Regional Disparity in China

von Wang Xiaolu, Gang Fan

Tendency and the Influential Factors

Wang Xiaolu und Fan Gang vom Nationalen Economic Research Institute der China Reform Foundation haben in einer von der Konrad-Adenauer-Stiftung geförderten Studie die regionalen Disparitäten in der Volksrepublik China anhand verschiedener Faktoren (Kapitalfluss, Arbeitproduktivität u.a.) untersucht. Der Beitrag fasst die Ergebnisse der Studie zusammen. KAS-Schriftenreihe China NR. 26, Beijing, Englisch, 37 Seiten.


General Report for the Research Project

Regional Disparity and Ways to Reduce It (1)

1. Changing in Regional Disparity in the Past 20 Years and the Analytical Framework

Economic disparities among the east coastal, central and west regions existed before the economic reform. In general, the level of economic development and per capita income in the coastal areas and part of the central provinces were higher than that in the west provinces. To reduce the gaps, the central government of China in the 1950s through 1970s practiced a fiscal transfer policy to balance the regional income discrepancies. In the 1950s, a number of large industrial projects were located in the west provinces; later under the concern of potential war, a strategic plan known as the “third front-line construction” to shift the west provinces into major industrial centers, a large amount of investment was poured into this region. It did reduce gaps in terms of average income, but it did not work on closing the gaps between the east and west regions in terms of production efficiency. Consequently, as a matter of fact, the economic growth of the west region relied heavily on the continuation of transfer of resource through fiscal channels to keep up in average income.

Since 1978, practices of the economic reform policies called for higher efficiency in business performances in local government and enterprises, by authorizing them autonomous management power and better economic incentive, therefore the central government cut down the rate in local revenue of the east provinces to be handed over to the Finance Ministry. It resulted in reduction of the magnitude of income transfer to the west region as a consequence on the one hand; while on the other, it gave a big push to the economic development in the east region. Among the beneficiaries, Guangdong Province in particular was the leading area, which enjoyed a record growth over the 1980s and 1990s. In the meantime, a progressive reform toward market economy was succeeding in almost all the provinces, resulted in a speeding economic growth and higher personal income on average.

Despite the fact that all the regions in China had higher rate in economic growth than the pre-reform time (prior to 1978), however in the coastal province in east China, the rates of economic growth in the last 20 years were much higher than that in the west region, and as a result, the regional gap in income was widened. In the 1990s in particular, the widening in regional income was a notable phenomenon. Table 1 in the following shows the annual growth rates of GDP by the east, central and west regions from two sources(2). There are discrepancies in the growth rates by the two sources, but it did not change the picture of disparity we are trying to illustrate. According the first source, in the 1980s the difference between the east and the central-and-west regions in GDP growth was around one percentage points, in the 1990s it was 2-3 percentage points. According to the second source, the differences were less than one percentage point in the 1980s, around two percentage points in the 1990s.

Table 1. GDP Growth Rate by Region (1981-2001)

Region1981-1990Difference Compare to the East Region1991-2001Difference Compare to the East Region

State Statistic Bureau

East 9.9812.94


West9.12-0.86 9.63-3.31

Meng and Wang


Center7.85-0.95 9.64-2.04

West8.71-0.09 9.69-1.99

Note: The computation by Meng and Wang was based on weighted average of provincial GDP (current price), deflated by a compound index which include CPI, industrial output price index, farm-gate agricultural price index, and fixed capital investment price index.

Source: Wang Xiaolu (2001); State Statistical Bureau (1999a, 2001, 2002).

The sizes of economy by the three regions (east, central and west) in current price GDP were listed in Table 2, in 1980, 1990 and 2000 respectively. It can be seen that in 1980 the economy of the west region was 40% of that of the east, in 1990 it went down to 39%, and further down to 30% in 2000 (by GDP in current price). In the case of the central region comparing to the east, it was 59% in 1980, 54% in 1990, and 45% in 2000. It says that the gaps widened faster in the 1990s, although disparities between the central and east regions started in the 1980s.

The regional discrepancies in economic growth lead to also a widening gap in per capita GDP and disposable income (or in average income in the case of rural residents). In 1980, the per capita GDP for the central and west regions were 65% and 53% as of that in the east region; in 2000, they became 53% and 41%, respectively (see table 3). Thanks to the nationwide transfer of income, the gaps in per capita disposable income was narrower compared with that in per capita GDP, however it was also widened in the past. In 1980, the proportion of average income in the central and west regions measured by that of the east was 78% and 70%, respectively; in 2000 it went down to 62% and 54%.

Table 2. GDP and Population of the Three Regions (Current Price, in 100 million Y)

1980 1990 19952000 2000

GDPPopulation (100 million)

East2202 9537 32344 55690 4.91

Center1301 5160 15075 24865 4.16

West871 3750 10471 16655 3.55

Comparison (East = 100%)Center/East59%54%47%45%85%


Source: State Statistic Bureau, 1999a, 2001.

Table 3. Per Capita GDP and Comparison, Three Regions

(Current Price, Y/Person)

1980 1990 19952000


East598 2240 7247 11334

Center391 1338 3708 5982

West308 1156 3035 4687

Comparison (East = 100%)Center/East65%60%51%53%


Source: State Statistic Bureau, 1999a, 2001.

The widening of income gap over the regions presents itself more clearly in the contrast of rural income gaps. The income ratio was 1:0.85 by east urban residents against west urban residents in 1980, which became 1:0.72 in the year 2000, 12 percentage points down; however in the case of east to west rural residents, it went from 0.70 to 0.53, a 17 percentage decrease over the same period of time.

Various studies on economic growth have covered all kinds of economies on the gaps in development and income. The neoclassical growth theory implies that, the low income economies tend to develop relatively faster in general, therefore the income gap between high and low income economies will narrow down (Solow, 1956; Swan, 1956), which is called “convergence theory”. On the other hand, the endogenous growth theory shows that the human capital has significant contribution to economic development, and developed countries usually have higher human capital stock compared with developing countries. This implies that less developed economies will not be able to catch up (Romer, 1986; Locas, 1988). Empirical studies using worldwide country data do not support the “convergence” theory in general. However other studies on grouped and categorized economies did show a tendency of “convergence”, such as the case among the Eastern Asian countries and within the OECD countries. Further studies show that if some variables such as human capital and saving rates are controlled in the analysis, then the “convergence” can be observed. Therefore these factors are considered to cause regional economic gap. This was called “club convergence” and “conditional convergence” theory (Barro and Sala-i-Martin, 1995).

In China, these theories of growth have been used in the study of regional economic gaps. According to Cai Fang and Du Yang (2000), “convergence” was observed within, respectively, the east, central and west regions in China on provincial basis; “conditional converge” was also shown across these three regions once the initial human capital stock and the rate of investment were controlled. Cai and Du’s study reveals that there existed internal and external factors to be the causes of the gaps between the regions in China. A number of factors were found to be responsible, but the questions are: Which are the main ones? How do these conditions change? And what causes these changes? It is critical to answer these questions in order to go into depth of the study on regional gaps. At least, we have seen that many conditions contribute to the regional gaps, such as migration and capital flow across the regions, the distribution of human capital over the regions, geographical location, urbanization, rural industrialization, infrastructure, institutions and policies, all are potential causes to shift the regional economic gaps. Only when these factors are thoroughly studied can we build a framework on the evolution of regional economic disparities.

In the following we will do a cross-regional and time-series analysis on the following factors: the distribution and flow of capital, labor and human capital, the difference in institution and policy reform, urbanization and rural industrialization, and changing regional comparative advantages. The positive and negative impact of these factors on the changing regional gaps will be discussed. We are trying to draw an overallpicture on the formation and changing of regional economic gaps, and to discuss on ways of narrowing these gaps. This study is based on several subject reports of the research project, “Regional Economic Disparity: Causes and Ways to Reduce It”.

2. Regional Capital Flow and Distribution of Foreign Investment

Allocation of capital is a main factor in regional economic development. Therefore the direction of capital flow across the regions will heavily impact the evolution of regional gaps. Prior to the economic reform, capital investment was totally determined by the economic planning under the central government. It was a non-market procedure. However since 1978, in addition to the fiscal appropriations and the allocation of bank funds that lead by central policies, capital investment has been led by the market. The regional distribution of foreign investment, if considered independently, has been influenced not only by the market but also the regional preferential policies. All these conditions, i.e., fiscal system, banking policy, market environment, preferential policy for FDI, are changing over time and affected the flow of capital across the provincial boundaries. The problems are: What is the general tendency of regional capital flow? Have there been any major changes in the picture? What is the main feature in its current movement? There has been a lack of overall study on this issue. We will analyze on the flow of capital from the following perspectives to draw a general picture on this matter.

A. Fixed Capital Investment: Distribution across the Regions

We keep the general term “regional” as defined above, i.e., to group the provinces of the whole country into three big geographical regions: coastal areas in the east, the central, and the west regions. Total fixed capital investment among the three regions since 1980 is listed in Table 4.

Table 4. Changes in Investment by Region (Current Price)


Total Capital Investment(100 million Y)

East391 1275 2345 7456 12226 14915 18752

Center238 704 1048 2494 4097 5454 7034

West189 458 792 2027 31674261 6111

Comparison (East=100%)



Investment ratio(Investment/GDP)




Source: The State Statistic Bureau, 1999a, 2001.

From the lower half of Table 4, we can see the ratio of capital investment, i.e., the proportion of capital investment in local GDP, in the west region in 1980 was higher than in the east and central regions (22% vs. 18%), despite the fact that the quantity of the capital invested to the west region was only 45% compared to the quantity invested in the east region. It implies a national policy favoring the west in capital investment in the 1980s, which draws the main feature of the economic policy.

In the last two decades, the ratio of investment in all the three regions has increased; however, relatively speaking, the ratio of capital investment was converging in the mid 1980s across the regions, but diverging in the 1990s. By the year 1990, the regional gap in the ratio of investment widened, lower in the central and west regions with the central region the lowest. The gaps had been growing up to the latter part of the 1990s; only by the end of the 1990s this tendency started to loose its momentum. In the year 2000, the investment ratio was higher in the west region than in the east once again. The implication was that over the period of 1990s, the east region was in a superior position in acquiring capital. It significantly contributed to the widening of regional gaps.

In order to dig deeper into the causes of the changes in the flow, we look into the main categories of investment across the regions in the following.

B. Transfer of Government Budgetary Funds

Transfer of government budgetary funds over the regions is an important source in local economic development. However, data for these are incomplete. In 2001, the amount of funds in total budgetary expenditure under the item “aids to support the underdeveloped areas” was only a limited Y13.4 billion nationwide (State Statistic Bureau, 2002). This number was only a small fraction of the total transfer of fiscal resources among regions. There has been large refund of tax revenue from the central to the local government, which is not proportional to the tax revenue collected from different provinces. The distribution of this refund across the regions is determined to a great deal by regional development policies, as well as a continuation of the previous structure of resource distribution. It is in fact carrying out a role of transfer payment of financial funds. In order to compute the rough amount of the transfer (see also Guo and Wang, 2002), we define the following:

i.Transfer of budgetary revenue across the provinces/regions is carried out mainly through the central government tax refund to provinces(3) ;

ii.The amount of tax refund from the central government equals the difference of budgetary expenditure and revenue of the local government in question; it is called “gross transfer payment” in this report(4) ;

iii. Due to the lack of data on the sources of the central government tax revenue, we assume the source of “gross transfer payment” that the central government collected from each province is proportional to its GDP;

iv.A province gets refund from the central government which can be greater o r smaller than the amount of revenue it handed in to the central government, which is defined as “net transfer payment” the province gets (or pays).

Based on the approximations defined above, the author computed the amount of net transfer payment the provinces acquired and its proportion to local GDP since 1990, and then grouped them into the three geographic regions. The net amount after balancing out the negatives is the net transfer payment by region. Table 5 shows the proportion of net transfer payment to local GDP. Details are in Table 10 and 11.

It can be seen in Table 5 that the net transfer payment favored the central and the west regions throughout the 1990s, especially the west; the trend was diminishing toward the end of the decade; however, it resumed the momentum obviously in 2000. It is easy to see that the diminishing trend in the 1990s was due to the decentralization of fiscal power down to the provinces; the resuming of the tendency favoring the west region was in the end of the 1990s thanks to the West Development strategy. We can reason that the intensity of transfer payment was stronger in the 1980s than in the 1990s, and even stronger during the pre-reform period (before 1978).

In the 1990s, transfer payment through the central government took one percent out of the GDP from the east region on average, which was up to nearly two percent in 2000; in the west region, income through transfer payment accounted for 2-3 percent to its local GDP on average, however in 2000 it was nearly 5 percent. The quantity of total transfer payment in 2000 was over ¥ 100 billion.

If consider the local financial revenue as the measuring scale, the transfer payment took a significant portion in the local budget. It can be seen in Table 6 that in the year 2000, in the east region, the outgoing transfer payment equaled to 25 percent of local revenue, while the amount of transfer payment the west region received was as much as 70 percent of its local sourced annual revenue.

Table 5. Net Transfer Payment across the Regions as percentage of local GDP (%)

Ratio of Net Transfer Payment to GDP(%)





Source: Computed on data from Guo and Wang (2000) and State Statistic Bureau (2002b, 1999a).

Table 6. Comparison of Net Transfer Payment across the Region (100 million Y)

Local Revenue(1)Local Expenditure(2)Gross Transfer Payment (2)-(1)(3)Net Transfer Payment(4)Ratio (in Expend.)(4)/(2)(5)

East526439771286 -1005 -25.3%

Center1332 2575 1243 220 16.5%

West1127 2598 1471 786 69.7%

Source: Table 5 and State Statistic Bureau (2002b, 1999a).

The following issues should be noticed. The transfer payment from central to local government in China covers multiple categories of expenditures on one hand: not only capital investment, but also public education and health, social welfare and others; on the other hand, the number here does no cover all channels of capital flow through fiscal funding, since there are some other direct investment projects and production sponsored and maintained by the central government not going through the fiscal refund. However, statistic analysis has shown that budget-sourced local government investment is closely related to the transfer payment. Provincial data of 1998-2000 show the correlation coefficient of 0.528 between the transfer payment (Y) to local capital investment expenditure (X). A panel data regression (fixed effect model) gave an estimated parameter of (X) to (Y) by 0.984 (t=4.437).

Therefore we have reason to consider that the main part of transfer payment was used by local government in investment, and we approximately take the net transfer payment as cross-regional capital flow through the fiscal channel(5) . It is used to indicate the direction, the changes and the rough sizes of capital flow through the fiscal channel.

C. Cross-Regional Transfer of Bank Funds

The band funds are transferring across the regions when a bank draws savings from one area but loans the money to other places. There are several channels through which the bank funds transfer: the headquarter moves funds among its branches, the People’s Bank (central bank) re-credit to the commercial banks, and the cross-region credit and debt among commercial banks, etc. In the calculation of bank fund cross-regional transfer, we may take the positive (or negative) balance of credit funds of banks (that is, year-end loans subtracting year-end deposit) at the provincial level to indicate the cumulated inflow (or outflow) of bank funds, and take the annual change in balance of credit funds to indicate the annual inflow (or outflow). However, the data of the banks show that, positive credit balance presented almost throughout the 1990s in the central and west regions, and also presented in the first half of the 1990s in the east region. Opposite situation took place only in the latte half of 1990s in the east region. At the same time, the overall credit balance as a national total turned from positive to negative. Obviously it was influenced by the macroeconomic situation and changing in monetary policy for the whole country.

To exclude the impact of these irrelevant factors from the analysis, the author defines a Standard Balance of Credit Funds by distributing the total balance of credit into the three regions according to their share in national total deposit as the weights. This standard balance is considered irrelevant to our analysis, which will be deducted from the real balance of regional record. The net balance thus computed is defined as the accumulated real cross-regional transfer of bank funds, we will call it “transfer of credit”. The annual increment in “transfer of credit” represents annual transfer of bank funds across the regional boundaries. Table 7 indicates the annual transfer of state bank funds as percentage of local GDP of the three regions; regional total is acquired as the summation of provinces(6). Positive and negative transfer of credit among provinces within each region has been cancelled mutually.

Table 7. Transfer of State Bank Funds and the Ratio to Regional GDP

1991 1993 1995 1997 1999 2000*

Incremental Credit Transfer(Y 100 million)

East-339 -243 -757 -734 -1044 317

Center239 145 411 592 1242 -556

West100 98 345 142 -199 240

Incremental Credit Transfer /GDP(%)




  • In the year 2000, the item National Development Bank loans within national total bank credit was split down to the three regions based on China Financial Almanac 2001.
Source: Guo and Wang (2002), State Statistic Bureau 2002b, 1999a, China Financial Almanac 2001.

From Table 7 it can be seen that, in the 1990s, the direction of bank funds movement was mainly from the east to the centerand west. The total amount in 1999 was Y 100 billion. This is a sharp contrast to the foreign investment directed by the market towards the east coastal areas. The transfer of net credit to the central region was strong, which accounted for 20-30 percent of total incremental bank loans in most years, and in 1999 it was equal to 5.5 percent of local GDP of that region. Possible reasons are the clustered SOEs (State Owned Enterprises) in this area, which absorb a lot of bank loans, and as main production area of food grains, the government purchasing agencies use large amount of bank funds. The amount of bank funds moving to the west has been smaller than to the central region, however it started to change in the year 2000. In the meantime, the negative transfer of bank credit into the east shifted into positive, totally contrary to the situation in the center. It seems to demonstrate that the movement of bank funds did not favor the west region as the “West Development” strategy tried to stimulate; or in other words, at the time when the central government is pushing the west region by non-commercial loans, the commercial loans favor the eastern enterprises. It may be caused by big amount of non-performing loans cumulated over the years, the commercial banks become more cautious in loaning money to high-risk areas. This can also be seen in the relative change of total deposits and loans of the banks. In the recent years, the increase in bank loan has been significantly slower than the increase in bank deposit .

The above analysis shows that the cross-region transfer of bank funds in the 1990s was mainly led by government policies and the institutional constraints, rather than by market.

D. Foreign Investment

Growth of foreign investment, mainly direct investment, accelerated dramatically since 1992. By 1996 the total amount exceeded $40 billion. Throughout the 1990s, 85% of the foreign investment poured into the east region (see Table 8), which continued up to the year 2000 and was non-responsive to the calling of “West Development”. Foreign investment accounted for over 5 percent in local GDP in the east region based on official exchange rate, and the highest was eight percent in the mid-1990s. Needless to say, the foreign investment is a very important factor in the widening of regional gaps. According to Wu Jian (2001), foreign investment is the second most influential factor contributing to regional gaps, only next to domestic investment in its regional difference in terms of quantity and efficiency.

Table 8. Regional Distribution of Foreign Investment

199019931995199720002000 Nationwide

Proportion(%,Total=100)($100 million)




Ratio to GDP (%,GDP=100,at Official Exchange Rate)




Source: State Statistic Bureau, multiple years. Data of 1990 covers only FDI.

There are two main reasons causing the concentration of foreign investment into the east region. First is the better physical condition for investment in the east region. Expected return from the east region is greater than in the west region. According to Wu Jian (2001), the efficiency ratio in the east, central and west regions were 1.517, 1.305 and 1.174, respectively. In addition to this strong attraction, there have been favorable policies in the four special economic zones and 14 open trading coast cities, which include lower tax rates and other privileges. The coverage of these favorable polices were expanded to a far broader area of the east coast region in later years. No doubt these policies are responsible to the widening of regional gaps in terms of foreign investment.

However, we can expect some changes in the near future. First, the favorable policy enjoyed by foreign investors in the east region will be eventually replaced by unified national treatment. Second, new policies favoring the west region along with the “West Development” strategy have been issued. Third, according to Cai Fang and Wang Dewen (2002), the marginal rate of return to capital in the East was twice as high as in the West in 1978, whereasv it has been close to convergence after 20 years. Under the market condition of perfect factor mobility, the convergence of marginal rate of return is consequent in the long run. Forth, in the central and west regions, the infrastructure and institutional environment is improving over time, which will lift the rate of return to capital. All of these will contribute to ease off the regional gaps in economic development.

E. Capital Market Financing

It is difficult to discuss capital market financing on its geographical sources due to lack of data. If computed by the outlet of the fund, the east region has been far ahead of the other two regions. However, the ratio of market financing to local GDP shows a different picture. The east region was dominating the central and west regions from 1991-1997 most of the time, while by the end of decade the gaps were closing down (Table 9). If we assume that the source of capital market financing is in proportion to local GDP, then we can say roughly that the capital market was conducting money flow from the west to the east in the early and mid 1990s, but this trend has basically terminated by the end of 1990s.

The capital market in China (bonds and stocks) has limited financing capacity if compared to other channels. If it did contribute to the regional gaps in the 1990s, the share is also insignificant. In addition, it should be pointed out that regional distribution of capital market financing is not totally determined by the market; to a great extent it is determined by the policy that rules and administers the listing of companies in capital market. It is therefore very likely that the changes in the trend revealed in Table 9 is due to launching of “West Development” strategy.

Table 9. Ratios of Capital Market Financing to GDP (%)


Capital Market Financing/ GDP(%,Regional GDP=100)Total(Y100 million)




Source: Guo and Wang (2002) and State Statistic Bureau (2001 and 1999a).

F. Capital Flow by Other Channels across Regions

In addition to the three main domestic channels, i.e., the fiscaltransfer, bank distribution and capital market reallocation, there are other informal or private channels through which capital flows across regions. It includes private and enterprise direct investment and lending, central government investment not via local budget, and bank loans that is excluded from the provincial statistics, etc. The sum of all these items in addition to the mentioned three major channels will give a general picture of cross-regional capital flow. Unfortunately statistical data of these items are either not exist or not region-specific; therefore the computation of capital flow through these channels cannot be done directly. The approach we adopted here is to estimate the total cross-region capital flow, and then subtract the amount of capital flow through fiscal, bank, and capital market channels from the total. The residual indicates the amount of capital flow through other channels.

G. Estimation of Overall Cross-Region Capital Flow

The first of the two approaches adopted in this section is to follow macro economic principle that net inflow of capital to, or outflow from, a province should be approximately equal to the difference between its investment and saving. There is no ready statistical information about overall saving rate by provinces; we define a weighted saving rate by the rates and the proportion of rural and urban residents in each province, to be used as the general provincial saving rate. Subtracting the savings from total fixed capital investment and distribution of errors, we reach an approximation of cross-region capital flow.

An additional computation is adopted to verify the outcome from the approach stated above. In China’s GDP accounting by expenditure, the GDP by province is broken down into the three general categories: final consumption, capital formation, and net export of goods and services. They are represented by C, I and X, respectively in the following (remember, “export” means trade across the provincial boundaries in this context). When C+I is greater (smaller) than GDP, then X is smaller (greater) than zero. If we can assume the final consumption C is not related to cross-provincial capital flow, then the formation I will include inflow (or outflow) of capital, the amount of which equals the negative net “export”. Taking out the net value of export to foreign countries, we will get an approximation of domestic net capital inflow (outflow) of the provinces. A simple calculation then will reach the cross-regional capital flow of the three geographic regions.

Subtract the capital flow through government budget, banking system and capital market from the computed cross-region capital flow stated above; the balance could be taken as the total amount of capital flow through “other” channels including the private ones. The sum of this estimation with foreign investment is the presentation of the distribution of increment of capital across the regions. Tables 10 and 11 display the computation output of the two approaches, respectively. Tables 12 and 13 are the ratios of capital flow to regional GDP (including foreign investment).

Table 10. Overall Cross-Regional Capital Flow

(Based on balance of investment and savings, Y 100 million)

East1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Cross-Region Capital Flow9 13 -151 -300 -408 -709 -586 -904 -595 -1227

Net Transfer Payment-173 -219 -260 -292 -329 -349 -380 -470 -680 -1005

Transfer of bank funds-339 -429 -243 -768 -757 -1064 -734 -105 -1044 1815

Capital Market Cross- Region Financing 5 31 74 27 9 12 100 -15 5

Other Capital Flow517 630 279 733 668 692 428 -313 1124 -2036

Foreign Investment216 544 1325 2446 2670 3010 3237 3349 2968 3022

Total Capital Inflow226 557 1175 2146 2261 2302 2652 2445 2373 1795

Center1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Cross-Region Capital Flow-114 -153 -199 -332 -375 -681 -519 -939 -567

Net Transfer Payment66 88 130 54 27 30 10 41 129 220

Transfer of bank funds239 274 145 419 411 571 592 169 1242 -1323

Capital Market Cross- Region Financing -3 -22 -37 -14 -15 2 -83 14 -5

Other Capital Flow-302 -454 -392 -658 -755 -978 -1199 -744 -2306 536

Foreign Investment10 41 135 222 282 325 396 364 311 305

Total Capital Inflow10 -73 -18 23 -50 -49 -285 -156 -628 -262

West1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Cross-Region Capital Flow112 101 303 499 740 1083 1267 1423 1534 1794

Net Transfer Payment107 131 130 238 303 319 370 429 551 786

Transfer of bank funds100 156 98 348 345 493 142 -64 -199 -491

Capital Market Cross- Region Financing -2 -9 -37 -13 6 -15 -17 1 0

Other Capital Flow-94 -177 113 -75 86 286 771 1057 1182 1500

Foreign Investment9 22 115 199 157 146 211 195 152 154

Total Capital Inflow121 122 418 699 897 1230 1478 1618 1686 1947

Source: Source: Guo and Wang (2002) and State Statistic Bureau (2001 and 1999a).

Table 11. Overall Cross-Regional Capital Flow

(Based on domestic cross-regional balance of trade, Y 100 million)

East1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Cross-Region Capital Flow-471 -614 -584 -157 -672 -379 -343 -954 -1626

Net Transfer Payment-173 -219 -260 -292 -329 -349 -380 -470 -680 -1005

Transfer of bank funds-339 -429 -243 -768 -757 -1064 -734 -105 -1044 1815

Capital Market Cross- Region Financing 5 31 74 27 9 12 100 -15 5

Other Capital Flow146 -185 449 920 729 635 248 765 -2435

Foreign Investment216 544 1325 2446 2670 3010 3237 3349 2968 3022

Total Capital Inflow73 711 1862 2513 2339 2859 3006 2014 1396

Center1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Cross-Region Capital Flow14 -1 107 -365 126 -236 -405 -66 16

Net Transfer Payment66 88 130 54 27 30 10 41 129 220

Transfer of bank funds239 274 145 419 411 571 592 169 1242 -1323

Capital Market Cross- Region Financing -3 -22 -37 -14 -15 2 -83 14 -5

Other Capital Flow-326 -240 -352 -788 -477 -754 -629 -1433 1120

Foreign Investment10 41 135 222 282 325 396 364 311 305

Total Capital Inflow55 134 329 -83 451 160 -41 245 322

West1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

Cross-Region Capital Flow457 616 477 522 546 615 747 1020 1609

Net Transfer Payment107 131 130 238 303 319 370 429 551 786

Transfer of bank funds100 156 98 348 345 493 142 -64 -199 -491

Capital Market Cross- Region Financing -2 -9 -37 -13 6 -15 -17 1 0

Other Capital Flow180 426 -97 -132 -252 119 381 668 1315

Foreign Investment9 22 115 199 157 146 211 195 152 154

Total Capital Inflow479 731 676 679 692 826 942 1172 1763

Source: Source: Guo and Wang (2002) and State Statistic Bureau (2001 and 1999a).

Table 12. Ratio of Regional Net Capital Inflow to GDP

(Based on balance of investment and savings, including foreign investment)





Note: Local GDP by region to be 100%.

Source: Based on Table 10 and State Statistic Bureau (2001 and 1999a).

Table 13. Ratio of Regional Net Capital Inflow to GDP

(Based on domestic cross-regional balance of trade, including foreign investment)




West 10.312.

Note: Local GDP by region to be 100%.

Source: Based on Table 11 and State Statistic Bureau (2001 and 1999a).

The results from the two computational approaches are similar in the general trend, although not in exact quantity. They both show domestic capital net outflow in the 1990s from the east region, mostly between Y 10 billion to Y 70 billion; in the year 1998 or 1999 the net outflow from the east reached the level of Y 90 billion, which was further lifted to Y 120 - 160 billion in year 2000. However, with foreign investment included, the east region had always had net capital inflow, with the amount over Y 100 billion in the early 1990s, and for more than half of the decade the east region had net capital inflow exceeding the amount of Y 200 billion. In the year 2000, the net capital inflow to the east region fell back to be less than Y 200 billion, because of the “West Development” strategy launched by the end of the 1990s.

The two computations also show net capital inflow to the west region in the 1990s. If excluding foreign investment, in the first half of the 1990s, the amount of inflow was between Y 10 to 70 billion; in the latter half of the decade the amount of capital inflow exceed Y 100 billion. When foreign capital investment was included, the amount increased only by Y 10 billion without changing the picture. In the year 2000, the total amount of capital inflow to the west region was Y 180-190 billion.

It was a more complex situation for the central region. According to the result from the first computational approach, net capital outflow has occurred in the central region for the whole study period. Foreign capital investment to this region did not change the direction of capital flow; nevertheless, the total amount of net outflow was not large. The second computation also shows net capital outflow in most of the study period, but the foreign capital investment, if included, would made it a net capital inflow case. Combine the two results, we consider that capital flow in and/or out of the central region could be in a range of ±Y 30 billion for most years in the 1990s.

From the perspective of the channels for capital flow, we can see that fiscal transfer payment and bank funds had been the main ways of capital flow from the east to the central and west regions. The major part of transfer payment went to the west region, below Y 40 billion till the year 1997; which grew up to Y 80 billion by the year 2000. The amount of transfer payment that flew to the central region was only a couple of billions of RMB in the most of 1990, and it went up to be over Y 20 billion in the year 2000.

The situation in transfer of bank funds is different. The amount of funds that went to the central was much greater than that to the west region, drawn from the east region ; in the peak year (1999) it reached Y 120 billion. Compare to the bank funds transfer to the west, it was only Y 50 billion in the peak year (1995). In the 1990s a great portion of the bank funds transferred to the central region, it seemed to show that the in central region the large number of SOEs absorbed huge amount of bank loans.

Contrary to the above two channels of capital flow, the movement of financial resource through the capital market went mostly to the east from the central and west regions,but the total amount is far smaller than that through the former channels. More obviously, 80% of the foreign investment went to the east region, the amount has exceeded Y 300 billion since 1996. This changed the situation in the east region from new capital outflow to inflow. In the year 2000, the amount of net capital inflow to the east region was somewhere between Y 140-180 billion.

The above analysis shows that in the 1990s, the movement of capital through market and government channels had been going different directions. The capital market favored the east region for profit-seeking; transfer payment that was conducted through government channels favored the west region for the purpose of reducing the development gaps. In the case of bank funds distribution, it should be noticed that the state banks tend to give the SOEs priorities as a tradition against profit-seeking, although they have become more market-oriented as the result of economic reform. This may explain that during the 1990s, bank funds were allocated to the central region disproportionately in its favor; and to some extent benefited the west region as well. This tendency was altered since 2000, more bank funds were moving eastward, which is more of a risk-avert behavior.

It should also be noticed that attracted by higher rate of return of investment in the east region, some loans allocated to the west region were transferred to the east region through non-official or private channels. At the same time, private funds could also flow to the east region through cross-region direct investment and lending. Capital flow under this category was captured in table10 and 11, under the “other” item. It can be seen that there was a net inflow of “other” capital to the east region and an outflow of capital from the central region in most years of the study period. In the case of the west region, in some years the “other” capital flow was negative (outflow) and in other years it was positive (inflow). This can be explained by the two parts of the “other” capital flow. One is private capital movement through market, moving from the west and central regions to the east; the other part is direct government investment not included in local government budget, and direct loaning skipped local bank branches. It was for large infrastructure projects which in some cases cover multiple provinces. In the recent years, such direct government investment and direct loaning has been growing fast, and in the majority of the cases the money was put to the west region, exceeding the outflow of private capital to mark a net inflow to the west region.

To summarize the above, in the 1990s capital was drawn out from the east region through fiscal transfer payment and the bank system, however greater amount of foreign direct investment and private funds through private channels and capital market established a net inflow of capital to the east region. In the mid 1990s the net inflow of capital accounted for 7-8% of local GDP in the east; but it went down to 3% by the end of the decade. Contrary to the case of east region, in the other two regions, the inflow of capital was conducted through government and the banks plus smaller amount of foreign investment, but private money had been flowing out of the two regions. In the case of the central region, the inflow and outflow of funds generally canceled out each other, left a small, and unstable, directional net balance accounting for no more than 2% of local GDP in most of the years. In the case of west region, the government transfer payment and policy-oriented bank loans greatly exceeded the outflow of private capital to cause a net inflow of capital into this region, which accounted for 5-10% of local GDP. Needless to say, that the base of local GDP in the west cannot match that of the rich east region.

3. Across-Region Labor Migration

Economic theory argues that the flow of labor between less developed and well-developed regions may reduce or moderate the total output of the less developed region, but it will lift the per capita GDP and average income in the less developed region; therefore it helps to reduce the regional gaps in wage and GDP per capita (for explanations see Yao Zhizhong, 2002). That is because under the market, the labor moves normally from areas with low marginal productivity to where it is higher, therefore to reduce the supply of labor where it is in surplus; in turn the marginal productivity and return to labor in that region will increase, which means raise of average income. Furthermore, the migration of labor force provides continued labor supply to keep the wage level from rising too fast in the target area. It helps to maintain a low domestic production cost and high competitiveness in the world market. The flow of labor in general can improve the allocation of resources and raise the productivity and total output. Cai Fang, Wang Dewen et al. observed that the development of labor market has direct impact on regional economic growth (Cai Fang, et al. 2002).

It is commonly observed that emigration from rural and under developed areas toward urban and developed regions normally make better use of the surplus labor from the former areas, it not only raises their wage level but also lift the income level of the home town through remittance. In some poor areas in China, remittance by migrant labor has become a major source of local income, helped capital accumulation and economic development. The following is a rough estimation on the flow of labor across the regions and its impact on regional disparity.

A.Magnitude of Labor Flow Across-Region

The estimation in this section is mainly based on the information on rural labor emigration came from the 2002 survey report by The Department of Training and Employment of the Ministry of Labor and Social Security, and the Rural Survey Brigade of the State Statistic Bureau, and also supplemented by their reports of other years (the mentioned report will be noted as “Ministry of Labor and Social Security, 2002” in the following). Their sample survey covered 60 thousand rural households in all provinces. According to the data from this report, by the year 2000, there were accumulated 113.4 million rural labor force transferred to non-agricultural sector, accounting for 23.6% of the rural labor force. The majority of the emigrants, 65.8% (74.6 million) were working in cities and small towns (46.8 million and 27.8 million, respectively). The number of people moving across the boarders of provinces was 28.2 million (living out of home village for more than 6 months), accounted for 5.89% of total rural labor. Within the total cross-provincial migrating labor force, 90% came out from the center and the west (56% and 34% respectively, or 25.4 million as a sum of the two regions), 82% (21.1 million) ended-up in the east region. Of the cross-provincial labor migrants, 74.5% moved from the central and west regions to the east, 15.4% moved across-province but within the same two regions, 7.5% was flowing within the east region but across-province, 2.5% was flowing from the east to the other two regions. Table 14 presented a matrix of rural labor migration among regions calculated from the above report. Data in columns are for labor outflow, and in rows are for labor inflow.

Table 14. Rural Labor Cross-Regional Migration Matrix (2000)


Amount (10,000 persons)





Total migrants as 100 (%)





Source: Ministry of Labor and Social Security (2002).

It can be seen from Table 14 that the orientation for rural labor emigration was basically east-ward. The single one biggest receiver of emigrants was Guangdong Province, which absorbed a half of total emigrants. The next 5 provinces/municipalities were Zhejiang, Shanghai, Beijing, Fujian, and Jiangsu, as a group they took a quarter of the total cross-provincial emigrants. The major base areas of the emigrants were: Sichuan (4.5 million estimated), Anhui (3.6 million), Hunan (3.3 million), Henan and Jiangxi (each 3.2 million), Guangxi and Chongqing (each 2.1 million).

To check the reliability of the above data, the author compared the 1999 data by Ministry of Labor and Social Security (2001) with the survey data of the same year done by the Policy Research Office of the Central Committee of CCP, and the Ministry of Agriculture (2001).

(the latter survey cover more than 20 thousand rural households). Statistical inference from the two surveys led to a total cross-provincial flow of rural labor by 21.15 million and 18.94 million, respectively. There was a 10% difference in the two results. This indicates that, although there are errors between the statistic inferences from the two sources, they are basically acceptable.

It should be noted that the above numbers of cross-regional migration was possibly under-estimated, because both surveys were indoor surveys in villages and therefore would not cover those families, which have entirely moved out. Still the relative proportions of the emigrants should be reliable. If we could assume that 10-20% of the cross-provincial flow of labor was emigrating with the whole family, then the total number of cross-provincial flow of labor in the year 2000 could have reached 31 million to 35 million. Proportionally, there would be 25 million people from the central and west regions to the east region

B.Contribution of Cross-Regional Labor Flow to Output

The economic impact of labor flow from the central west region to the east is estimated in this section. The first was to estimate the marginal output of the labor force in both the home region and target region. According to the State Statistic Bureau, in the year 2000, the net annual per capita income of rural residents for the three regions was Y 3476, Y 2075, and Y 1632, respectively, from the east to the west (simple average for all regions). Let us assume that return to labor generally accounted for 50% of the net income, the rest should be for the return to land and capital. We could estimate the average return to labor for the working age residents (as wage proxy) in the three regions, based on information about per capita income level, proportion of return to labor in total value of output, and the proportion of working age people in the total population. Take into consideration that the people who left home village for the cities were mostly young and healthy, the author adjust the original rate of return to out flowing labor by a factor of 1.2. Furthermore, it may be assumed that the average return to the out-flowing workers equals their marginal product at their home village, and ignore the changes of the wage level caused by cross-region employment (because they took only a small portion in total labor force, i.e., 5.89% in the year 2000).

By the same argument, let us assume that the marginal product of the inflow labor force in the target region equals their wage level. As a matter of fact, the average wage of the inflow labor force is lower than the average wage of the “native” workers in general. Due to lack of data, the author uses the local average wage of the collectively-owned enterprises in the east region as a proxy, which is the lowest among all the available wage levels of different category of businesses.

Table 15 shows the differences of the marginal product of labor in the three regions on average, according to the above calculation. We can see from the table that in the year 2000, the marginal product of out flowing labor at their home in the central and west regions were only Y 1613. Comparing to their marginal product of Y 7837 in the east cities as a hired worker, it was nearly a 5-fold difference, or Y 6224 increase per worker. Roughly there is a Y 155.6 billion increases in the value created by the 25 million cross-regional working people, equals 18% of the agricultural value added covering both the central and west regions.

Table 15. Marginal Product of Labor (MPL) in the Three Regions

Income Per Capita (Rural)(Y)Marginal Product of Labor (Rural)(Y)Marginal Product Of Labor(Urban)(Y)Urban-rural MPL Ratio




East urban/west rural4.86

Source: computation based on the State Statistic Bureau (2001).

Needless to say, the inflow of labor force would press on local employment. However, observations show that majority of the migrating workers were supplement to local labor market in the low-pay jobs under hard working condition. The rate of substitution to local workers should be far below 1. Even in the competitive jobs they can work harder therefore to raise labor productivity. In one word, the migrating workers contributed to the economic development of target regions.

C.Contribution of Cross-Regional Labor Flow to Reducing Regional Disparity

There are two ways the cross-regional labor flow can contribute to narrowing down of regional gaps. First, it will reduce the pressure of surplus labor on the limited farm land in the central and west regions. Because of the decreasing marginal return to input factors, emigration of labor force from these regions will lift marginal productivity of labor, which could be translated into higher per capita income, especially in where surplus labor widely existed. A good example is in crop cultivation. Reducing the labor force engaged in crop growing would help increase per capita land use therefore the size of farms, with the consequence to increase income. The problem is that the scale of outflow labor force from the central and west regions, although reached 25 million, accounted for only 8% of total local labor, not enough even to counter-balance the growth of population. Up to today, there are 310 million rural labor force crowded on the limited farm land in the central and west regions. According to the calculation by Yao Zhizhong (2002), in 1995, the covariance of the country-wide return to labor, C=0.405, Gini coefficient=0.221. If without the outflow of labor force, it could have been in a situation that C=0.599, Gini coefficient =0.315. It indicated that the outflow of labor force had brought about the change in a lower covariance by 19.4 percent, Gini coefficient by 9.4 percent. It means that the income gap between regions was brought down by labor migration, compared with that if there had been no labor flow.

In addition, a big amount of remittance was made by the outflow working people to the central and west regions. According to the Ministry of Labor and Social Security (2002), the average remittance per out flowing rural worker was Y 4522. A total of Y 512.8 billion was sent back to join the revenue of the workers’ rural households. The average remittance by the cross-provincial workers was Y 3273, then the total 25 million outflow workers from the central and west to the east region could have sent back a total remittance of Y 81.8 billion. Those working within the central and west regions but across provinces could have contributed in the remittance to make the total close to Y 100 billion. This was a great lift in annual income by 10% to the whole of 510 million rural population in the central and west regions. It was especially significant in the poor and backward areas in the two regions with higher proportion of emigrating workers.

D. The Potential of Cross-Regional Labor Flow

Despite the contribution of the flowing labor force to target regions, they have faced various constraints and discrimination, because of their lower education level and the pressure they brought in target areas on employment, infrastructure, the environment and social security. In fact, the magnitude of flowing population is relatively small if compare to the total population inthe target areas except Guangdong Province and a number of big cities. Some of the social problems that brought by labor flow in many places are exaggerated; some of them could be solved with reasonable effort. As a matter of fact, Guangdong Province alone has taken in one half of the nationwide cross province migrating labor force without social crisis or high unemployment rate. To the contrary, the unlimited labor supply supported economic development of the province to stay on top of the country. Product from this province has been highly competitive in the world market. Table 16 gives the distribution of migrating works in six major target provinces in the year 2000, calculated from the Ministry of Labor and Social Security (2002). It can be seen that these six provinces took 80% of total migrating labor in the whole country. In prospect, if in this decade a few provinces come up with Guangdong Province in this respect, we can hope a great expansion in cross-regional labor flow to reduce income disparity among the regions significantly.

Table 16. The Main Target Regions for Migrating Labor Across-Province


% in Cross-Province Flow51.

Estimated Inflow (10000 Persons)1458217167144121996192825

Source: Computed on data from Ministry of Labor and Social Security (2002).

On the other hand, the delay in urbanization in China has been one more constraint on the outflow of rural working people. By now, the rate of urbanization (by population) is 37%, lower than not only developed but also developing countries at the similar income levels. Compared with the large population, the number of cities is limited. In particular, large cities with more than one million population in China have taken less than 10% of the total population. The root of the problem lies in the urban-rural isolation policy prior to the economic reform (1978), and the administrative restrictions on development of big cities lasted in a longer period. In prospect, more effort is need in breaking the institutional barrier between the cities and the countryside, in systematic support to facilitate migration of labor force in job training, employment service, development of labor market and city infrastructure, etc, so that to contribute to the ultimate goal of integrated urban-rural development.

4.Regional Distribution of Human Capital

The difference in the rate of economic growth is a key issue in studies of regional gaps. According to neo-classical economic growth theory and the endogenous growth theory, changes in the saving rate, therefore capital formation, can affect the rate of growth only in the short run (Solow, 1956; Swan, 1956), but a change in human capital stock can directly affect the total factor productivity therefore to affect economic growth in the long run (Romar, 1986; Lucas, 1988). Therefore, when we look at the impact of capital and labor on regional gaps, the focus is on the flow; while looking at the impact of human capital, more attention should be paid to the levels of the stock across-regions.

A number of measures can be used for the stock of human capital, each has its own advantage and shortcomings. Romer (1986) used the number of R&D personnel for the measure of human capital. In the coverage of Chinese statistics, we have the data on professional technical personnel in state-owned enterprise and government R&D institutions (province-specific), but the data do not cover those in the private sector. An other indicator is the expenditure on R&D (province-specific), unfortunately historical data are not available. Lucas (1988) used the years of schooling of workers to represent human capital. In the Chinese statistics, there is the item of education of employees but covered only short period. Other data that could be taken as relevant are: number of students in schools, the rate of enrollment and graduation; however they can be used to calculate changes in human capital stock only under a constant condition which lasted over a relatively long time period. Otherwise they can only measure the changes in human capital in the future. Cai and Du (2000) used the rate of adult literacy to represent human capital, and in turn found that the initial disparity in the level of human capital stock had significant influence on local economic growth. Du and Wang (2002) also found that education (measured by school enrollment) and health level (doctor per 1000 persons) have positive impact on local economic development.

In the following, the author starts anal ysis with descriptions on human capital stock with several indicators.

A.Schooling of Workers and Scale of the Education Sector

The average education in schooling years of the working people in 1996-1999 of the three regions is listed in Table 17, in which we can see an average of 8.1 years of schooling for the working people in the east region, 7.7 years for the central region and 6.9 years in the west. There is a 1.2 year difference between the east and the west. There is no information about human capital carried by the flowing labor force; however observations strongly suggest that people with higher education tend to move toward the east region.

Table 17. Average Year of Schooling of Working People (Year)





West – East Difference1.

Note: Schooling year is a weighted average calculated from the working people in different education background. The regional average is a weighted average on a provincial basis.

Source: Calculated from The State Statistic Bureau (multiple years).

We can also use the following indicators to measure the scale and capacity in education across the three regions: the proportion of high school and university students in population, pthe proportion of teachers and staff numbers of high school and university in population (see Table 18), expenditure on education as a proportion of GDP, and per capita educational expenditure (see Table 19). We can see from the two tables that regional disparity in the scale of education existed; the secondary education in the west was weaker than in the other two regions, and that the higher education went down from the east to the west. However there was no huge disparity across the regions in proportions of students and teachers to total population. For high schools, these two proportions in the west were 87-88% relative to the east; and for universities, the two proportions in the west were 62-63% of that in the east. The difference revealed from these numbers matched that in the education level of working people. In the case of the proportion of educational expenditure in GDP, the west exceeded the east, but it lost again to the east in the per capita education funding (53:100). The implies that, the west region faces a strong limit in education funding constrained by the lower level of economic development.

Table 18. Ratio of Students and School Staff to Population (2001)

High/Vocational SchoolUniversity

Enrolled Students/Population




Education Staff/Population




Source: The State Statistic Bureau (1992, 2002).

Table 19. Education Expenditure Across the Regions (2000)

Education Expenditure/GDPPer Capita Education Expenditure(Y)




Source: The State Statistic Bureau (1992, 2002).

B.The Professional Technical Personnel and R&D Funding

There has been no statistics on the number of professionals technical personnel nationwide. The only relevant information was that about the professionals working in state-owned enterprises and government research institutions, including personnel in engineering, agriculture and medication, and scientists and teachers. To our knowledge, there are a certain amount of professionals working in private firms and agencies, but the majority of professionals are still engaged in state-owned organizations. We here took the statistic mentioned above as an approximation. In Table 20, it can be seen a relatively balanced regional distribution of professionals, the west/east ratio was 0.91; however the distribution of R&D funding as proportion in GDP was in disparity: the centerand west was 54% and 70% to the east.

Table 20. Proportion of Professionals in State-Owned Agencies to Population

Professionals/Population (%)R&D Funding /GDP (%)





West/East (%)91.0%70.3%

Source: The State Statistic Bureau (1992, 2002).

C.Marketization of R&D Output

The disparity in marketization of R&D output across the three regions is much larger compared with the disparities in the level of schooling, number of professional personnel and R&D inputs that indicated above. Table 21 shows in 2001, scientists/technicians in the east region applied 140 patents per 10 thousand professional and had 83 approved, that for the west region were 35 and 22, respectively, only a quarter of the east region. If look at the market value of sold patens and technologies per head of scientists and technical personnel, then the west region could match only 30% of the east region, in a gap that seems to be widening (Table 22).

Table 21. Patents Application and Approval

(Every 10 Thousand Scientist/Professionals)

Application/10000 PersonsApproval/10000 Persons





Source: The State Statistic Bureau (1992, 2002).

Table 22. Market Value of Sold Patents and Technologies (Y Per Person)






Source: The State Statistic Bureau (1992, 2002).

Above information shows regional disparity in education and human capital stock exist. In particular, lasting of the gap in education would further worsen the inferior position of the west in terms of human resources. However, a most significant gap between the east and the other regions are the poor performance of marketization of R&D output in the latter regions. Today, a major part of R&D output realizes its value through the market. However, scientists and technical persons in the central and west regions applied, and were granted, only 1/4 patents compared with those in the east, and the former produced only 1/3 R&D output in market values compared with the latter. The low level of marketization in R&D is apparently one of the major constraint to the development of the west region.

5.Regional Disparity in Marketization

Institutional factors and policies also play significant roles in regional disparities. In the past two decades, China achieved a high economic growth rate; this was greatly due to the market-oriented reform. However the marketization process over the whole country did not approach in the same speed. According to the study by the National Economic Research Institute (Fan, Wang and Zhu, 2003), the progress of marketization in the coastal provinces far exceeded that of the central and west regions. The Marketization Index for provinces (NERI Index for short) shown in Table 23 indicates that most of the provinces in the east areas were not only leading in economic grow but also were advanced in the marketization process. The central region showed a better performance than the west, still it was far behind the east region. The central and west regions are particularly dropped behind in the following two aspects: development of the private sector, and development of fact markets (see Table 24).

Table 23. Marketization Index by Province/Municipalities (2000)

East RegionMarketization IndexRankCentralRegionMarketization IndexRankWestRegionMarketization IndexRank






Shanghai7.046Jiangxi5.4620Inner Mongolia4.7624






RegionalAverageEast7.16Regional AverageCenter5.47Regional AverageWest4.71

Note: It is a relative index system of marketization, computed on the basis of 25 factors. It shows the relative positions of the provinces in the process of marketization.

Source: Fan, Wang and Zhu (2003).

Table 24. Five Aspects in Marketization in the Three Regions

Government-MarketRelationshipNon-StateSectorDevelopment Product MarketDevelopmentFactor MarketDevelopmentMarket Intermediary& Legal Environment





East-west gap2.744.102.593.452.13

Source: Fan, Wang and Zhu (2003).

Fan and Chen (2002) found that marketization of provincial economy made significant contribution to total factor productivity (TFP) growth. They carried out econometrical regressions and found that both the achievement of marketization (represented by the NERI index) and technical progress (represented by patents in technical innovation) have positive and significant impact on TFP (see Table 25). They pointed out that TFP growth can be lifted by reforms in the following areas: 1) to retrench the government sector and transform the government function to reduce its direct intervention on factor allocation, and to enhance the role of market in resource allocation; 2) to encourage development of the private sector; 3) to facilitate development of the goods market by enhancing the role of market in price determination, and eliminating local trade barriers in the markets; 4) to facilitate development of factor markets, i.e., financial market, labor market, technology market, etc., and 5) to improve legal environment for business and to heighten services of market intermediaries (e.g., lawyers, independent accountants, etc.).

The above finding confirmed that the different regional performance in marketization was an important cause for regional gaps.

Table 25 Impacts of Marketization and technological progress on TFP growth, 1997-99



Technological progress0.00039(4.76)0.0006(6.089)0.0006(6.41)0.001(3.73)0.0009(6.29)0.0009(5.855)

xt test0 chi2 value74.2175.23

xt haus chi2 value42.331.51



Note: RE - random-effects; FE - fixed-effects; t-student in parentheses. The null hypothesis of exogeneity of marketization can not be rejected by the Davidson-McKinnon exogeneity test.

Source: Fan and Chen, 2002.

6. Regional Disparity in Urbanization and Rural Industrialization

There has been a big gap in regional rural industrialization and urbanization. Since the beginning of economic reform, the coastal areas in east China experienced a rapid rural industrialization. The TVEs (township and village enterprises) in the whole 1980s and the first half of 1990s had been growing at a speed exceeding any other sectors in the economy, and were the leading sector in acceleration of economic growth. Studies show strong ties between the level of per capita rural income and development of TVEs in the rural areas. In the meantime, the higher urbanization level in the east areas was also a supporting factor in the economic development (Wang, 2001). The following is a brief description on the regional gaps in rural industrialization and urbanization.

In 1978, the beginning year of economic reform, the Commune and Brigade Enterprises (the predecessors of TVEs) played a minor role in rural economy. It hired only 20 million people from the farming sector, less than 10% of total rural labor force at the time. This sector grew at a very high speed in the next 17 years. At its peak in 1996, it hired a total 136 million people. It can be seen in Table 26 that in the east region in 1996, 39% of rural labor force was working in the TVEs. It went down in the following years, in 2000 it was 37% in the east region, but in the west region it was only 18%.

Table 26. TVE Employment and Its Proportion in Rural Labor Force


Employment in TVEs (mil.persons)96.1135.1128.2

As % of rural labor




Source: The State Statistic Bureau (multiple years).

The TVEs have been contributing to increases in income of rural residents. According to the sampling survey by the State Statistic Bureau, the per capita income of rural residents from production was Y 2332 in 2001, in which 48%, or Y 1066, came from the industry and service sectors. Take the wage income for comparison, the nation average per rural resident was Y 772, compared to only Y 430 in the west 12 provinces. It was mentioned earlier in this report that the gap in income from the east to the west mainly lies in the disparity of rural residents’ earning, and the key to the rural people’s income is in the development of non-farming sector. In other words, it lies in the chances to transfer rural labors into rural and urban non-agricultural sectors. It is logical to say that an important cause of the regional gap is the lagging of rural industrialization in the central and west rural areas.

Lagging behind in development means not only late in the time schedule. Due to changes in market environment, it is imporssible for the west region to simply copy the model of rural industrialization from the east region. A number of built-in weaknesses in the TVE sector have shown up in its 20-year growth. Lack of human resources and technology supply, the poor infrastructural conditions, difficulties in access of information, and disadvantages in their geographical location, all contributed to the more and more failures of the less competitive TVEs. As a matter of fact, all these constraints and limitations had existed from day one in the TVEs history, but in a supply-constrained market, there were a lot of blanks for them to fill in, and the SOEs, which dominated the industrial sector at that time, were always slow in responding to market demand and lack in competitiveness. It was a rare opportunity to the TVEs to grow and they did mushroom on their low labor cost and high flexibility in marketing strategy. Today the market has been more competitive; the consumer preference has been reoriented to quality, brand name, technology containing and customer service. The TVEs are challenged on their fundamental weakness. It is unlikely that the TVEs can resume the glory of expanding at a 20% plus annual growth rate.

Today, urbanization might be the only way to replace the TVEs as the main channel conducting the process of transferring rural population into non-agricultural sectors. Compare to the rural environment, the cities can offer not only better supply of transportation, communication, electricity and utilities, etc., but also human capital and higher technology supplies, plus big market centers and great room for service industries to grow. There are also the conditions in the cities to provide the needed services. Financing, consultancy, information technology, education and training, all are necessary condition for modern economic developmenThese form the agglomeration effect of cities.

During the last two decades, urbanization in China has been speeding. The rate of urbanization measured by the proportion of urban population in total has been growing from 17.9% in 1978 to 37.7% in 2001, a 20 percentage points lift. However due to the institutional constraint inherited from the planned economy, the rate of urbanization in China is lower than not only the developed economies but also those on the same development scale. Along with the market-oriented economic reform, urbanization will be playing a particularly important role in the process of China’s economic development (Wang Xiaolu and Xia Xiaolin, 1999). In prospect, further rural industrialization and urbanization in the central and west regions will play positive roles in reducing the regional gaps.

Nevertheless, the central and west regions are lagged behind in the process of urbanization. Table 27 shows the ratio of urban residents to population, and ratio of urban workers to total working people by region. It shows a widening gap between the east and the west in the rate of urbanization. In 1982, the gap between the east and the west was 7.5 percentage points, it grew up to 17.4 percentage points in 2000. During this period, the rate of urbanization in the east region increased by 22 percentage points, however in the west region it increased only by 12 percentage points. The gap in the proportion of urban workers in total between the east and west regions is also widening.

Table 27. Urbanization Rate (%)

Urban/total populationUrban/total employment





Note: Data of recent years are not included due to obvious errors in the region-specific data and the overall data, which add up to be an 800 million difference in the year 2000.

Source: The State Statistic Bureau (1999a, 2001).

Lacking of large and medium sized cities is in the center of urbanization gap across regions. China has a population close to 1.3 billion but with only 92 cities in a population scale of 500 thousand population and above, containing 9.6% of the national population in them (Table 28). The gap here between the east and the west regions is striking. Within the 92 cities mentioned above, 42 are in the east region; only 14 are in the west, a 3:1 ratio. In the east region, 65 million people (13.3% of total population) live in cities with population of 500 thousand (and above), while in the west region only 20 million people (5.6% of population) live in cities of the same scale.

Table 28. Large and Medium Sized Cities and Population (500,000 and above)

No. of citiesCity populationTotal populationPopulation ratio





Source: The State Statistic Bureau (2001c).

7. Changes in Regional Comparative Advantages

The allocation and transfer of capital, labor and human capital across the regions are the basic factors in determining the disparity in regional economic development. In a market economy, allocation and transfer of the factors across the regions are directed by regional comparative advantage, as well as the shifting of comparative advantage over time. In this section, comparitive advantage of different regions, and its shift, is investigated.

A.Regional Comparative Advantage by Factors

According to Cai and Wang (2002), in the case of comparative advantage in labor resource, the west region dominates the central and then the east region. The gaps among the three regions have been widening since the economic reform.

It was observed that the average wage of workers and per capita income of rural residents in the west and central regions have been generally lower than that in the east region. As a matter of fact, take the case of per capita net income of the rural residents, the ranking from high-to-low is east, centerand west; but if take the case of average wage of urban workers, then the high-to-low list sees the east, the west and the central region. The average wage level in the west region is higher than in the central region, which phenomenon might be caused by the legacy from central government policy that favored the boarder regions, as well as fiscal transfer payment among the three regions. But this issue did not change the ranking in the overall cross-regional pattern in income distribution as the east, central, and west regions from high to low, mainly because rural population takes the majority of population in the two regions, the centerand the west. Table 29 gives the weighted average earning of the workers and the rural labor force, as well as the weighted average income per capita covering both urban and rural residents. The proportional relationship of the cross-regional earning is 100:59:53, for the per capita income the ratio is 100:62:54. It indicates the existence of comparative advantage in labor resource in the central and west regions, especially in the west region. This result matches the finding by Cai and Wang (2002).

Table 29. Comparison of Average Earning and Per Capita Income (2000)

EarningPer Capita Income

AverageWage ofWorkersAverageEarning ofRural LaborForceWeighted AverageUrban ResidentRuralResidentWeighted Average

Per Capita Earning/Income(Y)

East10800 2862 5330 7692 3187 4588

Center7270 1773 3159 5181 2073 2856

West8368 1434 2860 5622 1690 2498




Note: Rural labor earning is computed on the net per capita income and proportion of working age people in population. Rate of return to labor in the income is 0.5; a full labor is assumed to be 1.2 times of average working age people.

Source: The State Statistic Bureau (2001, 2002).

In terms of physical capital endowment, the east region has its comparative advantage over the other two regions. At the beginning of economic reform, the central region was close to the east in per worker capital stock. The west was lagged behind. Over the last two decades, the speed of capital accumulation in the east has been far ahead of the other two regions. Foreign capital investment into this region was not the single most important cause, which came in mainly in the 1990s; the accumulation of capital in the east region was accelerating in the 1980s. In 1998, the value of physical capital per worker in the east region exceeded Y 20 thousand, however in the central and the west regions, the number was somewhere between Y 5-6 thousand, only a quarter to that of the east region (Cai and Wang, 2002). The “West Development” strategy implemented in the late 1990s assisted capital flow westward, this may result in a shrink of the relative regional disparity in capital stock over time, although the absolute difference could be growing for a longer time.

The east region also has strong advantage in human capital over the other two regions. As discussed in Section 4 of this report. It shows mainly in the number of patens applied and approved, and market values of technology sold per person, rather than in schooling years of workers, number of professionals and technicians, and R&D funding, etc.

The central and west regions have strong comparative advantage in land resource endowment over the east. Despite the fact that per capita arable land resource in all the three regions has been declining due to commercial development and increase of population, it did not alter the pattern of comparative advantage in the issue. In addition, there is a big potential in the wasteland in the west to be developed into forest and grazing ground.

There are two basic approaches to dig up the potential regional comparative advantages. The first is to develop local industries based on these potentials. For example, the central and west regions can further develop their labor-intensive, and relative land intensive, industries. For the east region, however, with the increasing capital stock and advantage in human resources, may gradually lift capital intensity as well as the technological intensity of the local industries.

Second, the abundant resources can be “exported” to other regions to earn higher returns, and in the mean time to facilitate the development of target region. One example is to “export” labor resource from the central and west regions to the east. By the same token, the east region can “export” its tangible capital and human capital to the central and west regions to make better use of local labor and land resources, and to lift the technical standard of the local business.

The success of applying the above approaches depends on not only the abundance of resources but also the local marginal factor productivity. When a local resource is plentiful, usually its marginal productivity tends to be low, and “export” of this resource is hopefully to earn higher return. In the real world things are more complex; so changing in marginal productivity of a factor is the key to understand the shift of regional comparative advantage, as well as that in the trend of resource allocation.

B.Shift of Comparative Advantage

According to Cai and Wang (2002), two decades ago the capital productivity in the east region was slightly higher than that in the central region, and significantly higher than the west. Twenty years after, the gap has been converging. To the contrary, the disparity in labor productivity 20 years ago across the regions was minor but has been widening since. It is similar in the issue of human capital except that the initial disparity was big. These are shown in Table 30.

Table 30. Changes in Marginal Productivity of Factors (Y)

YearPhysical CapitalLaborHuman Capital









Source: Cai and Wang (2002).

It is obvious that in the last two decades, the disparity in marginal productivity of capital, labor and human capital was the driving force in conducting the east-ward flow of foreign and private capital, labor and human capital. Because of the diminishing marginal return to factors, the large amount of inflow of factors into the east region has resulted in convergence in marginal productivity of factors. Considering the possible errors in quantitative analysis, we cannot confirm if this processhave been completed. However, at least we can say that the regional gap of marginal productivity of capital has been reduced significantly, therefore the tendency of east-ward flow of capital through private channels is likely to decline in the future, and the opportunities for foreign and private capital to invest in the central and west regions will expand.

The transfer of labor and human capital to the east region so far has not reduced the disparity in their marginal productivity across the regions, to the contrary the gap is widening. This can be explained by a few reasons. First, the magnitude of cross-region labor flow is minor compare to its huge total stock. Currently the majority of migrating labor is from the countryside. It has been mentioned in Section 3 of this report that in the year 2000, the central and west regions provided a total of 21 million labor force to the east region, which accounted for less than 5% of the total in the origin regions and 8% in the target region. This is not to result in convergence of marginal productivity of labor across regions. Same situation may apply to human capital transfer. Second, the continued higher per capita investment in the east than other regions, added by huge amount of foreign investment, causes a decline in the marginal productivity of capital; while at the same time it can lifted the marginal productivity of labor, in turn to widen the regional gap in the issue. This is a short-run phenomenon in a dynamic process, although it could last for years. Third, the east region enjoys higher technological growth as well as better performance in marketization, both could facilitate an increase in the marginal productivity of labor.

8. Conclusion

The causes that had dominated the changes in regional disparities are the focus of this study, which include the distribution of factors (capital, labor and human capital) and their flow across regions; institutional changes (mainly marketization) and structural evolution (mainly urbanization) in each region. Regional comparative advantage and its tendency is also examined.

The author can conclude our findings from the study on regional distribution of factors and the changing comparative advantage into the following.

1.Under economic reform during the last two decades, foreign capital and private capital had flowed into the east region in great magnitude. It accelerated economic growth in the east region, and also widened the gap between the east and the other regions. In the meantime, government budgetary payment and bank funds transferred to the center and west irresponsive to the market; however these funds were not large enough to counter-balance the inflow of foreign capital and private capital to the east region. The ease region has been a net receiver of capital.

2.The lasting inflow of capital to the east region has diminished the marginal rate of return to capital, therefore it will reduce the east-ward trend of capital flow and will be translated into a slow down of the widening of income gaps. The “West Developemnt” policy is in favor of this transition, but in any case the disparity between the east and the west will not be eliminated soon, it will exist for a fairly long period.

3.Migrating of labor force occurred mainly within each region in the past, mainly from rural to urban. However, across-region migration from the west to the east is expanding, which is attracted mainly by wage difference and chances of employment. It has facilitated the economic growth in the east regions and at the same time kept the wage level down. It also has helped reducing the pressure of unemployment in the central and west regions, and moderated the low income of farmers there through the remittance from the migrating workers. Because the labor productivity gap is large, the labor migration from the west to east will continue in a considerable long period in the future, and will be facilitated by further removal of institutional barriers on labor flow. This will improve allocation of resource and reduce income gap among regions

4.There is a significant disparity in the distribution of human capital stock across the regions. However more im portant is the low level in the marketization of R&D output in the central and west region; this seriously hindered the efficient use of human capital. To push the development of a technology market will contribute to narrowing the gap in human capital productivity and therefore the regional gaps.

5.In the evolution of marketization, there is a general difference between the east and the other two regions especially the west, mainly in the development of non-state economy and factor market. It will be critical to push the process of marketization and improve the institutional environment for business for developing the west.

6.The west region is far behind the east region in urbanization. Especially, relative large cities with population over 500 thousand are short in the west region. Cities will play an increasingly important role in economic development in the future. Therefore facilitating the process of urbanization is a necessary condition for a better investment environment in the west.

7.The central and west regions have a comparative advantage in its cheap labor, however due to their low capital and technology intensity, and the poor condition in management and market environment, the west region has a low labor productivity, and the gap to the east region is widening. It canceled the comparative advantage in labor of the west region, and resulted in the eastward flow of labor. Changes in this situation will depends on changes in the pattern of regional capital flow, however, improvement of the market environment in the west is also important.

Generally speaking, the gaps between the east coastal regions and the other two regions, in terms of economic development, has been widening in both absolute and relative measures. This was a result of the difference in comparative advantage, and flow of capital and labor therefore induced. However the continued capital flow has led to a reduction in capital productivity disparity, together with the “West Development” strategy, the pattern of capital flow will likely to change gradually in the future to place the central and west regions in a better position. However, there is no sign for the convergence of labor productivity across the regions, there is a long way to go to finally close down the regional gaps in this aspect. Along the process, improving physical and institutional environment for business in the central and west regions, including further marketization, facilitating development of the non-state sector, speeding up the process of urbanization, are all needed to reinforce development of the central and west regions towards eliminating regional disparities.


(Original In Chinese)

Cai Fang and Du Yang, 2000, Convergence and Divergence in China’s Regional Growth. Economic Research, issue 10, Page 30-37, Beijing.

Cai Fang and Wang Dewen, 2002, Regional Comparative Advantages: Disparity, Evolution and Impact on Cross-Regional gaps.

Cai Fang, Du Yang, and Wang Meiyan, 2002, Technical Efficiency, Allocative Efficiency and Distortion in the Labor Market: An explanation to the Institutional Cause in Differed Growth.

Du Yang, Wang Meiyan, 2002. Investment in Human Capital and its Impact on the Cross-Regional Difference.

Fan Gang, Wang Xiaolu, and Zhu Hengpeng, 2003, Marketization Index for China’s Provinces: A Report for Relative Achievement in Marketization. The Economic Science Press.

Guo Jinlong and Wang Hongwei, 2002, Cross-Regional Capital Flow and Regional Disparity.

Meng Lian, Wang Xiaolu, 2000, An Estimation on Creditability of China’s Growth Statistics. Economic Research, issue 10.

Wang Xiaolu, 2001, Rural Industrialization, Urbanization, and their Roles in Regional Economic Development.

Wang Xiaolu and Xia Xiaolin, 1999, Optimizing City Size, Driving Economic Growth. Economic Research, issue 9.

Wu Jian, 2001, Regional Distribution of FDI and Its Economic Effect.

Yao Zhizhong, 2002, Labor Mobility and Regional Disparity.

(Original in English)

Barro, R. and X. Sala-i-Martin, 1995, Economic Growth, McGraw-Hill, London.

Locas, 1988: On the Mechanics of Economic Development, Journal of Monetary Economics, 22, 3-42.

Maddison, A., 1998, Chinese Economic Performance in the Long Run, OECD publication, France.

Romer, 1986: Increasing Returns and Long-run Growth, Journal of Political Economy, 94, 1002-37.

Solow, 1956: A Contribution to the Theory of Economic Growth, Quarterly Journal of Economics, 70, February, 65-94.

Swan, 1956: Economic Growth and Capital Accumulation, Economic Record, 32, 334-61.

Wang, Xiaolu and Lian Meng, 2001: “A Reevaluation of China’s Economic Growth”, China Economic Review, 12, 4, 338-346.

Source of Data:

Department of Employment and Training of the Ministry of Labor and Social Security,and the Rural Survey Brigade of the State Statistic Bureau, 2002, Rural Employment and Labor Mobility in China in 2002.

Editorial House for China Financal Almanac, 2001, China Financial Almanac, China Finanal Press.

Policy Research Office of the Central Committee of CCP, and the Ministry of Agriculture (2001), National Rural Social-Economic Survey Data Collection, China Agricultural Press.

The State Statistic Bureau, multiple years, China Statistic Yearbook; The State Statistic Press.

The State Statistic Bureau, 1999a: A Collection of 50 Years’ Statistical Data for New China, The State Statistic Press.

The State Statistic Bureau, 2002b: Abstract of China’s Statistics. The State Statistic Press.

The State Statistic Bureau, 2001c: Statistical Yearbook for China’s Cities, The State Statistic Press.