Strategic choice of leverage and strategic cost of diversification

I examine the interaction between firms' financial and operating decisions, and the effects of this interaction on firms' optimal capital structure choices and on their valuation. In the first chapter, I extend the limited liability literature by studying the relation between firms' optimal leverage choices and the competitive structures of industries in which they operate. A firm's optimal leverage is related to the degree to which its operating strategy affects and is affected by its rivals' strategies. This relation is positive, regardless of whether the competition in output markets is in strategic substitutes or in strategic complements. I test the model's prediction using two proxies for the extent of competitive interaction among firms. The empirical evidence, obtained using 52 years of Compustat data, shows that the competitive structure of an industry is an important determinant of its participants' leverage ratios.; In the second chapter I analyze the behavior and valuation of multi-segment firms, operating divisions in different industries. The model applies the idea of shareholders' limited liability affecting firms' financial and operating choices, and demonstrates that conglomerates can be discounted relative to stand-alone firms, without relying on the usual arguments of ex-ante inferiority of diversifying firms and/or non-profit-maximizing behavior of firms' decision makers. The inability of conglomerates to commit to unconstrained optimal operating strategies, following from the lack of flexibility in choosing their capital structures; reduces their value. The model relates the valuation of diversified firms and the market reactions to announcements of diversifying acquisitions to the variation in the extent of competitive interaction in industries within which conglomerates operate. In addition, the model allows reconciling positive market reactions to announcements of conglomerate acquisitions with the existence of diversification discounts. The model's predictions are tested empirically using both the Compustat Industry Segment database and a sample of conglomerate acquisitions; and are generally supported by the data.