I introduce new measures of household uncertainty from the Michigan Survey of Consumers and new methods of identifying uncertainty and first-moment confidence shocks from qualitative survey data using sign restrictions on impulse responses. Household uncertainty shocks explain almost none of the fluctuations in real activity in the U.S. Moreover, the endogenous response of household uncertainty is unimportant in the transmission of other macroeconomic shocks. I conclude that household uncertainty plays essentially no role in U.S. business cycles. By contrast, I find evidence of larger roles for firm uncertainty and household confidence.