China - Regional Disparities

CONFIDENTIAL Report No. 14496-CHA CHINA REGIONAL DISPARITIES June 30, 1995 Country Operations Division China and Mongolia Department East Asia and Pacific Regional Office CURRENCY EQUIVALENTS Currency Unit: Yuan (Y) $1.00 = Y 8.5 Y 1.00 = $0.1176 FISCAL YEAR January 1 - December 31 WEIGHTS AND MEASURES Metric System ACRONYMS AND ABBREVIATIONS CASS Chinese Academy of Social Sciences FDI Foreign Direct Investment GDP Gross Domestic Product HH Household IDA International Development Association IBRD International Bank for Reconstruction and Development NMP Net Material Product SOE State-Owned Enterprise SPC State Planning Commission SSB State Statistical Bureau TVEs Township and Village Enterprises ACKNOWLEDGMENTS This report was prepared by Albert Keidel (EA2CO). Florencia Castro- Leal (EA2CO Young Professional) provided valuable assistance with household budgetary calculations and the literature review. Meetings in China in November 1993 and April 1994 with the State Statistical Bureau (SSB), the State Planning Commission (SPC), and the Chinese Academy of Social Sciences (CASS) gave the author valuable insignts. Special thanks are due to Wang Xiaoqing, Tang Ping, and Zhang Chaohuang (SSB), Jiang Runyu and Jin Hong (SPC Price Bureau), Song Yuansheng and Zhang Jian (SPC Department of Spatial Planning and Regional Economy), and Liu Shucheng, Zhou Fang, Fan Mingtai (CASS Institute of Quantitative and Technical Economics), and Lin Huaxiong (CASS). Many persons within the World Bank helped make this report possible. Special thanks are due to Sultan Ahmad, Michael Hee (IECSE), David Dollar (EAICO), William McCarten, Tina Almero-Siochi (SA2CI), Andy Hamer (EA2EU), Janet Hohnen, John Quinley (EA2HR), Shahid Javed Burki, Nicholas Hope, Shahid Yusuf, Eliana Cardoso (EA2DR), David Pearce, Zafer Ecevit, Peter Harrold, Vikram Nehru, Anthony Ody, Tom Tsui, David Rix, Tamar Manuelyan Atinc, Carlos Elbirt, Hongjoo Hahm, Bert Hofman, Anjali Kumar, Rajiv Lall, Julia Li, Tejaswi Raparla, Emily Fung (EA2CO), Ramgopal Agarwala, E.C. Hwa, Dusheng Liu, and Susan Su (EA2CH). Tang Li kindly gave valuable time and effort, together with Julia Li, and Susan Su, to statistical work and mapping for analysis of disparities within Hebei and Sichuan provinces. CONTENTS Acknowledgments............................................................................................................ i Executive Summary........................................................................................................ iv 1. Regional Introduction and Brief Literature Summary .................................................. 1 A. Economic Regions in China................................................................................... B. Academic Literature on China's Regional Disparities ............................................. 7 2. Regional Disparities.................................................................................................... 9 A. Regional Output Disparities .................................................................................. 9 B. Regional Income and Consumption Disparities.................................................... 16 C. Regional Social Indicator Disparities................................................................... 20 D. Intra-provincial Disparities.................................................................................. 22 E. Comparisons with Indian Regional Disparities ..................................................... 25 F. Recommendations ............................................................................................... 27 3. Causes of Regional Disparities.................................................................................. 29 A. Land, Climate and Demographic Change............................................................. 29 B. National and Regional Policies ............................................................................ 33 C. Investment and Trade Disparities......................................................................... 36 D. Fiscal and Financial Transfers.............................................................................. 39 E. Conclusion and Recommendations-Natural and Unavoidable Regional Disparities42 Annex 1: Statistical Annex............................................................................................ 47 Annex 2: Literature Review .......................................................................................... 53 A. Methodologies in the Academic Literature on China's Regional Disparities.......... 53 B. Findings in the Literature..................................................................................... 56 C. Overall Literature Assessment............................................................................. 57 Bibliography.................................................................................................................. 58 TABLES Table 1.1: China's Economic Regions ............................................................................. 3 Table 1.2: 1990 Population by Region ............................................................................ 6 Table 2.1: China's Regional GDP by Sector, 1992......................................................... 10 Table 2.2: Change in Regional Disparities, by Sector, 1980-92 ..................................... 13 Table 2.3: China's Real Growth of National Income (NMP), by Region, 1978-91.......... 15 Table 2.4: Urban Household Regional Deviation from National Mean, 1981-92............ 16 Table 2.5: Rural Household Regional Deviation from National Mean, 1981-92............. 17 Table 2.6: Urban Household Change in Income and Expenditure, 1981-92 ................... 18 Table 2.7: Rural Household Change in Income and Expenditure, 1981-92 .................... 19 Table 2.8: Regional Population and Infant M ortality, 1990 ........................................... 20 - iii - Table 2.10: Henan Regional Disparities, 1992............................................................... 22 Table 2.11: Sichuan Regional Disparities, 1992............................................................. 25 Table 2.12: India Regional D isparities............................................................................ 26 Table 3.1: Annual Growth of China's Population, by Region, 1978-91.......................... 30 Table 3.2: Net Inter-provincial Migration by Region, 1985-1990 .................................. 31 Table 3.3: Fixed Capital Investment by Various Measures, 1992, by Region ................. 37 Table 3.4: Foreign Trade and Investment Indicators, 1991-1993, by Region ................. 38 Table 3.5: China's Budget Surplus (Revenue-Expenditure), by Region, 1978-91 ........... 40 Table 3.6: Fiscal and Banking Indicators, 1990-9 1, by Region ...................................... 42 Table A.1: China's Regional and Provincial GDP by Sector, 1992................................. 47 Table A.2: Per Capita GDP Sector Deviation from National Average, 1980-1992......... 48 Table A.4: Urban Household Real Regional Disparities, 1981-92.................................. 52 Table A.5: Rural Household Real Regional Disparities, 1981-92................................... 52 FIGURES IN THE TEXT Figure 1: Lowest GDP per Capita in the Most Populous Regions................................... vi Figure 2: Investment on the Coast Overshadows Weak Interior Effort.......................... vii Figure 3: Exports soaring on the Coast Show Transport and FDI Benefits..................... ix Figure 4: 1992 Budget Transfers from the Coast Skip Over the Poorest Regions............ x Figure 2.1: South-East Coast and South-Central Interior Disparities............................. 1 Figure 2.2: India's Regional D isparities......................................................................... 26 Figure 3.1: Investment per Capita, by Region, 1992...................................................... 36 MAPS IN THE TEXT M ap 1: Coastal and Interior Regions............................................................................... v M ap 1.1: C hina's R ivers.................................................................................................. 2 M ap 1.2: Coastal and Interior Regions............................................................................ 4 Map 1.3: China's Economic Regions with Geographical Detail........................................ 5 Map 2.1: Coastal, Central Core, and South Hinterland Regions .................................... 12 M ap 2.2: Regions in H enan Province............................................................................ 23 M ap 2.3: Regions in Sichuan Province.......................................................................... 24 - iv - ExEcuTIVE SUMMARY 1. This report examines regional disparities in China and their trends since reforms began in 1978. It finds that: * Since 1978, per-capita output disparities between China's coast and interior have increased to nearly 2-to-1, with the increase accelerating since 1985. * Two interior regions, South Hinterland and Central Core, have 530 million persons at 320 GDP dollars per-capita versus 740 dollars for 230 million persons on the South and East coast (see Map 1 and Figure 1 on pages v and vi). * In contrast, disparities along the coast have lessened' as growth on the South Coast has nearly eliminated the East-Coast lead in per-capita indicators. * For consumption, rural households have significantly larger disparities between coast and interior than do urban households; South and East coastal regions enjoy 20-to- 70- percent advantages in both urban and rural areas. * Social-indicator disparities (education and health) closely match output and consumption disparities between coast and interior. * In the large poor provinces of Henan and Sichuan (populations 90 and 110 million, respectively) per-capita GDP disparities between hrge sub-regions (15-to-20 million persons each) are again more than 2-to-1. See Chapter 2. * Disparities appear to stem from competitive advantages in natural coastal centers of modern transport and communications, which have developed more rapidly than isolated locations and more rapidly than compensating migration flows. * Gaps of nearly 4-to-1 in investment-per-capita are the most direct explanatory correlates of coastal-interior disparities-matching even starker coastal advantages in foreign investment and trade (see Figures 2 and 3 on pages vii and ix). * Census evidence shows clear but slow permanent migration from interior regions to the coast-while surveys suggest much larger scale migrations. See Chapter 2. * Fiscal transfers to poor regions have stagnated at low levels. See Figure 4, page x. * Regional patterns of bank credit expansion dramatically overshadow fiscal transfers but are especially weak in the poorest interior region. 2. The two most important conclusions from a review of regional economic disparities in China are that (1) the disparities between coast and interior regions are large and worsening, and (2) these trends may be a natural phenomenon for a large country like China. China has a dense population distributed in inherited spatial patterns reflecting location advantages for agriculture rather than for modem industry and commerce. In China's case, official policy has accentuated market-oriented disparities by (a) encouraging investment and foreign economic activity in special coastal zones, and (b) by maintaining a population registration system which makes it more difficult for migrant workers to compete for coastal urban jobs. The major outstanding issue with regard to China's regional disparities is how to anticipate the pressure for migration out of regions less naturally suited for modern economic activity. Statistical evidence shows that migration pressures are already building. -v- MAP 1: COASTAL AND INTERIOR REGIONS Russia Russia Mongolia P c c Korea China EMFar West Shanghai M North Hinterland South Hinterland and/Central.Coreaare Central Core Hong North Coast Vi Kong Pa ic East Coast LOcean South Coast') China's Economic Regions. Thismap presents economic regions which highlight China's poorest and richest areas. The South Hinterland and Central Core are poorest, with 531 million people at roughly 320 dollars average per-capita GDP. The East Coast and South Coast are richest, with 228 million persons at 740 dollars per capita. See Figure 1 and more detailed analysis in Chapters 2 and 3. Source: For more geographical detail, see Map 1.3 in Chapter 1. Note: This map is a rough free-hand drawing useful only for notional regional identification. Its boundaries, denominations, and any other information shown do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 3. This report reorganizes China's plentiful provincial-level data into seven economically significant regions, three on the coast and four in the interior (see Map 1 and Figure 1). Coastal regions have roughly 440 million persons, compared to roughly 740 million in the interior. In the interior, the poorest regions (Central Core and South Hinterland regions) together have 530 million persons. The richest regions are the East and South Coast regions, with 228 million persons. For a more detailed map and a fuller description of China's economic regions, see Chapter 1. 4. Virtually all measures of economic disparity confirm the gap between coast and interior and its accelerated worsening since the mid-reform year of 1985. Coastal GDP per capita in 1992 was 680 dollars; in the south-east coastal regions it averaged 743 dollars. In the south- central interior, however, GDP per capita averaged only 316 dollars, and in the South Hinterland - vi - FIGURE 1: LOWEST GDP PER CAPITA IN THE MOST POPULOUS REGIONS US S per Capita or Milon Persos ., 781 695 800 60s GDP per Capita 600 398 430 S00 94 332 4oo 300 T Population 100 S 147 0 -305 209 125 10 Far North South Central North East South West Hinter- Hintcr- Core Coast Coast Cos land land Coast Interior Coast ...o0i. W Population (million) 0 GDP/Capita (1992 USS') China's Much more Populous Interior has Half the Coastal Productivity. The front columns in this diagram show that China's most populous regions are in the interior, and that the South Coast and East Coast regions have relatively small populations. The back columns show that the largest interior populations also have the lowest GDP per capita levels,' averaging roughly half the levels in coastal regions. Map: For a map of these regions, see Map 1, page v. Sourcc: China State Statistical Bureau, official data. For details, see Chapter 2. only 294 dollars' (with a population of 226 million--equivalent to the population of the two south-east coastal regions combined). See Figure 1. 5. Over time, real measures of both national output and household income show faster growth in coastal regions and worsening disparities between coast and interior. This trend is especially clear for the current period of economic expansion, which began in 1991-92, but it 'Dollar GDP values in this report are based on official data for total GDP, adjusted for domestic underreporting and non- market pricing and converted to dollars at a trade-weighted average of China's two legal exchange rates in 1992 using the World Bank Atlas method in this way, per-capita GDP for all of China in 1992 was 470 dollars (see World Bank 1994, China: GNP per Capita). - vii - FIGURE 2: INVESTMENT ON THE COAST OVERSHADOWS WEAK INTERIOR EFFORT us S --- - per Cap' 193 1s6 200 19 8 Investment 160 12per Capita 140 - 9 120- 100 -53 57 so- - FDI 4o0 per Capita 2o Sto 1s 34 Far North South Central North Es ot West Hinter- Hinter- Core ot East South land land Coast Coast Coast Interior Coast XI.noU. 0 FDI per Capita (USS) 0 Investment per Capita (1992 USS') Low Interior Investment Affects Regional Disparities. Interior investment per capita (back row of columns) is less than a third of the level on the coast. The FDI skew is even more severe. Source: State Statistical Bureau, provincial statistical yearbooks, and Ministry of Foreign EconomicCooperation and Trade. See Chapter 3. 'Note: US dollars for national accounts are consistent with 470 dollars per capita GNP for all China. See Footnote i, p. v also characterized the previous period of expansion in 1987-88. Real income and consumption levels in rural households in the Central Core region actually declined in the 1985-92 period.' 6. The causes of China's regional disparities take several forms. The fundamental cause of China's disparities and their worsening is the natural response of certain regions with location advantages for new commercial and investment opportunities. Furthermore, official government policy in the 1980s also encouraged pilot programs and special economic zones in ways which gave coastal regions additional advantages--over and above their already significant natural ones. More directly, large regional differences in per-capita investment are the clearest cause of large disparities in regional growth and output. Investment in China's densely populated interior regions was barely one fourth of per-capita levels on the coast. See Figure 2 on page vii. 2Serious flooding in the lower Yangtze River valley in 1991 is also a factor, but floods alone are unlikely to explain the 7-year performance of such a large region, with its population of more than 300 million. - viii - 7. Disparities are even starker for foreign direct investment (FDI), which provides not only funding but also modem scientific and management technology. FDI disparities closely match enormous coastal-interior disparities in foreign trade, which are also large along the coast. See Figure 3 on page ix. 8. Budgetary transfers from richer to poorer regions have not been large enough to make a significant impact on investment disparities and are quite small in comparison to the regional deposit and lending flows through banks. The only significant budget surplus is in the East Coast region (from Shanghai's profitable industrial base), and on a per-capita basis most of it transfers to the Far West and the North Hinterland. Very little goes to the poorest and most populous regions-Central Core and South Hinterland. See Figure 4 on page x. 9. The clearest regional issue facing Chinese policy makers is anticipation of migration from poorer interior regions to the coast. Migration has already begun and is known as China's "floating population" because it has a large seasonal component. Reliable statistics on its nationwide scale are unavailable. Survey data from one coastal province (Fujian) indicate the floating population is 10 percent of the provincial total. See Chapter 2. - ix - FIGURE 3: EXPORTS SOARING ON THE COAST SHOW TRANSPORT AND FDI BENEFTrs Billion USS 3 S 45 19 Total Exports S 30 $25 - i $S207 3 FD1- Related SL- Exports Z043 52 f .4 We t South Central North East South Hinter- Hinter- Cor Coast Coast Coast land land Interior Coast xnole 1993 FDI Exports (Billion 5) 0 1993 Total Exports (Billion USS) South Coast Dominates Exports. Customs data show that exports (back row of columns) come heavily from coastal regions (except for oil exports from the North Hinterland). Erport from Foreign Direct Investment enterprises in 1993 (front row of columns) are even more heavily concentrated in the South Coast region. FDI exports from the interior were negligible. Source: China Customs Statistics. See Chapter 3. FIGURE 4: 1992 BUDGET TRANSFERS FROM THE COAST SIGP OVER THE POOREST REGIONS US S U Deficit/Surplus (USS/Capita) 0 Expenditures (USS/Capita) per Capita Per Capita so- 41 Al 4o- _oBudget 30 - Expenditus 2o-- 10 -4 .Io- -8Surplus/ -2o. -22 Deficit Far North South Central North East South West Hinter- Hinter- Core Coast Coast Coast land land Interior Coast Shanghai's Budget Surpluses Go to the Northwest: The surplus from East Coast industrial profits (front row of columns), together with central government budgetary funds, transfer to the Far West and the North Hinterland On a per capita basis, the poorest regions (South Hinterland and Central Core) benefit little and have the lowest per-capita budget expenditures of any region (back row of columns). Source: State Statistical Bureau, Provincial budget statistics; see Chapter 3. 1. REGIONAL INTRODUCTION AND BRIEF LITERATURE SUmmARY 1.1 This chapter presents China by regions and briefly summarizes what other studies have concluded about regional disparities in China. Its major points are: * China has seven identifiable economic regions (please refer to Map 1.2 on page 4). * Three Coastal regions (North, East, and South) all have good ports, but only two, the North and East Coast regions, have transport corridors with access to the interior. * The Far West region has large land area and good mineral resources, but it is inaccessible and has relatively few people because of the inhospitable climate. * In between, hilly-to-mountainous Hinterland regions (North and South) and a Central Core of non-coastal lower-altitude alluvial lands are relatively isolated but densely populated The lower Yangtze River gives the Central Core some coastal access. * Most of the literature on spatial inequality in China has less of a regional focus than an emphasis on measuring whether inequality has increased or decreased in the country as a whole. * Regional considerations in the literature are based almost wholly on comparisons of China's 30 provincial-level units (including municipalities), an approach which detracts from the policy usefulness of the results. * The literature finds that coastal-interior disparities have increaseg and more so for consumption than for output. * Exceptions to these conclusions generally reflect an interest only in "how many" provinces have improved or fallen back, without weighting them by population. * Causal analysis in the literature is limited mostly to "coastal-zone strategy" and market-reform initiatives. * Few studies address the impact of shifts in central-government fiscal redistribution measures, with conflicting conclusions about whether transfers continue to be effective or have deteriorated * Migration pressures emerge as the most important implication of increasing regional disparities. A. ECONOMIC REGIONS IN CHINA 1.2. This section explains the report's division of China into different economic regions. Because data are largely available on a provincial-level basis, each region comprises several provincial-level entities (provinces'). While it is true that natural economic regions cross provincial borders and that one province can belong to two or more distinct economic regions, data and other resource limitations make it necessary to start with provinces when building 3 In addition to provinces proper, China has three municipalities with provincial-level status and six autonomous regions, where most ethnic minority populations are concentrated (e.g. Inner Mongolia Autonomous Region). For convenience, the report frequently refers to all of these as "provinces." -2- economic regions. The report does analyze regional disparities within two provinces-Henan and Sichuan-as an illustration of the variation not captured by national regions. But no attempt is made to link these sub-regions with sub-regions in neighboring provinces or with national regions. MAP 1.1: CHiNA's RIVERS Yellow River Yangtze River China: Major Riven Pearl River Ocean-., Note: This map is a rough free-hand drawing useful only for notional regional identification. Its boundaries, denominations, and any other information shown do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 1.3. Chinese statistics and academic analysts have a variety of schemes for dividing China into economic regions, but the most usual approaches have limited analytical usefulness for policy purposes. At one extreme, China's 30 provincial-level administrative units are too numerous and heterogeneous for national regional analysis, while at the other extreme, the official regional grouping of all provinces into three "belts" (coastal, central and western) is too broad for all but the most general analysis. Use of provincial-level data leads to inappropriate comparisons, such as between Beijing, Tianjin, and Shanghai metropolitan areas on the one hand and large complex provinces like Sichuan and Guangdong. Use of the three regional belts obscures the disparities between significantly different parts of the country, such as the south coastal region around Hong Kong and the east coast, centered on Shanghai. Both of these approaches, however, have the advantage that data are readily available by these categories, making analysis convenient. 1.4. Geography and climate are the ultimate bases for China's economic regions. Beginning on the Pacific coast in the East, much of China stretches west along major river systems (the Yellow, Yangtze and Pearl rivers) through low-elevation alluvial land into upper- river watersheds until reaching the far western regions, which are high-elevation and in many parts extremely mountainous. Over the millennia, most of China's population has settled in the alluvial -3- areas, where rainfall, lower elevation and moderate temperatures support denser populations through agriculture and commerce. Mountainous, arid, and colder regions to the west and northwest are sparsely populated, but some locations are rich in minerals. 1.5. Although China's long coastline stretches from Korea to Vietnam with many fine harbors, access to the interior is uneven. Roughly speaking, North of the Yangtze delta and Shanghai, coastal areas enjoy easy access to fertile plains, but from just South of Shanghai all the way to Vietnam, coastal mountain ranges generally block access to the rest of China. On the South Coast, the Pearl River (at Hong Kong and Guangzhou) leads west from its delta near Hong Kong, rather than north into China's interior (see Map 1.1 on page 2 for China's most important rivers). 1.6. Rainfall patterns also vary systematically, beginning with generous monsoons in the south and south-east regions, shifting into lower and less consistent rainfall patterns to the north and west. Drought and monsoon irregularity are frequent north of the Yangtze River, and to the far north-away from the coast and in the far west--China has a largely semi-arid if not desert climate, inhospitable to agriomlture. Overall, mountains, fertile land, rainfall, rivers, and outlets to the sea have combined with agricultural and commercial technology over the millennia to shape China's economic regions in ways that favor the southern and eastern provinces over northern and western areas. 1.7. The regions chosen for this report are designed to facilitate economic analysis in light of the interaction between climate, geography, and technology. Technology is especially important for shaping economic regions, because it determines the nature of economic activity and the degree to which a region can support its population density. In China's case, the transition from agriculture to industry and the shift from inward-oriented self-sufficiency to international commercial integration is changing the economic significance of China's various regions. These transitions also influence the patterns of regional disparities which emerge from statistical analysis. 1.8. Overall, the report's major analysis depends on seven national economic regions, several of which are subdivisions of larger regional categories (see Table 1.1, Map 1.2 on page 4 and Map 1.3 on page 5). For example the single coastal category has three regions--North Coast, East Coast and South Coast. The inland Central Core is treated as one region, but the hilly-to- mountainous hinterland category is divided into two regions--North Hinterland and South TABLE 1.1: CHINA'S EcoNoMIC REGIONS Region Provinces, Municipalities and Autonomous Regions Far West Xin.iang, Tibet, inghai Gansu Ningxia North Hinterland Heilongjiang, Jilin, Inner Mongolia, Shanxi, Shaanxi South Hinterland Sichuan, Guizhou, Yunnan, Guangxi Central Core Henan, Anhui, Jiangxi, Hubei, Hunan North Coast Liaoning, Hebei (including Beijing and Tianjin), Shandong East Coast Jiangsu (including Shanghai), Zhejiang South Coast Fujian, Guangdong, Hainan -4- MAP 1.2: COAsTAL AND INTERIOR REGIONS Russia. Russia M ongfi Korea Coina Far West Shanghai FONorth Hinterland South Hinterland / ian Central Core Hong North Coast Vi Kong Pa ic East Coast LOcean South Coast ' China's Economic Regions. This map presents economic regions which highlight China's poorest and richest areas. The South Hinterland and Central Core are poorest, with 531 million people at roughly 320 dollars average per-capita GDP. The East Coast and South Coast are richest, with 228 million persons at 740 dollars per capita. See Figure 1.1 and more detailed analysis in Chapters 2 and 3. Source: For more geographical detail, see Map 1.3 on page 5. Note: This map is a rough free-hand drawing useful only for notional regional identification. Its boundaries, denominations, and any other information shown do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. Hinterland. The seventh region is the Far West, with its vast expanse but small population. 1.9. In some cases it is difficult to determine where to place a provincial-level entity. The two most difficult cases of regional ambiguity are Inner Mongolia in the North and Guangxi in the South (see Map 1.3). Inner Mongolia extends from China's northeast to its far western grasslands and deserts. However, since much of its population is along its southeast border with the rest of China or near the upper reaches of the Yellow River, this report considers Inner Mongolia part of the hinterland regions rather than part of the Far West. Guangxi Autonomous Region appears to be coastal, but much of its population is cut off from the sea by hills and mountains and is concentrated along the upper reaches of the Pearl River, flowing East into Guangdong. Also the size of Guangxi's ethnic minority population further argues for Guangxi's inclusion in the hinterland category rather than the coastal category (see Table 1.1 for the provincial-level components of each region). 1.10. Each region has distinctive economic characteristics: * The Far West is largely isolated from the rest of China and is commercially oriented toward central Asia. It is rich in minerals, but its difficult climate severely limits farming. Expansive grasslands are a significant resource supporting a large animal husbandry component. 彥 必 寶 & 的 閂 n O Z O 昌 n 胃 O 日 Z 的 絕 韋 曾 O 髦 田 n 》 州 中 閃 呵 誅 邑 j 口 -6- TABLE 1.2: 1990 POPULATION BY REGION (Million Persons) 1990 1992 Region Census Official Interior Regions 710 729 Far West 49 51 North Hinterland 144 147 South Hinterland 220 226 Central Core 297 305 Coastal Regions 430 437 North Coast 206 209 East Coast 123 125 South Coast 101 103 China Total 1,140 1,166 Source: 1990 population census and SSB, Statistical Yearbook, various years. " The North Hinterland stretches from the upper reaches of the Liao River in continental Manchuria across Inner Mongolia and south through the mountainous areas straddling the upper reaches of the Yellow River. It has China's most important energy extraction centers (coal and petroleum) and related heavy industrial centers, but both climate and geography limit agriculture. " The South Hinterland encompasses the upper reaches of the Yangtze and Pearl River systems. Southern temperatures and good rainfall compensate for the large share of hilly and mountainous terrain surrounding large pockets of rich arable land (especially the Sichuan plain). It has significant concentrations of non-Han-Chinese ethnic minority populations, and like the North Hinterland, sustains Maoist-era heavy-industrial complexes concentrated there for strategic reasons. " The Central Core contains China's non-coastal alluvial lands--essentially those in the lower non-coastal reaches of the Yellow and Yangtze River systems. Rainfall is generally good, although it can weaken along the Yellow River, and agriculture is highly productive, although monsoon flooding is a regular concern. The navigable Yangtze is an important transportation link, supporting Wuhan, one of China's earliest industrial centers. " The North Coast sweeps from the Liaoning Peninsula to the Shandong Peninsula, with the Yellow River delta and many fine harbors, including Dalian, Tianjin and Qingdao. Rainfall is fair, and drought is a regular problem, affecting both agriculture and urban water supplies. Commerce and industry benefit both from transport conveniences and the influence of national government sponsorship based in Beijing. " The East Coast is centered on the Yangtze Delta and China's largest port of Shanghai. Its excellent soil and rainfall, river access to the deep interior, and Shanghai's fine harbor represent a combination of natural circumstances not enjoyed by other coastal regions. It was pre-reform China's most important industrial center, with a commensurate concentration of skilled engineering and scientific labor. " The South Coast stretches from the Taiwan Straits to Hainan Island and includes most notably the Pearl River delta, with coastal lands adjacent to Hong Kong. It has excellent rainfall, but except for river deltas and narrow valleys, it is a relatively hilly and mountainous -7- region with difficult transport access to the central interior. Its pre-1978 industrial base was not strong. 1.11. According to China's most recent census, the coastal regions account for more than 400 million persons, the Far West roughly 50 million, with a remaining population of 660 million in the Central Core and Hinterland regions (see Table 1.2). The large population in the coastal regions, with their good commercial access to world markets, justifies paying special attention to them as regions separate from the rest of the country. -The even larger population in the Hinterland and Central Core regions, however, recommends equally careful scrutiny. 1.12. In conclusion, China is a vast and highly differentiated continental economy which confounds easy regional division. The seven regions introduced here provide an imperfect but nevertheless useful segmentation for analyzing differences in output, consumption, basic needs provisions, and related economic policy. Furthermore, these regions provide a statistical basis for regional analysis which contrasts with analysis based either on the complete set of 30 provincial- level entities or the official aggregation of these entities into three regional belts. B. ACADEMIC LITERATURE ON CHINA'S REGIONAL DISPARITIES 1.13. Academic literature on China's regional inequality has generally reflected an overall interest in global measures of inequality rather than policy-oriented interregional comparisons. The literature has examined statistical measures of spatial inequality, their constituent components (e.g. industry or agriculture, rural or urban, within or between provinces), their trends over time, and their regional significance-that is, which parts of China are affected in which ways. The data and methods employed, however, vary significantly, and the variation frequently affects research conclusions. (For a fuller discussion of topics in this section, see the literature review in Annex 2.) 1.14 Evaluation of existing literature as a body of research leads to the conclusion that regional disparities in per-capita output, income and consumption have increased in the post-1978 reform period compared to the pre-reform period, and that in the course of post-1978 reforms, the increase before 1985 is not as significant as that after 1985, when the widening of interregional disparities accelerated. See Auguinier (1988), Denny (1991), Lyons (1991), Dali Yang (1989), Weiming Yang (1991), Tsui (1991), and Keidel (1993). Not all research findings in the literature conform uniformly with this conclusion (see Denny 1991 and Weiming Yang 1991), but when data and methods are taken into consideration, it nevertheless appears justified. The most important correlates of these shifts in inequality measures are widening rural-urban gaps for various economic indicators, given uneven regional concentration of cities, differential growth in agriculture and industry-whose output is unevenly distributed by province and region-and special advantages coastal areas enjoy regarding foreign trade and investment. 1.15 Conclusions regarding growing inequalities in consumption are in general better supported by the literature than inequality in output, in light of the availability of disaggregated household survey data which show growing inequality within provinces and regions as well as between them. These studies show that the growing urban-rural gap is even more important for -8- explaining spatial inequality within provinces than between provinces (Lyon 1991), and that official data understate the degree of spatial inequality compared to more carefully designed special survey results (Khan, et al. 1992). Several studies document how regional consumption inequality has become more serious for rural areas over time, but not so much for urban areas, in that coastal areas have experienced a rapid increase in rural household income throughout the reform period, while the increase slowed significantly after 1985 for the rest of the country's rural regions (Khan, et al. and Keidel). 1.16 In sun, the literature's findings are that regional disparities favoring the coastal areas appear significant and have increased markedly since the outset of reforms, that inter- regional consumption disparities for rural areas are particular noteworthy, that poverty is concentrated in particular interior regions, and that although fiscal transfers continue to perform an important corrective role, the regionally differential impact of credit expansion and foreign trade and investment appear overpowering by comparison. 1.17 We can evaluate the literature on China's regional economic disparities by noting that it presents a good description of disparities between provinces, and in a few cases between groupings of provincial-level units, although in most cases the statistics are somewhat out-of-date. The literature identifies important related issues--especially rural-urban differences and uneven industrial expansion--and indicates some of the most obvious factors behind these trends, such as foreign-trade and foreign-investment imbalances by region. The literature is less helpful for considering the policy background relevant for these patterns and trends and applying an evaluative framework. Nor does it do a good job of indicating the implications of these trends for further economic reforms--although several studies draw strong conclusions with regard to their significance for migration pressures. In particular, the literature's interest in migration warrants thorough expansion when considering the significance of emerging regional disparities and the policy directions they imply. -9- 2. REGIONAL DISPARITIES 2.1 This chapter presents the scale and nature of increasing regional disparities in China. Its major conclusions are: * Coastal-Interior disparities are roughly two-to-one and have worsened The greatest disparities are between the Coast regions (437 million persons) and the combined Central Core and Southern Hinterland regions (531 million). * 1992 GDP per capita for the Coast ($679) is more than double that for the combined Central Core and South Hinterland ($316). * Compared to the national average, from 1980 to 1992, the coast significantly increased its lead in per-capita GDP and industrial output, while the interior's relative position uniformly deteriorated from already-low levels. * Within the Coastal regions, the South Coast dramatically improved its GDP- per-capita and industry-per-capita positions vis a vis the East Coast, but only surpassed it for service output. * Regional growth rates for 1978-92 show that all regions shared in the 1983- 85, 1987-88 and 1991-92 booms, but the Coast and especially the South Coast did much better. * Regional household data show that urban income and consumption (in real terms) kept steady growth throughout the 1981-92 period, with the South Coast pulling ahead dramatically; however, rural income and consumption growth decelerated markedly after 1985 and actually declined in the Central Core region in 1985-92. * Selected social indicators by region show significant disparities for literacy, school attendance and infant mortality to the advantage of the coast, and especially the East Coast, although the South Coast has the lowest illiteracy rates. * Within-province disparities in Henan and Sichuan (89 and 109 million persons) appear from limited data to be quite serious, with 99 million persons in regions averaging under $250 per-capita GDP. A. REGIONAL OuTPUT DISPARITIES 2.2 According to official data, regional output disparities in China are stark, have worsened since 1980, and appear to be worsening at an accelerating pace. Disparities stand out because of the large populations in poor interior regions--for example, 531 million in the South-Central interior (the South Hinterland and Central Core regions). Provincial-level statistics aggregated by regions in Table 2.1 show that GDP per capita in 1992 was much higher for the 228 million people in the South Coast and East -10- Coast regions (from Guangdong to Jiangsu provinces). Compared to these coastal areas, the South-Central interior regions have per-capita GDP levels one-half as high. Per-capita industrial output is only one-third as high. These same poorest regions have fallen further behind the national average since 1980, especially in industry, and have shared less in the national growth surges since 1985 than they did in the early and middle 1980s. 2.3 Regional comparisons of total GDP, per-capita GDP, and the sectoral structure of GDP for 1992 appear in Table 2.1. The data show most importantly how much per-capita GDP varies by region and the degree to which the coast dominates the rest of the country (see per-capita data in column 4). There is a quantum disparity between the coastal and non-coastal regions. Each coastal region has more than double the per-capita GDP in the poorest region (the South Hinterland) and none of the interior regions has per-capita output levels anything close to coastal levels. 2.4 For industrial 6DP, disparities are even greater. Interior regions have per- capita industrial output roughly half of what it is on the coast. The South-Central interior regions have especially low per-capita levels of activity; the South Hinterland has barely one fourth the per-capita industrial output on the East Coast. Casual inspection of columns 4 and 5 reveals that per-capita levels of non-industrial output are also higher in coastal regions, a finding which reflects both the higher-valued crops grown on the coast and the much higher coastal levels of per-capita service-sector output. These disparities also appear in data on the structure of output in columns 6, 7 and 8. The two South- Central interior regions have significantly higher shares of GDP output in agriculture, with correspondingly lower shares in industry and services. TABLE 2.1: CHINA'S REGIONAL GDP BY SECTOR, 1992 2 3 4 5 6 7 8 Population. 1992 GDP 1992 GDP per Capita Sector Shares (1992) (Million) (Bil. 1992 US$') (1992 US)()e(Percent) 1992 Total Industry Total Industry Agriculture Industry Services China Total 1,166 548 263 470 226 24.2% 48.1% 27.7% Far West 51 20 8 398 159 27.6% 39.9% 32.5% North Hinterland 147 63 32 430 220 23.1% 51.2% 25.7% SouthHinterland 226 66 26 294 115 35.1% 39.3% 25.5% Central Core 305 101 43 332 140 32.0% 42.3% 25.7% North Coast 209 127 66 608 317 19.2% 52.0% 28.8% East Coast 125 98 56 781 450 16.2% 57.5% 26.2% South Coast 103 72 32 695 307 22.9% 44.2% 32.9% Output and Structural Disparities. Interior regions produce more than 40 percent of GDP and industrial GDP (cols. 2-3), but huge per-capita disparities (cols. 4-5) reflect underlying population differences (coL 1). The East Coast is especially dominant. Structural patterns (cols. 6-8) confirm heavier reliance on agriculture in the largest and poorest interior regions. Based on official data for total GDP adjusted for domestic underreporting and non-market pricing converted to dollars at a trade-weighted average of China's two legal exchange rates in 1992 (see World Bank 1994, China: GNP per Capita); industrial data based on adjusted GDP data and official sector shares (column 7). b Based on official data and reporting. Sources: State Statistical Bureau, Statistical Yearbook 1993; Statistical Abstract 1994. [rg_gdp5e!main] - 11 - FIGURE 2.1: SouTH-EAsT COAST AND SOUTH-CENTRAL INTERIOR DIsPARrrIEs 800-- 743 700- 600- 531 500 - 385 316 300 200 130 100 0 Population ODP/Capita Industry ODP/Capita (million) (1992S) (1992S) E South-East Coast I South-Central Interior Stark Disparities Between Coast and Interior. The South Coast and East Coast regions, with 228 million persons, have less than half the population of the South and Central interior regions (531 million), but they have twice the GDP per capit, and three times the industrial GDP per capita Source: Calculated from data in Table 2.1. 2.5 The clearest disparities are between the South-East coastal regions and the South-Central interior, in that the South-Central population is well more than twice that in the South-East coastal region (531 million vs. 228 million in 1992), but its GDP per capita is less than half and its industrial GDP per capita is barely one third the level on the coast. Figure 2.1 presents these comparisons in graphical form, and Map 2.1 summarizes the location of these regions. Coastal-interior disparities in GDP per capita and especially in industrial output might not be so noteworthy were it not for the very large interior population. The combination of population and output disparities makes it difficult to imagine that the coastal regions, as they urbanize and industrialize, could rapidly absorb the large interior rural population if it tried to move to the coast. 2.6 Certain characteristics of poor interior regions will influence future industrial and urban development, especially the scale of their existing industrial centers, their significant degrees of isolation from coastal industrial areas, and their very large populations. With regard to interior levels of industrial activity, even though industrial output in South-Central regions is relatively low in per-capita terms, in absolute terms these regions have a substantial industrial base (see Table, col. 3). This is in large part because the Central Core region includes the industrial base in Wuhan (Hubei Province), and the South Hinterland region includes the industrial base in Sichuan Province. In other words and to important degrees, China's poorest regions appear to form self-contained economies with better transport and other logistical linkages within the regions than between them and the coastal areas. The weaker outward transport linkages appear to be -12- largely a function of geography--in particular mountain ranges. Both the Central Core and South Hinterland regions are cut off from the South Coast region by mountains, with the exception of some access along the partially navigable Pearl River (see Map 1.1 on page 2). As for access to the East Coast, the relatively low-lying Central Core Region faces few geographical barriers via the Yangtze River basin, but South Hinterland provinces are largely isolated, except for Sichuan's long but potentially important river link, also via the Yangtze. 2.7 In general, the coastal-interior industrialization disparity has worsened significantly since reforms began, especially for the two poorest regions, the Central Core and the South Hinterland. Table 2.2 presents data on the shift in regional disparities between 1980 and 1992. They show that although total GDP per capita differences changed little (except for the South Coast), their structure changed a great deal. Cols. 1 and 2 show that for these two regions, GDP-per-capita deviations from the national average changed hardly at all from 1980 to 1992. Per-capita industrial output (cols 5-6), however, changed dramatically for both poor interior regions. In 1980, their activity level was roughly 10 percent behind the nation as a whole, but by 1992, they produced 38 percent and 48 percent less than the national average, respectively. Given the importance of manufacturing for absorbing surplus labor in these regions, such a comparatively poor industrial performance in the South-Central interior over the course of China's reforms deserves policy-makers' attention. 2.8 In an opposite trend, disparities between different coastal regions have lessened considerably since 1980. By far the most noticeable coastal trend is the narrowing of the large gap between the East Coast and South Coast. In GDP terms, the 1980 gap MAP 2.1: COASTAL, CENTRAL CORE, AND SOUTH HINTERLAND REGIONS China China 4 Economic 2 Hinterland Zones Sub-Zones Contrasting Coast and Interior. The cross-hatched Central Core region in the left-hand map and the highlighted South Hinterland in the right-hand map are China's largest and poorest regions. They contrast starkly with coastal regions highlighted in the left-hand map, especially the South and East Coast regions from Guangdong Province (facing Hong Kong) to Jiangsu Province (including Shanghai) See Figure 2.1. Source: See Map 1.3 on page 5 for a more detailed rendering of China's regions and their constituent provinces. Note: This map is a rough free-hand drawing useful only for notional regional identification. Its boundaries, denominations, and any other information shown do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. -13- TABLE 2.2: CHANGE IN REGIONAL DISPARITIES, BY SECTOR, 1980-92 (Per-capita % Deviation from National Average) 1 2 3 4 5 6 7 8 Total GDP Agriculture GDP Industry GDP Services GDP 1980 1992 1980 1992 1980 1992 1980 1992 China Total 0% 0%1 0% 0% 0% 0% 0% 0% Far West -9% -12%1 6% -5% -11% -28% -13% 0% North Hinterland 5% -10% 2% -12% -9% -3% 14% -15% SouthHinterland -35% -38% -31% -9% -10% -49% -53% -42% Central Core -23% -28% -29% -6% 6% -38% -37% -34% North Coast 29% 28% 24% 3% -3% 39% 50% 35% East Coast 68% 61% 64% 12% 16% 98% 101% 58% South Coast -3% 59% 19% 39% 10% 41% -20% 79% Shifts in Regional Disparities. Based on deviations from the GDP/capita national mean (cols. 1- 2), poorer regions (South Hinterland and Central Core) in 1980-92 only fell slightly more behind the national average. However, the South Coast shifted from below the national average to rival the East Coast. In industry, poor interior regions fell even more decisively behind (cols 5-6). Sources: State Statistical Bureau, Compendium of Historical Provincial-level Statistics 1949-89, Statistical Yearbook, 1993. See Annex for Provincial-level data. lrg_gdp4e!mainj. had all but disappeared, and although the South Coast's improvement in per-capita GDP national standings is broad-based, it is most impressive in the service sector. In particular, in overall GDP terms (cols. 1-2 in Table 2.2), the South Coast increased its per-capita output level from below the national average in 1980 to nearly 60 percent over the national average in 1990, to roughly the same level as the East Coast. In industrial output terms, however, the East Coast maintained its dominance, even though all coastal regions improved their national status significantly (cols. 1-2 in Table 2.2). 2.9 It is worth emphasizing the persistence of a coastal-interior disparity even in agriculture. In agriculture, the trend over the 1980-92 period shows a reduction in disparities, from a pattern of strong per-capita output superiority on the coast in 1980 to one in 1992 in which the coastal lead has narrowed significantly (excepting the South Coast, which doubled its performance deviation to nearly 40 percent over the national average by 1992). Nevertheless, all the coastal regions in 1992 had per-capita farm output above the national average, while interior regions remained at sub-national per-capita production levels. 2.10 For services, the large disparities in 1980 between the East Coast and much of the rest of the country had also narrowed by 1992, but only slightly (see Table, cols. 7- 8), and the gap between coastal and interior regions remained very large, especially between the South-East Coast regions and their neighbors in the South-Central interior (see Map 2.1 on page 12). The most noticeable shift in performance is for the South Coast, which during the course of reforms shifted from a per-capita level of production 20 percent below the national average in 1980 to one nearly 80 percent above in 1992. -14- 2.11 Although the South-Central interior regions are the largest and the poorest, China's two other interior regions--the North Hinterland and Far West--also show per- capita output gaps which in general worsened between 1980 and 1992. The North Hinterland, however, contains many old industrial bases, such as those in Northeast Provinces such as Heilongjiang and Jilin, where China has its largest developed oil field and the descendants of a large Japanese industrial investment program there in the 1930s, focused on metals and machinery. Other old industrial areas are to the west of Beijing near areas rich in coal. Similarly, the sparsely-populated Far West is the focus of significant central-government fiscal transfers to support scientific, strategic, and long- term energy programs (rocket launching sites, nuclear test facilities, and potentially large coal, oil and gas fields). These two regions, therefore, in spite of relative isolation from commerce on the coast, enjoy mitigating policy circumstances which help moderate the gap between their economic performances and those on the coast. Both Tables 2.1 and 2.2 show their per-capita output at intermediate levels between the coast and the South- Central interior. From 1980 to 1992, however, both regions dropped to 10 percent or more below the national GDP-per-capita average (Table 2.2, cols 1-2). Per-capita output in all three GDP sectors also remains at or below the national average, and the North- Hinterland position in both agriculture and service-sector output deteriorated markedly through 1992 (Table 2.2). 2.12 Not only have disparities worsened between 1980 and 1992, the worsening appears to have accelerated after 1985. Table 2.3 presents regional time series for real output growth (Chinese Net Material Product rather than GDP--see note a in Table 2.3). The most important patterns in Table 2.3 show that through 1985, interior regions generally shared rather fully in nationwide growth, but after 1985, their growth rates failed to match those on the coast, especially if one compares the South-Central interior (cols 4- 5) with the South-East coast (cols 7-8). 2.13 Coastal growth superiority appeared in surges. National trends in col. I of Table 2.3 highlight China's three major surges since reforms began in 1978: in 1983-85, in 1987-88, and beginning nationwide in 1992. The clearest comparisons of regional. growth differences before and after 1985 appear in panels A, B, C and E of Table 2.3. During the first surge (panels A and E), growth was generally comparable in coastal and interior regions (except, perhaps, for the Far West). During 1985-92, however, interior regions averaged annual growth generally a percentage point or more below that on the coast (panel E) and even lower when compared with the South-East Coast regions (cols. 7-8). 2.14 Coastal acceleration continued to build in the latter two surge periods (panels B and C in Table 2.3). Coastal growth in 1988 was 1-to-8 percentage points higher than in the interior (panel B), and 3-to-8 percentage points ahead in the South-East coast compared to the South-Central interior (cols. 6-8 and 2-5, respectively). Comparisons from 1991-92 show even more rapid coastal acceleration (panel C), again largely due to trends in South-East coastal regions (cols 7-8), which began their growth surges in 1991, ahead of the rest of the country. In 1992, both South-East coastal regions -15- TABLE 2.3: CHINA'S REAL GROWTH OF NATIONAL INCOME ( MP), BY REGION, 1978-91 1 2 3 4 5 6 7 8 China North South NMP a Far Hinter- Hinter- Central North East South Total West land land Core Coast Coast Coast A. 1979 7.0% 3.9% 6.2% 8.2% 11.4% 7.5% 13.3% 6.0% 1980 6.4% 7.5% 4.0% 7.1% 5.3% 9.0% 7.5% 14.3% 1981 4.9% -1.7% 2.5% 4.5% 7.6% 2.1% 8.2% 12.2% 1982 8.2% 8.1% 10.9% 11.6% 7.4% 5.2% 8.4% 8.6% 1983 10.0% 12.2% 11.8% 7.9% 10.3% 11.1% 9.4% 5.0% 1984 13.6% 15.2% 13.8% 12.9% 14.5% 15.9% 17.5% 15.8% 1985 13.5% 14.9% 8.7% 12.6% 13.8% 13.1% 17.0% 18.0% B. 1986 7.7% 9.6% 6.1% 6.0% 6.8% 5.5% 8.4% 8.3% 1987 10.2% 8.2% 8.4% 9.7% 10.2% 10.9% 9.9% 15.1% 1988 11.3% 10.4% 10.8% 8.6% 7.9% 11.8% 11.9% 16.0% C. 1989 3.6% 6.3% 2.0% 2.9% 3.6% 2.7% 1.2% 7.1% 1990 5.1% 5.7% 4.4% 5.1% 3.1% 3.3% 3.9% 9.2% 1991 7.6% 8.9% 4.7% 7.8% 4.1% 7.9% 10.0% 17.1% 1992 14.4% 9.3% 9.7% 12.7% 14.6% 14.9% 22.6% 22.9% D. 1978-92b 9.0% 9.1% 7.8% 8.4% 8.6% 8.7% 10.2% 12.2% E. 1978-58b 8.7% 8.1% 8.3% 9.0% 9.5% 8.5% 10.8% 10.9% 1985-92b 8.0% 8.1% 6.2% 7.0% 6.4% 7.5% 8.3% 12.9% F. 1978-82b 6.4% 4.1% 5.3% 7.4% 7.6% 5.8% 9.0% 10.8% 1982-88b 11.0% 11.8% 9.5% 9.8% 10.7% 11.2% 12.6% 13.3% 1988-92b 7.4% 7.5% 5.0% 6.9% 5.7% 6.8% 8.7% 13.7% G.NAfP. 19920 1,051.2 37.9 120.2 126.5 206.6 282.4 238.0 130.1 Coastal Growth Outpaces the Interior. All regions had high sustained real growth (row D). However, after 1985 (row E), coastal growth (cols. 6-8) was generally higher than in other regions, especially when compared to the Hinterland and Central Core (cols. 3-5). aGDP data are not available for all provinces and all years m this table. NMP (Net Material Product) is the measure of national output in China's traditional material-product system of national accounts. Unlike GDP, it excludes most services. b Average Growth rates are from logarithmic regression trend lines, hence, 1978-92 figures are not averages of sub-penods ' 1992 NMP in billion current yuan. Sources: State Statistical Bureau, Compendium of Historical Provincial-level Statistics 1949-89, 1993 Statistical Yearbook, an various provincial statistical yearbooks. For provincial data in this format, see the Annex. [p nmpg4e!main_1] had real growth of nearly 23 percent, 8-to-13 percentage points ahead of other regions. The North Coast region, however, showed growth more in line with the national average in 1991-92 (col. 6, panel C) and appears not to have shared in the accelerated coastal growth phenomenon. 2.15 Overall, however, and in spite of accelerating output divergence between regions, the most significant feature of China's regional growth patterns since 1978 may be the generally rapid pace of growth in all regions for more than a decade. Panel D in Table 2.3 shows that for the entire 1978-92 period, no region averaged lower than 7.8- percent real growth. Even at the provincial level, the two lowest average growth rates over this period are 6.5 percent and 7.1 percent--in Heilongjiang and Qinghai, respectively (see provincial-level growth data in Table A.3 of the Annex). Even if China's statistical -16- system systematically exaggerates real growth rates by several percentage points, these are respectable long-term growth rates for any economy, especially since growth-rate exaggeration appears most serious in areas with heavy concentrations of township and village enterprises (TVEs), which tend overwhelmingly to be in coastal regions. 2.16 In sum, regional output disparities in China are significant and appear to be worsening in an accelerated fashion. Coastal regions enjoy significantly higher GDP-per- capita output levels, especially for industrial GDP. These disparities have worsened since 1980, and regional growth rates show that the coastal areas after 1985 have generally enjoyed more rapid growth than regions in the interior. However, generally high growth rates in all regions since reforms began in 1978 indicate that China's regional disparities result not from development failure in the interior, but from initial output disparities (Table 2.2) and from subsequent growth rates inferior to those on the coast (Table 2.3). B. REGIONAL INCOME AND CONSUMPTION DISPARITIES 2.17 China's regional output disparities express themselves as regional income and consumption disparities as well, for both urban and rural households. Overall, the South Coast and East Coast regions enjoy much higher per-capita incomes, expenditure levels and food consumption than other regions. The scale of regional disparities among urban households on the one hand and among rural households on the other are overshadowed by the much larger disparities between urban and rural households in any region. Rural incomes for regions with the highest rural average were still below urban incomes in regions with the poorest urban households, and nationally rural incomes barely TABLE 2.4: URBAN HOUSEHOLD REGIONAL DEVIATION FROM NATIONAL MEAN, 1981-92 (Percent Deviation from National Averages, based on 198 1-Yuan-per-Capita Data) 1 2 3 4 5 6 7 8 9 Urban HH Income Urban HH Expenditure Urban Food Consumption 1981 1985 1992 1981 1985 1992 1981 1985 1992 National Deviation +0% +0% +0% +0% +0% +0% +0% +0% +0% Far West -5% +3% -7% -9% -2% -9% -13% -9% -13% North Hinterland -12% -13% -22% -13% -13% -20% -29% -16% -25% South Hinterland -8% -6% +0% -6% -1% +2% -2% -1% +4% Central Core -5% -10% -10% -4% -8% -9% -4% -8% -8% North Coast +7% +4% +4% +7% +4% +4% +2% +4% +2% East Coast +16% +23% +21% +16% +22% +19% +14% +20% +21% South Coast -2% -1% +43% -5% +2% +40% +7% +12% +40% Urban Average' 436 591 790 428 572 721 246 301 384 Three Exceptions to Modest Urban Household Disparities. Except for high urban incomes in the South and East Coast regions and unusually low incomes in the North Hinterland, col. 3 shows that urban incomes do not vary more than 10 percent from the national average. South Coast's urban income radicallv changed its ranking during 1985-92. a Per-capita urban household income, expenditure and food consumption, 1981, 1985, 1992, at constant 1981 prices. -17- TABLE 2.5: RURAL HOUSEHOLD REGIONAL DEVIATION FROM NATIONAL MEAN, 1981-92 (Percent Deviation from National Ave ages, based on 1981-Yuan-per-Capita Data) 1 2 3 4 5 6 7 8 9 Rural HH Income Rural HH Expenditure Rural Food Consumption 1981 1985 1992 1981 1985 1992 1981 1985 1992 National Deviation +0% +0% +0% +0% +0% +0% +0% +0% +0% Far West -26% -23% -28% -30% -26% -27% -23% -21% -24% North Hinterland -12% -13% -17% -13% -16% -21% -11% -15% -15% South Hinterland -14% -22% -24% -15% -18% -18% -4% -6% -10% Central Core -5% -6% -19% -7% -7% -14% 4% -2% -10% North Coast -1% +5% +2% -4% +2% +1% -16% -8% -6% East Coast +14% +26% +43% +25% +33% +50% +18% +21% +38% South Coast +14% +14% +38% +14% +14% +37% +16% +23% +35% Rural Average 242 356 404 200 290 331 118 164 186 Large Regional Disparities in Rural Income. Regional disparities in rural household income between China's South-East coastal regions and all of the interior (col 3) are larger than for urban households (see Table 2.4) and seem to have worsened in the Central Core and North Hinterland regions during 1981-92. * Per-capita anral household income, expenditure and food consumption, 1981, 1985, 1992 at constant 1981 prices. averaged half the national urban level. Furthermore, interregional disparities in 1992 were larger for rural households than for urban households, although household disparity patterns apparently did not change very much from 1981 to 1992. 2.18 For urban households, the South-East coastal regions enjoy significantly higher incomes than urban households in other parts of the country, which conforms with the lead those regions have in per-capita output. However, by far the poorest urban households are in the North Hinterland region, with its old industrial bases, and not in the generally low-productivity South-Central interior regions. The 1992 pattern of regional disparities in urban household income appears in col. 3 of Table 2.4, which shows the degree to which each region's urban household income deviated from the national average. In general, with the exceptions mentioned above, no regions vary from the national mean by more than 10 percent. This same observation holds for urban household expenditures and food consumption levels (shown in cols. 6 and 9 of Table 2.4). 2.19 Table 2.4 also shows that with two notable exceptions, these patterns of disparities in urban household income and spending have changed very little since 1981. Cols. 1-3 in Table 2.4 show that urban deviations from the national mean have hardly changed at all, except for North Hinterland and the South Coast. North Hinterland's urban incomes deteriorated rapidly from 1985 to 1992, just when South Coast surged from virtually no difference to a position more than 40 percent higher than the national average. These two exceptions aside, such urban household financial stability characterizes expenditures and food consumption as well (cols. 4-6 and 7-8, respectively). 2.20 For rural households, incomes nationwide average roughly half those in the cities (see the bottom rows in Table 2.4 and Table 2.5), and rural interregional disparities are -18- larger than for urban households. Col. 3 in Table 2.5 shows that the South- and East- Coast regions enjoy incomes significantly higher than the national average. Rural income levels in the East Coast region are especially elevated, much more elevated above the national average than urban incomes in the same region (see Table 2.4, col. 3). Such high rural incomes in the South-East coastal regions are matched by relatively low rural incomes throughout the interior, with lowest incomes in the South Hinterland and Far West regions. Household expenditure and food consumption patterns (Table 2.5, cols. 6 and 9) in general mirror regional income disparities (col. 3). Based on these income data, urban-rural income disparities as measured by ratios are also worse in the interior than on the coast, while urban-rural food disparities are not, especially when measured in absolute terms (see Tables A.4 and A.5 in the Annex for relevant data). 2.21 Over time, regional disparities in rural-household income have worsened significantly. In particular, as South- and East-Coast regional incomes improved during 1981-92 relative to the national average, those in the interior all deteriorated by the same measure (see Table 2.5, cols. 1-3). Central-Core rural incomes appear especially to have lagged behind the national trend, and although many interior regions kept up the national pace in terms of food consumption, rural households in the South-Hinterland and Central- Core regions did not (see cols. 7-9). 2.22 Growth rates showing real changes for benchmark years during this period confirm worsening interregional disparities for rural household incomes, but the pattern is not so clear for urban incomes. For most regions, real growth of urban income was quite high, especially in the South-Hinterland region (see col. I in Table 2.6); the only notable exception is the North Hinterland, which together with the Far West showed especially slow income, expenditure and food-consumption growth in 1985-92 (see Table 2.6, cols. 3, 6 and 9). For rural areas in 1981-92 (Table 2.7, col. 1), the slower growth of household income in the South-Central interior (the South Hinterland and Central Core TABLE 2.6: URBAN HOUSEHOLD CHANGE IN INCOME AND EXPENDITURE, 1981-92 (Percent Real Change in Per-ca lita Flows, 1981 Base-year Prices) 1 2 3 4 5 6 7 8 9 Urban HH Income Urban HH Expenditure Urban Food Consumption 1981-92 1981-85 1985-92 1981-92 1981-85 1985-92 1981-92 1981-85 1985-92 China Average 81% 36% 34% 68% 34% 26% 56% 22% 28% Far West 78% 47% 21% 68% 43% 18% 56% 28% 22% North Hinterland 62% 34% 21% 57% 33% 17% 65% 44% 15% South Hinterland 97% 38% 43% 82% 41% 29% 66% 24% 34% Central Core 71% 29% 33% 60% 28% 25% 48% 17% 27% North Coast 77% 33% 34% 64% 30% 26% 56% 24% 26% East Coast 90% 44% 32% 72% 39% 24% 65% 29% 28% South Coast 163% 37% 91% 149% 44% 74% 104% 28% 60% Growth in Urban Income Decelerates only Moderately. Urban incomes, spending, and food consumption in all regions (cols. 2-3, 5-6, and 8-9) had strong growth after 1985, and, except for the North Hinterland and Far West, grew much faster than rural income and spending. Source: Household Survey data and provincial consumer pnce deflators -19- TABLE 2.7: RURAL HOUSEHOLD CHANGE IN INCOME AND EXPENDITURE, 1981-92 (Percent Real Change in Per-capita Flows, 1981 Base-ye r Prices) 1 2 3 4 5 6 7 8 9 Rural HH Income Rural HH Expenditure Rural Food Consumption 1981-92 1981-85 1985-92 1981-92 1981-85 1985-92 1981-92 1981-85 1985-92 China Average 67% 48% 13% 65% 45% 14% 58% 39% 13% Far West 62% 54% 5% 72% 53% 12% 55% 42% 9% North Hinterland 58% 45% 10% 51% 41% 7% 50% 33% 13% South Hinterland 47% 33% 11% 59% 40% 14% 49% 36% 9% Central Core 43% 45% -2% 53% 46% 5% 50% 43% 5% North Coast 72% 56% 10% 75% 53% 14% 77% 52% 16% East Coast 109% 63% 28% 100% 55% 29% 84% 43% 29% South Coast 103% 48% 37% 100% 45% 38% 84% 47% 25% Sharp Deceleration in Rural Household Income Growth. Cols. 2-3, 5-6, and 8-9 show that slowed real growth in rural income, expenditure, and food consumption was especially serious in poor interior areas. Central-Core rural real income declined in 1985-92 (cot. 3). Source: Household Survey data and provincial consumer price deflators regions) is not evident for expenditure and food-consumption trends (cols. 6 and 9), implying perhaps that rural households in other areas have higher rates of saving. 2.23 Growth rates in Tables 2.6 and 2.7 also confirm that in general urban-rural differences worsened more quickly in interior regions than on the coast, and in some coastal regions--the North and East Coast--urban-rural differences actually declined for household expenditures and food consumption (see cols. 1,4 and 7 in Tables 2.6 and 2.7). The best example is the South Hinterland region, where real growth in urban incomes was much higher than for rural incomes. The same is true for expenditures and food. Urban- rural differences in South Hinterland must have worsened rapidly. As for the South Coast, even though urban incomes, spending, and food consumption all grew faster in real terms than they did for rural households, the real growth was so rapid in both rural and urban areas that worsening disparities as measured by urban-rural ratios appear to be less of a concern. 2.24 In conclusion, rural household disparities show larger and more seriously worsening disparities than urban households, although there is clear deterioration in relative standing for urban households in the North Hinterland region. Deceleration and even stagnation in rural household incomes and spending after 1985 are significant, and rural-urban disparities within regions appear to be more serious than disparities between regions for either rural or urban households. These findings therefore direct attention to the condition of rural households in all interior regions and to that for urban households in the North Hinterland. - 20 - C. REGIONAL SOCIAL INDICATOR DISPARITIES 2.25 Regional disparities in selected social indicators correlate roughly with disparities in output and household income. Literacy rates, school attendance, and infant mortality are all significantly worse in poorer interior regions and better on the coast. However, causality is difficult to determine, because the worst indicators also correlate with high population shares in national minorities (non-Han-Chinese nationalities in provincial-level Regions); cultural factors may also be important. 2.26 Shares of minority populations are highest in the Far West and South Hinterland regions, which also show the highest rates of infant mortality. Table 2.8 presents regional demographic and infant-mortality statistics based on China's most recent (1990) population census. The data do not reveal significant patterns in natural increase rates (col. 2), because faster population growth appears both in the fastest-growing coastal region (South Coast) as well as in poorer interior regions (Central Core and Far West). However, natural increase rates are not especially high in any region; the highest is 1.7 percent. 2.27 Infant mortality patterns show clear distinctions between coast and interior and between poorer and richer regions (col. 4), and although their correlation with minority population shares appears close, correlation with poverty in non-minority regions is also strong. Coastal regions have the lowest infant mortality rates, but within coastal regions, the poorest of the three has the best rate of all, and the worst rate is in the South Coast region, which has grown most rapidly since 1985. Among interior regions, the worst infant mortality rates of all are in poor regions with large minority populations, but the relatively poor Central Core region has an insignificant minority population and a significantly worse infant-mortality rate compared to the coast, while the somewhat better- off North Hinterland region has a rate only marginally different from that in the South TABLE 2.8: REGIONAL POPULATION AND INFANT MORTALITY, 1990 1 2 3 4 Census Population Minority Population Natural Rate Population Infant 1990 of Increase Share, 1990 Mortality (Million) (Percent) (Percent) (per thousand) China Total 1,140 1.4% 8.1% 34.5 Far West 49 1.6% 34.0% 52.8 North Hinterland 144 1.4% -6.3% 26.6 South Hinterland 220 1.3% 20.6% 60.1 Central Core 297 1.7% 2.8% 36.4 North Coast 205 1.1% 4.8% 16.0 East Coast 122 1.1% .3% 19.5 South Coast 100 1.7% 1.9% 24.8 Infant Mortality Higher in Poor Minority Regions. Ifant mortality is highest in the Far West and South Hinterland, where ethnic minority shares are also highest (cols. 3-4). The poorer Central Core rate is also high. Sources: State Statistical Bureau Yearbook, 1992 and Population Census 10% Survey, 1990. See annex for provincial-level data. [rgpople - 21 - Coast region. In other words, both poverty and minority populations in the interior appear significant for explaining disparities in infant mortality, which in turn frequently reflects a wide variety of other social indicators, such as nutrition, health care, shelter, water supply, and status of women. 2.28 Educational indicators also reflect economic disparities between regions, but again, the correlation is not clear-cut because of the relatively good performance of the North Hinterland region. Literacy and school attendance are both worse in poorer interior regions, especially those with larger minority population shares. Table 2.9 (col. 1) shows that compared to an adult illiteracy rate of 18-to-22 percent on the coast, its range in interior regions is 19-to-25 percent (excluding the Far West). These are not large differences, but the poorer and larger regions (Central Core and South Hinterland) have the worst rates, and the better-off North Hinterland has the second best illiteracy rate in the country. The relatively high illiteracy rate in the Far West probably reflects large ethnic minority populations. 2.29 Just as functional illiteracy rates show only modest interregional disparities, data on school attendance in Table 2.9 (cols. 3-4) also show little variation in primary- school attendance except in regions with large minority populations. With these significant exceptions, therefore, China's delivery of elementary school services is regionally relatively uniform. For high-school education (col. 4), however, the attendance generally rates favor the coast, although once again, the South Coast has a less impressive record-high-school attendance rates are lower there than for the North Hinterland, which has a high-school attendance record more similar to those on the coast than elsewhere in the interior. In the poorest interior regions (Central Core, South Hinterland and Far West), high school attendance is significantly lower. This interregional disparity in high- TABLE 2.9: REGIONAL EDUCATION INDICATORS, 1990 1 2 3 4 Rate of Functional Share of Children High School Illiteracy - Over Population under 6-14 yrs old Students % of 14 yrs Old 15 years of age In School Age Cohort (Percent) (Percent) (Percent) (Percent) China Total 22.3% 28.2% 81.0% 17.1% Far West 34.2% 30.2% 69.7% 13.4% North Hinterland 18.5% 28.9% 84.1% 21.6% South Hinterland 25.2% 28.7% 73.5% 9.3% Central Core 23.9% 29.2% 82.5% 12.8% North Coast 19.2% 27.1% 86.6% 26.5% East Coast 21.7% 23.1% 84.9% 31.5% South Coast 17.9% 30.8% 80.1% 15.2% Educational Disparities are Mixed Basic Education disparities are smaller than for high- school education (cols. 3-4), and coastal-interior differences are weaker, with the Northern Hinterland outperforming the South Coast. Ethnic minority regions fare worst. Sources State Statistical Bureau Yearbook, 1992 and Population Census 10% Survey, 1990 See annex for provincial- level data. - 22- school education is perhaps the most significan for future economic development in these various regions, because a more mature educat: :nal background will almost certainly facilitate higher-productivity employment and e isier migration to future concentrations of such jobs in coastal cities. 2.30 In sum, although basic educatio, tal services are regionally relatively uniform, China's coastal and North-Hinterland regions have better health conditions and better access to opportunities in middle and hig ier education, which in turn should support future productivity gains ahead of thos - in the rest of the country. D. INTRA-PROVINC:AL DISPARITIES 2.31 Disparities within regions and p ovinces are as great if not greater than disparities between regions, based on analysis cf two provinces-Henan in the Central Core region and Sichuan in the Southern Hinterland. These two provinces have populations of roughly 90 million and 110 million, respectively (see Tables 2.10 and 2.11). Just as these provinces have average per-capita GDP levels well below the national average, each has large sub-regions of tens of millions of persons with per-capita GDP levels well below the provincial averages. The most readily available social indicator, doctors per 10,000 persons, also shows that very large poor sub-provincial regions have significantly limited access to health care. Since the sub-provincial regions appearing in Tables 2.10 and 2.11 have large expanses of agricultural and urban land and populations in excess of a large number of countries, they are in some ways regions of China themselves. Disparities within these provinces shows that interregional disparities in China are more significant than interregional data alone reveal. TABLE 2.10: HENAN REGIONAL DISPARITIES, 1992 Popu- GDP per Doctorsa lation Capita per (million) (US Sb) 10,000 China Total 1,166 $470 12 Henan Province Total 89 314 11 1 - North River 18 412 14 2 - Central 21 410 15 3 - East 17 201 8 4 - Southeast 15 173 8 5-South 10 263 8 6 - West 8 409 15 Henan's Poorest Subregions. The three poorest (3, 4 and 5) have 42 million persons and average GDP per capita under 210 dollars. b a "Doctors" here refers to what Chinese statistical yearbooks call "Western Senior and Junior Doctors"; it excludes doctors trained only in traditional Chinese medicine. b Statistically adjusted and converted from yuan to US dollars at a trade-weighted exchange rate consistent with the national average of 470 dollars per capita for 1992. Sources provincial yearbooks and SSB Statistical Yearbooks, various years. -23 - 2.32 Disparities between subregions in Henan Province follow natural barriers and transportation lines, especially in terms of the province's major river systems-the Yellow River and the Huai River. The province's three poorest regions have 42 million persons and taken together average under 210 dollars per-capita GDP (Table 2.10, subregions 3, 4 and 5). They are located in south and southeast Henan, deep in the North China Plain's rural reaches, away from the railway and highway hub at the provincial capital of Zhengzhou in the provincial northwest (see Map 2.2). Their closest neighbors are other poor rural provincial subregions in northwestern Anhui and northern Hubei Provinces. Table 2.10 shows that these three Henan subregions also have much weaker access to healthcare, with roughly half the number of physicians per-capita as the three much better-off subregions to the north and northwest. 2.33 As for the three better-off Henan subregions, Table 2.10 shows that even though their populations (47 million) average roughly double the per-capita GDP and double the access to physicians in poorer subregions, their GDP-per-capita levels are still far below the national average, not to mention the much higher average in the coastal regions. In sum, in one of the poorest regions--the Central Core-in one of its poorest provinces, Henan, enormous disparities describe differences between vast contiguous subregions with more than 40 million persons each, such that average per-capita productivity in one is double that in the other. 2.34 For Sichuan Province, China's most populous province, the sub-regional per-capita-GDP disparities are even greater than for Henan (see Table 2.11). The poorest subregions (1, 2, 4, 6 and 9 in Table 2.11) altogether have 67 million persons with an MAP 2.2: REGIONS IN HENAN PROVINCE .ource: Base .: F n L s epbi o dat in.. Tab 2.10. on... pae. Noe This..... m is ar g fealu.: inifa Soudgen Bson eg stagufages Peior, tlof tny enopesen raeplice of h 198,undais da . ... Tabl 2 . on. page 22... . . ..... . No: This .. ma is. a. rog rebn rw sfl nyfrntoaxeinlidniiain t onais deoiaios.n.ayohr.no.tinson.o.ti.l,o.tepr.o h ordBn Gop n judgment~ ~ ~ ~ ~ ~ ~~~~~~~. on. the.. lea sttso.n ertr,o n noseeto etneo uhbude.... -24 - overall average GDP per capita of only 215 dollars. The remaining 42 million persons average much higher GDP per head (458 dollars), which is still under the national average, not to mention coastal levels. Furthermore, even among these better-off regions, the data show significant urban-rural disparities between urban regions around Chongqing and Chengdu (5 and 7 in Table 2.11) and non-urban subregions in the western end of the province (rows 3 and 8). Data from Sichuan thus show more clearly than for Henan that urban-rural disparities are important components of sub-regional spatial disparities. 2.35 In Sichuan, just as in Henan, healthcare disparities between subregions correlate highly with GDP-per-capita disparities. In two mountainous subregions with a total of 16 million people and per-capita GDP under 210 dollars (Table 2.11, rows 2 and 9), there are barely any doctors at all. Other low-income regions generally have less than half the doctors per capita in the wealthier metropolitan regions, although not significantly less than the national average of 12 per ten thousand. Again, to the degree spatial regions correlate with urban-rural differences, interpretation of these disparities is less straightforward. Large urban-rural healthcare disparities are common in many countries, and a statistic like doctors per ten thousand persons may show disparities larger than would more public-health-oriented measures. MAP 2.3: REGIONS IN SICHUAN PROVINCE .Y:x .. .. .. :'12408 . ..... ... .. ......- Source: Based on: Foreign Languages Press, Atlas of the People's Republic of China, 1989, and data in Table 2.11 on page 25. Note: This map is a roughi free-hand drawing useful only for notional regional identification. Its boundaries, denominations, and any other information shown do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. - 25 - TABLE 2.11: SICHUAN REGIONAL DISPARITIES, 1992 Popu- GDP per Doctors' lation Capita per (million) (USS*) 10,000 China Total 1,166 $470 12 Sichuan Total 109 309 12 1 - East Mountains 8 170 11 2 - North Mountains 10 190 1 3 - West Minority 6 353 8 4 - Northern Basin 22 221 9 5 - Chongqing Basin 15 462 20 6 - Central Basin 21 238 10 7- Chendu Basin 13 590 23 8 - Western Basin 8 317 14 9 - South Mountains 6 207 2 Sichuan Subregions Show Urban-Rural Gap. Sichuan's poorest regions (rows 1,2,4,6,9) have 67 million persons and 215 dollars" average GDP per capita. The two richest (5,7) are urban centers. S'Doctors" here refers to what Chinese statistical yearbooks call "Western Senior and Junior Doctors"; it excludes doctors trained only in traditional Chinese medicine. b Adjusted and converted to dollars consistent with the national average of 470 dollars. Sources: provincial yearbooks and SSB Statistical Yearbooks, various years. E. COMPARISONS wrrIH INDIAN REGIONAL DISPARITIES 2.36 A comparison of China's regional disparities with India's in 1990-91 reveals striking similarities. Each country has densely populated interior regions isolated from the coast with very low per-capita GDP output. At the same time, India also has coastal regions, especially the West Coast centered on the port of Bombay, with nearly double the per-capita GDP found in the poorer interior. Figure 2.2 and Table 2.12 show that the Middle Ganges region, with 235 milion persons but a per-capita income of only 192 dollars at official exchange rates, has barely half the GDP per capita of the West Coast region around Bombay, with 120 million persons but 374 dollars per person. 2.37 In India's case, the regional configuration is affected by special circumstances. For example, in the upper reaches of the Ganges and Indus Rivers, near the Indian border at a particularly productive region in Pakistan, the Punjab has many characteristics of a coastal region in China. At the other extreme, because of the sparsely populated regions adjacent to the Himalayan mountain range and because of the country's boundary's relationship with Bangladesh, several small regions in the Far East and Far North do not belong with their larger adjacent regions and have been left separate. 2.38 Most significantly, however, in 1990-91, India's program of economic reforms emphasizing entrepreneurship and more market-oriented relationships both domesticaly and internationally, had not yet begun. Many of the coastal areas show only -26 - slightly higher per-capita productivity compared to the national average and to some interior regions. It will be interesting to monitor India's regional disparities as data on the reform periods become available. FIGURE 2.2: INDIA's REGIONAL DISPARYTIES 450 374 40 1990-91 350 GDP per Capita 300- 222527 234 (SM n 192 226 235 2502- Source: See Table 2.12. TABLE 2.12: INDIA REGIONAL DISPARITIES Population GDP/Capita (Million) (US Dollars) Total 844 257 Greater Punjab 37 442 West Coast _ __ 120 374 Middle Ganges 235 192 West-Central 110 229 Southwest Coast 75 253 Southeast Coast 122 270 Northeast Coast 100 234 Far East 32 226 Far North 13 235 Source: Staff Calculations from Government of India, Central Statistical Office statistics on state-level GDP per capita and population; end-1990 exchange rate from IMF, International Financial Statistics. -27- F. RECOMMENDATIONS 2.39 Investing in Capabilities for Regional Analysis. China's framework of regional policy goals requires significant investment in flexible analysis of the nature, causes, impact, and likely tendencies of regional disparities, based on an ever-improving statistical and qualitative information base. The following recommendations are intended to encourage rapid improvement in China's measurement and analysis of its regional disparities and their underlying causes. (a) Strengthen local national accounting. Significant resources should be added to national and local statistical offices to improve the quality and comparability of local GDP statistics and related statistics, such as employment and inflation. Ultimately, the shifting structure of local economies will explain regional trends in economic fundamentals. These data need to be collected on a locational basis, not on the old rural-urban basis (nongye renkou-feinongve renkou basis). Hence, local statistical offices need reform and investments along these lines. This should begin immediately-or, if already begun, speeded up significantly. (b) Make major household survey reforms. China needs to accelerate its planning and implementation of reforms to the household survey system so it covers all households with one survey, regardless of whether the household is so-called "Agricultural," "Non-agricultural," or something else (a second type of urban registration, migrants, or any other status). Without fundamental survey reform as soon as possible, China's regional realignment policies will be frequently misinformed-and frequently in error. (c) Collect and report regional prices. China should accelerate funding of regional price comparators, especially regional cost-of-living differences, among other things to reduce misinterpretation of statistics on regional disparities reported without price corrections. (d) Identify new qualitative patterns. China should fund an increase in qualitative research based on open-ended investigations and non-leading survey elements to identify new household, enterprise, finance, and government activities for inclusion in statistical and causal analysis. Government should fund contracts to relevant agencies' research bodies, to CASS, and to other researchers to compete for designs and implementation of such qualitative, exploratory, non-leading investigations, for the purpose of enhancing standard data collection methods. (e) Focus on migration. Statistical and analytical efforts by a variety of agents should intensify their focus on migration-seasonal and permanent-to identify all the relevant incentives, costs, contributions, and various social and economic side-effects involved. (f) Assess ability to pay. An especially important quantitative and qualitative research effort is the kind which assesses a local (rural or urban) population's ability and willingness to pay for a range of public and quasi-public services and utilities. Such research has two purposes: -28- (a) to determine what a local population is already paying for water, electricity, sanitation, and other similar services-including costs in lost time for family members and (b) how better delivery of such services can be financed by community cost-recovery institutions. Government should fund the design and implementation of such assessments, first in pilot programs and then nationwide, as soon as possible. (g) Monitor interconnected trend patterns. Provide adequate resources to monitor all aspects of regional realignment in China, especially underlying and spontaneous trends which give information about underlying market forces, to avoid decisions based on partial or compartmentalized knowledge. It is especially important to register trends which emerge in spite of or counter to the purposes of government policy. Are trends aberrations resulting from even deeper distortions in the system? Or do they inform policy makers about what must in the long run naturally come to pass? The difference is critical for designing a strategic and policy framework. (h) Analyze disparities. China's government should fund short-term and long-term competing analyses of regional disparities, their causes, historical trends in other countries, and theories predicting the emergence and resolution of regional realignment. Analysis of other countries should find what ultimate regional realignments look like and what principals they reflect. Research should also identify dangers other countries have encountered, policy successes and failures they have experienced, and the degree to which these lessons are applicable to China. - 29 - 3. CAUSES OF REGIONAL DIsPARIEs 3.1 This chapter explores the causal forces responsible for China's emerging regional disparities/ Its major findings are: * The over-arching cause of China's increasing regional disparities is its shift from a closed and planned agricultural economy to an open, market-oriented industrial economy-resulting in high returns to natural and geographical advantages less useful for agriculture (like harbors and transport corridors on the East Coast). * Migration data for 1985-90 verify that movements ofpeople are heavily in a direction from the Central Core and Hinterland regions to Coast. * National policies in the 1980s promoted coastal development and international commercial relations, but policies in the 1990s are regionally more balanced * Investment disparities are the most important direct cause of regional performance differences; on a per-capita basis, the South Coast and East Coast have nearly triple the investment levels in the Central Core and South Hinterland. * The East Coast's investment structure is most advanced, because collective, individual and TVE (Township- Village Enterprise) investment have displaced SOE (state-owned enterprise) investment more than anywhere else, including the South Coast. * The most striking causal disparities are for foreign investment and trade; FDI (foreign direct investment) in the South Coast accountedfor more than 25 percent of total fixed investment in 1992, but less than 8 percent elsewhere on the coast, and less than 3 percent for interior regions. * Foreign trade disparities follow foreign investment differences, with nearly half of all imports going to the South Coast and another 38 percent to other coastal regions. * Foreign-funded-firm exports originated overwhelmingly from the Coast in 1993. * Regionalfiscal data show that even though reforms changed the appearance of regional deficits after 1985, transfers to interior regions continued and essentially stagnated in terms of their effect on interregional disparities. * Banking-flow disparities during a credit expansion year (1991) favor the coast, and the large loan volumes relative to deposit growth implies expansion of credit on a scale which dwarfs fiscal transfers. A. LAND, CLIMATE AND DEMOGRAPHIC CHANGE 3.2 The most basic explanation for China's regional disparities and their worsening since 1978 reflects China's rapid shift from a closed and planned economy to an open, market- oriented economy, a shift which is rapidly changing the geographic basis for supporting large -30- populations with a higher standard of living. China's current population distribution by region appears largely to reflect regional productivity in agriculture; warm rain-fed regions with good river-bottom soil have the densest populations. However, with industrial and market-oriented commercial growth-especially with commercial expansion to international markets--the locus of economic dynamism appears to have shifted rapidly to centers of internationally-oriented commerce and manufacturing. For China, these are on the coast, especially the southern and eastern coastal areas with easy access to Chinese-speaking international communities in Hong Kong and Taiwan. In recent years, migration and Malthusian fertility shifts have been much slower than the expansion of modern industrial and commercial production patterns associated with new global market-oriented trade links. In other words, new technologies are generating new regional productivity patterns faster than demographics can adjust. The result is rapid divergence in interregional per-capita output and income. 3.3 China's regional population growth rates show only marginal adjustments to newly emerging differences in regional productivity, but under-measurement may be significant. Table 3.1 (row E) presents average annual population growth rates for China's seven regions. Over the thirteen years reported, growth-rate differentials are very small in comparison to the emerging disparities in output and income shown in Chapter 2. The highest regional population growth rate is on the South Coast, but it is only marginally higher than that in one of the poorest interior regions, the Central Core region. Other coastal regions had population growth rates TABLE 3.1: ANNUAL GRowTH OF CHINA'S POPULATION, BY REGION, 1978-91 North South China Far Hinter- Hinter- Central North East South Total West land land Core Coast Coast Coast 1979 1.3% 1.6% 1.3% 1.2% 1.5% 1.2% 1.3% 1.5% 1980 1.2% 1.7% 1.1%: 1.0% 1.3%:: 1.1% 0.9% 1.6% 1981 1.4% 1.3% 1.2% 1.4% 1.4% -0.5% 1.2% 1.7% 1982 1.5% 1.7% 1.4% 1.4% 1.5% 1.7% 1.4% 1.8% 1983 1.2% 0.6% 0.8% 0.8% 0.9% 1.0% 0.9% 1.5% 1984 1.1% 1.7% 0.7% 0.9% 0.9% 1.0% 0.7% 1.5% 1985 1.1% 1.3% 0.9% 1.1% 1.0% 0.9% 0.8% 1.4% . ..-.-.... . .. ....... ......................-. 1986 1.4% 1.6% 1.0% 1.4% 1.3% 1.2% 1.0% 1.4% 1987 1.4% 2.6% 2.0% 1.7% 1.9% 2.3% 1.2% 1.7% 1988 1.4% 0.9% 1.3% 1.4% 1.5% 1.3% 1.3% 1.6% 1989 1.4% 1.8% 1.4% 1.4% 1.8% 1.1% 1.3% 1.7% 1990 2.8% 3.9% 2.7% 1.8% 3.6% 3.9% 2.1% 5.1% 1991 1.3% 1.5% 1.2% 1.2% 1.4% 0.9% 0.9% 1.5% Ave.'78-91 1.4% 1.7% 1.3% 1.3% 1.5% 1.3% 1.1% 1.8% End-'91' 1,158 50 146 223 301 208 124 102 Slow Demographic Adjustment The next-to-last row shows relatively insignficant regional variations in spite of under-registered births, which may account for the 1990 census surge. 1990 census patterns may also reflect under-registered migration. * Units: end-1991 population in millions; Source: Official administrative data for all years except 1990, which is based on census data; all data from national and provincial statistical yearbooks, various years -31- below the national average. 3.4 The population regional time series in Table 3.1 also reflect measurement uncertainties. The large surge in apparent fertility for all regions in 1990 represents adjustments to administrative data for information from the 1990 population census, which revealed that population registration procedures (births, deaths and residence changes) underreport actual trends, with underreporting most serious in the South Coast region and nearly as serious elsewhere, including the Central Core. China's strict population control policy creates incentives for underreporting births (and perhaps for over-reporting infant deaths), but even so, monitoring is close enough to ensure that distortions in vital statistics are moderate in terms of their influence on assessing regional disparities. The same cannot be said for migration statistics. 3.5 The expanding scale of interprovincial migration in China appears to be a response to interregional disparities at the same time that migration itself weakens the quality of interregional per-capita comparisons. Partial information on the scale of migration to coastal provinces from the interior shows that it can be as much as 10 percent of a coastal province's total population, but there is no complete statistical record. For example, interviews with officials reveal that one coastal province (Fujian), with a population of 31 million, had 3 million migrants in 1993. However, neither Fujian nor China's other provinces publishes migrant population figures regularly, even though most reportedly count migrants in periodic ad-hoc surveys. Consequently, official annual population statistics fail to report migrants fully. Regular population statistics rely on household registration data, and most migrant workers apparently stay registered in their home provinces. Most migration is also apparently seasonal to one degree or another, although many workers reportedly come to the same province annually, returning to home provinces only briefly for the Chinese New Year holiday. 3.6 Available national data show net migration flows from poor interior regions to the TABLE 3.2: NET INTER-PROVINCIAL MIGRATION BY REGION, 1985-1990 Net Migration from: North I South (Units: Thousands) China Far Hinter- Hinter- Central North East South Net Migration to: Total West' land land Core Coast Coast Coast Otherb China Total .. 64 378 1,381 518 -907 -455 -1,006 27 Far Weste -64 .. -21 60 16 -69 -41 -10 0 North Hinterland -378 21 58 4 -434 -10 -18 1 South Hinterland -1,381 -60 -58 -176 -251 -228 -616 8 Central Core -518 -16 -4 176 -112 -274 -289 2 North Coast 907 69 434 251 112 .. 40 -5 6 East Coast 455 41 10 228 274 -40 .. -62 5 South Coast 1,006 10 18 616 289 1 5 62 .. 5 Other_ -27 0 1 -8 -2 -6 -5 -5 Migration from Poor Regions to the Coast Reading down each column, all interior regions (especially South Hinterland) show net out-migration to the coast. Reading across rows, the South Coast and North Coast receive the most migrants. ' Data on movement destinations exclude Tibet. b "Other" migration figures reflect census data by destination only, and hence are gross inflows. -32- coast. China's most recent population census provides national migration data for the 1985-90 period, but the volume of reported migration flows seems low in comparison to reports from individual provinces, like Fujian. The census' 10-percent sample survey in 1990 recorded where respondents had resided five years earlier, and the results show that the direction and relative scale of migration flows correspond to interprovincial disparities--the largest flows are from the poorer interior provinces to the nearby coast. Table 3.1 presents these flows in the form of a matrix of net migration flows. Table 3.1 (looking down its columns) shows that all the net flows are from interior regions to the coast, and that the largest flows are from the poorest region (South Hinterland), even though its population is not the largest (226 million in 1992, compared to 305 million for the Central Core region). In other words, poverty--even more than size and proximity-seems important for understanding the scale and direction of migration. 3.7 The July, 1990 census probably understates the scale of migration flows in the late 1980s and early 1990s. The most important reason is that the summer of 1990 was at the end of a year of severe recession in China which had entered its most difficult phase after June, 1989. By July 1990, credit contractions and suspension of construction projects had reduced employment opportunities for migrants, and many had reportedly returned to their home provinces. In other words, the timing of the census made it unrepresentative of longer-term demographic trends. Second, it is uncertain how successful the census was in counting unregistered residents, especially unregistered migrant residents residing in coastal cities or other large urban employment centers. Many migrants live in relatively permanent shanty villages on the outskirts of major cities, where census organization may have been less systematic than in formal urban settings. 3.8 It is important to gauge the migration-reporting problem's seriousness. Is the unreported migrant population insignificant? Or is it 10 percent of the coastal population? Or, in the extreme, 50 percent? Underreporting of migrant populations in coastal regions undermines data on regional disparities, so that measures of disparities in Chapter 2 may distort actual conditions in China. How important could this distortion be? If, as in the case of Fujian Province in 1993-94, the migrant population in all coastal provinces (official population roughly 400 million) is 10 percent of the total, and if all is unreported, actual per-capita figures would need a 9-percent downward adjustment for the coast, and for interior regions (official population roughly 800 million) they would need a 5-percent upward adjustment. Adjustments on this scale would not significantly influence measures of regional disparities. In the extreme, however, if the floating population is as large as some unsubstantiated speculations--say, 200 million--and if all of it is unreported in coastal regions, then in order to be accurate, coastal per-capita indicators need a reduction of one-third, and interior indicators need a one-third increase. These kinds of adjustments would virtually eliminate significant GDP-per-capita disparities between coastal and interior regions--and would even reverse some. 3.9 A more realistic estimate of unreported interior migrants living in coastal provinces is most likely much closer to 10 percent of the official coastal population than to 50 percent. Consequently, it is highly likely that official-data measures of regional disparities like those in Chapter 2 are broadly descriptive of actual conditions. Nevertheless, more accurate and systematic reporting of migration flows would contribute to our understanding of regional -33- disparities and of the population movements they engender. Most importantly, adjustments to population data for underreported migration are unlikely to change the general conclusion that migration flows and differential fertility patterns have been inadequate to compensate for the rapidly emerging differentials in regional productivity associated with modem market-oriented commercial and economic reforms, especially when combined with promotion of rapid expansion in foreign trade and international financial activities favoring transportation centers on or near the coast. B. NATIONAL AND REGIONAL POLIcIEs 3.10 Since the onset of economic reforms under Deng Xiaoping in 1978, Chinese policies important for regional disparities have been of two basic kinds: (1) general nation-wide policies which have nevertheless indirectly induced regional disparities, and (2) policies directly targeting particular regions. Both kinds of policies have contributed to large and growing disparities between coastal and interior regions. Between different coastal regions, however, policies specifically targeting the South Coast have reduced its earlier disparities vis A vis the East Coast. For this reason, policies targeting the coast appear to have had the most dramatic regional influence. However, both direct and indirect policy effects favored the coast, and even if direct policy interventions shift increasingly to favor the interior, natural coastal advantages are likely to dominate. 3.11 Nationwide Policies with Indirect Regional Consequences. Since 1978, governments at all levels have introduced reforms and policies which, although they do not have explicit regional purposes, nevertheless have induced dramatic regional shifts and developments because of the natural economic forces they set in motion. Many of the central government's policies seek to change China's nationwide system of economic management and operation. These policies reflect a shift in philosophy and a national decision to substitute significant market mechanisms for central planning, to engage in international specialization, and to use decentralized domestic systems of economic discipline based on profitability and consumer choice. Implementing this shift in philosophy has had and will continue to have significant regional consequences which are by-products of the new economic system. 3.12 The clearest example of general policies with indirect regional consequences is the broad strategy after 1978 to strengthen international commercial relations. Once China decided to increase exports, imports, technology transfer, foreign borrowing and foreign direct investment as rapidly as possible, certain regions immediately benefited. Coastal regions with good harbors, convenient transport corridors to other Chinese regions, and good information about foreign markets naturally flourished in comparison to regions isolated from foreign commerce and contacts. Good harbors in Guangdong and Shanghai, good links to the interior at Shanghai, Tianjin and Dalian, and especially good contacts and information flows in Guangdong (adjacent to Hong Kong) became determinants of accelerated coastal development. 3.13 A second set of general policies with specific regional consequences was the program of fiscal and enterprise reforms in the early and middle 1980s, reforms which separated enterprise profits from the government budget, sought to improve enterprise management -34 - incentives, and prepared fiscal and financial institutions for new market-oriented roles. For all their imperfections, these reforms encouraged enterprises to focus more on profits and hence more on markets and on sales. This shift in enterprise incentive structure appears to have been more effective closer to natural domestic and international markets, such as in Shanghai, Guangdong, Beijing, Tianjin, and Dalian. An additional consequence of fiscal and financial reforms was a change in the incentive structure for local governments and their constituent enterprises in ways which allowed natural location advantages to speed township and village enterprise (TVE) development more in some regions than in others. Regions near existing industrial and financial centers-especially those near Shanghai-had location advantages providing them with technology, capital, management assistance, and markets for subcontracted parts and other intermediate inputs. Proximity to foreign markets and capital through coastal ports amplified metropolitan-area advantages for TVEs. 3.14 A third general nationwide feature of China's reform program after 1978 was its use of pilot programs. Pilot programs in a few selected sites where favorable conditions promised fruitful learning experiences have been part of virtually every Chinese policy program. For example, the Ministry of Agrieulture introduced pilot programs for grassland restoration in parts of Inner Mongolia with good potential for success of experimental policy steps. China applied this same pilot-program strategy to learn about measures necessary for successful international commerce. Coastal areas offered the best locations for such pilot programs for international commerce and finance, and the government introduced a variety of experimental programs for coastal areas. Even explicit pro-coastal policies, therefore, have in a sense been regional manifestations of a more general national rule to experiment with new policies through pilot programs in regions where experiments are likely to succeed. 3.15 Policies Directly Targeting Specific Regions. Chinese central government policies targeting specific regions have gone through a transition-from policies explicitly favoring the coast in the 1980s to policies with a more balanced coastal-interior emphasis in the early 1990s. For ten years, from 1978 through 1988, pro-coast policies actively emphasized the primacy of coastal provinces and counties in China's opening to the outside world, reversing policies begun during the Vietnam War in the middle 1960s which sought to invest in defense- related and other heavy industries in China's deep interior--the so-called third-front strategy. The pro-coast policies begun in 1978 represented a return to normalcy in diplomatic and commercial relations with the non-communist world, especially with the US and Japan. Beginning in the late 1980s and early 1990s, however, China's central leadership adjusted this pro-coast bias, and elements of a more even-handed treatment of all regions began to emerge. The new policy direction encourages foreign investment and related trade evenly, on the coast and in the interior. At the same time, it concentrates this encouragement in certain strategic sectors--infrastructure, energy, raw materials, and industrial sectors with special export or technology significance. 3.16 The pro-coastal policy program began in 1979-80 by granting special "open" investment and trade status to Guangdong and Fujian Provinces in China's South Coast region and by establishing "open" special economic zones on their coasts near Hong Kong, Macao and Taiwan. By "open," these policies meant a status which allowed provincial and local governments more freedom than other regions to approve large-scale foreign investments without -35- seeking additional approval from the central government. Through the 1980s, as the government awarded "open" privileges of varying degrees to other coastal areas, South-Coast privileges generally remained superior. In some cases, these "open" policies by the early 1990s relieved county governments of the need even to seek approval from provincial governments. Throughout the 1980s, the permitted scale of locally-approved foreign investments on the coast increased regularly, and other forms of foreign commercial discretion expanded as well, including access to foreign exchange. 3.17 The expanding openness proceeded in stages after the initial opening of Guangdong and Fujian. In 1984 China identified 14 coastal "open cities," and quickly continued the expansion with "opening" for major river delta areas and additional coastal towns and counties. Finally the 1986-90 five-year plan declared that all coastal provinces would make up a "coastal belt" which would receive priority in foreign investment. This "coastal-belt" program explicitly relegated "central-belt" provinces (essentially Central Core and Hinterland regions) to a foreign-investment "trickle-through" status. It also declared that foreign investment and technology transfer to the western belt (Sichuan Province and the Far West) were "postponed" for the time being. 3.18 The coastal belt program strengthened further in 1988 when China announced a "coastal zone strategy" promoting export-oriented foreign direct investment. The strategy sought to emulate other newly-emerging economies in Asia. One of its special steps made the South-Coast island of Hainan a province in its own right and a special economic zone in its entirety; it had been part of Guangdong Province. In sum, beginning in 1978, all of these policy steps essentially reversed the "third front" investment strategy of the 1960s, which had promoted heavy-industrial investment deep in China's interior (Hinterland) regions. 3.19 By 1990, both the successes and excesses of the coastal strategy were apparent, and China adjusted its regional preferences to introduce more balanced treatment of coastal and interior regions. By the late 1980s, China's coastal strategy had contributed to superior coastal growth, income and foreign commerce, but disparities with the interior had become increasingly obvious (see Chapter 2). By 1992, the government had clarified many components of this more even strategy, which emphasized at least six interior zones in addition to three coastal regions. First, it officially promoted seven regional development zones, four of which are in the interior. Additional interior regions received emphasis, especially in Sichuan Province and in Xinjiang far to the northwest. Second, the new strategy placed special emphasis on the Yangtze River, with its mouth at Shanghai. Eight Yangtze River-port cities-from Sichuan to the coast--opened to international shipping in 1991 for the first time in 40 years. Large-scale investment plans to develop Shanghai's new port zone in Pudong prepared the city to service its upstream hinterland. Other large-scale projects, such as the Three-Gorges Dam, add to transport and industrial infrastructure along the Yangtze. 3.20 By the middle 1990s, China's regional strategy acknowledges no regional preferences at all. Chinese officials describe its purpose as promoting investment in certain categories, wherever they are. These categories include energy, raw materials, basic infrastructure, and certain strategic industrial sectors such as electronics and motor vehicles. -36- Preferences for these sectors apparently have no regional dimension, but officials comment that the interior naturally has more such investment opportunities than the coast. C. INVESTMENT AND TRADE DisPARITIEs 3.21 The investment and foreign-trade manifestations of various economic policies provide the clearest image of the causal mechanisms responsible for China's worsening regional disparities. In both absolute and per-capita terms, investment and foreign commercial transactions show patterns clearly favoring the Coast. Table 3.3 on page 37 presents information showing that total investment levels in the three coastal regions is much larger than investment levels in the interior (Table 3.3 col. 1)-so much so that investment effort out of GDP is more than ten percentage points greater in southeast coastal regions than in the Central Core. It is not surprising that richer regions have higher savings and investment rates out of GDP, but in China's case the differences are large. 3.22 Per-capita investment levels show the regional gaps most clearly--especially between the South-East Coast and the South-Central interior (see Figure 3.1 on page 36). South Coast and East Coast levels of per-capita investment are nearly quadruple the levels in the poorest interior regions (Central Core and South Hinterland). Given this very large coastal-interior disparity in per-capita investment levels, it is not surprising that per-capita output and growth levels also show large disparities (see Chapter 2). Looking to the future, such low per-capita investment levels in heavily-populated poor regions imply relatively lower growth in non- FIGURE 3.1: INVESTMENT PER CAPITA, BY REGION, 1992 China Total Far West North Hinterland South Hinterland Central Core North Coast East Coast South Coast 0 200 400 600 800 1,000 1,200 [rg_nv5c s] Per Capita Fixed-Asset Investment, 1992 Enormous Per-capita Investment Gap. Lower investment shares and lower GDP per Capita imply very low per-capita investment in the interior, especially for the Central Core and South Hinterland regions. Levels there are roughly one-fourth what they are on the South and East Coast. Souce: Investment data from Table 3.3. -37- TABLE 3.3: FIXED CAPITAL INVESTMENT BY VARIOUS MEASURES, 1992, BY REGION 1 2 3 4 5 6 7 Investment Share TVE' Share of Fixed Total Fixed Investment by Type of Enterprise Investment Units: 1990 State- Rural & % of Non- Yuan Value % Share Owned All Urban % of Total SOE' (Billion) (Bil. yuan) of GDP Enterprises Collectives Individuals Investment Investment China Total 758 32% 66% 18% 16% 13% 39% Far West 33 38% 85% 6% 9% 4% 29% North Hinterland 83 30% 81% 6% 13% 4% 22% South Hinterland 76 32% 70% 12% 18% 7% 25% Central Core 110 25% 65% 13% 22% 8% 24% North Coast 179 32% 71% 19% 10% 14% 49% East Coast 154 36% 46% 32% 21% 26% 48% South Coast 122 39% 66% 19% 15% 12% 35% Coastal Investment Lead Coastal regions have significantly larger investment flows than interior regions (col. 1) and much higher investment shares in TYE? (col. 6-7). Curiously, only the East Coast has signicantly lower SOl? investment shares (col. 3). a TVE = Township & Village Enterprise; SOE = State-Owned Enterprise. See Annex for Provincial-level data. Sources: State Statistical Bureau, Statistical Yearbook, 1992. [rg_inv5e!main] agricultural jobs, and lower labor productivity in whatever new jobs do appear. In sum, whatever the reasons-indirect effects of general market reforms or direct effects of regionally targeted policies-China's pattern of development in the 15 years since reforms began in 1978 has left policy makers with a serious mismatch between the location of its labor force and the location of modern investments. This per-capita investment gap may be the single most fundamental characteristic of China's overall pattern of regional disparities. 3.23 Other aspects of regional investment patterns reinforce the impression of large disparities. Investment patterns by ownership status show significant regional variety (Table 3.3 cols. 3-5). The East Coast displays the only real exception to domination of investment by state- owned enterprises (SOEs), because of its high levels of investment on the part of both collectives and individually-owned enterprises. In the East Coast, nearly half of non-SOE investment comes from township and village enterprises (TVEs), similar to the share on the North Coast, but much higher than in any other region (Table 3.3 cols. 6-7). Curiously, the significance of SOE investment in the South Coast region is as high as the national average, and much higher than on the East Coast. If high levels of non-SOE investment are an important gauge of success in market-oriented reforms, then the East Coast is far and away China's leader in this regard. Except for the East Coast stand-out performance, however, there is surprisingly little coastal- interior variation in the SOE share of investment-which is generally high everywhere else. 3.24 In contrast to only slight variation in SOE investment, TVE investment disparities are large. TVE investment is nearly half of non-SOE investment in most coastal regions, but less than a quarter in the interior (Table 3.3 col. 7). The gap is even larger in terms of the TVE share of all investment (Table 3.3 col. 6), and considering actual levels of investment (Table 3.3 col. 1), -38- TABLE 3.4: FOREIGN TRADE AND INVESTMENT INDICATORS, 199 -1993, BY REGION 1 2 3 4 5 6 7 8 Trade Ministry Data Customs Data 1993 Foreign-Funded- Foreign Direct Investment 1991 (Billion US $) 1993 (Million US $) Enterprise Trade Share 1992 Level % of Fixed Import Surpl./Defici Imports Surpl/Def Exports Imports (mil. US $) Investment China Total 63,79 8,120 103,950 -12,188 28% 40% 11,292 8.2% Far West 15 662 1,143 -262 5% 7% 1 .0% North Hinterland 1L44 2 477 5,213 1,761 4% 18% 252 1.7% South Hinterland 1,08 1,516 3,534 -275 7% 26% 343 2.5% Central Core 103 3323 -359 8% 32% 543 2.7% North Coast 10,45 8,392 19,478 -4,066 21% 32% 2,091 6.4% East Coast 12,24 4,272 19,664 -1,834 24% 42% 2,197 7.9% South Coast 11,79 5,708 49,964 -7,153 40% 48% 5,577 25.2% NotRegional 25,57 -18,230 0 0 __ _ 288 Coast Dominates Trade and Foreign Funds Over 85 percent of imports go to coastal regions (col 3), with large shares of exports coming from foreign-funded enterprises (col 5). However, shares of imports to foreign-funded enterprises show a slight shift to the interior (col 6). The South Coast is the only region with investment levels heavily dependent onforeignfunds (col. 8). " Trade Ministry data only cover Trade Ministry transactions, resulting in underreported imports in particular. [rg_trd5elmain] Sources: State Statistical Bureau, China Statistical Yearbook, 1993, and Customs Statistics, 1993, No. 12 the shares in col. 6 imply large absolute coastal-interior disparities in TVE investment, not to mention per-capita disparities (see below). 3.25 Foreign trade and investment disparities complement regional inequalities in overall investment. The coastal advantage in foreign trade and foreign investment is overwhelming. Customs data show that in 1993, the three coastal regions received more than 85 percent of all China's imports (Table 3.4 col. 3).4 Of the heavily populated interior regions, the South Hinterland had the least imports of all in 1993, while the North Hinterland had a significant trade surplus in a year of large overall deficits (due mostly to petroleum exports-see Table 3.4 col. 4). 3.26 Large foreign-trade disparities between coast and interior are natural to some degree, given coastal-interior transport difficulties, but performance disparities along the coast- in favor of the South Coast-are unusual, given the many fine harbors and inland transport facilities in the North and East Coast regions, and given the high degree to which South-Coast harbors are cut off from the interior by mountains. Imports to the South Coast are greater than to the other two coastal regions combined and more than 150 percent greater than imports to the East Coast (Table 3.4 col. 3). Given the East Coast's traditional advantage, with what is arguably China's greatest economic center at Shanghai, how did the South Coast's foreign trade do so well after 1978? A combination of geographical and policy differences offers the best explanation, and (as mentioned in para. 3.14) the South Coast's natural geographical advantages may have 4Good foreign-trade data by region are difficult to come by. China's customs data began reporting trade by location of consignor and consignee only in 1993; Trade Ministry data have provincial-level categories for imports and exports, but these data are an incomplete record of China's trade and refer to ports of entry and export rather than domestic destination and origin (see Table 3.4 cols 1-2 and related note). -39- influenced government decisions to award the South Coast special policy status (see paras. 3.15 to 3.18). The South Coast's two major provinces, Guangdong and Fujian, are geographically adjacent to Hong Kong and Fujian, respectively, and perhaps even more importantly, they are linguistically compatible. Cantonese-speaking Hong Kong's economic links to Guangdong province (traditionally called Canton) have been close for decades, and beginning in the late 1980s, Taiwanese Chinese-whose indigenous dialect is the Fujian dialect-began traveling to Guangdong, Fujian, and other mainland provinces in large numbers. Even though slow progress in relations between authorities in Beijing and Taipei (Taiwan Province) have prevented direct transport and communication links, Taiwan-based businesses received Chinese government encouragement to use Hong Kong as a legal intermediary, and since 1990, trade and investment ties to the whole South Coast have expanded rapidly. 3.27 Not only are foreign trade levels high on the South Coast, their links to foreign investment are stronger there than anywhere else. Nearly half of imports to the South Coast are to foreign-funded firms. In general, the role of foreign-funded firms in foreign trade is more important in all coastal regions than in the interior, especially for exports (Table 3.4 col. 5-6). For imports, the data show that foreign-funded shares of imports in interior regions are respectable compared to those for exports-possibly signaling an early developmental phase of foreign-direct- invested (FDI) projects in interior regions. However, even though the foreign-funded share of imports is respectable in the interior, the low absolute levels of both foreign trade and FDI in interior regions places them at a serious disadvantage. (Table 3.4 cols 3 and 7). 3.28 Interior disadvantages show even more clearly in foreign investment data. FDI flows have a regional pattern similar to that for trade and strongly favor the South Coast; the South Coast received roughly half of all of China's 1992 FDI (Table 3.4 col. 7). Furthermore, the South Coast has become by far the most dependent on foreign direct investment to fund its fixed investment-more than 25 percent of all of the South Coast's 1992 investment came in the form of direct foreign participation (Table 3.4 col. 8). The corresponding shares for other coastal regions are much lower (less than 8 percent), and for the interior they are hardly significant at all. D. FISCAL AND FINANCIAL TRANsFERS 3.29 Although China's national policies after 1978 explicitly favored coastal areas with policies which "opened" them for foreign investment and greater degrees of local autonomy, other policies-most notably fiscal policies-appear to have compensated for this bias with measures which transferred significant resources from richer regions to poorer regions, especially from the East Coast to the Far West and North Hinterland regions. However, although interregional fiscal transfers continued into the 1990s, their significance relative to the scale of provincial economic activity seems to have declined, and interregional fiscal transfers have generally shown an inability to adapt to the pace and scale of emerging interregional disparities in output, income and investment. Nevertheless, the scale of transfers is still substantial-in the case of the East Coast amounting to outflows of a third of revenues (half of expenditures). The regional situation for bank loans, however, is quite different. Recent financial transfers through the banking system are much larger than fiscal transfers and seem to have greater potential to -40 - TABLE 3.5: CHINA'S BUDGET SURPLUS (REVENUE-EXPENDITURE), BY REGION, 1978-91 1 2 3 4 5 6 7 8 Units: China's North South Billion Regions Far Hinter- Hinter- Central North East South Currant Yuan Total West land land Core I Coast Coast Coast A. 1978 36.1 -1.5 2.0 -1.7 1.1 16.4 18.6 1.2 1979 27.9 -2.0 0.1 -2.8 0.6 13.7 18.1 0.3 1980 30.4 -2.6 -2.7 -1.6 1.9 14.0 20.3 1.1 1981 34.1 -2.7 -2.8 -1.2 3.6 14.9 21.2 1.1 1982 28.5 -2.6 -3.5 -1.1 2.4 12.1 20.7 0.6 1983 23.0 -3.5 -3.4 -1.5 2. 1 9.8 19.5 0.1 1984 15.8 -4.9 -5.0 -3.3 1.0 9.6 18.8 -0.3 1985 12.7 -5.3 -6.1 -3.4 -0.7 8.9 19.6 -0.4 B. 1986 -5.9 -6.2 9.0 -6.8 -4.5 5.1 17.1 -1.6 1987 4.3 -5.5 -7.2 -5.5 -1.5 6.8 17.9 -0.8 1988 -5.3 -6.3 -7.1 -5.5 -2.8 3.2 15.5 -2.1 C. 1989 -6.8 -6.2 7.2 -6.2 -2.9 2.4 15.1 -1.9 1990 -14.8 -6.8 -8. 1 -7.1 -2.6 -0.1 14.4 -4.4 1991 -12.4 -7.1 -7.7 -5.5 -6.6 2.8 13.0 -1.2 D. Surplus/Capita' -11 -142 -53 -24 -22 13 105 -12 E. % of E:.b -5.2% -47.1% -20.7% -13.8% -15.5% 5.7% 46.4% -4.4% F. 1991 Exp. 238 15 37 39 43 49 28 27 Interregional Budget Transfers Continue. After 1985, tax and enterprise-finance reforms cut regional budget revenues as expenditure growth accelerated (cot. 1). However, the central govern- ment continued to distribute somewhat reduced North- and East-Coast surpluses (col. 6-7), as well as other central-government monies, to poorer regions. The East-Coast distributed surpluses amounting to nearly a third of its revenues-half of its expenditures (cot. 7, row E). * Per-capita surplus data are in 1991 yuan. b "O/o of Exp." means 1991 regional surpluses as a share of budget expenditures, and "1991 Exp." is in billion yuan. Sources: provincial yearbooks, statistical yearbooks, economic yearbooks, and "Compendium of Provincial Historical Statistics," 1990, Beijing: SSB. match the scale of emerging regional economic disparities, because their regional influence overwhelms the effect of fiscal transfers in periods of central-government expansionary stimulus. 3.30 Regional fiscal patterns in China have been relatively stable, in spite of appearances. Interregional fiscal transfers still reflect a pre-reform pattern which taxed more advanced industrial-base regions in order to subsidize priority development regions. However, reforms have obscured annual trends in such-transfers and reduced their significance. In 1983-86, China undertook fiscal and enterprise-finance reforms which reduced provincial-level revenues by redefining enterprise profit remittances to the budget as retained earnings. Budget revenues, by definition, were lower. As a result, regional government budget revenues grew more slowly than expenditures, and what had been a net surplus transferred to the central government became a net deficit (Table 3.5 col. 1). At roughly the same time, reforms shifted some tax-revenue rights from provincial-level governments to the national government, helping it fund the net regional deficits which emerged. As a result of these and other changes in central-government fiscal behavior and -41- because of difficulties in knowing the regional significance of direct central government spending, regional fiscal trends in Table 3.5 provide only a partial indication of actual fiscal transfers. 3.31 The overall impression is that by the early 1990s, fiscal transfers had essentially stagnated, as transfers from East-Coast and central-government sources continued traditional patterns of fiscal subsidies to the interior and, in a much more limited way, to the South Coast. The limited change in South-Coast fiscal balances, given the region's extraordinary growth and commercial performance since 1978, is one indication of inertia in China's interregional fiscal relations. Annual data in Table 3.5 show the actual sizes of surpluses (and deficits). For 1991, funding for deficit regions (col. 2-4) and the Central Core (col. 5) came from a large surplus in the East Coast (col. 7), a small surplus in the North Coast (col. 6), and large transfers from the central government equivalent to the remaining net regional deficit (col. 1). Local-government borrowing in China was negligible throughout this period. The reforms in the middle 1980s make it impossible for these data to show how much the pattern has changed over time, but looking only at chronic deficit regions (cols 2-4) and at the Central Core region (col. 5), apparent outside budgetary support changed dramatically in the 1983-85 reform years but much less if at all thereafter. The general conclusiorris that in a period when China's overall real GDP growth averaged close to 9 percent and was even higher in coastal regions, the system of fiscal adjustments between regions remained rather closely tied to pre-reform patterns, rather than adapting to the new needs of worsening regional disparities between coast and interior. 3.32 In spite of what appears to be effective stagnation in interregional transfers, the inherited pattern nevertheless involves substantial flows-albeit not at all large enough to affect China's regional disparities directly. Via the central government, the East Coast transfers funds to other regions equivalent to roughly half (46 percent) of its own budgetary expenditures (Table 3.5 col. 7, row E)-implying a transfer equivalent to roughly one third of its revenues. For poor interior recipient regions, the transfer inflows amount to a 14-to-47-percent contribution to total regional budget expenditures (Table 3.5 cols 2-5, row E). While these are significant flows in a budgetary sense and could help fund strategic public expenditures in poor regions, they are by themselves quite small in relation to the scale of China's interregional output and income disparities. On a per-capita basis in both the poorest and richest regions, the transfers are well less than 10 percent of per-capita GDP levelss, which in richer regions are themselves roughly double those in poorer regions (see Chapter 2). 3.33 In contrast to relatively moderate and stable interregional fiscal flows, China's regional banking flows are both large and volatile. In a period of credit contraction, such as 1990, new bank loans were so low that they barely match the scale of deposit creation, and nationwide they are less significant than regional budget deficits (compare Table 3.6 col. 5 with cols 3-4). When economic policy changes, as it did toward the end of 1991, the volume of credit creation is much larger than levels of budgetary expenditure-by a factor of more than three-to-one in many regions (compare Table 3.6 cols 6 and 1). These comparisons indicate that in considering causes of regional disparities and measures to address them, regional financial policies deserve as much if Based on official data, per-capita fiscal transfers to the Far West amounted to 7.3 percent of per-capita GDP in 1991. For the East Coast, the per-capita fiscal-transfer outflow was 3.3 percent of per-capita GDP. -42 - TABLE 3.6: FISCAL AND BANKING INDICATORS, 1990-91, BY REGION 1 2 3 4 5 6 Units: 1991 Regional 1991 New Budget Surplus/Deficit Net New Bank Deposits Billion Budget Regional (Revenues - Expenditures) (New Deposits - Loans) Current Yuan Expenditures Bank Loans 1990 1991 19901 1991 China Total 238.1 1,804.4 -14.8 -12.4 -10.8 -604.3 Far West 15.1 73.9 -6.8 -7.1 -1.9 -24.9 North Hinterland 37.5 264.5 -8.1 -7.7 -16.5 -132.4 South Hinterland 39.5 180.8 -7.1 -5.5 2.4 -61.4 Central Core 42.6 307.5 -2.6 -6.6 -10.5 -147.8 North Coast 48.7 406.0 -0.1 2.8 1.6 -98.4 East Coast 28.0 249.4 14.4 13.0 4.5 -74.0 South Coast 26.9 188.8 -4.4 -1.2 11.4 0.8 Not Regional -1.8 -66.1 Credit Creation Dwarfs Fiscal Outlays. 1991 's policy-oriented net credit expansion (col. 6) out-scaled budget expenditures (col. 1) by roughly three-to-one, except for the South Coast, which had net deposits. The contrast with the recession year of 1990 is strong (col. 5), and the lower level of credit creation in South Hinterland (cols 5-6) is striking. Sources: State Statistical Bureau, Statistical Yearbook and People's Bank of China Financial Yearbook, various years. not more attention than regional fiscal policies, especially when the central government uses state- run banks to target loans to strategic sectors. 3.34 In addition to scale and volatility, loan and deposit data also reveal significant regional disparities in credit creation, which appear to reinforce output and income disparities. The poor but densely populated South Hinterland and Central Core regions received smaller shares of the 1991 nationwide credit expansion relative to their populations, while the East Coast, North Coast and North Hinterland received disproportionately large shares (Table 3.6 cols 2 and 6). Credit creation on the South Coast is different, because although the scale of new loans is large relative to population, deposits are also large enough so that even in a year of nationwide credit expansion, the South Coast has deposit growth even greater than new loan volume-possibly reflecting FDI and inflows of credit from other domestic regions (cols 5-6). Overall, and unlike the situation for interregional fiscal flows, bank credit flows are large, volatile, and tend to reinforce regional disparities rather than ameliorate them. In this regard, China's regional disparities appear even more to be the result of natural geographical and market-oriented advantages rather than unbalanced national development policies. E. CONCLUSION AND RECOMMENDATIONS-NATURAL AND UNAVOIDABLE REGIONAL DISPARITIES 3.35 Two major themes stand out for explaining China's worsening regional economic disparities, and two statistical measures strongly reflect both themes. The two themes are (1) the natural advantages of coastal areas and related transport corridors in a reform period emphasizing market-oriented development, and (2) specific reform policies since 1978 which have favored coastal regions and encouraged experimental exploitation of natural coastal advantages-to an -43 - important degree as preparation for broader management and international commercial liberalization in the 1990s. The workings of these broad causal influences have appeared most clearly in the form of (1) very large regional investment disparities and (2) very large regional disparities in foreign trade and direct foreign investment-which are closely correlated to one another and to transfers of foreign technology. In other words, reform-era advantages for coastal regions have resulted in superior investment behavior, both in quantity and quality. These advantages have appeared faster than population can adjust through migration or differential fertility, even though partial evidence indicates that significant migration flows have already begun to cut into regional per-capita productivity differences. 3.36 In sum, the causes for China's regional disparities are natural and unavoidable. The important issue for policy purposes is the speed of China's adjustment to market-oriented location advantages throughout the country. Will a more balanced program of regional investment preferences slow the worsening of regional disparities by correcting unnatural distortions in regional productivity patterns? Or will accelerating migration flows reduce regional per-capita disparities and replace them with growing income size-distribution disparities within individual regions, as interregional household mobility reduces geographical distances between China's rich and poor? The evidence in this chapter indicates that the location, scale, and nature of investment (e.g., SOE, TVE or FDI) is highly correlated with the appearance of regional economic disparities, and it is just as likely that these same variables will be critical for managing their future evolution and the response of policy makers to the secondary economic and social difficulties they generate. 3.37 The analysis presented here points to a broad set of policy guidelines, which are summarized in the following paragraphs: 1. Overarching Strategic Recommendations 3.38 A number of broad strategic recommendations are important for successful management of regional disparities and the regional realignment of China's capital and labor. These strategic recommendations not only support regional realignment, they are themselves broad objectives in their own right and lead to a range of more concrete recommendations. (i) Anticipate and facilitate orderly flows of labor and capital to rapidly-growing cities in market-oriented locations. National plans should anticipate both the positive and negative aspects of rapidly changing regional concentrations of labor and capital. Because pre-modern labor and capital served agriculture, new urban market concentrations are an essential aspect of modem economic growth. Furthermore, international experience shows that disparities between leading and lagging regions are inevitable, but that per-capita disparities decline in the long term, as population densities shift. Policies meant to manage this natural process in an orderly way must be careful not to generate long-term distortions in the locations of capital and labor. (j) Encourage modern jobs of appropriate labor-intensity in strategic locations. China's planning must concentrate on the rapid generation of large numbers of new jobs appropriate -44 - for China's modem growth-that is, jobs with appropriate capital intensity, located in strategic market centers, where they have long-term commercial viability. The central contradiction in this recommendation is that investment incentives based on low interest rates also encourage inefficient use of expensive equipment and structures. Incentives should be indirect, in the form of accelerated infrastructure, health, education, and cultural investments, rather than direct, and the enterprise cost of raising capital should reflect both expected inflation and expected high productivity and sales-not expected government or financial subsidies. (k) Avoid counterproductive subsidies and other unintended distortionary incentives or disincentives. The government must design direct and indirect policy instruments, such as subsidies and welfare programs, so they do not inadvertently distort the process of regional realignment-either by promoting new urban concentrations in inappropriate locations, or subsidizing residents of lagging, rural, regions so they continue indefinitely in dead-end low- income activities. International experience gives many examples of policies to support traditional employment in lagging regions, which only end up-at high public cost- indefinitely preventing necessary outmigration. (1) As much as possible, influence voluntary migration decisions to avoid migration's most costly and undesirable side effects. Migration flows are a natural aspect to changing economic opportunities, but they are such an important part of regional realignment that they deserve special strategic attention. Labor movements should be voluntary as much as possible. At the same time, they should result neither in large pools of unemployed workers in leading regions, nor in shortages of skills needed in lagging regions. The basic strategy is to ensure accurate expectations of income opportunities relative to the costs ofjob search in new cities, by discouraging urban wages out of line with productivity, and by encouraging costs for housing, food, and-public facilities which reflect urban scarcities. Similarly, moderation in suppressing gains in rural-urban terms of trade should allow rural incomes in lagging areas to adequately reflect China's scarcity of arable land. (m) Medium-term and long-term steps should prevent the worst pollution and congestion problems likely to emerge from rapid urbanization. Rapid urbanization involving large inflows of capital and labor invariably means that congestion and pollution will worsen in expanding urban centers. China's planning should allow for well-designed pollution regulations, urban zoning, transport patterns, and other infrastructure distributions so that from the outset pollution and congestion remain at manageable levels. 2. Locating Market-Oriented Factor Concentrations 3.39 The strategic recommendations above imply the need for several sets of more concrete planning requirements. The first of these sets has recommendations about the locations of emerging urban concentrations. Over the long run, they must emerge in locations appropriate for relevant markets. Market forces are best for achieving these goals, if they operate in an appropriate incentive framework. However, planning based on experience in industrialized countries can help China avoid unnecessary trial and error efforts to generate well-located investment and employment concentrations. This involves enabling investment in manufacturing -45 - capacity at convenient freight transport hubs with broad market access (direct or indirect), as well as engineering a more widely disbursed set of such hubs. 3.40 This primary goal implies a sequence of policy steps. (a) Model the regional future: Model various quantitative scenarios of future regional development patterns. Within 2-to-3 years, if it is not already prepared, government should model the likely future regional development patterns, especially the likely scale and nature of business establishments and the size and structure of populations. The projections should draw on international experience, from both developed and developing. countries, and have different scenarios for variations in infrastructure investments, especially long-distance containerized transport investments designed to encourage urban growth in appropriate locations. One goal of such modeling is to investigate possible urban location patterns which are different from what might emerge from a long series of short-term investment responses to immediate bottlenecks. (b) Open the interior: Accelerate transport and communications investments to connect the most important and accessible interior market centers to coastal ports, especially with transport facilities suitable for containerized freight. Starting immediately in the new five-year plan, investment priorities should concentrate public funding from national sources on new or improved waterways, rail lines, and above all, limited-access highways from provincial-level transport hubs to ocean ports-especially new or newly expanded ports. Initial highway designs should plan for more lanes than current traffic indicates are necessary. All related standards and inter-modal terminals should be suitable for containers. (c) Plan interior hubs: Wherever possible, encourage investments on transport corridors to the interior so that different transport modes intersect in locations most suitable as local, provincial-level, transport hubs. All current plans for long-distance transport corridor investments should be reviewed and revised in light of this criterion. These investments will create or strengthen natural market centers in the interior, which can function both as interior regional factor concentrations and as specialized hub links to international commerce via containerized ocean freight traffic. (d) Promote secondary satellite hubs: Design regional transport patterns to encourage secondary urban centers at natural transport corridor intersections at some distance from but with good access to primary regional transport hubs. Rather than encouraging a mono-polar urban area spreading at each major interior (or coastal) transport hub, encourage secondary hubs by placing transport, communications and other infrastructure investments so that satellite cities emerge naturally where there are local market advantages-including lower labor and land-use costs for service-sector centers less dependent on freight transport convenience than on good communications. (e) Promote competitive containerized freight: Within 1-to-2 years launch pilot experiments in transport enterprise reform separating containerized trucking from management of other containerized transport modes, especially shipping, and extending the experiments nationwide by the end of the five-year plan. At the same time, special emphasis should be placed on -46 - containerized freight facilities. Containers are critical for opening to the outside world. Competition among trucking companies is the best channel for promoting containers in all regions. Competitive containerized trucking services promote container use by other modes (rail, inland water, air). They will also speed development of the related logistical (information) systems necessary to improve China's link with world markets, which insist on predictable deliveries to reduce the risk of unwanted inventory fluctuations. (f) review the regional implications of all economic policy decisions. Virtually all economic policies influence regional trends, especially public expenditure policy. The location of public investments in roads, rail, ports, communications, health, and education can both benefit and harm regional development patterns. Central and regional governments should provide adequate resources to specialized departments which analyze the locational impact of public policies, to determine whether they serve regional strategies. -47 - ANNEx 1: STATISTICAL ANNEX TABLE A.1: CHINA'S REGIONAL AND PROVINCIAL GDP BY SECTOR, 1992 1 2 3 4 5 6 7 8 1992 1992 GDP 1992 GDP/Capita Sectoral Shares (1992) Population Total Industry Total Industry Agriculture Industry Services (million) (Bil. 1992 US$&) (1992 US$) (Percent) China Total 1,166 547.5 263.4 470 226 24.2% 48.1% 27.7% Far West 51 20.2 8.0 398 159 27.6% 39.9% 32.5% Xinjiang 16 8.8 3.4 554 214 30.0% 38.6% 31.4% Tibet 2 0.8 0.1 335 45 49.8% 13.4% 36.8% Qinghai 5 1 .9 0.8 419 180 23.1% 43.1% 33.9% Gansu 23 6.9 2.9 299 127 24.6% 42.7% 32.7% Ningxia 5 1.8 0.8 370 160 23.3% 43.3% 33.4% N. Hinterland 147 63.3 32.4 430 220 23.1% 51.2% 25.7% Heilongjiang 36 19.6 11.3 543 313 20.0% 57.6% 22.4% Jilin 25 11.8 5.9 466 233 25.4% 49.9% 24.6% Inner Mong. 22 8.7 3.5 393 158 33.5% 40.3% 26.2% Shanxi 30 11.9 6.6 400 222 16.0% 55.7% 28.4% Shaanxi 34 11.3 5.1 332 150 25.7% 45.1% 29.1% S. Hinterland 226 66.3 26.1 294 116 35.1% 39.3% 25.5% Sichuan 110 33.9 14.0 309 127 32.4% 41.1% 26.5% Guizhou 34 7.6 2.8 226 84 36.5% 37.0% 26.5% Yunnan 38 11.7 5.0 305 131 36.6% 42.9% 20.4% Guangxi 44 13.1 4.3 299 98 40.0% 32.8% 27.2% Central Core 305 101.2 42.8 332 141 32.0% 42.3% 25.7% Henan 89 27.8 12.5 314 141 29.2% 44.9% 25.9% Anhui 58 16.6 7.6 285 131 31.8% 45.9% 22.3% Jiangxi 39 12.8 4.6 328 117 35.9% 35.6% 28.5% Hubei 56 22.9 10.2 411 183 30.3% 44.4% 25.3% Hunan 63 21.1 8.0 336 127 35.2% 37.8% 27.0% North Coast 209 127.3 66.2 608 317 19.2% 52.0% 28.8% Liaoning 40 29.7 17.0 740 423 15.0% 57.2% 27.8% Greater Hebei 83 52.1 26.3 628 317 14.8% 50.5% 34.7% Shandong 86 45.4 22.9 527 266 27.0% 50.4% 22.6% East Coast 125 97.6 56.1 781 449 16.2% 57.5% 26.2% Jiangsu+S'hai 83 69.6 41.2 843 499 14.1% 59.1% 26.8% Zhejiang 42 28.0 15.0 660 353 21.5% 53.5% 25.0% South Coast 103 71.7 31.7 694 307 22.9% 44.2% 32.9% Fujian 31 15.9 6.7 511 214 28.5% 42.0% 29.5% Guangdong 65 52.5 24.1 805 370 20.3% 45.9% 33.8% Hainan 7 3.2 0.9 473 128 38.2% 27.1% 34.7% Based on official data for total GDP adjusted for domestic underreporting and non-market pricing converted to dollars at a trade- weighted average of China's two legal exchange rates in 1992 (see World Bank 1994, China: GNP per Capita); industrial data based on adjusted GDP data and official sector shares (column 7). -48- bBased on official data and reporting.rgg mi Souces: State Statistical Bueau, Statistical Yearbook 1993; Statistical Abstract 1994. [rSJdp5eli1n TABLE A.2: PER CAPITA GDP SECTOR DEVIATION FROM NATIONAL AVERAGE, 1980-1992 1 2----- 3 4 5 6 7 Total Agriculture Induga Services (Per-capita % (Per-capita % (Per-capita % (Per-capita % Deviation) Deviation) Deviation) Deviation) 1980 1992 1980 1992 1980 1992 1980 1992 China Total 0% 0% 0% 0% 0% 0% 0% 0% Far West .9% -12% 6% -5% -11% -28% -13% 0% Xinjiang _9% 21% -15% 46% 22% -6% -26% 30% Tibet #N/A -27% #N/A 47% #N/A -80% #N/A -8% Qighai 4% -9% 41% -15% -4% -21% -6% 6% Gansu -15% -36% -2% -35% -38% -44% -7% -27% Ningxia -11% -15% 2% -24% -21% -30% -10% -8% N. Hinterland 5% -10% 2% -12% 29% -3% 14% -15% Heiongjiang 50% 8% 15% 4% 29% 37% 78% -1 Jilin -1% -6% -9% 4% -3% 2% 3% -15% i.er Mong. -23% -18% -7% 16% -30% -31% -26% -23% Shanxi -4% -15% 2% -44% -40% -3% 15% -15% ShGanxi -26% -29% -29% 5% -27% -34% -24% -28% CentrCe -23% -1 -31 --9 -7% -49% -53% -42% Sichuan -31% -390/ -29% -12%% -43 Guizhou -52% -52% -57% -27% -35% -63% -61% -55% Yunnan -41% -35%1 -51% -2% -18% -43% -52% -54% Guangxi -38% -34% -24% 6% -10% -57% -62% -39% Central Core -23% -28% -29% -6% 6% -38% -37% -34% Henan -31% -33% -39% -19 -7% -38% 142% -39% Anhui -38% 40% 48% -20% 4% 43% -54% -53% Jiangxi -25% -31% -28% 4% 7% 49% -44% -30% Hubei -6% -13% -16% 10% 10% -20% -12% -22% Hunan -20% -29% -32% 4% -12% -44% -34% -32% North.... Coa.t i9 28-4%3 3% 39% 50% 35% Liaoning 53% 37% 5% -2% 02% Greater Hebei 50% 36%1 69% -18% -14% 39% 82% 62 Shandong -11% 10% -29% 25% 6% 16% -14% -11% East Coast 68% 61% 64% 12% H r16% 98% 101% 58% ianSu+S'ha 98% 74% 91% 5% 7% 118% 156% 68% Zhejiang 3% 29% -16% 25% 23% 54% -1% 23% South Coast -3% 59% 19% 39% 10% 41% -20% 79% Fujian -25% 12% -23% 28% -10% -6% -34% 12% Guangdong 4% 78% 27% 44% 15% 62% -13% 103% H-ainan #N/A 4% #N/A. 59%1 #N/A -44%, #N/A 22% sources: State Statistical Bureau, Compendium of Historical Provincial-level Statistics 1949- 89, Statistical Yearbook, 1992. - 49 - Table A.3: China's Real NMP Growth, by Region and Provincial-level Entity, 1978-91 Nort Chin Fa Hinter Heilng Inne Year Tota Wes Xnjian Tibe Qingha Gans Ningxi Ian jian Jili Mongli Shnx Shaanx 1978 1979 7.0 3.9 11.6 0.0 0.4 -0.1 3.5 6.2 1.6 4.3 4.6 16.0 9.0 1980 6.4 7.5 7.7 22.1 2.8 7.4 8.0 4.0 13.4 4.9 0.7 -3.9 -5.0 1981 4.9 -1.7 9.0 22.9 -8.0 -10.5 -0.7 2.5 -0.1 4.4 8.7 3.9 0.8 1982 8.2 81 8.7 -1.8 11.1 8.6 5.0 10.9 6.7 7.3 19.6 18.3 10.5 1983 10.0 12.2 11.7 -4.3 9.9 14.5 14.0 11.8 10.1 24.4 6.9 11.5 6.2 1984 13.6 15.2 12.8 62.2 15.2 13.3 15.8 13.8 8.5 12.1 15.5 21.7 17.1 1985 13.5 14.9 15.3 11.8 14.4 15.1 15.0 8.7 4.7 7.5 14.1 8.1 15.5 1986 7.7 9.6 9.1 -9.2 7.6 12.4 10.8 6.1 8.8 6.6 1.5 4.1 6.3 1987 10.2 8.2 9.1 11.4 6.4 7.8 6.2 8.4 7.9 16.3 6.3 3.1 8.6 1988 11.3 10.4 9.7 7.8 10.9 10.7 12.2 10.8 6.6 13.0 12.2 7.7 18.7 1989 3.6 6.3 5.7 2.7 0.5 8.1 8.1 2.0 3.8 -4.6 2.4 4.4 3.8 1990 5.1 5.7 8.4 4.5 2.1 5.0 2.2 4.4 3.6 4.2 6.7 4.9 4.0 1991 7.6 8.9 13.8 12.0 3.2 6.5 3.0 4.7 2.5 3.6 8.0 3.1 9.0 1992 14.4 9.3 9.8 5.8 7.1 10.0 6.9 9.7 6.0 12.7 9.4 12.1 10.8 1978-92* 9.0 9.1 10.2 10.2 7.1 8.7 9.1 7.8 6.5 9.0 8.8 8.3 8.8 1978-85* 87 81 10.5 14.2 5.9 6.1 8.1 8.1 6.6 9.7 10.2 10.5 6.8 1985-92* 80 81 8.9 5.2 5.0 8.3 7.3 7.3 5.5 6.3 6.4 4.8 8.6 1992* 1051. 37. 14. 1. 3. 15. 3. 120. 36. 23. 15. 22. 23. South China Hinter Central Total land Schuan Guizh. Yunn. Gngxi Core Henan Anhui Jngxi Hubei Hunan 1978 - - - - - - - - - - - - 1979 7.0% 8.2% 11.5% 6.6% 2.3% 4.9% 11.4% 10.2% 9.0% 12.7% 15.9% 9.5% 1980 6.4% 7.1% 7.5% 2.6% 7.4% 8.4% 5.3% 10.0% 3.7% 1.4% 5.3% 3.9% 1981 4.9% 4.5% 3.0% 4.2% 7.1% 7.3% 7.6% 7.8% 10.5% 4.9% 9.6% 4.5% 1982 8.2% 11.6% 8.4% 15.5% 15.0% 15.7% 7.4% 2.0% 8.1% 7.7% 10.1% 10.2% 1983 10.0% 7.9% 8.4% 13.2% 8.3% 2.6% 10.3% 20.6% 7.7% 6.1% 6.6% 7.4% 1984 13.6% 12.9% 14.2% 18.5% 13.5% 4.5% 14.5% 11.1% 19.7% 11.5% 20.5% 9.7% 1985 13.5% 12.6% 14.6% 7.8% 11.5% 11.0% 13.8% 12.6% 16.2% 14.5% 14.9% 11.4% 1986 7.7% 6.0% 5.3% 7.6% 3.5% 10.0% 6.8% 4.3% 10.4% 6.2% 6.4% 7.5% 1987 10.2% 9.7% 9.3% 10.4% 11.7% 8.6% 10.2% 15.4% 8.2% 9.6% 7.8% 9.2% 1988 11.3% 8.6% 8.2% 6.6% 17.8% 2.3% 7.9% 7.7% 8.2% 10.8% 6.3% 7.8% 1989 3.6% 2.9% 1.5% 3.7% 5.7% 3.9% 3.6% 3.7% 2.9% 6.8% 3.1% 3.0% 1990 5.1% 5.1% 3.4% 3.6% 9.4% 7.2% 3.1% 3.9% 3.5% 4.0% 1.5% 3.0% 1991 7.6% 7.8% 7.0% 9.9% 5.7% 11.5% 4.1% 6.7% -4.5% 7.4% 4.0% 6.9% 1992 14.4% 12.7% 12.4% 6.2% 9.6% 21.6% 14.6% 14.2% 24.0% 14.4% 12.3% 9.9% 1978-92* 9.0% 8.4% 8.2% 9.1% 9.9% 7.5% 8.6% 9.3% 9.3% 8.3% 9.0% 7.6% 1978-85* &7% 9.0% 8.8% 10.1% 9.7% 7.9% 9.5% 10.2% 10.1% 7.5% 11.0% 7.7% 1985-92* 8.0% 7.0% 5.6% 6.5% 9.7% 6.4% 6.4% 7.0% 5.0% 7.6% 4.8% 6.0% 1992* 1051.2 126.5 66.1 14.4 23.1 22.9 206.6 54.8 39.1 26.7 46.2 39.8 - 50 - Table A.3 (Continued) Chin Nort Liao- Great Shan Eas Jiangs Zhe Sout Guang Hai Tota Coas ning Hebe dong Coas +S'ha jian Coas Fujia don na 197 - 197 7.0 7.5 4.6% 9.3 6.7%13.3 13.0 14.1 6.0 5.5 7.7 -7.6 198 6.4 9.0 8.5% 8.6 10.3% 7.5 5.2 16.5 14.3 15.1 16.3 -11.6 198 4.9 2.1 -3.0% 1.6 8.0% 8.2 7.2 11.4 12.2 16.5 8.8 41.3 198 8.2 5.2 3.9% 4.3 8.1% 8.4 7.2 12.2 8.6 7.2 8.5 15.6 198 10.0 11.1 12.6% 10.4 11.2% 9.4 10.4 6.4 5.0 4.8 5.1 5.1 198 13.6 15.9 15.8% 14.5 18.7%17.5 16.0 22.3 15.8 16.6 15.3 18.4 198 13.5 13.1 12.3% 14.9 10.3%17.0 15.2 22.8 18.0 18.5 18.5 11.2 198 7.7 5.5 7.9% 3.4 7.5% 8.4 7.2 12.0 8.3 6.1 9.1 7.6 198 10.2 10.9 9.9% 8.3 16.8% 9.9 8.9 12.7 15.1 12.1 16.5 11.1 198 11.3 11.8 10.4% 11.4 13.7%11.9 12.0 11.8 16.0 15.7 16.7 9.8 198 3.6 2.7 2.1% 2.4 3.8% 1.2 1.5 0.4 7.1 9.0 6.6 5.2 199 5.1 3.3 -0.7% 4.1 5.0% 3.9 4.2 2.9 9.2 6.5 10.2 8.1 199 7.6 7.9 4.6% 6.6 12.6%10.0 7.7 16.3 17.1 14.6 18.5 9.1 199 14.4 14.9 11.7% 14.2 -18.2%22.6 23.5 20.3 22.9 22.3 23.2 21.9 1978-92 9.0 & 7 7.8% 8.0 10.7% 10.2 9.3 12.6 12.2 11.5 12.1 10.5 1978-85 8.7 8.5 7.2% 8.2 10.5% 10.8 9.8 14.0 10.9 11.6 10.7 11.3 1985-92 & 0 7.5 5.6% 6.2 9.7% &3 6.8 8.3 12.9 10.8 12.6 8.4 1992 1051. 282. 57.1 136. 88.7 238. 173. 64. 130. 29. 93. 6. Note: Average growth for Tibet and the Far West are for 1979-91, for lack of 1978 data. * Average growth from logrithmic regression trend line; hence 78-91 is not ave. of sub-periods. - 51 - TABLE A.4: URBAN HOUSEHOLD REAL REGIONAL DISPARITIES, 1981-92 (1981 Yuan per Capita) 1 2 3 4 5 6 7 8 9 Urban HH Income Urban HH Expenditure Urban Food Consumption 1981 1985 1992 1981 1985 1992 1981 1985 1992 China Average 436 591 790 428 572 721 246 301 384 Far West 413 608 738 391 560 659 214 273 334 North Hinterland 382 512 620 371 495 580 175 252 289 South Hinterland 402 554 793 403 569 734 240 298 398 Central Core 414 534 710 412 526 659 237 278 352 North Coast 465 616 824 456 594 748 252 313 393 East Coast 504 726 956 498 695 858 280 362 463 South Coast 429 588 1,126 405 581 1,009 263 336 536 TABLE A.5: RURAL HOUSEHOLD REAL REGIONAL DISPARITIES, 1981-92 (1981 Yuan per Capita) 1 2 3 4 5 6 7 8 9 Rural HH Income Rural HH Expenditure Rural Food Consumption 1981 1985 1992 1981 1985 1992 1981 1985 1992 China Average 242 356 404 200 290 331 118 164 186 Far West 178 275 289 140 214 241 91 129 141 North Hinterland 213 308 337 174 245 263 105 140 158 South Hinterland 209 277 308 171 239 271 113 154 168 Central Core 229 333 327 186 271 284 113 161 168 North Coast 239 373 412 192 295 335 99 151 175 East Coast 276 449 576 249 386 498 139 199 256 South Coast 0 275 407 557 228 330 455 137 201 252 - 52 - ANNEX 2: LITERATURE REVIEW A. METHODOLOGIES IN THE ACADEMIC LITERATURE ON CHINA'S REGIONAL DIsPARmHEs 1.18. Academic literature on China's regional inequality has gnerally reflected an overall interest in global measures of inequality rather than policy-oriented interregional comparisons. The literature has examined statistical measures of spatial inequality, their constituent components (e.g. industry or agriculture, rural or urban, -within or between provinces), their trends over time, and their regional significance--that is, which parts of China are affected in which ways. The data and methods employed, however, vary significantly, and the overall literature effort is more useful as a evaluation of trends in national inequality than as an analysis of geographical regions, their economic trends, and the policy steps implied by emerging regional disparities. 1. Nature of the Data A. 19. Researchers on Chinese spatial economics generally base their empirical analysis either on economic aggregates (e.g., national accounts, fiscal and financial data, and foreign trade and investment statistics) or on household survey data--both aggregated and disaggregated. Most research uses data in forms most readily available in official publications, but a few have had access to raw official survey data or special survey data. The kind of data used usually sets limits on the kinds of analysis possible, and in general, data most useful for testing hypotheses over time are less useful for the kinds of spatial analysis which survey data make possible. A.20. By far the most frequently used data are sets of official statistics for China's 30 provincial level entities (provinces). China publishes large volumes of such data, which include national accounting aggregates, household survey provincial averages, and most other national statistics. The convenience of provincial data lies in their ready availability and in the many statistical results which 30 data points make possible. The shortcomings of regional analysis based on provincial data, however, are significant, and the best research based on provincial data modifies the data before subjecting them to analytical inspection or statistical manipulation. A.21. The most serious shortcoming of provincial-level data results from the inclusion of three provincial-level municipalities (Beijing, Tianjin and Shanghai) in the group of 30. These municipalities are atypical, in that they have very little agriculture, small rural population components, and a high concentration of industrial, commercial, and cultural activities--merely by nature of their administrative treatment as provincial-level units. Conversely, the "provinces" to which these cities originally belonged (Hebei and Jiangsu), have atypically smaller urban components, weakening their comparative usefulness. A few researchers (Tsui 1991) and Keidel 1993) in the literature modify their data by reducing the number of provinces to 27 through inclusion of Beijing and Tianjin in Hebei ("Greater Hebei") and Shanghai in - 53 - "Greater" Jiangsu. Some (World Bank 1992 and Keidel 1993 ) also combine selected provinces into a smaller number of meaningful economic regions. A.22. A second shortcoming of provincial data is the variation in population size from province to province. Some research (Denny 1991) considers "how many" provinces have improved their relative status over the relevant years in the research, but such treatment compares provincial populations in the Far West of 2, 4 and 5 million persons (Tibet, Qinghai, and Ningxia) with provinces elsewhere of 85, 86 and 108 million persons (Shandong, Henan and Sichuan)." Other research includes population weighting in its techniques. A.23. Related to issues of population size by province is the growing significance of undocumented migration and hence the growing inaccuracy of provincial population figures (see Weimin Yang 1991). Unreported migrant workers (the "floating population") are not counted where they live and work but rather where they are officially registered citizens (frequently in poorer rural areas). More accurate provincial population data would probably offset measures of regional disparities by necessitating a larger denominator in all per-capita calculations for relatively well-off regions. A.24. The final difficulty with raw provincial data is the absence of corrections for pricing variation--over time and from one province to the next. Howes and Lanjouw (1991) analyze provincial price indexes for 1986 to show that the variation can be as large as the difference between an index of 0.93 in the South Hinterland (Sichuan) and A.32 on the South Coast (Guangdong). In general, however, China publishes no regional price surveys, although the State Statistical Bureau (SSB) reports that it would like to if it had more resources. Because poorer provinces in the interior are also generally known to have lower prices for the same goods and services, better data on regional price variations would probably reduce measured regional disparities in output, income and consumption. A.25. The second major data category common in regional analysis of China includes household survey data and some national-income statistics at the sub-provincial level. These are usually raw survey data or county-level aggregates. Research using these data has the advantage of exploring a wide range of hypotheses about the correlates of inequality, and in one case, Khan et al. (1992), use of a one-time fuller household income and expenditure survey (1988) enhanced analysis of survey data and summary results for other years. These more detailed data have the difficulty that they are usually only available for urban data or for rural data, or they are limited to a subset of provinces, or they exist only for one year. The advantages more than offset these shortcomings, however, and the richness of the resulting analysis makes important contributions to the overall literature. A.26. Finally, the third source of data compares very large regional aggregates, either by using published data on China's "three belts" (see paragraph A.3) or by comparing data for one province (such as Guangdong) with corresponding data for all of China. These data have the advantage of convenience and simplicity, and they generally indicate regional-disparity 6 1990 census data. - 54 - levels and tendencies with the broadest significance, such as for growing coastal-interior disparities. Such broad analysis is less useful for understanding causal mechanisms, which must rely on less aggregated statistics. In conclusion, the existing literature on regional economic disparities in China relies on a range of different kinds of statistics, and each kind makes important contributions. 2. Hypotheses and Research Techniques A.27. Research in the literature is of two basic sorts: (a) global statistical measures of inequality for the nation as a whole and (b) analysis of differences between identified regions. In a few studies the two approaches merge, when disaggregated data are plentiful enough to allow decomposition of inequality measures into between-region and within-region components, and nearly all studies of global inequality make some inferences about the regional significance of their results. A.28 In general, studies measuring global inequality test the hypothesis that inequality has worsened or improved over a particular time period--from the pre- to post-1978 era or within the reform period. They usually make little reference to the regions or regional phenomena associated with these trends. The techniques include calculation of basic dispersion statistics, such as the standard deviation, coefficient of variation, Theil entropy measure, Atkinson index, and gini coefficient (Lyons 1991 and Tsui 1991). The conclusions are at times sensitive to the statistic chosen; Lyons (1991) finds that "absolute" inequality (standard deviation) has increased but "relative" inequality (coefficient of variation) has lessened. Decomposition of national data at intra-provincial and sectoral levels of disaggregation permits some analysis of how rural-urban differences and sector-output specialization contribute to overall national levels of inequality, but data availability has limited the robustness of such research. A.29 Investigation of regional differences by comparing provinces and regions relies heavily on inspection and shifts in ratios between leading and lagging provinces or regions. Many studies compare economic indicators between coastal and interior provinces or coastal regions (or "belts") and interior ones. Implicit in these methodologies is the significance attached to changes in ratios or percentage differences between provinces and how these change over time. One popular technique (Denny 1991, World Bank 1992 and Keidel 1993) compares percentage deviation from the national mean of certain key indicators for relevant provinces or regions. For this kind of analysis, the treatment of cities like Beijing and Shanghai and of low-population provinces like those in the Far West significantly influences the results (see paragraphs A.21 and A.22). A.30 Overall, statistical techniques for measuring national levels of inequality for various indicators are more precise and more analytically focused than comparisons between identified provinces and regions based on inspection of ratios and percentage differences. They also permit useful decomposition to gauge the contribution to inequality of such characteristics as ownership (e.g., state-owned or township-and-village enterprises, known as TVEs). However, by themselves, they have limited usefulness for reaching conclusions about regional - 55 - disparities. Methods employed for gauging regional disparities are cruder and require inspecting indicator levels for specific locations in comparison to others. What both usually lack as methodologies is an accompanying standard about what constitutes significant inequality or significant disparities and what constitutes noteworthy shifts over time. In this regard, poverty lines and established minimum values for social indicators become important, as they are in World Bank (1992). B. FINDINGS IN THE LITERATURE A.31 Evaluation of existing literature as a body of research leads to the conclusion that regional disparities in per-capita output, income and consumption have increased in the post-1978 reform period compared to the pre-reform period, and that in the course of post- 1978 reforms, the increase before 1985 is not as significant as that after 1985, when the widening of interregional disparities accelerated. See Auguinier (1988), Denny (1991), Lyons (1991), Dali Yang (1989), Weiming Yang (1991), Tsui (1991), and Keidel (1993). Not all research findings in the literature conform uniformly with this conclusion (see Denny 1991 and Weiming Yang 1991), but when data and methods are taken into consideration, it nevertheless appears justified. The most important correlates of these shifts in inequality measures are widening rural-urban gaps for various economic indicators, given uneven regional concentration of cities, differential growth in agriculture and industry-whose output is unevenly distributed by province and region--and special advantages coastal areas enjoy regarding foreign trade and investment. A.32 Conclusions regarding growing inequalities in consumption are in general better supported by the literature than inequality in output, in light of the availability of disaggregated household survey data which show growing inequality within provinces and regions as well as between them. These studies show that the growing urban-rural gap is even more important for explaining spatial inequality within provinces than between provinces (Lyon 1991), and that official data understate the degree of spatial inequality compared to more carefully designed special survey results (Khan, et al. 1992). Several studies document how regional consumption inequality has become more serious for rural areas over time, but not so much for urban areas, in that coastal areas have experienced a rapid increase in rural household income throughout the reform period, while the increase slowed significantly after 1985 for the rest of the country's rural regions (Khan, et al. and Keidel). A.33 In terms of more detailed findings regarding different regions which go beyond global measures of inequality, most of the literature is limited by its data foundations, but several works have regionally significant conclusions. The most common observations are that the coast has improved its position by a variety of indicators vis A vis the interior, that certain coastal provinces (e.g., Guangdong) have outperformed coastal metropolitan areas (Beijing, Tianjin and Shanghai), and that certain Western regions (e.g., Xinjiang) have also reduced apparent disparities vis A vis the rest of the country. A few studies point out that disparities have lessened between the South Coast and other coastal areas (Weiming Yang 1991 and Keidel 1993). World Bank (1992) presents a thorough analysis of the incidence of poverty and a variety of related national indicators by named region, indicating that hinterland (especially - 56 - the South Hinterland) and central core (especially Henan) regions are much the most seriously affected. Keidel (1993) concludes that a "structural" gap reflecting industrial expansion on the coast and agricultural dependence in the interior is a particularly important kind of emerging regional disparity. A.34 Numerous studies also examine fiscal, financial and foreign commercial disparities by region in ways which complement the analysis of output and consumption. The findings concerning fiscal flows are that even though since 1978 central government revenues have declined significantly as a share of GDP, significant central government transfers in support of fiscal-deficit provinces have continued and increased through the end of the 1980s, although not necessarily in per-capita terms (Denny 1991, Tsui 1991, World Bank 1992 and Keidel 1993). Keidel (1993) shows that the scale and regional incidence of credit creation heavily overshadows the differential impact of fiscal transfers in a year of rapid national credit growth (1991) compared to a year of credit austerity (1990). For foreign trade and investment, after discounting analysis based on growth rates from very small starting values (Denny 1991), the literature generally supports the strong conclusion that foreign trade and investment have overwhelmingly favored the coast, especially the South Coast, although trade data for foreign-funded firms in the interior began to show some countervailing activity in the early 1990s (Weiming Yang 1991, Kato 1992, Keidel 1993). A.35 In sum, the literature's findings are that regional disparities favoring the coastal areas appear significant and have increased markedly since the outset of reforms, that inter- regional consumption disparities for rural areas are particular noteworthy, that poverty is concentrated in particular interior regions, and that although fiscal transfers continue to perform an important corrective role, the regionally differential impact of credit expansion and foreign trade and investment appear overpowering by comparison. C. OVERALL LITERATURE ASSESSMENT A.36 Overall, the literature on China's regional economic disparities presents a good description of disparities between provinces, and in a few cases between groupings of provincial-level units, although in most cases the statistics are somewhat out-of-date. The literature identifies important related issues--especially rural-urban differences and uneven industrial expansion--and indicates some of the most obvious factors behind these trends, such as foreign-trade and foreign-investment imbalances by region. The literature is less helpful for considering the policy background relevant for these patterns and trends and applying an evaluative framework. Nor does it do a good job of indicating the implications of these trends for further economic reforms-although several studies draw strong conclusions with regard to their significance for migration pressures. A.37 More than anything, the literature is weak in providing an evaluative framework for considering which trends in regional disparities are desirable from an economic- development-policy perspective. In particular, the tone of most studies appears to consider increasing disparities as a problem and lessening disparities as a good sign. There is very little analysis of under what circumstances growing disparities might be a promising sign and under - 57 - what circumstances they would be a sign of trouble (although as a descriptive benchmark, some studies do mention Kuznets' "inverted-U" hypothesis). In other words, the attention might best be place not on whether disparities are increasing, but on the ways and degrees in which disparities are increasing that are relevant for development goals. A.38 In terms of policy evaluation relevant for explaining regional disparities, most of the literature concentrates on economic reforms in general and on the "coastal strategy" of the 1980s (e.g., Aguinier 1988, Brudvig 1993, Kato 1992, Keidel 1993, Linge and Forbes 1990b, Dali Yang 1990, 1991). Keidel (1993) treats the shift to a more balanced regional-interior strategy in 1991. A few studies mention policies promoting administrative decentralization (e.g., Aguignier 1988) and rural TVE industrialization (World Bank 1992 and Keidel 1993). There is little treatment, however, of the degree to which natural advantages favor coastal areas and other transport centers, especially as decentralized market reforms proceed. In other words, there is little counter-historical analysis concerning what trends in disparities might have been like in the absence of special promotional coastal-zone policies. A more balanced treatment in this regard would strengthen the literature's analysis of the overall nature of China's regional disparities phenomena. A.39 In sum, the existing literature is a good basis for considering many dimensions of China's regional trends and differences, although one of its most immediate needs is a statistical update for the 1992-93 years of credit expansion and rapid growth following the retrenchment of 1989-91. In addition, the literature would be greatly strengthened by analysis of disparities between specific sub-provincial regions--which in numerous cases represent sub- economies with as many as 15-to-25 million persons each. Finally, the literature's interest in migration warrants thorough expansion when considering the significance of emerging regional disparities and the policy directions they imply. BIBLIOGRAPHY Aguignier 1988. P. Aguignier, "Regional Disparities Since 1978," in Stephen Feuchtwang, Athar Hussain and Thierry Pairault, eds., Transforming China's Economy in the Eighties, Volume H: Management, Industry and the Urban Economy, (Boulder: Westview Press, 1988) pp. 93-106. Bannister 1991. Judith Bannister, "China's Population Changes and the Economy" in Congress of the United States, Joint Economic Committee, China's Economic Dilemmas in the 1990s: The Problems of Reforms. Modernization, and Interdependence, Vol. 1 (Washington, D.C., April 1991), pp. 234-269. Brudvig 1993. Lee A. Brudvig, "The Fifth Dragon" in The China Business Review, July-August 1993, pp. 14-20. Cannon 1989. Terry Cannon, "National Minorities and the Internal Frontier" in Goodman 1989. pp. 164-179. - 58 - Cannon 1990. Terry Cannon, "Regions: Spatial Inequality and Regional Policy" in Cannon and Jenkins 1990. pp. 28-60. Cannon and Jenkins 1990. Terry Cannon and Alan Jenkins (eds.), The Geography of Contemporary China: the Impact of Deng Xiaoping's Decade (New York: Routledge, 1990). Cannon and Jenkins 1990b. "Postscript: the Impact of China's 1989 Crisis on Regional Development and Spatial Change" in Cannon and Jenkins 1990 pp. 291-305. Denny 1991. David Denny, "Regional Economic Differences During the Decade of Reform" in Congress of the United States, Joint Economic Committee, China's Economic Dilemmas in the 1990s: The Problems of Reforms. Modernization, and Interdependence (Washington, D.C., April 1991). Derbyshire 1990. Edward Derbyshire, "Environment: Understanding and Transforming the Physical Environment" in Cannon and Jenkins 1990, pp. 80-101. Ferdinand 1989. "The Economic and Financial Dimension" in Goodman 1989 pp. 39-56 Garnaut and Liu 1992. Ross Garnaut and Liu Guoguang, eds., Economic Reform and Internationalization: China and the Pacific Region (St. Leonards: Allen & Unwin, 1992 Goldstein 1993a. Carl Goldstein, "Into the Black" in Far Eastern Economic Review. April 1, 1993, pp. 66-67. Goldstein 1993b. Carl Goldstein, "Get Off Our Backs" in Far Eastern Economic Review. July 15, 1993, pp. 68-70. Goodman 1989. David S.G. Goodman, ed., China's Regional Development (New York: Routledge, 1989). Howard 1989. Maurice Howard, "Transport and Communications" in Goodman 1989, pp. 77-93. Howard 1990. Maurice Howard, "Industry, Energy and Transport: Problems and Policies" in Cannon and Jenkins 1990 pp. 168-199. Howes and Lanjouw 1991. Stephen Howes and Peter Lanjouw, Regional Variations in Living Standards in Urban China (London: London School of Economics, The Development Economics Research Programme, October 1991) Hua and Jing 1993. Hua Juxian and Jing Wei, "Border Trade in Minority Regions" in Beifin Review March 22-28, 1993, pp. 13-17. Huang 1993. Huang Wei, "Blueprint for Yangtze Valley Progress in 90s" in Beijing Review, June 28-July 4, 1993, pp. 13-17. Hull 1992. T. H. Hull, "Spatial Aspects of the Natural Increase of China's Population" in Linge and Forbes 1990, pp. 109-128. Jing 1993. Jing Wei, "Border Trade in Minority Regions" in Beijing Review, March 22-28, pp. 13-17. Jowett 1990. John Jowett, "People: Demographic Patterns and Policies" in Cannon and Jenkins 10 pp. 102-132. Kato 1992. Hiroyuki Kato, "Regional Development in the Reform Period," in Garnaut and Liu i , pp. 116-136. Keidel 1993. Albert Keidel, "China's Regional Economic Performance," (Washington, DC: Rock Creek Research, Inc.) Draft Report, U.S. Department of State. Khan, et.al. 1992. Azizur Rahman Khan, Keith Griffin, Carl Riskin and Zhao Renwei, "Household Income and its Distribution in China," China Ouarterly, December 1992, pp. 1029-1061. - 59 - Leeming and Powell 1990. Frank Leeming and Simon Powell, "Rural China: Old Problems and New Solutions" in Cannon and Jenkins 1990, pp. 133-167. Li 1992. Wen-yan Li, "Contemporary Spatial Issues" in Linge and Forbes 1990, pp. 59-84. Li, B. 1993. Li Bin, "County-level Reform to be in Full Swing" in Beijing Review Fegruary 15- 21, 1993, pp. 20-21. Li, R. 1993. Li Rongxia, "Free Trade Zones in China" in Beijing Review August 2-8, 1993, pp. 14-19. Linge and Forbes 1990. G.J.R. Linge and D.K. Forbes, China's Spatial Economy: Recent Developments and Reforms (Hong Kong: Oxford University Press, 1990). Linge and Forbes 1990b. G.J.R. Linge and D.K. Forbes, "The Space Economy of China" in Linge and Forbes 1990. pp. 10-34. Lockett 1989. Martin Lockett, "Foreign Trade" in Goodman 1989, pp.57-76. Lyons 1991. Thomas P. Lyons, "Interprovincial Disparities in China: Output and Consumption, 1952-1987," Economic Development and Cultural Change, 1991, pp. 471-506. Manguno 1993. Joseph P. Manguno, "A New Regional Trade Bloc in Northeast Asia?" in The China Business Review, March-april 1993, pp. 6-11 Mooney 1992. Paul Mooney, "At China's Crossroads: Neglected Wuhan City Enters the Spotlight" in Far Eastern Economic Review November 19, 1992, pp. 66-68 Phillips and Yeh 1989. David R. Phillips and Anthony G.O. Yeh, " Special Economic Zones" in Goodman 1989. pp. 94-134 Phillips and Yeh 1990. David R. Phillips and Anthony Gar-On Yeh, "Foreign Investment and Trade: Impact on Spatial Structure of the Economy" in Cannon and Jenkins 1990. pp. 224-248 Ross and Rosen 1992. Madelyn C. Ross and Kenneth T. Rosen, "China's Real Estate Revolution" in The China Business Review November-December 1992, pp. 44-51. Sender 1993. Henny Sender, "Convert Your Renminbi: Money Rushes Between Hongkong and China" in Far Eastern Economic Review, April 1, 1993, pp. 67-68. Shen 1992. Wei-cheng Shen, "Development and Problems of China's Seaports" in Linge and Forbes 1990. pp. 96-108. Sung 1991. Yun-wing Sung, The China-Hong Kong Connection: The Key to China's Open-door Poliy (Cambridge: Cambridge University Press, 1991). Sung 1992. Yun-wing Sung, "The Economic Integration of Hong Kong, Taiwan and South Korea with the Mainland of China," in Garnau and Liu 1992, pp. 149-181. Tsui 1991. Kai Yuen Tsui, "China's Regional Inequality, 1952-1985" in Journal of Comparative Economics, No. 15, March 1991, pp. 1-2 1. Vogel 1989. Ezra F. Vogel, One Step Ahead in China: Guangdong Under Reform (Cambridge: Harvard University Press, 1989). Wong 1991. Christine P.W. Wong, "Central-Local Relations in an Era of Fiscal Decline: The Paradox of Fiscal Decentralization in Post-Mao China" China Quarterly, December 1991, pp. 691-715. Wu 1993. Wu Naitao, "Reform Boosts Foreign Trade" in Beijing Review March 1-7, 1993, pp. 20-22. World Bank 1992. World Bank, China: Strategies for Reducing Poverty in the 1990s (Washington, D.C.: World Bank, 1992). - 60 - Wu 1993. Wu Naitao, "Reform Boosts Foreign Trade" in Beijing Review March 1-7, 1993, pp. 20-22. Wu and Xu 1992. Chung-Tong Wu and Sueqiang Xu, "Economic Reforms and Rural to Urban Migration" in Linge and Forbes 1990, pp. 129-143. Yang, W. 1991. Weimin Yang, "Analysis of Changes in the Income Gaps Among Different Regions in China" (China, State Planning Commission, unpublished photocopied paper, undated, but with data which indicate it was written in 1991-92). Yang, D. 1990. Dali L. Yang, "Patterns of China's Regional Development Strategy," in CLn& Quarterly June, 1990, pp. 230-57. Yang, D. 1991. Dali L. Yang, "Reforms, Resources, and Regional Cleavages: The Political Economy of Coast-Interior Relations in Mainland China," Issues and Studies, Vol. 27, No. 9, September 1991, pp. 43-69. Yao 1993. Yao Jianguo, "Yangtze River Valley: A Soaring Dragon" in Beijing Review February 15-21, 1993; pp. 14-19. Zhang 1993. Zhang Zhiping, "Border Trade Reaches New Stage" in Beijing Review, 31 May to 6 June, 1993, pp. 18-21. Zou 1992. Zou Jiahua, "Plans for Regional Economy" in Beijing Review August 3-9, 1992, pp. 18-22. t CATALOGUERS/FILE CONFIDENTIAL Report No: 14496 CHA Type: SR