Do U.S. migrants and China migrants
move in a different way?
An empirical comparative study of migration
Sun Qian, University of Groningen Supervisor: Doctor. Jianhong Zhang
Co-assessor: Leeuwen, dr. E.H. van Final accepted, July 2006
IE&B SID Program( Shortened Initial Degree), Economics Faculty Landleven 5
Postbus 800
9700 AV Groningen, the Netherlands, Student number: s1375334
Do U.S. migrants and China migrants move in a different way?
An empirical comparative study of migration
Abstract
This is an empirical comparative study of internal migration patterns.
Using internal migration data from U.S. and China’s 2000 censuses, this paper
attempts (a) to empirically investigate the key economic, non-economic
driving forces behind the migratory behavior in China and U.S. (b) to examine
the influences and the degree of sensitivity of each migration factor on each
country pattern (c) to reveal differences of internal migration patterns
between a highly industrialized country and a less developed country (d) to
test the general hypothesis that “internal migration pattern shifts when a
higher economic level is attained”.
Content
1. Introduction ………...3
2. General facts about China internal migration and U.S. internal migration ……….5
2.1 Relatively low mobility in China ……….………..5
2.2 Evenness of regional development and internal migration ………6
2.3 Migration reason and motivation differences ……….8
3. Literature review ………10
3.1 Step migration theory ………..10
3.2 Two major fields of migration study ………..12
3.2.1 Labor mobility migration study ………...12
3.2.2 Social Demographic migration study ………..14
3.3 Theoretical model and hypothesis ………..18
3.3.1 Theoretical model ………..16
3.3.2 Empirical models ………..17
3.3.3 Hypothesis ……….18
4. Methodology ………24
4.1 Econometric equations ……….24
4.2 Definition and measurement of variables ………..24
1. Introduction
Internal Migration is one of the key determinants of population change and the most important cause of regional demographic shifts. It is recognized now not only as a problem-creating phenomenon, but also as a problem solving social process, especially in those emerging economies. As natural population growth rate started to slow down globally1, migration became the main factor to influence regional population variation and demographic distribution.
Besides of the impacts on demographic and social factors, migration also exerts important influences on more purely economic factors, such as local GDP, employment, wage rate etc,. A great deal of studies examined the causality between economic growth and internal migration. Bal Kumar KC (2003) pointed out that gross mobility is positively associated with (economic) development. Cat Moondy (2006) argued that migration may contribute directly to economic growth in both the host and source countries. Chapple and Yeabsley (1996) used the neoclassical economic theory with some short-run Keynesian interactions successfully identified the channels through which migration affects national welfare.
Once the correlation relationship is widely recognized and confirmed both empirically and theoretically, there comes up an interesting question: Does a corresponding relation exist between the population mobility pattern and the economic development level? Or equally, how does migration pattern interact with economic development? This empirical research paper aims to address the following questions: (1) What are the factual differences between the recent migration patterns of China and U.S.? (2) What are the key motivating factors and discouraging factors of migratory behavior in these two countries? (3) Do these determinants exert the same influences on Chinese migrants and American movers? (4) If not, what are the reasons behind?
A so called place-to-place model will be established to examine the pull-factors and push-factors in highly industrialized country and less developed country separately. Six sub-hypotheses are formulated under an OD (originality-destination model), which enables
2. General facts about China internal migration and U.S. internal migration 2.1 Relatively low mobility in China
Compared to most capitalist economies, China’s population mobility is low. Between the years 1985 and 1990, the rate of interstate migration in the U.S. was 9 percent, while its counterpart in China- rate of interprovincial migration - was only 1 percent (see Figure 1). Recently, the interprovincial migration rate climbed up till 3%, the gap in mobility between China and U.S. becomes narrower, but still remain large in absolute term.
Figure 1. U.S. migration rate2 and China migration rate comparison 1985-1990, 1995-2000
9% 8% 1% 3% 0% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% 1985-1990 1995-2000 U.S. China
Resource: U.S. Census Bureau 1992 & 2003; State Statistical Bureau 1993, 20, 72; National Bureau of Statistics 2002, 1813-1817; )
To understand why China has such a low migration rate, we have to realize that the central planning is a key factor of mobility control system during China’s socialist period. In fact, like many other distorted market failures, migration can be regarded as a political economic phenomena in China. It is used as a tool of political and economic planning rather than free individual choice based on rational economic utility maximization that prevail in those highly industrialized country. The hukou3, implemented in the late 1950s,
has been the cornerstone of the government’s control over population mobility and a key component of socialist planning more generally (Cheng and Selden, 1994). This personal
administrative system, is similar to an internal passport system, curtailed self-initiated moves and limited migration from intra-provincial and inter-provincial movement. In China, an “illegal” migrant who does not possess the legal status to locate in the receiving place, has to bear high non-pecuniary cost and huge opportunity cost. Without hukou, it would be extremely difficult to find a formal job, neither are migrants entitled to any social welfare benefits and other privileges4 equally as local residence. Other related restrictions also sharply reduce the benefits and raise the costs of migration, particularly in large cities. The political prohibiting migration period lasted until the middle of 80s, since the late 1990s the restrictions on migration were started to loose gradually. In July, 2000, Ministry of Labor and Social Security, Ministry of Agriculture and State Council jointly issued a official document “Guidelines on strengthening experimentation to develop rural labor employment”, mainly aimed to narrow the rural-urban gap and abolish unreasonable restriction on employment of rural migrant laborers in cities (Huang and Frank N.2003). Publicly, this document is regarded as the turning point of central government’s attitude toward the sharp increase of migration flows. Though most recent policies and trends are towards gradual migration relaxation, in fact, these migration policy reforms of the hukou only partially apply to China’s largest cities that have the discretion to limit and select immigrants. The basic provisions remain in place even today. This means there still remains a large wedge between the social and private net benefits of migration. Thus, I argue that hukou now is still being enforced as a barrier to hinder and discourage cross-province migration in China. And we have to always bear in mind that unlike other country cases, the Chinese migration pattern is distorted by political reasons.
2.2 Evenness of regional development and internal migration
At individual level, migration can be regarded as a personal behavior of interacting with the push and pull factors between original residence place and destination place. From a macro perceptive, we can treat a country’s national mobility pattern as a response to regional differences, and over a sufficient period, migration reduces the difference in wage rates, cultural amenities or whatever. This reciprocal relation had been widely recognized
by in many migration and regional development researches. Greenwood (1981) identified a bi-directional causality relationship between regional economic development and migration by simultaneous equations. A great deal of succeeding regional economic studies on China also help to confirm this interesting relationship. As Zhu (2003) argued that the inter-provincial migration is essentially motivated by socio-economic disparities in conjunction with unbalanced development strategy. Fan (1996) point out that economic growth, job opportunities, and higher wages in more developed regions exert a strong pull to migrants from poorer regions. Especially those populous Central Western provinces5 are the main sources of surplus labors and out-migration flows. Each year, mass migrants are crowding into more developed Eastern costal cities to seek better economic opportunities. Thus, Eastern regions continuously gain a large amount of cheap laborers which are the key input for further industrialization and manufacturing expansion. Consequently, divergence in the levels of economic development among regions increases.
In China, economic reforms and uneven regional development have driven population movement. Economic reforms have widened the development gaps between regions. As a populous developing country, huge amount of surplus labor in conjunction with low arable land per capita, cause large-scale internal migration surge corresponding to the wide regional disparity, even with the presence of government control. In a decennial census research, Fan (2004) demonstrated the recent spatial direction of migration flow is also consistent with the uneven regional development level - from the poor South Central and Southwestern China to the most developed Eastern-regions. Moreover, her research shows the relationship between migration and regional development is getting stronger over time. Comparing to China, American society is becoming more homogeneous. Franklin D.Wilson’s (1988) conducted a empirical research about migration pattern of U.S.. The results supports the hypothesis that “the migration patterns of an advanced industrial society reflect a process of regional convergence in which the benefits of socioeconomic development are spread evenly throughout the society.” He argued that socio-economic convergence reduces the impact of push and pull forces on migration to a minimum. Kris James Mitchener and Jan W.McLean (1999) further pointed out that “U.S. economic
growth over the last century has been accompanied by convergence in nominal state personal income per capita” in a recent research. Thus, as the divergence decreasing, fewer people are stimulated to move in U.S..
2.3 Migration reason and motivation differences
Classical economic theory defined a rational migrant as the one who expected to be driven by better economic opportunity which is unrealizable in his original place. In the early industrial revolution era, economic opportunity was discovered to be the key driving force of internal migration in the capitalist economies. Likewise, nowadays, in those emerging markets, economic utility maximization motivation - typically income disparity and job reasons - are still the strongest pull factors of population mobility.
Although the importance of economic motive has been once again emphasized in modern migration studies, however, it is not the only one that stimulates internal migration. Noticeably, migration reasons became far diversified in highly industrialized countries in the last few decades. Non-economic reasons - such as the QOL-related (quality of life) factor (includes housing, climate, environment, etc) - are found to become the main considerations instead of pure economic-profiting motivations. U.S. census 2000 conducted a national survey to investigate the reasons-for-moving. Results show 52 percent movements are for housing related reasons, followed by family (26 percent) and work related reasons (16 percent). Contrary to the huge amount of job-seeking Chinese migrants, the American’s motives for job reasons are surprisingly low - “To look for work/lost job” only accounts for 1.3 percent.
Whereas we see a completely different picture in Chinese migration motivation survey. Employment in industry and business is absolutely the dominant reason to move. The percentage of this reason reaches 54.7% in non-hukou migrants (or called floating people) and even 100% for rural labor migrants. Industrial workers and farmers are the top two occupational statuses of those migrants who are in workforce6. This general difference can be further confirmed by exploring the age composition of American and Chinese internal migrants. The high concentration on 20 to 29 age of Chinese migrants, in conjunction with
the occupation status characteristics described above, can be explained as young labor, who move across province to pursue better economic opportunities. On the contrast, U.S. migrants’ age composition are more evenly spread over all different age groups due to more diversified migration reasons.
Figure 2. China cross-province and U.S. cross-state migrants by age composition
0% 10% 20% 30% 40% 50% 60% U.S. China
Resource: National Bureau of Statistics. 2002. (Tabulation on the 2000 Population Census of the People’s
3. Literature review 3.1 Step migration theory
Ravenstein (1885) first introduced us the concept of “step migration” in his two seminal papers titled “The Laws of Migration”. By using census data on individuals’ places of birth versus their places of current residence, he identified the short-distance as the key characteristic of the structures in the streams of population movement under way in the British Isles during the industrial revolution. Afterwards, migrants moved from farms to villages, from villages to towns,from towns to counties, and then to large cities - step by step, more and more farther.
Recently, two recent migration study researchers, D. A. Plane, C. J. Henrie, and M. J. Perry (2005) categorized the mobility transition into four stages according to the historic evolution of migration pattern (see Table1). They argued that rural-to-rural movement should initially predominate but then give way to rural-to-urban net migration when industrialization draws workers to manufacturing jobs in cities. Subsequently, after an integrated urban system has developed, migration should become primarily urban-to-urban. Based on these, a possible fourth stage, so called “counter urbanization” , “down-the-size-hierarchy” migration flows (Wardwell, 1980), “sub-urbanization migration” (Matthew E.Kahn, 2000) came up. Franklin D.Wilson’s (1988) pointed out that at this highest stage of development, a population equilibrium is reached, as well as a convergence in the social and economic structure of metropolitan and no metropolitan areas. Moreover, migration becomes essentially a random process, no longer substantially affected by area differences in resources, labor, or capital.
Table 1. D.A.Plane, C.J.Henire, and M.J.Perry’s Mobility Transition Matrix
William H.Frey (1988) proposed two perspectives on population distribution, named “regional restructuring explanations” and “de-concentration explanations”, provide us another important insight in understanding the different migration patterns of less developed and highly industrialized countries. He argued that less developed country continues to pervasive “up-the-size-hierarchy” migration flows due to the economies of agglomeration effects. However, the longstanding historical relationship between industrial development and urban population concentration has begun to erode in many of the world's highly industrialized countries, like United States, Great Britain and Canada. Factually, since the late sixties there has been an apparent shift in the direction of population movement in the majority of the American states, contrary to the conventional migration stream - mainly from rural to urban settings. This new trend has taken a new direction from metropolitan urban to non-metropolitan rural areas, reversing the traditional pattern of population movement (Figure 3). The recent migration processes of United States further implies a redistribution from core to peripheral regions and migration stream exchange that downward within the urban hierarchy trend (D.A.Plane, 2005). However, step migration today still remains widespread throughoutmuch of the developing world.
Figure 3. U.S. Metropolitan, non-metropolitan migration
5,000,000 5,200,000 5,400,000 5,600,000 5,800,000 6,000,000 6,200,000 6,400,000 6,600,000 6,800,000 1975 to 1980 1985 to 1990 1995 to 2000 Metropolitan to Nonmetropolitan Nonmetropolitan to Metropolitan
Resource: Migration and Geographic Mobility in Metropolitan and Non-metropolitan America: 1995 to
3.2 Two major fields of migration study
The study of migration can be divided into two major fields, namely labor mobility and social demographic studies. The studies in the first field are treated for the most part from an economic perspective, mainly focus on labor reallocation in response to market needs. However, the general framework for social demographic studies of migration is the push-pull model, as Lee (1966) argued that the decision to migrate includes consideration of positive and negative factors at sender and receiver areas, intervening obstacles, and personal factors. For instance, Human capital formation model of internal migration and
Harris and Todaro’s less developed country migration model are developed in labor
mobility study frame.
3.2.1 Labor mobility migration study
Human capital formation model of internal migration
Gene Laber and Richard X.Chase studied the interprovincial migration in Canada within a human capital investment framework. They test the interprovincial migration data of 1961 Census by purely focusing on money motivator.
(
)
0 1 ˆ ˆ 2
ij j i ij
M =α α+ W W− − ×α D
discount rate and the geographic distance.
There are few limitations in this model. First, this so called human capital formation model narrowed the definition of a migrant as a labor who regard economic utility maximization as the unique objective, which means wage difference is the sole driving force in this extremely simple model. However, nowadays people’s motivations to move are far diversified than pure economic reasons - examples of migration forces in developed countries are housing condition, population density, climate etc, - therefore I argue it is inappropriate to apply a simple cost-benefit economic model onto migrant study mechanically. Moreover, I argue that average annual wage earnings might be a less appropriate income measurement than household income, because dividends, investment return, property rent are also important income resources that cannot be neglected.
2. Harris and Todaro’s migration model for less developed countries
Harris and Todaro (1970) developed a migration model for less developed countries focusing on equating expected earnings between the rural and urban sectors. He specified an aggregate labor supply equation as below:
( )
( )
(
( )
)
ˆS t t F t
S = +β π α
ˆS represents net rural urban migration, S is the existing size of the urban labor force;
( )
ˆS t
S is the rural-urban migration rate during period t; is the natural rate of increase of the urban labor force, α(t) indicates the percentage urban-rural real income differential, π(t) refers to the probability of being employed, and π(t)F(α(t)) is the expected income
differential7- the income differential adjusted for the probability of finding an urban job,
hence it reflects the rate of urban labor force increases a result of migration.
This classic model is regarded as the corn stone of many migration studies later. The model starts from the assumption that migration proceeds in response to urban – rural
differences in expected income rather than actual earnings. Todaro (1969) emphasized the importance of the probability of obtaining employment on the decision to migrate. The key point of his argument is that the earnings variable should be an expected value, determined by both earnings and the probability of obtaining a job. The model does provide a useful framework for theoretical studies on rural – urban migration. However, in practice, the model has critical limitations, as it does not consider the effects of changes in industrial structure and the existence of rural industries on migratory behavior. As Todaro’s model was built within an the thyroidal framework of labor surplus, thus it is more appropriate for the primary “rural-urban” migration stage. When apply to study modern migration, it needs to be modified and extended.
Thus, in labor mobility migration study, wage gap and employment possibility are deemed as the effective leverage mechanisms to equalize the labor demand-supply through self-adjusted migration flow. Nevertheless, these simple cost-benefit models are over focus on the pure economic factors, dealing economic utility maximization as the sole driving force. Hence, in practice, these primary models have critical limitations, as they do not consider the effects of non-economic factors on migratory behavior, which are argued to be crucial for modern migration study.
3.2.2 Social Demographic migration study
The study towards social demographic direction become so fruitful nowadays. Place-to-place migration model is established within this framework. This type of model has been widely developed in market economies, in both developed and developing countries (Rogers, 1967; Greenwood, 1969; Wadycki, 1974; Fields, 1982; Schultz, 1982), since it can be used to capture the importance of push factors at origins and pull factors at destinations. Apart of this advantage, it also allows the user to trace the spatial direction of the migration flow. The model can been extended and adapted to different countries by drawing different factors.
3. Economic transformation model
agriculture-dominated economies into industrial ones - a common phenomenon in developing countries.
0 1 2 3 4
ij ij ij i j
M =β β+ Y +β C +βU +βU +ξ
(i, j represent receiving and sending places respectively)
The regression model took in-migration rate as dependent variable, relative income of receiving places to sending places (Y), costs of migration (C), unemployment rate of urban formal sector sending and receiving places (U) as explanatory variables.
There are several drawbacks in this model. First, it is well known that migration is an important factor of population redistribution and that it is strongly related to regional economic development. A large body of literature has examined these relationships (e.g., Greenwood 1981). China’s spatial patterns of migration is especially closely related with regional disparity. The booming coastal region is highly urbanized and benefits most from early openness of their economies. Nowadays, the coastal provinces have become the leader in the national economy and the major destinations of labor flows. Census 2000 further demonstrated that shares of migrations between central and western regions decreased, while inter-regional migration from central and western to eastern region constantly increased. Thus, government’s preference regional policy does exert direct and effective influences on the spatial migration flow. However, policy factor is missing in this transformation model. Except for this, Chinese migrants are mainly urban-towards, large population metropolitan areas do possess strong pull effect due to economies of
agglomeration. Or in the other words, like other developing country, Chinese migration pattern is typically “up-hierarchy” or “urban-wards”. On the contrary, depopulation of metropolitan is a phenomena widely perceived in many highly industrialized countries. 4. Place-to-place migration model
1 2
ij i j m mj n nj ij ij
migration model to examine the interprovincial migration pattern. ij
M is the gross migration rate from region i to region j, Pi and Pj are populations in region i and j, D is the geographic distance between pairs of regions,ij X , mj X are the nj socioeconomic variables in region i and j, which are further extended to consumption per capita (C C ), agricultural share (,i j A A ), migration stock ( MS ), foreign direct investment ,i j (FDI ) and land-labor ratio (j, LL LL ) as below: ,i j
(
, , , , , , , , , ,)
ij ij i j i j i j j j i j
M = f D P P A A C C MS FDI LL LL
Similarly, Fu and Gabriel (2002) built a more complicated place-to-place model which included more explanatory variable, such as origin-destination ratio of
unemployment-adjusted urban wage rates, origin-destination log ratio of average
consumer spending, origin total TVE (town-village-enterprise) income per rural worker, origin agricultural density, log of urban work force, etc.
However, above two models more relevant for agricultural developing countries, due to the particularity of those explanatory variables. In this paper, I will adapt the
place-to-place model to examine the push-pull effects of the several key driving factors on population mobility pattern in highly industrialized country and less developed country.
3.3 Theoretical model and hypothesis 3.3.1 Theoretical model
diversified research needs.
Functionally, place to place models link the gross migration rate between origins and destinations to an array of factors, such as demographic, socioeconomic variables at origins and destinations, and geographic distance. In addition to the population, distance and socioeconomic variables such as income, unemployment or employment are often included in the model (Mueller, 1982). Thus, the determinants of internal migration are classified into three categories: (1) economic factors (2) non-economic factors and (3) country-specific factors. Thus, the general theoretical model is:
, , , , , ,
( )
ij ij i j j ij
M = f D P P UR IncomeGap γ δ
3.3.2 Empirical models
To suit the actual situation, a formal model is revised based on the theoretical model. Using the same economic and non-economic variables, we will conduct regression analyses and compare the regression signs and the significance of the coefficients in each country equation respectively. By doing so, we can determine the influence and the degree of sensitivity of each factor in the highly industrialized country pattern and less developed country pattern respectively. Additionally, concerning the different social demographic background of the internal migration, country specific factors are introduced into equations as dummy variables. Standard migration pattern model is established as below:
Mij = + ×C θ1 DISTij+ ×θ2 PopDensityj+ ×θ3 Gapij+ ×θ4 URj+ +υ ωj
ij
M : the number of migrants from place i to place j in t period
ij
DIST : the geographic distance between pairs of places (originality i place and
destination j ) j
ij
Gap : the relative income gap between i and j, computed as (
i j j P P P − × 100%). j
υ : the region dummy variable of receiving place j
3.3.3 Hypothesis
Six sub-hypotheses(H1-H6) are formulated corresponding to this set of six independent variables in the empirical model. Each sub-hypothesis is divided into two country
hypothesis - China (a), U.S. (b). The results of the sub-hypotheses will lead us to test the general hypothesis of this paper -“internal migration pattern shifts as when a higher economic level is attained”.
(1) geographic distance
H1a: Increasing geographic distance between pairs of provinces discourages internal migration in China.
H1b: Increasing geographic distance between pairs of states discourages internal migration in U.S..
The importance of geographic distance between original and destination places is strongly emphasized in migration studies. Increase in distance not only raises the transportation costs but also increases the uncertainty and risks of migration, because information about more distant places is generally less readily available (Greenwood, 1975). Though migration cost is the sum of pecuniary and non-pecuniary cost, in practice, the latter one is hard to measure. Thus, for simplicity and precision, I take geographic distance between pairs of places as the estimator of migration cost - based on the rational assumption that cost goes up as geographic distance increases due to transportation fee, language obstacles, information quality, etc.
(2) population density
H2a: Higher population density in the destination province attracts more in-migration in China.
Population is another essential non-economic factor that should not be missed in migration research. Generally, population size is used as a kind of proxy variable to economic opportunities caused by agglomeration economies, so people will move from less populated to more populated places (Chan, Liu 1999). Greenwood (1971) proposed that migration would be positively associated with destination population size, and negatively with origin population size. In this paper, I use population density as a relevant variable of population size. Due to the unevenness of the population distribution and big differences in land area at provincial or state level8 (especially in China), I argue that population density is a more accurate measurement of a region’s urbanization level than pure population measurement. Lower population density is more rural, and people are less concentrated. By testing the coefficient of this variable, we might discover the differences in migratory preferences between Chinese migrants and American migrants, if any. (3) relative income gap
H3a: Increasing relative income gap between destination and original provinces causes more in-migration in China.
H3b: Increasing relative income gap between destination and original states causes more in-migration in U.S..
I choose relative income gap as the proxy for economic gap between original and destination places, rather than GDP per capita, wage disparity or consumption per capita, since it directly measures the economic disparity at micro level. Though a high GDP region usually has higher income, however I argue that GDP per capita itself is not an appropriate migration motivation measurement - money opportunity stimulates people to move, not
GDP. On the other side, as the economic structure evolving, wage is no longer the sole
resource of individual income, and high consumption does not necessarily mean high income - could be caused by high life cost. Thus, wage disparity and consumption capita are not selected as economic gap measurement.
and rural population ratios vary among provinces, thus the average income level of certain province is calculated as the sum of weighted urban per capita income and weighted rural house households at province level9. Based on this, I calculate relative income gap between each pair of provinces; However, U.S. Census Bureau defines income as the amount of money people or households receive during a calendar year. Calculate the differential between two states’ median household income in 2000 year (income valued in current dollars), we obtain the relative income gap.
(4) unemployment rate
H4a: High unemployment rate of destination province discourages in-migration in China. H4b: High unemployment rate of destination state discourages in-migration in U.S..
The importance of the probability of obtaining employment on the decision to migrate has been strongly emphasized by Todaro (1969) and succeeding researches. Apart of income disparity, job-related determinant is another key economic determinant of individual decision-making which should not be missed in any migration study. In this paper, I use unemployment rate of destination place as the explanatory term, which is expected to discourage in-migration. Alternatively, we can also use unemployment rate differential between origin and destination places.
China’s unemployment rate statistics only covers urban areas. The so called registered unemployment rate in urban areas refers to the ratio of the number of the registered unemployed persons to the sum of the number of employed persons and the registered unemployed person10. U.S. Census Bureau defines unemployment as all civilians 16 years old and over are classified as unemployed if they (1) were neither "at work" nor "with a job but not at work" during the reference week, and (2) were actively looking for work during the last 4 weeks, and (3) were available to accept a job. Also included as
those civilians who did not work at all during the reference week, were waiting to be called back to a job from which they had been laid off, and were available for work except for temporary illness. Relevantly, unemployment rate is the ratio of number of unemployment
9 Income level of province N= urban population ratio* urban per capita income+ rural population ratio* per
capita net income of rural households.
10 Income level of province N= urban population ratio* urban per capita income+ rural population ratio* per
persons to total labor force.
(5) Region dummy variables
H5(a) Higher Regional Policy Preference exerts positive pull effect on China internal migration pattern;
The reason to incorporate regional development index as province dummy variable in China case is that migration is strongly related to regional development in this country. Governments directly intervene the development of regions and areas’ urbanization by entitling advantageous economic policy to attract foreign capital and priority of financial support for transport and public infrastructure construction. Thanks to the favorable policy preference and natural geographic advantages, coastal area11 became the first option of FDI (foreign direct investment) and manufacturing industry location. These economic leading areas continuously attract large flows of migrants with more available job opportunities, higher income level and huge market potential. Subsequently, due to economies agglomeration effects, these regions become even more advanced, far ahead of other areas- further widened regional development gap. Thus, I argue that the importance of regional preference policy (as well as geographical location) does matter in the Chinese population mobility pattern.
The regional preference policies are divided into four levels, corresponding to investment priority, government financial support, public transportation construction, other financial treatment etc. The highest level is CDZ (China Development Zone), assigned by 3; second level, ETDZ (Economic and Technological Development Zone), FTZ (Free Trade Zone), BECZ (Border Economic and Cooperation Zone), assigned by 2; other preference policy zones, assigned by 1; No open zone is assigned by 012.
H5(b) Sun-belt State attracts more in-migration in U.S. internal migration pattern.
State’s Sun-snow belt attribute is introduced as state dummy variables into the U.S. model. In U.S., Sun-belt refers to the stats or parts of states that lie south of the 37th degree latitude (the Snow belt: verse versa) As we known, at higher income level, demand for good living
environment13 increases, people care more about QOL (qualify of life) factors, particularly those related to climate, air and water of the residence place. . In the last half century, the movement from the cold Snow Belt to warm Sun Belt has become more widespread in U.S.. As the special census notes pointed out that “The Sunbelt now contains a disproportionate number of people relative to its share of national land area. In 2000, the Sunbelt accounted for 40 percent of the nation’s population”. Thus, the Snow towards Sun Belt has already become a national movement pattern, which can be dealt as a general
preference of Americans. In this paper, I assign value 1 to the Sun belt states, value 0 to the
Snow belt states.
(6) FDI (foreign direct investment)
H6a Higher FDI (foreign direct investment) attracts more in-migration in China.
There are numerous literature identified the positive effect of FDI on local wage rate and unemployment. David O.Cushman(1987) pointed out that FDI raises the foreign capital stock, improves labor’s productivity, and lowers the unemployment rate by its effect on foreign income. As a labor-abundant country, China has a competitive labor cost, but lack in capital, FDI directly helps to relieve domestic capital supply bottlenecks and to promote employment and economic growth (Y.Y.Kueh,1992). Introducing more capital inflow enables Chinese manufacturers to significantly increase marginal productivity, thus ready for further expansion. Further production expansion call for more labor input - FDI
creates jobs. Under such an interactive growing labor demand and supply model, only
migrants can make it happen.
Not only FDI helps to absorb surplus labor, lower unemployment rate, it also provokes a surge in local wage rate. No doubt, increasing wage disparity is another motivation stimulate migration. Solomon Polachek(2005) argued that “the wage rate in the home country is increasing in the amount of capital invested in the home country and on the amount of human capital”. Once the wage rate reaches certain level, income disparity becomes wide enough to compensate the migration cost, thus this strong economic driving force motivates low-productivity agricultural people divert to profitable industry sector. Therefore, we formulated an extended model for China based on the standard equation,
include in FDI (foreign direct investment) as an additional explanatory variable.
4. Methodology
4.1 Econometric equations
Based on the theoretical framework, detailed econometric equations are formulated as below:
US inter-state migration equation:
Mij = + ×C β1 DISTij+ ×β2 PopDensityj+ ×β3 Gapij+ ×β4 URj+ + γ ξj
China inter-province migration equation (1):
1 2 3 4
ij ij j ij j j
M = + ×C β DIST + ×β PopDensity + ×β Gap + ×β UR + +δ ζ
China inter-province migration equation (2):
1 2 3 4 5
ij ij j ij j j j
M = + ×C α DIST +α ×PopDensity + ×α Gap +α ×UR + ×α FDI + + δ ζ
4.2 Definition and measurement of variables
The independent term M is the number of migrants from place i to place j in t period. ij Where DIST is the geographic distance between pairs of places (originality i place and ij destination j ), PopDensity and j UR refer to the population density and the j
unemployment rate of the receiving place. Gap indicates the relative income gap between ij
i and j .γj is the Sun-Snow belt region dummy variable of the receiving state in the U.S. model, I assign Sun belt state as 1, Snow belt state as 0. δjdenotes the China regional preference policy index of the receiving province. Value of this variable is assigned as below:
j
δ =3: CDZ
China Development Zone
j
δ =2: ETDZ
FTZ Free Trade Zone
BECZ Border Economic and Cooperation Zone
j
δ =1: other preference policy Zones
j
δ =0: No open zone
In China inter-province migration equation (2), FDI is the Actually Used Foreign j
Direct Investment. China National Census Bureau defines Direct Investment by Foreign
Entrepreneurs as the investments inside China by foreign enterprises and economic organizations or individuals (including overseas Chinese).
4.3 Data
This empirical study will be conducted by establishing internal migration models including carefully selected variables to test China and U.S.’s 2000 population censuses’ migration data separately. Two censuses are conducted on the comparable definitions14, operation techniques and scale. U.S. and China 2000 census provide 2550 inter-state gross migration stream data and 930 inter-province gross migration stream in an OD (origin-destination) matrix form, which provides an ideal database for comparison study in terms of sufficiency and accuracy .
4.4 Model specification
Before to test the significance of the equation, several techniques would be applied to diagnose the data and the correctness of model. First, as I use large sample matrix cross-section data, heteroskedasticity might exists. Simply by plotting the least squares residuals against each explanatory variable to see if those residuals vary in a systematic
14 China Census 2000 define migrant as: an individual who on 1 November 2000 resided in a sub
patterns can provide a good estimation of heteroskedasticity. Then Goldfeld-Quandt test can help to confirm this problem in a statistical sense15. Results show that the null hypothesis of homoskedasticity should be rejected, which means there does exist heteroskedasticity. In case of this, I apply White Heteroskedasticity technique to get a better estimator with lower variance.
Second, to detect the multicollinearity problem, I have examined the correlations among the independent variables. Results show that the correlation coefficients between pairs of explanatory variables are satisfactorily lower than 0.8, which means there is no existence of harmful collinear problem.
Table 2. Correlations Matrix (1) China equation DIST INCOMEG AP LAND REGIONAL UR DIST 1.0000 -0.0110 0.2596 0.0704 0.0047 INCOMEGAP -0.0110 1.0000 -0.1471 0.3587 -0.1241 POPDENSITY 0.2596 -0.1471 1.0000 -0.2850 0.2963 REGIONAL 0.0704 0.3587 -0.2850 1.0000 -0.1683 UR 0.0047 -0.1241 0.2963 -0.1683 1.0000 China equation (2) U.S. equation
DIST INCOMEGAP LAND SUNSNOW UR
DIST 1.0000 -0.0107 0.1786 -0.0514 0.1679 INCOMEGAP -0.0107 1.0000 0.0850 -0.3602 -0.1999 POPDENSITY 0.1786 0.0850 1.0000 -0.1279 0.4040 SUNSNOW -0.0514 -0.3602 -0.1279 1.0000 0.2130 UR 0.1679 -0.1999 0.4040 0.2130 1.0000 US equation
15 GQ values of two data sets are far above than the critical value from the F-distribution with (T-K) degrees
(3) China FDI equation 16
DIST FDI INCOMEGAP LAND REGIONAL UR
DIST 1.0000 -0.0686 0.0011 0.3964 0.1227 -0.0710 FDI -0.0686 1.0000 0.3177 -0.2296 0.5131 -0.1380 INCOMEGAP 0.0011 0.3177 1.0000 -0.1215 0.3528 -0.0817 POPDENSITY 0.3964 -0.2296 -0.1215 1.0000 -0.0269 -0.2171 REGIONAL 0.1227 0.5131 0.3528 -0.0269 1.0000 -0.0271 UR -0.0710 -0.1380 -0.0817 -0.2171 -0.0271 1.0000
China FDI equation
5. Empirical results
Table 3. Cross-section estimates of OD migration model
Dependent Variable: MOVER (numbers of migrants)
Techinique: White Heteroskedasticity-Consistent Standard Errors & Covariance
U.S. equation China equation China FDI equation17
Independent variables: Coefficients t statistics Coefficients t statistics Coefficients t statistics Constant 11.02 *** 8.54 6.66 *** 5.20 6.19 *** 5.03 DIST -2.29 *** -10.28 -3.80 *** -5.36 -3.76 *** -5.82 INCOMEGAP 5.45 *** 4.44 1.53 *** 5.03 0.90 ** 2.62 POPDENSITY -5.30 *** -3.99 0.02 *** 3.80 0.12 *** 4.18 SUNSNOW 8.03 *** 7.14 REGIONAL 4.82 *** 3.87 0.43 0.82 UR 5.73 0.17 -204.33 *** -3.78 -76.52 ** -2.10 FDI 0.27 *** 3.22 R-squared 0.08 0.11 0.22 Adjusted R-squared 0.08 0.10 0.22 Durbin-Watson stat 1.59 1.89 1.87 F-statistic 42.35 21.78 38.21 Sample Size: 2550 930 812 Included observations 2548 930 812 *** Significance level: 1% ** Significance level: 5% * Significance level: 10%
In general, the results are consistent with the theoretical prediction:
In both equations, the estimated coefficients of DIST are significant at 1% and have
correct signs. The results support the hypothesis (H1a) and (H1b) - in both countries internal migration is discouraged by geographic distance.
!
The population density is a key variable to reveal the country migration patterns. As I
expected, population density of receiving area shows different signs in two equations at 1%. The negative sign of the population density in U.S. equation strongly supports the fact that the moving pattern of U.S. is down-the-size-hierarchy. In postindustrial societies, more disposable income allows people to pay more attention to QOL (quality of life factors). It might be economically feasible as well as desirable to escape from
17Two provinces (Tibet, Ningxia) are excluded from the FDI extended equation due to data
heavily congested high population density areas in favor of smaller town or even rural living environments. The so called fourth stage - downwards move from more to less populated areas mobility pattern came up. Thus, hypothesis (H2b) is rejected - higher population has averse effect on attracting in-migration in U.S..
On the contrast, hypothesis (H2a) is supported in China model. The positive sign in China equation helps to confirm that the internal migration pattern is still conventional upward hierarchy. The rural-urban migration has been growing spectacularly in China in the last two decades. The dominance of rural-to-urban migration can be also observed from the authoritative statistics records. According to the official resources from the National Statistical Bureau, in the last two decades, the number of rural-urban migrants increased from merely two million in the mid-1980s to 70 million in the mide-1990s, and the flow continued to grow to 94 million in 2002. Thus, we can conclude that economies of agglomeration is a key factor in determining developing country‘s upward hierarchy internal migration pattern, but it starts to exert certain kind of adverse effect on internal migration in more developed country. However, it highly depends on the economic development level that has been achieved in that country.
"#
Income gap show positive signs at 1% significance level- large income gap between
sending and receiving area does exert a strong pull effect on in-migration. Thus, both of
hypothesis (H3a) and (H3b) are confirmed.
$&%
Unemployment rate show positive sign in U.S. equation, but insignificantly. This result
can be explained as job is relatively a less important determinant in the migration decision for Americans. In fact, as I explained before, American are less intended to move for job reasons. On the contrast, unemployment of the destination exerts a strong down-ward effect on in migration flow in Chinese case. High unemployment rate discourage people to move in, or equally to say that low unemployment rate, more job opportunities are the main motivations to move. Thus, hypothesis (H4a) is supported in China model at 1% significance level.
'
The signs of two dummy variables are also in line with my prediction. Americans move
China, advantageous regional preference policy cause manufacturing sectors cluster highly concentrated in coastal regions, attracting migration flows in for better economic opportunities which is not realizable in their originality. (H5b) is confirmed.
( %
The estimated coefficient of FDI is significantly positive at 1% level. Thus, H6a Higher FDI (foreign direct investment) attracts more in-migration in China is correct.
6. Conclusion
This paper investigated the key factors laid behind the current China and U.S. internal migration pattern with census 2000 data. The common findings in two country patterns can be summarized as:
First, other things being equal, distance is always a hindrance to move, since the relevant migration costs increase with distance, also increase the related pecuniary and non pecuniary cost to move.
Second, no matter to highly industrialized country’s migrants or less developed country‘s migrants, income gap is the a strong pull-effect, motivate people to move for better economic opportunities.
The paper also reveals a large number of interesting findings about the differences in two country patterns:
U.S. migrants and China migrants show different responses to the unemployment rate of destination place. For Americans, job is no longer a crucial reason to move, income disposal income allow people to pay more attention to quality of life factors.
between the pull and push effects between originality and destination places. However, the importance of this classic golden migration rule loses the glamour if we fail to recognize that (1) different types of economies have different internal migration driving forces; (2) even the same factors do not necessarily exert same influences on migrant patterns of highly industrialized country and less developed country.
As comparative study unraveled that income gap, population agglomeration, regional development and job opportunities are the pull factors in China internal migration pattern. In U.S. case, higher income level, quality of life factors are the key considerations of migration decision, unemployment rate of the destination are less important.
Based on, the general hypothesis “internal migration pattern shifts as when a higher economic level is attained” is supported. Less developed country’s internal migration are more primitive, conversely, highly developed country has already stepped into the fourth stage.
There are a few limitations in this paper. First, as I mentioned that the population movements depend on many complex variables, however this study mainly focus on main economic and non-economic variables, but does not include qualitative migration
determinants or individual preference factors. Second, migration restriction is one
References
Brian J.L. Berry .(1993). Transnational Urban ward Migration, 1830-1980. Annuls of the
Association of American Geographer 389-405
Cai Fang. (2000). The Invisible Hand and Visible Feet: Internal Migration in China. Working Paper Series ( 5)
Cat Moody (2006). Migration and Economic Growth: A 21st Century Perspective. New
Zealand Treasury working paper 06/02
C.Cindy Fan. (2004). Internprovincial Migration, Population Redistribution, and Regional development in China: 1990 and 2000 Census Comparisons. The Professional
Geographer: Blackwell Publishing
D.A.Plane, C.J.Henrie, and M.J.Perry. (2005). Migration up and down the urban hierarchy and across the life course. The National Academy of Sciences of the USA
Dapeng Hu. (2000). Trade, rural-urban migration, and regional income disparity in developing countries: “A Spatial General Equilibrium Model Inspired By The Case Of China, Regional Science and Urban Economics (32): 311-338
David O.Cushman(1987). The Effects of Real Wages and Labor Productivity on Foreign Direct Investment. Southern Economic Journal 54(1): 174-185
Dudley Baines .(1998). The Economics of Migration-Nineteenth-Century Britain, Refresh 27
E. G. Raven stein. (1885). The Laws of Migration. Journal of the statistical society. Frank Hobbs. (2002). Demographic Trends in the 20th Century. Census 2000 Special Reports.
U.S. Department of Commerce Economics and Statistics Administration, U.S. Census Bureau
Franklin D.Wilson’s (1988). Aspects of Migration in and Advanced Industrial Society. American Sociological Review. 53(1): 113-126
Gene Laber, Richard X.Chase. (1971). Interprovincial Migration in Canada as a Human Capital Decision. The Journal of Political Economy 79: 795-804
Huang and Frank N. (2003). China Migration Country Study. Series of regional conference
Population Survey. U. S. Census Bureau, Special Studies
Jason P.Scharchter, Rachel S.Franklin, and Marc J.Perry. (2003).Migration and
Geographic Mobility in Metropolitan and Nonmetropolitan America: 1995 to 2000.
Census 2000 Special Reports. U.S. Department of Commerce, Economics and
Statistical Administration, U.S. Census Bureau
Joochui.Ki. (1980). Characteristics of Migrants Within the Framework of Current
Migration Direction in the United States: Some Evidence From Micro-Data Anaylysis.
Policy Sciences: 355-370
Kam Wing Chan.(1999). Hukou and Non-hukou Migrations in China: Comparisons and Contrasts. International Journal of Population Geography. (5): 425-448
Kevin McQuillan. (1980). Economic Factors and Internal Migration: The Case of Nineteenth-Century England. Social Science History 4 (4): 479-499
Kevin Honglin ZHANG a,Shunfeng SONG. (2003).Rural – urban migration and urbanization in China: Evidence from time-series and cross-section analyses. China
Economic Review (14): 386 – 400
Kris James Mitchener and Jan W.McLean. (1999). U.S. Regional Growth and Convergence, 1880-1980. The Journal of Economic History 59(4): 1016-1042 Lorene Yap. (1976). Internal Migration and Economic Development in Brazil. The
Quarterly Journal of Economics90 (1):119-137
Lorene Yap. (1976). Internal Migration and Economic Development in Brazil. The
Quarterly Journal of Economics90 (1):119-137
Marc.J.Perry. (2003). State-to-State Migration Flows: 1995-2000. Census 2000 Special
Reports. U.S. Department of Commerce, Economics and Statistical Administration,
U.S. Census Bureau
Michael P.Todaro. (1969). A Model of Labor Migration and Urban Unemployment in Less Developed Countries. The American Economic Review 59: 138-148
Michael.J.Greenwood. (1975). Research on Internal Migration in the United States: A Survey. Journal of Economic Literature 13: 397-433
Nong Zhu. (2002). The Impacts of Income Gaps on Migration Decisions in China. China
Michael C.Seeborg.(2000).The new rural-urban labor mobility in China: Causes and implications. Journal of Socio-Economics (29):39–56
Rachel S.Franklin. (2003). Domestic Migration Across Regions, Division, and States: 1995 to 2000. Census 2000 Special Reports. U.S. Department of Commerce, Economics and Statistical Administration, U.S. Census Bureau
Robert E.Land and Kristopher M.Rengert. (2001). The Hot and Cold Sunbelts: Comparing State Growth Rates, 1950-2000. Fannie Mae Foundation Census Notes 02
Sylvie Démurger, Jeffrey D.Sachs. (2002).Geography, Economic Policy, and Regional Development in China. Asian Economic Papers
William Petersen (1958). Internal Migration and Economic Development in Northern America. Annals of the American Academy of Political and Social Science 316: 52-59 Y.Y.Kueh (1992). Foreign Investment and Economic Change in China. The China
Quarterly (131): 637-690
Data Resources
U.S. China
Migration data State of Residence in 2000 for the Population 5 Years and Over by State of Residence in 1995 2000, special tabulation)(U. S. Census Bureau, Census
Tabulation on the 2000
Population Census of the People’ s republic of China. Beijing: Zhongguo tongji chubanshe (China Statistic Press) Geographic distance interstate distance:
www.50states.com interprovincial distance: thegeographic distance between
each two provinces’ capital cities by a digital map
Population density Caculated by State Population
Data and State Land Caculated by Province Data andProvice Land
Income data U.S. Census Bureau, Housing and
Household Economic Statistics Division 2005(Historical Income Tables)
China Statistical Yearbook 2001
Unemployme rate U.S. Census Bureau, Census
2000 Summary File China Statistical Yearbook - 2001
Foreign direct investment China Statistical Yearbook - 2001
