Collateral booms and information destruction

We analyze theoretically and empirically an overlooked aspect of credit booms: their effect on the economy's production of information. We develop a model in which entrepreneurs have productive investment opportunities about which they are privately informed. When internal funds are scarce, entrepreneurial access to financial market requires costly screening, which produces information about the project's quality. A crucial aspect of this information is that it is public and long-lived, making it possible for the project to be traded -- and used for further investment -- in the future. Thus, screening in our economy produces "informational capital", which is combined with physical capital to produce output. An important assumption in our model is that informational capital is slow moving due to diminishing returns to screening. We use this environment to explore the effects of boom and busts in asset prices. In particular, we consider two types of shocks: (i) "sentiment shocks", modeled as the appearance and disappearance of rational bubbles on the price of investment projects, and; (ii) productivity shocks, modeled as an increase in the productive capacity of projects. Our main result is that booms driven by sentiment shocks lead to large crashes and slow recoveries: specifically, the crash of a bubble is followed by an overshooting of economic activity, which then recovers gradually in the transition to the new steady state. The reason for this result is that bubbles provide entrepreneurs with (bubbly) collateral, effectively enabling them to borrow without the need of screening. Although this raises the economy's stock of projects and expands economic activity during the boom, it also depletes the economy's stock of informational capital. In this sense, bubbles and informational capital are substitutes. This depletion of informational capital is not noticeable while asset prices are high, since the bubble sustains borrowing and investment. Once the bubble crashes, however, economic activity contracts for two reasons: the collateral provided by the bubble disappears, directly restricting investment, and the stock of informational capital is low because it was depleted during the boom. The effects of productivity driven booms, however, are very different. Periods of high productivity raise entrepreneurial incentives to invest and, thus, to generate information via screening. In this sense, productivity and informational capital are complements. The end of a productivity boom is therefore characterized by an undershooting of economic activity, since the resulting fall in investment is tempered by the informational capital that was accumulated during the boom. Our model thus suggests that the aftermath of a credit boom depends crucially on the nature of the boom itself.