1、标题： Giving Green to Get Green: Incentives and Consumer Adoption of Hybrid Vehicle Technology 原文： Abstract： Federal, state and local governments use a variety of incentives to induce consumer adoption of hybrid-electric vehicles. We study the relative efficacy of state sales tax waivers, income tax c
2、redits and non-tax incentives and find that the type of tax incentive offered is as important as the value of the tax incentive. Conditional on value, we find that sales tax waivers are associated a seven-fold greater increase in hybrid sales than income tax credits. In addition, we estimate the ext
3、ent to which consumer adoption of hybrid-electric vehicles (HEV) in the United States from 2000-2006 can be attributed to government incentives, changing gasoline prices, or consumer preferences for environmental quality or energy security. After controlling for model specific state and time trends,
4、 we find that rising gasoline prices are associated with higher hybrid sales, although the effect operates entirely through sales of the hybrid models with the highest fuel economy. In total, we find that tax incentives, rising gasoline prices and social preferences are associated with 6, 27 and 36
5、percent of high economy hybrid sales from 2000-2006. Data and Methodology We use quarterly state-level data on sales of each hybrid model to distinguish the impact of three factors which potentially speed consumer adoption of hybrid technology: (1) government incentives, including single occupancy H
6、OV lane access and income or sales tax incentives (2) changes in gasoline prices (3) changes in consumer travel intensity or preferences for environmentalism or national security. Our study period runs from Q1 2000 to Q4 2006 during which automakers introduced eleven hybrid models. The date at which
7、 domestic sales first appear in the JD Power data are: Honda Insight (2000Q1), Toyota Prius (2000Q3), Honda Civic (2002Q1), Honda Accord (2004Q4), Ford Escape (2004Q3), Lexus RX400h (2005Q1), Toyota Highlander (2005Q2), Mercury Mariner (2005Q3), Lexus GS450h (2006Q2), Toyota Camry (2006Q2) and Satur
8、n VUE (2006Q4). To allow for each model to have its own time trend and different attractiveness to each states population, we include both state*model and time*model fixed effects, and estimate the model using within-state*model variation in trend of sales. We purchased our data on hybrid sales from
9、 JD Power and Associates. The sales data is based on their proprietary Power Information Network data. The Power Information Network (PIN) collects real-time transaction level data from approximately 6,000 dealers. For our analysis, JD Power and Associates aggregated consumer purchases (as opposed t
10、o fleet purchases) in the PIN transaction data up to quarterly quantities for each model at the state level. Based on actual purchases, the transaction data provides more accurate timing information than vehicle registration data. For our analysis, JD Power limited the sample to consumer purchases.
11、This excludes private and government fleet purchases which at best are unrelated to consumer preferences and at worst are correlated with state government incentives. The exclusion of these sales ensures that we will not misattribute government purchases to state consumer incentives. We classify mod
12、els as high fuel-economy hybrids if the EPA-rated fuel economy is 50 percent greater than other vehicles in its class. Using this criterion, we classify the Honda Insight and Civic, Toyota Prius and Camry and the Ford Escape as high fuel-economy vehicles. Figure 1 graphs quarterly domestic sales of
13、high and low fuel-economy hybrids as well as the average quarterly retail price of gasoline reported by the Energy Information Administration (EIA) from Q1 2000 until Q4 2006. Honda and Toyota introduced high-economy hybrids first, beginning with the Insight, Prius and Civic. Due to earlier introduc
14、tion dates, sales of high-economy hybrid models account for the vast majority of total hybrid sales from Q1 2000 to Q4 2006.Over 2005 and 2006, though, sales of lower efficiency models increase substantially accounting for 28 percent of the 182,000 hybrid sales in 2005 and 25 percent of the 235,000
15、hybrid sales in 2006. Interestingly, sales of all models fall in the final quarter of 2005. Although only anecdotal, this may be the result of the change in federal tax incentives on January Consumers may have chosen to delay hybrid purchases in Q4 of 2005 in anticipation of the more generous federa
16、l tax credits. Our dataset of state and federal incentives programs were collected by contacting regulators in each jurisdiction. Although a large number of incentive programs exist, we focus specifically on incentives related to high-occupancy vehicle (HOV) lane access and tax incentives. We omit l
17、ocal incentive programs such as parking fee waivers and state vehicle registration incentives since both are likely of insufficient magnitude to affect state-level sales of hybrid vehicles substantially. We construct quarterly average tax-inclusive retail gasoline price for each state, based on mont
18、hly gasoline price data from the Energy Information Administration. We also include state-level demographic controls from the Current Population Survey. We use the percent of residents with a high school or four-year college diploma, per capita income, mean age and proportion female. Finally, we con
19、struct proxies capturing consumer travel behavior and preferences for environmental quality and energy security. As a proxy for travel intensity, we use annual per-capita vehicle-miles traveled in each state. We also use average commuting time as an alternative the choice of proxy for travel intensi
20、ty does not change our empirical conclusions. We use state-level per-capita Sierra Club membership as a proxy for the proportion of a states population with a high preference for environmental activism. As proxies for the salience of global warming and anti-war sentiment, we use quarterly deviation
21、from 100-year average temperatures, per-capita armed forces participation, and per-capita casualties in Iraq and Afghanistan. Table 2 presents the summary statistics for the hybrid sales data, gasoline prices, state and federal incentives, and socioeconomic measures. We use the unbalanced panel to g
22、enerate the summary statistics for hybrid sales and state incentives. For demographics and gasoline prices, we treat each state-quarter as a single observation and report the summary statistics for the balanced panel. Conditional on positive sales, mean model sales in a particular state-quarter are
23、121.1, with a high of 8,871 for Prius sales in California in Q3 2006. Approximately 94 percent of hybrid sales over the study period are eligible for a federal tax incentive with a mean value of $1,073. Twelve percent of hybrid sales are eligible for either a state income tax or sales tax incentive,
24、 with average values of $2,011 and $1,037 dollars respectively. During the sample period, mean tax-inclusive retail gasoline is $1.80 gallon with a low of $0.994 a gallon (Georgia, Q4-2001) and a high of $3.10 a gallon (Hawaii, Q3-2006). In the empirical section, we estimate the effect of government
25、 hybrid incentives, retail gasoline prices, and social preferences on hybrid sales. We first estimate and compare the effects of government incentives designed to stimulate purchases and market incentives created by changed gasoline prices. We evaluate a number of testable hypotheses related to relative sales of high-economy and low-economy hybrids. In particular, we would expect rising gasoline prices to have an especially
