japa Phoe: ; Fax: iagi-a@ip.civil.t.u-tokyo.ac.jp 3 Overseas Cosultig Admiistratio, Trasportatio Departmet Divisio, Developmet Plaig Departmet, Nippo Koei, Co. Ltd. 4 Kojimachi 5-chome, Chiyoda-ku, Tokyo , Japa Phoe: ; Fax: saito-z@-koei.jp 4 Departmet of Civil Egieerig, Yokohama Natioal Uiversity 79- Tokiwadai, Hodogaya-ku, Yokohama , Japa Phoe: d08sc93@yu.ac.jp Abstract. This paper aalyzes the feasibility of itroducig a bus rapid trasit (BRT) system ad the restructurig of the bus route etwork to improve the performace of the local bus service i Yago, Myamar. First, we summarize the characteristics of the curret urba bus service i Yago. We poit out that the poitto-poit bus route etwork results i may overlappig routes alog mai roads ad this leads to serious traffic cogestio. The, we examie the feasibility of two policy optios. The first case (Case ) itroduces a ew BRT lie to the existig bus etwork; while the secod case (Case 2) restructures the bus route etwork whilst itroducig a ew BRT lie. Five types of BRT fares are cosidered for each case. To evaluate these policy optios, a travel demad forecast system, icludig the estimatio of a origi-destiatio table ad a bus routechoice model, is developed. The results of the evaluatio show that Case -400, i which a ew BRT with a flat fare of 400 Kyats, itroduced to the existig bus route etwork offers the most desirable outcome. Word Cout: 5,475 words + 6 figures ad 2 tables

2 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu INTRODUCTION Yago is the most importat commercial ceter i Myamar with about 4. millio people i a area km 2 (). There are four mai types of trasportatio available i Yago: private car, rail, taxi, ad bus. The vast majority of urba trasportatio i Yago relies o bus trasportatio. Accordig to Zhag et al. (2), the modal share of bus trasportatio i Yago is 84%. This stems from the difficulties associated with car owership, the low quality of the rail service, ad regulatio of motorbikes ad the para-trasit service i the cetral busiess district. However, the performace of the bus service icludig its speed, reliability, capacity, ad image is poor i the city. The existig buses ru at low speeds, maily due to poor vehicle performace. The service is ot puctual because of serious traffic cogestio. The capacity of each vehicle is limited as the majority of buses i Yago are truck buses. Furthermore, the vehicles are ofte old ad this makes for a poor public trasportatio image. It is imperative that the quality of the bus service be improved. This paper aalyzes the feasibility of itroducig a bus rapid trasit system ito the bus trasportatio market ad the restructurig of the bus route etwork to improve the performace of the local bus service i Yago. The paper is structured as follows. The ext sectio summarizes the characteristics of the bus trasportatio service i Yago. The, sectio 3 presets the policy optios for improvig the problems associated with the local bus service. Sectio 4 shows the methods to be used i the evaluatio aalysis. This icludes a travel demad forecast ad a cost-beefit aalysis. The evaluatio results are discussed i Sectio 5. Fially, the achievemets are summarized ad outstadig issues are preseted. URBAN BUS SERVICE IN YANGON Kato et al. (3) showed that there are the two types of local bus services i Yago: the bus service provided by bus compaies ad the bus service cotrolled by the bus cotrol committees (BCCs). The bus compaies are large-scale, private bus operators who ow the vehicles, hire the drivers/coductors, ad operate the urba bus service i Yago. Two bus compaies ow provide services i Yago: the Golde City Lik Co. (GCL) ad the Uio of Myamar Ecoomic Holdigs Limited (UMEHL). The BCCs are oprofit orgaizatios cotrollig small-scale, idividual bus owers. Idividuals ow the buses ad lease them to the drivers ad coductors. The idividual bus owers must belog to oe of the BCCs. There are a umber of idividual bus owers i Yago. Accordig to our observatios ad iterviews with local people, the buses ca be categorized ito six types accordig to their fuctio, desig, ad capacity: city bus (air-coditioed); city bus (o-air-coditioed); mii bus; Dya/Cater; Hilux; ad others. Dya/Cater ad Hilux are trucks that have bee redesiged for passeger-use by the additio of a roof ad seats. They are widely referred to as a truck bus by the locals. The majority of buses i operatio are truck buses. The vehicles cotrolled by the BCCs are maily small-scale buses, icludig the Mii bus, Dya/Cater, ad Hilux. There are 5 orgaizatios operatig or cotrollig the local bus market i Yago. I total, 283 routes are operated i the city. TABLE shows the distributio of bus routes i the city across the various orgaizatios. The total umber of vehicles i operatio o a daily basis is 5,039. TABLE Orgaizatios operatig bus services i Yago (2008) Daily Average Orgaizatio No. of lies Ruig vehicle Trip Passeger YDBCC Easter District Wester District Norther District Souther District YCDC Pyi Khaig Phyo Kadayawadi Tha Mya Thu Parami Shwe Iwa Golde Guest, Shwe Myi Pya Myama Thar GEC Total

3 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu FIGURE Typical bus routes i the south-wester part of Yago City The total daily umber of bus passeger trips is 3,305,726. The size of the orgaizatios rages from the Yago Developmet Bus Cotrol Committee (YDBCC) cotrollig 2,470 buses to the Shwe Iwa cotrollig 4 buses. The majority of bus routes ru directly betwee suburba areas ad the cetral busiess district, alog the mai roads. The bus etwork i Yago is characterized as a poit-to-poit etwork. The poit-to-poit bus etwork i Yago results i the overlappig of may bus routes alog the major roads coectig the suburba areas with the cetral busiess district. A example of the bus routes is show i FIGURE ad the fact that may bus routes overlap alog the same road is highlighted. Sice a umber of buses o various routes ru alog the same mai roads, they icrease the traffic flows, particularly at peak hours. This results i serious traffic cogestio o the mai roads at peak times. Furthermore, the poit-to-poit etwork results i excessive competitio amog bus drivers o the same roads. Although bus services are moitored by the BCC, the moitorig of oe particular route is quite idepedet of the moitorig of aother route. Therefore, bus drivers ofte race agaist oe aother to reach the ext stop o the route, clearly raisig safety issues. It is essetial that the performace of bus services i Yago eeds to be improved. IMPROVEMENT OF BUS NETWORK IN YANGON Bus rapid trasit As Erst (4) shows, may developig coutries have commo characteristics i terms of city structure ad urba trasportatio: high populatio desities, a sigificat existig modal share for public bus trasportatio, ad fiacial costraits providig a strog political impetus to reduce, elimiate, or prevet cotiuous subsidies for public trasit operatios. These issues are relevat to Yago. The bus rapid trasit (BRT) has bee highlighted recetly as a potetial solutio to these problems. BRT is a high-quality bus based trasit system that delivers fast, comfortable, ad cost-effective urba mobility through the provisio of segregated right-of-way ifrastructure, rapid ad frequet operatios, ad excellece i marketig ad customer service (5). BRT ca reduce travel times, attract ew riders, ad iduce trasit-orieted developmet. It ca be more cost-effective ad provide greater operatig flexibility tha rail trasit (6). The BRT system will typically cost four to 20 times less tha a light rail trasit (LRT) system ad 0 to00 times less tha a metro system (7). Kowledge of the success of BRT throughout the world, alog with the low cost ad staged developmet possibilities has made BRT popular i developig coutries (8). 3

4 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu Desig of bus rapid trasit i Yago We examie the feasibility of itroducig a bus rapid trasit (BRT) system i Yago City. It is assumed that a ew BRT lie is itroduced o oe of the roads coectig Isei Garde ad Kyoe Gyi Road, as show i FIGURE 2. The total legth of the BRT lie is 5.6 kilometers. The roads iclude the mai road, which has six laes. It is assumed that two laes are coverted to laes exclusively for use by the BRT service. This eables the BRT vehicles to ru faster tha other buses because o other vehicles hamper the operatio of the BRT. It is also assumed that ew BRT vehicles with a higher capacity tha city buses will be itroduced. To simplify the fare trasactio, a fixed fare system will be itroduced. The service frequecy is assumed to be almost the same as that for existig city-bus services. The assumed details of the BRT desig are as follows: Average speed: 28 km/h icludig the ruig time betwee stops ad time spet at bus stops. BRT stops/termials: six BRT stops icludig Isei Garde, Thamie Juctio, Butar Yoe, Hlaig Zay, Halpi, ad Kyoe Gyi. BRT capacity: a BRT vehicle has the capacity for 20 passegers. Fare system: a fixed-fare system. Service frequecy: 20 services per hour (peak). BRT operator: a sigle bus compay. Ivestmet: stops/termials, exclusive laes, pedestria access, parkig facilities, etc. Restructurig of the bus etwork The bus route etwork must be restructured to solve the problems caused by the poit-to-poit etwork. As Vuchic (9) shows, bus services are improved if the truk is upgraded to a idepedet rail or BRT service. I particular, a etwork that icludes a truk lie with feeders ca icrease passeger umbers. It should be oted that services o separate rows, attractive stops, ad steady headway are required to attract passegers because a etwork that icludes a truk lie with feeders requires passegers to trasfer at termials. The, we examie the impact of restructurig the bus etwork i Yago. The case studies compare two optios: BRT itroductio without restructurig the bus etwork, ad BRT itroductio i cojuctio with restructurig the bus etwork. METHOD OF ANALYSIS The cost-beefit aalysis is applied to examiig the feasibility of the BRT itroductio. First, we forecast the demad for bus travel. The bus route demad is estimated with a bus-route choice model, after the bus-use origi-destiatio matrix is estimated. Next, the operatig cost of the bus operators is estimated based o the cost data collected i our survey. The, the cost of itroducig the BRT is calculated o the basis of the experiece of FIGURE 2 BRT routes examied i the study: deoted by the thick lie o No.4 Mai Road 4

5 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu other coutries. Fially, the cost-beefit aalysis is coducted. Travel demad aalysis Zoig system First, we defie the 29 zoes i the target area o the basis of towship boudaries. Sice the boudaries of some towships are much larger tha i other zoes, they are divided ito two or more zoes based o the road etwork. The, the odes i each zoe are defied ad are maily located at major itersectios i the correspodig zoe. The defied zoes are show i FIGURE 3. The, the bus trasport etwork is defied i the target area o the basis of the above zoes. Oce the 29 zoes are defied, the 29 odes are coected by a bus trasport etwork. The 42 liks i the simplified etwork are defied based o the real road etwork i the target area. Preparatio of travel data i the etwork First, the jourey time of each lik is defied o the basis of lik legth. Although the travel speed varies amog liks, we assume that the average travel speed durig the morig is 20 km/hour i ay lik. This speed is estimated usig the data collected from the bus travel survey. Next, we prepare data o the trasfer time at each lik. The trasfer time is defied as the time it takes to walk from the place where a bus passeger gets off oe bus to the place where the passeger boards the ext bus. O the basis of the surveyor s observatios, we estimate that the trasfer time is 5 miutes at ay stop. Data collectio for bus-use travel demad ad operator s cost structure First, the research team coducted local iterview surveys from September 5 to 23, 2008 ad November 9 to 2, 2008, relatig to local bus operatios i Yago. We prepared a iterview sheet for the survey. The iterview sheet requested the bus operators aswer questios about the curret situatio, icludig basic iformatio about the orgaizatio, labor system, buses, operatig facilities, routes, service frequecy, fuel, travel time, ad cost. Two bus compaies, four bus cotrol committees, 22 idividual bus owers, ad local bus authorities were iterviewed. The details of the survey are show i Kato et al. (0). Next, we coducted two types of local survey ito urba bus-user s behavior. Oe was a Origi- Destiatio (OD) survey, which collected data o sample-based idividual origis ad destiatios i Yago City. The other was a bus travel survey, which collected data o sample-based route-choice i Yago. The data collected i the OD survey will be aggregated to estimate the zoe-based OD travel patters, whereas the data collected i the bus travel survey will be used to estimate a discrete bus route-choice model. The OD survey was implemeted from November 9 to 2, 2008, with the bus travel survey coducted from November 4 to 9, CBD area FIGURE 3 Zoig system used i the aalysis 5

6 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu Fially, data from 2,524 idividuals was collected i the OD survey ad from 2,372 idividuals i the bus travel survey. Further detail of these surveys is show i Kato et al. (0). Estimatio of the origi-destiatio table We estimate the OD table for bus users. As o data regardig the OD table was available, the data was estimated usig socio-demographic data ad the OD survey origially coducted by our study team. Estimatio of zoe-based trip productio First, we prepared zoe-based socio-demographic data, icludig data o the populatio by geder ad the umber of households. This statistical data was collected from the Govermet. Ufortuately, data o zoebased umber of employees was ot available. Secod, we estimated the sample-based worker s rate ad the sample-based modal share of bus use by commuters. Accordig to the data collected i the OD survey, the total umber of people i households to which the respodets belog is 0,457, of which the total umber of workers was 6,095. Thus, workers accout for 58 percet of each household. Next, accordig to the data collected i the OD survey, the total umber of household commuters usig buses was 5,72. Thus, the average modal share of bus use amog workers was Third, we estimated the sample-based workig rate i days. The average commutig rate was 5.99 days. The, we estimated the populatio distributio i some towships give that two or more zoes are sometimes icluded i a towship. We simply assumed that populatio desity is costat i a towship. This meas the zoal populatio is i proportio to the zoal area i the correspodig towship. Fially, we calculated the trip productio from each zoe usig the followig formula: TP i t g g t g = PO WK BS WR PD () i t g where TP i t : the umber of trips geerated from zoe i i towship t ; PO t : populatio of towship t by geder i towship g ; WK : average worker s rate i the household; BS : average modal share of bus use g amog workers: WR : workig rate by geder g ; ad PD i t : the zoal populatio rate of zoe i belogig to towship t. Estimatio of OD table We estimate the OD matrix o the basis of the estimated trip productio. First, we estimate the sample-based proportio of trips from a origi to a destiatio for each zoe. The, we simply multiply the sample-based proportio of trips by the zoe-based trip productio. This is summarized as follows: T od = r od TP o (2) where T od deotes the umber of trips from origi zoe o to destiatio zoe d ; r od, the sample-based proportio of trips from origi zoe o to destiatio zoe d ; ad TP o, the umber of trips geerated from zoe o. The OD matrix shows the daily demad from oe zoe to aother. Note that the OD matrix oly icludes commuters who use buses. We assume that this OD matrix is costat durig the project period. Bus route demad forecast We use a bus route choice model for forecastig bus route demad. The route choice model of bus users i Yago was estimated by Kato et al (0). This model outputs the probability of choosig the bus route for a give OD pair. The explaatory variables i the model are travel time, travel cost, ad the type of bus vehicle. Our aalysis uses the estimated multiomial logit (MNL) model. The probability of choosig the r th route is chose from zoe o to d for idividual is derived as follows: exp p od, r = (3) exp where od r od r r R od ( V od, r ) ( V ) od, r p, deotes the probability of choosig the r th route to move from zoe o to d for idividual, ad V, is the uiversal compoet of the idirect utility of a idividual whe the r th route is chose from zoe o to d. Kato (0) proposed the followig estimated utility fuctio: V od, r TTod, r TCod, r BTod, r = (4) 6

7 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu od r where TT, deotes the total travel time for the r th route chose to move from zoe o to d for idividual ; od r TC,, the total travel cost; ad BT od, r, the dummy for bus type. The total travel time icludes the i-vehicle ridig time, waitig time at the stops, trasfer time at the stops, ad the access/egress time to ad from the stops. The dummy for bus type is defied as if the bus is a city bus, ad 0 if it is ot. If the total volume of traffic flow from zoe o to d is give ast od, the expected volume of traffic flow whe the r th route is chose to trasfer from zoe o to d t. is expressed as follows: od r [ tod, r ] = Tod pod r exp( Vod, r ) = Tod exp( Vod, r ) E, r R od (5) where p od. r deotes the probability of choosig the r th route to move from zoe o to d. Note we assume a represetative idividual for each OD pair i the above aggregatio process. Evaluatio of the policy optios The policy optios were evaluated cosiderig the demad for the BRT, the profit for the BRT operators, the users beefit, the supplier s beefit, the social beefits, ad a cost-beefit aalysis. To calculate the operatig costs for the BRT ad the bus operators, we used the results from the cost structure aalysis show by Kato (0). The cost fuctios icorporatig the vehicle km ad the fleet size are used for calculatig the operatig cost. Kato (0) proposed the followig cost fuctios: OC bc = 47.5VKM (6) OC cb = 65.6VKM NV (7) OC tb = 36.4VKM NV (8) where OC bc deotes the operatig cost of a bus compay (Kyats/day); OC cb, the operatig cost to a idividual bus operator owig a city-bus (Kyats/day); OC tb, the operatig cost to a idividual bus operator owig a truck-bus (Kyats/day); VKM, vehicle kilometers; ad NV is the fleet size. The reveue from the bus ad BRT operatios is simply assumed to stem from the fares paid by bus passegers. The fleet size o each route is also calculated based o the assumptio that the bus operators will adjust their fleet size to satisfy demad. Assumptios regardig the types of vehicles used are assumed o the basis of the curret situatio. The, the supplier beefits are defied as the total profit for all bus operators i the market. The user beefit is calculated o the basis of cosumer surplus theory. The cosumer surplus of the represetative bus user from zoe o to d is formulated as: CS od = l exp( Vod, r ) (9) κ c r R od where κ c is the coefficiet for the travel cost i the utility fuctio of the bus route-choice model (Williams, ). The, the user beefit is calculated as: = 0 ( ) = ( ) ( ) 0 UB Tod CSod CSod Tod l exp Vod, r Tod l exp Vod, r (0) od κc od r R od 0 od r Rod od where CS is the cosumer surplus for the represetative bus users travellig from zoe o to d i policy 0 optio, ad CS od is the cosumer surplus i the do-othig optio. The social beefit is defied as the sum of the user s beefit ad the supplier s beefit. The cost-beefit aalysis is applied o the basis of the social beefit ad the ivestmet cost. The lifetime of the project is assumed to be 5 years. The social discout rate is assumed to be 20 percet. 7

8 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu CASE STUDIES Study cases We examied two cases regardig policy optios o the bus route etwork. The first case (Case ) itroduces a ew BRT lie o the existig bus etwork. This meas that the existig bus routes remai eve after the itroductio of the ew BRT lie. The BRT operator must compete with the existig bus operators i Case. The secod case (Case 2) restructures the bus route etwork whilst itroducig the ew BRT lie. It is assumed that the bus routes are restructured ito a truk-ad-feeder etwork. The feeder routes are categorized ito five blocks, amely A, B, C, D, ad E, based o their locatio. I Case 2, the BRT lie adopts the role of the truk lie coectig the major bus termials, while the feeder bus lies ru betwee the bus stops ad the bus termials. The BRT operator ca ru a moopolistic busiess coectig the major bus termials. Case 0 is defied as the do-othig case. Evaluatios were made by comparig Case or Case 2 with Case 0. The bus route etworks assumed i each case are depicted i FIGURE 4. The followig five levels of BRT fares were examied i each case: 300, 400, 500, 600, ad 700 Kyats. It was assumed that the bus services provided by the existig bus operators i Case have the same fare system ad service frequecy as the existig system. I Case 2, the fare system adopted was a distace-based system, but this was ot adopted for the BRT lie. The service frequecy of the restructured feeder routes was defied based o expected demad. Results of case studies The evaluatio results for the case studies are show i TABLE 2, which iclude (i) the operatioal vehicles, aual reveue, aual cost ad aual profit of o-brt operators ad the BRT operator, ad (ii) the aual supplier s beefit, aual user s beefit, aual social beefit ad total social beefit over 5 years. First, the existig bus operators lose some reveue i Case. This is maily because may bus passegers shift from the existig bus routes to the ew BRT lie. Reveue icreases as the BRT fare icreases simply because bus passegers choose the cheaper bus route. The umber of operatioal buses ad the operatig costs are also reduced i Case. Cosequetly, existig bus operators do ot experiece a serious profit impact i Case. Secod, feeder operators ear more reveue i Case 2 tha i Case 0. This is because bus passegers are forced to use the feeder services twice, for access ad egress to/from the BRT termials. As the fare system is distace-based with the iitial fare for feeder service, the passegers must pay the iitial fare twice. O the other had, the feeder operators ca also reduce their operatig costs i Case 2. This reflects the reductio i vehicle kilometers because the feeder routes are shorter tha the existig bus routes. Cosequetly, the feeder bus operators ca ear a positive et profit i Case 2. Third, i Case, the BRT operator ears the most whe the BRT fare is set at 400 Kyats. If the fare is too low, reveue decreases although passeger umbers icrease. If the fare is set at too high a level, reveue is adversely affected by decliig passeger umbers. Operatig costs are lower as the BRT fare icreases because demad for the BRT decreases accordigly. Fourth, i Case 2, the BRT operator ears positive et reveues whe the BRT fare is set at 300, 400 or 500. This is because BRT passeger umbers are higher i Case 2 tha i Case. I Case 2, the bus passegers have to use the BRT as o other bus routes coectig the surburb area with the cetral busiess district exist. Operatig costs fall as the BRT fare rises. Cosequetly, the BRT operator ears positive et profits whe the BRT fare is 300, 400 or 500. Fifth, the supplier s beefit is positive i all cases. The case i which the supplier s beefit is highest is Case However, the user s beefit is oly positive i Cases -300 ad The user s beefit i Cases - 500, -600, ad -700 is egative because the BRT fare is expesive for bus users. The user s beefit i Case 2 is egative first because the jourey time icreases due to icreased trasfers, secodly because bus passegers have to pay more because of these multiple trasfers, ad thirdly, because the BRT fare is expesive for passegers. 8

9 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu FIGURE 4a Simplified bus etwork FIGURE 4b Bus route etwork (Case 0) B B D B C C A D B 8 9 D E E E E FIGURE 4c. Bus route etwork (Case ) FIGURE 4d. Bus route etwork (Case 2) Sixth, iterestigly, the social beefit is positive i all cases. Case offers the highest social beefit; followed by Case -300 ad the Case Cost-beefit aalysis The ivestmet cost assumed is based o previous experieces whe the BRT was itroduced to other lowicome coutries. We use the cost data from the experiece of itroducig the BRT system i Bogota (2). Accordig to Hook (3), the cost per kilometer was US$ 6.9 millio i Bogota. This icludes the costs for the studies ad desigs, exclusive routes, geeral traffic, public space, statios, pedestria termials, parkig, properties, etwork services, maiteace ad so o. Note that the cost of BRT itroductio i other cities was 9

10 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu TABLE 2 Evaluatio results of policy optios No-BRT BRT fare Operatioal vehicles Aual reveue Aual cost Aual profit Kyats Vehicles Millio Kyats/year Millio Kyats/year Millio Kyats/year Case 0.a Case Case Case Case Case Case Case Case Case Case BRT BRT fare Operatioal vehicles Reveue Cost Profit Kyats Vehicles Millio Kyats/year Millio Kyats/year Millio Kyats/year Case 0.a..a..a..a..a. Case Case Case Case Case Case Case Case Case Case Beefit BRT fare Supplier's beefit User's beefit Social beefit Social beefit i 5 years Kyats Millio Kyats/year Millio Kyats/year Millio Kyats/year Millio Kyats Case 0.a..a..a..a..a. Case Case Case Case Case Case Case Case Case Case cheaper tha i Bogota. For example, the cost per kilometer of itroducig BRT systems was US$ 2.5 millio i Curitiba, US$.2 millio i Quito (Eco-Via Lie) ad US$ 0.5 millio i Taipei. However, the reasos why the cost i those cities was so cheap should be carefully examied. The results of the cost-beefit aalysis are summarized i FIGURE 5. The cost-beefit ratio is more tha oe i Cases -300, -400, -500 ad Cases 2-300, 2-400, ad The highest cost-beefit ratio is.479 i Case Discussios First, the cost-beefit aalysis shows that Cases -300, -400, -500, ad Cases 2-300, 2-400, ad have cost-beefit ratios of more tha oe. This meas that they are viable from the viewpoit of the socio-ecoomic efficiecy of the ivestmet. Secod, TABLE 2 shows that the social beefit is positive i all cases. Case offers the highest social beefit, followed by Case -300 ad the Case Case offers the highest supplier s beefit although it imposes a egative user s beefit, maily because of etwork restructurig. Although Case is viable from the viewpoit of social beefits, it may be uacceptable from the user s viewpoit. Case -300 ad 0

11 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu Case 300 Case Case Case Case FIGURE 5 Cost-beefit ratios for each case BRT Passegers/BRT capacity Case 300 Case Case Case Case FIGURE 6 Load factors for BRT vehicles i each case offer positive beefits from both the supplier s ad users viewpoits. Moreover, the existig bus operators make some profit gais i Case -300 ad Case -400, makig both acceptable to them. Third, a techical problem should be poited out. FIGURE 6 shows the average load factors for BRT vehicles i these cases. This shows that the capacity of BRT vehicles does ot satisfy BRT demad i Cases - 300, -400, -500 or Cases 2-300, 2-400, ad However, if passegers are permitted to stad i vehicles, the maximum load factor could be.5 to 2. Note that the average umber of passegers per vehicle i Curitiba s BRT is 275, although oly 57 seats are provided (Evas, 4). This meas that the Curitiba BRT carries a large umber of stadig passegers. Sice may passegers have to stad i the existig bus services i Yago, this may be acceptable. If stadig passegers are acceptable, Cases -400, -500 ad Cases ad might also be feasible. Cosequetly, it is cocluded that Case -400 is the most viable. The cost-beefit ratio passes the social-ecoomic viability test. The supplier s beefit is positive. The et profit for existig operators is also positive. The load factor for BRT vehicles is quite high, but physically feasible if stadig passegers are permitted. The user s beefit is also positive. CONCLUSIONS This paper aalyzes the feasibility of itroducig the BRT system ad restructurig the bus route etwork to improve the performace of the local bus service i Yago, Myamar. First, we summarize the characteristics of the curret urba bus service i Yago. We poit out that the poit-to-poit bus route etwork causes may overlappig routes alog the mai roads ad this leads to serious traffic cogestio. The, we examie the

12 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu feasibility of two policy optios. The first case (Case ) itroduces a ew BRT lie to the existig bus etwork while the secod case (Case 2) restructures the bus route etwork i additio to itroducig the ew BRT lie. Five types of BRT fare are cosidered for each case. To evaluate these policy optios, a travel demad forecast system, icludig the estimatio of origi-destiatio table ad the bus route-choice model, was developed. The results of the evaluatio show that Case -400 i which the ew BRT with a flat fare of 400 Kyats is itroduced to the existig bus route etwork is the most viable ad beeficial optio. Topics for further research are set out as follows. First, the method for evaluatig the policy optio should be improved. For example, the bus route-choice model used i our paper does ot explicitly icorporate i-vehicle cogestio or traffic cogestio. The i-vehicle cogestio may seriously ifluece the route choice made by bus passegers. The itroductio of a exclusive lae reduces road capacity for other traffic ad this causes traffic cogestio i other laes. It is importat to cosider these factors whe evaluatig the real impact of itroducig a BRT system. Secod, it is importat to examie the social impact ad evirometal aspects of BRT itroductio. For example, the impact of itroducig the BRT o the participatio of low icome people i social activities must be ivestigated. The cotributio made by a BRT system to the local ad global eviromet should be also evaluated. Third, it is critical to examie the egieerig feasibility of itroducig a BRT system. For example, traffic cotrol at itersectios alog the exclusive lae must be carefully examied to avoid traffic accidets. The desig of the BRT stops/termial, icludig the facilities for quick boardig ad exits, should be ivestigated. It is also ecessary to cosider the itroductio of advaced techology ito the BRT service such as a smart card system, the provisio of passeger iformatio, ad a computer-based schedule cotrol system. Fiacial issues are also importat. Although the paper does ot explicitly assume the orgaizatio operatig the BRT, it may be ecessary to itroduce fiacial support for the BRT start-up. Fially, cosesus-buildig must be cosidered. The participatio of citizes, the private sector ad the public sectors is likely to be critical whe promotig the itroductio of BRT. ACKNOWLEDGEMENTS The Myamar survey was coducted uder the support of Urba Public Trasport Policy Framework i ASEAN-Japa Trasport Framework with the fiacial support of the Govermet of Japa. Mr. Tadashi Nishimura ad Mr. Ryo Masutomo (Japa Trasport Cooperatio Associatio), Mr. U Aug Myit (Road Trasport Admiistratio Departmet), Mr. Maug Maug Sue ad Mr. Ko Ya Aye Tu (SM Tours & Trasport Co., Ltd) supported our local surveys. We thak them all for their kid support. REFERENCES. Uited Natios. World Urbaizatio Prospects, 2007 revisio. 2. Zhag, J., Fujiwara, A. ad Thei, S. Capturig traveler s stated mode choice prefereces uder ifluece of icome i Yago city, Myamar. Joural of Trasportatio Systems Egieerig ad Iformatio Techology, Vol.8, No.4, 2008, pp.49 62, Kato, H., Saito, N. Iagi, A. ad Myit, U. A. Regulatory framework ad operatioal system of urba bus trasportatio i Yago, Myamar. Submitted to the 89th Trasportatio Research Board Aual Meetig, Erst, J. P. Iitiatig bus rapid trasit i Jakarta, Idoesia. I Trasportatio Research Record: Joural of Trasportatio Research Board, No.903, Trasportatio Research Board of the Natioal Academies, Washigto, D.C., 2005, pp Hesher, D. A. ad Golob, T. F. Bus rapid trasit systems: a comparative assessmet. Trasportatio, Vol.35, 2008, pp Levisom H. S., Zimmerma, S., Cliger, J. ad Gast, J. Bus rapid trasit: Sythesis of case studies. I Trasportatio Research Record: Joural of Trasportatio Research Board, No.84, Trasportatio Research Board of the Natioal Academies, Washigto, D.C., 2003, pp.. 7. Wright, L. ad Hook, W. Bus Rapid Trasit Plaig Guide, 3 rd ed. Istitute for Trasportatio ad Developmet Policy, New York, Sivalumar, T., Yabe, T., Okamura, T. ad Nakamura, F. Study desig to grasp ad compare user s attitudes o bus rapid trasit (BRT) i developig coutries. IATSS Research, Vol. 30, No.2, 2006, pp Vuchic, V. R. Bus semirapid trasit mode developmet ad evaluatio. Joural of Public Trasportatio, Vol. 5, No.2, 2002, pp

13 Hiroori Kato, Akihiro Iagi, Nozomi Saito, ad Phyo Thet The Htu 0. Kato, H. Improvemet of Urba Bus Trasport System i Yago, Myamar, Fial Report, ASEAN-Japa Trasport Partership Project, 2009, pp. 50, Prepared for Miistry of Lad, Ifrastructure, Trasport ad Tourism of Japa.. Williams, H. C. W. L. O the formatio of travel demad models ad ecoomic evaluatio measures of user beefit. Eviromet ad Plaig A, Vol. 9, 977, pp Targa, F. ad Rodriguez, D. A. Aalysis of Bogota s bus rapid trasit system ad its impact o lad developmet. Carolia Plaig Joural, Witer, , pp Hook, W. Fiacig bus rapid trasit: Optios for Chia. Preseted at the Iteratioal Mayors Forum o Sustaiable Urba Eergy Developmet, 2004, pp Evas, J. J. Capacity ad cost comparisos of rapid trasit modes. Preseted at the Istitute of Trasportatio Egieers Aual Meetig, Melboure, August