|
| |
Fighting for Equality
From: London School of Economics and Political Science
| By:
David Metcalf |
EDITOR'S INTRODUCTION |
 Unions have traditionally been seen as the guardians of egalitarian pay structures. This is achieved mainly through pay dispersion and rigid pay structures that reward jobs, not individuals. David Metcalf, director of the Centre for Economic Performance at the London School of Economics and Political Science, argues that the effects of such policies are even more profound: they compress the pay structure between women and men, between blacks and whites, and between the healthy and those with health problems. |
 here are three routes to this greater equality of pay in the organised, or unionised, sector. First, unions reduce pay dispersion within establishments. Unions prefer a single rate of pay for each occupational group, whereas in non-union firms with individual pay determination supervisors decide pay levels within a range. Unions also prefer seniority-based progression of rates. These preferences stem from unions' desire for objective standards, where pay goes with the job, because of concerns about favouritism, discrimination and the measurement of the "true" contribution where pay is subjectively related to the "merit" of the individual. Workers' solidarity is also likely to be greater when workers receive roughly the same pay than when they get different amounts. |
Union wage policies also lead to a narrower wage dispersion in the organised sector compared with the unorganised sector. However, the strength of this channel, which operates across firms and workplaces, has been weakened as collective bargaining has become more decentralised. At the beginning of the last century (in 1902), the Webbs introduced the concept of "the common rule" into the vocabulary of industrial relations and used it to define trade union objectives. The fixing of a standard rate of pay--the rate for the job--was the pivotal common rule. There are two rationales for the common rule. When firms compete in the same market, both employer and worker interests can be expected to favour a standard rate. The firm does not want a labour contract that is more expensive than its competitors. For the worker, it takes the wages out of competition--there will be no undercutting. Union solidarity is also more difficult to maintain if some workers are paid markedly more than others for the same job. |
The strength of this common rule has ebbed away as the locus of collective bargaining in the private sector has switched from national multi-employer agreements to firm or workplace agreements. Around a quarter of private sector employees are covered by collective agreements. Of those, less than a tenth are covered by national agreements. Consequently, the dispersion of pay in the organised sector is likely to be higher now than it was two decades ago. |
Nevertheless, the organised sector is still likely to have less dispersion, other things equal, than the unorganised. Multi-employer bargaining still exists in parts of printing, textiles, clothing and construction. But even in sectors where national bargaining has disintegrated, unions' own internal organisation often permits or encourages pay comparisons across companies. This comes via research support for collective bargaining, the way lay officers are taught to formulate pay claims, industry-level forums, and the involvement of full-time officials in local pay negotiations. |
Unions also contribute to lower pay dispersion because they operate a de facto minimum-wage policy via collective bargaining. In local government, for example, the lowest basic rate in the 1997 national agreement was £ 4 per hour; currently, public sector unions are looking to a £ 5 per hour minimum. The Transport and General Union reports a similar lower bound negotiated for some industrial cleaners, knitwear operatives, bar staff and retail workers. |
More equal than others
The upshot of these union wage policies--which operate within and among firms and which cut off the bottom end of the pay distribution--is that we expect unionised workers to have a more equal distribution of wages than their non-union counterparts. Some evidence on this is given in Table 1 , taken from the Labour Force Survey covering more than 16,000 employees each quarter. |
Two measures of the dispersion of earnings are presented in Table 1: the standard deviation (a measure of the variation from the average) and the 90th percentile minus the 10th percentile (i.e., the pay of the person 10 percent from the top of the wage distribution minus the pay of the person 10 percent from the bottom of the distribution). Consider initially the raw earnings data. On both measures, the union pay distribution is more concentrated than the non-union pay distribution. Of course, this could be nothing to do with unions--it might simply reflect the fact that unionised workers and jobs are more similar than non-union ones. To allow for this possibility the bottom panel of the table presents evidence adjusted to take account of this. |
There are two clear findings from these adjusted figures. First, considerable pay dispersion remains: the adjusted figures show pay dispersion which is about two-thirds of that in the "raw" figures. Second, it remains the case that, even when we compare essentially homogenous employees in similar workplaces and jobs, the pay dispersion is lower for union members than for non-union employees. Thus union wage policies do matter--the lower dispersion in pay among unionists is not just because they are more similar than non-unionists. |
Jobs, not people
One reason for this lower dispersion in the organised sector is because in unionised firms pay is based more on jobs and less on individual characteristics. Hence the characteristics of individual workers which raise pay (such as experience or qualifications) have a smaller effect in unionised workplaces, and so, in turn, inequality is smaller in union than in non-union workplaces. Consider, for example, the payoff to extra human capital. The returns to labour market experience indicate that, as compared with younger workers, those union members aged 25-54 (and thus with greater labour market experience) earn only 37 percent more, while such prime age non-union members earn 48 percent more. Likewise, the returns to higher qualifications are greater and the profile is steeper among the non-union group than for union members. Unions may nevertheless wish to ponder the implications of this finding: it could be that workers sort themselves, or are sorted by their potential employers, such that the more able prefer to work in non-unionised firms where they can get a higher payoff to their costly investment in human capital. |
Unions and low pay
The 1998 Workplace Employment Relations Survey asked the nearly 2,000 respondent workplaces what percentage of employees earned below £ 3.50, a convenient benchmark, as the NMW of £ 3.60 was introduced a year later. Cross tabulations confirm that union recognition is associated with a lower incidence of low pay. |
In those workplaces recognising unions, only one in 50 had a quarter of their workforces earning below £ 3.50 per hour. The incidence of such clusters of low pay is eight times greater in non-union workplaces. |
These results are confirmed by more sophisticated statistical analysis. The mean or average percentage of workers earning below £ 3.50 is 10.5 percent. Other things being equal, union recognition reduces the incidence of low pay by 3.9 percentage points. And where coverage of collective bargaining is above 60 percent, this reduces the incidence of low pay by a further 3.8 percentage points. Thus recognition coupled with extensive coverage reduces the percentage of low-paid workers by nearly 8 percentage points, or some three-quarters of the mean. |
Unions and equality
So far we have shown that unions reduce the spread in the pay distribution. But they also do something more profound: they compress the pay structure between women and men, blacks and whites, and those with health problems and the healthy. |
This impact of unionisation on the pay structure by gender, ethnicity, health and occupation is set out in Table 2 . The results are quite remarkable. Unionisation narrows the wage structure for each group. The fraction of employees who are union members is rather similar for each of the groups, at around a third. Thus the impact of unionisation on the pay structure comes via the premium associated with union membership--i.e., how much better they do in terms of pay because of their union membership. In each case, the lower-paid group--females, nonwhites, those with health problems and manual workers-- receive a higher premium from union membership than do their higher-paid counterpart groups. |
For example, female union members earn 8.7 percent more than a woman with the same characteristics who is not in a union. By contrast, male union members earn no more than their non-union counterparts. Similarly, the payoff to union membership for nonwhite workers is 8.4 percent, more than double the 3.9 percent premium for white workers. Likewise, union membership provides a bigger return to those with health problems (5.3 percent) and manual workers (12.9 percent) than the healthy (3.9 percent) and non-manuals (3.0 percent). |
Consequently, the average wage structure is compressed by the following percentages (in Table 3 ) compared with what it would be if there were no unions. |
If there were no unions, the gender pay gap would be 2.6 percent wider and the race pay gap 1.4 percent bigger. These are very substantial effects. They can be put into perspective by a comparison with the impact of the national minimum wage (NMW). This had a specially favourable effect on female pay--two-thirds of those affected were women--and it narrowed the gender differential by a little under 1 percent. Compare this with unions' impact: unions compress the gender pay differential by 2.6 percent--treble the impact of the NMW. |
It should be remembered in this context that many people think of unions as operating to the advantage of able-bodied, white, male workers. Indeed, Frances O'Grady, head of the TUCs New Unionism project, recently described unions as "male, pale and stale." If this were the case, the wage structures in Table 2 would be widened, not narrowed. |
Trade unions temper the inequality in pay, reduce the incidence of low pay and narrow pay differentials by race and gender as Table 4 indicates. Thriving unions are vital: they wield the sword of justice in the British labour market. |
|
| |
