National systems of public sector pay:

implications for ‘welfare outcomes’ and economic stability

Damian Grimshaw

(Draft, please do not quote without permission)

Introduction

The UK public sector workforce has witnessed a battery of reforms in recent years, inspired by an ideological rejection of the role of government in economic activity. These reforms have attempted to replace traditional employment practices in the public sector with new, stylised principles of labour organisation found in the private sector. Strategies include the devolution of decision-making, the decentralisation of pay, the weakening of job security and the application of closer links between pay and productivity. The reforms are not supported by empirical evidence that shows a clear relationship between improved economic performance and any of the stylised private sector practices. Rather, the policies tend to be underpinned by the assumption that decentralisation and devolution allow local managers to exercise greater flexibility in pay and employment strategies, from which all performance improvements follow (true to the conventional economic wisdom that sound incentive structures at the micro-level are a sufficient condition for good macroeconomic performance). Over the years, however, a growing body of evidence points to the adverse impact of strategies of pay decentralisation and greater employer autonomy on socio-economic indicators, such as the level of wage inequality, the gender pay gap and economic stability.

Alongside the range of studies which trace the implications of the radical restructuring for organisational and economic performance, a lively debate has also emerged around the appropriate union response. In a number of studies, reflecting a wide range of research, Fairbrother (eg. 1994a, 1994b, 1996) argues that the devolution and decentralisation of management structures in the public sector ought to be accompanied by a similar restructuring of union structures, reflecting the lesser need for national co-ordinated union action and the greater need for local action, and the opportunities this generates for ‘union renewal’. The thesis centres on the opportunities for increasing the participation of local union members, whether in workplace decision-making structures or in the coordination of local union action in response to management policy. However, in providing a relatively optimistic outlook of the implications of decentralisation for ‘union renewal’, the analysis is in danger of downplaying the potentially adverse consequences for ‘welfare outcomes’ and economic stability.

This paper seeks to contribute to the debate on the benefits of local versus national bargaining structures in the public sector, by drawing on cross-national research evidence. Some of the recent international studies on changing characteristics of wage bargaining raise parallels with Fairbrother’s work. In particular, the breakdown of centralised arrangements for the negotiation of wage equality in Sweden and Italy has led to recent calls for the transformation of union strategy, to emphasise workplace, rather than national-led activity (Baccaro and Locke 1998). This paper begins by providing a critical assessment of these prescriptions and outlines some of the major problems. This is followed by a characterisation of public sector pay systems in Europe and the US, and an attempt to map the different national systems against ‘welfare outcomes’ (wage structure) and economic stability (macro-economic coordination). The paper concludes by reaffirming the role of centralised public sector pay as an important factor in the equation between equality and economic stability.

1. National vs. workplace bargaining

In a critical assessment of Fairbrother’s work, Gall argues that the ‘union renewal’ thesis tends to be presented as "an end in itself" (1998: 151) and thus downplays the importance of union tactics, the context of employer resources and the overall gains for labour (op. cit.). While greater workplace participation in union activities is to be welcomed, Gall points out a number of potential weaknesses in promoting local bargaining as a union strategy. For example: local deals may be prone to ‘partnership’ agreements; local members may be less informed where they are not able to draw upon the knowledge and support of national officers; and devolution of management structures may be an illusion, conjured up to veil the fundamental management decisions which are still made at corporate level, outside of union penetration. Given the reality of decentralisation in many areas of the economy, the proper prescription may be to pursue a two-pronged strategy - for union renewal within the workplace, and to restore national bargaining (op. cit.).

For the UK public sector, the observation that employers may still retain fundamental managerial decisions and resources at a central level (Marginson et al. 1987) has particular relevance. A number of studies which trace the impact of decentralised budgets and decision-making identify the constraints on local ‘flexibility’ imposed by the retention of control at the central level (Walsh 1993, see also the review by Bailey 1994). The centralised determination of the level of public sector expenditures, and, critically, the overall paybill, for each sub-sector of public services defines the essential range for supplementary decentralised bargaining. If unions fail to confront the central determination of the total budget, then local negotiations may involve little more than reaching a compromise over a trade-off between job cuts and pay rises to meet the "efficiencies and other economies" required to meet local budget targets. Hence, it would seem that unions must devise strategies on both fronts, both to counter the current macroeconomic policy logic that prioritises a balanced budget (leading to the maintenance of a paybill freeze since 1994), and to counter the "new managerialist" policies at the workplace level.

In addition, the observation that local deals are prone to partnership agreements raises the question of whether public sector organisations might negotiate differential pay settlements for similar occupational groups, in response both to differential external circumstances, such as the state of the local labour market and the level of regional economic performance, and to internal characteristics, such as the ability of the organisation to generate surplus funds. Decentralised bargaining arrangements may thus foster an increase in wage inequality, both across organisations providing similar services and across workers with similar skills and experience. In addition, where the labour market for skilled and professional staff is tight (for example, qualified nursing and teachers), local bargaining may encourage a pay spiral as organisations compete for scarce labour. In the context of fixed local budgets, a pay spiral at one end of the job hierarchy is likely to be accompanied by job cuts or wage cuts for less skilled workers. Hence, if wage equity is an objective, then unions still need to design and implement strategies at a national level, where similar wage increases can be established across all organisations and all occupational groups.

However, the presumption that unions ought to pursue the goal of wage equity as an end itself has recently come under attack in a detailed assessment of the rise and fall of wage equity policies in Italy and Sweden (Baccaro and Locke 1998). Put briefly, the authors argue that policies of wage equality, implemented through national-level pay determination, are unsustainable since, on the one hand, they generate increasing conflict between different groups of the workforce and, on the other, they relegate issues of production to second place and thus lessen the ability of unions to adapt to the transformations taking place within the workplace. The alternative, they argue, is for unions to embrace a strategy focused on production issues (such as the promotion of training, teamwork and rotation of job tasks) at the workplace level.

The pursuit of equality through easily-made distributional formulae like the Italian scala mobile and the Swedish ‘earning development guarantees’ is short-lived and ulitmately self-defeating. If trade unions are to regain their capacity to foster solidarity among different groups of workers, they need to construct equality through new, supply-side strategies aimed at equalising the conditions and opportunities for both training and employment so that all workers have more challenging and rewarding jobs. Perhaps unions can turn to their own advantage recent managerial efforts to engage workers in various forms of workplace reorganisation (Baccaro and Locke 1998: 303; italics in original).

Again, however, there are problems with this thesis. In particular, by treating pay equity as a goal in its own right, the thesis gives insufficient attention to the wider social and economic role of pay equity. First, it is claimed that excessive compression of the wage structure discourages workers from investing in training and education since the material gains to skills acquisition are small (Baccaro and Locke 1998: 302). While this may be true, the incentives for employers to provide training may increase. Alternative evidence for Italy, for example, suggests that efforts to reduce wage differentials led to pressures on employers to introduce job rotation and enlargement of jobs so that semi-skilled workers could develop a wide range of skills to match their improved earnings position (Regalia and Regini 1995: 141-147).

Secondly, the impact of wage compression on intra-workplace structures may be less important than inter-organisational effects. In Sweden, for example, the principle of wage solidarity was an important component of the ‘Rehn-Meidner’ model, since it ensured that the wage rate earned by an individual was equivalent to the value of the work done, independent of the profitability or success of the particular sector in which a person happens to be employed (Wise 1993: 78). The well-known effect is, theoretically, to force low paying, low performing, firms out of business and to provide dynamic firms with a surplus for continued investment (Standing 1988). Importantly, pay differentials do not provide the only ‘signalling mechanism’ for labour mobility in the Scandinavian model, since active labour market policies enable displaced labour to transfer into more advanced industries; hence, the relationship between pay differentials and employee-based incentive structures is, in any case, likely to be fairly loose.

Thirdly, as well as potentially playing an integral role in the development of skills and improvements in firm performance, greater wage equality may increase the prospects for economic stability. Comparative data show a weak, but positive, relationship between a country’s level of income inequality and economic instability, measured as the standard deviation of growth in GDP over the 1980s (Corry and Glyn 1994: 212-214). Also, there is some evidence that a redistribution of income to the low paid would reduce the growth of imports and thus improve the balance of payments (Barooah 1988). Finally, the ability of social partners at the national level to negotiate the level of employment and pay of a substantial fraction of the workforce may facilitate a degree of macroeconomic co-ordination otherwise absent. A centralised, national system of public-service sector pay determination may enable social partners to draw upon the same principles of redistribution which traditionally underpin the allocation of public services, when negotiating the allocation of public funds to be paid to the workers employed. Esping-Andersen, for example, argues that:

... there may exist a trade-off between equality and efficiency in countries where the welfare state is large and very redistributive but in which the collective bargaining system is incapable of assuring wage moderation and stable non-conflictual industrial relations. ... as long as a large and redistributive welfare state is matched by neocorporatist-style political exchange mechanisms, equality and efficiency are compatible ... (Esping-Andersen 1994: 725).

This final point highlights the role of "connective" wage bargaining institutions in consolidating the associative behaviour of trade unions and employers in conjunction with the contradictory economic role of the state. In the UK, the lack of a "connective" institutional context means that the level of pay and employment in the public sector has always been dependent upon steady economic growth, rather than part of the institutional structures which underpin the economic model of production and reproduction. The continuing shift away from national-level institutional arrangements to the workplace may be expected to exacerbate this situation.

In an attempt to identify some of the consequences of a shift towards greater workplace bargaining in the UK public sector, the following sections draw on the evidence from a range of countries in an attempt to clarify the relationship between the different institutional forms of public sector pay systems, the wage structure and the degree of economic stability. The next section characterises different systems of public sector pay across a broad selection of OECD economies for which information was available - namely, the fifteen member states of the European Union, Norway and the US. An understanding of the range of typical features of national systems of public sector pay provides a basis to assess firstly, the implications for "welfare outcomes" and, secondly, the implications for economic stability.

2. Characterising national systems of public sector pay

The following discussion distinguishes national systems of pay regulation in the public services across three dimensions: first, the degree of integration of the system of pay determination across all employees in public services (see Rubery and Fagan 1994, 1995); second, in light of current trends towards decentralisation, the degree of local wage negotiation either as part of national pay settlements or as a substitute to them (see OECD 1993); and thirdly, the extent to which collective bargaining is subject to unilateral employer regulation (see Marsden 1994). The aim of the analysis is not to provide a detailed investigation of the system of pay regulation in each country, but rather to use the method of cross-country comparison in order to detail the range of features that characterise diverse systems of pay determination in the public-services sector.

Integrated vs. fragmented public sector pay systems

If an abstract continuum exists between a highly integrated and a relatively fragmented system of pay structures in the public sector, then Ireland and the UK would stand out as having relatively fragmented systems, and Italy and the Netherlands might be placed in between the two poles. For the majority of countries, pay structures are highly integrated, which in some cases involves a single pay spine covering all public service staff (Rubery and Fagan 1995: Table 2).

In Norway and Denmark, for example, a uniform pay scale covers all government employees (Marsden 1994: Appendix). In Germany, non-manual public servants without Beamte status (Angestellte) are located on the same pay scale, with the exceptions of magistrates and university lecturers. Importantly, however, this scale is divided into four sub-scales which imposes a "pyramidal rule of promotion" from one step to another (Meurs 1993). Collective bargaining is regulated predominantly by the "BAT" agreement (Bundesangestellten-tarifvertrag) for non-Beamte white-collar staff (Angestellte) and by three umbrella agreements at the federal, state and Länder levels for manual workers (Arbeiter) (Maier et al. 1993: 46-49). In France, the entire public sector workforce is bound by a single grid of pay scales (grille de la fonction publique). Four general pay scales exist, distinguishing between managerial positions of responsibility, clerical staff, specialised posts and non-specialised posts. Moreover, centralised pay increases (applied either to the pay scale or to the rate of premia) may be limited to certain categories of staff such as the low paid, and rates of advancement may vary within each pay scale according to variations in the significance of seniority in determining the number of index points (Marsden 1994: Appendix, Meurs 1993: 82).

Less integrated systems of pay scales are found in Italy and the Netherlands. In Italy, a 1983 agreement reduced the previously high degree of fragmentation by restricting bargaining to eight separate wage agreements, divided by sector of activity: schools, universities, the health sector, local authorities, ministries, autonomous enterprises, non-commercial public bodies and public research bodies. For example, within the health sector, all employees share an eleven-point pay scale, but this is split into four parts according to occupational groups (Bettio and Villa 1993: 50, 61-62). There is also, in principle, an inter-sectoral arrangement in Italy which specifies guidelines for the separate agreements. For example, a ceiling for wage settlements is set in line with the government’s projection for inflation, thus providing the condition for more closely integrated settlements. In practice, these upper limits have often been surpassed. For example, according to the agreed ceilings, civil service pay should have increased by 26% between 1987 and 1991, but in fact increased by 46% (Meurs 1993). In the Netherlands, separate arrangements exist for employees in health and education. In the health sector, pay is settled according to nine different collective agreements distinguishing between site of employment (such as hospital or old persons’ home), and, because of the number of different job evaluation systems, pay scales are not well integrated. In education, on the contrary, the eight collective bargaining agreements are reasonably well integrated. Also, the two pay scales which traditionally applied in public administration were recently harmonised (Plantenga and van Velzen 1993: 125).

Ireland and the UK are the clearest examples of highly fragmented public sector pay structures. In both countries, a voluntarist tradition of industrial relations underpins the establishment of separate pay scales for different groups of employees such as police, teachers and civil servants. Nevertheless, there are important differences. In Ireland, the centralised system of arbitration procedures strengthens the notion that some public service jobs are comparable. Also, there are strong pressures for greater comparison in pay determination across sectors in Ireland, and, in practice, civil service grades are used as a basis for wage comparisons (Barry and Roche 1993). In contrast, the fragmentation of pay determination in the UK has paved the way for greater localisation and individualisation of pay (see below). Moreover, in Ireland, a centralised agreement covering the years 1988 to 1990 was negotiated in 1987, and was followed by an agreement under the Programme for Economic and Social Progress for the years 1991 to 1993. Despite these attempts to coordinate public and private sector wage settlements, however, free collective bargaining at the firm level in the private sector continues to dominate wage agreements (Barry and Roche 1993, Marsden 1994: Appendix).

Decentralised vs. centralised pay systems

Centralised wage settlements vary by country according to the various provisions which enable national wage rates to be supplemented by locally-negotiated pay supplements. In Germany, the Netherlands, Ireland and Italy the scope for local variation in pay is relatively insignificant. In Germany, there is some scope for managerial discretion (at the Länder level) in cases where centralised job descriptions are vague, thus providing an opportunity to grade comparable staff at different levels (Rubery and Fagan 1994: 93). In the Netherlands, evidence for the early 1990s confirms the minimal practice of local wage bargaining. In the health sector, for example, 21 out of 758 institutions implemented supplementary wage agreements at the local level (Plantenga and van Velzen 1993). In Ireland, the scope for local variation in pay is weak for non-manual employees, but far stronger for manual workers (Barry and Roche 1993). Finally, in Italy, few negotiations take place at the local level, and these are largely within limits defined by national agreements. It is clear from Table 1, for example, that local level variations in Italy are far less significant in the public sector compared to the private sector. Average earnings in education and central government do not include "firm-specific allowances", contrary to earnings in the banking and engineering sectors which include average shares of 9% and 21% respectively. Unlike Germany and the Netherlands, however, there are moves in Italy to introduce greater scope for managerial discretion at the organisational level in the public sector. Recent legislation has removed the need for local managers to seek state approval for each organisational decision in a number of areas of workforce planning including overtime, training, incentive pay for work groups, and turnover (Bettio and Villa 1993: 59-61).

Other countries make more systematic use of local pay flexibility, or have recently introduced significant legislation in this direction. In Denmark, for example, "local wage pools" were introduced in 1989 marking a reversal of past practice whereby all collective bargaining took place at the central level. The reforms have enabled local managers to introduce individualised merit payments (Knudsen and Pedersen 1993: 57). In France, centralised agreements allow for the use of "non-established" personnel (contractuels) to cope with shortages of staff at the local level. By the late 1980s, employment of contractuels constituted around 18% of the total public sector workforce. In the case of the education sector, where 40% of all contractuels were employed in the early 1980s, contractuels are generally used to meet the demand for specialised staff. Although such strategies relieve local organisations of the rigidity and time lag involved in filling posts at a centralised level, contractuels do not enjoy job security or career opportunities (Meurs 1993: 87-88). Finally, in the US, legislation in the 1990s has aimed to reduce the national uniformity of centralised wage awards for employees in federal government. The Federal Employees Pay Comparability Act of 1990 established the notion that pay of federal employees should be compared to the average earnings of non-federal employees, by locality. However, partly due to the costs and expertise required to establish the necessary earnings database, full implementation of this policy is not expected until the year 2004 (Katz and Krueger 1993: 67-68).

The UK and Sweden provide the strongest illustrations of policies to establish local pay negotiations as the main feature of public sector pay. The practice of setting aside a fraction of the public sector paybill for local negotiations has been a feature of Swedish policy for almost 20 years, following the agreement in 1978 to reserve 1% of the paybill for these purposes. Legislation in 1985 officially recognised the steady erosion of the national integrated pay scales, although they were to survive formally until 1989, and also promoted the process of decentralisation of pay determination (Schager 1993: 115-116). Attempts to imitate the pattern of collective bargaining in the private sector, however, have raised a number of conflicting interests, both among employer representatives and among trade unions of different areas of the public sector. Part of the problem lies with the expectation that local supplements will be used as a means of topping up earnings of those professional staff in demand from private sector employers rather than of those for whom the state acts as a near monopsonist employer. Also, the ability of public utilities to generate revenue is at odds with public service organisations which must rely solely upon public funds. The difference in funding arrangements creates a tension between a preference for local determination of pay, funded by the "ability to pay", in the case of the public utilities, and for national increases in pay that are funded entirely by central agreement in the case of the public-services sector (op. cit.: 117-120).

In the UK, under the past Conservative government, the drive to decentralise pay determination was associated with the introduction of internal, quasi-markets in education, health and the civil service. Despite their determination, national pay structures remained resilient for the majority of occupational groups; there are, however, important differences across sectors. In health and education, there is little evidence of the replacement of national with local pay arrangements. A national survey of the health sector in 1995 found that only 11 out of 482 Trusts in Britain had implemented local pay structures for the majority of staff (IDS 1995), and, in 1996, a survey of 87 Trusts identified just 10 Trusts that had introduced a new local pay scale for nursing staff (IDS 1996). In education, grant-maintained schools have the option of implementing local pay; however, only a small proporiton of schools have applied for such status. In contrast, developments in the civil service have been more radical. Employment in "executive agencies" accounted for around 64% of all employees in the civil service by July 1994, and, in 1995, 36 agencies (57% of all civil servants) conducted local pay negotiations and around half of these agencies introduced new local pay and grading arrangements (own calculations, Elliott 1995).

Free collective bargaining vs. unilateral employer regulation

Free collective bargaining for the entire public-service sector workforce exists in only a few countries. In Denmark, bi-annual agreements are directly applicable for all public sector staff, whereas in Germany free collective bargaining is only guaranteed (according to Article 9 in the Constitution) for employees without official civil service status (Beamte) (Marsden 1994: Appendix Table). In the majority of countries, for those groups covered by collective bargaining, it is usual for the wage agreement to be subject to parliamentary approval or explicitly constrained by Ministerial assessment of the implications for the state budget. The former condition is notable in Austria, the Netherlands, Sweden, and the UK and the latter condition applies in Finland, Italy and Norway (op. cit.). In Spain, pay is not negotiated by collective bargaining but by "pacts" according to legislation introduced in 1984. The settlements reached must then be approved by government (Mólto 1993: 41).

The polar extreme to free collective bargaining is unilateral employer regulation, in which the state freely imposes its own decision regarding the appropriate pay settlement reflecting its requirements for the planned fiscal, labour market, and managerial objectives. This situation is exemplified by the position of Beamte in Germany, who represent around 40% of the public sector workforce. Beamte are excluded from national collective bargaining agreements, and have their pay determined unilaterally by state regulation. Although legislation specifies that an "appropriate standard of living" must be maintained, the criteria for recognising this is left in the hands of the government. Hence, pay may be adjusted according to changes in average earnings, some measure of general living standards, or the state’s fiscal constraints (Meurs 1993). In practice, Beamte benefit from the wage-bargaining power of public sector trade unions as collective bargaining agreements usually precede wage awards to the Beamte. Nevertheless, the state has often exercised its control by delaying pay increases for several months, thus saving a significant sum to meet fiscal policy objectives (Maier et al. 1993: 59). France provides an interesting case as it is in a period of transition from unilateral employer regulation of public sector pay towards a system of collective bargaining. The principles of pay negotiations were established for civil servants in 1983. Nevertheless, there are no signs that the pressures to achieve the objectives of budgetary control are any less strict. As Gauvin et al. argue, "although bargaining between the social partners does take place", the level of the multiplier applied to the indices on the single integrated pay structure is still decided unilaterally by government (1993: 70). Other examples of unilateral pay regulation include senior officials in Sweden, civil servants in Greece and Switzerland, and all staff except manual workers in Luxembourg (Marsden 1994: Appendix, Rubery and Fagan 1994).

Between the polar systems of free collective bargaining and unilateral employer regulation, there exist alternative systems of pay determination which limit the potential for direct trade-union or employer intervention. In some countries, particular occupational groups are covered by independent pay review bodies, which make recommendations to the government based upon evidence collected from employer and staff-side representatives, as well as an official statement from the government itself. This form of pay setting is typically applied to senior high paid public-service sector employees. In the UK, the practice also applies to nursing and medical groups and teachers, in the form of Pay Review Bodies.

Pay regulation may also be enforced by the state according to an established fixed rule, typically negotiated or imposed as a means of providing compensation for those employees who have relinquished, or are not granted, the right to strike. Such predetermined rules may adjust the pay scale according to changes in the retail price index, for example, or to changes in the average earnings of all employees. The latter represents a clear application of the principle of pay comparability, ensuring that the pay of public sector employees does not fall behind that of the private sector workforce. In the UK, despite considerable resistance by the Conservative government to the notion of pay comparability during the 1980s and 1990s, the pay scales of the police, who do not have the right to strike, have been indexed to the underlying increase in the average earnings index since 1978. In the US, employees of the federal government are prohibited from collective bargaining on wages. In principle, federal pay scales, covering around one in six public sector workers, are fixed centrally in accordance with "comparability legislation" which makes adjustments to pay scales based on a survey of private sector establishments (Katz and Krueger 1993: 44).

The analysis of different systems of public-service sector pay from a cross-national perspective highlights the diverse principles of wage determination. In the following two sections, these principles provide a basis upon which to investigate two issues central to the changing political economy of public-service sector employment. A first issue concerns how the particular features of the systems of pay regulation, in particular, the degree of integration and centralisation of the pay system, influence the ‘welfare outcomes’ in society. The ‘outcome’ considered here is the shape of the wage structure, including the level of wage inequality and patterns of wage growth. A second issue involves the extent to which a country’s system of public sector pay acts as an important factor in the maintenance of economic stability. The question considered here is whether more centralised and integrated pay arrangements facilitate central macroeconomic coordination in the context of pressures to restrict public expenditures.

3. Wage structure

The notion that there is a strong relationship between the institutional structure of a country’s system of public sector pay and the wage structure has, within an expanding literature, recently gained currency as an important competing explanation to the conventional focus of neoclassical economists on the influence of labour supply and demand (for examples of the influence of this alternative thesis among the mainstream camp, see Goldin and Margo 1992, Leslie and Pu 1996, OECD 1993). A number of studies present quantitative evidence which suggests that the greater the degree of decentralisation of pay determination, the wider the dispersion of pay among all employees and between male and female employees (Blau and Kahn 1992, 1994, Marshall 1995, Rowthorn 1992, Whitehouse 1992, Zweimüller and Barth 1994). Rowthorn (1992), for example, demonstrates that the relatively centralised pay systems of Norway, Denmark and Sweden have a low level of wage inequality, whereas the relatively decentralised systems of the US, Canada and Japan are associated with high wage inequality. In the spirit of these findings, this section examines the evidence for a relationship between the characteristics of a country’s system of public sector pay and the structure of pay in the public sector. No attempt is made at a comprehensive cross-country comparison due to lack of pay data. Instead, the aim is to collect together fragments of evidence from a sample of countries, in order to draw preliminary observations and raise questions for further research.

Two of the three characteristics described above are assessed here: integrated vs. fragmented systems; and centralised vs. decentralised systems. First, comparison of a relatively integrated pay system against a relatively fragmented system reveals evidence of a clustering of pay rises among the public sector workforce in the former system, and divergence of earnings growth in the latter. For example, in Germany, data for earnings growth of different industrial sectors between 1975 and 1985 show that different sub-sectors of the public sector are clustered together (Rubery and Fagan 1994). It is likely that this pattern reflects the tight linkages among public sector employees paid according to a relatively centralised and integrated pay and grading structure.

In contrast, pay scales in the UK are determined separately for different occupational groups within different public service activities. For example, separate pay scales exist for clerical staff and ancillary staff, and the pay scale of each group also differs according to the sector of employment, such as the civil service, local authority or the NHS. In the UK, it is clear that the absence of an integrated pay scale has resulted in large differences in the growth of average real earnings of different occupational groups (Table 2, see also Brown and Rowthorn 1990). Between 1980 and 1992, whereas the average real earnings growth of full-time teachers in primary and secondary schools proceeded at a faster pace than non-manual earnings in the private sector, teachers in further education colleges, and, most markedly, male academics in universities, experienced slower real earnings growth than the private sector. Over the same period, female secretarial staff gained less within the NHS than their counterparts employed by the local authority. More dramatically, female kitchen helpers in full-time work experienced real wage improvements of either 6.5% or 30.5% depending on the sector of employment.

Concerning the second characteristic of a pay system (centralised vs. decentralised), while there is strong cross-national evidence for a positive relationship between decentralised pay systems and wage inequality, evidence of the impact of a shift towards greater decentralisation on the level of wage inequality is less clear. For example, greater public sector pay decentralisation is associated with an increase in wage inequality in Sweden and the UK, yet levels of wage inequality in the public sector outpace those in the private sector in Sweden, while the reverse is true in the UK. In Sweden, data for the 1980s show that pay dispersion in the public sector increased from 0.18 to 0.28, compared to an increase from 0.37 to 0.42 in the private sector (Table 3). A wage stabilisation agreement was implemented during the years 1991 and 1992, counteracting the effect of pay decentralisation and leading to a compression in pay. From 1993 to 1996, however, pay dispersion again increased, from 0.286 to 0.313 (Schager 1993, Schager and Andersson 1996).

For the UK, there is a similar increase in wage inequality in the public sector (a rise in the coefficient of variation from 0.407 to 0.425), yet this is far less than the change in the private sector (from 0.396 to 0.468) (Table 4). The UK evidence has to be interpreted in the context of the relatively hesitant shift towards greater pay decentralisation in the public sector, combined with the collapse of inter-industry bargaining arrangements in the private sector (see, for example, Brown 1993). One might expect, therefore, a larger rise in wage inequality in the private sector compared to the public sector. Also, if some of the rise in wage inequality can be attributed to a decentralisation of pay in both sectors, then the UK data provide some evidence for the argument that a continuation of policies to decentralise pay determination in the UK public sector may exacerbate trends towards a widening of wage inequality, which has proceeded at a faster pace in the private sector.

The difficulty in assessing the implications of policies of pay decentralisation on the wage structure underlines the need to distinguish the societal context of trends towards greater pay decentralisation. If the direction of change in an economy’s system of pay determination, towards either greater centralisation or decentralisation is a function of the particular institutional structure, then it is clear that it is not possible to assess pay decentralisation as a homogeneous process, but rather as a range of practices with diverse potential outcomes. In each country, the practice of public sector pay decentralisation is influenced by two major factors: the given system of public sector pay determination; and the changing patterns of wage determination in the private sector. For example, in the UK - unlike most other countries - the decentralisation of public sector pay has been occurring within a pay structure that already facilitated relatively strong variations in pay between different organisations and between groups of employees as a result of the high degree of fragmentation in national pay determination. Also, in the majority of countries where decentralisation has occurred in the private sector, firm-level bargaining operates as a supplement to prior settlements made at a national or sectoral level, which set minimum rates for different grades of staff. In the UK, by contrast, recent changes mean that firm-level bargaining in the private sector now operates as a distinct alternative to industry-level bargaining (Rubery 1993). Hence, in the UK, pay decentralisation in the public sector implies moving from a relatively high level of wage regulation to a relatively unregulated position (or, "employer-led self-regulation"), where local pay bargaining is unconstrained by either minimum wage legislation or minima set by prior collective bargaining agreements. Hence, both the institutional structure of public sector pay and the characteristics of private sector collective bargaining provide additional important indicators of potential differences in the significance of policies of pay decentralisation.

4. Macroeconomic coordination

There is an emerging body of research which contends that the more integrated and centralised the system of public sector pay, the greater the potential for the state to achieve flexible, macroeconomic coordination between improvements in economic performance and earnings growth. In particular, the state may be better able to fulfil its target level of public sector expenditures by exercising control over the public sector paybill - an important element in overall government spending. With the general tightening of fiscal policy throughout the European Union, those countries with more integrated public sector pay structures may, therefore, be better placed to impose constraints with more flexibility and with greater effect. Alternatively, in economies where pay arrangements are relatively fragmented or decentralised, some macroeconomic control is surrendered. Hence, the particular system of public sector pay determination may shape the form and effectiveness of strategies to establish sustained economic growth in the context of pressures to control public spending. This section examines the evidence for such claims.

First, where pay systems differ along the axis of unilateral pay-fixing and free collective bargaining, it is likely that the greater the degree of unilateral pay regulation, the greater is the strategic emphasis by the state on controlling costs to meet budget requirements because of the greater central control enjoyed. In countries where the pay system is based on free collective bargaining, or systems of independent review bodies, the primary strategic aim is more likely to be the achievement of greater organisational efficiency, by implementing policies of pay flexibility and the decentralisation of decision-making (Marsden 1994: 17).

Nevertheless, it may be misleading to make such a strong distinction between strategies of budgetary control, on the one hand, and organisational efficiency improvements, on the other. The two may be closely inter-related. The latter objective may, in some cases, be pursued as a means to achieve cuts in public sector expenditures by, for example, reducing the national influence of trade unions and eliminating public sector pay from the national political agenda. Also, it is possible that where both objectives are pursued, there may be a conflictual relation between pay devolution, which implies a disintegration of economy-wide pay coordination and a decentralisation of decision-making, and fiscal demands for macroeconomic coordination.

Second, where public sector pay systems are characterised as relatively centralised and integrated, strategies of restricting public spending appear to favour pay restraint, either through capping the overall paybill or holding back growth in wage rates. France is a good example of the association between a centralised and integrated pay system and strategies of wage restraint, having implemented a variety of policies of wage restraint over the years. However, these policies have enjoyed mixed success. Attempts in the late 1970s to impose a ceiling on the overall paybill were resisted by trade unions. Then, until 1982/83, public sector pay scales were informally linked to changes in prices. The government soon rejected this method, however, due to the failure to control for seniority increments, promotions and growing qualification levels among employees, which caused the paybill to increase at a faster rate than the level of prices (Marsden 1994: 26). Despite resistance from trade unions, the government implemented a pay strategy that accounted for "career effects", which consequently came to account for a greater proportion of annual increases in the overall paybill (increasing from 13% to 40% between 1979 and 1988). Again, however, success was limited - only three agreements were signed between 1984 and 1992, and a series of public sector disputes took place in 1989 (Meurs 1993: 94-95).

In the UK, on the other hand, the relatively fragmented system of public sector pay has been associated with three different, and conflicting, policy strategies to restrain public spending. First, in common with countries with centralised pay systems, the government has imposed a policy of wage restraint. The total wage bill has been restrained by the government through a policy of freezing public sector running costs at 1993-94 cash levels. However, during the early to mid-1990s, this policy conflicted with a second strategy - to encourage public-service sector employers to replace national pay bargaining with local pay arrangements. This conflict may explain much of the slow progress towards local pay determination; the government did not surrender macroeconomic control of spending limits, and therefore did not deliver financial autonomy to local employers in the operationalisation of local pay structures. In contrast to France and Germany, however, where the centralised, uniform system of pay at least facilitates the implementation of a uniform strategy of wage-rate restraint (despite the fact that it does not guarantee its success), the UK government has not imposed an upper limit to pay rises since 1993. Each fragmented pay settlement must, instead, award pay increases within a fixed budget, so that any improvements in earnings are offset by "increased efficiency and other economies". As a result, the payment of differential wage increases between groups of public-services sector employees is encouraged, exacerbating the past pattern of differential growth in gross earnings. Also, employers in the UK public-services sector are achieving a number of "other economies" in order to finance the wage demands necessary for the recruitment and retention of certain groups of staff. Examples include attractive offers of early retirement packages, in order to substitute young inexpensive staff for their older counterparts, as well as the introduction of performance-related pay in place of automatic seniority-linked annual pay increments. Moreover, there is a fundamental conflict between the two policies since wage bargaining among a greater number of local units increases problems of competitive wage bargaining leading to inflationary pressures (because each unit has little incentive to stick to a moderate settlement because of the small overall impact on inflation; see, for example, Paloheimo 1990) and thus makes it more difficult to maintain a centralised policy of wage restraint.

Perhaps, in response to the weakening of control over macroeconomic co-ordination associated with the relatively fragmented and increasingly decentralised pay system, the UK has increasingly strengthened its commitment to a third strategy, to privatise areas of the public-services sector in order to reduce public expenditures. This method has been applied most extensively through the introduction of compulsory competitive tendering (CCT), with the aim of introducing competition for the provision of public services and thus driving down costs. Although there are problems in comparing the cost-effectiveness of public services before and after the introduction of CCT, due to changes in the specification of job tasks which may have reduced the level or quality of provision of services, it is clear that the government has achieved considerable cuts in pay and employment across a range of public services. For example, a survey of 300 contracts in Local Government awarded in 1989/90 found that, of those which remained under control of the local government employer (90% of all tenders), 51% of contracts introduced job cuts, 10% imposed pay cuts, and 17% cut the average working time (Escott and Whitfield 1995).

An additional factor of interest here is the potential for countries with relatively centralised systems of public sector pay determination to increase opportunities for public protest and worker resistance against planned budget cuts, due to the greater degree of solidaristic organisation inherent in the wage-bargaining process. In recent years, such protests have been widespread in France, as well as in Germany - where the public protest in 1992 was the first in two decades (Ferner 1994: 59). In both countries, public sector pay is relatively centralised. This factor may explain the motives of other countries such as the UK and Sweden in seeking to reduce the power of trade unions at the national level by further fragmenting and individualising pay determination.

Discussion and conclusion

A review of the incomplete cross-country evidence provides some support for the claim that employment in the public sector, accompanied by a relatively centralised system of pay determination, is an important factor in the equation between equality and economic stability. The above analysis suggests that the more integrated and centralised the system of public sector pay, the greater the potential for the state to achieve flexible, macroeconomic coordination between improvements in economic performance and the growth of earnings. In Germany and France, where the systems of public sector pay are more centralised than in the UK, macroeconomic policies to meet the convergence conditions of European Monetary Union appear to be relatively politicised, in part, because of the greater degree of solidaristic organisation in the bargaining process. In the short term, therefore, there is greater potential in these countries for public protest and worker resistance against the planned budget cuts. In the longer term, however, it is possible that the eventual changes in economic policy are made on a relatively consensual platform, thus laying an important condition for future economic and political stability.

Within the UK, the significance of solidaristic wage arrangements and a more compressed distribution of income for achieving improvements in economic performance is yet to be acknowledged among policy-makers. The supposed positive benefits of introducing decentralised pay arrangements - improved links between pay and performance, on the one hand, or ‘union renewal’, on the other - are problematic. Authors such as Elliott (1995) and Marsden (1992), for example, both emphasise the viability of managerial reforms as the major mechanism through which to enhance efficiency in the public sector. Such reforms would, they argue, cultivate a greater sense of commitment among the workforce to the organisation and thus improve the quality of service provision. However, the greater decentralisation of pay arrangements and the increased capacity for employers to exercise discretion over employment practices may have unintended effects. In terms of the overall wage structure, greater pay decentralisation is likely to foster greater wage inequality, which may reduce prospects for economic stability. Also, radical changes in employment practices may adversely affect norms and expectations regarding pay and employment stability in the public services, and thus impede the ability of public service employers to recruit and retain highly-motivated staff under tight public spending constraints.

From the partial evidence presented here, there is clearly a need for greater research on the relative gains to be won through centralised as opposed to decentralised bargaining arrangements. As it stands, the relatively fragmented system of public sector pay in the UK does not appear to facilitate the degree of coordination between policies of public spending restraint and pay decentralisation that might prevent a spiralling of wage settlements. Closer integration of pay structures among the different occupations and sub-sectors of the public services might encourage a greater degree of macroeconomic coordination and establish a more solidaristic form of wage negotiation. These are essential elements in the struggle to attain lasting economic stability and prosperity on the basis of a more equal distribution of income.

Table 1. Make-up of gross annual earnings in Italy by sector, 1990.

Source: adapted from Bettio and Villa (1993: Table 22).

Table 2. Real earnings growth of full-time employees in selected occupations in the UK public sector, 1980 to 1992

Source: adapted from Elliott and Duffus (1996: Table 1).

Table 3. Pay dispersion in Sweden

Note: Pay dispersion is calculated as the difference between the upper and lower quartile divided by the median level of pay.

Source: Schager (1993: Table 1) and Schager and Andersson (1996: Table 3).

Table 4. Pay dispersion in the UK

Note: Pay dispersion refers to the coefficient of variation.

Source: Grimshaw (1998).

Bibliography

Baccaro, L. and Locke, R. M. (1998) The end of solidarity? The decline of egalitarian wage policies in Italy and Sweden, European Journal of Industrial Relations, 4 (3): 283-308.

Bailey, Rachel. (1993) Annual review article 1993: British public sector industrial relations, British Journal of Industrial Relations, (32) (1): 113-136.

Barry, Ursula and Roche, A. (1993) Wage Determination and Sex Segregation in Employment in Ireland, Report for the European Commission, Working Paper, EC Network on the Situation of Women in the Labour Market, Manchester School of Management, UMIST.

Bettio, Francesca and Villa, Paula. (1993) Wage Determination and Sex Segregation in Employment in Italy. Report for the European Commission, Working Paper, EC Network on the Situation of Women in the Labour Market, Manchester School of Management, UMIST.

Blau, F. and Kahn, Lawrence M. (1992) The gender earnings gap: learning from international comparisons, American Economic Review, 82 (2): 533-538.

---- and ---- (1994) Rising wage inequality and the US gender gap, American Economic Review Papers and Proceedings, 84 (2): 23-28.

Borooah, V. (1988) Income distribution, consumption patterns and economic outcomes in the UK, Contributions to Political Economy, Vol. 8.

Brown, William. (1993) The contraction of collective bargaining, British Journal of Industrial Relations, 31 (2): 189-200.

---- and Rowthorn, Bob. (1990) A Public Services Pay Policy, Fabian Tracts 542, London: Fabian Society.

Corry, Dan and Glyn, Andrew. (1994) The macroeconomics of equality, stability and growth, in A. Glyn and D. Miliband (eds.) Paying for Inequality: The Economic Cost of Social Injustice, London: IPPR/ Rivers Oram Press.

Elliott, Robert. F. (1995) Pay reform and pay dispersion in the public service: a study of three OECD countries, Public Management Occasional Papers, 2 (1), Paris: OECD.

Escott, K. and Whitfield, D. (1995) The Gender Impact of CCT in Local Government, Centre for Public Services, EOC Research Discussion Series.

Esping-Andersen, Gøsta. (1994) Equality and work in the post-industrial life-cycle, in D. Miliband (ed.) Reinventing the Left, Cambridge: Polity Press.

Fairbrother, P. (1994a) Politics and the State as Employer, London: Mansell.

---- (1994b) Privatisation and local trade unionism, Work, Employment and Society, 8 (3): 339-356.

---- (1996) Workplace trade unionism in the state sector, in P. Ackers, C. Smith and P. Smith (eds.) The New Workplace and Trade Unionism: Critical Perspectives on Work and Organisation, London: Routledge.

Ferner, Anthony. (1994) The state as employer, in R. Hyman and A. Ferner (eds.) New Frontiers in European Industrial Relations, Oxford: Blackwell.

Gall, G. (1998) The prospects for workplace trade unionism: evaluating Fairbrother’s union renewal thesis, Capital and Class, 66: 149-157.

Gauvin, A., Silvera, R. and Granie, C. (1993) Pay Determination and Sex Segregation in Employment in France, Report for the European Commission. Report for the European Commission, Working Paper, EC Network on the Situation of Women in the Labour Market, Manchester School of Management, UMIST.

Goldin, Claudia, and Margo, Robert, A. (1992) The Great Compression: the wage structure in the US at mid-century, Quarterly Journal of Economics, 57 (1): 1-34.

Grimshaw, Damian. (1998) Public sector employment, wage inequality and the gender pay ratio in the UK, 1986-1995, Mimeo.

IDS (Income Data Services).(1995) Pay in the Public Services: Review of 1994: Prospects for 1995, London: Incomes Data Services.

---- (1996) IDS Report, 725 (November).

Katz, Lawrence F. and Krueger, Alan B. (1993) Public sector pay flexibility: Labour market and budgetary considerations, in OECD (1993a).

Knudsen, R. and Pedersen, L. (1993) Wage Determination and Sex Segregation in Employment in Denmark. Report for the European Commission, Working Paper, EC Network on the Situation of Women in the Labour Market, Manchester School of Management, UMIST.

Leslie, Derek and Pu, Yonghao. (1996) What caused rising earnings inequality in Britain? Evidence from time series, 1970-93, British Journal of Industrial Relations, 34 (1): 111-130.

Maier, Frederike, Quack, Sigrid, Carl, A. and Strunk, S. (1993) Wage Determination and Sex Segregation in Employment in West Germany. Report for the European Commission, Working Paper, EC Network on the Situation of Women in the Labour Market, Manchester School of Management, UMIST.

Marginson, P., Edwards, P., Martin, R., Purcell, J. and Sisson, K. (1987) Beyond the Workplace: Managing Industrial Relations in the Multi-establishment Enterprise, Oxford: Blackwell.

Marsden, David. (1994) Public service pay determination and pay systems in OECD countries, Public Management Occasional Papers, Series No. 2, Paris: OECD.

Marshall, Adriana. (1995) Wage determination regimes and wage inequality: Latin American cases, paper presented at the 18th Conference of the International Working Party on Labour Market Segmentation, Tampere, Finland, July 1996.

Meurs, Dominique. (1993) The rationale for, and implications of, centralised pay determination systems in the public sector, in OECD (1993a).

Mólto, Marie L. (1993) Wage Determination and Sex Segregation in Employment in Spain. Report for the European Commission, Working Paper, EC Network on the Situation of Women in the Labour Market, Manchester School of Management, UMIST.

OECD. (1993) Pay Flexibility in the Public Sector, Public Management Studies, Paris: OECD.

Paloheimo, H. (1990) Between liberalism and corporatism: the effect of trade unions and government on economic performance in 18 OECD countries, in R. Brunetta and C. Dell’Arringa (eds.) Labour Relations and Economic Performance, London: Macmillan.

Plantenga, J. and van Velzen, S. (1993) Wage Determination and Sex Segregation in Employment: The Case of the Netherlands, Report for the European Commission, Working Paper, EC Network on the Situation of Women in the Labour Market, Manchester School of Management, UMIST.

Regalia, I. and Regini, M. (1995) Between voluntarism and industrialisation: industrial relations and human resource practices in Italy, in R. Locke, T. Kochan and M. Piore (eds.) Employment Relations in a Changing World Economy Cambridge: MIT Press

Rowthorn, R. E. (1992) Centralisation, employment and wage dispersion, The Economic Journal, 102: 506-523.

Rubery, Jill. (1993) Wage Determination and Sex Segregation in Employment:Report for the UK, Report for the European Commission, Working Paper, EC Network on the Situation of Women in the Labour Market, Manchester School of Management, UMIST.

Rubery, J. and Fagan, C. (in collaboration with D. Grimshaw). (1994) Wage Determination and Sex Segregation in Employment in the European Community, Social Europe Supplement 4/94, Luxembourg: Office for Official Publications of the European Communities.

Schager, Nils-Henrik. (1993) An overview and evaluation of flexible pay policies in the Swedish public sector, in OECD (1993a).

---- and Andersson, P. (1996) Continuity and reform in the public sector pay determination in the EU: analysis of the centralised and decentralised response to managing labour market change, Mimeo.

Standing, Guy. (1988) Unemployment and labour market flexibility: Sweden, Geneva: ILO.

Walsh, Janet. (1993) Internalisation v. decentralisation: an analysis of recent developments in pay bargaining, British Journal of Industrial Relations. 3 (3): 409-432.

Whitehouse, Gillian. (1992) Legislation and labour market gender inequality: an analysis of OECD countries, Work, Employment and Society, 6 (1): 65-88.

Wise, L. R. (1993) Whither solidarity? Transitions in Swedish public sector pay policy, British Journal of Industrial Relations, 31 (1): 75-93.

Zweimüller, Josef and Barth, Erling. (1994) Bargaining structure, wage determination, and wage disperion in 6 OECD countries, KYKLOS, 47: 81-93.