FDI Spatial Spillovers in China

• Author: Mi Lin,Yum K. Kwan
• Published date: 01 August 2017
• Date: 01 August 2017
• DOI: http://doi.org/10.1111/twec.12337

FDI Spatial Spillovers in China

Mi Lin

1

and Yum K. Kwan

2

1

Lincoln Business School, University of Lincoln, Lincoln, UK and

2

Department of Economics and

Finance, City University of Hong Kong, Kowloon, Hong Kong

1. INTRODUCTION

MULTINATIONAL enterprises (MNEs) have been a major influence on the world econ-

omy and the focus of considerable attention. Such enterprises have been at the centre of

wide-ranging debates both within home countries and in recipient nations. In recipient nations,

for example, apart from macroeconomic issues such as tax revenues, employment, exports

and economic growth, it is argued that foreign direct investment (FDI) may potentially influ-

ence indigenous firms through various spillover effects (Blomstr€

om and Kokko, 1998). Over

the past two decades, both academic researchers and policymakers have paid a great deal of

attention to FDI spillovers that occur particularly in developing countries. Based on the fact

that MNEs on average possess superior technologies or knowledge capital, policymakers in

developing countries seem to take it for granted that there will be beneficial spillovers from

MNEs to nearby indigenous firms. Relatively little attention has been paid to the issue of

potential negative spillovers. In this study, by exploiting very detailed location information of

manufacturing firms in China, we provide empirical evidence showing that negative FDI spil-

lovers can be substantial along the spatial dimension.

Many studies investigate FDI spillovers in the context of firm or industry agglomeration.

This is motivated by the Marshallian notion that firms tend to agglomerate in specific areas so

as to reduce transaction costs and enjoy external economies. It is well documented in the

existing literature that the location of MNEs exhibits similar self-reinforcing pattern not just

along the geographical but also along the temporal and sectoral dimensions (see, among

others, Head et al., 1995; Cheng and Kwan, 2000a, 2000b; Blonigen et al., 2005; Lin and

Kwan, 2011). While this line of literature tends to gravitate on the notion that agglomeration

promotes spillovers, the direction and magnitude of the association between geographical dis-

tance and spillovers, however, is not as clear as one would expect.

The knowledge diffusion literature has long documented the co-existence of both positive

and negative spillovers in the context of firm agglomeration. Audretsch and Feldman (1996) and

Gertler (2003) emphasise that knowledge diffusion is highly sensitive to geographical distance

and that proximity promotes knowledge spillovers. Boschma and Frenken (2010), Broekel and

Boschma (2012), and Cassi and Plunket (2013) provide empirical evidence for the prevalence of

the so-called proximity paradox, which states that while proximity in various dimensions (geo-

graphical, cognitive, organisational and social) may be a driver for agents to connect and

exchange knowledge, too much proximity in any of the dimensions could result in substantially

weakened or even negative innovative performance and spillovers. It has also been argued that

We would like to thank Professor Cheng Hsiao for technical advice on spatial dynamic panel economet-

ric method. Helpful comments for earlier versions of the paper have been received from Chia-Hui Lu,

Larry Qiu, Eden S.H. Yu and seminar participants at City University of Hong Kong, Tsinghua Univer-

sity, University of Waterloo and Wilfrid Laurier University. All remaining errors are our own. This work

is supported by the Research Center of International Economics of City University of Hong Kong

(Project No. 7010009).

©2015 John Wiley & Sons Ltd

1514

The World Economy (2017)

doi: 10.1111/twec.12337

The World Economy

congestion pricing may come along with agglomeration, as excess demand would boost local

cost for land, labour and even public goods, thereby generating negative pecuniary externalities

onto the firms. Consequently, while geographical proximity between MNEs and indigenous

firms may facilitate knowledge diffusion and hence increase the likelihood of positive spillover s,

it can also bring along negative spillovers that adversely affect the performance of indigenous

firms (G€

org and Greenaway, 2004; G€

org and Strobl, 2005).

Many empirical analyses in the literature treat each region as an isolated entity. The role of

spatial dependence is completely ignored, even though it has been found to be an important force

in the process of productivity growth (Rey and Montouri, 1999; Madariaga and Poncet, 2007).

In the context of FDI spillovers, there is no reason why MNCs located in a particular region

can only generate spillovers onto indigenous firms collocated in the same region. Similarly,

intra- and inter-regional spillovers can also occur among indigenous firms. The pitfalls of ignor-

ing these spatial interactions, if they actually exist in the data, have also been well documented

in the econometric literature (Anselin, 2001). These considerations motivate us to employ in this

study a spatial dynamic panel model with both time and spatial autoregressive terms, which

simultaneously accounts for spatial interactions among indigenous firms as well as that between

indigenous firms and MNEs. The model is estimated by spatial dynamic panel GMM method

(Kukenova and Monteiro, 2009) that accounts for endogeneity and spatial dependence problems.

One of the unsolved issues in the FDI spillovers literature is the lack of consensus on the

genuine mechanism of spillovers. In most early empirical studies, documented association

between domestic firms’ productivity performance and FDI presence at most reveal a com-

pounded effect summarising both positive and negative effects from various channels. To

uncover the genuine spillovers realisation process, this compounded net effect needs to be fur-

ther decomposed by adding more dimensions into analysis. Recent literature thus increasingly

looks at FDI spillovers through industry linkages, origin of foreign ownership and firm hetero-

geneities (see Aitken and Harrison, 1999; Javorcik, 2004; Crespo et al., 2009; Lin et al.,

2009; Abraham et al., 2010; Hale and Long, 2011; Xu and Sheng, 2012, among other contri-

butions). In this study, we attempt to identify different channels of FDI spillovers using

geography scope. More specifically, although it might be difficult to entirely decompose

mixed effects from various channels, we would like to propose that spillovers from different

channels have different geography scopes and there is a possibility that one source of spil-

lover would become dominant in either narrow or wider geography scope.

As documented in the literature, on the one hand, positive knowledge spillovers (skill

acquisition) are more likely through labour turnover (Kokko, 1996; Glass and Saggi, 1998).

Using worker flow data in Norway, Balsvik (2011) document that labour mobility from

MNEs to non-MNEs represents a true knowledge externality. Moreover, a significant part of

knowledge spillovers are through MNCs’ transfer of their techniques, quality systems and stan-

dardisation procedures to their local suppliers by providing training to local labour (Blomstr€

om

and Kokko, 1998). Markusen and Trofimenko (2009) document that the use of foreign experts

by MNCs has substantial, although not always immediate, positive effects on the value added

per worker of domestic firms. On the other hand, labour turnover could also gener ate negative

spillovers to domestic firms when foreign firms poach local talents from their domestic coun-

terparts. In terms of the geography scope of spillover through employment, however, the nega-

tive effect of poaching local talents is more likely to have narrow scope and is confounded

locally. Knowledge spillovers through customer and supplier linkage, imitation, or the transfer

of quality systems and standardisation procedures nevertheless are less likely to be contained

within a small area and thus may become dominant and be identified among wider geographic

©2015 John Wiley & Sons Ltd

FDI SPATIAL SPILLOVERS IN CHINA 1515