Industry differences in the persistence of firm-specific returns
研究为何某些企业能长期获得高于竞争对手的收益，发现行业经济结构（如移动壁垒）是导致企业特定租金持续性的关键因素，并检验了不同行业间持续性差异的显著性。
Some firms such as Kellogg, Nestle, Hewlett-Packard, and Boeing have consistently earned significantly higher returns than their immediate competitors. Why do intraindustry rents persist for these firms? Competition typically erodes these rents, as evidenced by the strong convergence of firm profitability to the industry mean, illustrated in Figure 1. I claim that the persistence of these firmspecific rents depends at least partly on industry economic structure (similar to Richard E. Caves and Michael E. Porter's [1977] mobility barriers), so the rents are more persistent in some industries than others. In this paper I test which of the theoretical industry factors are significant. To support my claim I also show that the persistence of abnormal returns differs widely and systematically across industries. For example, an industry with highly persistent profitability differences is the American automobile industry in the 1970's, as illustrated by Figure 2, which graphs differences in returns on assets (ROA's) over time. Throughout the 1970's, General Motors maintained a persistent, significant profitability advantage over Ford, while Ford held a persistent advantage over Chrysler. This situation did change later, but it is still true that, during the 1970's, there was no apparent convergence of profitability differences among the big-three auto makers. A comparison of Figures 1 and 2 suggests that the economy-wide result of Figure 1 is merely an average over very different industries with convergence rates ranging from almost zero (as in automobiles in the 1970's) to very rapid. My methods for estimating persistence of above-average returns improve on previous studies in several ways. First, I only consider the persistence of returns around the industry mean. The persistence of this intra-industry or rent rather than total rent has not been previously investigated. Yet the factors affecting firm-specific rents need not be the same as those affecting industry rents. Second, I estimate persistence in several ways. Finally, I use a wider range of explanatory variables than earlier studies. This study is structured as follows. Section I explains the concept of persistent firmspecific rents. Later sections specify the empirical model and estimate persistence by industry (Section II), theoretically relate industry factors to these varying persistence estimates (Section III), discuss cross-sectional regression estimates (Section IV), and provide conclusions (Section V).