Monetary Regime and the Co-Ordination of Wage Setting

Holden, Steinar Working Paper Monetary Regime and the Co-Ordination of Wage Setting Arbeidsnotat, No. 2000/1 Provided in Cooperation with: Norges Bank, Oslo Suggested Citation: Holden, Steinar (2000) : Monetary Regime and the Co-Ordination of Wage Setting, Arbeidsnotat, No. 2000/1, ISBN 82-7553-157-8, Norges Bank, Oslo, https://hdl.handle.net/11250/2500413 This Version is available at: https://hdl.handle.net/10419/209775 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/deed.no 2000/1 Oslo February 1, 2000 Arbeidsnotat Research Department Monetary regime and the co-ordination of wage setting. by Steinar Holden Arbeidsnotater fra Norges Bank kan bestilles over Internett: www.norges-bank.no/publikasjoner eller ved henvendelse til Norges Bank, Abonnementsservice, PB 1179 Sentrum, 0107 Oslo Telefon 22 31 63 83, Telefaks 22 41 31 05 Norges Banks arbeidsnotater inneholder forskningsarbeider og utredninger som vanligvis ikke har fått sin endelige form. Hensikten er blant annet at forfatteren kan motta kommentarer fra kolleger og andre interesserte. Synspunkter og konklusjoner står for forfatterens regning. Arbeidsnotater/working papers from Norges Bank can be ordered via Internet: www.norges-bank.no/english/publications or from Norges Bank, Subscription service, P.O.Box. 1179 Sentrum, 0107 Oslo, Norway. Tel. +47 22 31 63 83, Fax. +47 22 41 31 05 Norges Bank's Working papers present research projects and reports (not usually in their final form), and are intended inter alia to enable the author to benefit from the comments of colleagues and other interested parties. Views and conclusions expressed are the responsibility of the author alone. ISSN 0801-2504 ISBN 82-7553-157-8 1 Monetary regime and the co-ordination of wage setting by Steinar Holden Department of Economics University of Oslo and Norges Bank First version: 12 October 1999 This version: 21 February 2000 Comments are welcome Abstract International comparisons show that countries with co-ordinated wage setting generally have lower unemployment than countries with less co-ordinated wage setting. This paper argues that the monetary regime may affect whether co-ordination among many wage setters is feasible. A strict monetary regime, like a country-specific inflation target, to some extent disciplines wage setters, so that the consequences of uncoordinated wage setting are less detrimental than under a more passive monetary regime (eg a monetary union). Thus, the gains from co-ordination are larger under a passive regime. Under some circumstances a passive regime may induce co-operation in wage setting, and thus lower unemployment, when a stricter regime would not. JEL Classification: E24, J5, E52. Keywords: Wage setting, co-ordination, equilibrium unemployment, monetary regime, monetary union, The views expressed are those of the author, and do not necessarily reflect those of Norges Bank. I am grateful to George Bratsiotis, Kai Leitemo and Kalle Moene for useful comments to an earlier version. Department of Economics, University of Oslo, Box 1095 Blindern, 0317 Oslo, Norway. Email: [email protected], homepage: http://www.uio.no/~sholden/ 2 1 Introduction Over the last decades, explanations of the high European unemployment have increasingly focused on the structure of the labour market. The wage formation system has played a key role in many of the explanations. An important view, advocated by among others Layard, Nickell and Jackman (1991), is that strong unions per se have a negative impact on employment, by increasing wage pressure. Yet if wage setters co-ordinate their wage setting, taking into consideration the negative external effects identified in the literature (cf. surveys in Moene, Wallerstein and Hoel, 1993, and Calmfors, 1993), the negative impact of unions may be mitigated. Indeed, the findings of several international comparisons (e.g. Nickell, 1997, OECD, 1997) support the theoretical prediction, showing a marked tendency that countries with a high degree of co-ordination in the wage setting have lower unemployment. More recently, Blanchard and Wolfers (1999) show that co-ordination in wage setting is one factor that explains whether adverse economic shocks have persistent effects on unemployment. Given the large potential benefits of co-ordination in wage setting, suggested by theory and supported by empirical evidence, a crucial question is: Why are unions in some countries able to co-ordinate their wage setting, while unions in other countries are not? This issue has, however, not received much attention within the economics literature (two exceptions are Holden and Raaum, 1991, and Holden, 1991a). As observed by Flanagan (1999), most studies take the labour market institutions as exogenous. To some extent this contrasts the political science literature, where this issue has been discussed considerably, cf. Iversen (1999) and the references therein. The present paper argues that the monetary regime may be one factor that affects whether co-ordination of wage setting is feasible. The argument builds on a recent literature, including Bratsiotis and Martin (1999), Soskice and Iversen (1998, 1999) and Coricelli, 3 Cukierman and Dalmazzo (2000),1, where it is shown that a non-accommodating central bank disciplines large wage setters by making employment more sensitive to the real wage. In the present paper, the result of this literature is blended with the analysis of Holden and Raaum (1991) and Holden (1991a) on whether co-ordinated wage restraint is feasible. Somewhat simplified, the idea is as follows. Consider an economy with several large wage setters (e.g. industry unions), where uncoordinated wage setting will lead to an adverse outcome with high unemployment. Unions have an incentive to co-ordinate their wage setting to achieve wage restraint. However, a binding agreement is not possible, so any union may deviate from the co-ordination and obtain higher wages. The deviator will obtain a gain in the short run, but at the expense of a possible breakdown of the mutually beneficial co-ordination on wage restraint. Co-ordination is only sustainable if unions' discount factor is sufficiently high so that the long run costs of a breakdown of the co-ordination outweigh the potential short run gain from deviating. Now enter the effect of the monetary regime. If unions co-ordinate their wage setting, the effect of the monetary regime is small, because a good equilibrium, with low unemployment, can be achieved in any case. In contrast, if wage setting is unionised, but uncoordinated, the monetary regime is more important, because a strict central bank may discipline wage setters and thereby dampen the negative consequences of uncoordinated wage setting.2 Thus, the gains from co-ordination of wage setting are larger if the central bank is accommodating, providing unions with a greater incentive to co-ordinate. This implies that 1 See also Horn and Persson (1988), Holden (1999), Wibaut (1999), and Vartiainen (1999). In a related literature, Cubitt (1992, 1995), Skott (1997), Jensen (1997), Gruner and Hefeker (1999), Cukierman and Lippi (1999a,b), Guzzo and Velasco (1999), and others argue, that unions are also concerned about inflation. Under this assumption an accommodating central bank may induce wage moderation and lower unemployment by “exploiting” that unions do not want to cause inflation. 2 This distinction is made clear in Bratsiotis and Martin (1999), but without any remarks on the possible implications for whether co-operation is feasible. 4 for some levels of unions' discount factor, co-ordinated wage restraint is only feasible if the central bank is accommodating. The idea that large unions' confederations may strive to restrain wages, but that they do not always succeed, may seem strange to many economists. However, this view is not only consistent with the theories and evidence mentioned above, it is also consistent with the rhetoric of trade union leaders in several European countries. In countries like the Netherlands, Denmark and Norway, leaders of trade unions' confederations talk publicly about the merits of co-ordinated wage restraint. Occasionally, unions' confederations even make explicit agreements with employers and the government with the intention of restraining wages (some examples are: the Wassenaar agreement in the Netherlands in 1982; the Solidarity alternative in Norway 1992; several agreements in Ireland in the 1980s and 90s). However, union leaders may find it difficult to persuade the rank and file, and sometimes they do not succeed in their wage restraint. The analysis of the present paper also sheds light on the consequences of participation in a monetary union. As argued by Soskice and Iversen (1998) and Cukierman and Lippi (1999b), monetary policy will not have the same disciplining effect on wage setting in a monetary union as in a country with a country-specific inflation target. The reason is that even if the central bank in the monetary union is a vigorous inflation fighter, the monetary policy response to higher wages in a single country would be small, because the effect on the aggregate price level in the union of higher wages in one single country would be small. In contrast, in a country with a country-specific inflation target, the central bank would counteract the effects of higher wages by increasing the interest rate. Thus, Soskice and Iversen (1998) and Cukierman and Lippi (1999b) argue unions will be more reluctant to increase wages under a country-specific inflation target, so that a monetary union will involve higher real wages and higher equilibrium unemployment. However, in this argument the 5 degree of co-ordination in wage setting is taken as exogenous. If one takes into consideration that the monetary regime may affect the degree of co-ordination, the conclusion may be reversed. Precisely because EMU may involve higher unemployment, the incentives for unions to co-ordinate wage setting at the national level are higher in the EMU. The analysis of the present paper implies that under some circumstances, co-ordination of wage setting would not be possible under a country-specific inflation target, but it would be possible under membership in the EMU. Thus, under these circumstances membership in the EMU would lead to lower equilibrium unemployment. To focus sharply on the effect of the monetary regime on whether co-operation in wage setting is sustainable, the model is kept simple on other accounts. There are no shocks in the model, so stabilisation issues are neglected. Other possible effects of participation in a monetary union, like increased product market competition, or a change in the incentives for labour market reform (Bean, 1998, and Calmfors, 1998), are also neglected. The paper is organised as follows. In section 2, I present the basic economic model, adapted from Bratsiotis and Martin (1999). In section 3, I analyse whether co-ordination in wage setting is feasible, drawing upon the theoretical framework in Holden and Raaum (1991). In section 4, I explore whether the key empirical prediction is consistent with data. Section 5 concludes the paper. 2 The model We consider an economy consisting of n symmetric firms. The workers are organised in k < n symmetric unions, so that each union covers n/k firms. Note that the assumption that the economy is symmetric and completely unionised is not important, but mainly done to simplify the analysis. The aim is to capture two important elements of European labour markets: (i) There are several fairly large wage setters (typically industry unions). (ii) Most wage earners 6 have their wage set by these large wage setters (bargaining coverage in Western European countries was in 1994 in the interval from 70 to 98 per cent, with two exceptions, UK 47 per cent and Switzerland 50 per cent, see OECD, 1997, table 3.3). In the model, time horizon is infinite. We first consider behaviour in one time period, which I shall to refer to as a year. To save on notation, time subscript is suppressed whenever possible. Firms produce under diminishing returns to scale, with labour as the only input (1) 10<<= γγ ii ly where lower case denotes natural logarithm, y is production, l is labour, γ is the elasticity of output with respect to labour, and subscript indicates firm. Firms set the price of their product facing a downward sloping demand curve3 (2) 0,)()( >−−−+= ηαη pppmagy ii where gis a constant, m is the nominal money stock, p is the aggregate price level defined as (3) ∑= =n ii p n p1 1 Unions care about the real wages and the employment of their members. The annual loss function of union j is 3 As shown by Weitzman (1985), this demand curve can be derived from optimising household behaviour. 13 consistent with international comparisons and the rhetoric of trade union leaders. In the following, I shall assume that this is the case. Thus, if unions were to co-ordinate their wage setting, they would choose lower wages than they do when wage setting is uncoordinated. If wage setting is uncoordinated, the real wage and unemployment outcome are derived by inserting σN = 1/k in (12) to derive λN, and then λN and σN are inserted in (14) to derive the equilibrium rate of unemployment uN. The associated equilibrium real wage, ωN = wN -pN, is given from the wage curve (13) (substituting out for the equilibrium rate of unemployment uN from (14)). Union loss, denoted ΩN, is derived by substituting out for the equilibrium real wage (13) and the equilibrium rate of unemployment (14) in (4)). The outcome if unions co-ordinate their wage setting, setting the wage that minimises their loss, can be derived in the same way, employing that, under complete co-ordination, σC = 1. The loss under co-operation, ΩC, is obviously smaller than the loss associated with uncoordinated wage setting, ΩN. From the discussion above, unemployment and real wages are lower under co-ordination than under uncoordinated wage setting, i.e. uC < uN, and ωC < ωN. However, co-operation being desirable is not sufficient to ensure that it is accomplished, and this is the topic of the next section. 3 Is co-ordination of wage setting feasible? One possible way of ensuring co-operation would be if the unions could sign a binding agreement on their wage setting. This, however, does not seem very realistic; a possible reason is the legal difficulties that would be involved, in view of the fact that collective agreements in reality are much more complicated than what is captured by the single wage 14 rate in the present model.5 Another solution would be joint wage setting, i.e. that all unions bargain collectively with the employers. To some extent this does occur. For example, in Norway, all industry unions in LO occasionally bargain jointly. However, while economywide, joint negotiations might solve the co-ordination problem, it would also involve many additional complexities. One problem might be that the industry unions have different views on other issues than wages, for instance regulation of part time work, manning ratios, flexible working hours etc, that might be easier to solve in separate negotiations. A second problem is that there is usually some scope for wage increases after the wage settlements (wage drift), and the scope for wage drift varies among the unions. Thus not even joint negotiations would solve all co-ordination problems. A third problem is that joint negotiations might affect relative wages; there appears to be a robust empirical regularity that joint wage setting reduces wage dispersion, cf. Wallerstein (1999). Thus, high wage industries may not want joint wage negotiations. A third solution, suggested by Soskice and Iversen (1998), is that one union acts as a wage leader vis-à-vis the rest of the labour market. However, as will be further discussed in section 4 below, this solution begs the question of why other unions will follow the leader. In the formal model, I shall neglect all these alternative solutions, and consider whether cooperation can be accomplished when each union always have the option of deviating from the co-operative solution. As wage determination takes place simultaneously in all industries, any one union can deviate from the co-operative agreement and set a higher wage. This will not be discovered until the other unions have set their wage. Thus, if the wage setting in one year is viewed in isolation, an agreement among the unions that they will all set the lower wage associated with 5 However, on rare occasions, e.g. in Norway in 1988-89, unions may accept legal restrictions on wage and price growth imposed by the government. 15 the co-operative equilibrium will not be credible. Each union prefers a higher wage for itself irrespective of the wage set by the other unions. Thus a prior agreement on wage moderation among the unions does not affect the fact that the unique one-shot Nash equilibrium in the wage setting game among the unions is the uncoordinated equilibrium derived above. All unions set the wage specified in (13) above, in the correct anticipation that the other unions will to the same. However, the main issue of the present section is that unions are also concerned about the future consequences of their behaviour. To analyse this, we must specify the overall objective of a representative union j, which is to minimise the discounted sum of annual losses (15) 1,,0, 1 1 0 << + − =Ω ∑ ∞ = βδ β δδ q q where tjt t The discount factor δ consists of two elements. The pure rate of time preference, as measured by the discount factor 1/(1+ β ), and the probability that the game ends, and a new game starts, q . The start of a new game may be associated with the replacement of the leaders of the unions, or that an important exogenous event takes place that makes the unions re-consider any agreement on wage moderation, and if necessary resume a co-operative solution after a breakdown (cf. discussion below). The game can now be analysed as a repeated game, where the annual wage setting is repeated over an infinite horizon of years (subject to the exogenous probability that the game ends, and a new game starts). The assumption that the horizon is infinite is the appropriate in this setting, because the notion that there exists a final period, where unions can neglect the future consequences of their behaviour, seems implausible in this setting (cf. Osborne and Rubinstein, 1990, page 54). 16 To derive the equilibrium of the game, it is convenient to define some further notation. Let ΩD(ρ) = Ω(wD(ρ); wC(ρ), ρ) denote the annual loss of a deviating union that sets wD(ρ) > wC(ρ), while all other unions set the wage that minimises their joint loss, wC(ρ), and the rate of monetary accommodation takes the value ρ. (wD is to be thought of as the optimal deviation. I shall not derive wD explicitly, as this is not necessary for our purposes.) As observed above the annual loss of the unions is ΩN(ρ) = Ω(wN(ρ); wN(ρ), ρ) under uncoordinated wage setting, and ΩC(ρ) = Ω(wC(ρ); wC(ρ),ρ) under the co-operative agreement. From the discussion above, it is clear that ΩD(ρ) < ΩC(ρ) < ΩN(ρ). Consider now the following strategy for the unions: 1. Co-operate by setting wC(ρ) until some union alone deviates from co-operation. 2. If some union alone has ever deviated from co-operation, then never co-operate (that is, set wN(ρ)) until a new game starts. Thus, if one of the unions deviates from the agreement, all unions revert to the one-shot Nash equilibrium (this is the standard trigger strategy, cf. Friedman, 1986). This strategy will support a subgame perfect equilibrium if unions incur a higher (or equal) discounted total loss by deviating from the agreement, inducing a reversion to the uncoordinated one-shot Nash equilibrium, than by sticking to the agreement. This condition is equivalent to (16) δ δ ρρ ρρ − ≤ Ω−Ω Ω−Ω 1)()( )()( NC CD The numerator on the left hand side of (16) indicates the reduced loss of a deviation in the year of deviation, while the denominator indicates the higher annual loss in subsequent years. 17 The expression on the right hand side reflects that the gain of a deviation is immediate, but only once, whereas the subsequent loss is delayed, but may go on for a long time. Let δ*(ρ) denote the value of the discount factor for which (16) holds with equality. The following Proposition is immediate from the analysis of trigger strategy equilibria (cf. Friedman, 1986). Proposition 2: There exists a unique critical value δ*(ρ) ∈ (0,1) such that co-operation is sustainable as a subgame perfect equilibrium under the strategy defined above if and only if δ ≥ δ*(ρ). Thus, the critical value for the discount factor depends on the degree of monetary accommodation, ρ. Both the temptation to deviate, as measured by the numerator of the right hand side of (16), and the subsequent costs associated with a breakdown of co-operative wage moderation, depend on ρ. A number of effects are present. First, ρ affects wage setting in symmetric equilibrium: a lower value of ρ (stricter central bank) leads to lower wages and lower unemployment, raising union welfare. Thus, a lower value of ρ mitigates the detrimental effect of uncoordinated wage setting. Secondly, ρ affects the gain from a deviation, via three channels. A lower value of ρ increases λ, which implies that (i) a wage rise has stronger negative effect on employment (this reduces the gain from a deviation), (ii) a wage rise has smaller impact on aggregate prices (this increases the gain from a deviation). (iii) A lower value of ρ reduces real wages and unemployment under co-ordination, so that the real wage is farther from its target value, while unemployment is closer to its target value. This increases the gain from a deviation. Thirdly, ρ affects price setting: a lower value of ρ reduces the mark-up by increasing the price elasticity. This is advantageous for the unions irrespective of the degree of co-ordination. 18 Despite its simplifying assumptions, the solution of the model is fairly complex, and I have been unable to sign the total effect unambiguously analytically. The numerical simulations in Table 1 illustrate the properties of the model. Because of the highly stylised nature of the model, the magnitudes of the differences cannot be taken seriously. Yet the simulations provide a rough indication of the effects that are at work. The main results are apparent from the three columns to the left of the Table, which presents the equilibria for three different degrees of monetary strictness ρ. • Vertical comparisons within each column shows that co-ordination leads to lower unemployment, lower real wages, and lower union loss than does uncoordinated wage setting. The effect is large: as indicated by the left column (ρ = 0.5), unemployment under co-ordination, uC, is 4 percent compared to 19 percent under uncoordinated wage setting, uN. • Horizontal comparison shows that a stricter central bank (lower ρ), leads to lower unemployment, lower real wages, and lower union loss, if the degree of co-ordination is kept constant. The effect is small under co-ordination: E.g. uC = 0.0425 for ρ = 0.5 and uC = 0.0423 for ρ = 0. The difference is, however, somewhat larger under uncoordinated wage setting, e.g. uN = 0.191 for ρ = 0.5 and uN = 0.177 for ρ = 0. • The bottom row shows that the critical value of the discount factor, δ*, above which coordination is sustainable as a subgame perfect equilibrium, is decreasing in ρ. This is the main result of the paper. As an illustration, in the base case, if δ is in the interval between 0.144 and 0.165, co-ordinated wage restraint is sustainable if ρ = 0.5, but not if ρ = 0. (If δ is below 0.144, co-ordinated wage restraint is sustainable for both values of ρ, if δ is above 0.165, it is sustainable for neither.) This reflects that the gain from co-ordination is 19 greater for a higher value of ρ (the difference between ΩC and ΩN is greater for higher values of ρ). • The eight columns to the right in Table 1 show that the qualitative results are not sensitive to the specific parameter values. Additional simulations with other variations of the parameter values (not reported) always give the same qualitative picture. Basis Basis Basis α=0.5 α=0.5 n=10 n=10 σN= 0.05 σN= 0.05 η=3 η=3 ρ=0.5 ρ=0 ρ=-0.5 ρ=0.5 ρ=0 ρ=0.5 ρ=0 ρ=0.5 ρ=0 ρ=0.5 ρ=0 uC0.0425 0.0423 0.0421 0.0427 0.0425 0.0467 0.0443 0.0425 0.0423 0.0216 0.0215 uN0.191 0.177 0.167 0.199 0.191 0.209 0.186 0.198 0.194 0.062 0.061 ΩC1.010 1.001 0.991 1.015 1.010 1.220 1.094 1.010 1.001 0.260 0.259 ΩN1.065 1.046 1.030 10.077 1.065 1.286 1.144 1.071 1.058 0.264 0.263 ωC-1.418 -1.411 -1.405 -1.422 -1.418 -1.558 -1.476 -1.418 -1.411 -0.719 -0.718 ωN-1.396 -1.391 -1.386 -1.398 -1.396 -1.534 -1.455 -1.395 -1.389 -0.713 -0.712 δ*0.144 0.165 0.184 0.132 0.144 0.132 0.159 0.150 0.156 0.099 0.102 Table 1: Numerical simulations in Gauss. Unemployment (u), union loss (Ω), and the real wage (ω) under uncoordinated (N) and co-ordinated (C) wage setting, for different values of the degree of monetary accommodation, ρ. δ* is the critical value of δ that yields equality in equation (16). δ* is calculated based on the approximation that the numerator in (16) is approximated by ∆ ∂ Ω∂ j CC j w ww ),;( ρ , where ∆=0.01, except in the simulations with η=3, where ∆=0.001.6 The top row indicates difference from base model. The second row indicates degree of monetary accommodation ρ. The base model has the following parameter values: α=1, γ=0.85, σC=1, σN=0.2, n=100, η=1.5, ω*=0. Thus, the simulations indicate that under a strict central bank, unions must be more far-sighted for co-ordinated wage restraint to be sustainable, than if the central bank is accommodating. A main reason is that the consequences of uncoordinated wage setting are more detrimental if the central bank is accommodating. One implication is that the relationship between equilibrium unemployment and central bank strictness may be non6 Note that the optimal deviation is not wN, because if a union were to deviate from an equilibrium it would be the only deviator, so other unions would not set wN to which wN is the best reply. 20 monotonic. For a given degree of co-ordination in wage setting, a more accommodating monetary regime is likely to lead to higher unemployment. However, at some point a more accommodating regime may induce co-ordinated wage restraint, leading to much lower equilibrium unemployment. Discussion Proposition 2 and the numerical simulations should not be taken to imply that the monetary regime is the only, or even the most important, factor that determine whether co-ordination of wage setting is feasible. Holden and Raaum (1991) explore this question more generally, within a similar theoretical framework, and conclude that co-ordinated wage restraint is easier to achieve if, among other things, union density (and coverage) is high (because the consequences of uncoordinated wage setting are worse), and if the political climate is stable (because co-ordination is more likely to be long lasting; in the formal model, this corresponds to the probability that a new game starts, q, is small). An alternative view of co-ordination in wage setting among individual unions is that one union acts as a wage leader vis-à-vis the rest of the labour market (Soskice and Iversen, 1998). If one union is wage leader, and all other unions follow mechanically, the situation is essentially as if wage setting were completely centralised. However, this view begs the question of why other unions follow the line initiated by the leader. If the leader is not certain that the other unions will follow the leader, this will clearly affect the wage setting of the leader. The leader may then set higher wages than it would prefer in a co-ordinated equilibrium, because of a fear that if it sets a lower wage increase, this will be exceeded by subsequent wage setters. The problems union leaders may face when they try to restrain wages is illustrated by a statement of the leader of the Norwegian unions' confederation LO (the largest union confederation, which usually acts as a wage leader in Norwegian wage 21 setting), that the total wage increase had been too high (Halvorsen, 1986). Another example of this dilemma is from Swedish wage setting in 1995: During the negotiation process in the engineering industry (which usually is the leading sector), the negotiators were surprised by an agreement in the forest industry giving a substantially higher increase than the mediators had planned in the engineering industry (Elvander, 1997). Note that a wage setting sequence where one union acts as a wage leader is not incompatible with the analysis in the present paper, as long as the concern for whether subsequent wage setters set higher wages is a restriction on the wage setting of the leader. In other words, within the framework of the present paper, co-ordinated wage restraint with one union being the leader is sustainable as long as all the other unions do not gain from deviating from the line of the leader. On this explanation, wage restraint by one union being the leader might be facilitated under an accommodating central bank, as the costs of a deviation, inducing a breakdown of the wage restraint, would be more severe. Soskice and Iversen (1998) use Germany as their leading example that a strict central bank may induce co-ordinated wage restraint. In contrast, I would argue that because Bundesbank has disciplined wage setters, their incentive to further co-ordinate wage setting has been weakened. To what extent wage setting in Germany has been co-ordinated (relative to other European countries) is not obvious (see indices in section 4 below). But relatively speaking wage setting in Germany has not been restrained: over the period 1981 to 1998, unit labour costs in German manufacturing increased by 31 percent relative to other countries (OECD Economic Outlook June 1999, Annex Table 44). The result that there are circumstances under which co-ordination in the wage setting is feasible with an accommodating central bank but not with a strict central bank, may seem surprising and unrealistic. However, it is important to keep in mind that I compare monetary regimes with a common monetary target, and under precommitment. In contrast, in Holden 22 (1991a), I show that under an exchange rate target, co-ordination of wage setting may be more difficult to achieve if the exchange rate target is not credible. Proposition 2 is based on the presumption that union behaviour can be described by a trigger strategy equilibrium. However, trigger strategy equilibria have been criticised because they are not "renegotiation proof", in the sense that if a deviation were to occur, the unions would nevertheless have an incentive to co-operate in the following period. If unions were to anticipate that they would co-operate even after a deviation, a deviation would involve no costs, and in this case the co-operation would not be sustainable in the first place (Farrell and Maskin, 1989). This argument may seem compelling in an analysis of perfectly rational players. However, renegotiation proofness is probably not an appropriate requirement in models of union behaviour. It would be very difficult for the leadership of a union to suggest wage moderation if another union had defected the previous year. Arguments like "Last year was a mistake, but they promise that this year…." will probably have trouble convincing the rank and file. Thus, a deviator cannot be certain that a deviation will be "forgotten". I find it more realistic to assume that whether or not it is possible to resume an agreement after a breakdown is determined by other aspects, rather than being inferred directly from assumptions concerning union behaviour. The best would be to have an explicit model of these other aspects, but in the absence of this, I let this be determined by an exogenous random variable. The analysis can also be criticised from the opposite angle, that they assume too much rationality on behalf of the unions. It is true that in the real world, unions that compete for members, or union leaders that are pushed by a militant membership, may have limited scope for wage moderation. Yet it would be both unrealistic and in contrast to the standard approach in economics to assume from the outset that agents like unions neglect the future 29 Cukierman, A. (1992). Central Bank Strategy, Credibility and Independence: Theory and Evidence. Cambridge, MA: The MIT Press. Cukierman, A. and F. Lippi (1999a). Central bank independence, centralization of wage bargaining, inflation and unemployment: Theory and some evidence. European Economic Review 43, 1395-1434. Cukierman, A. and F. Lippi (1999b). Labor markets and monetary union. A strategic analysis. Mimeo, Tel-Aviv University. Elvander, N. (1997). The Swedish Bargaining System in the Melting Pot. In The Swedish Bargaining System in the Melting Pot, by N. Elvander and B. Holmlund. Arbetslivsinstitutet, Solna, Sweden. Farrell, J. and E. Maskin (1989). Renegotiation in repeated games. Games and Economic Behaviour 1, 327-360. Flanagan, R.J. (1999). Macroeconomic performance and collective bargaining: An international perspective. Journal of Economic Literature XXXVII, 1150-1175. Friedman, J.W. (1986). Game Theory with Applications to Economics. Oxford University Press. Golden, Miriam; Lange, Peter; and Michael Wallerstein. 1997. "Union 30 Centralization Among Advanced Industrial Societies: An Empirical Study." Dataset available at http://shelley.sscnet.ucla.edu/data. Version dated 11/03/98. Grüner, H.P. and C. Hefeker (1999). How will EMU affect inflation and unemployment in Europe? Scandinavian Journal of Economics 101, 33-47. Guzzo, V. and A. Velasco (1999). The case for a populist central banker. European Economic Review 43, 1317-1344. Halvorsen, T. (1986). "Address to meeting of the Union of Norwegian Electricians and Power Station Workers" (Foredrag på landsmøte i Norsk Elektriker og Kraftstasjonsforbund), in Norwegian, Oslo, 18.8.96. Holden, S. (1991a). Exchange rate policy and wage formation in an economy with many trade unions. European Economic Review 35, 1543-1557. Holden, S. (1991b). Economic policy in an economy with local and central wage bargaining. Memorandum 8, Department of Economics, University of Oslo. Holden, S. (1998). Wage drift and the relevance of centralised wage setting. Scandinavian Journal of Economics 100, 711-731. Holden, S. (1999). Wage setting under different monetary regimes. Mimeo, Department of Economics, University of Oslo. http://www.uio.no/~sholden/#wp 31 Holden, S. og O. Raaum (1991). Wage moderation and union structure. Oxford Economic Papers 43, 409-423. Horn, H. and T. Persson (1988). Exchange rate policy, wage formation, and credibility, European Economic Review 32, 1621-1636. Iversen, T. (1998). Wage bargaining, hard money, and economic performance: Theory and evidence for organized market economies. British Journal of Political Science, January. Iversen, T. (1999). Contested Economic Institutions. The Politics of Macroeconomics and Wage Bargaining in Advanced Democracies. Jensen, H. (1997). Monetary policy co-operation may not be counter-productive. Scandinavian Journal of Economics 99, 73-80. Moene, K.O., M. Wallerstein and M. Hoel. (1993). Bargaining structure and economic performance. I Flanagan, R.J., K.O. Moene and M. Wallerstein (eds.) Trade Union Behaviour, Pay-Bargaining, and Economic Performance. Oxford University Press. Nickell, S. (1997). Unemployment and labor market rigidities: Europe versus North America. Journal of Economic Perspectives 11, nr 3, 55-74. OECD (1997). Employment Outlook July 1997, Paris. 32 Osborne, M. J. and A. Rubinstein (1990). Bargaining and Markets. San Diego, California: Academic Press. Skott, P. (1997). Stagflationary consequences of prudent monetary policy in a unionized economy. Oxford Economic Papers 49, 609-622. Soskice, D. and T. Iversen (1998). Multiple wage-bargaining systems in the single European currency area. Oxford Review of Economic Policy 14, 110-124. Soskice, D. and T. Iversen (1999). The non-neutrality of monetary policy with large price or wage setters. Forthcoming in Quarterly Journal of Economics. Vartiainen, J. (1999). Relative wages in monetary union and floating. Mimeo, Labour Institute for Economic Research, Helsinki, Finland. Wallerstein, M. (1999). Wage-setting institutions and pay inequality in advanced industrial societies. American Journal of Political Science 43, no 3. Weitzman, M.L. (1985). The simple macroeconomics of profit sharing. American Economic Review 75, 937-953. Wibaut, Q. (1998a). Money as a co-ordination device. Mimeo, Department of Economics, UCL. Wibaut, Q. (1998b). Exchange rate regime and wage indexation in a two sector openeconomy. Mimeo, Department of Economics, UCL. 33 Figure 2: Sources: Iversen (1999), tables 1.2 and 1.3. Aus is Austria, NL the Netherlands, and NZ is New Zealand. w-p u p rices' p rices wa g es wa g es' ω* u N ' u N Figure 1: A stricter monetary regime increases the wage elasticity of employment, shifting the wage curve down and the price curve up, and decreasing the equilibrium level of unemployment. Centralisation of wage setting and monetary regime 1984-93 15 OECD countries 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.2 0.4 0.6 0.8 Centralisation of wage setting TI Hard currency index TI Den Nor Swe Fin Aus NL Ger Jap Swi Bel Ita Bri Fra US Can Centralisation of wage setting and monetary regime 1983-94; 15 OECD countries 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.2 0.4 0.6 0.8 Centralisation of wage setting TI I Hard currency index T Den Nor Swe Fin Aus NL Ger Jap Swi Bel Ita Bri Fra US Can 34 Figure 3: Sources: CBI, Cukierman and Lippi (1999a), table 1; co-ordination in wage setting 1990, OECD (1997) table 3.3. A is Australia. Finland and Sweden are at the same data point. Figure 4: Sources: Wage setting, Golden-Wallerstein-Lange Labor & Political Data. Hard currency index, Iversen (1999), tables 1.3. Co-ordination of wage setting and central bank independence, 1990, 19 OECD countries 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 01234 Co-ordination of wage setting OECD Central bank independence C&L US Can Bri NZ Ita Bel Spa Fra Fin A Por NL Den Swi Nor Jap Aus Ger Centralisation of wage setting and monetary regime 15 OECD countries 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 012345 Centralisation of wage setting GWL Hard currency index TI Jap Swi Ger NL Bel Aus Fra US Can UK Fin Swe Nor Ita Centralisation of wage setting and monetary regime 15 OECD countries 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 012345 Centralisation of wage setting GWL Hard currency index TI Jap Swi Ger NLBel Aus Fra US Can UK Fin Swe Nor Ita Den Co-ordination of wage setting and central bank independence, 1990, 19 OECD countries 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 01234 Co-ordination of wage setting OECD Central bank independenc e C&L US Can Bri NZ Ita Bel Spa Fra Fin/Swe A Por NL Den Swi Nor Jap Aus Ger Co-ordination of wage setting and central bank independence, 1990, 19 OECD countries 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 01234 Co-ordination of wage setting OECD Central bank independenc e C&L US Can Bri NZ Ita Bel Spa Fra Fin/Swe A Por NL Den Swi Nor Jap Aus Ger 35 Figure 5: Sources: Wage setting, Golden-Wallerstein-Lange Labor & Political Data. CBI, Cukierman and Lippi (1999a), table 1. Finland and Sweden are at the same data point. Centralisation of wage setting and monetary regime 16 OECD countries 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 012345 Centralisation of wage setting GWL Central bank independenc e C&L Swi Ger Aus Den Fra Bel Ja p US Can UK N L A Fin Swe Ita N or Centralisation of wage setting and monetary regime 16 OECD countries 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 012345 Centralisation of wage setting GWL Central bank independenc e C&L Swi Ger Aus Den Fra Bel Ja p US Can UK N L A Fin Swe Ita N or Steinar Holden:Monetary regime and the co-ordination of wage setting. Arbeidsnotat/Working Paper 2000/1 KEYWORDS: Wage setting Co-ordination Equilibrium unemployment Monetary regime Monetary union 20832