In 1992, Congress passed the Professional and Amateur Sports Protection Act (PASPA), which banned most sports betting throughout the United States. However, New Jersey challenged this law on the premise that Congress was “commandeering,” meaning that the federal government unconstitutionally passed a law that it expected the states to enforce. In 2018, the Supreme Court sided with New Jersey, allowing states to regulate their own gambling markets. Since the Supreme Court’s decision on sports betting, online sports gambling has been legalized in 30 states and Washington, DC.
This liberalization of sports betting has sparked criticism from gambling opponents who argue that the industry is predatory and causes significant financial harm. Charles Fain Lehman’s recent article in The Atlantic consolidates some of the more persuasive arguments against legal sports betting. However, his case for prohibition is at a loss. Lehman argues that sports gambling leads to measurable harm among society’s most vulnerable communities, and the tax revenue it generates is insufficient to offset these damages.
While it’s true that gambling addiction can lead to harm, the industry is too young, and available data are too limited to evaluate the industry’s impact accurately. Additionally, many people enjoy sports betting, and only about 1% of Americans are estimated to have a severe gambling problem. This figure is in comparison to the estimated 85% of Americans who have gambled at least once in their life (an estimated 60% in the past year).
Of the three academic working papers Lehman cites, one is of moderate quality, while the other two rely on convoluted methods that lead to questionable conclusions. In contrast to Lehman’s argument, the tax revenue from sports gambling is relatively substantial compared to its market size and other industries.
Lehman claims that the combined tax receipts of all 38 states amount to just $500 million per quarter—about $2 billion annually—which he describes as “anemic” compared to alcohol, tobacco, or marijuana. Yet, earlier in the article, he notes that sports gambling generates $10 billion annually in total revenue. According to Lehman’s figures, $2 billion in taxes represents 20% of the market, which is a significant rate. It’s unclear why, as a baseline assumption, sports betting should be expected to generate similar revenues to some of the most highly taxed products in the country. One of the primary justifications for taxing cigarettes and alcohol so highly is they generate costs to non-users via crime, car accidents, and healthcare. People who spend their money on sports betting do not impose similar costs on non-bettors.
Critics might argue that if the $10 billion of revenue is concentrated on society’s most vulnerable populations, the costs of legalization could be more significant than tax receipts justify. The most substantial evidence Lehman cites for this case is a working paper by Scott R. Baker, Justin Balthrop, Mark J. Johnson, Jason D. Kotter, and Kevin Pisciotta, which suggests that gambling reduces household savings, particularly among financially constrained households. However, the same could be said for some non-essential household consumer spending. By that logic, we’d need to ban virtually all forms of entertainment and recreation, including tickets for sporting events.
Even if we accept Baker et al.’s concerns at face value, the evidence is difficult to verify. The data are not publicly available, and the study doesn’t include a complete picture of household finances. The authors reviewed a nonrandom selection of individual bank and credit card accounts, leaving out how the total assets of any single household were impacted. A more comprehensive analysis might look at state-level income, spending, and savings trends. However, such a study would still struggle to link directly to gambling households, highlighting the current lack of quality data.
Lehman also references two other studies, both of which have methodological flaws. Brett Hollenbeck, Poet Larsen, and Davide Proserpio find that gambling legalization leads to increases in bankruptcy rates, debt collections, debt consolidation loans, and auto loan delinquencies after gambling legalization. Yet these effects are attributed to a mere 0.3 percent decrease in average credit scores following legalization—a change too benign to be causal.
Kyutaro Matsuzawa and Emily Arnesen, meanwhile, report a rise in intimate partner violence in states with legal gambling when a fan’s home team unexpectedly loses. However, their analysis only compares states with others that had “Sundays without ‘home’ games,” neglecting the potential impact of away games, which could produce similar effects among gamblers and bias the authors’ estimates. Their model focuses on a narrow set of games and fails to account for all gameplay possibilities, raising concerns about potential confounding factors. Because this study looks at population-level data, there also isn’t direct evidence that abusive partners bet on such games. Moreover, “significant” results only mean that the confidence intervals don’t include 0 and should not necessarily be interpreted as “relevant.” Papers for academic journals tend to focus on results that are most statistically significant rather than the most realistic. This is problematic because academics can try many statistical models until one produces confidence intervals that don’t include 0. Publishing significant results is fine, but these studies tend to leave out the results of models that didn’t work, which could undermine their findings.
Lehman’s broader implication by raising studies like Matsuzawa and Arnesen—that emotional triggers like losing a bet cause intimate partner violence—also provocatively misplaces responsibility for such violence on external events rather than on the abuser’s inability to manage emotional responses. Many studies have found a link between sporting events and domestic violence, even when gambling isn’t involved. However, few would argue that this correlation justifies banning sports altogether. As with any statistical model, results can be found that appear to correlate a harmful outcome with a chosen exposure, but correlation does not imply causation. Various factors that lead to adverse outcomes might also correlate with sports betting and could inappropriately imply that sports betting leads to domestic violence, lower credit scores, and reduced saving rates.
The United States lacks quality survey data to assess the full impact of gambling legalization. Lehman suggests that the previous prohibition on state legalization of sports betting under PASPA was relatively harmless. Still, it’s challenging to know what was occurring in the growing illegal gambling markets. Such unchecked markets don’t include consumer protections and could harm problem gamblers more. In a world where internet markets are challenging to control, it is wiser for states to regulate sports betting and benefit from the resulting tax revenue.
Lehman also claims legalization hasn’t shrunk the size of the illegal market, pointing to a survey from Massachusetts showing monthly bettors were just as likely to bet through unauthorized channels after legalization. The problem is that the study Lehman cites was fielded just two months after the launch of retail sports betting and less than three weeks after the launch of online sports, with data collection ending just 18 days later. The author cautioned that “many jurisdictions internationally have found, it can take a substantial period of time for sports bettors to migrate fully from non-regulated to regulated providers.” Contrast this with the American Gaming Association’s research claiming more than 77 percent of online sports bets were placed through legal channels in 2023 compared to 44 percent in 2019, and 78 percent of bettors placed all or most of their bets through legal operators. To say legal sports betting hasn’t shrunk the illegal market is likely incorrect.
States that have legalized sports betting now have the tools to regulate the industry, provide better consumer protections for what was previously outside the law and practically impossible to track, and generate tax revenue that could not exist under prohibition.