Is the Lottery a Tax on the Poor? What the Evidence Actually Says
It is the oldest accusation aimed at state lotteries: that they are a tax on the poor, and a voluntary one only in the narrowest sense. It usually arrives as a slogan, which is a shame, because underneath it sits forty years of genuine research - more specific and more interesting than the phrase suggests. Answering it properly means getting one definition right, looking at what economists actually measured, taking the counterarguments seriously, and then confronting a quieter problem that has nothing to do with who buys the tickets and everything to do with where the money ends up.
The Definition Almost Everything Hangs On
Before you can answer the question you have to know what 'regressive' means. In economics, a levy is regressive when it takes a rising share of income as income falls. What matters is the share - not the absolute amount.
This is exactly where most discussions go wrong. If a household on €20,000 a year and one on €200,000 both spend €300 a year on tickets, the spending looks identical. As a share of income, the burden on the first household is ten times heavier. Same euros, very unequal load - which is why any statistic without a denominator is worthless in this argument.
What Was Actually Measured
The book that started the argument
The serious version of this claim traces back to Selling Hope: State Lotteries in America, published by economists Charles Clotfelter and Philip Cook at Harvard University Press in 1989. It was the first systematic study of who actually buys tickets, and it set the terms nearly every argument since has borrowed - usually without the caveats the authors attached.
Roughly half of every stake is take-out
Most big lotteries return somewhere near half of stakes as prizes. The other half - covering the state's share, good causes, operating costs and retailer commission - functions as an implicit tax on the amount wagered. Whatever you call it, a take-out of that size dwarfs the rate on almost anything else people buy.
A minority of players carries the revenue
Sales are heavily concentrated: study after study finds that a relatively small group of frequent, high-spending players accounts for a large majority of turnover, while most participants buy occasionally and spend very little. That concentration is what makes averages so misleading here - the 'typical' player and the player who funds the system are not the same person.
The core finding has been remarkably stable across decades: ticket spending rises with income far more slowly than income itself. The share spent therefore falls as households get wealthier - which is precisely the definition of regressive. We've worked through where that 50 percent take-out comes from separately. Expected value in lotteries
The Counterarguments Worth Taking Seriously
Accepting the evidence above does not deliver a verdict. Four objections survive scrutiny - and each of them moves the question rather than closing it.
- 1It is genuinely voluntary. Nobody is compelled to buy a ticket, and the word 'tax' does real work it hasn't earned when applied to a purchase people choose. Calling it an implicit tax rate describes the arithmetic; calling it a tax describes something the law does not.
- 2Most players spend little. For the large casual majority, a few euros a month buys a few days of anticipation. Treating that spending as an economic harm requires arguing that those people are wrong about what their own money is worth to them.
- 3The counterfactual is not nothing. Before state lotteries, illegal numbers games served the same demand with worse odds, no consumer protection and criminal proceeds. 'No legal lottery' has historically meant 'an unregulated one', not an absence of play.
- 4Regressivity is a property of the funding, not a verdict on the product. Sales taxes on fuel and tobacco are regressive too, and societies still levy them for other reasons. Establishing that lottery funding is regressive is the start of the policy argument, not the end of it.
The Earmarking Illusion
The strongest reply to the regressivity critique is that even if the funding is skewed, the money goes to education, sport, culture, heritage. Lower earners may put in proportionally more, the argument runs, but they share in what comes out.
That argument has a problem which almost never surfaces in public debate: money is fungible. Repeated public-finance studies - especially of US states that explicitly earmarked lottery revenue for education - have found that education budgets did not durably grow by the expected amount. The earmarked money arrived, but general-fund allocations to the same purpose were pulled back at the same time and spent elsewhere.
The result is an accounting illusion: the lottery can report, entirely truthfully, that it has transferred billions to education, while the schools end up with no more money than before. Anyone justifying the regressivity by pointing at good causes therefore has to show first that the budget actually grew - not merely the headline.
What lotteries demonstrably builtWhere This Honestly Lands
The defensible answer is uncomfortable for both camps. Yes: measured as a share of income, lottery funding is regressive, and the implicit take-out is higher than almost any ordinary tax. No: that does not automatically make lotteries exploitative - voluntariness, the small sums most people spend, and the realistic alternative all carry genuine weight.
The sharper criticism therefore isn't aimed at the game at all, but at two edges: where it is aggressively marketed, and whether the good-causes promise is honoured in the accounts at all. Those are questions for regulators and budget-setters - and they can be answered without taking anyone's ticket away.
Five Points for the Next Argument
- 1Ask for share of income, not euros spent. Any claim about who the lottery falls hardest on is meaningless without that denominator - and most published claims quietly omit it.
- 2Separate the casual player from the heavy player. Policy aimed at the average participant will miss the minority who generate most of the revenue and carry most of the risk.
- 3Treat 'funds good causes' as a claim to check, not a conclusion. Ask whether the earmarked budget actually grew, or whether general funding quietly moved elsewhere.
- 4Judge advertising separately from the product. Where a game is promoted matters independently of whether the game itself is fairly run, and it is the part regulators can change fastest.
- 5Know your own numbers. Whatever the macro debate concludes, the household question is simpler: what share of your income goes in, and would you still be comfortable with that figure written down?
The One Number You Control
You don't get to settle the macro debate. Your own ledger is another matter: run a fixed set of numbers through years of real draws - across every major lottery we cover - and see soberly what would have come back.
Disclaimer: This article summarizes publicly available research and debate on the distributional effects of lotteries. Findings come largely from US and UK studies and do not transfer unexamined to every country; payout ratios and the use of proceeds differ by operator. Lottery remains a game of chance. Play responsibly. LottoROI is not affiliated with any lottery operator.