Determinants of internal migration control : explaining the degree of difficulty of acquiring Hukou

Determinants of Internal Migration Control:

Explaining the Degree of Difficulty of

Acquiring

Hukou

Jonathan Dul (10003517)

Bachelor thesis Political Science

Words: 8,106

26 June 2014

Table of contents

INTRODUCTION 2

THEORY 5

Frameworks for immigration control policy analysis 5 Factors that influence the degree of difficulty of acquiring hukou 7 Conceptual model and propositions 10

METHODOLOGY OF EMPIRICAL TESTING 13

Sample selection 13

Measurement 14

Data analysis 17

EMPIRICAL RESULTS 19

Descriptive statistics 19

Results of the hierarchical multiple regression model 19

DISCUSSION 22

Theoretical contributions 22

Limitations and future research 24

CONCLUSION 25

REFERENCES 26

Introduction

Migration exists in many different shapes and forms. The first and foremost form of migration that comes up to many people’s minds when talking about migration is international migration, people crossing state-borders. International migration is widely discussed both in the academic world and in society, where migration issues are strongly affecting people’s political (voting) behavior. However, internal migration, migration of people within a country, is also increasingly apparent since an increasing number of people in the world is moving from rural to urban areas. China is one of the countries where there is an enormous urbanization: in 1990 295 million Chinese people (26% of China’s population) lived in urban areas, and in 2012 this number had increased to over 702 million people (52% of China’s population) (World Bank data, 2013). These numbers indicate the vastness of China’s internal migration, and they also explain why it is not surprising that the Chinese government has implemented policies that enable them to control this internal migration.

The most important measure to control this internal movement of people in China is China’s

hukou system, a household registration system. In mainland China, all citizens have a hukou

registration, which is determined by a citizen’s place of residence, either rural or urban. One is only allowed to register in one city, town, or village, and to change a hukou status is a complicated and unequal process. To acquire an urban hukou, someone is most of the time required to be either rich or highly educated, which results in many rural migrants being denied an urban hukou from a Chinese city. Looking at the level of decision-making in the hukou system, a change has been noticed to be going on since the 1990’s. Local Chinese governments now have full authority to decide to whom they grant a hukou and how many hukous they want to grant per year, which results in hukou regulations that allow mostly the rich and the highly educated to enter the cities. By localization of power in the hukou system, the Chinese government managed to delegate the fiscal responsibility (or fiscal burden) for the non-agricultural hukou population towards the cities (Chan and Buckingham, 2008: 591). Now that the policymaking process with regards to the hukou system is localized towards the city governments, every city is allowed to set their own entry requirements, which has resulted in differences between cities in the degree of difficulty of acquiring a local hukou (Zhang & Tao, 2012). This localization of power in China is seen in more areas of Chinese politics, with the decentralization of economic and administrative power towards the local levels since the economic reforms in 1978 (Ma, 2002). Before the localization of

difficulty of acquiring hukou between cities. However, now that the whole decision-making process is localized, cities can decide for themselves what kind of people they permit to enter the city. The resulting difference between cities in difficulty of acquiring hukou is barely discussed in the literature, and only recently an index was constructed to measure these differences of degree of difficulty of acquiring hukou of a Chinese city (Zhang & Tao, 2012).

Although not much has been written on the localized hukou system, the hukou system in general and its consequences are a popular topic. Many scholars focus on the division it creates in Chinese society, with big differences between the rural and the urban population (among others: Chan, Liu, & Yang, 1999; Wu & Treiman, 2004; and Wang, 2005). In this context some authors relate to the South-African apartheid system which also resulted in a segregated population and a division between rural and urban population (Alexander & Chan, 2004; Chan & Buckingham, 2008). The consequences of the hukou system are also extensively discussed in the literature, especially in relation to the widening socioeconomic gap between rural and urban citizens (Fan, 2002; Wu & Treiman, 2007; and Wang, 2012). This thesis will not focus on these unequal outcomes of the hukou system, but will rather focus on the state (city) actors and the underlying factors that influence the earlier mentioned degree of difficulty of acquiring urban hukou. Why, if the outcomes of the hukou system are creating so many inequalities between Chinese citizens, do cities maintain severe hukou regulations? It is important to look deeper into the factors that influence the difficulty of acquiring hukou to understand why Chinese cities are supporting the resulting socioeconomic gap.

Some of the motivations for the Chinese national government to uphold the hukou system on a national scale are already being discussed in the literature. The hukou system originated from a system that was used to distribute resources among the Chinese people, and to provide the urban citizens with food from rural areas. However, soon after that it developed into a system that was used to control migration flows from China’s rural areas to the urban cities (Cheng & Selden, 1994). Nowadays, two functions of the hukou system are most important: regulation of internal migration and social control of targeted segments of the population (Wang, 2005: 86). Some scholars elaborate on the sociopolitical purposes of the hukou system, especially in modern-day China where the hukou system enables the PRC police to closely monitor people that they targeted as either counter-revolutionary, criminal, or a threat to the public order (Wang, 2004). However, economic motivations are most often recognized as the main reason for the Chinese government to sustain the hukou system (among others: Wang, 2004; and Chan & Buckingham, 2008). Some argue that the rural-urban gap, which is enhanced by the hukou system, provides

China with the necessary cheap labor that is the engine for the country’s economic growth (as discussed in: Chan, 2010). All these mechanisms, however, focus on the motivations of the national government to maintain the hukou system, but the motivations of city governments to implement strict or loose hukou policies are not discussed. What are the underlying motivations for many cities to set a high degree of difficulty of acquiring hukou, enabling only very rich and highly educated people to enter the cities?

From the above, several things can be derived. Since the localization of the hukou management, the power of decision-making in the hukou system belongs to cities now, and they are still rejecting many applications of poor rural migrants, while very rich and/or highly skilled and educated people are often granted an urban hukou. By doing this, Chinese city governments are upholding, or maybe even increasing the Chinese socioeconomic gap. However, some cities pose harder requirements on the hukou applicants than others, resulting in differences in the degree of acquiring hukou. This poses an interesting research question: which factors influence the degree of difficulty of acquiring hukou in Chinese cities?

This research will define several factors that can influence the degree of difficulty of acquiring a

hukou of a Chinese city. After presenting a theoretical conceptual model, a statistical analysis will

Theory

Immigration policy is a much discussed topic, but, as Eytan Meyers (2000) justly argues, a distinction has to be made between immigration policy and immigration control policy. Immigration policy consists of the conditions provided to immigrants, whereas immigration control policy is the policy that governs the selection and admission of foreigners (Meyers, 2000: 1246). Here, the focus will lie on immigration control policy, which can be influenced by all kind of factors.

Frameworks for immigration control policy analysis

In the field of immigration control policy multiple factors are being recognized as important in shaping policies. Different theoretical frameworks are used in identifying these factors and in analysing the determinants of immigration control policy. Meyers (2000) delineates six major approaches in the field of immigration control policy: Marxist approach, national identity approach, domestic politics approach, institutionalist approach, realist approach, and the liberal approach. Cornelius and Rosenblum’s findings (2005) largely overlap with Meyers’ article and recognize many of the same frameworks in analysing immigration control policy. Before focussing on the possible determinants of the hukou policy it is necessary to briefly elaborate on the different approaches in framing immigration control policy.

Marxist approach

The fundamental idea of the Marxist approach is that the ruling capitalist class shapes immigration control policies in order to increase their own profits. One of the mechanisms to do so is by encouraging a loose immigration control policy in order to create an influx of migrants to press wages, which results in increasing the capitalist class’ profits (Meyers, 2000: 1247). All explanations that explain immigration control policy as a means of exploitation of workers for the purpose of capital accumulation can be gathered under the Marxist or neo-Marxist approach.

National identity approach

The national identity approach does not recognize economic and situational factors as important in explain immigration control policy, since it argues that a country’s history and its feeling of citizenship is crucial in this process (ibid.: 1251). Nationalism, nativism, and xenophobia are all concepts that are related to one another and influence immigration control policy. It is argued that the difference between countries’ openness towards immigration is based on their differences in nationalist sentiments and the feeling of national identity. A strong feeling of

national identity, not infrequently accompanied by xenophobia and nativist sentiments, results in stricter immigration control regulations (Brubaker, 1995: 903).

Domestic politics approach

The domestic politics approach is also often used in explaining immigration control policy (Meyers, 2000: 1257; Cornelius & Rosenblum, 2005: 107). This approach focuses on domestic socioeconomic factors that influence domestic politics and therefore also influence immigration control policy outcomes. Interest groups can influence policymakers in the political arena, which results in the assumption that immigration policy is a form of client politics, with different groups looking for ways to steer policy in their preferred direction.

Institutionalist approach

Institutionalists focus on the state as a central actor, but they also recognize interactions within the state. As claimed by the institutionalist approach, political institutions shape policy accordingly with state interests, without being influenced by society and interests groups (Meyers, 2000: 1261). Immigration control policy can therefore be explained by looking at state interests, and how these interests are defined by the political institutions (Cornelius & Rosenblum, 2005: 108).

Realist approach

The realist approach is most often used in international relations, focusing on the state as a unitary actor that always acts rationally. With regards to immigration control policy, realists focus on security and military conflicts, and how both influence migration policy (Meyers, 2000: 1264).

Liberal approach

Unlike realists, liberals think that cooperation among states is possible, which is seen in the neoliberal thinking of international institutions to overcome problems related to migration. Also, liberal scholars focus on globalization and how this influences immigration policy making (ibid.: 1266).. They focus on the emergence of international regimes and how they are influencing local policies, diminishing the authority of national and local governments. Moreover, the importance of economic globalization and transnational citizenship are recognized as important factors that can help in explaining immigration control policy (ibid.: 1267).

Factors that influence the degree of difficulty of acquiring a city’s hukou

When defining factors that could possibly play a role in the difference between cities in the degree of difficulty of acquiring hukou, it is worthwhile to look at both international migration literature and to the specific literature on the hukou system. Just as the international migration control literature focuses on factors that shape national policies, this thesis focuses on factors that shape city policies, and therefore the causal mechanisms that are apparent in international migration control policymaking could apply to China’s internal migration control policymaking. One concept that is often recognized in the international migration literature as a factor that plays a role in shaping immigration control policies is a country’s national identity. Attitudes towards immigration often reflect nationalist sentiments, and so do immigration policies (among others: Brubaker, 1995; Higham, 2002; Wimmer & Glick Schiller, 2002: 309-321; O’Rourke, 2006). It is a conventional view that a strong feeling of national identity, not infrequently accompanied by xenophobia and nativist sentiments, results in stricter immigration control regulations (Brubaker, 1995: 903). This same mechanism could play a role in the Chinese hukou policy. Nativism is certainly apparent in Chinese city-societies: urban citizens look down on rural citizens and there is a strong sense of urban citizenship, partly defined by who has a hukou and who does not (Zhang, 2002). As shown in the international migration control literature, a stronger nationalist/nativist sentiment could lead to stricter migration control policies. Reflected on the

hukou system, it could mean that a strong feeling of urban citizenship in a Chinese city leads to a

stricter hukou policy in that city. This first factor, nativist sentiments, is easiest recognized through the national identity framework, and would, according to this approach, play an important role in explaining a Chinese city’s degree of difficulty of acquiring hukou.

Another factor that is often identified as important in explaining international immigration control policy is a country’s demand for labour. During the 1960’s and 1970’s this was clearly seen in Europe, with high labour demands and flexible immigration control policies in the 1960’s, and low labour demands and strict immigration control policies in the 1970’s (Zimmerman, 1996: 96). Employers often claim to need migrant labour in order to fill labour shortages, and therefore immigration control is often loosened in times of high labour demand (Ruhs & Anderson, 2010). To protect the economic interests of employers, interest groups can express their concerns with regards to a high labour demand and try to steer the government to implement immigration policies favourable to immigrants (Freeman, 1995). The same could account for the hukou system in Chinese cities. When a city’s industry is in need of labour, interest groups can influence policymakers in order to adjust the hukou policy to lower the barriers for city entry. This second

factor, a city’s demand for labour, fits very well in the domestic politics framework, by focusing on the interests of a city’s employers and the related interest group politics to steer policy.

A third factor that is often related to migration control in the international migration literature is security. The realist approach most often recognizes security reasons as an important factor in influencing migration control policy. By focusing on state interests in international relations, it is argued that countries use migration policy to create a strong population in times of a possible threat (Meyers, 2000: 1264). On the other hand, actual or potential conflicts could also lead to stricter immigration policies, when a country does not want to accept all refugees that wish to enter the country (ibid.). At the moment that a country is not equipped to handle mass migration, it has to implement stricter immigration control policies for security reasons (Cornelius & Rosenblum, 2005: 109). In a Chinese city this same mechanism could lead to a higher degree of difficulty of acquiring an urban hukou. When a city is not equipped to handle many migrants entering the city, it can pose a public security threat. Due to public security reasons some cities might have to tighten their immigration control policy to prevent mass migration of rural citizens into their areas. Although the realist framework takes security reasons into account when it comes to international relations, I think that the institutionalist approach is better equipped to analyse this mechanism in the Chinese context, since it acknowledges the state as a central actor, but also recognizes interactions within the state. In the case of China’s cities, it means that government institutions that are responsible for public security can influence their counterparts that are occupied with hukou policy, which can lead to stricter or more flexible hukou policy. In this case, if public security is threatened by mass migration of rural citizens into urban areas, the

hukou policy could become stricter.

The fourth factor derives from Neil Smith’s article New Globalism, New Urbanism: Gentrification as

Global Urban Strategy (2002). One of the arguments he makes in this article is that there is a

division between a city’s core and its periphery, resulting from global production systems that no longer require surplus populations. He argues that the core is embracing global capital, but is not so much embracing surplus populations because they do not need those populations. In the case of China, this would mean that cities embracing foreign direct investment (FDI) would not need the migrants anymore, which could lead to stricter hukou regulations. Identifying FDI as a factor which impacts hukou policy fits well into the Marxist framework. While the core’s economic interests are supported through accumulation of FDI, migrants are excluded from the cities by the hukou system.

The fifth factor that can influence migration control policy is globalization. It is argued that state sovereignty in immigration control is diminishing, due to transnationalization of economic spaces and politics (Sassen, 1996). The existence of an international human rights regime is also a result of globalization, which influences policymakers to establish humane immigration control policies. In this same way globalization could also influence cities’ hukou policy, which is often seen as a policy that diminishes migrants’ human rights. A city with a higher level of globalization could be more likely to be sensitive to the international human rights regime than a less globalized city, which could result is less strict hukou policy in order to meet the international norms. This focus on international cooperation and regimes, and their ability of these regimes to diminish authority of local governments, fits into the liberal framework. It shows how a city’s level of globalization and transnational efforts can influence the city-level hukou policies.

Besides the factors above, there are also factors that can influence the degree of difficulty of acquiring hukou deriving from literature that discusses China’s national hukou system in general. One of those factors is the level of education of a city’s population that already possesses the city’s hukou. It is known that city governments prefer to grant hukou to the highly educated applicants (Wang, 2004: 119; Chan & Buckingham, 2008: 595; Chan, 2010: 361), which increases a city’s human capital and is good for its economy. However, it is conceivable that cities which populations are not highly educated will have lower educational requirements for hukou applicants than cities that already have a highly educated population. This shows a possible factor that influences a city’s degree of difficulty of acquiring hukou: level of education. The institutionalist framework is able to recognize this mechanism, by focusing on how policy is shaped in order to meet the city’s interest. In this case, the city’s interests are served when a higher educated population is generated through the hukou policy.

The same can be said about a city’s average wage level, which is another factor that can influence a Chinese city’s hukou policy. Just as educated people are wanted by Chinese city governments, so are the wealthy people (Wang, 2004: 129; Chan & Buckingham, 2008: 595). Also for this factor, the same kind of mechanism could be explanatory: a city with a high level of average wage leads to city governments wanting only richer people to enter their cities and acquire a hukou, in order to protect the city’s economic interests. Also here, and with the same reasoning as above, the institutionalist framework suits best to identify this relation.

A third factor deriving from the hukou literature that can influence a city’s entry barrier are the costs of a city’s government provided benefits. A Chinese city has to provide certain public services to its inhabitants, but only to the inhabitants that possess a city’s hukou. It is argued that

many cities maintain strict hukou regulations and allow few people to enter their areas, in order to press the costs of government provided benefits (Chan & Buckingham, 2008: 596; Zhang & Tao, 2012: 2885). By doing this, a city government is able to protect its finances and can make sure not too much money has to be spent on migrants. Likewise, this protection of a city’s financial interests by maintaining a strict hukou policy is best recognized by the institutionalist approach, concentrating on the political institutions and their interests.

The last factor deriving from the hukou literature is the influence of elite groups on Chinese city politics. As argued before, highly educated and rich people are often welcomed by Chinese cities and are granted an urban hukou, while poorer and less skilled migrants are sacrificed in this process. The hukou system is known as a system that protects the interests of China’s elite, while at the same time being a tool for this elite to repress and control (Wang, 2005: 148). A city government that is influenced a lot by the city’s elite could be less likely to loosen its hukou policy, since it is one of the elite’s most important tools to maintain their status. This mechanism is best analysed through the domestic politics framework, which focuses on the influence of different interest groups on policy, and is well equipped to recognize the way the Chinese city elite is able to influence the degree of difficulty of acquiring hukou.

Conceptual model and propositions

Taking all theoretical assumptions as mentioned above into account provides a conceptual model that shows how several factors can influence the degree of difficulty of acquiring a city’s hukou. The conceptual model as seen in figure 1 shows which factors can influence the degree of difficulty of acquiring a hukou of a Chinese city. Following from this theoretical model, the following propositions are put forward:

P1: More elite group influence on a city’s policy leads to a higher degree of difficulty of acquiring that city’s hukou.

P2: A higher demand for labour in a city leads to a lower degree of difficulty of acquiring that city’s hukou.

P3: A higher level of education in a city leads to a higher degree of difficulty of acquiring that city’s hukou.

P4: More foreign direct investment in a city leads to a higher degree of difficulty of acquiring that city’s hukou.

P5: A higher average wage level in a city leads to a higher degree of difficulty of acquiring that city’s hukou.

P6: A higher public security threat in a city leads to a higher degree of difficulty of acquiring that city’s hukou.

P7: A higher level of globalization in a city leads to a lower degree of difficulty of acquiring that city’s hukou.

P8: More nativist sentiments in a city lead to a higher degree of difficulty of acquiring that city’s hukou.

P9: Higher costs of government provided benefits in a city leads to a higher degree of difficulty of acquiring that city’s hukou.

Figure 1: Conceptual model Elite group influence Demand for labor Level of education

FDI

Average wage level Degree of difficulty of _{acquiring a city’s hukou} Public security threat

Level of globalization Nativist sentiments Costs of government provided benefits +

-

-

Importance of economic factors in Chinese policy-making

The factors mentioned in this thesis that can influence the degree of difficulty of acquiring hukou in a Chinese city can roughly be divided into two categories: sociopolitical factors and economic factors. Elite group influence, level of education, public security threat, globalization, and nativist sentiments all fit into the sociopolitical category, whereas demand for labour, FDI, average wage level and costs of government provided benefits are all factors that fit better into the economic category.

In Chinese policy decision making, economic criteria and economic factors have become increasingly important since the 1980’s (Lieberthal & Oksenberg, 1988: 334). With the implementation of the Four Modernizations policy, focusing on modernization in agriculture, industry, national defense, and science and technology, the importance of economic development became central in Chinese policymaking. This leads to the expectation that in Chinese cities’

hukou policymaking, the economic factors as mentioned above are crucial as well.

Deriving from the assumption that states that economic determinants are crucial in shaping Chinese policy, it is expected that proposition 2, 3, 4, 5, and 9 will have a higher explanatory power than the other propositions. This means that the economic factors are more important in explaining the degree of difficulty of acquiring a hukou than the sociopolitical factors.

Methodology of empirical testing

The conceptual model as shown before in this thesis comprises of nine different factors that can influence the degree of difficulty of acquiring a city’s hukou. However, in the empirical part of this thesis, where the theory as built above will be tested, only propositions 1, 2, 3, 4, and 5 will be included, due the practical difficulty of the selection of all factors in the timeframe of writing this thesis. This means that the following hypotheses will be tested:

H1: More elite group influence on a city’s policy leads to a higher degree of difficulty of acquiring that city’s hukou.

H2: A higher demand for labour in a city leads to a lower degree of difficulty of acquiring that city’s hukou.

H3: A higher level of education in a city leads to a higher degree of difficulty of acquiring that city’s hukou.

H4: More foreign direct investment in a city leads to a higher degree of difficulty of acquiring that city’s hukou.

H5: A higher average wage level in a city leads to a higher degree of difficulty of acquiring that city’s hukou.

Before turning to testing the hypotheses above and answering the main question of which factors influence the degree of difficulty of acquiring hukou in Chinese cities, the methodology of this research will be discussed.

Sample selection

China had 655 cities by 2008 (Hong Kong and Macau not included), that all can implement their own specific hukou policy (Zhang & Tao, 2012: 2888). Four of those cities are province-level cities, the highest possible rank for a city: Beijing, Chongqing, Shanghai and Tianjin. These cities fall under direct control of the central government and can be seen as a province on their own. Then there are also subprovince rank cities (15), prefecture rank cities (268), and county rank cities (368), of which the first have the most prestige and the latter the least.

In selecting the sample for this research, the same criteria as Zhang and Tao used in their article

Barriers to the acquisition of urban hukou in Chinese cities (2012) will be used:

‘First, the sample of cities should cover all provincial-level regions of the country. Second, the sample cities should be desirable destinations for migrants and were in the list of top 100

migrant cities for 2009. These criteria resulted in a sample of forty-six cities. Though satisfying the two criteria, Lhasa, the capital city of Tibet, was excluded from our analysis owing to lack of data on important variables’ (Zhang and Tao, 2012: 2889).

Due to a lack of available data for three of the above mentioned 45 cities, the sample of this research consists of 42 cities and therefore covers 6.4% of the whole population of 655 Chinese cities. The average size of urban area population of the sample is 3.45 million, with Shanghai as the largest population 15.26 million and Jiujian as the smallest city with 0.60 million. For the list of names of the incorporated 42 cities, see Appendix I.

Measurement

Dependent variable

The dependent variable in this research is a Chinese city’s degree of difficulty of acquiring hukou, which is defined by Zhang and Tao and describes a city’s entry barrier (CEB). They have introduced the CEB index (see Appendix II), on which a higher value indicates a higher degree of difficulty of acquiring hukou (Zhang & Tao, 2012: 2891).

The CEB index ‘is interpreted as a comprehensive assessment of the difficulty that an applicant meets in acquiring the city’s hukou’ (2012: 2888). This assessment is based on the requirements that an applicant has to meet in order to qualify for a hukou. Education level and a person’s wealth are the main indicators for evaluation, but many other indicators play a role as well. For an extensive scheme of indicators for evaluation of an applicant’s qualification see Appendix III. The lowest score possible on the scale of the CEB index is 0 (no difficulty of acquiring hukou), while the highest score possible is 3.5 (extreme difficulty of acquiring hukou).

Independent variable

1. Elite group influence

The variable elite group influence has to measure the level of influence that a Chinese city’s elite has on the city’s policymakers. This concept is very hard to measure, and extensive research on China’s city-elites would be necessary to come up with an index that accurately measures the elite group influence on Chinese city’s policymakers. However, it is said that China’s elite group influence leads to a decline in transparency, meaning that less transparency and more elite group influence are correlating (Dittmer, 2001: 55). Therefore, measuring transparency of these cities is a worthy measurement that resembles elite group influence.

An index on fiscal transparency of Chinese cities (see Appendix IV), published by the School of Public Policy at Tsinghua University (2013), will be used to measure a city’s transparency, and therefore elite group influence. The maximum score of full transparency is 400, while the minimum score is 0. It is measured by looking at four criteria: ‘1) performance in publicizing information about structures and functions of government organs and government-financed agencies; 2) performance in announcing budgetary revenues and expenditures; 3) performance in releasing extra-budgetary funds and local debts; and 4) performance in reporting the financial situations of state-owned business’ (Zhang, 2013). Since the relationship between elite group influence and transparency is negative, it is necessary to reverse the scores for transparency into the new scores of the variable elite group influence (elite group influence score = 400 – transparency score).

2. Demand for labour

There are no direct data available that measure a Chinese city’s demand for labour. However, some argue that when unemployment rates rise, the need for migrant workers declines (Ruhs & Anderson, 2010: 2). This relationship makes it possible to measure a Chinese city’s demand for labour by looking at the city’s unemployment rates, assuming that the demand for labour is high when unemployment rates are low.

The unemployment numbers of Chinese cities are available in the latest accessible China City Statistical Yearbook 2008 (中国城市统计年鉴) (PRC Bureau of Statistics, 2008). When using the statistical yearbook, it is important to only look at the data of cities’ urban areas (市辖区), because these are the areas that are relevant with regards to acquiring urban hukou. In this Statistical Yearbook the unemployment numbers are measured by looking at the number of persons that are unemployed (for an example of data in this yearbook, see Appendix V). In this thesis unemployment rates will be measured by dividing the number of unemployed persons in a city by all persons in that city’s labour force. The city’s labour force can also be found in the China City Statistical Yearbook, by adding the number of people working in a danwei (work unit) to the private and individual employed people, and to the number of unemployed people. Since the relationship between demand for labour and unemployment rate is negative, it is necessary to reverse the unemployment rates into the new scores of the variable demand for labour (demand for labour score = 100 – unemployment rate).

3. Level of education

A city’s level of education should indicate how well educated a Chinese city’s population is. Since the exact data to measure this variable is not available, it is necessary to create a variable that measures a city’s level of education in the best possible way. In this research the number of enrolled students in a city will be used to measure a city’s level of education. It is shown that highly educated cities are often home to major universities, which makes the number of enrolled students a good indicator for a city’s level of education (Winters, 2011: 255).

Data on the number of enrolled students per city is available in the China City Statistical Yearbook 2008. In order to measure a city’s level of education, the number of students enrolled in colleges and universities will be divided by the city’s population.

4. Foreign direct investment

The variable foreign direct investment is rather straightforward and easily measured by using the China City Statistical Yearbook 2008. In this book the, actual used foreign direct investment numbers are available per city, measured in millions of yuan (1 yuan = 0.12 euro).

Since a city with a larger economy is more likely to have more incoming FDI, it is necessary to view a city’s foreign direct investment as a percentage of a city’s GDP in order to measure the effect of FDI on a city’s entry barrier: (FDI/GDP) * 100%.

5. Average wage level

Also a city’s average wage level is available in the China City Statistical Yearbook 2008. This variable is also measured in yuan.

Control variable

In order to be able to accurately determine the effect of the above mentioned independent variables on the dependent variable, it is important to take other variables into consideration that could influence the outcome of this research. This research will make use of a regression analysis, and therefore it is necessary to control for a variable that could also possibly produce the outcome of the dependent variable. In this case it is important to control for a city’s gross domestic product (GDP) per capita. GDP per capita is expected to have an effect on a Chinese city’s degree of difficulty of acquiring hukou since ‘the cities that set higher qualifications for urban hukou are those generally perceived as more prosperous economically’ (Zhang & Tao, 2012:

Data on Chinese cities’ GDP per capita is available in the China City Statistical Yearbook 2008, and is measured in yuan.

Data analysis

In order to test the hypotheses that derive from the theoretical conceptual model, a hierarchical multiple regression analysis was executed using SPSS. A multiple regression analysis is a tool to evaluate the relationship between several independent variables and one dependent variable. A multiple regression analysis provides an estimate of the size and direction of the contribution of an independent variable to the dependent variable (=standardized regression coefficient, Beta), while all the other independent variables are held constant. Also, a multiple regression analysis shows how much variance in the dependent variable is explained by the set of independent variables (=adjusted R square). A hierarchical multiple regression analysis is a multiple regression analysis in which groups of variables are entered successively in order to estimate the explained variance of each group of variables, but also controls for the effect of other independent variable(s) on the dependent variable. In the case of factors that influence Chinese cities’ hukou policies, a two-step hierarchical multiple regression analysis is used. In the first step the control variable is entered (first model), and in the second step the independent variables are added as well (second model). This division into two stages provides the possibility analyse how important all the factors that influence Chinese cities’ hukou policy, entered in the second stage, are in comparison to the control variable, by looking at the model’s R square change. The hierarchical multiple regression analysis will allow to determine the effect of the five independent variables on the degree difficulty of acquiring hukou of a Chinese city, while controlling for the effect of the city’s GDP per capita.

In a cross-sectional research like this, the famous phrase “correlation does not imply causation” also accounts. However, thanks to the theoretical assumptions that the different factors influence the degree of difficulty if acquiring hukou of a Chinese city, causation can be assumed, although it cannot be proved by regression analysis.

All the hypotheses as formulated before assume a specific direction of the effect of one factor on the degree of difficulty of acquiring hukou of a Chinese city. Therefore, the significance of these hypotheses was tested by doing a one-tailed test and a level of significance of 0.05 was used. This means that a hypothesis is accepted if f sig. <0.05 and the direction of the effect (positive or negative) is according to the expectation.

Several presumptions are needed to be able to perform a multiple regression analysis. One important presumption is that the independent variables are relatively independent. This can be tested by testing the multicollinearity of the independent variables, to make sure that two or more of the factors that influence Chinese cities’ hukou policy are not highly correlated, and thus to avoid incorrect estimation of the regression coefficients. Multicollinearity was tested by looking at the tolerance levels: It is commonly assumed that a tolerance of more than 0.20 indicates that there is no multicollinearity problem.

Empirical results

Table 1 shows the descriptive statistics of the independent and dependent variables of the 42 Chinese cities in the sample. Especially the variation in degree of difficulty (Mean: 0.958, Standard Deviation: 0.685), and the variation in FDI (Mean: 0.591, Standard Deviation: 0.390) are high, meaning that the data is more spread out among these variables. On the other hand, the data on elite group influence (Mean: 299.141, Standard Deviation: 37.670), and demand for labour (Mean: 0.966, Standard Deviation: 0.016) does not have a high degree of variation.

Table 1: Means, standard deviations, and Pearson correlations

Variable M S.D. 1. 2. 3. 4. 5. 6.

1. Degree of difficulty of acquiring hukou

0.96 0.69

2. GDP per capita (in yuan) 45365 20207 0.65*

3. Elite group influence 299 38 -0.66* -0.31*

4. Demand for labour 0.97 0.016 0.47* 0.37* -0.15

5. Level of education 0.091 0.044 -0.31* -0.64 0.26 -0.16

6. FDI 0.59 0.39 0.23 0.49* -0.12 0.076 0.021

7. Average wage level (in yuan) 28597 7062 0.86* 0.69* -0.57* 0.51* -0.29* 0.19 * p<0.05, N = 42

In table 1, also the Pearson correlation coefficients are presented to measure the degree of association among the variables. A correlation coefficient of 1.0 indicates perfect correlation, whereas a correlation coefficient of 0 indicates no correlation. According to the ‘rule of thumb’ a correlation of 0.70 – 0.90 is strong, a correlation of 0.40 – 0.60 is moderate, whereas a correlation of 0.1-0.3 is weak (Dancey & Reidy, 2007).

Results of the hierarchical multiple regression model

The results of the hierarchical multiple regression analysis are presented in Table 2. Model 1, with only GDP per capita as independent variable is significant (sig .000), and so is model 2, with all independent variables incorporated (sig .000).

Table 2: Results of hierarchical multiple regression analysisa

1* 2*

B (SE) Bèta B (SE) Bèta GDP per capita 2.19E-05

(0.00)

0.65* 5.21E-05 (0.00)

0.15

Elite group influence -0.005

(0.002)

-0.28*

Demand for labour 3.80

(3.97) -0.089 Level of education -0.93 (1.27) -0.060 FDI 0.029 (0.16) 0.017

Average wage level 5.14E-05

(0.00)

0.540*

R2 _{0.42 0.80}

Adjusted R2 _{0.40 0.76}

* p<0.05

a. Dependent variable: Degree of difficulty of acquiring hukou

Model 1 shows that the control variable GDP per capita has a positive effect on the degree of difficulty of acquiring hukou, the dependent variable (bèta: 0.65, and sig.<0.05). This model explains 40% of the variance in the degree of difficulty of acquiring hukou (adjusted R2_{: 0.40).}

Model 2 shows that, when incorporating all independent variables, the explained variance in degree of difficulty of acquiring hukou is 76% (adjusted R2_{: 0.76). This means that model 2}

explains 36% more variance than model 1, which only includes the control variable. Also, the importance of GDP per capita on the dependent variable strongly decreases when incorporating all the independent variables. Also, the results of model 2 show that elite group influence has a negative effect on the degree of difficulty of acquiring hukou (bèta: -0.28, and sig.<0.05). This rejects hypothesis 1: More elite group influence on a city’s policy leads to a higher degree of difficulty of acquiring that city’s hukou. The same model shows, that no statements can be made about hypothesis 2 to 4, due to a lack of significance of the variables demand for labour, level of education, and FDI (sig.>0.05). In model 2, average wage level has a strong positive effect on the degree of difficulty of acquiring hukou (bèta: 0.540, and sig.<0.05). This confirms hypothesis 5: A higher average wage level in a city leads to a higher degree of difficulty of acquiring that city’s

Table 3: Multicollinearity statistics

Tolerance GDP per capita 0.37 Elite group influence 0.63 Demand for labour 0.71 Level of education 0.87

FDI 0.71

Average wage level 0.29

Table 3 shows tolerance scores of the independent variables. All scores are higher than the tolerance level of 0.20, indicating that there is no multicollinearity problem in the research and that the independent variables are not highly correlated.

Discussion

Theoretical contributions

The first contribution of this research is that it provides a theoretical conceptual model of factors, both economic and socio-political, that can influence hukou policymaking in Chinese cities. Although many aspects of the hukou system are discussed in the literature, the motivations of city governments in hukou policy decision-making are not discussed. This research provides an overview of the nine possible factors (see figure 1) and how they can influence the degree of difficulty of acquiring hukou. Besides that those factors are placed in the broader theoretical frameworks of the immigration debate, which provides the approaches of analysing them. Moreover, this research provides an important first insight in the importance of several factors on the degree of difficulty of acquiring hukou. A strong positive effect of a city’s average wage level on that city’s degree of difficulty of acquiring hukou was found, accordingly with hypothesis 5. It shows that cities that have a high level of average wage, are setting high requirements for potential new citizens, which allows them to only let highly educated and wealthy people enter the city. On the other hand, if a city’s average wage level is not that high, the degree of difficulty of acquiring hukou will be lower. To increase that city’s average wage level, it does not need to set its requirements as high as a city with a very high average wage level. In that case, it is in the city’s interest to have a lower degree of difficulty of acquiring hukou.

Another result, contrary to hypothesis 1, is that elite group influence has a negative effect on a city’s degree of difficulty of acquiring hukou. It was expected that elite group influence would have a negative influence on a Chinese city’s entry barrier. This means that either the theoretical assumption that elite group influence leads to a higher degree of difficulty of acquiring hukou is wrong, or that there is a problem with the validity of the measurement of the variable elite group influence. Assuming that the theoretical assumption is right, the possibility of a lack of measurement validity arises. In this research, elite group influence is measured by looking at a fiscal transparency index of Chinese cities, deriving from the argument that transparency and elite group influence are strongly negatively correlated, as mentioned by Dittmer (2001: 55). The outcome of the research indicates that more transparency leads to stricter hukou regulations, contrary to the expectation that a more transparent government would be implementing less severe policies. When looking at the transparency index it is striking that the top 4 of the index consists of the four Chinese cities that have a very large economy (Shanghai, Beijing, Guangzhou,

level of transparency. In the literature a positive relationship is seen between nations’ economies and their level of governmental transparency (Kurtzman et al., 2004), and the same mechanism could apply for Chinese city governments. Table 4 shows the correlation statistics of transparency and GDP (China Statistical Yearbook, 2008) for Chinese cities.

Table 4: correlation between transparency and GDP

Variable 1. 2.

1. Transparency 1.00 0.618* 2. GDP 0.618* 1.00 * p<0.01, N = 42

The Pearson correlation between a Chinese city’s transparency and city’s GDP is quite strong, and the relationship is positive (=0.618). This indicates that transparency could partly be acting as GDP in the multiple regression model, which explains the effect that elite group influence has on the degree of difficulty of acquiring hukou. Wealthier cities attract more migrants, and therefore stricter hukou policies are needed in order to turn away these migrants. This shows how a high GDP can result in a higher degree of difficulty of acquiring hukou. It also explains why elite group influence (=reverse of transparency variable) has a significant effect with an unexpected negative direction on the degree of difficulty of acquiring hukou. Although the model already controls for a city’s GDP per capita, a city’s GDP is still playing a role in the model. This is definitely possible, since a city’s GDP and GDP per capita are not as highly correlated (=0.530, see table 5).

Table 5: correlation between GDP and GDP per capita

Variable 1. 2.

1. GDP 1.00 0.530* 2. GDP per capita 0.530* 1.00 * p<0.01, N = 42

Considering the importance of both average wage level and GDP in explaining the degree of difficulty of acquiring hukou of a Chinese city, the expectation that economic factors are most important in Chinese hukou policymaking seems correct. This is in line with other Chinese policymaking, in which the economy plays an important role. Although not all the possible factors that can influence the degree of difficulty of acquiring hukou were implemented in the model, it already explained a lot of variance in the degree of difficulty of acquiring hukou (76%), and especially the effect of average wage level was significantly large (bèta: 0.540). This indicates the importance of the economic factors of the theoretical conceptual model.

Limitations and future research

This research has some limitations. First, the conclusions are based on a small sample (N=42). Moreover, it is a specific sample of cities that are popular destinations for migrants. Further research should replicate this study with a different sample in the same population to see if the same results arise. Also, due to a lack of data, not all the factors as mentioned in the theoretical conceptual model could be included. With more resources (time and money) available, it should be possible to create a dataset that includes all variables.

Also, some issues with regards to measurement validity could have influenced the results of this research. The variables elite group influence and demand for labour were measured by using indicators that measured respectively transparency and unemployment rates. Using more adequate indicators could possibly result in different outcomes. Besides that, the reliability of the data is also contested in the case of China. Due to the fact that it is a single-party state with an authoritarian regime, the Chinese government is able to censor all information. For this research, the China City Statistical Yearbook 2008 was a very important and very useful source. However, the data in this Yearbook could be biased towards positive numbers for the Chinese government. For example, it is known that official unemployment rates in China tend to be terribly wrong (Solinger, 2001). The same manipulation of numbers could be happening with other official statistics of the Chinese government, which could therefore influence the reliability of this research. However, future research is likely to encounter the same problem, since it is very hard to find information that are not checked by the Chinese government.

Besides that, the lack of prior research resulted in this exploratory research, rather than an explanatory research. Although there are limitations as mentioned above, this research provides a starting point for future research, and shows which factors should be considered when discussing entry barriers of Chinese cities. Also, this research shows the importance of the level of average labour wage in explaining the degree of difficulty of acquiring hukou of a Chinese city. More thorough research to understand this mechanism, especially in relation to the qualification process for a hukou itself, is recommended. Also the other eight factors as mentioned in the theoretical model and their relationship to the hukou qualification process are interesting subjects for future research.

Conclusion

Internal migration in China is an enormous phenomenon, with yearly many Chinese citizens migrating to other places. China’s hukou system is controlling this internal migration, by deciding who belongs where, now that many Chinese people want to live in the urban areas. Since the 1990s the decision-making process of the hukou has been localized, and currently city governments can shape their own hukou policies, often favouring the wealthy and highly educated. Zhang & Tao’s (2012) city entry barrier (CEB) index measures the different degrees of difficulty of acquiring hukou of Chinese cities and ranks the cities accordingly. This differentiation poses and interesting question: what are the underlying factors that influence a Chinese city’s degree of difficulty of acquiring hukou? In this research nine factors are recognized that play an important role in shaping Chinese city hukou policy: elite group influence, demand for labour, level of education, FDI, average wage level, public security threat, globalization, nativist sentiments, and costs of government provided benefits. By executing a hierarchical multiple regressions analysis, the effect of five of these factors (elite group influence, demand for labour, level of education, FDI, and average wage level) on the variance in the degree of difficulty of acquiring hukou was measured. This research found that a city’s average wage level plays a very important role, which indicates that that cities that have a high level of average labour wages only want rich and highly educated people in their populations. Also GDP plays an important role, suggesting that economic factors might be crucial in hukou policy decision-making, just as it often is in other Chinese policy areas. More research on the underexposed topic of China’s localized city-level

hukou system, both qualitative and quantitative, is recommended. This system is affecting millions,

if not billions of people, and not much is known about the city governments motivations to uphold these strict regulations, indicating that future research is desirable.

References

Brubaker, R. (1995). Comments on" Modes of Immigration Politics in Liberal Democratic States". International Migration Review, 903-908.

Chan, K. W., Liu, T., & Yang, Y. (1999). Hukou and non-hukou migrations in China: comparisons and contrasts. International Journal of Population Geography, 5(6), 425.

Chan, K. W., & Buckingham, W. (2008). Is China abolishing the hukou system? The China

Quarterly, 195, 582-606.

Chan, K. W. (2010). The household registration system and migrant labour in China: notes on a debate. Population and Development Review, 36(2), 357-364

Cheng, T., & Selden, M. (1994). The origins and social consequences of China's hukou system. The

China Quarterly, 139, 644-668.

Cornelius, W. A., & Rosenblum, M. R. (2005). Immigration and politics.Annu. Rev. Polit. Sci., 8, 99-119.

Dancey, C. P., & Reidy, J. (2007). Statistics without maths for psychology. Pearson Education. Dittmer, L. (2001). The changing shape of elite power politics. The China Journal, 53-67. Fan, C. C. (2002). The elite, the natives, and the outsiders: Migration and labor market segmentation in urban China. Annals of the Association of American Geographers, 92(1), 103-124. Freeman, G. P. (1995). Modes of immigration politics in liberal democratic states. International

migration review, 881-902.

Higham, J. (2002). Strangers in the land: Patterns of American nativism, 1860-1925. Rutgers University Press.

Kurtzman, J., Yago, G., & Phumiwasana, T. (2004). The global costs of opacity —Measuring business and investment risk worldwide. MIT Sloan Management Review, 46(1), 38−44.

Lieberthal, K., & Oksenberg, M. (1988). Policy making in China: Leaders, structures, and processes. Princeton University Press.

Ma, L. J. (2002). Urban transformation in China, 1949-2000: a review and research agenda. Environment and planning A, 34(9), 1545-1570.

Meyers, E. (2000). Theories of international immigration policy-A comparative analysis. International Migration Review, 1245-1282.

O'Rourke, K. H., & Sinnott, R. (2006). The determinants of individual attitudes towards immigration. European Journal of Political Economy, 22(4), 838-861.

PRC Bureau of Statistics (2008) China Statistical Yearbook 2008 (中国城市统计年鉴). Beijing: China Statistics Press.

Ruhs, M., and Anderson, B. (2010). Who needs migrant workers? Labour shortages, immigration, and

School of Public Policy at Tsinghua University (2013) Chinese municipal fiscal transparency report. Beijing.

Smith, N. (2002). New globalism, new urbanism: gentrification as global urban strategy. Antipode, 34(3), 427-450.

Solinger, D. J. (2001). Why we cannot count the “unemployed”. The China Quarterly, 167, 671-688. Wang, F. L. (2004). Reformed migration control and new targeted people: China's hukou system in the 2000s. The China Quarterly, 177, 115-132.

Wang, F. L. (2005). Organizing through Division and Exclusion: China's Hukou System. Stanford University Press.

Wang, L. (2012). Social exclusion and education inequality: towards an integrated analytical framework for the urban–rural divide in China. British Journal of Sociology of Education, 33(3), 409-430.

Wimmer, A., & Glick Schiller, N. (2002). Methodological nationalism and beyond: nation–state building, migration and the social sciences. Global networks, 2(4), 301-334.

Winters, J. V. (2011). Why Are Smart Cities Growing? Who Moves And Who Stays. Journal of

Regional Science, 51(2), 253-270.

World Bank data (2013) Urban population (% of total). Available on:

<http://data.worldbank.org/indicator/SP.URB.TOTL.IN.ZS> [24 April 2014]

Wu, X., & Treiman, D. J. (2004). The household registration system and social stratification in China: 1955–1996. Demography, 41(2), 363-384.

Wu, X., & Treiman, D. J. (2007). Inequality and Equality under Chinese Socialism: The Hukou System and Intergenerational Occupational Mobility1. American Journal of Sociology, 113(2), 415-445. Zhang, L. (2002). Spatiality and urban citizenship in late socialist China. Public Culture, 14(2), 311-334.

Zhang, L., & Tao, L. (2012). Barriers to the acquisition of urban hukou in Chinese cities. Environment and Planning-Part A, 44(12), 2883.

Zhang, Y. (2013) New Report Ranks Chinese Cities by Fiscal Transparency, Raises Questions About Disclosure Practices. Tea Leaf Nation. Available on:

<http://www.tealeafnation.com/2013/07/fiscal-transparency-the-missing-piece-in-debates-on-chinas-budgets/> [25 June 2014]

Zimmermann, K. F. (1996). European migration: push and pull.International Regional Science

Appendices

List of cities in sample (alphabetical order)

CITY PROVINCE

Beijing 北京市 Beijing Municipality Changsha 长沙市 Hunan Province Chengdu 成都市 Sichuan Province Chongqing 重庆市 Chongqing Municipality Dalian 大连市 Liaoning Province Fuzhou 福州市 Fujian Province Guangzhou 广州市 Guangdong Province Guiyang 贵阳市 Guizhou Province Haikou 海口市 Hainan Province Hangzhou 杭州市 Zhejiang Province Harbin 哈尔滨市 Heilongjiang Province Hefei合肥市 Anhui Province

Hohhot 呼和浩特市 (or 呼市)

Inner Mongolia Autonomous Region

Nei Menggu

Jinan 济南市 Shandong Province Jiujiang 九江市 Jiangxi Province Kunming 昆明市 Yunnan Province Lanzhou兰州市 Gansu Province Luoyang 洛阳市 Henan Province Mianyang绵阳市 Sichuan Province Nanchang 南昌市 Jiangxi Province Nanjing 南京市 Jiangsu Province

Nanning 南宁市 Guangxi Autonomous Region Ningbo 宁波市 Zhejiang Province

Qingdao 青岛市 Shandong Province Qinhuangdao秦皇岛市 Hebei Province Shanghai 上海市 Shanghai Municipality Shenyang 沈阳市 Liaoning Province

Shijiazhuang 石家庄市 Hebei Province Suzhou 苏州市 Jiangsu Province Taiyuan太原市 Shanxi Province Tianjin 天津市 Tianjin Municipality

Urumqi 乌鲁木齐市 Xinjiang Autonomous Region Wuhan 武汉市 Hubei Province

Wuhu芜湖市 Anhui Province Xi’an 西安市 Shaanxi Province Xiamen 厦门市 Fujian Province Xiangtan湘潭市 Hunan Province

Xining西宁市 Qinghai Province

Yinchuan 银川市 Ningxia Hui Autonomous Region Zhengzhou 郑州市 Henan Province

Appendix II.

City entry barrier (CEB) index

Appendix III.

Appendix IV.

Appendix V.