)JWfgplJOA 2iLL * ... -. . } < .- £.-. I....-. -t.- .........-f.............. .... . .- .... . .U -Lt...... ... ........ ... -F - :. - .. . .... .... ....... ..- ......~~~~~~~~~~~~~~~~~~~~~~~~..... ............. ... ............ . ........ ....... ...... ........... . . . .... .W.. .... .... .. ..... ....... .. ~~~~~~~~~~~~~~~~~~~~~~~~~~~.. ,L.. ........ -.-.................. . ...tl~~~~~~~~~~~~~~~~~~ .....-...... S~~~~~~~~~~~~~~~~~~~~~~~~~. ............ ....... .- - :- :: ~~~~~~~~. ........ .......... ............... , ,, - .. ... . . ... . . ... .... . . . . --.- -.- - - ~~~~~~~~.................. .. - -... llZOi'- UW SE~OId- 'cllqnd jo pedoldI~1DVVJOqD1 ...~~~~~~~~~~~~~~~~~~... . . .. . . ..* .' ''.'..'.. . : v.'. ' ."" ' , ."'" . .'."."* .. ..... ' ' * i ".'. ' .' '~~~~.. ." " ..''' .' . . ... ..... * *,..'.,''..~~~~~~~~~. ....":. SeON QUILUQUOD uoioau ea!la§v U!i I 11190 AJIUNO . .. . . . . . .. . .. . .. . .. . . . . .. . . . . . Country Dept. I Latin America and the Caribbean Region Economic Notes Labor Market Prospects of Public Employees in Brazil: An Empirical Evaluation* Ricardo Paes de Barros Indermit S. Gill Miguel Foguel Rosane Mendonga June 17, 1997 The World Bank *This background paper was prepared for the World Bank report "Brazil: Stabilization, Fiscal Adjustment and Public Employment Reform", under the immediate supervision of Gautam Datta, Senior Economist, LAlCI. Ricardo Paes de Barros, Miguel Foguel, and Rosane MendonQa are at Instituto de Pesquisa Econ6mica Aplicada (Rio de Janeiro); Indermit Gill is at the World Bank (Washington). The views expressed here are those of the authors and should not be attributed to either the World Bank or IPEA. We would like to thank Homi Kharas, Chief, LAlC1, for useful comments. I Labor Market Prospects of Public Employees in Brazil I 1. INTRODUCTION As part of the adjustment process in Trends in employment, earnings, and the Brazilian public sector, it is expected that other indicators are analyzed using Pesquisa the level of public employment will decline. Nacional por Amostra de Domicilios In this paper we take this possibility as given, (PNAD) surveys for 1981, 1985, 1990, without speculating on its likelihood or 1992, 1993, and 1995. The latest PNAD desirability. The focus here is on evaluating survey (1995) is examined in greatest detail. its potential impact on the welfare of those This paper has three interrelated parts. The currently employed in the public sector. first constructs the profile of employment in the public sector (civil service and public The general objectives of this paper enterprises), and documents private and are first, to obtain a profile of public public employment in a representative employees and, second, to evaluate the sample of states, by type of worker (e.g., labor market prospects of these workers if age, sex, education, and occupation). The they were to move to the private sector. In second estimates the wage gap between other words, we answer questions such as: public and private employment by type of what is the composition of employment in worker and labor market characteristic (e.g., Brazilian public sector? how difficult would region of residence). The third and final part it be for a worker displaced from the public is devoted to investigate other dimensions of sector to find a similar job in the private the labor market, especially the aspect of sector? and how does the cost of adjustment permanency of jobs (e.g., as proxied by vary with the type of worker displaced? expected tenure). A statistical annex contains detailed tables; tables with even greater To answer these questions, this paper detail (e.g., state-level employment and examines employment and average earnings earmings trends) are available from the in the private and public sectors, the growth authors upon request. of the public sector wage bill in recent years, and the wage premium earned by public All the analysis was conducted both workers over their private sector for the entire country and for selected states. counterparts. It decomposes this gap into a In particular, public and private employment component due to differences in worker- in following states were examined: specific and locational attributes and that Pernambuco, Bahia (Northeast), Rio de which can be considered a purer premium. It Janiero, Sao Paulo, Minas Gerais also examines public-private differences in (Southeast), Rio Grande do Sul (South) and other job characteristics, primarily job the Federal District. security. Finally, we provide imputations of differences in pension levels and duration for public servants and private sector workers. 2 Labor Market Prospects of Public Employees in Brazil 2. EMPLOYMENT IN PUBLIC AND PRIVATE SECTORS 2.1 All-Brazil, 1981-1995 (2.2 million in public administration, more than half at the municipal level, 0.6 million in PNAD surveys indicate that total the judiciary, legislature and military, and 2.9 employment in Brazil was 62.5 million in million in the education and health sectors). September 1995 (see Table 1). Of this, The enterprise sector had 56.9 million public sector employment - other than public workers (16.1 million in agriculture, 8.4 enterprises - was approximately 5.6 million million in industry, and 28.4 million in services). Table 1 Sectoral Shares in Employment, 1981-1995 All Brazil 1981 1985 1990 1992 1993 1995 ...................................................................................................................................................................................................................................... EMPLOYMENT (in percent of total) Public Employment 7.2 8.4 9.4 9.0 9.3 9.0 Direct Administration 2.9 3.3 3.8 3.6 3.6 3.5 Federal 0.5 0.4 0.3 0.3 0.3 0.3 State 1.4 1.5 1.S 1.4 1.4 1.3 Municipal 1.0 1.4 2.0 1.9 1.9 1.9 Education & Health 3.9 4.2 4.6 4.5 4.7 4.6 Judiciary & Legislative 0.4 0.4 0.4 0.4 0.4 0.5 Military 0.6 0.5 0.6 0.5 0.4 0.4 Private Employment* 92.8 91.6 90.6 91.0 90.7 91.0 Agnculture 28.5 27.0 22.4 27.8 26.9 25.7 Industry 17.0 17.4 17.4 14.3 14.3 13.4 Services 38.3 40.8 44.4 42.4 42.9 45.4 Construction 8.1 5.7 6.1 6.1 6.4 6.1 Other private 0.3 0.5 0.4 0.5 0.4 0.4 Total (percent) 100 100 100 100 100 100 Total (millions) n.a.** n.a.** n.a.** 58.8 59.9 62.5 * Enterprise employment includes employment in public enterprises. ** Total employment figures before 1991 are not accurately estimated. Source: __Pes ~uisa Nacionalpor Amostra de Domicilios (selectedyer) The share of the "public sector" - stayed constant at about 0.27 million. civil servants and the military - as defined Judicial, legislative, education and health above rose from about 7% to 9% of the total sectors registered increases in employment between 1981 and 1990, and has stayed between 1992 and 1995. Judicial and above 9% since then. Between 1992 and legislative staff increased from 0.26 to 0.32 1995, the number in public administration million, and employment in education and increased from about 2.07 million to 2.16 health increased from 2.66 to 2.85 million. rnillion - almost all of this increase was accounted for by an increase in municipal In the enterprise sector, agricultural employment. State employment fell by 1%, employment fell, employment in industry and and federal employment rose about 7% over construction stayed constant, and these two years. The size of the military Labor Market Prospects of Public Employees in Brazil 3 employment in services (especially personal public enterprises and autonomous agencies services) rose between 1992 and 1995. (fundacoes). In contrast to the private sector, almost all government employees are Using the class of worker distinction, either public servants or have a signed work the share of the public sector was about 12% card. Workers in the public sector who do during the period 1992-1995 (see Table 2). not fall into this category form about 1.5% This definition thus includes workers in of total employment in the country Table 2 Share of workers, by contract type (percent), 1995 All Brazil Public Non Public Servants Total Share ..................................................................................................................... ....................................................................................................................... ........ With Card Without Card Public Sector Workers 11.5 Federal 0.9 0.7 0 1 1.7 State 3.5 1.1 0.4 5.0 Municipal 1.9 1.6 0.9 4.4 Military 0.4 0.4 Private Sector Workers 88.5 Salaried 28.9 21.1 50.0 Selfemployed 23.6 23.6 Employers 4.1 4.1 Unpaid 10.8 10.8 Total 6.7 32.3 61.0 100 Source: Pes uisa Nacionalp.or Amostra de Domicilios, 1995. 2.2 Employment in Selected States, 1992-1995 For the six states and the federal district, while the largest share of We analyzed employment patterns in government employment is of public servants the federal district and six states (Bahia, (estatutarios), many government workers do Minas Gerais, Pernambuco, Rio de Janeiro, have contracts without guaranteed tenure. Rio Grande do Sul, and Sao Paulo) to This ratio ranges from 3% of total examine if there are considerable state-level employment in Pernambuco, to more than differences in these patterns and trends. 6% in Bahia and the federal district. Other than in Rio de Janeiro, the public Government employees without a signed sector (as defined above) is about 8% of working card are about 1% overall, but are total employment. Rio de Janeiro's ratio is close to 2% of total employment in Bahia about 11%, and the federal district's twice and the federal district. Much of this that. Public administration as a ratio of total "informality" of government employment is employment is highest in Pemambuco, while at the municipal level. In the case of Rio has the highest ratios for education and municipal employees, the two northeastern health. In the private sector, agriculture has states appear to have a large number without the lowest share in Rio and the federal a signed working card. district, and the highest in the northeastern states. Sao Paulo and Rio Grande do Sul There is also considerable variation have the highest ratios for industry, while across states in the degree of formality of Rio, the federal district and Sao Paulo have private employment across states. the highest ratio for services. These ratios Northeastern states have largest proportions did not change much between 1992-1995. of private workers who fall into this 4 Labor Market Prospects of Public Employees in Brazil category. While the share of workers with a is 15-20% for the northeastern states. Minas signed card in private employment is 35-50% Gerais, with a ratio of 30%, is in the middle. for southern states and the federal district, it Table 3 Sectoral Shares in Employment, 1995 Selected States Distrito Bahia Pernam- Minas Rio Grd Rio de Sao Federal buco Gerais do Sul Janeiro Paulo .......................................................................................................................................................................................................................................... Public Employment 22.6 7.8 8.4 8.4 8.3 11.3 8.3 Direct Administration 9.2 3.1 3.7 3.2 3.1 3.5 3.4 Federal 3.9 0.2 0.2 0.2 0.2 0.6 0.2 State 5.2 1.0 1.5 0.9 1.1 1.5 1.2 Municipal 0.1 1.9 2.0 2.1 1.8 1.4 2.0 Education & Health 7.0 4.2 3.9 4.6 4.0 5.4 4.1 Judiciary & Legislative 3.4 0.4 0.4 0.4 0.4 0.6 0.5 Military 3.0 0.1 0.4 0.2 0.8 1.8 0.3 Private Employment* 77.5 92.2 91.6 91.6 91.7 88.7 91.7 Agriculture 2.4 44.7 32.6 29.9 29.1 4.2 8.6 Industry 4.5 6.5 8.7 11.6 16.2 12.7 20.3 Services 62.3 35.1 44.1 42.6 41.0 62.9 55.9 Construction 8.0 5.4 5.2 7.0 4.7 8.1 6.8 Other private 0.3 0.4 1.0 0.2 0.8 0.9 0.2 Total (percent) 100 100 100 100 100 100 100 Total (millions) 0.75 5.63 3.15 7.76 5.01 5.70 15.10 * Private employment includes employment in public enterprises. ** Total employment figures before 1991 are not accurately estimated. Soure: es ia Ncionl oAmostra de Domicilios, 1995. Table 4 Share of workers, by contract type (percent), 1995 Selected States Distrito Bahia Pernam- Minas Rio Grd Rio de Sao Paulo Federal buco Gerais do Sul Janeiro .......................................................................................................................................................................................................................................... Government* Public service 21.0 4.0 7.2 7.5 6.3 9.9 5.9 Others With signed card 6.0 4.2 1.8 2.3 3.7 3.6 3.5 Without signed card 1.7 2.2 1.3 1.2 0.8 0.8 0.8 Private Employment** With signed card 29.6 14.5 18.2 27.7 31.4 39.9 43.4 Without signed card 18.2 25.0 22.6 26.1 14.9 19.7 19.0 Selfemployed 16.9 28.8 29.3 22.6 23.9 20.7 18.5 Employers 3.8 2.5 2.7 4.8 4.9 3.9 4.8 Unpaid 2.7 18.8 16.8 8.0 14.0 1.4 3.9 Total (percent) 100 100 100 100 100 100 100 Total (millions~) ........... 0.75 5.63 3.15 7.76 5.01 5.70 15.10 * Includes federal, state, municipal, and military personnel. ** Private employment includes employment in public enterprises. Source: Pesauisa Nacional por Amostra de Domicilios, 1995. Labor Market Prospects of Public Employees in Brazil 5 3. EARNINGS AND WAGE BILL IN PUBLIC AND PRIVATE SECTORS 3.1 Earnings, 1981-1995 judicial and legislative employees, and the largest fall was for municipal employees. Average real monthly earnings by sector were computed using PNAD surveys In absolute terms, and unadjusted for and the INPC deflator. Between 1981 and worker characteristics, earnings were highest 1995, real monthly earnings for increased for judicial and legislative workers in 1995, about 29% for federal employees and 8% for who earned almost R$1500 per month. state employees, and fell by 12% for Earnings were 25% lower than this for municipal employees. Earnings for judicial federal workers, 50% lower for state and and legislative employees rose by more than military personnel, 62% lower for education 40% during these years, by about 7% for the and health workers, and 75% lower for military, and stayed roughly constant for municipal workers. In the private sector, education and health workers in the public earnings were highest in productive services sector. In the private sector, earnings in at about R$950 per month. Earnings in agriculture fell by about 10%, in industry by distributive services and manufacturing were 6%, but rose in distributive and productive about 45% lower than this, those in personal services by 6%, and in personal services and services and construction about 60% lower, construction about 16%. The largest and those of agricultural workers about 75% increase in real earnings was therefore for lower. Table 5 Average Monthly Earnings, 1981-1995 All Brazil, in constant September 1995 reals* 1981 1985 1990 1992 1993 1995 ~~~~~~~............................................................ ................. .. .... .. . ... . . . . ............................................................ Public Employment Direct Administration Federal 876 1155 1059 1056 1150 1126 State 675 780 783 555 621 728 Municipal 412 386 398 330 316 361 Education & Health 546 606 621 476 512 551 Judiciary & Legislative 1043 1247 1362 1058 1235 1473 Military 669 844 687 637 622 718 Private Employment* Agriculture 238 256 210 207 235 213 Manufacturing 557 547 483 491 518 525 Services Distributive 501 537 520 455 484 532 Productive 900 953 896 878 975 953 Personal & other 312 308 335 292 314 354 Construction 333 345 367 311 319 389 * The deflator used is INPC (Brazil) ** Private employment includes employment in public enterprises. Souce:Pesuis Naionl r Amostra de Domicilios (sel _eed_ar) Considerable differences exist in average 1995, average earnings in almost all sectors earnings in different parts of the country. In were highest in the federal district and Sao 6 Labor Market Prospects of Public Employees in Brazil Paulo than in other states, and earnings in Pernambuco and Sao Paulo. In absolute Pernambuco and Bahia were generally the terms, however, earnings of Sao Paulo's lowest. Earnings in Minas Gerais were municipal workers were more than double somewhat higher than for the two those in Pernambuco. northeastern states, and earnings in Rio Grande do Sul and Rio de Janeiro higher In the private sector, relative earnings still. Relative to salaries in the by sector are uniform across regions. The manufacturing sector, state employees only exception is that in Rio Grande do Sul, received roughly 67-80% more in Bahia and earnings of agricultural workers is Pernambuco, and 40-45% more in the other comparatively high. But this is hardly states. Municipal workers' earnings relative surprising; what is more striking is that to private manufacturing earnings, in sectoral earnings relative to those in contrast, were highest in Rio Grande do Sul manufacturing are similar for all states. and Rio de Janeiro, somewhat lower in Minas Gerais and Bahia, and lowest in Table 6 Average Monthly Earnings, September 1995 Selected States, in current reals* Distrito Bahia Pernam- Minas Rio Rio de Sao Federal buco Gerais Grande Janeir Paulo do Sul o .......................................................................................................I.................................................................................................................................. Public Employment Direct Administration Federal 1361 842 744 1046 839 972 1436 State 1034 551 549 560 651 782 981 Municipal - 267 204 314 453 502 444 Education & Health 1128 364 387 505 647 562 672 Judiciary & Legislative 2057 869 1189 1565 1804 1496 1349 Military I048 576 577 708 541 719 744 Private Employment* Agnculture 337 156 149 278 458 281 328 Manufacturing 627 330 304 387 467 538 683 Services Distributive 592 356 339 474 534 514 681 Productive 1246 565 724 743 1026 947 1121 Personal & other 423 222 232 290 360 365 472 Construction 535 247 232 297 337 380 560 * The deflator used is INPC (Brazil) ** Private employment includes employment in public enterprises. Source: Pels uisa-Nacionalpor Amostra de Domicilios, 1995. 3.2 The Wage Bill, 1992- 1995 subsectors in public employment. Education and health workers, who are about 50% of Using employment and average all government workers, absorb about 45% earnings for each sector, we computed the of the wage bill. Municipal employees are sectoral wage bill in 1995. Figure 1 graphs about 20% of total public employment, but the relative importance of each of the absorb only 12.5% of total wages; state Labor Market Prospects of Public Employees in Brazil 7 employees, who are about 15% of actually fell in agriculture. In general, only employment, absorb about 17% of total employment growth in services exceeded the wage. Federal and military workers are national average of about 2% during this relatively small fractions of both total period. In the public sector, on the other employment and wage bill. The most hand, employment of federal, municipal, dramatic difference between these two shares judicial and legislative, and education and is for judicial and legislative workers, who health workers rose faster than this average are about 5% of total employment, but rate. Only state and military employment almost 14% of the total wage bill. Thus, grew slower than the national average. while employment of judiciaVlegislative workers may be relatively small, they are In the public sector, the 20% annual potentially critical for fiscal reasons. growth of the wage bill for judicial and legislative workers dwarfs that of other To examine the trends in wage- workers. This growth in personnel related expenditures in the public sector, we expenditures was due to both an increase in computed the growth of the wage bill in the their number (about 7.5% annually) and public an private sectors. Because of increased average earnings (about 12% inaccurate weighting before 1991, wage bill annually). The growth of state wage bill figures are reliable only for the three PNAD exceeded 9 percent, despite a small decrease surveys for 1992, 1993, and 1995. Figure 2 in state employment. The wage bill for graphs the annual growth rate for education and health workers increased by employment, average earnings, and the wage 7.5% annually, due to increases in both bill during the period 1992-1995. average real earnings and employment. The federal wage bill also increased by 4.5% Average earnings grew in every part because of both new hires and higher wages. of the private sector, especially in personal Military employment stayed roughly and distributive services and construction. In constant, but earnings increased moderately. contrast, industry and agriculture registered modest earnings growth, and employment Figure 1 FISCAL IMPORTANCE OF GOVERNMENT SUBSECTORS, All Brazil, September 1995 60 00 50.00 . . ..... 13~~ Share of Employment c Share of Wage Bill 30 00 20.DO~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0.00 Education & Health State Judicial S Legislative Municipal Miliary Federal 2 .t 1,392 99~5 HoMPOMYhl.iT & Et G ANNUAL G-V.l f o 20~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 'a~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~t 05/ 9i;i00'' "'0' wD18$M~~~~~~~~~~~~~~~~~~~~~~ vaeB 9 i i | \ \ \ i \ \M 10 Labor Market Prospects of Public Employees in Brazil 4. THE PUBLIC-PRIVATE WAGE GAP Without knowing what the wage gap market segmentation, since they do not between the public and the private sector is, necessarily represent the actual difference however, it is difficult to determine whether in payment between equally productive earnings growth of government employees workers in the public and private sectors. has been excessive. Thus, for example, the In fact, the overall wage gap captures relatively high average earnings of judicial both differences in payment between and legislative employees may be because of equally productive workers in the two their higher skill levels relative to other sectors and differences in the public employees and workers in the private qualifications and characteristics of the sector. And the relatively rapid growth in labor force employed in the two sectors. their earnings may be because they have been On the one hand, wages may be higher in historically underpaid relative to their private the public sector simply because the sector counterparts. To examnine whether or labor force in the sector is older and not public sector employees are under- or better educated. On the other hand, overpaid relative to those in the private wages in the public sector may be lower sector, we compute the earnings gap in this due to the possible concentration of section. This would help in determining public employment in the Northeast and whether the solution to a reducing the public because women tend to be over- sector wage bill lies in reducing public represented in the public sector. sector employment, or average earnings and benefits, or both. * The third step is dedicated to analyze the impact of differences in the spatial 4. 1. Approach location of the public and private labor force on the wage gap between the sectors. Specifically, we investigate to In this section, we investigate the which extent the over-representation of wage gap between the private and public public employment in the northeast tends sectors in Brazil using PNAD data. The to make overall measures of the wage analysis consists of four steps: gap between the public and private sectors an under-estimate of the actual * We begin by estimating the overall wage gap faced by local labor markets. gap between the private and public sectors for the entire Brazilian labor * In the fourth step we estimate the wage market. gap between workers with identical observed characteristics in the public and * In the second step, we take into account private sectors. To make this analysis the heterogeneity within the public empirically feasible we limit the scope to sector. We desegregate the public sector the six major Brazilian Metropolitan using two alternative procedures. In each Areas and the Federal District. Since the case, we estimate the wage gap between wage gap between the public and private the segments of the public sector and the sectors is likely to differ according to the overall private sector. Although these type of worker, in this fourth step we estimates for the overall wage gap also investigate how the wage gap varies represent an important starting point, with the workers' main observed they have serious limitations as a characteristics. measure of the actual degree of labor Labor Market Prospects of Public Employees in Brazil 11 All the empirical analysis conducted need to introduce some notation. Let w, and in this section is based on the Pesquisa Wb be the average wages in the public and Nacional por Amostra de Domicilios - private sectors, respectively. Using this PNAD - surveys for September 1995. We notation, one can define the absolute wage used two measures of labor income. The difference (Go), two alternative versions for first measure is the total monthly income the relative wage difference (Gla e Glb), and normally received by a worker in his/her the log-wage difference (G2) as follows: current main job. The second measure seeks to standardized for the number of hours Go= W-Wb usually worked. It is defined as the total monthly income normally received in the main job, divided by the number of hours GI. = W. -Wb usually worked per week in main job and Ga multiplied by a standard work week of 40 hours. Since the average number of hours Glb = w, worked by public employees tends to be b significantly smaller than the corresponding average for workers in the private sector, it does make a difference for level of the wage G2 = ln(w,) - ln(wb) = l(w,/w,) gap between these sectors whether a standardization for hours of work is conducted or not. To ensure that the T he absolute wage difference, Go, standardized wage is a good approximation has a disadvantage compared to the relative of what worker would get if they work 40 difference in wages, G 2a and Gib, and the hours a week, workers working less than 20 difference in log-wages, G2, in that it is hours per week were excluded from the sensitive to the unit of measurement. The analysis. We also exclude workers with zero other measures are not. labor income. Gi, and Gib, have a disadvantage Before reporting the results, we relative to the differences in log-wages, G2. present a brief methodological discussion on To define a relative wage difference, it is how the wage gap between the public and necessary to select a baseline wage as private sectors is measured in this section. reference. So this measure inherits the inconvenience of being sensitive to the 4.2. Wage-gap concepts choice of the baseline wage. Notice the difference between Gma and Gib. Thus, for Empirical analysis of wage instance, if the average wage in the public differentials commonly use wage levels, log- and private sectors are 50 and 20 monetary wages, arithmetic means and geometric units respectively, then the wage gap can be means. In this section we briefly review the expressed in relative terms as 150% of the connection between these concepts and how private sector average wage or 60% of the they are going to be used to obtain measures public sector average wage. The need to for the wage gap. constantly having to refer to the chosen baseline wage makes the use of relative wage We begin by reviewing the differences cumbersome in detailed analysis advantages and disadvantages of using of wage differentials. A way to avoid this is differences in log-wages vis-a-vis absolute or to use measures of wage differentials that are relative differences in w,ages. To proceed we insensitive to the choice of a baseline 12 Labor Market Prospects of Public Employees in Brazil reference wage. The log-wage difference is conducted before taking averages. This is such an alternative. In the example above, particularly ubiquitous to studies of wage the log-wage difference between the public differentials based on regression analysis, and private sectors will be 0.92 since they invariably use log-wages as the independently of the monetary units chosen dependent variable. and will not require a choice of a baseline wage for reference. When logs are taking before the averages are obtained, comparisons between The log-wage difference, however, the logs of the averages are not any more has a major disadvantage in that it is more comparisons between the logs of two difficult to interpret. One way to facilitate its arithmetic means. The comparison becomes interpretation is to notice that, for small one between the log of two geometric variations, the log-wage difference is an means. This is a consequence of the fact that approximation for the relative change in the the arithmetic mean of the logs is identical to level of wages. For instance, if the wages are the log of the geometric mean. 50 and 51 the relative wage variations using 50 or 51 as the baseline references are Thus, there is a third commonly used 0.0200 and 0.0196, respectively. In this case technique for comparing the wages of two the log-wage difference is 0.0198. In the populations. This third possibility consists of general case, ifW,>Wbthen computing the log difference between the geometric means, which is equivalent to the W. -aWb _(WE) > Wa -Wb difference between the arithmetic mean of WC, In( Wb) Wb the logs. We refer to this measure as the gap in average log-wage and denote it by G3. To or make clear when we are using G2 and when we are using G3 we will always refer to G2 as G laŽ G22 GIb the gap in log-average-wages. In sum, the log-wage difference has To clarify the differences between two useful properties: (a) It eliminates the these altewsative measures for the wage gap need to keep track of the wage used as between two sectors, let (wi: i:l,...,n} and reference, and (b) it gives an estimate for the {Wbj: i: I,... ,m} be the wages of workers in wage gap that is between the two natural sectors a and b, respectively. Using this measures for the relative difference in wages. notation, the three altemative measures for The major disadvantage of the log-wage the wage gap introduced up to now can be difference is the fact that it is difficult to expressed as: interpret. a Up to now we are considering G = n m alternative methods for evaluating the a 1m relative difference between two average Wbi wages, where supposedly the average wage is simply the arithmetic mean of wages. In n m this case the question of using the level of - Wa- - wages or their logs only appears after the Gl= :_n = m average wage has been already obtained. -w However, in many circumstances, the n passage from wage levels to log-wages is Labor Market Prospects of Public Employees in Brazil 13 fact, it is possible to relate the levels of G2 n and G3. The connection between them is W . n Ewai intrinsically related to the relative degree of G2= In n JI wage inequality in the two sectors. m J=l The simplest way to consider this question is to notice that the arithmetic mean and is always greater than the geometric mean, tf n lln except when all wages are the same. This l n l m trI Wail fact led researchers to propose the log- G3=-Ylnw j--EInw =i/. difference between the arithmetic and n i=' m i=, fnwb=/ geometric mean as a measure of wage = J inequality. Hence, if we define the wage inequality in these two populations by their Above we show that G2 is always Theil indices (T): between GI,, and Gib, but what is the connection between the level of G3 and the level of the other measures? As a matter of T =ln w) - Int wnIn Wa ] = _ - nw I 1 m m I lX m Tb = In(± wbi) -lIn F(Hwbz = In ±T Wbj m i bi we obtain (G2= G3 + (Ta - Tb) Therefore, G2 and G3 will be the same if and average log-wages (G3) is greater than the only if the level of wage inequality, measured gap in log-average-wage (G2). Annex table 1 by the Theil index, is the same in both illustrates this, presenting estimates for the sectors. More generally, G2 will be similar to wage gap between the private and public G3 if the levels of inequality are similar. sector using all these measures. This table When inequality is greater in sector b than in reveals that, as a consequence of the greater sector a, we will tend to observe values for degree of wage inequality in the private G3 greater than corresponding values for G2. sector, the wage gap between the public and Unfortunately, this is precisely the situation private sectors tends to be much higher when when contrasting the public and private measured by the gap in average log-wages sectors. Wage inequality in the private (G3). sector tends to be greater, so the gap in 14 Labor Market Prospects of Public Employees in Brazil Table 7 Measures of the Public Sector Wage Premium Unadjusted for Sector, Re ion and Worker Attributes, Percent Measure Public-Total Private Public-Formal Private Adjusted* Not adjusted Adjusted Not adjusted ............................................................................................................................................................................................................................. Relative Wage Gap Baseline: Public Sector Wage 47 33 40 30 Baseline: Private Sector Wage 89 49 67 42 Logarithmic Measure of Gap** Log average-wages 64 40 51 35 Average log-wage 70 45 38 18 * Adjusted refers to earnings adjusted for differences in hours worked. ** These measures refer to the wage gap relative to a weighted average of the levels of wages in the two sectors, and have the advanta e of bein "ind ndent" of the baseline levels 4.3 The overall wage gap that those leaving government employment are more likely to seek employment in. This Table 7 presents estimates for the table reveals that this wage gap is just wage gap between the public and private slightly smaller than the overall gap between sectors for Brazil. Average earnings tend to the public and private sectors. More be much higher in the public than in the specifically, the wage gap between the public private sector. This table also shows that the sector and the private formal sector is 30% wage gap between the sectors is much of the public sector average wage. Hence, it greater when labor income is not is just 3 percentage points smaller than the standardized for the number of hours corresponding gap between the public and worked, indicating that workers in the public the overall private sectors. Alternatively, the sector work fewer hours per week than gap in log-average-wage is 0.35, which is workers in the private sector. The non- just 0.05 smaller than the corresponding gap standardized wage gap tends to be between relative to the overall private sector. 15% and 20% smaller than the corresponding gap in standardized wages. This table reveals that this wage gap The results using the standardized wages is smaller than the overall standardized gap indicate that the relative wage gap between between the public and private sectors. At the sectors is equivalent to 33% of the this stage of the analysis, it is very difficult to average wage in the public sector or 48% of interpret these large wage gaps. Three the average wage in the private sector. factors complicate such a comparison: first, there are large differences in average The results using the standardized earnings within the public (e.g., between wages indicate that the relative wage gap municipal and legislative workers) and between the sectors is equivalent to 33% of private sectors (e.g., between agricultural the average wage in the public sector or 48% and service sector workers); second, there of the average wage in the private sector. are regional differences in average earnings The two log-wage gaps are close to 0.40. (e.g., between state employees in The gap in log-average-wage is 0.39 whereas Pernambuco and Sao Paulo); and third, there the gap in average log-wages is 0.44. Table 7 are large differences across workers by also reports estimates of the gap between the individual attributes (e.g., workers in the wage of employees in the public sector and public sector tend to be better educated and employees in the private formal sector, older). Henceforth, only wage gaps adjusted which may be the part of the private sector for hours worked will be discussed and Labor Market Prospects ofPublic Employees in Brazil 15 reported in the paper. Readers can find are particularly higher in federal results for unadjusted differences in the administration, in particular, in the legislative tables at the end of the paper. and judiciary sectors. In these sectors, average wages are between 150% and 250% The natural next step is to estimate higher than the average for the private the wage gap between workers in the public sector. At the other extreme, workers in and private sectors with identical observed health and education, in particular those at characteristics. Nevertheless, before we the municipal level, have earnings that are proceed in this direction, we dedicate the close to the average for the private sector or next section to the analysis of the degree of even below it. The average wage in wage heterogeneity within the public sector. municipal administration is 16% lower than the average in the private sector. For state 4.4. Heterogeneity in the public sector administration and military personnel, wages are close to 70% above the average in the Thus, the natural next step is to private sector. estimate the wage gap between workers in the public and private sectors with identical Similar results are obtained using the observed characteristics. Nevertheless, alternative desegregation of the public sector before we adjust for worker-specific based on the nature of labor contracts and characteristics, we examine the degree of level of government. Wages are particularly wage heterogeneity within the public sector. high among federal public servants, and To do this, we use two alternative employees in federal enterprises. (For decompositions. The first uses a simplicity we identify workers in the public desegregation based on the sector of sector with signed working card with economic activity where we distinguish employees in public enterprises.) In both between the three levels of public cases the wage gap is greater than 150% of administration (federal, state, and municipal), the average wage in the private sector. For public health and education, and the military personnel, state public servants and legislative and judiciary activities. This employees of state enterprises, the average desegregation of the public sector is not wage gap is about 75% to 100% of the exhaustive since it cannot distinguish average wage in the private sector. The employment in public enterprises from average wage of municipal public servants private employment. The second and employees of municipal enterprises is decomposition, which is based on below the average for the private sector, information on both the type of labor with the wages of employees in municipal contract and on the level of government, is enterprises being particularly low. exhaustive and permits decompose the employment at each level of government into Overall, the average wages in all public servants and employees with and segments of the public sector tend to be without a signed work card. above those paid in the private sector, the only exceptions are the wages paid by Table 8 presents estimates for the municipal administrations and enterprises. wage gap between segments of the public But these comparisons also reveal a large sector and the overall private sector. This degree of heterogeneity in wages among table uses both desegregation procedures for subsectors within the public sector, with the the employment in the public sector. The average wage in federal administration being results indicate a great level of heterogeneity more than 200% higher than the within the public sector, with the average corresponding average paid in municipal wage being much higher in certain segments administrations. of the public sector than in others. Wages Labor Market Prospects of Public Employees in Brazil 16 Table 8 Measures of the Public Sector Wage Premium Unadiusted for Re ion and Worker Attributes -Measure Relative Wage Gap(%): Logarithmic Measure of Gap**: Baseline: Baseline: Log average Average log Public Wage Private Wage -wage .........g..... ............................................................................................................. . .................................... . ................. . .. ......................... ............................ .............. ...................................... By Sector of Activity: Federal Administration 61.9 162.5 1.0 1.2 State Adrninistrations 41.1 69.9 0.5 0.7 Municipal Administrations -18.8 -15.8 -0.2 -0.0 Judicial & Legislative 69.8 231.0 1.2 1.3 Military 41.1 69.9 0.5 0.7 Education and Health 16 18 0.2 0.2 By Class of Worker: Federal Public Servants 67.4 207.0 1.1 1.4 Federal other, with signed card 62.5 166.4 1.0 1.2 Federal other, no signed card 40.5 68.2 0.5 0.4 State Public Servants 37.7 60.5 0.5 0.7 State other, with signed card 48.8 95.3 0.7 0.8 State other, no signed card -1.2 -1.1 -0.0 0.1 Municipal Public Servants -2.9 -2.8 -0.0 0.1 Municipal other, with signed card -48.8 -32.8 -0.4 -0.2 Municipal other, no signed card -71.4 -41.7 -0.5 -0.7 Milituy 41.8 71.7 0.5 0.7 * All estimates are for earnings adjusted for differences in hours worked. This measure is the wa e a relative to a weightedavera e of the levels of waes in the two sectors 4.5 Regional differences in the public- consider the hours-standardized measure for private wage-gap wages. In previous sections, we considered As far as differences in the spatial the magnitude of the overall wage gap distribution of the labor force are concerned, between the public and private sectors. This Annex Table 1 reveals that public overall gap, however, captures both intrinsic employment tends to be only marginally differences in wages between these sectors over-represented in areas with lower average and differences in the characteristics of wage (the northeastern states). In fact, public workers in the two sectors. In this section employment is over-represented in the we consider the role of spatial differences in Federal District and in Rio de Janeiro, which the distribution of public and private are among the states with the highest employment. More specifically we average wage. Employment in the public investigate three topics: first, we investigate sector is slightly over-represented in the whether public employment is over- poorer areas: 49% of public employment and represented in the poorest states; second, we 46% of private employment are in states with investigate how the uneven distribution of average wage below the national average. public employment across states affects the The results indicate that the gap in average overall level of the wage gap between the log-wage would increase from 44% to 46% public and private sectors; and finally we if public employment had the same spatial investigate how the within-state wage gap distribution as private employment, i.e., between the public and private sectors varies public employment has a "locational across states. In this subsection we only disadvantage" of only 2%. Labor Market Prospects of Public Employees in Brazil 17 Annex Table 2 presents alternative urban areas. The contribution of federal and measures for the wage gap for the entire state employment is likely to be greater. country and for the states and metropolitan Moreover, municipal jobs in metropolitan areas included for study in this report. This areas are more likely to be better paid. As a table reveals that the wage gap tends to be consequence, the wage gap estimated for considerably higher for the metropolitan these areas is likely to over-estimate the gap areas than for the corresponding states. This for the entire urban labor market in the is especially true in the northeastern states of respective state. Pernambuco and Bahia, and in the State of Minas Gerais. Sao Paulo is the only state 4.6.1. Methodology where the wage gap for the metropolitan area is smaller than for the corresponding The basic methodology consists of gap for the entire state. estimating the wage gap between workers with identical observed characteristics in the The average gap between the public public and private sectors. This wage gap, and private sectors is much higher in Brasilia which will referred as the controlled wage than in any state. The average wage gap gap, is taken as an estimate of the wage tends to be high in the northeast. and lower advantage faced by workers in the public in the more developed states in the South sector. The difference between this gap and and Southeast of Brazil, particularly in Sao the overall gross wage gap is a measure of Paulo, Parana and Santa Catarina. But it is the impact of sectoral differences in the also very low in a few poor states like Ceara composition of their labor force on the and Alagoas (see tables in Statistical Annex). overall wage gap between the sectors. 4.6 Role of differences in worker The basic set of observed characteristics characteristics includes gender, race, schooling, and age. We also work with a In this section we examine the role of version of this set that includes tenure at the differences in the characteristics of workers current job. Workers in the public sector between the public and private sectors in tend, on average, to have been in their explaining the wage gap between the two current job for a longer period of time. Since sectors. The fundamental question being to a considerable extent this is not a result of answered is: What fraction of the wage gap any differential merit between employees in between the two sectors is simply due to the public and private sectors, but one of the sectoral differences in the characteristics of major advantages of jobs in the public sector, the labor force employed in the two sectors? it is unclear whether we should control for this characteristic. Hence, we present all To investigate this question we results including and excluding tenure from consider the labor market of the major six the analysis. Brazilian metropolitan areas and Brasilia. The restriction of the analysis to these seven The methodology used to compare well-defined local labor markets is useful to the wage of observably identical workers in isolate the wage gap between the public and the public and private sectors consists of private sectors from possible spatial three steps. In the first, we regress the log- differences among local markets. This wages of workers on their characteristics and restriction, however, also has some on an indicator of whether they are in the important disadvantages. The major public or in the private sector. In the second disadvantage is the fact that the nature of step, based on the results of these public employment in large metropolitan regressions, we compute what would be the areas is likely to be different from smaller average wage of public employees if they 18 Labor Market Prospects ofPublic Employees in Brazil were in the private sector, given their To further simplify the empirical personal characteristics. Finally, we obtain analysis, we assume that the dependency on the wage gap between workers with identical schooling and tenure are linear and the observed skills as the difference between the dependency on age is quadratic, i.e., we actual average log-wage in the public sector assume that and the average log-wage they would receive if they were in the private sector, as f3(e,p) = a3(p).e estimated in the second step. f5Q, p) = a5(p).t We implement this procedure using and that both the standardized and the non- standardized measure for wages. In both 2 cases, two alternative specifications for the f4(a,p)= +b(p.a log-wage regressions are used. To describe Moreover, we can, without any loss of the specification of these regressions, we generality, write the functions on gender and concentrate the attention to the regressions race also as linear functions, since they are that include tenure, the other regressions are dichotomous variables, i.e., we can, without obtained simply by omitting tenure. any loss of generality, write We begin by establishing some f (g p) = a (p).g notation. Thus, let w denote the wage and denote by h the regression function of log- and wages on gender (g), race (r), schooling (e), age (a), tenure (t), and an indicator for the f2(r, p) = a2(p).r public sector (p), i.e., Collecting these results, we can express the Efln(w)/g, r, e, a, t, p] = h(g, r, e, a, t, p) log-wage regression function as Then, the specification of the regressions can h(g, r, e, a, t, p) = ao(p) + a,(p).g + a2(p).r + be seen as a series of hypothesis about the a3(p).e + a4(p).a + b4(p).a2 + as(p).t functional form of h. Our basic assumption is that the regression function is separable on This expression is our first specification, and gender, race, schooling, age, and tenure, but will be referred as the general model. We not necessary on the indicator for the public also estimate and analyze a simplified version sector, i.e., of this model, that will be referred as the basic model. To obtain the basic model we h(g, r, e, a, t, p) = f1(g, p) + f2(r, p) + f3(e, p) + further assume that the regression is also f4(a, p) + f5(t, p) separable on p, the indicator for the public sector. This hypothesis is equivalent to To allow for the regression function to be assume that the impact of all personal non-separable on p, the indicator for the characteristics on the level of wages is the public sector, is central to the analysis in this same in the public and private sector. That is, study. This non-separability is a necessary we assume that all coefficients except the condition for the wage advantage of workers intercept are common to both sectors. In this in the public sector to vary with their case the regression model becomes characteristics. If the regression function were separable in p the wage advantage h(g, r, e, a, t, p) = ao(p) + a,.g + a2.r + a3.e + would be identical for all types of workers. a4.a + b4.a2 + a5.t Labor Market Prospects of Public Employees in Brazil 19 Since p is also a dichotomous variable, In the case of the basic model the without any loss of generality we can express average log-wage in the public sector is the function ao as a linear function, i.e., given by h(g, r,e,a, t,p)= a+aop+al.g+a2.r+a3.e+ a+ao+a,.±(g/l)+a2.p(rl)+a3.p(e/l)+ a4.a + b4.a' + a5.t a4ii(a/l) + b4.g(a2/l) + a5.p(t/l) In terms of estimation, the general and the average log-wage that would prevail model is estimated running regressions of in the private sector if the labor force in the log-wages on gender, race, schooling, age sector had the same characteristics as those and tenure for the public and private sectors of the public sector labor force is given by separately. As already mentioned, this specification has the advantage of permitting a + a1.R(g/1) + a2.p.(r/l) + a3.p.(e/l) + a4.p.(aIl) to evaluate how the wage gap between the + b4. g(a2/I )+ a5..L(t/1) public and private sectors varies with worker's characteristics. The basic model As a consequence, the log-wage gap assumes that the wage gap is the same for all between workers with identical observed types of workers. It is estimated by characteristics is given simply by ao. regressing log-wages on gender, race, schooling, age, tenure and an indicator for 4.6.2. Results public employment in a sample pooling together the public and private labor force. Annex tables 3a and 3b present, for each metropolitan area and concept of wage, In the case of the general model, the estimates of the log-wage gap between average log-wage in the public sector is: workers with identical observed characteristics (excluding and including ao(l) + al(l).g(g/l) + a2(l).g(r/1) + a3(l).p(e/l) tenure on current job, respectively). + a4(1).g(a/l) + b4(1).g(a2/1) + + a5(1).p±(t/l) These tables reveal that differences in the composition of the labor force are a where, j(x/l) denotes the average of the major explanatory factor of the log-wage gap characteristic x in the public sector. Based on between the public and private sector. the same model the average log-wage that Overall these differences in the composition would prevail in the private sector if the of the labor force are responsible for the labor force in the sector had the same average log-wage in the public sector being characteristics as those of the public sector is from 0.5 to 0.8 higher than in the private given by sector. As a consequence, these differences explain almost all the wage gap between the ao(O) + a1(O).g(g/l) + a2(0).p(r/l) + a3(0).g(e/l) public and private sectors, revealing that + a4(0).R(a/l) + b4(0).p(a2/I) + most of the overall wage gap between the + a5(0).g(t/1) public and private sector just reflects differences in the composition of their labor Therefore, the estimate of the log-wage gap force. Therefore, we reach the conclusion between workers with identical observed that overall measures for the wage gap that characteristics is given by does control for differences in the characteristics of the labor force are bounded (ao(l)-ao(O)) + (a,(l)-a,(O)).p(g/l) + (a2(1)- to be very misleading indicators of the actual a2(0)).g(r/1) + (a3(l)-a3(0)).i(e/l) + wage advantage of workers in the public + (a4(l)-a4(0)).4(a/l) + (b4(1)-b4(O)).p(a2/1) + sector. (aA(I)-aA(O))..LQ(/I) 20 Labor Market Prospects ofPublic Employees in Brazil In terms of the individual Annex table 4 reveals that differences contribution of each characteristic, around in the composition of the labor force are a 70% of the difference is explained by the major explanatory factor of the log-wage gap higher educational level of the labor force in between the public and private sector. A the public sector. The remaining 30% is large fraction of the overall wage gap explained by the fact that public employees between the public and private sector just tend to be older and have longer tenure, with reflects differences in the skill composition of each of these two factors being responsible their labor force. Therefore, overall measures for 10 to 20% of the differential. The impact for the wage gap that do not control for of the differences in the composition by differences in the characteristics of the labor gender and race is small. Differences in race force are misleading indicators of the actual composition between the private and public wage advantage of workers in the public sectors are small and, as a consequence, they sector. have a negligible impact on the wage gap. Contrary to the other characteristics, gender About 70% of the gross wage gap is has a small negative impact, since women explained by the higher educational level of have lower wages and are over-represented the labor force in the public sector. The in the public sector. remaining 30% is explained by the fact that public employees tend to be older and have In sum, differences in the sectoral longer tenure, with each of these two factors composition of the labor force are a major being responsible for 10% to 20% of the reason for the wage gap between the public differential. The impact of the differences in and private sector. One of the consequences the composition by gender and race is small. is that once the differences in composition Differences in race composition between the have been eliminated, the rather high wage private and public sectors are small and, as a gap between the two sectors, becomes rather consequence, they have a negligible impact small, with Brasilia being the notable on the wage gap. The impact of gender exception. differences is also small because though women have lower wages than men in either The methodology used to compare sector, they tend to be over-represented in the wage of observably identical workers in the public sector which pays higher wages. the public and private sectors consists of three steps. In the first, we regress the log- In sum, differences in the sectoral wages of workers on their characteristics and composition of the labor force are a major a dummy variable for whether they are in the explanation for the wage gap between the public or in the private sector. In the second public and private sector. Once the step, based on the results of these differences in worker attributes have been regressions, we compute what would be the eliminated, the high wage gap between the average wage of public employees if they two sectors becomes smaller but remains were in the private sector, given their considerable for most states: personal characteristics. Finally, we obtain the wage gap between workers with identical * In Brasilia, the wage gap declines to a observed skills as the difference between the still very large public-private wage gap of actual average log-wage in the public sector 63% among workers with identical and the average log-wage they would receive observed characteristics. if they were in the private sector, as * In Recife and Salvador, workers in the estimated in the second step. public sector receive salaries that are 32% and 21% higher respectively. Labor Market Prospects of Public Employees in Brazil 21 * The public-private wage gap is 17% in The table reveals large differences Belo Horizonte, 14% in Porto Alegre, among segments of the public sector with and 12% in Rio de Janeiro. respect to their wage advantage. The wage advantage, measured by the controlled log- o Only in Sao Paulo is the pattern reversed: wage gap, is larger at the federal level for among equally qualified workers, both among public servants and among workers in the public sector receive employees in public enterpnses. For these salaries that are 13% lower than private groups the average salaries are 35% to 65% sector salaries. higher than for comparable workers in the private sector. At the state level, the wage 4.7 Controlled wage gaps by sector and advantage is close to zero for public class of worker servants; for employees in public enterprises the wage advantage is positive and between In this section we estimate the wage 10% and 22%. At the municipal level, public gap between each segment of the public servants and employees in public enterprises sector and overall private sector controlling get lower wages than -workers with similar for differences in worker attributes. This is observed characteristics. Overall, the perhaps the best estimate of the "pure" evidence corroborates the existence of premium enjoyed by public sector employees significant wage advantage of some over what they would have earned had they segments of the public sector relative to the held private sector jobs. Table 9 presents private sector. these estimates for the country as a whole. Table 9 Measures of the Public Sector Earnings Premium, September 1995 -Adusted for Worker Characteristics*, by Sector/Class, Percent Measure Monthly" ~Hur!y** ......... .. .... ........._ . ............. ........... ................ ............................................ ....................................................................... ................................. By Sector of Activity: Federal Administration 23.1 28.9 State Administrations -7.8 -3.8 Municipal Administrations -31.9 -22.4 Judicial & Legislative 43.9 55.9 Military 2.3 5.7 Education and Health -31.1 -15.6 By Class of Worker: Federal Public Servants 40.8 46.3 Federal other, with signed card 27.0 36.3 State Public Servants -15.4 -5.5 State other, with signed card 10.3 18.0 Municipal Public Servants -30.5 -17.9 Municipal other, with signed card -31.5 -20.9 All workers without signed card -42.6 -19.7 Military 3.6 7.2 * Workers characteristics adjusted for are age, sex, race, tenure, and education. ** Monthly earnings premia are not adjusted for differences in hours worked. Hourly earnings comparisons are adjusted for differences in hours worked. As mentioned earlier, the main the basic model is to allow an analysis of advantage of the general model relative to variations in the controlled log-wage gap 22 Labor Market Prospects ofPublic Employees in Brazil across different types of workers. In this in Annex tables 6a and 6b. This table also section we explore this further. We indicates how this gap varies with race and investigate how the controlled log-wage gap gender, with the level of schooling, age and varies with the educational level of workers, tenure. with their age and tenure, and with their race and gender. The methodology consists of Except for the fact that the wage two steps: first, we compute the log-wage advantage of women in the public sector is gap for a baseline type of workers, and clearly greater than that for men, all other second, we vary each characteristic in tum results are not robust across regions. Annex and estimate how the wage gap changes as tables 7a and 7b illustrate this conclusion. In these characteristics change. this table we indicate which groups have greater wage advantage in the public sector The "standard" or baseline worker is in each region. This table reveals no general assumed to be a white male with 8 years of patterns. The profile of the groups with schooling, who is 35 years old and has greater wage advantage seems to be worked 5 years in his current job. This is the completely region-specific. prototypical worker in the private sector in metropolitan Brazil. The controlled log- wage gap for this type of worker is presented 5. PUBLIC-PRIVATE DIFFERENCES IN PENSIONS The calculation of public private differentials year's salary in the 36th year as the is somewhat complicated, and we present the estimated monthly pension (because technique only in outline here, focusing the pensions for public servants are based on discussion on results instead. To estimate the last month's salary). For com differences in pensions for public servants carteira salaried workers and the and those receiving social security benefits selfemployed, we use the 34th year's from the INSS Time of Service and Old Age salary as the estimated pension level schemes, we use the following technique: (because INSS time of service benefits are based on salar levels in the last 36 * First, using 1995 PNAD survey data for months of servic els in the lake the seven metropolitan areas used in this the assumption that selfemployed report, we estimate the profile of workers do actually contribute to and monthly earnings for male and females, receive social security. For sem carteira divided fuirther into four worker workers, we assume that these workers categories (public servants, private con will receive minimum pensions from carteira, private sel carteira, and other INSS programs: their pension private selfemployed) and three levels are assumed to be equal to one education groups (0-8 years, 9-11 years, minimum wage in 1995, which we and 12+ years of schooling). assume to be R$100. * Second, using these estimated earnings * Third, we make adjustments for the fact profiles, and assuming that all males that INSS pensions are capped at 10 retire after 35 years of service, and all minimum wages. This results in pensions women after 30 years, we estimate the for com carteira men with 12+ years of average pension levels for each group. education being restricted to equal 1000. For public servants, we use the 36th Labor Market Prospects of Public Employees in Brazil 23 * Fourth, using weights obtained from private differences are greater for men than nationwide PNAD data for 1995 on the for women, and greatest for the most shares of com carteira and sem carteira educated male workers. Thus, the average employees, and selfemployed workers, CLTista male worker with 0-8, 9-11, and we compute the average expected 12+ years of schooling will receive, pensions for private sector men and respectively, 75%, 72% and 50% of the women. We do not consider pensions monthly pension level of his estatutario from the complementary social security counterpart. The corresponding numbers for system (i.e., closed or company funds). women are 73%, 73%, and 71%. The findings on private pay and pension * Finally, these numbers are compared with differences, combined with the fact that expected monthly pensions for public pension contribution rates are 0-2% for servants with the same education level. public servants but range between 8-11% for Table 10 reports the results of these CLTistas, have important implications for estimations. the structure of optimal severance packages, and the likely success and full costs of The results show that due to the current voluntary severance programs. nature of the estimated profiles and differences in rules (basically the pension ceiling of 10 minimum salaries), public- Table 10 Estimated Monthly Pensions, September_1995 Reais Gender/Education Public Private With Card Without Card .my Total . ..... Males 0-8 years schooling 447 430 100 407 335 9-1l years schooling 772 758 100 686 558 12+ years schooling 1747 1793** n.a. n.a. 1000 Females 0-8 years schooling 264 226 100 213 190 9-11 years schooling 448 447 100 362 326 12+ years schooling 927 995 100 832 655 Source: PNAD Surveys, IPEA-World Bank calculations. Men are assumed to retire after 35 years of service, and women after 30 years. * Using weights derived from nationwide PNAD employment distribution data. ** Capped at 10 times the minimum monthly salary level, assumed to be R$100 in September 1995 n.a. sign_fistathsa le was too small to ri reliable estimation of ~ e nene-earnins profiles. 5. PUBLIC-PRIVATE DIFFERENCES IN JOB SECURITY Table 11 presents the average tenure worker-specific attributes. Not surprisingly, on the current job in months, for the seven it appears that public sector workers have metropolitan areas. The table presents the more job security, as measured by the tenure average tenure in the public and private at the current job, adjusting for sectors, the "raw" gap in average tenure, and characteristics such as age, education, and then the gap controlling for differences in gender. This is only a crude measure of job 24 Labor Market Prospects of Public Employees in Brazil security, because it does not measure This advantage of public sector expected tenure. Nevertheless, it suggests workers is strengthened by their lower that public sector workers, besides earning unemployment rates, classified by sector of higher wages in all states except Sao Paulo, previous employment. But this pattern is also have jobs that are more secure. reversed when worker-specific attributes are Estimates of public-private gaps in job controlled for. Combined with the finding security range between 50% for the federal for average tenure differences, this may district and 20-35% in the six states. indicate that while public sector workers are less likely to lose their jobs, they are more likely to stay unemployed if in fact they do. Table 11 Measures of the Public-Private Tenure and Unemployment Gap, By State, in Septmber 1995 Metropolitan Rego Public Private GpControlled Gap * ......................................... ....................... ....................... ......... ........................................................................ ................................... ...................................... ............................. Average Tenure (Months) Distrito Federal 119 57 63 46 Pernambuco 119 93 27 27 Bahia 108 96 12 16 Minas Gerais 104 78 26 18 Rio de Janeiro 132 68 64 49 Sio Paulo 114 66 48 31 Rio Grande do Sul 123 89 34 31 Average Unemployment Rates (Percent) Distrito Federal 1.0 4.9 -3.9 2.9 Pernambuco 1.2 5.2 -4.0 3.6 Bahia 1.8 5.4 -3.6 2.7 Minas Gerais 1.7 4.1 -2.5 1.2 Rio de Janeiro 1.1 3.7 -2.7 1.8 Sdo Paulo 1.5 5.2 -3.7 2.0 Rio Grande do Sul 2.3 5.0 -2.7 1.7 *Estimates are for differences for workers with identical characteristics.. 7. CONCLUSIONS In this paper, we have documented Gerais, Sao Paulo, Rio de Janeiro, and public-private differentials in employment Rio Grande do Sul) and federal district. and pay at both aggregate and disaggregated levels. This helps in identifying where public * For the country as a whole, and for each employment is relatively high and where, of these states (and the federal district), alternatively, the problem of high public we examined the share of public payroll costs is the result of private-public employment at different levels of differentials in compensation. government - federal, state, and municipal, and public enterprises. * Using PNAD surveys, we examined * For the country as a whole, and for these public-private employmem differences in selected states, we examined public and selected states (Pernambuco, Minas Labor Market Prospects of Public Employees in Brazil 25 private employment by type of contract closely with contributions, or reforms that (e.g., com carteira, or sem carteira). lower artificially high turnover rates and increase private sector job stability (e.g., Under the constraints imposed by the social security reforms that lower the 1988 Constitution, state and federal informality of employment or a redesign of governments have relied largely on the fundo garantia por tempo de servico - incentives to tenured public employees to FGTS) would reduce private-public voluntarily leave government employment differentials in earnings, pensions, and job (see Carneiro and Gill, 1997). In most cases, stability respectively, and make it easier to the size of "optimal" severance packages reduce civil service employment. Similarly, required for employees to leave depends not an administrative reform bill for the public only on the level of public earnings, but on sector that enforces longer working hours, public-private differences in earnings, requires civil servants to make contributions pensions, job stability, and other benefits. for pensions at the same rate as for INSS Much of the discussion surrounding fiscal benefits, or eliminates tenure, would also adjustment, however, has focussed only on reduce public-private differentials in the former. This paper's results can assist in earnings, pensions, and job stability and shifting the focus to the latter: make it easier to reduce public employment. - Based on nationwide household surveys, Focussing on public-private we compute public-private differences in differentials also helps to illustrate that there monthly earnings, adjusting for worker is more than one way to achieve the same characteristics such as education, age, objective. Thus, the incentives for public tenure, sex, and race. sector employees to give up government jobs * Using experience-earnings profiles, and can be made stronger by reducing public rules for determining pensions, we sector earnings, pensions, and job stability; compute public-private differentials in or by increasing private earnigs, formality pensions, of employment, or job stability. * Using tenure in current job, we estimate Finally, the analysis can also be used public-private differentials in job to distinguish the effects of policy measures stability. that require constitutional reforms, and those that are feasible under current legislative This approach lends itself easily to conditions. Thus, the severance packages policy discussions. Labor market reforms that are necessary and sufficient to that reduce the cost of labor and raise the compensate workers for possible losses in demand for labor (e.g., by lowering payroll earnings and pensions can be estimated. tax rates), or social security reforms that make INSS pension benefits conform more REFERENCES Carneiro, Francisco G., and Indermit S. Gill. 1997. "Effectiveness and Financial Costs of Voluntary Severance Programs in Brazil: 1995-1997." Economic Note no. 25, Country Department I, Latin America Region, The World Bank, Washington, DC, USA. Annex Table 1: The spatial distribution of the labor force, average wage and log-wages Average wage Loa-wanes Relative wage gap between Public and Private sectors State Share of Share of baseline: Wage baseline: Wage Gap in log- Gap in public sector private sector Public sector Private sector Public sector Private sector in public sector in private sector average-wages average-log-wages (Wa) (Wb) In(Wa) ln(Wb) (Gla) (Gib) (G2) (G3) Acre 0.54 0.14 647.62 463.26 5.95 5.58 0.28 0.40 0.34 0.37 Alagoas 2.05 1.48 440.74 254.61 5.32 4.94 0.42 0.73 0.55 0.38 Amnazonas 1.63 0.98 496.75 399.07 5.69 5.49 0.20 0.24 0.22 0.20 Amapa 0.43 0.13 643.56 393.87 6.15 5.57 0.39 0.63 0.49 0.58 Bahia 7.04 7.21 407.43 243.05 5.40 4.92 0.40 0.68 0.52 0.48 Ceara 3.90 3.93 453.16 217.68 5.24 4.81 0.52 1.08 0.73 0.43 Distrito Federal 2.80 1.02 1278.80 531.92 6.80 5.74 0.58 1.40 0.88 1.06 Espirito Santo 1.76 1.83 673.80 364.07 6.02 5.36 0.46 0.85 0.62 0.66 Goias 2.93 3.06 517.03 325.79 5.64 5.27 0.37 0.59 0.46 0.37 Maranbho 2.75 2.93 385.49 165.55 5.22 4.50 0.57 1.33 0.85 0.72 Minas Gerais 10.06 11.45 544.77 356.48 5.80 5.29 0.35 0.53 0.42 0.51 Mato Grosso do Sul 1.49 1.29 574.65 385.16 5.89 5.40 0.33 0.49 0.40 0,49 Mato Grosso 1.71 1.53 577.06 381.56 5.92 5.45 0.34 0.51 0.41 0.47 Para 2.14 1.66 578.72 313.76 5.72 5.24 0.46 0.84 0.61 0.48 Paraiba 2.74 1.69 489.01 220.80 5.42 4.82 0.55 1.21 0.80 0.60 Pernambuco 3.95 4.20 468.04 255.00 5.51 5.02 0.46 0.84 0.61 0.49 Piaul 1.90 1.30 411.03 163.02 5.31 4.54 0.60 1.52 0.92 0.77 Parana 5.70 6.08 610.80 436.67 5.93 5.51 0.29 0.40 0.34 0.42 Rio de Janeiro 10.42 9.42 748.31 468.98 6.16 5.61 0.37 0.60 0.47 0.55 RioGrandedoNorte 2.20 1.41 525.48 214.76 5.53 4.87 0.59 1.45 0.89 0.66 0 Rondonla 0.96 0.49 796.50 463.24 6.24 5.55 0.42 0.72 0.54 0.69 Roraima 0.34 0.09 507.97 554.52 5.98 5.86 -0.09 -0.08 -0.09 0.12 Rio Grande do Sul 6.70 6.85 674.54 452.47 6.11 5.57 0.33 0.49 0.40 0.54 Santa Catarina 2.62 3.67 701.70 482.64 6.12 5.70 0.31 0.45 0.37 0.42 Sergipe 1.17 0.82 438.94 231.66 5.44 4.92 0.47 0.89 0.64 0.52 Sao Paulo 20.04 25.35 760.47 610.52 6.22 5.92 0.20 0.25 0.22 0.30 Brazil 99.97 100.01 623.44 417.08 5.88 5.43 0.33 0.49 0.40 0.45 Counter-factual public ( ... ... 621.97 ... 5.89 ... 0.34 0.58 0.44 0.46 Counter-factual private. ... ... ... 403.91 ... 5.39 0.36 0.63 0.47 0.48 Source: Constructed based on information of Pesquisa por Amostra de Domicilios (PNAD) -1995. Note (1) - Counter-factual simulation: Overall gap if the spatial distribution of public employment were identical to the spatial distribution of private employment (using spatial private shares as weights). (2) - Counter-factual simulation: Overall gap if the spatial distribution of private employment were identical to the spatial distribution of public employment (using spatial public shares as weights). o C4 'I Annex Table 2: Alternative measures for the states and metropolitan areas - 1995 Average wage Loc-wages Relative wage aao between Dublic and private sectors P baseline: Wage baseline: Wage Gap in log- Gap In average o Public sector Private sector Public sector Private sector in public sector in private sector average-wages log-wages (Wa) (Wb) ln(Wa) In(Wb) (Gla) (Gib) (G2) (G3) Brazil 623.44 417.08 5.88 5.43 0.33 0.49 0.40 0.45 Distrito Federal 1278.80 531.92 6.80 5.74 0.58 1.40 0.88 1.06 Metropolitan area Pernambuco State 468.04 255.00 5.51 5.02 0.46 0.84 0.61 0.49 Metropolitan area (Recife) 650.61 323.16 5.96 5.25 0.50 1.01 0.70 0.71 Bahia State 407.43 243.05 5.40 4.92 0.40 0.68 0.52 0.48 Metropolitan area (Salvador 662.41 381.24 5.92 5.26 0.42 0.74 0.55 0.66 Minas Gerals State 544.77 356.48 5.80 5.29 0.35 0.53 0.42 0.51 Metropolitan area (Belo Hor 798.66 443.45 6.21 5.56 0.44 0.80 0.59 0.65 W Rio de Janeiro State 748.31 468.98 6.16 5.61 0.37 0.60 0.47 0.55 Metropolitan area (RiodeJ 816.36 511.19 6.27 5.70 0.37 0.60 0.47 0.56 Sao Paulo State 760.47 610.52 6.22 5.92 0.20 0.25 0.22 0.31 Metropolitan area (Sao Pau 838.70 730.46 6.33 6.11 0.13 0.15 0.14 0.22 Rio Grande do Sul State 674.54 452.47 6.11 5.57 0.33 0.49 0.40 0.54 Metropolitan area (Porto Al 884.18 557.38 6.37 5.78 0.37 0.59 0.46 0.59 Source: Constructed based on information of Pesquisa por Amostros de Domicillo (PNAD) - 1995. 28 Labor Market Prospects of Public Employees in Brazil Annex Table 3a: Estimates oT iog wage-gap controlled for defferences in observed characteristics composition effects Standardized Non-Standardized Standardized Non-Standardized Region General Basic General Basic General Basic General Basic (1) (2) (3) (4) (1) (2) (3) (4) DIstnto Federal Gender 0.00 0.00 0.00 0.00 0.28 0.22 0.41 0.33 race 0.02 0.02 0.02 0.02 2.76 2.57 2.63 2.44 education 0.54 0.54 0.50 0.52 79.28 77.91 77.13 76.19 age 0.12 0.13 0.13 0.14 17.68 19.30 19.83 21.05 total 0.68 0.69 0.65 0.68 100.00 100.00 100.00 100.00 Recife Gender -0.03 -0.03 -0.04 -0.04 -4.46 -4.40 -6.60 -6.48 race 0.01 0.01 0.01 0.01 1.67 1.70 2.46 2.25 education 0.50 0.52 0.49 0.50 88.32 87.87 87.70 87.26 age 0.08 0.09 0.09 0.10 14.47 14.83 16.44 16.97 total 0.57 0.59 0.56 0.58 100.00 100.00 100.00 100.00 Salvador Gender -0.04 -0.04 -0.06 -0.06 -7.01 -6.99 -10.50 -10.32 race 0.04 0.04 0.04 0.04 6.36 6.45 6.40 6.28 education 0.47 0.47 0.43 0.44 78.36 78.29 76.81 76.68 age 0.13 0.13 0.15 0.16 22.30 22.25 27.28 27.35 total 0.60 0.06 0.56 0.57 100.00 100.00 100.00 100.00 Belo Hodzont. Gender -0.06 -0.05 -0.08 -0.08 -9.13 -8.48 -14.25 -13.39 race 0.03 0.03 0.03 0.03 5.36 5.00 5.26 4.82 education 0.51 0.53 0.48 0.49 83.35 83.28 84.75 84.44 age 0.13 0.13 0.14 0.14 20.42 20.21 24.24 24.12 total 0.61 0.63 0.57 0.59 100.00 100.00 100.00 100.00 Rio de Janeinm Gender -0.01 -0.01 -0.01 -0.01 -1.18 -1.19 -1.96 -1.95 race 0.01 0.01 0.01 0.01 2.12 2.09 2.04 2.03 education 0.45 0.44 0.43 0.43 85.31 85.15 84.63 84.46 age 0.07 0.07 0.08 0.08 13.75 13.96 15.29 15.46 total 0.52 0.52 0.50 0.50 100.00 100.00 100.00 100.00 Sao Paulo Gender -0.06 -0.06 -0.09 -0.09 -11.82 -12.08 -20.09 -20.07 race 0.01 0.01 0.01 0.01 1.13 1.12 1.20 1.18 education 0.39 0.39 0.37 0.37 83.48 83.69 87.45 87.23 age 0.13 0.13 0.13 0.14 27.20 27.27 31.44 31.66 total 0.47 0.47 0.43 0.43 100.00 100.00 100.00 100.00 Porto Alegre Gender -0.03 -0.03 -0.05 -0.05 -5.31 -5.25 -9.32 -9.00 race 0.00 0.00 0.00 0.00 0.05 0.05 0.07 0.07 education 0.52 0.52 0.50 0.50 93.18 92.66 94.11 93.37 age 0.07 0.07 0.08 0.08 12.08 12.54 15.15 15.57 total 0.56 0.56 0.53 0.54 100.00 1 00.00 100.00 100.00 Source: Constructed based on infotrmation of Pesquisa por Amostra de Domicilio (PNAD) - 1995. Labor Market Prospects of Public Employees in Brazil 29 Annex Table 3b: Estimates of log wage-gap controlled for defferences in observed characteristics Standardized Non-Standardized Standardized Non-Standardized Region General Basic General Basic General Basic General Basic (1) (2) (3) (4) (1) (2) (3) (4) Dlstrlto Federal Gender 0.00 0.00 0.00 0.00 0.24 0.19 0.35 0.28 race 0.02 0.02 0.02 0.02 2.29 2.26 2.15 2.10 education 0.54 0.53 0.50 0.51 69.68 69.91 67.21 67.45 age 0.10 0.11 0.11 0.12 13.47 14.81 15.09 15.84 Tenure 0.11 0.10 0.11 0.11 14.33 12.82 15.20 14.32 Total 0.77 0.76 0.74 0.76 100.00 100.00 100.00 100.00 Recife Gender -0.03 -0.03 -0.04 -0.04 -3.98 -3.90 -5.83 -5.69 race 0.01 0.01 0.01 0.01 1.45 1.49 2.13 1.96 education 0.50 0.51 0.49 0.50 79.26 78.88 77.61 77.18 age 0.07 0.07 0.08 0.08 11.32 11.42 12.70 12.87 Tenure 0.08 0.08 0.08 0.09 11.95 12.11 13.38 13.68 total 0.63 0.65 0.63 0.65 100.00 100.00 100.00 100.00 Salvador Gender -0.04 -0.04 -0.06 -0.06 -5.88 -6.00 -8.96 -8.96 race 0.04 0.04 0.04 0.04 5.43 5.63 5.52 5.50 educabon 0.46 0.46 0.43 0.43 66.58 67.62 65.88 66.66 age 0.11 0.11 0.13 0.13 15.45 15.63 20.17 20.39 Tenure 0.13 0.12 0.11 0.11 18.42 17.03 17.39 16.41 Total 0.70 0.69 0.65 0.65 100.00 100.00 100.00 100.00 BelAo Horizonte Gender -0.05 -0.05 -0.08 -0.08 7.96 -7.35 -12.08 -11.32 race 0.03 0.03 0.03 0.03 4.75 4.47 4.50 4.18 educabon 0.50 0.51 0.47 0.48 73.30 72.92 71.77 71.30 age 0.11 0.11 0.12 0.12 16.24 15.63 18.44 17.91 Tenure 0.09 0.10 0.11 0.12 13.67 14.34 17.37 17.93 Total 0.68 0.71 0.66 0.68 100.00 100.00 100.00 100.00 Rio de Janeiro Gender -0.01 -0.01 -0.01 -0.01 -0.94 -0.96 -1.56 -1.58 race 0.01 0.01 0.01 0.01 1.68 1.70 1.58 1.63 education 0.44 0.43 0.42 0.41 70.22 70.98 68.68 69.49 age 0.06 0.06 0.06 0.06 9.62 9.72 10.63 10.70 Tenure 0.12 0.11 0.13 0.12 19.42 18.55 20.67 19.76 Total 0.62 0.61 0.61 0.60 100.00 100.00 100.00 100.00 Sao Paulo Gender -0.05 -0.05 -0.08 -0.08 -9.54 9.76 -17.19 -16.18 race 0.00 0.00 0.00 0.00 0.91 0.92 0.93 0.93 education 0.38 0.38 0.36 0.36 71.41 71.53 72.49 72.22 age 0.11 0.11 0.11 0.11 20.43 20.28 22.71 22.67 Tenure 0.09 0.09 0.10 0.10 16.78 17.03 20.06 20.36 Total 0.54 0.53 0.50 0.50 100.00 100.00 100.00 100.00 Porto Alegre Gender -0.03 -0.03 -0.05 -0.05 -4.09 -4.09 -7.38 -7.20 race 0.00 0.00 0.00 0.00 0.04 0.04 0.06 0.06 education 0.51 0.50 0.48 0.48 76.93 77.03 76.42 76.46 age 0.06 0.06 0.07 0.07 9.03 9.28 11.36 11.62 Tenure 0.12 0.12 0.12 0.12 18.09 17.74 19.54 19.07 Total 0.66 0.65 0.63 0.63 100.00 100.00 100.00 100.00 Source: Constructed based on information of Pesquisa porAmostra de Domicilio (PNAD) - 1995.. Labor Market Prospects of Public Employees in Brail Annex Table 4a: Estimates of the log wage-gap controlled for differences in observed characteristics Standardized Non-standardized General Basic General Basic (3)-1) (4)-(2) (1) (2) (3) (4) Distrito Federal 0.50 0.49 0.42 0.40 -0.08 -0.09 Recife 0.28 0.27 0.15 0.14 -0.13 -0.13 Salvador 0.19 0.19 0.09 0.08 -0.10 -0.11 Belo Horizonte 0.16 0.14 0.05 0.03 -0.11 -0.11 RiodeJaneiro 0.11 0.11 0.05 0.05 -0.06 -0.06 Sao Paulo -0.14 -0.14 -0.22 -0.21 -0.08 -0.07 Porto Alegre 0.13 0.13 0.05 0.05 -0.08 -0.08 Source: Constructed based on informabon of Pesquisa par Amostros de Domicilio (PNAD) - 1995. Annex Table 4b: Estimates of the log wage-gap controlled for differences in observed characteristics Standardized Non-standardized General Basic General Basic (3)-(1) (4)-(2) (1) (2) (3) (4) Distrito Federal 0.41 0.42 0.33 0.32 -0.08 -0.10 Recife 0.22 0.20 0.08 0.07 -0.14 -0.13 Salvador 0.09 0.10 0.00 0.00 -0.09 -0.10 Belo Horizonte 0.09 0.07 -0.04 -0.06 -0.13 -0.13 Rio de Janeiro 0.01 0.02 -0.05 -0.04 -0.06 -0.06 Sao Paulo -0.21 -0.20 -0.29 -0.28 -0.08 -0.08 Porto Alegre 0.04 0.04 -0.05 -0.04 -0.09 -0.08 Source: Constructed based on information of Pesquisa por Amostros de Domicilio (PNAD) - 1995. Labor Market Prospects of Public Employees in Brazil 31 Annex Table 5a: Direct comparison of public and private jobs The results of the counter-factuals Standardized Non-Standardized Public Private Wage-Gap Public Prvate Wage-Gap Distrito Federal 0.45 -0.10 0.55 0.46 -0.02 0.48 Recife -0.29 -0.58 0.29 -0.40 -0.55 0.15 Salvador -0.15 -0.36 0.21 -0.25 -0.34 0.09 Belo Horizonte -0.07 -0.23 0.16 -0.14 -0.19 0.05 Rio de Janeiro -0.15 -0.25 0.10 -0.17 -0.22 0.05 Sao Paulo 0.00 0.14 -0.14 0.00 0.22 -0.22 Porto Alegre -0.02 -0.18 0.16 -0.04 -0.12 0.08 Source: Constructed based on information of Pesquisa por Amostros de Domicilio (PNAD) - 1995. Annex Table 5b: Direct comparison of public and private jobs The results of the counter-factuals Standardized Non-Standardized Public Private Wage-Gap Public Private Wage-Gap Distrito Federal 0.43 -0.04 0.47 0.43 0.05 0.38 Recife -0.31 -0.55 0.24 -0.42 -0.51 0.09 Salvador -0.17 -0.29 0.12 -0.27 -0.28 0.01 Belo Horizonte -0.10 -0.18 0.08 -0.17 -0.12 -0.05 Rio de Janeiro -0.17 -0.19 0.02 -0.20 -0.16 -0.04 Sao Paulo 0.00 0.21 -0.21 0.00 0.29 -0.29 Porto Alegre -0.04 -0.11 0.07 -0.05 -0.04 -0.01 Source: Constructed based on information of Pesquisa por Amostros de Domicilio (PNAD) - 1995. Annex Table 6a: The controlled log-wage gap for the workerst in private sector in metropolitan Brazil Standardized Region Educaion Age Tenure 0 years 4 years 8 years 11 years 30 years 35 years 40 years 42 years 1 year 2 years 5years 10years Distrito Federal 0.42 0.40 0.38 0.37 0.32 0.38 0.43 0.47 0.43 0.42 0.38 0.33 Recife 0.12 0.19 0.25 0.30 0.20 0.25 0.28 0.29 0.25 0.25 0.25 0.25 Salvador 0.16 0.17 0.18 0.19 0.18 0.18 0.18 0.18 0.23 0.22 0.18 0.13 Belo Horizonte -0.30 -0.21 -0.12 -0.05 -0.10 -0.12 -0.13 -0.14 -0.15 -0.15 -0.12 -0.07 Rio de Janeiro 0.07 0.06 0.05 0.04 0.05 0.05 0.05 0.06 0.08 0.07 0.05 0.02 Sao Paulo -0.14 -0.16 -0.19 -0.21 -0.15 -0.19 -0.22 -0.25 -0.21 -0.20 -0.19 -0.17 Porto Alegre 0.09 0.08 0.06 0.05 0.01 0.06 0.11 0.16 0.09 0.08 0.06 0.02 Source: Constructed based on information of Pesquisa por Amostra do Domicilio (PNAD) - 1995. Note. "Typical worker": white male with 8 years of schooling, who is 35 years old and worked at 5 years in his current job. Annex Table 6a: Controlled log-wage gap (cont'd) Standardized Region Sex Race Sex and race male female white non-white female and non-white Distrito Federal 0.38 0.40 0.38 0.37 0.67 Recife 0.25 0.19 0.25 0.30 0.17 Salvador 0.18 0.17 0.18 0.19 0.14 Belo Horizonte -0.12 -0.21 -0.12 -0.05 0.07 Rio de Janeiro 0.05 0.06 0.05 0.04 0.04 Sao Paulo -0.19 -0.16 -0.19 -0.21 -0.22 Porto Alegre 0.06 0.08 0.06 0.05 0.12 Annex Table 6b: The controlled log-wage gap for the workers in private sector in metropolitan Brazil Non-standardized Region Education Age' Tenure 0 years 4 years 8 years 11 years 30 years 35 years 40 years 42 years 1 year 2 years 5years 10years Distrito Federal 0.16 0.19 0.21 0.22 0.15 0.21 0.26 0.30 0.23 0.23 0.21 0.18 Recife -0.03 0.02 0.07 0.11 0.01 0.07 0.12 0.14 0.07 0.07 0.07 0.07 Salvador -0.04 -0.01 0.03 0.05 0.00 0.03 0.04 0.06 0.06 0.05 0.03 -0.01 Belo Horizonte -0.40 -0.32 -0.25 -0.19 -0.24 -0.25 -0.25 -0.24 -0.28 -0.27 -0.25 -0.21 Rio de Janeiro -0.01 -0.03 -0.04 -0.06 -0.04 -0.04 -0.03 -0.01 -0.02 -0.02 -0.04 -0.08 Sao Paulo -0.22 -0.26 -0.31 -0.35 -0.28 -0.31 -0,.33 -0.35 -0.33 -0.33 -0.31 -0.29 Porto Alegre -0.04 -0.06 -0.07 -0.09 -0.13 -0.07 -0.02 0.04 -0.04 -0.05 -0.07 -0.12 Source: Constructed based on information of Pesquisa por Amostra do Domicilio (PNAD) - 1995. Note. "Typical worker": white male with 8 years of schooling, who Is 35 years old and worked at 5 years in his current job. Annex Table 6b: Controlled log-wage gap (cont'd) Non-standardized Region Sex Race Sex and race male female white non-white female and non-white Distrito Federal 0.21 0.52 0.21 0.23 0.54 Recife 0.07 0.03 0.07 0.11 0.07 Salvador 0.03 0.02 0.03 0.04 0.03 Belo Horizonte -0.25 -0.09 -0.25 -0.19 -0.03 Rio de Janeiro -0.04 -0.02 -0.04 -0.05 -0.03 Sao Paulo -0.31 -0.26 -0.31 -0.30 -0.24 Porto Alegre -0.07 0.04 -0.07 -0.02 0.09 Annex Table 7a: Estimnates for the Internal Comnposition of the Public Sector Standardized Non-standardized Region Log Log Proportions Log Log Log Log Proportions Log Log wage-gap Level (S) wage-p wage-gap wage-gap Lrvel (%) wage-gp wage-gap Dirito Federd 0.15 0.03 0.12 0.02 Public Servants - Federal 0.66 1.77 3.54 0.61 6.93 3.54 Non Public Servanta with a signed working card - Federal 0.41 1.52 3.47 0.31 6.63 3.47 Public Servant - Stale 0.54 1.65 10.32 0.42 6.74 10.32 Non Public Servanta with a uigned working card - State 0.50 1.61 2.64 0.40 6.72 2.64 Public Servants- Municipal 0.00 111 0.00 0.00 6.32 0.00 Non Public Servants with a signed working card - Municipal 0.00 1.11 0.00 0.00 6.32 0.00 Non Public Serants without a signed working card - Federal State, Municipal 0.27 1.37 1.75 0.06 6.39 1.75 Military Psonnel 0.07 1.13 2.61 0.06 6.33 2.61 Recife 0.04 0.02 0.02 0.01 Public Servants - Federl 0.63 1.23 1.54 0.60 6.42 1.54 Non Publc Servants with a signed working card - Federal 0.72 1.37 0.96 0.63 6.50 0.96 Public Scrvants - State 0.23 0.93 4.93 0.15 5.97 4.98 Non PubEic Servants with a signed working card - State 0.37 1.02 1.05 0.23 6.10 1.05 Publc Servants - Municipal 0.04 0.69 2.03 -0.13 5.69 2.03 Non Public Servants with a signed working card - Municipal 0.00 0.65 1.11 -0.20 5.62 1.11 Non Public Servants without a signed working card - Federal State, Municipal 0.04 0.69 1.50 -0.13 5.64 1.50 Mitary Pesonnel 0.32 0.97 0.82 0.27 6.09 0.32 Salvador 0.03 0.01 0.01 0.00 Publc Servants - Federal 0.66 1.55 1.53 0.53 6.63 1.53 Non Public Servants with a signed working card - Federal 0.64 1.53 1.45 0.60 6.65 1.45 Publc Servants - State -0.02 0.37 5.23 -0.12 5.93 5.23 Non Pubc Scrvants with a signed working card - State 0.11 1.00 2.16 0.03 6.13 2.16 Publc Servants - Municipal 0.11 1.01 1.30 0.00 6.05 1.30 Non Public Servants with a signed working card - Municipal 0.11 1.00 1.76 -0.04 6.01 1.76 Non Pubfic Servants without a signed working card - Fedcral State, Municipal 0.17 1.07 1.30 -0.12 5.93 1.30 Military Personnel 0.56 1.45 0.44 0.52 6.57 0.44 8do lorirzoat 0.02 0.01 0.00 0.00 Public Servants - Federal 0.42 1.43 1.33 0.34 6.52 1.33 o Non Public Sermnts with a signed working card - Federal 0.37 1.33 0.33 0.29 6.47 0.83 Public Servants - State 0.09 1.10 4.55 -0.04 6.14 4.55 Non Publc Semnts with a signed working card - State 0.23 1.29 1.55 0.26 6.44 1.55 % Publc Servants - Municipal 0.06 1.07 1.74 -0.06 6.12 1.74 C Non Public Servants with a signed working card - Municipal 40.04 0.97 0.33 -0.10 6.09 0.13 3i Non Public Servants without a signed working card - Federal State, Municipal -0.05 0.96 0.33 -0.30 5.89 0.33 Military Personnel 0.09 1.09 0.23 0.05 6.23 0.23 -. Antnex Table 7a (continued): Estimates for the Internal Compositfon of the PubUc Sector Standardihed Non-standardized d Region Log Log Psportionr Log Log Log Log proportions Log Leg wage-gap Level (%) wage-gp Wage-gap wage-gap Level (%) wage-gap Wup- a. *5* ,. *aa Rio de Janeiro 0.02 0.00 0°01 0-00 Public Serans - Federal 0.38 1.31 2.59 0.48 6.70 2.59 Non Public Servan4 with a signed working card - Federal 0.34 1.28 1.30 0.27 6.49 1.30 Public Servantb- State 0D06 0.99 3.64 -0.10 6.12 3.64 NonPublicServantswithasignedworkingcard-State 0.29 1.23 1.05 031 6.53 1.05 Public Servants - Municipal -0.15 0.79 2.42 005 6.27 2.42 Non Public Servants with a signed working card - Municipal -0.39 0.54 0.94 -0.1B 6.04 0.94 Non Public Servants without a signed working card - Federal, State. Municipal 0.08 1.02 0.77 .0.21 6.01 0.77 Militry Personnel 0.17 1.10 1.93 -0.14 6.08 1.93 Sb Peale 0.01 -0.01 -0.02 -.002 Public Servants - Federal 0,36 1.71 0.39 0.22 6.78 0.39 Non Public Sernnts with c signed working card - Federal 0.03 1.39 0.43 -0.08 6.48 0.43 Q Public Servants- State -0.16 1.19 3.63 -0.23 6.34 3.63 Non Public Servants with a signed working card - State 0.05 1.41 1.34 -0.01 6.55 1.34 r Public Serants - Municipal .0.25 1.10 1.72 -0.32 6.25 1.72 a NonPublicServants withasigniedworkingcard-Municipal .0.15 1.21 0.88 -0.21 6.36 0.88 N Non Public Servants without asigned working card - Fderal. State. Municipal -0.41 0.95 0.69 -0.55 6.02 0.69 Military Pesonnel 40.45 0.90 0.16 -0.46 6.10 0.16 Porto Alegr 0.02 0.01 0.01 000 Public Servnts - Federal 0.50 1.50 0.99 0.48 6.10 0.99 Non Public Servant with a signed workng card - Fedenl 0.37 1.37 1.88 0 27 6.49 1.88 Public Sernnts- State -0.01 0.99 3.29 -0.10 6.12 329 NonPublic Scmant withasignedworkingcard-State 0.36 1.36 1.77 0.31 6.53 1.77 Public Servants - Municipal 0.16 1.16 1.94 0.05 6.27 1.94 Non Public Servants withasigned working card -Municipal -0.07 0.93 1.47 -0.18 6.04 1.47 Non Public Servants without a signed working card - Federal. State. Municipal 4.09 0.91 1.33 4-21 6.01 1.33 Military Personnel -0.18 0.82 0.46 -0.14 6.08 046 Source: Constructed based on information ofPesquisa por Amostra de Domicilio (PNAD) - 1995. Note: * Estiuates obtained not using tenure as independent variable, 0 Using the own region public sector crnployment shares (propotions) as weights. * Usig Sb Paulo public scr employnent shares (proportion) a wights. Annex Table 7b: Estimates for the Internal Compositlon or the Public Sector^ Standardized Non-standardized Region LOg Log Proportions Log Log Log Log Proportions Log Log wage-gap Levl wage-gap wage-pp wage-gpP Level (%) wage-pp wge-p Di&feto Federal 0.12 0.03 0.09 0.02 Public Servanta - Federal 0.57 1.66 8.54 0.51 6.31 3.54 Non Public Servants with a signed working card - Federal 0.32 1.41 3.47 0.21 6.51 3.47 Public Servants - State 0.43 1.57 10,32 0.35 6.65 10.32 Non Publc Servants with a signed working card - State 0.44 1.52 2.64 0.33 6.63 2.64 Public Servants - Municipal 0.00 1.09 0.00 0.00 6.30 0.00 Non Public Servants with a signed working card - Municipal 0.00 1.09 0.00 0.00 6.30 0.00 Non Public Servants without a signed working card - Federal Stte, Municipal 0.27 1.36 1.75 0.07 6,37 1.75 Military Personnel -0.09 1.00 2.61 -0.11 6.19 2.61 Recife 0.03 0.02 0.01 0.00 Public Servantsn Federal 0.55 1.1 1.54 0.51 6.32 1.54 Non Publc Servants with a signed working card - Federal 0.64 1.27 0.96 0.5S 6.39 0.96 Publc Servants - State 0.20 0.34 4.93 0.07 5.S7 4.98 Non Public Servants with a signed working card - State 0.29 0.93 1.05 0.20 6.00 1.05 Public Servants - Municipal -0.03 0.61 2.03 -0.21 5.60 2.03 Non Public Servants witha signed workingcard-Municipal -0.03 0.61 1.11 -0.23 5.57 1.11 Non Public Servants without a signed working card - Federal State, Municipal 0.05 0.69 1.50 -0.17 5.64 1.50 Military Personnel 0.26 0.39 0.32 0.20 6.00 0.32 Salvada 0.02 0.00 0.00 40.01 t Public Servantn - Federal 0.53 1.42 1.53 0.47 6.51 1.53 O Non Public Servants with a signed working card - Federal 0.54 1.42 1.45 0.51 6.56 1.45 Public Servants - State -0.15 0.73 5.23 -0.24 5.81 5.23 Non Pubic Servants with a signed working card - Stte 0.03 0.92 2.16 0.01 6.05 2.16 Public Servants - Muricipal 0.03 0.92 1.30 -0.03 5.97 1.30 Non Publc Servants with a signed working card - Municipal 0.07 0.96 1.76 -0.07 5.97 1.76 Non Public Servants without a signed working card - Federal, State, Municipal 0.21 1.09 1.30 -0.09 5.95 1.30 Miltitry Personnel 0.45 1.33 0.44 0.42 6.46 0.44 Belo Horirenr 0.01 0.00 -0.01 -0.01 Public Servants - Federal 0.33 1.33 1.33 0.23 6.41 1.33 Non Publc Servants with a signcd working card - Federal 0.27 1.28 0.33 0.17 6.35 0.33 Publc Servants - State 0.00 1.00 4.55 -0.15 6.03 4.55 Non Publc Servants with a signcd working card - State 0.19 1.19 1.55 0.15 6.33 1.55 Publc Servants - Municipal 0.02 1.03 1.74 -0.11 6.07 1.74 Non Public Servants with a signed working card - Municipal -0.11 0.89 0.33 -0.19 5.99 0.3 Non Publc Servants without a signed working card - Federal, State, Municipal 40.04 0.96 0.33 -0.29 5.9 0.38 Military Personnel -0.04 0.96 0.23 -0.10 6.07 0.23 - . Annex Table 7b (continued): Estimates for the Internal Composition of the Public Sector * Standardized Non-standardized Region Log Log Proportions Log Log Log Log Proportions Log Log wage-gap Level (%) wage-gap wage-gap wage-gap Level (%) wage-gap Wage-upp 5* .** Y555 Rio de Janiro 0.00 0.00 40.01 -0.01 Public Servants - Federal 0.26 1.1 2.59 0.26 6.36 2.59 '0* Non Public Servants with a signed working card - Federal 0.24 1.16 1.30 0.24 6.34 1.30 Public Servants - State -0.04 0.88 3.64 -0.13 5.97 3.64 Non Public Servants with a signed working card - State 0.20 1.12 1.05 0.21 6.31 1.05 Public Servants - Municipal -0.22 0.70 2.42 -0.35 5.75 2.42 Non Public Servants with a signed working card - Municipal -0.43 0.49 0.94 -0.49 5.61 0.94 Non l'ublic Servants without a signed working card - Federal, State, Municipal 0.10 1.02 0.77 -0.07 6.03 0.77 Military Personnel O OS 1.00 1.93 0.05 6.15 1.93 SAO Paulo -0.02 -0.02 -0.03 -0.03 { Public Servants -Federal 0.31 1.64 0.39 0.17 6.71 0.39 Non Public Servants with a signed working card - Federal -0.06 1.27 0.43 -0.19 6.35 0.43 Public Scrvant - State -0.24 1.09 3.63 -0.31 6.23 3.63 Non Public Servants with a signed working card - State -0.03 1.30 1.34 -0.10 6.44 1.34 Z Public Servantsb Municipal -0.33 1.00 1.72 -0.41 6.13 1.72 N Non Public Servants with a signed working card - Municipal -0.15 1.g I.gg -0.21 6.33 0o.S Non Public Servants without a signed working card - Federal, State, Municipal -0.40 0.93 0.69 -0.54 6.00 0.69 MilitaryPersonnel -0.50 0.83 0.16 -0.52 6.01 0.16 Porto Alegrc 0.01 0.00 -0.01 -0.01 Public Servants - Federal 0.38 1.37 0.99 0.36 6.57 0.99 Non Public Servants witha signed working card - Federal 0.26 1.24 1.gg 0.16 6.37 1 BB Public Servants - State -0.13 0.86 3.29 -0.22 5.99 3.29 Non Public Scvants with a signed working card - State 0.20 1.19 1.77 0.15 6.36 1.77 Public Servants - Municipal O.08 1.07 1.94 -0.03 6.13 1.94 Non Public Seivants with a signed working card - Municipal -0.10 0.89 1.47 -0.21 6.00 1.47 Non Public Servants without a signed working card - Fcdcral, State, Municipal -0.06 0.93 1.33 -0. II 6.03 1.33 Military Personnel -0.26 0.72 0.46 -0.23 5.9S 0.46 Source: Constructed based on infornation of Pesquisa por Amostra de Doniicilio (PNAD) - 1995. Note: I Estirntes obtained using tenure as another independent variable. Using the own region public sector employment shares (proportions) as weights.