摘要：The proposition of a carbon tax has reinvigorated the discussion about the price elasticity of gasoline demand. This paper analyzes how consumers react to higher gas prices in a new setting, examining whether they choose Internet purchasing over in-store purchasing when gas prices increase. Prior studies of Internet purchases have focused mainly on local sales tax rates, traditional retail options, and prices. This is the first study to incorporate transportation costs directly, studying 155,000 households across the U.S. from 2008 through 2009. The impact of gas prices on individuals' transportation budget depends on the fuel efficiency of their vehicles, on the distance that they need to drive, and on the availability of alternate public transportation options. I take all those factors into account. The tax avoidance argument for online shopping implies that individuals who live in locations with higher sales taxes are more likely to purchase online. I analyze the latter at the zipcode level. I also explore how shopping opportunities impact online purchases, taking the number of brick-and-mortar stores in a particular zipcode into account. Another motivation for online shopping relates to consumers' time value. I study whether individuals with less free time due to working full time and time commitments such as small children purchase more online. I also analyze how individual's demographic characteristics, such as age, gender, and education, influence online shopping. I control for holidays on which there are sales and for holidays on which presents are exchanged. When gas prices are high, there is an income effect that leads to less purchases overall. I take this into account by computing the cost of commuting to work, which depends on the distance to the workplace, the retail gas price, and the fuel efficiency of the vehicle. I also include an interaction effect between income and gas prices. In addition, I use a dummy variable indicating whether respondents answered that one of the reason not to drive was ``high gas prices''. I find that individual demographic characteristics matter most as determinants of online shopping. People who purchase online more frequently tend to be female, work full time, and have small children. The availability of traditional shopping opportunities reduces the frequency of online purchases slightly, while there is no support for the tax avoidance hypothesis. Regarding the variable of most interest, I do not find a significant effect for the cost of filling the gasoline tank of an individual's vehicle on online purchases. Individuals whose work commute is costlier purchase less online, by way of an income effect. These results support prior studies about relatively inelastic gasoline demand.
