¹ Hibbs, Douglas A. Jr., ‘On the Political Economy of Long-Run Trends in Strike Activity’, British Journal of Political Science, VIII (1978), 153–77.CrossRefGoogle Scholar

² Shorter, Edward and Tilly, Charles, Strikes in France 1830–1968 (Cambridge: Cambridge University Press, 1974)Google Scholar; Korpi, Walter, ‘Industrial Relations in Sweden’ (Swedish Institute for Social Research, mimeo, 1975)Google Scholar; Pizzorno, Alessandro, ‘Political Exchange and Collective Identity in Industrial Conflict’, in Crouch, Colin and Pizzorno, Alessandro, eds., The Analysis of Class Conflict in Western Europe (London: Macmillan, 1978).Google Scholar

³ The classic theorists of the labour movement – Marx, Perlman, the Webbs – all dwelt on this issue, and so do their latterday counterparts. A strong normative element runs through most of this work, but for a more strictly analytical treatment by an industrial relations scholar, see Sturmthal, Adolph, ‘Industrial Relations Strategies’, in Sturmthal, Adolph and Scoville, James G., eds., The International Labor Movement in Transition (Urbana: University of Illinois Press, 1973), pp. 1–33.Google Scholar

⁴ Ross, Arthur and Hartman, Paul T., Changing Patterns of Industrial Conflict (New York: Wiley, 1960)Google Scholar. The proposition that socio-political conflict has ‘absorbed’ industrial militancy is also articulated in Kernhauser, Arthur, Dubin, Robert and Ross, Arthur M., eds., Industrial Conflict (New York: McGraw-Hill, 1954), p. 15Google Scholar; and Lester, Richard, As Unions Mature: An Analysis of the Evolution of American Unionism (Princeton, N.J.: Princeton University Press, 1958), p. 46CrossRefGoogle Scholar. For critiques of Ross and Hartman, see Eldridge, J. E. T., Industrial Disputes: Essays in the Sociology of Industrial Relations (London: Routledge and Kegan Paul, 1968)Google Scholar; and Ingham, Geoffrey K., Strikes and Industrial Conflict (London: Macmillan, 1974).CrossRefGoogle Scholar

⁵ Ross, and Hartman, , Changing Patterns, p. 68.Google Scholar

⁶ Knowles, K. G. J. C., Strikes – A Study in Industrial Conflict (Oxford: Blackwell, 1952), p. 96.Google Scholar

⁷ This emphasis is perhaps clearest in Snyder, David, ‘Institutional Setting and Industrial Conflict’, American Sociological Review, XL (1975), 259–78CrossRefGoogle Scholar. However, as Snyder emphasized, he was concerned with short-run fluctuations in strikes rather than the long-run drift.

⁸ Headey, Bruce W., ‘Trade Unions and National Wage Policies’, Journal of Politics, XXXII (1970), 407–39, p. 435 (emphasis added).CrossRefGoogle Scholar

⁹ Shorter, and Tilly, , Strikes in France, p. 330.Google Scholar

¹⁰ See Erbès-Seguin, Sabine, ‘Les Deux Champs de l'Affrontement Professionel’, Sociologie du Travail, XVIII (1976), 121–38.CrossRefGoogle Scholar

¹¹ This point has been recognized by both ‘pluralist’ and ‘Marxist’ students of industrial relations. See, respectively, Barbash, Jack, ‘Price and Power in Collective Bargaining’. Journal of Economic Issues, XI (1977), 847–59CrossRefGoogle Scholar; and Hyman, Richard, Industrial Relations: A Marxist Introduction (London: Macmillan, 1975).CrossRefGoogle Scholar

¹² In a personal communication, David Snyder has raised the possibility that the distinction between price and power is of no practical significance for Hibbs's argument if control over the two facets actually covaries. In practice, however, I doubt that this is the case. For example, it has often been remarked that workplace ‘job regulation’ is given much more emphasis in Anglo-American industrial relations than under economy-wide collective bargaining of the Northern European type. Indeed, it was precisely the gap between the distributional and workplace power of Sweden's LO (major union confederation) which encouraged it to seek comprehensive statutory reforms aimed at enhancing its role on the power dimension. See, for example, Casten von Otter's chapter on Sweden in Barkin, Solomon, ed., Worker Militancy and its Consequences, 1965–75 (New York: Praeger, 1975).Google Scholar

¹³ I do not wish to imply, as Clegg, Hugh has recently asserted in Trade Unionism under Collective Bargaining (Oxford: Blackwell, 1976)Google Scholar, that institutions of industrial relations in themselves explain behaviour. Nevertheless, institutional factors may be important as intervening variables, or as symptomatic expressions of causal forces of a structural nature. They also represent part of the historical configuration within which action takes place and by which it is limited. In short, institutional variables are treated here as illuminating rather than determining behaviour.

¹⁴ Like Ross and Hartman, Hibbs implicitly discounts the significance of ‘rank and file revolts’ for national strike patterns. This is not the case as far as Shorter and Tilly are concerned (see Strikes in France, p. 327)Google Scholar; and Pizzorno, , in ‘Political Exchange and Collective Identity’Google Scholar, devotes considerable attention to the problem of intra-union relations in a context of political bargaining by unions.

¹⁵ Woods, Noel S., The Industrial Relations Act 1973, Occasional Papers in Industrial Relations, no. 11 (Wellington, New Zealand: Victoria University of Wellington Industrial Relations Centre, 1974).Google Scholar

¹⁶ This point is brought out in many of the contributions to Colin Crouch and Pizzorno, Alessandro's two volume collection, The Resurgence and the Analysis of Class Conflict in Western Europe Since 1968 (London: Macmillan, 1978).Google Scholar

¹⁷ For two views, see Parkin, Frank, Class Inequality and Political Order: Social Stratification in Capitalist and Communist Societies (London: MacGibbon and Kee, 1971)Google Scholar, and Stephens, John D., The Transition from Capitalism to Socialism (London: Macmillan, forthcoming).CrossRefGoogle Scholar

¹⁸ The countries where strikes have more or less ‘withered away’ in the post-Second World War period are here deemed to be Austria, Germany, Switzerland, the Netherlands, Sweden, Norway and Denmark.

¹⁹ See especially Korpi, , ‘Industrial Relations in Sweden’Google Scholar. For a fuller exposition, see Korpi, Walter, ‘Industrial Relations and Industrial Conflict: The Case of Sweden’ (paper presented to the Conference on Contemporary International Labor Problems, Madison, Wisconsin, 1977).Google Scholar

²⁰ Pizzorno, , ‘Political Exchange and Collective Identity’.Google Scholar

²¹ The ‘volume’ of strikes is the number of man-days lost, normally adjusted for labour force size.

²² Hibbs, Douglas A. Jr., ‘Industrial Conflict in Advanced Industrial Societies’, American Political Science Review, LXX (1976), 1033–58, p. 1035.CrossRefGoogle Scholar

²³ Knowles, K. G. J. C., ‘Work Stoppages in the United Kingdom: A Comment’, Bulletin of the Oxford Institute of Economics and Statistics, XXVIII (1966), p. 61.Google Scholar

²⁴ Shorter, and Tilly, , Strikes in France.Google Scholar

²⁵ Forcheimer, K., ‘Some International Aspects of the Strike Movement’, Bulletin of the Oxford University Institute of Statistics, X (1948), 9–24.Google Scholar

²⁶ In addition to the work of Ross and Hartman, Shorter and Tilly, and Snyder see, for example, Erickson, Kenneth and Grofman, Bernard, ‘A Comparative Longitudinal Analysis of Strike Activity’ (paper presented to the International Political Science Association, Montreal, 1973)Google Scholar; Ellis, Lillian, Inter- and Intro-National Comparisons of Two Strike Models (Ph.D. dissertation, New School for Social Research, 1972)Google Scholar; Britt, David and Galle, Omer R., ‘Industrial Conflict and Unionization’, American Sociological Review, XXXVII (1972), 46–57CrossRefGoogle Scholar and ‘Structural Antecedents of the Shape of Strikes: A Comparative Analysis’, American Sociological Review, XXXIX (1974), 642–51.Google Scholar

²⁷ For example, I have calculated zero-order correlations over time between different strike indices, for an average of seventy annual observations for each of eighteen countries. These correlations are unstable (between countries), both in size and in sign, and they are never very large. The correlations between frequency and mean size range between -·24 and +·55; for frequency and mean duration the range is -·52 to +·43; and for size and duration it is -·47 to +·49. Clearly, a single theory could not explain the behaviour of all three indices in any one country, let alone in all countries.

²⁸ The geometric mean of n observations is defined as the nth root of the product of these observations. It can also be thought of as the arithmetic mean, after logarithmic transformation. The geometric mean minimizes the effect of extreme values, in much the same way as the median does.

²⁹ See especially Cronin, James Eugene, Strikes in Britain, 1888–1974: A Statistical and Historical Analysis (Ph.D. dissertation, Brandeis University, 1977)Google Scholar; Tilly, Charles and Shorter, Edward, ‘Les Vagues de Grèves en France, 1890–1968’, Annales, XXXV (1973), 857–87.CrossRefGoogle Scholar

³⁰ Gifford, Adam Sr., ‘The Impact of Socialism on Work Stoppages’, Industrial Relations, XIII (1974), 208–12.Google Scholar

³¹ The degree to which working-class parties (and union movements) are class-based and class-wide is clearly a variable, and one which will probably prove highly significant in explaining differing levels of strike activity cross-nationally.

³² Headey, , ‘Trade Unions and National Wage Policies’.Google Scholar

³³ In fn. 19 of ‘On the Political Economy…’, Hibbs appears to demonstrate that measures of proportional change perform nearly as well as the correlation between absolute differences. In fact, however, he is correlating the proportional change in strike volume with the absolute change in cabinet composition.

³⁴ See Fuguitt, Glen and Lieberson, Stanley, ‘Correlation of Ratios or Difference Scores Having Common Terms’, in Costner, Herbert L., ed., Sociological Methodology 1973–1974 (San Francisco: Jossey-Bass, 1974), pp. 128–44Google Scholar. For the present test I was able to utilize Hibbs's original data on cabinet composition (arithmetic means for 1918–38 and 1945–72). The arithmetic means for strikes in the same period were derived from my own figures for man-days lost per thousand members of the non-agricultural active population. Let S signify the strikes indicator and C the cabinet composition indicator, with subscripts 1 and 2 referring to the pre-war and post-war periods respectively. Estimate C₂ = a + bC₁ and set equal to the resulting predicted values of C ₂. Then let . Then the test equation is

³⁵ Apart from sources (too complex a matter to go into here), the differences between the present series and Hibbs's series are: (1) Hibbs attempted to estimate the number of employed non-agricultural employees, whereas I have used the more reliable (though conceptually less adequate) data for the non-agricultural active population. (2) Hibbs treats the problem of agricultural strikes in Italy (much more common than in other countries) by deflating total strike volume for Italy by the total number of employees. I have subtracted agricultural strikes from the Italian total, and deflated as usual by the non-agricultural active population.

³⁶ Sources for the ILO data were ILO Yearbook of Labour Statistics, various years, and Department of Employment Gazette (U.K.), various issues.

³⁷ The measures used are described in Table 1.

³⁸ As is clear from Hibbs's table, these effects are not adequately captured by measuring strike volume. The combination of Bourgeois Government and a small State Share in North America produces virtually the same volume as Bourgeois Government and a large State Share in France and Italy. The difference between the two types is in the form and function of industrial conflict.

³⁹ Since our ‘sample’ includes virtually the entire universe, the small number of cases does not detract from the generalizability (statistical ‘significance’) of the results. But with so few cases there is likely to be insufficient variation in the variables entered into multiple regression equations to adequately identify the separate effects of each explanatory factor.

⁴⁰ In Table 1 the index of Leftism was used in preference to the dummy variables so as to keep the number of independent variables to a minimum. However, results not reproduced here for the same set of equations, but using the dummy variables instead of the index, supported exactly the same interpretations.

⁴¹ It should be pointed out that the same orders of magnitude would be evident in Table 3 if the economy-wide measure of strike volume was used instead of the ILO measure. This is not surprising, since the zero-order correlation between them is ·974.

⁴² Hibbs's argument is that the French and Italian strike pattern is a consequence of the working class being excluded from power over a state apparatus which has considerable influence over the workers' conditions of life. On the face of it, this is a plausible interpretation. However, while it may be the result of using crude indicators, the data endow this hypothesis with much less plausibility. Results for the total State Share index can be examined in Table 5. Using the non-defence index, France and Italy would move down to a medium and low share respectively. The medium State Share group under bourgeois rule then comprises France, Germany, Canada and Ireland – of which only France seems to meet Hibbs's expectation of highly politicized strike activity.

⁴³ In an earlier article (‘Industrial Conflict in Advanced Industrial Societies’), Hibbs himself tried out several specifications of the short-run deterrent effect of electoral contests on strike activity, and brought in a verdict of not proven. On the other hand, his study was limited to a twenty-year period in ten countries, and used only strike volume as the dependent variable.

⁴⁴ Korpi, , ‘Industrial Relations and Industrial Conflict’.Google Scholar

⁴⁵ As Shorter, and Tilly, (Strikes in France, p. 325)Google Scholar point out, ‘in those countries where the transformation of the strike took place, the roots of change lay in the 1930s and the Great Depression, in fascist regimes and the warfare state, not necessarily in the post-war events as such’.