Bank of America: Bettors Get Back 75 Cents for Every Dollar Deposited

The bank can see what its own customers pay into betting apps and what comes back out, and across seven months the ratio never reached break-even. The number that should worry the industry more is the one about balances: households that bet online hold 59% of the deposit balance of households that do not.

• Bank of America Institute analysis of its own customer accounts found the online betting cash recovery ratio stayed below 1 for the whole period, with inflows averaging less than three-quarters of outflows
• In 2026 the median deposit account balance of households that bet online was 59% of the balance held by households that did not
• First-time online betting users in June and July were more than triple the January level, and Gen Z and Millennials accounted for 88% of online betting activity in July
• Gen Z had the highest recovery ratio of any generation at roughly 82 cents on the dollar, still well short of breaking even
• In a proprietary survey, 20% of respondents said they consider sports betting a form of investment, and across every generation event contracts were more likely to be seen that way than sports betting

The Bank Can See the Money Going In and the Money Coming Out

Bank of America has published an analysis of online betting drawn from its own consumer account data, and the headline finding is blunt. "Our analysis found the online betting cash recovery ratio has remained below 1, with total inflows less than three-quarters of total outflows on average for the duration of the series," wrote Taylor Bowey of the Bank of America Institute. "In other words, customers typically recover less than 75 cents for every dollar transferred to online betting platforms."

The report, titled Online betting platforms: What's the money line?, covers deposits into and withdrawals from betting platforms rather than wagering itself. The bank is explicit about the limitation: it sees only money moving through its own accounts, so winnings that stay on a platform are not captured. It also notes that the analysis covers sports betting and prediction markets, not other internet betting products or land-based casinos.

Within that caveat, the direction is consistent. Over the January to July period the ratio cleared 75 cents in only two months. The bank's conclusion is about income rather than entertainment: the figures "suggest online betting is not currently functioning as a meaningful source of income for most bettors". It is widely estimated that only 3% to 5% of sports bettors are profitable over the long term.

Adoption, meanwhile, is accelerating. The number of online betting users rose sharply over the summer, with first-time users in June and July more than triple January levels. Gen Z and Millennials made up 88% of online betting activity in July.

Those younger cohorts do not fare better. Gen Z recorded the highest recovery ratio of the four generations examined, at around 82 cents on the dollar, and still fell well short. "Despite seemingly recovering more than older generations, total inflows remained substantially below total outflows, suggesting that online betting is not a reliable or constant source of income," Bowey wrote.

The most striking figure is not about returns at all. In 2026, the median deposit account balance of households that participated in online betting was 59% of that held by households that did not.

The bank also ran a proprietary survey on how customers classify the activity. Some 20% of respondents said they consider sports betting a type of investment, though more said it was not. Gen Z was twice as likely as average to regard it as investing. Across all generations, buying event contracts on prediction markets was more likely to be considered an investment than sports betting was.

The 59% Balance Figure Is the One That Travels Beyond the Trade Press

A recovery ratio below 1 is arithmetically unavoidable: the house edge guarantees it, and no operator has ever claimed otherwise. The balance comparison is different in kind. It says households that bet online hold barely half the cash cushion of households that do not, and it comes from a bank rather than a campaign group. That figure does not establish causation, and the obvious alternative reading is that people with less cushion are more likely to bet in the first place. But it will be quoted in every affordability debate from now on, and the industry's answer cannot be that the number is wrong, because it is the bank's own ledger.

Prediction Markets Are Winning the Argument They Should Want to Lose

The survey line that matters most is the one showing event contracts are more likely than sports betting to be classed as investing, in every generation. That is the framing the exchanges have spent two years building, and it is working. It is also the framing that regulators from Michigan to France are citing as the harm: a product sold as a trade, bought as a bet, and consumed disproportionately by the under-35s who make up most of the volume. A judge in Michigan used almost exactly those words last week.

A Flat Handle and a Falling Recovery Ratio Point the Same Way

The American Gaming Association projected flat legal NFL handle at $29.5 billion for the season now starting, and blamed the exchanges. Bank of America's data suggests the customer base is simultaneously getting younger, larger by first-time users and no better at winning. Those two things together describe a market that is not growing revenue from existing customers but recruiting new ones into a product they lose at, which is a durable business and a fragile political position. The tax debate in the United States has so far been about rates. This is the kind of evidence that turns it into a debate about conduct.

Bank of America has not said anything the industry did not know. It has said it with the deposit data of its own customers, which is a different thing entirely.