¹ Economic variables do, of course, have an important influence on short-run fluctuations in strike activity. See Hibbs, Douglas A. Jr., ‘Industrial Conflict in Advanced Industrial Societies’, American Political Science Review, LXX (1976), 1033–58, and the studies cited therein.CrossRefGoogle Scholar

² Ross, Arthur M. and Hartman, Paul T., Changing Patterns of Industrial Conflict (New York: Wiley, 1960), pp. 4–5.Google Scholar

³ Forcheimer, K., ‘Some International Aspects of the Strike Movement’, Oxford University Institute of Statistics Bulletin, X (1948), 9–24Google Scholar; Knowles, K. G. J. C., Strikes: A Study in Industrial Conflict (Oxford: Basil Blackwell, 1952), Chap. IV, Section I, pp. 143–60Google Scholar; Goetz-Girey, Robert, Le Mouvement des grèves en France 1919–1962 (Paris: Editions Sirey, 1965)Google Scholar; and Shorter, Edward and Tilly, Charles, ‘The Shape of Strikes in France, 1830–1960’, Comparative Studies in Society and History, XIII (1971), 60–86.CrossRefGoogle Scholar

⁴ Since strikes rarely occur in the agricultural sector (and those that do are not recorded with great accuracy), I have excluded agricultural workers from the labour force data of all nations except Italy, a country in which there has been substantial strike activity by farm labourers during most of the twentieth century. The military as well as small proprietors, entrepreneurs, rentiers, and other self-employed persons have also been excluded from the labour force data because they contribute little to aggregate strike activity but comprise a significant fraction of the work force in many nations and time periods. International and inter-temporal comparisons of industrial conflict are therefore faciliated by using the number of non-agricultural civilian wage and salary workers to adjust the strike statistics for differences in the size of the labour force.

⁵ Notice that strike duration is calculated from the available aggregate data by dividing total man-days lost by the total number of strikers, which yields a ‘weighted’ average duration (as opposed to a simple arithmetical average computed from individual disputes) – the weights being proportionate to the number of workers involved in the strike. For example, if w₁, w₂,… w_n, are the number of workers involved in strikes 1, 2,… n, and if d₁, d₂,… d_n are the corresponding durations of these strikes (in days), the number of man-days lost m₁ m₂,…m_n= d₁w₁, d₂w₂,… d_nw_n. The total number of man-days lost is M = m₁+m₂+…+m_n, and the total number of workers involved is W= w₁+w₂+…+ w_n. The weighted average duration defined in the text is therefore

where the weights are the number of workers involved in each dispute. The practical significance of this is that the duration measure is heavily influenced by large-scale strikes.

⁶ Shorter, and Tilly, , ‘The Shape of Strikes in France, 1830–1960’.Google Scholar

⁷ Notice, for example, the reduction in strike duration but not in strike volume between Figs, 1 (a) and 1 (b). By using the composite measure of ‘strike volume’, one avoids making spurious conclusions about trends in the gross magnitude of industrial conflict that can arise by focusing exclusively on one strike dimension and mistaking changes in it for changes in overall strike activity.

⁸ Small differences in definitions and methods of collection of the strike data somewhat affect the accuracy of inter-temporal and international comparisons. However, they are not important enough to impair analyses of major changes through time and major differences across countries in aggregate strike activity. See the discussions in Fisher, Malcolm, Measurement of Labour Disputes and Their Economic Effects (Washington, D.C.: Organization for Economic Cooperation and Development, 1973)Google Scholar; Eldridge, J. E. T., ‘Explanations of Strikes: A Critical Review’, in Industrial Disputes: Essays in the Sociology of Industrial Relations (London: Routledge and Kegan Paul, 1968)Google Scholar; and Ross, and Hartman, , Changing Patterns of Industrial Conflict, appendix.Google Scholar

⁹ Germany was excluded from the sample of industrial societies from the outset of this project because the partitioning of ine Country makes long-run time-series analyses a problem.

¹⁰ Depressions are defined as periods in which unemployment increased and Gross National Product and industrial production decreased for two years in succession. The primary source used to identify depression periods was Mitchell, B. R., European Historical Statistics 1750–1970 (New York: Columbia University Press, 1975).CrossRefGoogle Scholar

¹¹ Actually, the strike wave of 1968 in France was not recorded in the usual way at all; the figures for man-days lost shown in the French volume plot was derived from the careful unofficial calculations of Durand, M. and Harft, Y., ‘Panorama statistique des grèves’, Sociologie du Travail, IV (1973), 356–75.CrossRefGoogle Scholar

¹² An excellent source for references on the events of May–June 1968 is Wylie, Laurence et al. , France: The Events of May–June 1968, A Critical Bibliography (Pittsburgh, Pa.: Council for European Studies, 1973).Google Scholar

¹³ See Shorter, Edward and Tilly, Charles, Strikes in France 1830–1968 (Cambridge: Cambridge University Press, 1974).Google Scholar

¹⁴ My model for short-run fluctuations (see Hibbs, Douglas A. Jr., Strikes: The Political Economy of Industrial Conflict in Western Industrial Societies (Princeton, N.J.: Princeton University Press, forthcoming)) in British strike activity suggests that two factors were responsible for the 1966–72 trend: (i) the rate of growth of real wages systematically lagged behind the rate of growth of labour productivity in the industrial sector; and (ii) the reaction of the British trade-union establishment to the Conservative government's Industrial Relations Act of 1971 (repealed by the subsequent Labour government), which stimulated the normally moderate Trades Union Congress (the peak union organization) to join the shop stewards in pressing the militant position. The latter factor, of course, only influenced the post-1970 strike rate. For an alternative interpretation of the up-turn in British strike activity see the discussion in the concluding section.Google Scholar

¹⁵ Figure 2 shows that the dramatic decline in strike activity in Sweden, as in other countries, took place when the Social Democrats assumed political power. I return to this important point ahead.

¹⁶ The classic summary and critical review of the sociological and industrial relations literature on strike activity is Eldridge, ‘Explanations of Strikes’. My longer monograph on which this article is based– Long-Run Trends in Strike Activity – also reviews critically much of the relevant literature. The recent books by Ingham, Geoffrey K., Strikes and Industrial Conflict: Britain and Scandinavia (London: Macmillan, 1974)CrossRefGoogle Scholar and Hyman, Richard, Strikes (Glasgow: Fontana/Collins, 1975)Google Scholar cover similar ground in greater detail.

¹⁷ Cf. Shorter, and Tilly, , Strikes in France 1830–1968, especially Chaps, 1, 12 and 13Google Scholar; Snyder, David and Tilly, Charles, ‘Hardship and Collective Violence in France 1830 to 1960’, American Sociological Review, XXXVII (1972), 520–31CrossRefGoogle Scholar; and especially Korpi, Walter, ‘Industrial Relations in Sweden’, Swedish Institute for Social Research (08, 1975).Google Scholar Although I read Korpi's unpublished paper after this article was drafted, the theory sketched here is in broad agreement with his analysis of the evolution of Swedish industrial relations. However, Korpi, Shorter and Tilly, and Snyder and Tilly attribute much more importance than I do to working-class political power per se, as opposed to its instrumental consequences for the political economy of distribution. See the discussion ahead.

¹⁸ Perhaps it is more accurate to describe the process of public sector/welfare state distribution as a reshaping or alteration of market distributional outcomes.

¹⁹ The nations with the highest inter-war strike means (Norway, Sweden, the United Kingdom) have the largest absolute changes (declines) in strike volume and contribute heavily to the high correlation in Fig. 4. However, the correlation between the proportional changes in mean strike volume and the change in leftist political power is almost as large; r = —·88.

²⁰ A discussion of the reasons underlying international differences in the electoral success and executive political power of Socialist, Labour, and Communist parties is beyond the scope of this article. It should be noted, however, that the ‘politicization’ of the strike is most pronounced in France and Italy, where the state is heavily involved in establishing wages and conditions of work in the private sector, and, also, where left-wing parties have commanded a sizeable share of the vote for thirty years or more but have been largely frozen out of positions of executive power. I return to this point ahead.

²¹ In the United Kingdom, for example, the public sector share of the GNP (exclusive of defence) expanded in three waves: (1) 1944–48, from less than 20 per cent to 35 per cent as a result of the post-war Labour government's creation of the welfare state and nationalization; (2) 1964–68, from 35 per cent to 45 per cent, during the second post-war Labour government; and (although it is beyond the time frame of this study) (3) 1973–75, from 45 per cent to 55 per cent, as the third Labour government tried to deliver its side of the social contract. See the analysis in the Economist, 21 02 1976.Google Scholar

²² I was unable to find data on general government expenditure in Belgium and Italy before the Second World War and so the correlations in Fig. 5 are based on nine rather than eleven countries. Sources of the government expenditure data were: Statistical Office of the United Nations, National Income Statistics (New York: United Nations, 1950)Google Scholar, and Oshima, Harry T., ‘Share of Government in G.N.P. for Various Countries’, American Economic Review, LVII (1957), 381–90Google Scholar; OECD, National Accounts of OECD Countries 1961–1972 (Paris: OECD, 1974).Google Scholar Sources of the data for the other variables are given in earlier footnotes.

²³ A subsequent empirical analysis testing rival explanations of the growth of the public sector's share of the GDP over the period 1960–74 in seventeen industrial democracies also supports the historical-causal sequence implied by the first half of Figs. 5 and 6. See Cameron, David, ‘Open Economies, Electoral Politics, and the Expansion of the Public Economy: A Comparative Analysis’ (Yale University, mimeo).Google Scholar

²⁴ Notwithstanding the larger political visions of many left-wing trade-union leaders, most workers are probably mobilized for strike activity not by slogans about the workers' seizure of political power but by the narrower economic incentives usually associated with American ‘business unionism’. As Val Lorwin (The French Labor Movement (Cambridge, Mass.: Harvard University Press, 2nd printing, 1966), p. 215Google Scholar) put it in his study of French labour relations: ‘When they received wage adjustments, workers, including most union members, showed little determination to press for the institutional content of agreements about which their leaders talked’. Even the massive strikes of May–June 1968 in France (which were viewed largely as spontaneous ‘political’ events in many popular accounts) centred in the overwhelming majority of cases around traditional demands for wage increases and came to an end in the wake of sizeable wage concessions from the government and employers. See Durand, and Harft, , ‘Panorama statistique des grèves’Google Scholar, and Ross, George, ‘French Working Class Politics After May–June 1968: A New Working Class?’ (paper delivered at the annual meeting of the American Political Science Association at New Orleans in 1973).Google Scholar

²⁵ Headey has argued that a similar configuration of factors underlies trade-union acceptance of incomes policies. See Headey, Bruce W., ‘Trade Unions and National Wages Policies’, Journal of Politics, XXXII (1970), 407–39.CrossRefGoogle Scholar

²⁶ In some situations investigated by Calmfors the tax system had perverse effects, i.e. increases in money wages lead to decreases in after-tax real wages. See Calmfors, Lars, ‘Inflation in Sweden’, in Krause, L. B. and Salant, W. S., ed., Worldwide Inflation: Theory and Recent Experience (Washington, D.C.: Brookings, 1977).Google Scholar

²⁷ Jackson, D., Turner, H. A. and Wilkinson, Frank, Do Trade Unions Cause Inflation? (Cambridge: Cambridge University Press, 1972).Google Scholar Actually, Jackson et al. find that very large increases in money wages were necessary for workers to realize modest net-of-tax increases in real wages. I return to this point in the next section.

²⁸ See OECD, Revenue Statistics in OECD Member Countries 1905–1972 (Paris: OECD, 1975).Google Scholar

²⁹ Although most trade-union energy in the United States is still committed to private market activity, the process outlined in the text has occurred to a limited extent, though not enough to make a great impact on the volume of industrial conflict. As a result of labour's favourable experience with the Democratic party during and since the Roosevelt era, there has been a perceptible shift in the allocation of American trade unions' organizational resources toward the political arena.

³⁰ Jackson, , Turner, and Wilkinson, (Do Trade Unions Cause Inflation?)Google Scholar have already completed some suggestive research on the British case. They found that since the 1960s ‘the combined effect of income movements and changes in tax policy was to increase post-tax inequality between earned incomes,’ (p. 79) and that ‘the ratio of benefits-received to taxes-paid was falling’ (p. 83) for most employee and wage-earner households.

³¹ See fn. 14.

³² Johnson, J. and Timbrell, M., ‘Empirical Tests of a Bargaining Theory of Wage Rate Determination’, The Manchester School of Economic and Social Studies, XLI (1973), 140–67.Google Scholar

³³ The conclusion is controversial. For an alternative account of the recent wage explosion that does not rely on the private consumption/tax effect argument, see Hibbs, Douglas A. Jr., Trade Union Power, Labour Militancy and Wage Inflation: A Comparative Analysis (Cambridge, Mass.: Center for International Studies, Massachusetts Institute of Technology, 1977).Google Scholar

³⁴ The words ‘might interpret’ are used advisedly. There have been too many other constraints and complexities to support a firm inference about labour's revealed preference.

³⁵ For an extensive analysis, see Borre, Ole, ‘The General Election in Denmark, January 1975: Toward a New Structure of the Party System’, Scandinavian Political Studies, x (1974), 211–16Google Scholar, and Rusk, Jerrold and Borre, Ole, ‘The Changing Space in Danish Voter Perceptions 1971–1973’, European Journal of Political Research, 11 (1974), 329–61.CrossRefGoogle Scholar

³⁶ In September 1974, Prime Minister Paul Hartling's Agrarian-Liberal minority government introduced a 7 billion kroner cut in income taxes and government expenditure which was passed in parliament with the support of Radical Liberals, Conservatives, Centre Democrats, Christian People's party, Single Taxers, and some members of the Progress party. I am tempted to infer, without support from survey evidence, that the decline in ‘new bourgeois’ electoral support in the 1975 election was associated with this significant policy change. In any case, public sentiment continues to run strong against further extensions of the Danish welfare state: New York Times (28 09 1975)Google Scholar reports polls taken in 1975 indicating that 63 per cent of the public felt that the burden of taxation was excessive and that the welfare state had gone too far. The minority Social Democratic government, which assumed office after the 1975 election, acknowledged this by submitting a budget that further reduced social spending.

³⁷ See, for example, the analyses of Valen, Henry and Rokkan, Stein, ‘Norway: Conflict Structure and Mass Politics in a European Periphery’, in Rose, R., ed. Electoral Behavior: A Comparative Handbook (New York: Free Press, 1974), pp. 315–70Google Scholar, and Petersson, Olof, ‘The 1973 General Election in Sweden’, Scandinavian Political Studies, IX (1974), 219–28.Google Scholar

³⁸ Särlvik, Bo, ‘Recent Electoral Trends in Sweden’, in Cerny, K., ed., Scandinavia At The Polls (Washington, D.C.: American Enterprise Institute, 1977).Google Scholar

³⁹ As readers undoubtedly are aware, a bourgeois coalition led by the Centre party is now in power. It was inevitable that the Social Democrats would eventually lose an election; it is not possible to say without survey data to what extent their defeat at this time was due to the ‘welfare state’ issue. The expansion of nuclear power, economic stagnation, a trade-union plan to gradually ‘expropriate’ large firms, and the simple fact that the Social Democrats had been in power for so long, were also issues in the campaign.

⁴⁰ I base this statement on data for manufacturing production workers with average incomes, reported in OECD, Revenue Statistics in OECD Member Countries 1965–1972 (Paris: OECD, 1975). Appendix, Tables 2, 3 and 6.Google Scholar