| There exists a substantial literature on how firms should manage inventory and capacity, but less is known about what firms actually do. In practice, there is considerable heterogeneity in inventory holdings, even across firms within a single industry. The main objective of this dissertation is to use field data on actual decisions to measure which are the main drivers of inventory and capacity decisions.; This dissertation is composed of three essays. The first essay develops a methodology to estimate the underlying costs of a Newsvendor model based on actual decisions. We present a structural estimation framework that captures the effect of uncertainty and mismatch costs on the decisions made by the newsvendor. This econometric model is applied to a hospital that balances the costs of reserving too much vs. too little operating room capacity. The results reveal that the hospital places more emphasis on the tangible costs of having idle capacity than on the costs of schedule overrun. Over-confidence and incentive conflicts are important drivers of the frequency of schedule overruns.; The second and third essays analyze finished-goods inventory holdings in the U.S. automobile industry. The second essay uses panel data to measure the effect of several factors on aggregate inventory holdings of new vehicles. Four of these factors---product variety, demand variability, production flexibility and sales volume---explain a large fraction of the differences in inventory across manufacturers.; The third essay estimates the effect of local market conditions on inventory holdings of automobile dealerships. We collected (via a web-crawler) an original dataset on inventory and sales of auto dealerships of a large manufacturer. Using cross-sectional variation of dealers in geographically isolated markets, we estimate the effect of market structure (number and type of competitors) and sales on inventory levels. The results suggest a strong positive non-linear effect of the number of rivals on the buffer stock carried by retailers, which increases inventory. Counterfactual experiments indicate that reducing the dealership network of this manufacturer (thereby reducing competition) could reduce the remaining dealers' days-of-supply from 14% to 27%. |
