1
The Impact of the National Wage Agreement on Firm-Level
Competitiveness and Wage Inequality
*Seamus McGuinness, Elish Kelly and Philip O’Connell
The Economic and Social Research Institute, Whitaker Square, Sir John Rogerson‟s Quay,
Dublin 2, Ireland.
Abstract
This paper uses a linked employer-employee dataset to analyse the impact of institutional wage bargaining regimes on levels of average labour costs and within firm wage dispersion in private sector companies in Ireland. The results show that while centralised bargaining reduced labour costs within both the indigenous and foreign-owned sectors, the relative advantage was greater among foreign-owned firms. The analysis suggests that there are potentially large competitiveness gains to multinational companies that choose to locate in countries implementing a centralised bargaining system. Furthermore, the results provide additional support to the view that collective bargaining reduces within firm wage inequality.
JEL Classification:. J51, J52, J58.
Key Words: Wage Bargaining Regimes, Labour Costs, Wage Dispersion, Ireland.
* Corresponding author: Tel.: +353-1-8632067; Fax: +353-1-8632100. Email address:
seamus.mcguinness@esri.ie (S. McGuinness). We are grateful to the Central Statistics Office (CSO) for access to the data used in this study.
Introduction
This paper uses a linked employer-employee dataset to analyse the impact of institutional wage bargaining arrangements on levels of average labour costs and within firm wage dispersion in private sector corporations in Ireland. Average labour costs are taken as a proxy measure for firm-level competitiveness, while dispersion measures the extent of wage inequality within the firm. Average labour costs can be thought of as a health measure relating to both the firm and, by extension, the wider economy, while dispersion relates more closely to employee wellbeing and, more particularly, the relative position of low paid workers in firms operating within one of the fastest growing OECD economies.
It is important to note that the questions addressed here go beyond that of the impacts of trade union membership on firm level performance and inequality as, due the nature of wage bargaining arrangements in Ireland, and continental Europe more generally, the proportion of workers covered by collective bargaining regimes far exceeds the proportion who are members of trade-unions (Hartog, 2002). While collective wage bargaining processes will be the primary channel through which trade unions impact wage levels, this is not to say that there are no additional influences deriving from firm trade-union density. However, within a multivariate framework these additional influences can be thought off as separate to those derived from the wage bargaining process itself. This paper adds to the limited literature in this area by considering theses issues within an Irish context. The study also provides some assessment of the extent to which the impact of wage bargaining regimes differs across privately-owned indigenous firms and multinational companies (MNC‟s).
Context and Literature
Since 1987, wage bargaining in Ireland has been centralised at the national level, through a process known as social partnership. This partnership approach, which
involves voluntary negotiations between the Government, main employer bodies and
1
2, was introduced by the Government at this time to assist it in moving the trade unions
country out of the bleak economic situation that it found itself in; a period characterised by high inflation, weak economic growth and, subsequently, considerable unemployment, mass emigration and unsustainable government
34borrowing and national debt. There have been seven agreements to-date, each of
which has been tailored to medium term national economic and social needs, and has often built on its predecessor. Initially, pay and wage issues were the core elements of the negotiated agreements; specifically moderate wage increases in exchange for reductions in income tax to boost take-home pay. However, as the country recovered and moved into the „Celtic Tiger‟ era, the partnership nature of the agreements became deeper and their coverage was extended to include various social and welfare issues that either emerged or become more prominent as the economy prospered.
The national pay agreements that have been negotiated through social partnership have been identified in a number of studies (Sexton & O‟Connell (1996), Lane (1998), Sweeney (1998), Aust (1999), O‟Donnell (1999), Hardiman (2000), Teague &
Donaghey (2004), Baccaro & Simoni (2007), O‟Donnell (2008)) as having played an important role in the remarkable revival in the Irish economy that has taken place in
5the last two decades. In particular, most of this research indicates that the wage restraint attained under the pay agreements enhanced the country‟s competitiveness, through lower labour costs, and this consequently led to both significant employment and economic growth. Figure 1 illustrates that real unit labour costs in Ireland have fallen in most years since the social partnership process began in 1987, and are well below those in the rest of the European Union (EU) and United States. Over the 1987-2002 period, unit labour costs fell by around 25 per cent in Ireland compared to, approximately, 10 per cent across the EU. Thus, this would seem to suggest that the
2 In 1996 the Government extended the list of social partners eligible to participate in the negotiation of national agreements to include various voluntary and community groups. 3 Ireland had one of the highest budgetary deficits in the European Community (EC) in 1987, 10.7 per cent of Gross National Product (GNP). In addition, it had a national debt in excess of 120% of GNP, an unemployment rate of 18.5 per cent, net emigration close to 30,000 (which was equivalent to the natural increase in the population at this time), GDP per capita that was only 64 per cent of the EC average and very high interest rates (Department of the Taoiseach, 1987). 4 An eight agreement is currently being negotiated. 5 In the Netherlands, a form of corporatism, similar to Ireland‟s social partnership model, has been identified with having contributed significantly to the „Dutch Miracle‟ (Visser & Hermerijck, 1997;
Visser, 1998).
2
wage increases that have been attained under social partnership have been modest enough to boost Ireland‟s international competitiveness.
Furthermore, a few authors (for example, Lane (1998) and Aust (1999)) have suggested that, in addition to Ireland‟s low corporation tax rate and highly educated workforce, some MNC‟s were attracted to locate in Ireland because of its improved cost competitiveness, along with the relatively stable industrial relations environment that social partnership gave rise to. The attraction of MNC‟s to Ireland through
6, has also been identified as having played a very foreign direct investment (FDI)
important role in Ireland‟s economic transformation (see, for example, Barry (1999),
Barry et al. (1999) and OECD (1999, 2008)), and it continues to be a significant contributor to the economy‟s performance through both output and employment. For
example, the number of individuals employed in manufacturing in MNC‟s in Ireland
has grown from almost 74,000 in 1986 to 109,000 in 2005, a 48 per cent increase. Thus, MNC‟s share of total manufacturing employment in Ireland has increased over
7this period from 35 to just under 50 per cent (Figure 2). Finally, Tansey (1998) and
Gunnigle & McGuire (2001) also raise the possibility that centralised bargaining has helped improve the competitiveness of multinationals in Ireland.
While the literature has alluded to the benefits of social partnership in Ireland‟s recent
economic success, specifically through the national wage agreement (NWA) component of the process, due mainly to a lack of data, there appears to be no
8previous empirical analysis of its direct impact on Ireland‟s competitiveness.
6 The term MNC‟s and FDI are used interchangeably throughout the paper. 7 This employment data can from the Census of Industrial Production (CIP), which is produced by the Central Statistics Office (CSO), and 2005 is the most recent year for which data is available. Unfortunately, we do not have data from this source on nationality of ownership for the services sector. However, data available from Forfás (Ireland‟s national policy and advisory board for enterprise, trade, science, technology and innovation) indicates that the number of employees employed in FDI-owned services companies increased from 2,945 in 1986 to 153,508 in 2007, a growth of 52 per cent (Forfás, 1995, 2007). This data, however, only relates to FDI firms that received government support and, therefore, will underestimate the total. 8 Fitzgerald (1999, 2000), however, looked at wage formation in Ireland between 1962 and 1994, using national aggregate data, and concluded that the slow down in real wage growth observed after 1980 was due to “market forces” as opposed to centralised wage bargaining. He argued instead that the main
contribution of social partnership to Ireland‟s economic turnaround has been in bringing about
industrial peace as opposed to moderating wage increases, and hence improving the country‟s competitiveness (see also Bradley et al., 1991).
3
However, there are grounds to believe that social partnership will lead to lower labour costs, particularly for MNC‟s. Leahy & Montagna (2000) develop a theoretical framework which predicts that MNC‟s may prefer to locate in countries with centralised bargaining, due to the gain in competitiveness associated with wage setting aimed at securing the interests of domestic firms. In the model, unions will choose to limit the rent extracted from the MNC in order to preserve employment levels in the less competitive domestic firms. Leahy & Montagna (2000) argue that there may be a welfare loss to the economy in situations where the MNC‟s compete directly with domestic firms; however, they suggest that this is unlikely to be the case in Ireland where FDI has mainly occurred in sectors that did not have a substantial domestic component. Baccaro & Simoni (2007) also argue convincingly that by linking wage growth in the high value added MNC‟s sector with that in the less competitive domestic component of the Irish economy, centralised wage bargaining, through the NWA, will have substantially enhanced the competiveness levels of MNC‟s. However, the authors only provide some descriptive support for their
9. argument
In relation to wage dispersion, a number of studies have shown that countries with centralised wage bargaining have significantly lower levels of earnings inequality (OECD, 1997). These results have led to some degree of acceptance of the hypothesis that coordinated wage bargaining limits growth in inequality. Thus, a potential secondary impact of social partnership, and more specifically its NWA component, may have been to reduce earnings inequality. However, the evidence for Ireland is somewhat mixed. Barrett et al. (1999) found that there was a large growth in earnings dispersion in Ireland between the commencement of social partnership in 1987 and 1994. The authors concluded that centralised wage setting was not sufficient to limit the growth in earnings inequality. Further analysis by Barrett et al. (2002)
demonstrated that wage inequality stabilised in the mid-1990s, a result they attributed in part to a marked increase in high-skilled inward migration. More recently, McGuinness et al. (2009) found that wage inequality fell for Irish men between 1994 and 2001, and for females between 1997 and 2001. The authors partially attributed the
9 While not explicitly reporting a link with competitiveness, Radulescu & Robson (2006) report some tentative evidence that co-ordinated wage bargaining systems can help reduce any deterrent effect of high levels of trade union density on MNC location decision.
4
female result to the introduction of the minimum wage in 2000, a policy measure that was introduced through social partnership. However, while McGuinness et al. (2009)
do not directly address the issue, it cannot be ruled out that the reduction in wage inequality observed over the period in question may be due, in part, to centralised wage bargaining.
From a wider international perspective, while there is a great deal of literature relating to trade union effects on earnings, both in the context of a wage premium (Freeman & Medoff (1981), Lewis (1986), Stewart (1987), Schumacher (1999), Forth & Millward (2002)) and wage dispersion (Freeman (1982), Blau & Kahn (1996) DiNardo et al.
(1996), Card et al. (2003), Dustmann & Schönberg (2004)), much less is known with respect to the impact of bargaining regimes on firm level wage dispersion and even
10. Existing evidence is mostly restricted to less on measures of firm competitiveness
European countries, presumably due both to the absence of such institutional arrangements elsewhere in the world and/or a lack of available data. Regarding wage inequality, for Germany, Gerlach & Stephan (2005) found, for blue collar workers,
11lower levels of dispersion in firms implementing industry-level agreements, while
Kohn & Lembake (2007) reported that industry and firm-level bargaining resulted in lower levels of wage dispersion relative to individual-level bargaining. Domínguez et
al. (2004) found that firm-level agreements had a negative impact on firm-level wage dispersion in Spain compared to industry-level agreements. Finally, Plasman et al.
(2007), who examined bargaining regimes in Belgium, Demark and Spain, reported higher (lower) levels of wage inequality associated with firm-level bargaining in Belgium and Denmark (Spain), and argued that this disparity in results may be accounted for by cross-country variations in the goals of trade unions.
Turning to the impact of bargaining regime on average labour costs, Kohn & Lembake (2007) found higher average wages in German firms implementing both firm and industry-level agreements. Again for Germany, Guertzgen (2006) compared the impacts of industry and firm-level agreements and found a small wage premium
10 There is also an extensive body of research that looks at the impact of collective bargaining regimes on macroeconomic performance indicators. Aidt & Tzannatos (2008) provide a recent systematic review of the empirical research in this area, along with the relevant theoretical literature. 11 See also Stephan & Gerlach (2005)
5
for industry-level bargaining in West Germany and firm-level bargaining in the East. Some further evidence can be taken from the impact of bargaining on individual wage outcomes; Card & De la Rica (2006) reported a premium to individuals covered by firm-level bargaining in Spain relative to those covered by industry agreements. Plasman et al. (2007) also found that in Belgium, Denmark and again for Spain, that firm-level agreements resulted in a wage premium relative to workers covered by industry-level bargaining.
Data and Methods
The data used in this study comes from the 2003 National Employment Survey (NES). The 2003 NES is a matched employer-employee workplace survey, covering both the public and private sectors, which was carried out by the Irish Central Statistics Office
12. The employer sample was drawn from the CSO‟s Central Business Register. (CSO)
Selected firms were then asked to extract a systematic sample of employees from their payrolls. Approximately 6,500 private sector employers and 300 public sector bodies
13were surveyed across the economy. Within this, a total sample of 60,000 employees
were included from the private sector and 29,000 from the public sector. In total 54,000 returns were received.
The employer questionnaire requested information on employee earnings, hour‟s
1415worked and occupation. Information was also obtained on ownership type, firm
size, industry, use of pay agreements, percentage of employees covered by the minimum wage and provision of certain employee benefits, specifically career breaks and work-sharing and/or part-time work. With respect to the pay agreements
12 While the NES was of enterprises with 3 plus employees, the results were calibrated to the Quarterly National Household Survey (QNHS) employment data for employees (excluding agriculture, forestry and fishing), which covers all employees. 13 Only employers with more than three employees were surveyed and the data were collected at the enterprise level. 14 The earnings information collected in the 2003 NES represents the gross monthly amount payable by the organisation to its employees, and relates to the month of March in 2003. This includes normal wages, salaries and overtime; taxable allowances, regular bonuses and commissions; and holiday or sick pay for the period in question. It does not include employer‟s Pay Related Social Insurance (PRSI), redundancy payments and back pay. 15 The “ownership” variable allows us to distinguish public and private sector organisations, and, inter alia, within the private sector between indigenously owned and subsidiaries of foreign companies.
6
information, employers were asked to indicate the proportion of their workforce covered by the following types of agreement:
i) Individual-level
ii) Business-level
iii) Industry-level
iv) National-level (NWA)
v) Other
Clearly, business, industry and national-level represent the various types of collective wage bargaining.
Employees were issued with a separate questionnaire within which they provided information on their age, gender, educational attainment, family status, employment status (part-time or full-time), length of time in paid employment, length of service with current employer and also other job-related characteristics (for example, trade union membership, shift-work, supervisory role, and flexi hours).
The dataset at hand does not allow an assessment of the impacts of bargaining agreements on individual level earnings as we cannot be certain (below a 100 per cent threshold) that a specific form of wage bargaining covers an individual. Given this, the key aim of this paper is to assess the impact of institutional wage bargaining on firm-level average labour costs and wage inequality. To assess the impact of bargaining on average labour costs, we retain one observation per firm and derive a number of variables that we use to explain both average labour costs and wage dispersion. We apply establishment-level weights to these firm-level observations to ensure that our results are representative. Given that the NWA is the governing wage bargaining strategy used in the public sector in Ireland, suggesting little variation in this component of the data, the study is restricted to private sector firms. Once this
adjustment is made to the data we are left with a sample of 3,755 firms employing 275,570 employees. Of the firm sample, 414 are MNC‟s and 3,341 indigenous firms, each respectively employing 93,025 and 182,545 individuals.
7
An advantage of focussing on firm-level impacts is that while it is not certain that statistically significant individual-level wage effects translate to comparable firm-level effects, we can be more confident that any firm-level impact will be indicative of the wage consequences for the majority of relevant workers within the firm. More importantly, from a policy perspective this approach also allows us to assess the extent to which average labour costs are affected by wage bargaining arrangements and therefore their potential impact on competitiveness.
In addition to average wage level impacts, we also assess the effect of various forms of bargaining on wage dispersion within the firm. A priori, we would expect firms implementing forms of collective bargaining to have more compressed wage distributions; however, it is less clear to what extent wage compression varies by the mode of collective bargaining. As a dependant variable in this analysis we opt for the coefficient of variation as such a dimensionless measure is preferable when making comparisons across firms with widely varying wage distributions.
Results
The following section provides a descriptive assessment of the extent to which modes of wage bargaining vary by firm size, ownership (foreign versus indigenous) and sector.
The first column in Table 1 looks at the average coverage rate of the various types of agreement across all private sector firms. Individual-level agreements constitute the dominant wage bargaining mechanism in the private sector with, on average, 49 per cent of employees within the typical firm covered by such agreements. The average coverage level of the NWA was 30 percent, while industry and business-level
16. However, agreements each had an average coverage level of less than 10 per centthe average in this case is misleading as it is not the case that firms typically implement a mix of different types of agreement. Specifically, as the second column of Table 1 indicates, within firms‟ one form of wage agreement tends to dominate.
16 Due to differences in the average employment size of firms implementing various forms of bargaining, these figures will not correspond to the percentage of the workforce covered by each type of agreement.
8
For instance, for those firms implementing the NWA to any degree, the average coverage rate within the firm is just below 90 per cent. This point is further illustrated by the fact that taking a 70 per cent employment coverage cut-off point, only 12.3 per cent of firms do not implement a single form of agreement above this level. The remaining distribution of firms implement the various forms of agreement at the 70 per cent cut-off point were as follows: 43 per cent for an individual-level agreement, 28 per cent the NWA, 7 per cent for both business-level and industry-level agreements and 2 per cent for other. Consequently, given that the overwhelming majority of firms adopt a particular form of wage agreement for more than 70 per cent of their workforce, we use cut-off points as the baseline for the analysis undertaken in the paper. The 12.3 per cent of firms not implementing a single bargaining strategy above the 70 cut-off point forms the comparison group for the study, which we consequently refer to as the „no majority‟ grouping.
Table 2 reports the average proportion of employees covered by each bargaining arrangement by firm size at the 70 per-cent employment coverage cut-off point. Perhaps, not surprisingly, individual-level bargaining is more heavily implemented by smaller firms, while collective bargaining arrangements, in particular the NWA are, on average, more heavily implemented by firms employing more than 50 workers. These results suggest that the share of the workforce employed under the NWA (individual agreements) will be higher (lower) than the proportions reported in Table 1.
With respect to sector, and adopting a similar interpretation as in Table 2, Table 3 shows that individual-level bargaining is dominant in the Business Services and Transport & Communications sectors, the NWA is somewhat over-represented in Manufacturing, while industry-level agreements are most important within the Construction sector.
Finally, Table 4 looks at the incidence of bargaining agreement by ownership type. Interestingly, while the generally accepted view suggests that MNC‟s are more likely to adopt individual-level bargaining and are, therefore, less inclined towards collective forms of wage determination, the results suggest that the opposite is in fact the case in the Irish context. MNC‟s were found to be somewhat less likely to
9