[Excerpt] 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, though there is no statistically significant difference between the predicted gains from autonomous versus non-autonomous 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, 2) the benefits of autonomous teams relative to those of non-autonomous teams are inflated.