Reviewed by: Japan's Network Economy: Structure, Persistence, and Change Edward J. Lincoln (bio) Japan's Network Economy: Structure, Persistence, and Change. By James R. Lincoln and Michael L. Gerlach. Cambridge University Press, Cambridge, 2004. xx, 409 pages. $80.00. For much of the last 50 years, social scientists have grappled with the question of why the Japanese economy performed so well from roughly 1950 through the 1980s. Economists, political scientists, and sociologists all brought different analytical tools to this question. Economists, for example, [End Page 239] have tended to view the corporation as simply a black box, an institution that converts inputs into outputs in a measurable way, while sociologists have been much more interested in the institutional and human details of both the inner workings of the corporation and the interaction among corporations. Michael Gerlach and James Lincoln are two sociologists at the University of California, Berkeley, who have labored long in this vineyard. This new study looks at a very specific aspect of corporate behavior: the existence and meaning of networks of long-term business relationships among firms, known as keiretsu. This volume is both an analytical expansion and an update of a previous book by Gerlach—Alliance Capital (University of California Press, 1992). While the authors spend some time dealing with vertical keiretsu (the vertical production chains linking finished product producers and their parts suppliers), the bulk of the analysis deals with the horizontal keiretsu. Their focus is mainly on the six broad, informal business groups that have been so visible in the economic landscape in the last half-century—Mitsui, Mitsubishi, Sumitomo, Sanwa, Fuyo, and DKB. While most of these groupings have prewar zaibatsu antecedents, they are not conglomerates in the usual sense of the word, since no lead firm has controlling ownership stakes in the other members of the group, but these groups have been characterized by a distinctive pattern of minority-stake cross-shareholding. Members of a group also use the group bank as their main bank (that is, the bank from which the firm has its largest single loan), and core members belong to presidents clubs that meet once or twice a month. After an introductory chapter and an initial historical chapter that provides a fairly standard review with little new information, the guts of the book come in three lengthy statistical chapters. These rely on a data set covering the 259 largest corporations, including some financial institutions and trading companies as well as manufacturing firms and others, over the period from 1968 to 1997. Chapter three establishes support for the hypothesis that keiretsu ties (both horizontal and vertical) exist but weakened in some respects over the period covered. Chapter four establishes how equity connections and dispatch of directors is connected to trading ties among firms—establishing that there is a real business impact of keiretsu ties. Chapter five asks what the implications of these ties are for corporate performance, lending support to the hypothesis that member firms gain stability at the expense of profitability. That is, firms that belong to keiretsu are less profitable on average than independent firms, but they have less variability in their profits over time, as weak firms are pulled up and leading firms constrained. None of the conclusions reached by this analysis is particularly startling, but they should be reassuring to readers who have believed that keiretsu have had real implications for corporate behavior. [End Page 240] The data set for these chapters is explored in many ways—including analyses over time, by industry, and by individual horizontal keiretsu. Indeed, the statistical analysis is so thorough and technical that those readers without a background in quantitative analysis will find these chapters heavy going. The one major limitation of their data set, which the authors acknowledge, is that many of their variables are binary (1 or 0). For example, company A owns stock in company B, or it does not. This means that a great deal of potential richness in the analysis is lost by not measuring the strength of various linkages. Even without that richness, though, the data seem to provide adequate support for the various hypotheses put forth by the authors. Despite the thoroughness of...