Using Chinese Household Income Project survey (CHIP) data, this study analyzes the impact of the Minimum Wage (MW) policy on average wage and wage distribution in urban China in the 1993-1995, the 1998-2002, and the 2007-2013 periods, and compared the MW effects between public and private sectors. Several major conclusions emerged from this study. First, comparatively, the overall impact of the MW on average wage in the 1993-1995 period is greater than the effect of the MW level on the average wage. Second, the MW effects on average wage exist in both the public and private sectors. However, holding the other factors consistent, the MW effect on average wage is greater for the private sector than for the public sector. Third, the overall effects of MW level on the low-wage group increased in the 2007-2013 period. Fourth, the effects of MW on wage distribution are greater for the private sector than for the public sector in the three periods. Holding the other factors consistent, the impact of MW on the low-wage group for the private sector increased greatly than for the public sector in the 2007-2013 period. Fifth, decomposition results indicated that the differences of distribution proportions, in regions with different MW levels between the public and private sectors helped reduce the wage gaps, as did the MW effects on wage, which were greater for the private sector than for the public sector in the 1993-1995 and 2007-2013 periods.
The rationale behind the Minimum Wage (MW) policy is to increase the wage level for low-income group reduce their poverty level and narrow income inequality between high- and low-wage groups. Thus, the MW implementation is an important labor policy in both developing and developed countries.
In China, income inequality increased during the economic transition period. Along with marketization reform progress, the Chinese government also implemented the MW policy. This policy addressed “market failure” caused by firms that set lower than equilibrium wages. MW was first promulgated as a law―Enterprise’s Minimum Wage Regulations―in 1993. The MW level is mainly determined by the regional government. In 2004, the government published a new MW regulation to enforce the implementation of MW policy in whole of China; therefore, there was a large increase in MW level since 2004.
In addition, although the labor market is reformed by Chinese government since 1978, it is pointed out that the wage determine systems are different between the public sector and the private sector; there exits wage gap between these two sectors even if worker’s human capital factors are consistent (Chen, Demurger, & Fournier, 2005; Zhang Juwei & Xue Xinxin, 2008; Ye Lingxiang, Li Shi, & Luo Chuliang, 2011; Demurger, Li, & Yang, 2012; Ma, 2015, 2016) . Thus, if the MW effects exist, the effects may be different by these two sectors.
There are some empirical studies on the effects of MW on employment, wage gaps, and income inequality for developed countries, but hardly any empirical studies to understand the effects of MW on wage distribution in China. Particularly, there is no empirical study on the comparison between the public sector and the private sector. In this study, we provide numerical evidence to these issues.
Concretely, this study attempts to answer the following questions through an empirical analysis using micro- data from Chinese Household Income Project survey (CHIPs). First, does the MW affect wage levels? Second, if it does, is the effect of MW on wage different by wage distribution? Third, is the MW effect on wage different for the public and the private sector? Fourth, does MW affect wage gaps between the public and private sectors? Considering that the MW primarily affects low-income groups, we employ different models to conduct an analysis on both average wage and different wage percentiles.
For the developed countries, Card and Krueger (1995), Neumark (2001), Gindling and Terrell (2005), Neumark, Cunningham and Siga (2006), Hohberg and Lay (2015) utilized OLS (ordinary least squares) and QR (quantile regression) models to analyze the effects of MW on average wage and different percentile wage groups based on the cross section data and time-series data1.
Considering empirical studies on the issue for China, Jia Peng and Zhang Shiwei (2013) utilized the Neumark, Schweitzer, Wascher (2004) model (which is names as Neumark, et al. 2004 ) to analyze the MW effect through- out wage distribution in China using 1997-2009 CHNS (Chinese Health and Nutrition Survey) survey data. They found that the MW effect increase can reach to 1.00 - 1.25 times of the MW level on wage distributions. Di Junpeng and Han Qing (2015) utilized OLS (Ordinary Least Squares), QR (Quantile Regression) and DID (Difference in Difference) models to analyze the effect of MW on urban residents income using 1996-2010 CHNS, and revealed that when the MW level increases per 1%, the average wage will increase by 0.6%, and the affect mainly concentrates in the low-income group. Moreover, they pointed out that the MW specially provide protection for the elderly and low-skill workers based on the results by the DID method analysis. Ma Shuang, Zhang Jie, & Zhu Xi (2012) utilized Chinese manufacturing firm-level data and city-level MW from 1997 to 2007 to analyze the effect of MW on the average wage in firms by OLS, and found that if the MW increase by 10%, the average wages in firms would pick up by 0.4% - 0.5%.
The main contributions of this study are as follows: First, considering the MW compliance problem in China, using CHIP 1995, 2002, 2007 and 2013 data, the effects of MW policy are analyzed for the three periods―the MW beginning period (1993-1995), the MW implementation period (1998-2002), and the MW enforcement period (2007-2013). This shows the disparities of MW effects on wage distribution, which is perhaps caused by the government’s enforcement behaviors. Second, considering that the wage determinants system and the MW policy compliance situations are different for the public and private sectors and that the proportions of low-wage groups may be different for these two sectors, the effect of MW on wage distribution might differ for the public and private sectors. Using the subsamples, the comparison between these two sectors is also employed in this study.
1) The model for MW Effects on average wage
The OLS model for MW effects on average wage is represented with the Equation (1.1).
In the Equation (1.1), i represents individual workers, t represents periods and j represents regions.
Comparing the difference between the public and private sectors in terms of the MW effect, public sector dummy is utilized in the OLS model, it is represented with the Equation (1.2). In the Equation (1.2),
2) The model for MW effects on wage distribution
To see the effects of the MW by wage distribution, we adopt the Quantile Regression model (Koenker & Baset, 1978) , which can be expressed as:
In Equation (2), i represents individual workers, and
The QR model is designed for estimation using the optimal method, which minimizes the two error terms. The coefficients of
3) The model for MW effects on wage gaps between the public and private sectors
Based on wage functions by OLS, the Oaxaca-Blinder decomposition model (Oaxaca, 1973; Blinder, 1973) can be derived as equations (3.1), (3.2).
In equations (3.1), (3.2),
public and private sectors in the individual characteristics factors―including human capital (e.g. education, experience years, occupation and industry), the MW level and regional distribution,
CHIPs 1995, CHIPs 2002, CHIPs 2007 and CHIPs 2013 are utilized for the analysis. These data are gained from the four surveys of the CHIPs (Chinese Household Income Project Survey) conducted by NBS (National Bureau of Statistics), Institute of Economic, CASS (Chinese Academy of Social Science) and Beijing Normal University in 1996, 2003, 2008 and 2014, including respective information about employment and wages of urban residents.
Using retrospective survey data of income in CHIPs 1995 and CHIPs 2002, and the survey data in CHIPs 2007 and CHIPs 2013, we can conduct time series data sets for the 1993-1995 period, the 1998-2002 period, and the 2007-2013 period. Because there are design similarities of the data in the questionnaire, we can use the same information for analysis for all three periods-MW beginning period (1994-1995), MW performed period (1998- 2002), MW enforcement period (2007-2013).
CHIP surveys cover the representative regions in China, including Beijing, Shanxi, Liaoning, Jiangsu, Anhui, Guangdong, Henan, Hubei, Sichuan, Yunnan, and Gansu in 1995, Beijing, Shanxi, Liaoning, Jiangsu, Anhui, Guangdong, Henan, Hubei, Sichuan, Chongqi Yunnan, and Gansu in 2002, Beijing, Shanxi, Liaoning, Jiangsu, Anhui, Guangdong, Henan, Hubei, Sichuan, Chongqi, Yunnan, Gansu, Shanghai, Zherjiang, Fujian, Hunan in 2007 and 2013. Utilizing the information of the regions, we merge the MW level2 based on the National Minimum Wage Dataset to the CHIP survey data to construct the new dataset―including the individual level data and the regional level MW data for the analysis.
The wage is defined as the total earnings from work (called “the total wage”)3. Here, it comprises the basic wage, cash subsidy, and no cash subsidy4. We use the CPI in 1995 as the standard, and adjust the nominal monthly wage and nominal MW level in every year.
The analytic objects of this paper are employees, excluding the self-employer and the unemployed. Considering the retirement system of the state-owned sector, to reduce the effect of that system on the analysis result, the analytic objects are limited in the groups to between the ages of 16 and 60.
In the wage function, the explained variable is the logarithm of the monthly wage, and the explaining variables are the variables likely to affect the wage, such as schooling years, experience years5, public sector dummy variable6, occupation dummy variable7, industry dummy variable8, and region dummy variables (the East, Central and West Regions).
The statistical description of variables for the three periods is summarized in
Total | Public sector | Private sector | |||||
---|---|---|---|---|---|---|---|
Means | S.D. | Means | S.D. | Means | S.D. | ||
Panel A: 1993-1995 | |||||||
Sectors: public sector | 0.807 | 0.395 | |||||
Male | 0.505 | 0.500 | 0.532 | 0.499 | 0.390 | 0.488 | |
Schooling year | 11 | 3 | 11 | 3 | 9 | 3 | |
Experience year | 28 | 10 | 28 | 10 | 28 | 10 | |
Occupations: blue-color workers | 0.381 | 0.486 | 0.348 | 0.476 | 0.520 | 0.500 | |
Industries: manufacturing | 0.399 | 0.490 | 0.372 | 0.483 | 0.516 | 0.500 | |
Regions | |||||||
West region | 0.262 | 0.439 | 0.263 | 0.440 | 0.256 | 0.436 | |
Central region | 0.272 | 0.445 | 0.284 | 0.451 | 0.221 | 0.415 | |
East region | 0.467 | 0.499 | 0.453 | 0.498 | 0.523 | 0.499 | |
Observations | 37,658 | 30,391 | 7,267 | ||||
Panel B: 1998-2002 | |||||||
Sectors: public sector | 0.667 | 0.471 | |||||
Male | 0.556 | 0.497 | 0.578 | 0.494 | 0.511 | 0.500 | |
Schooling year | 12 | 3 | 12 | 3 | 11 | 3 | |
Experience year | 29 | 10 | 29 | 10 | 28 | 10 | |
Occupations: blue-color workers | 0.286 | 0.452 | 0.278 | 0.448 | 0.303 | 0.460 | |
Industries: manufacturing | 0.254 | 0.435 | 0.216 | 0.412 | 0.328 | 0.470 | |
Regions | |||||||
West region | 0.265 | 0.441 | 0.278 | 0.448 | 0.240 | 0.427 | |
Central region | 0.265 | 0.441 | 0.274 | 0.446 | 0.248 | 0.432 | |
East region | 0.470 | 0.499 | 0.448 | 0.497 | 0.512 | 0.500 | |
Observations | 48,856 | 32,598 | 16,258 | ||||
Panel C: 2007-2013 | |||||||
Sectors: public sector | 0.661 | 0.473 | |||||
Male | 0.485 | 0.500 | 0.476 | 0.499 | 0.502 | 0.500 | |
Schooling year | 12 | 3 | 12 | 3 | 12 | 3 | |
Experience year | 29 | 13 | 29 | 14 | 29 | 10 | |
Occupations: blue-color workers | 0.127 | 0.333 | 0.098 | 0.297 | 0.184 | 0.387 | |
Industries: manufacturing | 0.129 | 0.335 | 0.088 | 0.284 | 0.208 | 0.406 | |
Regions | |||||||
West region | 0.231 | 0.422 | 0.240 | 0.427 | 0.214 | 0.410 | |
Central region | 0.214 | 0.410 | 0.214 | 0.410 | 0.214 | 0.410 | |
East region | 0.555 | 0.497 | 0.546 | 0.498 | 0.572 | 0.495 | |
Observations | 22,385 | 14,807 | 7,578 |
Source: Calculated using CHIPs 1995, CHIPs 2002, CHIPs 2007 and CHIPs 2013.
For example, the average schooling year is more for the public sector (11 years in the 1993-1995 period, 12 years in the 1998-2002 period) than for the private sector (9 years in the 1993-1995 period, 11 years in the 1998-2002 period). However, the average schooling year is similar for the public and private sectors in the 2007-2013 period. Moreover, the proportions of manufacturing industry workers are more for the private sector (51.6% in the 1993-1995 period, 32.8% in the 1998-2002 period, 20.8% in the 2007-2013 period) than for the public sector (37.2% in the 1993-1995 period, 21.6% in the 1998-2002 period, 8.8% in the 2007-2013 period).
Second, for the low and middle wage level groups (25 percentile wage group, 50 percentile wage group, 75 percentile wage group), the wage levels in the public sector are higher compared to the private sector (in the three periods).
Third, the minimum wage levels are lower for the public sector (1.609 in the 1993-1995 period, 2.575 in the 1998-2002 period) than for the private sector (1.872 in the 1993-1995 period, 2.703 in the 1998-2002 period). However, it is higher for the public sector (7.127) than for the private sector (7.038) in the 2007-2013 period.
Fourth, the maximum wage levels are higher for the public sector in the 1993-1995 and the 1998-2002 periods; however, they are higher for the private sector (11.521) than for the public sector (10.199) in the 2007- 2013 period.
Fifth, from 1993 to 2002, the standard deviation values are greater for the private sector (0.714 in the 1993- 1995 period, 0.691 in the 1998-2002 period) than for the public sector (0.575 in the 1993-1995 period, 0.613 in the 1998-2002 period). However, it is greater for the public sector (1.101) than for the private sector (0.880) in the 2008-2013 period. Although the wage gaps within sectors exhibit growth in both the public and private sectors from 1993 to 2013, it is greatest for the public sector in the 2007-2013 period.
Considering the effect of MW on wage distribution, it is pointed out in the previous studies that there exists a spike effect for the group with wage below or around the MW level. Does the spike effect exist in urban China? Wage distribution using kernel density estimates are calculated for the three periods. These results are shown in
Panel A: 1993-1995 | Panel B: 1998-2002 | Panel B: 2007-2013 | ||||
---|---|---|---|---|---|---|
Private | Public | Private | Public | Private | Public | |
Maximum | 9.050 | 11.157 | 9.900 | 10.441 | 11.521 | 10.199 |
Minimum | 1.872 | 1.609 | 2.703 | 2.575 | 1.904 | 1.680 |
Mean | 5.815 | 6.111 | 6.323 | 6.571 | 7.038 | 7.127 |
S.D. | 0.714 | 0.575 | 0.691 | 0.613 | 0.880 | 1.101 |
p25 | 5.454 | 5.831 | 5.901 | 6.239 | 6.558 | 6.805 |
p50 (Median) | 5.859 | 6.147 | 6.288 | 6.605 | 7.098 | 7.343 |
p75 | 6.194 | 6.443 | 6.725 | 6.949 | 7.609 | 7.763 |
p25/p50 | 0.931 | 0.949 | 0.939 | 0.945 | 0.924 | 0.927 |
p75/p50 | 1.057 | 1.048 | 1.069 | 1.052 | 1.072 | 1.057 |
Source: Calculated using CHIPs 1995, CHIPs 2002, CHIPs 2007 and CHIPs 2013.
There seems to be a small spike shapes in the wage distribution in 1994 for public sector and private sector, in 2007 for public sector, and in 2013 for the private sector. The spike-shaped wage distribution showed that the proportions of groups around the maximum values of the MW level, particularly for the group with the wage just a little more than the maximum values of the MW level are greater―it indicated that the MW increased the wage levels around the MW before the MW enforcement. As it is described from the above, the MW was first promulgated as a law―Enterprise’s Minimum Wage Regulations―in 1993, and it was enforced since 2004 in China, therefore these estimated results indicated that there exist the spike effects in urban China for the MW beginning period (1993-1995), and the MW enforcement period (2007-2013).
First (Estimation 1), the estimated coefficients of MW logarithmic values are the greatest for the 1998-2002 period and smallest for the 1993-1995 period. Comparing the impact of the MW system at the beginning to the present times, the effects of MW level on the average wage level is becoming greater.
In additions, holding the other factors consistent, the average wage level is higher for the public sector than for the private sector (17.4% higher in the 1993-1995 period, 16.2% higher in the 1998-2002 period, 9.3% higher in the 2007-2013 period). Wage gaps between the public and private sectors exist in urban China even though the human capital of workers is consistent in these two sectors from 1993 to 2013.
Second (Estimation 2 and Estimation 3), if the sample is divided between the public and private sectors, the effects of the MW are all positively significant in the three periods, and the coefficients of MW logarithmic values are greater for the private sector than for the public sector in the three periods. Results showed that although the MW effect exists in both public sector and private sector, it is greater for the private sector than for the public sector.
Third (Estimation 4), results of Estimation 2 and Estimation 3 might be caused by the human capital differential and distribution proportion differential in terms of occupation, industry, and regions between these two sectors. Therefore, Estimation 4 is employed to control these influences. The estimated coefficients of interaction items of
First, the estimated coefficients of MW logarithmic values for the low-wage group (5th, 10th, 20th, 30th percentile wage groups) are greatest in the 2007-2013 period; they are smallest in the 1993-1995 period. It is indicated that after the government has enforced the implementation of the MW policy since 2004, the effect of MW on low-wage group has become greater.
Second, the estimated coefficients of MW logarithmic values are greater for the low-wage group than for the middle- and high-wage group in the 2007-2013 period, whereas they are greater for the high-wage group than that for the middle-wage group in the 1993-1995 and 1998-2002 periods. It indicated that the spillover effects are relatively greater for the high-wage group than for the middle-wage groups from 1993 to 2002; however, the same is smaller in the 2007-2013 period.
Do the differences of MW effect between the public sector and the private sector exist? To answer the question, two types of estimations are employed; they are shown in
First, using the subsamples, the total samples are divided into two subsamples―the public sector group and the private sector group. Estimations by the two sectors are employed and the results are summarized as follows:
Panel A: 1993-1995 | |||||||||
---|---|---|---|---|---|---|---|---|---|
(1) Public + private | (2) Public | (3) Private | (4) Public + private | ||||||
coef. | t-val. | coef. | t-val. | coef. | t-val. | coef. | t-val. | ||
lnMW | 0.759*** | 28.90 | 0.685*** | 24.35 | 0.985*** | 14.45 | 1.011*** | 27.33 | |
Sector (private sector) | |||||||||
Public sector | 0.174*** | 13.73 | 2.008*** | 10.53 | |||||
lnMW*public | −0.330*** | −9.64 | |||||||
Male | 0.155*** | 15.86 | 0.136*** | 13.14 | 0.214*** | 8.22 | 0.152*** | 15.62 | |
Education | 0.041*** | 21.16 | 0.040*** | 19.66 | 0.046*** | 8.05 | 0.042*** | 21.49 | |
Exp. | 0.059*** | 24.43 | 0.063*** | 24.19 | 0.051*** | 8.28 | 0.060*** | 24.80 | |
Exp.-sq. | −0.001*** | −18.56 | −0.001*** | −18.19 | −0.001*** | −6.57 | −0.001*** | −18.78 | |
Occupations (no-blue-color workers) | |||||||||
Blue-color workers | −0.133*** | −11.60 | −0.139*** | −11.19 | −0.086*** | −3.11 | −0.128*** | −11.22 | |
Industries (no-manufacturing) | |||||||||
Manufacturing | −0.013 | −1.27 | −0.022* | −1.96 | 0.013 | 0.48 | −0.015 | −1.43 | |
Regions (east) | |||||||||
West | −0.078*** | −5.46 | −0.066*** | −4.39 | −0.154*** | −3.98 | −0.083*** | −5.84 | |
Central | 0.081*** | 5.08 | 0.079*** | 4.80 | 0.047 | 1.03 | 0.073*** | 4.60 | |
Year dummy | Yes | Yes | Yes | Yes | |||||
Cons. | 0.573*** | 3.38 | 1.099*** | 6.09 | −0.610 | −1.39 | −0.862*** | −3.83 | |
Obs. | 13,410 | 10,818 | 2,592 | 13,410 | |||||
Adj.R-sq. | 0.322 | 0.313 | 0.332 | 0.326 | |||||
Panel B: 1998-2002 | ||||||||
---|---|---|---|---|---|---|---|---|
(1) Public + private | (2) Public | (3) Private | (4) Public + private | |||||
coef. | t-val. | coef. | t-val. | coef. | t-val. | coef. | t-val. | |
lnMW | 1.048*** | 52.36 | 0.922*** | 40.99 | 1.298*** | 32.08 | 1.078*** | 43.25 |
Sector (private sector) | ||||||||
Public sector | 0.162*** | 27.80 | 0.424*** | 3.36 | ||||
lnMW*public | −0.046** | −2.07 | ||||||
Male | 0.189*** | 35.26 | 0.146*** | 23.51 | 0.275*** | 27.17 | 0.189*** | 35.26 |
Education | 0.063*** | 55.74 | 0.063*** | 46.90 | 0.060*** | 27.84 | 0.063*** | 55.70 |
Exp. | 0.039*** | 26.30 | 0.045*** | 26.32 | 0.025*** | 8.88 | 0.039*** | 26.32 |
Exp.-sq. | 0.000*** | −17.44 | −0.001*** | −17.62 | 0.000*** | −5.62 | 0.000*** | −17.46 |
Occupations (no-blue-color workers) | ||||||||
---|---|---|---|---|---|---|---|---|
Blue-color workers | −0.062*** | −9.27 | −0.065*** | −8.28 | −0.073*** | −5.92 | −0.062*** | −9.27 |
Industries (no-manufacturing) | ||||||||
Manufacturing | −0.071*** | −10.87 | −0.133*** | −17.20 | 0.031*** | 2.65 | −0.071*** | −10.88 |
Regions (East) | ||||||||
West | −0.023*** | −2.80 | −0.046*** | −5.00 | 0.023 | 1.36 | −0.023*** | −2.77 |
Central | 0.130*** | 14.16 | 0.113*** | 11.13 | 0.173*** | 9.10 | 0.130*** | 14.20 |
Year dummy | Yes | Yes | Yes | Yes | ||||
Cons. | −0.967 | −8.48 | −0.198*** | −1.54 | −2.098*** | −9.12 | −1.143*** | −8.04 |
Obs. | 46,740 | 31,502 | 15,238 | 46,740 | ||||
Adj.R-sq. | 0.259 | 0.259 | 0.218 | 0.259 |
Panel C: 2007-2013 | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
(1) Public + private | (2) Public | (3) Private | (4) Public + private | ||||||||||
coef. | t-val. | coef. | t-val. | coef. | t-val. | coef. | t-val. | ||||||
lnMW | 0.855*** | 52.50 | 0.719*** | 32.68 | 1.020*** | 42.42 | 1.055*** | 43.29 | |||||
Sector (private sector) | |||||||||||||
Public sector | 0.093*** | 8.10 | 2.269*** | 11.47 | |||||||||
lnMW*public | −0.339*** | −11.02 | |||||||||||
Male | 0.365*** | 33.42 | 0.400*** | 25.29 | 0.315*** | 21.44 | 0.363*** | 33.37 | |||||
Education | 0.105*** | 44.40 | 0.131*** | 37.06 | 0.082*** | 26.38 | 0.106*** | 45.19 | |||||
Exp. | 0.066*** | 26.80 | 0.092*** | 24.96 | 0.040*** | 12.14 | 0.067*** | 27.04 | |||||
Exp.-sq. | −0.001*** | −24.89 | −0.001*** | −23.00 | −0.001*** | −11.35 | −0.001*** | −25.04 | |||||
Occupations (others) | |||||||||||||
Blue-color workers | 0.027* | 1.73 | 0.166*** | 6.85 | −0.068*** | −3.33 | 0.029* | 1.83 | |||||
Industries (others) | |||||||||||||
Manufacturing | 0.083*** | 5.21 | 0.066*** | 2.60 | 0.112*** | 5.61 | 0.078*** | 4.92 | |||||
Regions (East) | |||||||||||||
West | 0.026** | 1.98 | 0.021 | 1.11 | 0.012 | 0.63 | 0.026** | 2.00 | |||||
Central | −0.046*** | −3.29 | −0.079*** | −3.92 | −0.027 | −1.44 | −0.042*** | −3.02 | |||||
Year dummy | Yes | Yes | Yes | Yes | |||||||||
Cons. | −0.865*** | −7.01 | −0.640*** | −3.87 | −1.279*** | −7.00 | −2.196*** | −12.73 | |||||
Obs. | 23,310 | 12,027 | 11,283 | 24,428 | |||||||||
Adj.R-sq. | 0.251 | 0.278 | 0.231 | 0.255 | |||||||||
Note: *, **, ***: statistically significant in 10%, 5%, 1% levels. Source: Calculated using CHIPs 1995, CHIPs 2002, CHIPs 2007 and CHIPs 2013.
lnMW | Public | lnMW*public | Constants | R2 | |||
---|---|---|---|---|---|---|---|
Panel A: 1993-1995 | |||||||
1st | 1.213*** | 3.200 | −0.488 | −5.638** | 0.172 | ||
5th | 1.021*** | 3.034*** | −0.483*** | −3.113*** | 0.176 | ||
10th | 0.925*** | 2.368*** | −0.378*** | −1.752*** | 0.186 | ||
20th | 0.896*** | 2.301*** | −0.378*** | −0.934*** | 0.190 | ||
30th | 0.879*** | 1.954*** | −0.317*** | −0.414* | 0.191 | ||
40th | 0.898*** | 1.939*** | −0.316*** | −0.240 | 0.194 | ||
50th | 0.882*** | 1.613*** | −0.260*** | 0.065 | 0.200 | ||
60th | 0.932*** | 1.703*** | −0.277*** | −0.027 | 0.207 | ||
70th | 0.987*** | 1.795*** | −0.295*** | −0.037 | 0.215 | ||
80th | 1.051*** | 1.926*** | −0.320*** | −0.136 | 0.228 | ||
90th | 1.176*** | 2.046*** | −0.350*** | −0.499* | 0.245 | ||
Panel B: 1998-2002 | |||||||
1st | 1.034*** | −0.544 | 0.141 | −3.936*** | 0.075 | ||
5th | 0.953*** | −0.075 | 0.053 | −2.206*** | 0.121 | ||
10th | 0.961*** | 0.032 | 0.035 | −1.616*** | 0.130 | ||
20th | 0.918*** | −0.068 | 0.054 | −0.930*** | 0.151 | ||
30th | 0.926*** | 0.061 | 0.029 | −0.634*** | 0.161 | ||
40th | 0.938*** | 0.229* | −0.004 | −0.467*** | 0.164 | ||
50th | 0.944*** | 0.500*** | −0.055** | −0.315** | 0.163 | ||
60th | 0.939*** | 0.426*** | −0.047** | −0.098 | 0.161 | ||
70th | 1.002*** | 0.520*** | −0.068*** | −0.245 | 0.160 | ||
80th | 1.111*** | 0.722*** | −0.111*** | −0.660*** | 0.163 | ||
90th | 1.209*** | 1.021*** | −0.171*** | −0.816*** | 0.164 | ||
Panel C: 2007-2013 | |||||||
1st | 1.063*** | 1.852 | −0.331 | −7.502*** | 0.166 | ||
5th | 1.121*** | 0.762* | −0.126* | −6.342*** | 0.137 | ||
10th | 1.152*** | 1.391*** | −0.228 | −4.803*** | 0.138 | ||
20th | 1.128*** | 2.167*** | −0.316 | −3.341*** | 0.155 | ||
30th | 1.098*** | 2.256*** | −0.327 | −2.635*** | 0.165 | ||
40th | 1.093*** | 2.661*** | −0.387 | −2.223*** | 0.167 | ||
50th | 1.043*** | 2.746*** | −0.400 | −1.667*** | 0.170 | ||
60th | 1.012*** | 2.693*** | −0.392 | −1.248*** | 0.173 | ||
70th | 0.995*** | 2.787*** | −0.409 | −0.909*** | 0.175 | ||
80th | 0.943*** | 2.386*** | −0.350 | −0.275*** | 0.177 | ||
90th | 0.922*** | 2.231*** | −0.331 | 0.234*** | 0.177 | ||
Note: 1) The other variables-male, education, experience year, occupations, indutries, region dummy, year dummy are also extimated in these models. 2) *, **, ***: statistically significant in 10%, 5%, 1% levels. Source: Calculated using CHIPs 1995, CHIPs 2002, CHIPs 2007 and CHIPs 2013.
1) The estimated coefficients of the MW logarithmic values indicated that except the lowest-wage group (1st percentile) in the 1998-2002 period, the estimated coefficients of MW logarithmic values are greater for all in the private sector than for the public sector in the three periods.
2) The differences of the estimated coefficients of MW logarithmic values between the public and private sector for the low-wage group are greatest in the 2007-2013 period.
These results showed that although the effects of MW on wage distribution are greater for the private sector than for the public sector in the three periods, the effect of MW on the low-wage group has increased in the 2007-2013 period.
Second, considering the human capital differentials between the public and private sectors, estimations holding the other factors―including the human capitals consistent―are employed, and the results are shown in
Results of the MW effects on the wage gap between the public and private sectors are shown in
First, overall, the influences of both unexplained differential and explained differential affect the wage gaps between the public and private sectors in the three periods. Moreover, the influences of unexplained differential differ by the periods. For example, the influences of unexplained differential are greater than the explained differential in both the 1993-1995 period and the 1998-2002 period; however, the influences of explained differential are greater than the unexplained differential in the 2007-2013 period. Concretely, the percentages of
Explained differentials | Unexpained differentials | |||||
---|---|---|---|---|---|---|
Actural value | Percentage (%) | Actural value | Percentage (%) | |||
Panel A: 1993-1995 | ||||||
Total | 0.047 | 22.1% | 0.167 | 77.9% | ||
Decomposition category: | ||||||
lnMW | −0.084 | −39.1% | −1.685 | −787.2% | ||
Education | 0.068 | 31.9% | −0.045 | −21.1% | ||
Experience year | 0.014 | 6.7% | 0.225 | 105.1% | ||
Male | 0.018 | 8.4% | −0.033 | −15.3% | ||
Occupation | 0.024 | 11.1% | −0.026 | −12.1% | ||
Industry | 0.004 | 1.8% | −0.018 | −8.6% | ||
Region | 0.003 | 1.3% | 0.019 | 8.9% | ||
Constant | 0.000 | 0.0% | 1.730 | 808.2% | ||
Panel B: 1998-2002 | ||||||
Total | 0.094 | 37.8% | 0.155 | 62.2% | ||
Decomposition category: | ||||||
lnMW | −0.031 | −12.6% | 0.153 | 61.5% | ||
Education | 0.094 | 37.8% | 0.021 | 8.4% | ||
Experience year | 0.004 | 1.6% | 0.322 | 129.3% | ||
Male | 0.009 | 3.7% | −0.074 | −29.9% | ||
Occupation | 0.002 | 0.8% | 0.002 | 0.9% | ||
Industry | 0.016 | 6.5% | −0.037 | −14.7% | ||
Region | 0.000 | 0.0% | 0.035 | 14.2% | ||
Constant | 0.000 | 0.0% | −0.267 | −107.5% | ||
Panel C: 2007-2013 | ||||||
Total | 0.151 | 85.4% | 0.026 | 14.6% | ||
Decomposition category: | ||||||
lnMW | −0.102 | −57.4% | −1.955 | −1102.4% | ||
Education | 0.223 | 125.7% | 0.560 | 316.0% | ||
Experience year | 0.018 | 10.2% | 0.710 | 400.6% | ||
Male | 0.023 | 13.1% | 0.043 | 24.5% | ||
Occupation | −0.008 | −4.3% | 0.048 | 26.9% | ||
Industry | −0.005 | −2.9% | −0.009 | −5.3% | ||
Region | 0.002 | 1.1% | −0.010 | −5.8% | ||
Constant | 0.000 | 0.0% | 0.639 | 360.2% | ||
Source: Calculated using CHIPs 1995, CHIPs 2002, CHIPs 2007 and CHIPs 2013.
unexplained differential are 77.9% in the 1993-1995 period and 61.5% in the 1998-2002 period, whereas the percentages are only 14.6% in the 2007-2013 period. It indicated that along with the transition of the economy systems, the differentials caused by wage determine systems decreased; the influences of explained differentials, including the human capital differences, increased.
Second, considering the influence of the MW on wage gap, (1) the estimated values of the MW as a component of explained differential are all negative values in the three period (−39.1% in the 1993-1995 period, −12.6% in the 1998-2002 period, −57.4% in the 2007-2013 period). It indicates that the differential of distribution proportions in the regions with different MW levels between the public sector and the private sector contributes to reduce the wage gap.
It can be explained that if the distribution proportions in the high MW level region are greater for the private sector (in the case when most private firms, individual firms, and foreign investment firms are concentrating in the east region where the MW levels are higher than other regions), the wage gaps between these two sectors might be reduced by the MW implementation. Based on the CHIPs data, the proportion of workers with wages lower than the MW are greater for the private sector (7.2% in the 1993 1995 period, 12.4% in the 1998-2002 period, 16.3% in the 2007-2013 period) than for the public sector (3.0% in the 19931995 period, 5.2% in the 1998- 2002 period, 11.4% in the 20072013 period).
(2) The estimated values of the MW as a component of unexplained differentials are negative in the 1993- 1995 (−782.2%) and 2007-2013 periods (−1102.4%). However, the value was positive (61.5%) in the 1998- 2002 period. It indicates that the MW effects on wage are greater for the private sector than for the public sector and contributes to reduce the wage gap in the MW beginning period and the MW enforcement period.
(3) Considering the influences of other factors, the human capital is the main factor to cause the wage gap. For example, the estimated value of education as a component of explained differential (31.9%) and of experience year as a component of explained differential is greatest (105.1%) in the 1993-1995 period. The estimated value of education as a component of explained differential (37.8%) and of experience year as a component of explained differential is greatest (129.3%) in the 1998-2002 period. The estimated value of education as a component of explained differentials (125.7%) and of experience year as a component of explained differentials is greatest (400.6%) in the 2007-2013 period. These results showed that the greater proportions of high-level education and the seniority wage system, which is mostly implemented in the public sector, are the main factors to cause the wage gaps between the public and private sectors.
Chinese government has been officially implementing the MW system since 1993 and enforced MW policy in all of China since 2004. It is thought that the implementation of the MW policy contributes to increasing incomes of low-wage groups and reducing their poverty. In China, does the MW affect average wage and wage distribution? Particularly, does the MW affect the wages for low-wage groups? Does the spillover effect of MW exist in China? If it does, does there exist the difference of MW effects between the public and private sectors? To answer these questions, this study employs empirical studies using cross-section survey data―CHIPs 1995, 2002, 2007 and 2013, and divides periods into three periods―the MW beginning period (1993-1995), the MW performed period (1998-2002), and the MW enforcement period (2007-2013). Several major conclusions emerge.
First, comparatively, the overall impact of the MW on average wage at the MW policy beginning period is greater than the effect of the MW level on the average wage.
Second, the MW effects on average wage exist in both the public and private sectors. However, holding the other factors consistent, the MW effect on average wage is greater for the private sector than for the public sector.
Third, the overall effects of MW level on the low-wage group increased in the 2007-2013 period.
Fourth, the effects of MW on wage distribution are greater for the private sector than for the public sector in the three periods. Holding the other factors consistent, the impact of MW on the low-wage group for the private sector increased greatly than for the public sector in the 2007-2013 period.
Fifth, decomposition results indicated that the differences of distribution proportions, in regions with different MW levels, between the public and private sectors helped reduce the wage gaps, as did the MW effects on wage, which were greater for the private sector than for the public sector in the 1993-1995 and 2007-2013 periods.
According to these empirical analyses, we can conclude that the MW implementation contributes to the rise in wage level for the low-wage groups in the public and private sectors. There are two points worthy of attention. First, the results revealed that the difference between the effects of MW on the public and the private sectors became greater for the low-wage group in the 2007-2013 period. The results might be caused by the reason that along with the MW implementation, the MW compliance problem was reduced. A comparison study on the MW compliance problem for the private sector and the public sector should be done in the future. Moreover, the results show that the effects of MW on average wage and wage distribution are greater for the private sector than for the public sector. Thus, the effects of MW level adjustment (rise in the MW levels) on labor force costs― which affect labor demand for the low-skill or low-education workers―might be greater for the private sector; hence, the negative effect on employment might be greater for the private sector than for the public sector. Thus, the MW effects on employment should be considered when the government decides to modify the MW levels.
Although this study utilized the repeated cross-section data (CHIPs) to analyze the impact of the Minimum Wage (MW) policy on average wage and wage distribution in urban China in the 1993-1995, the 1998-2002, and the 2007-2013 periods, and compared the MW effects between public and private sectors, there left the endo- geneity and heterogeneity problems in the study. Panel data should be utilized to address these econometric problems in the future. Moreover, the results also showed that the greater proportions of high-level education and the seniority wage system, which is mostly implemented in the public sector, are the main factors to cause the wage gaps between the public and private sectors. Thus using the firm level survey data, the researches related to the effects of firm wage and employment systems also should be done in the future.
This research was supported by JSPS KAKENHI Grand Number JP16K03611.
Data used for the study was provided by the Income Inequality Research Center, Beijing Normal University.
Xinxin Ma, (2016) Impacts of Minimum Wage Policy on Wage Distributions in Urban China: Comparison between Public and Private Sectors. Chinese Studies,05,45-61. doi: 10.4236/chnstd.2016.53006