We study the dynamics of an innovative industry when agents learn about its strength, i.e., the likelihood that it gets hit by negative shocks. Managers can exert risk-prevention effort to mitigate the consequences of such shocks. As time goes by, if no shock occurs, confidence improves. This attracts managers to the innovative sector. But, when confidence becomes high, less managers exerting low risk-prevention effort also enter. This accelerates the growth of the industry, while inducing a decline in risk-prevention. The longer the boom, the stronger the confidence, the larger the losses if a shock occurs. While the above dynamics arise in the first best, with asymmetric information there is excessive entry of inefficient managers, earning informational rents at the expense of efficient managers. This inflates the innovative sector and increases its vulnerability.

We model asset management as a continuum between active and passive: managers can deviate from benchmark indices to exploit noise trader–induced distortions, but agency frictions constrain these deviations. Because constraints force managers to buy assets that they underweight when these assets appreciate, overvalued assets have high volatility, and the risk-return relationship becomes inverted. Distortions are more severe for overvalued assets than for undervalued ones because trading against the former entails more risk and tighter constraints. We provide empirical evidence supporting our model’s main mechanisms. Using the data, we infer the constraints’ tightness and compute a measure of effective arbitrage capital.