Although reformers often claim Japanese firms appoint inefficiently few outside directors, the logic of market competition suggests otherwise. Given the competitive product, service, and capital markets in Japan, the firms that survive should disproportionately be firms that tend to appoint boards approaching their firm-specifically optimal structure. The resulting debate thus suggests a test: if Japanese firms do maintain suboptimal numbers of outsiders, then those with more outsiders should outperform those with fewer; if market constraints instead drive them toward their firm-specific optimum, then firm characteristics may determine board structure, but firm performance should show no observable relation to that structure. To explore the issue, we assemble data on the 1,000 largest exchange-listed Japanese firms from 1986-94. We first ask which firms tend to appoint which outsiders to their boards, and find the appointments decidedly non-random. Firms appoint directors from the banking industry when they borrow heavily, when they have fewer mortgageable assets, or when they are themselves in the service and finance industry. They appoint retired government bureaucrats when they are in construction and sell a large fraction of their output to government agencies. And they appoint other retired business executives when they have a dominant parent corporation or when they are in the construction industry and sell heavily to the private sector. Coupling OLS regressions with two-stage estimates on a subset of the data, we then ask whether the firms with more outside directors outperform those with fewer. They do not. Obviously, one cannot draw solid conclusions from the absence of significant results. Yet the regressions do suggest-exactly as the logic of market competition predicts-that board composition is endogenous: firms choose boards appropriate to them. Given that the composition does not change from the thriving 1980s to the depressed 1990s, optimal board structure also seems not (at least by the observable indices) to depend on the macroeconomic environment.