I estimate the effect of team production on labor productivity and product quality using a cross‐section of British establishments, finding that the typical establishment enjoys statistically significant increases in labor productivity (but not product quality) from using teams, although there is no statistically significant difference between the predicted gains from autonomous versus nonautonomous teams. I show that standard methodological approaches that treat teams and autonomy as exogenous induce biases of two forms: (1) the benefits from teams are inflated, and (2) the benefits of autonomous teams relative to those of nonautonomous teams are inflated.