Conference – 1998 Industrial Relations Reform and Labour Market Outcomes: A Comparison of Australia, New Zealand and the United Kingdom Mark Wooden and Judith Sloan

1. Introduction

The 1980s and 1990s have seen important changes in many countries' labour market institutions, particularly those that impinge on the determination of wages and working conditions. Nowhere has such change been more dramatic than in the three countries at the centre of discussion in this paper – Australia, New Zealand and the United Kingdom. Decentralisation of collective bargaining structures in the UK, for example, accelerated in the 1980s under the Thatcher administration. Moreover, the shift away from industry-level bargaining was accompanied by legislation designed to significantly weaken the power of trade unions. In New Zealand the decentralisation process has been even more radical. While the reform process can be traced back to the 1970s, the most significant changes occurred in one hit following the passing of the Employment Contracts Act in 1991. Under this legislation, the system of industry and occupation awards was dismantled, compulsory union membership and union preference outlawed, and scope for non-collective bargaining introduced. The process of industrial relations reform in Australia, on the other hand, has been very different. Indeed, the reform process began not with decentralisation of industrial relations structures, but with a return to a highly centralised form of wage determination, with variations in wages and conditions dependent on national wage case decisions, which in turn were directly linked to prices. Despite this, from the late 1980s on, most Australian governments pursued a reform agenda which was similar in thrust to that being pursued in both New Zealand and the UK. Unlike New Zealand and the UK, however, the reform process, at least within the federal arena, was undertaken within the ambit of the Accord process, and hence had the seal of approval of the trade union movement.

The objectives of this paper are twofold. First, to describe in more detail the process of industrial relations change that has taken place in each of these three countries during the past two decades. Second, and perhaps of greater importance, to examine whether or not these changes have had any significant impact on labour market outcomes.

2. The New Zealand Experience

The history of regulation of industrial relations in New Zealand has many similarities to the regulation of industrial relations in Australia. Both countries operated systems of compulsory arbitration for many years, with the principal statute in New Zealand being the Industrial Conciliation and Arbitration Act 1894. The core features of these systems have been described by Brooks (1995, p. 32) in the following way:

[They are] based on a philosophy of massive state intervention in and regulation of employer-employee relations in the public interest; while there is no compulsion to participate in the system but once participating the parties are subject to a range of coercive measures and sanctions; trade unions are actively encouraged to participate in the system, they are protected within the system and their legal status is equated to that of a trading corporation; the system attempts to equalise bargaining power between the parties; there is direct and, often, immediate access to permanent industrial tribunals; the predominant activity of the permanent industrial tribunals is not the prevention of strikes but the making of an award in settlement of an industrial dispute; the concept of an ‘industrial dispute’ has nothing to do with strikes or stoppages but is a highly legalistic concept…

The Industrial Conciliation and Arbitration Act 1894 lasted nearly one hundred years. Over time, there were some modifications, particularly in relation to the right to strike and whether unionism and arbitration should be compulsory or voluntary. Additionally, a raft of supporting legislation emerged in respect of holidays, minimum wages and leave entitlements (Rasmussen et al. 1995).

The effects of the New Zealand system of compulsory arbitration have been outlined by Rasmussen et al. (1995, pp. 1–2) and include the following: legalism; adversarial relations between parties; many weak unions flowing from the union registration provisions; blanket coverage of wages and conditions across industries and occupations; restriction of bargaining to narrowly defined ‘industrial issues’; specialist tribunals; and issues related to the individual employment contract being dealt with under common law. These effects have many parallels with the Australian experience of compulsory arbitration.

While the Employment Contracts Act 1991 (ECA) is generally seen as the key turning point in respect of the regulation of industrial relations in New Zealand, there had been a number of important developments in the two decades or so prior to the enactment of the ECA (Blandy and Baker 1987; Rasmussen et al. 1995). As Boxall (1990) has noted, the period from the mid 1960s to the early 1990s was characterised by a shift towards decentralisation and direct bargaining, although there were instances of attempted centralisation within the period. Indeed, the Industrial Relations Act 1973 was an attempt to control ‘second-tier agreements’ (over-award bargaining in Australia), an attempt which ultimately failed. Boston (1984) argued that direct state intervention in particular disputes, coupled with wage and price controls, ultimately undermined public support for the system of compulsory arbitration. By the same token, the Industrial Relations Act 1973 introduced the distinction between interest and rights disputes, a distinction common in jurisdictions in which direct bargaining is predominant (Boxall 1990).

New Zealand's Labour Government of 1984 to 1990 introduced a number of pieces of legislation affecting industrial relations. The Industrial Relations Amendment Act 1984 abolished compulsory arbitration, the Industrial Relations Amendment Act 1985 re-instituted compulsory unionism, subject to regular ballots of workers. Of more significance was the Labour Relations Act 1987 which incorporated a number of important reforms to the arbitration system, while strengthening some of its controls. In particular, direct bargaining was encouraged on the initiative of unions, with responsibility for the enforcement of agreements devolved to the parties themselves. Unions could opt for coverage of workers by either an award or registered agreement, with return to the parent award being subject to the consent of the employer. The one exception to this rule was the case of composite bargaining where a reduction in the number of bargaining units could be demonstrated. In effect, the statute sought to encourage the rationalisation of bargaining structures, which had been based on the highly fragmented, craft configuration of New Zealand trade unions. Registered unions were also required to have a minimum membership of 1,000, which led to a number of amalgamations between unions (Boxall 1990).

What was the effect of the Labour Relations Act 1987 on the incidence and pattern of bargaining? Boxall (1990, p. 533) stated that ‘by the end of 1989, it was becoming apparent that the evidence on the extent of bargaining reform … had to be seen as equivocal. There was evidence of conservatism … but there were also signs of valuable change in the direction of composite bargaining’.

While the ECA is generally seen as a ‘big bang’ reform, arguably the statute represented the culmination of a process of change, characterised by a shift towards bargaining and a diminution in the role of third parties. This said, the ECA ushered in a number of very significant changes, including the complete abolition of the award system and the removal of registration and regulation of trade unions. Indeed, as Kasper (1996) has noted, the word ‘union’ does not appear in the ECA. The ECA is divided into five parts. Part I deals with freedom of association; Part II deals with freedom to bargain; Part III deals with the treatment of ‘personal grievances’ (dismissal, discrimination, harassment and duress); Part IV deals with rights and obligations under employment contracts; and Part V provides for the right to strike and lock-out during the negotiation of (enterprise) contracts. The Act also provides for the Employment Court and Employment Tribunal. In terms of its immediate impact, the three key changes introduced by the ECA were: the abolition of the award system; the absence of legal rights for trade unions; and the bias against multi-employer negotiations arising from the illegality of industrial action undertaken in pursuit of multi-employer agreements (Kasper 1996).

Rasmussen et al. (1995, p. 4) have described the ECA in the following way:

[It] introduced substantial deregulation of the New Zealand bargaining structure. It rejected many of the principles associated with the arbitration model. It abolished the award system, union rights, and promoted an enterprise-bargaining model in both private and public sectors.

By the same token, the ECA fell short of a full deregulation model based solely on common law, as advocated by such authors as Brook (1990). In particular, the ECA provided for a strong role for legal precedence; personal grievance procedures for all employees, including in relation to unfair dismissal; and the preclusion of the ability to opt out of statutory minima (Bray and Neilson 1996; Rasmussen et al. 1995).

What have been the effects of the ECA? The most common forms of employment contract to emerge after the ECA were single-employer collective agreements and individual agreements (Harbridge and Moulder 1992; Rasmussen et al. 1995; Kasper 1996). On the most recent evidence (Hector and Hobby 1998), this remains the case, with 34 per cent of employees covered by single-employer contracts, 49 per cent by individual employment contracts and 11 per cent by multi-employer awards (see also Department of Labour 1998). Of the collective agreements negotiated, unions have acted as bargaining agents of workers in the vast majority of cases. Rasmussen et al. (1995, p. 11) conclude that ‘the ECA has accelerated and facilitated the change process by making bargaining structures, processes and outcomes more flexible’, but note that some of the major examples of workplace reform pre-date the ECA.

In Table 1, information on the coverage of individual and collective employment contracts for four separate periods after the enactment of the ECA is presented. Immediately after the ECA's enactment, a relatively high proportion of employees remained covered by multi-employer contracts. This proportion dwindled very rapidly, however, with small proportions of employees covered by multi-employer contracts. By the same token, collective contracts continued to cover a fair proportion of workers, with over one-third covered by single-employer collective contracts on the latest figures. Between May 1991 and August 1992, there was a very significant increase in the coverage of employees by individual contracts, with nearly half of employees on individual contracts on the latest figures. Overall, the figures in Table 1 point to a marked decentralisation in agreement-making in New Zealand following the ECA becoming law and the near-extinction of multi-employer contracts.

Table 1: Coverage of Individual and Collective Employment Contracts (IEC and CEC)
Per cent
IEC Multi-employer
(and awards)
Single employer Combined
IEC/CEC
Total CEC
May 1991 28 59 13 72
August 1992 52 8 35 5 48
August 1993 40 9 37 8 54
August 1996 49 11 34 4 49

Notes: These figures are taken from the three surveys of Labour Market Adjustment commissioned by the Department of Labour. It should be noted that the use of different bases, and the low response rate of the 1997 survey, may mean that the figures are not exactly comparable.

Source: Hector and Hobby (1998, p. 314).

In Table 2, data are presented on employee representation by sector and contract type in 1996. Not surprisingly, union representation is much more common for collective contracts than for individual contracts. In 42 per cent of private collective contracts and 61 per cent of government collective contracts, employees were represented by unions. Hector and Hobby (1998, p. 317) note some interesting changes over time, however. The percentage of employees represented by unions fell from 90 per cent in the public sector in 1993 to 61 per cent in 1996, and from 67 per cent in the private sector in 1993 to 42 per cent in 1996. While there has been a small move to employees representing themselves, the larger move has been towards representation by other agents such as groups of employees, lawyers, industrial relations consultants or combined representation.

Table 2: Employee Representation by Sector and Contract Type
Per cent; 1996
Negotiation outcome
Private IEC Private CEC Govt IEC Govt CEC
Employees represented self 71 18 58 3
Union representation 1 42 3 61
Other/combined employee representation 11 24 26 34
Don't know/no representative/no negotiation 10 5 6 1
Missing information 6 11 7 2

Source: Hector and Hobby (1998, p. 316).

Apart from the decentralisation of bargaining associated with the ECA, the number of unions has fallen as some unions went out of existence while others merged. The extent of union coverage has dropped noticeably – from an estimated 55 per cent in 1986, to 37 per cent in 1992 and to 25 per cent in 1996 (Table 5). The number of working days lost per worker in industrial disputes rose slightly between 1991 and 1992 but has fallen since (Figure 2). Gardner (1995, p. 55) has described the reaction of the New Zealand trade union movement to its changed circumstances in the following way: ‘Occupied with establishing the ground rules for union movement at the enterprise level and unlikely to face the labour market circumstances that will allow much change to their current position’.

While the impact of the ECA on bargaining was clearly deregulatory with no clear presumption in favour of collectivism, the effects of other parts of the ECA have been less clear-cut. Indeed, the opening up of personal grievance mechanisms to all employees, plus the establishment of the Employment Court and the Employment Tribunal, have led to a number of trenchant criticisms of these parts of the Act (Baird 1996; Robertson 1996; Epstein 1997; Kerr 1997). Indeed, it has been argued that the non-prescriptive or ‘enabling’ nature of much of the ECA has been responsible for Court-determined rules based on legal precedents (Anderson 1996; Rasmussen and Deeks 1997). This has led, in turn, to some winding back of aspects of the statute – in relation to bargaining rights of trade unions, for instance – and to a surge in applications for remedies against unfair dismissal.

Rasmussen and Deeks (1997, p. 294) have summarised the outcomes of the ECA in the following way:

These include the rapid acceleration of the move to individual employment contract, marked by the growth of individual employment contracts (IECs) and the reduction in collective employment contracts (CECs); the reduced role for unions in bargaining, particularly in the private sector; the decline in union membership and in union density; and the greater use of legal remedies in employment disputes and grievance resolution.

3. The United Kingdom Experience

While the regulation of industrial relations in the United Kingdom was never based on a model of compulsory arbitration, the degree of specific government involvement in this area has been substantial over the years. The intervention took a number of forms including: providing trade unions with immunity under tort for industrial action; requiring government contractors to comply with working conditions as determined by collective bargaining (Fair Wage Resolution Act 1949); obliging firms to pay wages not less favourable than those fixed by collective bargaining with industries (Employment Protection Act 1975); and establishing minimum wages through Wages Councils.

A marked change of direction occurred in 1980 with the election of the Conservative Thatcher Government. A series of reforms occurred (in 1980, 1982, 1984, 1986, 1988, 1989, 1993 and 1995) which Gregory (1997) has grouped under three headings: provisions affecting industrial action; unions at the workplace; and the internal organisation of unions. The key statutes were the Employment Act 1980, the Employment Act 1982, the Employment Act 1988, the Employment Act 1990 and the Trade Union Act 1984. In 1983, the Fair Wage Resolution Act 1949 was rescinded and in 1980, the Employment Protection Act 1975 was removed. The Wages Councils were abolished in 1993 (Brown and Wadhwani 1990; Gregory 1997).

Elgar and Simpson (1992) outline the importance of the removal of trade unions' immunity in terms of the scope for affected parties to seek damages to compensate for commercial loss caused by industrial action. ‘The potential importance of this goes beyond the very small number of claims actually made against trade unions in the decade after the law was changed’ (Elgar and Simpson 1992, p. 11). Other regulatory changes included: a narrower definition of industrial action; the requirement for secret ballots before industrial action is taken; a seven-day notice period of industrial action; and the proscription of secondary action, including picketing. The net effect of these changes, according to Gregory (1997), has been to substantially raise the costs and risks to unions of undertaking industrial action.

As far as trade unions at the workplace are concerned, statutory changes effectively outlawed pre- and post-entry closed shop arrangements; required periodic authorisation of deduction of union dues by workers, and removed the unions' access to the Advisory, Conciliation and Arbitration Service (ACAS) which, in turn, had provided assistance to unions to secure recognition by reluctant employers. Again, according to Gregory (1997, p. 3), the impact of these changes was to ‘increase the costs to the union of maintaining its membership and collecting union membership subscriptions’. Other changes affected the election of union officials and the requirement for secret ballots of members under specified procedures. A postal ballot was required every five years to elect union officials and a ballot of members every ten years was required if unions wished to use funds for political purposes. Members were able to pursue complaints against their union and unions could not discipline members for refusal to participate in industrial action (Brown and Wadhwani 1990).

In terms of the associated industrial relations outcomes in the United Kingdom, the two most significant were, first, the decentralisation of bargaining (although signs of decentralisation were apparent in the 1970s) and, secondly, the decline in the power of the trade unions. During the course of the 1980s and into the 1990s, there was a dramatic decline in the number of multi-employer agreements, replaced by either single-employer or plant contracts (Millward and Stevens 1986; Millward et al. 1992; Metcalf 1994). As Metcalf (1994, p. 5) stated in 1994:

[Twenty] years ago over 9 out of 10 workers in Great Britain had their pay determined either by collective bargaining (77%) or by minimum wages set by Wages Councils (15%). The fraction covered by these arrangements has halved since 1975.

Metcalf also noted that the United Kingdom experienced the most precipitous decline in the coverage of collective bargaining of all OECD countries.

The coverage of collective bargaining in the United Kingdom, as elsewhere, varies by firm size, with larger firms more likely to be covered. According to figures in the OECD Employment Outlook of 1994, the largest fall in collective bargaining coverage between 1985 and 1990 occurred among small firms (between 25 and 99 employees), from 53 per cent to 35 per cent. At the other end of the spectrum, collective-bargaining coverage among the largest firms (with 1,000 or more employees) fell from 89 per cent in 1985 to 77 per cent in 1990. In Table 3, we present figures on the dominant level of bargaining in 1984 and 1990. The figures clearly demonstrate the decline in multi-employer bargaining as well as the rise in the incidence of no collective bargaining. Gregory (1997) also notes that, in addition to the shift towards single-employer bargaining, the scope of bargaining has narrowed, with more issues unilaterally determined by employers. An example proffered of the latter is regulation of recruitment and staffing levels.

Table 3: Most Important Level of Collective Bargaining, United Kingdom
Percentage of British workers employed in plants with 25 or more employees
Manuals Non-manuals
1984 1990 1984 1990
Private manufacturing
Multi-employer 24 19 8 6
Single-employer, multi-plant 19 15 19 17
Plant 35 34 31 24
No collective bargaining(a) 22 32 42 53
Private services
Multi-employer 26 13 16 6
Single-employer, multi-plant 17 22 18 28
Plant 8 6 6 2
No collective bargaining(a) 49 59 60 64
Public sector
Multi-employer 74 72 87 76
Single-employer, multi-plant 20 17 11 12
Plant 2 1 1 0
No collective bargaining(a) 2 1 1 0

Note: (a) Includes ‘Other answer’.

Source: Metcalf (1994).

What was the impact of the changed industrial relations regulations for trade union density and activity? Table 5 presents figures on changes in unionisation rates in the three countries, Australia, New Zealand and the United Kingdom, which reveal remarkably similar patterns of decline. However, taking the United Kingdom alone, unionisation fell from 53 per cent in 1982 to 31 per cent in 1996. In absolute terms, union membership in the UK peaked at some 13.5 million in 1979, falling to around 8.5 million in 1997, with private sector unionisation standing at 23 per cent (Gregory 1997). Metcalf (1994) outlines five factors which he maintains are responsible for the decline in unionisation in the UK: changes in the structure and composition of the labour force and employment; macroeconomic conditions unfavourable to unions; regulatory changes hostile to collectivism and unionism; employer militancy, particularly the reluctance to grant recognition to unions in new establishments; and the strategy and tactics of the unions themselves. This holistic explanation contrasts with the conclusion of Freeman and Pelletier (1990), using time-series data, that an index of ‘favourableness of industrial relations laws to unionism’ is by far the most important explanation of union density.

Metcalf (1994) notes that there is no strong history in the UK, certainly compared with other EU countries, of regulation of enterprise numerical flexibility or working-time flexibility. While there are some laws governing mass redundancies (notice periods and minimum payouts), these laws are generally much weaker than in most EU countries. Metcalf (1994), however, points to the inclusion of rules governing employment separation in many collective agreements – hence, the decline of the coverage of collective agreements is relevant in this context. In addition, in 1985, the minimum qualifying period to claim unfair dismissal increased from one to two years, having risen from 26 weeks in 1979.

4. The Australian Experience[1]

The recent Australian experience differs markedly from that of the UK and New Zealand in at least three important ways. First, during the 1980s wage-determination structures initially became more centralised, not less. Second, the shift to more-decentralised structures that eventually took place occurred within the Accord policy framework, and hence with the co-operation of the trade union movement. Third, progress towards an enterprise- and individual-based system of bargaining has been much slower, with many of the traditional features of Australian industrial relations (notably the awards system) remaining intact.

4.1 Decentralisation of bargaining structures

Unlike the UK and New Zealand, the 1980s did not see a headlong rush towards decentralisation of bargaining structures. Indeed, with the introduction of the Accord between the ACTU and the Australian Labor Party as federal government policy in 1983, Australia returned to a highly centralised system. A key element of the Accord was the reliance on national wage adjustment of industry and occupation awards as the primary mechanism for wage determination. In its initial guise at least, the Accord required that trade unions exercise wage restraint and not seek additional wage increases outside of those provided by national wage case decisions. In return, the Accord provided the ‘guarantee’ that real wages would be maintained (though this would require the co-operation of the Conciliation and Arbitration Commission). Union members would also benefit more generally from increases in the ‘social wage’, as a result of the introduction of, and improvement in, a range of health and social welfare measures, and from the higher rates of economic growth that were presumed to flow as a consequence of improved co-operation between labour and capital and the exercise of real wage restraint.

Whether the Accord can be judged an economic success has been the matter of debate, though the consensus appears to be that the Accord was certainly successful in restraining real wage growth below what it would otherwise have been and thereby promoting employment growth, at least during the period 1983 to 1989 (see, for example, Chapman and Gruen 1990). Despite this apparent success, the nature of the Accord was to change markedly. As Dabscheck (1995, p. 143) has observed, the Accord Mark I ‘championed centralisation’ as the solution to Australia's industrial, economic and social problems, yet it was not long before the Accord partners were advocating more decentralised wages principles which ultimately were to lead to the emergence of a hybrid industrial relations system, with the traditional systems of awards reduced to benchmarks around which other wages and other conditions could be negotiated on an enterprise, workplace or even individual basis.

At one level, it can be argued that the recentralisation of wages and employment conditions under Accord Mark I was always doomed to fail in the longer term. It was, for example, incompatible with other government initiatives designed to increase the competitiveness of product markets and capital markets (such as reductions in tariffs, the floating of the dollar, the abolition of foreign exchange controls and general deregulation of the financial sector). The rise in competitive pressures was further heightened by the increased globalisation of world markets (resulting from, for example, advances in technology and the greater importance of global corporations). It thus was not long before the Accord partners were forced to acknowledge that the real-wage maintenance guarantee could not be sustained in the face of a deteriorating external position, leading to the 1986 National Wage Case decision being discounted for the price effects of currency depreciation.

Even more significant changes in the way the Conciliation and Arbitration Commission defined the operating principles for the federal wages system, however, were to follow in 1987 and 1988. Specifically, the focus shifted away from preserving real wages to encouraging productivity (which presumably would enable reductions in relative unit labour costs without the need for real wage reductions). Accord Mark III, for example, implemented by the Commission in March 1987, introduced a two-tier wages system where, under the Second Tier, wage increases (of up to four per cent) could result from direct negotiations between unions and employers at the enterprise level. While this process was highly regulated, with all increases requiring the endorsement of the Commission, and arguably delivered little in the way of sustainable productivity improvements (Frenkel and Shaw 1989; Reilly 1989), it represented an important break from the highly centralised wage determination practices of the recent past.

In 1988, with the prodding of both the ACTU and the Federal Government, the Commission backed away from this initial experiment with enterprise bargaining and returned to a system revolving around industry awards. Nevertheless, the intent was still to tie wages in some way to productivity and efficiency. Thus, at the August 1988 National Wage Case, the Structural Efficiency Principle (SEP), often referred to as award restructuring, was introduced. Under this principle, wage increases were to be contingent on the negotiation of award variations that would contribute to improving the competitiveness and efficiency of the industry. Again, the evidence suggests that this process was not particularly effective in achieving productivity gains. Almost by design, the SEP was implemented slowly and not widely adopted. Survey results reported in Sloan and Wooden (1990), for example, indicated a general failure of firms to implement the industry-agreed changes. Similarly, Still and Mortimer (1993) reported results which were broadly supportive of these findings, leading them to conclude that award restructuring had a minimal effect at the enterprise level.

The failure of both the Second Tier and the SEP to assuage macroeconomic pressures, together with rising pressure from business groups such as the Business Council of Australia (see, for example, BCA Industrial Relations Study Commission 1989; Hilmer et al. 1993), encouraged parties within the system to develop new approaches. Chief among these was the possibility of a more decentralised system which provided greater opportunities for employers and workers to negotiate directly over wages at the enterprise and workplace level. Increasingly, it was being recognised that delivering sustained and genuine productivity increases would require linking wages to the circumstances of individual enterprises, and by 1991 even the ACTU had embraced the concept of enterprise bargaining. The Commission, on the other hand, was far less supportive, initially rejecting the enterprise bargaining principle and then working to obstruct its implementation (by not approving enterprise agreements). Ultimately, it was the actions of the Commission which led first to amendments to the Industrial Relations Act 1988 (in July 1992), which reduced the ability of the Commission to become involved in the enterprise bargaining process, and then to the enactment of the Industrial Relations Reform Act in 1993. A key feature of the Act was the introduction of Enterprise Flexibility Agreements which, unlike the already existing Certified Agreements, could be negotiated at workplaces which had few, if any, union members.

Finally, and most recently, industrial relations reform has been provided with renewed vigour through the Federal Coalition Government's Workplace Relations Act 1996. The main development under this Act was that, for the first time, individual agreements (known as Australian Workplace Agreements or AWAs) could be struck directly between employers and workers (and without the intervention of unions if so desired) and would be recognised as legally binding before the Commission.

State Governments, unlike the Federal Government, are not subject to constitutional limitations in the area of industrial relations. This (comparative) degree of freedom in matters relating to industrial relations has meant each State has developed its own system of operation. It is, perhaps, for this reason that major legislative reforms to industrial relations systems have tended to occur in State jurisdictions well ahead of the federal system. In the Queensland system, formal individual-level agreements were introduced in 1987, while in New South Wales, the introduction of the Industrial Relations Act in 1991 represented the first comprehensive reform of industrial relations processes and practices. This greater freedom, however, has also contributed to greater flux and volatility within the State systems, with the legislative reforms of one government often being overturned by its successor. The NSW reforms, for example, were largely replaced by legislation introduced by the newly elected Labor Government in 1996. That said, the direction of long-term change in all systems, irrespective of the philosophical persuasion of governments, appears to be uniformly towards greater emphasis on enterprise- and workplace-level arrangements.

The clearest evidence of the rising importance of decentralised bargaining structures is provided by the growing number of workers covered by enterprise agreements. Within the federal system, the number of registered agreements has steadily grown since 1991 when formalised collective agreements first became possible. Between October 1991 and October 1997, around 15,000 federal agreements had been formalised by the Australian Industrial Relations Commission, with the number of employees estimated to be covered by these agreements reaching 1.74 million by late 1996, or 64 per cent of employees within the coverage of the federal awards system (Figure 1). Figure 1, however, also suggests that growth in coverage may be slowing. Indeed, in the twelve-month period to the September quarter of 1997, the number of employees covered by federal ‘wage’ agreements actually declined, falling by 191,700. It thus might seem reasonable to conclude that under the Workplace Relations Act 1996, the trend towards enterprise-based bargaining has slowed and possibly even reversed. There are, however, a number of reasons for suspecting that such a conclusion is likely to be very wide of the mark.

Figure 1: Employee Coverage of Federal Enterprise Agreements

First, the trend in the number of agreements ratified is still clearly upwards, especially in the private sector. Growth in the number of ratified agreements, together with declining employee coverage, must mean that many agreements are expiring and are not being renewed. This, however, does not in itself imply a fall in the level of interest in enterprise-based agreements. Instead, it may be that at many workplaces there is no need to renegotiate new agreements. Expiry of an agreement will typically not mean that wages and conditions revert to that specified in awards. Rather, the conditions that applied during the term of an agreement will continue to be applied. Given both satisfaction with the operation of the previous agreement and the absence of wages pressures (which can be expected given the low levels of price inflation), there may be no need for any of the parties to continue to formalise their ongoing relationship. The process of bargaining at many Australian firms, and especially those without well-developed human resource management functions or a well-organised union presence, may therefore be one of intermittent formalisation. That is, agreements are renewed on a needs basis and driven largely by the need for change and/or to deliver wage increases.

Second, part of the decline in employee coverage of agreements may reflect downsizing within the public sector. At the end of the September quarter 1996, 579,900 public sector employees were covered by certified wage agreements. One year on, employee coverage in the public sector had declined to 339,200 employees.

Third, at least some of the decline in the number of employees covered by formal collective agreements may reflect a shift towards individual agreements.

Fourth, the decline in coverage of ‘wage agreements’ may reflect growth in the number of agreements where it is difficult to quantify the associated wage increase.

This upward trend in the coverage of employees by enterprise agreements has also been apparent within the State systems, though the spread of formal enterprise agreements has developed more slowly within the State jurisdictions. As reported in Table 4, data for September 1996 indicate rates of coverage (of award-based employees) varying from just six per cent in Tasmania up to 39 per cent in Queensland (Joint Governments 1997, p. 91). In total, the coverage of award-based employees by enterprise agreements throughout Australia would appear to be close to 50 per cent.

Table 4: Employees Covered by Formal Enterprise Agreements by Jurisdiction
September 1996
Jurisdiction Estimated no. of employees covered (Thousands) Per cent of award-based employees covered
Federal 1,740 64
States:
NSW 360 31
Victoria n.a. n.a.
Queensland 265 39
South Australia 82 33
Western Australia 86 29
Tasmania 4 6
Total (excl. Victoria) 2,537 49

Source: Joint Governments (1997, pp. 89–91).

Overall, the shift towards enterprise-based bargaining arrangements would appear, at least on the surface, to be impressive. Enterprise agreements, however, can vary enormously in terms of what they deliver. For the majority of employees covered by enterprise agreements, the agreement still has to be read in conjunction with awards (Buchanan et al. 1997, p. 101). Indeed, it has been argued that in some sectors, not only do agreements not wholly replace the award, but those agreements often only touch upon a narrow range of work conditions. Growth in the incidence of enterprise agreements, therefore, does not necessarily convey any useful insights into their significance.

With respect to individual-based agreements, less is known. Australian Workplace Agreements (AWAs) only became operational on 12 March 1997 and hence, it is too early to pass judgment on their likely impact. Nevertheless, the evidence to date is suggestive of a modest uptake, with 15,750 AWAs (covering 765 employers) having been received by the Office of the Employment Advocate by 28 February 1998. Of these, around 7,500 had been approved. AWAs, however, do appear to be growing quite rapidly, with approximately half of the total number of applications being received in the three-month period between December 1997 and February 1998.

Moreover, there are good reasons to expect the rate of AWA lodgement to be slow. First, the process for having an AWA approved may still be perceived by many employers as complex and cumbersome. Indeed, it is possible that the availability of AWAs may stimulate parties to reach individual agreements which are then not formalised. Certainly, the incentives for employers to formalise agreements negotiated with individual employees do not appear to be great, especially given a labour market where, because of the presence of large-scale unemployment, management is typically in a much stronger bargaining position than are employees. Second, union opposition to non-collective forms of bargaining will almost certainly reduce interest in, and hamper attempts to introduce, AWAs in many enterprises, and especially those with a well-organised union presence. Third, some enterprises which may desire individual-based agreements will be locked-in to collective agreements reached under the previous Act.

Progress is likely to have been more rapid in some of the States, especially in Western Australia where legislation facilitating individual employment agreements or contracts dates back to 1993. According to Mulvey (1997, p. 236), around 80,000 employees and almost 2,300 employers in Western Australia had signed the new workplace agreements by the end of 1996. This figure, however, included both collective and individual agreements. Excluding collective agreements would reduce coverage to about 66,000 employees, still an impressive figure given there were only about 300,000 employees represented within the WA system at the time.

Overall, the incidence of individual agreements within the formal systems remains relatively uncommon. While the industrial relations climate for pursuing the implementation of such arrangements is clearly much more favourable today than just a few years ago, it is not easy to see why many employers currently operating with informal arrangements with individual workers would seek to formalise those arrangements, especially given that within the federal system, individual agreements are subject to external review.

4.2 Changing role of trade unions

There can be little doubt that the shape and structure of trade unions in Australia have changed enormously over the past decade or so. Trade union membership as a proportion of the employee workforce (or union density) has been falling continuously since the early 1980s and, in the most recent figures (August 1997), stood at just 30.3 per cent. As argued by Griffin and Svensen (1996), the explanation for this decline is multi-factorial, though numerous commentators (e.g. Berry and Kitchener 1989; Peetz 1990) have emphasised the importance of structural factors, such as changes in the industrial composition of employment, the growth of part-time and casual employment, and the relative decline of public sector employment. Other likely contributing explanations include the decline in the incidence of compulsory unionism (Peetz 1996), changes in management strategy (Peetz 1997), which in turn may be the result of changes in product markets (Bodman 1998), and even changing union strategy.[2]

Whatever the explanation, it would appear that unlike the UK and New Zealand, very little of the blame for the decline in trade union membership can be directly attributed to changes in legislation. Potentially, the area where legislation might have had most impact relates to union preference and compulsory union membership. Opinion poll data reported by Rawson (1992) and Peetz (1996), for example, indicate that compulsory union membership is on the wane, with the proportion of union members in ‘closed shops’ apparently declining from about 67 per cent in 1978 to between 33 and 36 per cent by 1996. Indeed, Drago and Wooden (1998) have calculated that the decline in the incidence of compulsory unionism has had the effect of reducing the overall unionisation rate since 1976 by almost five percentage points. It is, however, difficult to relate these changes to changes in legislation. In the federal arena, for example, union preference has only recently been outlawed (following the enactment of the Workplace Relations Act 1996), and even at the state level, it does not seem to follow that legal proscription has had much impact on actual practice (McGraw and Palmer 1994; Weeks 1995). Indeed, Weeks (1995, p. 92) has argued that legislation has not been a substantial factor influencing union membership, largely because the majority of compulsory union membership arrangements were the result of consensual agreements between employers and unions. Presumably, therefore, the decline in compulsory unionism has been mainly the result of employers withdrawing support for such arrangements.

Legislation, however, has been of much greater importance in facilitating changes in trade union structures. Traditionally, the trade union structure in Australia reflected craft and occupational distinctions, and hence there were a large number of unions, with the average union having a small number of members spread across a large number of workplaces. From the employer perspective, such arrangements were highly inefficient given it meant that many employers often had to deal with more than one union. More importantly, unions structured on an occupational basis, as distinct from an industry or enterprise basis, are less likely to take into account the circumstances of individual enterprises when bargaining. The union movement also saw significant weaknesses in the traditional union structure, and in the 1980s committed itself to ‘the creation of 20 large and efficient union federations’ (ACTU 1987). From the perspective of the ACTU, fewer and larger unions were required because of economies of scale in the provision of services and in order to eliminate inter-union conflict and rivalries that were fostered by the traditional craft- and occupational-based system. The intent of the ACTU plan, therefore, was to create a system of super unions organised along broad industry lines.

Compliance with this strategy was assisted by legislative changes within the federal system during the late 1980s and early 1990s which:

  1. effectively removed the requirement that amalgamations could not take place without the approval of 25 per cent of the membership;
  2. removed the requirement that members of a large union involved in an amalgamation had to be balloted, if membership of that large union was more than four times larger than the other parties to the merger;
  3. provided the Commission the power to give exclusive coverage to a single union when resolving demarcation disputes;
  4. increased the minimum size limit required for unions to maintain registration to first 1,000 members and then 10,000 members;[3] and
  5. made financial assistance available to unions intending to amalgamate.

The end result of this process has been a marked reduction in the number of active trade unions – from 299 in 1990 to 132 in 1996 – with the large majority of union members in 1996 (88 per cent) represented by just 24 unions.[4] Thus, while the level of bargaining has been gravitating towards the enterprise level, union structures have actually become more centralised.

5. Summary: The Australian, New Zealand and UK Experiences Compared

In all three countries, there has been a noticeable shift towards enterprise-based bargaining, away from either industry bargaining and/or national arrangements. In the UK, the trend towards decentralisation of pay determination was evident prior to the aggressive labour market reform measures initiated by the Thatcher Conservative Government. However, the trend accelerated during the 1980s and 1990s to the point that industry bargaining, in the private sector at least, is now relatively uncommon. Plant-level bargaining, by contrast, is relatively common in the UK, with a significant proportion of the workforce not covered by formal agreements. In the cases of New Zealand and Australia, the changing role of the award systems has been a central feature of their shift towards enterprise-based bargaining. The ECA in New Zealand effectively abolished all awards and the trend to individual and collective agreement-making was understandably rapid. Associated with this transformation was the rapid uptake of individual employment agreements, with the penalties for strike action taken in pursuit of multi-employer agreements an important inhibitor of collective agreement-making. In Australia's case, the shift away from national and sectoral bargaining to enterprise-based bargaining has been slower and less pervasive than in the other two countries. For most of the past decade, a mixed arrangement has existed, with national (award-based) pay rises forming part of the overall system of pay determination. In more recent times, there has been some take-up of individual bargaining, but most of the bargaining that does occur remains collective with trade unions as party.

Despite these differences, trends in other industrial relations outcomes, such as trade union membership and industrial disputes, appear to be remarkably similar. In all three countries, trade union density has fallen dramatically. According to the figures reported in Table 5, trade union membership as a proportion of employees has fallen from somewhere close to, or in excess of, 50 per cent at the start of the 1980s in all three countries, to 31 per cent in Australia and the UK, and to 25 per cent in New Zealand, by 1996. Moreover, the data collected for New Zealand will understate the actual level of trade union density given it is measured in terms of full-time-equivalent trade union members.[5] Consequently, the level of union density in New Zealand is in all likelihood, much closer to the level in Australia and the UK than the figures reported in Table 5 suggest. The New Zealand experience, however, is different from the other two countries in one important respect – virtually all of the decline in trade union membership occurred after the enactment of the ECA; that is, after 1990.

Table 5: Unionisation Rates
Australia, New Zealand and the United Kingdom; per cent
Year Australia New Zealand United Kingdom
1976 50(a) 50 54
1982 48(b) 55(c) 53
1986 46 55 47
1992 35 37 36
1996 31 25 31

Definitions: Australia: Trade union members in main job (from household survey) as a percentage of the number of civilian employees.
New Zealand: 1986–96 – Full-time-equivalent trade union members (as reported by trade unions) as a percentage of the number of civilian employees. 1976–82 – Registered union members in the private sector as a percentage of private sector wage and salary earner employment.
United Kingdom: 1992–96 – Trade union members (from household survey) as a percentage of civilian employees. 1976–86 – Number of trade union members (as reported by trade unions) as a percentage of the number of civilian employees.
Notes: (a) The 1976 publication reports a unionisation rate of 51 per cent, which does not square with the reported number of union members and the reported number of employees.
(b) The published figure in the 1982 survey (49.5 per cent) included as union members, employees who were members of the union in their second job but not their main job. The number of union members has been adjusted downwards (by 61600) to account for this.
(c) 1981 data.

Sources: Australia: Trade Union Members Survey, Australia, ABS cat. no. 6325.0, various issues.
New Zealand: Union membership data for 1986, 1992 and 1996 come from Crawford, Harbridge and Hince (1997). Employee data come from OECD, Labour Force Statistics, various issues. Data for earlier years come from OECD (1991, Chapter 4).
United Kingdom: Data for 1992 and 1996 come from Cully and Woodland (1997). Data for earlier years are from Waddington (1992).

As far as changes in trade union structures are concerned, in all countries there has been a trend towards fewer, larger unions. In all cases, legislation, as well as changes to the external environment, contributed to altering structures, governance and activities of trade unions. In the UK, for instance, the increased costs and difficulties of recruiting and maintaining members produced significant changes to union activities, with attention particularly directed towards securing recognition by employers and providing services to the rank and file. There has also been a marked decline in the overall financial position of the trade union movement in the UK (Metcalf 1994). A similar story applies a fortiori to New Zealand, with the demise of a number of unions after the enactment of the ECA and a shift of focus to representing workers at the enterprise level (Gardner 1995). In the case of Australia, the rationalisation of unions was mainly the consequence of a union-initiated strategy, aided briefly by supportive legislation. Prima facie, the focus of Australian unions would appear to be less member-centred than is the case in either the UK or New Zealand.

The trend in industrial disputation in all three countries over the course of the 1980s and 1990s has also been quite similar. Much has been made of the effect of the Accord in promoting a marked improvement in levels of industrial disputation in Australia (e.g. Beggs and Chapman 1987; Morris and Wilson 1995), yet Figure 2 demonstrates that significant sustained reductions in industrial action also occurred in both New Zealand and the UK during the 1980s and 1990s. Indeed, if the impact of the UK coalminers strike in 1984 is ignored, the UK experience would closely mirror that of the Australian experience. In other words, the aggressively anti-union stance adopted in the UK and the corporatist approach adopted in Australia appear to have resulted in very similar outcomes when measured in terms of long-run disputation levels.[6]

Figure 2: Industrial Disputes – Australia, New Zealand and the UK

6. Decentralisation and the Consequences for Labour Market Outcomes

6.1 Theory

The shift towards more decentralised bargaining structures is typically justified on the grounds that it is conducive to superior macroeconomic outcomes. As explained in OECD (1997, pp. 64–65), the principal argument here revolves around differences in elasticity of demand in the product market. With the exception of monopoly firms, the elasticity of demand for the output of an individual firm will typically be greater than that for an aggregation of firms (e.g. an industry). The simple reason for this is that there are more substitutes for the output of an individual firm than for the output of an industry. As a result, the trade-off between wages and employment will be much larger at the enterprise level than at the industry level. Consequently, enterprise-level bargaining is much more likely to be employment sensitive than is industry-level bargaining.

On the other hand, wage bargains may be associated with negative externalities which will not be internalised by bargaining agents when bargaining structures are highly decentralised. The types of externalities are summarised in OECD (1997, p. 65), and include the impact of higher wages for one group of workers on consumer prices, on input prices in other sectors, on government expenditure (given higher wages generate some additional unemployment), and on the morale of other workers not covered by the bargain, and the generation of inflationary pressures from ‘leap-frogging’ pay claims. With decentralised structures, bargaining agents will have little incentive to take into account such negative externalities given those that benefit represent only a small fraction of those harmed. In contrast, with highly centralised bargaining structures, the distinction between those who benefit and those who are harmed is less obvious, and hence bargaining agents will be more inclined to internalise the wider economic consequences of their bargaining behaviour.

On theoretical grounds, therefore, it is difficult to determine a priori whether decentralised wage systems will produce superior economic outcomes to more centralised systems. In a very influential article, however, Calmfors and Driffill (1988) argued that the net result of the interaction of these competitive effects and externality effects is a hump-shaped relationship between the degree of centralisation and real wages/unemployment, with the best-performing economies being those with either highly decentralised systems or highly centralised systems. This hypothesis has a lot of intuitive appeal. In a highly decentralised system, with bargaining occurring on a firm-by-firm basis with unions either non-existent or structured on an enterprise basis, wage outcomes are likely to be so fragmented that the sum total of negative externalities will be quite small. On the other hand, where bargaining takes place at the national level and involves a small number of large encompassing groups, the incentive for the bargaining agents to redistribute income towards themselves without taking into account the social cost will be quite small.

6.2 Labour market outcomes in Australia, New Zealand and the UK

At the start of the 1980s, Australia, New Zealand and the UK can all be described as occupying positions close to the top of the Calmfors-Driffill hump. This was particularly true of Australia and New Zealand which combined centralist elements, such as an industry-based awards system, with an occupational-based trade union structure and relatively poor levels of employer co-ordination. The Accord in Australia, therefore, can, at least initially, be viewed as an attempt to move Australia down the hump to the right (increased centralisation). From the late 1980s, however, policy direction reversed, and Australia has presumably moved back up the hump and down the other side. The Calmfors-Driffill hypothesis does not, therefore, generate an unambiguous prediction about the net impact of industrial relations change on macroeconomic performance in Australia over the course of the 1980s and 1990s. The New Zealand and UK economies, however, have clearly moved down the hump to the left (reduced centralisation) and hence, the Calmfors-Driffill hypothesis does suggest that the overall economic performance of both economies should have improved, though in the case of New Zealand such effects should be largely confined to the post-ECA period.

Identifying the economic consequences of changes in institutional arrangements within the labour market, however, is no simple matter, especially at the economy-wide level. As Dawkins (1997) has observed, linking changes in institutional arrangements to other changes is complicated by at least three factors. First, at the same time as changes in the institutional arrangements are taking place, other changes that may impact on outcomes will also be taking place. As a consequence, it will be difficult to isolate the contribution to economic performance from changing institutional arrangements. Second, institutional arrangements within the labour market may not be independent of economic performance; that is, rather than a cause of economic performance, change in institutional arrangements may be a response to the economic environment. Third, the process of labour market deregulation (and re-regulation) is itself a complex process which may affect different variables in different ways.

This problem in identifying the counterfactual is highlighted in Figure 3, which charts the course of the aggregate unemployment rate for Australia, New Zealand and the UK since 1978. This figure reveals that despite differences across the countries in the pace and scope of labour market reform, there is a common pattern in movements in unemployment rates across the three countries, highlighting the importance of international business cycles for variations in economic activity. Important differences, however, are still evident. Most obviously, and consistent with the Calmfors-Driffill hypothesis, during the 1980s the unemployment rate improved in both Australia (which was becoming more centralised) and in the UK (which was becoming more decentralised). On the other hand, in New Zealand, where industrial relations structures had not yet greatly changed, employment stagnated and the unemployment rate worsened well ahead of the downturn in the international business cycle that became apparent towards the end of 1990.

Figure 3: Unemployment Rates

The timing of this downturn suggests that the changes introduced as part of the 1987 Labour Relations Act might in some way be implicated in the rise in unemployment in New Zealand. This, however, seems highly unlikely given the argument advanced earlier that the effective impact of this legislation was quite limited, with most unions preferring to keep their members on awards. More likely explanations lie in other features of the economic reform process, including privatisation of public sector monopolies and restrictive monetary policy, and large increases in the minimum wage in 1985 and 1987.[7]

The international business cycle bottomed in 1991 and hence it is hardly surprising that the unemployment rate improved in all three countries in the years that followed. Nevertheless, it is of some interest that the improvement appears to have been most rapid in New Zealand, following the introduction of the ECA, and least marked in Australia, where the decentralisation process has been relatively modest. That said, there are signs in recent data that unemployment in New Zealand is now rising again.

Overall, the unemployment data presented in Figure 3 are consistent with a Calmfors-Driffill interpretation, but prevent definitive conclusions from being drawn. Data on employment-population (EP) ratios, on the other hand, do not lend themselves to a Calmfors-Driffill interpretation. As shown in Figure 4, there is no evidence at all of a sustained rise in the EP ratio in the UK or New Zealand. Only in Australia is the EP ratio in 1996 notably higher than in 1978.

Figure 4: Employment/Population Ratios

Data on growth in real earnings (Figure 5) also do not appear to be consistent with the Calmfors-Driffill hypothesis. Only the subsidence of real wages growth in Australia during the early years of the Accord is clearly consistent with the hypothesis. There appears to be a general reduction in the level of volatility in real earnings growth in recent years in all three countries, but this is a feature common to many other industrial nations.

Figure 5: Growth in Real Compensation per Employee

Unlike other outcomes, earnings dispersion is expected to be associated with the degree of centralisation of bargaining structures in a linear fashion. That is, decentralised bargaining structures are expected to be associated with a greater diversity in wage outcomes across enterprises and across individual workers, thus leading to greater earnings inequality. The data reported in Table 6 appear to be broadly consistent with this hypothesis. The degree of earnings dispersion was very similar in each of the three countries at the start of the 1980s. Over the next two decades or so, the degree of earnings dispersion increased. Moreover, the extent of the increase in dispersion is broadly in line with the extent to which the different countries have decentralised bargaining structures, with the gap between the highest paid and the lowest paid widening most in the UK and least in Australia.

Table 6: Trends in Earnings Dispersion, Full-time Employees
Australia, New Zealand and the UK; ratio of earnings in the 90th percentile to the 10th percentile (D9/D1)
1980 1984 1988 1992 1995 Annual
percentage change
Australia
Males 2.72 2.73 2.89 2.92 2.94 0.52
Females 2.54 2.70 2.64 2.56 2.53 −0.003
Total 2.84 2.89 2.89 2.82 2.92 0.19
New Zealand
Males 2.72 2.85 3.13 3.17(a) 1.54
Females 2.42 2.51 2.71 2.62(a) 0.80
Total 2.89 2.92 3.08 3.04(a) 0.51
United Kingdom
Males 2.51 2.77 3.03 3.20 3.31 1.86
Females 2.34 2.51 2.81 3.00 3.06 1.80
Total 2.79 3.04 3.24 3.31 3.38 1.29

Note: (a) 1994 data.

Source: OECD, Employment Outlook, July 1996.

Decentralised bargaining structures are also typically argued to be conducive to more flexible employment arrangements. One possible manifestation of such flexibility is part-time employment. As documented in Table 7, the prevalence of part-time employment has increased in all three countries. The rate of growth of part-time employment, however, appears to have been greatest in Australia where the progress towards decentralised bargaining structures has been slowest. Moreover, growth, at least in the case of Australia and New Zealand, has been greatest during periods of relatively high levels of centralisation. Overall, there are good reasons to suspect that the factors driving growth in part-time employment have very little to do with the changing industrial relations arrangements. Finally, it should be noted that the data provided in Table 7 do not permit direct comparisons of the level of part-time employment across the three countries since part-time employment is defined differently in each.

Table 7: Part-time Employment as a Percentage of Total Employment
Australia, New Zealand and the UK
1973 1979 1983 1990 1994 1996
Australia
Males 3.7 5.2 6.2 8.0 10.9 11.7
Females 28.2 35.2 36.4 40.1 42.6 42.6
Total 11.9 15.9 17.5 21.3 24.4 25.0
New Zealand
Males 4.6 4.9 5.0 8.4 9.7 10.4
Females 24.6 29.1 31.4 35.0 36.6 37.3
Total 11.2 13.9 15.3 20.0 21.6 22.4
United Kingdom
Males 2.3 1.9 3.3 5.2 7.1 5.6
Females 39.1 39.0 42.4 42.6 44.3 42.7
Total 16.0 16.4 19.4 21.3 23.8 22.1

Notes: Definitions of part-time:
Australia: Usually worked less than 35 hours a week and did so in the survey week.
New Zealand: 1986–90 – Actually worked less than 30 hours in survey reference (unless worked zero hours, in which case, determined by usual hours). 1973–90 and 1990–96 – Usually worked less than 30 hours per week.
UK: Based on respondent's classification.

Sources: van Bastelaer, Lemaitre and Marianna (1997); OECD, Employment Outlook, July 1997.

6.3 Wider international comparisons

Given the difficulty identifying the counterfactual from time-series data, many researchers have resorted to comparisons between countries with varying institutional arrangements. Indeed, such comparisons are used by Calmfors and Driffill (1988) to support their hypothesis and have spawned a number of similar studies. Most studies that have used a measure of centralisation have found at least some support for the hypothesis (e.g. Bleaney 1996; Freeman 1988; Golden 1993; Heitger 1987; Rowthorn 1992; Scarpetta 1996). On the other hand, some researchers have argued that it is not the level at which bargaining takes place, but the extent to which bargaining is co-ordinated. These researchers typically find a linear relationship between performance measures and the degree of co-ordination (e.g. Soskice 1990; Layard, Jackman and Nickell 1991; Bean 1994). Finally, in what is arguably the most extensive evaluation of the Calmfors-Driffill hypothesis to date, the OECD (1997, Chapter 3) could find no robust evidence for either a U-shaped relation between the structure of collective bargaining and employment, or a hump-shaped relation with unemployment.

Is this the final word? A critical determinant of the results obtained is how countries are ranked, or assigned to clusters, according to the degree of centralisation of bargaining, and there is clearly a good deal of subjectivity associated with this process. Examples of this are the OECD's treatment of New Zealand and Australia. The OECD assigned New Zealand to the decentralised category for all three of the periods considered in its analysis – 1980, 1990 and 1994 – and assigned Australia to the centralised category in 1980 and 1990, and to the decentralised category in 1994. The OECD study thus has Australia and New Zealand poles apart in 1980, when the reality is that the systems were very similar. Both operated with compulsory arbitration systems, with wages and conditions largely determined by industry and occupation awards, but with some informal over-award bargaining over and above this. In our view, at the start of the 1980s, Australia and New Zealand were the classic intermediate economies. The OECD classification also implies a radical change in Australian industrial relations structures in the early 1990s. Again, and as argued earlier, while a shift towards less centralised structures has been occurring, it is still very misleading to describe Australia as having embraced an enterprise-based bargaining structure. The reality is that the awards system is still vitally important, with enterprise agreements often mere supplements to awards. More importantly, around half of all employees on awards are still not covered by enterprise agreements.

Perhaps of even greater significance for the results, the OECD classifies two of the best-performing economies, Switzerland and Japan, as intermediate economies on the basis of their apparent high level of co-ordination of wage bargaining, even though Calmfors and Driffill treat these as decentralised economies. The treatment of Switzerland is understandable given the predominance of industry-level bargaining. However, it is difficult to see how Japan can be classified as an intermediate economy. It is one of the few developed economies (along with the US and Canada) where enterprise- and plant-level bargaining have been predominant over a sustained period of time, trade unions are relatively weak, and the level of coverage of collective bargaining is very low (at about 21 per cent of the workforce according to the OECD (1997)). It is this latter factor which is of crucial importance. While the evidence suggests that the bargaining process in Japan (Shunto) does deliver highly uniform settlement rates (Sako 1997), the wages of most workers are presumably determined by other mechanisms.

Using the data reported by the OECD (1997, p. 71), a measure of bargaining structure has been constructed that uses indices of both centralisation and co-ordination, but where the latter is adjusted for the extent of coverage of collective bargaining. The results are reported in Table 8. As is common practice, the sum of the scores on these two indices have been used to rank countries. These rankings are then used as an aid to dividing countries into three clusters: centralised, intermediate and decentralised economies. We have classified five countries as centralised and five as decentralised, with the remaining nine countries assigned to the intermediate category. The centralised cluster includes all four countries where national-level bargaining remains of large importance, as well as Germany, where sector-level bargaining is complemented by highly co-ordinated wage rounds and rates of coverage of collective agreements of in excess of 90 per cent. At the other extreme, the decentralised cluster includes only those countries where, by 1994, multi-employer bargaining was not the norm.

Table 8: Measures of Centralisation and Co-ordination of Wage Bargaining
OECD countries; 1994
Centralisation index Adjusted co-ordination index Total index
Centralised economies
Austria 2.5 2.9 5.4
Finland 2.5 2.4 4.9
Belgium 2.5 2.3 4.8
Germany 2 2.8 4.8
Norway 2.5 1.9 4.4
Intermediate economies
Italy 2 2.1 4.1
France 2 1.9 3.9
Sweden 2 1.8 3.8
Denmark 2 1.7 3.7
Spain 2 1.6 3.6
Netherlands 2 1.6 3.6
Portugal 2 1.4 3.4
Switzerland 2 1.3 3.3
Australia 1.5 1.2 2.7
Decentralised economies
United Kingdom 1.5 0.5 2.0
Japan 1 0.6 1.6
Canada 1 0.4 1.4
New Zealand 1 0.3 1.3
United States 1 0.2 1.2

Sources: The centralisation index is taken from OECD (1997, Table 3.3, p. 71). The adjusted co-ordination index is calculated as C x B where C is the OECD's co-ordination index and B is the bargaining coverage percentage (as reported in OECD 1997, Table 3.3, p. 71).

Following Calmfors and Driffill (1988), various measures of macroeconomic performance are then averaged across the countries within each cluster. Table 9 reports the results of this exercise using the unemployment rate, the employment to population ratio, the Okun index (the unemployment rate and the rate of price inflation summed), and a measure of earnings dispersion. The figures reported in this table do appear to be consistent with the Calmfors-Driffill hypothesis, with the intermediate economies performing worst on all measures, except earnings dispersion. Also, the data presented in Table 9 suggest that labour markets in decentralised economies are more successful in generating employment than those in centralised economies, but as expected, this greater employment has been associated with higher levels of earnings dispersion. That said, there are a number outliers in the data, and hence the figures reported in Table 9 can only be described as providing weak support for the Calmfors-Driffill hypothesis. A number of the intermediate economies (especially Switzerland, but also Denmark and the Netherlands), for example, appear to have been quite successful in generating relatively good employment outcomes.

Table 9: Indicators of Labour Market Performance
OECD countries
Unemployment rate (Per cent) Employment/population ratio(a) (Per cent) Okun index(b) Earnings dispersion (D9/D1)
1996 1986–96
Average
1996 1986–96
Average
1996 1986–96
Average
1996(c)
Centralised economies
Austria 4.4 5.2   68.1 66.3   6.3 7.9   3.66
Finland 15.7 10.2   62.1 68.0   16.3 13.6   2.53
Belgium 9.8 11.2   56.6 56.2   11.9 11.0   2.25
Germany(d) 9.0 7.3   64.0 64.9   10.5 9.4   2.32
Norway 4.9 4.6   75.7 74.7   6.2 8.7   1.98
Average 8.8 7.7   65.3 66.0   10.2 10.1   2.55
Intermediate economies
Italy 12.0 10.3   51.3 53.8   15.8 15.5   2.80
France 12.4 10.6   58.8 59.2   14.4 13.2   3.28
Sweden 10.0 4.5   69.8 76.3   10.8 9.5   2.13
Denmark 6.0 9.8   74.2 75.9   8.1 12.7   2.17
Netherlands 6.3 6.9   66.4 62.4(e)   8.4 8.7   2.59
Spain 22.2 19.8   47.1 47.2   25.9 31.5   n.a.
Portugal 7.3 6.0   67.4 67.4   10.4 14.7   4.05
Switzerland 3.5 2.2   79.3 n.a.   4.3 4.9   2.65
Australia 8.6 8.5   68.8 67.6   11.2 13.5   2.87
Average 9.8 8.7   64.8 63.7   12.1 13.8   2.82
Decentralised economies
Japan 3.4 2.6   74.6 72.8   3.5 3.7   3.02
United Kingdom 8.2 8.5   68.7 68.3   10.6 13.0   3.31
New Zealand 6.1 7.2   71.1 68.9   8.4 12.6   3.04
Canada 9.7 9.5   67.6 68.3   11.3 12.7   4.20
United States 5.4 6.2   73.6 72.2   8.3 9.7   4.35
Average 6.6 6.8   71.1 70.1   8.4 10.3   3.58

Notes: (a) Total employment divided by total population aged 15 to 64 years, as reported in OECD, Labour Force Statistics, 1973–1996.
(b) Sum of the unemployment rate and the consumer price inflation rate.
(c) Data for Belgium, Germany, Italy, Portugal and Sweden are from 1993, for Denmark from 1990, and for Norway from 1991.
(d) Up to and including 1992, data for Germany only apply to West Germany.
(e) Averaged over period 1987–97.

Sources: OECD, Employment Outlook, June 1997; OECD, Employment Outlook, July 1996; OECD, Labour Force Statistics, 1976–1996; OECD, Economic Outlook, June 1997.

Finally, and again following the OECD (1997), we examine whether changes over time in bargaining structures are associated with changes in macroeconomic indicators. The OECD examined the period 1980 to 1990 and claimed to find that, if anything, countries that moved to more decentralised arrangements had performed worse. The analysis undertaken by the OECD, however, does not provide a test of the Calmfors-Driffill hypothesis because it did not account for an economy's initial starting position on the hump. The OECD identified six countries as moving towards less centralised arrangements during the 1980s: Denmark, Finland, New Zealand, Spain, Sweden and the UK. According to the OECD's own centralisation rankings, four of these countries – Denmark, Finland, Spain and Sweden – possessed relatively highly centralised bargaining structures in 1980. In these cases, the Calmfors-Driffill hypothesis leads us to expect that any shift towards less centralised arrangements would, at least initially, be associated with a deterioration in macroeconomic performance.

Here we examine the period 1986 to 1996. In contrast to the OECD, we distinguish between economies moving down the hump and those moving up the hump. Drawing largely on the discussion provided in OECD (1994, 1997), the economies identified as moving down the hump are New Zealand, the UK, the US and Norway, with Norway being the only country of the group moving down the hump towards greater centralisation. The economies identified as movers up the hump are Finland, France, Italy, Portugal and Sweden, with Finland and Sweden moving up the hump from a highly centralised position, while the other three economies have moved up the hump from the other direction. With the exception of Australia and Denmark, the remaining economies are assumed to have had more or less stable bargaining structures over the period considered.[8] Industrial relations structures in Australia and Denmark, on the other hand, have clearly been changing. Nevertheless, since both of these economies were arguably positioned to the right of the hump during the mid 1980s, and were moving toward the other side of the hump (i.e. towards less centralised arrangements), it is impossible to predict the net impact on macroeconomic performance indicators.

As demonstrated in Table 10, the experience during the period 1986–96 is highly consistent with the Calmfors-Driffill hypothesis. The economies identified as moving away from the intermediate position with respect to bargaining structure performed better, on average, on all indicators, than the economies moving towards the intermediate position. The ‘movers up the hump’ typically experienced rising rates of unemployment, falling rates of employment and very little change in the Okun index. In contrast, the ‘movers down the hump’ experienced little change in unemployment and employment rates, but did experience a subsidence in inflationary pressures (as indicated by improving scores on the Okun index). Together, these indicators suggest a reduction in the NAIRU. The movers down the hump also fared better than the so-called stable economies on both unemployment and the Okun index. The stable economies, however, have fared just as well with respect to jobs generation.[9]

Table 10: Change in Economic Performance
OECD countries
Country Unemployment rate Employment/population ratio(a) Okun index(b)
1990–96 1986–96 1990–96 1986–96 1990–96 1986–96
Movers down the hump
New Zealand −1.7 +2.1   +3.6 −1.6   −5.4 −8.8
Norway −0.4 +2.9   +1.8 −2.0   −3.1 −3.0
United Kingdom +1.1 −3.0   −2.9 +2.9   −4.0 −5.8
United States −0.1 −1.5   +0.4 +4.1   −2.7 −0.5
Average −0.3 +0.1   +0.7 +0.9   −3.8 −4.5
Movers up the hump
Finland +12.2 +10.4   −12.0 −11.4   +6.7 +8.1
France +3.4 +2.0   −1.8 0.0   +2.1 +1.3
Italy +2.9 +1.5   −4.4 −2.6   −0.6 −0.8
Portugal +2.7 −1.1   −4.6 +3.7   −7.6 −9.8
Sweden +8.2 +7.2   −11.1 −9.3   −1.4 +3.8
Average +5.9 +4.0   −6.8 −3.9   −0.3 +0.5
Stable
Austria +1.5 +1.7   +2.6 +3.9   +0.1 +1.9
Belgium +3.1 −1.4   −0.6 +2.0   +1.3 −0.6
Canada +1.6 +0.2   −2.4 −0.2   −1.6 −2.4
Germany(d) +4.2 +2.6   −0.8 −6.0   +3.0 +4.2
Japan +1.3 +0.6   +2.0 +4.2   −1.7 +0.1
Netherlands +0.1 −3.6   +4.7 +8.4(e)   −1.6 −1.6
Spain +6.0 +1.4   −2.8 +2.3   +3.3 −3.7
Switzerland +4.2 +4.0   n.a. n.a.   −0.4 +5.0
Average +2.8 +0.7   +0.4 +2.1   +0.3 +0.4
Indeterminate
(movers across the hump)
Australia +1.6 +0.6   −0.6 +3.1   −3.0 −5.9
Denmark −1.7 +1.0   −3.0 −3.8   −1.3 −0.6
All countries +2.6 +1.6   −1.8 −0.1   −0.9 −1.0
Notes and sources: See Table 9.

7. Conclusions

This paper has two objectives: first, to describe and analyse the processes of industrial relations change in Australia, the UK and New Zealand; and, secondly, to trace the impact of industrial relations change on labour market outcomes. There are many similarities, both historic and current, between the three English-speaking countries in terms of labour market institutions and traditions. All three countries had moderate levels of unionisation and laws which favoured collectivism of the employment relationship and accorded special privileges to trade unions as representatives of workers. The process of transformation of industrial relations institutions was initiated first in the UK, with the election of the Thatcher Conservative Government. Over the next decade and a half, there was a steady stream of statute-based and other changes affecting industrial relations. There was no single ‘big bang’, although the changes taken together involve an almost complete change to the regulation of the labour market. By making the recruitment and the retention of trade unions' members more difficult and expensive, and by raising the costs of certain types of industrial action, the new laws paved the way for a transformation of British industrial relations. The key features of this transformation have been: the decentralisation of pay determination; the growth of plant and individual bargaining; the decline in the number of union members and union density; and the decline in industrial action.

While the change process in New Zealand is often seen to fall into the ‘big bang’ category – with the enactment of the Employment Contracts Act in 1991 – in point of fact, there had been a series of changes made to industrial relations institutions in the years leading up to the ECA. These changes included the abolition of compulsory unionism; the establishment of certification elections to determine union representation (a form of contestable unionism); a distinction between interest and rights disputes; and a form of ‘opt out’ bargaining. The ECA certainly accelerated the process of change, with the award system being abolished and trade unions losing all legal privileges. (The term ‘trade union’ does not appear in the ECA.) In the absence of a presumption of collectivism, individual employment contracts have grown rapidly, with single-employer collective agreements common in some sectors. In this latter case, trade unions most commonly represent workers. In addition to the decentralisation of bargaining, the ECA has also been associated with a decline in the number of trade unions, a decline in the number of union members and a decline in union density, the latter trend not being apparent prior to the ECA. Industrial action increased somewhat immediately after the enactment of the ECA but has fallen to historically low levels since then.

The case of Australia is somewhat different. For most of the past fifteen years, the elected Federal Government was favourably disposed to the trade union movement and disinclined to introduce hostile legislation. Nonetheless, the process of change occurred, including a lengthy experiment with corporatism and incomes policy. A managed experiment of enterprise-based bargaining was undertaken in conjunction with limited implementation of national pay rises for the low paid. The end result has been a degree of decentralisation of pay determination, especially covering larger, unionised enterprises. By the same token, a significant proportion of the workforce remain covered mainly by the award system. More recent legislative changes have included provision for individual bargaining, an option which is being taken up in a small number of cases. However, the mechanics of the system of compulsory arbitration remain in place – in contrast to the situation in New Zealand.

Despite these differences between the UK (slow-track, but now relatively deregulated), New Zealand (faster-track, also relatively deregulated) and Australia (slow-track, partially deregulated), any differences in the labour market outcomes of the three countries defy definitive conclusions. It is certainly true, for instance, that unemployment in New Zealand fell significantly after the enactment of the ECA (and is still significantly lower than in Australia); however, the overall rate of unemployment is higher than had been the case for most of the 1980s. Similarly, the employment to population ratio in New Zealand recovered after 1991, but only to levels apparent in the mid 1980s. In the case of the UK, unemployment remained stubbornly high through most of the 1980s and has declined to relatively low levels only recently. Similarly, the employment to population ratio has only recovered to levels apparent in the late 1970s. While the labour market outcomes in Australia have been inferior to those in the UK and New Zealand in recent years, the overall differences between the three countries have not been dramatic.

Of course, there are a number of explanations for this rather meek conclusion, including the multi-factorial explanation of labour market outcomes and the unknown length of lags between institutional change and observed outcomes. What this paper does to add interest to the findings is to broaden the comparisons to include other OECD countries, and to consider a variant on the Calmfors-Driffill hypothesis. This variant involves consideration of countries' base positions on the hump (measured along a centralised–decentralised spectrum) and the direction of change along the hump. Considering those countries which have moved down the hump in the direction of decentralisation for the period 1986 to 1996, those countries moving away from the intermediate position in respect of bargaining position performed better, on average, on all indicators. While the ‘movers down the hump’ (UK, US and New Zealand) experienced little change in unemployment and employment rates, they experienced a noticeable subsidence of inflationary pressures, suggesting a reduction in their NAIRUs.

Footnotes

Much of this section is drawn from Hawke and Wooden (1997). [1]

Bodman (1998), for example, reports evidence based on aggregate time-series data which suggests that the union-amalgamation process contributed to the decline in overall trade union membership. Wooden (forthcoming), however, has analysed panel data on changes in union density at Australian workplaces between 1989/90 and 1995 and can find no evidence that amalgamations reduced union membership levels within workplaces. [2]

The minimum size requirement was subsequently deemed in breach of ILO Conventions and repealed in 1993. [3]

See Trade Union Statistics, Australia, ABS cat. no. 6323.0, 30 June 1996. [4]

Unlike the Australian and UK data, the New Zealand data are based on information provided by unions. The Australian experience suggests that such data tend to overstate union membership relative to data collected from household surveys. This will tend to offset some of the overestimation of union density that results from treating part-time employees as only 0.5 of a union member. [5]

Extending the data period back to the early 1970s does not alter this conclusion with respect to the comparison between Australia and the UK. [6]

The size and direction of the employment impacts of minimum wage increases in New Zealand are not uncontentious (cf. Chapple 1997; Maloney 1995). [7]

The OECD (1997) identified Spain as having moved to more decentralised arrangements during the 1980s. These changes, however, appear to have been concentrated in the first half of that decade. [8]

Note, however, that the marked improvement in the employment to population ratio in the Netherlands (one of the ‘stable’ economies) is almost certainly a function of policies which have fostered rapid growth in part-time employment levels. [9]

References

Anderson, G. (1996), ‘Collective Bargaining and the Law: New Zealand's Employment Contracts Act Five Years On’, Australian Journal of Labour Law, 9(2), pp. 103–134.

Australian Council of Trade Unions [ACTU] (1987), Future Strategies for the Trade Union Movement, Doc. D229/87, ACTU, Melbourne.

Baird, C. (1996), The Employment Contracts Act and Unjustifiable Dismissal: The Economics of an Unjust Employment Tax, New Zealand Business Roundtable and New Zealand Employers' Federation, Wellington.

Bean, C. (1994), ‘European Unemployment: A Retrospective’, European Economic Review, 38(3/4), pp. 523–534.

Beggs, J. and B. Chapman (1987), ‘An Empirical Analysis of Australian Strike Activity’, Economic Record, 63(180), pp. 46–60.

Berry, P. and G. Kitchener (1989), Can Unions Survive?, Building Workers' Industrial Union (ACT Branch), Dickson, ACT.

Blandy, R. and M. Baker (1987), Industry Assistance Reform and the Labour Market: the New Zealand Experience, National Institute of Labour Studies, Adelaide.

Bleaney, M. (1996), ‘Central Bank Independence, Wage Bargaining Structure, and Macroeconomic Performance in OECD Countries’, Oxford Economic Papers, 48(1), pp. 20–38.

Bodman, P. (1998), ‘Trade Union Amalgamations, Openness and the Decline in Australian Trade Union Membership’, Australian Bulletin of Labour, 24(1), pp. 18–45.

Boston, J. (1984), Incomes Policies in New Zealand, Victoria University Press, Wellington.

Boxall, P. (1990), ‘Towards the Wagner Framework: Change in New Zealand Industrial Relations’, Journal of Industrial Relations, 32(4), pp. 523–543.

Bray, M. and D. Neilson (1996), ‘Industrial Relations Reform and the Relative Autonomy of the State’, in F. Castles, R. Gerritsen and J. Vowles (eds), The Great Experiment: Labour Parties and Public Policy Transformation in Australia and New Zealand, Allen and Unwin, Sydney, pp. 68–87.

Brook, P. (1990), Freedom at Work: The Case for Reforming Labour Law in New Zealand, Oxford University Press, Auckland.

Brooks, B. (1995), ‘100 Years of Compulsory Conciliation and Arbitration in Australia and New Zealand’, International Journal of Comparative Labour Law and Industrial Relations, Spring, pp. 31–41.

Brown, W. and S. Wadhwani (1990), ‘The Economic Effects of Industrial Relations Legislation since 1979’, National Institute Economic Review, 131, pp. 57–70.

Buchanan, J., K. van Barneveld, T. O'Loughlin and B. Pragnell (1997), ‘Wages Policy and Wage Determination in 1996’, Journal of Industrial Relations, 39(1), pp. 96–119.

Business Council of Australia Industrial Relations Study Commission (1989), Enterprise-based Bargaining Units: A Better Way of Working, BCA, Melbourne.

Calmfors, L. and J. Driffill (1988), ‘Bargaining Structure, Corporatism and Macroeconomic Performance’, Economic Policy, 6, pp. 14–61.

Chapman, B. and F. Gruen (1990), ‘An Analysis of the Australian Consensual Incomes Policy: The Prices and Incomes Accord’, in C. de Neubourg (ed.), The Art of Full Employment, Elsevier/North Holland, Amsterdam, pp. 475–504.

Chapple, S. (1997), ‘Do Minimum Wages Have an Adverse Impact on Employment?’, Labour Market Bulletin, 2, pp. 25–50.

Committee of Review into Australian Industrial Relations Law and Systems (Chairman: K. Hancock) (1985), Australian Industrial Relations Law and Systems: Report of the Committee, (3 volumes), AGPS, Canberra.

Crawford, A., R. Harbridge and K. Hince (1997), ‘Unions and Union Membership in New Zealand: Annual Review for 1996’, New Zealand Journal of Industrial Relations, 22(2), pp. 209–216.

Cully, M. and S. Woodland (1997), ‘Trade Union Membership and Recognition’, Labour Market Trends, 105(6), pp. 231–239.

Dabscheck, B. (1994), ‘The Arbitration System since 1967’, in S. Bell and B. Head (eds), State, Economy and Public Policy in Australia, Oxford University Press, Melbourne, pp. 142–168.

Dawkins, P. (1997), The Economic Effects of Deregulation and Decentralisation of Wage Determination, paper presented to the Conference in Honour of Professor Emeritus Joe Isaac, ‘Economics and Industrial Relations: Reappraising the Relationship’, Australian National University, 4–5 December.

Department of Labour (1998), Contract: The Report on Current Industrial Relations in New Zealand, Industrial Relations Service, Wellington, 24.

Drago, R. and M. Wooden (1998), ‘The Changing Role of Trade Unions in Australian Industrial Relations’, The Transformation of Australian Industrial Relations Project, Discussion Paper Series No. 3, National Institute of Labour Studies, Adelaide, forthcoming.

Elgar, J. and R. Simpson (1992), ‘The Impact of the Law on Industrial Disputes in the 1980s’, Centre for Economic Performance Discussion Paper No. 104.

Epstein, R. (1997), ‘Employment Law: Courts and Contracts’, California Western International Law Journal, 28(1), pp. 27–44.

Freeman, R. (1988), ‘Labour Market Institutions and Economic Performance’, Economic Policy, 6, pp. 63–80.

Freeman, R. and J. Pelletier (1990), ‘The Impact of Industrial Relations Legislation on British Union Density’, British Journal of Industrial Relations, 28(2), pp. 141–164.

Frenkel, S. and M. Shaw (1989), ‘No Tears for the Second Tier’, Australian Bulletin of Labour, 15(3), pp. 90–111.

Gardner, M. (1995), ‘Labour Movements and Industrial Restructuring in Australia, New Zealand and the United States’, in K. Wever and L. Turner (eds), The Comparative Political Economy of Industrial Relations, Industrial Relations Research Association, Madison, Wisconsin, pp. 33–69.

Golden, M. (1993), ‘The Dynamics of Trade Unionism and National Economic Performance’, American Political Science Review, 87(2), pp. 439–454.

Gregory, M. (1997), ‘Labour Market Deregulation: The UK Experience’, Australian National University Centre for Economic Policy Research Discussion Paper No. 376.

Griffin, G. and S. Svensen (1996), ‘The Decline of Australian Union Density – A Survey of the Literature’, Journal of Industrial Relations, 38(4), pp. 505–547.

Harbridge, R. and J. Moulder (1992), ‘Collective Bargaining and the Employment Contracts Act 1991: One Year On’, Industrial Relations Centre, Victoria University of Wellington, May.

Hawke, A. and M. Wooden (1997), ‘The Changing Face of Australian Industrial Relations’, The Transformation of Australian Industrial Relations Project, Discussion Paper Series No. 1, National Institute of Labour Studies, Adelaide.

Hector, J. and M. Hobby (1998), ‘Labour Market Adjustment Under the Employment Contracts Act: 1996’, New Zealand Journal of Industrial Relations, 22(3)/23(1), pp. 311–327.

Heitger, B. (1987), ‘Corporatism, Technological Gaps and Growth in OECD Countries’, Weltwirtshaftliches Archiv, 123(3), pp. 463–473.

Hilmer, F., M. Angwin, J. Layt, G. Dudley, P. Barratt and P. McLaughlin (1993), Working Relations: A Fresh Start for Australian Enterprises, The Business Library, Melbourne.

Joint Governments (1997), ‘Safety Net Review – Wages’, Department of Workplace Relations and Small Business, Canberra.

Kasper, W. (1996), Free to Work: The Liberalisation of New Zealand's Labour Markets, Centre for Independent Studies Policy Monograph No. 32, Sydney.

Kerr, R. (1997), ‘The New Zealand Employment Contracts Act: Its Enactment, Performance, and Implications’, California Western International Law Journal, 28(1), pp. 89–102.

Layard, R., S. Nickell and R. Jackman (1991), Unemployment: Macroeconomic Performance and the Labour Market, Oxford University Press, Oxford.

Maloney, T. (1995), ‘Does the Adult Minimum Wage Affect Employment and Unemployment in New Zealand’, New Zealand Economic Papers, 29(1), pp. 1–19.

McGraw, P. and I. Palmer (1994), ‘Human Resource Managers and the Closed Shop: Individuals and Organizations in NSW’, Journal of Industrial Relations, 36(4), pp. 519–529.

Metcalf, D. (1994), ‘Deregulation of the British Labour Market 1979–94’, paper prepared for the International Labour Organisation World Employment Report, Geneva.

Millward, N. and M. Stevens (1986), British Workplace Industrial Relations 1980–1984, Gower, Aldershot.

Millward, N., M. Stevens, D. Smart and W. Hawes (1992), Workplace Industrial Relations in Transition, Gower, Aldershot.

Morris, A. and K. Wilson (1995), ‘Corporatism and Australian Strike Activity’, Australian Bulletin of Labour, 21(2), pp. 153–173.

Mulvey, C. (1997), ‘Industrial Relations’, in Changing Labour Markets: Prospects for Productivity Growth – Workshop Proceedings, Industry Commission, Melbourne.

OECD (1991), Employment Outlook, July, OECD, Paris.

OECD (1994), Employment Outlook, July, OECD, Paris.

OECD (1997), Employment Outlook, July, OECD, Paris.

Peetz, D. (1990), ‘Declining Union Density’, Journal of Industrial Relations, 32(4), pp. 197–223.

Peetz, D. (1996), ‘The Accord and the Paradigm Shift in Union Membership’, paper presented at the End of the Accord Conference, Victoria University of Technology, Melbourne, 26–27 November.

Peetz, D. (1997), ‘De-unionisation and Unionisation: The Impact of Workplace Change, HRM Strategies and Workplace Unionism’, Labour and Industry, 8(1), pp. 21–36.

Rasmussen, E., P. Boxall, N. Haworth, S. Hughes, M. Powell, T. Maloney and M. Wilson (1995), ‘Industrial Relations and Labour Market Reforms in New Zealand’, paper prepared for PECC-HRD Task Force Meeting, Taipei, June.

Rasmussen, E. and J. Deeks (1997), ‘Contested Outcomes: Assessing the Impacts of the Employment Contracts Act’, California Western International Law Journal, 28(1), pp. 275–296.

Rawson, D. (1992), ‘Has Unionism a Future?’, in M. Crosby and M. Easson (eds), What Should Unions Do?, Pluto Press, Leichhardt, pp. 2–15.

Reilly, S. (1989), ‘An Analysis of the Factors Influencing the Second Tier and its Evolution’, Australian Bulletin of Labour, 15(2), pp. 200–222.

Robertson, B. (1996), The Status and Jurisdiction of the New Zealand Employment Court, New Zealand Business Roundtable and New Zealand Employers' Federation, Wellington.

Rowthorn, R. (1992), ‘Corporatism and Labour Market Performance’, in J. Pekkarinen, M. Pohjola and R. Rowthorn (eds), Social Corporatism: A Superior Economic System?, Clarendon Press, Oxford, pp. 82–131.

Sako, M. (1997), ‘Wage Bargaining in Japan: Why Employers and Unions Value Industry-level Co-ordination’, Centre for Economic Performance Discussion Paper No. 334.

Scarpetta, S. (1996), ‘Assessing the Role of Labour Market Policies and Institutional Settings on Unemployment: A Cross-country Study’, OECD Economic Studies, 26, pp. 43–98.

Sloan, J. and M. Wooden (1990), ‘The Structural Efficiency Principle in Action – Management Views’, Australian Bulletin of Labour, 16(3), pp. 199–223.

Soskice, D. (1990), ‘Wage Determination: The Changing Role of Institutions in Advanced Industrialized Countries’, Oxford Review of Economic Policy, 6(4), pp. 36–61.

Still, L. and D. Mortimer (1993), ‘The Effectiveness of Award Restructuring and the Training Levy in Providing a More Educated Workforce: A Comparative Study’, International Journal of Employment Studies, 1(1), pp. 77–102.

van Bastelaer, A., G. Lemaitre and P. Marianna (1997), ‘The Definition of Part-time Work for the Purpose of International Comparisons’, Labour Market and Social Policy Occasional Papers No. 22, OECD, Paris.

Waddington, J. (1992), ‘Trade Union Membership in Britain, 1980–1987: Unemployment and Restructuring’, British Journal of Industrial Relations, 30(2), pp. 287–324.

Weeks, P. (1995), Trade Union Security Law, The Federation Press, Annandale, NSW.

Wooden, M. (forthcoming), ‘Union Amalgamations and the Decline in Union Density’, Journal of Industrial Relations.