This paper studies how biases in managerial beliefs affect managerial decisions, firm performance, and the macroeconomy. Using a new survey of US managers I establish three facts. (1) Managers are not over-optimistic: sales growth forecasts on average do not exceed realizations. (2) Managers are overprecise (overconfident): they underestimate future sales growth volatility. (3) Managers overextrapolate: their forecasts are too optimistic after positive shocks and too pessimistic after negative shocks. To quantify the implications of these facts, I estimate a dynamic general equilibrium model in which managers of heterogeneous firms use a subjective beliefs process to make forward-looking hiring decisions. Overprecision and overextrapolation lead managers to overreact to firm-level shocks and overspend on adjustment costs, destroying 2.1 percent of the typical firm’s value. Pervasive overreaction leads to excess volatility and reallocation, lowering consumer welfare by 0.5 to 2.3 percent relative to the rational expectations equilibrium. These findings suggest overreaction may amplify asset-price and business cycle fluctuations.