Decomposing Male Inequality Change in East Germany During Transition

Gang, Ira N.; Yun, Myeong-Su Article Decomposing Male Inequality Change in East Germany During Transition Schmollers Jahrbuch – Zeitschrift für Wirtschaftsund Sozialwissenschaften. Journal of Applied Social Science Studies Provided in Cooperation with: Duncker & Humblot, Berlin Suggested Citation: Gang, Ira N.; Yun, Myeong-Su (2003) : Decomposing Male Inequality Change in East Germany During Transition, Schmollers Jahrbuch – Zeitschrift für Wirtschaftsund Sozialwissenschaften. Journal of Applied Social Science Studies, ISSN 1865-5742, Duncker & Humblot, Berlin, Vol. 123, Iss. 1, pp. 43-53, https://doi.org/10.3790/schm.123.1.43 This Version is available at: https://hdl.handle.net/10419/292037 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Decomposing Male Inequality Change in East Germany During Transition By Ira N. Gang and Myeong-Su Yun* Abstract This paper studies the cause of the changes in male wage inequality in East Germany during its transition from a socialist to a market-oriented economic system. We are interested in how much of the change in the dispersion of wages can be explained by the changes in the characteristics of workers and how much can be explained by the changes in returns to the characteristics of workers. JEL Classification: D 30, J 30 1. Introduction During the course of East Germany’s economic transition, there was an apparently substantial widening in wage dispersion. Various inequality measures indicate that the wage inequality increased between 25 % and 61 % relative to the level in 1990 (Gang/Yun, 2002). 1 This paper studies the causes of the changes in wage inequality measured in terms of variance of log-wages, asking what factors explain the change in wage dispersion. We are interested in how much of the change in the dispersion of wages can be explained by the changes in the characteristics of workers and how much can be explained by the changes in returns to characteristics of workers. There is a small literature on inequality change in East Germany. Franz/ Steiner (2000) and Burda/Hunt (2001) address changes in distribution of hourlywages from 1990–1997, finding that inequality increases. Some papers Schmollers Jahrbuch 123 (2003) 1 Schmollers Jahrbuch 123 (2003), 43–54 Duncker & Humblot, Berlin * Myeong-Su Yun’s work was supported by the CIBC Program in Human Capital and Productivity, Department of Economics, University of Western Ontario. Part of this work was done while Myeong-Su Yun visited the Department of Economics, Rutgers University. An earlier version of this paper was presented at the 5 th International German Socio-Economic Panel Users Conference, July 3–4, 2002, Berlin, Germany. We especially appreciate the comments of Jennifer Hunt. 1Gang/Yun (2002) examines the Gini coefficient, the coefficient of variation, the Theil index, the log-wage differentials between top and bottom 10 % and the variance of log-wages. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.1.43 | Generated on 2023-04-04 12:31:36 44 Ira N. Gang and Myeong-Su Yun (e.g., Abraham/Houseman (1995), Hunt (2002), Krueger/Pischke (1995)) discuss the effects of the transition in terms of wage inequality and the gender wage gap. Biewen (2000) extensively analyzes income inequality changes (based on net monthly household income), finding increased inequality in East Germany after unification. Gang/Yun (2002) analyze both wage growth and the changes in wage dispersion in a unified framework, making comparisons between East and West Germany from 1990–2000. Generally, all of these papers find a widening dispersion of household income and wages. In this paper, we employ a newly developed Blinder-Oaxaca type inequality decomposition method for analyzing the change in wage inequality since East Germany’s unification with West Germany. Our inequality decomposition method (see Yun, 2002) allows us to find not only the gross contribution of each factor to the changes in wage inequality, but also the price and quantitative effects of each factor by utilizing information contained in the earnings equations. The standard Blinder-Oaxaca (Blinder, 1973; Oaxaca, 1973) decomposition explains wage differentials in terms of differences in individual characteristics (characteristics effect) and differences in the coefficients of the wage equations (coefficients effect). The methodological innovation introduced by Yun (2002), which is based on Oaxaca’s decomposition methodology, allows us to derive these types of effects for changes in wage dispersion, and overcomes some difficulties in the earlier methodologies proposed by Juhn/Murphy/Pierce (1993) and Fields (2001). We employ the 1990 through 2000 waves of the German Socioeconomic Panel (GSOEP), a comprehensive panel of household and individual data. 2 Collection in East Germany began in May 1990. We restrict our sample to men aged between 20 and 60 who are not in school or in formal occupational training, with real before tax wages (in 1995 DM) less than 100 DM per hour. We exclude the self-employed, those on maternity leave, in agriculture, and who were originally in the sample but moved from East to West Germany. For each wave, we perform our analysis on all men meeting these criteria (unbalanced sample). Figure 1 presents the East German male mean hourly wage rates and the variance of log-wages for each year from 1990 to 2000, normalized to 1990 = 1.00 for comparison purposes. 3 The growth in mean wages stands out, especially the doubling of mean wages from 1990 to 1994. Wage inequality measured by the variance of log-wages increases from 1990 to 2000 by 58.1 %. It is interesting that most of the increase in wage inequality occurred between the first two years (1990 to 1991), while the wage growth was achieved from Schmollers Jahrbuch 123 (2003) 1 2We use the international version of the GSOEP, which is a 95 % sample of the German version. For a full description, see http://www.diw.de/soep/soep.htm 3The inequality measures are constructed using population weights provided in the GSOEP data. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.1.43 | Generated on 2023-04-04 12:31:36 Decomposing Male Inequality Change in East Germany During Transition 45 1. Mean hourly wage rates in terms of 1995 constant German Marks. 2. Population weights given in the GSOEP data are used for calculation. 3. Standardized measures, 1990=1.00. Figure 1: Wage Growth and Changes in Male Wage Inequality (East Germany) 1990 to 1994. From 1991 to 2000 wage inequality fluctuates without much overall change. Under socialism we would expect that the bias toward egalitarianism would have suppressed wage inequality. Our simple calculations show that among men wage dispersion as well as absolute wages have increased during the transition. In this paper, we will study the sources of the changes in wage inequality: Have the changes in workers’ characteristics caused the increase in wage inequality?; Have the changes in returns in workers’ characteristics due to the changes in the economic system caused the widening of wage inequality? In the next section we outline our methodology. Section 3 discusses our decomposition results, and Section 4 concludes. 2. Explaining Changes in Inequality using Earnings Equation We are interested in explaining the change in wage dispersion (inequality) in East Germany that has occurred since unification. Yun (2002) develops a Schmollers Jahrbuch 123 (2003) 1 1 1.5 2 2.5 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Wages Variance ofLog-Wages OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.1.43 | Generated on 2023-04-04 12:31:36 46 Ira N. Gang and Myeong-Su Yun new decomposition method for the changes in wage inequality measured in terms of variance of log-wages utilizing the information contained in the earnings equation. Let wages be generated from the following regression equations (earnings equations) yA   0A  X k  k ÿ 1 k  1  kAxkA  eAand  1  yB   0B  X k  k ÿ 1 k  1  kBxkB  eB ; where yt  log  Yt  , and xkt ; etare the k th exogenous variable and residuals, respectively, and t  A ; B. A feature of the newly developed decomposition method for wage inequality is that it uses the information contained in the earnings equation, i.e., it utilizes the coefficients of the earnings equation. The method explains the changes in wage inequality in terms of characteristics effect, coefficients effect and residuals effect similar to the Blinder-Oaxaca decomposition for wage growth. From equation (1), we find the following identity,  2 y  P K ÿ 1 k  1  kxk ; y   e ; y, where  e ; y   2 eif OLS is used for estimation of the equation (1). Fields (2001) defines the relative factor inequality weight for a factor kusing the OLS estimate of the coefficient of the earnings equation as sk   kxk ; y  2 y   k 1  xk 1  xk ; y  y ; where  xkis the standard deviation of xkand  xk ; y   xk ; y  xk  y. Fields applies the relative factor inequality weight to study changes in wage inequality over time. Fields’ method, however, does not decompose the changes in wage inequality in terms of characteristics, coefficients and residuals effects. On the other hand, Juhn/Murphy/Pierce (1993) explain changes in wage inequality in terms of characteristics, coefficients and residuals effects, and study the changes only at aggregate level without identifying the role of each variable. Yun (2002) unifies the methods of Fields (2001) and Juhn/Murphy/Pierce (1993). According to the unified method, the changes in the variance of log-wages may be decomposed as follows; Schmollers Jahrbuch 123 (2003) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.1.43 | Generated on 2023-04-04 12:31:36 Decomposing Male Inequality Change in East Germany During Transition 47  2 yA ÿ  2 yB  X k  K ÿ 1 k  1  skyA 1 2 yA ÿ skyB 1  2 yB   X k  K ÿ 1 k  1  sky 31 2 y 3ÿ sky 31  2 y 3  skyA 1 2 yA ÿ skyB 1  2 yB   X k  K ÿ 1 k  1   kB 1  xkA 1  xkA ; y 31  y 3ÿ  kB 1  xkB 1  xkB ; yB 1  yB   X k  K ÿ 1 k  1   kA 1  xkA 1  xkA ; yA 1  yA ÿ  kB 1  xkA 1  xkA ; y 31  y 3   2 aA ÿ  2 eB  ;  2  where y 3  0B  P k  K ÿ 1 k  1  kBxkA  eAand an index Krepresents error term. The first, second and last terms of the equation (2) respectively represent the characteristics effect, coefficients effect and residuals effect. These are based on the information contained in the earnings equation (1). 3. Analysis – Empirical Results We apply the unified inequality decomposition of Yun (2002) to analyze the coefficient and characteristics effects that lie behind the overall changes in wage inequality, using the variance of log-earnings as our inequality measure. In order to perform our wage inequality decompositions, we estimate wage equations for 1990 and 2000 using OLS. Table 1 presents the sample means for the variables we use in our analysis. We restrict ourselves to basic variables for our wage analysis: experience, education, occupation, firm size and industry. Table 1 shows some changes in East Germany over the decade since the unification. Education and experience increase slightly from 1990 to 2000. There is a stark change in occupation, with blue collar workers falling by 9 %, and scientist/managers and office/business/service job holders (presumably whitecollar workers) increasing by about 4.5 % each. There is a marked movement to smaller firm sizes. There are also industrial shifts, the most notable being the decline in the transportation/postal industries, and, notably, an increase in the construction sector by 10 % from 1990 to 2000. Table 2 reports the wage equation estimates for East German males. In both 1990 and 2000 experience is significant and follows a U-shape. Education adds to wages, with the return to an additional year of schooling increasing Schmollers Jahrbuch 123 (2003) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.1.43 | Generated on 2023-04-04 12:31:36 48 Ira N. Gang and Myeong-Su Yun from 1990 to 2000. Occupation, firm size, and industry differentially affect wages, and the effects seem to vary from 1990 to 2000. Table 1 Sample Means 1990 2000 Wages (constant 1995 DM) 8.292 (2.670) 18.080 (7.326) Experience 20.846 (11.274) 22.407 (9.845) Education (year) 12.258 (2.351) 12.554 (12.269) Occupation Scientist/Manager* 0.192 (0.394) 0.233 (0.423) Office/Business/Service 0.152 (0.359) 0.193 (0.394) Blue Collar 0.655 (0.475) 0.574 (0.494) Firm Size Size < 20 * 0.107 (0.309) 0.296 (0.457) Size 20 – < 200 0.206 (0.405) 0.415 (0.493) Size 200 – < 2000 0.347 (0.476) 0.129 (0.335) Size 2000+ 0.340 (0.474) 0.160 (0.367) Industry Energy/Water/Mining 0.074 (0.262) 0.033 (0.179) Chemicals/Synthetics 0.071 (0.257) 0.038 (0.192) Iron/Mechanical 0.167 (0.373) 0.134 (0.341) Electrical/Clothing 0.189 (0.392) 0.129 (0.335) Construction 0.131 (0.337) 0.238 (0.426) Sales 0.050 (0.219) 0.088 (0.283) Transportation/Postal 0.130 (0.337) 0.071 (0.256) Finance/Education /Health/Legal 0.114 (0.317) 0.150 (0.357) Service 0.016 (0.125) 0.032 (0.175) Public Administration* 0.058 (0.234) 0.088 (0.283) Sample Size 1011 663 Standard deviations are reported in parentheses. * indicates a reference group in the regression analysis. Schmollers Jahrbuch 123 (2003) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.1.43 | Generated on 2023-04-04 12:31:36 Decomposing Male Inequality Change in East Germany During Transition 49 Table 2 Regression Results of Earnings Equations 1990 2000 Constant 1.429* (0.100) 1.944* (0.122) Experience 0.017* (0.003) 0.015* (0.006) Exprience 2 /100 -0.030* (0.007) -0.027* (0.012) Education (year) 0.039* (0.005) 0.052* (0.007) Occupation Office/Business/Service -0.120* (0.033) -0.128* (0.043) Blue Collar -0.090* (0.030) -0.111* (0.041) Firm Size Firm Size 20 – < 200 0.068* (0.031) 0.142* (0.029) Firm Size 200 – < 2000 0.086* (0.029) 0.318* (0.042) Firm Size 2000+ 0.118* (0.030) 0.281* (0.041) Industry Energy/Water/Mining 0.064 (0.047) 0.140 (0.080) Chemicals/Synthetics -0.010 (0.047) -0.051 (0.077) Iron/Mechanical 0.009 (0.042) 0.045 (0.058) Electrical/Clothing -0.046 (0.041) 0.032 (0.060) Construction -0.005 (0.043) 0.047 (0.058) Sales -0.097 (0.050) -0.132* (0.060) Transportation/Postal 0.004 (0.042) -0.020 (0.063) Finance/Education /Health/Legal -0.063 (0.042) 0.005 (0.056) Service -0.256* (0.074) -0.142* (0.082) Adjusted R 2 0.223 0.306 F Value 18.11 18.13 Sample Size 1011 663 1. Standard errors are reported in parentheses, and * means statistically significant at 5%. 2. Reference groups are scientist/manager for occupation, size less than 20 for firm size and public administration for industry. Schmollers Jahrbuch 123 (2003) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.1.43 | Generated on 2023-04-04 12:31:36 50 Ira N. Gang and Myeong-Su Yun Table 3 Decomposition of Changes in Male Inequality (1990–2000) Earning Inequality a) Decomposition b) 1990 2000 Characteristics Effect Coefficients Effect Total 0.086 (100.0) 0.135 (100.0) 0.000 (0.6) 0.023 (46.6) Human Capital 0.013 (14.7) 0.020 (14.9) -0.001 (-1.0) 0.008 (16.2) Experience 0.005 (6.0) 0.003 (2.1) -0.001 (-2.5) -0.001 (-2.2) Experinece 2 /100 -0.002 (-2.7) -0.001 (-1.1) 0.000 (0.8) 0.001 (1.1) Education 0.010 (11.4) 0.019 (13.8) 0.000 (0.6) 0.009 (17.3) Occupation 0.004 (4.4) 0.005 (3.8) -0.001 (-1.0) 0.002 (3.7) Office/Business/ Service 0.001 (1.1) 0.000 (0.0) -0.000 (-0.8) -0.001 (-1.1) Blue Collar 0.003 (3.3) 0.005 (3.8) -0.000 (-0.3) 0.002 (4.8) Firm Size 0.002 (2.3) 0.015 (10.9) 0.000 (0.5) 0.013 (25.3) Size 20 – < 200 -0.000 (-0.4) 0.000 (0.0) 0.001 (1.2) -0.000 (-0.4) Size 200 - < 2000 0.000 (0.2) 0.008 (6.0) 0.000 (0.8) 0.008 (15.4) Size 2000+ 0.002 (2.6) 0.007 (4.9) -0.001 (-1.5) 0.005 (10.3) Industry 0.002 (2.3) 0.004 (2.7) 0.001 (2.1) 0.001 (1.4) Energy/Water/Mining 0.000 (0.5) 0.001 (0.9) -0.000 (-0.2) 0.001 (1.8) Chemicals/Synthetics -0.000 (-0.0) -0.000 (-0.0) -0.000 (-0.0) -0.000 (-0.1) Iron/Mechanical 0.000 (0.1) 0.000 (0.1) -0.000 (-0.1) 0.000 (0.1) Electrical/Clothing 0.000 (0.4) -0.001 (-0.1) -0.000 (-0.0) -0.000 (-0.8) Construction 0.000 (0.0) -0.000 (-0.5) 0.000 (0.0) -0.001 (-1.4) Sales 0.001 (0.7) 0.002 (1.7) 0.000 (0.6) 0.001 (2.8) Transportation/Postal -0.000 (-0.0) 0.000 (0.0) 0.000 (0.0) 0.000 (0.1) Finance/Education / Health/Legal -0.000 (-0.3) 0.000 (0.0) -0.000 (-0.7) 0.001 (1.4) Service 0.001 (1.0) 0.001 (0.6) 0.001 (2.5) -0.001 (-2.5) Residuals 0.065 (76.3) 0.092 (67.7) 0.026 (52.7) a) Shares of variance of log-wages in 1990 (0.086) and 2000 (0.135) are reported in parentheses. b) Share of differences in variance of log-wages between 1990 and 2000 (0.050) are reported in parentheses. Schmollers Jahrbuch 123 (2003) 1 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.123.1.43 | Generated on 2023-04-04 12:31:36