The Earnings Differential between Formal and Informal Employees in Urban China

Few studies in the field of labor economics have analyzed the earnings differential between formal and informal employees in urban China. Due to data limitations, previous studies on this subject have not yielded accurate and widely acceptable results. This study seeks to obtain more accurate results by analyzing the earnings differential based on the conceptual framework of informal employment developed by the 17th International Conference of Labor Statisticians. We analyzed data from the 2006 Chinese General Social Survey, using the Lee model to correct for selection bias and decomposing the earnings differential between formal and informal employees in urban China in terms of the effects of employee characteristics, employment, and working hours. The latter two of these comprise the segmentation effect. We found that only 33% of the observed earnings differential can be explained by employee characteristics; the remaining 67% is attributable to the segmentation effect. The working hours effect narrows the earnings differential. Based on the results of our analysis, we conclude that informal employees in urban China, especially female informal employees, suffer from segmentation and that policies to address segmentation issues are needed.


Introduction
In recent years, informal employment has become a significant phenomenon, not only in developing countries but also in developed countries, as reflected by increasing proportions of workers who are informally employed. Between 20% and 30% of all workers in developed countries and more than 40% of workers in developing countries are engaged in informal employment (Hu & Yang, 2001). However, informal employees suffer from low wages, lack of social protection, and other problems (International Labor Organization [ILO], 2002).
In China, informal employment is an important and urgent issue. When China transitioned to a marketoriented economy, to address the problem of overallocation of labor to the agriculture sector, the Hukou System was modified to permit more flexibility. The state policy concerning the control of rural worker mobility has gradually changed since 1978. According to Meng (2003), from 1958 to 1978, the movement of rural residents to urban areas was forbidden. This As there were not enough formal jobs for all of these people, informal employment came to exist and has experienced a steady increase in China since the 1990s.
Informal employment is becoming the main mode of employment in China (Hu & Yang, 2001). However, informally employed workers are at a disadvantage in various respects, such as in their work conditions and social welfare level, compared to formal workers (Wu & Cai, 2006). Therefore, labor market policy changes in China are called for to address informal employment issues.
Aspects of informal employment, including its definition (Hu & Yang, 2001;Wang, 2006), scale (Wu & Cai, 2006), and wage differential (Deng, 2009) have been researched by many Chinese scholars. Although several studies have addressed informal employment in China on a macro scale, few if any had sufficient micro data available to support their findings, and these studies were not structured in accordance with the internationally accepted framework of informal employment. These previous studies were thus not able to accurately characterize the earnings differential between formal and informal employment in China and were also unable to make comparisons with international research on informal employment.
To tackle the problems confronted by previous researchers and to provide more accurate and widely acceptable results, we analyzed the earnings differential between formal and informal employees in urban China, utilizing data from the 2006 Chinese General Social Survey (CGSS), a relatively new nationwide sample and four-stage stratified sampling scheme that provides abundant job information that can be analyzed using the International Conference of Labor Statisticians (ICLS) framework, using an approach based on this conceptual framework of informal employment developed by the 17th ICLS. Lee's model (1983) was employed to correct for possible selection bias, and the effect of working hours on the earnings differential is considered. Earnings differential decompositions have been widely used in previous studies on the existence and extent of labor market segmentation (Dickens & Lang, 1985) and specifically in studies on informal employment (Du, Cai & Wang, 2008). We follow the examples of previous studies.
The results of this study are presented as follows.
In Section 2, we present the results of a literature review of informal employment in China. In Section 3, we present the conceptual framework of informal employment developed by the 17th ICLS and describe the data and methodology. In Section 4, we present an overview of formal and informal employees in urban China. In Section 5, we discuss the results of the estimation and decomposition. We lastly offer concluding remarks in Section 6. Bernabè (2002) defined informal employees as individuals in any of the following situations in either their primary or secondary jobs: (1) self-employed workers and employees in household enterprises; (2) (unpaid) contributing family workers; (3) non-regular employees; (4) persons employed casually, temporarily or seasonally; or (5) employees engaged in "left-hand work" (earnings informal income at their formal workplaces). Günther and Launov (2011)  There have only been a few empirical studies on the earnings differential between formal and informal employees in China. The most noteworthy research in this area is that of Deng (2009), who used survey data to gain insights into informal employment in urban areas as China, for which the information required to conduct an empirical analysis on informal employment has long been lacking. Deng (2009) treated workers in private and individually owned enterprises (with fewer than 100 employees), the self-employed, and workers without long-term contracts as being informally employed. He found that the earnings differential between formal and informal wage earners was primarily attributable to unexplained factors rather than to the different characteristics of formal and informal workers. However, those working in individual household production or in small private enterprises with less than 100 workers were considered informal employees according to Deng's definition, which could result in overestimation of the extent of informal employment.

Literature review
As rural migrants are an important component of informal employment in China, studies of the earnings differences between rural migrants and urban residents yield some insights into the issue under study. Meng and Zhang (2001) utilized two comparable survey data sets, the Shanghai Floating Population Survey and the Shanghai Residents and Floating Population Survey, to analyze the degree to which segmentation between rural migrants and urban residents has occurred, in terms of occupational segregation and wage differentials. They found significant differences in occupational attainment and wages between rural migrants and urban residents, with most of the difference being unexplainable by productivity-related differences between the two groups.
The previous studies mentioned are very important to subsequent studies of informal employment. However, as mentioned before, the results obtained in previous studies were not very accurate or widely accepted and were not obtained using the ICLS framework.
The ICLS framework was used in this study to conduct a more comprehensive analysis of informal employment in urban China. Lee's model was used to correct for possible selection bias, and the effect of working hours on monthly earnings was determined from the unexplained factors.

Definition of informal employment
According to Hussmanns (2004), the international statistical definition of the informal sector was adopted by the 15th ICLS in 1993. Employment in the informal sector was defined as all jobs in informal-sector enterprises or employment of all persons who, during a given reference period, were employed in at least one China has thus far lacked the data necessary to analyze informal employment, the definitions used in previous studies are not totally satisfactory.
According to the ICLS framework, there are three standards by which to judge whether a worker is in formal or informal employment. The first standard is the production unit to which they belong. Production units are classified into three groups: formal-sector enterprises, informal-sector enterprises, and households. The second standard is their job status. According to the International Classification by Status in Employment 93 (ICSE-93, ILO), employment status is classified as follows: self-employed workers, employers, contributing family workers, employees, members of producers' cooperatives, and agricultural workers. The third standard is the nature of the job, i.e., whether the type of job a person holds is a formal or informal job.
Considering the differences in work conditions and social welfare levels, in this study, we consider the following to be components of the formal sector in terms of production units: party and governmental organizations; state-owned or state-shareholding enterprises; collective or collective-shareholding enterprises; state-owned or collectively owned non-profit institutions; enterprises in Hong Kong, Macao, or Taiwan;, foreign capital enterprises; and social groups. The self-employed, private enterprises, and other enterprises are considered part of the informal sector, in terms of production units. Following ILO (Hussmanns, 2004), workers in firms with fewer than 10 employees are considered to be working in the informal sector, and because of data limitations, these workers are treated as informal employees. We consider the nature of the job to be formal if the worker has an employment contract or if his employment is registered with the government. More details on the definition of informal employment are given by Zuo (2013).

Methodology
As some workers in the samples were unemployed, their earnings were not observed; therefore, selection bias could be an issue. In this study, we adopted the Lee (1983) approach to correct for this selection bias. Because the classification of employment status in this study was not binary, we adopted Lee's (1983) model rather than Heckman's model (1979) to correct for selection bias. The four employment status conditions considered are as follows: formal employees, informal employees, self-employed workers and the unemployed. Oaxaca-Blinder's decomposition approach is widely used to separate the earnings differential into two components. However, this approach cannot be applied to decomposing the earnings differential effects into three parts: difference in characteristics (e.g., age, sex, education and so on), differences in employment (i.e., formally or informally employed), and dif-The Earnings Differential between Formal and Informal Employees in Urban China ference in hours worked (i.e., monthly working hours).
To conduct such decomposition, the extension of the Oaxaca-Blinder approach proposed by Bourguignon, Fournier and Gurgand (2001)  Similarly, the monthly earnings of formal employees can be written as follows: The monthly earnings differential between formal and informal employees can be decomposed into three parts: (1) Pure differences in characteristics (e.g., sex, education): The latter two, the difference in employment and the difference in hours worked, comprise the segmentation effect. We report the mean of the alternative decomposition results.

Overview of Formal and Informal Employees in Urban China
We present an overview of formal and informal employees by indicating the proportion of workers in each group (see Table 1). There are six variables that detail the characteristics for all employed individuals: demographics, total years of education, location, firm size, occupation, and industry.
As Table 1 shows, males occupy larger proportions of both the formal and informal employee groups than females. Young workers below the age of 40 are more likely to be formal employees, while workers above age 40 are more likely to be informal employees. Most formal employees have more than 7 years of education, while 63% of informal employees have between 7 and 12 years of education.
Workers in the east account for the vast majority of both formal and informal employees (more than 45%), followed by workers in the middle and west regions. Approximately 56% of informal employees are not from large cities, and more than 80% of all of the workers are non-immigrants.
Nearly 40% of informal employees are in enterprises that employ fewer than 50 workers. In contrast, more than 57% of formal employees are in enterprises that employ more than 100 workers.
In terms of occupation, proportionally more professional and technical workers are formal employees than informal employees. A larger proportion of informal employees are clerical workers. The majority of formal and informal employees work in services, although the percentage is higher for informal employees.

Hourly earnings functions
The hourly earnings are summarized in Table 2. We only comment on the marginal effects that are statistically significant. Male employees enjoy hourly earnings that are approximately 9% higher than female employees, meaning that an hourly earnings differential by gender exists among both formal and informal employees.
Higher educational attainment leads to approximately 6% higher hourly earnings for formal employees, and it is also significant and slightly higher, at 7%, for informal employees. This indicates that education is rewarded for both informal and formal employees in urban China.
Hourly earnings differ greatly for employees from different regions. Hourly earnings are 40% for formal employees and approximately 30% higher for informal employees in the east than in the middle and western regions. Formal and informal employees from large cities receive hourly earnings more than 20% higher than employees from smaller areas. Immigrant employees receive higher hourly earnings than non-immigrants.
Informal employees in firms employing fewer than 50 workers earn significantly less than employees in the reference group; i.e., those in firms with more than 500 workers. Informal employees in firms with fewer than 10 workers earn 42% less per hour than those in the reference group. For formal employees, all other groups of workers earn significantly less than those in The Earnings Differential between Formal and Informal Employees in Urban China the reference group, with workers in firms employing fewer than 15 workers earning the least.
Formal employees in managerial and professional occupations earn higher hourly earnings than skilled agricultural workers. For informal employees, it seems that occupations do not make much difference in hourly earnings, except for professional workers. For-mal employees engaged in services earn approximately 12% higher hourly earnings than formal employees in the manufacturing industry.

The determinants of working hours
The hourly earnings are summarized in Table 3. In addition, we only comment on the marginal effects that   Among informal employees, immigrant employees work more than 15 hours per month more than nonimmigrants.
Firm size does not make much difference in monthly working hours for formal employees, but the situation is much different for informal employees working in firms employing fewer than 50 workers. These informal employees work approximately 20 hours more per month than informal employees working in firms employing more than 500 workers.
Occupation does not make much difference in monthly working hours for either formal or informal employees. (a) ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively; (d) represents a discrete change in the dummy variable from 0-1; data in parentheses are the Z values.

Decomposing the Observed Earnings Differential
Following the methodology presented in section 3, we use the earnings function and working hours function estimations to decompose the earnings differential into the pure difference-in-characteristics effect, the pure difference-in-employment effect, and the pure difference-in-hours-worked effect, with the latter two comprising the segmentation effect. The decomposition results are given in table 4.
As Table 4 shows, the majority of the observed total earnings differential between formal and informal employees can be attributed to the difference-in-employment effect and the difference-in-hours-worked effect, rather than the differences in characteristics.
Only 33% of the earnings differential between formal and informal employees can be explained by the differences in characteristics. The remaining 67% can be attributed to the segmentation effect, in which the pure employment effect dominates. In fact, longer hours worked by informal employees narrow the earnings differential between formal and informal employees, meaning that informal employees work more but earn less. This result confirms that the Chinese labor market is far from fully integrated, leading to a high earnings differential between formal and informal employees.
We also examined the earnings differential decomposition results by gender. For the male sub-sample, segmentation forces play a slightly less important role in explaining the earnings differential between formal and informal employees. However, in the female subsample, segmentation forces account for nearly 80% of the earnings differential between these two groups of employees, suggesting that female informal employees suffer most from segmentation.

Concluding remarks
The purpose of this research was to analyze the earnings differential between formal and informal workers in urban China based on the ICLS framework of informal employment, in an effort to obtain more accurate and widely acceptable results than previous studies.
Using the 2006 CGSS micro data, we employed the approach of Bourguignon et al. (2001) to decompose the earnings differential between formal and informal employees in urban China into characteristics, employment and working hour effects, with the latter two comprising the segmentation effect. The possible selection bias problem was addressed using the Lee model.  The main conclusions are that informal employees suffer from segmentation and work more but earn less, meaning that the earnings differential would be even larger if they worked the same number of hours as formal employees. This is especially true for female informal employees. Employment policies that address segmentation issues, e.g., gender segmentation, are needed in urban China. Segmentation, which prevents individuals from taking advantage of economic opportunities, must be identified and its adverse effects minimized. However, because we only have data for 2006 on hand, it was not possible in this study to assess how the situation has changed since 2006. This will be addressed in future research.