Mix-and-match compatibility and asymmetric costs
This thesis analyzes how the ability of consumers to buy components of a system from different firms affects prices, profit margins, RandD effort, and welfare. It also examines firms' incentives to make their products compatible, that is, to allow consumers to mix-and-match different brands of components into systems. Chapter 1 reviews the economic literature on product compatibility with motivating material drawn from the personal computer industry. Three strands of the literature study compatibility using definitions based on the ability of consumers to mix- and-match components, to capture externalities arising from networks, and to switch brands costlessly. The mix-and-match literature has found that compatibility raises prices compared with those under incompatibility in a variety of settings. In practice, however, compatible computers appear to be less expensive than incompatible computers, and computer buyers have promoted standardization. Chapter 2 develops models of mix-and-match compatibility which make predictions that are the opposite of the literature's. If many Bertrand competitors draw their component costs, qualities, or characteristics from independent random distributions, then expected prices and profit margins are lower under compatibility than under incompatibility, while expected consumer surplus is higher. In addition, the chapter examines the incentives of firms to form coalitions around competing standards. It is found that a subset of firms may become compatible with each other to attract customers away from other firms, creating excess incentives for firms to become compatible from the perspective of industry profits. However, compatibility raises welfare if it is costless and components are homogeneous, because incompatibility is a restriction on the technology for combining components into systems. Chapter 3 shows that shifting from incompatibility to compatibility has an ambiguous impact on RandD effort to reduce costs. In an industry with sufficiently many firms that faces elastic demand, compatibility lowers prices and raises output, and therefore leads to greater RandD incentives. If effort lowers costs without changing the shape of the cost distribution function, compatibility induces firms to choose RandD effort levels that are closer together than under incompatibility. Chapter 4 relaxes the assumption that consumers combine components in fixed proportions. With variable coefficients, compatibility does not necessarily raise the profits of duopolists. For instance, compatibility prevents a dominant firm from setting the price of either component above its competitor's cost. On the other hand, when two "mirror-image" firms each have the lowest cost in one component and demand is symmetric across components, the firms prefer compatibility, as they did in the fixed coefficients case. When sufficiently many firms draw their costs from discrete random distributions, this ambiguity disappears, and expected profits are higher under incom- patibility. Variable coefficients also allow analysis of quantity competition, by eliminating the problem of unmatched components when there are asymmetric quantity choices. In this case, firms with mirror-image costs prefer compatibility to incompatibility because they can specialize in their low-cost component. However, when each firm has the same cost across components, firms are indifferent between the two regimes.