The effects of British industrial relations legislation 1979-97.

Brown, William ; Deakin, Simon ; Ryan, Paul 等

The Review is pleased to give hospitality to CLARE Group articles, but is not necessarily in agreement with the views expressed; responsibility for these rests with the authors. Members of the CLARE Group are M.J. Artis, A.J.C. Britton, W.A. Brown, W.J. Carlin, J.S. Flemming, C.A.E. Goodhart, J.A. Kay, R.C.O. Matthews, D.K. Miles, M.H. Miller, P.M. Oppenheimer, M.V. Posner, W.B. Reddaway, J.R. Sargent, M.FG. Scott, Z.A. Silberston, S. Wadhwani and M. Weale.

The recent change of government brings to an end a sustained attempt to transform British industrial relations by legislative action. This article explores the consequences. It explains the cumulative effect of the legal changes since 1979, including the growing influence of the European Community, and examines the economic and social results. While legal intervention has had an impact on the institutions of industrial relations, most notably in reducing the power of organised labour, this cannot be isolated from wider structural changes in labour and product markets. A review of research on economic outcomes suggests an uneven and tenuous link between institutional change and economic performance.

Introduction

The use of legislation to reform industrial relations has been a central component of economic policy during the period of Conservative government that has now come to an end. It is widely considered to have made a major contribution to improving the performance of the economy since 1980. In 1990 the CLARE Group published a paper in this Review (Brown and Wadhwani, 1990) attempting to assess the nature and extent of this contribution. The purpose of the present paper is to return to this task after more time has elapsed, and more legislation has been passed. There is now an abundance of research evidence dealing with different aspects of these changes, including attempts to address the same topic, either in part, by Elgar and Simpson (1993) and by Addison and Siebert (1997), or as a whole, by Dunn and Metcalf (1996). As well as this research, we draw on recent interviews of our own with employers over a wide range of industry.

Throughout the period an explicit link between legal change and economic performance was made by government itself. But there were also broader economic changes which were particularly unfavourable to organised labour. Labour markets slackened and product markets tightened, both very substantially. Official measures of unemployment, which had averaged below 4 per cent of the workforce over the fifteen years before 1980, rose to over 9 per cent during the subsequent fifteen years. There was a massive increase in the exposure of the workforce to product market competition. In manufacturing, imports increased as a proportion of home demand from 26 per cent in 1980 to 45 per cent in 1995. The highly unionised public sector was greatly reduced in employment and subjected to a variety of market mechanisms.

The bargaining system continued to fragment. In the private sector, enterprise-based bargaining at plant or company level increasingly eclipsed the once dominant multi-employer, industry level agreements, the coverage of which fell from perhaps 60 per cent of the workforce in 1950, to 30 per cent by 1980, and 10 per cent in 1990. The principal explanation of the trend, evident in most industrialised countries, is the internationalisation of product markets, which diminishes the advantages of national industry-based collective bargaining (Brown and Walsh, 1994). As with the substantial decentralisation of the public sector after 1980, this fragmentation greatly reduced the industrial basis of trade union power, tying the interests of union organisations increasingly to those of individual employers. These changed economic and institutional circumstances had consequences for trade unions which were in many ways similar to those intended by the legislation, creating a further difficulty for its evaluation.

The experience of other industrialised countries, in most of which there were comparable changes in economic conditions, is of relevance and will be drawn upon later. Most of these countries adapted to a changing economic environment without undertaking similar adjustments to their systems of industrial relations to Britain (OECD, 1994). In some, greater labour flexibility was achieved and productivity enhanced through adjustments to legislation which left collective institutions in the field of employment largely intact (Deakin, 1991). A comparative perspective suggests that the reforms adopted in Britain represent just one possible path among many. However, we should not conclude from this that the legislative changes made in Britain were either irrelevant or unnecessary to the wider process of economic reform undertaken by the Conservative governments. The question which we have to address is, given the broad thrust of economic policy after 1979, did the changes made to industrial relations legislation have the economic impact which was envisaged for them?

2. Labour law and the industrial relations system before 1979

In order to assess the impact of changes in legislation in the period after 1979, it is necessary to consider the role previously played by the state within the industrial relations system. The essence of the system which Kahn-Freund (1959) described in terms of `collective laissez-faire' was that the state supported the principle of collective self-organisation, without prescribing the form, level or structure of collective bargaining. The statutory `immunities', which protected the right to organise and the right to strike, insulated trade unions and their members and officials from liabilities which would otherwise have arisen at common law. The common law was (and is) hostile to collective organisation. Trade unions, in so far as they seek to regulate wages and conditions and to control the supply of labour, are potentially acting in restraint of trade. As a result, at common law they cannot effectively hold property or enforce their own rule books. Industrial action almost inevitably involves the commission of one or more legal wrongs (torts) which arise where economic pressure is exercised collectively or with the effect that certain property or contract rights are infringed. These rights could be those of employers, non-members or non-strikers, or third parties such as consumers of services.

In the Conspiracy and Protection of Property Act 1875 and again in the Trade Disputes Act 1906, Parliament intervened to protect unions from such liabilities. The restraint of trade doctrine was disarmed, unions were exempted from all liability in tort, and individual organisers of strike activity were protected against certain torts where their action was taken `in contemplation or furtherance of a trade dispute'. This formula, in turn, was given a broad meaning which embraced secondary or sympathy action and action in support of the closed shop.

Within this framework, collective bargaining developed to the point where, by the 1950s, basic wage rates were set at industry level in most sectors. Court-based sanctions played almost no role in the process of resolving industrial disputes, nor in the enforcement of collective agreements. Disputes between unions over spheres of influence were settled under the auspices of the TUC, and not in the courts. Although the courts were neutralised, other organs of the state supported collective bargaining in various ways. From the mid-1890s on, conciliation and arbitration were used to encourage employers to settle disputes on the basis of agreements for the recognition of trade unions. The principles of freedom of association and the right to organise were adopted in the public sector after 1918. Both central and local government pursued fair wages policies which had the effect of requiring private contractors to recognise trade unions or to follow the going rate. Regulatory legislation was introduced to deal with areas of workplace relations where collective bargaining was limited in scope (in particular, health and safety) or in low-paying sectors where voluntary collective bargaining had failed to take root (as in the case of the industries covered by the trade boards and, later, wages councils, which in the late 1940s amounted to 25 per cent of the employed labour force).

Various attempts were made to modernise the system of collective laissez-faire from the early 1960s onwards including the ill-fated Industrial Relations Act of 1971, which offered unions a greater range of positive legal rights (including a right to recognition) in return for stricter legislative controls over union organisation, in particular in relation to industrial action. The unions rejected the Act and it was never effectively implemented before its repeal in 1974. Incomes policies during this period failed in part because efforts to impose greater control over wage increases from the centre were undermined by the increasing decentralisation of bargaining to company and establishment level which many unions favoured and which the reforms of the Donovan Report (1968) also encouraged. Individual employment protection legislation was introduced in the form of rights to a written statement of terms and conditions, the payment of redundancy compensation, protection against unfair dismissal, and rights to compensation for interruption to wages. Anti-discrimination laws also came in at this time. However, basic terms and conditions--in particular wages and working hours--remained essentially the preserve of collective bargaining. A policy of abolishing wages councils was pursued from the 1960s onwards, in the expectation (largely disappointed) that voluntary collective bargaining would emerge in sectors where legal regulation was removed.

The result was that by the end of the 1970s, British labour law lacked certain features which had been widely adopted elsewhere in Europe, in particular universal minimum standards on wages and hours. The state's support for the institutions of voluntary collective bargaining had produced a system in which trade union membership was high by international standards, and elements of union monopoly (the closed shop in its various forms) were well entrenched. But in many respects the system was more fragile than it appeared. Outside the core of union strength in manufacturing and the public sector, organ]sat]on was sporadic and the coverage of collective bargaining was patchy. Despite their involvement in the corporatist experiments of the mid-1970s, British trade unions had a more limited voice in economic decisionmaking at all levels than some of their continental counterparts. With the failure of the Bullock Committee (1977) there was no equivalent to the various forms of codetermination and company- and workplace-based employee representation which existed alongside union organ]sat]on in some mainland systems.

3. Implementing neoliberal policies

The British system was subject to substantial criticism by neoliberal theorists. Union organ]sat]on in general and the closed shop in particular were seen as stifling managerial initiative and entrenching `restrictive practices'. Hayek (1973) identified the Trade Disputes Act 1906 as conferring upon the unions `unique privileges' which protected them against the normal consequences of civil liability. He argued for a return to the common law of tort and contract which the statutory immunities had displaced. He also rejected most forms of regulatory legislation aimed at protecting the employment relationship, a criticism that had much less resonance for Britain in 1979 than it would have had for other European systems.

However, Hayek's critique by no means implied a reduced role for the law in relation to the labour market. A straightforward return to private law, through deregulation, proved to be insufficient. Although the trade dispute immunities were whittled away by successive legal changes, in other respects the body of labour legislation grew enormously to cope with the task of controlling collective labour organ]sat]on. The state, and legislation in particular, remained a powerful influence within the labour market.

What is of immediate interest here are the regulatory techniques by which the goal of restoring a functioning labour market was to be achieved. Three may be identified: the restoration of private law; the use of competition policy and informal government pressure to restructure systems of pay determination; and the restrictive re-regulation of collective action.

Return to private law

The Employment Act 1982 made it possible for the first time since 1906 for the trade union itself to be sued in tort. In what was in effect a partial restoration of the Taft Vale judgment of 1901, unions were made responsible for the acts of certain officials and committees in organising industrial action. The statutory change revived legal possibilities which, in the context of industrial relations, had lain dormant for decades. From 1982 on, unions engaging in tortious conduct faced the prospect of injunctions prohibiting industrial action. Failure to obey would result in fines for contempt being levied by the court and, if necessary, collected through sequestration, involving the seizure of union assets. This duly occurred in a number of major disputes in the mid- to late- 1980s. In this way private law rights were supplemented by the quasi-criminal procedures of contempt of court, which were applied and extended by the courts in innovative ways.

The substantive scope of the immunities was gradually reduced, so that by the end of the 1980s it was unlawful for industrial action to involve secondary action, secondary picketing, action in defence of the closed shop, and action in support of union recognition by third party employers. The legal concept of the `trade dispute' was also narrowed in such a way as to reinforce the tendency to confine lawful industrial action to individual employment units. Picketing was subject to new controls in the form of a limit on numbers and more effective exercise of police powers. These changes to the law were used by employers to tilt the balance of power in their favour during strikes. Employers could effectively redraw the limits of lawful industrial action through changes to corporate form, as appears to have happened during the Wapping dispute (Ewing and Napier, 1986).

Competition policy and pay bargaining

The second strand represents the growing encroachment upon the industrial relations system of competition policy. In the first wave of 1980s legislation, `props' to collective bargaining were removed. These included the complex recognition procedure which had operated through ACAS in the 1970s, and both the Fair Wages Resolution and the `Schedule 11' procedure for extending the effect of industry-level collective agreements to non-unionised firms and establishments. The powers of the remaining wages councils were severely reduced (in 1986) before they were removed completely in 1993, leaving only the Agricultural Wages Board as the last minimum wage setting body. Gradually, the focus of the law shifted from deregulation to active intervention, aimed at limiting the scope of collective regulation of wages and conditions. Compulsory competitive tendering to local authorities, and a prohibition from imposing minimum labour standards, were introduced in 1988. The abolition of the National Dock Labour Scheme in 1989 removed the basis of operation of the pre-entry closed shop in port transport.

Less formal government pressure was used to encourage changes in the system of pay determination. The public sector received official encouragement to decentralise bargaining, considerably constrained by Treasury concerns about loss of public expenditure control. Since industry level agreements in private industry were almost all based on voluntary arrangements, it was not necessary to introduce legislation in order to encourage continuing change there. Changes to industrial relations law nevertheless facilitated the break up of industrial agreements, by making it more difficult for trade unions to mount nationwide disputes.

Re-regulation of collective action

New forms of regulation sprang up to restrict the scope for collective action by workers. These included the imposition upon trade unions of complex procedural requirements, relating to the conduct of industrial action, the recruitment and organisation of members, and internal union governance. In relation to industrial action, the essential point is that the regulations dovetailed with the newly effective private law liabilities. Once the principle of the trade union's liability in tort was re-established, it provided the route to imposing upon the union the obligation to conduct a ballot in respect of industrial action. If the union failed to conduct an appropriate ballot, with a simple majority voting in support of the action, it would forfeit any further immunity it would otherwise have had. Following the initial 1984 Act, the procedural constraints on unions were gradually tightened. From 1988, it became necessary in certain cases to conduct separate ballots for each workplace and to obtain a separate majority in each one. After 1990 it was no longer possible for the union to organise a ballot in order to legitimise industrial action which had already begun. The ballot now had to precede the industrial action. This was the precursor to the introduction (in 1993) of a requirement that the union give seven days' notice of the commencement of industrial action. Notice of the ballot itself was made mandatory. The union was also required to provide the employer with a list of its members entitled to take part in the ballot.

These and related procedural restrictions are also closely linked to changes which were made to the law governing the dismissal of employees taking part in industrial action. At common law, an employee taking part in such action almost invariably commits a breach of his or her contract of employment. Prior to 1971, when unfair dismissal law was introduced for the first time, there was no legal protection for strikers whatsoever; they were liable to instant dismissal. With the arrival of statutory controls over dismissal, it became necessary to specify what the rights of strikers were. The Labour government of 1974-79 provided employers with an immunity from liability in unfair dismissal only if all those taking part in a strike or other industrial action at the relevant time were dismissed. Either selective dismissal, or selective re-engagement, would result in the loss of this employer's immunity.

From the early 1980s onwards, the employer's immunity was widened, most notably by allowing selective reengagement of strikers after a certain period had elapsed. In 1990 it became possible for an employer to dismiss on a selective basis individuals taking part in industrial action which the union had not authorised. Faced with industrial action which started prior to a ballot, a union now had two options, each unpleasant. It could either endorse the action, in which case it would automatically face legal consequences, most likely an injunction and claims for damages in tort; or it could repudiate it, leaving its members vulnerable to selective dismissals. If the union organised a protest action against the dismissal it would face further legal liabilities.

Changes to unfair dismissal law provided the principal means by which the closed shop was made effectively inoperable, or, in some cases, driven underground. Under the law put in place by the 1974 Labour government, it was automatically unfair for an employer to dismiss an employee on the grounds of his or her trade union membership, and enhanced compensation was payable. However, it was automatically fair to dismiss on grounds of an employee's non-membership where a `union-membership agreement', or closed shop, was in force. In the 1980s, the scope for fair dismissal on grounds of non-membership was gradually reduced, until in 1988 a principle of parity between members and non-members was established. Dismissal on the grounds of either membership or nonmembership was automatically unfair.

In 1990 a new, statutory civil wrong was also created, to prevent employers refusing to hire workers on the grounds of membership or non-membership. This helped put an end to the pre-entry closed shop. However, it did not affect the blacklisting of individuals on the basis of their past trade union activities. The increasing role of the law may be contrasted to the position prior to the introduction of unfair dismissal law in the 1970s, when employers had, in legal terms, complete freedom to adopt whatever strategy they liked, but where, in practice, they were constrained by the social force of union organisation to adopt closed shops. Further changes to the law undermined the principle of parity between membership and non-membership, and employers exploited the growing weakness of unions on the ground to pursue strategies aimed at discouraging union membership. For a while, judicial decisions kept in check attempts by employers to provide differential benefits to members in particular unions, or to non-unionists. Such practices infringed the principle that no `action short of dismissal' should be taken to penalise either membership or non-membership. However, following the decision of the House of Lords in the Wilson and Palmer case (1995), it became possible for employers to introduce `individualised' or personal contracts of employment outside the scope of collective agreements, without falling foul of the law, a result which Parliament itself sanctioned by amending the relevant legislation while the case was going through the courts (Deakin and Morris, 1995).

The law relating to internal union governance saw some of the most pronounced effects of restrictive regulation. Statutory controls over the admission, discipline and expulsion of members were introduced. Unions were prevented from taking disciplinary action against strikebreakers. Periodic ballots became mandatory for senior union posts, and for the purposes of maintaining a political fund. The organisation of members and the collection of union dues were made procedurally more complex and expensive, by requiring workers to give written authorisation every three years for employers' deducting membership dues from wages.

4. Countervailing influences: European Community law

The tide of legislation did not just flow in one direction. Throughout the period from 1979, the influence of European Community law grew steadily (Deakin, 1997). EC labour law, while far from representing a comprehensive body of labour legislation, nevertheless draws on a continental tradition in which the rights of employees are laid down in constitutional texts and legislative codes. This contrasts with the British tradition of collective laissez faire. Unlike standards laid down in conventions of the ILO, those contained in the European Community Treaty, and in directives and regulations, may form part of the domestic legal order. A series of court rulings established that European Community law took priority over domestic law in the event of a conflict, and that certain rights--those having `direct effect'--could be directly enforced by individuals in the UK courts. Legislation concerning equal pay and equal treatment between men and women, on the one hand, and employment protection, on the other, slowly began to take effect in Britain. Equal pay for work of equal value was introduced in 1983, and the principle of equal treatment was later applied to retirement and pension rights.

An influential EC provision was the 1977 Acquired Rights Directive, which had been implemented by a Conservative government in 1981 in the form of the Transfer of Undertakings (Protection of Employment) Regulations. The European Court of Justice made it clear in a series of judgements that the Directive could be invoked to preserve the rights of employees affected by outsourcing or where there was a change in the identity of the employer following privatisation. It also held that existing UK law failed to make adequate provision for consultation with employee representatives in the event of a transfer or large scale redundancies. A number of EC measures in the field of health and safety, in particular Directives on Working Time and Young Workers, also necessitated changes in UK law during this period.

The government's willingness to use its veto in the Council of Ministers effectively blocked a number of new initiatives and, to this extent, it was able to limit the influence of European level norms. This policy reached its logical conclusion in the opt-out from the so-called Social Chapter in the Maastricht Treaty. However, the UK's self-exclusion enabled the other member states to make progress on measures such as the European Works Councils Directive. All eligible UK-based multinationals appear to have extended EWC arrangements to their British operations, even though not legally obliged to do so. EWCs are so far in place in 53 UK based and 151 overseas based companies operating in Britain, and a further 127 UK-based companies will become subject to the Directive when Britain's opt-out is reversed (EWC Bulletin). The Maastricht Treaty also initiated `social dialogue' or discussion between social partners at transnational level, encompassing the British CBI and TUC, with agreements on parental leave and on the rights of part-time workers being a tangible result. It is a process with the potential to transform still further the relationship between European-level labour standards and the laws of individual member states.

5. Outcomes--the coverage of trade unions and of bargaining

As we have seen, the intention of the legislative programme was to reduce in various ways the impact of trade unions and the coverage of collective bargaining. Table 1 provides data on membership collected from unions, as well as that more recently available from individuals; their absolute figures differ but they point in the same direction. The steady growth in trade union membership of the 1970s turned to a sharp fall in the early 1980s which slowed in the late 1980s before turning to a sharper decline in the first half of the 1990s.

Table 1. Trade union membership in the UK, 1971-96

 Certification Officer Data

 Members (000s) % of employed

5-year annual average
1971-75 11,548 (+1.5) 50.0
1976-80 12,916 (+1.5) 55.1
1981-85 11,350 (-3.5) 53.2
1986-90 10,299 (-1.7) 46.1
1991-95 8,740 (-4.0) 40.2

annual
1989 10,158 (-2.1) 44.2
1990 9,947 (-2.1) 43.9
1991 9.585 (-3.6) 43.8
1992 9,048 (-5.6) 42.3
1993 8,700 (-3.8) 40.2
1994 8,278 (-4.9) 38.1
1995 8,089 (-2.3) 36.7
1996 n.a. n.a.

 Labour Force Survey Data

 Members(000s) % of employed

5-year annual average
1971-75 n.a. n.a.
1976-80 n.a. n.a.
1981-85 n.a. n.a.
1986-90 n.a. n.a.
1991-95 7,854 34.8

annual
1989 8,964 39.0
1990 8,854 (-1.2) 38.1
1991 8,633 (-2.5) 37.5
1992 7,999 (-7.3) 35.8
1993 7,808 (-2.4) 35.1
1994 7,553 (-3.30 33.6
1995 7,275 (-3.7) 32.1
1996 7,215 (-0.8) 31.3

Sources: Employment Gazette and Labour Market Trends (various). Figures in brackets indicate percentage change in membership as an annual rate.

It is helpful to place Britain's unionisation experience in an international perspective (Table 2). Many of the economic and structural factors that influence unionisation are common to most industrialised countries. There are considerable difficulties in interpreting absolute differences between countries, but the comparison of their experiences over time is useful. Many have experienced substantial declines in union membership during the 1980s and 1990s. Indeed, the UK's `league position' in this list barely changed, at about the midway point, between 1970 and 1995. Bearing in mind that several of the countries saw declining membership while governments were in power which were sympathetic to trade unions--France, Spain and Australia, for example--it is not immediately evident that Britain's legislative experience had a distinctive impact.

Table 2. Trade union membership by country as a percentage of all employees, 1970-95

Country 1970 1980 1990 1995

United Kingdom 45[9th] 50[10th] 39[10th] 32[11th]
Australia 50 48 40 33
Austria 62 56 46 43
Belgium 45 56 51 53
Canada 31 36 36 34
Denmark 60 76 71 82
Finland 51 70 72 81
France 22 17 10 9
Germany (West) 33 36 33 29(*)
Greece 36 37 34 n.a.
Ireland 53 57 50 38
Italy 36 49 39 38
Japan 35 31 25 24
Netherlands 38 35 25 26
New Zealand 41 56 45 22
Norway 51 57 56 56
Portugal 61 61 32 32
Spain 27 25 11 15
Sweden 68 80 82 83
Switzerland 30 31 27 26
United States 23 22 16 15

Trade union membership as a per cent of wage- and salary-earners. (*) 1993 Source for 1970-90: OECD Employment Outlook, July 1994; cf for details. Sources for 1995: various including ILO unpublished; Ebbinghaus and Visser (forthcoming); Ferner and Hyman (forthcoming); miscellaneous. UK rank order of highness of density.

The decline in membership in Britain has been traced to various factors. Although Freeman and Pelletier (1990) tried to link it directly to the early legal changes, other research suggests that more dominant influences have included the changing composition of employment, and limited bargaining opportunities at new sites (Geroski et al, 1995; Disney et al, 1996). These studies support the view implied by the international comparison that British legislative change has not exerted a major influence on union membership.

Coverage of collective bargaining

One of the startling findings of the 1990 Workplace Industrial Relations Survey (WIRS) was the extent to which the coverage of collective bargaining had contracted after the 1984 WIRS. The proportion of employees in workplaces of 25 or more covered by collective bargaining fell from 71 per cent to 54 per cent over the six year period (Millward et al, 1992). Dunn and Metcalf (1996) concluded that much of this arose from broader changes in the way employment is managed, most notably the shift from industry level to enterprise-based bargaining, rather than from legislative change.

There is, however, a deeper point, best made by international comparison. Table 3 shows the proportion of the employed workforce covered by collective bargaining or some form of enforceable industrial wage arrangement in ten OECD countries between 1980 and 1994. The British figures up to 1990 overstate as they combine the coverage of collective agreements and wages councils (abolished in 1993), ignoring a degree of overlap (Milner, 1995). The British figure for 1994 is the proportion of employees in workplaces with some union recognition and also overstates somewhat (Cully and Woodland, 1997). Several of these countries have experienced substantial declines in unionisation over the period. But none experienced so dramatic a fall in coverage of agreements as Britain. This is partly because, for several European countries, the law provides what are called `extension rules' whereby the terms and conditions of a collective agreement can be made legally binding over those non-unionised employees who are within the appropriate industry or bargaining unit (OECD, 1994). Thus, although the decline of trade union membership in Britain may not have been exceptional in international terms, the implications of it are. When taken with the abolition of statutory wage minima, it will be evident that Britain has, in a relatively short time, become exceptional within Europe in its low coverage of collective arrangements (Gregory and Sandoval, 1994).

Table 3. The proportion of the workforce covered by collective bargaining and statutory sectoral wage arrangements by country 1980-94

Country 1980 1985 1990 1994

United Kingdom 83 76 65 48
Australia 88 85 80 80
Canada n.a. 37 38 36
Finland 95 n.a. 95 95(**)
France 85 92 95 95(*)
Germany 91 91 90 92(**)
Japan 28 n.a. 23 22
Netherlands 76 76 71 81(**)
Spain n.a. 67 68 66(**)
United States 26 20 18 18

Source: UK: maxima derived from Milner (1995) and Cully and Woodland (1997). Non UK: For 1980, '85, '90: OECD Employment Outlook, July 1994. For 1994, OECD unpublished from Blanchflower (1996) (*) 1992; (**) 1993.

Closed shop

We can sharpen our focus by looking at two sources of trade union security which have been the subject of direct legislative action: the closed shop and union recognition by employers. The WIRS evidence is that the coverage of closed shops declined from over five million union members in 1980 to under half a million in 1990 (Millward et al, 1992). The pre-entry closed shop had previously been in retreat as its craft-based and other bastions shrank, but several industries with substantial pre-entry closed shops (notably in printing, newspapers, television, shipping, and docks) lost them in the 1980s as a direct result of disputes in which the outcome was critically dependent upon recent legal changes. The relevant legislation here was not the banning of the pre-entry closed shop per se, which did not become effective until 1991, but rather the earlier prohibition of the secondary action and mass picketing upon which it had depended for viability at industry level (Elgar and Simpson, 1993).

So far as the more common post-entry closed shop was concerned, however, action in its defence did not become illegal until the 1988 Act, and many employers who had routinely treated it as a source of procedural discipline continued subsequently to support it informally (Wright, 1996). Union membership was not substantially greater in 1990 where a closed shop survived than where the union had no more than strong endorsement by management, leading to the conclusion that `the decline of the closed shop had only a limited part to play in the overall drop in union membership' (Millward et al, 1992:101).

Recognition

Continuing decline in union membership in the 1990s suggests there may have been a change in employer policy toward recognition. The evidence of the 1980s was that union derecognition was unusual. The WIRS panel data suggested that withdrawal of recognition reflected employee apathy rather than employer initiative (Millward et al, 1992; Beaumont and Harris, 1995). But there was evidence that unions were being shunned in new premises, and only a minority of companies which already recognised unions automatically granted recognition at new sites (Marginson et al, 1996)

More recently there has been evidence of a shift towards more strategic derecognition. A survey of unions suggested that a step increase in derecognition occurred during 1987 (Claydon, 1996). Certainly there is growing anecdotal evidence that strategic derecognition has, during the 1990s, become a feature of many industries where once trade unions were strong, notably the newly privatised utilities, printing, publishing, television, docks, and chemicals. How far has this development been dependent upon legal change? Many employers who have derecognised unions have told us that they considered that the legal changes had been important. Rather than specify particular acts, they have generally said that it was the whole range of post-1980 legislation which had contributed to a change in attitude and made possible the new strategies of which &recognition was a part.

6. Trade union governance and finances

One of the stated intentions of the legislation was to make trade unions more accountable to their members. Research suggests that it strengthened the national leadership of unions (Undy et al, 1996; Heery, 1996). This was partly because of the necessity of controlling the conduct of strike ballots in order to avoid expensive procedural errors. In addition, the requirement to maintain a record of members' names and addresses and to publish financial details has stimulated improvements in their information, accounting and communication systems. Unions generally complied with the laws affecting their own governance. This had no consistent impact on their policies; it did not, for example, promote the emergence of more `moderate' leadership. The outcome of ballots depends in part upon the presence of factions which can organise the vote. More formal processes such as postal ballots do not generally increase turnouts, and increased turnouts do not generally lead to more moderate outcomes.

How far has the legislative change affected the financial resources of trade unions? Many unions suffered financially during the 1980s when they incurred the penalties for newly illegal actions. Such expensive conflicts have been unusual in the 1990s. The use of injunctions by employers involved in industrial disputes, which had looked like becoming an expensive field of litigation in the 1980s, has diminished in the 1990s with the number of applications made generally falling to single figures per year (Gall and McKay, 1996).

A potentially costly legal innovation encouraged individuals to take their own action against trade unions. The Commissioner for the Rights of Trade Union Members was established in 1988 with considerable investigative powers as a self-styled `union ombudsman', and with the duty to provide a form of `legal aid' for union members wishing to pursue a grievance. However, the fact that her agency has paid legal costs averaging only 125,000 [pounds sterling] per year for the five years since 1991 suggests that it did not generate substantial extra burdens for trade unions. In 1993 she was given additional duties as Commissioner for Protection Against Unlawful Industrial Action to help members of the public believing themselves to be financially damaged by unlawful strikes, but there has been minimal use of this facility (CRTUM/CPAUIA, 1996). Unions appear to be coping with these legal and financial minefields.

Apart from the penalties for newly unlawful actions, the legislation has given trade unions increased difficulties in raising funds and also increased running costs. The 1984 requirement to have a ballot authorising a political fund every ten years triggered a second round of ballots between 1993 and 1996, and this time they had to be postal. However, although this new requirement reduced the turnout substantially, the outcome was a very similarly supportive 4:1 vote in favour as had been delivered in the first 1985/86 round, and with a slightly increased number of unions participating (Leopold, 1997).

The requirement for postal ballots has brought additional costs to unions. From 1980 until 1996 they could reclaim some of these from the Government, and over that period a total of 24 million [pounds sterling] was paid back. By the mid 1990s, when the scheme was being widely used, the costs implied by the compensation paid averaged over 3 million [pounds sterling] per year, approximately one third of which was for strike ballots (Certification Office, unpublished data). This amounts to approximately half a per cent of trade unions' total annual expenditure. Although this cost, which unions since 1996 have had to bear themselves, may not seem particularly high, it is additional to the procedurally demanding balloting rules. It is significant that well over half of all injunctions sought by employers relate to alleged balloting infringements.

These increased costs should be put in the perspective of the broader financial circumstances of trade unions (Table 4). Although these aggregate data conceal variations between unions, it is clear that their overall finances have deteriorated since the 1950s and early 1960s, when unions had substantial reserves and adequate income, not least from membership subscriptions. They apparently failed both to invest in inflation-proof assets and to keep up the real value of subscriptions. Most unions have thus been left with few reserves for bargaining purposes. Undiminished competition for membership has prevented effective increases in subscriptions. Unions have improved and centralised their controls, but their financial performance continues to deteriorate (Willman and Morris, 1995). This deterioration long pre-dates both membership decline and the legislative change.

Table 4. Trade union financial ratios

Date (assets-liab) subs income
 total expenditure total expenditure

1951-55 4.13 1.04
1956-60 3.54 0.97
1961-65 3.60 1.03
1966-70 2.92 0.87
1971-75 n.a. n.a.
1976-80 1.47 0.96
1981-85 1.29 0.90
1986-90 1.25 0.85
1991-95 1.09 0.84

Date total income
 total expenditure

1951-55 1.18
1956-60 1.11
1961-65 1.19
1966-70 1.08
1971-75 n.a.
1976-80 1.14
1981-85 1.12
1986-90 1.05
1991-95 1.03

Source: Willman, Morris and Aston (1993), Certification Officer Annual Reports, (various). Data for 1971-75 are unavailable because of non-registration under the Industrial Relations Act (1971).

An important factor lying behind the depletion of union finances was that, with the gradual emergence of enterprise-based bargaining in the 1970s and 1980s, unions became increasingly reliant upon employers for resources. It became accepted in many industries that employers provided shop stewards with paid time for their union duties, deducted union dues for them, and gave them access to office facilities. By 1990, in the diminishing number of establishments where workplace representation continued to be recognised, such support still continued (Millward et al, 1992). This increased dependence of workplace union organisations upon resources under management control has made them potentially more malleable to the employer's will and increasingly vulnerable to the withdrawal of bargaining rights.

Thus, while legal change may have restored some authority to union leaderships, it is on the basis of considerably reduced resources. Furthermore, for reasons largely unconnected with legislation, the decentralised organisations into which their unions are increasingly, in effect, fragmenting have become heavily dependent upon the employers with whom they negotiate, and consequently more compliant in accommodating to change and accepting job losses.

7. Outcomes--industrial disputes

The steady decline in strikes of the 1980s continued into the 1990s (Table 5) as has the tendency towards smaller and shorter disputes. The decline in collective action has not, however, been matched by individual disputes. The early 1990s saw a sharp rise in cases going to ACAS conciliation. This did not reflect increasing claims against unfair dismissal, which have remained surprisingly constant at around 40,000 per year for the past twenty years, but against discrimination, breach of contract (by either side) and incorrect payment of wages. The rise suggests that the steadily widening jurisdiction of ACAS and the Industrial Tribunals is substituting to some extent for needs previously met by trade unions, and also that unions are increasingly resorting to ACAS as a means of supporting their members.

Table 5. Strikes and individual grievance conciliation in Britain: average annual statistics, 1990-96

 Stoppages Working days Per cent Individual
 reported lost per of stoppages concilia-
 1000 lasting not tion cases
 employees more than received by
 3 days ACAS (000s)

1960-64 2512 139 74 n.a.
1965-69 2380 168 65 n.a.
1970-74 2917 629 49 n.a.
1975-79 2345 509 43 43.5(*)
1980-84 1363 484 54 44.8
1985-89 895 180 58 45.7
1990-94 334 37 65 67.8
1995-96 240 39 67 96.0

Source: Employment Gazette, Labour Market Trends (various), ACAS Annual Report (various); (*)data for 1976-79.

Survey evidence suggests that legal changes have shaped but not eliminated the particular forms of industrial action against which they were targeted. Legal requirements notwithstanding, secret ballots were reported not to have been held in over a third of instances of strike action and in the majority of instances of non-strike action during 1989-90. While political strikes had all but gone, the equally unlawful sympathy strike and secondary picketing were still in evidence in 1990. Mass picketing had largely gone, but picketing in numbers greater than the recommended `six per entrance' was still common, though less so than in 1980 (Millward et al, 1992).

It is helpful to discuss the implications of the legislation in a theoretical context. Hicks/Nash bargaining theory suggests that the outcome of negotiations depends on the relative vulnerability of the two parties during a dispute. The legislation increased strike costs to labour, both to the individual employee faced with greater threat of dismissal, and to the trade union faced with increased costs of procedural compliance and larger penalties for noncompliance. Theory predicts that higher strike costs to unions lead to lower pay settlements. The incentive to avoid strikes, viewed as bargaining mishaps, will increase when joint strike costs increase. Bargaining theory also implies that disputes arise from informational asymmetries, concerning either the costs of a dispute to either party or the amount of surplus available to be shared by the union.

It is often argued that the balloting legislation may have reduced the incidence of disputes by providing clearer evidence of membership willingness to strike. The high proportion of ballots for industrial action that are carried suggests that union leaders choose their balloting occasions carefully. The proportion of strike ballots notified to ACAS which went in favour of industrial action has remained high at 92 per cent in both 1987-89 and 1994-95. Moreover, industrial action did not follow in the majority of cases where a ballot went in favour. Of the more intractable cases of majority support for action on which ACAS was asked to conciliate over the period 1994-96, only 20 per cent were followed by a strike, and only a further 16 per cent by action short of a strike (ACAS, unpublished data).

It is, however, not clear how a majority vote will be interpreted, unless it is for rejection. There is anecdotal evidence that strike ballots are used as cheap substitutes for action, with an element of bluff. It costs a union member little to vote for industrial action, especially as the enhanced anonymity of the ballot makes it possible to vote for action without such strong consequential obligation to join it. The costs to the individual of non-compliance in collective action were further reduced by the removal in 1988 of any formal union powers to discipline non-strikers. The information provided by ballots is thus of limited value.

Union officers consider that ballots provide a valuable source of credibility, although later legislative changes have made them more legally hazardous (Elgar and Simpson, 1996). The study by Undy et al (1996) concluded that strike ballots did make a minor contribution to preventing strikes because they signalled resolve without recourse to action and they encouraged the option of action short of a strike. Another factor was the fear of legal action if ballots were not properly conducted, a point underlined by employers' relatively heavy use of injunctions over their conduct.

As was the case with union membership, this cautious view of the effect of the legislation upon British strikes is reinforced by international comparison (Table 6). There has been a world-wide decline in strike incidence, suggesting economic and political influences common to many advanced economies. The British decline has been greater than average, moving the country from a middling to a low rank, and suggesting a distinctive role for influences specific to Britain, of which one might be legislation. However, the implications of the change in rank are limited at these low incidence levels. Had Britain merely maintained its 1986-90 rank, its strike incidence would still have had to fall from 137 to below 55 working days lost per 1,000 employees as opposed to the 24 recorded. The implications for any distinctive effect of British legislation are accordingly muted.

Table 6. International comparison of working days lost per 1000 employees, 1981-95

Country 1981-85 1986-90 1991-95

United Kingdom 440[15th] 137[11th] 24[5th]
Australia 386 224 130
Austria 2 2 6
Belgium n.a. (48) 32
Canada 532 429 159
Denmark 306 41 45
Finland 326 410 218
France 78 111 94
Germany 52 5 17
Greece 516 6316 1148
Ireland 474 242 109
Italy 774 315 183
Japan 10 5 (3)
Netherlands 24 13 33
New Zealand 408 425 55
Norway 58 142 62
Portugal 176 82 34
Spain 584 602 469
Sweden 40 134 50
Switzerland n.s. n.s. 1
United States 128 82 42

Notes: Data from Employment Gazette, June 1989 and Labour Market Trends, April 1997. () Brackets indicate averages based on incomplete rank order of lowness of days lost excludes Belgium data. n.a. = not available, n.s. = less than five days lost per thousand. UK

8. Economic outcomes--pay

Other studies have attempted to trace economic outcomes of the legislation with regard to pay, productivity, employment and profitability. Here we focus on the first two, as being more immediately open to legislative influence. We also focus on the microeconomic evidence, because causal links are even harder to trace at aggregate level. Space constraints prevent our referring to more than a selection of the relevant recent research. So far as pay is concerned, increases in the costs of union organisation and of industrial action might be expected to lower the wage which unions achieve. This would be expected to lead, first, to a fall in the pay differential between employees in unions (or covered by agreements) and others and, second, to a decline in inflationary wage pressure. The evidence for these is discussed in turn.

Union pay differentials

Econometric research into the union differential in pay is impeded by the inadequacy of the controls for other relevant characteristics, whether of individuals or workplaces. It also neglects the common practice of nonunion employers in paying higher wages specifically to exclude unions. With these serious reservations, the research of the last decade has painted a reasonably coherent but modest picture. It suggests that trade unions typically achieve only moderately higher pay for their members relative to comparable non-members. In the early 19805 the union differential is estimated from establishment data to have been no more than 10 per cent for any category of manual employee, and insignificant for skilled workers (Stewart, 1987; 1990). The results from individual employee data were very similar (Blanchflower, 1996; Main, 1996).

The effect of unions on pay suggested by the evidence from establishment data is varied (Table 7). The low water mark is associated with highly competitive product markets, where the differential is estimated at zero. The high water mark is where there is a pre-entry closed shop, where the union pay differential is variously estimated during the early 19805 as between 7 and 23 per cent, according to skill level. Only a pre-entry closed shop was associated with a positive differential in competitive product markets (Stewart, 1990). It was also associated with a wage gain additional to that linked to high membership density at the workplace, leading to the conclusion that it was `the one institutional form of trade unionism which really bit in terms of pay' (Metcalf and Stewart, 1992: 514).

Table 7. Estimated trade union net pay mark-up over nonunion pay in the UK, by occupation, 1980, 1984 and 1990

Union Occupational 1980 1984 1990
category category

All(a) skilled .030 .034 .015
 semi-skilled .075(*) .100(*) .063(*)
 unskilled n.a. .102(*) .072(*)
Pre-entry closed skilled .073 .174(*) .074
shop only(b) semi-skilled .139(*) .226(*) .080
 unskilled n.a. .191(*) .118(*)

Source: Stewart (1987), Table 2, (1995), Tables 3,4. (*) indicates significance at 5 per cent level.

Note: estimates refer to mean difference in log of weekly earnings of typical employee in each category in workplaces where (a) unions are recognised for bargaining purposes for manual workers and (b) at least some employees are covered by a pre-entry closed shop, relative to workplaces without union recognition for manual workers, after controlling for available workplace attributes, including sector, ownership, and workforce composition. The percentage union pay premium is approximately 100 times the coefficient.

From that starting point, the overall union pay differential is estimated to have fallen moderately, after 1984 and through to 1990 at least; and the differential associated with the pre-entry closed shop, to have fallen substantially (Gregg and Machin, 1991; 1992). By 1990 the wage effect attributable to the pre-entry closed shop, which was about to become legally indefensible, had become indistinguishable from that attributable to the post-entry variant, and both from that attributable for the de facto closed shops by then allowed by law (Stewart, 1995). The heterogeneity of union wage effects therefore fell during the 1980s.

There are difficulties of disentangling the effects of legal changes from those of increased competition in labour and product markets. The fact that the average union pay differential fell so little during the 1980s may reflect its low level at the outset, but it means that there is little for legal changes to explain (Stewart, 1991). Pay structures were shaped in the 1980s primarily by changes in product markets, which broadly affected union and non-union firms alike, tending to intensify non-union causes of segmentation in labour markets (Ingram, 1991; Blanchflower, 1991; Blanchflower and Freeman, 1994).

The decline of union-related pay was most marked under the pre-entry closed shop. Moreover, the coverage of pre-entry closed shops contracted sharply during the 1980s. These two changes accounted for 60 per cent of the fall in the overall union pay differential during 1984-90 (Stewart, 1995). It is here that legal changes can perhaps be most plausibly linked to the institutional changes, even after allowing for differences in timing, and the primacy of changes in technology and product markets. It is not possible yet to give an account of developments in the 1990s, but it appears likely that increasing derecognition of unions has taken over from the decline of the closed shop as the main factor eroding union pay differentials, though the influence of union recognition on pay appears to be weaker than that of the closed shop (Gregg and Machin, 1991).

Wage pressure and pay inequalities

At aggregate level, weakened trade unions might be expected to exert less of the wage pressure to which Britain's inflationary problems have in part been attributed, thereby reducing the unemployment rate at which inflation accelerates (Layard et al, 1991). The prospects for such an interpretation are, however, greatly impaired by the weakness of estimated declines in union mark-ups since 1979 and by the smallness of the average mark-up, and also by the lack of consensus on the existence, let alone the size, of equilibrium unemployment (Main, 1996; Henry and Lee, 1997). The continuation of wage pressure after 1979 has been widely attributed to employer pay strategies and insider power, in non-union and union companies alike, rather than to any crude effect of unionism per se (Blanchflower and Freeman, 1994; Brown et al, 1995; Brown et al, 1997). The downward trend in the share of employment income in national and household incomes since the 1970s may, however have been generated partly by the, weakening of trade unions (Ryan, 1996). Otherwise the., legislation appears to have been tangential to wage setting and macroeconomic performance.

Pay inequality has since the 1970s grown more rapidly in the UK than in any other advanced economy except the USA. While the general tendency for pay inequalities to rise in most advanced economies has been attributed to technical change and international trade, marked differences in trends between countries have been attributed primarily to changes in national institutions, with the deregulation of the UK economy as an outstanding instance (Freeman and Katz, 1995; OECD, 1996; Nickell and Bell, 1996). Between one-eighth and one quarter of the growth of earnings inequality in the UK during the 1980s has been attributed to the decline in bargaining coverage (Gosling and Machin, 1995; Schmitt, 1995).

Although we have judged legal changes to have been a secondary cause of declining unionisation in Britain, statutory wage protection constitutes an important area of influence upon economic outcomes. The weakening and eventual removal of wages councils was associated with increased pay dispersion in low wage sectors and, to a lesser extent, declining relative pay for, and growing pay inequality amongst, young workers, who were the first to be removed from coverage (Machin and Manning, 1994; Garonna and Ryan, 1991). In that respect at least, legislative change has contributed directly to the widening of pay inequalities in contemporary Britain.

9. Economic outcomes--productivity

Stronger effects might be anticipated for industrial relations legislation in relation to economic performance. It is widely accepted that Britain's system of industrial relations constrained productivity growth in the post-war period, albeit with disagreement over whether the prime responsibility lay with trade unions or employers, or both. To the extent that trade unionism was responsible, legislation since 1979, with its anti-union orientation, should have improved economic performance. As with pay, outcomes have moved in the intended direction. Although Britain's investment and employment record continues to cause concern, at least the post-war gap between productivity growth in Britain and other European economies has not re-emerged since 1979 (Crafts, 1997). The difficulty is to distinguish the contribution of anti-union policies in general, and the legislation in particular, from a wider recasting of British industrial relations (Metcalf, 1993; Nolan, 1996).

The productivity effects of trade unions are complex and the theoretical debates cannot be explored here. Their influence depends, inter alia, upon product market characteristics, occupational specificity, the nature of technological change, and the employer's response. No general prediction emerges concerning union effects on productivity. The matter is for any open-minded economist an empirical one, dependent on both the quality of management and the competitiveness of the product market.

Strong traditions of craft control and workplace bargaining have long implied adverse productivity effects in a range of industries. There is evidence that the end of the 1970s marked a turning point. The extent to which nonpay issues were negotiated at the workplace declined between 1980 and 1984 (Millward et al, 1992). A linking of productivity with pay negotiations become more frequent in the early 1980s (Marsden and Thompson, 1990). Productivity growth in unionised employment increased relative to that in its non-union counterpart between the 1970s and the early 1980s, by when, by most accounts, it had become absolutely higher (Wadhwani, 1990; Oulton, 1990; Nickell et al, 1992). The picture which emerges is the retreat of restrictions on working practices, and an association with higher productivity growth not of unionism per se, but of unionism at that juncture.

There is mixed evidence on the investment and innovation aspects of the union-productivity nexus. That suggestive of an adverse association between unionism and productivity includes a tendency for firms which derecognised unions to achieve higher than average productivity growth during the mid-1980s (Gregg et al, 1993); and a lower rate of investment, ceteris paribus, at unionised than at other manufacturing establishments in 1984 (Denny and Nickell, 1992). By contrast, Machin (1995) found that the extreme disinvestment of plant closure was no more common in union than at non-union workplaces during 1984-90; Daniel (1987) found widespread support amongst employees and union representatives for the introduction of technical change; and, applying multivariate analysis to the same data, Machin and Wadhwani (1991) found that union recognition was associated positively, albeit weakly, with the introduction of technical change. Similarly, the introduction of human resource management techniques appears if anything more extensive where unions are present (Millward, 1994; Guest and Hoque, 1996).

When the net is cast wider to consider particular attributes of economic performance, positive associations with trade union presence have been found in econometric studies of training (Green, 1993;), labour turnover (Elias, 1994) and health and safety at work (Sandy and Elliott, 1996). The dual importance of, first, market failure in the provision of training (Stevens, 1995) and, second, training for productivity (Steedman and Wagner 1988) point to a productivity-enhancing role for trade unions. Similar considerations apply to health and safety at work where the weakening of union influence since 1979 has encouraged the erosion of pre-1980 successes (OECD, 1989; Dawson et al, 1988).

These findings rule out any simple association between unionism and productivity and point to the other factors which mediate between the two. The potentially important factors are the characteristics of employers and product markets. The role of the employer is difficult to investigate because few data sets offer good measures of employer attributes, with the result that the tendency of economists to study the `effects of unions' rather than the `effects of industrial relations' is encouraged. There is ample evidence that management, under pressure to raise efficiency, has reshaped pay structures and greatly increased control over the internal organisation of the firm. Sometimes this has been done by marginalising trade unions. But in other cases, such as the steel and car industries, remarkable improvements in productivity have been achieved by employers working with unions, linking the interests of unions' workplace organisations to the interests of the enterprise.

In sum, the legislation appears to have contributed substantially to economic performance only in the few, highly publicised cases in which extensive pre-entry closed shops and associated union controls were removed after intense conflict. Elsewhere, the key factors have been heightened product market competition and more coherent management strategy, which have radically reshaped British industrial relations and removed many of the constraints on economic performance with which it was previously associated.

10. Conclusions

Since about 1980 there have been fundamental changes in the way labour is managed throughout the industrialised world. Collective bargaining and trade union membership have retreated under the combined pressures of intensified competition in product markets and sustained high unemployment. The British experience of these changes coincided with the adoption by successive Conservative governments of a policy of hostility towards collective bargaining. In many ways this represented a reversion from collective to individual laissez faire, although the Conservative laws went further in applying a deregulatory competition policy to the labour market. In this review we have attempted to identify how far these legal interventions had an effect upon the industrial relations system over and above the changes which would have been expected to follow from the wider transformation of the economic environment.

For trade unions, the legislation has undoubtedly produced new financial and procedural burdens, and obstacles to the recruitment and retention of members. It also provided employers with powerful new sanctions at both the collective and individual levels, and in this way helped to shift power further away from organised labour. It is, however, doubtful whether even these sanctions could have been deployed so effectively under different economic conditions. The apparent success of the Conservative laws cannot be seen in isolation from the rise in unemployment and from wider changes to the structure of labour and product markets which occurred during this period.

So far as the wider economic effects of the legislation have been concerned, the difficulty has been one of disentangling any distinctive consequences in circumstances where employers already had powerful competitive reasons to reduce the scope of collective bargaining. The impact of the legal change has certainly been very uneven across industries, with most effect in those industries which were both confronted by an unprecedented competitive shock and where uncompetitive employment practices had been allowed to build up in the past. But while a legislative attack on collective bargaining may have served its purpose in these cases, in many others competitive shocks were met equally effectively by maintaining collective bargaining, albeit in somewhat different forms from before. In judging between these strategies in the longer term, a crucial question is whether productive efficiency can be maintained without employees having some expression of collective voice on such matters as health and safety, training, and employment flexibility.

From a comparative perspective, the Conservative years mark a less dramatic point of departure than it seemed at the time. The declines in union density and in the incidence of industrial action were broadly in line with the experience of other countries. Moreover, certain distinctive features of the British system--most notably the weakness of centralised bargaining mechanisms and the absence of a comprehensive statutory floor of rights--were reinforced by the policies of the 1980s and 1990s. The new laws also confirmed an underlying assumption of adversarialism and separation of interests in British labour-management relations. New forms of co-operation in the workplace emerged, but this was not because the old assumptions of adversarialism were rejected. If anything, they were strengthened by the draconian new legal sanctions made available to employers and the courts. Rather, it was because unions were weakened by high unemployment and falling membership. The true test of the new laws will only come at a point if, under changed economic circumstances, there is some revival in the bargaining power of organised labour.

During the period of Conservative government in Britain, a quite different route for industrial relations law, involving institutionalised support for labour-management co-operation, was being mapped out within the systems of mainland Europe and by the emerging body of EC labour law. A central issue now is whether Britain will continue to drift towards a system of individualised employment relations within a largely union-free environment, as in the United States, or will join the European mainstream in acknowledging a role for collective employee representation and statutory support for minimum employment standards. The arrival of a Labour government committed to the Social Chapter may provide the answer.

References

Addison J.T. and W.S. Siebert (1997), `Union security in Britain', forthcoming in Journal of Labour Research.

Beaumont, P.B. and R.I.D. Harris (1995), `Union de-recognition and declining union density in Britain', Industrial and Labour Relations Review, vol. 48, no. 3.

Blanchflower, D. (1991), `Fear, unemployment and pay flexibility', Economic Journal, no. 101.

Blanchflower, D. and R.E Freeman (1994), `Did the Thatcher revolution change the British labour market? in R. Barrell (ed.), The United Kingdom Labour Market, Cambridge University Press.

Blanchflower, D. (1996), `The role and influence of trade unions in the OECD', Centre for Economic Performance Discussion Paper no. 310.

Brown D., P. Ingrain and J. Wadsworth (1997), `Free to choose'. Centre Piece, vol. 2, no. 1.

Brown, W.A. and S. Wadhwani (1990), `The economic effects of industrial relations legislation since 1979', National Institute Economic Review, no. 131.

Brown, W.A. and J. Walsh, `Corporate pay policies and the internationalisation of markets', in J. Niland, R. Lansbury and C. Verevis (eds), The Future of Industrial Relations, Thousand Oaks: Sage.

Brown, W.A., P. Marginson and J. Walsh (1995), `Management: pay determination and collective bargaining' in P. K. Edwards (ed.), Industrial Relations: Theory and Practice in Britain, Blackwell.

Bullock Committee of Inquiry (1977), Report on Industrial Democracy, Cmnd.6706, HMSO.

Claydon, T. (1996), Union derecognition: a re-examination, in I. Beardwell (ed.), Contemporary Industrial Relations, Oxford University Press.

Commissioner for the Rights of Trade Union Members and Commissioner for Protection Against Unlawful Industrial Action (1996), Annual Reports 1995-96, Commissioners' Office.

Crafts, N. (1997), Britain's Relative Economic Decline, 1870-1995, London, Social Market Foundation. Cully, M. and S. Woodland (1997), `Trade union membership and recognition: an analysis of data from the 1996 Labour Force Survey', Labour Market Trends, June.

Daniel, W.W. (1987), Workplace Industrial Relations and Technical Change, Pinter. Dawson, S., P. Willman, M. Bamford and A. Clinton (1988), Safety at Work: the Limits of Self-Regulation, Cambridge University Press.

Deakin, S. (1991), `Legal change and labour market restructuring in Western Europe and the US', New Zealand Journal of Industrial Relations, no. 16.

Deakin, S. (1997), `Integration through law?', in J.T. Addison and W.S. Siebert (eds.), Labour Markets in Europe. Dryden.

Deakin, S. and G. Morris (1995), Labour Law, Butterworths.

Denny, K. and S. Nickell (1992), `Unions and investment in British industry', Economic Journal, no. 102.

Disney, R., A. Gosling and S. Machin (1996), `What has happened to union recognition in Britain?', Economica, no. 63.

Donovan (1968), Royal Commission on Trade Unions and Employers' Associations, Report, Cmnd 3623, HMSO.

Dunn, S. and D. Metcalf (1996), Trade union law since 1979, in Beardwell, op cit.

Ebbinghaus B. and J. Visser (forthcoming), The Development of Trade Unions in Western Europe, 1945-95, Campus.

Elgar, J. and R. Simpson (1993), The impact of the law on industrial disputes in the 1980s', in D. Metcalf and S. Milner,(eds), New Perspectives on Industrial Disputes, Routledge.

Elgar, J. and R. Simpson (1996), Industrial Action Ballots and the Law, IER.

Elias, P. (1994), `Job-related training, trade union membership and labour mobility, Oxford Economic Papers, vol. 46, no. 4.

Ewing, K. and B. Napier (1986), `The Wapping dispute and labour law', Cambridge Law Journal, no. 45.

European Works Councils Bulletin, various.

Ferner, A. and R. Hyman (forthcoming), Industrial Relations in the New Europe, second edition, Blackwell.

Freeman, R.B. and J. Pelletier (1990), `The impact of industrial relations legislation on British union density, British Journal of Industrial Relations, vol. 28, no. 2.

Freeman, R.B. and L.E Katz (1995), Differences and Changes in Wage Structures, University of Chicago Press.

Gall, G. and S. McKay (1996), `Injunctions as a legal weapon in industrial disputes', British Journal of Industrial Relations, vol. 34, no. 4.

Geroski, P., P. Gregg and T. Desjonqueres (1995), `Did the retreat of UK trade unionism accelerate during the 1990-1993 recession?', British Journal of Industrial Relations, vol. 33, no. 1.

Gosling, A. and S. Machin (1995), Trade unions and the dispersion of earnings in British establishments, 1980-90', Oxford Bulletin of Economics and Statistics, vol. 57, no. 2.

Green, E (1993), `The impact of trade union membership on training in Britain', Applied Economics, no. 25.

Gregg, P. and S. Machin (1991), `Changes in union status, increased competition and wage growth in the 1980s', British Journal of Industrial Relations, vol. 29, no. 4.

Gregg, P. and S. Machin (1992), `Unions, the demise of the closed shop and wage growth in the 1980s', Oxford Bulletin of Economics and Statistics, vol. 54, no. 1.

Gregg, P., S. Machin and D. Metcalf (1993), `Signals and cycles: productivity growth and changes in union status in British companies', Economic Journal, no. 103.

Gregory, M. and V. Sandoval (1994), `Low pay and minimum wage protection in Britain and the EC', in Barrell (ed.) op cit. Guest D. and K. Hoque (1996), Human resource management and the new industrial relations', in Beardwell op cit.

Hayek, E (1973), Rules and Order, Routledge.

Heery E. (1996), `The new new unionism' in Beardwell, op cit.

Henry, B. and K. Lee (1997), `Unemployment, wage dispersion and labour market flexibility', in J. Michie and J. Grieve-Smith (eds), Employment and Economic Performance, Oxford University Press.

Ingrain, P. (1991), `Ten years of manufacturing wage settlements: 1979-89', Oxford Review of Economic Policy, vol. 7, no. 1.

Kahn-Freund, O. (1959), `Labour law', in M. Ginsberg (ed.), Law and Public Opinion in Britain in the Twentieth Century, Stevens.

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

Leopold, J.W. (1997), `Trade unions, political fund ballots and the Labour Party', British Journal of Industrial Relations, vol. 35, no. 1.

Machin, S. and A. Manning (1993), `The effects of minimum wages on wage dispersion: evidence from UK Wages Councils', Industrial and Labour Relations Review, vol. 47, no. 2.

Machin, S. (1995), `Plant closures and unionisation in British establishments', British Journal of Industrial Relations, vol. 33, no. 1.

Machin, S. and S. Wadhwani (1991) `The effects of unions on investment and innovation', Economic Journal, no. 101. Main, B. (1996), `The union relative wage gap', in D. Gallie, R. Penn and M. Rose (eds), Trade Unionism in Recession, Oxford University Press.

Marginson, P., P. Armstrong, P. Edwards and J. Purcell (1996), `Facing the multinational challenge', in P. Leisink, J. Van Leemput, and J. Vilrokx (eds), The Challenges to Trade Unions in Europe, Elgar.

Marsden, D. and M. Thompson (1990), `Flexibility agreements and their significance in the increase in productivity in British manufacturing since 1980', Work, Employment and Society,, vol. 4, no. 1.

Metcalf, D. (1993), `Industrial relations and economic performance, British Journal of Industrial Relations, vol. 31, no. 2.

Metcalf, D. and M. Stewart (1992), `Closed shops and relative pay', Oxford Bulletin of Economics and Statistics, vol. 54, no. 4.

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

Millward, N. (1994), The New Industrial Relations?, Policy Studies Institute.

Milner, S. (1995), `The coverage of collective pay-setting institutions in Britain, 1895-1990', British Journal of Industrial Relations, vol. 33, no. 1.

Nickell, S. and B. Bell (1996), `Changes in the distribution of wages and unemployment in OECD countries', American Economic Review, no. 86.

Nickell, S., S. Wadhwani and M. Wall (1992), `Productivity growth in UK companies, 1975-86', European Economic Review, no. 36.

Nolan, P. (1996), `Industrial relations and performance since 1945', in Beardwell, op cit OECD (1989), Employment Outlook, OECD.

OECD (1994), Employment Outlook, OECD.

OECD (1996), Employment Outlook, OECD.

Oulton, N. (1990), `Labour productivity in UK manufacturing in the 1970s and in the 1980s', National Institute Economic Review, no. 132.

Ryan, P. (1996), `Factor shares and inequality in the UK', Oxford Review of Economic Policy, vol. 12, no. 1.

Sandy, R. and R.E Elliott (1996), `Unions and risk', Economica, no. 63.

Schmitt, J. (1996), `The changing structure of male earnings in Britain, 1,974-88, in Freeman and Katz (eds) op cit.

Steedman, H. and K. Wagner (1989), Productivity, machinery and skills, National Institute Economic Review, no. 128.

Stevens, M. (1996), `Transferable training and poaching externalities' in A. Booth and D. Snower (eds), Acquiring Skills, Cambridge University Press.

Stewart, M.B. (1987), `Collective bargaining arrangements, closed shops and relative pay', Economic Journal, no. 97.

Stewart, M.B. (1990), `Union wage differentials, product markets and the division of rents', Economic Journal, no. 100.

Stewart, M.B. (1991), `Union wage differentials in the face of changes in the economic and legal environment', Economica, no. 58.

Stewart, M.B. (1995), `Union wage differentials in an era of declining unionisation', Oxford Bulletin of Economics and Statistics, vol. 57, no. 2.

Undy R., P. Fosh, H. Morris, P. Smith, and R. Martin (1996), Managing the Unions, Clarendon.

Wadhwani, S. (1990), `The effects of unions on productivity, growth, investment and employment', British Journal of Industrial Relations, vol. 28, no. 3.

Willman, P., T. Morris and B. Aston (1993), Union Business, Cambridge University Press.

Willman, P. and T. Morris (1995), `Financial management and financial performance in British trade unions', British Journal of Industrial Relations, vol. 33, no. 2.

Wright, M. (1996), `The collapse of compulsory unionism?', British Journal of Industrial Relations, vol. 34, no. 4.

William Brown, Simon Deakin and Paul Ryan, Correspondence should be addressed to the authors at the Faculty of Economics and Politics, University of Cambridge, Austin Robinson Building, Sidgwick Avenue, Cambridge, CB3 9DD. The authors wish to thank Maria Hudson and Cliff Pratten, with whom they are currently engaged in fieldwork on the individualisation of employment contracts. Thanks also to members of the CLARE Group for helpful comments, and to Greg Bamber, Bill Callaghan, Bernard Ebbinghaus, Anthony Ferner, Simon Goldstone, Mark Hall, Bob Hepple, Richard Hyman, Bente Ingebrigtsen, Steve Jefferys, Ian McAndrew, Paul Marginson, John Monks, Graham Osborne, Jim Scoville, Priya Sharma, Steen Sheuer, Tony Smith, Bob Simpson, Mike Terry, and Paul Willman for miscellaneous help.