An empirical evaluation of the 1986 Semiconductor Trade Arrangement

In the paper, we aim to empirically evaluate the extent to which the 1986 Semiconductor Trade Arrangement (STA) affected the Japanese producers' behavior in the market for the 1 M Dynamic Random Access Memory (DRAM) chips. Based on the Euler equation of non-Japanese firms' learning-curve optimization, we estimate the marginal cost function and then calculate Japanese price-cost margins by using nonparametric estimation. Our estimation results show that the price-cost margins of Japanese firms were set far above the learning-curve optimization margins but still lower than the marginal costs on average. Hence we can infer that the STA could not push the price above the average Japanese marginal cost but facilitate the Japanese firms to collude to set the price-cost margins by taking account of learning-by-doing effects. This empirical finding and no significant spillover effect are combined with other observations to imply that the STA might have enabled later-entering non-Japanese firms to rapidly expand their market shares and get a foot in the door of the R&D competition for new generational DRAM.