1. INTRODUCTION

Foreign investment is increasinglymoving towards the service sector. Manycountries are affected by the rise of serviceforeign investments and the broad basedgrowth of service TNCs (TransnationalCorporations). The efficiency productivityand supply capacity of the industries of host-

countries can be enhanced either directly orindirectly through foreign investment inservices, thereby benefiting the economy asa whole. An attempt has been made toexamine the role of foreign investmentinflows in promoting growth in the threemain economic sectors, namely primary,secondary and tertiary. The growth effects offoreign investment inflows should be sector

FOREIGn InVESTMEnT InFLOWS AnD SECTORAL GROWTH

PATTERn In InDIA- An EMPIRICAL STUDY

Mousumi Bhattacharya*

Army Institute of Management, Judges Court Road, AliporeKolkata, West Bengal, Pin: 700027, India

(Received 19. June 2012; accepted 3 December 2012)

Abstract

The paper aims to study the causal relationship between foreign investment inflows and sectoralgrowth pattern of the three main sectors namely primary, secondary and tertiary sector of the Indianeconomy during the post liberalization period (1996-97:Q1 to 2009-10:Q2). Using Granger causalitytest conducted in a multivariate VAR framework, bidirectional causality is observed between foreigninvestment inflows and secondary sector growth. It is seen that tertiary sector growth causes foreigninvestment inflows. Regression analysis is also done with the same dataset and it reveals that bothsecondary sector growth and tertiary sector growth have led to foreign investment inflows to theeconomy after 1990 reforms.

Keywords:Foreign investment inflows, Granger causality, Sectoral growth pattern

* Corresponding author: mousumi.ma@gmail.com

S e r b i a n

J o u r n a l

o f

M a n a g e m e n t

Serbian Journal of Management 8 (1) (2013) 39 - 51

www.sjm06.com

DOI: 10.5937/sjm8-2929

specific as the productivity spillovers iswidely believed to differ across sectors. Thestudies conducted so far either dealt with onesector without due consideration of all thethree sectors of the Indian economy.

It is true that greater foreign investmentinflows bring unambiguous benefits to theeconomy by generating employment, raisingproductivity, transferring foreign skills andtechnology and contributing positively togross domestic product, gross fixed capitalformation and balance of payments. Apartfrom stimulating export markets andproducing foreign exchange revenue, foreigninvestment inflows can also contributetowards servicing debt repayments. About athird of total global exports are generated bythe subsidiaries of MNCs (MultinationalCorporations), which bring the vast portionof the FDI (foreign direct investments). Theimpact of foreign investments (both FDI andportfolio flows) depends on what form ittakes and includes the type of foreigninvestments, sector, scale, duration andlocation of business and secondary impactson the economy. In order to reap the benefitsof foreign investments it is crucial torefocuss on the perspective apart frommerely enhancing the availability of foreigninvestments and better application of foreigninvestments for attaining sustainableobjectives.

Foreign investment inflow, where itgenerates and expands business enhancessocial development by stimulatingemployment, raising wages and activatingdeclining market sectors. The benefits offoreign investment may only be felt by asmall section of the population, e.g., whereemployment and training is given to themore educated. In case of wealthy elites orwhere there is urban emphasis wagedifferentials between income groups will be

exacerbated. Cultural and social impactsmay also occur when investment isparticularly directed at non-traditionalgoods. Huge foreign currency inflowsresulting from increased foreign investmentinflows under a flexible exchange rateregime where market forces determineexchange rate, essentially result inappreciation of the local currency.

Apart from stimulating export marketsand producing foreign exchange revenue,foreign investment inflows can alsocontribute towards servicing debtrepayments. The impact of foreigninvestment inflows depends on what form ittakes and includes the type of foreigninvestment inflows, sector, scale, durationand location of business and secondaryimpacts on the economy. In order to reap thebenefits of foreign investment inflows it iscrucial to refocus on the perspective apartfrom merely enhancing the availability offoreign investment inflows and betterapplication of foreign investment inflows forattaining sustainable objectives. The impactof foreign investment inflows depends notonly on the characteristics of the sectors butalso on the linkage potential of these sectorswith the rest of the economy.

2. REVIEW OF LITERATURE

It has been observed from the theoreticalwork of Findlay (1978) and Wang andBlomstrům (1992) that the importance ofFDI as a conduit for transferring technology,relates to foreign investment inflows tomanufacturing and services sector ratherthan primary sector. In his seminal book oneconomic development Hirschman (1958)described that FDI’s potential to createlinkages might vary across sectors.

40 M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

Hirschman (1958) also emphasized that notall sectors have the same potential to absorbforeign technology or to create linkages withthe rest of the economy and noted thatlinkages are weak in agriculture and mining.Kokko (1994) observed that linkages varyacross industries and agrues that spilloverscould not be expected in all kinds ofindustries because foreign companiessometimes operate in “enclaves” that offerlittle scope for the local economy to benefit.

In UNCTAD World Investment Report(2001) it is argued that in the primary sector,the scope of linkages between foreignaffiliates and local suppliers is often limitedwhereas the manufacturing sector has abroad variation of linkage intensive activitiesand in the tertiary sector the scope fordividing production into discrete stages andsubcontracting out large parts to independentdomestic firms is also limited. Rodriguez-Clare (1996) shows that multinationals’intensive use of intermediate goods enhancesproduction in host economies and withincreased demand for inputs lead to apositive externality to other producers owingto an increase in variety. Greater variety ofinputs seems to be more relevant to themanufacturing than to the agricultural sector.Alfaro (2003) conducted a study for theperiod 1981-1999 for 47 countries and cameto the conclusion that FDI exerts anambiguous effect on growth. FDI in theprimary sector, tends to have a negativeeffect on growth while investment inmanufacturing sector have a positive effecton growth of the economy. However, theeffect on growth from the investment in theservices sector is ambiguous.

In Indian context, Kathuria (2002) wasthe first to estimate the impact of foreigninvestment on productivity and spillovers todomestic firms in Indian manufacturing

sector in the liberalization period 1989-90 to1996-97. Chakraborty and Nunnenkamp(2008) assess the proposition that FDI inIndia will promote economic growth bysubjecting industry specific FDI and outputdata to Granger causality tests with a panelcointegration framework and it turns out thatthe growth effects of FDI vary widely acrosssectors. Lo and Liu (2009) demonstrate whyChina has been so successful indisproportionately attracting foreignoffshore manufacturing activities, whileIndia has been engaged mainly in offshoreservice activities.

Assessing the role of foreign investmentinflows in the context of various sectors(agriculture, industry, service) andexamining output growth of those sectors(measured by share of those sectors in GDP)over the years have not been adequatelyexamined at various sectoral levels duringthe post reform period in India. The numberof empirical studies that cast the relationshipof foreign investment inflows in differentsectors and its impact on growth is limiteddue to unavailability of data. This paper aimsto study the causal relationship betweenforeign investment inflows and sectoralgrowth pattern of the three main sectorsnamely primary, secondary and tertiarysector of the Indian economy during the postliberalization period (1996-97:Q1 to 2009-10:Q2).

3. METHODOLOGY

3.1. Data Source

This study used quarterly data on foreigninvestments inflows (both FDI and portfolioflows) in India covering the period 1996-97:Q1 to 2009-10:Q2. The shares of the three

41M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

broadly classified sectors (primary,secondary and tertiary) in the quarter- wiseGDP estimates at constant (1999-2000)prices are taken. The share of primary sector(including agriculture, forestry and fishing)in the GDP, measuring growth of therespective sector over the years is denoted byPRI. The share of secondary sector(including mining & quarrying,manufacturing, electricity, gas & watersupply and construction) in the GDP,measuring growth of the respective sectorover the years is represented by SEC. Theshare of tertiary sector (including trade,hotels, transport & communication,financing, insurance, real estate and businessservices along with community, social andpersonal services) in the GDP, measuringgrowth of the respective sector over the yearsis denoted by TER. The data have been takenfrom various issues of the Reserve Bank ofIndia bulletin.

3.2. Estimation Issues

Regression Analysis

A simple regression model designed bythe author is used to predict which of thethree sectors is contributing more towardsforeign investment inflows in the country forthe time period 1996-97:Q1 to 2009-10: Q2.

The linear regression equation used heremay be represented as:

(1)

where is the constant term. LnPRIrepresents the primary sector’s share in GDP(in logarithmic form), LnSEC denotes thesecondary sector’s share in GDP (inlogarithmic form), LnTER denotes thetertiary sector’s share in GDP (in logarithmic

form) and LnFI represents foreigninvestment inflows (in logarithmic form). isthe error term. are the coefficientsto be estimated.

The study uses a Granger causality test ina multivariate Vector Autoregressive (VAR)framework to examine causality betweenforeign investment inflows and sectoralgrowth pattern for the time period 1996-97:Q1 to 2009-10:Q2 (Table 1).

Tests for Stationarity

To test the stationarity of the variablesADF test, PP test and KPSS tests areperformed.

Tests for Causality

The models used for testing Grangercausality in a VAR framework at level form:

(2)

(3)

(4)

42 M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

1 2 3LnFI LnPRI LnSEC LnTER

1 2 3, ,

LnFIt = p

1j11,jLnFIt-j +

p

1j12,jLnPRIt-j +

p

1j13,jLnSECt-j +

p

1j14,jLnTERt-j + 1t

LnPRIt = p

1j21,jLnPRIt-j +

p

1j22,jLnFIt-j +

p

1j23,jLnSECt-j +

p

1j24,jLnTERt-j + 2t

LnSECt = p

1j31,jLnSECt-j +

p

1j32,jLnFIt-j +

p

1j33,jLnPRIt-j +

p

1j34,jLnTERt-j + 3t

(5)

where LnFI, LnPRI, LnSEC and LnTERare the time series of foreign investmentinflows, share of the primary sector, share ofthe secondary sector and share of the tertiarysector respectively in GDP (measuringgrowth of the respective sectors) and they arein the logarithm form. arewhite noises. p is the lag length of VAR.

The Cointegration test is not feasible inthis study. It is feasible only if the variablesare non-stationary at their levels. From Engeland Granger’s original definition,cointegration refers to variables that areintegrated of the same order. In this study thevariables are either I(0), I(1) or I(2) and notintegrated of the same order. In other words,before testing the cointegration (i.e., toestablish an existence or otherwise of a long-term equilibrium relationship) between twoeconomic time series, say X and Y, it is firstnecessary to test whether they are integratedof the same order. This suggests that it wouldnot be feasible to consider the cointegrationanalysis, which implies that a long-runrelationship does not exist between FI andthe growth of the primary, secondary andtertiary sectors for the period of the study1996-97: Q1 to 2009-10: Q2. Theprerequisite for the Cointegration test is thatthe given variables should be I(1). In theabsence of the above prerequisite in India’sdata series pertaining to FI and sectoralgrowth, the cointegration test has beenomitted and short run Granger causality testis done.

4. DISCUSSION OF RESULTS

The results in Table 2 reveal that thesecondary sector’s share in GDP issignificant at 1% level in explaining FIinflows and tertiary sector’s share in GDP issignificant at 1% in explaining FI inflows inIndia during the period 1996-97: Q1 to 2009-10: Q2. This signifies that growth in thesecondary sector and tertiary sector affectsforeign investment inflows in the hostcountry. The R2 value of 0.961714 issignificant in explaining measurement ofgoodness of fit of the regression model. Thesmall p value (0.00001) of the F statisticreveals that the regression is significant. Thestudy has employed various diagnostic testsviz., Jarque Bera normality test,Autoregressive ConditionalHeteroskedasticty (ARCH) LM test andRamsey RESET specification test toexamine the validity and realiability of theregression model.

Jarque Bera test statistic (Table 3) is usedfor testing whether the residuals of the seriesare normally distributed. The null hypothesisis of a normal distribution – a smallprobability value leads to the rejection of thenull hypothesis. Here the null hypothesiscannot be rejected (p value=0.514513) so itcan be concluded that the residual series isnormally distributed.

Ramsey’s RESET test is general test formisspecification of functional form. The nullhypothesis that the functional form iscorrectly specified is tested and theconsequent F statistic and the log likelihoodratios are reported. Both F and versions ofthe test (Table 4) shows that there is no-apparent non-linearity in the regressionmodel and it can be concluded that the linearmodel for FI is appropriate.

43M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

LnTERt = p

1j41,jLnTERt-j +

p

1j42,jLnFIt-j +

p

1j43,jLnPRIt-j +

p

1j44,jLnSECt-j + 4t

1t, 2t, 3t and 4t

2

The ARCH test is a Lagrange multiplier(LM) test for autoregressive conditionalheteroskedasticity (ARCH) in the residuals.The ARCH test is one of a joint nullhypothesis that all q lags of the squaredresiduals have co-efficient values that are notsignificantly differently form zero. The Fstatistic is an omitted variable test for thejoint significance of all lagged squaredresiduals. The Obs*R-squared statistic isEngle’s LM test statistic, computed as thenumber of observations multiplied by the co-efficient of multiple correlation and isasymptotically distributed as a . (Table5) Both the F-version and the LM-statistic

are not quiet significant and there is no archeffects in the residuals of the estimatedmodel. GDP sector share is presented inFigure 1.

Finally CUSUM and CUSUMSQ tests areused to check the stability of the parametersin the model. The null hypothesis ofparameter stability cannot be rejected at the5% level of significance as the cumulatedsum stays inside the 95% confidence band incase of both CUSUM and CUSUMSQ tests.The CUSUM test indicates stability in theequation during the sample period becausethe line (blue) lies within the 5% criticallines (Figure 2). The CUSUM of squares test

44 M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

Table 1. VAR Lag Order Selection Criteria (LnFI, LnPRI, LnSEC and LnTER)Lag Log L LR FPE AIC SC HQIC

0 98.67392 NA 2.66e-07 -3.786957 -3.633995 -3.728708 1 234.0005 243.5878 2.26e-09 -8.560020 -7.795210 -8.268776 2 350.1641 190.5083 4.16e-11 -12.56656 -11.18991 -12.04233 3 407.5529 84.93543 8.22e-12 -14.22212 -12.23361* -13.46488 4 437.7782 39.89739* 4.96e-12* -14.79113* -12.19078 -13.80090*

*indicates lag order selected by the criterion

Table 2. Regression Results for the Time Period 1996-97:Q1 - 2009-10:Q2

Constant LnPRI LnSEC LnTER 2R

2R F-statistic

LnFI -49.6586*** (2.81381) [-17.6481]

0.008062 (0.19019) [0.04238]

7.57938*** (1.011597) [7.492502]

2.3852*** (0.82719) [2.88357]

0.9617 0.9594 418.657 (p value =0.00001)

*** indicates significant at 1% level. Standard errors in ( ) & t-statistics in [ ]

Table 3. Jarque Bera testDiagnostic

test

Purpose Test Statistic Probability Conclusion

Jarque Bera test

Normality 1.329070 0.514513 Normally distributed

Table 4. Ramsey RESET TestF -statistic 1.914629 Prob. F (1,49) 0.1727

Log likelihood ratio 2.069819 Prob. Chi-Square(1) 0.1502

Table 5. Heteroskedasticity Test – ARCHF -statistic 0.114530 Prob. F (1,51) 0.7364

Obs*R-squared 0.118755 Prob. Chi-Square(1) 0.7304

2 ( )q

45M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

8

9

10

11

12

13

14

5 10 15 20 25 30 35 40 45 50

LNFI LNPRI

LNSEC LNTER

LnFI,LnPRI,LnSEC,LnTER

Year

Figure 1. Logarithmic values of foreign investment inflows, primary sector’s share to GDP,secondary sector’s share to GDP, tertiary sector’s share to GDP

- 2 0

- 1 0

0

1 0

2 0

1 0 1 5 2 0 2 5 3 0 3 5 4 0 4 5 5 0

C U S U M 5 % S i g n i f i c a n c e

Figure 2. Diagrammatic representation of CUSUM Test

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

5 10 15 20 25 30 35 40 45 50

CUSUM of Squares 5% Significance

Figure 3. Diagrammatic representation of CUSUMSQ Test

shows that the cumulative sum of the squaresis within the 5% significance lines,suggesting that the residual variance is stable(Figure 3).

Test for Stationarity

Table 6 and Table 7 represent the resultsof ADF, PP tests of unit root respectively bylag length chosen based on minimum valuesof Schwarz information criterion (SIC).

KPSS test is also performed on both the leveland first differences of the lagged variables.

In Table 6 the unit root test is doneassuming the data set having a trend andTable 7 represents the data not having atrend. It is seen as per Table 6 that thevariable foreign investment inflows is an I(1)process according to ADF, PP and KPSStests. The variable PRI is an I(1) processaccording to ADF, KPSS test but is an I(0)process according to PP test. The variable

46 M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

Table 6. Test of Unit Root Hypothesis without Trend (1996-97:Q1 to 2009-10:Q2) ADF Statistic PP Test KPSS Series Test Statistic Lags Test Statistic Lags Test Statistic Band

width Level -0.44574 0 -0.44574 0 0.83457*** 6

LnFI First Difference

-8.52138*** 0 -8.52138*** 0 0.061344 1

Level 0.108848 8 -85.6542*** 8 0.94408*** 4 LnPRI First

Difference -2.884514** 7 -209.758*** 7 0.345592 11

Level 0.417082 4 0.778940 4 0.86455*** 6 LnSEC First

Difference -2.174470 3 -44.2354*** 3 0.272891 11

LnTER Level 1.415755 4 -0.404592 4 0.87989*** 6

First Difference

-2.473573 4 -105.256*** 4 0.257110 11

(a)The critical values are those of McKinnon (1991).(b)***, ** and * represent the rejection of null hypothesis at 1%, 5% and 10% levels of significance respectively.

Table 7. Test of Unit Root Hypothesis with Trend (1996-97:Q1 to 2009-10:Q2) ADF Statistic PP Test KPSS

Series Test Statistic Lags Test Statistic Lags Test Statistic Bandwidth

Level -2.896910 0 -2.896910 0 0.13232*** 5 LnFI First

Difference -8.46169*** 0 -8.46169*** 0 0.043105 1

Level -1.570167 8 -51.7343*** 8 0.14112*** 11 LnPRI First

Difference -2.871645 7 -253.478*** 7 0.278100 11

Level -1.921142 4 -2.806372 4 0.22206*** 5 LnSEC First

Difference -2.263210 3 -41.2156*** 3 0.148913 11

LnTER Level -3.557074** 10 -39.6120*** 10 0.12215*** 14

First

Difference -3.957665** 3 -28.5945*** 3 0.25597*** 11

(a)The critical values are those of McKinnon (1991).(b)***, ** and * represent the rejection of null hypothesis at 1%, 5% and 10% levels of significance respectively.

SEC is an I(2) process according to ADF testbut is an I(1) process according to PP andKPSS test. The variable TER is an I(2)process according to ADF test but is an I(1)process according to PP and KPSS test.

In Table 7 it is seen that the variableforeign investment inflows is an I(1) processaccording to ADF, PP and KPSS tests. Thevariable PRI is an I(2) process according toADF test, is an I(0) process according to PPtest and is an I(1) process according to KPSStest. The variable SEC is an I(2) processaccording to ADF test, is an I(1) processaccording to PP test and KPSS test. Thevariable TER is an I(0) process according toADF and PP test, is an I(2) process accordingto KPSS test.

Causality Test

Although it has been concluded that thereis no cointegration (i.e. long run relationshipbetween the variables), it does not meanabsence of causality or relation in the shortterm. In cases where FI inflows and growthin primary, secondary and tertiary sector donot move together in the long run (i.e., nocointegration) it is possible for the variables

to affect each other in the short run. Since thevariables in a VAR are not cointegrated theGranger causality tests can be conductedusing a standard F distribution. Since thevariables are either I(0), or I(1) or I(2), theGranger causality test can be performed atthe level form in the framework of VARmodel. The VAR estimates are computedtaking each one of them once as thedependent variable and the other three as theindependent variables. The null hypothesis isaccepted or rejected on the basis of “Fstatistic” to determine the joint significanceof the restrictions under the null hypothesis.The lag length is justified by a minimum ofFinal prediction error (FPE), Schwarzinformation criterion (SIC) and Likelihoodratio test statistics. The test result suggestslag order of 4 as optimal lag as seen in Table1. In Table 8 the null hypothesis “LnPRIdoes not Granger cause LnFI” cannot berejected (p value 0.2802), which indicatesthat the coefficients of the lags of LnPRI arejointly zero in the equation for LnFI and itcan be concluded that growth in the primarysector growth does not cause foreigninvestment inflows. The null hypothesis“LnFI does not Granger cause LnPRI” may

47M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

Table 8. Pairwise Granger Causality TestsSample: 1-54 Lags: 4 Null Hypothesis: Obs F-Statistic Prob. LnPRI does not Granger Cause LnFI 50 1.31581 0.2802 LnFI does not Granger Cause LnPRI 2.27202 0.0779 LnSEC does not Granger Cause LnFI 50 5.93664 0.0007 LnFI does not Granger Cause LnSEC 3.02407 0.0283 LnTER does not Granger Cause LnFI 50 3.68539 0.0118 LnFI does not Granger Cause LnTER 1.15126 0.3463 LnSEC does not Granger Cause LnPRI 50 4.76831 0.0030 LnPRI does not Granger Cause LnSEC 20.8061 0.0000 LnTER does not Granger Cause LnPRI 50 4.75154 0.0030 LnPRI does not Granger Cause LnTER 5.09583 0.0020 LnTER does not Granger Cause LnSEC 50 16.5118 0.0000 LnSEC does not Granger Cause LnTER 4.84521 0.0027

be rejected (p value 0.0779) and it can beconcluded that foreign investment inflowscauses growth in the primary sector. The nullhypothesis “LnSEC does not Granger causeLnFI” may be rejected (p value 0.0007) andit can be concluded that growth in thesecondary sector causes foreign investmentinflows. The null hypothesis “LnFI does notGranger cause LnSEC” may be rejected (pvalue 0.0283) and it can be concluded thatforeign investment inflows causes growth inthe secondary sector. It is seen thatbidirectional causality is observed betweensecondary sector growth and foreigninvestment inflows. The null hypothesis“LnTER does not Granger cause LnFI” maybe rejected (p value 0.0118) and it can beconcluded that growth in the tertiary sectorcauses foreign investment inflows. The nullhypothesis “LnFI does not Granger causeLnTER” cannot be rejected (p value 0.3463)and it can be concluded that foreigninvestment inflows does not cause growth inthe tertiary sector. The null hypothesis“LnSEC does not Granger cause LnPRI”may be rejected (p value 0.0030) and it canbe concluded that secondary sector growthcauses growth in the primary sector. The nullhypothesis “LnPRI does not Granger causeLnSEC” may be rejected (p value 0.0000)and it can be concluded that primary sectorgrowth causes growth in the secondarysector. So, birectional causality is observedbetween the growth of these two sectorsnamely secondary and primary sector.

The null hypothesis “LnTER does notGranger cause LnPRI” may be rejected (pvalue 0.0030) and it can be concluded thattertiary sector growth causes growth in theprimary sector. The null hypothesis “LnPRIdoes not Granger cause LnTER” may berejected (p value 0.0020) and it can beconcluded that primary sector growth causes

growth in the tertiary sector. So,bidirectional causality is observed betweenthe growth of primary and tertiary sectors.The null hypothesis “LnTER does notGranger cause LnSEC” may be rejected (pvalue 0.0000) and it can be concluded thattertiary sector growth causes growth in thesecondary sector. The null hypothesis“LnSEC does not Granger cause LnTER”may be rejected (p value 0.0027) and it canbe concluded that secendary sector growthcauses growth in the tertiary sector. So,bidirectional causality is observed betweenthe growth of secondary and tertiary sectors.

5. CONCLUSION

Bidirectional causality is observedbetween foreign investment inflows andsecondary sector growth and it is seen thattertiary sector growth causes foreigninvestment inflows. The impact of foreigninvestment inflows depends not only on thecharacteristics of the sectors but also on thelinkage potential of these sectors with therest of the economy. Foreign investmentinflows to the agricultural sector (primarysector) is mostly capital intensive and thescope and linkages between foreigncompanies and the rest of the economy isoften limited, whereas foreign investmentinflows in manufacturing sector may have alarger impact in the economy through abroad range of potential linkage intensiveactivities. Services sector (tertiary sector)includes a wide range of different activitiessuch as finance, infrastructure (electricity,water and telecommunications), wholesaleand retail, real estate as well as tourism.Foreign investment inflows to this sector ismostly to serve the domestic market andhence create potential forward linkages for

48 M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

the sectors, which are quite strong, whilebackward linkages may vary by industry.

Major part of the foreign investmentinflows in agricultural sector comes as mega-projects with huge capital inflows to acountry. These projects have limitedlinkages to domestic economy as theyusually use few local intermediate goods andare mostly export-oriented. Not only doforeign investment flows to this sector tendto be more volatile but also most investmentsare large and sensitive to world commodityprices. Foreign investment inflows inducesgrowth of some of the primary sector such aspetroleum, mining and agriculture whererevenue can be generated and fewpolicymakers feel that investment in theprimary sector is conducive to creatingemployment . It is seen from the results thatgrowth of the primary sector does not seemto influence foreign investment inflows andprimary sector investment fails in countrieswith no centralized political control orfundamental macroeconomic dislocation.

Foreign firms prefer to invest in themanufacturing sector, rather than export to acountry for either efficiency seeking ormarket seeking or a combination of both. Incase of efficiency seeking, foreigninvestment is more likely to bring in thetechnology and know-how that is compatiblewith the country. It usually generatessignificant employment and providestraining. Foreign firms usually use somelevel of local intermediate products. Hence,foreign investment has significant horizontaland backward linkages. In addition, foreigncompany exports increase the total exports asalso the foreign currency receipts of thecountry. Foreign investment inmanufacturing is expected to havesignificant impact in the recipient economy.Not only the growth of the secondary sector

causes foreign investment inflows to thecountry, but also the foreign investmentinflows enhances growth of the secondarysector. It seems that foreign investmentinflow in this sector will yield statisticallyidentifiable benefit. Transfer of technologyand management know-how, introduction ofnew processes and employee training tend torelate to the manufacturing sector rather thanthe primary sector. Policymakers tend tofocus on secondary sector investmentbelieving that manufacturing investment canbring jobs to absorb labor from lower-productivity sectors, an especially importantamong emerging markets where agriculturalemployment remains important.

Output is not tradable in the service sectorlike that of the agricultural andmanufacturing sectors; so much of theforeign investment in the form of FDI in theservices sector is market seeking whereforward linkages for foreign investment arewell defined and potential impact of foreigninvestment in the sector is immense. In thehighly capital-intensive infrastructure sector,foreign investment can provide the necessaryfunding and technology to improve capacityto meet increasing demand, as well asimprove the quality and lower the cost of theservices. Foreign investment in the bankingsector can also have an important impact onboth the efficiency and stability of thebanking system through increasedcompetition and increased access to globalfinancial markets. If foreign investment inthe sector improves services in the country,all other associated sectors will be positivelyaffected. India is yet to realize the benefits offoreign investment inflow in the servicessector and gradually more capital flows arehappening in this sector. The growth of thetertiary sector (services) influences foreigninvestment inflows and draws more foreign

49M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

capital to the country. However, foreigninvestment in the services sector may alsohave significant negative effects. As themarket structure is less competitive and thereis capital intensity of many industries in thissector, foreign investors command superiormarket power. So much of the benefit offoreign investment inflows in the tertiarysector is yet to be realized.

While openness did not appear to matterfor primary or secondary sector, it is veryimportant for foreign investment inflows inthe services sector when the real effectiveexchange rate is controlled. Apart fromopening up trade and investment thegovernments may need to set newcompetition rules. Strong, independentregulatory agencies and competition controlauthorities are required for performing thesetasks although such institutions may becostly to set up and run. In developingcountries technical assistance and capacitybuilding are important, as these institutions

require sophisticated skills, which brings agreat challenge in front of these countries.

It is also observed that growth in India’sindustrial sector growth and growth in theservices sector influence each other. India’smanufacturing sector growth to a large extentdepends on growth in the services sector suchas transportation and electricity. Moreover,India’s world-class knowledge-based servicesare also increasingly assuming importance forthe inputs they provide for other sectors of theIndian domestic economy (such asmanufacturing, retail and logistics) therebyimproving productivity in those sectors andfurther enhancing growth prospects. The role ofthe qualitative and institutional is alsoimportant. Liberalizing labor markets andmeasures to increase financial deepening couldattract more secondary foreign investmentinflows into emerging markets, though theseeffects are weaker among advanced economies.A more independent judiciary and betterinfrastructure appear to attract more servicesforeign investment inflows to both types ofeconomies.

50 M.Bhattacharya / SJM 8 (1) (2013) 39 - 51

ПРИЛИВ СТРАНИХ ИНВЕСТИЦИЈА И РАСТ ПО СЕКТОРИМА

У ИНДИЈИ - ЕМПИРИЈСКА СТУДИЈА

Mousumi Bhattacharya

Извод

Овај рад се бави одређивањем односа између прилива страних инвестиција и начина растапо секторима у три основна привредна сектора у Индији, током периода након либерализације(1996 до 2009). Одређена је дводирекциона повезаност између страних инвестиција исекторског раста, применом Грангеровог теста корелације у вишеваријантном окружењу.Примећено је да раст терцијалног сектора генерише прираст нових страних инвестиција.Такође је урађена регресиона анализа на истој полазној бази података, која је такође показалада раст терцијаног сектора доводи до пораста директних страних инвестиција након реформаиз 1990-тих.

Кључне речи:Прилив страних инвестиција, Гренгеров тест, Начин раста по секторима

References

Alfaro, L. (2003). Foreign DirectInvestment and Growth: Does the SectorMatter? Boston, MA, Harvard BusinessSchool, Mimeo.

Chakraborty, C., & Nunnenkamp, P.(2008). Economic Reforms, FDI andEconomic growth in India: A Sector LevelAnalysis. World Development, 36(7), 1192-1212.

Findlay, R. (1978). RelativeBackwardness, Direct Foreign Investment,and the Transfer of Technology: A SimpleDynamic Model. The Quarterly Journal ofEconomics, 92(1), 1-16.doi:10.2307/1885996

Hirschman, A.O. (1958). The Strategy ofEconomic Development. New Haven, Conn:Yale University Press.

Kathuria, V. (2002). Liberalisation, FDI,and productivity spillovers--an analysis ofIndian manufacturing firms. OxfordEconomic Papers, 54(4), 688-718.doi:10.1093/oep/54.4.688

Kokko, A. (1994). Technology, MarketCharacteristics and Spillovers. Journal ofDevelopment Economics, 43, 279-293.

Lo, C., & Liu, B. (2009). Why India ismainly engaged in offshore serviceactivities, while China is disproportionatelyengaged in manufacturing? China EconomicReview, 20, 236-245.

MacKinnon, J.G. (1991). Critical Valuesfor Co-integration Tests. Pp 267-276 in R.F.Engle and C.W.J. Granger (Eds.), Long RunEconomic Relationships: Readings inCointegration, Oxford: Oxford UniversityPress.

Rodriguez-Clare, A. (1996).Multinationals, Linkages and EconomicDevelopment. American Economic Review,86, 852-873.

United Nations Conference on Trade andDevelopment [UNCTAD] (2001). WorldInvestment Report 2001: PromotingLinkages, Geneva and New York:UNCTAD.

Wang, J.Y., & Blomstrom, M. (1992).Foreign Investment and TechnologyTransfer; A Simple Model. EuropeanEconomic Review, 36(1), 137-155.

51M.Bhattacharya / SJM 8 (1) (2013) 39 - 51