Internal Migration in Contemporary China Donald J. Treiman April 2009 1.

1

Internal Migration in Contemporary China

Donald J. TreimanApril 2009

2

Outline

• Introduction: brief history of migration in China.

• Describe our just-completed migration survey and outline our plans for the future.

• Show some preliminary results.

3

The hukou system

• China built an urban welfare state on the backs of the peasants.

• In 1955 established an internal registration (“hukou”) system.– Overarching agricultural vs. non-agricultural

(“rural” vs. “urban”) status, acquired from mother and very difficult to change (Wu and Treiman 2004).

– Local vs. non-local status.

4

Hukou system (2)

• Separate welfare provisions for rural and urban populations; inferior or non-existent for rural population: health care, housing, education, jobs, unemployment, disability, and retirement benefits.

• Many services restricted to those with local registration, or require high non-resident fees. Example: education in Beijing. Also, health care. Until recently, housing, etc., connected to danwei (work unit).

5

Migration trends

• Severe restrictions on migration from 1961-1978 (end of Great Leap Forward to beginning of Economic Reform) [hotel room example].

• Increasing migration since then, due to– Push: “family responsibility system,” resulting

agricultural labor surplus.– Pull: economic expansion in urban areas,

resulting in need for low-level labor (factory, construction, service, and sales).

6

Migration trends (2)

• Currently 150 million migrants (people living other than where they are registered), 12% of Chinese population.

• Migration is complex:– Rural-to-urban labor migration; but also rural-to-

rural (often replacing rural-to-urban migrants), urban-to-urban.• “Permanent” migration (resulting in 2nd generation?).• “Circular” migration.

7

Migration trends (3)– Marriage migration.– Formal migration, with hukou mobility (mainly

through education).• More work needs to be done (by me) to

create a viable migration typology (due to complexity of migration, errors in reporting).

• For now, I am working with a 9-category typology based on registration, residence, and whether the respondent has a local hukou.

• But I collapse this typology in various ways for the results I will show shortly.

8

Migration trends (4)

Rural registration Yes NoLocal registration Yes No Yes NoRural residence (1) (2) (4) (4)Urban residence (3) (2) (4) (4)

(see the next slide for the four resulting categories)

9

Migration trends (5)

10

Migration trends (6)

• Nearly half (49%) the urban population lacks urban privileges (=urban peasants + rural migrants living in cities and towns). China is willy-nilly creating a two class urban population, similar to the U.S. (inner city Blacks and Latinos); more about this later.

• The current Chinese saying, “the difference between urban and rural status is like the difference between heaven and earth,” needs to be expanded to take into account migration dynamics.

11

The 2008 Chinese Migration and Health Survey

• Overall goal: analyze determinants, dynamics, and consequences of internal migration for health and well-being.

• Specific aims:– Who moves and why: for work, for schooling, for

marriage, or for other reasons?– How selective is the migration stream?• The “healthy migrant” hypothesis.• Other kinds of “migrant quality” selection.

12

Survey (2)– Migration dynamics:• Is the “floating population” (liudong renkou) truly

floating? How is the process of labor migration organized, in terms of advance arrangements for jobs, housing?• The pattern of movement—nearby or long distance?• Who moves? Individuals or families?• The extent of “circular” migration.• Who returns home? The “salmon” hypothesis.

– Migration outcomes: • The migration experience (working conditions,

housing, health care, trouble with authorities, loss of social support, etc.)

13

Survey (3)

• Material well being (employment, occupational advancement, income, durable goods, etc.)• Health consequences.

– Consequences for families/villages left behind:• Risk diversification.• Remittances.• Loss of labor.• Loss of social support.• Infusion of capital (material and human).

14

Survey (4)

– Consequences for receiving communities:• Development of an urban underclass (like Blacks and

Latinos in the U.S.)? Possible consequence of lack of adequate schooling for migrant children.

15

Survey (5)

• Sample design– So far a single cross-section of 3,000 adults.– Seeking new funding to:• Create a panel study, with new data every 3 years.• Expand the sample through a combination of

– A larger probability sample;– Targeted samples: families left behind, returned migrants,

factory workers, construction workers, entrepreneurs, and tertiary-educated people (to compare urban-origin high status people with formal hukou changers).

16

Survey (6)

• Design of current sample– National probability sample, with oversample of high

in-migration areas and high out-migration areas, stratified by education level and township type (jiedao, zhen, xiang).

– Primary sampling units: 150 townships (average population size 25,000).

– Within each township, 4 small areas randomly chosen, completely enumerated; households randomly selected; adults within households randomly selected.

17

Survey (7)• Doing sample surveys in contemporary urban China

is very difficult.– Small area statistics no longer reliable; incomplete

coverage, particularly of migrants. Thus our decision to enumerate small areas.

– Many people live in non-standard housing (at their work place, in rooms added to buildings, etc.)

– Change in housing stock. Nearly half (47%) of the urban population lives in recently built (since 1995) buildings; probably most are controlled access.

– Rise of distrust.• Thus, response rates are falling in China, as

elsewhere.

18

• Here are some of the kinds of dwellings enumerators and interviewers face in China today.

“Village in a city” in Guangzhou

Dwelling for road repair workers

21

Boat dwelling near Suzhou

22

The guard at the gate may not permit entry.

23

Even downscale apartments are often difficult to get into

24

Survey (8)• Questionnaire includes extensive information

on education, work, and migration histories; characteristics of family members; life experiences; etc.

• Distinctive feature: the collection of health histories, biometric data (height, weight, blood pressure, lung capacity, etc.) and also dried blood samples, from which to do bioassays of disease markers: we plan to measure blood sugar, anemia, cholesterol, and c-reactive protein.

25

Survey (9)

• Emphasis on health reflects growing interest of demographers, students of social stratification on health. In some sense, health disparities are the ultimate indicator of inequality.

• Collecting blood necessitated distinctive data collection procedure: pairs of interviewers—a survey interviewer and a community doctor.

• Now seeking funding to do biomarker work.

26

Survey (10)• The remainder of the study team:

– HU Peifeng (Perry HU), MD, PhD, Assistant Professor of Medicine, UCLA

– LU Yao, PhD, Assistant Professor of Sociology, Columbia University

– William M. MASON, PhD, Professor of Sociology and Statistics, Emeritus, UCLA

– QI Yaqiang, PhD, Assistant Professor of Sociology, People’s University, Beijing

– SONG Shige, PhD, Associate Professor, Institute of Sociology, Chinese Academy of Social Sciences

– WANG Wei, MD, PhD, Dean, School of Public Health and Family Medicine, Capital Medical University, Beijing

– Amy ZENG, Account Manager, MillwardBrown/ACSR, Beijing

27

Don Treiman, Bill Mason, HU Peifeng, SONG Shige, and CMU graduate student J.J. QIU

28

LU Yao and QI Yaqiang

29

WANG Wei and Don Treiman

30

Amy ZENG

31

Survey (11)

• Funding Sources– U.S. National Science Foundation (NSF)– U.S. National Institute of Child Health and Human

Development (NICHD)– China, “863” grant program– Capital Medical University, Beijing

32

First results• We are still in the process of cleaning our

data and trying to figure out what we have. So I am simply going to offer – an analysis of what increases the odds that

peasants go out for work (discrete-time hazard rate models for people 14-58);

– preliminary analysis of the consequences of migration, contrasting current rural migrants with those who have never migrated, adjusting for sample selection bias using propensity scores.

33

34

35

36

37

38

39

40

41

42

43

44

45

46

Conclusions (2)

• Remaining questions:– Disentangling cause and consequences of

migration.– Assessing “Salmon hypothesis”—unsuccessful or

unhealthy migants return home.– The many topics mentioned at the outset,

including:• Dynamic of migration.• Consequences for those left behind.• Consequences for receiving communities

47

Conclusions

• Our analysis is just beginning. Many intriguing questions:– Selectivity of migrants (for health, efficacy, etc.);

some hints (parental education, larger villages, better childhood diets, less frailty (although causal order ambiguous).

– Huge return in terms of income.– Cost: difficult physical and social working and

living conditions.

48

Descriptive comparisons

• Let’s consider:– Socio-economic outcomes– Working conditions– Other troubles– Hygiene– Diet– Physical health– Emotional health

49

50

51

52

53

54

55

56

57