HOW FUTURE PLUG-IN ELECTRIC VEHICLE BUYERS DIFFER FROM PRESENT OWNERS

0 HOW FUTURE PLUG-IN ELECTRIC VEHICLE BUYERS DIFFER FROM PRESENT OWNERS Dr. Jonn Axsen Simon Fraser University University Ave., Burnaby B.C., VA S Tel: + Fax: + ; Email: jaxsen@sfu.ca Suzanne Goldberg Simon Fraser University University Ave., Burnaby B.C., VA S Tel: + Email: sgoldber@sfu.ca Word count: + ( tables/figures x 0) =, words [excludes bibliography] Submission Date: August st, 0

Axsen, Goldberg 0 ABSTRACT The entry of various plug-in electric vehicles (PEVs) into the passenger vehicle sector provides novel opportunities to learn about the potential future PEV markets. However, early PEV buyers (or Pioneers ) can substantially differ from buyers of conventional gasoline vehicles that have interest in purchasing PEVs in the future (or the potential Early Mainstream buyers). To directly compare Pioneers and Early Mainstream buyers in terms of characteristics, preferences and motivations, we draw data from the Canadian Plug-in Electric Vehicle Study, which implemented a three-part mixed-mode survey with samples of PEV owners (n = ) and conventional new vehicle buying households (n = ). We identify several statistically significant differences in household characteristics, including income, education, and recharge access. In terms of preferences, Pioneers express extremely high valuation of PEVs and prefer pure battery electric vehicle (BEV) designs over plug-in hybrid electric (PHEV) designs. In contrast, Early Mainstream respondents greatly prefer PHEVs. Both Pioneer and Early Mainstream respondents are similarly cautious about controlled charging programs, but Pioneers place five times as much value on using electricity generated from renewable sources. Pioneers also tend to have different motivations, being significantly more likely to have higher levels of environmental concern, and to engage in environment- and technology-oriented lifestyles. Policymakers, automakers, and electric utilities that anticipate a transition to electric mobility ought to consider how future PEV buyers may significantly differ in their technology preferences, usage and motivations relative to current PEV owners. Keywords: plug-in electric vehicles; consumer behavior; choice model; controlled charging; renewable electricity

Axsen, Goldberg 0 0 0. INTRODUCTION Despite much effort and policy support, alternative fueled vehicles have repeatedly failed to displace gasoline and diesel powered passenger vehicles in the vast majority of countries. In recent years, plug-in electric vehicles (PEVs) have shown promise in several markets, including pure battery electric vehicles (BEVs) and plug-in hybrid electric vehicles (PHEVs) that can be powered by gasoline and plugged in to recharge. In 0, PEV sales made up less than one percent of new light-duty vehicle sales in the U.S. () and Canada (), though a few regions have achieved much greater sales rates, including Norway where market share was percent in 0 () and California where market share was over three percent () Increased PEV uptake may be signaling the beginning of a transition to widespread electric mobility, and many stakeholders want to better understand PEV buyers, including preferences, motivations and potential PEV usage patterns. Specifically, national and sub-national policymakers wonder how best to incentivize and support further growth in the PEV market; automakers want to understand what PEV designs and attributes will be valued by consumers; and electric utilities want to know how widespread adoption of PEVs might impact electrical load, and how PEV charging might be integrated with intermittent, renewable sources of electricity. This study explores several aspects of the potential PEV market, focusing on a comparison between current PEV owners (who we call Pioneers ) and conventional new vehicle buyers that might purchase PEVs in the future (the Early Mainstream ). While there may be lessons to learn from current Pioneers, it is arguable that Early Mainstream buyers are a particularly important target for analysis. Acceptance must occur in this Early Mainstream segment for PEVs to be successful in displacing fossil fuels and effectively decarbonizing the passenger vehicle sector. While there is little research that directly compares Pioneers and Early Mainstream buyers, initial insights suggest that important differences exist in terms of demographics and context (, ), and in preferences for PEV and renewable electricity programs (). This study uses samples of new vehicle buying households in Canada, with a focus on the Canadian province of British Columbia a region with PEV-supportive policy (subsidies and deployment of recharge infrastructure) but with PEVs making up only one percent of new market share in 0. Through in-depth data collected as part of the Canadian Plug-in Electric Vehicle Study, or CPEVS (), we compare Pioneers and Early Mainstream respondents according to household demographics and context, PEV awareness and preferences, purchase and use motivations, as well as interest in controlled charging programs that could manage how and when electricity flows from the grid to a PEV (which could support the use of intermittent, renewable sources of electricity). We next summarize previous literature on these topics, explain our methods and results, and conclude with implications for policy.. BACKGROUND. Distinguishing Consumers by Time of Purchase Consumers can vary widely in their perceptions, tastes, preferences and motivations. It is therefore useful to identify different groups or segments of PEV buyers. Conventional approaches to PEV market segmentation tend to focus on innovativeness as the key segmentation parameter (). Perhaps the most popular innovation model is Rogers () diffusion of innovations model, which separates potential buyers into innovators, early adopters, early majority, late majority, and finally laggards. However, we find that Rogers model is too limited in its representation of human motives a full critique is provided in elsewhere (), but in short it has been shown that PEV buyers can be driven by many motives other than technological innovativeness (). Thus, we find it more useful to segment the actual and potential PEV market according to behavior, specifically

Axsen, Goldberg 0 0 0 the act or likelihood of purchasing a PEV, rather than segmenting based on a single motivation (e.g. innovativeness) that hasn t been shown empirically to be a central driver of purchase behaviour. We thus consider three segments in this paper:. Pioneers are the very first buyers of PEVs;. The potential Early Mainstream are conventional vehicle buyers that have interest in PEVs and are the most likely to buy one in the next - years (if supply becomes available); and. The Late Mainstream (or non-buyers) do not have interest in PEVs and are not likely to buy one in the next - years, and substantial changes in policy, costs, technology, or cultural norms will be required for them to become interested in PEVs.. Previous Literature on PEV Pioneers versus the Mainstream We identify only three previous studies that directly compare PEV Pioneers and Mainstream buyers, including a survey of 0 U.S. consumers (), comparisons of Pioneer (n=,0) and Mainstream (n=,00) survey respondents in California (), and an online survey of Pioneers and Mainstream respondents (n=,) in Germany (, ). Other literature on PEV demand which is largely comprised of studies that look at only Pioneers (-) or only Mainstream buyers (, ). As a starting point of comparison, studies consistently find that Pioneers are demographically distinct from Mainstream car buyers. In particular, Pioneers tend to have significantly higher income and education levels than Mainstream buyers (,, ), and tend to be older (,, ). Pioneers also tend to be male (, ), accounting for up to % of Pioneer samples (). In terms of household characteristics, Pioneers in California and Germany tend to be more likely to live in multi-car households (, ) and U.S. and California Pioneers are more likely to live in detached homes (, ). While about half of U.S. Mainstream buyers have Level (0/0 V) access at home (), almost all U.S. Pioneers have access to Level charging at home (). About, % of San Diego Mainstream respondents have relatively easy Level (0/0 V) home charging potential, and a further % could install Level at a higher cost (). As a point of reference, over 0% of Pioneers in California have installed Level home chargers () though home Level installation is about twice as common among BEV owners relative to PHEV owners (). In terms of PEV perceptions, there is evidence that U.S. Mainstream buyers have much lower levels of PEV familiarity and awareness, and are generally confused about the idea of a PHEV and how to refuel it (, 0). Pioneers and Mainstream buyers also differ in their preferences for PEVs, where U.S. Pioneers are up to three times more likely to want to purchase a PEV as their next vehicle (), and German Pioneers are willing-to-pay up to twice as much for a PEV design (). Further, these segments differ in terms of the type of PEV they want; once the concept of PHEVs and BEVs are explained, U.S. Mainstream buyers express relatively more interest in PHEVs (,, ), while most Pioneers express more interest in BEVs (, ). Relatively less research has explored Pioneer and Mainstream preferences for renewable electricity and controlled charging programs relating to PEVs. Exploratory research suggests that U.S. Pioneers are almost three times as likely to be willing to pay extra to power their PEV with some form of green or renewable electricity (, ). Pioneers are also more likely to have installed home solar panels, accounting for % of California Pioneers compared to less % of the general California population (, ). Exploratory research on demand for controlled charging and vehicle-to-grid programs suggests that many U.S. Mainstream buyers may be reluctant to enroll due to inconvenience (), though no comparison to Pioneers has been conducted. In terms of consumer motivations, many studies find that consumer interest in PEVs and

Axsen, Goldberg 0 0 0 renewable electricity is associated with concern for the environment, and pro-environmental attitudes and lifestyles (,,,, ). Other studies also find that PEV interest is associated with technology interest or engagement in a technology-oriented lifestyle (,, ). Lastly, some studies find a connection between financial savings and PEV interest (, ); however, the importance of costs savings may vary across Pioneers owning different PEV models, for example with California Volt owners being more likely to be motivated by cost savings relative to Leaf owners ().. METHODS. Data Collection and Survey Data were collected as part of the Canadian Plug-in Electric Vehicle Study (CPEVS) which is described in full elsewhere (). CPEVS collected data from two surveys: the British Columbia Pioneers survey and the Canadian Mainstream survey. We presently focus on respondents from the province of British Columbia here as a case study. The target population for the Mainstream survey was new vehicle buying households in English-speaking Canada. The web-based survey recruited a representative sample of households in 0 and included an oversample of respondents from British Columbia (n = ) to permit regional analysis. Our Pioneer survey targeted PEV owners in British Columbia specifically households that had purchased or leased a commercially-available PEV. We did not include owners of after-market conversions. Participants were recruited from 0 to early 0 via invitations sent to a list of recipients of PEV purchase or home charging installation rebates from the British Columbia Government s Clean Energy Vehicle Program, members of the Vancouver Electric Vehicle Association (VEVA) and subscribers to a PEV promotional campaign in British Columbia (called Emotive BC ). In total, respondents completed the entire Pioneer survey. CPEVS implemented similar three-part survey instruments with both the Pioneer and Mainstream samples. Part of the CPEVS survey was a web-based survey that collected background details of each respondent s vehicle fleet, energy usage, and knowledge and perceptions about PEVs. Part was a mail-out and web-based survey package that included a technology primer with basic information on the vehicle technologies and charging scenarios discussed in Part, as well as a driving diary and a charging assessment. Mainstream respondents were asked to complete a -day driving diary and indicate potential charging opportunities at home and the other parking locations, while Pioneer respondents were asked to complete a -day driving diary and provide a log of charging activity. Part was a web-based survey that elicited respondent interests and preferences for PEV technology and controlled charging programs via stated choice experiments and design space exercises (described in Sections. and.).. PEV Design Games The CPEVS survey uses design space exercises to elicit respondent interest in PEVs. These exercises provide respondents with a space or design envelope a series of design options that respondents could select to create their preferred design given a particular context, which has been demonstrated in previous research (, -). The PEV design space exercise was nearly identical for Pioneer and Mainstream respondents. Each respondent first selected a currently available base vehicle (conventional gasoline powered) that represented the vehicle they would buy next if they were limited to a conventional vehicle, including its body type, price and fuel economy. In the design space, respondents selected from four vehicle drivetrains: (i) conventional vehicle version (base vehicle); (ii) a hybrid version (with % improved fuel economy over the base); (iii) a PHEV version powered with electricity for the first,, or kilometres

Axsen, Goldberg 0 0 (achieving the same % increase in fuel economy until the battery is recharged); and (iv) an BEV version powered only by electricity for 0, 0, 0, 00, or 0 kilometres of range (the Pioneer survey included an additional 0km option). Respondents completed higher price and lower price scenarios for the vehicle design exercises, with increment prices based on a simple battery cost model constructed by Axsen and Kurani (), and summarized in Table. We select the 0% of Mainstream respondents that selected a PEV design (BEV or PHEV) in the lower price scenario as the Early Mainstream. TABLE PEV Design Space Exercise Options and Prices (Prices Incremental to Respondents Next Anticipated Conventional Vehicle) Higher Price Scenario Lower Price Scenario Vehicle type Compact Sedan Mid-SUV Full-SUV Compact Sedan Mid-SUV Full-SUV HEV $0 $0 $00 $0 $0 $0 $00 $0 PHEV- $0 $0 $0 $0 $0 $ $0 $0 PHEV- $0 $0 $ $00 $ $0 $0 $0 PHEV- $0 $0 $0 $0 $0 $0 $0 $0 BEV-0 $00 $0 $0 $0 $0 $0 $00 $00 BEV-0 $0 $0 $0 $00 $0 $00 $0 $0 BEV-0 $0 $00 $ $0 $0 $0 $0 $0 BEV-00 $0 $ $0 $00 $0 $0 $0 $0 BEV-0 $0 $0 $0 $0 $0 $0 $0 $00 Pioneer survey (only) BEV-0 $0 $0 $00 $0 $0 $00 $0 $0 Note: HEV=hybrid electric vehicle (e.g. Toyota Prius), PHEV=plug-in hybrid, BEV=pure battery electric vehicle. Stated Choice Experiments In addition to design space exercises, respondent interest in PEVs can also be assessed via discrete choice models. Part of the CPEVS survey included two stated choice experiments, which we used to estimate discrete choice models for vehicle preference and controlled charging. Discrete choice models quantify consumer preferences and are frequently utilized to model consumer demand for alternative-fuel vehicles (0-). Consumer utility is represented by a vector of coefficients weighted to the specified attributes of the product in question, e.g. purchase price and fuel costs for a PEV. The alternative specific constant represents the observable utility of each choice that is not captured by attributes specified in the model, and represents intangible vehicle benefits or drawbacks like symbolic values, safety concerns, or perceived inconveniences. In addition to determining the coefficients, we can also estimate the average willingness-to-pay for certain attributes that are significant at a % confidence level or greater. Base vehicle data (make, model, purchase price, fuel costs) from the design exercise were then used to present six customized choice sets to each respondent, where each choice set presented four different vehicle options: a conventional vehicle (respondent s base vehicle), as well as a hybrid, plug-in hybrid and pure battery electric version of that vehicle. Respondents were asked to select one of the four vehicle options in each of the six choice sets. Table depicts the experimental design for the choice sets, which included incremental purchase price premiums, fuel cost differences, and electric-powered driving ranges as well as the access to slower (Level ) or faster (Level ) vehicle charging at respondents homes.

Axsen, Goldberg TABLE PEV Choice Model Experimental Design ( Choice Sets per Respondent) Attributes Next anticipated conventional vehicle Hybrid vehicle Plug-in hybrid vehicle Battery electric vehicle Purchase price premium Selected by respondent Conventional price % more 0% more 0% more Conventional price % more 0% more 0% more Conventional price % more 0% more 0% more Weekly fuel cost Selected by respondent 0% less 0% less 0% less % less 0% less 0% less 0% less 0% less 0% less 0% less 0% less 0% less Electric-driving range n/a n/a km km km 0 km 0 km 00 km 0 km Home recharge access n/a n/a Level ( kw) Level ( kw) Level ( kw) Level ( kw) 0 Recharge time a n/a n/a Calculated Calculated a The discrete choice experiment showed recharge time to respondents to help them understand the recharging needs of the PHEV or BEV. Recharge time was calculated as the time required for the respondent to fully recharge a depleted battery using their home charger. This time is a function of the vehicle s electric driving range, the base vehicle type (where larger vehicle bodies are assumed to require more electricity consumption or have a higher kwh/km), and the speed of the home charger (Level or Level ). We also used a choice experiment to assess respondent interest in controlled charging for the purpose of increased renewable integration. Controlled charging was framed as participants electric utility potentially delaying charging of a PEV until later in the evening or taking electricity from the vehicle s battery when it is plugged-in. We used four attributes to represent controlled charging in the choice exercises: percentage of green electricity, source of green electricity (wind, solar, small hydro or mixed ), electricity costs and guaranteed minimum charge. Each attribute was assigned four levels that varied between the alternatives depicted in each choice set (Table ). To represent the potential inconvenience associated with controlled charging, we use the attribute guaranteed minimum charge (GMC), described as: the minimum level of charge that your battery would have after a night of being plugged-in. We represented this GMC as a percentage of charge (e.g. 0% charged) as well as resulting electric driving range in km (dependent on their vehicle design). As with the vehicle design exercises, respondents completed a series of six choice sets which were customized to their current home charging access and electricity costs. Pioneer respondents completed the choice experiment for the PEV design that they currently own. Mainstream respondents completed up to two scenarios: i) with the PEV they selected in the PEV design exercise (if they selected one), and ii) with a BEV-0km version of their base vehicle.

Axsen, Goldberg TABLE Controlled Charging Choice Model Experiment Design ( Choice Sets per Respondent) Electricity system powering respondent home and vehicle Attributes Status Quo Alternate Design Alternate Design Percentage of Green Electricity: Percentage of current electric supply powering the respondents home and vehicle. Current Green Electricity% % 0% % 0% % 0% % 0% Source of Green Electricity: The source of the green electricity to supply the respondents home and vehicle. Existing grid supply mix Wind Solar Small Hydro Mixed Wind Solar Small Hydro Mixed Guaranteed Minimum Charge: The amount of charge that the vehicle s battery would have the next morning. This was displayed to the respondent as both percentage charge and electric range in km. 0% charge 0% charged 0% charged 0% charged 0% charged 0% charged 0% charged 0% charged 0% charged 0 Monthly Electricity Bill: Current bill or with green electricity, user s current electric bill plus the expected cost of charging a vehicle multiplied by a scalar. Current bill provided by respondent 0% of current bill 0% of current bill 0% of current bill 0% of current bill 0% of current bill 0% of current bill 0% of current bill 0% of current bill. Respondent Motivations We assess consumer motivation through several question scales included in Part of the CPEVS surveys. Engagement in an environment- or technology-oriented lifestyle was assessed via questions of respondent engagement in activities with -point response categories ranging from never to very frequently, using an approach similar to Axsen et al. (). The five environment-oriented activities included engaging in environmental conservation activities, attending environmental meetings and promoting environmental conservation (talking to people about the environment). The five technology-oriented activities included researching new technology, shopping for new technology and working on or tinkering with technology. For each lifestyle category, we create a single composite score by summing their responses to the five questions. We also assessed respondent liminality or openness to change via a nine question scale, which has been found to be associated with PEV interest (). CPEVS also included a measure of environmental concern using a brief, eight-item version of the New Environmental Paradigm (NEP) scale (). Finally, the survey included a -item scale that assessed respondent motivation according to biospheric, altruistic, egoistic and traditional values ().

Axsen, Goldberg 0. RESULTS. Household Characteristics and Context Table summarizes household demographic and context data from our sample of Pioneers and Mainstream respondents residing in British Columbia, the Canada-wide sample of Mainstream respondents, and census data for the British Columbia and Canada general populations. As a starting point, our sample of Mainstream vehicle buyers seems to be representative of the population, in that respondents tend to have higher income and higher education than the general census population to the degree found in previous studies (-). It is more difficult to determine the representativeness of the Pioneer sample as we do not have data on the full population of PEV owners, and that population is changing regularly. TABLE Comparing Pioneer and Mainstream Demographics in British Columbia and Canada Demographics and context British Columbia Samples Early Pioneer Mainstream Late Mainstream Census Canada Sample Mainstream a Census Sample Size 000 Femal (respondent) *** % % % % % % Age (respondent) *** - % 0% % 0% 0% % - % 0% % % % % - % 0% % 0% 0% % - % % % % % % + % % % % % % Household income (pre-tax) *** Less than $0,000 % % % % % % $0,000 to $, % % % % % % Greater than $,000 % % % % % % Education (respondent) *** Other 0% % % % % 0% Diploma or some university % % % % % % University degree (Bachelor) % 0% % % % % Graduate or professional degree 0% % % % % % Residence type *** Detached House % % % % % % Attached House % % % % % % Apartment % % % % % 0% Mobile Home % % % % % % Own residence *** % % % 0% % % Vehicle ownership *** % % % N/A N/A N/A % % % N/A N/A N/A or more % % % N/A N/A N/A Level or home charging access *** % % N/A N/A N/A Note: Data on household size, sex, age, and residence type are from the 0 Canada Census. Data on work status, education, and income are from the 00 Canada Census. Data on home ownership are from the Canadian Mortgage and Housing Corporation: http://www.cmhcschl.gc.ca/odpub/esub//_0_a0.pdf?fr=0 a Overall Canada sample is unweighted. Survey data includes only English-speaking Canada Quebec was excluded due to language translation costs. Census data includes Quebec. *** Differences indicated between Pioneers, Early Mainstream and Late Mainstream British Columbia samples using chi-squared analysis at % confidence level

Axsen, Goldberg 0 0 All differences in demographics and household context between the three segments of British Columbia respondents are significant at a % confidence level. Overall, Pioneers report significantly higher incomes (being five times more likely to earn over $k/year), and education (being about three times more likely to have a graduate degree) than the Mainstream. A large proportion of Pioneers are male (%), and Pioneers are more likely to be in the - age range (%) compared to Mainstream respondents (-%). Pioneers are also significantly more likely to own their homes, and to live in single-family detached homes. Further, % of Pioneer respondents reported having home charging access, compared to % of Mainstream respondents with Level charging access (access to a 0/0V outlet). Pioneers are also more likely to be multi-vehicle households (%) compared to Mainstream respondents (-%). The Early and Late Mainstream samples are not as substantially different from one another; age, income, education, gender and household type are not significantly different at a 0% confidence level.. PEV Awareness Mainstream respondents have much lower PEV familiarity and awareness than Pioneer respondents (Figure ). Only a minority of Early Mainstream respondents reported that they were familiar or very familiar with the Nissan Leaf (%), Chevrolet Volt (0%) and Toyota Prius (%), while most Pioneer respondents reported a high degree of familiarity (-%). The majority of Early Mainstream respondents also demonstrated confusion with each of the three vehicle models operation. Only 0% successfully indicated that the Toyota Prius is fuelled only by gasoline, and one-third successfully identified the correct fueling method for the Chevrolet Volt and the Nissan Leaf. In contrast, the vast majority of Pioneer respondents (0-%) were able to identify the correct fuelling method for the Nissan Leaf and Chevrolet Volt. Late Mainstream respondents demonstrated very similar familiarity and awareness as the Early Mainstream respondents (see light blue bars in Figure ). % of respondents 0% 0% 0% 0% 0% 0% Early Mainstream Late Mainstream Toyota Pruis Chevrolet Volt Nissan Leaf Toyota Pruis Chevrolet Volt Nissan Leaf Familiar with... Pioneer Identified correct fuel for... FIGURE Comparison of PEV Familiarity and Refuelling Between Mainstream (N=, Blue Bars) and PEV Owners (Pioneers, N=, Green Bars) From the British Columbia Samples.

Axsen, Goldberg 0 0. Preferences for PEV Designs (Design Space and Choice Exercises) Both the design space exercises and choice experiments indicate that most Pioneer respondents tend to prefer BEV designs, while Early Mainstream respondents prefer PHEV designs. Across the higher and lower price scenarios, just over one-third of Mainstream respondents selected a PEV in the design game, with to % selecting a PHEV and only -% selecting a BEV (Figure ). In contrast, Pioneer respondents had very high interest in PEVs and almost all participants selected some type of PEV in both price scenarios (-0%). Pioneers were almost evenly split between BEV and PHEV designs in the higher price scenario, but not in the low price scenario where more than two-thirds selected a BEV. However, we must be careful in comparing the results of the Pioneer and Mainstream respondents because the Pioneer version of the design space exercise included one additional BEV option (one with 0km of range). That said, BEV interest is much higher among Pioneer respondents even when ignoring the BEV-0. % of Respondents 0% 0% 0% 0% 0% 0% 0% % 0% PHEV km km km BEV 0 km 0 km 00 km 0km 0 km 0 km PHEV BEV PHEV BEV PHEV BEV Mainstream Respondents Pioneer Respondents High Price n= Low Price n= High Price n=0 Low Price n= FIGURE Comparing PEV Designs Among Pioneer and Mainstream Respondents (Higher and Lower Price Scenarios). The multinomial logit model estimated from the PEV choice experiment revealed similar results as the design exercise (Table ). Note that the Mainstream model is estimated from the full Canada sample (not just British Columbia respondents), as this larger sample provides a better statistical model (with no apparent differences in comparison to the British Columbia-only model). For Mainstream respondents, the alternative specific constants indicate that when holding all other attributes constant, the HEV is the most desirable drivetrain (relative to the conventional vehicle) followed by the PHEV, then a conventional vehicle and lastly the BEV. The willingness-to-pay calculations in Table indicate that Mainstream respondents, on average, would be willing to pay an additional $ for a PHEV, and would have to be compensated over $,000 to purchase a BEV. With Level (0/0 V) home charging access, however, both vehicle designs are more valuable to participants, increasing willingness to pay by $, for PHEVs, and by $, for BEVs. Interestingly, valuation of BEV and PHEV range was not statistically significant.

Axsen, Goldberg TABLE PEV Discrete Choice Model Accompanied by WTP Estimates and Model Summary Statistics for Early Mainstream and Pioneer Samples Attributes Mainstream Pioneer (Canadian Sample) (BC Sample) Alternative Specific Constants HEV constant 0.0 *** -0. PHEV constant 0. **.0 *** BEV constant -.0 *** 0. ** Base = Conventional Vehicle Vehicle Attributes Vehicle price (CAD$) -0.000 *** -0.0000 *** Fuel cost (CAD$/week) -0.00 *** -0.00 ** PHEV range (km) -0.000 0.0 *** BEV range (km) 0.000 0.00 *** PHEV x Level charging at home 0. *** 0. ** BEV x Level charging at home 0. *** 0.0 *** Implied willingness-to-pay a Saving $00/year in fuel $, $, HEV b $, N/A PHEV $ $, BEV -$, $, km PHEV range b N/A $ km BEV range b N/A $ PHEV with Level charging $, $, BEV with Level charging $, $, PHEV- km b N/A $, BEV - 0km b N/A $, Model Number of observations Number of individuals R-square 0. 0. Log-likelihood - - * Significant at 0% confidence level ** Significant at % confidence level *** Significant at % confidence level a We only depict willingness-to-pay calculations where the coefficient estimates are significant at a % confidence level or greater. b The coefficient estimate for this variable was not significant and therefore no willingness-to-pay calculation was estimated. Note: HEV=conventional hybrid vehicle (e.g. Toyota Prius); WTP=Willingness-to-pay In contrast, Pioneers greatly preferred PEVs, with PHEVs and BEVs having similarly high values. When considering the electric driving range coefficients (which were significant and positive for the Pioneer sample), this valuation translates into an average willingness-to-pay of $, for a PHEV-, and over $0k for a BEV-0. While these values are perhaps too extreme to be taken literally, they demonstrate that Pioneers are highly enthusiastic for both PHEV and BEV designs. Finally, Pioneers place a much higher value on fuel cost savings relative to

Axsen, Goldberg 0 Mainstream buyers, though it is not clear if this is motivated by financial motivations, or other motivations associated with fuel savings, such as reduced fuel or energy consumption, or reduced environmental impacts.. Preferences for Controlled Charging Table presents a multinomial logit model based on responses to the controlled charging choice experiment. Controlled charging, electric bill, guaranteed minimum charge (GMC) and percentage of renewables coefficients are all significant at a % confidence level and of the expected sign. Respondent willingness-to-pay for controlled charging indicates that, on average, Pioneers and the Early Mainstream would require similar levels of compensation to participate in such a program. However, the two samples differ in two important ways. First, Pioneers value renewable electricity five times more than the Early Mainstream, particularly electricity generated by solar power. Secondly, while both samples value a higher guaranteed charge in the morning (and thus less sacrifice in charge), Pioneers are about twice as sensitive to changes in the level of charge. Results suggest that Pioneers and the Early Mainstream are incentives by different attributes in terms of their participation in a controlled charging program where renewable electricity is relatively influential for Pioneers, and costs savings are relatively influential for the Mainstream. TABLE Controlled Charging Discrete Choice Model Accompanied by WTP Estimates and Model Summary Statistics for Early Mainstream and Pioneer Samples Attributes Alternative specific constants Early Mainstream (Canadian Sample) Pioneer (BC Sample) Utility controlled charging -0.0 *** -0.0 *** Model attributes Monthly electric bill (CAD) -0.0 *** -0.0 *** Guaranteed minimum charge in the morning (%) 0.0 *** 0.0 *** Percentage of green electricity (%) 0.00 *** 0.0 *** Type of green electricity (base = mixed sources ) model attributes Dummy - if wind. -0. *** -0.0 Dummy - if small hydro. -0.0 0.0 Dummy - if solar. -0. ** 0.0 *** Monthly WTP ($CAD) For a % increase in GMC $. $. For a % increase in % of renewables $. $.0 To adopt controlled charging -$ -$ Model Log Likelihood -.00 - Number of Individuals 0 Number of Observations 0 Overall Pseudo R 0. 0. * Significant at 0% confidence level ** Significant at % confidence level *** Significant at % confidence level Note: GMC = Guaranteed Minimum Charge; WTP=Willingness-to-pay

Axsen, Goldberg 0 0. Respondent Motivations As a final comparison, we identify significant differences between Pioneer and Mainstream respondents motivations, as measured in terms of lifestyles, attitudes and values (Table ). Pioneer respondents have significantly higher engagement in both technology- and environment-oriented lifestyles, higher environmental concern, higher levels of traditional values, and lower levels of egoistic values (all at a % confidence level). The Early and Late Mainstream samples are more similar to each other, though the Early Mainstream is more likely to have higher engagement in environment-oriented lifestyles (% confidence level) and technology-oriented lifestyles (% confidence level), and a higher level of environmental concern (% confidence level). Scores for liminality and values are nearly identical between Early and Late Mainstream respondents. TABLE Lifestyles, Attitudes and Values for Pioneers, Early Mainstream and Late Mainstream (British Columbia Samples, Average Scores for Each Question Scale) Sample Pioneer Early Mainstream Late Mainstream Sample size Lifestyle Engagement Technology Orientation (0-) ***.0.. Environmental Orientation (0-) ***... Attitudes Liminality/Openness (- - +) (ns) 0.. 0. Environmental Concern (- - +) ***.0.. Values Traditional Values (0-) ***... Egoist Values (0-) ***... Biospheric Values (0-) *... Altruistic Values (0-) (ns)...0 Note: Differences indicated between BC Pioneers, Early Mainstream and Late Mainstream samples using ANOVA analysis: ns = not significant; * Significant at 0% confidence level; ** % confidence level; *** % confidence level.. DISCUSSION AND CONCLUSION To inform how the future PEV market may differ from the present one, this study provides a comprehensive characterization of Pioneers and potential Early Mainstream buyers, using in-depth survey data collected in British Columbia, Canada. Many of our results are consistent with previous research from the U.S. and Germany, indicating that Pioneers are more likely to have higher income and education (,, ), more likely to be middle-aged, and much more likely to be male (). We also find that Pioneers are more likely to be multi-vehicle households which can allow PEV buyers to adapt to limited range BEVs (). Pioneers are also more likely to have charging access at home, which corresponds with U.S. data (, ), indicating that charging access is unlikely to be a constraint among Early Mainstream buyers. While previous comparisons tend to focus solely on demographic and household characteristics often as proxy measures of consumer preferences or motivations we directly assessed preferences via an in-depth questionnaire that included PEV design exercises and choice models. As a starting point, we find that Mainstream awareness of PEV technology is very low, particularly regarding PHEVs and how they can be refueled which is consistent with early U.S.

Axsen, Goldberg 0 0 0 market research on potential PHEV demand (, 0). Thus, it seems unlikely that most Mainstream have clear preferences for PEVs if they are not yet aware of the basic functions of PEVs (i.e. how they are refueled). Once the idea of a PHEV was explained to respondents in the survey, both the design space and choice experiment results indicate that Early Mainstream respondents are substantially more likely to prefer a PHEV over a BEV design, as has been found in samples drawn from San Diego () and the U.S. (). In contrast, Pioneers tend to prefer BEV designs. We also find that Pioneer preferences for PEVs are more likely to relate to environmental concern, as well as engagement in technology-oriented and environment-oriented lifestyles. These same motivations also exist amongst the Early Mainstream respondents relative to the Late Mainstream, albeit to a lesser degree. These results regarding Eearly Mainstream consumer perceptions and preferences hold important implications for a number of stakeholders. Policymakers seeking to support PEV adoption and automakers seeking to sell PEVs (or to comply with PEV supportive policies) should be aware of the unique qualities of the Early Mainstream. For one, these future buyers are more likely to be interested in PHEVs designs than BEV designs. It therefore seems important for automakers to produce a wide range of PHEVs and for policy makers to implement policies that support both PHEV and BEV designs; for example providing PHEV subsidies or providing credits for PHEVs as part of a Zero-Emissions Vehicle mandate. Second, given the observed confusion among Mainstream respondents regarding how PEVs work, it seems important for future information or marketing campaigns to focus on clarifying the differences between PHEVs and BEVs. Results also suggest that the Early Mainstream and Pioneers may have different patterns and preferences for charging. In terms of general charging patterns, the fact that Mainstream buyers prefer different types of PEVs (notably PHEVs) and tend to have different levels of home charging access indicates that overall time of day recharge (or electricity demand) profiles could differ significantly from those of Pioneers. When specifically looking at enrollment in a controlled charging program, Pioneers place five times more value on supporting renewable energy than Early Mainstream buyers, as has also been found in U.S. research (, ). Thus, electric utilities seeking to recruit future PEV buyers to enroll in controlled charging programs might want to use different incentives for Mainstream buyers, potentially by relying more on cost savings rather than the incentive of recharging with renewable electricity (though the latter might be effective with Pioneers). ACKNOWLEDGMENTS This study was funded by Natural Resources Canada, the Social Sciences and Humanities Research Council of Canada (SSHRC), the Pacific Institute of Climate Solutions (PICS), and the Government of British Columbia. Special thanks to Joseph Bailey for assistance with designing and estimating the discrete choice models, to Paulus Mau for programing the CPEVS survey instruments, and to Zoe Long for assisting with preparing this manuscript. REFERENCES. EV Sales, Accessed June, 0, 0, ev-sales.blogspot.com, 0. Klippenstein, M. Canadian plug-in electric vehicle sales, 0.. Yang, Y. Electric Vehicle Markets Have their Ups and Downs (0 YTD Update), International Council for Clean Transportation, 0.. Axsen, J. and K. S. Kurani. Connecting plug-in vehicles with green electricity through

Axsen, Goldberg 0 0 0 consumer demand, Environmental Research Letters, Vol., No., 0, pp. -.. Tal, G. and M. A. Nicholas. Studying the PEV Market in California: Comparing the PEV, PHEV and Hybrid Markets. Presented at th International Battery, Hybrid and Fuel Cell Electric Vehicle Symposium (EVS), Barcelona, Spain, 0.. Axsen, J., S. Goldberg, H. J. Bailey, G. Kamiya, B. Langman, J. Cairns, M. Wolinetz and A. Miele. Electrifying Vehicles: Insights from the Canadian Plug-in Electric Vehicle Study, Sustainable Transportation Research Team, Simon Fraser University, 0.. Peters, A. and E. Dütschke. How do Consumers Perceive Electric Vehicles? A Comparison of German Consumer Groups, Journal of Environmental Policy and Planning, Vol., No., 0, pp. -.. Rogers, E. Diffusion of Innovations, Free Press, New York, NY, 00.. Axsen, J. and K. S. Kurani. Social influence, consumer behavior, and low-carbon energy transitions, Annual Review of Environment and Resources, Vol., No. 0, pp.. Heffner, R. R., K. S. Kurani and T. S. Turrentine. Symbolism in California's early market for hybrid electric vehicles, Transportation Research Part D-Transport and Environment, Vol., No., 00, pp. -.. Plötz, P., U. Schneider, J. Globisch and E. Dütschke. Who will buy electric vehicles? Identifying early adopters in Germany, Transportation Research Part A: Policy and Practice, Vol., No. 0, pp. -.. Tal, G., M. A. Nicholas, J. Woodjack and D. Scrivano. Who Is Buying Electric Cars in California? Exploring Household and Vehicle Fleet Characteristics of New Plug-in Vehicle Owners, Institute of Transportation StudiesUniversity of California-Davis, 0.. California Center for Sustainable Energy. California PEV Owner Survey, California Center for Sustainable Energy (CCSE), 0.. Pierre, M., C. Jemelin and N. Louvet. Driving an electric vehicle. A sociological analysis on pioneer users., Energy Efficiency, Vol., No., 0, pp. -.. Jensen, A. F., E. Cherchi and S. L. Mabit. On the stability of preferences and attitudes before and after experiencing an electric vehicle, Transportation Research Part D: Transport and Environment, Vol., No. 0, pp. -.. Axsen, J. and K. S. Kurani. Early U.S. market for plug-in hybrid electric vehicles: Anticipating consumer recharge potential and design priorities., Transportation Research Record: Journal of the Transportation Research Board, Vol., No. 00, pp. -.. US Department of Energy. EV Project Electric Vehicle Charging Infrastructure Summary Report, US Department of Energy, 0.. Axsen, J. and K. S. Kurani. Who can recharge a plug-in electric vehicle at home?, Transportation Research Part D: Transport and Environment, Vol., No., 0, pp. -.. Axsen, J. and K. S. Kurani. The Early U.S. Market for PHEVs: Anticipating Consumer Awareness, Recharge Potential, Design Priorities and Energy Impacts, UCD-ITS-RR-0-, University of California, 00. 0. Caperello, N. and K. S. Kurani. Households' stories of their encounters with a plug-in hybrid electric vehicle, Environment and Behavior, Vol., No., 0, pp. -0.. Carley, S., R. M. Krause, B. W. Lane and J. D. Graham. Intent to purchase a plug-in electric vehicle: A survey of early impressions in large US cites, Transportation Research Part D: Transport and Environment, Vol., No. 0, pp. -.. Axsen, J. and K. S. Kurani. Hybrid, plug-in hybrid or electric: What do car buyers want?, Energy Policy, Vol., No. 0, pp. -.

Axsen, Goldberg 0 0 0. Turrentine, T., D. Garas, A. Lentz and J. Woodjack. The UC Davis MINI E Consumer Study, Institute of Transportation StudiesUniversity of California, Davis, 0.. Parsons, G. R., M. K. Hidrue, W. Kempton and M. P. Gardner. Willingness to pay for vehicle-to-grid (VG) electric vehicles and their contract terms, Energy Economics, Vol., No. 0, pp. -.. Egbue, O. and S. Long. Barriers to widespread adoption of electric vehicles: An analysis of consumer attitudes and perceptions, Energy Policy, Vol., No. 0, 0, pp. -.. Krupa, J. S., D. M. Rizzo, M. J. Eppstein, D. B. Lanute, D. E. Gaalema, K. Lakkaraju and C. E. Warrender. Analysis of a consumer survey on plug-in hybrid electric vehicles, Transportation Research Part A: Policy and Practice, Vol., No. 0, pp. -.. Lee-Gosselin, M. The dynamics of car use patterns under different scenarios: A gaming approach, Developments in dynamic and activity-based approaches to travel analysis, Vol. No. 0, pp. 0-.. Kurani, K. S., T. Turrentine and D. Sperling. Testing electric vehicle demand in 'hybrid households' using a reflexive survey, Transportation Research Part D, Vol., No.,, pp. -0.. Axsen, J. and K. S. Kurani. Early U.S. market for plug-in hybrid electric vehicles: Anticipating consumer recharge potential and design priorities, Transportation Research Record: Journal of the Transportation Research Board, Vol., No. 00, pp. -. 0. Bunch, D. S., M. Bradley, T. F. Golob, R. Kitamura and G. P. Occhiuzzo. Demand for clean-fuel vehicles in California: A discrete-choice stated preference pilot project, Transportation Research Part a-policy and Practice, Vol., No.,, pp. -.. Hidrue, M. K., G. R. Parsons, W. Kempton and M. P. Gardner. Willingness to pay for electric vehicles and their attributes, Resource and Energy Economics, Vol., No., 0, pp. -0.. Potoglou, D. and P. S. Kanaroglou. Household demand and willingness to pay for clean vehicles, Transportation Research Part D-Transport and Environment, Vol., No., 00, pp. -.. Axsen, J., J. TyreeHageman and A. Lentz. Lifestyle practices and pro-environmental technology, Ecological Economics, Vol., No. 0, 0, pp. -.. Cordano, M., S. Welcomer and R. Scherer. An analysis of the predictive validity of the New Ecological Paradigm, The Journal of Environmental Education, Vol., No., 00, pp. -.. Stern, P. C., T. Dietz and G. A. Guagnano. A Brief Inventory of Values, Educational and Psychological Measurement, Vol., No.,, pp. -0.. Axsen, J. and K. S. Kurani. Anticipating PHEV energy impacts in California: Constructing consumer-informed recharge profiles, Transportation Research Part D: Transport and Environment, Vol., No., 0, pp. -.. Inc., H. D. What does 0 hold for the auto industry in Canada?, 0.. Busse, M. R., C. R. Knittel and F. Zettelmeyer. Are Consumer Myopic? Evidence from New and Used Car Purchases, The American Economic Review, Vol., No., 0, pp. 0-.