Spatial agglomeration, FDI, and regional growth in China: Locality of local and foreign manufacturing investments

Journal of Asian Economics 17 (2006) 691–713

Spatial agglomeration, FDI, and regional growth in China: Locality of local and foreign manufacturing investments Linda Fung-Yee Ng 1,*, Chyau Tuan 1 Faculty of Business Administration, The Chinese University of Hong Kong, New Territory, Shatin, Hong Kong Received 16 December 2005; received in revised form 1 June 2006; accepted 20 June 2006

Abstract While China remains as the largest foreign direct investment (FDI) host country among all developing nations, the nearby Hong Kong and Taiwan are the two dominating sources of manufacturing FDI among all FDI origins. FDI absorption in the case of China could be well explained by Krugman [Krugman, P. (1991a). Increasing returns and economic geography. Journal of Political Economy, 99, 483–499] core (Hong Kong)–periphery (Guangdong, China) system and the critical role of spatial agglomeration in directing foreign investments. This paper aims at understanding the ‘‘spatial dimension’’ of firm concentration and its economic interactions with growth in China as well as how firm locality is related to institutional factors, such as regional policy on FDI, investment source, gravity of the core city, and natural geographic aspects. The effects of spatial agglomeration and FDI on regional output growth and their structural relations with endogenous FDI are addressed. Using a micro-(firm) level data consisting of 55,348 local Chinese and foreign manufacturing firms investing in Guangdong, China, research results show that other than institutional forces, spatial agglomeration and their synergies as well as gravity have directed the patterns of inward FDI and further, induced regional GDP growth. # 2006 Elsevier Inc. All rights reserved. Keywords: FDI; Spatial agglomeration; Firm locality; Institutional factors; Economic growth JEL classification: F21; O19; R12; R11

* Corresponding author. E-mail address: [email protected] (L.F.-Y. Ng). 1 Research fellows at Shanghai Academy of Social Sciences, Shanghai, PR China. 1049-0078/$ – see front matter # 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.asieco.2006.06.008

692

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

1. Background After China’s accession to WTO, China remains hosting the largest share of foreign direct investment (FDI) receipts and has also become the top FDI destination among all developing countries. Among various sources of FDI flowing into China, the two nearby overseas Asian Chinese economies, Hong Kong and Taiwan, have represented the two largest FDI sources to be followed by USA and other OECD countries. After China’s economic reform in 1979 and her implementations of the preferential FDI policies in Guangdong, other coastal regions in China were further opened up stepwise to receive FDI via carefully designed FDI promotion policies.1 Guangdong which contains three of the four Special Economic Zones (SEZs) and the Pearl River Delta region (PRD) has been able to attract ample inward FDI from other Asian economies, especially from the nearby Hong Kong and Taiwan. It remained as the largest FDI recipient sharing at least one-third of the national total FDI as of 2000 and more than a quarter (25.7%) of the total in 2003.2 Both PRD in Guangdong in southern China and Yangtze River Delta region (YRD) in eastern China represented only 1.4% (or 4% at the provincial level) of the total area in China were, respectively, recorded to account for, 26.27 and 26.05%, on an average, of the total national FDI receipts, or in sum, more than one-half of the national total FDI during the period of 1985–2003. The continuous huge influx of manufacturing FDI during the past two decades has made PRD in Guangdong distinctively known as the ‘‘world’s largest manufacturing base’’. The dominance of the inward FDI in Guangdong by Hong Kong and Taiwan, despite of the vigorous competition arising from other regions in China and with the rapid rise of YRD, in particular,3 posted a significant question: ‘‘why has this region (PRD) in China, other than being FDI preferential policy-led, been able to attract and mobilize continuously the largest share of inward FDI and particularly from these two overseas Chinese economies?’’ This paper aims at investigating the critical roles of spatial agglomeration that have directed the FDI patterns in China with particular reference to manufacturing in Guangdong and based on Krugman’s (1991a) view of core–periphery (CP) system. Until recently, very limited efforts have been devoted to study how spatial agglomeration, FDI, and regional output growth are related and how these effects would interact. In this paper, theoretical perspectives regarding the critical roles of spatial agglomeration (that is, the ‘‘critical masses’’ phenomenon) in terms of firms’ physical 1 Guangdong province, located at the southern coast of China and immediately north of Hong Kong, was first designed as a showroom to receive FDI via the establishments of Special Economics Zones (SEZs) in 1979. Subsequent steady and rapid FDI growth were induced via the economic opening of its Pearl River Delta (PRD) region in 1987 and the reconfirmation of China’s commitment to economic reform by Deng Xiaoping’s speech in early 1992. A brief review of China’s FDI promotion policy and assessment of its effectiveness on FDI absorption were performed by Ng and Tuan (2001). 2 In terms of FDI receipts, Guangdong (PRD) was surpassed by the Yangtze Delta region (YRD) in 2002 both in terms of volume and growth rates. For the period of 1985–2003, the three coastal regions, PRD (Guangdong), YRD, and Bohai, respectively, shared 26, 26, and 13%, on the average, of the total realized FDI representing about two-third of the total national FDI receipts. 3 The competition between PRD in south China and YRD in east China in FDI and the critical importance of the institutional perspective on regional growth in China were investigated by Tuan and Ng (2004b). Shanghai together with other 28 cities and eight areas along the Yangtze River were, respectively, opened up in 1984 and 1992 to receive FDI (Ng & Tuan, 2001). The subsequent development of YRD (recently known as 15 + 1) to consist of Shanghai and its nearby 15 cities/counties owing to their outstanding record in FDI attraction posted YRD another vivid case, after PRD, in demonstrating the power of a core–periphery economy (Tuan and Ng, 2004b). Visual comparisons and examinations of PRD and YRD of their fundamental economic and investment environments depicted by Geographic Information System (GIS) maps using basic economic indicators at the municipality and city/county/district levels can be viewed and studied via the following website: www.jlgis.cuhk.edu.hk/business (Tuan, Ng, & Lin, 2006).

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

693

location and especially that of foreign firms and the implications of such localities on regional growth were reviewed. The research objectives of this paper examine: (1) the locality effects of manufacturing firms including firm spatial agglomerations particularly on the impacts of institutional (FDI-led policy) characteristics; (2) the role and gravity effect of the center core; and (3) the natural geographic effects in directing FDI flow patterns. Furthermore, the postulated structural relations of manufacturing spatial agglomeration in attracting international inward investment (FDI) and inducing (regional) output growth were tested. Empirical investigations using micro-(firm) level data of 55,348 local and foreign firms operating in China were performed. 2. Spatial agglomeration, inward FDI, and regional growth: theoretical considerations 2.1. New economic geography and spatial agglomeration The essence of Krugman’s (1990, 1998) ‘‘new economic geography’’ with emphases on the ‘‘spatial dimension’’ of firms’ economic interactions has focused on the economic process of firm behavior from a geographic (space) perspective. This involves firms’ physical locations related to institutional characteristics such as frictional factor (distance effect), mechanism of networking, organizations, degrees of information and knowledge sharing, etc. This conceptualization of ‘‘space’’ suggests the persistent effects of ‘‘space’’ on firms’ identities and behavior, their interaction patterns, and individual and collective performance to imply different drivers of agglomeration and their sectoral specificities (Bottazzi, Dosi, & Fagiolo, 2002). 2.1.1. Spatial agglomeration: space, locality, and agglomeration economies The early ideas of the pattern of land use in the city-suburban relation (Alonso, 1964), the forces behind location and the urban system (Isard, 1956; Henderson, 1974), and the significance of urban agglomeration economies/diseconomies in spatial development (Henderson, 1988; Richardson, 1995) have provided strong arguments for the existence of spatial concentration and diversity. The rediscovery of the ‘‘new economic geography’’ in the study of spatial economics and its revolution beginning in the 1970s has emerged as the ‘‘fourth wave of the increasing-returns revolution in economics’’ (Fujita, Krugman, & Venables, 1999). It has also given a new dimension to the study of the ‘‘spatial’’ perspective of firm locality and ‘‘agglomeration’’ due to scale economies. Scale economies is considered, according to the ‘‘new economic geography’’, as the incentives of agglomeration as well as the force to sustain concentration (Dixit & Stigitz, 1977; Fujita & Thissa, 1996; Krugman, 1991a,b, 1996). Externalities in the form of agglomeration economies are taken as the major elements in the location decision of every firm (Fujita & Thissa, 1996). Hence, the spatial concentration pattern generated is envisaged as the outcome of the trade-off of the dispersion forces. More recent work in this area has emphasized on the theoretical perspectives of the dynamism of economies/diseconomies of agglomeration, spatial agglomeration phenomena via location or spatial patterns, interactions of such economic activities (Authur, 1994; Bottazzi et al., 2002), and the complexity of industrial districts and their interactions in a dynamic system (Curzio & Fortis, 2002).4 An overview of the recent and future 4

Following the ‘‘new economic geography’’, spatial agglomeration is particularly important in describing firms’ investment behavior in that the concept of ‘‘space’’ or their choice of destinations by locality and agglomeration is to aim at exploiting the agglomeration economies such derived. An overview on the developments of the theories of economic agglomeration and a theoretical literature review on the location and agglomeration behavior and their dynamics were provided by Maggioni (2002, Chapter 3). A survey of the literature of ‘‘economic geography’’ was also performed (Lee & Willis, 1997) and referenced (Bottazzi et al., 2002; Martin, 1999).

694

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

developments of the ‘‘new economic geography’’ and agglomeration (clustering) of economic activities, agglomeration economies, location space, trade, transportation costs, and growth is provided by Fujita and Mori (2005) and Fujita and Thissa (2002). 2.1.2. Spatial agglomeration and international investment flows Krugman’s (1990, 1998) ‘‘new economic geography’’ highlights the role of agglomeration externalities in determining investment locations and FDI flow behavior through which ‘‘geographic concentration’’ is formed via the interactions of increasing returns, transportation costs, and factor mobility. As far as the process or dynamism of regional growth of the economic activities is concerned, the existence of increasing scale economies in production would tend to direct the patterns of spatial concentration and the locality of investments (Krugman, 1991a,b). The agglomeration economies derived from the core–periphery relation further act as the dominant force of specific regional investment and trade flows.5 Following the ‘‘new economic geography’’ approach, the formation of industrial clustering by means of capital mobility and investment dynamics (Baldwin, 1997), industry spillovers via manufacturing activities diffusion among regions (Puga & Venables, 1996), and the influence of industry forward and backward linkages (Venables, 1996) are stressed.6 2.2. Spatial agglomeration and regional growth The evolution of the spatial economics with emphases on the significance of market forces, spatial concentration and agglomeration economies/diseconomies, and localization helps to understand economic growth from a new perspective. Following the spirit and the new perspectives, the later development of the integration of new trade and new growth theory (Grossman & Helpman, 1991; Helpman & Krugman, 1985) is believed ‘‘to synthesize each field into a coherent whole’’ (Fujita et al., 1999). The recognition of the existence of agglomeration economies arising from economies of production and location choice has not only helped to explain the location pattern of industries, but further envisages the dynamics of industry activities leading to city and regional growth.7 The focus of the ‘‘new economic geography’’ on the spatial agglomeration of industry and the longrun convergence of regional incomes is further argued as the ‘‘reworking (or re-invention) . . . of traditional location theory and regional science’’ (Martin, 1999). A collection of papers (Giersch, 1995) has adequately presented the significance of urban agglomeration on spatial development and its impacts on economic growth. Along this line, urban and regional growth via agglomeration effects in a metropolitan economy (Black & 5

The recent development of the concept of agglomeration economies and their significance were further reviewed (Fujita & Krugman, 2004; Fujita & Thissa, 2002). The concept of Krugman’s (1991a) core–periphery system with the function of the city core and its relations to the periphery region in China was first demonstrated by using the Hong KongPRD case (Tuan & Ng, 1995). The gravity relations and effects from the city core to the periphery market center were further illustrated and tested (Ng & Tuan, 2003; Tuan & Ng, 2001b, 2003a, 2004a). 6 Curzio and Fortis (2002) presented a collection of papers to investigate the complexity of industrial districts and its dynamics from theories to practice. 7 However, a recognized disadvantage of Krugman’s view of economies of scale and production externalities in forming spatial concentration is the drawbacks of the nonexistence of congestion dynamics and diseconomies. Following this argument, the author thank an anonymous reviewer for spelling out the part of the diseconomies that ‘‘I(i)ncreasing agglomeration may well lead to increasing environment degradation . . . and to limitations on available infrastructure services such as power and water (problems common in many developing countries)’’.

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

695

Henderson, 1999) and the significant role of the agglomerative implications of city size and its diversity on city growth (Quigley, 1998) are also well investigated. Moreover, spatial agglomeration is also believed to contribute to endogenous growth via knowledge accumulation (Quah, 2002) and human capital (Baldwin, 1997; Palivos & Wang, 1996; Wolff, 1985). 2.3. Spatial agglomeration, international investment (FDI), and growth: further evidence Recently, more extensive empirical literature investigating the issue of spatial agglomeration and its relations with city/urban and regional growth has been performed. From the point of view of location decision, the main objective of any profit-maximizing firm in choosing favorable investment sites is whether the locations are capable of generating positive net benefits. Such net benefits include: (1) geographical benefits or benefits derived from the quality of the location to depict the location outcome of the process in forming clusters and (2) agglomeration benefits or benefits derived from other firms in the same location. These agglomeration benefits may be iterated as agglomeration economies, localized source of inputs and/or consumers, information cascades and reduction of location search costs, market power, and firm clustering behavior generated by a set of relevant location factors to include various types of economic and institutional effects (Maggioni, 2002). A summary of the empirical literature on the formation of agglomeration economies and firm clustering is found in Maggioni (2002, Chapter 5). Scale economies is considered as the pre-condition of agglomeration economies as well as the main centripetal force in determining firms’ location. Empirical studies following Krugman’s ideas on scale economies have provided strong supports for such effects on location decisions and hence, subsequent patterns of inward FDI.8 These studies include the investigations of spatial behavior and location choice of U.S. financial and professional services transnationals (Nachum, 1998), U.S. and Swedish FDI and multinationals (Braunerhjelm & Ekholm, 1998),9 Japanese FDI in East Asia (Kohno, Nijkamp, & Poot, 2000), worldwide location decisions made by Japanese manufacturing FDI (Blonigen, Ellis, & Fausten, 2005), and China’s manufacturing and services joint ventures (Tuan & Ng, 2001a, 2003a). Regional industry agglomerations driven by a three-tier agglomeration system would also serve as an incentive and critical driving force to direct inward FDI in China (Tuan & Ng, 2003b, 2004b). In considering the effects of agglomeration on output, strong industry production agglomeration and globalization (measured by export and FDI) were proved to cause increasing regional income disparity in coastal China (Fujita & Hu, 2001). Moreover, agglomeration economies and FDI acquisition were closely associated with market size and export propensity in the case of European integration (Girma, 2001). The relation of FDI and growth has been widely examined since the early 1990s. The impacts of FDI on growth with the role of technology transfer and market integration were emphasized 8 Relevant factors affecting firm locality contained a set of nine groups of indicators to include institutional/structural characteristics, environmental factors, and market competition, etc. (see Maggioni, 2002, Chapter 5 for details). Empirical evidence show that the causes of agglomeration and the determinants of location decision of firms in the case of U.S. manufacturing included factors such as proximity to consumers (or market), production externalities, and specialized inputs (LaFountain, 2005). An empirical study of the electronics joint ventures operating in PRD further support the significance of agglomeration economies in directing FDI flows in the form of firm clustering pattern by investment locations and in the formation of networked clusters within the core–periphery economy (Tuan & Ng, 2001b). 9 A collection of papers in Braunerhjelm and Ekholm (1998) addressed the question of location and geographical dispersion of Swedish multinationals based on empirical data. These studies illustrated that geographical specialization and proximity advantages, scale economies, agglomeration in geographical location, and strategic location of production were the major factors, among others, affecting the patterns of FDI activity.

696

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

(Barro & Sala-I-Martin, 1995, 1997; Grossman & Helpman, 1991, 1995). The significant role of FDI and its spillover effects including human skills, employment, technology transfer, and trade, to promote economic growth were demonstrated by the ASEAN-5 economies (Bende, 1999). Foreign capital inflows and imports had affected growth in four ASEAN countries (Marwah & Tavakoli, 2004). With emphases on the role of institutions in growth, FDI had also significantly benefited the growth of those countries with well-developed financial markets (lfaro, Chanda, Kalemli-Ozcan, & Sayek, 2004). Nevertheless, growth regressions performed at the macroeconomic level had not provided any support on the effects of FDI on economic growth (Borensztein, De Gregorio, & Lee, 1998). Further in the recent growth literature, a new perspective on FDI-led growth and their causal relations were investigated (Borensztein et al., 1998; De Mello, 1999; Marwah & Tavakoli, 2004). FDI is considered as a main transmission mechanism of advance technology and fosters GDP growth especially in developing countries. Co-integration and causality studies on the effects of FDI on long- and short-run growth support both uni- and bi-directional causalities between FDI and GDP growth (Bengoa & Sanchez-Robles, 2003; Basu, Chakraborty, & Reagle, 2003; Liu, Burridge, & Sinclair, 2002; Nair-Reichert & Weinhold, 2001). The positive link of FDI and growth and particularly, the causal direction from FDI to long-run growth have been well supported by these literatures. Little effort so far has been made to understand how spatial agglomeration, FDI flows, and output growth are inter-related. This is particularly true in regard to empirical investigations at the micro-(firm) level due to the limitation of demanding for a huge data set of microscopic nature. This paper attempts to empirically examine at a micro-(firm) level the above theoretical perspectives by focusing on: (1) the critical roles of spatial agglomeration of investments (both local and foreign) or ‘‘critical masses’’ phenomenon; (2) how spatial agglomeration affects FDI flowing patterns; (3) the role and effects of the gravity center in a core–periphery economy; and (4) the structural relations of spatial agglomeration with inward investment (FDI) and regional growth together with the endogeneity of FDI addressed. 3. Research methodology Bearing the above perspectives in mind, six major research hypotheses would be postulated. The impacts of institutional characteristics (including gravity effect, FDI-led policy, and natural geographic hurdle) in directing firm locality and FDI patterns, spatial agglomerations and regional FDI flows, and the impacts on regional output growth were examined and tested. 3.1. Research hypotheses Hypothesis 1. Firm locality to be reflected by the spatial distributions and concentration patterns of investments is governed by the institutional characteristics confronted by the firms. Hence, firm locality and its diversity behavior are directed by investment types by source, government preferential or regional FDI-led policy, and natural geographic characteristics in such a way that (1.1) firms incline to ‘‘stay closer together’’ to take advantage of the economies arising from agglomeration, that is, the ‘‘critical mass’’ phenomenon so as to gain scale economies. This is particularly true in the case of foreign firms which intend to exploit the advantages generated from business networking (agglomeration) and production supporting facilities so as to increase efficiencies and competitiveness;

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

697

(1.2) investment locations equipped with governmental preferential (FDI) policies aiming at FDI absorption are the favorable investment destinations; (1.3) firm locality responds positively to favorable geographic characteristics so as to meet their needs in terms of production supporting facilities as well as transaction and other production costs; (1.4) firm concentrations by investment types (regardless of local or foreign firms) are closely associated by locality in order to gain economies by ‘‘staying closer together’’ and exploit the advantages of the ‘‘critical mass’’ such that the spatial agglomerations of local and foreign firms are positively related; and (1.5) given different institutional effects including governmental FDI-policy and natural geographic features, the same is true as presented by Hypothesis 1.4. Hypothesis 2. In choosing investment sites, firms incline to select locations with higher spatial agglomeration in order to exploit the economies generated from agglomeration by simply ‘‘staying closer together’’ in both the horizontal and vertical directions. That is, spatial agglomeration by investment types would positively affect FDI absorption. Hypothesis 3. The strategic interactions of the investments between local and foreign firms would create synergies contributing to agglomeration economies and production efficiencies, which further attract FDI into the region. That is, strategic interactions of investments would positively affect FDI absorption. Hypothesis 4. Institutional forces would significantly direct FDI inflows such that the frictional (gravity) effect of the center core is critical to affecting FDI flows in a core–periphery economy.10 Hence, gravity (or frictional distance) would negatively affect FDI absorption. Hypothesis 5. Spatial agglomeration promote regional output growth in that (5.1) higher firm agglomeration (by investment types) would directly induce higher GDP growth; and (5.2) spatial agglomeration would enhance regional GDP growth via FDI absorption; that is, FDI is endogenous in the determination of regional growth. Hypothesis 6. Institutional factors such as natural geographic (hurdle) would affect firms’ location decision and hence, FDI patterns and regional GDP growth. 3.2. Measurements of spatial concentration and agglomeration 3.2.1. Diversities due to spatial concentration Suppose a population of N firms will distribute over R regions (or locations) with NR firms or establishments in R such that N = SNR where R = 1 to r. Locality or the choice of

10

According to Krugman (1991a), the two co-existed factors of economies of scale and agglomeration economies are mutually reinforcing. On the other hand, the major centrifugal force is transportation costs. The effects of the transportation costs on FDI flows and firm agglomeration were tested with reference to a core–peripheral economy (Tuan & Ng, 2004a). While transportation cost was positively affected by the geographic distance (friction) between the gravity center (city core) and the investment destinations, FDI flows were negatively affected by friction (Ng & Tuan, 2001; Tuan & Ng, 2003a, 2003b).

698

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

investment sites leading to certain density (concentration) of firms in R could simply be measured by: (1) total firm number in R (NR) and (2) taking into the geographical space of R into account, location density (concentration) in terms of number of firms (N) per area space (N/Area), 8R, that is, NR/AreaR. The spatial concentration or diversity behavior (pattern) can be measured by the simple statistics of the coefficient of variation (CV), that is, CV = s/Navg where s is the standard variation of the distribution and Navg is the mean number of firms, 8R. 3.2.2. Diversities due to spatial agglomeration: spatial agglomeration index To further reflect the extent of diversity in terms of the strength of agglomerations or cohesiveness derived by a collection of firms located in R in both the horizontal and vertical directions, the concept of Herfindahl index is adopted for the construction of the spatial agglomeration index (I) in R; that is, IR = (NR/N)2 and hence, I = SIR. Thus, the spatial agglomeration effect of firms in R is defined as the squared term of the share of R’s total firm number in the total population N. In terms of measuring the strength of firm spatial agglomeration in each location (region) R, IR being ranged from ‘‘0’’ to ‘‘1’’ will indicate, respectively, minimal to maximal agglomeration effect. Alternatively, ‘‘0’’ stands for highest diversity (that is, absolute heterogeneity or lowest cohesion) and ‘‘1’’ for least diversity (that is, absolute homogeneity or highest cohesion). Furthermore, the synergies of spatial agglomerations derived from the strategic interactions of agglomerations by firm (investment) types is simply the cross (joint) effects of IR by investment types in each R. The three types of institutional characteristics studied are: (1) investment flows by source; (2) FDI policy by policy region; and (3) natural characteristics by geographic region. 3.2.3. Data source and management To validate the above hypotheses calls for a fairly large-scale micro-(firm) level information to be identified via various types of institutional forces in association with regional characteristics. The population used for this study is the Guangdong Industrial and Commercial Enterprises Database, 1998 containing a total of 55,664 local Chinese manufacturing and manufacturing joint ventures (JVs) operating in Guangdong in 1998. The database originally consisted of firm-level information published in the form of a spreadsheet intended to provide business information for conducting businesses in Guangdong. Hence, raw data entries first by converting all the firm-level information into computer readable and computable format is necessary before all subsequent statistical computations and analyses can be performed. The relevant variables in the database used were executed and recorded as follows: 1. the location of a firm in region R (1 to 105 cities/counties/district) was identified by each firm’s self-reporting address via postal code and allocated in accordance with the 105 administrative/ municipal zones in Guangdong; 2. the gravity or frictional (distance) effect of a firm was measured by the highway distance in kilometers from the core (Hong Kong) to R; 3. investment type by source, that is, local versus foreign origin, was identified as local Chineseowned (LC) or foreign-owned (FF) by the investment classification self-reported by each firm; 4. governmental FDI policy was defined by the three FDI-led policy regions directed by the Central government: (1) Special Economic Zones (SEZs), (2) Pearl River Delta (PRD) region

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

699

equipped with FDI-led policies, and (3) non-FDI policy regions (Non-PRD). Each R was allocated to one of the three policy regions; and 5. the natural geographic region of each R was assigned as either east (PR-east) or west (PR-west) by the natural geographic partitioning of the Pearl River (PR) of Guangdong into the eastern and western sub-regions. The gravity center core (Hong Kong) is found naturally located at the east-bank. All the secondary data including land area, city/county level, and sectoral GDP, and realized FDI in volume by the 105 cities/counties in Guangdong were obtained from various issues of Guangdong Statistical Yearbook. 4. Statistical methods and estimations 4.1. Diversity and agglomeration of locality behavior: correlation analysis Hypothesis 1 state that the diversities of firm location in terms of spatial concentration (density) and agglomerations (cohesive forces) are directed by institutional characteristics (including FDI-policy and geographical characteristics) was first examined. Three measurements were used in the estimations of the diversity behavior of R: (1) total number of firms (NUM); (2) location density (NUM_A); and (3) spatial agglomeration (I). The simple statistical concept of coefficient of variation (CV) used to measure the diversity of spatial density (concentration) was also computed. Further, the associations of the spatial agglomeration effects between local Chinese (LC) and foreign firms (FF) by different institutional characteristics were examined by simple correlation (Pearson) (gx.y) analyses. 4.2. Effects of spatial agglomeration on FDI: regression analyses Following Hypotheses 2–4, FDI absorption in region R at time t is assumed to be determined by three effects, namely, (1) spatial agglomeration effects by investment type, that is, LC (ILC) versus FF (IFF); (2) synergy effect generated by the strategic interactions of the firm agglomerations between LC and FF (ILC
IFF); and (3) institutional effect of frictional distance of R from the center core (Dist). FDI absorption is measured by: (1) current FDI at time period t (FDIt); (2) future (planned) FDI at time period t + 1 (FDIt+1); and (3) FDI adjustment or its accelerated change (DFDI = FDIt+1  FDIt). With the time subscripts (on the right hand side of the equations) and R omitted for simplicity, the following equations are postulated: FDIt ¼ f ðI LC ; I FF ; I LC
I FF ; DistÞe

(1)

FDItþ1 ¼ f ðI LC ; I FF ; I LC
I FF ; DistÞe

(2)

DFDI ¼ f ðI LC ; I FF ; I LC
I FF ; DistÞe

(3)

4.3. Effects of spatial agglomeration and FDI on growth: 2SLS regression To test Hypothesis 5, regional output (GDP) growth are hypothesized to be affected by both the spatial agglomeration effects by investment types (ILC and IFF) and FDI (that is, FDI-led). To address the endogeneity property of FDI and the inter-relationship among spatial agglomerations

700

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

and institutional gravity effect with regional output growth, the structural relations can be modeled by the following three sets of simultaneous equations: ðiÞ

ðiiÞ

ðiiiÞ

FDIt ¼ f ðI LC ; I FF ; I LC
I FF ; DistÞe

(4)

GDPt ¼ f ðFDIt ; I LC ; I FF Þe

(5)

FDItþ1 ¼ f ðI LC ; I FF ; I LC
I FF ; DistÞe

(6)

GDPtþ1 ¼ f ðFDItþ1 ; I LC ; I FF Þe

(7)

DFDI ¼ f ðI LC ; I FF ; I LC
I FF ; DistÞe

(8)

DGDP ¼ f ðDFDI; I LC ; I FF Þe

(9)

Eqs. (4), (6), and (8) are simply the replications of Eqs. (1)–(3), respectively. Eqs. (4)–(9) would be estimated in log-linear form by 2SLS. The estimated partial regression coefficients are partial elasticities expected to carry positive (‘+’) signs (that is, dFDI/dI > 0; dGDP/dFDI > 0) except that of gravity (distance) effect with negative (‘’) sign (that is, dFDI/dDist < 0). 4.4. Effects of regional diversity on FDI and growth: 2SLS regression Regional diversities derived from firms’ location behavior being examined by Hypothesis 1 support that investments incline to concentrate in the FDI-policy led regions and/or avoid those regions countering natural geographic hardships. Such effects of regional diversities of investment as proposed by Hypothesis 6 would be examined by estimating again the three sets of structural equations (Eqs. (4)–(9)) by sub-regions of the east bank (PR-east) against the west (PRwest). Consistent statistical results in terms of the expected signs of the regression parameters as discussed in the above hypotheses are expected. 5. Statistical results 5.1. Spatial disparity and agglomerations by institutional characteristics: investment source, FDI-policy led, and geographic directed patterns Spatial distribution of firms and hence, firm locality is directed by institutional characteristics. Table 1 presents the behavior of firms’ spatial distributions by investment types (LC versus FF) and institutional characteristics (FDI-policy and natural geographic characteristics). Table 2 presents investment behavior in terms of spatial diversity and agglomeration by investment types and institutional characteristics. 5.1.1. Spatial distribution and disparity The simple descriptive statistics in Table 1 show high regional disparities in the spatial distributions of LC versus FF and high concentration in the FDI-policy regions (SEZs/PRD) and in PR-east. Of the total population of 55,348 firms operating in Guangdong in which 72.06% (39,882) was LC and 27.94% (15,466) was FF, almost 60% of them (59.41%; 32,878) were found located in the two FDI-policy led regions (SEZs and PRD). By geographic characteristics, PReast accounted for a total of 73.39% (35,084) and PR-west only 36.62% (20,264), making a ratio of about 2:1.

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

701

Table 1 Spatial distributions of the population of local Chinese and foreign manufacturing firms in Guangdong (1998) Total (N)

Total (%)

Investment origin (%)

Geographic region (%)

55348 39882 15466

100.00 72.06 27.94

100 100 100

100 – –

1809 3120 4929

3.27 5.64 8.91

4.54 20.17 –

36.70 63.30 100.00

PRD NUMLC NUMFF Sub-total

19946 8003 27949

36.04 14.46 50.50

50.01 51.75 –

71.37 28.63 100.00

Others NUMLC NUMFF Sub-total

18127 4343 22470

32.75 7.85 40.60

45.45 28.08 –

80.67 19.33 100.00

23839 11245 35084

43.07 30.32 73.39

59.77 72.71 –

67.95 32.05 100.00

16043 4221 20264

28.99 7.63 36.62

40.23 27.29 –

79.17 20.83 100.00

Population NUMLC NUMFF FDI-policy region SEZs NUMLC NUMFF Sub-total

Geographic region PR-east NUMLC NUMFF Sub-total PR-west NUMLC NUMFF Sub-total

Notes: NUM is the total number of firms with subscripts LC and FF to denote local Chinese and foreign firms, respectively; SEZs and PRD denote Special Economic Zones and Pearl River Delta Region (PRD), respectively.

Again from Table 1, the spatial distributions of firms by investment types (LC versus FF) show that FF were more heavily concentrated in PR-east (72.71%; 11,245) than PR-west (27.29%; 4221), while LC were less diverse (PR-east, 59.77% or 23,839 versus PR-west, 40.23% or 16,043). When considering FDI policies implemented by region, FF were found mainly concentrated in SEZs (20.17%; 3120) and PRD (51.75%; 8003), representing a total of 71.92% of FF. For LC, however, the respective figures were only 4.54% (1809) and 50.01% (19,946), about one-half (54.54%) of LC population. Such figures suggest that the majority of FF had been directed by FDI-preferential policy into SEZs/PRD while LC were somewhat indifferent. It is, therefore, reasonable to conclude that FF were more responsive to FDI-preferential policies than LC. Further, a common locality behavior by geographic characteristics (PR-east versus PR-west) observed is that the majority of the firms were found mainly located at PR-east (73.38%) rather than PR-west (36.63%) with FF exhibited a much higher proportion (43.07%) than LC (30.32%) (Table 1). 5.1.2. Spatial diversity and agglomeration Table 2 presents the statistics showing spatial diversity and agglomeration by investment types and institutional characteristics measured by total number (NUM), location density (NUM_A),

702

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

Table 2 Spatial diversity and firms’ agglomerations in Guangdong: local Chinese vs. foreign manufacturing firms (1998) Variable

Mean

S.D.

CV

823.51 543.48 328.39

1.553 1.422 2.221

1 1 0

6227 3861 2366

411.02 193.88

661.68 410.08

1.610 2.115

1 0

3861 2366

356.51 93.80

354.79 174.92

0.995 1.865

35 2

1628 924

Firm density in number (NUM) NUM 530.13 382.25 NUMLC NUMFF 147.89 Geographic region PR-east NUMLC NUMFF PR-west NUMLC NUMFF

Location density (NUM_A) NUM_A 0.305 0.215 NUM_ALC NUM_AFF 0.090

Minimum

Maximum

0.463 0.322 0.184

1.518 1.498 2.044

0.006 0 0

2.526 0.960 0.178

Geographic region PR-east NUM_A NUM_ALC NUM_AFF

0.342 0.234 0.108

0.508 0.346 0.213

1.485 1.479 1.972

0.006 0.003 0

2.526 1.591 0.960

PR-west NUM_A NUM_ALC NUM_AFF

0.258 0.191 0.067

0.400 0.292 0.140

1.550 1.529 2.090

0.015 0.009 0.001

2.178 1.741 0.567

5.366E4 2.004E4

3.797 4.829

3.221E10 0

48.022E4 18.033E4

1.173E4 4.089E4

0.010E4 3.386E4

0.082 0.969

1.110E4 1.075E4

1.293E4 8.444E4

PRD ILC IFF

2.207E4 0.363E4

4.501E4 0.873E4

2.039 2.409

3.221E10 3.221E10

23.03E4 4.802E4

Non-PRD ILC IFF

1.009E4 0.277E4

5.867E4 2.202E4

5.815 7.948

0.093E10 0

48.022E4 18.033E4

1.930E4 0.653E4

7.049E4 2.651E4

3.652 4.060

3.221E10 0

48.000E4 18.000E4

0.806E4 0.125E4

1.694E4 0.470E4

2.102 3.760

0.004E10 0.129E10

8.538E4 2.750E4

Spatial agglomeration index (I) 1.413E4 ILC IFF 0.415E4 FDI-policy region SEZs ILC IFF

Geographic region PR-east ILC IFF PR-west ILC IFF

Notes: Population (N = 55,664); number of city/county = 105; S.D. and CV represent standard deviations and coefficient of variations, respectively; the subscripts LC and FF represent local Chinese and foreign firms, respectively.

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

703

and index of spatial agglomeration (I). From Table 2, the coefficient of variation (CV) of the three measures of locality density (concentration) demonstrated the concentration and cohesion of investments. The following major features were observed. Hypothesis 1.1. FF consistently revealed a more diverse pattern of spatial distributions than LC when measured by all the three above designated concentration indicators: (a) NUM (CV = 2.22 versus 1.42, respectively); (b) NUM_A (CV = 2.04 versus 1.50, respectively); and (c) I (CV = 4.83 versus 3.80, respectively). These statistics support FF’s behavior to choose ‘‘stay closer together’’ than that of LC. Hypothesis 1.2. For FDI-policy implementations by regions, high disparities in firm locations among regions were observed. Both LC and FF would first choose SEZs and PRD as their preferred investment sites. FF consistently exhibited higher regional diversities. For FF, the CVof I in the three FDI-policy regions, namely, SEZs, PRD, and non-PRD were 0.969, 2.409, and 7.948, respectively. The respective figures for LC were much lower (0.082, 2.039, and 5.815). Hypothesis 1.3. By geographic regions (PR-east versus PR-west), both LC and FF had concentrated more at PR-east. The unique locality pattern of FF to locate nearer the city core (at PR-east) is again proven necessary and critical for their efficient production. The diversities of spatial patterns measured by all the three indicators were much lower for FF at both PR-east and PR-west. For FF, the CVs of the three respective concentration indicators at PR-east were 2.115, 1.972, and 4.060, respectively; and at PR-west were 1.865, 2.090, and 3.560, respectively. The respective CVs for LC at PR-east were 1.610, 1.479, and 3.652; and at PR-west were 0.995, 1.529, and 2.102, respectively. 5.2. Associations of spatial agglomerations of local Chinese versus foreign firms by locality: correlation analyses The relations between the spatial agglomerations by localities of LC versus FF among R investment destinations were studied by the correlation analyses between ILC and IFF, 8R. Table 3 provides the statistical results of the simple correlation (Pearson) coefficients showing the spatial distribution patterns between FF and LC (gLCFF) for the three concentration indicators. Hypothesis 1.4. The hypothesis is strongly supported by the high correlation coefficients obtained for the three measures—firm density (NUM, g = 0.77; p < 0.01), location density (NUM_A, g = 0.65; p < 0.01), and spatial agglomeration index (I, g = 0.86; p < 0.01). Such a Table 3 Associations of spatial agglomerations between local Chinese and foreign manufacturing firms by FDI-policy and geographic region: correlation analysis gLCFF

NUM NUM_A I

All population

#

0.770 0.645# 0.860#

FDI-policy region

Geographic region

PR-east

PRD

PR-east

PRD

Non-PRD #

0.531 0.424# 0.406**

#

0.923 0.882# 0.994#

#

0.814 0.634# 0.873#

PR-west #

0.553 0.675# 0.440#

PR-west Non-PRD

**

0.501 0.295 0.380*

#

0.973 0.904# 0.998#

PRD

Non-PRD **

0.557 0.611# 0.424*

0.549# 0.268 0.432**

Notes: # and ** indicate statistical significance at p < 0.01 and p < 0.05, respectively; the correlation (Pearson) coefficients on first row are based on firm density in total firm number (NUM), second row firms’ location density per land area (NUM_A), and third row spatial agglomeration index (I) between local and foreign firms.

704

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

result implies that both LC and FF chose to ‘‘stay close together’’ or form ‘‘critical masses’’ to take advantage of the economies of scale in production and other agglomeration economies such generated. Hypothesis 1.5. Given the institutional effects by FDI-policy and geographic regions, consistent localities behavior for both LC and FF as described by Hypothesis 1.4 is also true, according to the statistics presented in Table 3. (a) By FDI-policy regions, LC and FF locating at non-PRD (that is, in the absence of preferential-FDI policies) displayed high associations measured by all types of indicators (gLCFF > 0.88; p < 0.01), while the firms at PRD were modestly related (gLCFF > 0.44; p < 0.01). Such a result seemingly implies the importance of agglomeration or forming ‘‘critical mass’’ to gain efficiencies in the absence of any governmental (FDI) policy aids. (b) By natural geographic regions, LC and FF located at PR-east consistently exhibited higher associations of agglomerations (gLCFF > 0.80; p < 0.01; except DEN_A) than that of at PR-west. (c) By both FDI-led policy and geographic regions, the high correlation coefficients found in non-PRD at PR-east (gLCFF > 0.90; p < 0.01), once again, illustrate the critical importance of spatial agglomeration. It provides the means for firms to stay competitive in production and to exploit the agglomeration economies, given the highly competitive environment in resource (labor and land) costs and in the absence of governmental policy aids. However, such an effect is not as critical at PR-west where resources are still relatively abundant. Yet, the associations of spatial agglomerations between LC and FF were still relatively important in determining firm locality (gLCFF > 0.42; p < 0.01). The above correlation results suggest that the associations of firm spatial agglomeration by investment source and regional characteristics were strong and critical to production efficiencies. Similar patterns were also true given governmental policy and geographic characteristics. In general, the research results imply that both LC and FF would have to rely on the power of a collective (agglomeration) or ‘‘critical mass’’ effect, regardless of policy-aids. In this connection, special attention should be devoted to those regions without any FDI policy aids and particularly, PR-west. Enhancing higher growth by developing PR-west and the crucial role of the gravity center (Hong Kong) to attract FDI into Guangdong were illustrated (Tuan & Ng, 2003b, 2004a). 5.3. Effects of spatial agglomeration, strategic interactions, and gravity on FDI: regression analyses Hypotheses 2–4 which postulated the effects of spatial agglomeration (I) by investment types (LC versus FF), the strategic interactions of the spatial agglomerations between LC (ILC) and FF (IFF), and friction (geographic distance from the center core) on FDI flows were estimated by OLS regression (Eqs. (1)–(3)). The corresponding regression results are presented in Table 4 (Eqs. (I.1)–(I.3)). All the estimated regression relations were highly statistically significant (F-stat; p < 0.01) with satisfactory goodness-of-fit (adj-R2 > 0.32). The estimated partial regression coefficients were all statistically significant showing the correct expected signs, that is, dFDIt/dILC > 0, dFDIt/dIFF > 0, dFDIt/dILC
IFF > 0, and dFDIt/dDist < 0. The same results had applied to future (planned) FDI (FDIt+1) and FDI growth (DFDI) except that the effects of ILC and Dist were unimportant to DFDI (Table 4, Eq. (I.3)). Thus, Hypotheses 2–4 are generally supported.

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

705

Table 4 Effects of manufacturing spatial agglomerations, gravity, and FDI on GDP growth: results from 2SLS regression Variable 2SLS (Stage I) (I.1) FDIt b0 FDI-hata ILC IFF ILC
IFF Dist Adj-R 2 F-Stat RMSE DF

2SLS (Stage II) (I.2) FDIt+1

#

#

0.574# 0.653# 0.034# 1.135# 0.738 62.86# 0.979 84

(0.181) 0.554# (0.135) 0.599# (0.011) 0.032# (0.253) 1.074# 0.731 62.09# 0.925 86

15.317 (1.894) 14.538 (1.810) (0.174) (0.129) (0.011) (0.242)

(I.3) DFDI #

(II.1) GDPt

(II.2) GDPt+1

(II.3) DGDP

6.676 (0.551) 6.411 (0.609) 5.538# (1.301) 0.386# (0.091) 0.426# (0.104) 1.123** (0.544) 0.414 (0.312) 0.297# (0.036) 0.278# (0.039) 0.384** (0.179) 0.569** (0.228) 0.061 (0.050) 0.070 (0.055) 0.220 (0.238) 0.032* (0.018) 0.590 (0.497) 0.328 0.784 0.739 0.284 7.09# 107.63# 86.05# 7.46# 1.341 0.537 0.575 1.688 45 85 87 46 7.871 (3.242)

#

#

Notes: #, **, and * represent statistical significance at p < 0.01, p < 0.05, and p < 0.10, respectively; all estimations in loglinear form. a Indicates the corresponding endogenous variable appearing in Stage I of 2SLS regression.

Besides, it is also logical to find that FF (IFF) had exerted higher impacts (elasticities) on FDI inflows than that of local firms (ILC) (Table 4, Eqs. (I.1)–(I.3)). The significant positive joint (interaction) effect (ILC
IFF), though modest, further demonstrated the fact that the synergies generated from firm strategic interactions are important in determining regional FDI absorption.11 The importance of the institutional gravity (distance) effect on both current (FDIt) and planned (FDIt+1) FDI is further supported by the statistical significance of the respective partial elasticities (1.135 and 1.074; p < 0.01; Table 4, Eqs. (I.1) and (I.2)). 5.4. Spatial agglomerations, FDI, and regional growth: 2SLS regression 5.4.1. Effects of spatial agglomeration and FDI on GDP growth: 2SLS Table 4 shows the 2SLS statistical results of the three sets of simultaneous relations as postulated by Hypothesis 5 (Eqs. (4)–(9)). Other than the statistically significant agglomeration and distance effects (Stage I of 2SLS; Eqs. (I.1)–(I.3)) being found earlier, the estimation results from Stage II (2SLS) show that the endogenous FDI had significantly affected regional GDP and its growth (Table 4, Eqs. (II.1)–(II.3)). The corresponding partial elasticities on GDP measured by GDPt, GDPt+1, and DGDP due to FDI were 0.386 ( p < 0.01), 0.426 ( p < 0.01), and 1.123 ( p < 0.05), respectively. Among the three measurements of FDI flows, FDI adjustment (DFDI) had the highest impact on GDP growth (b = 1.123, Eq. (II.3)). In considering the contribution to growth by investment types, only the spatial agglomeration of LC, but not FF, had directly affected GDP growth (Table 4, Eqs. (II.1)–(II.3)). The critical effects generated by domestic investment (LC) were reflected by its spatial agglomerations (ILC) both through its direct impact on GDP (Table 4, Eqs. (II.1)–(II.3)) and indirect impact via endogenous FDI (Table 4, Eqs. (I.1)–(I.3)). Furthermore, the significant negative distance effect (Dist) (Table 4, Stage I, 2SLS) supports gravity as a critical frictional factor in FDI absorption in 11 Eqs. (1)–(3) with the synergy (interaction) effect omitted were also estimated and with the two sets of estimation results compared. The inclusion of strategic interaction effect facilitated much higher partial regression coefficients (elasticities) on the part of spatial agglomerations by investment types.

706

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

Table 5 Effects of manufacturing spatial agglomerations on FDI and GDP by geographic region (PR-east vs. PR-west): results from 2SLS regression Variable

2SLS (Stage I) (I.1) FDIt

Region: PR-east 16.149# b0 FDI-hata ILC 0.723# IFF 0.790# ILC
IFF 0.048# Dist 1.061# 2 Adj-R 0.736 F-Stat 32.30# RMSE 1.044 DF 41 Variable

2SLS (Stage II) (I.2) FDIt+1

(2.501) 14.665# (2.006) (0.237) 0.631# (0.168) 0.693# (0.014) 0.039# (0.337) 0.943# 0.784 45.55# 0.857 45

(I.3) DFDI 5.889# (3.167)

Region: PR-west 13.829# (1.932) b0 FDI-hata ILC 0.160** (0.077) IFF 0.321# (0.078) Dist 1.506# (0.424) 0.791 Adj-R2 F-Stat 53.86# RMSE 0.801 DF 39

6.176# (0.730) 0.480# (0.146) 0.156** (0.062) 0.012 (0.050)

(0.192) 0.343 (0.291) (0.137) 0.597# (0.189) (0.011) 0.030* (0.015) (0.270) 0.216 (0.509) 0.590 0.809 7.56# 64.34# 0.993 0.580 21 42

2SLS (Stage I) b (I.1) FDIt

(II.1) GDPt

(II.2) GDPt+1

(II.3) DGDP

6.061# (0.733) 5.374# (0.873) 0.489# (0.146) 0.985* (0.503) ** 0.130 (0.052) 0.217 (0.159) 0.021 (0.076) 0.072 (0.216)

0.833 82.45# 0.505 46

0.603 13.64# 1.004 22

2SLS (Stage II)b (I.2) FDIt+1 13.262# (2.461)

(I.3) DFDI 6.947 (4.870)

(II.1) GDPt

(II.2) GDPt+1

(II.3) DGDP

8.640# (0.338) 8.536# (0.425) 8.588# (2.782) 0.076* (0.041) 0.042 (0.056) 0.668 (0.471) 0.435# (0.034) 0.426# (0.042) 0.499** (0.237)

0.153 (0.099) 0.246** (0.104) 1.586# (0.543) 1.736** (1.816) 0.659 0.133 0.882 26.80# 4.53** 157.83# 0.982 1.654 0.326 37 22 40

0.803 82.29# 0.399 38

0.206 3.99** 1.623 21

Notes: #, **, and * represent statistical significance at p < 0.01, p < 0.05, and p < 0.10, respectively; all estimations in loglinear form. a Indicates the corresponding endogenous variable appearing in Stage I of 2SLS regression. b Indicates the corresponding insignificant interactions (ILC
IFF) in Eqs. (1)–(3) as well as agglomeration (ILC and IFF) effects in Eq. (I.3) for the case of PR-west were omitted from the estimations.

a core–periphery economy.12 Such a result might well demonstrate the merits of a gravity center (core) and its critical role in promoting FDI and output growth in a core (Hong Kong)–periphery (PRD/Guangdong) relation. 5.4.2. Regional effects on spatial agglomerations, FDI, and GDP growth: 2SLS The statistical analyses presented in Tables 1 and 2 support the existence of regional diversities due to firm locality and the distinctive concentration patterns due to institutional effect particularly by geographic regions (PR-east versus PR-west). The patterns of FDI absorption and its impacts on regional GDP growth were examined by estimating again Eqs. (4)–(9) (that is, PR-east versus PR-west) using 2SLS. The corresponding estimation results are given in Table 5. 12

The significance of the gravity effect for FDI absorption in a core (Hong Kong)–periphery (Guangdong) system was first examined at the aggregate macro-level for the period of 1988–1992 (Tuan & Ng, 1995), disaggregate micro-(firm) level by cities/counties (Tuan & Ng, 2003a), and micro-(firm) level by industry location and clustering (Ng & Tuan, 2003).

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

707

The pattern of FDI absorption and its impacts on regional GDP growth at PR-east were found consistent with the main results presented previously, but that of PR-west showed different findings. From Table 5, the following distinctive major features are observed at PR-west. (1) Spatial agglomeration effect—ILC was found a major factor in determining GDP and growth (PR-west: Eqs. (II.1)–(II.3)). However, its effect on FDI was only limited to the current period (PR-west: Eq. (I.1)). On the other hand, IFF was found more important in determining both current and future FDI (PR-west: Eqs. (I.1) and (I.2)). At PR-east, on the contrary, ILC played major role in both current and planned FDI as well (PR-east: Eqs. (I.1)–(I.3)). (2) Strategic interaction effect—Contrary to that of at PR-east, the strategic interaction effect was not at all a significant factor in determining FDI. (3) Gravity effect—Friction remains a critical factor in affecting FDI. The high (elastic) partial elasticities (jbj > 1.5; Eqs. (I.1)–(I.3)) suggest that reducing the distance between PR-west and the center core would definitely facilitate higher FDI inflows. (4) FDI effect—Contrary to PR-east, endogenous FDI had not affected GDP except a rather weak effect was found in the determination of current GDPt (b = 0.076; p < 0.10) (PR-west: Eq. (II.1)). At PR-west, the spatial agglomeration of LC, rather than FDI, had played a more significant role in affecting GDP growth (b > 0.42; p < 0.05). The above statistical findings provide strong supports to the arguments that in researching the issue of economic growth, the distinctive institutional features and/or the regional characteristics itself seemingly call for the deliberation of regional perspectives at a less aggregate level. That is, regional growth rather than macroeconomic growth should serve as a better unit of analysis in the search of China’s growth. 5.5. Spatial agglomerations and sectoral growth—industry linkages effects: regression analysis In view of the distinctive features observed from the above, it is of further interests to investigate if different effects of spatial agglomerations by investment types would affect sectoral GDP (that is, manufacturing versus services). The critical role of the center core in affecting sectoral output at each investment destination (R) would further be examined. The regression relation formulated for the estimation is Qi = f (ILC, IFF, Dist)e where Qi = output of ith industry (i = 1 to 2 for manufacturing and services, respectively, and with the subscript R omitted for simplicity). The regression was estimated in log-linear form and the statistical results are presented in Table 6. 5.5.1. Manufacturing versus services From Table 6, both ILC and IFF significantly affected sectoral (industry) output while the gravity (distance) effect was only important to manufacturing production (Eqs. (1.1) and (1.2)). Besides, the partial impacts of ILC on manufacturing outputs were much higher than that of IFF in both cases of manufacturing and services (for LC, 0.318, 0.386; p < 0.05 versus for FF, 0.134, 0.097; p < 0.05, respectively). Such a result suggests the higher importance of local industries (manufacturing investments) in facilitating output production and industry (forward) linkages.

708

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

Table 6 Effects of manufacturing spatial agglomerations on sectoral industry growth: regression results Variable

Y = GDP b0 ILC IFF Dist Adj-R2 F-Stat RMSE DF

All regions

PRD_east

PRD_west

(1.1) GDP(M)

(1.2) GDP(S)

(2.1) GDP(M)

(2.2) GDP(S)

(3.1) GDP(M)

(3.2) GDP(S)

22.862# (1.278) 0.318# (0.068) 0.134** (0.057) 0.755# (0.260) 0.677 61.86# 0.909 84

17.844# (0.964) 0.386# (0.054) 0.097** (0.043) 0.067 (0.205) 0.723 75.87# 0.674 83

22.268# (1.799) 0.002 (0.105) 0.358# (0.086) 0.733** (0.365) 0.714 39.34# 0.945 43

17.944# (1.580) 0.213** (0.104) 0.195** (0.074) 0.173 (0.341) 0.693 34.87 0.802 42

25.015# (1.525) 18.037# (1.086) 0.567# (0.070) 0.501# (0.050) 0.019 (0.063) 0.053 (0.045) 1.018# (0.340) 0.001 (0.242) 0.791 0.826 51.33# 62.05# 0.642 0.457 37 37

Notes: Dependent variable (Y) = Sectoral GDP; GDP(M) and GDP(S) stand for manufacturing and services GDP, respectively; all estimations by OLS in log-linear form; #, **, and * represent statistical significance at p < 0.01, p < 0.05, and p < 0.10, respectively.

5.5.2. PR-east versus PR-west Again from Table 6, distinct effects on output behaviors by geographic regions (PR-east versus PR-west) were observed. At PR-east, manufacturing output was more affected by IFF and gravity while that of PR-west by ILC other than gravity (Table 6, Eqs. (2.1) and (3.1)). Such results further support the dominance of FF at PR-east but that of LC at PR-west in manufacturing production. At PR-east, both the agglomerations of LC and FF offered linkages effects in Table 7 Estimated elasticities by FDI, gravity, and spatial agglomeration effects: summary results Effects

Foreign investments FDI

Gravity Distance (Dist)

Region

(1) All (2) East (3) West

FDI

GDP

Sectoral output

FDIt

FDIt+1

DFDI

GDPt

GDPt+1

DGDP

– – –

– – –

– – –

0.386 0.480 0.076

0.426 0.489 0

1.123 0.985 0

– – –

– – –

– – –

GDP(M) – – –

– – –

(1) All (2) East (3) West

1.135 1.061 1.506

1.074 0.943 1.586

(1) All (2) East (3) West

0.574 0.723 0.160

0.554 0.631 0

0 0 0

0.297 0.156 0.435

0.278 0.130 0.426

0.384 0 0.499

0.318 0 0.567

0.386 0.213 0.501

Foreign firms (FF)

(1) All (2) East (3) West

0.653 0.790 0.321

0.599 0.693 0.246

0.569 0.597 0

0 0 0

0 0 0

0 0 0

0.134 0.358 0

0.097 0.195 0

Strategic interaction (LC
FF)

(1) All (2) East (3) West

0.034 0.048 0

0.032 0.039 0

0.032 0.030 0

– – –

– – –

– – –

– – –

– – –

Spatial agglomerations Local Chinese (LC)

0 0 1.736

0.755 0.733 1.018

GDP(S)

0 0 0

Note: Figures (elasticities) were extracted from Tables 4–6; ‘‘0’’ value denotes statistically insignificant partial regression coefficient.

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

709

enhancing service production Eq. (2.2). This is only true, however, for LC at PR-west Eq. (3.2). Moreover, the gravity effect had not affected service production regardless of regions. A summary of the estimated partial elasticities is provided in Table 7 for easy comprehension. The results from Table 7 demonstrate, once again, the critical feature and significance of the ‘spatial’ dimension and ‘regional’ perspective in study growth especially when the partial elasticities by institutional characteristics and geographic regions (PR-east versus PR-west) were compared and contrasted. 6. Reflections The research results seemingly support the postulated hypotheses regarding firm investment behaviors (both local and foreign) with specific reference to their location decisions and spatial concentration patterns as well as spatial agglomeration effect on FDI and regional GDP growth in China. The major research findings are summarized as follows: 1. The spatial concentration patterns of both local Chinese and foreign investment were directed by institutional characteristics. Highly diverse FDI patterns were found in the form of concentrations mainly in the FDI-policy regions at SEZs/PRD and especially at PR-east. Although local Chinese manufacturing firms were less diverse in concentration, they shared the common preferential locality behavior by concentrating in PRD and PR-east. 2. Both spatial agglomeration and strategic interaction (synergy) between local and foreign firms were critical to attract FDI, and hence firms’ location decisions. 3. Spatial agglomerations had not only strongly affected FDI absorption and its growth, the same effect is true on GDP. Furthermore, local manufacturing agglomerations had created much higher industry (forward) linkages than that of foreign firms. 4. The simultaneous structural relations of spatial agglomerations, FDI, and GDP demonstrated that spatial agglomerations had directly affected FDI flows and regional GDP growth via FDI after controlling for the effects of strategic interactions and friction. Manufacturing agglomerations of local firms was more important in affecting directly GDP while that of foreign firms in FDI. 5. Reducing the frictional distance of an investment destination from the center core was essential to promoting FDI and output growth in the (core–periphery) region. Bearing the above research findings in mind, the behavior of the inward FDI in China and its significant impacts on output growth from a regional perspective can be clearly deduced. 1. Given the critical roles of spatial agglomerations and the synergy effects arising from the strategic interactions (such as from demonstration effects, supporting networks, information sharing, etc.) between local and foreign firms, manufacturing investments would incline to capitalize on this advantage of agglomeration economies via their choice of investment destinations. That is, PRD, the ‘‘world’s largest manufacturing base’’, would continue to dominate as a prime investment site (Tuan & Ng, 2004b). However, the trade-off between the two regions, PR-east versus PR-west, or among the cities/counties in the region would much depend upon the transaction costs and agglomeration economies/diseconomies in production (Ng & Tuan, 2003; Tuan & Ng, 2004a). 2. With the absorption of FDI from a core–periphery perspective where Hong Kong serves as the center core and PRD/Guangdong as the manufacturing periphery (Tuan & Ng, 2003a), further

710

3.

4.

5.

6.

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

inward FDI would continue to realize at those locations lying closer to the center core, in general, due to the gravity effect of a shorter distance and the agglomeration economies such generated.13 Along this line, the provision of better infrastructural supports, particularly by connecting PRwest with the center core, would direct more FDI inflows by the virtue of decreased transaction costs via reducing the travel distance and transportation time from the core in this metropolitan economy (Tuan & Ng, 2004a). The direction of inward FDI demonstrated in this study envisaged the crucial role of the structural (institutional) factors in FDI promotion. Building a better investment environment is essential for further FDI promotion in this regard (Ng & Tuan, 2002).14 FDI and FDI driven institutional reform made rapid growth in PRD during the 1980s and subsequently, YRD during the 1990s. The dynamic developments of both global delta economies demonstrate the strategic directions of economic growth in both regions. Such dynamism definitely would raise expectations especially on the part of international investment seeking for business opportunities. Hence, any further search of a more meaningful interpretation of economic growth in China should be focused along a regional perspective via FDI absorption. After the WTO accession, China will have to abide with the WTO rules by opening up impartially to all FDI types. To leverage on the synergies of manufacturing spatial agglomerations and the economies such derived, the existing pattern of FDI inflows is expected to continue (Tuan & Ng, 2004b). Moreover, a more complete form of plant relocation to the peripheral areas in the core–periphery economy should be observed for those manufacturing industries (e.g., textiles and clothing) which quota restrictions would have been removed by then (Ng & Tuan, 2003). Other than the potential change in structural factors arising from the WTO accession, the institutional factor on upgrading the image and credibility of the local governmental administration was recently listed as a new top national agenda.15 The ‘‘globalization process’’ of the epidemics (SARS outbreak in 2003) shocked China and its famous manufacturing base (PRD) in terms of both local economic activities and foreign investment. Such an episode reinforces the critical concerns of foreign investors regarding the institutional government feature as a political risk factor of investment (Ng & Tuan, 2001), which is especially true in developing economies.

After carefully controlling for political risk and the possible detriments toward the economic and political performance of China, the year 2003 might have served as a critical turning point for 13 The effectiveness of building a ‘‘connection’’ (a new infrastructure such as a bridge) to connect PR-east and PR-west aiming at FDI absorption and economic growth in Guangdong was assessed (Tuan & Ng, 2003b). Such a construction would provide a more complete linkage effects toward the transportation networks in the region (core–periphery economy). 14 To improve the investment environment, some recent implemented major economic policies attributable to the utilization of FDI in China were the stepwise openings of the service sector and fewer restrictions imposed on technology transfer (Tuan & Ng, 2000). The top three dimensions which accounted for almost one-half (47.8%) of the factor components of an ideal investment environment in China were government and its administration, lesser restrictions on business operations, and provision of infrastructural supports (Ng & Tuan, 2002). 15 China government has attempted to continuously improve the business environment for foreign investors especially for multinationals. The decisive resolution and immediate dismissals of those incompetent and incredible high-ranking governmental officials definitely project an image of a more credible government, which can be well perceived as a reduction of political risk as far as FDI promotion is concerned.

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

711

the global economy in considering a new wave of FDI in China and particularly in the world’s largest manufacturing base (PRD). China’s huge domestic demand in an open market and its potential growth have also become critical determinants in directing the global economic activities and future FDI flows in both manufacturing and services in the region. With further improved and credible controls of some institutional factors, China’s economic presence and emergence as a super-economic power in the today’s mutually economic dependent global economy should also act as a new engine of future global economic growth. In any case, the results from this research reinforces the reflection that ‘‘spatial’’ dimension is essential in considering investment flows while ‘‘regional’’ perspective is critical in studying the dynamics of economic growth in a huge country like China. Acknowledgment The authors thank an anonymous reviewer for helpful comments. All remaining errors are ours. References Alfaro, L., Chanda, A., Kalemli-Ozcan, S., & Sayek, S. (2004). FDI and economic growth: The role of local financial markets. Journal of International Economics, 64(1), 89–112. Alonso, W. (1964). Location and land use. Harvard University Press. Authur, W. B. (1994). Increasing returns and path-dependency in economics. Ann Arbor: University of Michigan Press. Baldwin, R. E. (1997). Agglomeration and endogenous capital. European Economic Review, 43(2), 253–280. Barro, R., & Sala-I-Martin, X. (1995). Economic growth. New York: McGraw Hill. Barro, R., & Sala-I-Martin, X. (1997). Technology diffusion, convergence, and growth. Journal of Economic Growth, 2, 1–26. Basu, P., Chakraborty, C., & Reagle, D. (2003). Liberalization, FDI, and growth in developing countries: A panel cointegration approach. Economic Inquiry, 41(3), 510–516. Bende, N. A. (1999). FDI, regionalism, government policy, and endogenous growth: A comparative study of the ASEAN-5 economies, with development policy implications for the least developed countries. UK: Ashgate. Bengoa, M., & Sanchez-Robles, B. (2003). Foreign direct investment, economic freedom and growth: New evidence from Latin America. European Journal of Political Economy, 19(3), 529–545. Black, D., & Henderson, V. (1999). A theory of urban growth. Journal of Political Economy, 107(2), 252–284. Blonigen, B. A., Ellis, C. J., & Fausten, D. (2005). Industrial groupings and foreign direct investment. Journal of International Economics, 65(1), 75–91. Borensztein, E., De Gregorio, J., & Lee, J. W. (1998). How does foreign direct investment affect economic growth? Journal of International Economics, 45, 115–135. Bottazzi, G., Dosi, G., & Fagiolo, G. (2002). On the ubiquitous nature of agglomeration economies and their diverse determinants: Some notes. In A. Q. Curzio, & M. Fortis (Eds.), Complexity and industrial clusters (pp. 168–191). New York: Physica-Verlag. Braunerhjelm, P., & Ekholm, K. (Eds.), The geography of multinational firms, economics of science, technology and innovation (Vol. 12). Boston: Kluwer Academic. Curzio, A. Q., & Fortis, M. (Eds.). (2002). Complexity and industrial clusters. New York: Physica-Verlag. De Mello, L. (1999). Foreign direct investment led growth: Evidence from time series and panel data. Oxford Economic Papers, 51, 132–151. Dixit, A. K., & Stigitz, J. E. (1977). Monopolistic competition and optimum product diversity. American Economic Review, 67, 297–308. Fujita, M., & Hu, D. P. (2001). The Effects of globalization and economic liberalization. Annals of Regional Science, 35(1), 3–37. Fujita, M., & Krugman, P. (2004). The new economic geography: Past, present, and the future. Papers in Regional Science, 83, 139–164. Fujita, M., Krugman, P., & Venables, A. J. (1999). The spatial economy: Cities, regions and international trade. Cambridge, MA: MIT Press.

712

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

Fujita, M., & Mori, T. (2005). Frontiers of the new economic geography. Discussion paper no. 27. Institute of Developing Economies, JRTRO: Japan. Fujita, M., & Thissa, J. F. (1996). Economics of agglomeration. Journal of the Japanese and International Economics, 10, 339–378. Fujita, M., & Thissa, J. F. (2002). Economics of agglomeration: Cities, industrial location, and regional growth. Cambridge: Cambridge University Press. Giersch, H. (Ed.). (1995). Urban agglomeration and economic growth. Berlin: Springer-Verlag. Girma, S. (2001). The process of European integration and the determinants of entry by non-EU multinationals in UK manufacturing. GEP research paper: 2001/20: University of Nottingham. Grossman, G., & Helpman, E. (1991). Innovation and growth in the world economy. Cambridge: MIT Press. Grossman, G., & Helpman, E. (1995). Technology and trade. In Grossman, G. M., & Rogoff, K. Eds. Handbook of international economics (Vol. 3). New York: North Holland. Helpman, E., & Krugman, P. (1985). Market structure and foreign trade. Cambridge: MIT Press. Henderson, J. V. (1974). The sizes and types of cities. American Economic Review, 64, 640–656. Henderson, J. V. (1988). Urban development: Theory, fact, and illusion. New York: Oxford University Press. Isard, W. (1956). Location and space economy. Cambridge: MIT Press. Kohno, H., Nijkamp, P., & Poot, J. (Eds.). (2000). Regional cohesion and competition in the age of globalization. MA: Elgar Publishing. Krugman, P. (1990). Rethinging international trade. Cambridge, MA: MIT Press. Krugman, P. (1991a). Increasing returns and economic geography. Journal of Political Economy, 99, 483–499. Krugman, P. (1991b). Geography and trade. Cambridge, MA: MIT Press. Krugman, P. (1996). The self-organizing economy. Cambridge: Blackwell. Krugman, P. (1998). Space: The final frontier. Journal of Economic Perspectives, 12(2), 161–174. LaFountain, C. (2005). Where do firms locate? Testing competing models of agglomeration. Journal of Urban Economics, 58(2), 338–366. Lee, R., & Willis, J. (1997). Geographies of economies. London: Arnold. Liu, X., Burridge, P., & Sinclair, P. J. N. (2002). Relationships between economic growth, foreign direct investment and trade: Evidence from china. Applied Economics, 34, 1433–1440. Maggioni, M. A. (2002). Clustering dynamics and the location of high-tech-firms. New York: Physica-Verlag. Martin, R. (1999). The new ‘‘geographical turn’’ in economics: Some critical reflections. Cambridge Journal of Economics, 23, 65–91. Marwah, K., & Tavakoli, A. (2004). The effect of foreign capital and imports on economic growth: Further evidence from four Asian countries (1970–1998). Journal of Asian Economics, 15(2), 399–413. Nachum, L. (1998). Economic geography and patterns of international business activity. ESRC Center for Business Research working paper: WP114, University of Cambridge. Nair-Reichert, U., & Weinhold, D. (2001). Causality tests for cross-country panel: A new look at FDI and economic growth in developing countries. Oxford Bulletin of Economics and Statistics, 63, 153–171. Ng, L. F. Y., & Tuan, C. (2001). FDI promotion policy in China: Governance and effectiveness. The World Economy, 24(8), 1051–1074. Ng, L. F. Y., & Tuan, C. (2002). Building a favorable investment environment. The World Economy, 25(8), 1095– 1114. Ng, L. F. Y., & Tuan, C. (2003). Location decisions of manufacturing FDI in China: Implications of China’s WTO accession. Journal of Asian Economics, 14(1), 51–72. Palivos, T., & Wang, P. (1996). Spatial agglomeration and endogenous growth. Regional Science and Urban Economics, 26(6), 645–669. Puga, D., & Venables, A. J. (1996). The spread of industry: Spatial agglomeration in economic development. CEPR discussion paper series, 1354, London: CEPR. Quah, D. (2002). Spatial agglomeration dynamics. American Economic Review, 92(2), 247–252. Quigley, J. M. (1998). Urban diversity and economic growth. Journal of Economic Perspectives, 12(2), 127– 138. Richardson, H. W. (1995). Economies and diseconomies of agglomeration. In H. Giersch (Ed.), Urban agglomeration and economic growth. Berlin: Springer-Verlag. Tuan, C., & Ng, L. F. Y. (1995). Hong Kong’s outward investment and regional economic integration with Guangdong: Process and implications. Journal of Asian Economics, 6(3), 385–405. Tuan, C., & Ng, L. F. Y. (2000). China’s changing business environment: Some current issues. In C. M. Lau, & J. F. Shen (Eds.), China review, 2000 (pp. 175–190). Hong Kong: The Chinese University of Hong Kong Press.

L.F.-Y. Ng, C. Tuan / Journal of Asian Economics 17 (2006) 691–713

713

Tuan, C., & Ng, L. F. Y. (2001a). FDI in China and regional development: From institutional reform to agglomeration economies’ perspectives. Paper presented at nation states and economic policy: Conflict and cooperation. The first international conference of the Japan economics policy association: Japan. Tuan, C., & Ng, L. F. Y. (2001b). Regional division of labor from agglomeration economies’ perspective: Some evidence. Journal of Asian Economies, 12(1), 65–85. Tuan, C., & Ng, L. F. Y. (2003a). FDI facilitated by agglomeration economies: Evidence from manufacturing and services joint ventures in China. Journal of Asian Economics, 13(6), 749–765. Tuan, C., & Ng, L. F. Y. (2003b). A miracle via agglomerations: The HK-PRD connections, monograph (UNDP#CPR/91/ 161) (pp. 1–45): The Chinese University of Hong Kong. Tuan, C., & Ng, L. F. Y. (2004a). FDI and industrial restructuring in post-WTO greater PRD: Implications on regional growth in China. The World Economy, 27(10), 1609–1630. Tuan, C., & Ng, L. F. Y. (2004b). Manufacturing agglomeration as incentives to Asian FDI in China after WTO. Journal of Asian Economics, 16(1), 673–693. Tuan, C., Ng, L. F. Y., & Lin, H. (2006). The globalized delta economies in China (Yangtze-River Delta and Pearl-River Delta): Visualized economic indicators for investment environment. Institute of space and earth information science. The Chinese University of Hong Kong. Website: www.jlgis.cuhk.edu.hk/business. Venables, A. J. (1996). Equilibrium locations of vertically linked industries. International Economic Review, 37, 341–359. Wolff, R. (1985). Efficient growth of an agglomeration regional economy. Regional Science and Urban Economics, 15(4), 555–573.