FDI and Outsourcing in a Service Industry: Welfare Effects ... · FDI in the service sector are more personal than FDI in the goods sector, such as nationality quotas for managers

FDI and Outsourcing in a Service Industry: WelfareEffects of Liberalising Trade and Investment*

TE-CHENG LU ANDYAN-SHU LIN

Department of Economics, National Dong HwaUniversity, Hualien, Taiwan

JIN-LI HU

Institute of Business and Management, NationalChiao Tung University, Hsinchu, Taiwan

This study examines a foreign firm’s entry decision and its effectson the host country’s welfare in a model with a composite good inwhich both commodity and service generate utility for consumers.Along with the commodity it produces, a producer can provide theservice by itself or outsource the service. The result shows that theincentive for foreign direct investment (FDI) in the service sectorincreases under liberalising trade in the final-good market. More-over, there exist policy combinations of trade and investmentliberalisation, whereby the domestic firms’ profitability is tradedoff with the host country’s social welfare when the foreignfirm provides a service through FDI or through outsourcing,respectively. Finally, the welfare after simultaneously liberalisingtrade and investment is not necessarily greater than that underautarky.

I IntroductionAccording to the World Trade Organization

(WTO), services have grown from 55 per cent oftotal GDP in the world to 70 per cent from 1977 to2007, with the added value of the service industrycontributing to 75 per cent of GDP in 2007 among

Organisation for Economic Co-operation andDevelopment (OECD) countries.1 Foreign directinvestment (FDI) and outsourcing are two majortypes of entry mode for a producer’s service,2 andservice outsourcing has been a widely usedstrategy for many enterprises in the global mar-ket. Service outsourcing includes business pro-cess outsourcing and knowledge processoutsourcing (Rajan & Srivastava, 2007). TheInternet has also turned service into a tradablegood instead of a non-tradable good.

*We are grateful to an editor and two anonymousreferees of this journal, seminar participants at theAustralian Conference of Economists in Perth in July2013, and Chao-Cheng Mai and Sheh-Jye Wu at theTaiwan Economic Association conference for theirvaluable comments, leading to substantial improve-ments of this study. Financial support from Taiwan’sNational Science Council (NSC-102-2410-H-259-003)is gratefully acknowledged. The usual disclaimerapplies.

JEL classifications: F13, F15, F21Correspondence: Yan-Shu Lin, Department of Eco-

nomics, National Dong Hwa University, No. 1, Sec. 2,Da Hsueh Rd., Shoufeng, Hualien 97401, Taiwan.Email: [email protected]

1 The data are from Francois and Hoekman (2010).2 GATS (General Agreement on Trade in Services)

classifies these four modes as follows: Mode 1, directcross-border trade in services (OS); Mode 2, movementof the customer to the country of the provider (Con-sumption Abroad); Mode 3, sales of services through anoffshore affiliate (FDI) and Mode 4, temporary entry ofnatural persons.

ECONOMIC RECORD, VOL. 90, SPECIAL ISSUE, JUNE, 2014, 74–86

74

© 2014 Economic Society of Australiadoi: 10.1111/1475-4932.12112

In addition to commodities, manufacturersnowadays also provide services called ‘producerservices’ (Markusen, 1989). Multinationalcompanies provide Internet, computer and tele-communications services to foreign customers byFDI or outsourcing. One Taiwanese Internet-related company, D-Link, provides networkservices worldwide to over 60 countries.Hewlett–Packard distributes its own brand ofpersonal computers through local sellers world-wide. In Taiwan, Nokia outsources its servicethrough Senao International, a local company. Toenter an overseas market, a multinational firmcan either directly invest in that market oroutsource producer services (Markusen et al.,2005).In the USA, FDI in the service industry is the

most important channel to serve overseas cus-tomers. Although FDI in the service industry isrestricted by discriminatory and non-discrimi-natory barriers, the amount of FDI stood at US$15 trillion in 2007 worldwide, of which 60 percent was contributed by FDI in the serviceindustry. Discriminatory and non-discriminatoryentry barriers include limits on equity owner-ship shares and limits on the number ofmultinational firms.3 Whereas the USA, UKand many developed economies are usuallyservice outsourcers, developing economies usu-ally play the role of service providers (Amiti &Wei, 2005).In 2006, almost one half of all services relates

to business and ICT (Internet, computer andtelecommunications).4 Distribution contributesto nearly one-fifth of all services. This studytherefore focuses on these major types (business,ICT and distribution) of services (i.e. producerservices) and ignores finance and insuranceservices.5

A large share of the empirical literaturefocuses on a firm’s productivity and profitability.Jabbour (2010), for example, finds that outsourc-ing firms are more efficient than vertical inte-grated (FDI) firms because outsourcing firmsmight concentrate on core activities and reduce

organisational cost. Tomiura et al. (2011) con-clude that outsourcing (vertical integrated) firmsprovide incentives to supply a labour- (capital-)intensive good by sharing labour (investment)cost.Markusen (1997, 2002) notes that liberalising

trade and investment at the same time is verydifferent from applying each policy separately.These two types of liberalisation complementa-rily affect social welfare. Markusen et al. (2005)find that liberalisation of inward FDI in producerservices will increase the domestic total-factorproductivity and real wage rate of domesticskilled labour. Egger et al. (2004, 2007) use athree-country knowledge–capital model of tradeto check the effect of liberalising trade andinvestment in welfare, noting that trade andinvestment liberalisation are not substitutiveunder vertical FDI. After completely liberalisedtrade, both similar endowment countries gainfrom liberalising investment. Following theMarkusen (1997, 2002) model, Amiti and Wakelin(2003) verify that liberalising investmentincreases (decreases) exports when trade costsare low (moderate to high). Srivastava (2006)presents that India has attracted a lot of FDI in theICT services industry and shows a positive (orcomplementary) relationship between FDI inservices and service exports when trade is liber-alised. In contrast, our research focuses on FDI inservices and exports in the final good through atheoretical model.There is still not much theoretical literature on

trade and investment liberalisation in the pro-ducer service industry. To the best of ourknowledge, the literature most closely relatedis as follows. Ishikawa et al. (2010) investigatethe connection between FDI in services andprice competition in the final good. Tradeliberalisation in the final good may decrease itssocial desirability if it is not combined withliberalisation of FDI in services. Mukherjee andSuetrong (2012) devise a model of home countryexport platform FDI to find a negative relationbetween trade cost and FDI. Collie (2011)discovers a ‘regional tariff jumping’ model thatcan explain the increases in both the amount ofFDI and the volume of world trade, under whichthese increases are motivated by multilateraltrade liberalisation. Beladi et al. (2009) showthat trade liberalisation increases the incentivesof FDI if the foreign firm undertakes jointventures. In addition, Horiuchi and Ishikawa(2009) find that a tariff reduction on the final

3 Further details on trade and policy for services canbe found in Francois and Hoekman (2010).

4 Abramovsky and Griffith (2006) find that ICT-intensive firms have more incentives to purchaseoffshore services than less ICT-intensive firms.

5 The data are from IMF and OECD.

© 2014 Economic Society of Australia

2014 FDI OUTSOURCING SERVICE INDUSTRY 75

good attracts FDI in the presence of a North–South technology transfer. Neary (2009)discusses the relation between vertical (andhorizontal) FDI and trade costs.6

Trade liberalisation has received much atten-tion in the past few decades, but not manytheoretical studies have combined trade andinvestment liberalisation at the same time.7

Ederington and McCalman (2010) use a politicaleconomy model and find that the median voterprefers a simultaneous liberalisation of trade andinvestment rather than liberalising trade followedby liberalising investment. The recent relatedliterature still focuses on the manufacturingindustry and ignores the relation between domes-tic firms’ profitability and social welfare. As aresult, this study aims to study the welfare effectsof liberalising both trade and investment in theservice industry.Most barriers to FDI are not in goods, but in

service sectors (Arnold et al., 2011). Severalaspects of service liberalisation are discussed,such as the presence of foreign providers,privatisation and the level of competition.Results from the Czech Republic show apositive relation between service sector reformand the performance of domestic firms inmanufacturing sectors. A number of studies

trace the issue of FDI policy and find that thehost country prefers subsidising inward FDIfrom different aspects. Skaksen (2005), forinstance, recommends that the WTO should notoppose host government subsidies to multina-tionals that produce complementary goods in thehost country.There are some studies that recommend that a

FDI policy should be more flexible and variedunder different situations. Demekas et al. (2007)suggest that tax holidays and domestic corrup-tion do not seem to have a statistically signif-icant impact on FDI. Qiu and Wang (2011) findthat the domestic country adopts flexible FDIpolicy to maximise social welfare throughchanging the foreign firm’s entry decision.Kayalica and Lahiri (2007) discover that domes-tic firms lobby for lower (higher) subsidies whena discriminatory (uniform) subsidy on FDI isapplied. Chang (2004) argues that the benefits ofliberalising foreign investment outweigh thecosts when the domestic industry is complexand competitive. A range of instruments candevelop the industry by limiting ownership,transferring technology or local procurement,focusing on joint ventures with local firms andrestricting brownfield investments through merg-ers and acquisitions. Although regulations onFDI in the service sector are more personal thanFDI in the goods sector, such as nationalityquotas for managers of affiliates, this kind ofdiscriminatory barrier eventually affects invest-ment cost and entry decision. This study high-lights the importance of timing on liberalisingtrade and investment to maximise a domesticcountry’s welfare.This study novelly explores the importance of a

foreign firm’s entry selection and suggests asignificant policy implication of liberalising bothtrade and investment in the service industry.Comparing with the previous literature, we dis-cover that liberalising trade in the final goodmarket induces a foreign firm to undertake FDI inthe service industry for ‘service fee jumping’,which is in parallel to Mukherjee and Suetrong(2012), Collie (2011), Beladi et al. (2009) andHoriuchi and Ishikawa (2009). Furthermore, thedomestic firms’ profitability is not consistent withthe host country’s social desirability in somepolicy combinations of liberalising trade andinvestment when the foreign firm provides aservice by FDI or outsourcing respectively.Finally, the welfare after simultaneously liberal-ising trade and investment is not necessarily

6 The following articles are in regards to FDI oroutsourcing: Ishikawa and Horiuchi (2012) use aNorth–South trade model to discuss strategic FDI invertically related markets. Lin and Saggi (2011) discussthe policy competition for FDI and the welfare effect ina vertical-related industry. Arya et al. (2008) compareCournot competition and Bertrand competition when anupstream firm outsources an intermediate good to avertically integrated rival. Buehler and Haucap (2006)find that outsourcing softens competition in the finalgood market through colluding or raising rivals’ costs iffirms foresee the effect of outsourcing prices. Shy andStenbacka (2003) demonstrate how the competitionamong input suppliers makes components not onlyavailable but also firms take advantage of economies ofscale. Grossman and Helpman (2005) find offshoreoutsourcing is determined by the thickness of the inputsuppliers’ markets and the nature of the two parties’contracting environment. Aronsson and Koskela (2011)study optimal income taxation under FDI and outsourc-ing. Grossman and Helpman (2003) consider how theextent of contractual incompleteness and the relativewage rate affect FDI and outsourcing.

7 For instance, trade liberalisation has been discussedin Chen et al. (2004) and Fujiwara (2011).


76 ECONOMIC RECORD JUNE

improved, which is in contrast to Ishikawa et al.(2010).The remainder of the study is organised as

follows. Section II provides the basic model.Section III discusses the welfare effect andpolicy implications of the host country. Sec-tion IV extends the basic model. Finally, Sec-tion V concludes this article.

II The ModelFirm h and firm f both produce a homogeneous

final good X, located in the home country andforeign country respectively. The final goodrequires services before it can be sold in thehome market, and the services must be providedby the home market – that is, the consumersconsume composite goods consisting of bothgoods and services. To enter the home market,firm f exports the final good to the home countryand incurs a specific tariff t. Assume that oneunit of the service is required for one unit of thefinal good. The two firms’ production costs arezero without loss of generality. Because firm fwants to provide a service in the home country,firm f can choose either FDI or outsourcing asthe entry mode. If firm f chooses FDI for theservice, then it incurs a fixed investment cost Fand a unit service cost c. If firm f choosesoutsourcing, then it incurs a service price r thatis charged by an independent service firm s. Firms is located in the home country. The service costc is identical, as firms h, f and s use the sameresource to provide the same service in the homecountry.This study constructs a Cournot model and

rationalises the timing of the game. The timing ofthis three-stage game model is as follows. In thefirst stage, firm f decides on FDI or outsourcing(OS). In the second stage, firm s decides on theservice price r if firm f undertakes OS. In the thirdstage, firms h and f compete over quantity in thefinal good market.8

We present the inverse linear demand functionin the home country as p(xh + xf), where p is theprice of good X = xh + xf and the derivative ofinverse demand p0 < 0, p″ = 0, and xh and xf

denote the production of firm h and firm frespectively.

(i) Cournot Competition Under FDIIn the last stage of the FDI regime, firm h and

firm f maximise their profits respectively.

Maxxh

ph ¼ ðp� cÞxh; ð1Þ

Maxxf

pf ¼ ðp� c� tÞxf � F: ð2Þ

We derive the first-order conditions due toprofit maximisation by both firms.

phh �@ph

@xh¼ pþ p0xh � c ¼ 0; ð3Þ

piii � @2pi=@x2i ¼ 2p0\0; i ¼ h; f ð4Þ

Denote the equilibrium outputs in the FDIsubgame as xi(c, t), i = h, f. The second-orderconditions require piii � @2pi=@x2i ¼ 2p0\0,i = h, f. The following stability conditionsare satisfied and there exists a unique equilib-rium:

piij � @2pi=@xi@xj ¼ p0\0; i; j ¼ h; f ; i 6¼ j; and

D � phhhpfff � phhfp

ffh ¼ 3ðp0Þ2 [ 0:

We have the following comparative staticeffects:

@xh@t

¼ �phhfD

¼ �p0

D[ 0;

@xf@t

¼ phhhD

¼ 2p0

D\0: ð5Þ

Needless to say, these results are very stan-dard.

(ii) The Outsourcing RegimeUnder outsourcing, the profit function and

first-order condition of firm h are identical toEquations (1) and (3). The foreign firm’s profitfunction and first-order condition are respec-tively:

Maxx̂f

p̂ f ¼ ðp̂� r � tÞx̂f ð6Þ8 It is more reasonable that a foreign firm makes the

entry decision, and then a service firm adjusts its price.However, Ishikawa et al. (2010) reverse the order ofthe above timing of the game.



@p̂ f

@x̂f¼ p̂þ p0x̂f � r � t ¼ 0: ð7Þ

Denote the equilibrium outputs in the outsourcingsubgame as x̂iðr; c; tÞ; i ¼ h; f . By solving theseequations, we have:

@x̂h@r

¼ @x̂h@t

¼ �p̂hhfD̂

¼ �p0

D̂[ 0; ð8Þ

@x̂f@r

¼ @x̂f@t

¼ p̂hhhD̂

¼ 2p0

D̂\0:

These comparative static results imply that tradeliberalisation in the final good and/or a decreasein the service price decreases x̂hðr; c; tÞ, butincreases x̂f ðr; c; tÞ.In the second stage of the OS regime, firm s

decides the service price to maximise profits.

Maxr

p̂s ¼ ðr � cÞx̂f ðr; c; tÞ: ð9ÞThe first-order condition of firm s is as follows:

p̂sr �@p̂s

@r¼ ðr � cÞ @x̂f

@rþ x̂f ¼ 0: ð10Þ

The equilibrium service price r(c, t) is a functionof c and t. We therefore obtain the followinglemma.

Lemma 1 A decrease in the tariff rate tincreases the service price r.

Proof The second-order condition requiresp̂srr ¼ 2@x̂f =@r\0. Totally differentiatingEquation (10), we obtain dr=dt ¼ �p̂srt=p̂

srr,

where p̂srt ¼ @x̂f =@t\0, and therefore dr/dt = � 1/2.9

The logic behind Lemma 1 is as follows. Bothan increase in service price r and a decrease in thetariff t raise the profit margin of independentservice firm s. Intuitively, trade liberalisation inthe final good induces an increase in the deriveddemand for firm s, and therefore firm s maximisesprofits by raising service price r.

(iii) Firm f Undertakes FDI or OSIn the first stage, firm f undertakes FDI (OS) if

pf � p̂f [ ð\Þ0. In other words, firm f undertakesFDI if F\F � ðp� c� tÞxf � ðp̂� r � tÞx̂f . Totallydifferentiating the critical value F, we obtain:

d �F

dt¼ dp f

dt� dp̂ f

dt

¼ ½ @p f

@xh

@xh@t|fflfflfflffl{zfflfflfflffl}

indirectðFDIÞ

þ @p f

@t|{z}directðFDIÞ

�

� ½ @p̂f

@x̂h

@x̂h@t|fflfflfflffl{zfflfflfflffl}

indirectðOSÞ

þ @p̂ f

@t|{z}directðOSÞ

�

� ½ð@p̂f

@x̂h

@x̂h@r

þ @p̂ f

@rÞ drdt|fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl}

outsourcingcost

�: ð11Þ

This study highlights the above findings asLemma 2.

Lemma 2 The incentive for FDI in the serviceindustry increases when liberalising trade inthe final good market.

Proof By using Equations (5) and (8), it is

easy to derive ð@p f =@xhÞð@xh=@tÞ þ ð@p f =@tÞ ¼�4xf =3 and ð@p̂ f =@x̂hÞð@x̂h=@tÞ þ ð@p̂ f =@tÞ ¼�4x̂f =3. Here, xf [ x̂f due to r > c, and thesum of the first and second terms is negative.Together with the outsourcing cost effect beingpositive in Equation (11), we verify that Equa-tion (11) is negative.

Trade liberalisation in the final good sectoraffects the entry decision of firm f through thefollowing effects: direct effect and indirect effectunder FDI (OS) and outsourcing cost effect, whichcan be, respectively, explained by Equation (11).We know that both the direct and indirect effectsare negative under FDI or OS, and the magnitudeof the negative effect under FDI dominates thelatter one (OS). Therefore, the sum of the first andsecond terms is also negative. Taking into accountthese direct and indirect effects, it implies thattrade liberalisation makes firm f more willing toundertake FDI. By using Lemma 1, the positiveoutsourcing cost effect also increases the incen-tives for firm f to undertake FDI as a reduction in traises r, which in turns increases the outsourcingcost of firm f. Combining these three effects, wenotice a negative relation between the tariff rate tand the incentive for FDI in the service (F).

9 Throughout the article, we demonstrate the resultsin a linear demand function. The Appendix discussesthe robustness of our lemmas and propositions undergeneral demand.



We analyse the entry decision of firm f from theviewpoint of cost. The effective marginal cost offirm f in the FDI regime is (c, t), and in the OSregime it is (r, t). Because the service price isalways greater than the service cost, the operatingprofit of firm f in the FDI regime is bigger than thatin the OS regime. As the effective marginal cost islow under FDI, why does firm f undertake OS whenthe trade barrier is high? The reason is that firm ftrades off the loss of sunken investment cost Fwith the benefit of a low effective marginal cost inthe FDI regime. The loss of sunken investmentcost F outweighs the benefit of a low effectivemarginal cost in the FDI regime when the tradebarrier is high, which is why firm f undertakes OS.By contrast, the benefit of a low effective marginalcost outweighs the loss of a sunken investmentcost in the FDI regime when the trade barrierdrops. Therefore, firm f undertakes FDI. As aresult, the incentive for FDI in the service industryincreases when liberalising trade in the final good.Figure 1 (by assuming linear demand

p = a � xh � xf) show the intuition behind theforeign firm’s decision. We know that firm fcannot earn any positive profits at prohibitivetariff �t (at point A, Figure 1a) in which theequilibrium outputs xf (c, t) equal zero. When theinvestment cost F is zero, the profits under FDIare always greater than those under OS. Further-more, the curve of profits under FDI shiftsdownward and intersects the curve of profitsunder OS at point B when the investment costbecomes larger, say F1. Given F = F1, firm fprefers OS to FDI if t1 � t\�t, whereas firm f

prefers FDI to OS if 0 < t < t1. The foreign firm’sentry decision can be depicted as the ABF boldcurve in Figure 1a. Similarly, the foreign firm’sentry decision can be illustrated by the ACEcurve in the situation of F = F2.This study suggests that the foreign firm’s

decision features a negative relation between Fand t, as shown as Figure 1b. The incentive for FDIin the service �F can be characterised by the ADcurve. Given F = F1, firm f is indifferent betweenchoosing FDI or OS at point B when t = t1. It isanalogous that point C makes firm f also indifferentbetween undertaking FDI or OS in the situation ofF = F2 and t = t2. As a result, firm f is indifferentbetween FDI and OS on the locus of the AD curve.If the pairs of (F, t) are below (above) the ADcurve, then firm f prefers FDI (OS).We conclude the foreign firm’s entry decision as

follows. If firm f does not incur any investment cost,then firm f definitely undertakes FDI as the effectivemarginal cost under FDI is small. When F isrelatively small, the benefit of trade liberalisationeasily dominates the sunken investment cost.WhenF is relatively large, the sunken investment cost isdifficult to overcome from the benefit of tradeliberalisation. Therefore, firm f trades off the lossof a sunken investment cost with the benefit of a loweffective marginal cost under FDI.

III The Domestic Country’s WelfareThis section considers the welfare of the domes-

tic country. We analyse the profitability of thedomestic firms before the analysis of domesticcountry’s desirability. Similar to the definition of

FIGURE 1(a) Foreign Firm’s Decision. (b) Foreign Firm’s Decision Under Different Values of F and t



Chaudhuri and Benchekroun (2012), trade liber-alisation is said to increase (decrease) the desir-ability of the foreign firm’s entry at t = t0 if andonly if @W=@tjt¼t0

\ð[ Þ0. The analysis of profit-ability is analogous to the analysis of desirability.

(i) Profitability of the Domestic FirmsThe equilibrium profit of firm h under FDI

decreases with trade liberalisation (GH curve inFigure 2). In other words, a reduction in tdecreases the profitability of firm h if firm fundertakes FDI. By contrast, there are two firms, hand s, in the OS regime. Accordingly, a tariffreduction decreases the profitability of firm h (GIcurve), but increases the profitability of firm s (ADcurve). Using the envelope theorem, we have:

dðp̂h þ p̂sÞdt

¼ @p̂h

@x̂fð@x̂f@t

þ @x̂f@r

dr

dt|fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl}strategic effect

Þ

þ ½@p̂s

@x̂f

@x̂f@t

ð1þ dr

dtÞ|fflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflffl}

quantity effect

þ x̂fdr

dt|ffl{zffl}price effect

�ð12Þ

Substituting Equation (8) and Lemma 1 intoEquation (12) shows dðp̂h þ p̂sÞ�dt ¼ x̂h=3� x̂f ,and d2ðp̂hþ p̂sÞ�dt2¼ð1=3Þðdx̂h=dtÞ�dx̂f

�dt[0:

From Equation (12), the profits of domestic firmsare minimised at t = tP (GJ curve). Therefore, thisstudy suggests that the domestic firms’ profitabil-ity is as follows.

Proposition 1 (i) Trade liberalisationdecreases the total profits of the domestic firmsif firm f undertakes FDI.

(ii) For tp\t\�t ð0\t\tpÞ, trade liberalisationdecreases (increases) the total profits of thedomestic firms if firm f undertakes OS.

Figure 2 shows the results. From the viewpointof firm h, liberalising trade makes firm h lessprofitable as trade liberalisation induces firm f toproduce more. However, firm s can charge ahigher service price, which benefits firm h, as ahigher service price induces firm f to produce less.The former direct effect dominates the latterindirect service–price effect, which is why tradeliberalisation decreases the profitability of firm h,

FIGURE 2Profitability of the Domestic Firms



which is summed up as the strategic effectin Equation (12). From the viewpoint of firm s,the direct and indirect effects are opposite, sum-ming up as the quantity effect. Except for thequantity effect, firm s directly benefits from a highservice price, the price effect, which is induced byliberalising trade. Considering the quantity andprice effects, trade liberalisation increases theprofitability of firm s. As firm h and firm s areinvolved in the OS regime, trade liberalisationdecreases the total profits of the domestic firms if tis relatively large (tp\t\�t ) and increases the totalprofits of the domestic firms if t is relatively small(0 < t < tp).

(ii) Trade Liberalising Effect on ConsumerSurplus Under OSConventional wisdom tells us that trade

liberalisation is beneficial for consumer surpluswithout OS. We check this viewpoint in theoutsourcing regime by:

dðx̂h þ x̂f Þdt

¼ @x̂h@r

dr

dtþ @x̂h

@t

� �þ @x̂f

@r

dr

dtþ @x̂f

@t

� �

¼ 1þ dr

dt

� �p0

D̂

� �:

ð13Þ

By using Lemma 1, we know that a decrease inthe tariff rate t increases consumer surplus.

(iii) Welfare Under Autarky, FDI and OSFirst, we take the welfare of the domestic

country under autarky (superscript A) as abenchmark case:

WA ¼ CSA þ phA; ð14Þwhere CS denotes consumer surplus.Second, the welfare of the domestic country

under FDI is as follows:

WFDIh ¼ CSþ ph þ txf ð15Þ

where txf is the tariff revenue. We derive the first-order condition as:

@WFDIh

@t¼ �p0ð@xh

@tþ @xf

@tÞxf|fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl}

consumer surplus effect

þðp� cÞ @xh@t|fflfflfflfflfflfflffl{zfflfflfflfflfflfflffl}

profit effect

þ ðxf þ t@xf@t

Þ|fflfflfflfflfflfflffl{zfflfflfflfflfflfflffl}tariff revenue effect

¼ 0:

ð16Þ

Substituting Equation (5) into Equation (16)

shows @WFDIh

�@t ¼ ð2x f þ xh � 2tÞ�3, and

@2WFDIh

�@t2 ¼ ½2ð@x f

�@tÞ þ ð@xh

�@tÞ� 2��3\0.

Therefore, the welfare under FDI is maximisedat t = t1. Figure 3 shows the relation betweenthe domestic country’s welfare and t. Theconsumer surplus effect in Equation (16) impliesthat trade liberalisation is beneficial for con-sumer surplus. The profit effect points towardstrade liberalisation being harmful to the profitsof firm h. The tariff revenue effect is ambiguous.Third, the welfare of the domestic country

under OS is as follows:

WOSh ¼ CS

^þp̂h þ p̂s þ tx̂f : ð17Þ

The first-order condition is analogous to Equa-tion (16) and goes as follows:

@WOSh

@t¼ �p0ðdx̂h

dtþdx̂f

dtÞx̂f|fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl}

consumer surplus effect

þðp̂�cÞdx̂hdt|fflfflfflfflfflffl{zfflfflfflfflfflffl}

profit effect

þ½ drdtx̂f þðr�cÞdx̂f

dt|fflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflffl}outsourcing profit effect

�þ ½x̂f þ tdx̂fdt

�|fflfflfflfflfflffl{zfflfflfflfflfflffl}tariff revenue effect

¼0:

ð18ÞSubstituting Equation (8) and Lemma 1 into

Equation (18) shows @WOSh

�@t ¼ ð�x̂ f þ x̂ h

�2tÞ�6, and @2WOSh

�@t2 ¼ ð�@x̂ f

�@t þ @x̂ h

�@t�

2Þ�6\0. The welfare under OS is maximisedat t = t2. The effects in Equation (18) are anal-ogous to Equation (16). Note that trade liberali-sation is beneficial for the profits of firm s,which is represented by the outsourcing profiteffect.

(iv) Discussion and Policy ImplicationNow let us consider how the domestic coun-

try’s welfare is affected by the foreign firm’sentry decision. Given F equals F1 in Figure 1b,firm f prefers OS when the tariff drops from �t to t1and the domestic country gets its welfare underoutsourcing, which is AB’ in Figure 3. On theother side, firm f prefers FDI when the tariff dropsfrom t1 to zero and the domestic country gets itswelfare under FDI, which is BD in Figure 3. As aresult, the domestic country’s welfare can bedepicted as the bold curve AB’BD. Similarly,given F equals F2, the domestic country’s welfarecan be presented as curve ACC’D.



After applying Equations (16) and (18), it isnow straightforward to show that the respectiveslopes of curves AB and AB’ near point t ¼ �t are@WFDI

h

�@t��t¼t

and @WOSh

�@t��t¼t

, which present thatthe welfare of the domestic country increasesquicker in the FDI regime than that in the OSregime when liberalising trade near point A.Recalling Lemma 1, we note that both consumersurplus and tariff revenue increase quicker, aswell as producer surplus decreases quicker in theFDI regime. However, the effect on consumersurplus and tariff revenue dominates theeffect on producer surplus. Therefore, the welfareof the domestic country increases quicker inthe FDI regime than that in the OS regime nearpoint A.Subtracting Equation (17) from Equation (15)

yields:[ 0 t[ tw

WFDIh �WOS

h ¼ 0 if t ¼ tw\0 t\tw:

ð19Þ

Given all the possible values of t (0\t\�t), weexplore the welfare discrepancy between FDI andOS by the following effects: consumer surplus,profits of firm h (and s) and tariff revenue. First,consumer surplus under FDI is definitely largerthan that under OS due to severe competitionunder FDI. Second, domestic firms’ profits underFDI are definitely smaller than those under OS(see Figure 2). Third, the tariff revenue effect isambiguous. In the beginning (end) of trade

liberalisation, tariff revenue under FDI is larger(smaller) than that under OS. Eventually, thepositive effects induced by trade liberalisationwill be balanced by the negative effects at pointt = tw, shown at point W in Figure 3.Figure 4 illustrates the relation between the

foreign firm’s decision and the welfare of thedomestic country by considering different valuesof t and F. Because Equations (15) and (17) are nota function of F, the domestic country’s welfare isindifferent between FDI and OS at the vertical lineof t = tw, and it is shown as a vertical line inFigure 4. First, firm f undertakes FDI (OS) beneath(above) the AD curve as the investment cost isrelatively small (large). Second, we knowWFDI

h [ ð\ÞWOSh if t[ ð\Þtw from Equation (19).

Furthermore, the social welfare that the domesticcountry can get depends on the entry mode of firmf. As a result, the entry mode chosen by firm f cangenerate the welfare that domestic governmentpursues in regions O1 to O5.10 In contrast, thewelfare cannot be maximised in regions N1 andN2. Our discussions focus on the reasonableregions of O1 to O5.In the beginning of trade liberalisation,

t1\t\�t, firm f undertakes FDI within region O1,and liberalising trade decreases the profitability,but increases the desirability. In region O2,tw < t < t1, firm f also undertakes FDI, andliberalising trade decreases both the profitabilityand the desirability. In region O3 (O4) t2 < t < tw(tp < t < t2), firm f undertakes OS, and tradeliberalisation decreases the profitability, butincreases the desirability (decreases both theprofitability and the desirability). In the neigh-bourhood of free trade, region O5, 0 < t < tp,foreign firm f also undertakes OS, and tradeliberalisation increases the profitability, butdecreases the desirability. The effect of tradeliberalisation on profitability and desirability ineach region can be represented as follows:Regions O1 (�, +), O2 (�, �), O3 (�, +), O4(�, �) and O5 (+, �). There are conflictsbetween social desirability and the firms’profitability in regions O1, O3 and O5. Summa-rising the above results of the profitability andthe desirability, we obtain the followingproposition.

FIGURE 3Relation Between the Domestic Country’s Welfare and t

10 The definitions of regions O1 to O4 go according tothe welfare functions’ slope in Figure 3, and region O5 isdefined by the slope of domestic firms’ profit in Figure 2.



Proposition 2 (i) For t1\t\�t (t2 < t < tw),liberalising trade decreases the profitability,but increases the desirability in the FDI (OS)regime.

(ii) For tw < t < t1 (tp < t < t2), liberalisingtrade decreases both the profitability and thedesirability in the FDI (OS) regime.

(iii) For 0 < t < tp, trade liberalisationincreases the profitability, but decreases thedesirability in the OS regime.

We further derive policy implications of liberal-ising both trade and investment in the serviceindustry from Proposition 2. The domestic govern-ment can hence use different policy combinations(e.g. both trade and FDI liberalisation or tradeliberalisation with a strict FDI policy) to induce theforeign firm to adopt FDI or outsourcing to improvethe domestic welfare. For example, in Figure 4 ifthe current combination of tariff rate t and FDI costF lies in region N1, then the optimal entry mode forthe foreign firm is outsourcing. In this case thegovernment may use FDI liberalisation to inducethe outcome to be in region O1, while further usingtrade liberalisation to improve the domestic wel-fare. This complementary policy combinationencourages firm f to replace outsourcing with FDI,leading to larger social welfare, while being harm-ful to the domestic firm h. On the contrary, inFigure 4 if the current combination of tariff rate tand FDI cost F lies in region N2, then the optimal

entry mode for the foreign firm is FDI. In this casethe government can apply a stricter investmentpolicy (such as a more rigorous environmentalstandard) to induce the outcome to be in region O3,while further adopting trade liberalisation that willimprove the domestic welfare. This substitutivepolicy combination encouragesfirm f to replaceFDIwith outsourcing, leading to a higher level ofdomestic welfare. Our results are in contrast toIshikawa et al. (2010), where complementary pol-icies lead to domestic welfare improvement. Wefind that a stricter FDI and a looser trade policycombination may be welfare improving when theproducer service is taken into account. These resultssuggest important policy implications as follows.

Proposition 3 For tw\t\�t, the governmentmay use FDI liberalisation, while further usingtrade liberalisation, to improve the domesticwelfare. For 0 < t < tw, the government canapply a stricter FDI policy, while furtheradopting trade liberalisation, to improve thedomestic welfare.

IV Model Extension: Firm f Outsources a Serviceto Firm h

In previous sections, the model assumes thatthe foreign firm f outsources its service to a pureservice provider, firm s. However, firm f can alsooutsource its producer service to the final goodcompetitor, the domestic firm h. The model set-upis the same as that in Section II, except that firm hprovides the service for firm f and also determinesthe price of the outsourced service. In this casethe domestic pure service provider s is notinvolved in this market. As a result, firm h’sprofit function becomes ~ph ¼ ðp� cÞ~xhþ ðr � cÞ~xf ,where the superscript ~ indicates the situationwhen firm f outsources the service to firm h.The solution of quantity competition in the third

stage of the game is the same as that in Section II.In stage 2, firm h maximises its profit by choosingthe price of the outsourced service under theincentive constraints of firm f in the regime of OSinstead of FDI: ~pf ðr; c; tÞ � pf ðc; t;FÞ� 0.11 The

FIGURE 4Foreign Firm’s Decision and the Welfare of the

Domestic Country

11 Firm h maximizes the following objective functionin the second stage:Max~p h ¼ðp� cÞ~xhðr; c; tÞþðr � cÞ~x f ðr; c; tÞ s:t ~p f ðr; c; tÞ�p f ðc; t;FÞ� 0. The incentive constraints of firm f arebounded at �r � ~pf ðr; c; tÞ � pf ðc; t;FÞ ¼ 0. By assuminglinear demand p = a � xh � xf, we have

�r ¼ ½ðaþ c� 2tÞ � ðð2t � aþ cÞ2 � 9FÞ1=2��2, with

@�r=@t [ 0.



optimal price for firm h on the outsourced serviceis r, where r is an increasing function in t.Intuitively, without taking into account the incen-tive constraint for firm f, firm h will charge anextremely high price for the outsourced service, todrive firm f out of the market and to become adomestic monopoly in the final good market.However, as long as the outsourced service’s priceis high enough, firm f will provide its own servicevia FDI and hence make the domestic finalgood market structure a duopoly. As a result, theoptimal outsourced service price of firm h is tomake firm f feel indifferent between OS andFDI.This finding differs from that in Lemma 1.

The economic intuitions are as follows. Whenthe firm service provider is firm s, its profit-maximising price always maintains a positivequantity of firm f. Therefore, when a drop in tincreases firm f’s exports, an increase in thederived demand for the service firm s willincrease the outsourced service price, resultingin a negative relation between the tariff rateand service price. However, as firms h and fcompete in the final good market, firm h has totake into account the incentive constraint forfirm f to keep using the OS strategy. Therefore,when firm h provides the service to firm f, theequality of firm f’s incentive constraintsbetween OS and FDI should always hold.When t decreases, assuming the service pricestays the same, firm f’s profit always increasesunder both OS and FDI strategies; moreover,the increase in firm f’s profit under FDI islarger than that under OS. In order for firm fto still have an incentive to outsource theservice, firm h has to reduce the service priceas the tariff rate decreases, creating a positiverelation between the tariff rate and serviceprice.Summarising the above discussion, we find that

trade liberalisation in the final good market alwayslowers the domestic firm’s profit when the foreignfirm outsources the service to it. The results aredifferent from those in Proposition 1. However,there is an inverse U-shape relation between thetariff rate and domestic welfare. As a result, in thissituation there is a trade-off between the domesticfirm’s profit and social welfare with respect totrade liberalisation, as Proposition 2 shows. Aslong as the FDI cost is positive, the domestic firmh always charges a service price to induce firm f tooutsource the service and not to conduct FDI. As aconsequence, there is no policy substitution

between trade and FDI liberalisation, as Proposi-tion 3 discusses.12

V ConclusionThe composite good (with both final good and

producer service) framework presented in thisstudy highlights the importance of a foreignfirm’s entry selection and suggests a strongpolicy implication for liberalising both tradeand investment in the service industry. Collie(2011) explains that trade liberalisation induces aforeign firm to undertake FDI when the inter-regional transport cost is sufficiently large.Ishikawa et al. (2010) discover that trade liber-alisation in the final good may decrease socialdesirability if not combined with liberalisation ofFDI in services. In our study we find thatliberalising trade in the final good market inducesa foreign firm to undertake FDI in the serviceindustry. However, the domestic firms’ profit-ability is not consistent with the host country’ssocial desirability in some policy combinations ofliberalising trade and investment when the for-eign firm provides a service by FDI or outsourc-ing respectively. Finally, the welfare aftersimultaneously liberalising trade and investmentis not necessarily greater than that under autarky.Our research results show that there are many

welfare trade-offs between liberalising trade andinvestment. Policy-makers therefore have to bal-ance the welfare gains and losses between liberal-ising trade and investment at the same time. Forexample, Thailand and Philippines offer tax incen-tives to compete for inward FDI, whereas Vietnamoffers a 10-year tax holiday to win over thecompetition (Bjorvatn & Eckel, 2006).Australia’s import tariff on motor vehicles hasbeen reduced from 57.5 to 17.5 per cent during thepast few decades, but the FDI inflow has increasedduring this time (Lloyd, 2008; Sadleir & Mahony,2009; and Crotti et al., 2010). Taiwan signed anFTA with New Zealand to drive trade liberalisa-tion, while also offering tax incentives to attractFDI (Claus, 2009). These policies should becautiously checked under different situations.Future work will endogenise the choice of out-sourcing services from a domestic firm or anindependent service firm. Policy coordination

12 If the domestic government can use subsidies toattract FDI, then firm f may still choose FDI in thissituation. Future research may study this case.



among governments on both FDI and serviceoutsourcing is also worth more in-depth research.

Appendix

Recall Lemma 1, we know dr=dt ¼ �p̂srt=p̂srr,

where p̂srr\0, and

p̂srt ¼ ðr � cÞ @2x̂f

@r@tþ @x̂f

@t: ðA1Þ

From Equation (A1), we derive

@2x̂f@r@t

¼�8ðp0Þ3p00 � ðp00Þ2½1þ 7ðp0Þ2�x̂h � p0p00

�p0ðp00Þ3ðx̂h þ x̂f Þx̂hD̂

: ðA2Þ

Equation (A2) is negative if p″ < 0. In summary,Lemma 1 holds if p″ < 0, p″ = 0 and p″ > 0 is notsufficiently large.Recall Lemma 2, this study verifies

d �F

dt¼ ½x̂f � xf � þ3ðp0Þ3½ðp

0 þp00x̂hÞx̂f �ðp0 þp00xhÞxfDD̂

�

þ ðp0Þ2p00½ðp0 þp00x̂hÞx̂f X�ðp0 þp00xhÞxf X̂

DD̂�

þ ½p0ðp0 þp00x̂hÞx̂f

D̂þ x̂f �dr

dt:

ðA3Þ

Four terms in Equation (A3) are all negative if dr/dt < 0. Therefore, if dr/dt < 0, Lemma 2 holds.The robustness of propositions 1 and 2 are

analysed as follows:

dðp̂h þ p̂sÞdt

¼

2ðp0Þ2x̂h þ p0p00xhx̂h þ 2p0ðr � cÞ þ p00ðr � cÞxhD̂

" #

ð1þ dr

dtÞ þ x̂f

dr

dt

ðA4Þ

@WFDIh

@t¼

�ðp0Þ2xf �p0ðp�cÞ�p00ðp�cÞxhþ2p0tþp00txhD

þxf

ðA5Þ

@WOSh

@t¼

ðp0Þ2x̂f þ p0ðp̂� cÞ þ p00ðp̂� cÞx̂h�2p0ðr � cþ tÞ � p00ðr � cþ tÞx̂h

D̂

26664

37775 dr

dt

þ 1þ dr

dt

� �x̂f

ðA6Þ

The sign of above Equations also depends on thesign of dr/dt. In summary, all lemmas andpropositions hold, if p″ < 0, p″ = 0 and p″ > 0are not sufficiently large.

REFERENCES

Abramovsky, L. and Griffith, R. (2006), ‘Outsourcingand Offshoring of Business Services: How ImportantIs ICT?’, Journal of the European Economic Asso-ciation, 4, 594–601.

Amiti, M. and Wakelin, K. (2003), ‘Investment Liber-alization and International Trade’, Journal of Inter-national Economics, 61, 101–26.

Amiti, M. and Wei, S.J. (2005), ‘Fear of Service Outsourc-ing: Is It Justified?’, Economic Policy, 20, 308–47.

Arnold, J.M., Javorcik, B.S. and Mattoo, A. (2011),‘Does Services Liberalization Benefit ManufacturingFirms? Evidence from the Czech Republic’, Journalof International Economics, 85, 136–46.

Aronsson, T. and Koskela, E. (2011), ‘RedistributiveIncome Taxation, Outsourcing and Foreign DirectInvestment’, Economics Letters, 110, 86–9.

Arya, A., Mittendorf, B. and Sappington, D.E.M.(2008), ‘Outsourcing, Vertical Integration, and Pricevs. Quantity Competition’, International Journal ofIndustrial Organization, 26, 1–16.

Beladi, H., Marjit, S. and Chakrabarti, A. (2009),‘Tariff Jumping and Joint Ventures’, SouthernEconomic Journal, 75, 1256–69.

Bjorvatn, K. and Eckel, C. (2006), ‘Policy Competition forForeign Direct Investment between Asymmetric Coun-tries’, European Economic Review, 50, 1891–907.

Buehler, S. and Haucap, J. (2006), ‘Strategic Outsourc-ing Revisited’, Journal of Economic Behavior andOrganization, 61, 325–38.

Chang, H.J. (2004), ‘Regulation of Foreign Investmentin Historical Perspective’, The European Journal ofDevelopment Research, 16, 687–715.



Chaudhuri, A.R. and Benchekroun, H. (2012), ‘WelfareEffect of Mergers and Multilateral Trade Liberaliza-tion’, Review of International Economics, 20, 119–33.

Chen, Y., Ishikawa, J. and Yu, Z. (2004), ‘TradeLiberalization and Strategic Outsourcing’, Journal ofInternational Economics, 63, 419–36.

Claus, I. (2009), ‘New Zealand’s Economic Reformsand Changes in Production Structure’, Journal ofEconomic Policy Reform, 12, 133–43.

Collie, D.R. (2011), ‘Multilateral Trade Liberalisation,Foreign Direct Investment and the Volume of WorldTrade’, Economics Letters, 113, 47–9.

Crotti, S., Cavoli, T. and Wilson, J.K. (2010), ‘The Impactof Trade and Investment Agreements on Australia’sInward FDI Flows’, Australian Economic Papers, 49,259–75.

Demekas, D.G., Horv�ath, B., Ribakova, E. and Wu, Y.(2007), ‘Foreign Direct Investment in EuropeanTransition Economies—The Role of Policies’,Journal of Comparative Economics, 35, 369–86.

Ederington, J. and McCalman, P. (2010), ‘The Liber-alization of Trade and Foreign Direct Investment: APolitical Economy Analysis’, Journal of EconomicPolicy Reform, 13, 225–40.

Egger, P., Larch, M. and Pfaffermayr, M. (2004),‘Multilateral Trade and Investment Liberalization:Effects on Welfare and GDP Per Capita Conver-gence’, Economics Letters, 84, 133–40.

Egger, P., Larch, M. and Pfaffermayr, M. (2007), ‘Onthe Welfare Effects of Trade and InvestmentLiberalization’, European Economic Review, 51,669–94.

Francois, J. and Hoekman, B. (2010), ‘Services Tradeand Policy’, Journal of Economic Literature, 48,642–92.

Fujiwara, K. (2011), ‘Tariffs and Trade Liberalisationwith Network Externalities’, Australian EconomicPapers, 50, 51–61.

Grossman, G.M. and Helpman, E. (2003), ‘Outsourcingversus FDI in Industry Equilibrium’, Journal of theEuropean Economic Association, 1, 317–27.

Grossman, G.M. and Helpman, E. (2005), ‘Outsourcingin a Global Economy’, Review of Economic Studies,72, 135–59.

Horiuchi, E. and Ishikawa, J. (2009), ‘Tariffs andTechnology Transfer through an IntermediateProduct’, Review of International Economics, 17,310–26.

Ishikawa, J. and Horiuchi, E. (2012), ‘Strategic ForeignDirect Investment in Vertically Related Markets’,Economic Record, 88, 229–42.

Ishikawa, J., Morita, H. and Mukunoki, H. (2010), ‘FDIin Post-production Services and Product MarketCompetition’, Journal of International Economics,82, 73–84.

Jabbour, L. (2010), ‘Offshoring and Firm Performance:Evidence from French Manufacturing Industry’, TheWorld Economy, 33, 507–24.

Kayalica, M.O. and Lahiri, S. (2007), ‘DomesticLobbying and Foreign Direct Investment. The Roleof Policy Instruments’, Journal of InternationalTrade and Economic Development, 16, 299–323.

Lin, P. and Saggi, K. (2011), ‘Foreign Direct Invest-ment in a Two-Tier Oligopoly: Coordination, Verti-cal Integration, and Welfare’, InternationalEconomic Review, 52, 1271–90.

Lloyd, P. (2008), ‘100 Years of Tariff Protection inAustralia’, Australian Economic History Review, 48,99–145.

Markusen, J.R. (1989), ‘Trade in Producer Services andin Other Specialized Intermediate Inputs’, AmericanEconomic Review, 79, 85–95.

Markusen, J. R. (1997), ‘Trade versus InvestmentLiberalization,’ NBER Working Paper No. 6231,Cambridge, MA.

Markusen, J.R. (2002), Multinational Firms and theTheory of International Trade. MIT Press,Cambridge, MA.

Markusen, J.R., Rutherford, T.F. and Tarr, D. (2005),‘Trade and Direct Investment in Producer Servicesand the Domestic Market for Expertise’, CanadianJournal of Economics, 38, 758–77.

Mukherjee, A. and Suetrong, K. (2012), ‘Trade CostReduction and Foreign Direct Investment’, EconomicModelling, 29, 1938–45.

Neary, J.P. (2009), ‘Trade Costs and Foreign DirectInvestment’, International Review of Economics andFinance, 18, 207–18.

Qiu, L.D. and Wang, S. (2011), ‘FDI Policy, GreenfieldInvestment and Cross-border Mergers’, Review ofInternational Economics, 19, 836–51.

Rajan, R.S. and Srivastava, S. (2007), ‘Global Out-sourcing of Services: Issues and Implications’, Har-vard Asia Pacific Review, 9, 39–40.

Sadleir, C. and Mahony, G. (2009), ‘Institutional Challengesand Response in Regulating Foreign Direct Investment toAustralia’, Economic Papers, 28, 337–45.

Shy, O. and Stenbacka, R. (2003), ‘Strategic Outsourc-ing’, Journal of Economic Behavior and Organiza-tion, 50, 203–24.

Skaksen, M.R. (2005), ‘Should Governments SubsidiseInward Foreign Direct Investment?’, ScandinavianJournal of Economics, 107, 123–40.

Srivastava, S. (2006), ‘The Role of Foreign DirectInvestment in India’s Services Exports: An EmpiricalInvestigation’, The Singapore Economic Review, 51,175–94.

Tomiura, E., Ito, B. and Wakasugi, R. (2011), ‘OffshoreOutsourcing Decision and Capital Intensity: Firm-Level Relationships’, Economic Inquiry, 49, 364–78.



Documents

FDI and Outsourcing in a Service Industry: Welfare Effects ... · FDI in the service sector are more personal than FDI in the goods sector, such as nationality quotas for managers