The dynamics of firm profitability, growth, and exit
This thesis analyses the dynamics of firm profitability, growth, and exit across different industries. Chapter 1 documents a striking empirical regularity in the joint distribution of firm profitability and firm size which varies systematically across industries: In industries with a high intensity of R&D investments, there is a strong, systematic "negative tail" of small loss-making firms in the profits-size distribution, whereas this "negative tail" is much less pronounced in industries with low R&D intensity. The chapter also proposes a simple reduced form dynamic model which explains the main empirical features by combining two key mechanisms: a real option effect at the business level and a diversification effect at the firm level. The second part of the thesis takes a structural approach. Its focus is on estimating the dynamic evolution of firm productivity which is an unobserved state variable in an underlying structural model. In this model, firms make exit decisions and investment decisions in physical capital and in R&D. Chapter 2 extends the model in Olley & Pakes (1996) to include R&D decisions that stochastically affect future productivity realisations and proves that their invertibility approach still applies. It estimates the distribution of future productivity conditional on current productivity and R&D investments, which is the key stochastic primitive in theoretical models of firm dynamics. Chapter 3 introduces knowledge capital as a second unobserved state variable into the model and extends the invertibility idea and the estimation strategy to the case of two unobserved state variables. Knowledge allows for lagged effects of R&D on productivity while simultaneously accounting for the stochastic nature of R&D. This reconciles the knowledge capital view in the tradition of Griliches (e.g. 1998) with the stochastic approach in the recent literature on firm dynamics.