Taxing the Poor

Policy Reports | Taxes

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

Section III: Government-Sanctioned Gambling


Taxes on gambling are similar to excise taxes on tobacco and alcohol products, to a degree:  Each commodity comes in different shapes and sizes, and in most locales different types are taxed in different ways  However, gambling taxes are more complex because there are so many varieties, sponsors and venues of games, from billion-dollar commercial casinos to charity bingo tournaments.  This section considers the effects of gambling taxes on lower-income households.

Although this section looks at many different types of gambling, it focuses primarily on state-sponsored lotteries, both because they constitute possibly the single largest source of gambling revenue and taxes and because of the availability of data on lotteries that lends itself to analysis.  However, the evidence shows that regardless of the game, gambling taxes place a disproportionate burden on lower-income households. 

Gambling: Everyone Is Doing It

“Regardless of the game, gambling taxes place a disproportionate burden on lower-income households.”

A quarter-century ago, only three states allowed legalized gambling; now, every state except Utah and Hawaii allows some form of gambling.  Today, Americans spend more money on more forms of legal gambling than at any other time in the nation's history, and gambling has become a fixture in today's popular culture.  Casinos are opening at a record pace, state lotteries are recording record sales and shows like the "World Series of Poker" air during prime time on ESPN.  Gambling is big business:

  • In 2005, almost 900 casinos (including commercial, tribal and racetrack casinos) were open in 36 states; consumer spending at casinos, which has almost doubled over the past decade, totaled $30.3 billion.
  • Also in 2005, state-sponsored lotteries in 41 states raked in a record $52 billion, up from just $19 billion in 32 states in 1989.
  • Americans spend $33 billion on slot machines annually and wager $13 billion a year on Internet gambling.

Gambling opportunities are continually expanding.  The quest to bring casino gambling to dog and horse racetracks - or "racinos" - has become a new trend.  Eleven states already allow racinos and six more are considering them.  Meanwhile, Arkansas, Alabama and Wyoming are considering joining the 42 states that offer lotteries.  And online gambling - the most rapidly expanding form of gambling - is wildly popular.  [See the sidebar "Betting on Online Gambling."]

“Two-thirds of Americans gambled in the past year.”

Who Plays?   Two-thirds of Americans say they have gambled in the past year and another 12 percent say they gambled in previous years.  Of those who gambled, 52 percent bought a lottery ticket, 29 percent visited a casino and one-fourth played a slot machine.  Others bet in office pools, played cards for money, played video poker and so forth.  Gambling participation cuts across all races, sexes, education levels and household incomes. 

Interestingly, the games of choice are changing.  The percentages of Americans who visited a casino or played a slot machine rose over the past two decades (from 20 percent to 29 percent for casinos, and from 19 percent to 24 percent for slot machines), while betting on sports and horse racing have become less popular (from 22 percent to 14 percent for sports, and from 14 percent to 5 percent on horse racing).

Sidebar: Betting On Online Gambling

Attitudes about Gambling .  Oddly, as availability and access increased, approval of gambling - while still quite high - has lessened:

  • In a recent poll by the Pew Research Center, 71 percent say they approve of lotteries, down from 78 percent in 1989.
  • About 51 percent approve of legalizing casino gambling as a way for states to raise money, down from 54 percent in 1989.

The slip in popularity seems to stem not from a return to America's Puritanical roots, but from a concern that gamblers will become addicts: 70 percent said legalized gambling leads people to wager more than they can afford, compared to 62 percent in 1989.

Taxing Gambling

“The lottery is the most heavily taxed form of gambling, yielding $52.4 billion in total revenues in 2006.”

Most gambling taxes are collected by state governments.  By far the largest source of gambling tax revenues is state lotteries, which last year yielded some $52.4 billion in total revenues and $16.4 billion in state proceeds (that is, total revenue less prizes and administrative costs). The lottery is also the most heavily taxed form of gambling:  On average, states keep about 33 cents of each dollar collected after subtracting prizes and operating costs.  If state proceeds were treated as an implicit tax, the tax rate would be 49 percent (33 cents as a percentage of the 67-cent cost of operation and prizes).

However, to produce an ad valorem tax rate comparable to excises on tobacco and alcohol products, the calculation should include the full cost of providing the product; that is, both state proceeds and administrative costs should count toward the lottery's cost to consumers.  Stated another way, the tax rate calculation would be state proceeds (33 cents) plus administrative cost (6.1 cents) as a percent of prizes (60.9 percent, which is total revenue after state proceeds and administrative costs are subtracted).  Thus, the implicit tax rate paid by state lottery participants in the aggregate actually exceeds 64 percent. 

This rate is much higher than any state's sales tax and higher than many state excise tax rates on other "sinful" products.  For instance:

  • One of the lowest lottery tax rates, in Massachusetts, is still twice the state's sales tax rate.  [See Table III-1 .]
  • One of the highest lottery tax rates is found in Montana, which doesn't even collect sales taxes.

Income taxes collected on lottery winnings could also be included in a calculation of lottery taxes, driving the implicit tax rate higher.  However, reliable estimates do not exist and are not considered in this paper.

State Monopolies .  Virtually all gambling enterprises (with the exception of online gambling) are either state monopolies or state franchises.  The government determines how much and where people gamble, and how many opportunities exist.  For example, the monopoly status of state-sponsored lotteries allows states to charge whatever rates they please; after all, they are the only game in town.  Recall that the states retain some 33 cents of each dollar of revenue (after subtracting administrative costs). 

“Massachusetts has one of the lowest lottery tax rates, but it is still twice the state's sales tax rate.”

Casino Gambling.  Commercial casinos, while highly regulated, operate in a much more competitive environment.  The casino industry's house "take" is but a fraction of the state's profit from the sale of lottery tickets.  By one recent estimate, casinos retain about 4.4 cents of each dollar collected - the rest is redistributed in winnings.

The Regressive Nature of Gambling Taxes

There is virtually unanimous agreement among leading researchers that excise taxes on gambling, particularly implicit taxes on lottery sales, are regressive.  Overall, lower earners spend a greater percentage of income on gambling - and in some cases spend greater real dollar amounts - than higher earners.  Thus, by default, lower earners pay a higher rate of tax as a percent of income.

“Lower earners gamble more real dollars and a larger percentage of their incomes than higher earners.”

Lower-Income Gamblers Bear a Heavier Load .  Surveys show participation in the lottery is broad-based, but the amount of play varies - sometimes quite a bit - between different groups.  One of the most insightful examinations of lottery play was conducted for the National Gambling Impact Study Commission by researchers at Duke University. Using the results of an extensive survey by the National Opinion Research Corporation, the researchers determined:

  • The lottery participation rate generally increases with income, although it falls for people with incomes higher than $100,000; the highest rate of play is for those with incomes between $50,000 and $100,000.  [See Table III-2 .]
  • By contrast, the annual amount spent, or per capita play, by gamblers is actually highest for lower-income households ($597 per year), exceeding any other income category, and more than double the amount for the highest earners ($289 per year, on average).

Since many households have more than one adult, and since higher-income households have more adults per household than lower-income households, the study also examines per household spending:

  • Even after the adjustment, households earning just $10,000 spend twice the amount on gambling as households earning $90,000.
  • Put another way, the lowest-earning households spend about 10.8 percent of income on gambling, versus 0.7 percent of income for the highest earners.

Clearly, lottery expenditures are regressive:  Lower earners not only spend larger percentages of their incomes gambling but also spend larger real dollar amounts.  As a result, lottery taxes are also regressive, since the implicit lottery tax is incurred as a percent of total purchases, and lower earners pay a greater share of their income in lottery taxes.

Who Are the Heaviest Gamblers?  The Duke study examined above also shows that despite broad-based lottery participation, most spending comes from a small subset of frequent players.  The top 5 percent of players account for 54 percent of total sales, while the top 20 percent of players account for 82 percent of total sales.  If all players spent the same as the "typical" median player, sales would fall 87 percent!  Clearly, these high-frequency players push lottery expenditures up.  So, who are they?

  • The largest amounts are spent by high school dropouts, who represent 20.3 percent of players but only 12.3 percent of U.S. adults, and households with incomes under $10,000 (9.7 percent of players versus 5 percent of adults). 
  • The median age of the heaviest players is slightly higher than that of the overall U.S. population (47.5 years versus 43.0 years).

These results reinforce the previous findings: Lottery revenue comes disproportionately from those with lower incomes. 

Corroborating Evidence .  Other survey results are generally similar to the Duke study.  A recent survey by the Pew Research Center suggests a similar rate of lottery play, in that participation trends upward for lower- to medium-income earners (from 44 percent to 60 percent), but drops off for the highest earners (57 percent for those earning more than $100,000). Participation rates for all types of gambling combined (not just the lottery) is much higher, ranging from 59 percent among those earning less than $30,000, and steadily increasing to 79 percent for the highest earners.  The survey did not collect data on gambling expenditures.

Another recent survey, by analysts at the Research Institute on Addictions, suggests participation in the lottery is slightly higher than the Duke results, ranging from 58 percent to 69 percent for the first four income quintiles, but dipping to 61 percent for the highest income quintile (those with an average income of $115,000). Like the Duke study, this survey shows that lower earners, in general, put more money on the line than higher earners when it comes to lottery play. 

“Lower-income players spend more than double the amount of the highest earners each year.”

How Regressive Is the Lottery Tax?  Most research into the burden of lottery taxes focuses on specific states.  Researchers Donald Price and Shawn Novak examined the regressivity of three lottery games in Texas (Lotto Texas, Pick 3 and an instant-win game), a state with one of the largest amounts of lottery revenue. The authors determined that lottery taxes are the most regressive of all Texas taxes, and the instant-win game is most regressive of all.  

“The instant win game imposes a tax more than twice as regressive as a general sales tax.”

One measure of regressivity employed by the authors is the "Suits Index," which compares the cumulative portion of the tax paid by a cumulative portion of the population.  The Index varies from +1 to -1; positive values indicate tax progressivity and negative numbers indicate tax regressivity.  The study shows the lottery tax is more regressive than virtually any other tax, including the sales tax, payroll taxes or personal property taxes.  In fact, the instant-win game imposes a tax more than twice as regressive as a general sales tax.  [See Table III-3 .]

Consumer Welfare and Paternalism

Most states earmark lottery revenues for programs that provide benefits to lower-income populations.  If earmarked gambling revenues supplement rather than substitute for existing funding, and if the projects bestow relatively larger benefits to the poor, that could be taken into account when determining the regressivity of gambling taxes. 

Earmarks are often adopted to boost the popularity of legalizing or expanding gambling opportunities.  But it is unclear whether the earmarks, say, for education, truly increase funding for schools or whether they simply replace money that would otherwise have been spent on education but can now be funneled to other state priorities.  Ross Rubinstein and Benjamin Scafidi examined the state-sponsored lottery in Georgia to shed light on this issue. In Georgia, as in other states, higher-income households spend a smaller fraction of their income on the lottery than lower earners, and they also tend to spend smaller dollar amounts.  Almost all lottery revenues are devoted to the Helping Outstanding Pupils Educationally (HOPE) program, which provides scholarships for college and universal pre-kindergarten.  Even after the researchers accounted for the benefits of the programs, they determined that the lowest-income households experience a net loss of $160.55, on the average, from the Georgia lottery, while families earning three times that amount receive a net gain of $113.81.   

“Although most lottery profits are earmarked for welfare programs, the revenues typically substitute rather than supplement existing funding.”

Regardless, as a hypothetical exercise, assume all of a state's profit from a lottery was somehow earmarked to welfare programs that directly benefit the poor.  Assume also that the earmarked funds actually increase the state's expenditures on those programs, rather than substitute for money that would have otherwise been spent on them.  Would this justify even higher taxes on gambling, such as lowering the payout percentage going to lottery winners?  Higher tax rates would either force lower-income households to spend more on gambling or reduce their gambling.  In the first case, lower-income families would be worse off because they would have less to spend on other priorities.  In the second case, they would be worse off because they would be unable to pursue an activity they enjoy.

Should the government paternalistically "protect" lower earners from their own decisions?  That is a question that cannot be answered by scientific analysis.

Additional Costs and Benefits of Gambling

Several issues should be addressed that may affect the relative effect of gambling taxes on lower-income earners, possibly making the taxes more or less regressive.

Legalization, Taxation, Utility and Regressivity .  The taxation of gambling should not be analyzed without considering the question of legalization.  If states did not legalize gambling, people would not get the enjoyment of playing (ignoring, for a moment, illegal gambling opportunities).  Economists examine people's behavior to determine the value they place on a product or service:  Those who play the lottery most are likely get the most enjoyment out of it.  Since lower-income people spend more on gambling in proportion to their income, they are relatively heavier users of the product and are presumably better off because of its legalization and availability.

Thus, some may argue that even in the face of regressive tax rates, the disproportionate use of the product by lower earners actually makes the legalization-cum-taxation function pro-poor.  As researcher Charles Clotfelter says, in this context, regressivity "becomes relevant only to the question of raising or decreasing the tax on a form of gambling already legalized."

Of course, it has also been established that lottery taxes are quite high compared to state sales tax rates on other forms of entertainment and - in many cases - higher than excise tax rates on other products.  Because of the high rates, lottery taxes are even more regressive than would otherwise be the case.  As a result, the same lower-income people who benefit from gambling would benefit even more if the taxes were in line with the rates on other products.

Social and Economic Costs of Gambling .  Much of the research on gambling focuses on the social and economic costs and benefits to the surrounding communities, gamblers and state revenues, which may have a bearing on estimates of regressivity.  For example, if lower-income populations receive a disproportionate economic or social benefit from the legalization and spread of gambling opportunities, it might mitigate the regressivity of the taxes. 

Unfortunately, after years of study, the National Gambling Impact Study Commission - a blue-ribbon commission authorized by Congress in the late 1990s - concluded it was "currently impossible to obtain even a rough approximation of a true cost-benefit calculation concerning the economic impact of legalized gambling."  The commission also noted that high-quality research is sparse.   However, empirical analyses have shown the most prominent results of legalized gambling on communities, gamblers and American Indian tribes.  For example:

  • The obvious winners are the gambling businesses themselves and the consumers who enjoy consuming the product; generally, state revenues tend to increase, as well. 
  • Opening new American Indian casinos seems to increase employment and wages in the surrounding community, with measurable positive economic effects on American Indian tribes who operate casinos; their mortality rates seem to improve, perhaps due to the prevalence of new jobs.
  • On the other hand, new casinos bring an increase in the prevalence of violent crimes, auto thefts, larceny and other property crimes, as well as a rise in bankruptcies.

“On the average, households displace 1.9 percent of their quarterly consumption with state lottery purchases.”

Others have attempted to study how gambling affects household spending.  A recent study published by the National Bureau of Economic Research found household expenditures on state lotteries do, indeed, crowd out nongambling expenditures. On the average, households displace 1.9 percent of their quarterly consumption, or $38 per month, with state lottery purchases.  However, the lowest earners displace 2.7 percent of their consumption, or $46 per month, from other household expenditures.

With both positive and negative economic effects - varying by location, type of gambling and makeup of the surrounding community - it is difficult to discern a net economic effect. 

Further, even if the net economic effects are positive, there are also social effects to consider, such as the harms a gambler does to the people around him.  Granted, it is difficult to identify a causal relationship between gambling and other problems, as problem-gamblers also often have other behavioral disorders.  Thus, nearly every study on the costs and benefits of gambling comes to the same conclusion:  More research is needed.  This paper, too, calls for more study on these issues in the hope that future decision-makers can be better informed by more detailed economic analyses.

“Although most gamblers lose most of the time, 71 percent say they play for pure enjoyment, not to get rich quick.”

Are Gamblers Rational or Just Rash?   When asked to consider all the money spent on gambling over the past year, less than a third of gamblers think they are ahead for the year.  So if gamblers lose most of the time, why do they keep playing?  For the entertainment value:  Some 71 percent of gamblers say they play for pure enjoyment, not to get rich quick.

But while a great many people enjoy gambling, a great many people also enjoy exercising, collecting butterflies, watching videos and any number of other leisure activities.  Judged against other forms of diversion, gambling has a significant but not dominant share of the market.  However, unlike these other activities, gambling is subject to a wide range of state and federal taxes. 

This strikes at the very heart of the purpose of taxes on gambling.  If gambling is perceived as a harmless alternative to other forms of entertainment, it should be taxed at a similar rate as other activities.  But if gambling is perceived as a "sin" with deplorable social side effects, governments should weigh the effects of legalization against the costs (both social and financial) of high tax rates on lower-income populations. 

Gamblers' Responsiveness to Changes in Price .  Economists generally measure responsiveness to price using elasticity of demand.  [See discussion in Section I: Excise Taxes and Tobacco.]  In the case of state lotteries, elasticity would be the percentage change in lottery tickets purchased after a 1 percent increase in price (that is, a 1 percent reduction in the payoff). 

Gamblers' responsiveness to price increases has important implications for determining the appropriate level of gambling taxes.  For example, if one believes gambling has serious social consequences - and if elasticity is greater than one - higher taxes on gambling serve the useful purpose of inducing people to stop or reduce their play.  However, if elasticity is less than one, gambling will not be abetted and gamblers' costs will rise.  Thus, there would be less justification for imposing regressive taxes on the poor if the demand will not be stemmed by high taxes. 

Sadly, there is general consensus that little reliable data exists on the price sensitivity of gambling.  There is also a general consensus that most forms of gambling are unlikely to be sensitive to changes in price.  First, unlike normal consumer goods, the price of gambling is not readily apparent.  Second, it is unlikely lotteries could continue to remain so popular at such high tax rates (49 percent) if high prices led people to flee the market. 26   However, there is some contrary evidence presented in a few recent studies.  For example, leading researchers Charles Clotfelter and Philip Cook find price sensitivity for the lottery is very high (price elasticity of 2.55), meaning higher prices, or lower payouts, encourage people to gamble less. 27   It is possible that both results can be true, in that it is possible the demand for lotteries is sensitive at very high prices, but not at lower prices. 

Of course, if one believes gambling is more like other harmless recreational activities - with few social consequences - there is no justification for taxing gambling more than other recreational activities, whether elasticity is high or not.


Households with lower incomes spend more on gambling than higher-income households, both as a percentage of income and in real dollar amounts.  Thus, proportionately, the poor bear a heavier burden of gambling taxes, meaning gambling taxes are regressive.  Lawmakers who are considering raising gambling taxes at the federal, state or local level should consider the disproportionate burden their lower-income constituents will bear.

References: Table III-1 , Table-III-2 , Table-III-3


1 Consider, for example, the different tax treatment between cigarettes and smokeless tobacco.  While smokeless tobacco is, on balance, "healthier" than cigarettes, it is taxed by many states at a higher rate.  For more information, see the "Excise Taxes and Tobacco" section.

2 In one additional state, Pennsylvania, racetrack casinos have been legalized but are not yet operational. See the American Gaming Association, "Industry Information: Tax Payments," Fact Sheet, April 2006. Available at Also see "State of the States: The AGA Survey of Casino Entertainment," American Gaming Association, 2006. Available at

3 North Carolina added a lottery in March 2006, bringing the current total to 42 states.  Paul Taylor et al., "Gambling: As the Take Rises, So Does Public Concern," Pew Research Center, Social Trends Report, May 23, 2006.  Available at http://pewre

4 Ibid.

5 Kavan Peterson, "48 States Raking in Gambling Proceeds,", May 23, 2006.  Eleven states allow racinos:  Delaware, Florida, Iowa, Louisiana, Maine, New Mexico, New York, Oklahoma, Pennsylvania, Rhode Island and West Virginia; the six states considering them are Georgia, Massachusetts, Minnesota, Mississippi, Ohio and Texas. 

6 Ibid.

7 Ibid.

8 Paul Taylor et al., "Gambling: As the Take Rises, So Does Public Concern."

9 This section draws heavily from Charles T. Clotfelter, "Gambling Taxes," in Sijbren Cnossen, ed., Theory and Practice of Excise Taxation (New York, N.Y.: Oxford University Press, 2005), pages 84-119.

10 North American Association of State and Provincial Lotteries.

11 Indeed, some of the academic and policy community literature express the implicit lottery tax rate in this way.  Tax Foundation, "Implicit Tax Rates on State Lottery Sales FY 2004," March 9, 2006. Available at:

12 To demonstrate the complexity of assessing taxes on other forms of gambling, consider taxes on casino games.  Dissecting this figure is a daunting task since the calculation includes implicit excise taxes, such as taxes on admissions, gross revenue, license fees and negotiated payments for casinos operated by American Indian tribes.  Last year, according to the American Gaming Association, casinos paid $4.93 billion in federal and state property and gambling taxes.  As mentioned in the text, total spending at casinos totaled $30.3 billion, meaning casino gamblers as a whole paid a tax rate of roughly 16.3 percent (assuming corporate taxes paid by casinos are passed on to consumers through higher prices).  American Gaming Association, "Industry Information: Tax Payments," Fact Sheet, April 2006.

13 See Clotfelter, "Gambling Taxes."

14 The survey, by the National Opinion Research Corporation in 1998, includes 4,358 households.   Since surveys often suffer from individuals' underreporting of the true level of their activity, the paper adjusts for this by comparing the true level of lottery spending with the amount that would be consistent with the reporting levels.  See Charles T. Clotfelter et al., "State Lotteries at the Turn of the Century: Report to the National Gambling Impact Study Commission," Duke University, June 1999.

15 Paul Taylor et al., "Gambling: As the Take Rises, So Does Public Concern," Pew Research Center, May 23, 2006.

16 John W. Welte et al., "Gambling Participation in the U.S. - Results from a National Survey," Journal of Gambling Studies , Vol. 18, No. 4, December 2002, pages 313-97.

17 Donald I. Price and E. Shawn Novak, "The Income Redistribution Effects of Texas State Lottery Games," Public Finance Review , Vol. 28, No. 1, January 1, 2000, pages 82-92.

18 Ross Rubenstein and Benjamin Scafidi, "Who Pays and Who Benefits? Examining the Distributional Consequences of the Georgia Lottery for Education," National Tax Journal , Vol. 55, No. 2, June 2002, pages 223-38.

19 Clotfelter, "Gambling Taxes."

20 Part of the reason implicit taxes on state lotteries are so high is that state lotteries are essentially monopolies.  Recall that the house "take" for casinos, a competitive industry, is about 4.4 cents of each dollar collected, while states retain about 46 cents from state lotteries.  See Clotfelter, "Gambling Taxes."

21 "National Gambling Impact Study Commission Final Report," National Gambling Impact Study Commission, June 1999, pages 7-29.  Available at

22 Readers are encouraged to examine the results of a number of studies that examine different aspects of the positive and negative externalities of gambling.  Among them: Chapter 7 of the National Gambling Impact Study Commission's Final Report; Melissa S. Kearney, "The Economic Winners and Losers of Legalized Gambling," National Bureau of Economic Research, Working Paper No. 11234, March 2005; William N. Evans, "The Social and Economic Impact of Native American Casinos," National Bureau of Economic Research, Working Paper No. 9198, September 2002; and Dean Gerstein et al., "Gambling Impact and Behavior Study: Report to the National Gambling Impact Study Commission," National Opinion Research Center, April 1, 1999; available at

23 Melissa Schettini Kearney, "State Lotteries and Consumer Behavior," National Bureau of Economic Research, Working Paper No. 9330, November 2002.

24 Ibid.

25 "National Gambling Impact Study Commission Final Report," National Gambling Impact Study Commission, June 1999.

26 "Australia's Gambling Industries," Productivity Commission of Australia, Report 10, November 26, 1999.   Appendix D provides an overview of price sensitivity of gambling using evidence from around the world.

27 Charles T. Clotfelter and Philip J. Cook, "On the Economics of State Lotteries," Journal of Economic Perspectives , Vol. 4, No. 4, Fall 1990, pages 105-19.

Read Article as PDF