This empirical paper investigates the relationship between the dynamic strategic interactions among competitors in a component market and demand factors in the market for the end product. The structure of competition in the US microprocessor (MPU) industry is analyzed using data on prices and sales in both the MPU market as well as the market for personal computers. The pattern of dynamic strategic interaction between competing firms in this market on a key decision variable, price is studied. Non-nested model comparison tests based on equilibrium solutions derived for specific differential games are applied to identify the mode of competitive strategy between pairs of competing brands. The empirical fit to the longitudinal and cross-sectional data, of alternative models of competition, independent (Bertrand–Nash), Stackelberg leader–follower, and Collusion, is used to determine which dynamic model best describes actual competitive behavior over the life of each MPU. Demand for the product market which is downstream from microprocessors, that for personal computers, is estimated using a generalized diffusion model with price effects. Data from the markets for desktop and laptop computers are analyzed at the level of computer vendor and internal microprocessor. Patterns are uncovered, linking downstream demand parameters with upstream competitive strategy. There is evidence to suggest that when there are strong diffusion effects driving sales of both the competing computer brands, there is a higher likelihood of Bertrand–Nash competition among MPU firms. However, when there are higher cross-price effects (substitutability) among personal computer brands there is a greater chance of Stackelberg leader–follower price competition. When self-price effects are relatively high, the likelihood of Bertrand–Nash competition among MPU firms increases. Furthermore, when the potential demand for the computer product category is high, there is a higher likelihood of Bertrand–Nash pricing in the MPU market.
Read full abstract