Research summary:We replicate one of the first studies on the dynamics of diversification, Chang (1996). Our sample of public and private firms from 1992 to 2001 yields findings similar to the earlier research. Firms tend to enter new markets that have human resource profiles that are similar to the firms' existing businesses, and exit markets that have dissimilar human resource profiles. This general finding is robust to alternative specifications and controls, including the timing of market entry and technological diversification. However, we also find that entry and exit by private firms is much more prevalent than that by public firms. Differences in how public and private firms diversify based on knowledge similarity invite further research in evolutionary economics.Managerial summary: Chang (1996) revealed that large, public manufacturing firms in the 1980s tended to enter new businesses that leveraged the firm's knowledge, some of which is embedded in routines among employees sharing similar occupations. The new businesses provide the basis for further expansion, while a firm divests businesses that do not stay profitable, or which no longer match the firm's new portfolio of businesses. Our data show that these patterns continued in the 1990s. Our sample also includes private firms, which conduct the majority of market entry and exit. Private firms enter new businesses with different proportions of scientific and marketing personnel than their existing operations. Thus, diversification still reflects knowledge resources for smaller, private firms, but differently than for the largest firms in the economy. reductions in firm scope over time. 1 In his landmark paper, he made three main contributions to an evolutionary economics view (Nelson and Winter, 1982) of how diversification develops. First, he considered market entry and exit together, which 1 Starting with Gort's (1962) seminal book, a few other dynamic analyses of diversification predate Chang (1996), primarily focused on explaining firm financial performance, not behavior. In the management literature, Palepu (1985) tested one longitudinal hypothesis about related diversification, Hill and Hansen (1991) examined changes in diversification in the U.S. pharmaceutical industry, and Pennings, Barkema, and Douma (1994) analyzed "expansion projects" (i.e., market entry) in terms of organizational learning. 2325 2 Chang (1996) used the change in ROA and operating cash flows (OCF) over a 14-year period to measure performance. In our view, later research that controlled for riskiness of cash flows and self-selection of diversification strategy greatly improved the specification of models for testing the relationship between diversification and financial performance.
