Conflits du Travail, Changement Social et Politique en France depuis 1950.

By Monique Borrel (Paris: L'Harmattan. iv plus 266pp.).

Beneath a paragraph making large claims for its content, the back cover of Conflits du Travail bears a little passport-style photograph of Monique Borrel. She peers out at readers quizzically as if to say, "See if you can figure out what I did, why I did it, and what it means!" Even for readers who are familiar with recent French history, measurement of industrial conflict, econometric modeling, and analysis of contentious politics, her book sets quite a challenge. Between a four-page introduction in the format of a dissertation abstract and a twelve-page conclusion offering general reflections on the place of strikes in France's political and economic history since 1950, the book lays out three dense slabs of unequal size: 1) 35 pages of definitions, procedures, quarterly counts, and principal events from the roughly 125 thousand strikes (the book never tells us exactly how many) occurring in France between 1950 and 1987 that Borrel assembled in a catalog drawn from governmental reports and her own reading of Le Monde; 2) a 130-page section alternating repeatedly among hasty reviews of one relevant literature or another (theories of collective action, econometric models of industrial conflict, measurement problems, and so on), compressed statements of Borrel's own arguments in these regards, and descriptions of her procedures for representing those arguments econometrically; 3) 46 pages consisting entirely - repeat, entirely! - of results from Ordinary Least squares (OLS) regressions among variables specified in the previous section, definitions of the variables, data tables, and notes on the tables.

A reader stumbles through a badly lighted labyrinth far from the equally databased, technically more sophisticated, and much more vivifying worlds of recent, relevant (and here uncited) monographs by Samuel Cohn on French miners' strikes 1890-1935, Roberto Franzosi on Italian postwar industrial conflict, and Olivier Fillieule on French demonstrations 1979-1993. Borrel's book takes us even farther from the world of Michelle Perrot's magisterial Les Ouvriers en Greve, which a generation ago used statistical descriptions of nineteenth-century strikes to set problems for social-historical analysis of workers' experiences. Where are we? In the midst of sixteen OLS regression equations reported on a mere three pages, pp. 185-187. The surrounding text does not say from how many statistical trials, or by what rules, Borrel selected the final sixteen; the variability of predictors included in different equations, the absence of obvious multicollinearity, among predictors, comforting proximity to the desired value of 2.0 for most Durbin-Watson serial-correlation tests, and explained variances ranging from 0.43 to 0.98 bespeak huge winnowing of unsatisfactory results, not to mention some instances in which the ostensible independent and dependent variables intertwined so closely as to be almost indistinguishable. The absence of zero-order correlation coefficients - one of the few operational details missing from this massive compendium - makes it impossible to check suspicions in these regards. The surviving equations, in any case, a) analyze quarterly or annual variation in national aggregate statistics, most often for six to eight years in the cases of quarterly data and a dozen years or so in the cases of annual data, b) treat as causes and/or as effects seven different segments or summaries of the strike catalog: strike waves; national strikes involving several industries; one-industry national strikes; regional multi-industry strikes; single-industry private-sector strikes; CGT-organized strike demonstrations; national workers' action days; c) treat as further causes or effects sixteen national political-economic variables for the intervals in question: not only such standards as unemployment, unemployment compensation rates, workers' wages, and union membership, but also such political-organizational matters as electoral success of the left, governmental policy toward labor, and alliances among major labor federations. (We begin to see why it took Borrel ten years to collect her data and conduct her analyses.) To the extent that we trust simple linear regressions for short periods as representations of cause and effect in labor-management-state interactions, Borrel's statistical results give some comfort to conventional expectations that unemployment dampens strike activity, while price increases, worker mobilization, and left political power promote it. They suggest that both government benefits and wage levels increased in times of labor militancy. But they fall calamitously short of identifying, as the back cover trumpets, "the mechanisms by which strikes played a central part in France's social and political evolution, determining not only the establishment of a large social-welfare system during the post-war period but also rapid unionization and the emergence of a left electoral majority during the 1970s."

Charles Tilly Columbia University