| This thesis develops and estimates a structural model of production and inventory behavior for manufacturing plants in two Colombian industries, metal products and plastic products, from 1977 to 1991. Plant level decisions on production, inventories, sales, entry, and exit are developed using an equilibrium industry evolution model. Estimates of the structural parameters are used to quantify the effect of a plant productivity realization on its production smoothing behavior. In contrast to previous inventory studies, which usually investigate a collection of producers from various industries, this paper characterizes how the full set of dynamic decisions by each plant in a single industry interacts with its own productivity change and the industry states. This allows the model to provide a detailed set of pathways connecting inventory behavior, idiosyncratic productivity shocks, the aggregate demand process, and industry turnover patterns observed in the data.;The empirical model is estimated in two steps. In the first step, we recover some of the parameters of the demand process including: the root and the variance of the aggregate demand process. In the second step, a simulated method of moments estimator is used to estimate the storage unit cost of inventory, the fixed operation cost, the magnitude of entry cost, and the productivity process parameters.;Three counterfactual experiments are implemented. We increase the persistence level of the idiosyncratic productivity process of the plastic products sector. The results show that there is an increase in the fraction of firms that smooth production. The other two experiments are related to the aggregate demand process. The results show that the aggregate demand process does not strongly influence the firm's production smoothing behavior. |
