Ni

3% WWIIWIW||§1llllHWIIHNIWW

7,
mm

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the
thesis entitled

POPULATION MIGRATION AND LABOR MARKET

SEGMENTATION: EMPIRICAL EVIDENCE FROM XINJIANG,

M.S.

NORTHWEST CHINA

presented by

ANTHONY J. HOWELL

has been accepted towards fulfillment
of the requirements for the

degree in GEOGRAPHY

 

 

/jr4AY}—/&/
9’ Major Professor’s Signature

011- 09 ~67:
/

 

Date

MSU is an Affirmative Action/Equal Opportunity Employer

 

.-.-n-c-------‘a-I---n-

.l-a-O-o-1--I-I-I-l---a-n-l-n-a-.-0-¢-O--O-c-h-o-I-I- _..

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProj/Aco&Prele|RC/Date0ueiindd

 

POPULATION MIGRATION AND LABOR MARKET SEGMENTATION:
EMPIRICAL EVIDENCE FROM XINJIANG, NORTHWEST CHINA

By

Anthony J. Howell

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Master’s of Science

Geography

2009

ABSTRACT

POPULATION MIGRATION AND LABOR MARKET SEGMENTATION:
EMPIRICAL EVIDENCE FROM THE XINJIANG, NORTHWEST CHINA.

By
Anthony J. Howell

Population migration and emerging labor market segmentation are two
phenomena that are dramatically shaping the spatial, economic, and social relationships
in Chinese cities. While considerable bodies of research have been published with
respect to each phenomenon, few studies have focused on the relationship linking
migration and market segmentation (Li, 1997). In fact, there has been no previous
research linking the two phenomena together using Urumqi, the capital city of the
Xinjiang, China as a case study. In order to better understand the linkages between
migration and labor market segmentation and income disparities within labor market
sectors, this thesis utilizes field survey data collected by the author in Urumqi in 2008.

The major findings from this paper show that minority Uighur migrants and non-
mi grants obtain lower wages compared to majority Han migrants and non-migrants.
Furthermore, after employing cluster analysis, Urumqi’s service economy is found to
consist of three labor market segments, which are identified as the primary-independent-
sector, primary-subordinate sector, and the secondary sector. The results from the
discriminant analysis reveal that migrant status and ethnicity have the highest
contribution effect in placing respondents into one of the three labor market segments
listed above. This research concludes that mi grant-status, as well as ethnicity, reinforce

labor market segmentation in Urumqi’s service sector.

Cepyright by
ANTHONY J. HOWELL
2009

DEDICATION

This work is dedicated to my loving parents who have been instrumental in the
completion of this thesis through their unconditional love and support. You are loved

and appreciated.

iv

ACKNOWLEDGEMENTS

First and foremost 1 want to acknowledge my family. In particular, my parents, as
it is through their love and unconditional support that has inspired me to reflect the love
and care that I received growing up to help others in need and to be service-oriented.

I want to thank God for allowing the power and favor of Christ to work in my life
in a very special way, not for my own glory, but so that He may be glorified through me.
To say it simply, without God intervening into my life in February 2004, I would never
have had the opportunity to make it this far in life. I am forever humbled and grateful
that he considered me to be His servant.

I would like to express my deepest gratitude to my advisor, Dr. Guo Chen, as well
as my thesis committee members, Dr. Bruce Pigozzi and Dr. Kyle Evered, for their
thoughtful and constructive comments on this thesis and for all their combined help and
support. I would also like to acknowledge Dr. Jiaguo Qi for his introduction to Dr. Qing
Dong Shi at Xinjiang University, and to all the students at Xinjiang University who aided
my field research while in Urumqi in the Summer of 2008. This thesis could not have
been written without their support, guidance, and friendship. I wish them all the best of
luck in their future endeavors.

The field study for this research was supported by the Geography Department and
the Graduate School at Michigan State University.

Although not directly related to my thesis research, the Asian Studies Center

awarded me two consecutive academic year Foreign Language and Area Studies (FLAS)

fellowships, which enabled me to learn Chinese during my graduate studies. My
language acquisition was most helpful during my field research in Urumqi.

Special thanks to my pastor at Life Christian Church, Pastor David Stephens, and
my whole church family, all of whom have helped me to lay a strong spiritual foundation
over the past five years of my life, which has transcended through all of my other

academic- and work- related roles. I am forever grateful.

vi

 

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... viii
LIST OF FIGURES ........................................................................................................... ix
GLOSSARY ....................................................................................................................... x
1. INTRODUCTION .......................................................................................................... 1
1.1 Research Questions ................................................................................................... 2
1.2 Methods ..................................................................................................................... 2
1.2.1 Research Problem ............................................................................................... 3
1.2.2 Site of Study ....................................................................................................... 5
1.2.3 Field Research .................................................................................................... 9
1.2.4 Major Findings ................................................................................................. 11
1.2.5 Limitation of Research Approach and Methods ............................................... 1 1

1.6 Organization ............................................................................................................ 14

2. LITERATURE REVIEW AND STUDY AREA ......................................................... 16
2.1 Occupational Structural Change and Sectoral Labor Shifts .................................... 16
2.2 Segmented Labor Market Theory ........................................................................... 19

’ 2.3 Minority Placement in the Labor Market ................................................................ 20
2.4 Mi gration’s role in reinforcing labor market segmentation .................................... 23
2.5 Studying a Contested Region: Xinjiang Minorities Considered ............................. 27
2.5.1 Xinjiang under Qing Rule, 1759-1911 ............................................................. 28
2.5.2 Xinjiang During the Chinese Republican Era, 1911-1949 ............................... 30
2.5.3 Xinjiang in the Mao Era, 1949-1978 ................................................................ 31
2.5.4 Xinjiang since the Reforms, 1978-Present ....................................................... 32

2.6 Sources of Minority Discontent in Xinjiang ........................................................... 35
2.6.1 Minority Education Attainment ........................................................................ 35
2.6.2 Occupational Stratification ............................................................................... 36
2.6.3 Migration patterns in Xinjiang: 1949 — present ................................................ 38

3. ANALYTICAL FRAMEWORK AND EMPIRICAL ANALYSES ............................ 40
3.1 Migrant Respondent Characteristics ....................................................................... 41
3.2 Labor Market Earnings Determination ................................................................... 44
3.3 Employment-Level Related Differentiation ............................................................ 46
3.4 Occupation—Related Earnings Differentiation ........................................................ 49
3.5 Detecting Labor Market Segmentation in Urumqi .................................................. 52
3.5.1 Predictions ........................................................................................................ 53
3.5.2 Principal Components Analysis ........................................................................ 56
3.5.3 Cluster Analysis ................................................................................................ 57
3.5.4 Discriminate Analysis ....................................................................................... 61

vii

4. CONCLUDING REMARKS ........................................................................................ 65

4.1 Conclusion ............................................................................................................... 68
5. APPENDICES .............................................................................................................. 71
6. BIBLIOGRAPHY ......................................................................................................... 88

vii

 

LIST OF TABLES

Table 1: Average income earnings by migrant status, nationality, and gender. ............... 45

Table 2: Average income earnings for occupational level by gender, migrants status,

and nationality ........................................................................................ 47
Table 3: Average income earnings for job type by nationality and gender ...................... 50
Table 4: Prediction for variables relationship to labor market segment ........................... 55
Table 5: Cluster means by factor ...................................................................................... 59
Table 6: Variable means for each cluster .......................................................................... 60
Table 7: Discriminate Analysis- Lambda test statistic ..................................................... 62
Table 8: Standardized Canonical Discriminate Functions ................................................ 63
Table 9: Canonical DA Classification Matrix .................................................................. 64
Table 10: Hypotheses and Outcomes ................................................................................ 66
Table 1 1: Employee-Level Survey Questionnaire ............................................................ 72
Table 12: Migration Reasons and Information (%) ............................................. 80
Table 13: Variable Definitions and Sample Means .......................................................... 82
Table 14: Principal Components Analysis Results ........................................................... 84

viii

Figure 1:
Figure 2:
Figure 3:
Figure 4:
Figure 5:
Figure 6:

Figure 7:

LIST OF FIGURES

Composition of Workforce in Urumqi .............................................................. 75
Migrant Origin (Province) ................................................................................. 76
Location Map of Urumqi Urban Districts ......................................................... 77
Location Map of Xinjiang Uighur Autonomous Region .................................. 78
Education Level by Nationality for Xinjiang Province (ratio) .......................... 79
Cluster Tree Produced by Hierarchical Cluster Analysis .................................. 86
Scores Mapped in Factor Component Space by Cluster ................................... 87

ix

GLOSSARY
CASS - Chinese Academy of Social Sciences
CCP - Chinese Communist Party
DA - Discriminate analysis
ETR - East Turkestan Republic
GDP - Gross domestic product
GMD — Guomindang (the Nationalist Party)
PCA - Principal component analysis
PI - Primary-independent
PS - Primary-subordinate
PRC —- People’s Republic of China
SLM - Segmented labor market
SOE - State-owned enterprises
XPS - Xinjiang provincial statistical yearbook

XUAR - Xinjiang Uighur Autonomous Region

1. INTRODUCTION

Since the first economic reforms were enacted in 1978, China has been in the
process of transition from a centrally planned economy towards a market oriented one,
resulting in enhanced labor mobility and the emergence of labor markets (Xu, 2006).
Population migration and emerging labor market segmentation are two phenomena that
are dramatically reshaping the spatial, economic, and social relationships in Chinese
cities. While considerable bodies of research have been published with respect to each
phenomenon, few studies have focused on the relationship linking migration and market
segmentation (Li, 1997). In fact, to this author’s knowledge there has been [no previous
research linking the two phenomena together using Urumqi, the capital city of the
Xinjiang Uighur Autonomous Region (XUAR), China as a case study. In order to better
understand the linkages between migration and labor market segmentation and income
disparities within labor market sectors, the author utilizes a household-level field survey
collected by the author in Urumqi, Xinjiang in 2008.

Labor market segmentation is important to understand since segmented markets
are a major cause of economic inefficiency. For example, excluding a certain group of
people from specific labor markets leads to wasteful utilization of human resources, and
incurs a reduction in the flexibility of the labor market to handle economic change (Xu,
2006). Migration is important to understand in the context of studying Segmented Labor
Market (S LM) theory, since research has shown that rural-urban migration plays an
important role in reinforcing segmentation of the labor markets (Gordon, 1995; Li, 1999;

Fan, 2002; and Fan, 2003).

1.1 Research Questions

The major research questions of this paper are: (1) to what extent do descriptive
statistics provide support for a new urban wage structure emerging in Urumqi’s service
sector; and (2) to what extent does labor market segmentation exist in Urumqi’s service
sector?

A key hypothesis of this paper is that in the case of Urumqi, Han natives and Han
intra- and inter- provincial migrants are all given preference in the labor market over both
Uighur urban Hukou holders and Uighur migrants. This differs from the social and
occupational stratification seen in other Chinese provinces where urban natives are given
job preference over temporary migrants. Under the key hypothesis, Han migrants,
including temporary migrants, are given preference over Uighur urban natives in
Urumqi’s service sector. Therefore my paper recognizes a new urban labor hierarchy
forming in Urumqi based on nationality1 and migration, as follows: (1) Han migrants; (2)
Han natives; (3) Uighur migrants; and finally (4) Uighur natives. In addition, I
hypothesize that migration reinforces labor market segmentation in Xinjiang as migrants

with similar characteristics are channeled into similar job types.
1.2 Methods

To obtain the research objectives outlined above, employee-level survey data

collected in Urumqi are used to: (1) provide descriptive statistics that explain migration

 

' For the purposes of this research, nationality and ethnicity are used interchangeably to distinguish the
majority Han Chinese from minority groups in China, such as the Uighur and Hui. ‘Nationality’ stems
from the Chinese word ‘minzu,’ which is used to denote different nationalities or ethnic groups within
China. See chapter 3 in Gladney (2004) for a more in-depth discourse on China's approach to classify
minority groups.

2

characteristics and reveal labor earnings for all respondents. Using average monthly
wages respondents are disaggregated by migrant status, nationality and gender; and (2)
determine whether labor markets exist in Urumqi. This includes running principal
component analysis (PCA) and cluster analysis. Discriminate analysis (DA) is also
employed in order to assess whether labor market segments identified by the cluster
analysis are statistically significant from each other, as well as to determine which
variables are most influential in placing respondents into the labor market segments.

For the purpose of this research, labor-market segments are defined operationally
as clusters of jobs associated with certain combinations of occupation, industry and
respondent characteristics. The primary-independent (PI) segment is characterized as
free from elaborate rules and procedures and places a premium on creativity and problem
solving. Jobs in the PI segment are high paying, offer prestige, and provide workers high
chance of internal advancement. Employment usually requires college education. The
primary-subordinate (PS) segment consists of relatively high paying jobs and offer fringe
benefits. However, jobs follow a routine, task-oriented schedule and require more direct
supervision compared to the PI sector. The secondary segment consists of low-paying
jobs with few benefits, little opportunity for job advancement, and generally work long
hours in poor or unsafe working conditions. The jobs in this sector are unskilled, highly

unstable, and require little formal education (Fichtenbaum et. al., 1994).

1.2.1 Research Problem

In the case of Xinjiang, researching wage inequality and migration issues is

considerably more complicated, relative to other regions in China, due mainly to the

presence of the ethnic variable, as well as the Chinese Politburo’s sensitivity toward the
region. The presence of the ethnic variable disqualifies research findings elsewhere in
China that do not control for minority populations. Due to the Chinese govemment’s
sensitivity towards Xinjiang, it is nearly impossible to obtain substantial economic data
on the region.

At present, there are only three major sources for minority data, each with its own
drawback: population censuses, ethnic statistical yearbooks, and household surveys
undertaken by the Chinese Academic of Social Sciences (CASS). Population censuses
do not give minority socioeconomic data and the 2000 census does not give data on
literacy or educational attainment by nationality. Ethnic statistical yearbooks are
infrequent and fail to capture the whole poor minority population as sampling is done in
autonomous areas only. This is problematic because in 1998, 55% of minorities in China
lived outside these areas and therefore may not be impacted by preferential policies
implemented by autonomous administrations. Household surveys taken by the Chinese
Academy of Social Sciences (CASS) in 1988 and 1995 use a narrow definition of income
and do not estimate the rent value of housing (Bhalia and Qiu, 2006).

The current CASS data on Xinjiang fail to disaggregate migration data along
nationality types, which is problematic since migration patterns include Han-dominated
inter-provincial migration as well as minority and Han intra-provincial migration. Data
are needed to delineate such migrant groups, based on nationality type, in order to
recognize to what extent push/pull factors, destination desirability, and reception in
destination areas, may be significantly different for different nationality groups and

migrant types.

Although not explicitly explored in this paper, there is an undeniable link that
underscores the importance of this research. The interrelated factors of economic
hardship, discrimination, and inequality, combined with Chinese policy and Xinjiang’s
political structure are major drivers of minority-led separatist movements (Bovingdon,
2004). Thus choosing to study labor market segmentation and income disparities in
Xinjiang is particularly timely and imperative considering unintended consequences of
market reforms are already being seen in other places in China and the world, e. g. market
discrimination, occupation stratification and a rising gap in social and income inequality
(Wu, 2004). If these unintended consequences of market reforms that exist in Xinjiang
are allowed to intensify, they may lead to future surges in protests and acts of terror

against the communist state.

1.2.2 Site of Study

The survey site for this research is Urumqi, the capital city of Xinjiang Province,
located in Northwest China. Xinjiang is China’s largest province, making up 1/6 of the
total land (Milward, 1998). The author of this research was drawn to Xinjiang mainly
due to the uniqueness of the region in terms of its presence of Muslim minorities, ethnic
conflict and contestation of space, geopolitical significance and abundance of natural
resources.

While Xinjiang is inhabited by various ethnic groups that have added to the rich
cultural and historical dynamics of the region, e. g. Kazahks, Krgyz, and Tajiks, the
Uighur people are the primary focus of this research because they are the major victims

of discrimination and human rights violations (Amnesty International, 1999). The

Uighur people are of Turkish descent and claim to have ancestral ties to the indigenous
peoples of the Tarim Basin. Uighurs’ ancestral claims to the land equate to the fact that
the Chinese government is operating as a colonialist state that is occupying the Uighur
motherland. In the eyes of Uighurs, their claim to ancestral ties to the land justifies their
right for self-rule and has in a large way fueled the Uighur resistance movement (Dillon,
2004). There are over eight million Uighur minorities currently living in Xinjiang,
accounting for approximately 45% of the population. As a result of their large population
and history of rebellion, Uighurs have been often been viewed as a legitimate threat to the
Chinese rule in the region (Moneyhon, 2004).

Xinjiang is arguably one of China’s most restive regions and has had a long
history of dissent. Recently, Xinjiang has gone through rapid economic development due
to its substantial oil and gas deposits and geopolitical significance. Despite an abundance
of natural resources and high per capita GDP, Xinjiang is rife with regional disparities,
growing income inequality between Han and ethnic minorities, and simmering ethnic
discontent (Hannum and Xie, 1998; Pannell and Schmidt, 2006). A further analysis of
Xinjiang’s history and present minority issues are further discussed in the proceeding
chapter.

Studying income inequality and migration patterns is deeply important Xinjiang
in order to reduce the level of ethnic discontent in the region. This is due, in part, to the
perception by Uighur that Han migrants are taking the good jobs and making life more
difficult for Uighur, in terms of social mobility, discrimination, and earnings capacity
(Pannell and Schmidt, 2006). To what extent and on what significance level such

inequality exists, what drives and sustains Han migration to Xinjiang, particularly from

6

provinces like Gansu, and how that might affect the labor market outcomes in Xinjiang
are research questions that are crucial to current debates in the policy arena.

In Xinjiang, migration has impacted labor market outcomes in two key ways: (1)
relative to other parts of China, the Chinese government has encouraged massive flows of
Han migrants to the region, particularly during the 19503 and 19608 and during the 10th
5-year plan (2001-2005) (Pannell and Schmidt, 2006). The last state-orchestrated mass
migration of Han “volunteers” occurred in 1964. After that, the majority of in-mi grants
to Xinjiang were either out of work or looking for a better way of life (Toops, 2004); (2)
Han migrants traditionally have been and still are given priority over minorities in
obtaining urban employment, which has led to significant inequality between Han and
non-Han peoples (Su et. al., 2001; Gladney, 2004; Wiemer, 2004; Pannell and Schmidt,
2006).

The outcome of these two inter-related points leads to a new urban stratification
arising in Xinjiang that differs from the Eastern coastal cities. In the East, local urban
Hukou holders, including permanent migrants and natives, are given preference over
temporary migrants, whereas the incorporation of nationality type as a variable may
disrupt that order, with Han migrants given preference over minority urban natives and
migrants. For the purposes of this research, Han migrants are further divided into two
categories: inter-provincial migrants and intra-provincial migrants. It is important to
distinguish between the two groups because the intra—provincial Han migrants tend to be
children of inter-provincial migrants and therefore represent a first generation of

Xinjiang-bom Han Chinese. It is important to determine whether significant differences

in earnings exist between the two different Han migrant groups to serve as a proxy as an
indicator for Han migrant social mobility.

The city of Urumqi was chosen because it has the most mature labor market in
Xinjiang, has the highest percentage of non-agricultural work, and has also experienced a
recent boom in the service economy (Pannell and Schmidt, 2006). More than 70% of
Urumqi’s workforce is non-agricultural labor. This percentage is considerably higher
compared to the province as a whole, where only 30% of the work is non-agriculture.
Approximately 55-65% of Urumqi’s urban workforce is employed in the service sectorz.

In addition, Urumqi’s service sector was selected as a case study because it
employs large populations of both Uighur and Han migrants. Previous research contends
that minorities in Xinjiang are highly underrepresented in technical, administrative and
professional jobs (Hannum and Xie, 1998). However, the sectoral shifi model indicates
that minorities are representative in Xinjiang’s service economy and thus provide
adequate representation of both Uighur and Han (Pannell and Schmidt, 2006).

According to the 2005 Xinjiang Provincial Statistics, 2,081,834 people live in
Urumqi (XPS, 2005). Urumqi’s demographics are comprised mostly of Han and Uighur
nationalities, as well as some Hui, Kazak, Tajik, and other Chinese minority groups.
While Uighur represents the provincial majority in Xinjiang at 45%, Han is the majority
nationality group in Urumqi. There are 1,567,562 Han in Urumqi, compared to only
266,342 Uighur. Of the migrants currently living in Urumqi, there are approximately
774,336 people whose household registration is registered in a province other than

Xinjiang, which indicates the number of official inter-provincial migrants currently

 

2 See Figure 1 in Appendix B to view the breakdown of employment categories (XPS, 2005).
8

residing in Xinjiang; however, the number of actual migrants is much higher when
informal migrants are included in the count. The three provinces that contribute the most
in-coming migrants to Xinjiang are Sichuan, Henan, and Gansu Province.3

Urumqi is divided into 7 districts, 4 of which are included in the survey sample.4
The four districts included in the sample are: Tianshan, Shayibake, Xinshi, and

Shuimoguo. The last two districts are more recently developed as a result of urban

expansion, whereas Tianshan and Shayibake make up the city core.

1.2.3 Field Research

In order to fill in the gaps left by current official data on minorities, primary data
was collected by the author in Urumqi in summer 2008. The data disaggregates
respondents by nationality and migratory status. The survey questionnaire includes
questions related to earnings (monthly income), personal endowment characteristics
(gender, age, marriage, number of children, and education), mobility characteristics
(place of origin, number of places lived, length of stay, household registration status),
occupation characteristics (type of employment, type of firm, and secondary
employment), and social capital characteristics (how did you find employment, where did
your father work, and how many languages do you speak).5

Before carrying out the survey in Urumqi, two Uighur students at the University
of Michigan were interviewed. They provided advice on how to flame questions in a

culturally appropriate manner, which helped reduce respondent bias and subsequently

 

3 See Figure 2 in Appendix C to view migration flows by province (XPS, 2005).
4 See Figure 3 in appendix D for a location map of Urumqi’s urban districts
5 See Table 11 in Appendix A to view the survey instrument.

9

increased the consistency of the data findings. The sample frame consists of all working
individuals employed in Urumqi’s labor market. In order to obtain a large sample size
for both Han and Uighur, a stratified sampling method was adopted. Only urban
employees who were employed in Urumqi’s service economy were sampled.
Respondents were typically employed in visible, commercial and low scale retail and
wholesale jobs, including informal work. Oversampling Urumqi’s retail and wholesale
industry introduces additional bias that hinders the survey results from being
representative across Urumqi’s service sector.

Prior to the start of surveying, approximately 30 survey sites were selected by
looking at an urban district map. Sites were as evenly distributed as possible across each
district. Survey sites included several large, medium, and small commercial markets that
span across 4 different urban districts. To the greatest extent possible, respondents were
selected on a random basis by selecting every 10th store location on-site to survey.

Respondents were marked as either employee, employer, or self-employed;
whoever was available at that time was surveyed. There was approximately a 60%
refusal rate, in which case we would then walk to the next store until we found a willing
participant. In total, a sample size of 595 was obtained. To carry out the survey
questionnaires, two groups of university students from Xinjiang University were
employed to conduct interviews in designated urban commercial areas in Urumqi. Each
member of the two groups received on-site training. They were told how to introduce
themselves and the survey and they were briefed on the research project. Each day, the
author accompanied one of the two groups; every other day the author alternated which

group to shadow. At least one of the members in each group contained one student that

10

can speak Uighur and Mandarin. The surveys were administered in Mandarin, with each

survey taking approximately 20 minutes to complete.

1.2.4 Major Findings

Before carrying out descriptive statistics, respondents were broken into five sub-
groups: Han-natives, Han inter-provincial migrants, Han-intra-provincial migrants,
Uighur natives, and Uighur migrants. The major findings from the descriptive analyses
show that Uighur migrants and non-mi grants obtain lower wages compared to Han
migrants and non-migrants. Women within every nationality-mi grant status sub-group
were all found to earn less money on average than their male counterpart.

From the cluster analysis, three labor market segments were detected, which were
identified as the PI sector, PS sector, and the secondary sector. After running the DA,
the variables that have the most impact on placing respondents into one of the labor
market segments are found to be migrant status and Uighur, showing that migrant status
and nationality play an important role in reinforcing labor market segmentation in

Urumqi’s service sector.

1.2.5 Limitation of Research Approach and Methods

The field research used to inform this research offers a rare look at employee-
level information on urban workers employed in the service sector in Urumqi. While this
research attempts to supplement the lack of data from the highly contested region of
Xinjiang, the survey design is not without considerable drawbacks that need to be
addressed here. First, the survey results are not generalizable beyond Urumqi’s service

sector. This survey offers no insight into other sectors of the economy. In fact,

11

considerable bias is also introduced within the service sector because the majority of
respondents were low-skill employees, 95% of the respondents were employed in retail,
transportation, or service work. Only 5% of respondents were employed in a hi gh-skill
professional or technical job. Respondents employed in a high-skill position include
professors, police officers, lawyers, and doctors.

Another limitation on the survey data stems from issues of respondent
classifications. For example, there was no direct question that asked respondents whether
they were employed in formal or informal work. While most workplaces were obviously
one or the other, there were quite a few respondents that could have been either. We
took each respondent case by case and visually inspected the workplace to help determine
whether a respondent was formally or informally employed. Another classification
problem existed with the college students who were responsible for carrying out the
survey questionnaire. Some of them would mark a respondent as only one employee
type, e. g. self-employed or employer, even if a respondent was in fact both self-employed
and an employer. So the results on this variable are not completely accurate. Also, when
it came to asking sensitive questions, such as how many taxes are paid per month and
monthly earnings, some respondents would not want to answer, however the surveyors
would mark down an answer anyways for fear of handing in an incomplete survey.

When this fact was discovered, we reiterated the importance of leaving unanswered
questions blank; however, this point was not emphasized until well within the second
week of surveying.

Of the 596 respondents surveyed, only 94 were Uighur minorities. Unfortunately,

we did not receive permission to interview Uighur until late in the collecting process and

12

did not have sufficient time to collect more survey data on Uighur. For various reasons,
the majority of Uighur minorities were also only surveyed in two different locations.

Despite these important limiting factors, the biggest problem with the dataset is
that it relies purely on quantitative data. While hard economic data is lacking in the
region, greater attempts should have been made to systematically incorporate open-ended
questions for respondents, particularly for Uighur minorities and Han inter-provincial
migrants. The neglecting of qualitative data was, in part, due to the short amount of time
allotted in the field, only three weeks. However, it is imperative that future research in
this region adopt a mixed-method approach so that qualitative information can be used to
fill in any gaps left by quantitative analyses.

Many of the quantitative methods applied to the dataset were not adequate enough
to provide the results in this paper. For example, although ordinary least squares
regression was used in an attempt to determine wage disparities for nationality and
migratory status, many of the variables collected in the survey data were binary. As a
result, the assumptions of ordinary least squares regression were not satisfied. Therefore,
findings from the linear regression model were not valid. With respect to another
methodological issue, the results from the DA are included in this research, however, are
suspect. This is because the same variables used in the PCA were also used in the DA
and the results are thus tautological. Despite this fact, the results from the DA still lend
evidence that indicates which variables played an important role in placing respondents

into one of the segments found using cluster analysis.

13

1.6 Organization

The organization of this thesis is as follows. Chapter two examines labor market
segmentation theory and migration, as well as establishes a link between the two
phenomena. The beginning of chapter two first offers a general framework of the
institutional changes and sectoral shifts that have transpired in China. This offers insight
into explaining many of the changes that have occurred within Xinjiang, particularly the
surges in Han inter-provincial migration to Xinjiang. Chapter two also offers a brief
analysis of Xinjiang’s recent historical, social, and economic dynamics that continue to
impact the present day challenges facing the region, e. g. regional disparities, ethnic
fragmentation, labor market segmentation, and rising income inequality.

Chapter three lays out the research design and restates the major objectives of the
paper. A specific list of hypotheses is also provided. Chapter three also includes the
methods sections and reports the major findings of the descriptive statistics and
quantitative analyses. In this chapter, descriptive statistics are provided for each of the
five nationality-mi grant status groups created for comparisons. These sub-groups are:
Uighur natives, Uighur migrants, Han natives, Han intra-provincial migrants, and Han
inter-provincial migrants. Descriptive statistics are used as the basis for revealing the
urban wage structure that exists in the service sector of Urumqi. PCA and cluster
analysis are employed in order to detect segments in the labor market and place
respondents into their corresponding segment. DA is later used to determine whether
segments are statistically significant from each other and to identify which variables are

most important in determining which segment respondents are placed.

14

Chapter four sums up the major findings from the empirical analyses, tie in
several policy implications, and offer a conclusion that re-states the relationship between

population migration and labor market segmentation.

15

2. LITERATURE REVIEW AND STUDY AREA

Before using the dataset collected in Urumqi to test for labor market segmentation
in Urumqi’s service sector, the proceeding section will provide a brief synthesis of the
current literature on segmented labor market (SLM) theory, as well as to posit SLM’s
central hypotheses. Although most of the SLM literature has been developed and applied
in a western context, more recently, Chinese scholars have applied and modified SLM
theory to better fit the case of China. Distinct from Western scholars, Chinese research
has placed a greater emphasis on the roles of institutional forces than traditional SLM
theory (Solinger, 1999; Zang, 2002; Fan, 2003), whereas Western scholars have focused
on minority outcomes in the developed nations (Reich et. al., 1973; Hayter and Barnes,
1992; Mclafferty and Preston, 1992; Hiebert, 1999; Bauder, 2001; Constant and Massey,
2005; Hudson, 2007; Gordon, 2008). First, a brief explanation of the occupational
structural changes in the Chinese labor market will be provided in order to better

understand emerging labor market segmentation in Xinjiang.
2.1 Occupational Structural Change and Sectoral Labor Shifts

Since the economic reforms in 1978, and particularly since the 19903 China has
pursued a course of development described as an emerging “socialist market economy”
or “capitalism with Chinese characteristics” (Hsing, 1998). The market transition debate
has arisen to explain how economic reforms have benefited private firms at the expense
of the redistributive economies in China (Zang, 2002). Economic reforms combined with
streams of rural-urban migrants have led to pervasive changes in the urban labor market.

The resulting increase in labor mobility at regional and interregional levels has not

16

produced a wage convergence, rather a multi-tiered labor market (Fan, 2002; Xu et. al.,
2006).

The structural change framework, advocated by Syrquin and Chenery (1989),
examines how a nation’s output economy changes over time (Pannell and Schmidt,
2006). The sectoral labor shift model, as applied to industrialized nations, is outlined by
the following phases: 1) Agricultural productivity increases and the industrial sector
attract surplus rural labor; 2) industrial output reaches its peak, coupled with
technological advancements, which reduces demand for industrial laborers; 3) a tertiary
sector emerges to absorb excess industrial labor supply (Pannell and Schmidt, 2006).

With regards to the newly emerging structures of change within China, ecOnomic
reforms have led to: 1) decreased role of the state in job provision. Specifically, the state
sector no longer assigns jobs to graduating students, rather young adults are now
encouraged to seek employment in the labor markets; 2) the state sector has experienced
substantial decline in terms of labor employment and productive output; and 3) non-state
sectors, such as collective enterprises, private firms, and joint ventures, play a more
predominate role in society (Li, 1997); and 4) returns to political capital have declined,
while returns to human capital have increased (Zang, 2002).

Adverse effects associated with structural changes, induced by neoliberal
economic reforms, have varied in scale and scope. For example, despite official
unemployment rates listed as low 3.1% in 2001 , several prominent Chinese scholars have
estimated unemployment rates to be 7.5% in 1997 or even as high ale.4% in 1998 (Wu,
2004). Fan (2003), gives a detailed account into the nature of gendered labor market

segmentation that has emerged from rural-urban migration. Her research shows that as a

17

result of socio-cultural traditions and social networks, male and female migrants are
channeled into distinct sectors of the labor market. Young women are selected for
factory work, while men are recruited into hard labor, such as construction. Furthermore,
middle-aged women are often discriminated against on the basis of their age due to
employers’ association between youth and productivity (Fan, 2003). There has been little
effort by local governments to prevent labor market discrimination based on age and
gender. If left unmitigated the rural-urban migration flows will continue to reinforce
socio~cultural traditions and increase the gaps between labor market sectors.

In addition, Zang (2002) examines how economic reforms have led to structural
changes in the social stratification hierarchy. Permanent migrants are at the top, followed
by urban non-mi grants, and finally temporary migrants at the bottom (Fan, 2002; Zang,
2002). In addition to rural migrants, other marginal groups have emerged in Chinese
cities: 1) laid-off SOE employees who are unable to find a new job; 2) people who have
never entered into the work-unit system, c. g. disabled and the widowed elderly; and 3)
those who retired before reforms brought about new forms of in-kind welfare guarantees
(Wu, 2004).

To varying extents the aforementioned economic reform impacts can be seen
throughout China. However, Deng Xiaoping’s uneven development strategy has fostered
economic development along the eastern coastal provinces at the expense of China’s
interior provinces. Therefore, research findings from Eastern provinces on emerging
labor markets, as well as market segmentation and income determination, can not
necessarily be generalized in places like Xinjiang. The sectoral shift model for Urumqi

indicates that minorities and inter-provincial migrants are representative in the service

18

economy, but are underrepresented in other sectors of the economy. In this case the
service sector represents industries such as insurance, government, tourism, banking,
retail, education, and social services. As a result, the collection of field data for this
project focused primarily on Urumqi’s service sector in order to ensure that both Uighur
and Han inter-provincial migrants were adequately represented in the sample. Given the
different political and institutional frameworks within Xinjiang, this research project will
help contribute to sub-national regional models with regards to labor market
segmentation, thus enabling local economic geographies to be established and compared

to other regions within China, as well as abroad (Pannell and Schmidt, 2006).
2.2 Segmented Labor Market Theory

Generally speaking, Western scholars have used SLM theory to explain how
minorities, women, and the working class become economically marginalized in society
(Bauder, 2001). Because of the combined result of emerging labor markets and increased
rates of migration, with a strong minority presence, studying labor market segmentation
in Urumqi is both timely and imperative in order to identify potential inequality and
discrimination issues in the labor market. This thesis will determine to what extend
Uighur minorities are discriminated against in wages and occupational concentration.

According to Leontaridi (1998) there are several key tenets central to
segmentation theory: 1) the labor market consists Of a few clearly identifiable segments;
2) mobility barriers exist and prevent individuals from obtaining jobs in other segments;
3) each segment is subject to a different set of employment and wage setting

mechanisms; and 4) neo-classical theory for returns on human capital is not applicable

19

for the lower segment of the labor market (Leontaridi, 1998; Zang, 2002). In his research
in developing countries, Fields (2008) concludes that labor market segmentation exists if
one of both of the following conditions exist: (1 ) jobs for individuals with the same skill
level differ in terms of wages of other characteristics; and (2) access to good jobs is
limited in that people who want better jobs are unable to obtain them (Fields, 2008).

In general, conventional SLM theory focuses on supply-side factors, which
include workers’ personal endowments (education and work experience), constraints
(language), and preferences (working environment) (Xu, 2006). Contrary to neo-
classical labor economists who posit that wage differentials are primarily the result of
differences in acquired human capital, SLM theorists argue that market segmentation is
the result of institutional rules that differ across labor market segments and have thus
replaced the market processes of supply and demand (Leontaridi, 1998). SLM theory
often describes a dual labor market; all jobs fall into one of two separate sectors, e. g.
primary and secondary. The primary, sector represents good jobs in the labor market,
which are marked by high negotiated wages, fiinge benefits, and high employment
security; whereas the secondary sector represents bad jobs, marked by low skill
requirement, low wage rates, and little or no access to career advancement opportunities

(Zang, 2002).
2.3 Minority Placement in the Labor Market

Vulnerable groups, such as minorities and women, become trapped in the lower
segment of the labor market due to mobility barriers, e. g. place of residence, poor work

histories, and discrimination, which reduces inter-sectoral job transfers while

20

occupational stratification increases (Bauder, 2008; Gordon, 2008). Discrimination
theory was developed in order to explain why minorities tend to be concentrated in lower
segments of society. Discrimination theory asks why employers “choose not to employ
some minority workers (such as immigrants and women) in certain jobs where these
workers could have the same value of marginal product as the other workers” (Constant
and Massey, 2005). Essentially discrimination theory posits that even after controlling
for human capital factors, minority workers will earn less money than members of the
majority and are less likely to be hired for employment (Becker, 1957).

In the United States and other industrialized countries, many researchers have
found support for discrimination theory and labor market segmentation theory. Many
experts conclude that minority groups are overrepresented in the secondary sector, face
wage and other forms of discrimination, and are often times unable to achieve inter-
sectoral mobility (Reich et. al., 1973; Hayter and Barnes, 1992; Mclafferty and Preston,
1992; Hiebert, 1999; Bauder, 2001; Constant and Massey, 2005; Hudson, 2007; Gordon,
2008).

In the United States, Hudson (2007) found that Hispanics of both sexes, Black
and Native American men, were all at an increased risk of being employed in the
secondary sector compared to Whites. Furthermore, a substantial proportion of blacks
were found to remain in the secondary sector over the course of their working life
(Hudson, 2007). In Canada’s three largest metropolitan areas, female groups and
minorities, particularly immigrants, were all found to be overrepresented in poorly paid,
vulnerable jobs (Hiebert, 1999). In addition to discrimination practices, McLafferty and

Preston (1992) find that labor market segmentation is deepened due to job mismatching.

21

Job mismatching includes a spatial component where minorities, including women, have
a lack of access to the location of primary sector jobs; therefore they are forced into low
paying, ‘bad jobs’ in the secondary sector. Even in non-industrialized countries,
minorities have been found to be overrepresented in the secondary sector. Telles (1993)
analyses shows that women in Brazil tend to be concentrated in “unprotected work and
paid domestic service. . .Black and mixed-race workers earn less than whites, and they are
overly represented in informal sector employment” (Telles, 1993).

When vulnerable groups such as minorities and women become trapped in the

lower segment of the market, differing wage setting mechanisms between the primary
and secondary sectors are created. In the secondary sector, wages are based on supply
and demand; however, the primary sector’s wages are insulated from supply and demand
forces. The differing wage mechanisms are created, in part, by the segmentation gap
between primary and secondary sectors, where excess labor supply outside the primary
sector is unable to move across occupational strata and the concentration of labor supply
becomes trapped in the lower sector. The effect of this segmentation gap, therefore,
works to keep the primary sector’s labor supply low and also keeps wages artificially low
in the secondary sector (Leontaridi, 1998). Hence, some researchers argue that different
sets of wage earning mechanisms exist for both the primary and secondary sectors.
While human capital variables, such as education and work experience, are useful for
explaining wage variations in the primary sector, they do not explain wage variation in
the secondary sector (Zang, 2002).

There is not so much a debate on whether labor market segmentation exists; rather

the real problem occurs when trying to delineate segments. Due to truncation bias, which

22

 

occurs as a distortion of results due to the omission of values that fall outside a given
range, sectors cannot be a prior defined by wage earnings or occupational types
(Leontaridi, 1998). Other alternative classification systems must be implemented in
order to reduce such biases as much as possible. An interesting statistical procedure has
been the use of hierarchical clustering, which organizes occupational types into separate
groups according to a priori selection of variables. While this procedure is not without
its bias, interesting results have been found by other researchers and hence a similar

process is carried out in chapter three (Drago, 1992, Anderson et. al., 1987).
2.4 Migration's role in reinforcing labor market segmentation

Much of the labor market segmentation research in post-reform China, as well as
in other third world countries, has focused on the rural migrants who were unable to
obtain formal work in the city and consequently were forced into the informal sector,
which is characterized by labor-intensive production, unskilled labor, low productivity
and income, and poor job security (Fan, 2002). According to Fan (2003) labor market
segmentation in China stems from employers’ desire to minimize costs and maximize
efficiency, which promotes strict regulations and compartmentalization of skills and
results in peasant migrants being funneled into a “ narrow selection of gender-segregated
jobs” (Fan, 2003).

Other researchers recognize the impact institutional factors, such as the
household-registration system, have in creating labor market segmentation, via migration.
State policies, urban bureaucracies, and urban rationing regimes influence the likelihood

of migration as well as the type of life migrants will face upon entering the city without

23

an official urban Hukou (Solinger, 1999; Xu, 2006). Social networks (guanxi) are
another factor that influences labor market segmentation in China, which many migrants
depend on for knowledge of urban employment opportunities (Fan, 2003). These social
networks promote labor segmentation as it reinforces the already existing cultural and
institutional forces that facilitate market segmentation in the first place.

Most of the current migration theory literature on China is based on capitalist
market economies in advanced industrialized countries, which fail to fully recognize the
role of institutions and labor market processes, which are of particular importance in the
case of understanding migration in China (Fan, 2002). The Chinese Central Party
Committee (CCP) reform policy focused on urban development much at the expense of
the agricultural sector and in particular, rural peasants. By 1993, cities had been
receiving the largest shares of investment (China Statistical Yearbook, 1995). With over
a quarter century of biased urban policy, China’s urban areas have developed at a much
greater pace than in rural areas. Ultimately a large rural-urban divide has been created
and a sharp increase in economic disparity has continued to rise between urban and rural
areas. Due to the high inequality between rural and urban areas, rural peasants have
become increasingly attracted to migrating to urban areas in search of a better life (U.N
Population Estimates and Projections, 2000). Fan (2003) quotes, “migrant work is the
best, perhaps the only, option to make ends meet, and is widely perceived to be the key to
improving a peasant household’s wellbeing” (Fan, 2003). Unfortunately, the massive
flows of migrants (91% of which are unskilled labor) create unstable conditions in the

cities, and many migrants are unable to find housing or jobs (People’s Daily, 2004).

24

Of particular interest for the scope of this research is the relationship between
labor migrants and labor market segmentation. The enactment of the ‘Reforrn and Open
Policy’ in 1978, gave way for a new philosophy of development in China, which called
on industrialization and urbanization to jump start China’s economic progression. At the
same time, the central government’s monopoly over migration controls slowly began to
unravel and economic dynamics emerging in China began to have a greater influence on
individuals’ choice to migrate (Liang and White, 1996). The dual outcome of the reforms
brought about the emergence of the labor market, while simultaneously enabling massive
flows of rural peasants to migrate to cities. Due to certain policy restrictions still
enforced regarding migration controls, the majority of migrants has entered into cities
without receiving official permission from the government, and thus is considered
informal, or illegal, migrants. The informal, rural-urban group of migrants is collectively
referred to as blindly migrating people (mong liu) and peasants who come to city for
work (nong min gong). Both terms denote an illegal status for rural-urban migrants (UN-
Habitat, 2003).

Due to their illegal status in cities, migrant populations face assimilation barriers
and are subjected to discrimination in terms of employment and public goods.
Limitations that exist in the labor market included migrants being restricted from over 36
white-collar occupations, such as accountants and administrators (Wing-Shing, 2002).
Such institutional barriers hinder many migrants from assimilating into the city.
Consequently migrants are often marginalized from other urban residents, thus creating a
dualistic city comprised of first and second class citizens. Migrants fiom rural areas are

channeled into jobs that pay extremely low wages, subject to harsh working

25

environments that increase risk to personal safety, and are offered little chance of
promotion or wage raises. Examples of jobs include: factories, transportation, service or
illegal sector, or self employed peddler, cobbler, or repairman (UN—Habitat, 2003).
Because of the institutional (i.e. Hukou system) and other social barriers that have
remained intact, political and economic institutions still control urban permanent
residence and entitlements (Fan, 2002). Migrants without Urban Hukou therefore are
more easily funneled into specific sectors of the labor market, which in turn reinforces
labor market segmentation.

In addition to government policy, employers in labor markets play a large role in
channeling migrants into different segments of the market. Western scholars have
emphasized the role of employers in deepening labor market segmentation (Peck, 1996;
Gordon, 2008). Employee applicants’ production potentials are not readily known to the
employer and their work performance depends on a combination of factors such as
incentives and personal perceptions; therefore employers are likely to rely on judgments
regarding applicant’s personal responsibility, commitment, and development potential.
Ultimately, employers must rely on stereotypes to a large degree in deciding whether or
not to hire an applicant. Stereotypes are based on such indicators as age, gender, marital
status, ethnicity, and work history. Historical evidence shows that disadvantaged groups,
such as ethnic minorities, find it difficult to be hired into “good” jobs (Zang, 2002;
Gordon, 2008). According to Peck (1996), the ‘social nature of labor,’ is a social
phenomenon in which employer stereotypes are constructed outside the market and affect
employment relationships to the point of contributing to labor market segmentation

(Peck, 1996).
26

 

Research in places outside of Xinjiang has found that rural-urban migration has
acted to reinforce the segment gap within China’s labor markets (GordOn, 1995; Li, 1999;
Fan, 2002; and Fan, 2003). This fact leads us to question what role migration plays in
deepening the segmented gaps within the labor market in Xinjiang, particularly for
employers who give preference to Han inter-provincial migrants over local and migrant

minorities (Hannum and Xie ,1998).

2.5 Studying a Contested Region: Xinjiang Minorities Considered

Studying labor market segmentation and income disparities in Urumqi is both
timely and imperative considering that negative consequences of emerging labor market
segmentation have resulted in other Chinese provinces, especially rising inequality (Wu,
2004). Because Uighur minorities are likely to be the most discriminated against, labor
market segmentation and subsequent income inequality present a serious risk for fueling
ethnic discontent in the future, which may eventually turn into further Uighur-led
uprisings against the state. Thus, this research presents a first step to provide hard
economic data that reveals the extent to which segmentation and wage disparity exists in
Urumqi’s service sector.

In order to obtain a sense of appreciation for this research, it is important to
provide an introduction to the critical contested region of Xinjiang, briefly highlighting:
(I) the historical context of how Xinjiang was brought under Chinese rule; (2) Xinjiang’s
economic role for China’s national plan of sustainable development; and (3) several key

issues that fuel minority discontent at present.

27

The region that is at present officially called the Xinjiang Uighur Autonomous
Region (XUAR) has an incredibly long, complicated, and controversial history.6
Depending on the source of information, how and when Xinjiang was subjugated under
Chinese rule is disputable. For the purposes of this paper, only Xinjiang’s recent history

starting from the Qing Dynasty (1644-191 1) will be included in the summary.

2.5.1 Xinjiang under Qing Rule, 1644-1911

In 1759, Xinjiang was officially brought under Qing rule. The main motivation
for this Qing expansionism was not necessarily to annex a new province; rather the
region was intended to serve as a buffer zone to protect the Chinese people from nomadic
threats from Central Asia (Perdue 2005). The Qing government implemented a local
system of rule, which relied on a military settlement and local tribal elders called begs to
avoid local Muslim resistance (Milward, 1998). The Qing dynasty initially adopted a
non-interference policy with local Muslim customs and laws. While few policies of
assimilation were forced upon the indigenous minority groups at this particular point in
time, the region experienced sporadic uprisings against the Qing Empire, which shows
the initial local opposition to Chinese rule over the region (Milward and Perdue, 2004).
For example, in 1765 a small rebellion in Ush occurred. However, this rebellion was
quickly quelled by Qing forces. The significance of this uprising stems in the power to
alter future policies, which after the uprising the local government put into effect policies

that segregated Han from the Muslim population. This separation, combined with

 

° See Figure 4 in Appendix E for a location map of Xinjiang.

28

 

favorable economic policies towards Muslim locals led to relative stability and peace
until the 1820’s (Milward 1998; Perdue, 2005).

In the 1820’s, the Kokand Khanate threatened the Qing rule in the region through
the use of militant rebellions. Although the Qing were able to defeat the uprisings, they
ended up acceding to trade demands made by the Kokand forces, which has been called
“China’s First Unequal Treaty” (Milward and Perdue, 2004). The Kokand invasion
inspired Qing government to change policy in order to secure Xinjiang. The new policies
orchestrated large-scale Han-migration into the region. The goals of this policy change
were two-fold: first, Han migrants would elevate Xinjiang to be economically sufficient;
second, Xinjiang would be populated by Qing loyalists (Milward, 1998).

However, Han migration to Xinjiang only exacerbated tension between Muslim
groups and the new Han migrants. In addition to local factors in Xinjiang, the costs
associated with fighting the Opium Wars and the Taiping rebellion made it impossible for
Beijing to financially support Xinjiang. As a result, cumbersome taxes were levied on
the local population in order to raise revenue for the region. The accumulation of ethnic
tension and new tax burdens created strong anti—Qing resentment among Muslim groups,
which provided the necessary support for a successful uprising against Qing rule. In
1862, an uprising led by Yaqub Beg took control of the southwestern region and
established an independent Kashgar state. The region remained independent until
Yaqub’s death in 1877. In 1884, Xinjiang was made into a province and established a
full bureaucratic system of government similar to all other Chinese provinces (Milward

and Turson, 2004).

29

2.5.2 Xinjiang During the Chinese Republican Era, 1911-1949

In 1911, both Han and Muslim groups strongly opposed Qing rule of the region,
and successfully ousted the government leaders from the region. As the last Qing
governor, Yuan Da Hua, was forced out of the region, he appointed Yang Zengxin as the
new governor. From this point until 1944, Xinjiang was ruled by three consecutive
quasi-dictators, Yang Zengxin, J in Shuren, and Sheng Shicai, all of whom more or less
served the national government (Millward and Tursun, 2004). Yang consolidated power
by giving opposition leaders political positions. Yang allowed Muslim minorities to hold
lower positions of power and also refrained from implementing any policies that forced
Muslim assimilation (Forbes, 1996). Yang’s successor, J in Shuren enacted unpopular
minority and economic policies that resulted in a rebellion by Muslim minorities. This
rebellion in 1933 turned into a short lived period of independence, in which the first East
Turkestan Republic (ETR) was established. However, Sheng Shicai quickly regained
control over the lost territory and firmly established his rule with the help and financial
support of the Soviet Republic (Milward and Turson, 2004).

By 1944, the economy had downturned and Shicai had severed his relationship
with the Soviets. Sheng was eventually removed from power and the GMD took over
power. The GMD adopted antagonistic policies towards minorities, which included
removing all Muslim minorities from political power and enacting large taxes. A
rebellion led by the ETR forces resulted near the city of Yining, called the “Lli
Rebellion.” From 1944-1949, the East Turkestan Republic was established and ruled the
north areas of Xinjiang (McMillen, 1979). However, at the end of the Chinese civil war

in 1949, the Chinese army marched into the region and reports say they, “peacefully
30

 

liberated” Xinjiang (Shichor, 2004). Thus in 1949, Xinjiang was annexed into the

People’s Republic of China, where it has remained ever since.

2.5.3 Xinjiang in the Mao Era, 1949-1978

Beijing undertook a slow reform process to socialism in Xinjiang. Minority
collectives were enacted, but the day-to-day lives of minorities were relatively
unchanged. In 1955, Xinjiang was turned into an autonomous region that was based on
the largest ethnic population represented by the “autonomous” region. This led to Kazak
Autonomous Prefectures, Uighur Autonomous Prefectures, and Krygyz Autonomous
Prefectures. This system of autonomous areas successfully prevented different minority
groups from uniting (McMillen, 1979). Minority complaints about the system arose,
particularly due to the fact that true autonomy was not given to the region and that Han
government officials did not learn the local languages. As a result of these “local
nationalist” complaints, numerous minority nationals belonging to the Chinese
Communist Party, who held government positions, were removed from their positions
(Ibid., 1979).

During the 1960’s policies allowed minorities to use their own local languages
and mandated that local customs were to be respected by the Han Chinese. However,
political discussions also ensued on how to remove Islam influence in Xinjiang (Ibid,
1979). During the Cultural Revolution in China, various forces including the Xinjiang
Red Guards and Xinjiang CCP elements committed various abuses against minorities
including defacing minority art and public beatings of suspected local national separatists

(Milward, 1998). Ethnic tensions continued to simmer throughout the 19703.

31

2.5.4 Xinjiang since the Reforms, 1978-Present

When Deng Xiaopeng took over the PRC, Xinjiang along with the rest of China
witnessed an opening up of its borders to the outside world. Trade began to increase,
especially with Pakistan, and new policies enacted in Xinjiang provided greater freedoms
to the Muslim communities (Haider, 2005). During the 1980’s, personal freedoms
continued to expand and the government allowed Uighur language to return to Arabic
text and encouraged the use of minority texts for educational purposes (Benson, 2004).
However, after Tiananmen Square in 1989, many of the liberties granted during the
1980’s were retracted and a clamp was placed on Muslim minorities’ right to express
their culture (Fuller and Lipman, 2004).

Since then, various acts of protest have been carried out against the Chinese rule
in the region, which have for the most part been led by Muslim Uighur people. The
newly gained independence of the Central Asian states in 1991, e. g. Kazakhstan,
Kyrgyzstan, Tajikistan, and Uzbekistan, further fueled Uighurs’ aspirations of an
independent Uighur state, called Uighurstan (Milward, 1998). Over the course of the
19903, thousands of violent actions were carried out against the Chinese government in
an attempt to ward off the Chinese imperialists. In 1999, the chairman of Xinjiang,
Abulahat Abdurixit, claimed that “since the 19903, if you count explosions,
assassinations and other terrorist activities, it comes to a few thousand incidents”
(Becquelin, 2000). In 1996, China carried out a “Strike Hard” campaign which focused
on ‘solving major crimes, capturing fugitives, and crushing gangs,’ and the targets were
terrorism, robbery, murder, organized crime, and the manufacture of and trafficking of

drugs (Dillon, 2004). However, in 1997, the year Deng Xiaoping died, three large bus
32

bombings occurred in Urumqi and one in Beijing. These bombings marked the high
point in the Uighur resistance movement against the Chinese rule over the region
(Moneyhon, 2003). While there had been a sense of cohesion among the Uighur
resistance, there were also sharp divisions within the Uighurs, which prevented any large
scale united front to prevail against the Chinese state. As Moneyhon (2003) writes, “The
Uighurs are divided by religious conflicts, territorial loyalties, linguistic discrepancies,
commoner-elite alienation, and competing political loyalties.”

In addition to the internal divides within the Uighur resistance movement, the
Chinese government responded with massive troop buildup and effectively militarized
the region, which all but crushed the resistance movement. After 9/11, the Chinese
government brought the “War on Terror” to the region and claimed that the East
Turkestan Islamic Movement, a Uighur Resistance group, had links to Al Qaeda (Dillon,
2004). In 2002, the Chinese government deployed 40,000 People’s Liberation Army
troops into the region. In addition to the troop buildup, the “War on Terror” enabled the
Chinese government considerable flexibility in the way they handled their “Uighur
terrorist” problem. Human rights groups, e. g. Amnesty International and Human Rights
Watch, have reported on various human rights violation including torture of political
suspects, lack of fair and transparent political trials, and various executions of political
prisoners (Amnesty International, 1999).

Part of the reason why China has carried out such repressive acts of violence and
policies of assimilation in Xinjiang stem from two major underlying reasons: The
Chinese fear of a domino effect and the presence of natural resources. With regard to the

former, the Chinese government fears that the new found independence of Central Asia

33

will spill over into China’s borders and provide the impetus for substantiated claims and
rights to independence not only in Xinjiang, but also in Tibet, Inner Mongolia, and
Taiwan (Moneyhon, 2004). Besides China’s fear of an ensuing ‘domino effect’, it is
Xinjiang’s abundance of natural resources, in particular large deposits of oil, which have
led the Chinese government to solidify their hold on the region at any cost.

Xinjiang plays a significant geopolitical and economic role for the protection and
supply of natural resources that are extracted from Xinjiang and transported to East
China. In 2000, Xinjiang produced 11.3% of China’s crude oil and 13% of its natural
gas. From 2000-2005, natural gas output doubled and by 2004 oilfields in the Junggar
and Tarim Basins were producing a combined total of 15 million metric tons annually.
Xinjiang has the highest per capita GDP of all non-coastal provinces, and has the only
positive net migration rate in the western interior (Wiemer, 2004; Pannell and Schmidt,
2006).

Since the 1990’s the CCP has paid particular attention to border cities in Xinjiang
and in 1992, Yining, Bole, and Tacheng, along with Urumqi, were given the status of
Open Economic Zones with rights for direct border trade. By promoting cross-border
trade with Central Asia, the CCP has effectively helped to raise per capita GDP in border
cities, thus reducing overall inequality in Xinjiang. In addition, during the 10’h 5-year
plan (2001-2005), the “Go-west” campaign led to improvements in infrastructure, e. g. the
railroad from Korla via Aksu to Kashgar was expanded, as well as building modern
transcontinental highways, railroads, and telecommunications lines in order to revive the

Old Silk road (Pannell and Schmidt, 2006).

34

Uighurs makes up approximately 45% of the population in Xinjiang and is the
majority group in the region. Distinct from most other ethnically homogeneous places in
China, inter-provincial Han migrants that migrate to Xinjiang represent a minority group
who bring with them an onslaught of cultural customs alien and sometimes adverse to
Xinjiang minority culture and customs. During the 1990s a number of Uighur-led
uprisings were organized against the Chinese Communist Party and demands for
independence from Beijing were made (Gladney, 2004). This last point is particularly
important because the current literature describes Uighur as facing the hardest challenges
to obtaining employment in urban and rural areas, compared to other minorities such as
Hui, due to discriminatory hiring practices (Pannell and Schmidt, 2006; Gladney, 2004).
To better understand these social ailments and sources of social instability the remainder

of this section underscores the sources of ethnic fragmentation within Xinjiang.
2.6 Sources of Minority Discontent in Xinjiang

There are many local dynamics in Xinjiang that produce discontent and
fragmentation along social, ethnic, and gender divides. The following is not a
comprehensive analysis of the causes of conflict that have erupted in Xinjiang,
particularly in the 19903. Only a few relative major factors will be discussed in this
section- education, occupation stratification, and migration trends- as they pertain to

influencing labor market outcomes in Urumqi.

2.6.1 Minority Education Attainment
A major source of social unrest is the issue of education for minorities. Many

Western scholars and others, argue that minority marginalization is increasing as a result

35

of failed policies that attempt to improve education graduation rates for minorities.
Despite preferential policies for minority educational advancement, some data shows that
the effects of such policies have had little impact on the larger amount of mass minorities.
Only secondary schools had a steady increase from 1979-1998 (Bhalia and Qiu, 2006).
Furthermore, enrollment rates for minorities in primary school were also stagnant over all
autonomous regions.

These results are seen as a consequence of two factors: first, since the 19803, the
emphasis from non-formal schooling began to shift to formal schooling, which placed
disadvantaged minority students in competition with Han children. The Mandarin
language is an obvious barrier to success for most minority students. Second, market
privatizations of school institutions have created special tuition fees, which exclude many
poor minority students (Bhalia and Qiu, 2006).

When comparing the ratios of education levels using the 2005 Xinjiang Provincial
Statistics yearbook, Uighur have a higher proportion of uneducated compared to Han.7

Han are almost 3.5 times more likely than Uighur to have at least some college.

2.6.2 Occupational Stratification

As a result of Han migration to Xinjiang, there has been growing ethnic tensions
because Han are perceived by minority groups to be obtaining good jobs compared to
local minority groups, especially Uighur (Su et. al. 2001). The occupational stratification
in Xinjiang has in a large part been shaped by context-specific factors, including physical

geography, urban/rural divisions, and ethnic fragmentation. Most of the development in

 

7 See Figure 5 in Appendix F to view education level by nationality for Xinjiang Province

36

 

Xinjiang during the past four decades has occurred in the Northern Jungar basin, which
has created large regional disparities between the more developed northern counties and
the Uighur-dominated southern counties. With regards to ethnic fragmentation,
minorities, particularly Uighur, have typically been excluded from the industrial job
market and the energy service sector. Many of the minorities who have migrated to the
more developed north, e. g. Urumqi, have been concentrated in low paying service jobs,
including petty vendors and informal sector employment. Consequently, increasing
inequalities between Han and Uighur groups have arisen (Pannell and Schmidt, 2006).

When controlling for educational differences between Han and minorities, Han
were found to be over-represented in technical, governmental, and administrative jobs
(Su et. al., 2001). This has led to increased occupational stratification for minorities,
particularly Uighur, who possess few economic opportunities outside of agriculture and
services. As seen from 1990 to 1995, 55.3% of employed minorities work in agriculture
compared to only 21.3% for Han. Only 22.3% of minorities worked in industry
compared to 45.7% for Han (Su et. al., 2001). These statistics show that minorities are
2.5 times more likely than Han to work in agriculture, while Han are more than 2 times
more likely to be employed in industry.

Other empirical research on occupational stratification in Xinjiang corroborates
that minorities are less likely than Han to be employed in non-agricultural sectors.
Hannum and Xie (1998) reveals that in 1982 minorities represented 52.8% of the labor
force compared with 53.8% in 1990. During this time period, minorities were found to
be overrepresented in agricultural work: 69.4% in 1982 and 76.7% in 1990. In

comparison, minorities were underrepresented in all other occupational categories.

37

Although underrepresented, from 1982-1990, the share of minorities engaged in
agricultural work underwent a significant sectoral shift (primary to tertiary) sector in
urban areas, particularly industries of growth include retail, transportation, and the
restaurant business. This sectoral shift was driven due to the vast economic development
and investments into urban infrastructure, combined with minorities desire to improve
their quality of life by moving to the city. As a result, minority representation among
service workers increased from 25.8% in 1982 to 41.2% in 1990 (Hannum and Xie,

1998).

2.6.3 Migration patterns in Xinjiang: 1949 - present

Since the founding of the People’s Republic of China in 1949, the first major
govemment-supported program that facilitated flows of Han migrants to Xinjiang
occurred in the late 19503 and early 19603 following the fallout between China and the
former Soviet Union. This program, in effect, represented a systematic effort by the state
to increase the proportion of ethnic Chinese in the region. Han migration peaked in the
period from 1960-1964, representing 8.65% of the total inter-provincial flows in China,
second only to Heilongjiang. Migration to Xinjiang began to decline in the early 19703
(Liang and White, 1996). From 1985-1995, Xinjiang received the fourth highest net
migration rate among all other provinces, behind Beijing, Shanghai, and Guangdong (Su
et al., 2001). For example, inter-provincial immigrants numbered 345,365 in 1990 and
increased to 1 million by 1995, 95% of which were Han (Bhalia and Qiu, 2006). Also
part of the 10th 5-year plan (2001-2005) was the state facilitation of migration of Han into

the region in order to bring additional skills to the labor force (Pannell and Schmidt,

38

 

2006). Consequently, Han migrants often obtain government, administrative, and
managerial jobs, which have led many Uighur and other minorities to view Han migrants
as taking their “good” jobs (Su et. al., 2001). Hence, Han immigration is a major driver
of minority discontent.

Migration policy enacted from 1949-1985, in part, explains some of the present
day spatial inequality within Xinjiang, because Han migrants were directed by the state
into places that were not occupied by Uighur (Liang and White, 1996, Zang 2002). From
1957-1967, 2 million Han migrants settled into Xinjiang, most of whom had very little
contact with the Uighur. The new Han settlements often emerged as transportation
arteries, which fostered faster economic development, compared to non-Han migrant
settlement areas (Zang, 2002). The impact of segregating migrants from local minority
groups has led to greater inequality between minority concentrated areas and Han
concentrated areas within Xinjiang.

As a result of the large spatial inequality, Xinjiang Uighur peasants’ income
averaged 732 Yuan in 1993, compared to Xinjiang Han peasants who earned an average
income of 2,680 Yuan. This rate of inequality is stated by minority scholars, “to create
an ethnic psychological imbalance that can emerge as an unfavorable factor for unity and
stability” (Sautman, 1999). Furthermore, Sautrnan (1999) also shows that GDP in
minority counties declined from nearly 26% of the provincial total in 1987 to about 18%
in 1994. This trend is also reflective of other areas as well. Xinjiang rural net annual per
capita income in 1994 for minorities was 76.6% of national average and fell to 74% in
1998. Such spatial inequality between Uighur and Han counties have led to discontent

among Uighur, as well as acted as a factor for Uighur peasants to migrate to urban areas.

39

 

3. ANALYTICAL FRAMEWORK AND EMPIRICAL ANALYSES

The analyses in this paper are geared towards explaining population migration
and identifying wage labor discrimination in Urumqi’s service sector by carrying out a
set of descriptive statistics and advanced quantitative methods.

The major research questions of this paper are: (1) To what extent do descriptive
statistics methods provide support for a new urban wage structure emerging in Urumqi’s
service sector, using average monthly wages by migrant status nationality and gender
divides? (2) What are the wage differentials for respondents in different industries,
according to nationality, migrants’ status, and gender? (3) To what extent does labor
market segmentation exist in Urumqi, in terms of nationality, gender, and household
registration status (i.e. migrant or not)? The answer to these questions will help develop
economic policies that will effectively protect vulnerable groups from being further
marginalized in the labor market. This is essential to reducing ethnic conflict and
diminishing the chance of ethnic uprising in the future.

The questions stated above provide the foundation for stating the following three
objectives:

Objective (1): Provide descriptive information for labor migrants and identify the

emerging urban wage structure in Urumqi’s service sector using average monthly

wages, by migrant status, nationality, and gender

Objective (2): Test whether labor market segmentation exists in Urumqi. This
includes running PCA, cluster analysis, and DA;

A full set of corresponding hypotheses for each objective are listed as follows:
Objective (1) Hypotheses: Descriptive Statistics

1.1 Han earn more average monthly wages than Uighur

4O

1.2 Within each nationality group, natives earn more average monthly
wages than migrants

1.3 Between Han and Uighur nationalities, Han migrants earn more
average monthly wages than Uighur natives.

1.4 For each hypothesis listed above, females earn less
average monthly wages than their male counterparts.

Objective (2) Hypotheses: Cluster analysis and DA
2.1 Two segments exist in Urumqi, primary and secondary
2.2 ‘Migrant status type,’ ‘nationality,’ and ‘gender’ play the greatest role

in discriminating respondents into their corresponding sector.

3.1 Migrant Respondent Characteristics

Descriptive analyses of migrant characteristics and patterns are presented in this
section. In total, 412 migrants are included, 325 of which are Han, 75 are Uighur, and 12
are Hui. Of the 412 migrants that were surveyed, 209 are male and 203 are female.8 A
migrant is defined as anyone not born in Urumqi. Unfortunately, the distinction between
permanent migrants with urban Hukou and short-term migrants with no urban Hukou is
not demarcated. However, to serve as a proxy for permanent migration versus short-term
migration, the duration of migration period is broken down into 3 categories, short-term
(less than 1 year), mid-term (1-3 years), and long-term (more than 3 years). Furthermore,
the distinction between contracted migrant with urban Hukou and informal migrant with
no urban Hukou are not distinguished either. However, fi'om Table 2 in appendix G, more
than 75% of Uighur and Han indicated that they obtained their job themselves or with the

help of social networks. For this reason, plus the fact that 95% of respondents were

 

8 See Table 12 in Appendix G.

41

employed in low level jobs, e. g. sales, services, and transportation, we assume that the
migrants collected in this dataset are non-contracted, voluntary migrants.

Next, in order to detect for heterogeneous patterns of migration, the migrants
surveyed were further disaggregated by gender and nationality. This disaggregation is
important in filling the gap left by census data, which treat migration as ethnically
homogenous (Bhalia and Qiu, 2006). From the survey results, a majority of the
respondents, regardless of gender or nationality, indicated that their primary reason for
migrating to Urumqi was to find employment.

Researchers on Chinese migration often refer to social relationships (guanxi) as
playing a pertinent role in providing critical information on job opportunities for peasants
looking to migrate to the city. However, from the survey results, a majority of
respondents from all 4 groups of migrants indicated that they themselves were
responsible for obtaining their current employment. 9 Only 24% of Uighur and 22% of
Han migrants received help from a fiiend of relative to find employment; 25% of male
and 26% of female migrants received help from a friend/family member. Either there
was major survey bias for this question or Guanxi does not play an important role in the
case of Urumqi. With regards to length of migration period, the largest response category
for all 4 groups was 3 years or more, indicating that they have lived in Urumqi since
before July, 2005. The majority of respondents have lived in Urumqi for 3 years or more;
however, the average length of employment was only between 4 and 5 years. Male

migrants were found to be almost 80% more likely than females to be an employer.

 

9 See Table 12 in Appendix G.

42

Furthermore, the 56% of women, which represents the largest response category, are
employees; whereas the largest percentage of males, 48% is found to be in the self-
employed category.

It is surprising that only 69% of Uighur migrants are fluent in Mandarin. This
result suggests that Mandarin language attainment is not necessary for Uighurs to migrate
to the city and to obtain employment; however, the quality of that employment and the
wage earning potential will be analyzed later on in the paper. Due to the absence of
qualitative interview data, a deeper exploration of these results is not possible. Instead,
these results should serve as the basis for future exploratory questions to be asked for
subsequent field research in Urumqi.

The distribution of employment provides some support to suggest that Uighur
migrants experience greater occupational stratification relative to Han migrants in
Urumqi’s service sector. 77% of all Uighur surveyed were either in services or were food
or cultural specialty venders; whereas Han were more evenly distributed among the 8 job
classifications.

When looking at education by nationality and gender the results are mixed. First,
Uighur are more likely than Han to have no primary education, 10.8% for Uighur
compared to only 3.4% for Han. Moreover, Han are almost 4 times more likely to attend
technical school. Contrary to this trend is that Uighur are almost twice as more likely to
attend college than Han migrants. Comparing males to women, the percentage of
education attainment is fairly equal between the two groups, except at the higher
education level. Surprisingly, 17% of female migrants attended technical school,

compared to only 10% of male migrants; 10% of female migrants attended college,

43

compared to only 6% of male migrants. Although female migrants tend to be slightly
more educated on average than male migrants, the average wage outcomes for female
migrants are less than male migrants.

The average remittances sent home on a monthly basis provides a somewhat
interesting result. Uighur send far less remittances relative to earnings back to their
hometown compared to Han migrants. Further qualitative analysis into this result is
needed to determine what cultural or social/economic factors exist that reduces the
likelihood of Uighur migrants to send remittances back to their hometown. Although
some of these findings fall outside the initial scope of this paper, they are nevertheless
interesting information that may be usefirl in guiding future research endeavors in the

region.
3.2 Labor Market Earnings Determination

Survey results show (Table l) Uighur migrants earn almost 40% more than
Uighur natives, while Han natives earn almost 57% higher incomes than Uighur natives.
Han migrants, both intra- and inter-provincial, earn almost 54% and 36% more than
Uighur migrants, respectively. In other Chinese provinces, because of the Hukou system,
urban Hukou holders earn more money on average than temporary or informal migrants
(Xu, 2006). However, in the service sector in Urumqi, this is not the case. Based on the
survey results, Han migrants, both inter- and intra-provincial migrants, earn more money
on average than Uighur natives, despite the fact that Uighur natives have urban Hukou

and are legal employees.

44

Table 1: Average income earnings by migrant status, nationalit , and gender.

”nu-fw-

 

 

 

 

 

 

 

 

 

 

 

 

Total Avg. Wage Male avg. wage Female avg. wage
Wuan) (Yuan) (Yuan)
(N) (N) (N)
Uighur natives Y1076 Y1260 Y883
(N=39) (N=20) (N=l9)
Uighur intra- Y1503 Y1569 Y 1459
provincial (N=55) (N=32) (N=36)
migrant
Han natives Y1685 Y1722 Y1645
(N=93) (N=48) (N=45)
Han intra- Y1656 Y1801 Y1520
provincial (N=58) (N=28) (N=30)
migrant
Han inter- Yl46l Y1715 Y1220
provincial (N=251) (N=121) (N=I 30)

 

Gender reveals a much larger income inequality gap compared to the income
inequality between Uighur and Han. Male migrants earn on average almost 33% more
than female migrants. By disaggregating the average wages by gender, an urban
stratification system is created that places male Han intra-provincial migrants at the top,
earning 1,801 Yuan. The bottom three wage groups are male Uighur natives, followed
by female Han inter-provincial migrants and female Uighur natives at the bottom. As can
be seen, Uighur and female groups earn lower wages than their respective Han and male
counterparts. The only exception is that female Uighur migrants earn more than female
Han inter-provincial migrants.

As seen from the results provided above, reviewing of the average wages shows
that both Han natives and Han migrants earn almost 58% and 41% more than Uighur
natives, respectively. These results support the notion that Urumqi’s urban wage structure

for labor is constructed differently than in the East Coast, where temporary migrants earn

45

less money than natives. In Urumqi, Han natives are at the top, followed by Uighur
migrants and then Han migrants. Uighur natives, however, are the bottom of the
hierarchy, which suggest that nationality may play a deeper role in wage outcomes than

migrant status.
3.3 Employment-Level Related Differentiation

In order to identify the earnings inequality by occupational status, all respondents
have been classified into 4 employment types of occupation (Table 2) that are not entirely
mutually exclusive. The two hi gher-social-strata occupations are SOE employees and
employers who are in the private sector. Some of the respondents, who identified
themselves as both self-employed and employers may fit better into Xu (2006)
classification of household-enterprise owners. Although some respondents in the survey
marked that they were both self-employed and an employer, because of a standardizing
error with the survey, we had to treat these two categories as mutually exclusive. When
running statistical procedures, if the respondent indicated both categories, we treated that
respondent as an employer. Therefore, within the employer category there is an unknown
percentage of respondents who are self-employed and an employer, and then an unknown
percentage who are only an employer but not self-employed. This has possibly led to an
underrepresentation of self-employed respondents and an over representation of
employers in the private sector Respondents that are employers experience management

freedom and flexibility, yet are small and usually employ a maximum of 8 people.

46

 

Table 2: Average income earnings for occupational level by gender, migrant status, and

 

 

 

 

 

 

 

 

 

nationality.
Respondent Employee Self- Employer SOE
group Avg. employed Avg. Earnings Employees
Earnings Avg. Earnings (Yuan) Avg. Earnings
(Yuan) (Yuan) (N) (Yuan)
(N) (N) (N)
All Y1369 Y1464 Y2156 Y1482
respondents (N =295) (N =1 77) (N =70) (N =62)
Male Y1543 Y1555 Y2347 Y1351
(N=135) (N=94) (N=45) 01:32)
Female Y1222 Y1362 Y1814 Y1622
(N=160) (N=83) (N=25) (N=30)
Migrant Yl41 1 Y1444 Y2082 Y1 709
(N=191) (N=143) (N=47) (N=33)
Native Y1292 Y1550 Y2308 Y1225
(N=104) (N=34) (N=23) (N =29)
Han Y1411 Y1457 Y2307 Y1469
(N=206) (N=l47) (N=54) (N=43)
Y1237 Y1532 Y1459 Y1679
Uighur (N=63) (N=21) (N=l 1) (N=12)
Hui Y1410 Y1424 Y2307 Y1242
(N=23) (N=9) (N=5) (N=7)

 

 

 

 

 

 

 

Respondents in the self-employed category are self-employed, but are the sole
workers; they do not employ any other workers except on some occasions the spouse also
worked with them. Self-employed may or may not be considered in the higher social
strata, depending on the type of service/product provided, and whether the respondent is
formal or informal. The respondents were not asked directly whether they were formal or
informal, but if we were unable to see their certificate of employment, we marked the
respondent as informal. Respondents in the employee category are those who are neither
employers nor self-employed. Employees make up the lower-social-strata occupations

and include retail, drivers, clerks, and bankers.

47

 

There is considerable inequality in average earnings among residents in the 4 t
different categories. As expected employers in the private sector receive the highest
income of 2,1 56 Yuan per month, which is almost 58% higher than employees cam on
average. After disaggregating the self-employment category into formal and informal
employment, formal self-employed respondents earn the next highest average income,
1,871 Yuan per month. It is clear that informal work does not provide more income than
any of the formal employment types. This finding supports the notion that informal work
is not a preferred choice, rather a forced one for individuals who are unable to find work
through the proper channels. Personal interviews that were conducted with several
willing respondents who are employed in the informal sector also indicated that they
could not find any formal work, so they were forced to find informal ways to earn money.
Approximately 90% of the informal respondents were Han inter-provincial migrants,
therefore a complete comparison of informal workers by nationality and migrant status is
not possible.

When the respondents are disaggregated by gender, migrant status, and
nationality, the earnings differentiation within and between occupation status types is
even more revealing. Except for the SOE category, females earn considerably less than
males. The largest income gap between male and female occurs in the employer category,
where males earn almost 30% more than females. Natives earn almost 8% and 11% more
than migrants in the self-employed and employer groups, respectively. This result
highlights the importance of disaggregating wages by nationality type, because this result
alone would indicate that natives do in fact earn more than migrants, similar to other

provinces in China. Controlling for nationality therefore, is an important factor when

48

considering wage determination, especially for policy making. However, many national
data sets, i.e. population census data, do not disaggregate average wages migration status
by nationality (Bhalia and Qiu, 2006).

The role that state-owned enterprises play in wage determination is also worth
mentioning here. While females and Uighur receive lower wages in almost all other job
groups, the state sector actually pays them more. In the SOE’s, women earn almost 21%
more than males, and Uighur earn almost 15% more than Han. While the private
employment in the labor market is the focus of this paper, it is interesting to note the
important role that SOEs have in paying high wages to traditionally underrepresented
groups such as females and Uighur.

In sum, when disaggregated according to gender and nationality, the employer
category represents the largest offender in promoting wage disparities. Stronger
monitoring of this group and implementation of fair wage laws should be enacted to
protect women and minority groups. Surprisingly, the SOEs provide better protection for
minorities and women compared to the developing labor markets in the service sector in

Urumqi.
3.4 Occupation-Related Earnings Differentiation

All respondents have been classified (Table 3) into one of eight job types, according to
industry type. Because respondents employed in the retail industry makes up 68% of the
dataset, the retail industry is further broken into five sub-categories based on the product

that is sold. Respondents in Retail-Food/beverages sell snack- items and beverages.

49

Respondents in Retail-Elect/apps. sell or repair electronics and/or home appliances.
Respondents in Retail-Clothes/acc. sell clothes and accessories.

Respondents in Retail-firrn/home sell home firrniture. Respondents in Retail-other sell
other items such as cultural specialty foods or goods, such as knives or carpets.
Respondents in Transportation are either taxi or truck drivers. Respondents in Service are
employed in service jobs, such as security guards, street cleaner, construction workers,
restaurant server, barber, or receptionists. Respondents in Professional/technical are
employed in white-collar jobs, such as professors, doctors, engineers, etc.

Respondents are broken up separately by nationality and gender (Table 3). Uighur
clearly earn less money in every job type, except in the transportation and
professional/technical categories. While these results are informative, they are not
representative throughout the job types due to the small sample size for some of the
indicators. For example, only one Uighur respondent worked in Retail-fum/home and
only three worked in retail-elect/apps. However, for the other job types that have a
sample size of at least 10 Uighur surveyed, Uighur earn substantially less than their Han
counterpart.

Women earn less money than males in all 8 job types. Retail-home/furniture
shows the largest wage gap, where males earn 88% more than females. Prof/Tech. show
the smallest wage income gap between male and female, where males earn .08% more
than females. Although Prof/Tech. have a small number of sampled respondents (n= 24),
the average wage inequality appears to be more exacerbated in jobs that do not require

high skill or educational level attainment.

50

Table 3: Average income earnin, s for job type by nationality and gender

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Respondent Retail — Retail — Retail- Retail-
group Food/bev. elect/apps clothes/ace furn/home
Avg. Earnings Avg. Earnings Avg. Earnings Avg. Earnings
(Yuan) (Yuan) (Yuan) (Yuan)
(N) (N) (N) (N)
All Yl317 Y1720 Y1342 Y2144
respondents (N=1 1 1) (N =52) (N=58) (N=34)
Han Y1378 Yl695 Y1481 Y2270
(N=83) (N=42) (N=46) (N=30)
Uighur Y1007 Y1200 Y781 Y1000
(N=20) (N=3) (N=11) (N=1)
Hui Y1639 Y2091 NA Y1267
(N=7) (N=7) (N=3)
Male Y1421 Y1847 Y1468 Y3079
(N=6l) (N=37) (N=14) (N=12)
Female Y1,191 Y1406 Yl302 Y1635
(N=50) (N=15) (N =44) (N=22)
Table 3 (continued)
Respondent Retail —Other Transportation Service Prof/technical
group Avg. Earnings Avg. Earnings Avg. Earnings Avg. Earnings
(Yuan) (Yuan) (Yuan) (Yuan)
All Y1346 Y2222 Y1326 Y2233
respondents (N=l 02) (N =26) (N=1 35) (N =24)
Han Y1423 Y1915 Y1415 Yl762
(N=81) (N=20) (N=89) (N=16)
Uighur Y1008 Y3795 Y] 121 Y3514
(N =1 7) (N=4) (N=32) (N =7)
Y1225 Y2150 Y1204 Y800
Hui (N=4) (N=2) (N=l3) (N=l)
Male Y1421 Y1834 Y1353 Y1938
(N=67) (N=2) (N=40) (N=3)
Y1223 Y1890 Yl,207 Y2154
Female (N=71) (N=2) (N=51) (N=13)

 

 

 

 

 

 

51

 

 

3.5 Detecting Labor Market Segmentation in Urumqi

In order to test whether labor market segmentation exists in Urumqi, a set of
quantitative techniques must be employed in order to expose labor market segments
captured in the dataset, group respondents into their appropriate sector, and then test
whether sectors are statistically different from each other. Labor market segmentation
theory is used to determine which variables should be included in the PCA”).

PCA is used to delineate trends within the dataset by creating statistically independent
principal components (O’Sullivan and Unwin, 2003).

PCA is first employed in order to identify discrete labor market segments (Biles
and Pigozzi, 2000). In addition, PCA is also used in order to reduce the data and
essentially diminish the problem of multi-collinearity, which was found to exist among
many of the variables collected in the survey (Rummel, 1967). Running PCA reduces
multi-collinearity by orthoganolizing the dataset. The multiple score vectors that are
produced were saved as a new variable and were inserted into the hierarchical cluster
analysis to identify the number of discrete groupings in the dataset. K-means testing is
next used to classify respondents into their corresponding cluster (Dragon, 1992,
Anderson et. al., 1987). DA is used to determine whether the sectors are statistically
different from each other as a whole, using Wilks’ lambda F-test. In addition, DA will
assess the relative importance of the independent variables’ role in dictating which labor

market sector a respondent will be placed (Klecka, 1980; Biles and Pigozzi, 2000).

 

to See Table 13 in Appendix H for a listing of key variables, along with their definitions and sample means.

52

3.5.1 Predictions

To obtain an ideal demarcation among segments in society, an extensive list of
key segment characteristics would focus on various aspects of workers, jobs and firms
(Drago, 1992). However, such an extensive dataset does not exist, especially in the case
of Urumqi. Therefore, select variables are pertaining to job and employment type and
respondent characteristics are used in the PCA. Predictions are provided (Table 4)
according to which labor market segment a variable is most likely to correspond with. A
“+” denotes the characteristic should be most highly represented within the particular
segment, a “-” indicates a negative predicted relationship, and a “0” indicates either an
uncertain relationship or a projected intermediate ranking (Drago, 1992). Employees are
expected to be most pronounced in secondary sector and least pronounced in the PI
sector; a ranking that is reversed for both self-employed and employer. ‘Retail-food/bev.’
is expected to be lower in the PI sector and highest in the secondary sector, same as
‘retail-clothes,’ ‘retail-other,’ and ‘services.’ This is because respondents in these jobs
sell common goods that are usually inexpensive, indicative of secondary sector jobs.
Professional/technical jobs are expected to be highest in P1 jobs and lowest in secondary
jobs.
The remaining job types are difficult to predict, because some respondents may
sell/repair hi gh-priced or low-priced items, and therefore some may have good jobs,
representative of the primary sector, and others may have bad jobs, representative of the
secondary sector. F orrnal jobs, job tenure, education, Mandarin fluent, Han, and inter-
provincial migrants are all expected to be highest in either of the primary sectors and

lowest in the secondary sector. Gender, native, and infra-provincial migrants are expected
53

to be highest in secondary sector and lowest in the primary-independent. Wage
impoverishment, which represents respondents who earn at least 50% less than the
average wage of all respondents, is expected to be highest in the secondary sector and
lowest in the PI sector. Wages are expected to be highest in the PI sector and lowest in

the secondary sector.

54

Table 4: Hypothesized variables’ relationship with the labor market segments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Labor Market Segment
Primary Primary Secondary
Independent Subordinate
Employee - 0 +
Self-Employed + O -
Employer + 0 _
Retail - 0 +
Transportation 0 0 0
Services - 0 +
Prof/Tech + + -
Formal + + -
Job Tenure + + -
Age 0 0 0
Gender - + +
Education + + -
Mandarin Fluent + + -
Han + + -
Uighur + + -
Native - + +
Intra-Prov. - + +
Inter-Prov. + + -

 

 

 

 

 

 

55

3.5.2 Principal Components Analysis

PCA is used to delineate trends within the dataset by creating statistically
independent principal components (O’Sullivan and Unwin, 2003). PCA produces a score
vector for each principal component extracted. The score vectors are then used in the
cluster analysis in order to classify respondents into the group they most readily identify
with.

Twenty-four variables, including industry, occupational, and human capital
variables, are included in the PCA.ll Varimax rotation is used in order to maximize the
sum of the variances of the squared coefficients within each eigenvector, and the rotated
axes remain orthogonal (Rummel, 1967). In effect, this will enable a more
straightforward interpretation of the component loadings. A loading of .7 or greater is the
threshold used to indicate whether a variable is highly correlated to its corresponding
factor.

To begin with, five components were chosen and the corresponding rotated
component loadings were analyzed. This process was repeated until there were only two
components produced by the PCA. When extracting 5, 4, and then 3 components, only
the first two components explained more than 5% of the variation in the dataset. After an
analysis of the component loadings, running PCA with only two components offered the
best outcomes in terms of high loadings and interpretable factors. Loadings on
employee, self-employed, and formal all have absolute value of .7 or higher. As such,

Factor 1 is labeled employment type. The interpretation of this factor is that employees

 

" See Table 14 in Appendix J to view each variable’s relationship to its respective component factor.

56

are at the negative end of the factor, while non-employees (i.e. self-employed and
employer) are found at the positive end of the factor. For factor 2, only Han and Uighur
have loadings of absolute value .7 or higher. As such, factor 2 is labeled nationality. Han
are found to be at the positive end of the factor, while Uighur are found at the negative
end. Factor 1 explains 13.5% of the total variance of the dataset, while Factor 2 explains
12.0% of the total variance of the dataset.

Overall, approximately 25% of the total variance is explained by the two factors,
which is rather low. With 75% unexplained variation, it is possible that the dataset may
not suffer from extreme problems of multi-collinearity, as originally thought.
Nevertheless, the score vectors are still used in the hierarchical clustering analysis to
ensure that the problem of multi-collinearity is diminished to the greatest extent possible.
The following section uses the scores produced by the PCA to group respondents into

cluster groups.

3.5.3 Cluster Analysis

To determine how jobs are grouped in the labor market, cluster analysis is carried
out (Anderson, et. al., 1987). Using the scores created from PCA in the last section,
hierarchical cluster analysis, using Ward’s minimum variance, is employed to determine
how many distinct cluster groups are within the dataset. Hierarchical clustering is used to
classify a set of observations that are similar to each other, while relatively different from
other observation sets. The hierarchical clustering process starts with linking similar
respondents into small clusters and then each small cluster is placed into larger groupings

higher up in the hierarchy (O’Sullivan and Unwin, 2003). Using the cluster tree

57

produced from the cluster analysis, three groups are determined to be adequate to classify
respondents. ‘2 Furthermore, labor market segmentation theory affirrns that three
segments exist in the labor market: PI, PS , and secondary. Therefore, for the rest of the
analyses, this work will adopt this classification of labor market segments.

Using the scores from the PCA output, K-means cluster analysis is next used to
classify respondents into 3 groups. When the clusters are mapped in component space,
there is an obvious delineation by nationality and employment type. '3 Each factor’s
cluster means (Table 5) are used to label clusters according to which sector in the labor
market they most closely represent based on other labor market segmentation theory
research. Cluster 1 contains 167 cases. Cluster 1’s mean for factor 1 is -1.l9, and mean
for factor 2 is .522. Cluster 1, therefore, represents Han respondents who are not
employees. Cluster 2 has a mean of .68 for factor 1 and a mean of .53 for factor 2. Cluster
2, therefore, represents Han respondents who are employees. Cluster 3 has a mean of .133
for factor 1 and a mean of—1.52 for factor 2. Cluster 3, therefore, represents Uighur

respondents who are employees.

 

'3 See Figure 6 in Appendix S to view the cluster tree produced by the cluster analysis.
'3 See Figure 7 in Appendix T to view groupings of the three clusters.

58

Table 5: Cluster means by factor

 

 

 

Cluster Cluster Cluster
1 2 3
(Mean) (Mean) (Mean)
Factor I: -I.19 .68 .133
‘employment
type’
Factor 2: .522 .53 -1.52
‘Nationality’

 

 

 

 

 

 

Based on careful inspection of each cluster’s mean values, Cluster 1 appears to
most closely represent the PI sector; cluster 2 appears to most closely represent the PS
sector; cluster 3 appears to most closely represent the secondary sector (Table 6).
However, when comparing the means for variables within each group, cluster 2 has the
highest wages compared to the other clusters and has the lowest proportion of
respondents who are considered impoverished.l4 Furthermore, cluster 2 has the highest
average education level relative to the other cluster groups. These results are indicative

of the PI sector. Cluster 3 on the other hand, has the highest proportion of respondents

 

'4 A respondent is in considered to be in poverty if their earnings are less than 50% of the average income
of all respondents in the dataset.

59

Table 6: Variable means for each cluster

 

 

 

 

 

 

 

 

 

 

 

 

 

Cluster 1 Cluster 2 Cluster 3
Formal .70 .82 .82
Job Tenure .13 .16 .15
Gender .51 .48 .60
Mandarin .94 .98 .78
Fluent
Education 2.4 2.9 2.6
Native .1 l .33 .34
Intra-Prov. .07 .20 .59
Inter-Prov. .83 .47 .06
Avg. Wage 1,477 1,514 1,489
Nationality 1.1 1.1 2.1
W- .20 .16 .27
Impoverish*
Prof/Tech .03 .05 .06

 

 

 

 

 

 

who are impoverished, the highest concentration of females and Uighur respondents, and
the lowest average education level attainment. These results are indicative of the
secondary sector. For most of the variable means, Cluster 1 falls in between cluster 2 and

cluster 3 and hence most closely represents the PS sector.

60

3.5.4 Discriminate Analysis

Canonical DA is used in order to test whether the labor market segments created
using PCA and clustering analysis are significantly different from one another. Canonical
signifies that the dependent variable is a categorical variable, and therefore regular DA
cannot be used. Moreover, canonical DA is used in order to see how
well the model accurately predicts the observed categories of the dependent variable.
Lastly, the canonical DA model is used to assess the relative importance of the
independent variables in classifying the dependent variable, labor market sectors (Klecka,
1980). The dependent variable consists of three groups, PI, PS, and secondary, therefore
two orthogonal functions will be listed in the output. The first firnction will provide the
most overall discrimination among the three groups and the second function provides the
second most overall discrimination among groups. Because the firnctions are
independent, the functions’ contribution to the discrimination among groups does not
overlap (Starsoft, 2008).

The results show (Table 7) a large Wilks’ lambda F-statistic at 10.25, which
indicates that the DA model does a good job at discriminating among the three labor
market segnent groups. The canonical correlation on the first discriminate function is
.658, meaning that 65.8% of the variation in the dependent variable is discriminated by
the set of independents. The canonical correlation for the second discriminate function
explains 34.9% of the remaining unexplained variation in the dependent variable. This
overall high percentage shows that much of the variance in the discriminate scores can be

attributed to group differences.

61

Table 7: Discriminate Analysis- Lambda test statistic

 

 

 

 

 

Value Approx. p-value
F-Value
Wilks’ Lambda .489 10.24 0.000
Canonical Function 1 Function 2
Correlations .658 .349

 

 

 

 

Standardized discriminate function coefficients are listed (Table 8) in order to
interpret the unique contribution of each variable to the discriminate function. In the first
discriminate function, native and inter-provincial migrant have the highest coefficients.
This result indicates whether a person is born in Urumqi and whether a person has
migrated from outside of Xinjiang province as having the highest contribution effect on
discriminating respondents into one of the three segments. Surprisingly, education,
gender, nor nationality plays a large role in the first discriminating function. This would
suggest that a person’s migrant status type is a more important discriminator in
determining which sector an individual is likely to be placed. In the second discriminate
function, Uighur has the highest coefficient, followed by native and interprovincial-
migrant. This second discriminate function indicates that nationality does play a strong
discriminating role in dictating in which sector an individual will find employment. The
results from the second function also further reinforce the importance of migrant status in

discriminating which sector of the economy respondents enter.

62

Table 8: Standardized Canonical Discriminate Functions

 

1

 

Employee

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.511 .474

Serf-Employed -_ 142 -.126
Employer _783 .252
Retail. -.286 -.127
Transportation _205 .081
Services __3 53 -.024
Professional/Technical -. 1 76 -. 126
Formal .049 -.063
Job Tenure “047 _107
Age -.009 -.057
Gender .027 -,O83
Education '1 77 -,057
Mandarin Fluent .084 -.189
Han .106 .400
Uighur -,277 .975
Native _1_513 _709
Migrant-

(Intra-Prov.) -0'603 .349
Migrant —

—1.441 .708

(Inter-Prov.)

 

 

 

63

 

The classification matrix (Table 9) is used to assess the performance of the DA.

The rows indicate the observed categories of the dependent variable and the columns
indicate the predicted categories of the dependents. The classification matrix shows that
the DA model does the best job at predicting occurrences in the secondary sector, a
percentage of 81%. The PI sector shows a 66% correct prediction and the PS sector
shows the least correct prediction of only 57%. It is not surprising to see PI and PS
sectors have lower prediction scores relative to the secondary sector, as many jobs in
these two sectors tend to share similar job characteristics, which make job classification

difficult (Fichtenbaum et. al., 1994; Hudson, 2007).

Table 9: Canonical DA Classification Matrix

 

 

 

 

 

Primary- Primary-

independent Subordinate Secondary % Correct

(Cluster 1) (Cluster 2) (Cluster 3)
Primary-
independent 1 07 47 9 66
Primary-
Subordinate 77 I47 34 57
Secondary 2 26 1 16 81
Tom 186 220 159 65%

 

 

 

 

 

 

 

The findings from the DA provide support that labor migration reinforces labor
market segmentation. Furthermore, the high loading on Uighur in the second
discriminate function suggests the possibility of mobility barriers due to hiring practice
discrimination, or the existence of guanxi or ethnicity-based job selection, as nationality

type plays a larger role in placing respondents than any of the human capital variables

64

captured in the dataset, which include job tenure, education, and level of Mandarin
fluency. Additional job mobility data is necessary to acquire in order to sufficiently
conclude whether minorities voluntarily choose to stay within a certain sector, or are

forced to stay in this sector.

65

4. CONCLUDING REMARKS
Using the findings described above, the following section will highlight which of the
hypotheses stated in chapter three are supported and which of the hypotheses are rejected

(Table 10).

Table 10: Hypotheses and Outcomes

 

List of Hypotheses Outcome*

 

1.1 Han earn more average monthly wages than Uighur +

1.2 Within each nationality group, migrants earn more average
monthly wages than natives *

1.3 Between Han and Uighur nationalities, Han migrants earn
more average monthly wages than Uighur natives. +

1.4 For each hypothesis listed above, females earn less

 

average monthly wages than their male counterparts. +
2.1 Two segments exist in Urumqi, primary and secondary *
2.2 migrant status type, nationality, and gender play the *

greatest role in discriminating respondents into their
corresponding sector.

 

 

 

 

Note: The ‘+’ symbol indicates that the hypothesis under question is supported
by the survey results; The ‘-’ symbol indicates the opposite; ""’ indicates the
hypothesis is partially supported.
From the descriptive results, hypothesis 1.1 is confirmed. Han respondents earn
more than their Uighur counterpart. Hypothesis 1.2 is partially correct, in that Uighur

migrants earn more than Uighur natives; however, Han migrants are found to earn less

than Han natives. From hypothesis 1.3, Han migrants earn more than Uighur natives.

66

Hypothesis 1.4 is also confirmed, as females within every nationality-migrant status sub-
group earned fewer wages than their male counterpart.

Based on the descriptive statistics, the urban wage structure is as follows: Han
natives, Han intra-provincial migrants, then Uighur infra-provincial migrants, followed
by Han inter-provincial migrants and finally Uighur natives at the bottom. This finding
supports the Uighur notion that Han migrants are taking away good jobs from the local
Uighur population (Pannell and Schmidt, 2006). These results support that Uighur
natives, who have urban Hukou, earn less money than Han migrant without urban Hukou.
This fact differs from other Chinese provinces where natives earn more money on
average than informal migrants. Specific regional policy measures, therefore, need to be
implemented to protect Uighur Hukou holders against increasing wage inequality,
particularly as the labor markets continues to develop in Urumqi. If protective measures
are not enacted, hiring practice discrimination and wage inequality will likely increase,
which may play a role in deepening minority discontent and possibly even cause protests
and uprisings in the future.

The hypotheses for objective 2 were both found to be not completely supported
by the results revealed from the cluster analysis in chapter three. Instead of two
segments, hierarchical clustering produced three sectors. This result is not unusual
considering dual-labor theorists have also produced results that include three sectors: two
primary sectors and a secondary sector (Drago, 1992). The results from the DA are
crucial in order to understand what variables are most responsible for reinforcing labor
market segmentation. The DA provides support for two of the three hypothesized

variables (migrant status type and nationality) that are most important in placing

67

respondents into one of the three sectors. However, the third hypothesized variable,
gender, was not found to be among the most discriminating variables. This result shows
that human capital factors do not play as strong a role in deepening labor market
segmentation as migrant status and nationality. Furthermore, this finding provides
support for the argument that migration reinforces labor market segmentation. The
results from the DA suggest greater attention should be placed on Uighur natives and Han
inter-provincial migrants, and particularly women within each group, because they are
the ones that appear to be channeled into the secondary sector and are experiencing the

greatest income inequality.
4.1 Conclusion

China’s economic reform has led to many changes in China’s urban labor market.
This paper focuses on two major phenomena occurring in Xinjiang: labor migration and
the creation of urban labor markets. This paper posits that there is a strong link between
migration and urban labor markets and specifically provides support that migration
reinforces labor market segmentation. In addition, Uighur natives and women were
found to be most likely to earn the least amount of money and be placed into a secondary
sector job.

Xinjiang was chosen as a case study for several distinguishing factors: (1)
Xinjiang borders 8 other countries, Mongolia, Russia, Kazakhstan, Kyrgyzstan,
Tajikistan, Afghanistan, Pakistan, and India, and thus plays a critical geopolitical role in
national security and foreign trade; (2) in 2000, Xinjiang produced 11.3 % of China’s

crude oil and 13% of its natural gas (Pannell and Schmidt, 2006). Hence, Xinjiang’s

68

natural resource base is of crucial importance for China’s overall economic growth and
development; (3) Xinjiang has been the destination for mass flows of government
sponsored Han migrants originating from Eastern provinces. From 1984-1994, Xinjiang
received the fourth highest migrant reception rate, behind Beijing, Shanghai, and
Guangdong (Liang and White, 1996); (4) the demographic population in Xinjiang is
much more diverse compared to the Eastern provinces. 45% of the total population
consists of the Uighur ethnic group, followed by Han Chinese who comprises a little less
than 45% of the total population. The remaining 10-12% of the population consists of
various ethnic groups, including Hui, Tajik, Kyrgyz, and Kazak (Toops, 2004); and lastly
(5) during the 19903, a rash of protests, separatist uprisings, and bombings broke out in
Xinjiang as a result of minority discontent over claims of cultural genocide, growing
inequality between Han and minorities, unfair policies, and lack of official autonomy
(Bovingdon, 2004). For example, in 1997, at National Peoples’ Congress (NPC)
meetings, minority-area delegates demanded greater autonomy from central officials
(Sautrnan, 1999). Specifically, minority leaders were quite discontented over the Law on
Regional Autonomy (LRA) enacted in 1984. The law stated that “autonomous areas are
empowered to adapt, modify, or supplement national laws according to local conditions”
(Sautrnan, 1999). Many leaders in autonomous regions, including minorities, however,
felt that little official power is given to autonomous regions, because all laws for the five
autonomous regions must first be approved by the NPC before being enacted. Also, the
provincial authorities must approve any law for the prefectures and the counties as well.
Some of the discontentment among minority leaders stems from the lack of

respect offered by central government departments. Previous administrations in Beijing

69

have been unwilling to acknowledge the existence of ethnic discrimination. In 1987,
Deng Xiaoping quoted, “since New China was founded in 1949, there had never been any
ethnic discrimination in the country.” In 1998, CCP Chief Zhao Xiyang noted that racial
discrimination was frequent in every other country, but not in China (Sautrnan, 1999).
Despite 280 laws and regulations on minority preferential treatment created in 1979-
1995, many minorities believe that the enacted policies have failed to: 1) increase ethnic
regional autonomy; 2) offset growing inequality between Han and minority groups; and
3) reduce growing anti-minority bias among Han Chinese (Sautrnan, 1999).

The uniqueness of Xinjiang highlights the need for regional geographies of scale
in order to examine the impact of China’s economic reform policies. Exploring China’s
economic reforms at the national scale will produce crude over-generalizations that 1)
lead to misspecification of social and economic relationships, particularly hiring practice
discriminations and patterns of migration; and (2) fail to recognize local and regional
anomalies that are imperative to understand in order to advocate informed, effective local
policies aimed at alleviating social unrest seen in various places throughout China. The
author hopes that the major findings from this paper will lead to local policies to be
enacted that will protect local and migrant minorities in Xinjiang from becoming further
excluded in the urban labor market. This will lead to greater social stability and harmony
and ultimately reduce the threat of future uprisings in the region. Additional attention

should be paid to females within both Han and Uighur nationality groups.

70

5. APPENDICES

71

Appendix A

Table 11: Employee-Level Survey Questionnaire

 

Descriptive Data

 

1. ID Number:

2. Day/ District

3. Self Employed: Y/N
4. Formal/Informal

5. SOE: Y/N

6: Low Skill: Y/N

 

Employment Indicators

 

 

1. What is the title of your current occupation?

2. How long have you worked at your current place of employment

3. How were you able to find your current job (i.e. walk-in, family member,
fiiend, government)?

4. How many hours/day do you work

5. How many days/wk do you work

6. Did you work at your current job in 2005? Yes or No

If yes do you make more, same, or less wages now?

7. Have you worked anywhere else? Yes or No

If yes, what location did you work (i.e. what province, county, or prefecture)
and what was your title?

8. Were you employed in the year 2000 (this job or another one)? Yes or No
If yes, do you make more or less money now

9. Have you ever lost your job and became unemployed? Yes or No

If you have ever lost your job please answer the next three questions a-c:

a. why did you lose your job;

 

72

 

 

b. how long were you unemployed for; c
c. How did you find another job?
10. If you could find employment in another place would you? If so, ideally

where would you like to work?

 

Descriptive Information

 

 

1 1. What year were you born in?
12. Are you male/female?
13. What is your highest attainment of Education?
14. Do you believe in a religion? Yes or No
If yes, what is it?
15. Can you speak Mandarin fluently? Y/N
16. Do you speak any other languages than your native language? Y/N
If Yes, please list:
17. Are you married?
18. Do you have a child?
19. What is your Nationality
20. Were you born in Urumqi? Y/N
If you were not born in Urumqi please answer the next 5 questions a-e:
a. what County/prefecture, province were you born:
b. Approximately what year did you migrate to Urumqi
c. Did you migrate with your father or mother?
(1. What was the main reason you (or your parents) decided to migrate to
Urumqi?
e. when you migrated did the government help you (or your parents) to find
employment?
21. What does your father and mother do for a living?
22. Was either your mother or father born in Urumqi? Y/N
23. Was any of your grandparents born in Urumqi?
24. How long have you resided at your present location in Urumqi?

25. What district of Urumqi do you presently reside?

 

73

 

 

Economic Indicators:

 

 

26. How often do you eat at restaurant per week?

27. Do you have at least 1 friend who owns his/her own business? Y/N

28. Do you have at least 1 fiiend who has a professional job (i.e. Lawyer,
Doctor, Professor)? Y/N

29. Do you own a car? Yes or No

30. Do you own a bicycle? Yes or No

31. Do you own or rent your apartment? Yes or No

32. Approximately how many square meters is the apartment you currently live
in?

33. If you rent your apartment how much money do you pay for rent per month
34. How much money do you spend on food per day,

35. How much money do you spend on transportation per day, ?

36. How much money do you pay for taxes per month,

37. How much money do you spend on clothes per month

38. How much money do you currently earn per week?

 

74

 

 

 

Appendix B

Figure 1: Composition of Workforce in Urumqi

 

 

40

 

Urumqi City Labor Composition

 

 

 

 

Product. Prof/Tech. Clerical Sales & Other
& Services
Transport.

75

Appendix C

Figure 2: Migrant Origin (Province)

 

 

r _.
l
Province of Orrgrn (%) l
i l
i I
l l
1 l
l I
l w -- -5 1
l
i —— .
l
i I
I t
l 1
i l

 

 

 

 

 

nsucg
Inquv

1qu —
uopuqu . ‘
H

 

 

76

 

Appendix D

Figure 3: Location Map of Urumqi Urban Districts

 

 

   

LEGEND
* CBD
C: 1949 Urban area

: 2004 Urban Area N
I

 

Source: Mappoint

77

 

Appendix E

Figure 4: Location Map of Xinjiang Uighur Autonomous Region

  
    
     

 

; P 5 N [I I l 9" jwnhnc Mark
I . i .-" - .
: 3 l ,3? 1mm" . ,
’ j ! l " Nunavut-‘- ' I
‘ l 4 1 . .f
._ _,_ l ___' 1 l 'Qron‘lo :1; 7

 

4.
4.-—

l Jim-vat . ‘ l . I .4‘, 1
oval-amen. , .. (32. b0 '1 os";==—*—J”"

Source: Mappoint

78

Appendix F

Figure 5: Education Level by Nationality for Xinjiang Province (ratio)

 

 

 

Education Level by Nationality (Provincial Level)

 

IHan

I We

 

 

No Primary Secondary College
Schooling

79

 

Table 12: Migration Reasons and Information (%)

Appendix G

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Uighur Han Hui Women Men

Reason for

Migration 44 59 85 76 76
Employment 10 6 5 9 9
Go to School 16 14 5 12 11
Parents 30 21 5 3 4
Other

Introduction to

present

employment 13 17 18 18 19
Friend 68 72 62 69 72
Self 11 5 8 8 6
Family 5 6 12 4 2
Government 3 0 0 1 1
Other

Destination district
Tianshan 62 23 39 34 30
Shayibake 35 26 31 26 27
Shuimo 1 38 8 30 32
Xinshi 2 13 22 10 11

Have held

Previous 26 59 58 53 53

Employment

Length of

migration 28 1 l 48 15 16

Short-terrn (< 1

year) 19 13 15 13 15

Mid-term (1-3

years) 54 76 37 72 69

Long-term (3+

years)

Employee type
Employee 57 48 46 56 44
Self-Employed 37 46 40 40 48
Employer 15 12 14 9 16

Occupation Type
Ret. Food/Bev 24 24 23 23 26

 

80

 

 

 

 

 

 

 

 

Ret. Elec/App 4 10 18 5 13
Ret. Clothes 7 10 0 1 7 4
Ret. Furn/Home O 7 8 7 4
Ref. Other 15 18 10 24 11
Transportation 4 4 5 O 7
Services 38 21 33 19 30
Professional/Tech. 8 5 3 5 5
Employed in
Formal Work 77 74 83 77 71
Employed 1n SOE 10 8 17 9 8
Education
No schooling 11 3 3 4 5
Primary 20 18 15 19 18
Secondary 31 32 23 32 31
High School 20 23 28 18 28
Technical 4 16 23 1 7 10
College 14 8 8 10 8
Fluent in
Mandarin 69 95 98 93 89
Gol’emmem 26 7 38 8 11
assrstance
Father Occupation
Farmer 59 74 50 7O 77
Employee 6 5 8 7 4
Retired n/a 6 19 5 6
Self-employed 24 3 1 3 l 0 3
Prof/tech. 10 12 10 8 10

 

 

 

 

 

 

81

 

Appendix H

Table 13: Variable Definitions and Sample Means

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Definition Mean
LWages_mo Monthly wages of Yuan in natural 7.2
logarithms
Nationality
Han I for Han, O for others .72
Uighur 1 for Uighur, 0 for others .21
Other 1 for Other (Hui), 0 for others .07
Migrant status 0 for Native, 1 for not native .29
Gender 1 for Male, 0 for female .52
Education
No schooling 1 for no schooling, 0 for others .04
Primary school 1 for primary school, 0 for others .16
Secondary school I for secondary, O for others .28
High school 1 for high school, 0 for others .26
Technical school 1 for technical, 0 for others .16
College 1 for college, 0 for others .10
Age
Age Years 33
Age squared Years squared 1193
Nationality-Education (Interaction Term)
Han-education N/A
Uighur-Education
Mi grant-Education (interaction Term)
Migrant-Education N/A
Native-Education
Married 1 if married, 0 for others .71
Children 1 if have children, 0 for others .60
Employment Type
Employee 1 if employee, 0 for others .51
Self-Employed 1 if self-employed, O for others .4]
Employer 1 if employer, 0 for others .14

 

 

 

82

 

Industry/ Occupation

 

 

 

 

Type 1 if in sell food/beverages, 0 for other .21
Retail- Food/Bev. 1 if sell electronics/appliances, 0 for other .09
Retail- 1 if sell clothes/accessories, 0 for other . 11

Electronics/app. 1 if sell fumiture/housing supplies, 0 for .06
Retail- clothes/ace. other .18
Retail- furniture I if sell other things (cultural specialties), .05
Retail- Other 0 if other
Transportation- 1 if taxi/truck driver, 0 if other .25
Service-

Professional/1‘ ech. I if in service job (security guard, street
cleaner, construction, restaurant barber, .04
receptionist), 0 if others
I if professional or business, 0 if other

Job Tenure 1 for 10 or more years of job tenure with .15

current job, 0 for others

District Type

CBD .67
Tianshan 1 if surveyed in Tianshan, 0 for others .39
Sha I if surveyed in Sha, 0 for others .28

Sprawl .33
Shuimo 1 if surveyed in Shuimo, 0 for others .23
Xinshi 1 if surveyed in Xinshi, 0 for others. .10

Market Type 1 if surveyed in large marketplace, 0 for .29
Big Market others .41
Small Market 1 if survey in small market, 0 for others .08

Alley I if surveyed in alleyway, 0 for others .05
Restaurant 1 if surveyed in restaurant, 0 for others .16

Other - school, hotel,
bank, etc.

 

1 if surveyed in professional place

 

 

(school, hotel, bank), 0 for others

83

 

Appendix I

Table 14: Principal Components Analysis Results

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Factor 1* Factor 2
Employee + *alnlr _
Self-Employed J” +
Employer - +
Retail-Food/Bev - -
Retail-Elect/App + +
Retail-Clothes + +
Retail-Furn/Home + +
Retail-Other + +
Transportation + -
Services + -
Professional/Tech. + +
F ormal + * ** +
Job Tenure - -
Age - +
Gender - -
Education + +
Mandarin Fluent + +

 

 

 

84

 

 

 

 

 

 

 

 

Han - + ***

Uighur + - ***
Nationality-other + -
Native + -
Migrant-Intra_Prov. + -
Migrant-Inter_Prov. - +
Wage_Pov - -

Tot. Var. Explained (%) 13.5 12.0

 

 

 

 

 

 

* ‘+’ indicates the variable loads positive with the factor, ‘-’ indicates
the variable loads negatively with the factor,

‘***’ indicates a variable loading with absolute value of at least .70.

 

85

 

Appendix K

Figure 6: Cluster Tree Produced by Hierarchical Cluster Analysis

 

 

Cluster Tree
’1

1 l I I
50 100 150 200

Distances

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O

 

 

86

Appendix L

Figure 7: Scores Mapped in Factor Component Space by Cluster

 

Cluster SPLOM

 

 

I HOIDVJ

Z HOLDVJ

 

 

 

 

.2498»: '3!
£51.“? 3 ‘3’...
. ’ « *8 ~*"~ r.‘ , CLUSTER
e W .3 o . I

 

 

I‘M.»
.115.
" v!

HINDU»)

 

FACTOR l FACTOR 2

 

87

 

6. BIBLIOGRAPHY

88

Amnesty International. “People’s Republic of China: Gross Violations of Human Rights
in the Xinjiang Uighur Autonomous Region.” Al Index: ASA 17/18/99 (1999).

Anderson, Kathryn, John Butler, and Frank Sloan. "Labor Market Segmentation: A
cluster analysis of job groups and barriers to entry." Southern Economic Journal, Vol. 53
(1987) '

Bauder, Harald. "Culture in the labor market: segmentation theory and perspective of
place." Progress in Human Geography, Vol. 25 (2001).

Becker, Gary. “The Economics of Discrimination.” Chicago, University of Chicago Press
(1957)

Becquelin, Nicolas. “Xinjiang in the Nineties.” China Journal. Vol. 44 (2000).

Benson, Linda. “Education and Social Mobility among Minority Populations in
Xinjiang”, in Frederick Starr, ed., Xinjiang: China '3 Muslim Borderland, (Arrnonk, NY:
2004),

Bhalia, Qiu, and Shufang Qiu. “Poverty and Inequality among Chinese Minorities.” New
York: Routledge (2006).

Biles, James, and Bruce Pigozzi. "The Interaction of Economic Reforms, Socio-economic
Structure and Agriculture in Mexico." Growth and Change, Vol. 31 (2000).

Bovingdon, Gardner. "Autonomy in Xinjiang: Han Nationalist Imperatives and Uighur
Discontent." Policy Studies 11, East-West Center (2004).

Constant, Amelie and Douglas Massey. “Labor Market Segmentation and the Earnings
of German Guest workers.” Journal Population Research and Policy Review, Vol. 24
(2005).

Dillon, Michael. “Xinjiang: China's Muslim Far Northwest.” London and New York:
Routledge (2004).

Drago, R. "Divide and Conquer in Australia: a study of labor segmentation." National
Institute of Labor Studies Working Paper, 121 (1992).

Fan, Cindy. "The Elite, the Natives, and the Outsiders: Migration and Labor Market
Segmentation in Urban China." Annals of the Association of American Geographers, Vol.
92(2002)

Fan, Cindy. "Rural-Urban Migration and gender Division of Labor in Transitional
China. " International Journal of Urban and Regional Research, Vol. 27 (2003).

89

F ichtenbaum, Rudy, and Kwabena Gyimah-Brempong and Olson Paulette. "New
Evidence on the Labor Market Segmentation Hypothesis." Review of Social Economy,
Vol. 52 (1994).

Fields, Gary. "Segmented Labor Market Models in DeveIOping Countries." Cornell
University ILF schooleaper forthcoming Harold Kincaid and Don Ross, The Oxford
Handbook of the Philosophy of Economic Science, Oxford University Press (2008).

Forbes, Andrew. “Warlords and Muslims in Chinese Central Asia: A Political History of
Republican Sinkiang 1911-1949.” New York (1996).

Graham E. Fuller and Jonathan N. Lipman, “Islam in Xinjiang”, in Frederick Starr, ed.,
Xinjiang: China ’3 Muslim Borderland, (Arrnonk, NY: 2004).

Gladney, Dru. “Dislocating China: Muslims, Minorities, and Other Subaltem Subjects.”
Chicago: University of Chicago P (2004).

Gordon, Ian. "Migration in a Segmented Labor Market." Transactions of the Institute of
British Geographers Vol. 20 (2008).

Guo, Fei, Naran Bilik, and Robyn Iredale. “China's Minorities on the Move.” New York:
ME. Sharpe, INC (2003).

Haider, Zaid. “Sino-Pakistan Relations and Xinjiang’s Uyghurs: Politics, Trade, and
Islam along the Karakoram Highway”, Asian Survey, Vol. 45 (2005)

Hannum, Emily, and Xie Yu. "Ethnic Stratification in Northwest China: Occupational
Differences between Han Chinese and National Minorities in Xinjiang, 1982-1990."
Demography, Vol. 35 (1998).

Hayter, R and TJ Barnes. “Labour Market Segmentation, flexibility, and recession: A
British Columbian case study.” Environment and Planning C (1992).

Hiebert, Daniel. “Local Geographies of Labor Market Segmentation: Montreal, Toronto,
and Vancouver.” Economic Geography, Vol. 75 (1999).

Hsing, You-Tien. “Making Capitalism in China: The Taiwan Connection.” Oxford
University P (1998).

Kim, Hodong. “Holy War in China: the Muslim Rebellion and State in Chinese Central
Asia, 1864-1877.” Stanford (2004).

Klecka, William. “Discriminate Analysis.” Sage publications (1980).

90

Leontaridi, Marianthi. "Segmented Labor Markets: Theory and Evidence." Journal of
Economic Surveys, Vol. 12 (1998).

Li, Si-Ming. "Population Migration, Regional Economic Growth and Income
Determination: A Comparative Study of Dongguan and Meizhuo, China." Urban Studies,
Vol. 34 (1997).

Liang, Zai, and Michael White. "Internal Migration in China, 1950-1988." Demography,
Vol. 33 (1996).

McLafferty, Sara and Valerie Preston. “Spatial Mismatch and Labor Market
Segmentation for African-American and Latina Women.” Economic Geography, Vol. 68
(1992)

McMillen, Donald. “Chinese_Communist Power and Policy in Xinjiang, 1949-1977.”
Boulder, Colorado: Westview Press (1979).

Millward, James. “Beyond the Pass: Ethnicity and Economy in Chinese Central Asia.”
Stanford (1998).

Milward, J arnes and Peter Perdue. “ Political & Cultural History Through the Late 19‘h
Century” in Frederick Starr, Ed Xinjiang: China ’3 Muslim Borderland. Studies of
Central Asia and the Caucasus. Armonk, NY: ME. Sharpe (2004).

Milward, James and Nabijan Tursun. “Political History and Strategies of Control, 1884-
1978” in. Frederick Starr, Ed_Xinjiang.' China ’3 Muslim Borderland. Studies of Central
Asia and the Caucasus. Armonk, NY: ME. Sharpe (2004).

Moneyhon, Matthew. “Taming China’s “Wild West”: Ethnic Conflict in Xinjiang.”
Peace, Conflict, and Development, Vol. 6 (2004).

Oshima, Harry. “Economic Growth in Monsoon Asia: A Comparative Survey.” Tokyo:
University of Tokyo P (1987).

O’Sullivan, David, and David Unwin. “Geographic Information Analysis.” John Wiley
and (2003) Sons, Inc. (2003).

Pannell, Clifton and Phillipp Schmidt. "Structural Change and Regional Disparities in
Xinjiang, China. " Eurasian Geography and Economics, Vol. 47 (2006).

Peck, Jamie. “Work-place: the social regulation of labor markets.” New York: Guilford P
(1996)

Perdue, Peter. “China Marches West: The Qing Conquest of Central Eurasia.”
Cambridge, Massachusetts (2005).

91

People’s Daily Online. "200 Million Peasant Workers to be Trained by 2010." [Beijing]
(2004).

Reich, Michael, David Gordon, and Richard Edwards. “A Theory on Labor Market
Segmentation.” American Economic Association, Vol. 63 (1973)

Rummel, Rudolph. “Understanding Factor Analysis.” The Journal of C onflict
Resolution, Vol. 11 (1967).

Sautrnan, Barry. "Ethnic Law and Minority Rights in China: Progress and Constraints."
Law and Policy, Vol. 21 (1999).

Shichor, Yitzhak. “The Great Wall of Steel: Military and Strategy in Xinjiang” in
Frederick Starr, Ed_Xinjiang: China '3 Muslim Borderland. Studies of Central Asia and
the Caucasus. Armonk, NY: ME. Sharpe (2004).

Solinger, Dorothy. “Contesting Citizenship in urban China: Peasant Migrants, the State,
and the Logic of the Market.” University of California P (1999).

StarSoft, Inc. “Discriminate Function Analysis.” Copyrighted 1984-2008.
<http://www.statsoft.com/textbook/stdiscan.html.>

Su, Wang, Naran Bilik, and Robyn Iredale. “Contemporary Minority Migration,
Education and Ethnicity in China.” Northampton, MA: Edward Elgar Limited (2001).

Syrquin, Moshe, and Hollis Chenery. "Patterns of Development, 1950-1983."
Washington, DC: The World Bank (1989).

Telles, Edward. "Urban Labor Market Segmentation and Income in Brazil.” Economic
Development and Cultural Change, Vol. 41 (1993).

Toops, Stanley. “The Demography of Xinjiang” in Frederick Starr, Ed_Xinjiang: China '3
Muslim Borderland. Studies of Central Asia and the Caucasus. Armonk, NY: ME.
Sharpe (2004).

UN-Habitat. “Strategies to Combat Homelessness.” The United Nations Human
Settlements Program, Environmental Publications from UNEP and Key lntemational
Organizations (2000).

XPS, Xinjiang Provincial Statistics. China Data Center, University of Michigan (2005).

Xu, Wei. "Segmented local labor markets in post reform China: gender earnings
inequality in the case of two towns in Zhejiang province." Environment and Planning A,
Vol. 38 (2006).

92

Wiemer, Calla. “The Economy of Xinjiang” in Xinjiang: China ’3 Muslim Borderland.
Studies of Central Asia and the Caucasus. Frederick Starr, Ed. Armonk, NY: ME.
Sharpe (2004).

Wu, Fulong. "Urban Poverty and Marginalization under Market Transition: The Case of
Chinese Cities." International Journal of Urban and Regional Studies, Vol. 28 (2004).

Wu, Zhengzhang. "China’s Industrial Restructuring in the Twenty-First Century." Tokyo
Club Foundation for Global Studies: Promoting studies of the global economy (2001).

Zang, Xiaowei. "Labor Market Segmentation and Income Inequality in Urban China."
The Sociological Quarterly, Vol. 43 (2002).

93

lillll

3
9
2
1
3

H
E“
fl
”
"I
”