John Paul Helveston 1, Yimin Liu 4, Elea McDonnel Feit 3, Erica Fuchs 1, Erica Klamfl 4, Jeremy Michalek 1,2

Transcription

1 Consumer Preferences for Hybrid and Electric Vehicles in China and the United States John Paul Helveston 1, Yimin Liu 4, Elea McDonnel Feit 3, Erica Fuchs 1, Erica Klamfl 4, Jeremy Michalek 1,2 October 7, 213 Department of Engineering and Public Policy 1. Department of Engineering & Public Policy, CMU 2. Department of Mechanical Engineering, CMU 3. Department of MarkeDng, The Wharton School, University of Pennsylvania 4. Ford Motor Company

2 China is the world s largest vehicle market...and growing Annual New Vehicle Sales (Mil) China 212 U.S October 7, 213 Vehicles per 1, people China StaDsDcal Yearbook, 212 NaDonal Bureau of Economic Research U.S. Department of TransportaDon, Federal Highway AdministraDon, Highway Sta+s+cs 211

3 EVs transition vehicles from gas to electricity All Gas All Electric Conven&onal Hybrid- Electric Plug- in Hybrid Electric Ba8ery Electric AbbreviaDon CV HEV PHEV BEV Power Convertor Engine Engine & Motor Engine & Motor Motor BaUery Pack - - Small Medium Large Gasoline - - Electricity October 7, 213

4 EVs have potential to reduce oil consumption & life cycle emissions Oil Consumption Problem: U.S. and China consume ~1/3 of global oil consumpdon. Impact: PHEVs expected to use about 4-6% less petroleum than convendonal vehicles (ANL 21). BEVs use no oil! 3 Emissions Problem: In U.S., cars contribute to 2% of GHG, 4% of VOC, 77% of CO, 5% NOx. In China, similar, even worse for NOx / CO. Impact: Depends on grid mix; PHEVs have potendal to reduce GHG by 32%. 1 Life cycle GHGs (g CO2 eq/km) CV HEV PHEV3 PHEV6 BEV Low carbon U.S. Average China Average Natural gas Coal Life cycle electricity GHG intensity (g CO2 eq/kwh) 1) Samaras & Meisterling (28), Environmental Science & Technology. (Figure adapted to include BEV). 4 October 7, 213

5 Research Questions How do existing preferences shape adoption of electrified vehicles in America & China? o Under what conditions would mainstream adoption of electrified vehicles occur? o What are the implications of these preferences for policy? 5 October 7, 213

6 Measuring Consumer Preferences How do you get data on vehicles that don t exist yet? Our Approach: Choice-Based Conjoint Surveys 6 October 7, 213

7 Past Automotive Demand Modeling Paper Electrified Year Model Form Vehicles Conjoint Data years Loca&on Boyd & Mellman 198 MNL, MXL USA Goldberg 1995 NL, MNL USA Berry, Levinsohn, & Pakes 1995 MNL, MXL USA McCarthy 1996 MNL USA Golob et al P x x 1994 USA Goldberg 1998 NL, MNL x 1993 USA Brownstone & Train 1999 MNL, MXL, P x x 1993 USA Brownstone, Bunch, & Train 2 MNL, MXL x x 1993 USA McFadden & Train 2 MXL x x 1993 USA Sudhir 21 MXL USA Choo & Mokhtarian 24 MNL x 1998 USA Berry, Levinsohn & Pakes 24 MXL 1993 USA SanDni & Vyas 25 MNL x x USA Train and Wilson 27 MXL 2 USA Dagsvik and Liu 29 NL x 21 Shanghai Axsen, Mountain, Jaccard 29 MNL x x 26 USA & Canada Axsen, Kurani x 27 USA Ziegler 212 P x x Germany This Study 213 MXL x x 212 USA & China MNL = MulDnomial Logit, NL = Nested Logit, MXL = Mixed Logit, P = Probit 7 October 7, 213

8 Conjoint Design 8 October 7, 213

9 China Fielding: Jul. 16 Aug. 16, 212 On the ground Partnered with State Information Center to field on the ground in China Targeted 4 major cities: Beijing, Shanghai, Shenzhen, & Chengdu Tier 1 cities Largest car markets Not rural Gov t focus for EVs Geographically diverse 35% of 21 total sales Chengdu 448 Respondents 15 Choice QuesDons Each 6,72 Choice QuesDons Beijing Shenzhen Shanghai 9 October 7, 213

10 U.S. Fielding: Sept. 212 & Feb. 213 & Pi8sburgh Auto Show 283 Online Respondents 11 Auto Show Respondents 15 Choice QuesDons Each 5,76 Choice QuesDons Online car buyers identified through short screener survey, compensated $2.. Auto show car buyers taken at random, compensated with ipod raffle. Full Sample Auto Show Sample Urban Suburban Rural 34.7% 29.6% 35.7% Urban 26.3% Suburban 41.1% Rural 32.6% 1 October 7, 213

11 Weights added to match car-buying demographic Sample demographics compared to large new car buyer survey by Maritz (supplied by Ford) Variable U.S. China Our Sample Weighted Sample Maritz Sample Our Sample Weighted Sample Maritz Sample Household Income Age Num Children Num Vehicles Daily VMT Annual VMT Household Size Years EducaDon Percent Female Percent Married Percent with No Children Percent College Graduates Percent First Time Buyers n 57.3 (29.3) 74.3 (28.7) 74.8 (27.3) 33.9 (12.7) 51 (14.8) 53.1 (15.4).6 (1.1) 1.4 (1.4).4 (.8) 1.8 (.8) 2 (.7) (1.4) 23.3 (11.4) ,2 (4,8) 12,5 (4,6) 11,4 (6,4) 2.7 (1.3) 2.7 (1.2) 2.5 (1.2) 7.2 (1.9) 7.9 (2.3) 7.2 (2.3) 35.3% 32.6% 39.3% 44.6% 68.9% 73.5% 72.1% 4.3% 75.% 52.3% 71.2% 53.7% 4.4% 1.3% , (15.7) 26.1 (18) 26.1 (17.6) 33.3 (1.6) 34.8 (7.8) 35.1 (7.8).6 (.6).7 (.6).7 (.6).4 (.6).5 (.7) ,6 (6,) 3.3 (1.1) 3.3 (1.2) 3.2 (1) 5.9 (1.9) 6 (1.8) 5.9 (2) 39.4% 41.1% 28.7% 55.1% 7.2% 85.6% 52.2% 36.5% 36.4% 3.6% 33.1% 34.4% 65.4% 59.2% , September 29, 214

12 Discrete Choice Models: Linking Attributes to Choice 12 October 7, 213

13 Automotive Demand Modeling Random UDlity Model: Es&mated partworth u&lity weight Observed a8ributes U&lity to person n from choosing alterna&ve j =U nj = β ' x nj +ε nj Unobservables Goal of inference: To understand the rela&onship between product a8ributes and purchase proclivity 13 October 7, 213

14 Use mixed logit model to link attributes to choice Logit Model Estimates average weights for entire population. P ni = eβ x ni j eβ x nj. Mixed Logit Model Estimates distribution of weights across population. P ni = ( e β x ni x j eβ nj ) φ(β b, W ) dβ, P ni β x ni = Probability of choosing an alternadve from the set of choices. = RelaDve weight of auribute. = AUribute of alternadve. *Berry, Levinsohn, & Pakes (1995) Brownstone & Train (1999) McFadden & Train (2) 14 October 7, 213

15 U.S. more opposed to BEVs than China WTP = Increasing bauery size Increasing electric range β att β price Vehicle Technologies Vehicle Technology Baseline: CV HEV PHEV1 PHEV2 PHEV4 BEV75 BEV1 BEV15 China U.S Willingness- to- Pay WTP ($1,) U.S. consumers sensidve to limited electric range (Axsen & Kurani, 212). Most Chinese are first- Dme car buyers. Chinese more familiar with plug- in vehicles (2 3 million e- bikes currently on the ground). 15 October 7, 213

16 WTP = Brand, Fuel Economy, & Acceleration Important β att β price Brand Brand Baseline: German American Japanese Chinese SKorean Performance Performance Willingness-to-pay ($1,) Performance PHEV Fast Charge Ability BEV Fast Charge Ability Reduce Op. Cost by 5 cents / mile Reduce -6 mph Accel. Time by 4 sec Willingness-to-pay ($1,) 16 October 7, 213

17 Under what conditions would a plug-in vehicle gain comparable market share against it s gasoline counterpart? A8ribute CV Focus BEV Focus Price $19, $4, Type CV BEV1 Brand American American Op. Cost (cents/mile) - 6 mph AcceleraDon Time Subsidy ($1, USD) China Subsidy U.S. Subsidy BEVs PHEVs Subsidy (1, RMB) Federal Subsidy - - U.S.: $7,5 China: $9, Battery capacity (kwh) 17 October 7, 213

18 PHEVs attractive in both countries, current subsidy plays small role CV / HEV Prius HEV C- Max HEV PHEV Prius PHEV1 C- Max PHEV2 Plug in vehicle share (%) PHEV1 vs. HEV (Toyota Prius) Subsidy ($1, USD) Current Subsidies U.S. China Plug in vehicle share (%) PHEV2 vs. HEV (Ford C Max) Subsidy ($1, USD) Current Subsidies U.S. China CV CV PHEV4 PHEV4 Plug in vehicle share (%) PHEV4 vs. CV (BYD F3) Current Subsidies U.S. China Plug in vehicle share (%) PHEV4 vs. CV (Volt vs. Cruze) Current Subsidies U.S. China Subsidy ($1, USD) Subsidy ($1, USD) 18 October 7, 213

19 Chinese more willing to adopt BEVs, China subsidy important CV Versa CV BEV Leaf BEV75 Plug in vehicle share (%) BEV75 vs. CV (Leaf vs. Versa) Current Subsidies U.S. China Subsidy ($1, USD) Focus CV Focus BEV1 Plug in vehicle share (%) BEV1 vs. CV (Ford Focus) Current Subsidies U.S. China Subsidy ($1, USD) 19 October 7, 213

20 Interaction of Preferences & Policy Preferences Preferences + Federal Subsidy Impacts (rela&ve to CVs) U.S.: PHEVs neutral. BEVs hard sell. PHEVs slightly more auracdve. BEVs sdll hard sell. Oil consumpdon ê Emissions ê China: No strong preference towards PHEVs vs. BEVs. Far less opposed to BEVs than U.S. BEVs become more auracdve than PHEVs. Oil consumpdon ê Emissions (é ê?) 3 Life cycle GHGs (g CO2 eq/km) CV HEV PHEV3 PHEV6 BEV Low carbon U.S. Average China Average Natural gas Coal Life cycle electricity GHG intensity (g CO2 eq/kwh) 2 October 7, 213

21 Acknowledgements This study was funded by: Ford Motor Company. The NaDonal Science FoundaDon. Special thanks to:. Jiang Zhijie, Zheng Wei, and Zang Ye at the State InformaDon Center in Beijing, China. Center for Climate and Energy Decision Making. 21 October 7, 213