Three essays on capital adjustment, reallocation and aggregate productivity

This dissertation consists of three chapters. Chapter one estimates the capital adjustment costs at the plant level in a model entry and exit. We find that the estimated variance of plant-specific productivity shock is larger than obtained from balanced panel estimation. Estimation using the unbalanced panel generates a larger irreversibility cost, a smaller disruption cost, and a smaller convex cost, all compared with the estimates by Cooper and Haltiwanger (2006). In chapter two, we study how much of the aggregate productivity changes can be accounted for by the capital reallocation. We also study the impact of capital reallocation on the productivity dispersion across firms. We find that the capital reallocations accounts for roughly 12 percent of the labor productivity and capital productivity are reduced as the reallocation activity increases. When the economy-wide technology has a positive change, the reallocation increases temporarily then drops to its original level. After a short transition, the economy settles down with an increased labor productivity. Chapter three further studies the quantitative role of allocation, entry and exit in the growth of aggregate productivity. We find that, without including in the model the forces that drive the entry and exit changes, the model economy has a modest increase in the aggregate productivity as a result of decrease in the fixed reallocation cost.