Driving Electric with Equity: A Survey of Current Electric Vehicle Infrastructure, Policies and Planning

Thesis (Ph.D.)--University of Washington, 2022Electric vehicles (EV) and the electrification of the transportation sector is critical to reduce carbon emission. Governments are implementing large scales of public policies, regulations, incentives, and disincentives to build more infrastructure to promote vehicle electrification. However, the equity impacts of these policies, especially the non-financial, planning-related policies, are rarely discussed and evaluated in existing academic literature and professional reports. This dissertation aims to explore different quantitative methods to evaluate the equity impact of EV policies. The dissertation is based comprised of three parts: (1) the theoretical and philosophical contexts of equity and justice in planning, transportation, and vehicular electrification, and a discussion of existing EV policies around the world; (2) three sets of research that evaluate the equity impact of EV policies from different angles; and (3) the conclusions. The first study is to explore how different socioeconomic groups weigh the importance of three non-financial incentives, namely charging infrastructure, free parking, and HOV lane access, and how they influence the EV purchase decision-making in California. An Ordered Logistic Regression and its time-modifying effects between 2017 and 2019 were applied with the data, the California Vehicle Survey from California Energy Commission. Both logistic regression model and the time-varying model suggest that the EV buyers from the marginalized group, including the low-income and non-single family homeowner groups, consider public charging availability the most crucial incentive to their EV purchase, while the more advantaged drivers find HOV lane access to be the most critical. The second study aims to estimate the carbon reduction amount of each driver, and to forecast the carbon mitigation cost, as well as the transition cost to EV by socioeconomic groups under the scenario of a full transition to EV. We create the two scenarios using the assumptions that consumers switch to EV immediately with a new Tesla Model 3 and a used Chevrolet Bolt. Multiple regression tree and boosted regression tree algorithms were applied. The 2019 California Vehicle Survey is used as the database to provide existing driving behavior as input to the forecasting model. The result finds that a greater variations of carbon emission amount among racial groups, housing types, income levels, employment status followed by education achievement. It suggests that marginalized groups show higher quantity carbon emissions than the higher-income and white-only groups. The number of vehicles per household and annual household income are the two most important variables to impact the potential carbon mitigation cost for both scenarios, with education, race, and housing types less relevant. The lower-income drivers face a higher financial cost to switch to EV. The third study aims to investigate the spatial association between current electric vehicle supply equipment (EVSE) availability and socioeconomic status of different areas within metropolitan areas. We choose six metropolitan areas in the U.S.: Los Angeles-Long Beach-Anaheim, San Francisco-Oakland-Berkeley, Portland-Vancouver-Hillsboro, Seattle-Tacoma-Bellevue, Detroit-Warren-Dearborn, and Atlanta-Sandy Springs-Alpharetta. The EVSE location data are from the Alternative Fuels Data Center, and the demographic and socioeconomic data are from the 2020 Census and the 2019 American Community Survey. After testing several geospatial statistics methods, we found that Geographically Weighted Regression (GWR) models offer better performance than Ordinary Least Square models. Results show heterogeneity among cities and neighborhoods. In general, we found that current EVSE density distributions are spatially correlated with certain socioeconomic variables in various cities: housing density (except San Francisco), poverty (except Los Angeles and San Francisco), and employment rate (except San Francisco and Detroit) are mainly negatively associated with EVSE densities, while rental rate (except San Francisco) and median household incomes (except Portland and Atlanta) are positively associated. White-only variable shows no significant relationship in all six metropolitan areas. Overall, availability and accessibility of EVSE in marginalized communities needs to be improved. These studies and the methodologies could shed light for EV policymakers to better understand the equity implications of vehicular electrification, to equip with different quantitative methods to evaluate the policies present, and to improve the policies to achieve equity and climate goals in the future