Impacts of social welfare system on the employment status of low-income groups in urban China

Purpose – Aiding employment is an important poverty reduction strategy in many countries ’ social welfare systems, as this strategy can help empower the recipients with a better living standard, development and social inclusion. The purpose of this paper is to identify the most significant individual and systematic variables for the employment status of low-income groups in urban China. Design/methodology/approach – The data of this study are drawn from “ Social Policy Support System for Poverty-stricken Families in Urban and Rural China 2015 ” report. The Ministry of Civil Affairs of the People ’ s Republic of China appointed and funded the Institute of Social Science Survey (ISSS) at Peking University to deliver the related project and organize a research team to write the report. Multiple binary logistic regression analysis is adopted to identify both individual and systematic factors that affect the employment status among low-income groups in urban China. Findings – According to the results of the binary logistic regression model, individual factors, including: gender; householder status; education; and self-rated health status, play a significant role in determining the employment status of low-income groups in urban China. Clearly, the impacts of individual factors are more influential to marginal families than to families entitled to receive Basic Living Allowance. In contrast, compared with marginal families, systematic factors are more influential to families entitled to receive Basic Living Allowance. Originality/value – This study highlights the importance of precise poverty reduction strategy and the issue of “ welfare dependence ” among low-income groups in urban China. Policy recommendations derived from the findings are hence given, including: the promotion of family-friendly policies; the introduction of a smart healthcare system; the establishment of a Basic Living Allowance adjustment mechanism; and the provision of related social services.


Introduction
The relationship between poverty and unemployment has always been a controversial topic, especially for developing countries (Visaria, 1981). Employment empowers the poor to adapt to development of the society, and therefore, the employment of the poor is always at the core of poverty alleviation strategies (Corcoran and Hill, 1980). However, some scholars now argue that globalization presents new challenges to traditional poverty alleviation strategies because of intense global competition, the technological displacement of low-skilled workers and slimmer chances of upward social mobility for younger generations (Ukpere and Slabbert, 2009). Generally, the unemployment rate of low-income groups is higher than medium-and high-income groups. The reasons for adverse employment status among low-income groups include irregular employment, lack of social security, long working hours, low income and a lower chance of changing job (Wu, 1994;Yin and Wang, 2015). Worse still, unemployment has more adverse impacts on low-income groups and is now one of the main drivers of poverty and inequality in Chinese cities (Xue and Zhong, 2003).
This problem is not unique to China, and so, in the hope of alleviating poverty, increasing labor participation and lowering unemployment among low-income groups have become standard policy and practice across many countries. Some studies have shown that an increase in public capital, especially on infrastructure, can reduce the unemployment rate by improving the quality of human capital (Akinbobola and Saibu, 2004). Governments around the world have experimented with several innovative measures, that differ from traditional social security systems. For example, the government of Botswana has long made use of its informal sector economy to ease the uneven regional development, like unemployment and poverty (Hope, 1996). The government of India introduced the National Rural Employment Guarantee Act (NREGA) in 2006, which stipulates that federal and local governments should provide at least 100 days of work and pay for the rural poor families with predetermined remuneration. Therefore, poor farmers are provided with a sustainable and Basic Living Allowance, which increases their chances of improving their socio-economic conditions (Haque, 2011). It has also been shown that there is strong relationship between income distribution and the growth of the employment rate in Pakistan (Aurangzeb and Asif, 2013). In South Africa, the government has introduced proactive measures which enable the young poor to select a grant with pro-employment arrangements, such as grants conditional on accepting training and education, as well as, opportunity vouchers for employers (Altman et al., 2014). For the sake of dealing with the issue of welfare dependence, Orszag and Snower proposed to adopt unemployment accounts rather than traditional unemployment benefits (Orszag and Snower, 2002). Most programs of employment promotion or poverty alleviation here can be considered a social investment and are operated under the principle of cost-effectiveness (Hope, 1992;Hope, 1996).
Due to the transition from a centrally planned economy to a market economy, the Chinese Central government has already adopted various measures to ease conflicts that have arisen from the transition, such as the New Labor Contract Lawintended to balance labor market flexibility, labor protection, and assistance to laid-off workers at state-owned enterprises (SOEs) following reforms. As a result, the relationship between social assistance and employment in China could be different from those of other countries.
China's social assistance system Being a responsible government of a developing economy, the Chinese government has always paid attention to the employment situation of low-income groups, and has traditionally maintained a reasonable living standard for low-income groups through measures like employment assistance and aid. However, the government has recently experimented with some proactive measures, including the introduction of public posts and the enrichment of a spectrum of social assistance measures (Han, 2016). In order to fulfill China's precise poverty reduction strategy, the government has carried out family budget investigations and then designed employment aid, monetary support and employment services based on their findings, and offered social assistance to help maintain their living standards (Han, 2016). However, several studies indicate that these policy attempts have achieved little progress, and the low employment rate of low-income groups still regularly 126 PAP 22,2 captures public attention (Wang, 2009(Wang, , 2017. Other studies further prove that the proportion of the related policy coverage (low-income groups with working capacity receiving Basic Living Allowance in urban China) was 75 percent in 2007 (Wang, 2009) and 63 percent in 2015 (Wang, 2017). To address this gap, this study focuses on how poverty alleviation strategy in China should switch from a traditional welfare system, to instead, one that identifies the barriers to job-seeking low-income groups, in addition to promoting the effectiveness of poverty reduction measures.
Individual factors that affect the employment prospects of low-income groups have already been identified in several studies, which include: age, gender, education, health condition and attitude. For example, the relatively older people have stronger desire to seek work (Wang, 2011). Males are more likely to have a stronger aspiration to be reemployed than females (Wang, 2012). Individuals with higher education tend to have more aspiration to seek work (Huang, 2007). Middle-aged and older adults, due to their inferior working capacity, have fewer opportunities to be employed (Lin and Lu, 2012). Finally, some studies demonstrate that the awareness of the responsibilities of the unemployed or the responsibility to work could affect the likelihood of being reemployed (Ci, 2003;Huang, 2007).
Systematic factors can also weaken the effectiveness of poverty reduction measures, with the two main issues being welfare dependence (Hong, 2005;Bian, 2014) and a significant degree of fragmentation among poverty reduction measures (Huang, 2007;Qiao, 2009). First, the current system guarantees beneficiaries a relatively stable expectation of payments. As long as there is no significant income increase, they can expect a monthly fixed payment given by the government. This has been shown to fuel dependency among recipients (Hong, 2005). This form of compensation also discourages beneficiaries from seeking better paid jobs because it means that earning more from employment reduces the amount of subsidy payments (Bian, 2014;Li and Xiao, 2007). Second, some studies have found that the irrational design of the tied aid systemsuch as the high level of Basic Living Allowance and corresponding benefitsreinforces the incentives of low-income groups to stay in the social security system (Xiao and Li, 2016). Replacement ratio is an important indicator of measuring the relation between social aid and working motivethe lower replacement ratio is, the higher the motivation to work is (Qiao, 2009). However, the gap between the current minimum wage and Basic Living Allowance is too narrow, leading to a very high replacement ratio, which adversely impacts working motive (Zhou, 2012). Finally, there is no positive correlation between Basic Living Allowance and reemployment. As a result, the current employment service cannot assist low-income individuals in exiting the social security system, and other services like employment and training agencies cannot be put into effective practice (Qin, 2017). Although many cities have adopted policies to "withdraw aid step by step," their execution time has been too short and the outcomes have been unsatisfactory (Li and Xiao, 2007).
In 2007, China's central government enacted the New Labor Contract Law to balance labor market flexibility and labor protection. However, some studies have shown that the law has increased the probability of laying-off workers with more than 10 years of employment (Akee et al., 2019). Furthermore, due to SOE restructuring in the late 1990s, around 30m laborers were laid off, but they maintained a relationship with their work unit and hence entitled to receive Basic Living Allowance. The laid-off SOE workers were not counted as unemployed, because they are not eligible for social assistance benefits (Günter, 2009). Moreover, many workers without urban hukou (household registration) are also not qualified for unemployment benefits (Cai and Chan, 2009). The Chinese government focuses on the development of labor-intensive industries, like automobile manufacturing and machine-building, to absorb low-skilled laborers and low-income groups (Cai and Chan, 2009). Such labor policies can mediate the impacts of social assistance system on the employment prospects of low-income groups.

127
Impacts of social welfare system Aside from discussing how the effectiveness of poverty alleviation strategies is affected by individual and systematic factors, limitations which also raise concerns include the ignorance of marginal or low-income groups at legal working age and with working capacity. In this context, on the basis of differentiating the beneficiary and marginal groups of the current security system, this study explores the internal mechanisms that affect the employment of low-income groupsspecifically citizens at legal working age with working capabilityin regard to both individual and systematic factors.

Sampling
The data of this study is drawn from the "Social Policy Support System for Poverty-stricken Families in Urban and Rural China 2015" report. The Ministry of Civil Affairs of the People's Republic of China appointed and funded the Institute of Social Science Survey (ISSS) at Peking University to deliver the related project and organize a research team to write the report. As the target population is low-income groups, purposive sampling is adopted. The sampling frame is provided by the Ministry of Civil Affairs, while ISSS executes administrative procedures, such as proportional population sampling, interviewer training and data collection by Computer Assisted Personal Interviewing (CAPI) (Peking University Open Research Data, 2016). The sample size is made up of more than 170,000 individuals in 29 provinces (excluding the Tibet Autonomous Region, Xinjiang Uygur Autonomous Region, Taiwan, Hong Kong and Macau) between July and December 2015. The report collected data from 4,232 rural families, 7,338 urban families, and 2,609 migrant families. Since the target population of the study is urban families, the authors have excluded ineligible respondents, such as students, retired people, people with no formal education and those lacking working capability. This means the sample size of the study is narrowed down to 2,973. The data have previously been adopted in some studies, like medical financial assistance program (Liu et al., 2017) and the situation of low-income groups (Han and Tang, 2018).
Variables and measurement Dependent variable. "The Status of Employment of Respondents" serves as the dependent variable in this study. This variable is further divided into two categories, namely, the employed and the unemployed. The employed include establishment, contract and temporary workers, self-employed people and farmers; the unemployed include but not limited to laid-off workers, jobless people with working capability, and long-term housekeepers.
Independent variable. The selected independent variables, referenced to the study conducted by Han and Guo (2012), are divided into both individual and systematic variables. Individual variables include gender, age, marital status, education and health condition. The authors have also added a variable by listing the role of respondents in their family so as to examine the impact of family role on the process of job seeking. Systematic variables are divided into three categories, namely, subsistence security system, tied aid system and employment system. The characteristic variables of the subsistence security system include low-income family, the amount of Basic Living Allowance, the length of time to receive Basic Living Allowance, and the evaluation of a sense of stigma to receive Basic Living Allowance. The characteristic variables of the tied aid system include types of per capita tied aid services and per capita government assistance income. The characteristic variables of the employment system include types of per capita employment and entrepreneurship services. The definition of evaluating dependent variable and independent variable can be further seen in Table I.

PAP 22,2
Analytical methods This study adopts Stata 14.0 as a tool to do the data analysis. When it comes to the statistic model, this study adopts the binary logistic regression model and concludes its mathematical expression as follows: where p represents the probability of the occurrence of an event and also represents the probability of the respondents to be employed. logit( p) represents the natural logarithm of the ratio of the probability of occurrence and nonoccurrence. In other words, the more positive the value of logit( p) the higher the likelihood for the respondents to be employed. β 0 is the intercept term of the regression; whereas β 1 and β 2 are the coefficients of RWC and MIGC, respectively. RWC is the individual variables and MIGC is the systematic variable.

Descriptive statistics
As shown in Table II, the employment proportion of the respondents is 62.5 percent, among which the proportion of urban beneficiaries (people entitled to Basic Living Allowance) is 53.0 percent. Urban marginal families (people disqualified from Basic Living Allowance) make for 72.2 percent. It is more likely for beneficiaries to be unemployed.

129
Impacts of social welfare system Table III demonstrates the current amount of Basic Living Allowance received in urban China is relatively low overall with per capita allowance of only 246 yuan per month, while the length of receiving the allowance is close to six years on average and relatively long compared with western countries, in which citizens spend around one year on equivalent allowances (Immervoll et al., 2015). Compared with unemployed beneficiaries, the length of time during which employed people receive the Basic Living Allowance and the amount of Basic Living Allowance per capita they get are lower by 8.4 months and 80 yuan, respectively. The gap of Basic Living Allowance is closely associated with the implementation of the corresponding compensation system and the gap of time to receive allowance may result in beneficiaries without jobs having a higher degree of dependence on the system. Table IV shows which types of current low-income families in urban China enjoy tied aid services of 0.40 per capita. Meanwhile, employment and entrepreneurship services were only 0.13 per capita, indicating an obvious deficiency of employment and entrepreneurship services provided for low-income families. Such a condition may help conclude that the Basic Living Allowance beneficiaries enjoy more types of tied aid services than the marginal ones, and this finding reflects that the procedure of combining social welfare system in urban China with tied aid services significantly constrains marginal beneficiaries from receiving tied aid services.

Regression model
This study establishes three models. Model 1 focuses on the overall sample of low-income groups. Model 2 concentrates on the sample of marginal groups. Model 3 centers on the sample of families entitled to Basic Living Allowance.
Among all individual factors, the three models have similar significant factors for promoting the likelihood of being employed, including being male, having a high self-rated health status, and being a householder. Having a high school education lowers the likelihood of being employed among families entitled to Basic Living Allowance. As a whole, the magnitude of the coefficients in the model of marginal families is higher than that in the model of the families entitled to Basic Living Allowance.
Among all systematic factors, the three models do not share the same set of significant systematic factors. Generally, the higher the income from government assistance, Basic Living Allowance and entitled tied aids, the smaller the likelihood of being employed. However, those who use employment and entrepreneurship services increase the likelihood of being employed. Among marginal families, those entitled to employment and entrepreneurship services have a higher likelihood of being employed. Among families entitled to Basic Living Allowance, the likelihood of being unemployed is increased by the higher per capita of Basic Living Allowance they receive, a higher perceived level of social stigma; and the higher per capita of tied aids. The identified significant factors in the full model may simply consist of all the factors in the model of marginal families and the families entitled to Basic Living Allowance. Without doubt, systematic factors have significant effect on families entitled to Basic Living Allowance. This data demonstrates that individual factors are more influential to the likelihood of being employed among marginal households, whereas more systematic factors affect the likelihood of being employed among the families entitled to Basic Living Allowance (Table V ).

Conclusion
The challenges of social assistance in China include decreasing the number of beneficiariesvia stringent requirementsand introducing target methods, derived from more precise and comprehensive information about applicants, so that some low-income families are dropped from China's social assistance system (Yang, 2018). It is therefore necessary to examine the differences among the different low-income groups, such as marginal families and families  Impacts of social welfare system entitled to Basic Living Allowance. In the midst of reforms in China's social assistance system, the shrinking demographic dividend and the narrowing wage gap between unskilled and skilled workers is leading to lower incentives for schooling and a higher replacement ratio of social assistance (Cai and Du, 2011), as well as, multidimensional objectives of poverty reduction strategies (Wang et al., 2016). Expectedly, poverty reduction should be approached differently when it comes to marginal families and those entitled to Basic Living Allowance. The impacts of individual factors are more influential in marginal families than families entitled to Basic Living Allowance, which are more affected by systematic factors. This study also concludes that the identified individual factors, including being a householder, male and good self-rated health status, are similar to the finding of previous studies. Age and marriage status are not identified as significant factors, perhaps because people across the spectrum in these categories may also be both a householder (which is less likely to be adopted as an independent variable in previous studies) and male. The results also echo research showing that single-mother families face lower chances of employment and a higher risk of poverty (Damaske et al., 2017;Wu, 2011), because female householders spend more time on raising their children and caring for their families. This study also finds that, among families entitled to Basic Living Allowance, having a higher education leads to lower likelihood of being employed. This conflicts with the commonly accepted belief that higher education leads to higher incentive to find work. One of the possible explanations to this is that respondents with higher education often demand better jobs. In other words, it is difficult for the respondents with higher education to find their jobs (usually non-labor-intensive jobs). Perhaps it is more likely for the respondents with higher education to be unemployed intentionally. Indeed, Zhang (2018) verifies that current employment assistance has limited power to encourage young recipients to join the labor market.
As for the most significant systematic factors, our models indicate that families entitled to Basic Living Allowance are less likely to be employed than marginal families. This difference in the two models highlights the adverse effects of "welfare dependence" (among families entitled to Basic Living Allowance) brought about by China's poverty reduction strategy. The finding that employment and entrepreneurship services (wider range of related services) lead to a higher likelihood of being employed among marginal families demonstrates that, compared to families entitled to Basic Living Allowance, employment and entrepreneurship services provide stronger incentives or more relevant services for marginal families to join the labor market.
Compared with labor protection laws of most western countries, China has stronger employment protection regulations, like higher dismissal requirements and more stringent working conditions, i.e. wage and working time (Xie, 2017). According to a recent study, employers avoid the adverse impacts of the Labor Contract Law, by being more likely to dismiss formal-contract workers (Akee et al., 2019). This results in low-skilled workers and poorly educated residents being forced to rely on the social assistance system. Because of this, vocational training is an important element of poverty reduction strategy (Chakravarty et al., 2019) as it equips laborers with suitable and sought-after skills. Therefore, it is more likely for the recipients of employment and entrepreneurship services to be employed, and hence to leave the safety net. Chen and Funke (2009) recognize that the present Labor Contract Law encourages companies to innovate and motivate themselves to employ more high-skilled and highly educated workers. Highly educated workers have a higher expectation of their jobs, and are willing to spend more time on job searching, therefore, higher productivity can lead to further unemployment in urban China (Liu, 2013).

Policy recommendations
As gender and householder status are significant variables among low-income groups in urban China, the patriarchy must have a large effect on the employment prospects of 132 PAP 22,2 family workforce. In a male-dominated society, wage labor takes precedence over house work and people are mostly assigned to these two kinds of work according to their gender. Females need to take care of chores and are therefore forced to be jobless, while males are expected to be wage labors outside of their homes. Clearly, the equality of men and women should be actively advocated so as to solve the problem of gender discrimination (especially invisible discrimination) in the labor market. Family-friendly policies should be developed with the goal of building a harmonious working and family relationship, i.e., binding government, community and family together to help promote women's employment.
Health is one of the most crucial factors influencing the employment prospect of the workforce (Luo et al., 2010). Illness derived from poverty and poverty caused by illness are both common phenomena that greatly hinder the employment prospects of many low-income people in urban China (Liu and Zhang, 2018). As a result, the improvement of healthcare systems, including, but not limited to, basic medical insurance, and medical assistance and healthcare, would improve the health of the young workforce in urban China in poverty. In particular, medical assistance policies should be strengthened and earnestly implemented, in order to provide low-income workers with better and more precise medical care (He and Nolen, 2019).
The current social assistance system reduces low-income people's incentive to work and encourages welfare dependence among its beneficiaries. It is necessary to note that "welfare dependence" can have different meanings in China and the West. In Western welfare states, "welfare dependence" is mainly measured by indicators, such as the length of time welfare is received, the number of instances a recipient has claimed welfare and the amount of subsidizations. It is usually concluded that subsidization is too high (Barrett, 2000). However, welfare dependence in China, as discussed in this study, places greater focus on the adverse incentive impacts generated by welfare systems. The authors make no suggestion that welfare dependence is caused by the current subsistence security system in urban China being too generous. Furthermore, "welfare dependence" is often employed by neo-conservatives in Western welfare states to attack their own welfare systems. Thus, from the perspective of the Chinese government, a balanced development of the welfare system is certainly a critical key to implementing its poverty alleviation strategy. In the midst of a diminishing demographic dividend and the rising burden of an aging population, the welfare system's adverse incentive impacts on employment should be highly stressed, but not the benefit level. Borrowing the findings of this study, an increase of per capita Basic Living Allowance, would help generate more obvious adverse incentive impacts on low-income groups. As a solution, a scientifically based Basic Living Allowance adjustment mechanism in urban China should be established (Zhong, 2012). Followed by changes in both the domestic and the international economic environments, such a standard can be adjusted in a comprehensive way with the per capita income of residents as the main factor and expenditure as a supplement. In this context, "welfare dependence" resulting from a very high standard, and lack of benefit deriving from a very low standard are, as a result, to be practically prevented (Wong et al., 2014).
A reduction in stigma to receiving Basic Living Allowance as expected would lower the likelihood of being employed among the families entitled to Basic Living Allowance (Contini and Richiardi, 2012). The finding indicates that among welfare recipients, the "stigmatization effect" of social assistance is diminishing; in other words, the welfare recipients believe that it is reasonable for them to depend on the welfare system. Moreover, the tied aid system is still considered as an important cause of unemployment to its beneficiaries. Most of them are therefore reluctant to seek jobs and exit from this system, as they rely too much on their fringe benefits, especially housing and medical assistance. Furthermore, the current employment and entrepreneurship service system cannot work effectively to encourage its beneficiaries to work. This is because the beneficiaries have 133 Impacts of social welfare system working capability but are lacking adequate education and working skills, or the channel and function of their investment in human capital would still have certain restrictions. For example, some projects in the employment and entrepreneurship service system fail to meet all the needs of low-income groups and employers. Being a main form of the supply of employment and entrepreneurship service, occupation programs can only offer menial and labor-intensive jobs, like security guards and cleaners, who are paid the minimum wage or slightly above (Wang, 2018). Worse still, the income of such jobs is not steady, so that the impetus of systems on employment is relatively weak. For the sake of being entitled to Basic Living Allowance, few beneficiaries would turn down their potential job offers three times in a row. Even though they have to take up the job offered by the system, they would normally quit the jobs by very similar reasons of poor health or having no related skills, and again, they become unemployed in order to benefit from Basic Living Allowance and the corresponding fringe benefits. To effectively overcome such a serious policy dilemma, targeted occupational assistance and guidance should be further introduced by the Chinese government, in order to help current beneficiaries in accordance with their education, working skills, service hours and workplace (Duckett and Hussain, 2008). In particular, vocational training and education should be provided to recipients who lack working skills to strengthen their competitiveness in the job market; counseling services should be given to those who are reluctant to work to increase their willingness to work; and social services should be provided to the female recipients to ease their barriers to join the labor market again (Guan, 2014;Solinger, 2005). In China, the balance between work and family life has become one of the most crucial aspects of its poverty reduction strategy, which can be considered as a Pareto improvement to cater for the needs among different stakeholders.