The aim of the thesis was to investigate the causes of union membership decline in Britain between 1980 and 1998. It sought to answer two key questions. First, to what extent was union membership decline the result of structural factors (that is, economic, political, legal, attitudinal, and compositional changes) compared to the failure of unions to cope with a changing environment? Second, if structural determinants were important, what was the relative importance of each in accounting for membership decline? The methodology chosen to answer these questions was analysis of micro data from workplaces and individuals, using a modification of the stocks and flows model proposed by Freeman (1988). For reasons of expediency and convenience, the period was split into three time periods: 1980 to 1984, 1984 to 1990, and 1990 to 1998, and the workplace-level processes of decline were investigated for each period in turn. Finally, the individual-level processes of decline were analyzed over the period 1991 to 1997. By investigating the individual- and workplace-level processes of decline, I hoped to make more informed judgements about the causes of membership decline. The results and analysis were firmly rooted in the context provided by the existing rich quantitative and qualitative literature on union decline.

Summary of the Results

1980Ð1984

     Between 1980 and 1984 the lion's share of union decline was attributable to the declining proportion of workers working in workplaces with union recognition and closed-shop agreements. These changes were most likely the result of the closure of large, highly unionized workplaces in the production sector during the recession of the early 1980s. As such, union decline in this period can be traced back to the consequences of the macroeconomic and industrial policies of Margaret Thatcher's Conservative government.

1984Ð1990

     Around half of the decline in this period can be attributed to the decline of union recognition. The decline in union recognition was mainly the result of lower levels of recognition amongst workplaces established after 1980. This can partly be attributed to the abolition of the statutory recognition procedure and the withdrawal of other state supports for collective bargaining, combined with other legislative changes that restricted the ability of trade unions to coerce employers into recognition. However, it is difficult to disentangle the impact of the law from the impact of secular economic changes: increasingly international product markets, technological change, deregulation, and privatization. Union failure to invest in new organizing also may have played a role. Declining collective bargaining coverage and union derecognition probably accounted for one quarter of the decline. The decline of the closed shop, predominantly the result of legislation, accounted for around one fifth of the decline. Compositional change, predominantly the decline of the public sector's employment share, accounted for around one tenth of the decline. This compositional change can partly be attributed to government policies of privatization and cuts to public sector budgets.

1990Ð1998

     Decline of union coverage explained around half of the decline in union membership over this period. This was a result of a lower incidence of union recognition in workplaces established after 1980 and the abandonment of collective bargaining in continuing workplaces that recognized unions (likely to be indicative of managerial policies of union marginalization). Once again, this change is likely to have been the result both of legislative restrictions on trade unions and secular economic changes that changed the costs and benefits of union recognition and collective bargaining for firms. Behavioral change, likely to indicate increased free-riding, accounted for most of the remainder.

     The importance of behavioral change over this period (in contrast to the 1980s, when behavioral change accounted for a minimal amount of the decline in density) can probably be attributed to two factors. First, there was declining union effectiveness in workplaces where management chose to pursue union marginalization strategies, resulting in the social custom of union membership broke down. Managers had an increased incentive to pursue union marginalization strategies because they faced more intense product market competition and more demanding shareholders. The secular economic changes that created this incentive also provided the means to weaken unions, for example, through labor-saving new technologies and the threat of moving production overseas. Legal restrictions on the trade unions also helped to facilitate marginalization. Workers and their unions often lacked the ideological resources that would have allowed them to challenge union marginalization and work restructuring. The generally low level of inflation and positive wage growth also lessened the incentive to challenge management for higher pay. Second, in new workplaces with union recognition, low levels of density suggest that unions struggled to establish the social custom of union membership.

1991Ð1997 (Individuals)

     Results from a panel of individuals were reassuringly similar to results from workplaces over a similar period. Around half of the decline in union density was attributable to the decline of union coverage, with most of the remainder attributable to increased free-riding. Changes in perceptions of the efficacy of strong unions accounted for a minimal amount of the decline in membership. New workers were less likely to work in a union job and more likely to free-ride than the workers they replaced, but continuing workers also became more likely to free-ride. Most continuing workers who left union membership did so when they changed jobs.

Evaluation

     It is important to be aware of the potential shortcomings and weaknesses of the analysis. First, Disney (1990) showed that unless econometric methods that account for the unobservable preferences of workers and employers that lead workers to be sorted into union and non-union jobs (for example, tobit or interval regression) are used when estimating the determinants of union membership, results will be biased. However, the results of interval regression do not lend themselves to use in decomposition analysis, so I have used the simpler, but strictly technically inappropriate, weighted least squares/linear probability regression.

     My justifications for adopting this approach were twofold. First, others have done the same thing (Green 1992, Machin 2004, Bryson and Gomez 2005). Second, even if the technically correct econometric methodology had been employed, the results may still be biased by omitted variables that are not randomly distributed. The work of Arulampalam and Booth (2000) suggests that omitted variables that are correlated with the individual fixed effects do bias estimates of the determinants of union membership. In any case, following from the critical realist critique of quantitative methodologies in positivist research, survey data like those analyzed here present at best crude approximations of complex underlying social processes and realities, so any results, regardless of the "correctness" of the methodology, would need to be treated with caution. The results should be read as indicators of underlying trends in the general population rather than a precise diagram of exactly how union membership decline happened.

What Caused Union Decline: Structural Change
or Union Failure?

     The charge that unions themselves were at least partly responsible for union membership decline has been made most clearly by Kelly and Heery (1989) and Kelly (1990). The central charge was that unions failed to invest sufficient resources in organizing and recruiting nonmembers; when investment was made, union tactics were poor so the campaigns were ineffectual. Yet to maintain the number of members that they had in 1980, unions would have had to recruit an extra 293,000 members per year between 1980 and 1998. I estimated the number of campaigns needed to achieve this level of membership increase each year and the financial resources needed to finance these campaigns. The numbers are necessarily rather arbitrary and imprecise, but they suggest that unions would need to increase organizing activity by at least a factor of ten, with at least 10 percent of union income from members needing to be spent on organizing. Current union organizing activity will have at best only a marginal impact on membership levels. I argue that both theory and evidence suggest that unions are unable to make that level of investment.

How Important Were the Different Elements
of Structural Change?

     Which elements of structural change (workforce composition, the business cycle, secular changes to economic organization, the legal regulation of industrial relations, the wider political climate, and employee attitudes and values) best explain membership change? The element of structure that can be dismissed most easily is the business cycle. Purely on empirical grounds, business cycle models failed to predict the severity and extent of the downturn in union membership. Union membership decline continued through recessions (which we would expect) and through periods of strong growth and inflation, which, other things being equal, we would expect to be associated with union growth. Carruth and Disney (1988) and Disney (1990) attempted to save business cycle theory by arguing that the business cycle label was a misnomer because the key influence on employee desire for unionization was the level of real wage growth; thus "macroeconomic conditions" would be a better label for the approach. Guided by economic theory, Disney (1990) argued that if real wage growth averaged 3 percent per annum, "steady state" union density would be 27 percent because there would be little demand for union membership amongst workers largely satisfied with their economic progress. As the 1980s and 1990s were a period of positive real wage growth, the 29 percent membership density reached in 1998 may well represent Disney's "steady state."

     The key problem with this argument is a theoretical one. Disney argues that strong real wage growth "causes" union membership decline because workers no longer desire union membership. However, real wage growth is ultimately dependent on productivity growth, and economic theory posits that unions will affect productivity; strong unions will lower productivity growth by choking off investment and implementing restrictive practices. Therefore, strong real wage growth may be a symptom of union weakness and decline rather than a cause of it. The interesting question then is, What has caused strong real wage growth, and how is it related to union decline? My argument is that both strong real wage growth and union decline are bound up with secular changes to economic organization and wider political and legal changes that resulted in radical alterations in the system of industrial relations. Therefore, while change in real wage growth may successfully predict change in union membership, it is not the cause of union membership change.

     It is also relatively straightforward to dismiss attitudinal change as a cause of decline. The British Social Attitudes Survey (BSAS) has charted changes in the wider political attitudes of the workforce since 1983, and the British Household Panel Survey has charted perceptions of the efficacy of strong trade unions since 1991. The BSAS shows a small "rightwards" shift in attitudes, with workers becoming slightly less collectivist in outlook. However, Bryson and Gomez (2005) have shown that the impact of this change on union membership has been minimal. Since 1991 perceptions of the efficacy of strong trade unions actually increased among continuing employees. Employees who joined the workforce between 1991 and 1997 were less likely to have positive perceptions of the efficacy of strong unions but also less likely to have negative perceptions. This greater indifference to unions among new workers does explain a proportion of union decline—but only a small proportion.

     Crouch (2001) and Towers (1989) both argued that compositional change was an important contributory factor in explaining union membership decline because the decline of the manufacturing industry, in particular, depleted unions "core membership reserves" (Crouch 2001). The results reported in chapters 3 to 5 of this thesis, which come from workplaces rather than individuals and include measures of union coverage, typically find that a much smaller proportion of decline was attributable to compositional change. Just one tenth of the overall decline in union density between 1980 and 1998 in workplaces with more than twenty-five employees could be attributed to compositional change.

     So if attitudinal shifts, compositional change, and the business cycle cannot adequately account for union decline, what structural factors were important? My explanation of union membership decline rests on three factors: (1) secular changes to the economic environment, partly related to changing trade patterns and partly the result of technological change (Freeman 1995); (2) political changes that delegitimized trade unions, stripped them of ideological resources, made the workforce more pessimistic about the prospects of effective collective action, and exposed unions to the full force of secular economic change; and (3) legal changes that were the result of political change and restricted the ability of trade unions to mobilize their members and use the strike weapon. All of these changes are attributable, at least in part, to the agency of Margaret Thatcher's Conservative government.

Secular Economic Change

     Over the last thirty years increased trade with the developing world and rapid technological change as a result of advances in information and communications technology have wrought far-reaching changes on the economies of advanced capitalist economies like Great Britain. The impact of these changes for unions have been far reaching.

     As product markets became more competitive, so firms became less able to pass on increases in labor costs to consumers; the wage elasticity of demand for labor must have increased, thereby reducing the benefits and increasing the costs of unionization for workers. At the same time, the costs and benefits of union recognition for firms changed. It was no longer viable for wages to be taken out of competition through collective bargaining, and the risks associated with unions raising labor costs became greater. Technological change has made it easier and cheaper for firms to replace labor with capital. These technological changes have disproportionately impacted workers performing routine but skilled jobs (Autor, Levy, and Murnane 2003; Goos and Manning 2003)—for example, skilled machinists in the engineering industry, who have been replaced by computer controlled robots—and possibly lower-skilled workers (Machin 2001). The impact of these changes has been disastrous for trade unions because it was precisely the skilled manual workers whose bargaining position has been most weakened by these changes who once formed the vanguard of the labor movement. The resulting environment of low inflation has removed one of the key collective grievances that acted as a spur to unionization throughout the twentieth century: the threat of real wages being eroded by inflation. Finally, these economic changes have made the economic interests of the working class more heterogeneous, a change that reinforced and was reinforced by a revolution in political and economic ideas that, discredited and delegitimized ideas of Marxism, socialism, and social democracy—ideas that once provided space for trade unions to operate in and equipped workers and their unions with the ideological resources to mount a challenge to managerial authority.

     These changes have affected unions in all advanced capitalist economies. However, British unions suffered more severely than most other union movements because the traditions of voluntarist, decentralized collective bargaining left British unions peculiarly exposed to political and economic changes, and because the Conservative government, elected in 1979, rushed to embrace economic change as a way of weakening trade unions and making the British economy more competitive (Western 1995, Ebbinghaus and Visser 1999, Turner 2003).

Political Change

     Political change affected unions in four ways. First, changes to economic policy initiated by the Thatcher government exposed the U.K. economy, labor market, and unions to the full force of secular economic change. Second, changes to the way in which the public sector was managed impacted negatively on unions. The size of the public sector shrunk as a result of budget cuts, compulsory competitive tendering, and privatization. Budget cuts led to the erosion of terms and conditions and work intensification. When workers sought to rebel against these changes through industrial action, the government resolved to face down demands in order to secure "demonstration effects" that would deter other groups of workers from striking. Consequently, union attempts to mobilize workers against collective grievances became fewer, leading to the corrosion of the social custom of membership, with the result that free-riding increased. Third, as discussed above, the government changed the terms of ideological debate in ways that disabled to unions. Fourth, radical changes to labor law were introduced, which also placed serious restrictions on the ability of unions to mobilize workers.

Legal Change

     It is difficult to disentangle the effects of changes to the legal regulation of industrial relations from the wider aspects of government hostility to trade unions, of which the legal changes were an integral part. However, it is apparent that restrictions on the ability of unions to wield the strike weapon and the increased bureaucratic hurdles that unions had to jump through before calling a strike deprived workers of any element of surprise, thereby handling tactical advantage to the employer. This made union officials less willing to resort to industrial action and workers less willing to vote for it or participate. This will have contributed to the erosion of the social custom of membership, causing free-riding to increase. The outlawing of the closed shop will also have contributed to this increase.

     Taken together these changes amounted to a fundamental change to the system of industrial relations in Britain, and it was this systemic shock that caused the erosion of union membership. In part this systemic shock was the result of secular economic change, but it was mainly the result of the agency of the Conservative government that held power between 1979 and 1997.

Retheorizing Union Membership Change

     The thesis concluded by retheorizing union membership change in the light of the preceding results and analysis. Following the approach of Bain (1970), it argued that the key determinants of union membership density are the level of union recognition, the elasticity of demand for labor, the ideological resources available to unions, and the nature and stock of grievances (that is, are grievances predominantly collective in nature, for example, related to the erosion of pay by inflation), with union recognition being determined by the elasticity of demand for labor, ideological resources, the stock of grievances, plus the level and nature of product market competition and the level of state support for collective bargaining.

Implications for the Future of Unions

     As the key argument of the thesis is that union membership is structurally determined and that union decline was structurally determined, it follows that the actors in the system of industrial relations, particularly trade unions, will have only limited ability to bring about their own renewal and revitalization. Unless the wider environment changes in ways that favor trade unions—for example, some reduction in product market competition—the re-birth of serious inflationary pressures, or a more interventionist economic approach by government, perhaps linked to a revival in social-democratic and socialist ideologies—it is difficult to envisage any revival in union membership.

References

Arulampalam, W., and A. Booth. 2000. PPersistence in Union Status: A Longitudinal Study of Young Men in Britain." Journal of Applied Econometrics, Vol. 15 (August), pp. 289Ð310.

Autor, D., F. Levy, and R. Murnane. 2003. "The Skill Content of Recent Technical Change: An Empirical Investigation." Quarterly Journal of Economics, Vol. 118, no. 1, pp. 273Ð333.

Bain, G. S. 1970. The Growth of White-Collar Unionism. Oxford: Clarendon.

Bryson, A., and R. Gomez. 2005. "Why Have Workers Stopped Joining Unions? Accounting for the Rise of Never-Membership in Britain." British Journal of Industrial Relations, Vol. 43, no.1, pp. 67Ð92.

Carruth, A., and R. Disney. 1988. "Where Have Two Million Trade Union Members Gone?' Economica, Vol. 55, no. 1, pp. 1Ð19.

Crouch, C. 2001. "The Snakes and Ladders of Twenty First Century Trade Unionism." Oxford Review of Economic Policy, Vol. 16, no. 1, pp. 70Ð83.

Disney, R. 1990. "Explanations of the Decline in Trade Union Density in Britain: An Appraisal." British Journal of Industrial Relations, Vol. 28, no. 2, pp. 165Ð177.

Ebbinghaus, B., and J. Visser. 1999. "When Institutions Matter: Union Growth and Decline in Western Europe." European Sociological Review, Vol. 15, no. 2, pp. 135Ð158.

Freeman, R. 1988. "Contraction and Expansion: The Divergence of Private Sector and Public Sector Unionism in the United States" Journal of Economic Perspectives, Vol. 2, no. 2, pp. 63Ð88.

Freeman R. 1995. "Are Your Wages Set in Beijing?" Journal of Economic Perspectives, Vol. 9, no. 3, pp. 15Ð32.

Goos, M., and A. Manning. 2003. "Lousy Jobs and Lovely Jobs: The Rising Polarization of Work in Britain." Centre for Economic Performance Discussion Paper 604, London School of Economics.

Green, F. 1992. "Recent Trends in British Trade Union Density: How Much of a Compositional Effect?" British Journal of Industrial Relations, Vol. 30, no. 3, pp. 445Ð458.

Kelly, J. 1990. "British Trade Unionism 1979Ð89: Change, Continuity and     Contradictions." Work, Employment and Society, (May), pp. 29Ð65.

Kelly, J., and E. Heery. 1989. "Full-Time Officers and Trade Union Recruitment." British Journal of Industrial Relations, Vol. 27, no. 2, pp. 196Ð213.

Machin, S. 2001. "The Labour Market Consequences of Technological and Structural Change." Oxford Bulletin of Economics and Statistics, Vol. 63, no. 5, pp. 753Ð766.

Machin, S. 2004. "Factors of Convergence and Divergence in Union Membership." British Journal of Industrial Relations, Vol. 42, no. 3, pp. 423Ð438.

Towers, B. 1989. "Running the Gauntlet: British Trade Unions Under Thatcherism, 1979Ð1988." Industrial and Labor Relations Review, Vol. 42, no. 2, pp. 163Ð188.

Turner, L. 2003. "Reviving the Labor Movement: A Comparative Perspective." In D. Cornfield and H. McCammon, eds., Labor Revitalization: Global Perspective and New Initiatives: Research in the Sociology of Work. Greenwich, CT: JAL, vol. 11, pp. 23Ð58.

Western, B. 1995. "A Comparative Study of Working Class Disorganization: Union Decline in Eighteen Advanced Capitalist Countries." American Sociological Review, Vol. 60, no. 2, pp. 179Ð201.