Lien-Klein Paper on Relatedness
| Peter Klein |
Lasse and I have a new paper on the measurement of relatedness, the degree to which a diversified firm’s markets or industries are “close” to each other. Relatedness is a key concepts of corporate strategy, but it is difficult to define and measure consistently. We discuss a new, “survivor-based” approach and compare it to conventional measures. The survivor-based approach lets the competitive process and the knowledge of local decision makers replace the judgment of the researcher (or the SIC system) in determining what is related to what, giving it a Hayekian flavor. Specifically, we measure the relatedness between a pair of industries by comparing how often they are actually combined to what one would expect if diversification patterns were random. Industries are related when this difference is large and positive, and they are unrelated if it is negative. This concept was originally suggested by Teece, Rumelt, Dosi and Winter (1994) who used it to illustrate persistent patterns of “coherence” among US firms.
The paper, “Measuring Inter-Industry Relatedness: SIC Distances versus the Survivor Principle,” is available on SSRN. Here’s the abstract:
The conventional approach to measuring inter-industry relatedness uses the SIC system to capture the “distance” between industries. While relatedness measures based on SIC codes (or equivalent classifications) are readily available and easy to compute, they do not screen effectively for the conditions under which related diversification creates value. This paper constructs an alternative, survivor-based measure of inter-industry relatedness and compares it to similar measures based on distances between SIC codes. We find that survivor-based measures consistently outperform SIC-based measures in predicting firms’ decisions to enter new markets, even when herding tendencies and motives related to mutual forbearance are taken into account.