Foreign Direct

8/2/2019 Foreign Direct

http://slidepdf.com/reader/full/foreign-direct 1/19

ournal of Economic Studiesmerald Article: Foreign direct investment and employment in manufacturingnd services sectors: Fresh empirical evidence from Singapore

oi Nyen Wong, Tuck Cheong Tang

rticle information:

cite this document: Koi Nyen Wong, Tuck Cheong Tang, (2011),"Foreign direct investment and employment in manufacturing and

rvices sectors: Fresh empirical evidence from Singapore", Journal of Economic Studies, Vol. 38 Iss: 3 pp. 313 - 330

rmanent link to this document:

p://dx.doi.org/10.1108/01443581111152427

ownloaded on: 27-03-2012

ferences: This document contains references to 43 other documents

copy this document: [email protected]

ccess to this document was granted through an Emerald subscription provided by STAMFORD COLLEGE

r Authors:

you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service.

formation about how to choose which publication to write for and submission guidelines are available for all. Additional help

r authors is available for Emerald subscribers. Please visit www.emeraldinsight.com/authors for more information.

bout Emerald www.emeraldinsight.com

ith over forty years' experience, Emerald Group Publishing is a leading independent publisher of global research with impact in

siness, society, public policy and education. In total, Emerald publishes over 275 journals and more than 130 book series, as



Foreign direct investment andemployment in manufacturing

and services sectorsFresh empirical evidence from Singapore

Koi Nyen WongSchool of Business, Sunway University, Selangor Darul Ehsan, Malaysia, and

Tuck Cheong Tang Department of Economics, School of Business, Monash University,

Selangor, Malaysia

Abstract

Purpose – This paper aims to examine both the cointegrating and causal relationships amonginward FDI and the host country’s employment in manufacturing and services sectors.

Design/methodology/approach – This paper applies autoregressive distributed lag (ARDL)framework to test the cointegration and causality patterns using Singapore as a case.

Findings – Apart from the presence of a unique long-run relationship, the findings also showevidence of long-run causality, running from employment in manufacturing and services to FDIinflows, and from FDI inflows and services employment to manufacturing employment. Furthermore,there is evidence of short-run causality showing strong FDI-employment and employment linkages,predominantly from the manufacturing to services.

Research limitations/implications – One likely area of future research is to extend this paper byusing disaggregated data, e.g. FDI inflows by sector (manufacturing and services), and employmentby the respective sectors.

Practical implications – Manufacturing and services have been regarded as the “twin engines” of growth for the Singapore economy. As the economy is moving up the value chain from downstream toupstream activities, a significant proportion of foreign direct investment (FDI) has been attracted tothe manufacturing and services sectors. The present study provides useful policy implicationstowards promoting foreign investment in emerging areas of and manpower development in bothsectors of the economy.

Originality/value – This paper explores the possible interactions between FDI inflows andemployment in manufacturing and services sectors as well as the employment linkages betweenmanufacturing and services in Singapore.

Keywords Causality, Foreign direct investment, Employment, Singapore

Paper type Research paper

1. IntroductionSingapore is one of the most globalised nations in the world as a result of the adoptionof open foreign investment policy and liberalised trade regime over a long period of time[1]. In 2007, Singapore was the third largest recipient of foreign direct investment(FDI) in the Asian region with flows valued at US$24.1 billion after China (US$83.5billion) and Hong Kong, China (US$59.9 billion) (see Table I). By the same token,Singapore was also one of Asia’s largest foreign sources of direct investment (DI), and

The current issue and full text archive of this journal is available at

www.emeraldinsight.com/0144-3585.htm

Foreign directinvestment

313

Received 6 January 2009Accepted 19 April 2010

Journal of Economic StudiesVol. 38 No. 3, 2011

pp. 313-330q Emerald Group Publishing Limited

0144-3585DOI 10.1108/01443581111152427



the bulk of Singapore’s stock of DI abroad went to financial services andmanufacturing (Singapore Department of Statistics, 2008a). According to the officialreport, although the 12 per cent growth in the country’s stock of FI abroad (which

amounted to S$226.3 billion in 2006) was comparable to the 12.4 per cent increase in itsstock of inward FDI in 2006 (which stood at S$363.9 billion), the latter is significantlylarger than the former in terms of magnitude (Singapore Department of Statistics,2008a, b).

In the course of globalisation, a large part of the manufacturing and services sectorsin the Singapore economy have been dominated by multinational corporations (MNCs),which use the host country as a hub for global operations and international business(Islam and Chowdhury, 1997). Thus, both manufacturing and services sectors are themajor recipients of FDI. In 2006, out of the total FDI stock, financial and insuranceservices had the largest share of 38.7 per cent followed by manufacturing (29.9 percent) (see Table II). Moreover, the distribution of FDI stock in financial and insuranceservices sector in 2006 was 81.5 per cent in investment holding companies, 7 per cent in

banks and 4.9 per cent in insurance services (see Table III). In terms of share of manufacturing FDI stock in 2006, pharmaceuticals accounted for 35.2 per centcompared to electronics (29.7 per cent), petroleum (13.2 per cent) and chemicals (7.1 percent) (Table III).

In recognition of the strategic importance of both sectors in terms FDI promotion,employment generation and contribution to GDP (Gross Domestic Product) growth, theSingapore Economic Development Board (2004) noted manufacturing and services will

Asia 2007 (billions of US dollars)

China 83.5Hong Kong, China 59.9

Singapore 24.1India 23.0 Japan Thailand 9.6Malaysia 8.4Indonesia 6.9

Source: UNCTAD (2008)

Table I.FDI inflows to majorAsian countries, 2007

Industry 2005 (%) 2006 (%)

Financial and insurance services 37.6 38.7Manufacturing 32.0 29.9

Wholesale and retail trade, hotel and restaurants 17.5 18.2Transport and storage 5.5 6.2Professional and technical, administrative andsupport services 3.4 3.0Real estate, rental and leasing services 2.6 2.8Information and communications 1.1 0.8Total 100 100

Source: Singapore Department of Statistics (2008b)

Table II.Distribution of inwardFDI by industry, 2005and 2006 (stock as atyear-end)

JES38,3

314



continue to be the “twin engines” of growth for the nation’s economy. For instance, theservices industry cluster is Singapore’s leading contributor to some of the keymacroeconomic aggregates. In 2007, services’ contributions to real GDP andemployment were 63.3 and 67.7 per cent respectively, while the manufacturing

sector accounted for 26.1 and 20.8 per cent share for the former and the latterrespectively (see Table IV). However, with reference to investment commitments,which include both fixed assets investments and business spending by both local andforeign firms, the manufacturing sector tends to attract the bulk of the totalinvestments as compared to the services industry cluster (see Table IV). In 2008,foreign investments still remained as the largest contributor to investmentcommitments in both the manufacturing and services sectors in Singapore with 89.7per cent share (S$16,183 million) compared with 10.3 per cent (S$1,863 million) indomestic investment (Ministry of Trade and Industry, Republic of Singapore, 2008).

In fact, manufacturing and services could mutually reinforce each other. Forinstance, diversifying the manufacturing sector to high value chain activities could

generate intermediate (also known as derived) demand for services ranging fromlogistics and ports to business and finance. Conversely, a strong complementaryservices sector could encourage more manufacturing activities; thus creating asmutually reinforcing growth cycle[2].

Broadly speaking, this study contributes to the existing literature on FDI byexploring the long run relationships and causal linkages among FDI inflows, andemployment in manufacturing, and in services, for a small open economy namelySingapore. In general, the existing studies (i.e. Feenstra and Hanson, 1995; Mickiewiczet al., 2000; Ernst, 2005; Jenkins, 2006; Jayaraman and Singh, 2007) use aggregate andsectoral data to examine the relationships between FDI inflows and employment. Forinstance, Feenstra and Hanson (1995) found that inward FDI increased the relativedemand for skilled labour in Mexican manufacturing. Mickiewicz et al. (2000)concluded that the role of FDI in employment creation/preservation had been mostsuccessful in Hungary attributable to its structural similarity to developed economies.Ernst’s (2005) findings showed that Mexico, except for Argentina and Brazil,experienced a strong employment creation as a consequence of a surge inmanufacturing FDI. Using Fiji as a country case study, Jayaraman and Singh (2007)

2005 (%) 2006 (%)

Financial and insurance services 100 100Investment holding companies 78.4 81.5Banks 8 7

Insurance services 4.7 4.9

Manufacturing 100 100Pharmaceutical products 37.3 35.2Electronics products 28.7 29.7Petroleum products 13.4 13.2Chemical products 7.2 7.1

Source: Singapore Department of Statistics (2008b)

Table III.FDI in financial and

insurance services as wellas manufacturing sector,

2005 and 2006 (stock as atyear-end)


315



R e a l G D P

E m p l o y m e n t

I n v e s t m e n t c o m m i t m e n t s

S e c t o r

2 0 0 5

2 0 0 6

2 0 0 7

2 0 0 5

2 0 0 6

2 0 0 7

2 0 0 5

2 0 0 6

2 0 0 7

P e r c e n t a g e c o n t r i b u t i o n f r o m : m a n u f a c t u r i n g

a

( % )

2 5 . 7

2 6 . 6

2 6 . 1

2 0 . 5

2 0 . 7

2 0 . 8

9 0 . 6

8 5 . 4

9 3 . 6

P e r c e n t a g e c o n t r i b u t i o n f r o m : s e r v i c e s p r o d u c i n g

i n d u s t r i e s ( % )

6 3 . 4

6 3 . 0

6 3 . 3

6 8 . 7

6 8 . 4

6 7 . 7

9 . 4

1 4 . 6

6 . 4

N o t e : a T h e m a n u f a c t u r

i n g s e c t o r i n S i n g a p o r e c o m p r i s e s t h e f o l l o w i n g s u b s e c t o r s : e l e c t r o n i c s , c h e m i c a l s ,

b i o m e d i c a l m a n u f a c t u r i n g , p r e c i s i o n

e n g i n e e r i n g , t r a n s p o r t e n

g i n e e r i n g a n d g e n e r a l m a n u f a c t u r i n g

i n d u s t r i e s ( M i n i s t r y o f T r a d e a n d I n d u s t r y ,

R e p u b l i c o f S i n g a p o r e ,

2 0 0 8 )

S o u r c e : M i n i s t r y o f T r a

d e a n d I n d u s t r y ,

R e p u b l i c o f S i n g a p o

r e ( 2 0 0 8 ) a n d M i n i s t r y o f M a n p o w e r ,

R e p u b l i c o f S i n g a p o r e ( 2 0 0 8 )

Table IV.Percentage contributionsto real GDP, employmentand investmentcommitments frommanufacturing andservices sectors, 2005 to2007

JES38,3

316



found that there was unidirectional long-run causality running from FDI toemployment.

However, this study explores the possible interactions between FDI inflows andemployment in manufacturing and services sectors as well as the employment linkages

between manufacturing and services using Singapore as a case. It is important tohighlight that this study aims to examine if there is a cointegrating relationship as wellas causality patterns among these variables of interest from the macroeconomicperspective, rather than providing empirical evidence of what determine the FDIinflows into Singapore, in particular using firm level data such as by Greenaway et al.(1994); Blomstrom and Kokko (1994); Lipsey (1994); Urata (2006). Moreover, this studyapplies the error-correction version of autoregressive distributed lag (ARDL) approachto cointegration and causality, which has the advantage of avoiding the stationarypre-testing for small sample size (Pesaran et al., 2001).

The rest of this paper is organised as follows. Section 2 describes the analyticalframework exploring the inter-linkages among FDI inflows, employment inmanufacturing and in services. This section also discusses the data andmethodology – determination of the stationarity of each variable based on unit roottests and the Granger causality (Granger, 1969, 1988) test procedure based on the errorcorrection adaptation of the ARDL approach. Results are reported in section 3.Concluding remarks and policy discussions are documented in section 4.

2. Analytical framework, data, and autoregressive distributed lag (ARDL)approach 2.1 Analytical frameworkThe analytical framework to the linkages between FDI inflows and employment inmanufacturing and services can be conceptually constructed based on the microfoundation of macroeconomics, and analysed using the causality approach in

macroeconomic perspective. In this relation, the literature has at least conceptually andempirically documented that the key factors for attracting FDI inflows are to augmentdomestic capital, create employment, and create access to foreign technology, globalnetwork and also the organization structure. On the other hand, the wage differentialsbetween local and foreign firms, human capital developments, R&D activities, andspillovers from multinational activities are also important factors determining FDIinflows (Blomstrom and Kokko, 1994; Lipsey, 1994; Chen, 1996; Khawar, 2003;Thangavelu and Owyong, 2003; Urata, 2006). More recent study considers productionfragmentations and outsourcing on employment as the other key factors for firms toinvest abroad and link with the global production chain (Greenaway et al., 1994). Inprinciple, the literature has provided little macroeconomic explanations on theinter-linkages between FDI inflows and employments via conventional production

function, i.e. Y ¼ A F(L, K, H, N) (where Y is the quantity of output, A is the level of technology, L is quantity of labour, K is physical capital, H is human capital, and N isnatural resource). Thus, FDI inflows could enhance the factors of production via thevarious channels of transmission to the level of technology (A), physical capital (H),and human capital (H), and increase the quantities of output (Y) and employmentcreation (or increasing employment) in the economy as a whole.

However, the possible inter-linkages (i.e. causality) among FDI inflows, employmentin manufacturing, and services can be perceptively elaborated from the microeconomic


317



perspective. In view of the fact that FDI inflows are highly concentrated inmanufacturing and services sectors in Singapore as well as there is evidencesupporting interdependence between different sectors of the economy (Greenhalgh andGregory, 2001; Franke and Kalmbach, 2005; Guerrieri and Maliciani, 2005; Thangavelu

and Tan, 2008), the present study empirically explores the causality pattern betweeninward FDI and the host country’s employment in manufacturing and services as wellas the employment linkages between these two sectors. Thus, using relativelyconventional vector autoregressive (VAR) based Granger causality test procedures tostudy these linkages, we can ascertain the extent of the causal-linkages between theinflows of FDI and Singapore’s employment in manufacturing and services sinceinward FDI can potentially contribute to the nation in terms of technology transfer,usage of local inputs, increased job opportunities and other spill over effects. By thesame token, the causality test results can provide inputs for formulating appropriatepolicies for the development of labour market in Singapore that can help attract foreignfirms to locate and grow in emerging areas of manufacturing and services that arerelevant to global businesses. By and large, Singapore represents a suitable case studyfor such an investigation because it is an export-oriented economy that actively courtFDI to advance domestic industrialisation and economic development along withmanufacturing and services as the key growth drivers of the economy.

In principle, causality can run in either or both directions, i.e. from FDI inflows toemployment in manufacturing and/or services and, conversely, from employment inmanufacturing and/or services to FDI inflows. For example, if there is a boost ininward FDI in manufacturing sector, this could directly create employment in thesector itself due to higher capital investment by MNCs. The total employment effect of inward FDI could be greater given multinational trade and investment activities alsoinvolve backward and forward integrations[3]. If the manufacturing and servicessectors are mutually supporting, this could also lead to increased employment in

services too. However, if inward FDI is seen as a substitute for domestic investment, itis expected to have little and sometimes negative effects on employment owing to thedisplacement of local firms by relatively more efficient foreign firms. As pointed out byBellak (2004), the crowding out of weak domestic firms by foreign entry could lead tosome short run job losses as a result of competition effect. Eventually, Dunning (1994)argued that the engagement in backward linkages by MNCs in the host economy couldresult in some job losses in domestic supporting industries due to low local sourcing byMNCs. For the range of potential effects including empirical studies of the impacts of inward FDI on employment, see Baldwin (1994), UNCTAD (1994), Ernst (2005),

Jayaraman and Singh (2007) and Jenkins (2006).On the other hand, an increase in employment in manufacturing or and services

could also encourage inward FDI. A rapid growth of skilled workforce[4] in the host

country is instrumental in attracting foreign firms to set up their bases there ashigh-tech start-ups or to perform some of the core activities (e.g. research anddevelopment, production, marketing, delivery, and provision of after-sale services), orsupporting activities (e.g. procurement of inputs, technology, human resources, andother infrastructure, like management and finance)[5]. Dunning (1993) noted that someof these activities can readily be performed cheaper or better by suppliers locatedelsewhere, which can lead to the determination of cross-border transaction flows inintermediate goods and services by multinational hierarchies.

JES38,3

318



2.2 DataThe data for FDI inflows is obtained from International Financial Statistics,International Monetary Fund while the source of the time-series on employment inmanufacturing (EM) and services (ES) is from CEIC Asia Economic Database. All the

raw data are transformed into natural logarithm ( ln ) with the exception of FDI inflows,which is converted into real terms before the logarithmic transformation occurs[6]. Theraw data are provided in Table V, while Figure 1 presents the time series plot of eachvariable (in logarithmic forms). A visual inspection of the time series plots fails todetect any seasonal pattern for both EM and ES, while a cyclical component is presentfor the EM series. As expected, the FDI inflows variable is fluctuating from period toperiod, and an outlier appears in 2002Q3. The estimation period starts from the second

Employment: manufacturing(‘000)

Employment: services(‘000)

FDI(US$ million)

Jun, 1997 437.0 1,224.0 625Sep, 1997 439.5 1,262.3 6,927Dec, 1997 442.8 1,262.3 3,115Mar, 1998 441.8 1,265.4 1,524 Jun, 1998 433.8 1,266.3 2,192Sep, 1998 423.2 1,260.7 3,055Dec, 1998 415.3 1,270.9 485Mar, 1999 412.0 1,275.2 8,487 Jun, 1999 416.1 1,289.9 2,233Sep, 1999 418.5 1,300.0 1,990Dec, 1999 419.7 1,324.1 3,859Mar, 2000 419.2 1,340.6 4,617 Jun, 2000 425.7 1,360.9 2,824Sep, 2000 436.8 1,378.2 3,865

Dec, 2000 445.5 1,404.6 5,174Mar, 2001 444.4 1,432.9 3,177 Jun, 2001 439.3 1,445.2 2,619Sep, 2001 435.5 1,441.4 3,762Dec, 2001 430.3 1,442.1 6,021Mar, 2002 429.2 1,442.5 5,414 Jun, 2002 430.0 1,446.2 1,002Sep, 2002 428.5 1,447.3 2751Dec, 2002 425.0 1,458.6 1,403Mar, 2003 422.4 1,463.7 4,094 Jun, 2003 415.9 1,451.6 2,567Sep, 2003 415.9 1,454.6 1,916Dec, 2003 420.0 1,466.6 3,094Mar, 2004 426.1 1,478.0 6,568

Jun, 2004 432.5 1,486.0 5,818Sep, 2004 441.0 1,493.1 3,404Dec, 2004 447.2 1,521.3 4,025Mar, 2005 452.7 1,531.2 3,973 Jun, 2005 462.0 1,550.4 3,111Sep, 2005 470.2 1,568.9 4,506Dec, 2005 476.7 1,595.2 3,415

Source: CEIC Asia Economic DatabaseTable V.Raw data


319



quarter of 1997 to the fourth quarter of 2005. The choice of this sample period is basedon the availability of the data especially the series for EM and ES.

2.3 ARDL approach to Granger causality testsConventionally, in order to perform the Granger causality test (Granger, 1969, 1988),specification of the possible causal linkages between the endogenous variables, i.e.inward ln FDI, ln EM and ln ES can be tested in terms of a VAR system framework. Oneof the typical settings in the standard VAR causality test is to ensure all the variablesincluded in the system are stationary ( I (0)). The commonly used tests for unit root (i.e.ADF and PP) suggest that ln ES, and ln EM are non-stationary, or I (1), with theexception of ln FDI, which is stationary in levels, I (0) (the results are tabulated inTable VI)[7]. Thus, the VAR specification for the Granger causality test procedureincludes the stationary I (0) variables such as Dln ES, and Dln EM (which aredifferenced once), and ln FDI. Generally speaking, combining the level ( ln FDI) anddifferenced variables ( Dln ES, and Dln EM) in VAR framework may cause possiblemisspecification by considering the fact that the specification of the VAR model isshort run, and the long run adjustment is introduced through the vector autoregressiveerror correction form (VECM) by an error correction term, which is a level laggedresidual variable (or error correction version of ARDL form by a level laggeddependent variable on the right hand side). Besides, first difference (i.e. use changes)

Figure 1.Plots of time seriesvariables (in log forms)

JES38,3

320



the nonstationary variables can eliminate the long run information in the VARestimation for causality tests.

As advocated by Pesaran et al. (2001), this unrestricted error correction specificationcan be applied irrespective of whether the regressors are stationary ( I (0)) or

nonstationary ( I (1)), and it is able to avoid the pre-testing problems associated withstandard cointegration analysis which requires the classification of variables into I (1)and I (0). Hence, Pesaran et al. (2001) had tabulated a set of appropriate critical valuesfor all I (0), I (1), or both integration regressors. Narayan (2005) re-tabulated the criticalvalues for different numbers of observations as small as 30. In particular, thestatistically significance of error correction term (i.e. the level lagged dependentvariable) of ARDL model (which is adapted with error correction) depicts long runcausality, while the joint significance of the lagged first differenced variables capturesshort run causality. Tang (2006) applied the so-called error-correction version of ARDL(Pesaran et al., 2001) to investigate the direction of causality between exportsexpansion and economic growth in China.

Formally, before the ARDL model is estimated, the order of lag length for the VARhas to be identified. The optimal lag length is chosen based on a set of statisticalselection information criteria viz. final prediction error (FPE), Akaike informationcriterion (AIC), Schwarz criterion (SC) and Hannan-Quinn information criterion (HQ).Table VII provides the appropriate lag length for the estimated VAR models. Theoptimal lag length for the chosen VAR models is two as indicated by FPE and SC, AIC,and HQ, respectively.

Lag FPE AIC SC HQ

0 7.65e-09 210.17534 210.03657 210.130111 1.25e-11 216.59568 216.04059 216.414732 6.53e-12 *

217.26451 *216.29310 *

216.94785 *

3 7.17e-12 217.21847 215.83074 216.766114 9.82e-12 216.99870 215.19465 216.41062

Notes: *Indicates lag order selected by the criterion; Conventionally, a maximum lag order of four isconsidered since the data used in this study are quarterly

Table VII.VAR lag order selection

criteria

ADF PP I (d)

ln FDI 24.426971 [1] (0.0067) 24.884826 [5] (0.0020) ln FDI , I (0)

ln ES2

1.861177 [10] (0.6428)2

1.851293 [3] (0.6573)Dln ES 22.551784 [8] (0.1161) *

24.802364 [3] (0.0005) Ln ES , I (1)ln EM 22.118626 [6] (0.5137) 20.817778 [2] (0.9539)Dln EM 23.130351 [1] (0.0342) 21.995321 [5] (0.2874) * * ln EM , I (1)

Notes: FDI ¼ Foreign Direct Investment; ES ¼ Employment for services; and EM ¼ Employment inmanufacturing. Unit root equation for the variable in levels includes a constant and a trend while forthe variable in first-differenced, only a constant is included. The optimum lag order is in [.] which isbased on AIC from a maximum of 12 lags. The value in (.) is p-value. *The p-value is 0.0018 for thecase without constant and trend; * *The p-value is 0.0467 for the case without constant and trend

Table VI.Results of ADF and PP

unit root tests forstationarity


321



Thus, the error correction version of ARDL framework can be written as equations (1)to (3). The next step is to estimate these equations to determine whether there is acointegrating relationship (via bounds testing procedure proposed by Pesaran et al.,2001), and causal-linkages (i.e. long run and short run causality via t -test and F -tests,

respectively) among inward FDI, employment in services (ES), and employment inmanufacturing (EM):

D ln FDI t ¼ a0 þ a1 ln FDI t 21 þ a2 ln ES t 21 þ a3 ln EM t 21 þX2

j¼1

a4D ln FDI t 2 j

þX2

j¼1

a5D ln ES t 2 j þX2

j¼1

a6D ln EM t 2 j þ ut ð1Þ

D ln ES t ¼ b0 þ b1 ln FDI t 2

1 þ b2 ln ES t 2

1 þ b3 ln EM t 2

1 þX

2

j¼1b4D ln FDI t

2 j

þX2

j¼1

b5D ln ES t 2 j þX2

j¼1

b6D ln EM t 2 j þ u‘t ð2Þ

D ln EM t ¼ c0 þ c1 ln FDI t 21 þ c2 ln ES t 21 þ c3 ln EM t 21 þX2

j¼1

c4D ln FDI t 2 j

þX2

j¼1 c5D

ln ES t 2

jþX

2

j¼1 c6D

ln EM t 2

jþ

u

"

t ð3Þ

More precisely, we can test the long run relationship (cointegration) among thevariables, i.e. FDI, ES, and EM in equation (1) for example, under the null hypothesis of no long run relationship (H0: a1 ¼ a2 ¼ a3 ¼ 0) against the alternative hypothesis of long run relationship ((HA: a1 – 0, a2 – 0, a3 – 0) using the F -statistic (Wald test). If the computed F -statistic is greater than the upper I (1) critical value as tabulated inNarayan (2005), then the null hypothesis can be rejected, suggesting cointegration.However, if the computed F -statistic is less than the lower I (0) critical value, the testfails to reject the null hypothesis and we conclude that there is no cointegration. If thetest statistic lies within the lower ( I (0)) and upper ( I (1)) critical bounds, conclusiveinference can only be made once the order of the integration of the underlyingregressors is known (see Pesaran et al., 2001). Narayan’s (2005) provided a set of criticalvalues for small sample study i.e. 30 observations, which is applicable for this paper,which gives effectively 35 observations.

On the other hand, the lagged dependent variable in level (i.e. ln FDIt-1, ln ESt-1, andln EMt-1 ) can be interpreted as error correction term in this ARDL equation (oraugmented error correction model). Conversely, even if the F -statistic rejects the null of no cointegration, we still cannot draw a conclusion because a more powerful test forcointegration in this setup is the coefficient obtained from error correction term (see

JES38,3

322



Kremers et al., 1992). As documented by Bahmani-Oskooee and Brooks (1999), if thelagged error correction term turns out to be negative and significant, cointegration issupported. According to Bahmani-Oskooee and Rehman (2005, p. 775):

However, these results [ F -test for cointegration] are considered as preliminary at this stageand more evidence of cointegration is looked for in the second stage when an appropriate lagselection criterion is employed.

Kanioura and Turner (2005) found that the F -test has higher power than theEngle-Granger test but lower power than the t -form of the error-correction test. Thus,the significance of the estimated coefficients of ln FDIt-1, ln ESt-1, or ln EMt-1 (or errorcorrection term) appears to be robust indicating that there is a cointegratingrelationship among these three variables (see for example, Bahmani-Oskooee andBrooks, 1999).

By the same token, the statistically significance of ln FDIt-1, ln ESt-1, or ln EMt-1

(error correction term in error correction version of ARDL framework) indicates thedirection of (long run) causality running from the right hand side variables (i.e. ES and

EM jointly) to the left hand side variable (i.e. FDI). The Wald test, which follows thechi-square distribution, is computed to test the causal relations among FDI, ES, andEM based on the ARDL framework. Conventionally, the jointly significance of theestimated coefficients of the lagged first differenced variables reveals the direction(short run) causality via the joint significance test (chi-square statistics).

3. ResultsFigure 2 presents the ordinary least square (OLS) estimates of ARDL-error correctionmodel as in Pesaran et al. (2001) for the VAR(2) as suggested by the lag length selectioncriteria reported in Table VII. The Ramsey RESET statistics (with large p-value) showno evidence of model misspecification in equations (1) to (3). Again, the plots of

CUSUM statistics reveal no structural break(s) for the underlying equations. Thesediagnostic test results suggest that the two outliers (such as a sharp decline in FDI in1998 and 2001 due to the Asian financial crisis and the September 11 terrorists’ attackrespectively) did not affect the stability of the ARDL estimates of equations (1) to (3).

The bound testing statistics ( F -statistics) for equations (1) and (3) as reported inTable VIII are 3.018, and 1.265 respectively, which are below the lower bound criticalvalue (3.437), while the test statistic for equation (2), 4.460 is closer to the upper boundcritical value of 4.470. Clearly, as highlighted by Bahmani-Oskooee and Brooks (1999)and Bahmani-Oskooee and Rehman (2005), no final conclusion can be made about acointegrating relationship among FDI, EM, and ES without looking at other test forcointegration based on the error correct term. The estimated coefficients and p-valuesfor the error correction terms in equations (1) to (3) are reported in Figure 2, which

shows that the estimated coefficients of the error correction terms for equations (1) and(2) are statistically significant (at 0.10 level of significance), implying there is acointegration among FDI, EM, and ES for Singapore. They have correct sign withconsiderable variation in the speed of adjustment (e.g. how long does FDI take torestore long-run equilibrium after some short-run changes in its regressors). Forexample, Dln FDIt takes about one quarter (one divided by the estimated coefficient of the error correction term) to converge to long-run steady state, while Dln EMt requiresapproximately thirteen quarters to restore long-run equilibrium. On the other hand, the


323



results of Johansen’s multivariate cointegration tests suggest at least one cointegratingvector (not reported here but are available from authors upon request), which also

support the earlier findings on cointegration on the grounds that an equilibriumrelationship is present among the variables on interest.

The first panel in Table IX provides the chi-square statistics and p-values for theerror correction terms for equations (1) to (3). These statistics show that both EM andES do jointly Granger-cause FDI inflows in the long run (equation 1), while there is alsoa long-run causality running from FDI inflows and ES to EM (equation 2). Theseresults corroborate the arguments that employment in both manufacturing andservices can encourage FDI inflows into Singapore in the long run owing to its

Figure 2.ARDL estimates

Dependent variableDln FDIt

(equation 1)Dln EMt

(equation 2)Dln ESt

(equation 3)

F -statistic 3.018 4.460 1.26510 per cent critical value I (0) 3.437(Narayan, 2005, p. 1988, n ¼ 30 and k ¼ 2) I (1) 4.470

Table VIII.Cointegration test(bounds testing)

JES38,3

324



well-endowed skilled workforce in both sectors. On the contrary, inward FDI as well asemployment in services can potentially promote employment in manufacturing in thelong run due to higher capital investment by MNCs in the manufacturing sector on onehand, and the strong complementary services sector on the other, hence, this can leadto more employment in manufacturing.

The second panel (in Table IX) reports the results of joint significant test of thelagged first differenced variables which can be used to determine the direction of causality in the short run. The summary of the direction of the short-run causalityamong FDI, EM, and ES for Singapore is given by Figure 3, which shows that there is ashort-run causal relationship running from FDI inflows to employment in

manufacturing, which in turn Granger causes employment services but not in thereverse direction. This implies that as the inward FDI is moving up the value chainfrom downstream to upstream activities, an expansion of skilled workforce in thissector can potentially create employment in the manufacturing, which could lead toincreased employment in services ranging from logistics and information technologyto business and finance, for example. This empirical observation is well in line with thestudies by Thangavelu and Tan’s (2008) findings, who ascertained the employmentlinkages in the Singapore economy using input-output framework for 18 industrialsectors at 2-digit industrial classification. This unique causality pattern advocates the

Figure 3.Short-run inter-linkage

between inward FDI,manufacturing

employment and servicesemployment

Dependent variable Dln FDIt (equation 1) Dln EMt (equation 2) Dln ESt (equation 3)

Error-correction term 21.027 * * * (0.008) 20.136 * * (0.034) 20.019 (0.330)

X2

j¼2

b jD ln FDI t 2 j – 5.839 * (0.054) 1.101 (0.577)

X2

j¼2

b jD ln EM t 2 j 0.109 (0.947) – 9.109 * * (0.011)

X2

j¼2

b i D ln ES t 2 j 1.844 (0.398) 3.518 (0.172) –

Notes: The error correction term refers to the coefficient of ln FDIt-1, ln ESt-1, and ln EMt-1,respectively. The value in (.) is p-value. The joint significant test statistics (Chi-square statistics) arereported in the second panel and the value in (.) is p-value. * * *, * *, and * denote 1, 5, and 10 per cent,respectively

Table IX.Summary of causality

tests


325



argument that the high value-added FDI, i.e. in manufacturing and services sectors, iscapable of creating demand for employment in view of the fact that these two sectorscould form the “twin-engines” of growth for Singapore economy. In turn, a significantproportion of FDI tends to attract skilled workforce from domestic manufacturing, and

employment in manufacturing causes employment in services. The empirical evidenceof FDI-employment in manufacturing-services linkage would be consistent withmultinational trade and investment activity that involves in forward linkages. As aresult of sustained FDI inflows into the country, Singapore is in the position to promotehighly educated and well-trained labour force (i.e. in both manufacturing and services)to undertake high value-added manufacturing activities and provide professional andtechnology services locally and internationally.

4. ConclusionsThe present study empirically investigates the inter-linkages between FDI inflows,manufacturing and services employment for a small open economy – Singapore,

which is well endowed with human resources, offers attractive environments andincentives for foreign investors and, both manufacturing and services sectors are seenas future engines of growth. The basis for such an investigation is that an increase inFDI inflows could lead to higher employment in the host country if it complementswith domestic investment. The employment effect of FDI inflows is greater if it isconcentrated in labour-intensive industries ( Jenkins, 2006). Conversely, the growth inthe skilled workforce in emerging industries could attract FDI, which can potentiallylead to technology transfer.

The cointegration results indicate that there is a unique long-run relationship amongFDI inflows, employment in manufacturing and in services. Besides, the findings showthat there is unidirectional causation that runs from employment in manufacturing andservices to FDI inflows in the long run, implying apart from her strategic location in the

region, sophisticated infrastructure, and attractive investment incentives, an expansionof skilled workforce in these sectors is instrumental in drawing FDI inflows into thecountry, thus, opening up a channel for important source of technology and betterinternational business network. On the other, there is also evidence of causation runningthe other way, from FDI inflows and employment in services to employment inmanufacturing, suggesting the presence of MNCs has lower propensity to establishlinkages with local industries, and the services employment can be seen as a complementto manufacturing employment given these sectors are mutually supporting. Thepossible explanations for low local sourcing by MNCs are MNCs have massive resourcesand international network, and local firms’ business methods are not suited to globalmarkets. To forge linkages between MNCs and domestic firms, the government shouldassist local companies to tie up with foreign partners (e.g. forming strategic alliances

with MNCs), which can increase the transnationality of domestic firms. Furthermore, thegovernment should continue to develop programmes for manpower developmentfocusing on critical skills needed by the wholly owned subsidiaries of foreigncorporations in order to increase local sourcing.

The findings also provide robust evidence of short-run causation that runs from FDIinflows to manufacturing employment and in turn, employment linkage is runningfrom manufacturing to services, which supports the view that as the manufacturingindustries move up the value chain, they tend to generate employment spill-over effects

JES38,3

326



to the services sector. Hence, in light of the fact that both the manufacturing andservices are regarded as the “twin engines” of growth for Singapore economy and thesetwo sectors are interlinked, the appropriate strategy to increase the job opportunities inthe services sector is to diversify the manufacturing base and increase high

value-added activities in emerging manufacturing industries.Even though this study provides fresh empirical evidence on the patterns of

causation between overall FDI inflows and employment in manufacturing, andservices in Singapore in both short and long run, one likely area of future research is toextend the present study by using disaggregated data on FDI inflows (e.g. FDI inflowsspecifically in manufacturing as well as services instead of using the aggregated FDIas employed in this paper), which not only may reveal some interesting asymmetriesand comparable results but also are able to allow one to determine whether the findingsin this paper can be generalized or not.

Notes

1. Accessed at: www.atkearney.com/shared_res/pdf/Globalization-Index_FP_Nov-Dec-06_S.pdf (accessed 1 July 2008).

2. Accessed at: www.edb.gov.sg/edb/sg/en_uk/index/news_room/news/2005/important_to_ keep.html?showMode¼ printable (accessed 1 July 2008).

3. Backward integration is achieved when inputs are being imported from abroad or homecountries of MNCs for value added in the host country while forward integration is achievedwhen the intermediate or final outputs are being produced and exported back to their homecountries or affiliates elsewhere for assembly and distribution (Sieh-Lee, 2000).

4. Other pull factors that attract FDI may include strategic location, sophisticatedinfrastructure, attractive investment incentives etc.

5. According to Porter’s (1985) generic value chain model, it is unusual that one single firmperforms all the core and supporting activities by itself.

6. In September 2002, Singapore’s inward FDI had a negative value of US$751 million. In orderto remove the negative magnitude for logarithmic transformation, a constant value of US$2,000 million had been added to the whole FDI time series in real terms. The price indexused to deflate the FDI variable into real terms is consumer price index (CPI).

7. Given that both the ADF and PP test statistics suggest the FDI inflows in Singapore isstationary in levels (i.e. I(0)), it is interesting to do an ERS point-optimal unit root test (Elliottet al., 1996) to ascertain whether this variable of interest is stationary. However, this test failsto reject the null hypothesis of a unit root for this variable given that its test statistic is 4.092,which is less than the 0.10 critical value (6.77). However, the KPSS (Kwiatkowski et al., 1992)test statistic shows that the FDI inflows is stationary, in which the null of trend stationaritycannot be rejected (i.e. the test statistic is 0.0695, and the 0.10 level critical value is 0.119). Unitroot tests with a structural break (i.e. specifications with impulse dummy, and shift dummy)

(Lanne et al., 2003) were also considered. The tests indicate there is a structural break in2003Q3, and both tests specifications reject the null of a unit root for Singapore’s FDI in levelsimplying stationarity. The test statistics are23.3554 (0.10 level critical value is22.58), and23.7853 (0.10 level critical value is 22.76) for impulse dummy, and shift dummy settings,respectively. Apart from the ERS point-optimal unit root test, clearly, the majority of the unitroots test statistics support theFDI inflowsin Singaporeare stationary. In addition, Wong andTang (2007) also found that the FDI inflows in Malaysia is stationary in levels (I(0)). If ERSpoint-optimal unit root test is considered, then thefirst-differenced operation on FDI will causeinformation loss as highlighted by most of the empirical research.


327



References

Bahmani-Oskooee, M. and Brooks, T.J. (1999), “Bilateral J-curve between US and her tradingpartners”, Weltwirtschaftliches Archiv (Review of World Economics), Vol. 135 No. 1,pp. 156-65.

Bahmani-Oskooee, M. and Rehman, H. (2005), “Stability of the money demand function in Asiandeveloping countries”, Applied Economics, Vol. 37 No. 7, pp. 773-92.

Baldwin, R.E. (1994), “The effects of trade and foreign direct investment on employment andrelative wages”, OECD Economic Studies, No. 23, pp. 7-54, Winter.

Bellak, C. (2004), “How performance gaps between domestic firms and foreign affiliates matterfor economic policy”, Transnational Corporations, Vol. 13 No. 2, pp. 29-55.

Blomstrom, M. and Kokko, A. (1994), “Home country effects of foreign investment: the case of Sweden”, in Globerman, S. (Ed.), Canadian-Based Multinationals, University of CalgaryPress, Calgary, pp. 341-64.

Chen, C. (1996), “Regional determinants of foreign direct investment in mainland China”, Journal of Economic Studies, Vol. 23 No. 2, pp. 18-30.

Dunning, J. (1993), Multinational Enterprises and the Global Economy, Addison-Wesley,Wokingham.

Dunning, J. (1994), “Re-evaluating the benefits of foreign direct investment”, Transnational Corporations, Vol. 3 No. 1, pp. 23-52.

Elliott, G., Rothenberg, T.J. and Stock, J.H. (1996), “Efficient tests for an autoregressive unit root”, Econometrica, Vol. 64 No. 4, pp. 813-36.

Ernst, C. (2005), “The FDI-employment link in a globalizing world: the case of Argentina, Braziland Mexico”, Employment Strategy Papers 2005/17 , pp. 1-45.

Feenstra, R.C. and Hanson, G.H. (1995), “Foreign investment, outsourcing and relative wages”,NBER working paper no. 5121, NBER, Cambridge, MA.

Franke, R. and Kalmbach, P. (2005), “Structural change in the manufacturing sector and itsimpact on business-related services: an input-output study for Germany”, Structural Change and Economic Dynamics, Vol. 16 No. 4, pp. 467-88.

Granger, C.W.J. (1969), “Investigating causal relations by econometric models and crossspectralmethods”, Econometrica, Vol. 37 No. 3, pp. 424-38.

Granger, C.W.J. (1988), “Some recent developments in a concept of causality”, Journal of Econometrics, Vol. 39 Nos 1/2, pp. 199-211.

Greenaway, D., Hine, R. and Milner, C. (1994), “Country-specific factors and the pattern of horizontal and vertical intra-industry trade in the UK”, Weltwirtschaftliches Archiv,Vol. 130 No. 1, pp. 77-100.

Greenhalgh, C. and Gregory, M. (2001), “Structural change and the emergence of the new serviceeconomy”, Oxford Bulletin of Economics and Statistics, Vol. 63, pp. 629-46.

Guerrieri, P. and Maliciani, V. (2005), “Technology and international competitiveness:

the interdependence between manufacturing and producer services”, Structural Changeand Economic Dynamics, Vol. 16 No. 4, pp. 489-502.

Islam, I. and Chowdhury, A. (1997), Asia-Pacific Economies: A Survey, Routledge, London.

Jayaraman, T.K. and Singh, B. (2007), “Foreign direct investment and employment creation inPacific Island countries: an empirical study of Fiji”, pp. 1-17, Asia-Pacific Research andTraining Network on Trade working paper series no. 35.

Jenkins, R. (2006), “Globalization, FDI and employment in Viet Nam”, Transnational Corporations, Vol. 15 No. 1, pp. 115-42.

JES38,3

328



Kanioura, A. and Turner, P. (2005), “Critical values for the F -test for cointegration in amultivariate model”, Applied Economics Letters, Vol. 37 No. 3, pp. 265-70.

Khawar, M. (2003), “Productivity and foreign direct investment – evidence from Mexico”, Journal of Economic Studies, Vol. 30 No. 1, pp. 66-76.

Kremers, J.J.M., Ericsson, N.R. and Dolado, J.J. (1992), “The power of cointegration tests”, Oxford Bulletin of Economics and Statistics, Vol. 54 No. 3, pp. 325-48.

Kwiatkowski, D., Phillips, P., Schmidt, P. and Shin, Y. (1992), “Testing the null hypothesis of stationarity against the alternative of a unit root”,Journal of Econometrics, Vol. 54 Nos 1-3,pp. 159-78.

Lanne, M., Lutkepohl, H. and Saikkonen, P. (2003), “Test procedures for unit roots in time serieswith level shifts at unknown time”, Oxford Bulletin of Economics and Statistics, Vol. 65No. 1, pp. 91-115.

Lipsey, R. (1994), “Outward direct investment and the US economy”, National Bureau of Economic Research (NBER) working paper, no. 4691.

Mickiewicz, T., Radosevic, S. and Varblane, U. (2000), “The value of diversity: foreign direct

investment and employment in central Europe during economic recovery”, The ESRC“One Europe or Several?” programme working paper 05/00, Sussex European Institute,University of Sussex.

Ministry of Manpower, Republic of Singapore (2008), “Manpower statisitics in brief, Singapore2007”, available at: wwwmom.gov.sg (accessed 1 June 2009).

Ministry of Trade and Industry, Republic of Singapore (2008), “Economic survey of Singapore:second quarter 2008”, available at: www.mti.gov.sg (accessed 1 June 2009).

Narayan, P.K. (2005), “The saving and investment nexus for China; evidence from cointegrationtests”, Applied Economics, Vol. 37 No. 17, pp. 1979-90.

Pesaran, M., Shin, Y. and Smith, R. (2001), “Bounds testing approaches to the analysis of levelrelationships”, Journal of Applied Econometrics, Vol. 16 No. 3, pp. 289-326.

Porter, M.E. (1985), Competitive Advantage: Creating and Sustaining Superior Performance,The Free Press, New York, NY.

Sieh-Lee, M.L. (2000), Taking on the World: Globalization Strategies in Malaysia, McGraw-Hill,Malaysia.

Singapore Department of Statistics (2008a), “Singapore investment abroad, 2006”, available at:www.singstat.gov.sg/pubn/business/sia2006.pdf (accessed 1 June 2009).

Singapore Department of Statistics (2008b), “Foreign equity investment in Singapore, 2006”,available at: www.singstat.gov.sg/pubn/business/fei2006.pdf (accessed 1 June 2009).

Singapore Economic Development Board (2004), “Annual report 2003/04”, available at: www.sedb.com/edb/sg/en_uk/index/about_us/annual_report_/annual_report_2003.html(accessed 1 July 2008).

Tang, T.C. (2006), “New evidence on export expansion, economic growth and causality in China”, Applied Economics Letters, Vol. 13 No. 12, pp. 801-3.

Thangavelu, S.M. and Owyong, D.T. (2003), “The impact of export growth and scale economieson productivity in Singapore’s manufacturing industries”, Journal of Economic Studies,Vol. 30 No. 6, pp. 623-35.

Thangavelu, S.M. and Tan, Z.L. (2008), “Employment linkages and structural changes inSingapore”, Economic Survey of Singapore 2007 , Ministry of Trade and Industry,Singapore, pp. 131-45.


329



UNCTAD (1994), World Investment Report 1994: Transnational Corporations, Employment and the Workplace, United Nations, Geneva.

UNCTAD (2008), “World investment report 2008: transnational corporations and theinfrastructure challenge”, available at: www.unctad.org/en/docs/wir2008_en.pdf

(accessed 1 June 2009).Urata, S. (2006), “The creation of regional production network in Asia-Pacific: the case of

Japanese multinational corporations”, East Asian Bureau of Economic Research (EABER), June, pp. 1-11.

Wong, K.N. and Tang, T.C. (2007), “Foreign direct investment and electronics exports:exploratory empirical evidence from Malaysia’s top five electronics exports”, Economics

Bulletin, Vol. 6 No. 14, pp. 1-8.

Corresponding authorTuck Cheong Tang can be contacted at: [email protected]

JES38,3

330

To purchase reprints of this article please e-mail: [email protected] visit our web site for further details: www.emeraldinsight.com/reprints

Foreign Direct

Documents

Transcript of Foreign Direct