¹ The well-known political leftism of miners is paralleled by their atypical militancy in economic confrontations with management. See, for example, Kerr, Clark and Siegel, Abraham, “The Interindustry Propensity to Strike — An International Comparison,” in Industrial Conflict, ed. Kornhauser, Arthur, Dubin, Robert and Ross, Arthur (New York: McGraw-Hill, 1954)Google Scholar.

² See Forchheimer, K., “Some International Aspects of the Strike Movement,” Oxford University Institute of Statistics Bulletin, 10 (January, 1948)Google Scholar; Knowles, K.G.J.C., Strikes — A Study in Industrial Conflict (Oxford: Basil Blackwell, 1952)Google Scholar; Goetz-Girey, Robert, Le Mouvement des Grèves en France 1919–1962 (Paris: Editions Sirey, 1963)Google Scholar; and Shorter, Edward and Tilly, Charles. “The Shape of Strikes in France, 1830–1960,” Comparative Studies in Society and History, 13 (January 1971), 60–86CrossRefGoogle Scholar.

Small proprieters, entrepreneurs, rentiers and other self-employed persons are on the whole not relevant to industrial conflict but constitute a significant fraction of the civilian labor force in some nations and time periods. Therefore they have been excluded from the frequency ratio. The self-employed do of course occasionally strike (witness, for example, the strikes by independent French shopkeepers over the government's profit and price controls in 1973, or the strikes by independent truckers over fuel pricing issues in the United States). However, such events are not included in the industrial conflict data reported to the I.L.O. and are in any case not germane to the line of analysis pursued here.

³ Edward Shorter and Charles Tilly, “The Shape of Strikes in France.”

⁴ Knowles, Strikes, and Peterson, Florence, Strikes in the United States, 1880–1936 (Washington: U.S. Department of Labor, Bureau of Labor Statistics, Bulletin No. 651, 1937)Google Scholar.

⁵ The contrast between the traditional, war-of-attrition strike familiar to most Americans and the guerrilla warfare strategy described in the text is put into even sharper focus by the description of the latter pattern given by Blumenfeld: “Italian unions, which have no strike funds, have developed a whole range of disruptive activities. Among these are the ‘chessboard’ strike, involving only selected departments; the ‘paybook’ strike, in which every worker whose paycard carries an odd number engages in disputes on odd days of the week, while workers with even numbers fight out their claims on the even days; and strikes in which blue-collar workers lay down their tools in the morning but return to work after lunch, only to find that the white-collar clerks are out — thus stopping work for an entire day with the loss of only a half a day's pay.” See Blumenfeld, Yorick, “Industrial Strife in Western Europe,” Editorial Research Reports, 21 (June 1971), 409–426Google Scholar. Sartori provides an interesting, although exaggerated, assessment of the leverage that “guerrilla warfare” strike strategies (known as scioperi articolati in Italy and débrayages or grèves tournantes in France) gives labor in service-dependent, “postindustrial” societies. See Sartori, Giovanni,” “Il Potere Del Lavoro Nella Societa Post-Pacificata,” Revista Italiana Di Scienza Politica, ANNO III (Aprile 1973), 31–81Google Scholar.

⁶ An extensive analysis of twentieth century trends in industrial conflict is given in my paper “Long-Run Trends in Strike Activity in Comparative Perspective,” mimeo., 1976. Aaron's thoughtful review of the evidence for six European nations, and Erickson and Grofman's preliminary analyses of long-term trends in fifteen industrial societies also square with the conclusions drawn here. See Aaron, Benjamin, “Methods of Industrial Action: Courts, Administrative Agencies, and Legislatures,” chapter 2 in Industrial Conflicts: A Comparative Legal Survey, ed. Aaron, Benjamin and Wedderburn, K. W. (London: Wm. Clowes & Sons, Ltd., 1972)Google Scholar; and Erickson, K. and Grofman, B., “A Comparative Longitudinal Analysis of Strike Activity: Applications to Industrialized and Industrializing Economies,” Paper delivered at the International Political Science Association Meeting, Montreal, Canada, August, 1973Google Scholar.

⁷ Touraine, Alain, The Post-Industrial Society (New York: Random House 1971), p. 16Google Scholar.

⁸ See Bouvier, Jean, “Mouvement Ouvrier at Conjonctures Economiques,” Le Mouvement Social (Juillet-Septembre, 1964), pp. 3–30Google Scholar; Robert Goetz-Girey, Le Mouvement des Grèves en France; Knowles, Strikes; O'Brien, F. S., “Industrial Conflict and Business Fluctuations: A Comment,” Journal of Political Economy, 73 (1965), 650–654CrossRefGoogle Scholar; Albert Rees, “Industrial Conflict and Business Fluctuations,” in Kornhauser, Dubin and Ross, Industrial Conflict; and Weintrub, Andrew, “Prosperity Versus Strikes: An Empirical Approach,” Industrial and Labor Relations Review, 19 (January 1966), 231–238CrossRefGoogle Scholar.

⁹ It is not really necessary that employers actually replace strikers with great regularity — it is the psychological threat that is important, and this is heightened during periods of high unemployment. For example, after the 1970–71 recession in the United States, The Wall Street Journal ran a front page article on the “conciliatory mood” of U.S. workers. The story read in part: “… across the country at large and small companies, workers are choosing to be more conciliatory when faced with the threat of losing their jobs. This is in sharp contrast to the labor scene of recent years (i.e., when unemployment was low), both union and corporate officials agree. Not long ago, they say, rank-and-filers … would probably have been angered by the thought of concessions … many manufacturing executives have openly complained in recent years that too much control had passed from management to labor. With sales sagging … they feel safer in attempting to restore what they call “balance” (January 26, 1972, p. 1, cited in Boddy, Raford and Crotty, James, “Class Conflict and Macropolicy: The Political Business Cycle,” Review of Radical Political Economics, 7 [Spring, 1974], 1–19)CrossRefGoogle Scholar.

¹⁰ For a contrary argument which contends that strikes have greater disutility for firms when economic activity is low (unemployment high) see Vanderkamp, John, “Economic Activity and Strikes in Canada,” Industrial Relations, 9 (February 1970), 215–230Google Scholar.

¹¹ SirHicks, John, The Theory of Wages (New York: St. Martin's Press, 1963), p. 156CrossRefGoogle Scholar.

¹² “Real” wages refer of course to money wages deflated for price increases, that is, R = W/P where R, W, and P denote real wages, money wages, and a price index, respectively. A following section deals with the “money illusion” thesis that labor militancy responds to money wage increases alone.

¹³ “A Comparative Study of Civil Strife,” in Violence in America, ed. Graham, Hugh Davis and Gurr, Ted Robert (New York: New American Library 1969), p. 590Google Scholar.

¹⁴ Davis, James C., “The Curve of Rising and Declining Satisfactions as a Cause of Some Great Revolutions and a Contained Rebellion,” in Graham, and Gurr, , p. 671Google Scholar.

¹⁵ Krech, David and Crutchfield, R. S., Theory and Problems of Social Psychology (New York: McGraw-Hill, 1948), p. 542CrossRefGoogle Scholar. Economists and industrial relations specialists also have used expectational explanations of worker unrest and militancy. See, for example, Trade Union Government and Collective Bargaining: Some Critical Issues, ed. Seidman, Joel (New York: Praeger Publishers 1970)Google Scholar, particularly the essays of Barbash and Wirtz.

¹⁶ Even the massive strikes of May–June 1968 in France (which were viewed largely as spontaneous “political” events in many popular and academic accounts) centered in the overwhelming majority of cases around traditional demands for wage increases and were called, promoted, and directed by the usual labor organizations. (See Durand, Claude, “Revendications Explicites et Revendications Latentes,” Sociologie du Travail (October–Décembre 1973), pp. 394–409Google Scholar and Ross, George, “French Working Class Politics After May–June 1968: A New Working Class?” Paper delivered at the 1973 Annual Meeting of the American Political Science Association, New Orleans, Louisiana, September 1973Google Scholar. Such strikes of course have important political implications in the sense that they challenge state economic policies designed to hold down the rate of increase of wages and prices.

¹⁷ In the current period Δ R is likely to be influenced by SV and so the model in (1) should be viewed as an abbreviated form of a larger system within which Δ R and Δ R* are lagged endogenous variables.

¹⁸ See Arrow, Kenneth J. and Nerlove, Marc, “A Note on Expectations and Stability,” Econometrica, 26 (April 1958), 297–305CrossRefGoogle Scholar; Cagan, Phillip D., “The Monetary Dynamics of Hyperinflation,” in Studies in the Quantity Theory of Money, ed. Friedman, Milton (Chicago: University of Chicago Press, 1956)Google Scholar; Enthoven, Allan C. and Arrow, Kenneth J., “A Theorem on Expectations and the Stability of Equilibrium,” Econometrica, 22 (April 1956), 288–293CrossRefGoogle Scholar; Hadar, Josef, “On Expectations and Stability,” Behavioral Science, 13 (November 1968), 445–454CrossRefGoogle Scholar; Nerlove, Marc, “Adaptive Expectations and Cobweb Phenomena,” Quarterly Journal of Economics, 72 (May 1958), 227–240CrossRefGoogle Scholar; and Turnovsky, Steven J., “Empirical Evidence on the Formation of Price Expectations,” Journal of the American Statistical Association, 65 (December 1970), 1441–1454CrossRefGoogle Scholar.

¹⁹ For those who prefer elegance, the result in (11) follows directly from lag algebra. Let B equal a back-shift operator such that B ⁱY _t, = Y _t-i and rewrite (9)

may be viewed as a convergent geometric series with solution 1/1−(λ + δ)B, and thus (a) implies:

Clearing (b) of the denominator 1−(λ + δ)B yields the estimating equation given in (11).

²⁰ This model is adopted from Ashenfelter, Orley and Johnson, George E., “Bargaining Theory, Trade Unions, and Industrial Strike Activity”, American Economic Review, 50 (March 1969) 35–49Google Scholar. I am grateful to Robert Solow for helping me derive the implications of various lag functions which led to the interpretation given here.

²¹ Almon, Shirley, “The Distributed Lag Between Capital Appropriations and Expenditures,” Econometrica, 33 (January 1965), 178–196CrossRefGoogle Scholar.

²² A fixed, dummy variable approach to the pooled estimation problem is taken here in preference to the alternative random variables, “error components” approach because of computational ease, and also because an important assumption necessary to preserve the consistency of the latter method — that “specific ignorance” be independent of regressors — seems unreasonable. Extensive analyses of pooled estimation strategies are given by Maddala, G. S., “The Use of Variance Components Models in Pooling Cross Section and Time Series Data,” Econometrica, 39 (March 1971), 341–358CrossRefGoogle Scholar; Nerlove, Marc, “Further Evidence on the Estimation of Dynamic Economic Relations from a Time Series of Cross Sections,” Econometrica, 39 (March 1971), 359–382CrossRefGoogle Scholar; and Wallace, T. D. and Hussain, Ashig, “The Use of Error Components Models in Combining Cross Sections with Time Series Data,” Econometrica, 37 (January 1969), 55–72CrossRefGoogle Scholar. Note also that the long-run increase expression (1-ρ) L in (15) is necessarily embedded within the nation-specific intercepts of (17c) and therefore is effectively “lost” for inferential purposes (i.e., these parameters are under-identified).

²³ Marquardt, Donald W., “An Algorithm for Least Squares Estimation of Non-Linear Parameters,” Journal of the Society for Industrial and Applied Mathematics 11 (June 1963), 431–441CrossRefGoogle Scholar.

²⁴ Note that even if the disturbance of the original adaptive expectations function in (9) is “white noise,” application of the Koyck transformation introduces a first-order moving average error process in the estimating equation (17b). As it turned out, however, the errors displayed negligible serial correlation in all equations and so instrumental variables as well as the rather complicated estimation procedures required to secure efficiency in a pooled data model proved unnecessary. A discussion of the estimation problems posed by auto-correlated disturbances and lagged endogenous variables is provided in Hibbs, Douglas A. Jr., “Problems of Statistical Estimation and Casual Inference in Time-Series Regression Models,” in Sociological Methodology 1973–74, ed. Costner, Herbert (San Francisco: Jossey-Bass, Inc., 1974), pp. 252–308Google Scholar.

²⁵ The best fitting model was a second degree polynomial constrained to drop-off to zero after a five period lag. Similar results are reported by Ashenfelter and Johnson, “Bargaining Theory,” and Pencavel, John H., “An Investigation into Industrial Strike Activity in Britain,” Economica, 37 (August 1970), 239–256CrossRefGoogle Scholar, for quarterly models of strike frequency in the United States and Great Britain respectively. The estimation technique used here was initially proposed by Shirley Almon, “The Distributed Lag.” A comprehensive review of distributed lag formulations generally is given by Dhrymes, Phoebus J., Distributed Lags: Problems of Estimation and Formulation (San Francisco: Holden Day, 1971)Google Scholar.

²⁶ A plot of the lag distribution coefficients is presented in a following section, after several additional hypotheses are incorporated into the strike volume model. The real wage data used here are average hourly earnings in the manufacturing sector deflated by the consumer price index. Movements in manufacturing wages are presumed to reflect closely economy-wide changes (cf. Ashenfelter, Orley, Johnson, George E. and Pencavel, John H., “Trade Unions and the Rate of Change of Money Wages in United States Manufacturing Industry,” The Review of Economic Studies, 39 (January 1972: Appendix)CrossRefGoogle Scholar. Analyses using total annual employee compensation per civilian labor force member produced results virtually identical to those in Table 1.

²⁷ Albert Rees, “Industrial Conflict and Business Fluctuations.”

²⁸ Although the profits result is theoretically plausible, the statistical insignificance of this variable may stem from excessive measurement error. Profits data are easily subjected to the artistry of corporate accountants, are well-known to be “hidden” as costs via accelerated depreciation tax laws, and are difficult to determine for unincorporated businesses. Furthermore, a significant fraction of the labor force in all economically advanced societies is employed in nonprofit activities. The profits-to-wages ratio is also correlated with real wages and unemployment. However, excessive collinearity is not a problem here: the multiple R² of profits/wages with the remaining independent variables is .17.

²⁹ Ross, Arthur M. and Hartman, Paul T., Changing Patterns of Industrial Conflict (New York: John Wiley & Sons, 1960), pp. 49–50Google Scholar.

³⁰ Malles, Paul, Trends in Industrial Relations Systems of Continental Europe (Ottawa: Economic Council of Canada, Task Force on Labour Relations, 1969)Google Scholar.

³¹ Sturmthal, Adolf, Comparative Labor Movements: Ideological Roots and Institutional Developments (Belmont, Calif.: Wadsworth Publishing Co., 1972)Google Scholar.

³² A great many studies were consulted in developing these categories. Among the most useful of those not already cited were: Albeda, W., “Recent Trends in Collective Bargaining in the Netherlands,” International Labour Review, 103 (March 1971), 247–268Google Scholar; Banks, Robert F., “The Pattern of Collective Bargaining,” chapter 3, Industrial Relations: Contemporary Problems and Perspectives, ed. Roberts, Ben C. (London: Methuen and Co., Ltd. 1968)Google Scholar; Blanpain, Roger, “Recent Trends in Collective Bargaining in Belgium,” International Labour Review, 104 (July–August 1971) 111–129Google Scholar; Cullen, Donald E., “Recent Trends in Collective Bargaining in the United States,” International Labour Review, 105 (June 1972), 507–529Google Scholar; Delamotte, Yves, “Recent Trends in Collective Bargaining in France,” International Labour Review, 103 (April 1971), 351–377Google Scholar; Giugni, Gino, “Recent Trends in Collective Bargaining in Italy,” International Labour Review, 104 (October 1971), 307–328Google Scholar; Heady, Bruce W., “Trade Unions and National Wages Policies,” Journal of Politics, 32 (1970), 407–439CrossRefGoogle Scholar; Hogberg, Gunnar, “Recent Trends in Collective Bargaining in Sweden,” International Labour Review, 107 (March 1973), 223–238Google Scholar; Horowitz, David L., The Italian Labor Movement, (Cambridge: Harvard University Press, 1963)CrossRefGoogle Scholar; Lorwin, Val R., The French Labor Movement (Cambridge: Harvard University Press, 1954)Google Scholar; Lorwin, Val R., “Labor Organizations and Politics in Belgium and France,” in National Labor Movements in the Postwar World, ed. Kassalow, Everett (Evanston: North-western University Press, 1963)Google Scholar; Mitsufuji, Tadashi and Hagisawa, , “Recent Trends in Collective Bargaining in Japan,” International Labour Review, 105 (February 1972), 135–153Google Scholar; Roberts, Ben C. and Rothwell, Sheila, “Recent Trends in Collective Bargaining in the United Kingdom,” International Labour Review, 106 (December 1972), 543–571Google Scholar; and Sturmthal, Adolf, Contemporary Collective Bargaining in Seven Countries (Ithaca, N.Y.: Cornell University Press 1957)Google Scholar.

³³ Analogous tests for parameter stability were undertaken for equations (17a) and (17b), and for various other subsets of nation-year observations, but the results add little to what already has been reported.

³⁴ Knowles, K. G. J. C., Strikes —A Study in Industrial Conflict (1952), p. 227Google Scholar. The idea that workers suffer from a money illusion apparently originated with Keynes, John M., The General Theory of Employment, Interest and Money (New York: Harcourt, Brace, Jovanovich, 1936)Google Scholar.

³⁵ See Ashenfelter and Johnson, “Bargaining Theory”; Pencavel, “Strike Activity in Britain”; Vanderkamp, “Economic Activity and Strikes in Canada.”

³⁶ Actually the continuous time equality

is only approximated by the discrete time relation ΔR = ΔW − ΔP. However, for our purposes the difference is negligible.

³⁷ It is worth noting here that Kramer found percentage changes in prices and per capita monetary income to act as “mirror images” in their effects on fluctuations in the partisan division of the vote for the U.S. House of Representatives during the 1896–1964 period. Kramer interpreted this, quite properly, as evidence of rational mass electoral behavior. See Kramer, Gerald, “Short-Term Fluctuations in U.S. Voting Behavior, 1896–1964,” American Political Science Review, 65 (March 1971), 131–143CrossRefGoogle Scholar.

³⁸ Ross, and Hartman, , Changing Patterns and Industrial Conflict, p. 69Google Scholar.

³⁹ Cited in Kassalow, Everett M., Trade Unions and Industrial Relations: An International Comparison (New York: Random House 1969), p. 52Google Scholar.

⁴⁰ Przeworski, Adam and Sprague, John, “Concepts in Search of Explicit Formulation: A Study in Measurement,” Midwest Journal of Political Science, 15 (May 1971), 183–218CrossRefGoogle Scholar.

⁴¹ Ross, and Hartman, , Changing Patterns of Industrial Conflict, p. 69Google Scholar.

⁴² See Blackmer, Donald L. M., “Italian Communism: Strategy for the 1970's,” Problems of Communism, 21 (May–June 1972), 41–56Google Scholar, and Ross, George, “French Working Class Politics After May–June 1968: A New Working Class?”, Paper delivered at the 1973 Annual Meeting of the American Political Science Association, New Orleans, Louisiana, September, 1973Google Scholar.

⁴³ Similar conclusions with regard to Great Britain are reached by Eldridge, John E. T., Industrial Disputes: Essays in the Sociology of Industrial Relations (London: Routledge & Kegan Paul, 1968)Google Scholar; and Turner, Herbert A., The Trend of Strikes (Leeds: Leeds University Press, 1963)Google Scholar. It should be understood that the results associated with the political party terms apply only to nations experiencing oscilliations in Labor-Socialist rule. The effects of long-term leftist rule on labor militancy (e.g., in Sweden, Norway and Denmark where the Left has governed almost continuously during the postwar period) are absorbed by the country-specific constant terms. The consequences of such long-standing political and institutional differences are discussed in great detail in my paper “Long-Run Trends in Strike Activity in Comparative Perspective.”

⁴⁴ Cited in Kassalow, , Trade Unions and Industrial Relations, p. 50Google Scholar.

⁴⁵ Several of these alternatives are worth mentioning. First, since the translation of electoral votes into legislative seats is by no means direct [see Rae, Douglas W., The Political Consequences of Electoral Laws (New Haven: Yale University Press, revised edition, 1971)Google Scholar, for an exhaustive analysis], the incentive function of Labor-Socialist Party elites in (22) was also defined in terms of seat shares as opposed to vote shares. Second, Equation (23) was revised to incorporate the capability as well as the motivation of party elites to discourage strike activity; where capability was measured in terms of the size of the Labor-Socialist political constituency (voters). Third, the Labor-Socialist vote share distance in (21), as well as the seat share variant described in the first point above, was modified to include (a) the vote or seat share of the U.S. Democratic party — following Greenstone's argument [Greenstone, J. David, Labor in American Politics (New York: Alfred A. Knopf, 1969Google Scholar] that American labor and the Democrats have become interpenetrated in a way that is at least partially equivalent to party-union alliances in much of Western Europe — and, (b) the vote or seat share of the French Communist Party (PCF) during the period of the left alliance. As noted in the text, none of these modifications led to empirical outcomes significantly different from those reported in Table 4.

⁴⁶ Goodman, J. F. B., “Strikes in the United Kingdom: Recent Statistics and Trends,” International Labour Review, 95 (May 1967), 465–481Google Scholar; and Silver, Michael, “Recent British Strike Trends: A Factual Analysis,” British Journal of Industrial Relations, 11 (March 1973), 66–104CrossRefGoogle Scholar. The situation was reversed, however, after the passage of the Industrial Relations Act of 1971, which was bitterly resented by the British trade union establishment and which stimulated the normally moderate Trades Union Congress (the peak union organization) to join the shop stewards in pressing the militant position. Consequently, nearly 80 per cent of the man-days lost in strike activity in the years after 1971 were due to “official” disputes. A detailed analysis is given by Dorfman, Gerald, “An End to Producer Group Politics in Britain?: The Industrial Relations Act of 1971,” paper delivered at The 1973 Annual Meeting of the American Political Science Association, New Orleans, Louisiana, September, 1973Google Scholar. The dynamics underlying differential responsiveness of union leadership and the rank-and-file to the wider political interests of Labor-Socialist parties is of course an important problem, which the simple models developed here cannot address with any authority.

⁴⁷ See Greenbaum, John S., “The Rebellious Rank and File,” Personnel, 49 (March–April 1972), 20–25Google Scholar; Jamieson, Stuart M., Times of Troubles Labour Unrest and Industrial Conflict in Canada, 1900–66, (Ottawa: Economic Council, Task Force on Labour Relations, 1968)Google Scholar; Seidman, Trade Union Government and Collective Bargaining; and Weir, Stanley, “Rebellion in American Labor's Rank and File,” in Workers' Control, ed. Hunnius, Gary, Garson, G. David and Case, John (New York: Vintage Books 1973)Google Scholar.

⁴⁸ Ross, and Hartman, , Changing Patterns of Industrial Conflict, p. 58Google Scholar.

⁴⁹ The t-statistics in this nonlinear model should be interpreted with caution since strictly speaking they have meaning only in terms of the linearized equation at the solution point. There is no question, however, that the exponential form given in Table 4 outperforms all reasonable alternative functions.

⁵⁰ Hobsbawm, E. J., “Economic Fluctuations and Some Social Movements Since 1800,” Economic History Review, Second Series, 5 (1952), 1–25CrossRefGoogle Scholar.

⁵¹ Since the real wages variable is undoubtedly measured with significant error, one should not place great confidence in the exact values of the individual lag coefficients or in the lag distribution. This follows from the results of a recent paper by Grether and Maddala which shows that measurement error in an exogenous variable typically leads to the appearance of a longer lag distribution than is actually the case, but at the same time produces a better estimate of the total response. In the case at hand, this means the β₁Σμ₁ = − 59.2 in (28c) of Table 4 is probably a good estimate of the total response of SV _t, to ΣΔR _t-1, (the total multiplier), although the lag distribution index i = 1 … 5 may be too long. See Grether, David and Maddala, G. S., “Errors in Variables and Serially Correlated Disturbances in Distributed Lag Models,” Econometrica, 41 (March 1973), 255–262CrossRefGoogle Scholar.

⁵² The work of Hines and Ashenfelter et al., are important exceptions to the restrictive treatment of the employment-inflation problem found in the bulk of the empirical economic studies. See Hines, A. G., “Trade Unions and Wage Inflation in the United Kingdom 1893–1961,” Review of Economic Studies, 31 (October 1964), 221–252CrossRefGoogle Scholar; and Ashenfelter, Orley, Johnson, George E. and Pencavel, John H., “Trade Unions and the Rate of Change of Money Wages in United States Manufacturing Industry,” The Review of Economic Studies, 39 (January 1972), 27–54CrossRefGoogle Scholar. Also, see my forthcoming paper “Labor Militancy and Wage Inflation: A Comparative Analysis.”