Do Unions Affect Employer Compliance with the ... - Wiley Online Library

4 downloads 11836 Views 111KB Size Report
Unions, Employer Compliance, and the LawBritish Journal of Industrial Relation. Mark Harcourt is at ... Middlesex University Business School in London. Sondra ...
Blackwell Publishing Ltd.Oxford, UKBJIRBritish Journal of Industrial Relations0007-1080Blackwell Publishing Ltd/London School of Economics 2004September 2004423527541Articles Unions, Employer Compliance, and the LawBritish Journal of Industrial Relation

British Journal of Industrial Relations 42:3 September 2004 0007–1080 pp. 527–541

Do Unions Affect Employer Compliance with the Law? New Zealand Evidence for Age Discrimination Mark Harcourt, Geoffrey Wood and Sondra Harcourt

Abstract Over the last thirty years, collective rights to organize into unions, bargain collectively and strike have been weakened in both New Zealand and the UK. At the same time, individual rights to due process and to protection from discriminatory or unjust management decisions have been strengthened, leading some to conclude that collective and individual rights are unrelated, incompatible or mutually exclusive. On the contrary, we use evidence of employer compliance with anti-age provisions in the New Zealand Human Rights Act to show that the two sets of rights can be highly complementary: the presence of unions strengthens individual protection from discriminatory treatment.

1. Introduction This is a study of the effects of unionization on compliance with labour legislation. Over the last thirty years, the international trend in labour law has been towards a reduction in the scope and depth of the collective rights of unions. In New Zealand, for example, the 1991 Employment Contracts Act removed union rights to access the workplace, to strike for multiemployer agreements and to be the exclusive bargaining agents for entire industries and occupations of workers. The 2000 Employment Relations Act restored some of these rights, but left most of the changes in place. This international trend, most pronounced perhaps in Britain and New Zealand, reflects the reduced political clout of organized labour and governments’ attempts to secure a greater degree of pay and workforce flexibility, often as part of a general quest for ‘international competitiveness’, through the Mark Harcourt is at Waikato University, Hamilton, New Zealand. Geoffrey Wood is at the Middlesex University Business School in London. Sondra Harcourt is an independent researcher in Hamilton, New Zealand. © Blackwell Publishing Ltd/London School of Economics 2004. Published by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.

528

British Journal of Industrial Relations

individualization of industrial relations. The de-collectivization of industrial relations has encouraged many to think that industrial relations have been ‘deregulated’. Yet, as unions’ collective rights have been weakened and reduced, employees’ individual rights have been strengthened and expanded. Smith and Morton (2001) argue that the expansion of individual rights has come at the expense of collective ones. We take the contrary view, that collective and individual rights are, in fact, largely complementary. We hypothesize that unions are active defenders and enforcers of individual rights; the weakening of collective rights effectively undermines mechanisms for enforcing individual rights. Using New Zealand as an example, we test this hypothesis by estimating the relationship between unionization rates and employers’ likelihood of violating individual rights to non-discriminatory treatment by asking legally prohibited questions about a job applicant’s age. We posit a negative relationship.

2. The age discrimination problem Evidence from Australia (Bennington and Wein 2000), Canada (Saunders et al. 1992), New Zealand (Harcourt and Harcourt 2002), and the USA (Jolly and Frierson 1989) suggests that age discrimination is one of the more common forms of discrimination. Moreover, the effects of age discrimination are not limited to a small minority. Anyone can be a potential victim of it, since everyone ages. Any age group could be the target of age discrimination. However, most of the evidence suggests that the problem mainly affects older workers, aged 45–65 (Bendrick et al. 1996, 1999; McGoldrick and Arrowsmith 1993; Oswick and Jones 1991; Taylor and Walker 1994). Employers generally prefer to make older workers redundant (Gardner 1995; Polsky 1999; Taylor and Walker 1994), but prefer to hire younger workers (Bendrick et al. 1996; Finch 1998; Manning 1995; Payne and Payne 1994; Polsky 1999; Westergaard et al. 1989). Many employers feel, despite evidence that suggests otherwise, that older workers have obsolete skills, that they will be harder to train, and that any investment in training will be wasted, with retirement just a few years away (Gibson et al. 1993; Oswick and Jones 1991). In the circumstances, many older, laid off workers drop out of the work force, preferring to call themselves ‘retired’ rather than ‘unemployed’ (Mazerolle and Singh 1999). Yet many, perhaps most, workers would prefer to continue to work (Bazzoli 1985; Daniels and Daniels 1991; Holtmann et al. 1994; Ruhm 1995; Schell et al. 1989). Canadian evidence suggests that involuntary ‘retirement’ of this kind is more than twice as common as genuine, early retirement for the 55–64 year old age group, and more than fifteen times as common in the 45–54 year old age group (Osberg 1993: 507). Obviously, results of this nature are likely to vary from country to country. Nevertheless, they do suggest that older workers have a potentially large unemployment problem, one that is not accurately measured by official, unemployment statistics that © Blackwell Publishing Ltd/London School of Economics 2004.

Unions, Employer Compliance, and the Law

529

report only the number of people actively seeking work (Statistics New Zealand 2001).

3. Individual rights and collective voice Individual rights provide workers with the right individually to complain or grieve about how they have been treated by their employer. In other words, individual rights give workers an ‘individual voice’, both for expressing their discontent with some aspect of managerial behaviour, practice or policy and for seeking some form of redress. Collective rights provide workers with the entitlement to representation by a union in their dealings with management. In other words, collective rights secure workers a ‘collective voice’ in using a union to articulate demands for changes in management behaviour, practice or policy, often through the formal route of collective bargaining. The exercise of ‘individual voice’ has costs and benefits. The costs are many and potentially onerous for the individual acting alone. There are the imputed costs of the time spent preparing, writing, filing and presenting a complaint or grievance. If the complaint or grievance proceeds to a court, arbitration or adjudication hearing and attorney representation is necessary, there may also be heavy legal costs. Finally, and perhaps most importantly, there may be costs associated with employer reprisals, manifested, for instance, in the denial of a promotion or pay increase. The employee could even be sexually or racially harassed or dismissed. In contrast, the benefits of ‘individual voice’ may involve relatively small, subtle improvements in working conditions, which have little or no immediate impact on the worker’s welfare. A slight change in company safety procedures would be one such example. If costs ordinarily exceed benefits, a worker will reasonably decide that filing a grievance or complaint is unlikely to be personally worthwhile. Since other workers will reach this same conclusion most of the time, few, if any, of the employer’s decisions are likely to be openly criticized or protested. The addition of a union’s ‘collective voice’ to the worker’s ‘individual voice’ can improve the cost–benefit trade-offs in initiating a grievance or complaint. By providing research expertise, professional representation and informal advice, a union can greatly reduce many of the worker’s costs. It can also help to protect the worker from employer reprisals; in particular, it can often provide the worker with the protection of total anonymity by assuming direct responsibility for making the complaint or by making the complaint on behalf of a group of workers. Moreover, the involvement of a union increases the perceived benefits of taking action. These benefits result from the ‘public goods’ characteristics of any improvements in working conditions arising from a successful complaint or grievance. ‘Public goods’ are goods or services that you cannot prevent other people from using, and for which one person’s use does not diminish another person’s (Mankiw 1998). Similarly, good (or bad) working conditions that are made available to one worker must also be made available to his or her co-workers, and one co-worker’s enjoyment (or © Blackwell Publishing Ltd/London School of Economics 2004.

530

British Journal of Industrial Relations

suffering) under these working conditions in no way affects another coworker’s. Safety conditions, lighting, heating, the speed of the production line, the firm’s formal grievance procedure, pension plan, and policies on matters such as layoffs, work sharing, cyclical wage adjustment, and promotion obviously affect the entire workforce in the same way that defence, sanitation, and fire protection affect the community at large (Freeman and Medoff 1984: 8–9).

Policies governing discrimination, including discrimination on the basis of age, are no different. A union will consider the positive externalities of these improvements for all workers who are, or may become, union members. Unlike a worker acting alone, a union is likely to conclude that there are significant potential benefits to be had from making a complaint or grievance. As a result, complaints and grievances are much more likely to be made in a union than in a non-union work environment. Furthermore, knowing that this is true, unionized employers are more likely to be proactive than nonunion employers in avoiding potentially unlawful, unfair and/or discriminatory policies that are likely to precipitate costly complaints or grievances. The empirical evidence supports the theory. In general, unionized workers are more likely than non-union workers to exercise their individual rights, and more likely to exercise them in a thorough and comprehensive fashion. For instance, in Weil’s (1991, 1992) studies of health and safety inspection activity in US manufacturing and construction firms, unions substantially increased the probability of inspection, the probability of employee participation in each inspection, the duration of each inspection, the number of violations detected per inspection and the penalty per violation, and these were more pronounced the larger the firm. Furthermore, this more active use of ‘individual voice’ makes unionized employers more likely than their nonunion counterparts to comply with their statutory obligations in protecting workers from health and safety hazards (Dawson et al. 1988; Sandy and Elliott 1996), and in providing them with written contract clauses covering issues such as pay and sick leave (Brown et al. 2000).

4. Methods The data were collected from several sources. Information about age-related questions on job application forms was obtained directly from 174 organizations that had advertised a job in the Wellington Post, Wellington Dominion, or New Zealand Herald newspapers, or on the internet during September, October and November 1999. They were also collected from 55 organizations participating in a large-scale survey of small and medium-sized enterprises in New Zealand. Information about union membership numbers for each employer in the sample was obtained directly from the relevant unions. A list of unions was provided by the Department of Labour’s Registrar of Unions. Membership © Blackwell Publishing Ltd/London School of Economics 2004.

Unions, Employer Compliance, and the Law

531

figures were compared with aggregate employment levels, supplied by New Zealand Who’s Who Aotearoa1 and Kompass: Australia,2 to determine the unionization rate (union density level) of each organization in the sample. Where employment levels were not available from either of these sources, organizations were contacted directly for this information. Data were also collected on select committee submissions concerning the 1991 Human Rights Commission Amendment Bill. This Bill added age to the list of prohibited grounds for unlawful discrimination. The New Zealand Parliamentary Library supplied photocopies of all of the submissions made either in support of, or in opposition to, this proposed Act. Four of these submissions were from unions, all of which were in favour of adding age to the list of prohibited grounds. Information about the presence of human resource departments was obtained from two sources: organizational websites and directly from the organizations themselves. The sample consists of 227 job applications from New Zealand organizations. These organizations collectively employ approximately 200,000 people, in excess of 10 per cent of the New Zealand labour force. They represent a cross-section of New Zealand organizations.3 Descriptive statistics for the dependent and independent variables are outlined in Table 1. The dependent variable, AGE, is a dichotomous variable, and indicates whether or not a job application form contains a question regarding the applicant’s age. AGE assumes the value of 1 if there is at least one agerelated question on a form; it assumes the value of 0 if there are no such questions. Section 23 of the New Zealand Human Rights Act indicates that asking for an applicant’s age shall be taken as evidence of an intention to discriminate on this ground. There are three types of unlawful, age-related question. First, there are questions that simply require the applicant to indicate his or her age. Second, there are questions that ask the applicant to give his or her date of birth. Third, there are questions that ask the applicant to indicate the year he or she left secondary school. Applicants are entitled, under New Zealand case law, to assume that requests for information about the year they left school are requests for information about their age, given that virtually all students

TABLE 1 Descriptive Statistics for the Variables Frequency (% of the sample) AGE UNION DENSITY DENSITY SQUARED UNION SUBMISSION PUBLIC SECTOR HR DEPARTMENT * Mean values. © Blackwell Publishing Ltd/London School of Economics 2004.

114 (50%) 33%* 2.089* 76 (34%) 107 (47%) 154 (68%)

532

British Journal of Industrial Relations

leave school when they are 16, 17 or 18 years old.4 Of the organizations in the sample, 114 (50 per cent) asked at least one of the proscribed questions on their job application form: 84 asked one question, 25 asked two and 5 asked three questions. The independent variables predict whether or not a job application form contains at least one question pertaining to age. The independent variables of primary interest are UNION DENSITY, DENSITY SQUARED and UNION SUBMISSION. UNION DENSITY indicates what percentage of the workforce belongs to a union; this varies from a low of 0 per cent to a high of 100 per cent, with a mean value of 33 per cent. DENSITY SQUARED is the square of UNION DENSITY divided by 1000.5 UNION SUBMISSION indicates whether an employer’s employees are represented by a union that made a submission concerning the 1991 Human Rights Commission Amendment Bill, favouring the introduction of age as one of the prohibited grounds for unlawful discrimination. This variable assumes a value of 1 for employers whose employees are represented by one or more of these unions and 0 otherwise. There were four unions that made such a submission; 76 (34 per cent) of the employers in the sample have employees represented by one of these four unions.6 UNION DENSITY is used as a proxy measure of collective voice. There is some empirical evidence to support the notion that collective voice depends at least in part on union density. For instance, Reshef and Lam (1999) found that unions were more likely to try to influence management-initiated, quality improvement programmes as union density rose. Collective voice is expected to increase with union density for several reasons. First, a union clearly cannot influence organizational policies, unless it represents at least some of the organization’s workers. Second, a union’s grievance- and complaintrelated activity should vary directly with the number of workers it represents in the workplace. Third, under the Employment Relations Act a collective agreement, and any provisions it may have that relate to discrimination issues, covers only union members. Fourth, the union’s negotiating power to influence human resource policies increases with its capacity to strike, which, in turn, clearly depends upon the proportion of workers who are union members and thus are legally permitted to strike. DENSITY SQUARED is added to capture any potentially nonlinear effects of union density. For instance, it is possible that increases in density have a larger impact on employer compliance at lower levels of unionization. UNION DENSITY and DENSITY SQUARED measure the potential for collective voice. However, this potential may remain unrealized if a union is unconcerned by age discrimination or makes other issues its main priorities. One indication of the seriousness with which a union and its leadership regard age discrimination is their willingness to speak out in favour of banning it. In 1991 four New Zealand unions spoke out against age discrimination by writing submissions supporting the Human Rights Commission Amendment Bill. In contrast, the great majority of New Zealand unions, whether large or small, public or private-sector, did not write a submission. © Blackwell Publishing Ltd/London School of Economics 2004.

Unions, Employer Compliance, and the Law

533

The other independent variables are PUBLIC SECTOR and HR DEPARTMENT. PUBLIC SECTOR is a dichotomous variable that indicates whether an organization is part of the broader New Zealand public sector. This includes public schools, public hospitals, universities, polytechnics, state agencies, state-owned enterprises and civil service departments and ministries. It assumes the value of 1 for organizations of this type and 0 otherwise. Of the organizations in the sample, 107 (47 per cent) belong to the public sector. HR DEPARTMENT is a dichotomous variable which indicates whether an organization has a human resource department. It assumes the value of 1 for organizations with a human resource department and 0 otherwise. Of the organizations in the sample, 154 (68 per cent) have such a department. Public-sector organizations are expected to comply more frequently than private-sector organizations with the anti-discrimination provisions in the Human Rights Act and the Employment Relations Act. Unlike their privatesector counterparts, public-sector organizations are required by legislation to be ‘good employers’, with equal employment opportunity (EEO) programmes to promote the equitable treatment of customarily disadvantaged groups and to monitor their progress through the organization.7 Although EEO obligations do not specifically encompass discriminatory questions on job application forms, they do encourage organizations to vet their human resource practices, including those related to hiring, for potentially discriminatory effects. Similarly, a higher level of compliance is expected from organizations with a human resources department than from organizations without such a department. Higher levels of compliance might be explained by several factors. Most importantly, human resource departments are typically staffed by trained professionals who know the law and how to comply with it. Even if untrained in discrimination law, human resource specialists may have had direct experience in dealing with personal grievances or complaints involving discrimination. Furthermore, as part of their professional ethics, human resource specialists may place a high value on equal employment opportunities in their workplace. In contrast, general managers, who typically handle human resource matters when there is no human resource department, are unlikely to be as trained, experienced or committed to equal employment opportunities, especially given the many other work roles they may perform. Collinearity diagnostics were run on all five of the independent variables. Tolerance statistics were relatively high for UNION SUBMISSION (0.68), PUBLIC SECTOR (0.47) and HR DEPARTMENT (0.73), indicating that multicollinearity was not a problem for these variables. In contrast, the tolerance statistics for UNION DENSITY (0.070) and DENSITY SQUARED (0.079) were both very low, obviously reflecting the high degree of collinearity between these two similar variables. However, with the removal of DENSITY SQUARED, the tolerance statistic for UNION DENSITY rose to a high level (0.63), again suggesting no general problem with multicollinearity. The data are analysed using binomial logistic regression. This technique applies maximum likelihood to estimate coefficients for the independent © Blackwell Publishing Ltd/London School of Economics 2004.

534

British Journal of Industrial Relations

variables that best predict the odds of an organization asking the applicant’s age on a job application form. The logistic model estimated has the following form: log[ probability (questions) probability (no question)] = b 0 + b1X 1 + b 2 X 2 + ... + b nX n The ‘question’ referred to in the model is an unlawful question on the job application form about the applicant’s age. ‘No question’ in this case represents the absence of such a question. The probability of ‘question’ divided by the probability of ‘no question’ represents an odds ratio, in this case the odds of being asked a question about one’s age rather than not. The log of this odds ratio is predicted by the independent variables, indicated by Xs in the model statement. The betas ( b ) in the model statement are the coefficients and indicate the effect of a one-unit change in the relevant independent variable on the log odds ratio.

5. Results Results for the logistic regression model are provided in Table 2. The estimated coefficients for the model and their associated standard errors are outlined in the first column of this table. Changes in the odds of an application form asking a question about the applicant’s age rather than not asking such a question are calculated by taking the inverse log of each coefficient. These results are outlined in the second column of the table. Negative (positive) and statistically significant coefficients indicate that, as the value of the independent variables rise, the odds ratio falls (rises as well). The bottom of Table 2 outlines goodness-of-fit statistics for the model. The chi-square statistic (37.142) shows that at least one of the coefficients in the model, other than the intercept, is not zero (probability value = 0.00). The Hosmer–Lemeshow test statistic (4.653) indicates that the model cannot be TABLE 2 Binary Logistic Regression Results for Job Applications Requesting Information about Age

INTERCEPT UNION DENSITY DENSITY SQUARED UNION SUBMISSION PUBLIC SECTOR HR DEPARTMENT

AGE (Std errors)

Odd ratio

0.991*** (0.289) 0.032* (0.018) -0.385** (0.193) -0.749** (0.363) -1.089*** (0.410) -0.718** (0.361)

2.694 1.033 0.680 0.473 0.337 0.488

* Statistically significant at the 0.10 level; ** Statistically significant at the 0.05 level; *** Statistically significant at the 0.01 level (2-tailed tests). Chi-square statistic = 37.142 with 6 degrees of freedom (probability value = 0.00). Hosmer–Lemeshow test = 4.653 (probability value = 0.79). Nagelkerke R2 = 0.0.201. Cox–Snell R2 = 0.151. © Blackwell Publishing Ltd/London School of Economics 2004.

Unions, Employer Compliance, and the Law

535

rejected at the 10 per cent level (probability = 0.79) and thus fits the data at least adequately. The Cox–Snell R2 (0.151) and the Nagelkerke R2 (0.201) both indicate that at least some of the uncertainty in the data can be explained by the model. The classification table shows that the model correctly predicts 79 (69 per cent) of the 114 job application forms that contained at least one question about the applicant’s age.8 It also correctly predicts 74 (66 per cent) of the 113 forms that did not contain any questions about the applicant’s age. In total, the model accurately classifies 153 (67 per cent) of the 227 observations. The coefficient for the intercept is positive and statistically significant, and indicates that the odds of an application form having, rather than not having, a question about the applicant’s age is close to 3:1 when all of the independent variables equal zero. Contrary to expectation, the coefficient for UNION DENSITY is statistically significant and positive, and indicates that the odds of asking an agerelated question rise by 3 per cent for every 1 per cent rise in union density. However, the negative and statistically significant coefficient for DENSITY SQUARED runs in the opposite direction, indicating that this effect is more than offset at higher union densities. The odds of an employer asking an agerelated question initially rise, reaching a maximum increase of nearly 100 per cent at a union density level of 42 per cent. At density levels higher than 42 per cent the size of this increase diminishes. At density levels higher than 84 per cent, the odds of an employer asking an age-related question actually decline. Once an employer is 100 per cent unionized, its odds of asking an age-related question are only half as high as an employer that is 1 per cent unionized. The coefficient for UNION SUBMISSION is also statistically significant and negative. The odds of an organization asking an age-related question on its job application form are nearly 55 per cent lower at each union density level, if it negotiates with a union that made a submission concerning the 1991 Human Rights Commission Bill. Compared with a non-union employer, the odds of such an employer asking an age-related question are 51 per cent lower for an employer with a 1 per cent (and 84 per cent) density level, 7 per cent lower for an employer with a 42 per cent density level, and 74 per cent lower for an employer with a 100 per cent density level.9 Overall, UNION DENSITY, DENSITY SQUARED and UNION SUBMISSION taken together imply the following. First, non-union employers are less likely to have age-related questions on their job application forms than employers that are less than 84 per cent unionized and deal with a union that did not make a submission concerning the Human Rights Commission Bill. Second, non-union employers are more likely to have age-related questions on their job application than all employers, no matter how unionized, dealing with a union that did make a submission concerning the Human Rights Commission Bill. Public-sector organizations are less likely than private-sector organizations to ask discriminating questions about age. The coefficient for PUBLIC © Blackwell Publishing Ltd/London School of Economics 2004.

536

British Journal of Industrial Relations

SECTOR is statistically significant, and indicates that being a public-sector organization reduces the odds of asking a discriminatory question about age by approximately 65 per cent. Organizations that have a human resource department are also less likely than other organizations to contravene the Human Rights Act by asking questions about age on their job application forms. The coefficient for HR DEPARTMENT is statistically significant, and indicates that having a human resource department decreases the odds of an age-related question on a job application form by roughly 50 per cent.

6. Discussion Like Weil (1991, 1992), we found considerable evidence for the positive influence of unions on employer compliance with the Human Rights Act. More specifically, increasing rates of unionization eventually reduce the odds of employers asking unlawful questions about age on their job application forms. For instance, an increase in union density from 1 per cent to 100 per cent reduces the odds of such a question by approximately 45 per cent. This result is consistent with a ‘collective voice’ perspective of trade unions and their influence on employer legislative compliance. In particular, the ‘public goods’ nature of employer compliance with anti-discrimination legislation makes it much more likely that a union acting for a group of affected employees would take enforcement action than an individual acting alone. Faced with greater enforcement activity, we would expect the more heavily unionized employers to respond more rapidly and readily to employee criticism of their discriminatory policies, and to make more of an effort to avoid such criticism by not discriminating in the first place. Contrary to expectation, the UNION DENSITY and DENSITY SQUARED findings, taken together, suggest a curvilinear relationship between unionization and employer compliance. Unionization appears initially to increase the odds of an employer asking a discriminatory question, although this effect is reversed at higher levels of union density. What could account for such an effect? At low union density levels, the relationship between employer compliance and unionization may reflect reverse causation, with the former driving and determining the latter. Bad employer behaviour could lead to unionization, rather than unionization encouraging good employer behaviour. For instance, a few brave employees who are being mistreated by a discriminating boss might well opt to join a union for protection and moral support; however, at a low density level, that union is unlikely to have the power to demand major changes in employer policies — in New Zealand it might not even be able to negotiate a collective agreement with the employer. Furthermore, at low density levels a poorly established union is likely to be preoccupied with organizing workers and institutionalizing collective bargaining, rather than protecting the rights of job applicants who are not even union members. © Blackwell Publishing Ltd/London School of Economics 2004.

Unions, Employer Compliance, and the Law

537

In a departure from previous studies, we also found that the collective voice effect was not even across all unions. Employer compliance was clearly greater where a union had articulated a strong statement against age discrimination by writing a submission in support of the 1991 Human Rights Commission Bill, banning age discrimination. Where the employer negotiated with such a union, the odds of an employer asking a discriminatory question about age were more than 50 per cent lower. In many respects, this finding should not be surprising. Union members often have discriminatory attitudes. Given that union leaders are elected by the membership and are often chosen from the membership, we would expect union policies and practices to reflect some grass-roots prejudice. At the very least, member apathy towards, or ignorance of, discrimination may encourage union leaders to be equally complacent about the issue. Just as easily, the reverse could apply. Empirical evidence confirms that unions vary considerably in their stance on discrimination, including their stance on age discrimination (Duncan et al. 2000). Some studies have even documented union support for profoundly discriminatory practices and policies, especially prior to the 1970s (Ashenfelter 1972; Gould 1977; Sass and Troyer 1999). Nowhere was this perhaps more apparent than in South Africa, where white-dominated unions were strong apartheid supporters until the 1970s (Wood and Harcourt 1998). All this suggests that incorporating measures of union policy is essential to any future analysis of the effects of collective voice. Overall, our findings underscore the importance of inter-firm institutions, including unions and social conventions, in reinforcing the employment transaction (Marsden 1999: 72). As Marsden notes, any rules, whether in the form of laws or informal conventions governing the employment relationship, are likely to be more robust when widely invoked (p. 72); this promotes uniformity, and increases the number of people exposed to them. Moreover, the enforcement of rules is easier, and opportunism less likely, when parties to a dispute can call on outside support to back them up; this makes for both greater stability and sustainable flexibility, given the existence of a nonnegotiable minimum position to fall back on (Marsden 1999: 72).

7. Implications We found some support for the idea that high levels of unionization encourage employer compliance with legislation. This finding has several important implications. First, it indicates that unions should promote their enforcement role as a major advantage to potential members. The protection of individual rights should not be seen as a second-best substitute for collectivism. Rather, it represents one of several mechanisms through which unions can advance the interests of their entire membership. This underscores the need for unions to re-evaluate the desirable balance between collective and individual interventions, without regressing into the kind of service unionism that narrowly focuses on enhancing the benefits accruing to individual members, and/or © Blackwell Publishing Ltd/London School of Economics 2004.

538

British Journal of Industrial Relations

shackles the union into vague partnership deals with management. Whatever the external pressures a union movement faces, the strategic choices it makes remain critical (Wood and Psoulis 2001). Few unions today are in a position to wage the protracted trench warfare of the 1960s and 1970s; the most feasible strategy is one of flexible response. This can encompass such diverse interventions as rallying support around community campaigns and policing the acts of employers, enforced through carefully chosen legal interventions in areas of collective concern. In an age of insecure employment contracts and stagnant wage levels, any victories, no matter how gained, may be a source of inspiration to existing and prospective union members. In short, rather than the Luxembourgian assumption, still tacitly held by many contemporary writers, that real advances can be made only through mobilization and action on a large scale, we would suggest that significant gains can be secured through strategies focused on individual cases, carefully targeted interventions designed to secure maximum benefits for members, and publicity, even in the face of adverse circumstances and temporary setbacks. By resorting to sustained legal action in a carefully chosen number of test cases, unions can enhance the rights of all workers. Hence, unlike Smith and Morton (2001), we believe that unions should seize opportunities as and when they arise; clear victories — and unambiguous support for members — may both energize the labour movement and improve the general lot of employees by reinforcing desirable transaction rules underpinning the employment contract (c.f. Marsden 1999). Again, the exposure of ham-handed actions by employers heightens awareness of the relevance of unions in a fractured and individualized age. Second, we believe that legislators should incorporate unions more directly and formally into their employment law enforcement mechanisms (c.f. Ewing and Hendy 2002). As a first step in this direction, the New Zealand government is planning to provide for greater union involvement in health and safety regulation with its 2002 Health and Safety in Employment Amendment Act. This approach could be carried further into other aspects of employment law. For instance, unions could be permitted to take personal grievances and complaints of discrimination against employers, without having to have them initiated by courageous individual employees. Again, this represents a mechanism for bringing unions back into the employment relationship, within the very different political and ideological context of the early twenty-first century. Final version accepted 15 March 2004.

Acknowledgments We would like to thank Professors Yonatan Reshef, Helen Lam and Dave Jobson for their helpful comments on earlier drafts of the paper; we would also like to thank Kate Young and Cathy Horrocks for their help with research assistance. Finally, we wish to thank all the other people — govern© Blackwell Publishing Ltd/London School of Economics 2004.

Unions, Employer Compliance, and the Law

539

ment officials, union presidents, human resource managers and parliamentary librarians — who have been involved in one way or another with supplying data. This article would not have been possible without their support.

Notes 1. New Zealand Who’s Who Aotearoa lists New Zealand companies, provides their contact details and provides information about some of their basic characteristics, including staff numbers. 2. Kompass: Australia is a searchable database from Kompass International, which provides information about companies, products, brand names and executives’ names for companies in more than sixty nations. 3. Specifically, 59 public-sector ministries, departments or agencies; 17 health care organizations; 35 primary, secondary and tertiary educational institutions; 15 financial services firms; 6 transportation and 3 communication firms; 16 business or professional service firms, 21 retailers and/or wholesalers; 9 hotels, motels or restaurants; 10 manufacturers, 3 firms in forestry, agriculture or fishing, and 4 volunteer organizations. 4. McCready v. Hawkins, and ors, unreported, Complaints Review Tribunal, CRT Decision 10/97. 5. Dividing the square of union density does not change the overall results. However, it does increase what would otherwise be a tiny coefficient by 1000 times, making it easier to interpret and report in Table 2. 6. The four unions are: the Public Service Association (PSA), the New Zealand Educational Institute (primary school teachers’ union), the Financial Sector Union (FinSec), and the Police Association. 7. For example, the civil service has EEO obligations under the ‘good employer’ provisions of the 1988 State Sector Act. The State Services Commission consistently and thoroughly monitors and enforces civil service EEO compliance. In particular, the Commission annually appraises the EEO performance of each civil service chief executive. An executive who consistently under-performs in EEOrelated matters risks not being awarded an annual bonus and even not being reappointed at the end of his or her three- or five-year contract. 8. The model ‘predicts’ the asking of an age-related question, whenever the estimated probability of a question being asked is greater than 50 per cent. 9. Seventy three (32 per cent) of the employers have union density levels of 0 per cent; 66 (29 per cent) have union density levels of 1–42 per cent; 68 (30 per cent) have union density levels of 43–84 per cent; 20 (9 per cent) have union density levels in excess of 84 per cent.

References Ashenfelter, O. (1972). ‘Racial discrimination and trade unionism’. Journal of Political Economy, 80: 435–64. Bazzoli, G. J. (1985). ‘The early retirement decision: new empirical evidence on the influence of health’. Journal of Human Resources, 20: 214–34. © Blackwell Publishing Ltd/London School of Economics 2004.

540

British Journal of Industrial Relations

Bendrick, M., Jackson, C. and Romero, J. H. (1996). ‘Employment discrimination against older workers: an experimental study of hiring practices’. Journal of Aging and Social Policy, 8(4): 25–46. ——, M., Brown, L. and Wall, K. (1999). ‘No foot in the door: an experimental study of employment discrimination against older workers’. Journal of Aging and Social Policy, 10(4): 5–23. Bennington, L. and Wein, R. (2000). ‘Anti-discrimination legislation in Australia: fair, effective, efficient, or irrelevant?’ International Journal of Manpower, 21(1): 21–33. Brown, W., Deakin, S., Nash, D. and Oxenbridge, S. (2000). ‘The employment contract: from collective procedures to individual rights’. British Journal of Industrial Relations, 38: 611–29. Daniels, C. E. and Daniels, J. D. (1991). ‘Factors affecting the decision to accept or reject a golden handshake’. Benefits Quarterly, 7(2): 33–46. Dawson, S., Willman, P., Bamford, M. and Clinton, A. (1988). Safety at Work: The Limits of Self-Regulation. Cambridge: Cambridge University Press. Duncan, C., Loretto, W. and White, P. (2000). ‘Ageism, early exit, and British trade unions’. Industrial Relations Journal, 31: 220–34. Ewing, K. and Hendy, J., (2002). A Charter of Workers’ Rights. London: Institute of Employment Rights. Finch, J. (1998). ‘The re-employment experiences of former aerospace employees within a local economy’. Regional Studies, 32: 421–33. Freeman, R. and Medoff, J., (1984). What Do Unions Do? New York: Basic Books. Gardner, J. (1995). ‘Worker displacement: a decade of change’. Monthly Labor Review, 118(4): 45–57. Gibson, K., Zerbe, W. and Franken, R. (1993). ‘Employers’ perceptions of the reemployment barriers faced by older job hunters’. Relations Industrielles, 48: 321–34. Gould, W. (1977). Black Workers in White Unions: Job Discrimination in the United States. Ithaca, NY: Cornell University Press. Harcourt, S. and Harcourt, M. (2002). ‘Do employers comply with civil/human rights legislation? New evidence from New Zealand Job Application Forms’. Journal of Business Ethics, 35: 207–21. Hirsch, B., Macpherson, D. and Hardy, M. (2000). ‘Occupational age structure and access for older workers’. Industrial and Labor Relations Review, 53: 401–18. Hirschman, A. O. (1970). Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, Mass.: Harvard University Press. Holtmann, A. G., Ullmann, S. G., Fronstin, P. and Longino, C. F. Jr (1994). ‘The early retirement plans of women and men: an empirical application’. Applied Economics, 26: 591–601. Jolly, J. and Frierson, J. (1989). ‘Playing it safe’. Personnel Administrator, 34(6): 44–50. Mankiw, N. G. (1998). Principles of Economics. New York: Dryden Press. Manning, N. (1995). ‘Do older workers drop out of the labour market? Some recent British evidence’. Bulletin of Economic Research, 48: 167–71. Marsden, D. (1999). A Theory of Employment Systems. Oxford: Oxford University Press. Mazerolle, M. and Singh, G. (1999). ‘Older workers adjustment to plant closures’. Relations Industrielles, 54: 313–33. McGoldrick, A. and Arrowsmith, J. (1993). ‘Recruitment advertising: discrimination on the basis of age’. Employee Relations, 15(5): 4–65. Osberg, L. (1993). ‘Is it retirement or unemployment? Induced “retirement” and constrained labour supply among older workers’. Applied Economics, 25: 505–19. © Blackwell Publishing Ltd/London School of Economics 2004.

Unions, Employer Compliance, and the Law

541

Oswick, C. and Jones, P. (1991). ‘The age factor in work performance: age-ism or realism?’ Management Services, 35(12): 12–15. Payne, J. and Payne, C. (1994). ‘Recession, restructuring and the fate of the unemployed: evidence in the underclass debate’. Sociology, 28(1): 1–19. Polsky, D. (1999). ‘Changing consequences of job separation in the United States’. Industrial and Labor Relations Review, 52: 565–80. Reshef, Y. and Lam, H. (1999). ‘Union response to quality improvement initiatives: factors shaping support and resistance’. Journal of Labor Research, 20(4): 111–31. Ruhm, C. (1995). ‘Secular changes in the work and retirement patterns of older men’. Journal of Human Resources, 30: 362–85. —— (1996). ‘Do pensions increase the labor supply of older men?’ Journal of Public Economics, 59: 157–75. Sandy, R. and Elliott, R. (1996). ‘Unions and risk’. Economica, 63: 291–309. Sass, T. and Troyer, J. (1999). ‘Affirmative action, political representation, unions, and female police employment’. Journal of Labor Research, 20: 571–87. Saunders, D., Leck, J. and Marcil, L. (1992). ‘What predicts employer propensity to gather protected group information from job applicants?’ New Approaches to Employee Management: Fairness in Employee Selection, 1: 105–30. Schell, B., Lebrasseur, R. and Renaud, R. (1989). ‘Predictors of acceptance of early retirement offers for workers’. Relations Industrielles, 44: 376–92. Smith, P. and Morton, G. (2001). ‘New Labour’s reform of Britain’s Employment Law: the devil is not only in the detail but in the values and policy too’. British Journal of Industrial Relations, 39: 119–38. Statistics New Zealand (2001). Labour Market Statistics. Wellington, NZ: Statistics New Zealand. Taylor, P. E. and Walker, A. (1994). ‘The ageing work-force: employers’ attitudes towards older people’. Work, Employment and Society, 8: 569–91. Walker, A. (1985). ‘Early retirement: release or refuse from the labour market?’ Quarterly Journal of Social Affairs, 1: 211–29. Weil, D. (1991). ‘Enforcing OSHA: the role of labor unions’. Industrial Relations, 30(1): 20–36. —— (1992). ‘Building safety: the role of construction unions in the enforcement of OSHA’. Journal of Labor Research, 13(1): 121–32. Westergaard, J., Noble, I. and Walker, A., (1989). After Redundancy. Oxford: Polity Press. Wood, G. and Harcourt, M. (1998). ‘The rise of South African trade unions’. Labor Studies Journal, 23(1): 74–92. —— and Psoulis, C. (2001). ‘Globalization, democratization, and organized labor in transitional economies’. Work and Occupations, 28: 293–314.

© Blackwell Publishing Ltd/London School of Economics 2004.