The Welfare and Employment Effects of Centralized Public Sector Wage Bargaining

• Date: 01 April 2017
• Author: GABRIELE CARDULLO
• Published date: 01 April 2017
• DOI: http://doi.org/10.1111/jpet.12200

THE WELFARE AND EMPLOYMENT EFFECTS OF CENTRALIZED

PUBLIC SECTOR WAGE BARGAINING

GABRIELE CARDULLO

University of Genoa

Abstract

In many countries, the government pays almost identical nominal

wages to workers living in regions with notable economic disparities.

By developing a two-region general equilibrium model with endoge-

nous migration and search frictions in the labor market, I study the dif-

ferences in terms of unemployment, real wages, and welfare between

a regional wage bargaining process and a national one in the public

sector. Adopting the latter makes residents in the poorer region better

off and residents of the richer region worse off. Private sector employ-

ment decreases in the poorer region and it increases in the richer one.

Under some conditions, the unemployment rate in the poorer region

soars.

1. Introduction

In many countries, public sector wages are very similar in nominal terms for employ-

ees of regions with different private sector productivities and costs of living. The spa-

tial distribution of public sector earnings is very compressed in the five largest EU

economies (Germany, France, the UK, Italy, and Spain) (see Elliott, Mavromaras, and

Meurs 2007).1Albeit in a weaker form, even the U.S. federal government regional pays

are substantially unaffected by local market conditions, while different is the case for

state and local public employees (Katz and Krueger 1991). Among the several explana-

tions for the greater compression of public sector wage structures compared to private

1In Italy, Spain, and Germany this is accompanied by a pronounced income disparity between re-

gions (see, respectively, Dell’Arringa, Lucifora, and Origo 2007; Garcia-Perez and Jimeno 2007; and

Heitmueller and Mavromaras 2007). For France, see Meurs and Edon (2007). Some of these papers

look at real wage spatial distributions but, because they use a national price index, their results also

apply to nominal pay variations.

Gabriele Cardullo, Department of Economics, University of Genoa, Genoa, Italy (cardullo@

economia.unige.it).

I thank the associate editor, Giorgio Brunello, Maurizio Conti, Bruno Decreuse, Thomas Gall, An-

drea Ichino, Christian Holzner, Andrey Launov, Claudio Lucifora, Fabien Postel-Vinay, Bruno Vander

Linden, Klaus W¨

alde, two anonymous referees, and the participants to the EEA conference in Toulouse,

the LAGV conference in Marseille, the Mainz Workshop in Labour Economics, the SAM conference in

Bristol, and the “Brucchi Luchino” conference in Rome for useful comments and suggestions.

Received April 28, 2014; Accepted October 1, 2015.

C2016 Wiley Periodicals, Inc.

Journal of Public Economic Theory, 19 (2), 2017, pp. 490–510.

490

Welfare and Employment 491

sector ones, the highly centralized structure of the public pay systems is one of the most

relevant (Elliot et al. 2007).2

In times in which many governments face the twofold challenge of improving the

efficiency of the public sector and restraining spending, the poor responsiveness of civil

servants to local labor market conditions has become an important issue to deal with

for economists and policy makers.3

This paper enters the debate by contrasting a centralized public sector pay system

with a regional one and looking at the differences in terms of welfare, employment, and

real earnings.

In both scenarios, public sector employment and wages are determined through

efficient bargaining between a public authority that aims to maximize revenues net of

production costs and a union.

The two frameworks differ in that under the national bargaining both public sector

vacancies in each region and the common wage are bargained over by the central gov-

ernment and a union whose members are all the civil servants in the country, whereas

in the regional bargaining process the actors are the local government and a union that

only cares about the utility of public sector employees of the region.

These two different public sector wage settings are nested into a two-region general

equilibrium model in which private tradable and public nontradable goods are pro-

duced and the labor market exhibits search and matching frictions. Migration between

regions is endogenous and unemployed workers are free to apply either for a private

sector job or for a public sector one. Individual bargaining is assumed in the private

sector.4The model is analytically tractable in steady-state and the conclusions on wage,

employment, and welfare are obtained via comparative statics.

The first result of the paper is perhaps not surprising. A centralized public sector

pay system is a redistributive tool that shifts resources from the richer regions to the

poorer ones. What is perhaps less expected is the extent of such a redistribution, as it

does not involve only civil servants but all workers. Under centralized bargaining, both

real wages and the welfare of all residents in the richer region decrease, whereas the

opposite occurs in the poorer region.

The rationale behind these results is the following. Under a national negotiation,

the public sector union must account for the marginal productivities and fall-back po-

sitions of workers of all regions. So, it ends up accepting a nominal wage that is lower

(resp. higher) than the one that would accrue to employees in the richer (resp. poorer)

region under a decentralized bargaining scheme. This has an impact on labor supply.

In comparison with a regional bargaining system, public sector jobs become more entic-

ing for workers in the poorer region while more residents of richer region will apply for

private jobs. In turn, the difficulties in filling public sector vacancies in the richer region

raises the production cost of the public nontradable good and, in turn, the cost of living.

Therefore, in the richer region real wages and welfare are lower under a centralized pay

system than with a regional bargaining scheme. On the contrary, in the poorer regions

2The differences between private and public wage structures may be the result of worker self-selection.

However, recent research shows that the wage compression associated with public sector pay scales has

a large causal component (Melly and Puhani 2013). The higher degree of unionization in public sector

also plays a role. On the role of unions in compressing wage inequality, see Kahn (2000) and Lemieux

(1998).

3See the discussion in the next section.

4To account for the differences between the U.S. and the European labor markets, in the working

paper I consider both the case of individual bargaining and the case of decentralized collective negoti-

ation in the private sector. The analytical results do not change. See Cardullo (2015).