Taxing the Poor

Studies | Taxes

No. 300
Friday, June 22, 2007
by the National Center for Policy Analysis Task Force on Taxing the Poor


Section II: Excise Taxes and Alcohol

Introduction

Alcohol taxes, like tobacco taxes, are designed to raise revenue by taxing a behavior that society has deemed "sinful."  Also like tobacco taxes, alcohol taxes have become a popular, convenient way for state legislators to raise money without resorting to explicit increases in income or sales taxes.  Unlike expenditures on tobacco products, alcohol expenditures rise with income.  Even so, alcohol taxes are still regressive, since they constitute a larger percentage of income for lower earners than higher earners. 

Alcohol Taxes

“Alcohol taxes are a popular, convenient way to fill government
coffers.”

Alcohol taxes have been a feature of the U.S. tax system since the country's founding.  In fact, the first tax implemented under the new U.S. Constitution in 1789 was a levy on imported spirits.  Congress followed with a tax on domestically produced alcohol in 1791.  Since then, alcohol taxes have impacted everything from government bureaucracy to stock car racing. 1

Federal treatment of alcohol has varied widely over time, from repeal of the federal excise tax on liquor, due in large part to the Whiskey Rebellion of 1794, to outright prohibition of alcohol by 1919, and still later to repeal of Prohibition.  For much of the last half of the 20th century, the federal government largely ignored alcohol as a source of new revenue, and between 1951 and 1990, taxes on alcohol remained mostly unchanged.  However, in recent years, federal and state governments have turned to excise taxes to fill government coffers:

  • The 1990 Budget Act, which raised taxes on a number of so-called "luxury" goods to increase federal revenues, doubled tax rates on beer and wine. 
  • Legislative efforts at the state level have accelerated over the past several years, although they have been largely unsuccessful. 
  • In 2005 alone, 13 states proposed increases in alcohol taxes and fees. 2

Variations in Alcohol Taxes .  Today, taxes are much lower on beer and wine than on distilled spirits, and they are calculated in different ways.  At the federal level: 3

  • Beer, measured by the barrel, is currently taxed at $18 per barrel or about 10 cents per ounce of alcohol (assuming an alcohol content of 4.5 percent); however, some microbrews are taxed at only $7 per barrel.
  • Wine is taxed at $1.07 per gallon, or about 8 cents per ounce of alcohol (assuming an average alcohol content of 11 percent).
  • Distilled spirits, measured in proof gallons, are taxed at $13.50 per proof gallon, or about 21 cents per ounce of alcohol. 4

“Taxes on alcohol vary widely by state and follow no coherent pattern.”

In addition, each state levies its own excise taxes on alcohol based on its own unit of measure - including gallons, barrels, fluid ounces and percent of retail price.  Some states also charge varying rates for different concentrations of alcohol.  This significantly complicates calculation and comparison of alcohol taxes between states.  To aid in this effort, Table II-1 presents the varying federal and state charges converted to a comparable index called "tax per drink."  A drink equals a 12-ounce beer with less than 5 percent alcohol, a 5-ounce glass of wine with less than 14 percent alcohol, or 1-ounce of 100-proof alcohol.  The averages indicate that tax rates on spirits are greater than the tax rates on wine and beer; however, averages can be deceiving.  The real story is how tax rates on alcohol - like taxation in general - follow no coherent pattern, with widely different rates charged on the same basic activity.

  • Nationally, the average tax is 2.3 cents for beer, 2.9 cents for wine and 2.9 cents for spirits.
  • States' taxes per drink range from a low of 0.002 cents on beer in Wyoming to a high of 8.8 cents on wine in Florida.
  • The federal government's tax, which is added to the state rate, is 5.3 cents for beer, 4.2 cents for wine and 10.6 cents for spirits.

“State taxes per drink range from 0.002 cents on beer in Wyoming to 8.8 cents on wine in Florida.”

Further, regulations in some states impose an even larger implicit tax on alcohol.  For example, in some states certain types of alcohol can only be purchased from a state-sponsored or state-operated store, which imposes a large cost in convenience as well as price.  These state regulations also complicate cross-state comparisons of alcohol tax rates. 

Alcohol Taxes from the Consumer's Perspective .  Perhaps a more meaningful way to examine alcohol taxes is from the consumer's perspective.  Table II-2 expresses excise taxes by the unit with which consumers might be most familiar: a six-pack of beer (72 ounces), a bottle of wine and a fifth of spirits (the latter two of which are each 750 milliliters or 25.36 ounces).  Again, the tax levies follow no coherent pattern from state to state:

  • Nationally, the average tax is 14 cents for a six-pack, 15 cents for a bottle of wine and 73 cents for a fifth of spirits. 
  • States' taxes per unit of purchase range from a low of 1 cent on beer in Wyoming to $1.29 on spirits in Florida.
  • Federal taxes amount to 32 cents for beer, 21 cents for wine and a whopping $2.69 for spirits.

In all, the federal government collected about $8.9 billion in alcohol taxes last year, including $3.6 billion on beer, $800 million on wine and $4.4 billion on spirits. 5   States collectively raised almost $5 billion. 6

Who Drinks What?   Given the variation in the way different types of alcoholic beverages are taxed, it is important to explore the characteristics of those who buy different types of alcohol. 7   Beer is the type most often purchased in America today.  Although higher-income households buy beer more often than lower-income households, beer makes up a much greater portion of alcohol purchases among lower earners than higher earners:

  • Overall, 29 percent of American households purchase alcohol; beer makes up more than half (52.5 percent) of purchases, wine is 28.1 percent and spirits are 19.3 percent.
  • Only 13.1 percent of the lowest-earning households purchase alcohol; of their total alcohol purchases, 60.2 percent is beer, 25 percent is wine and about 15 percent is spirits.
  • Forty-three percent of the highest-earning households purchase alcohol; among them, beer and wine purchases are relatively equal (41.7 percent and 38 percent, respectively), while spirits constitute 19.3 percent. 8

Alcohol Taxes and the Price of Alcoholic Beverages

In most locales, taxes are levied directly on the producers of alcoholic beverages.  The question is whether these alcohol taxes, like tobacco taxes, are passed on to consumers in the form of higher prices. 

The standard economic textbook model assumes a relatively simple good sold in a reasonably competitive market.  In such a market a tax increase will, in the short run, cause the price to rise by less than the full amount of the tax.  As the market adjusts over time, more of the tax is reflected in the price; eventually, the price increase will be fairly similar to the tax increase. 

But if the market is not a typical competitive environment, the impact of a tax increase is much less predictable.  For example, if the market is controlled by a monopoly (a market with a single seller), a tax increase will, after a period of adjustment, cause the price of the good to increase by more than the tax.  In an oligopolistic market (a market with relatively few sellers) a tax increase can also lead to a price increase larger than the tax.    

A tax increase may also have an "echo" effect:  If sales drop in response to the initial price increase, the seller's unit costs go up and may lead to further price increases.  It is also possible that excise tax increases will cause firms to change the types of products sold, substituting higher-quality or higher-priced goods.  This last effect is especially likely for alcohol taxes, which are based on volume and not value, since the tax represents a lower portion of the total cost for higher-valued goods.

“When alcohol taxes increase, consumer prices rise by more than the amount of the tax increase.”

So, are alcohol taxes passed through to the consumer?  In an extensive study of excise taxes and prices, researchers Douglas Young and Agnieszka Bielinska-Kwapisz found a surprising degree of "over-shifting."  That is, prices rose significantly more than the amount of the tax, and the change occurred within three months of the tax increase: 9

  • Wine prices increased 1.2 times the amount of any tax increase in states without liquor monopolies and 2.1 times the tax increase in states with a liquor monopoly.
  • Beer prices increased 1.7 times the amount of any tax increase.
  • Spirit prices increased 1.6 times the amount of any tax increase in all states, with no difference in states with liquor monopolies. 

A more recent examination of price and tax data in Alaska by Donald Kenkel confirmed this price over-shifting: 10

  • In nearly all markets and for all alcohol products, the rate of over-shifting was greater than the tax increase.
  • In both on-premise (bars) and off-premise (liquor store) beer sales, and for many different brands of beer, the price increase was around double the tax increase.
  • For on-premise sales of wine and spirits, the price increase was almost four times the tax increase. 

Granted, the study is of a specific market, but it allows an examination of the price-tax relationship without concern that differences between markets may distort the true picture. 

The Regressive Nature of Alcohol Taxes

The percentage of people purchasing alcohol - as well as the average dollar amount spent - increases with income, according to the Consumer Expenditure Survey, which presents data from 2004 in quintiles of income.  However, although lower earners spend far fewer dollars than higher earners on alcohol products, the amount they spend constitutes a much greater percentage of their income. 11

  • Twenty-three percent of the lowest-earning quintile (average income $9,169) purchased alcohol, compared to 31 percent of the middle quintile (average income $34,647) and 43 percent of the highest earners (average income $112,158). 
  • The highest-earning quintile spent about $876 on alcoholic beverages, on the average, compared to only $194 for the lowest earners.
  • The bottom quintile of income earners spent 2.1 percent of income on alcohol products, on the average, twice the middle quintile and more than three times the highest earners.

Obviously, however, not everybody in an income quintile drinks alcoholic beverages.  Examining the expenditures of actual drinkers provides a more meaningful comparison of the effects of alcohol taxes on different income groups.  Data from the 2003 Consumer Expenditure Survey Anthology provides such a comparison. 12

  • Lower-income earners who actually purchase alcohol spent $1,158 per year. [See Table II-3.]
  • The highest-income earners who actually purchase alcohol spent $1,583 per year.
  • The lowest earners spent some 13 percent of total income on alcohol products, compared to 1.3 percent for the highest earners.

Table 2-3

“Lower earners spend a much greater percentage of their income on alcohol than higher earners.”

Of course, it is possible that alcohol taxes - particularly on spirits - are more regressive than reported in this paper.  Taxation of spirits is a function of alcohol content; the federal government, for example, charges taxes based on proof gallons, not as a percentage of price.  Suppose the rich and poor spend the same proportion of income on alcohol, but the wealthy buy higher-quality products or more-expensive brands.  In this instance, the tax could be considered even more regressive than presumed since the affluent get more for their money.

“The lowest earners spent almost 13 percent of total income on alcohol, compared to 1.3 percent for the highest earners.”

Some researchers have concluded that while alcohol taxes are regressive with respect to current household income, they are less regressive over time for an important reason:  incomes change.

  • First, income is subject to short-term fluctuations; a higher-income household might be classified as low-income if the primary wage-earner was transitioning to a new job or temporarily unemployed.
  • Second, income varies over a lifetime; a household of two married medical students would be classified as low-income even though their incomes will most definitely rise in future years.
  • Third, expenditures vary over a lifetime, leading to an incorrect conclusion about the lifetime burden of certain taxes. 

It has been established that alcohol taxes are regressive in a given year.  But how does the regressivity change over longer periods?  Perhaps the most thorough examination of the burden of alcohol taxes over a longer period was developed by Andrew Lyon and Robert Schwab of the University of Maryland. 13   Analyzing data from the Survey of Income Dynamics, they determined that alcohol taxes are definitely regressive when studying annual income. 

However, over longer periods - including a five-year horizon and lifetime consumption - the regressivity of the alcohol tax is reduced relative to the annual data.  This occurs because income data covering multiple years reduces the variation in the income distribution and the distribution of alcohol consumption.  But the alcohol tax's regressivity, while reduced, is not eliminated.  As the authors note, "The alcohol tax remains firmly regressive."

Alcohol Consumption and External Costs

Alcohol consumption creates externalities - costs paid by people who do not enjoy the benefit.  Expressing external costs as a tax rate on alcoholic beverages is complex - even more so than for tobacco products - since externalities are largely due to abusive consumption.  Readers should note that the vast majority of consumers of alcoholic beverages consume the product responsibly. 14   Even so, external costs do exist and should be addressed. 

Societal Costs of Alcohol Abuse .  A study conducted for the National Institute on Alcohol Abuse and Alcoholism estimated the annual costs of alcohol abuse as high as $184 billion, including lost earnings, property damage, traffic accidents, costs to the criminal justice system, and medical costs associated with illness, premature death and fetal alcohol syndrome. 15   Granted, not all of this cost is external, since, for example, lost earnings are included in the calculation.  Regardless, as stated earlier, total excise tax revenues on alcohol at the federal and state levels total only about $13.9 billion annually, far shy of the total supposed external cost.  Alcohol abuse also imposes other social costs: A number of studies show strong correlation between domestic violence and alcohol consumption.

“Moderate consumption of alcohol has positive effects on health.”

Positive Effects of Moderate Consumption .  However, consumption of alcohol can have positive effects.  Moderate consumption - a maximum of two to three glasses of wine or beer a day for men and one to two glasses a day for women - has been shown to have a greater impact on reducing heart disease than any other factor aside from the cessation of smoking.  Psychologically, alcohol can make one more relaxed, able to talk more easily and feel less tired.  Some even claim it helps people socialize.

Take the case of beer.  Since the 1990s, more than 100 studies have shown it is rich in protein, B vitamins and certain minerals, such as magnesium, cadmium and iron.  For subjects consuming one to two beers a day, beer provides 14 percent of calories, 11 percent of dietary protein, 12 percent of dietary carbohydrates, nine percent of dietary phosphorus, seven percent of dietary riboflavin and five percent of dietary niacin.  Some of its nonalcohol components, called phenols, appear to have anti-oxidant properties that reduce LDL cholesterol ("bad" cholesterol) oxidation. 

Other studies suggest red wine seems to offer increased cardiovascular protection - particularly in areas with cold climates - and speculate the effect results from polyphenols, chemical components contained in grape skin.  And still others find a significant inverse association between the risk of heart disease and the drinking of spirits.  Moreover, alcohol has been found to act as a blood thinner, which can lower rates of heart disease in both men and women.

Taxing Alcohol .  One way to recoup the negative costs - or reduce the incidence of the behavior, and thus its external cost - is to increase the excise tax on alcoholic beverages.  But problems arise.  First, as discussed in this paper, alcohol taxes are regressive.  Placing the burden on those least able to pay may not be the best public policy approach.  Second, the taxes fall not only on problem drinkers, but also on those who pose no costs on society, including light drinkers for whom alcohol consumption may produce health benefits.  Third, if the goal is to reduce consumption - and thereby reduce the external costs - heavy taxation may discourage light or moderate drinkers, but it is unlikely to deter heavy drinkers, who may resort to spending even more of their household resources on alcohol.  Finally, policymakers should remember, just as there are social costs to alcohol consumption, extreme policies designed to discourage alcohol consumption can result in costs that far exceed benefits, as evidenced by Prohibition.

Where the Money Goes .  An examination of recently proposed alcohol tax increases in the states shows some of the money would have been targeted to alcohol awareness and abuse prevention programs.  But much of the money was slated for myriad projects, such as offsetting an income tax reduction, pay raises for teachers and education. 18    If the new money generated by higher taxes on alcohol simply funds general budget items not related to alcohol abuse prevention programs, then state legislatures are simply punishing alcohol consumers - and unduly burdening lower-income people - to add cash to state coffers.

Conclusion

“Higher alcohol taxes penalize moderate drinkers and lower-income families.”

Alcohol taxes are like a patchwork quilt in that tax rates vary by product and vary widely from state to state with no coherent pattern.  The bottom line is they impose a burden on those least able to afford them:  Lower-income individuals sacrifice more of their income to pay alcohol taxes than more affluent drinkers.  Unlike tobacco taxes, studies show the external costs of abusive alcohol consumption are not fully recouped by excise tax revenues.  But policymakers shouldn't rush to raise the taxes.   Increasing the taxes to fill the gap penalizes moderate drinkers and lower-income families.

References: Table II-1 , Table-II-2

Notes

1 The first Civil Service Act in 1883 was a response to the involvement of government officials abetting alcohol tax evasion by the "Whiskey Ring" in 1875.  The first stock car racers in the 1930s began as "moonshiners" seeking to avoid revenue officials by driving high-performance cars while delivering untaxed alcohol.

2 "Proposed Increases in State and Local Taxes and Fees, 2005," Alcohol Policies Project, Center for Science in the Public Interest, Web page, updated November 3, 2005.  Available at http://www.cspinet.org/booze/taxguide/2005/2005TaxStateUpdate.htm.

3 "Revenue Option 47 - Increase All Alcoholic Beverage Taxes to $16 per Proof Gallon," in Congressional Budget Office, Budget Options (Washington, D.C.: Congress of the United States, February 2005).

4 A proof gallon equals one gallon of 100-proof alcohol (or 50 percent alcohol by volume).

5 "Federal Excise Taxes Reported to or Collected by the Internal Revenue Service, Alcohol and Tobacco Tax and Trade Bureau, and Customs Service, by Type of Excise Tax, Fiscal Years 1998-2005," Internal Revenue Service, 2005, Table 21.  Available at http://www.irs.gov/pub/irs-soi/histab21.xls.

6 "State and Local Alcoholic Beverage Tax Revenue, Selected Years 1977-2004," Urban Institute-Brookings Institution Tax Policy Center.  Available at http://www.taxpolicycenter.org/TaxFacts/TFDB/TFTemplate.cfm?Docid=399.

7 Much of the evidence on who buys what kind of alcohol is anecdotal or comes from studying specific groups of buyers or specific products.  For example, a campaign in 2002 to roll back the federal tax on beer generated a wealth of information about who paid the tax.  This roll-back campaign came 12 years after Congress raised taxes on luxury items such as expensive cars, fur coats, jewelry, yachts and private airplanes.  For reasons probably best understood by the politicians involved, beer was included among these "luxuries" and the federal excise tax doubled to $18 per barrel (32 cents per six-pack). By the early 1990s it was obvious that this luxury tax was devastating certain industries, such as yacht construction, and it was quickly repealed on most goods.  But since beer drinkers lacked the concentrated political clout of other industries, the beer tax remained.  Finally, in 2002, a coalition of brewers and wholesalers organized an unsuccessful attempt to repeal the tax.  They were opposed largely by public interest groups who viewed the tax as a way to discourage drinking.

8 Of course, the difference between higher- and lower-earning households would be mitigated - though not completely eliminated - by the average number of people in higher- and lower-earning households.  For example, in a recent survey, households with less than $10,000 in annual income have, on average, 1.8 adults per household.  Households with annual incomes greater than $100,000 have, on average, 2.3 adults.  See Charles T. Clotfelter et al., "State Lotteries at the Turn of the Century: Report to the National Gambling Impact Study Commission," June 1, 1999.

9 Douglas J. Young and Agnieszka Bielinska-Kwapisz, "Alcohol Taxes and Beverage Prices," National Tax Journal , Vol. 55, No. 1, March 2002, pages 57-73.

10 Donald S. Kenkel, "Are Alcohol Tax Hikes Fully Passed Through to Prices? Evidence from Alaska," American Economic Review , Vol. 95, No. 2, May 2005, pages 273-77.  In 2002 Alaska experienced some of the most dramatic alcohol tax increases in recent history; taxes increased from 35 cents to $1.07 per gallon for malt beverages, 85 cents to $2.50 per gallon for wine, and $5.60 to $12.50 per gallon for distilled spirits.

11 "Consumer Expenditures in 2004," U.S. Department of Labor, Bureau of Labor Statistics, Report No. 992, April 2006.

12 Geoffrey Paulin, "Consumer Expenditures for Alcohol in 2000," in "Consumer Expenditure Survey Anthology, 2003," U.S. Department of Labor, Bureau of Labor Statistics, Report No. 967, September 2003.

13 Andrew B. Lyon and Robert M. Schwab, "Consumption Taxes in a Life-Cycle Framework: Are Sin Taxes Regressive?" Review of Economics and Statistics , pages 89-406.  The Panel Survey of Income Dynamics (PSID) is an annual survey of households that began in 1968. Lyon and Schwab reported the expenditures on alcohol and cigarettes by income group (including the expenditures on cigarettes yields an interesting comparison of alcohol to a tax that, as we have seen, is highly regressive).  Households are grouped by quintiles, with the first quintile being the 20 percent of households with the lowest income.  From the PSID it is possible to calculate the total amount earned by all households in the sample and also the total amount earned by those households at or below each income quintile.

14 As a result, an efficient alcohol tax may not recoup all external costs.  That is, abusive drinkers will likely consume alcohol even in the face of high taxes, even if it means shifting to cheaper products.  This suggests a higher tax will do more to discourage consumption by responsible drinkers who generate little external cost.  Thus, an ideal tax rate may not generate tax revenue as large as the external cost. 

15 Henrick Harwood, "Updating Estimates of the Economic Costs of Alcohol Abuse in the United States: Estimates, Update Methods and Data," report prepared by the Lewin Group for the National Institute on Alcohol Abuse and Alcoholism, December 2000, based on estimates, analyses, and data reported in Henrick J. Harwood, Douglas Fountain and Gina Livermore, "The Economic Costs of Alcohol and Drug Abuse in the United States 1992," report prepared for the National Institute on Drug Abuse and the National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, U.S. Department of Health and Human Services, Publication No. 98-4327, September 1998. Available at http://pubs.niaaa.nih.gov/publications/economic-2000/alcoholcost.PDF.

16 Lenore E. Walker, The Battered Woman Syndrome (New York: Springer Publishing Company, 1984),  cited in Sara Markowitz, "The Price of Alcohol, Wife Abuse and Husband Abuse," Southern Economic Journal , Vol. 67, No. 2, October 2000, pages 279-303.  In addition, Markowtiz, of Rutgers University, found that a 1 percent increase in the price of pure alcohol decreases the probability of wives being the victims of spousal abuse by about 5 percent.  A Swedish study found a strong correlation between suicide and alcohol abuse.  Thor Norstrom, of Stockholm University, found that a one-liter increase in alcohol consumption per capita is associated with a 10 percent increase in the suicide rate.  He also found that alcohol consumption influenced the incidence of suicide even more than divorce or unemployment.  See Thor Norstrom, "The Impact of Alcohol, Divorce and Unemployment on Suicide: A Multilevel Analysis," Social Forces , Vol. 74, No. 1, September 1995, pages 293-314.

17 For more information see, Alfredo C. Cordova et al., "The Cardiovascular Protective Effect of Red Wine," Journal of the American College of Surgeons , Vol. 200, No. 3, March 2005, pages 428-39; Dean Edell and David Schrieberg, Eat, Drink and Be Merry: America's Doctor Tells You Why the Health Experts Are Wrong (New York: HarperCollins, 1999), pages 191-92; Pete du Pont (National Center for Policy Analysis), "A Drink to Your Health," Washington Times , June 26, 2001; Pete du Pont (National Center for Policy Analysis), "Belly Full of Benefits," Knight-Ridder Tribune Newswire, June 28, 1999; Eric B. Rimm et al., "Review of Moderate Alcohol Consumption and Reduced Risk of Coronary Heart Disease: Is the Effect Due to Beer, Wine, or Spirits?" British Medical Journal , Vol. 312, No. 7033, March 23, 1996.

18 "Proposed Increases in State and Local Taxes and Fees, 2005," Center for Science in the Public Interest. 


Read Article as PDF