77465 Regulated Efficiency, World Trade Organization Accession, and the Motor Vehicle Sector in China Joseph F. Francois and Dean Spinanger This article is concerned with the interaction of regulated efficiency and World Trade Organization (WTO) accession and its impact on China’s motor vehicle sector. The analysis is conducted using a 23 sector–25 region computable general equilibrium model. Regulatory reform and internal restructuring are found to be critical. Restruc- turing is represented by a cost reduction following from consolidation and rationaliza- tion that moves costs toward global norms. Without restructuring, WTO accession means a surge of final imports, though imports of parts could well fall as production moves offshore. However, with restructuring, the final assembly industry can be made competitive by world standards, with a strengthened position for the industry. Producing automobiles has often been a symbol of economic prestige in the developing world. Brazil, China, Indonesia, Malaysia, and others have all promoted and sometimes even showcased the development of a domestic motor vehicle industry. In China, with its huge population and a surface area roughly as large as the United States and almost 15 percent larger than Brazil (table 1), almost every province has its own motor vehicle factory and satellite factories. But despite all the factories, China has the largest number of people per vehicle among major economies in the world. Even Indonesia, with a 30 percent lower per capita income, has only half as many people per automobile. China’s status as a country with one of the highest number of people per vehicle is the outcome of a series of policy measures, dating as far back as the inception of the People’s Republic of China, that have distorted the structure of the automobile industry (table 2). Internal measures limited and even pro- hibited trade through local protectionism (analogous to former interprovincial trade restrictions in Canada). The government has also set prices and limited competition through a barrage of import restrictions, which have included Joseph F. Francois is professor of economics and research fellow at the Tinbergen Institute and research fellow at the Centre for Economic Policy Research; his e-mail address is francois@few.eur.nl. Dean Spinanger is senior research economist at the Institute for World Economies in Kiel; his e-mail address is dspinanger@ifw.uni-kiel.de. The authors thank Zhang Wenkui for help with data, Will Martin and Alan Winters for detailed comments on an earlier draft, and three anonymous referees who offered valuable suggestions. They also thank participants in a World Bank–sponsored conference in Beijing for helpful discussion. THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1, Ó The International Bank for Reconstruction and Development / THE WORLD BANK 2004; all rights reserved. DOI: 10.1093/wber/lhh034 18:85–104 85 86 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1 T A B L E 1 . GNP, Population, and Stocks of Automobiles in China and Selected Countries, 2000 GNP per Vehicle stock, capita Population, 2001 (millions) (2002 2001 People per Surface area Economy PPP $)a (millions) Automobiles Trucks automobile (1,000 sq. km) Low and middle 4,682 3,274.4 140.6 54.9 23.3 49,263 income India 2,570 1,032.4 6.3 5.9 163.2 3,287 Indonesia 2,990 209.0 3.0 2.4 68.8 1,905 China 4,390 1,271.8 8.5 15.4 149.0 9,598 Colombia 5,870 43.0 1.8 0.8 23.4 1,139 Turkey 6,120 66.2 4.5 1.6 14.6 775 Thailand 6,680 61.2 2.9 4.1 21.4 513 Brazil 7,250 172.4 15.8 4.0 10.9 8,547 Russia 7,820 144.8 21.2 5.1 6.8 17,075 Malaysia 8,280 23.8 4.2 1.0 5.6 330 Mexico 8,540 99.4 12.2 5.6 8.2 1,958 South Africa 9,870 43.2 41.0 2.5 1.1 1,221 Argentina 9,930 37.5 5.4 1.6 7.0 2,780 High income 29,248 742.7 351.6 129.4 2.1 21,937 Korea, Rep. of 16,480 47.3 8.9 4.0 5.3 99 Taiwan, China 17,730 22.4 4.8 0.9 4.6 36 Spain 20,460 41.1 18.2 4.2 2.3 506 Italy 25,320 57.9 33.2 3.8 1.7 301 United Kingdom 25,870 58.8 27.8 3.4 2.1 243 Japan 26,070 127.0 53.5 19.9 2.4 378 France 26,180 59.2 28.7 5.9 2.1 552 Germany 26,220 82.3 44.4 3.6 1.9 357 Canada 28,070 31.1 17.1 0.7 1.8 9,971 United States 35,060 285.3 128.7 88.0 2.2 9,629 a PPP is purchasing power parity. Source: World Bank, various years, World Development Indicators ; Verband der Automobilindustrie, various issues. quotas, high tariffs, and differential taxes favoring local suppliers. The restric- tions on trade have encouraged inefficient production and allowed for market segmentation. China’s integration into the World Trade Organization (WTO), and thus into most favored nation principles, has important implications for its economy, especially the motor vehicle sector. Accession agreements define major changes in tariffs, quotas and local content requirements, and rules on foreign investment. There has already been a change in market perceptions by outside investors, as the application of WTO rules on treatment of foreign firms has reduced uncertainty about the general economic climate, inducing notable increases in investment and prompting new decisions about entering the market. Francois and Spinanger 87 T A B L E 2 . Summary of Developments in the Chinese Automotive Sector Period Characteristics 1953–65: Self-reliance policy Roughly 60,000 vehicles produced per year. Relied on Soviet technologies. No other international contacts. Provincial governments set up production units. By 1960, 16 auto producers and 28 assembly companies. 1966–80: Security oriented Government invested heavily in western regions (Sichuan, Shanxi, and Hubei). Remote locations caused severe problems and overcapacity. Focus on heavy military vehicles. Car demand increased rapidly and capacities expanded to 160,000 units a year. By 1980, 58 carmakers, 192 assembly companies, and 2,000 spare parts producers. 1981–98: Initial fruits of open-door policy Open-door policy in 1978 kick-started industry. VW already started in 1978. Number of companies almost doubled during 83–85, from 65 to 114 units. By 1998, roughly 2,500 production units. Provincial governments further regionalized production. Major international firms begin to invest and then stop rapidly. Joint ventures accounted for about 60 percent of production. 1999–present: opening up and beyond Major investments by foreign companies. All major Japanese and German companies in China. French, Italian, and U.S. producers nominally present. Rapid expansion; capacity now near 2.5 million units. Growing capacity in costal areas. Source: Summary produced by authors from various sources. This article is concerned with the impact of these broad changes on the Chinese motor vehicle sector. It emphasizes the role of administratively imposed inefficiencies (‘‘regulated efficiency’’) within the sector and the role of such regulated efficiency in structural adjustment. The industry itself anticipates significant change. In recent years the sector has grown rapidly, with output expanding at an annualized rate of 13 percent in the four years ending in 1999, at a rate of 26 percent in the three years to 2002, and at more than double that in 2003. With modern plants having come on line in 2001 and 2002, and 88 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1 additional facilities expected to increase capacity by more than 150 percent from 2002 to 2005, a large, discrete change in production levels is expected. At the same time, WTO membership implies lower prices and steeper foreign competition in the sector. Response to this shift in the competitive landscape will be shaped by continuing problems with local government protection, lack of automobile infrastructure (roads, parking, service facilities), and related factors that act as constraints on growth of the sector. Even so, the industry itself expects continued strong growth.1 Notwithstanding industry expectations, what can realistically be expected once the competitive landscape has changed in critical ways? This question is explored here using a computable general equilibrium (CGE)model. I. THE AUTOMOBILE INDUSTRY IN CHINA National and regional policies in China have resulted in a highly fragmented and inefficient motor vehicle industry by global standards. This was not only the result of the introduction of Soviet-style industrialization beginning in the 1950s, with firms viewed merely as production units producing according to plan, making questions about efficiency irrelevant.2 It was also the result of import substitution policies and cooperation agreements with foreign com- panies beginning in the 1980s that were meant to fill the increasing gap between supply and rapidly expanding demand for automobiles. The major thrust of policies was to build trucks, not passenger cars (figure 1). Motor vehicle companies are thus operating with cost structures that are well within the global frontier, with plants that are producing considerably below global standards for efficient scale (table 3). For plants producing a single model, minimum efficient scale for final assembly of cars has been estimated 1. See, for example, China Online (2000). As WTO membership approached, the opinions of the industry and related ministries, as reflected in the Chinese press, hinged critically on whether restructur- ing of the domestic industry would be allowed to proceed. Thus a report in Touzi Yu Hezuo (summarized in China Online 2000) stressed expected injury to the industry, whereas the industry itself was expressing optimism that it could realize significant cost reductions and remain competitive with imports (Feenstra and others 2001). In the meantime, price cuts by foreign producers in China are becoming common, some of them induced by increased import competition and others by more intense domestic competition. Buick, for instance, reduced prices on its domestically produced models by 12 percent, and Volkswagen lowered Passat prices by 6.5 percent (indiacar.net, May 3, 2002). Even more important, nearly all major foreign producers have announced plans to establish or sizably increase production capacities. A recent major manufacturer to do so was DaimlerChyrsler in September 2003, finally ratifying plans to establish facilities to produce C and E models in China (International Herald Tribune, September 9, 2003). 2. As noted by Zhang and Taylor (2001, pp. 261ff.), First Automobile Works (FAW) provides ample evidence of the impact of various policies over the past 50 years. Between 1959 and 1981 FAW produced a mere 1,542 units, an average 67 units annually. In 1970 the production cost of a particular model (the CA72) was 220,000 yuan, but ‘‘the sales price was only 40,000 yuan. . . . In the absence of competition, all production units ran at low levels of productivity and efficiency. . . . By 1980 the number of automotive enterprises had risen to 2,379, consisting of 56 vehicle manufacturers . . . [producing] 5,418 cars.’’ Francois and Spinanger 89 F I G U R E 1 . China’s Production of Motor Vehicles, 1984–2002 (thousands) 3,500 3,000 2,500 2,000 Total motor vehicles 1,500 1,000 Trucks and buses 500 Passenger cars 0 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Source: Bessum (2002) and Verband der Automobilindustrie (various issues). at more than 200,000 units per plant per year (Huang 2002, p. 543). China’s entire sedan production in 1998 was 507,000 vehicles produced in 13 factories. Of these, only one factory produced more than 200,000 sedans. Several plants had production runs of fewer than 20,000 sedans.3 In 1998 China had 122 motor vehicle manufacturing plants, 520 automobile refitting factories, 130 motorcycle factories, 62 car engine factories, and 1,589 auto- mobile and motorcycle spare parts factories. Annual production capacity now exceeds 2.3 million motor vehicles and 10 million motorcycles. Since 1995 the general pattern has been to shut down smaller plants (generally relegated to the ‘‘other’’ category in table 3), and expand production runs in the larger plants. With foreign investment and rapid growth in the industry, the number of plants producing at least 25,000 vehicles rose from 3 in 1995 to 11 in 2002. Import and domestic shipment data in value terms for 1997, the preaccession reference point, are summarized in table 4. Reflecting relative tariff differences, imports are concentrated in parts rather than in vehicles. China’s preaccession average tariff on automobile products (vehicles and parts) was 35 percent (table 5). The rate for vehicles averaged 70 percent, with sedans subject to tariffs of 80–100 percent. Parts were subject to an average tariff of 23 percent. Import shares were 3. There are strong parallels to the situation in Mexico before the North American Free Trade Agreement (Lopez de Silanes and others 1994), where protected, inefficient factories operated well within the global technology frontier. T A B L E 3 . Number of Passenger Cars Produced by Plants in China, 1995–2002 Rank 2002/1995 Plant 1995 1996 1997 1998 1999 2000 2001 2002a 1/1 Shanghai-Volkswagen 160,070 200,222 230,443 235,000 230,946 221,524 230,378 248,000 2/4 FAW-Volkswagen 24,553 44,825 46,405 66,000 81,464 94,147 101,622 131,000 3/NA Shanghai-General Motors — — — — — 30,024 58,548 106,000 4/2 Tianjin Xiali (Daihatsu) 65,258 88,232 95,155 100,021 101,828 81,951 41,703 93,000 5/5 FAW-Audi-Hongqi 19,350 15,000 15,731 31,225 52,667 78,000 6/9 Shenlong (Citroe¨ n) 3,797 9,228 30,035 36,240 40,200 53,900 52,850 68,000 7/6 Chang’an (Suzuki) 17,770 16,420 35,160 36,239 44,583 48,235 50,573 64,000 8/NA Guangzhou-Honda — — — 2,246 10,008 32,228 51,153 60,000 9/NA Shanghai-Qirui — — — — — 2,767 30,085 47,000 10/NA Geely Group — — — — — 14,594 21,702 38,000 11/NA Dongfeng Fengshen — — — — — 3,159 8,000 32,000 12/NA Haima (Nainan-Mazda) — — — — — 3,059 7,800 20,000 13/NA Yuedo-Kia — — — — — 2,423 6,210 16,000 14/NA Qinchuan — — — — — 5,380 5,686 16,000 15/NA Nanya — — — — — 1,000 8,000 13,500 90 16/3 Beijing (Jeep) 25,127 26,051 19,377 8,344 9,294 4,867 4,663 4,400 17/7 Guizhou Yunque (Subaru) 7,105 798 1,000 — — 859 1,253 2,100 18/NA Tianjin-Toyota — — — — — — — 2,000 NA/8 Guangzhou-Peugeot 6,698 2,416 1,557 — — — — — Other 22,570 — 22,479 8,013 31,312 17,930 1,900 Total 352,298 388,192 481,611 507,103 565,366 649,272 732,883 1,040,900 Number of plants 3 4 5 5 5 8 9 11 producing > 25,000 cars Number of plants 2 2 2 3 3 4 7 8 producing > 50,000 cars Number of plants 1 1 1 2 2 1 2 3 producing > 100,000 cars NA, not applicable. —, Not available or plant did not exist. a Values are based on company projections. Source: Bessum 2002; Chinese Motor Vehicle Documentation Center 2002. Francois and Spinanger 91 T A B L E 4 . China’s Motor Vehicle Industry before World Trade Organization Accession, 1997 (millions US$) Sector Amount Imported motor vehicles and parts, world prices 3,607.7 Imported motor vehicles and parts, internal prices 4,849.3 Imported parts 3,239.5 Imported motor vehicles 1,609.9 Domestic motor vehicles, intermediates, and parts 32,812.5 Intermediates and parts 10,896.2 Industry consumption of motor vehicles 21,625.5 Final consumption of motor vehicles 290.8 Source: McDougall 2001. T A B L E 5 . Tariffs on Motor Vehicles in China (percent) Item 1997 rate Final rate Finished motor vehicles 71 25 Motor vehicle parts 23 10 Electronic parts 12 10 Average motor vehicles and parts 35 15 Source: China WTO accession schedule, GTAP data, and Office of the U.S. Trade Representative. relatively low, averaging perhaps 3 percent during 1995–2002. Officially, only 20,000 sedans were imported, though many more were likely smuggled into the country.4 Official policy encouraged the use of domestic parts and favored locally (regionally) produced parts. Domestic content rules applied to new investments, stipulating 80 percent domestic content by the third year. The effects of these policies are reflected in the low share of imported automotive parts imports in total production. Even after China’s completion of WTO accession, foreign ownership will be limited to 50 percent.5 Tariff rates are scheduled to come down substantially as part of the WTO accession process: by 25 percent on vehicles and 10 percent on parts on a most favored nation basis. Quotas will be phased out by 2006 and will be reduced by 15 percent a year until then. Domestic content requirements have already been removed. (Both of these nontariff barriers violate basic WTO rules.) Other WTO obligations imply free movement of imported automobiles (free of import quotas) 4. Unofficial estimates (based on interviews) are that 100,000 or more sedans have been imported in recent years. Many smuggled cars are luxury models. 5. In the past, finding partners often meant having to go to provinces other than those on the coast. These provinces often tried to ensure that ‘‘buy local’’ conditions prevailed. In the case of taxis in Shanghai, regulations stipulated specifications that could be filled only by a Volkswagen model. 92 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1 within the Chinese market. The internal barriers to trade simply cannot be sus- tained if China’s new WTO obligations are to be taken seriously. All these changes in the structure of protection imply tremendous pressure for a breakdown of internal barriers for domestic production and for rationalization of the domestic industry. The government realizes the implications for the structure of the automotive sector. Official and industry sources indicate an intention to support only a small number of domestic production groups, perhaps including the Shanghai group (Volkswagen), China First Auto Works (Volkswagen), Shanghai GM (Buick), and the Dongfeng Group (Citroe ¨ n). These groups with their foreign partners already account for more than 70 percent of production in China. Such a sharp rationalization will undoubtedly be painful, but it could allow the industry to consolidate production and work its way down the average cost curve for vehicle production. II. THE MODELING FRAMEWORK A CGEmodel is used to assess the possible impact of China’s accession to the WTO. (More technical details and references for the model are provided in Francois and Spinanger 2001 and in the technical annex available for download with the model files.6) CGE models have become a standard approach for analysis of multisector policy initiatives, such as the accompanying WTO acces- sion (Francois 2000). Although the exercises are hampered by both the neces- sary assumptions and the quality of available data, their estimates of the direct and indirect impact of broad policy changes have proved helpful for assessing existing economic policies and formulating new ones. The Model Data The data come from a number of sources. They are organized into 23 sectors and 25 regions (table 6). Details on the value-added chain linking fibers into textiles and clothing production are included to better capture the initial impact on the base scenario of the Agreement on Textiles and Clothing (ATC), which is scheduled to phase out the remaining textile and clothing quotas by 2005. Data on production and trade are based on national accounting data linked through trade flows and drawn directly from the Global Trade Analysis Project (GTAP) version 5 dataset (McDougall 2001). The dataset is benchmarked to 1997 and includes detailed data on national input-output, trade, and final demand structures. The basic database was updated to better reflect actual import protection for goods and services. Basic data on current tariff rates come from UN Conference on Trade and Development and WTO data on the schedules of applied and bound tariff rates. 6. The model files, along with the technical annex describing the model, can be downloaded from www.intereconomics.com/francois. The model is implemented in GEMPACK. Francois and Spinanger 93 T A B L E 6 . The Regional and Sectoral Breakdown of the CGE Model Region Sector Primary Hong Kong, China Wool People’s Republic of China Natural fibers Taiwan, China Primary food production Japan Other primary production Korea, Rep. of Sugar ASEAN5a Processed food, tobacco, and beverages Vietnam Manufacturing India Textiles Bangladesh Wearing apparel Other South Asian economiesb Leather products Australia Chemicals, refinery products, rubber, plastics New Zealand Steel refinery products Canada Nonferrous metal products United States Motor vehicles and parts Mexico Electronic machinery and equipment Brazil Other machinery and equipment MERCOSURc Other manufactured goods Caribbean Basin Initiative economiesd Services Andean Trade Pact economiesd Wholesale and retail trade services Chiled Transportation services (land, water, air) Other Latin Americad Communications services European Union, 15 economies Construction Turkey Finance, insurance, and real estate services Africa and the Middle East Other commercial services Rest of world Other services a Indonesia, Malaysia, Philippines, Singapore, and Thailand. b Nepal, Pakistan, and Sri Lanka. c Includes Argentina, Paraguay, and Uruguay. Brazil is represented separately. d Not treated in tables and diagrams. Source: Database aggregation produced by authors. These are integrated into the core GTAP database. They are supplemented with data from the Office of the U.S. Trade Representative and the U.S. International Trade Commission on regional preference schemes in the Western Hemisphere. Data on agricultural protection, as integrated into the GTAP core database, are based on estimates by the Organisation for Economic Co-operation and Devel- opment and U.S. Department of Agriculture. Estimates on tariffs and nontariff barriers are further adjusted to reflect remaining Uruguay Round commitments, including the phase-out of textile and clothing quotas under the ATC. Data on post–Uruguay Round tariffs are from recent estimates reported by Francois and Strutt (1999), which come primarily from the WTO’s integrated database, with supplemental information from the World Bank’s recent assessment of detailed pre– and post–Uruguay Round tariff schedules. All this tariff information has been matched to the current model sectors. Services trade barriers are based on 94 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1 the estimates described in the technical annex and are shown in table 7 (the basic GTAP database includes no information on trade barriers for services, for example). The basic GTAP dataset is benchmarked to 1997 and reflects applied tariffs in place in 1997. Because the interest here is with the post–Uruguay Round world, a ‘‘pre-experiment’’ was run on the model to implement the remaining Uruguay Round tariff cuts. Most of these cuts were already in place in the 1997 benchmark T A B L E 7 . China’s Tariff Rates before and after World Trade Organization Accession, as Modeled (percent) Sector Model base rates Accession rates New bound rates Merchandise Wool 14.8 42.0 38.0 Natural fibers 3.1 17.4 13.6 Primary food production 58.8 58.1 46.8 Other primary production 0.5 6.9 5.0 Sugar 29.5 30.0 20.0 Processed food, tobacco, 37.7 40.7 23.2 and beverages Textiles 25.1 25.4 10.2 Wearing apparel 31.8 32.8 16.1 Leather products 12.1 20.9 17.0 Chemicals, refinery products, 12.6 14.9 7.2 rubber, plastics Steel refinery products 9.7 8.9 5.1 Nonferrous metal products 7.8 8.2 5.5 Motor vehicles and parts 34.4 38.7 15.4 Motor vehicles 70.5 70.5 25.0 Parts 23.4 23.4 10.0 Electronic machinery and equipment 11.9 16.9 9.6 Other machinery and equipment 12.8 15.4 10.1 Other manufactured goods 14.5 22.0 16.3 Services Wholesale and retail trade services 0.0 NA 0.0 Transportation services 4.0 NA 2.0 (land, water, air) Communications services 9.2 NA 4.6 Construction 13.7 NA 6.8 Finance, insurance, and real 8.1 NA 4.0 estate services Other commercial services 48.0 NA 24.0 Other services 25.7 NA 13.0 NA, not applicable. Note: Service barriers are based on gravity equation estimates. Accession rates reflect an assumed 50 percent drop in cross-border trading costs. Source: China WTO accession schedule, GTAP data, and Office of the U.S. Trade Representative. Gravity estimates are based on trade and macroeconomic data and cross-country regressions; see Francois and Spinanger 2001. Francois and Spinanger 95 dataset. The data were also adjusted to reflect regional preference schemes in Latin America (not represented in the core GTAP database). The dataset used for actual policy experiments is therefore a representation of a notional world economy (with values in 1997 dollars) with full Uruguay Round tariff cut implementation. Experiments consider both the ATC phase-out and China’s WTO accession, with reference to this post–Uruguay Round tariff benchmark. Model Structure Except for the automobile sector, the CGE model structure is standard. On the production side, firms in all sectors minimize costs, employing domestic factors of production (capital, labor, and land) and intermediate inputs from domestic and foreign sources to produce goods and services. Technologies are modeled as constant elasticity of substitution processes defined over primary inputs and Leontief processes defined over intermediate inputs. Products from different regions are assumed to be imperfect substitutes in accordance with the Armington assumption. Prices on goods and factors adjust until all markets are simul- taneously in (general) equilibrium—all markets clear. Although changes are modeled in gross trade flows, changes in net international capital flows are not (this does not preclude changes in the level of gross capital flows). Trade liberalization in the goods sectors involves reduction of tariffs and a shift from model base rates to the new bound rates. The new bound rates are generally quite close to the calculations of average accession rates. Liberalization in the service sector is modeled as a reduction in trading costs, reflecting the barrier reductions in barriers reported in table 7. These are Samuelson iceberg costs. To reflect the status quo in the motor vehicle sector in a stylized, though representative way, one option was to implement imperfect competition in the model. This was rejected, however, because it does not adequately reflect the primary issue at hand. Government policy has certainly resulted in market segmentation, but there is also price setting and regulation. The choice was made to focus on realized cost efficiency for the sector. The cost structure of the industry reflects the net effect of a basket of policies. Like clothing in India or automobiles in Mexico before the North American Free Trade Agreement, the structure of the automobile sector in China reflects regulated efficiency—the impact of the general regulatory and administrative environment. The critical issue is thus these collective inefficiencies, which follow from the full set of industrial policies. At the same time, an implication of intended public policy seems to be restructuring and consolidation, leading to an improvement in regulated efficiency. What shape will the gains from changes in regulated efficiency take? Through rationalization, the industry may collectively move down relevant cost curves. Although minimum efficient scale for some models is about 200,000 units per plant (Huang 2002), a global norm, based on comparisons with plant sizes in North America and Europe, is closer to 350,000 units. Further comparison of current plant scale in China (see table 3) with such a norm implies that average 96 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1 costs are roughly 20 percent higher simply because of inefficient scale.7 Data from interviews with industry representatives (Feenstra and others 2001) point to similar cost savings, with expectations of even higher cost reductions in the range of 25–30 percent. A World Bank study (1993, p. 57) describes quite succinctly the expected gains from reaching minimum efficient scale (MES): ‘‘If this cost-volume relationship is applied to the Chinese automotive industry, the passenger car segment has a cost disadvantage of 20 to 30 percent compared with the international producers having MES. This cost disadvantage could be an understatement, however, as there are already eight producers in the market.’’ This net cost effect is stressed here and sets the treatment of motor vehicles apart from that of other sectors in the model. The lower bound of these cost effect estimates is used. In particular, the focus is on potential cost savings in the final assembly of automobiles due to a higher regulated efficiency level for the industry, resulting from consolidation and rationalization of policy.8 In addition, the differ- ential treatment of parts and finished vehicles in the tariff schedule is also tracked. That large gains can be achieved by rationalizing production and reducing costs was clearly demonstrated in the United States at the beginning of the twentieth century (figure 2). In 1914, ‘‘13,000 workers at Ford were producing 260,720 cars. By comparison, in the rest of the industry, it took 66,350 workers to make 286,770’’ cars.9 Such dichotomies also exist across the spectrum of production possibilities in China today, with new foreign-built modern plants coexisting with Mao-era facilities. Similar demand factors also prevailed. As a result of Ford’s new production methods, cars in the United States moved from being scarce goods to goods affordable by large segments of the population. China is already moving into this phase. The similarity between China’s motor vehicle production from 1984 to 2002 and that of the United States between 1900 and 1924, as shown in figure 2, would seem to justify such an analogy. III. EXPERIMENTS AND RESULTS The experiments involve full accession for China and Taiwan, China. The basic accession package involves the changes in tariffs detailed in table 7. For auto- mobiles the following effects are modeled: . Tariffs on motor vehicles decrease to 25 percent. . Tariffs on automobile parts are phased down from an average of 23.4 percent to an average of 10 percent. 7. The 20 percent figure is based on the distribution of current plants shown in table 3. An average cost index for the industry can be calculated by applying the formula Dln(Average Cost) = CDR Dln(Quantity), where CDR is the inverse elasticity of scale, defined as CDR = (Average Cost À . Marginal Cost)/Average Cost, and is between 0.125 and 0.135 (the range of values found in engineering studies). If the index is 100 at 350,000 units per plant, current plant structure yields a cost index of roughly 120. 8. In other words, cost savings are modeled at the assembly level. 9. See www.wiley.com/products/subject/business/forbes/ford.html. Francois and Spinanger 97 F I G U R E 2 . Annual Motor Vehicle Production in China and the United States (thousands) 4,500 3,500 4,000 3,000 3,500 2,500 3,000 2,500 2,000 2,000 1,500 1,500 1,000 1,000 500 500 0 0 1901 1905 1909 1913 1917 1921 1925 1984 1988 1992 1996 2000 United States 1901-1925 China 1984-2002 Source: Bessum (2002) and Verband der Automobilindustrie (various issues). . The industry is rationalized. Implicitly, this involves elimination of internal regional barriers and consolidation within the domestic market. Small, inefficient factories close. To quantify this effect, sedan production is taken as representative. Given the typical scale of domestic production, automobile plants are assumed to realize a 20 percent cost savings in assembly if they move to efficient scale. (See note 2 and the discussion in section II). This savings is modeled at the assembly level. The overall sectoral impacts of the experiments are presented in table 8, which reports changes in the quantity of output under alternative scenarios. Extending the ATC phase-out to China and Taiwan implies a dramatic expansion in the textile and clothing sectors, with textiles growing 14 percent and clothing 50 percent. There are important general equilibrium effects, as the resources needed for this experiment are drawn from other parts of the economy, includ- ing the motor vehicle sector. Especially important for the motor vehicle sector are the results reflecting the incremental impact of China’s market access commitments made as part of WTO accession and shown in columns B and C of table 8. Column B is a business as usual scenario, without restructuring. It reflects a domestic motor vehicle 98 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1 T A B L E 8 . Impact on Output of World Trade Organization Accession by China and Taiwan, China (percentage change) Accession Total impact Total Elimination without Accession without impact with of textile automobile with automobile automobile automobile and apparel sector sector sector sector quotas restructuring restructuring restructuring restructuring Sector A B C D=A+B E=A+C Wool 12.8 18.3 16.8 33.4 31.8 Other natural fibers 12.1 17.9 16.4 32.1 30.5 Primary food À0.4 À1.0 À0.9 À1.5 À1.3 Other primary À2.6 À3.6 À3.3 À6.1 À5.8 production Sugar À2.3 À7.9 À8.5 À10.0 À10.6 Processed foods À1.0 À4.7 À4.7 À5.6 À5.7 Textiles 13.9 32.0 30.6 50.4 48.8 Clothing 50.3 75.5 73.0 163.7 160.0 Leather goods À7.2 5.4 3.5 À2.2 À3.9 Chemicals, rubber, À2.0 À4.5 À4.3 À6.5 À6.2 and refineries Primary steel À4.0 À9.1 À7.9 À12.8 À11.5 Primary nonferrous À5.4 À9.2 À8.9 À14.2 À13.9 metals Motor vehicles À4.1 À36.7 8.0 À39.3 3.5 and parts Electronics À5.1 À3.9 À4.4 À8.8 À9.3 Other machinery À3.8 À5.4 À4.8 À9.0 À8.5 and equipment Other manufactures À2.2 À0.3 0.1 À2.5 À2.0 Wholesale and À0.3 1.4 1.9 1.1 1.7 retail trade Transport services À1.9 À2.0 À1.4 À3.9 À3.3 Communications À0.5 0.1 1.0 À0.5 0.5 Construction 0.8 2.8 4.2 3.6 5.0 Finance, insurance, À0.7 À0.4 0.2 À1.1 À0.4 and real estate Commercial services À0.8 À5.9 À5.4 À6.6 À6.2 Other services 0.0 0.5 1.2 0.5 1.2 Source: Model estimates; see table 7. industry that continues to be fragmented, with favored producers in each region, small production runs, and high costs. Such an industry is simply unable to compete with imports. It is hit very hard by imports, with domestic production falling 37 percent. Combined with the initial impact of the ATC phase-out, there is a dramatic retrenchment of the uncompetitive domestic industry in the face of imports (column D). By contrast, the scenario with elimination of internal barriers, rationalization of plants (with smaller plants being closed), and an efficiency gain of roughly Francois and Spinanger 99 20 percent as scale economies are realized, production rises slightly (3 percent) and the industry emerges as relatively competitive, despite the loss of protection (columns C and E). The most striking difference between the two scenarios is in the impacts on intermediate parts production and final automobile production (table 9). Under the scenario of business as usual, imports of parts rise slightly, whereas their share of the domestic market rises substantially. There is a dramatic surge in imports of motor vehicles, which displace more than a third of domestic production. There is a drop in the overall market for parts because of the decline in domestic vehicle production. Under the second scenario of rational- ization of the final assembly sector, which allows the sector to compete more directly with imports, there is a shift to imported intermediates (rising to a market share of more than 50 percent), a fall in domestic parts production (displaced by imports), and steady overall demand for parts. Although ground is lost to parts imports, sales of domestic vehicles remain steady in the face of imports. China’s WTO accession also affects value added and trade. It is logical to expect some export response, both because of the general liberalization in trade T A B L E 9 . Impact of World Trade Organization Accession by China and Taiwan, China, on China’s Motor Vehicle Market Accession without Accession with automobile automobile Benchmark sector sector Item 1997 restructuring restructuring Value (millions of 1997 U.S. dollars) Imported motor vehicles 3,607.7 10,595.7 6,968.0 and parts, world prices Imported motor vehicles 4,806.4 12,080.7 7,995.7 and parts, internal prices Imported parts 1,609.9 2,827.9 5,535.2 Imported motor vehicles 3,196.5 9,252.8 2,460.5 Domestic automobiles, 32,812.5 19,401.9 24,249.6 intermediates, and parts Intermediates and parts 10,896.2 4,494.0 5,189.1 Industry consumption of motor 21,625.5 14,698.8 18,785.0 vehicles Final consumption of motor vehicles 290.8 209.2 275.4 Index and share Import share of total 12.9 38.6 51.6 automobile parts (percent of value) Index of vehicle production 100.00 68.0 102.8 Index of parts production 100.00 41.2 56.3 Source: McDougall 2001 (baseline) and authors’ model estimates (impact). 100 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1 T A B L E 1 0 . China’s Export Shares, Baseline and Two Scenarios Total impact Total impact without with automobile automobile Baseline sector sector Sector 1997 restructuring restructuring Primary 0.046 0.033 0.033 Textiles 0.084 0.098 0.097 Wearing apparel 0.102 0.303 0.298 Motor vehicles and parts 0.006 0.004 0.019 Electronic machinery and equipment 0.133 0.100 0.099 Other machinery and equipment 0.146 0.104 0.103 Other manufactured goods 0.397 0.294 0.290 Services 0.087 0.062 0.062 Source: McDougall 2001 (baseline) and authors’ model estimates (impact). and because pressure from imports may force firms to seek other markets. China exports less that 4 percent ($1.3 billion of production of $32 billion) of its production in the sector based on 1997 values. To put this in perspective, Australia has a comparable level of exports with an industry only one-third the size of China’s. The Republic of Korea’s export share is 10 times as large. China’s trade is therefore well below global integration standards, measured by exports. The model experiments show that restructuring accelerates the export orien- tation of the automobile industry, with a rapid growth in exports (table 10). Exports rise by roughly $3.8 billion, or 300 percent, reaching roughly 10 percent of production by value. Although this seems dramatic, it needs to be kept in perspective. Automobiles and parts are a small share of exports (0.6 percent in 1997) and remain small (up to 2 percent) even with the growth in automobile exports. Most of the restructuring remains focused on the domestic market. IV. SUMMARY AND CONCLUSIONS Regulatory reform and internal restructuring are critical to the impact of WTO accession on China’s motor vehicle sector. Such restructuring is represented here by a cost reduction following from consolidation and rationalization. This representation is supported by a comparison of scale in a typical automobile plant in China to that in typical plants in North America or Europe and also by firm survey responses. It is also supported by earlier estimates of the benefits from achieving minimum and efficient scale and radical restructuring to improve production efficiency. The net result is a movement of costs toward global norms. With restructuring, the final assembly industry can become competitive by world standards, while the parts industry further integrates Francois and Spinanger 101 with the global industry through exports (and through higher imports). Without such restructuring, however, the domestic industry remains uncompetitive, and WTO accession means that imports of final vehicles will surge, though imports of parts will fall as production moves offshore. Viewed in total, what do the results show? They highlight the importance of incorporating the impact of regulatory regimes on costs when assessing the impact of changes in trade policy. For China, restructuring within the domestic market results in a qualitatively different impact from tariff reductions. Without such restructuring, the industry fails to compete and contracts dramatically. However, with restructuring, the final assembly industry can be made inter- nationally competitive. In addition, the industry shifts to local assembly, with high import content for domestic vehicles. Two additional issues need to be raised. The first concerns China’s popula- tion to motor vehicles ratio, which is far higher than that in many other countries with similar income levels (see table 1). Because this reflects the impact of existing policies, a significant change in policies could shift demand closer to a normal pattern of consumption, given China’s geographic attributes. The second issue concerns further strengthening of demand for cars through better access to financing. Roughly 75 percent of U.S. and European automobile purchases are financed through loans, but only 15 percent of automobile purchases in China are financed this way. Although China’s protocol of acces- sion to the WTO stipulates that automobile finance will be liberalized, only draft legislation has been presented to date.10 To the extent that this potential demand can be tapped, the pressure on firms to be more productive and thus more competitive will be all the greater. This would be another factor helping ensure that the calculated welfare gains will come about. The shortcomings of the analysis also need to be highlighted. The model applied here is very stylized, although it widely captures the real world. Although restructuring has positive overall implications for the industry, there are bound to be adjustment costs that are not pointed to in the model. Even if value added is preserved within the sector, there will likely be a dramatic relocation of jobs toward a limited number of plants, with job losses in the other, smaller plants. The current regional scattering of final automobile pro- duction (table 11) will be replaced by a more geographically concentrated pattern. Parts production will also tend to concentrate. To the extent that parts suppliers are able to supply regional markets, this is likely to mean that existing clustering in coastal regions will intensify, with parts shipments to Japan, the Republic of Korea, the United States, and other regional centers of 10. Nonetheless, some major car companies (Volkswagen and Ford) did reach agreements with Chinese banks earlier this year (KPMG 2003, p. 7). According to the International Herald Tribune (October 6, 2003) China has opened up this sector in line with its WTO commitments. 102 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1 T A B L E 1 1 . Location of Automobile Production in China, 2002 Foreign production Producer Foreign producer Capacity (cars/year) Production 1 SAIC VW Volkswagen 450,000 278,890 2 SAIC GM General Motors 100,000 111,623 3 FAW VW Volkswagen 270,000 158,654 4 FAW Toyota Toyota/Mazda 70,000 30,165 5 Dongfeng PSA ¨n PSA/Citroe 150,000 84,378 6 Dongfeng Honda Honda 60,000 59,024 7 Dongfeng Yulong Nissan/Yulong 60,000 38,897 8 Tianjing Toyota Toyota 30,000 2,147 9 JIangsu Nanya Fiat 100,000 23,393 10 SAIC Chery Daewoo 60,000 49,397 11 Zehjiang Jili Daewoo (geplant) 150,000 47,443 12 Chongqing Chang’an Suzuki Suzuki/Yanjin 150,000 67,846 13 Chang’an Ford Ford 50,000 NA 14 Dengfeng Yueda Kia Kia 50,000 20,080 15 FAW Hainan Mazda 50,000 11,989 16 Beijing Hyundai Hyundai 30,000 1,356 17 China Guizhou Aviation Ind. Wanhong/Chenchang 10,000 1,831 18 Shenyang Brilliant Junbei BMW (by mid-2003) 200,000 NA 19 Harbin Hafei Mitsubishi 30,000 14,577 20 Shangdong Yantei General Motors 50,000 NA 21 Southeast Zhonghua 60,000 16,935 22 Beijing Jeep Daimler-Chrysler 85,000 9,052 23 Jinbei General Motors General Motors 30,000 3,751 24 Hunan Changfeng Mitsubishi 30,000 15,067 25 Zhengzhou Nissan Nissan 30,000 NA 26 Rongcheng Huatai Hyundai 20,000 NA 27 Jiangxi Fuqi Golden Lion 20,000 NA 28 Tianjing Huali Golden Lion 20,000 NA 29 SAIC GM Wuling General Motors 150,000 NA 30 Sanjiang Renault Renault 30,000 NA 31 Chengdu FAW Toyota 5,000 NA 32 Yizhong SAIC/RDS 10,000 NA (Continued) production.11 From an employment perspective, output and value added results closely track the impact on employment. The results point to a range of effects on employment, from À40 percent without restructuring to À3 percent with restructuring. This range of effects highlights the importance of rationalizing the structure of plants. 11. European manufacturers have already established 12 plants in China, and one large U.S. company (Delphi) is shifting from Mexico. Francois and Spinanger 103 T A B L E 1 1 . Continued Production capacity in provinces Province Capacity (cars/year) Production Anhui 60,000 49,397 Bejing 115,000 10,408 Fujian 80,000 16,935 Guandong 120,000 97,921 Guangxi Zhuang 150,000 NA Guizhou 10,000 1,831 Hainan 50,000 11,989 Heilongjang 30,000 14,577 Henan 30,000 NA Hubei 180,000 84,378 Jiangsu 130,000 38,460 Jilin 340,000 188,819 Liaoming 230,000 3,751 Shandong 80,000 NA Shanghai 550,000 390,513 Shanxi 50,000 20,080 Sichuan 205,000 67,846 Tianjing 50,000 2,147 Zehjiang 150,000 47,443 Total 2,380,000 1,046,495 Other foreign companies Number of employees Number of plants Bosch 3,600 6 Kolbenschmidt 1,500 2 Michelin 4,000 2 ZF/Sachs 2,100 2 Total 11,200 12 NA, actual production was not yet available. Source: Bessum 2002; Chinese Motor Vehicle Documentation Center 2002. REFERENCES Bessum, F. 2002. ‘‘Global Car Production Statistics.’’ Online document available at www.geocities. com/motorcity/speedway/4939/carprod.html. China Online. 2000. ‘‘How WTO Membership Could Affect China’s Auto Industry.’’ January 5, 2000. Available at www.chinaonline.com. Chinese Motor Vehicle Documentation Center. 2002. Catalogue of the Present Chinese Motor Car Production, 2nd ed. Aldeboarn, Netherlands. Feenstra, R., D. Sperling, L. Branstetter, E. Harwitt, and W. Hai. 2001. ‘‘China’s Entry to the WTO: A View from the Auto Industry.’’ Working Paper, University of California, Davis.. Francois, J. F. 2000. ‘‘Assessing the Results of General Equilibrium Studies of Multilateral Trade Negotiations.’’ Policy Issues in International Trade and Commodities Study 3. UN Conference on Trade and Development. Available online at http://192.91.247.38/tab/pubs/itcdtab4_en.pdf. 104 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1 Francois, J. F., and D. Spinanger. 2001. ‘‘Greater China’s Accession to the WTO: Implications for International Trade/Production and for Hong Kong.’’ Paper prepared for the Hong Kong Trade Development Council. Francois, J. F., and A. Strutt. 1999. ‘‘Post Uruguay Round Tariff Vectors for GTAP Version 4.’’ Working Paper, Erasmus University, Rotterdam, Netherlands. Huang, Y. 2002. ‘‘Between Two Coordination Series: Automotive Industrial Policy in China with a Comparison to Korea.’’ Review of International Political Economy 9 (3):538–73. KPMG.2003. ‘‘China Automotive and Component Parts Market.’’ Online document available at www.kpmg.com.cn/Press%20releases/KPMG_China_auto_industry_report03.pdf. Lopez de Silanes, F., J. Markusen, and T. F. Rutherford. 1994. ‘‘Complementarity and Increasing Returns in Intermediate Inputs.’’ Journal of Development Economics 45:133–51. McDougall, R., ed. 2001. ‘‘The GTAP Database Version 5.’’ Working Paper, Purdue University, Global Trade Analysis Center, West Lafayette, Ind. Verband der Automobilindustrie. Various issues. Tatsachen und Zahlen. Frankfurt, Germany. World Bank. 1993. ‘‘China Industrial Organization and Efficiency Case Study: The Automotive Sector.’’ 12134-CHA. Washington, D.C. World Bank. Various years. World Development Indicators. Washington, D.C. Zhang, W., and R. Taylor. 2001. ‘‘EU Technology Transfer to China: The Automobile Industry as a Case Study.’’ Journal of Asia Pacific Economy 6(2):261–74.