Psychology and behaviour
Economists have a serious model in which buying a ticket is rational. It works — at small, fixed spend. The honest bound is worth stating precisely.
The standard objection to lottery play is that expected value is negative, so buying is irrational. That objection is weaker than it sounds, and it is worth steelmanning the other side properly before drawing any conclusion.
John Conlisk, The Utility of Gambling, Journal of Risk and Uncertainty, volume 6 (1993), pages 255–275, sets out the formal version. Conlisk shows that adding a very small utility term for the act of gambling itself to an otherwise conventional expected-utility framework preserves the standard results almost everywhere, while explaining why people take small unfair bets. You do not need to abandon rational choice to explain lottery play. You need to allow that the gamble is partly a consumed good.
David Forrest, Robert Simmons and Neil Chesters, Buying a Dream, Economic Inquiry, volume 40, issue 3 (2002), pages 485–496, argued the same thing empirically for lotto specifically: demand responds to the jackpot beyond what effective price alone explains, which is what you would expect if part of what is purchased is the size of the dream rather than the expected return.
Put plainly: from Wednesday to Saturday you own a live claim on an enormous sum. That claim is worth almost nothing in expectation and something real in experience. You think about it in the queue, you talk about it at work, you run the fantasy. If you would pay for a film that made you feel something for two hours, it is not obviously incoherent to pay for three days of a specific, personal daydream.
That is the strongest form of the argument, and it is a genuine one.
Suppose a $2 ticket in a game returning about 50% to players — the typical figure for major draw games, compared game by game in return to player by game.
| Item | Value |
|---|---|
| Ticket price | $2.00 |
| Expected return | about $1.00 |
| Net expected cost | about $1.00 per draw |
So the price of the experience is roughly one dollar, not two. The rest is a lottery ticket's expected value, returned to the player population.
One dollar for three days of anticipation is, on its face, cheap. Cheaper than a coffee, much cheaper than a cinema ticket, and it lasts longer than either.
If the analysis stopped there, the case would be closed in favour. It does not stop there.
The entertainment framing has one hard requirement: the spend must be small and it must not grow. That requirement fails in three specific and well-documented ways.
1. It scales with the jackpot. The person spending $2 on an ordinary draw is often spending $20 or $50 on a record rollover. But the anticipation is not fifty times better, and the sharing penalty means the expected receipt does not scale with the headline either (why jackpot size drives sales). If the entertainment is the product, the correct response to a bigger jackpot is to buy exactly the same amount.
2. It scales with near misses. Matching three numbers reliably increases the urge to play again, for reasons covered in near-miss effects. Entertainment spending does not normally respond to the outcome of the last unit consumed. When it does, it is no longer being priced as entertainment.
3. It scales with financial stress. Melissa S. Kearney, State Lotteries and Consumer Behavior, Journal of Public Economics, volume 89, issues 11–12 (2005), pages 2269–2299, found that household lottery spending is financed primarily by reducing non-gambling expenditure rather than by substituting away from other gambling — the introduction of a state lottery was associated with an average decline of about $46 per month, or 2.4%, in household non-gambling spend, with larger proportional reductions among low-income households.
That third finding is the sharpest challenge to the entertainment story. Entertainment spending normally displaces other entertainment. Lottery spending displaces the rest of the budget.
Blalock, Just and Simon, Hitting the Jackpot or Hitting the Skids, American Journal of Economics and Sociology, volume 66, issue 3 (2007), pages 545–570, tested the substitution story directly using sales data from 39 states over 10 years. They found a strong positive relationship between lottery sales and poverty rates, and no such relationship for cinema ticket sales — another inexpensive entertainment good. If lotto were simply cheap entertainment, the two should behave alike. They do not.
The second bound is arithmetic rather than behavioural. Small, fixed and regular is still cumulative.
Two draws a week at $2 is $208 a year. Invested instead at a 5% real return:
| Weekly habit | Annual outlay | Value after 40 years at 5% real |
|---|---|---|
| $2 per draw, 2 draws/week | $208 | about $25,100 |
| $50 per draw, 2 draws/week | $5,200 | about $628,200 |
The first row is a real cost and a defensible one — $25,000 over a working life is a meaningful sum, but so is forty years of a pleasant weekly ritual, and reasonable people can price that trade differently.
The second row is not entertainment spending under any framing. It is the difference between retiring and not.
Run your own numbers with the lifetime spend calculator and the invest instead calculator; the fuller comparison is in investing instead of buying tickets.
The entertainment case is valid within a narrow band:
The test is not how much you spend. It is whether the amount moves. A fixed line every week is a subscription to a daydream. A line that grows with the headline is something else, and it is worth knowing which one you have.
Setting the number in advance is the entire content of setting a lottery budget. If you want the rare cases where the expected value genuinely turns positive, they are in positive-EV lotteries — and they are rarer, and more constrained, than the internet suggests.
Last verified: 2026-08-29