Empirical work suggests the presence of a public sector wage premium. This paper investigates the theoretical reasons for the presence of such a premium. The results of the paper are consistent with the higher premium paid to women and with the fact that the premium decreases with skills. The key insight of the paper is that job security undermines the incentive to work hard and forces the public sector to pay higher wages. One implication of the paper is that the public sector wage premium can be used as an indicator of inefficiency of the public sector.
In many countries the public sector offers a pleasant and noncompetitive work environment and a level of job security and fringe benefits (Bellante and Long, 1981 and Poterba and Rueben, 1998) that cannot be matched in the private sector. Furthermore, Quinn (1982) finds that private sector workers tend to be subject to more work-place disamenities (pace of work, supervision, and danger) than their public sector counterparts. Given the less stressful and safer work environment, we would expect that the public sector should pay wages that are substantially lower than the wages paid in the private sector. This is not the case. There is mounting evidence that wages paid in the public sector are not lower and are often significantly higher than wages paid in the private sector (Gregory and Borland, 1999, Panizza and Qiang, 1999, Panizza, 2001). In essence, there is reason to suspect that a "public sector wage premium" might exist.
The existing theoretical literature on the public sector wage premium consists mainly of verbal explanations and emphasizes the role of the higher level of unionization and the soft budget constraint faced by the public sector (Ehrenberg and Schwarz, 1986, and Gregory and Borland, 1999, are two excellent surveys). Holmlund (1993) presents a formal model in which the public sector wage premium arises from the fact that, while the private sector unions fully internalize (trough the decrease in employment) the cost of any increase of the wage bill, the public sector unions are able to discharge part of the burden on the private sector.
This paper offers an alternative explanation for the public sector wage premium. The model of Section 3 focuses on the fact that the public sector is not able to offer its workers the package of incentives available to the private sector. The paper adopts an efficiency wage model where the main incentive for working hard is the probability of being fired if caught shirking. In this setting, the higher level of firing costs faced by the public sector tightens the no-shirking constraint for the public sector and leads to higher wages. Hence, the higher job security offered by the public sector, instead of being compensated by lower wages, is the main cause of the public sector premium. This paper does not contradict but rather complements the political considerations raised by the previous literature. Here the political power of the public sector unions is reflected by the public sector's relative difficulty (with respect to the private sector) in firing workers who shirk. In countries with a large public sector, unstable political majorities, and weak institutional systems, public sector unions will have a strong influence on policy-makers and force the latter to pay higher rents to public sector workers.
An implication of the paper is that a high wage premium should be associated to an inefficient public sector. The idea that the public sector premium can be used as an indicator of inefficiency of the public sector goes against the common wisdom --mainly based on the experience of one country: Singapore-- that higher public sector wages would decrease corruption and therefore increase the efficiency of the public sector (Nunberg and Nellis, 1995). Using a panel of 31 countries Van Rijckeghem and Weder (2001) do not find a strong correlation between public sector wages and corruption.(1) Rauch and Evans (2000) survey the relationship between the bureaucratic structure and performance of 35 less developed countries. They find a positive correlation between the quality of bureaucracy on the one hand and meritocratic recruitment and promotion systems on the other but do not find any clear correlation between the former and the level of public sector pays. La Porta et al. (1999) find a negative correlation between public sector wages and various indicators of government efficiency and conclude that countries in which bureaucrats have much power they collect both high wages and significant bribes. In a sample of 17 Latin American countries, Panizza (2001) finds no correlation between public-private wage differentials and bureaucratic quality but does find that meritocracy in the public sector is associated with bureaucratic quality.
These results are consistent with the idea of this paper: high public sector wages do not derive from a mechanism that tries to attract the best workers to the public sector; they rather arise from the government's inability to solve its principal agent problem. Higher pays do not seem to be the best method to provide public sector workers with the right set of incentives. Meritocratic recruitment and promotions are the policies that offer the right incentives and have a positive effect on the quality of the public sector.
The paper is organized as follows: Section 2 discusses the empirical evidence regarding the public/private wage differential. Section 3 presents the basic model that generates the public sector wage premium. Section 4 shows that the presence of a premium is not affected by a mechanism where the private sector can skim the labor force and hire the best workers. Section 5 concludes.
Is There a Public Sector Premium?
This section reviews the existing empirical literature on the public/private wage differential. While there exists a wide literature testing for the presence of a public sector wage premium in the United States (Ehrenberg and Schwarz, 1986, survey 23 studies that support the presence of a public sector wage premium) considerably little work has been done at the cross-country level. Blanchflower (1996) computed the public sector wage premium for 15 OECD countries. His results are reported in Table 1. According to Blanchflower's estimations, 11 countries have a positive and statistically significant public sector wage premium and only one (Norway) has a statistically significant public sector wage penalty. In most countries, public sector employees earn between 4 and 13 percent more than workers with similar characteristics employed in the private sector (Japan, however, has a 21 percent public sector premium).
The results of Table 2 were obtained by running wage regressions on household surveys data for 14 Latin American countries (Panizza and Qiang, 1999 and Panizza, 2001). In this sample of less developed countries we also find that, on average, the public sector pays more than the private sector. Out of 14 countries, 4 show a public sector penalty and 8 a public sector premium (the coefficient is not statistically significant for the two remaining countries). It should be recognized that as the estimations reported in Tables 1 and 2 do not distinguish males from females, they may suffer from aggregation bias and just proxy for the lower gender gap in the public sector (Gregory and Borland, 1999, Gornick and Jacobs, 1998, and Bardasi, 1998).
The data used in the estimations of Tables 1 and 2 do not include non-wage benefits. Hence, the results are likely to underestimate the rent enjoyed by public sector workers. Using U.S. data, Quinn (1982) finds that public sector employees receive pension contributions that are 30 to 50 percent greater than the pension contributions paid by private employers. Braden and Hydland (1993) estimate that one-third of the raw differential in total labor cost between public and private sector employees can be attributed to non wage benefits. Brunelli and Cox (1992) estimate that only if a federal...