Sociology of Strikes Research Paper

1. Strike Statistics

Systematic collection and publication of strike data started in most countries between 1870 and 1900. Until then, strike data were kept secret in the hands of Ministries of the Interior. After all, strikes were viewed as crimes under Combination Acts (Perrot 1974). Only when the conception of strikes shifted from one of crime to one of ‘social illness’ did strikes become the concern of specialized Labor Ministries. It is these Ministries that initiated systematic recording and publishing of strike data and that standardized definitions and collecting procedures. Early statistics provided rich information on each individual strike: the immediate causes of strikes, their outcomes, their organizational bases, the occurrences of violent incidents and the number of establishments affected by any given strike. There is none of that richness in today’s strike statistics. Furthermore, strike data are highly aggregated (Fisher 1973). Typically, they do not go much beyond three aggregate measures of strike activity: number of strikes (S ), number of workers involved (W ), and number of hours lost (H ).

On the basis of these three indicators, several different measures have been constructed in the literature (Forcheimer 1948, Knowles 1952, Shorter and Tilly 1974, Hibbs 1976): frequency or number of strikes per thousand employed workers; size or average number of strikers per strike (W/S ); duration or average number of hours lost per strikers (H/W ); and volume, defined as frequency × size × duration = (S/thousand employed workers) × (W/S) × (H/W) = (H/thousand employed workers); the Galambos’ and Evans’ composite strike index based on an unweighted average of S, W, and H (Galambos and Evans 1966). Graphically, these different components of strike activity can be represented in a three-dimensional space as a parallelepiped—the ‘shape’ of strikes— whose sides are given by the size, duration, and frequency of strikes and whose volume corresponds to the product of the three sides (Shorter and Tilly 1974, Hibbs 1976, Franzosi 1995).

2. Data Reliability

The reliability of strike statistics varies from country to country, but it is generally low (Shalev 1978). Some countries carry out data collection through specialized offices within Departments of Labor; others rely on police departments (e.g., Italy). Some use newspapers as sources of data; others rely on self-reporting, and still others base their statistics on police records. Different countries adopt different definitions of strikes, leading to difficulties in cross-national comparisons. It is generally difficult to gauge the bias of strike statistics under any data collection arrangement. Strike size can be just as easily overestimated as underestimated. If employers report strike size, they may be inclined to play down the strike to embarrass the unions or to portray a public image of peaceful labor relations (particularly vis-a-vis clients), but they may also wish to magnify the strike in order to justify future retaliatory actions or police/government intervention, or to contest the imposition of penalties by clients for late delivery. Union reported figures are more likely to be overestimated in a show of organizational strength. If police offices themselves estimate the size of the strike it may be no more than pure guesswork. Strike frequency, on the other hand, is likely to be always greatly underestimated regardless of the efficiency of collection: small, short, shop-level spontaneous work stoppages, that represent the bulk of all strike occurrences, probably escape recording, as scores of field investigations have shown (Turner et al. 1967, Korpi 1978, 1981, Batstone et al. 1978).

Bias in the data is not without consequences. When single strikes across a city or an industry are recorded separately, the number of strikes is inflated but the average strike size is reduced. To the extent that strike size is related to the workers’ organizational capacities, a bias in this measure would misrepresent the organizational basis of strikes (Batstone et al. 1978, p. 22). Under-reporting of small, plant-level strikes obscures the role of informal leadership and the existence of diffused, grass-root discontent, often directed against official unions as much as against employers (Korpi 1981). The systematic lack of information on strike tactics (e.g., walk-out, sit-down, checker-board, with demonstrations through the shop, marches to city hall), and on the broader spectrum of working class actions (such as slowdowns, work to the rule, overtime ban, rallies, demonstrations), fundamentally biases the nature and extent of working-class mobilization. In any case, exclusive emphasis on what workers do and on just one type of action (strike) distorts the broader social relations in which strikes occur. ‘It takes two to tango,’ Franzosi argues, ‘it takes at least two to fight’ (Franzosi 1995, pp. 15–18). Strikes emerged in the nineteenth century out of the incessant interaction among three main actors—workers, employers, and the state; it is this interaction that ultimately defined the rules and the acceptable forms of protest (Tilly 1978, p. 161). It is not possible to understand what workers do in isolation of what employers and the state do.

Data aggregation adds problems of its own. There is no way of telling which combination of strike frequency and duration produces any given figure of number of hours lost (Batstone et al. 1978, p. 23). A yearly total of 100,000 strikers could refer to 100,000 different workers or to the same 10,000 going on strike 10 different times (Batstone et al. 1978, p. 21). ‘Average’ measures, such as size and duration, may mask wide differences in the distribution of actual strike sizes and durations, in particular when used for international comparisons. The same ‘average’ size, in fact, could be the result of a few, large and official strikes for the renewal of industry-wide, collective agreements and a large number of small, plant-level strikes (such as the Swedish or the Italian case), or of a great deal of intermediate size strikes, as in the United States, where plant-level bargaining is the rule but workforce concentration is higher (Edwards 1981, p. 231). The same ‘volume’ can mask marked differences in each of its constituent components. Standardized indicators, such as frequency, have been computed using different deflating procedures: total civilian labor force, non-agricultural labor force, unionized labor force, employment figures, etc., leading to an unnecessary proliferation of strike indicators that stand in the way of comparability of findings (Stern 1978).

3. Patterns Of Strikes: What We Know

The ready availability of strike data for over a century has resulted in a rich scholarly tradition of (mostly quantitative) research on strikes, focusing on the interindustry propensity to strikes (e.g., Kerr and Siegel 1954), but notably on the temporal dynamics of strikes. From as early as the beginning of the 20th century there has been a concern with the relationship between strikes and the short-and medium-term fluctuations in the level of economic activity. Several statistical studies have shown that strike frequency follows the business cycle and the movement of unemployment in particular—the higher the level of unemployment, the lower the number of strikes (see the seminal work by Ashenfelter and Johnson 1969). As American historian David Montgomery wrote: ‘Large scale strikes erupted [in the early 1900 in America] whenever the level of unemployment fell off sufficiently to give the strikers a ghost of a chance of success’ (Montgomery 1980, p. 94).

Indeed, the dependence of strike frequency on the business cycle is one of the clearest findings of econometric strike research. Yet, economic theories of strikes have not gone unchallenged. For one thing, economic determinants of strikes can be statistically significant without being the main or exclusive influence (Vanderkamp 1970). Furthermore, while there is strong statistical evidence in favor of the relationship between strike frequency and the business cycle, it is not so for the other strike measures: size, duration, and volume. While strike frequency is negatively related to the level of unemployment, there is evidence supporting a positive relationship for duration (e.g., Turner et al. 1967). If economic determinants seem to play a significant role only on frequency, and each strike component measures a different aspect of strike activity, what are the determinants of strike size, duration, and volume?

No doubt, it is workers’ organizational capacities that have most deeply affected the overall shape of strikes, size and duration, in particular. As early as 1911–13, French economist March showed the close statistical connection between strike size and unionization. But it is Shorter and Tilly (1974) who most systematically investigated the effect of organization on strike shape in what to date remains a landmark in strike research. For one thing, national, large-scale strikes are only possible in the presence of national, large-scale labor organizations. But even at the local level, the organizational basis of trade unions helps to bring out on strike a larger portion of the workforce, leading to larger average strike sizes. Labor organizations have also contributed to reduce the average strike duration. The rise of labor organizations has typically shifted the timing of strikes from economically ‘bad’ times to ‘good’ times. Through organization, as Hobsbawm put it, industrial workers learned ‘the common sense of demanding concessions when conditions are favorable, not when hunger suggests’ (Hobsbawm1964, pp. 126, 144). By timing strikes during periods of prosperity unions also contributed to shortening the average duration of strikes. In fact, strikes are more likely to be long and drawn-out during periods of recession when employers are under no pressure to seek a quick settlement.

Changes in the timing of strikes and the consequent shortening of their average duration have also brought about changes in the likelihood of successful settlements and of the occurrence of violence during a strike. Long disputes are likely to be unsuccessful for two reasons: (a) because they tend to occur in economically unfavorable times and (b) because they have a greater likelihood of being violent (and violent strikes decrease the chances of success) (Snyder and Kelly 1976).

Yet the statistical significance of a unionization variable in econometric models of strike size hardly offers a causal explanation. After all, the reverse causal order has produced statistically significant results in union growth models, where unionization becomes the ‘dependent’ variable and number of strikers the ‘independent’ one (Ashenfelter and Pencavel 1969, Bain and Elsheikh 1976). In the end, we may have done no more than observe a strong correlation (Goetz-Girey 1965, pp. 146–7, Edwards 1981, p. 9). The exclusive reliance on highly aggregated strike and unionization figures in econometric work has obscured the causal mechanisms in the relationship between these different aspects of mobilization processes; it has also exaggerated the role of formal organization at the expense of informal shop-floor leadership (Batstone et al. 1978). Indeed, the use of temporally disaggregated firm-level data shows how rises in unionization follow (rather than precede) surges in the level of strikes, successful strikes in particular (Franzosi 1995, pp. 124–34).

The political position of labor in national power structures has also deeply affected the shape of strikes. Both strike volumes (Hibbs 1978) and strike sizes (Korpi and Shalev, 1980) have gone down in countries where labor-oriented social-democratic parties acquired stable and lasting control over governments, such as in the Scandinavian social democracies and in Austria (Korpi and Shalev, 1980 p. 320). Direct access to political power offers labor alternative and less costly means to achieve a more favorable distribution of resources: the government machinery itself, rather than the strike (Shorter and Tilly 1974, Hibbs 1976, Korpi and Shalev, 1980). When the working class has control of the government, the locus of conflict over the distribution of resources, national income in particular, shifts from the labor market and the private sector, where strike activity is the typical means of pressure, to the public sector, where political exchange prevails. Short-lived labor governments, on the other hand, or outright exclusion from the government of communist parties, are associated with higher, rather than lower, levels of strike activity (Hibbs 1976, Paldam and Pedersen 1982). Franzosi’s work on postwar Italy shows that the ‘historic compromise’ strategy of the Italian Communist Party drastically changed the shape of strikes between 1975 and 1978 during the Governments of National Solidarity (Franzosi 1995).

4. The Withering Away Of Strikes And Cycles Of Strikes

In the 1950s arguments about the ‘embourgeoisement’ of the working class and the ‘end of ideology’ were popular. Postwar prosperity and the institutional mechanisms for conflict resolution and collective bargaining, the arguments went, have resulted in permanent industrial peace, ‘the withering away of the strike, the virtual disappearance of industrial conflict in numerous countries’ (Ross and Hartman 1960, p. 6, Dahrendorf 1959, pp. 257–79, Kornhauser 1954, pp. 524–5). History was to prove these arguments wrong, with the resurgence of conflict in the 1960s to unprecedented levels in those same numerous countries (Crouch and Pizzorno 1978). But once again the ‘virtual disappearance’ of strikes in the 1980s and 1990s rekindled the debate on the withering away of strikes, with a novel, and perhaps more powerful trump to the argument: the very working class has now virtually disappeared, with the drastic shrinking of manufacturing and the soaring of services. Since service workers are largely white collar, with a high percentage of women—both actors being traditionally little strike prone—changes in the occupational structure should result in the ‘withering away of strikes’ once and for all.

These arguments are compelling. No doubt the change in the location of conflict from industry to services has deeply changed the nature of labor conflicts. For one thing, it has changed the very nature of the two traditional actors involved in labor conflicts: workers and employers. The typical worker is no longer male and blue-collar, and the typical employer is public rather than private: the state has become the largest service employer in most Western countries (for comparative evidence, see Tiziano 1987). Historically, the state intervened in labor disputes in- directly, regulating conflict via legislative means, only rarely taking a more direct role (either to repress or to arbitrate). As employer the state cannot escape being drawn directly into service disputes as a principal bargaining actor. Second, the ‘essential’ nature of many services (transportation, health care, etc.) and the monopoly conditions under which most services are provided has dramatically increased strikers’ disruptive power (i.e., their ability to inflict severe losses to the counterpart at minimal cost to strikers). The more centralized and monopolistic a service is, and the larger the number of users affected by a service strike. Disruption is always intrinsic to service disputes, regardless of the intentions of the actors. The law-and-order potential of highly disruptive strikes further draws the state into service disputes, even when these disputes are located in the private sector.

But from here to argue in favor of the ‘withering away of strikes’ is to ignore fundamental characteristics of strikes. First, strikes, as we have seen, are closely linked to wage labor. The nature of employment (wage labor) does not fundamentally change with changes in the occupational structure. Technicians, teachers and college professors, nurses and doctors, pilots and air controllers have shown to be as willing to go on strike as blue-collar workers. Second, strikes are a cyclical phenomenon. Strike research has made that quite clear (Franzosi 1995, pp. 345–8). Short-term seasonal fluctuations weave in and out of medium-term fluctuations of some years duration, mostly related to the business cycle, which in turn ziz-zag around long-term fluctuations of some 40–50 years in length. Research on these long-term cycles of strikes suggests that these ‘strike waves’ are linked to the long-term movements of capitalist economies (the so-called Kondratieff’s cycles; Mandel 1980, Cronin 1980). During wave years the characteristics of strikes are different than in surrounding years: strikes tend to be longer; they involve a higher proportion of the workforce, both at the plant and at the industry level, and a higher number of establishments (Hobsbawm 1964, p. 127; Shorter and Tilly 1974, pp. 142–6; Franzosi 1995). Strike tactics become more radical and innovative during strike waves, being mainly plant-based and in the hands of informal leaders, outside the control of the union and the institutionalized rules of the game. Demands become radicalized as well. Strike waves also often bring about legislative reforms (Shorter and Tilly 1974, p. 145; Franzosi 1995, pp. 302–6). Finally, the resurgence of conflict after long periods of peace is typically sudden and overwhelming. It takes employers, the state, and even the unions by surprise (French May of 1968 docet or the Italian autunno caldo of 1969).

If these theories are correct, the decline in strike activity in recent decades is only the result of a long- term cycle rather than a permanent downward trend leading to the withering away of strikes. If the past is anything to go by, the mobilization process will catch us by surprise with its sudden, sharp, and sweeping bouts of conflict: 1840s, 1880s, 1920s, 1960s … we may not have to wait long to be surprised.

Bibliography:

1. Ashenfelter O, Johnson G E 1969 Bargaining theory, trade unions, and industrial strike activity. American Economic Review 59(40): 35–49
2. Ashenfelter O, Pencavel J H 1969 American trade union growth: 1900–1960. Quarterly Journal of Economics 83: 434–48
3. Bain G S, Elsheikh F 1976 Union Growth and the Business Cycle. Blackwell, Oxford, UK
4. Batstone E, Boraston I, Frenkel S 1978 The Social Organization of Strikes. Blackwell, Oxford, UK
5. Cronin J E 1980 Stages, cycles and insurgencies: The economics of unrest. In: Hopkins T K, Wallerstein I (eds.) Processes of the World-system. Sage, Beverly Hills, CA, pp. 101–18
6. Crouch C, Pizzorno A (eds.) 1978 The Resurgence of Class
7. Conflict in Western Europe Since 1968. Holmes and Meier, New York
8. Crouzel A 1887 Etude historique, economique, juridique sur les coalitions et les greves dans l’industrie. A. Rousseau, Paris
9. Dahrendorf R 1959 Class and Class Conflict in Industrial Society. Stanford University Press, Stanford, CA
10. Edwards P K 1981 Strikes in the United States 1881–1974. Blackwell, Oxford, UK
11. Fisher M 1973 Measurement of Labour Disputes and Their Economic Eff OECD, Paris
12. Forcheimer K 1948 Some international aspects of the strike movement. Bulletin of the Oxford Institute of Statistics 10: 9–24
13. Franzosi R 1995 The Puzzle of Strikes: Class and State Strategies in Postwar Italy. Cambridge University Press, Cambridge, UK
14. Galambos P, Evans E W 1966 Work-stoppages in the United Kingdom, l95l–l964: A quantitative study. Bulletin of the Oxford University Institute of Economics and Statistics 28: 33–55
15. Goetz-Girey R 1965 Le mouvement des greves en France. Editions Sirey, Paris
16. Hibbs D Jr 1976 Industrial conflict in advanced industrial societies. American Political Science Review 70(4): 1033–58
17. Hibbs D Jr 1978 On the political economy of long run trends in strike activity. British Journal of Political Science 8(2): 153–75
18. Hobsbawn E 1964 Labouring Men. Basic Books, New York
19. Kerr C, Siegel A 1954 The interindustry propensity to strike. An international comparison. In: Kornhauser A, Dubin R, Ross A (eds.) Industrial Conflict. McGraw-Hill, New York, pp. 189–212
20. Knowles K G J C 1952 Strikes, BA Study in Industrial Conflict. Blackwell, Oxford, UK
21. Kornhauser A 1954 The undetermined future of industrial conflict. In: Kornhauser A, Dubin R, Ross A (eds.) Industrial Conflict. McGraw-Hill, New York, pp. 519–26
22. Korpi W 1978 The Working Class in Welfare Capitalism. Routledge and Kegan Paul, London
23. Korpi W 1981 Unofficial strikes in Sweden. British Journal of Industrial Relations 19(1): 66–86
24. Korpi W, Shalev M 1980 Strikes, power, and politics in the Western nations, 1900–1976. In: Zeitlin M (ed.) Political Power and Social Theory. JAI Press, Greenwich, CT, Vol. 1, pp. 301–34
25. Mandel E. 1980 Long Waves in Capitalist Development. Cambridge University Press, Cambridge, UK
26. Montgomery D 1980 Strikes in nineteenth-century America. Social Science History 4(1): 81–104
27. Paldam M, Pedersen P 1982 The macroeconomic strike model: A study of seventeen countries, 1948–1975. Industrial and Labor Relations Review 35(4): 504–21
28. Perrot M 1974 Les ouvriers en greve (France 1871–1890). Mouton, Paris
29. Ross A H, Hartman P T 1960 Changing Patterns of Industrial Conflict. Wiley, New York
30. Shalev M 1978 Lies, damned lies, and strike statistics: The measurement of trends in industrial conflicts. In: Crouch C, Pizzorno A (eds.) The Resurgence of Class Conflict in Western Europe Since 1968. Volume 1. National Studies. Holmes and Meier, New York, pp. 1–20
31. Shorter E, Tilly C 1974 Strikes in France, 1830–1968. Cambridge University Press, Cambridge, UK
32. Snyder D, Kelly W R 1976 Industrial violence in Italy, 1878– 1902. American Journal of Sociology 82(1): 131–62
33. Stern R N 1978 Methodological issues in quantitative strike analysis. Industrial Relations 17(1): 32–42
34. Tilly C 1978 From Mobilization to Revolution. Addison-Wesley, Reading, MA
35. Tiziano T (ed.) 1987 Public Service Labour Relations: Recent Trends and Future Prospects. International Labour Organisation, Geneva
36. Turner H A, Clack G, Roberts G 1967 Labour Relations in the Motor Industry. Allen & Unwin, London
37. Vanderkamp J 1970 Economic activity and strike activity. Industrial Relations (University of California, Berkeley) 9(2): 215–30