Sports injuries are a major concern to athletes, medical experts, and policy makers. Scholars in sports medicine and sports management have made multiple proposals to minimize the injury cost of participating in sport, such as prevention, treatment, protective equipment mandate, or changes to a sport’s rules (Finch, 2006). This existing approach tends to view injuries as unfortunate yet largely unavoidable accidents.
In this work, we investigate whether some sports injuries occur because athletes take risk in the pursuit of the pinnacle of success. According to the risk-escalation hypothesis, injuries should be more prevalent as competition stakes increase because athletes are willing to bear additional risk. Demonstrating that game stakes influence injury prevalence opens new horizons to frame injury policy in term of incentive design, under the general premise that it is possible to influence risk-taking choices (Cisyk & Courty, 2023a, 2023b).
We investigate whether injury costs depend on game stakes by using the incidence of concussions sustained during National Football League (NFL) regular-season games. Because concussions have a high potential of lifetime consequences for their victims, they are of particular concern in most sports of the contact or collision varieties, as best exemplified by American football (DeHaven & Lintner, 1986).1 Sports organizations have implemented policies to increase player safety (Deubert et al., 2017), legislators have passed laws to standardize concussion detection and diagnosis (Rotolo & Lengefeld, 2020), and players have initiated legal battles that threaten the financial viability of some professional sports (Bull, 2021).
Despite being the world’s most valuable sports league with massive investments in injury detection, prevention, and treatment, the NFL still grapples with concussions. Moreover, the NFL offers rigorous reporting of injury events (at least relative to other sports) subject to strict monitoring and public-disclosure rules, and a mature literature studying confounding factors. Finally, the NFL offers an exemplary data laboratory: each of the league’s 32 teams play 16 games during the regular season, some with high stakes, and others not.