Asymmetric Incidence of Sales Taxes
One sentence summary: Both sales-tax increases and decreases are under-shifted to consumer
prices, although the under-shifting of sales-tax decreases is much
higher, suggesting evidence for asymmetric incidence of sales taxes.
The corresponding paper by HakanYilmazkuday titled "Asymmetric Incidence of Sales Taxes: A Short-Run Investigation of Gasoline Prices" has been accepted for publication at Journal of Economics and Business.
Abstract
This paper investigates the shifting of sales taxes to consumers through retail prices in the short run. Retail data on gasoline prices are used at the station level within the U.S., including observations from all fifty states and the District of Columbia. A difference-in-differences approach is employed to identify the short-run effects of the changes in state taxes as of January 1st, 2015, when five states have increased their gasoline sales taxes, while five other states have decreased theirs. States experiencing such changes in sales taxes (between December 31st, 2014 and January 1st, 2015) are analyzed as the treatment group of a natural policy experiment, where the control group consists of states with no changes in their sales taxes. The results show that both sales-tax increases and decreases are under-shifted to consumer prices, although the under-shifting of sales-tax decreases is much higher (i.e., the asymmetric incidence of sales taxes). The pass-through measures also differ significantly across states, showing the importance of having a nationwide analysis. The results are robust to the consideration of retailer characteristics, wholesale prices, retail brand effects and hourly price changes within each day.
Non-technical Summary
The incidence of gasoline sales taxes is a fundamental concept in public economics, because it determines how economic welfare is distributed between gas stations and consumers due to changes in taxes. Since gasoline accounts for about 5% of consumer spending and sales taxes are determined by policy makers, the measurement of the incidence is an essential concern of politicians as well. However, there are only a few studies that have attempted to measure the effects of gasoline sales taxes at the station (i.e., retail-firm) level. Having an investigation at the retail level is especially important for the gasoline market, because each gas station can pass the effects of taxes on to consumers differently (by over-shifting or under-shifting taxes to consumer prices), which leads a distribution of tax incidence among gas stations and thus a redistribution of economic welfare even within consumers purchasing gasoline from different stations or among stations located in the same political district.
This paper achieves such an investigation at the gas-station level in the short run. Using retail prices of regular gasoline obtained from gas stations within the U.S., including observations from all fifty states and the District of Columbia, we investigate the effects of state-level sales tax changes (on retail prices) that have become effective on January 1st, 2015, when five states have increased their sales taxes, while five others have reduced theirs.
Accordingly, these ten states experiencing changes in their sales taxes (between December 31st, 2014 and January 1st, 2015) are analyzed as the treatment group of a natural policy experiment, where the control group consists of states with no changes in their sales taxes. Since all sales tax changes are due to earlier state laws (rather than market conditions), using a difference-in-differences approach is a compelling way to study the effects of tax changes on retail prices, and it is robust to any identification/endogeneity problem. Within this context, the main assumption is that the retailers would re-optimize their pricing decision according to changes in tax rates, since they already know the timing of such changes.
The results of a difference-in-differences approach show evidence for asymmetric incidence of sales taxes. In particular, although both sales-tax increases and decreases are under-shifted to consumer prices, the under-shifting of sales-tax decreases is much higher. In the case of sales-tax increases, the under-shifting corresponds to an increase in gasoline prices less than the increase in sales taxes due to the estimated pass-through coefficients less than one, while in the case of sales-tax decreases, it corresponds to an increase in prices despite the decrease in sales taxes due to the estimated pass-through coefficients less than zero. The latter result is interesting in an environment of gasoline prices decreasing nationwide, because it implies that retailers have either kept their prices constant or have reduced their prices less than the national average, which has potentially resulted in higher rates of return on capital (in a perfectly competitive market) or higher markups (in an environment with imperfect competition). When we further investigate the retailers facing tax reductions, we in fact observe that the average retailer have reduced its price by only 0.1% after the average tax reduction of 0.85%, while the retailers in the control group (with no tax changes) have experienced an average price reduction of 0.71%. Finally, the short-run pass-through measures differ significantly across states, showing the importance of having a nationwide analysis.