Heterogeneity and firm growth in the pharmaceutical industry

In this chapter we investigate some properties of the patterns of firms’ growth in the pharmaceutical industry. As a preliminary step towards a more comprehensive analysis of the determinants of firms’ growth, we focus on a smaller – but crucial - set of issues that are suggested by the peculiar nature of the process of innovation and competition in this industry. In particular, innovation in this industry has often been described and conceptualized as a pure “lottery model”, where previous innovation (in a particular submarket) does not influence in any way current and future innovation in the same or in other submarkets (Sutton, 1999). To the extent that firms’ growth is driven by erratic innovation, it should also be erratic. Moreover, the pharmaceutical industry is composed by a small and persistent core of large innovative corporations and by a vast fringe of smaller companies that conduct incremental and imitative R&D and manufacture and sell drugs on local markets.
Against this background, we test if firms’growth is erratic, i.e. we test Gibrat Law, using a Bayesian estimation approach that – differently from previous studies – allows us to control for firm’s heterogeneity both in the intercept and in the slope of the underlying statistical model and nests previous approaches.
We find that: (i) there seems to be strong evidence against the Gibrat’s law on average (ii) estimated steady states differ across units, and firm sizes and growth rates do not converge within the same industry to a common limiting distribution; (iii) there is only weak evidence of mean reversion, i.e. initial larger firms do not grow relatively slower than smaller firms. Rather, differences in growth rates and in size steady state are firm-specific, rather than size-specific;
(iv) differences in growth rates do not seems to disappear over time.
These results suggest that firms’ growth in pharmaceutical is linked to systematic, firm-specific factors among a heterogeneous population of companies. More generally, these findings suggest that previous, standard tests of Gibrat Law are probably incorrect because they are based on models do not adequately control for potential heterogeneity.