JB asked in the comments to an earlier post what we know about who bears the burden of cigarette taxation. A brief exploration of the economic literature reveals the following:
(1) The Heterogeneous Geographic and Socioeconomic Incidence of Cigarette Taxes: Evidence from Nielsen Homescan Data
Author: Harding, Matthew; Leibtag, Ephraim; Lovenheim, Michael F.
Author Affiliation: Stanford U; USDA; Cornell U
Source: American Economic Journal: Economic Policy, November 2012, v. 4, iss. 4, pp. 169-98
Publication Date: November 2012
Abstract: We use Nielsen Homescan data to examine who bears the economic burden of cigarette taxes. We find cigarette taxes are less than fully passed through to consumer prices, suggesting consumers and producers split the excess burden of these taxes. Using information on consumer location, we show the availability of lower-tax goods across state borders creates significant differences in the pass-through rate. Tax avoidance opportunities also have a sizable effect on purchasing behavior by altering consumer search, prices paid and quantities purchased. Finally, we demonstrate that the incidence of cigarette taxes and the border effect varies by household income and education.
WINTERCOW: here is a link for those who can get access. On p. 192 we find that the poor don’t actually experience as serious an impact of the tax burden as you would think, because although the poor smoke more, they also search more for alternatives (roll your own, out of state purchases, etc.) than wealthier people do. You can see Table 8 in the paper for specific information on how much they actually bear.
(2) Consumer Response to Cigarette Excise Tax Changes
Author: Chiou, Lesley; Muehlegger, Eric
Author Affiliation: Occidental College; Harvard University
Publisher Information: Harvard University, John F. Kennedy School of Government, Working Paper Series
Publication Date: 2010
Abstract: We use a rich dataset of weekly cigarette sales to examine how consumers adapt their behavior before and after excise taxincreases–whether by reducing demand, stockpiling, traveling to low-tax jurisdictions, or substituting towards lower-cost brands. Consumer response varies substantially for different types of cigarettes. Stockpiling primarily occurs for discount cigarettes and is most pronounced at stores far from lower-tax jurisdictions. Border-crossing is greatest at stores close to low-tax jurisdictions and occurs primarily for cigarettessold by the carton. Finally, we find modest short-run substitution towards lower-cost brands following a tax-increase, consistent with consumers smoothing the transition to higher cigarette taxes. These differences in consumer behavior lead to meaningful differences intax incidence–pass-through is higher for discount cigarettes which have more inelastic demand. Pass-through is lower near low-taxborders, especially for cigarettes sold by the carton for which cross-border evasion is greatest.
WINTERCOW: here is a copy of the paper. Table 4 has some relevant data. It does not appear that stockpiling behavior varies by income class.
(3) Who Pays Cigarette Taxes? The Impact of Consumer Price Search
Author: DeCicca, Philip; Kenkel, Donald S.; Liu, Feng
Author Affiliation: Unlisted; Unlisted; Unlisted
Publisher Information: National Bureau of Economic Research, Inc, NBER Working Papers: 15942
Publication Date: 2010
Abstract: We conduct an empirical study of the impact of consumer price-search on the shifting of cigarette excise taxes to consumer prices. We use novel data on the prices smokers report actually paying for cigarettes. We document substantial price dispersion. We find that cigarette taxes are shifted at lower rates to the prices paid by consumers who undertake more price search–carton buyers, and especially, smokers who buy cartons of cigarettes in a state other than their state of residence. We also find suggestive evidence thattaxes are shifted at slightly higher rates to the prices paid by non-daily smokers, less addicted smokers, and smokers of light cigarettes.
Publication Type: Working Paper
WINTERCOW: It appears that we don’t have a lot of direct research on just how much of the burden is passed on to the poor. Here is a key extract from the end of the paper: ”
Another direction for future work is to explore the implications of price search
for the incidence of cigarette taxes across income groups. Standard analyses conclude
that because smokers have lower incomes in the U.S., cigarette taxes are regressive
(Lyon and Schwab 1995, Colman and Remler 2008). These standard analyses are
incomplete if taxes are differentially passed through to the prices paid by consumers
with different incomes. Indeed, Heiding, Leibtag and Lovenheim (2010) find that
cigarette taxes are fully shifted to the prices middle-income consumers pay, but less
than fully shifted to the prices paid by both low- and high-income consumers. Future
work could further explore the role of consumer price search in these patterns of tax
incidence and income.
(4) Tax Incidence When Individuals Are Time-Inconsistent: The Case of Cigarette Excise Taxes
Author: Gruber, Jonathan; Koszegi, Botond
Author Affiliation: MIT; U CA, Berkeley
Source: Journal of Public Economics, August 2004, v. 88, iss. 9-10, pp. 1959-87
Publication Date: August 2004
Abstract: One of the most cogent criticisms of excise taxes is their regressivity, with lower income groups spending a much larger share of their income on goods such as cigarettes than do higher income groups. We argue that traditional quantity-based measures of incidenceare only appropriate under a very restrictive “time-consistent” model of consumption of sin goods. A model that is much more consistent with existing evidence on smoking decisions is a time-inconsistent formulation where excise taxes on cigarettes serve a self-control function that is valued by smokers who would like to quit but cannot. This self-control function benefits lower income groups more, since they have a significantly higher price sensitivity of smoking. Calibrations show that, as a result, cigarette taxes are much less regressive than previously assumed, and are even progressive for a wide variety of parameter values.
WINTERCOW: So we have suggestions that cigarette taxes are possibly progessive. One important reason for this is that lower income consumers are much more sensitive to price changes than higher income ones and it is not just in search for lower priced cigarettes but also their propensity to quit.
(5) Consumption Taxes in a Life-Cycle Framework: Are Sin Taxes Regressive?
Author: Lyon, Andrew B.; Schwab, Robert M.
Author Affiliation: U MD; U MD
Source: Review of Economics and Statistics, August 1995, v. 77, iss. 3, pp. 389-406
Publication Date: August 1995
Abstract: We construct measures of tax incidence over the life-cycle and compare these measures to traditional measures based on annual data. Annual measures of the incidence of taxes on consumption goods may differ from life-cycle measures for three reasons. First, annual measures of income reflect transitory components which should have smaller effects on consumption than permanent changes in income. Second, income measured in a single period differs from lifetime income due to age-related differences in earnings. Third, consumption of certain items follows life-cycle patterns independent of changes in income. Surprisingly, we find that these effects cause little change in the assessment of the incidence of taxes on cigarettes. For alcohol, we find that a tax on its consumption is slightly less regressive when measured with respect to lifetime income than when measured with respect to annual income.
Publication Type: Journal Article
WINTERCOW: On p. 398 and 399 of the paper you will find their evidence that cigarette taxes, over the 1968-72 period, appear to be regressive.
There are a few other scattered papers that tangentially address the question. But I would say based on the limited evidence that the jury is out on the actual incidence.