Is Brazilian society's perception of the difference between public and private salaries valid?

[vc_row][vc_column][vc_column_text]

Responsible researcher: Eduarda Miller de Figueiredo

Article title: SALARY DISTRIBUTION AND THE PUBLIC-PRIVATE DIFFERENTIAL IN BRAZIL

Article authors: Walter Belluzzo, Francisco Anuatti-Neto and Elaine T. Pazello

Location of intervention: Brazil

Sample size: 110,433 observations

Sector: Job Market

Type of intervention: Interview on perception of the political system

Variable of main interest: Salary

Assessment method: Others- Quantile Regression

The perception of Brazilian society in the year of the study was that salaries offered in the public sector were lower than those in the private sector, which were accepted because public servants considered the expectation of retiring with full benefits an element that would compensate for such a difference. According to Foguel et al (2000), observing the characteristics of each sector, the wage differential between sectors is reduced, but remains significantly favorable in favor of the public sector. 70% of the total salary differential corresponds to the level of education.

Given this, the research aimed to analyze the conditional distribution of salaries, reconciling the perception of Brazilian society about salaries in the private sector being higher than those in the public sector.

Assessment Context

The research used data from the National Household Sample Survey (PNAD) to observe the salary differential between the public and private sectors. During the article, the authors brought a comparison of data between the years 1995 and 2001.

Data between the two years already mentioned showed that there was an increase in older workers in the public sector, that is, a more experienced workforce. An educational advance was also observed among employees in the private sector, with workers with 12 years or more of education accounting for 35.5% in the public sector and 10.8% in the private sector.

An increase in the salary differential between 1995 and 2001 was also noted, in which the gross salary had a difference of 0.28 in 1995 and increased to 0.40 in 2001 and the standardized salary ^[1] changed from 0.45 to 0.57 between the years.

Methodology Details

The study used the PNAD database, including all individuals living in the urban area of ​​the country, aged 16 or over and who had an occupation at the time of the interview. Among those employed, only those employed in non-agricultural, civil activities and working between 20 and 70 hours per week were selected. This resulted in a total sample of 110,433 observations, of which 17,028 (15.42%) are public sector workers and 93,405 (84.58%) are from the private sector.

To measure the average salary differential between sectors, the authors followed the literature by using the conditional average model with the logarithm of the salary, a vector of individual characteristics, in addition to a binary variable to indicate whether it is in the public sector or not. However, this approach by Foguel et al (2000) imposes the hypothesis that the variables only affect the position of the salary distribution and not the dispersion or shape.

In view of this, the article discussed here sought to use a different approach, aiming to enable covariates to affect the position, dispersion and format of wage distribution. To do this, quantile regression methods are used, where the basic model provides the measurement of the salary differential directly, with the effect of the covariates fixed for each quantile. While in the general model, the measurement of the wage differential occurs through a counterfactual analysis, following the work of Foguel et al (2000).

Furthermore, the study estimated independent models for each region of Brazil: north, northeast, center-west, southeast and south. Each model has a sequence of quantiles between 0.05 and 0.95, with intervals of 0.05, totaling a total of 38 models per region.

Results

The results for the standardized salary differential between the private and public sectors at the municipal level demonstrate that the differential is favorable to public employees and that this occurs in all regions. However, this only occurs in the lower tail of the salary distribution, that is, as the salary increases, the advantage decreases and can even become negative. Thus, it is suggested that individuals with low wages have a greater advantage in relation to the private sector. The regressions for the gross salary at the municipal level demonstrate that the advantage of the public sector disappears across the entire salary distribution, except for the North and Northeast regions.

At the state level, the results continue to demonstrate an advantage of the public sector over the private sector in practically the entire salary distribution. However, when observing the results for the gross salary of the state public sector, a shift to the left of the curves was found, however, the shift is much smaller in magnitude compared to the municipal public sector.

In relation to the federal level, the results demonstrate that the difference is always significantly positive and greater than those observed at the municipal and state levels. However, it was noted that the salary differential tends to be smaller for the upper tail of the distribution, except for the South region.

The counterfactual analysis was carried out only at the national level, with no separate results for each region of Brazil.

Figure 1: Marginal densities x counterfactuals - Standardized x Raw

Figure 2: Differentials marginal distribution x counterfactual

The counterfactual analysis was carried out only at the national level, with no separate results for each region of Brazil. The results demonstrated that when moving from standardized to gross wages, the counterfactual densities shift mass a little to the right, while marginal densities shift mass a little to the left. This suggests that there is a reduction in the level of differentials, however, it remains in the same quantile of the salary distribution.

The results obtained, therefore, confirm that the perception of Brazilian society that salaries in the public sector are lower than those in the private sector is valid for the highest salaries at the state and municipal levels, especially when observing gross salaries without standardization of working hours. However, for federal civil servants the differentials are positive for all quantiles of the salary distribution, gross and standardized. A possible answer provided by the authors for this perception about salary differentiation between sectors is that, when observed from a political economy perspective, the groups that have the highest salaries also have the greatest capacity for expression and influence over public opinion.

Public Policy Lessons

The common sense that public employee salaries are lower than those of private sector employees is only valid at the state and municipal levels. This perception is refuted for the salaries of public servants at the federal level. This may be a consequence of the fact that such employees have greater influence over society's opinion, making it possible to encourage the common sense that public salaries are lower than private ones.

Reference
BELLUZZO, Walter; ANUATTI-NETO, Francisco; PAZELLO, Elaine T. Wage distribution and the public-private differential in Brazil. Brazilian economics magazine, v. 59, no. 4, p. 511-533, 2005.

^[1] Standardized monthly salary for a 40-hour work week.

[/vc_column_text][/vc_column][/vc_row]