Product diversification of U.S. manufacturing firms: A theoretical and empirical investigation
MetadataShow full item record
This dissertation consists of two studies. One is about the level of product diversification of multiproduct firms. The other is about the product switching decision of multiproduct firms. Most of the existing studies rely on economies of scope or strategic considerations to explain why firms produce multiple products. However, only a few explain why the level of product diversification varies across multiproduct firms. I extend the Dixit-Stiglitz monopolistic competitive model to analyze product diversification of heterogeneous multiproduct firms, focusing on the role of firm-specific productivity. The model shows that the firm's level of product diversification is determined by and positively related to its productivity at the equilibrium. Using data from the Thomas Register of American Manufacturers and Compustat, I test the prediction of the theoretical model. To overcome simultaneity bias associated with productivity, I implement a IV estimation where I use the number of product varieties offered by the upstream industries of a firm as the instrumental variable to identify exogenous changes in the firm's productivity. This IV strategy is motivated by Romer (1990) who argues that technological changes can take the form of increases in the variety of intermediate inputs. I find strong and robust empirical evidence consistent with the prediction of the theoretical model—firm productivity is a causal factor for product diversification. To understand product switching decision, I further expand the Dixit-Stiglitz model to allow for product-specific shocks. In this model, firm-specific productivity determines the range of products, and product-specific relative consumer's preference determines the composition of products offered by the firm. This shows that, in the absence of productivity shocks, product-specific shocks induce product switching—adding the product experienced a positive demand shock and dropping the product undergone a negative shock from its original product portfolio. Assuming that a product experiences positive shocks before the market shakeout and negative shocks after shakeout as it goes through its life cycle, this result suggests that the incidence of product switching is positively correlated with the proportion of the firm's products that are in the early (growing) stage of their life cycles. This prediction is confirmed in the empirical estimation. In addition to product-specific demand shocks, I also find that larger and older firms as well as firms facing greater market competition are more likely to switch products.
Showing items related by title, author, creator and subject.