What do mean impacts miss? Distributional effects of corporate diversification

Corporate diversification is one of the most debated topics in finance over the past two decades. While it is widely believed that there exists a discount in the stock market valuation of conglomerate firms, the extant research based on least squares methods points to different directions. We argue that the existing empirical analyses ignore some important data features, especially cross sectional heterogeneity, predicted by both theories and casual observations on corporate diversification, and thus cannot provide a complete picture of the diversification discount. Using a quantile regression analysis on U.S. public firms, we investigate the importance of heterogeneity of diversification as well as other firm characteristics. Estimated quantile treatment effects exhibit substantial heterogeneity as predicted. Thus mean impacts miss a great deal. We also tie back differences in the effect of diversification in high-valued and low-valued firms to observable agency characteristics; the most interesting finding is that CEOs seem to play vastly different roles in high-valued and low-valued firms. We conclude that the effect of diversification is likely more varied and more extensive than has been recognized.