Ege Y. Ercan

I develop an equilibrium theory of asset management firms as a nexus of relationships between investors and managers. Investors and managers can meet directly. But when either side departs, the remaining side must search again. In a firm, the remaining capital and managers can instead be reassigned to each other without new search. Anticipating such departures, investors and managers come together inside firms to reduce their need for future external search. This dynamic consideration makes firm value exhibit increasing returns to relationships. Estimated on U.S. data on firms' entry, size, and growth, the model attributes 58% of the value of asset management firms to reduced dependence on future external search.

Growth-oriented firms borrow more, innovate more, and default more. We build a Schumpeterian growth model in which firms differ in size, leverage, and growth potential, and banks set firm-specific interest rates and debt limits, calibrated to U.S. supervisory loan-level data and Compustat. High growth-potential firms innovate roughly four times as intensively as low-potential firms and borrow more to finance this investment. Within type, more debt lowers innovation. The positive leverage-growth correlation therefore reflects selection, not a causal effect of debt. A temporary credit-cost shock triggers a default wave concentrated among levered high-potential firms and leaves productivity persistently below its no-shock path, and bankruptcy reallocation and survivor selection drive only a partial recovery. Temporary interest subsidies cushion the short-run loss, but the marginal firms they save are low-potential: a moderate subsidy improves the productivity path at every horizon, whereas an aggressive one, by stopping the default wave that fuels reallocation, eventually leaves the economy worse off than no intervention.

Using a novel dataset of U.S. buyout and VC investments, we study the informativeness of fund managers' interim valuation reports about portfolio companies on final outcomes. We find that when investors assess the performance of individual portfolio companies, they can do better than just relying on the most recent reported valuation. The history of reported valuations is also informative. Investments with greater past staleness or more frequent markdowns tend to perform more poorly in the future than other investments. We also find that the timing of an investment's exit is predictable. Moreover, the combined knowledge of interim realized and unrealized returns, past staleness, and past markdown frequency can help predict whether an investment will end up in the left or right tail of all investments. These predictions are informative as early as the first year of the investment.