matthew kimball lottery

Matthew kimball lottery

Research-based policy analysis and commentary from leading economists

  • Columns
    • By Topic
    • By Date
    • By Reads
    • By Tag
  • Covid-19
    • Economics in the Time of Covid-19
    • Mitigating the COVID Economic Crisis
    • Covid Economics
  • Vox Multimedia
    • Video Vox
    • VoxTalks
  • Publications
    • Books
    • CEPR Reports
    • Policy Insights
  • Blogs&Reviews
  • People
    • A
    • B
    • C
    • D
    • E
    • F
    • G
    • H
    • I
    • J
    • K
    • L
    • M
    • N
    • O
    • P
    • Q
    • R
    • S
    • T
    • U
    • V
    • W
    • X
    • Y
    • Z
  • Debates
    • Economics in the Time of Covid-19
    • Populism
    • WWII
    • Digital Money
    • Euro Area Reform
  • Events
    • Workshops & Conferences
    • Courses
    • Discussion Meetings
    • Submit an Event
  • About
    • About Vox
    • Help
    • Subscribe
    • Contact

Using the labour supply responses of lottery winners to evaluate tax and transfer policies

David Cesarini, Erik Lindqvist, Matthew Notowidigdo, Robert Östling 24 January 2016

Cash welfare programmes are widely thought to discourage work because unearned income reduces the labour supply even when it does not alter work incentives. This column discusses recent evidence from Swedish lottery players suggesting that this ‘income effect’ is economically significant, but modest in magnitude and surprisingly similar across various demographic groups. Introducing ‘unconditional basic income’ programmes in developed countries may reduce the labour supply across a broad cross-section of the population.


Recently, citizens and politicians in Switzerland began considering a universal basic income programme, which is currently scheduled for a voter referendum in late 2016. The programme attracted widespread media attention; the Washington Post described the proposal as ‘universal and unconditional’ with every citizen receiving the guaranteed income no matter what, with “no work requirements, no means-testing, and no restrictions on how the money is used”.

In many countries, cash welfare programmes are designed instead as conditional transfer programmes and phased out as income increases. As is taught in many introductory labour economics and public economics classes (e.g. Chapter 2 in Gruber 2011), conditional transfer programmes generate both income and substitution effects due to higher ‘effective’ marginal tax rates for low-income workers. As a result, a common criticism is that conditional transfer programmes discourage work.

By contrast, the universal and unconditional aspect of the Swiss policy does not change effective marginal tax rates or work incentives. However, the programme could still reduce employment and average hours worked through the income effect. 1 Most existing empirical evidence points towards leisure being a normal good, so that increases in (unearned) income reduce labour supply (Keane 2011).

Empirical evidence on income effects in labour supply

A large empirical literature has provided estimates of income effects using a variety of different empirical strategies, but disagreement over the magnitude remains. Although some agreement exists that large, permanent changes in real wages induce relatively modest differences in labour supply, Kimball and Shapiro (2008) write that “there is much less agreement about whether the income and substitution effects are both large or both small”. As a result, it is difficult to precisely forecast the labour market consequences of basic income programmes like the one discussed above.

A key empirical challenge in estimating income effects in labour supply is isolating variation in unearned income or wealth that is unrelated to other determinants of labour supply. One way to address this challenge is to use lotteries, which randomly assign wealth. This allows researchers to study the effect of unearned income on labour supply using the well-understood framework of a randomised controlled trial. In many lotteries, participants purchase different numbers of tickets, which means that the participants do not all have the same chance of winning. However, by conditioning on the number of tickets, one can provide credible estimates of the effect of lottery wealth on labour supply.

The use of lotteries to study labour supply is pursued in a classic paper by Imbens et al. (2001), who study a small sample of 496 lottery players in Massachusetts by linking survey data on the lottery players to administrative data containing information on annual labour earnings. They find modest reductions in labour earnings suggesting every dollar of universal basic income would reduce labour earnings by roughly $0.11.

New evidence from Swedish lottery winners

In recent work, we build on and extend the work by Imbens et al. (2001) using a very large, newly collected sample of lottery players in Sweden, drawing from a sample of several million lottery players who appear fairly representative of the national population. In our main analysis, we follow the lottery players for 5-10 years after winning a lottery prize. Figure 1 shows the annual earnings reduction to a lump-sum lottery prize in our main sample. Overall, we find broadly similar results to Imbens et al. (2001).

Figure 1. Effect of wealth on individual gross labour earnings

Notes: This figure reports estimates obtained from equation (2) estimated in the pooled lottery sample with gross labor earnings as the dependent variable. A coefficient of 1.00 corresponds to an increase in annual labor earnings of 1 SEK for each 100 SEK won. Each year corresponds to a separate regression and the dashed lines show 95% confidence intervals.

We also provide several additional results that may prove useful for policy.

  • First, we estimate the effect of winning the lottery on both pre-tax and after-tax earnings.

Most of the existing empirical literature has not emphasised this distinction, even though both after-tax and pre-tax earnings responses are relevant. For example, most models of individual labour supply make predictions regarding how after-tax earnings respond to unearned income. By contrast, pre-tax earnings responses are useful for understanding how widespread increases in unearned income affect aggregate economic activity. This distinction is especially important in countries with high tax rates, since taxes create a significant wedge between pre-tax and after-tax earnings. In Sweden, we find that the lottery-induced reduction in taxes and social security contributions combined is larger in magnitude than the estimated reduction in after-tax earnings, highlighting the fiscal costs of the labour supply reductions. On average, each $100 won reduces annual after-tax earnings by about $0.60 while taxes and social security contributions fall by $0.80, implying a total annual productivity loss of $1.40.

  • Second, we are able to measure earnings responses of both the lottery winners themselves as well as their spouses.

We find that spouses also reduce their labour earnings, although by less than the winner. As a result, the overall household labour earnings reduction is larger than the individual response of the winner, suggesting that focusing on just winners understates the aggregate effects of unearned income shocks. 2

  • Lastly, we compare the dynamic response to a lump-sum lottery prize with the dynamic response to ‘instalment prizes’, which pay out a constant amount every month for a pre-determined (and randomly assigned) number of years.

This comparison is useful because any differences between these two types of prizes would suggest important departures from ‘textbook’ economic models and could also indicate the importance of liquidity constraints or myopia. In the Swedish data, however, we find no evidence of significant differences between the responses, which suggests a limited role for liquidity constraints in determining the magnitude of income effects in labour supply.

  • Our preferred estimates suggest that every dollar won in a lottery reduces lifetime after-tax labour earnings of winners by $0.10-$0.20.

The reduction in lifetime earnings is largest for younger winners. The reduction in pre-tax earnings and social security contributions (which is a broader measure that more accurately captures the decline in economic activity) is more than twice as large. Additionally, we find that the overall earnings reduction comes from both intensive margin and extensive margin changes, meaning that winning a lottery causes both of a reduction in work hours as well as a reduction in the probability of being employed.

Lessons for economists and policymakers

We conclude by discussing some broad lessons for tax and transfer programmes. Returning to textbook labour economics and public economics, many policies generate both income and substitution effects. The lottery estimates described above directly inform income effects. Under the assumption that the income and substitution effects are similar in magnitude, income effects can also indirectly inform the magnitude of substitution effects. The similarity of income and substitution effects implies that the long-run labour supply elasticity is small in magnitude, which in turn suggests that permanent changes in tax rates are unlikely to have large long-run effects on aggregate employment and average hours worked. 3

A more subtle lesson from the lottery estimates is that labour supply effects from transitory changes in wages and taxes are also unlikely to be large. Standard analysis of labour supply responses to transitory changes in wages and taxes focuses on the Frisch elasticity, which in standard models is related to the income effect, the substitution effect, and the intertemporal elasticity of substitution (IES). 4 Unfortunately, there no consensus on the appropriate value of intertemporal elasticity of substitution in the literature. For example, Hall (1988) argues for a low value, whereas Gruber (2006) provides empirical evidence suggesting a higher value of the elasticity of substitution. However, none of the values in this wide range — when combined with the estimated income effects from Swedish lottery data — can produce large implied Frisch elasticities. This may provide useful guidance to macroeconomists regarding the role of labour supply shifts in understanding business cycle fluctuations, which often use the Frisch elasticity when calibrating macroeconomic models.

Lastly, the lottery estimates described in this column complement several recent papers that emphasise non-negligible income effects in labour supply for prime-age adults in developed countries. Picchio et al. (2015) estimate income effects using the Dutch State Lottery; Autor et al. (2015) estimate large income effects from disability compensated for US veterans; Gelber et al. (2015) estimate income effects from US Social Security Disability Insurance programme. Although the magnitudes vary across these papers, taken together the results provide fairly clear evidence to policymakers that unconditional cash transfer programmes in developed countries will likely reduce aggregate labour supply by a meaningful amount. 5 Although the ‘textbook’ treatment of income effects in labour supply emphasises that unearned income does not distort the labour-leisure choice, we emphasise that this does not mean that income effects can or should be ignored by policymakers. Our estimates of the income effect can be used to quantify the labour supply effect and thereby to better understand the fiscal consequences of the introduction of universal basic income programmes.


Browning, M (2005), “A working paper from April 1985: Which demand elasticities do we know and which do we need to know for policy analysis?”, Research in Economics 59(4), 293–320.

Cesarini, D, E Lindqvist, M Notowidigdo, and R Östling(2015), “The Effect of Wealth on Individual and Household Labour Supply: Evidence from Swedish Lotteries”, NBER Working Paper No. 21762.

Gruber, J (2006), “A Tax-Based Estimate of the Elasticity of Intertemporal Substitution”, NBER Working Paper No. 11945.

Gruber, J (2013), Public Finance and Public Policy, 4th Edition, New York: Worth Publishers.

Hall, R E (1988), “Intertemporal Substitution in Consumption”, Journal of Political Economy, 96(2), 339-357.

Imbens, G W, Rubin, D B and Sacerdote, B I (2001), “Estimating the Effect of Unearned Income on Labour Earnings, Savings, and Consumption: Evidence from a Survey of Lottery Players”, American Economic Review 91(4), 778–794.

Keane, M (2011), “Labour Supply and Taxes: A Survey”, Journal of Economic Literature 49(4), 961–1075.

Kimball, M S and Shapiro, M D (2008), “Labour Supply: Are the Income and Substitution Effects Both Large or Both Small?”, NBER Working Paper No 14208.

Pencavel, J(1986), “Labour Supply of Men: A Survey”, in O Ashenfelter and R Layard., eds, Handbook of Labour Economics, Vol. 1, North-Holland, pp. 3–102.

Picchio, M, S Suetens, and J C Van Ours (2015), “Labour Supply Effects of Winning a Lottery”, CEPR Discussion Paper No. 10929.

Ziliak, J P and Kniesner, T J (1999), “Estimating Life Cycle Labour Supply Tax Effects”, Journal of Political Economy 107(2), 326–359.


1 Following Pencavel (1986), the income effect in labour supply is also called the marginal propensity to earn out of unearned income.

2 Although not the focus of this column, the findings for spouses also helps determine appropriate models of household labour supply. The fact that winners consistently reduce labour earnings more than their spouses is consistent with some — but not all — of the standard household labour supply models in the literature, as described in more detail in Cesarini et al. (2015).

3 This captures the view of many labour economists according to a recent interview with David Card, who says “the long-run labour supply elasticity is pretty small, and probably negative.”

4 The relationship between the income effect and Frisch elasticity is given by Ziliak and Kniesner (1999) and Browning (2005), and we use this relationship in Cesarini et al. (2015) to calculate plausible magnitudes of Frisch elasticity implied by our estimates

5 Additionally, it is conceivable that universal income transfers can have additional ‘social multiplier’ effects, for example due to changes of work-related norms, which are not incorporated in our estimates of the income effect.

Cash welfare programmes are widely thought to discourage work because unearned income reduces the labour supply even when it does not alter work incentives. This column discusses recent evidence from Swedish lottery players suggesting that this ‘income effect’ is economically significant, but modest in magnitude and surprisingly similar across various demographic groups.

Are lotteries legitimate means of financing public needs?

William D. Oswald, an attorney and regional representative. Lotteries are a form of gambling in which people buy numbered tickets. Winners are determined at random. Absolutely no skill is involved in picking the winning numbers and only a few—sometimes only one individual—will win a prize. The prizes, however, are big because the chances of winning are so small—sometimes no more than 1 in 3.5 million!

Because of the large revenues generated by lotteries, even public lotteries designed to support charities have been used for private greed. In the past, criminals have used lottery money to bribe public officials, to corrupt newspapers, to control banks, and to suppress opposition with lavish payoffs. In the U.S., for example, the last big lottery was so corrupt and such a national disgrace that federal laws were enacted to stop it.

There is now a trend by governments, especially those which find themselves in financial trouble, to legalize this form of gambling as a means of financing certain public purposes. These state-supported lotteries are set up to collect monies for such public purposes as meeting the needs of the elderly or financing education. Colorado’s lottery profits go to state parks, Pennsylvania’s go to senior citizens, and Arizona’s go to transportation. Lotteries have now been made legal in twenty-two states in the United States, and proposals have been introduced in Congress for a national lottery, which, advocates claim, could quickly and painlessly raise several billion dollars a year to reduce the growing federal deficit.

With the lottery issue becoming more and more a topic for public debate, Latter-day Saints should ask themselves: Is it now appropriate for government to finance public needs through lotteries, when for most of our history gambling was punished as a vice?

The words of Jesus remind us that before undertaking certain activities, it is appropriate for one to sit down first, and count the cost. (See Luke 14:28.) Before supporting any legislation which would legalize or extend lotteries as a means of financing the needs of the public, Latter-day Saints should address the following questions.

1. What Is the Cost to Society?

The Doctrine and Covenants teaches us “that governments were instituted of God for the benefit of man; and that he holds men accountable for their acts in relation to them, both in making laws and administering them, for the good and safety of society.” (D&C 134:1.)

It doesn’t take much research to show that a state-sponsored lottery doesn’t benefit mankind or protect the good and safety of society; it hurts them. Indeed, the cost of state sponsored lotteries, in moral and human terms, is extremely high.

All gambling, even legalized gambling, is parasitic because it creates no economic goods and no real wealth. At most, it merely takes money away from many and gives it to a few. Lotteries discourage thrift, and the publicity given the winners encourages even more people to throw their money away in pursuit of the illusion of instant wealth. Thinking they can get what they want without working for it, people lose sight of the work ethic and begin looking for the mythical “easy road to wealth.” There’s some indication, in fact, that legalizing gambling increases its illegal forms by stimulating the demand for all types of gambling. 1

Ultimately the taxpayers pick up the real costs of government-sponsored lotteries by paying the price for lost jobs, broken families, and the impoverishment, crime, and violence often brought on by compulsive gambling.

2. What Is the Cost to the State?

Lotteries are an extremely expensive way to raise money. In one state, it is estimated that up to two-thirds of the total lottery sales are taken out before anything goes to finance public needs; thus, only one-third of the dollars of lottery sales are used to meet public needs. 2

Furthermore, study after study indicates that it is the poor who most heavily patronize lotteries. The economically disadvantaged—those who feel trapped in a cycle of unemployment and welfare—are the first to look to lotteries as their one big chance to reverse the fortunes of life. The very people who can least afford to buy tickets are the very ones who purchase most of the tickets. In this sense, lotteries transfer the cost of government services to those who can least afford it.

One of America’s foremost students of gambling concludes: “As a revenue source, the state lottery is one of the most regressive taxes known and imposes by far the heaviest relative burden on those least able to pay.” 3

The landmark United States Government study on gambling entitled “Gambling in America” concurs: “The Lottery is one of the more regressive forms of gambling—that is, people in the low income categories spend proportionately more on it than those in the highest income brackets. The money for the lottery comes most from those who can least afford it, worsening their conditions and making them more dependent on aid from the taxpayers.” 4

How self-defeating for government to help the disadvantaged on the one hand, and then take money from them by exploiting their hopes on the other hand.

One of the ironic features of state lotteries is that if they were subject to consumer protection laws, most would be declared illegal. The Wall Street Journal carried an article not long ago entitled “State Lotteries: The Only Legal Swindle.” After dissecting the misleading nature of lottery advertising, the article ended with this penetrating comment: “It is ironic that today not even the sleaziest moneylender is permitted to do things that state lotteries do as a matter of routine.” 5

Senator Durenberger of Minnesota, chairman of a United States Senate Committee which held hearings on the conduct of state lotteries, recently concluded: “You can’t run a successful lottery by telling the whole truth. You need hard sell promotion, often vague and misleading about the odds and the prizes. That enterprise of parting the sucker from his dollar is questionable enough in the free marketplace; it’s no business for a state or federal government whose purpose is to serve and protect the people.” 6

3. What Is the Cost to the Individual?

It should come as no surprise that the growth of lotteries has been accompanied by a sharp rise in the number of compulsive gamblers. 7 Like any activity that produces hopes and anticipation, gambling can become addictive.

In 1979, Johns Hopkins University became the first major medical center in the United States to establish a Compulsive Gambling Counseling Center. Psychiatrists report an increasing number of pathological gamblers who are seeking medical help for habits that have gotten out of control. “Gambling,” says one person who counsels compulsive gamblers, “is an equal opportunity destroyer.” 8

There are no verifiable figures on the total number of compulsive gamblers in the U.S., but all reliable estimates suggest that at least two and maybe as high as eight million people in the United States are addicted to gambling. It cuts across all lines, and experts say it’s getting worse.

Lotteries only exacerbate the problem. The president of the National Foundation on the Study and Treatment of Pathological Gambling says that lotteries may serve to introduce gambling to those who otherwise would shun it. “People who have never bet before, seeing a state-run lottery with the imprimatur of government upon it, might buy a ticket,” he says. “Buying the first lottery ticket might be compared to a future drug addict taking his first puff on a cigarette. It’s a starting point.” 9

4. What Have the Prophets Said about the Cost of Gambling?

Latter-day prophets such as President Brigham Young and President Lorenzo Snow have spoken out against gambling. 10 President Joseph F. Smith raised his voice against the evils of gambling, 11 and in 1925, President Heber J. Grant and his counselors warned the Saints against this vice when they said, “The Church has been and now is unalterably opposed to gambling in any form.” 12

As recently as October 1985, President Gordon B. Hinckley reaffirmed the position of the Church against lotteries when he said, “There can be no question about the moral ramification of this practice. A lottery is a form of gambling.” 13

Ultimately, the decision about a lottery is a question of moral values and involves decisions regarding the kind of religious, social, and cultural environment in which we want to live.

If we look at people through the eyes of Jesus of Nazareth, we must see our fellow citizens not as potential lottery customers to be exploited, but as children of a common father whose “worth … is great in the sight of God.” (D&C 18:10.)

Jesus loved the individual. His teachings are founded upon the principle that each individual has merit in the eyes of God and should be treated with dignity and compassion. Jesus had compassion on the poor and needy. He healed their wounds and restored the fallen. He taught that human life is to be used to further God’s work, which is “to bring to pass the immortality and eternal life of man.” (Moses 1:39.)

Gambling degrades human dignity and saps moral strength as it promotes a philosophy of getting something for nothing. Gambling takes more out of society than it puts in, impoverishing the many and enriching the few.

Lotteries have no place in an enlightened society. It is morally wrong for a state or a nation to exploit the weaknesses of its citizens through sponsorship of lotteries. On the contrary, government should act to restrict gambling, not encourage or sponsor it. That some governments now promote what they once enforced as being unlawful or illegal is symptomatic of a syndrome of greed and covetousness and reflects a deterioration of public and political morality. It is time for members of the Church to voice their concern to legislators and government leaders to say that the whole lottery scheme is so gross in its human toll that public policy requires that lotteries be prohibited.

Are lotteries legitimate means of financing public needs? ]]>