the sensitivity of u.s. multinational enterprises to political and macroeconomic uncertainty: a...

ARTICLE IN PRESS

international

business

review

International Business Review 16 (2007) 732–750

0969-5931/$ -

doi:10.1016/j

�Tel.: +32

E-mail ad

www.elsevier.com/locate/ibusrev

The sensitivity of U.S. multinational enterprisesto political and macroeconomic uncertainty:

A sectoral analysis

Rodolphe Desbordes�

Department of Economics, University of Strathclyde, 130 Rottenrow, Glasgow G4 0GE, UK

Received 15 October 2004; received in revised form 20 December 2006, 12 July 2007; accepted 31 August 2007

Abstract

This paper confronts two alternative approaches for explaining U.S. foreign direct investment

(FDI) pattern in developing countries. According to the real options (RO) approach, FDI in capital-

intensive industries should be particularly deterred by political and macroeconomic uncertainty. On

the other hand, the supply chain risk management (SCRM) approach puts forward that

multinational enterprises in vertically integrated industries are unlikely to locate their foreign

activities in risky countries. Thanks to the use of sectoral data, it is demonstrated that the SCRM

approach explains much better the pattern of U.S. FDI in developing countries than the RO

approach.

r 2007 Elsevier Ltd. All rights reserved.

JEL classification: F23

Keywords: Country risk; FDI; Multinational enterprises; Uncertainty

1. Introduction

Few economists or laymen would deny that political events can have an important,sometimes even overwhelming, impact on economic decisions in general, andinvestment decisions in particular. (Stevens, 2000, p. 179).

see front matter r 2007 Elsevier Ltd. All rights reserved.

.ibusrev.2007.08.006

10 473 508.

dress: [email protected]

www.elsevier.com/locate/ibusrev

dx.doi.org/10.1016/j.ibusrev.2007.08.006

mailto:[email protected]

ARTICLE IN PRESSR. Desbordes / International Business Review 16 (2007) 732–750 733

Many studies have put forward the strong negative impact of political risk and, to alesser extent, of macroeconomic uncertainty, on aggregate foreign direct investment (FDI)in developing countries.1 At the disaggregated level, the evidence seems to be less clear-cut;Wheeler and Mody (1992) and Raff and Srinivasan (1998) do not find that FDI in the U.S.electronic industry is deterred by politically related uncertainty and Campa (1994) fails touncover a link between capacity expansions by multinationals in the chemical industry andmacroeconomic uncertainty. In Aizenman and Marion (2004), supply-side macroeconomicuncertainty exerts no impact on horizontal FDI but strongly influences the location ofvertical FDI. These mixed results suggest that the impact of political and macroeconomicuncertainty on FDI is largely governed by industry-specific characteristics.

This paper contributes to the literature by confronting two alternative approaches forexplaining U.S. FDI pattern in developing countries. According to the real options (RO)approach, FDI in capital-intensive (CI) industries should be particularly deterred bypolitical and macroeconomic uncertainty. On the other hand, the supply chain riskmanagement (SCRM) approach puts forward that multinational enterprises (MNEs) invertically integrated (VI) industries are unlikely to locate their foreign activities in riskycountries. Thanks to the use of sectoral data, available for 19 developing countries over the1984–1996 period, it is demonstrated that the SCRM approach explains much better thepattern of U.S. FDI in developing countries than the RO approach. Whereasfragmentation of the production process makes an MNE highly vulnerable to internationalsupply chain disruptions, replication of identical production facilities provides ‘‘opera-tional flexibility’’ (Kogut, 1985b), by reducing the risk exposure of an MNE confronted topolitical or economic uncertainty. Hence, not all organisational forms of multinationalitygenerate an instrument hedge against risk and the impact of country-specific uncertaintyon the investment of an MNE cannot be assessed without knowing beforehand the globalstrategy of the firm.

This paper is organised as follows. Section 2 reviews two alternative approaches forexplaining the impact of uncertainty on FDI; they both underline that the sensitivity of FDIto uncertainty is industry specific, though for different reasons. Section 3 presents the modelspecification and data used to confront these two theories. In Section 4, empirical results aregiven. Section 5 provides a discussion of the results by highlighting that the manufacturingpresence of U.S. MNEs in the developing world is largely shaped by industry-specificcharacteristics. Section 6 concludes and puts forward two implications for managers.

2. The theoretical impact of uncertainty on FDI

2.1. Real options approach

The option approach to investment2 is based on the idea that the investment of an MNEexhibits three important features: a partial irreversibility, an uncertainty over its futurereturn and a possibility to choose its timing (Rivoli and Salorio, 1996). Uncertainty isassumed to be fully exogenous to the investment decision of the MNE and thereforeproceeding now to a Kogut and Kulatilaka (1994)’s ‘‘platform investment’’ will not expand

1See for instance Mody and Srinivasan (1998), Wei (2000), Benassy-Quere et al. (2001) or Globerman and

Shapiro (2002, 2003).2The reader is referred to Dixit and Pyndick (1994) for an exhaustive discussion on this subject.

ARTICLE IN PRESSR. Desbordes / International Business Review 16 (2007) 732–750734

investment opportunities by resolving uncertainty. In fact, because the future is uncertainand the investment irreversible, an MNE may prefer to wait for new information beforeinvesting in a risky country. In other words, it holds an option whether to invest or not andthe opportunity cost of exercising it should be included in overall investment expenditures.This can drastically change the decision outcomes based only on the standard net presentvalue rule, i.e. invest if current value of expected profits exceeds current value ofinvestment costs. In a one country and two periods model, it can be shown that taking intoaccount the impact of uncertainty on the timing of an irreversible MNE investment leadsto the following decision rule (Serven, 1997):

p0 � rI4ð1� pÞrI � p1l

r

� �,

p04rI �rþ ð1� pÞ

rþ ð1� pÞ � a,

p04rI � risk premium, ð1Þ

where ð1� pÞ is the probability of a loss-making situation, corresponding to futurerevenues ðp1l Þ being lower than the user cost of capital (rI) and a is the ratio of future loss-making revenues to current revenues ðp0Þ, with ao1. Since ðrþ ð1� pÞÞ=ðrþ ð1� pÞ � aÞmust be greater than 1, the above equation indicates that the MNE will invest in a riskycountry only if it earns a current variable profit which exceeds the user cost of capital bythe multiple ðrþ ð1� pÞÞ=ðrþ ð1� pÞ � aÞ. In the absence of sunk costs, the firm wouldsimply invest if p04rI .Industries diverge in their capital intensity. Table 1 shows that the capital to labour ratio

is much higher for U.S. FDI in the chemical, metal and transport industries than for U.S.FDI in the food, machinery and electronic industries. CI industries incur large industry-specific sunk costs which should make them particularly sensitive to political andmacroeconomic uncertainty. In Eq. (1), first line, a larger irreversible investment increasesthe profitability threshold above which investment is undertaken, since the required returnon investment to invest now must increase to compensate for the rising size of potentialfuture per-period losses rI � p1l . As put by Rivoli and Salorio (1996) ‘‘the less reversible theFDI, the greater the rationale for postponing the decision (p. 347).’’The higher the business environment uncertainty, the higher the risk premium and

therefore the higher the return on investment required by MNEs to invest in a given country.The high-risk premia diminish the range of projects deemed attractive to foreign investorsand therefore the amount of FDI inflows. Hence, if the RO approach is valid to explain thepattern of FDI in developing countries, political and macroeconomic uncertainty shouldexert a negative impact on FDI, with FDI in CI industries being much more sensitive topolitical and macroeconomic uncertainty than FDI in other industries.

Hypothesis 1. According to the RO approach, FDI in CI industries should be much more

sensitive to political and macroeconomic uncertainty than FDI in other industries.

2.2. The supply chain risk management approach

The SCRM approach finds its roots in the minimisation of international supply chaindisruptions. Many MNEs offshore an increasing part of their production to take

ARTICLE IN PRESS

Table 1

U.S. manufacturing industries abroad, by capital intensity and international integration

Industry Type Property, plant and

equipment, per

employee

Percentage of sales

exported to U.S. parents

(and other foreign

affiliates)

Percentage of goods

imported from U.S.

parents for further

processing

All manufacturing

industries

— 46 526 12 (30) 10

Food and kindred

products

— 47 759 2 (12) 2

Chemicals and allied

products

CI 78 197 3 (22) 6

Primary and

fabricated metals

CI 59 292 7 (17) 6

Industrial machinery

and equipment

VI 33 701 18 (46) 10

Electronic and other

electric equipment

VI 21 342 16 (31) 21

Transportation

equipment

CI/VI 55 948 27 (47) 17

Notes: Data come from Mataloni and Fahim-Nader (1996) and cover the whole universe of MOFAs. CI, Capital-

intensive industry: VI, Vertically integrated industry. Exports and imports expressed as a share of total sales.

R. Desbordes / International Business Review 16 (2007) 732–750 735

advantage of international differences in factor prices (Feenstra, 1998; Kotabe and Swan,1994; Moxon, 1975). Such phenomenon is particularly salient in the manufacturing sector,especially in equipment manufacturing, as the production process of many final goods canbe physically divided in a number of discrete stages, varying in their factor intensities(Hanson et al., 2001). This fragmentation of manufacturing activities leads to the locationof each stage of production where it can be produced at the lowest cost: labour-intensivegoods in labour-abundant (developing) countries and capital/skilled-intensive goods incapital/skilled-abundant (developed) countries (Helpman, 1984; Kogut, 1985a). However,international vertical specialisation can be a risky and costly strategy if disruptions in theinternational supply chain are common, as a consequence of political or macroeconomicuncertainty. As shown by Aizenman and Marion (2004), the limited substitutabilitybetween the outputs produced in the geographically distinct production stages makes avertical FDI much more vulnerable to uncertainty than a horizontal FDI, which entails theproduction of similar products in multiple countries. For instance, assembly of the finalgood in the home country of the MNE will be jeopardised if production of intermediategoods in host countries is perturbed by political risk. Likewise, it is unlikely that effectiveinternational cost minimisation can be achieved by locating activities in countriescharacterised by volatile economic conditions. Without stability, the ‘‘relational costs’’generated by the geographic separation of various production stages may outweigh thelocational benefits of offshoring, such as labour costs savings (Levy, 1995; Prater et al.,2001). Conversely to fragmentation, replication of identical production facilities provides‘‘operational flexibility’’ (Kogut, 1985b), by reducing the risk exposure of an MNEconfronted to political or economic uncertainty. Under an horizontal structure,perturbations of production in one country or a rise in foreign costs can be compensatedby shifting production from one plant to another. In addition, investing in excess capacity


in plants in multiple countries reduces the bargaining power of a government which may betempted to renegotiate previous agreements, in an ‘‘obsolescing bargain’’ fashion(Aizenman and Marion, 2004; Janeba, 2000; Kogut, 1983, 1985b; Vernon, 1971). Asexplained by Campa (1994, p. 574), ‘‘This does not mean that multinational corporationsare not affected by volatility, but rather that they endogenously take it into considerationwhen determining their expansion plans in different countries.’’ Hence, in face ofuncertainty, horizontal FDI appears to be a more suitable expansion strategy than verticalFDI. Such relationship between uncertainty and the expected profit ðE½

Q�Þ associated with

each type of FDI (vertical or horizontal) can be expressed as

qE½Q

vert�

qUncertaintyo

qE½Q

horiz�

qUncertaintyp0. (2)

Industries diverge in their international vertical integration. Table 1 shows that affiliateexports as a share of affiliate total sales are much higher for the machinery, electronic andtransport U.S. industries abroad than for the food, chemical and metal industries, whoseforeign affiliates mainly aim at serving the local market. In addition, production ofequipment industries involve much more trade in intermediate products than otherindustries. If the SCRM approach is valid, FDI in these VI industries3 should be muchmore deterred by political and macroeconomic uncertainty than FDI in other industries.

Hypothesis 2. According to the SCRM approach, FDI in VI industries should be much more

sensitive to political and macroeconomic uncertainty than FDI in other industries.

3. Model specification and data

In order to discriminate between the RO approach and the SCRM approach, thefollowing model will be estimated:

LnðU :S:CPX ticÞ ¼ a0 þ a1 LnðGDPt�1

c Þ þ a2 LnðGDPPCt�1c Þ þ a3 LnðAGGt�1

c Þ

þ a4 LnðUWCt�1c Þ þ a5 LnðTAX t�1

c Þ þ a6 LnðTROPt�1c Þ

þ a7 LnðDISTcÞ

þ a8 LnðUNCERt�1c Þ þ a9½LnðUNCERt�1

c Þ �Di�

þ a10Di þ a11Tt þ �tic ð3Þ

where �tic is the error term. The dependent variable (U.S.CPX) corresponds to the industry-

specific i capital expenditures in the food, chemical, metal, machinery, electronic andtransport industries4 carried out by manufacturing U.S. majority-owned foreign affiliates(MOFAs) in developing country c at period t, in millions of U.S.$. A MOFA is a foreignaffiliate in which the combined direct and indirect ownership interest of all U.S. parentsexceeds 50%, whereas capital expenditures include all expenditures that are charged tocapital accounts and are made to acquire, add to, or improve property, plant andequipment (Fahim-Nader, 1994). In 1994, 89% of U.S. foreign affiliates were MOFAs(Mataloni and Fahim-Nader, 1996) and therefore the variations in the fixed assets of

3The classification used in this paper echoes the findings of Kobrin (1991) on the transnational integration of

various U.S. industries.4The corresponding ISIC3d categories are, respectively, 31-, 35-, 37- & 380/381, 382, 383 and 384.


MOFAs should give a good picture of the investment activity of U.S. MNEs. This measureof MNE foreign presence is preferred to FDI flows, which are a poor proxy for MNEs’activities (Hausmann and Fernandez-Arias, 2000). Data come from benchmark andannual sample surveys of U.S. direct investment abroad carried by the U.S. Bureau ofEconomic Analysis (BEA).5

The volume of U.S. manufacturing capital expenditures is assumed to depend on thefollowing host country determinants: market size (GDP), host country wealth (GDPPC),agglomeration forces (AGG), labour costs (UWC), corporate taxes (TAX), trade openness(TROP), distance from the United States (DIST) and uncertainty (UNCER).

The market size (GDP) is approximated by the Gross Domestic Product in millions of2000 constant U.S.$. Data come from the World Bank Global Development NetworkDatabase (GDND).6 Host country wealth (GDPPC), measured by GDP per capita in 2000constant U.S.$, should capture aspects of the business climate which are well correlatedwith development, such as the availability of physical infrastructure or the extent to whichlabour and goods markets are heavily regulated.7 The agglomeration forces (AGG) aretaken into account through the inclusion of the lagged stock of overall U.S. financial FDI,expressed in 2000 constant U.S.$. Foreign investors locate close to each other in order tobenefit from inter-firm technological spillovers, the creation of a pooled market for skilledworkers or the promotion of specialised intermediate inputs (Head et al., 1995; Markusen,1990). In addition, size of the stock of past investments is likely to send a strong signalabout operating conditions and economic prospects to potential foreign investors,especially in the limited information context characterising developing countries(Kinoshita and Mody, 2001). Data come from the U.S. BEA.8 Labour costs (UWC)correspond to unit wage costs in the manufacturing sector. They are calculated by dividingaverage wage per worker, including supplements, by value added per worker. Data comefrom UNIDO (1997). Since data on nominal wages and value added are only availableevery five years, missing values have been filled through linear interpolation. Instead ofstatutory tax rates, which may give a poor picture of the effective fiscal burden faced byMNEs as they are often eligible to a number of fiscal incentives, average tax rates (TAX)paid by U.S. MNEs are used. They have been calculated by Grubert and Mutty (2000), onthe basis of tax returns of more than 500 large U.S. manufacturing MNEs. In order toreduce measurement error, average effective tax rates correspond to the average of valuesfor the current year and last year. In addition, because data are only available for evenyears, values for odd years are the average of values for the previous year and the followingyear. Data have been kindly provided by the authors. The trade openness variable (TROP)is an indicator developed by Gwartney et al. (2001). It is based on the comparison betweenthe actual trade openness ððX þMÞ=GDPÞ and the expected trade openness, given thestructural characteristics of a country (population, area, length of coastline and remotenessfrom world demand). An actual trade openness much greater than its expected value isassumed to imply trade policies favourable to international trade. The indicator rangesbetween 0 and 10 (worst to best). Annual data have been kindly provided by the authors.As a proxy for transaction and transportation costs, the distance (DIST) between the host

5http://www.bea.gov/bea/di/di1usdop.htm6http://devdata.worldbank.org/wbquery/7The World Bank (2004) has shown that it is the poorest countries which regulate the most.8Data retrieved from http://www.bea.doc.gov/bea/di/di1usdbal.htm

http://www.bea.gov/bea/di/di1usdop.htm

http://devdata.worldbank.org/wbquery/

http://www.bea.doc.gov/bea/di/di1usdbal.htm


country and the United States is also included; the geodesic distances are calculatedfollowing the great circle formula, which uses latitudes and longitudes of the mostimportant cities/agglomerations (in terms of population). Data come from the CEPII.9

It is expected that market size, host country wealth and agglomeration forces will have apositive impact on U.S. manufacturing capital expenditures abroad, whereas labour costsand taxes should exert a negative effect. It is ex ante unclear whether trade openness anddistance will have a positive or negative impact on U.S. FDI. Horizontal FDI, aimed atserving the local market, is usually motivated by excessive trade costs, i.e. high transportcosts and trade barriers, which make exports an expensive strategy. On the other hand,vertical/export-platform FDI, which involves international trade of intermediate and finalgoods, should be concentrated in relatively open countries. ‘‘Pure’’ vertical FDI, i.e. whenthe production of foreign affiliates is exclusively exported towards the home country, alsoneeds to be located in countries geographically close to the United States. Finally in bothcases, distance may make coordination of worldwide activities of an MNE more difficult,which could particularly hinder the management of an integrated international supplychain. Nevertheless, since horizontal and vertical FDI tend to be attracted by opposingfactors, a high volume of FDI can coexist in both closed/distant and open/near countries.Hence, signs of the trade policy and distance variables should depend in fine on theprevailing FDI type in the country sample.Beyond these control variables, proxies of country-specific political and macroeconomic

uncertainty are included. The political uncertainty variable (POLRISK) corresponds to theInternational Country Risk Guide indicator, which aggregates numerical evaluations of 12dimensions of political risk.10 The indicator ranges from 0 (high political risk) to 100 (nopolitical risk) and is available on a monthly basis. For the purpose of this paper, valuescorrespond to the annual average of monthly values and the index has been rescaled suchas 0 corresponds to no political risk and 100 corresponds to high political risk. Theuncertainty of the fluctations of the following macroeconomic variables is assessed: output(INCERGDPG), consumer price index (INCERINF), real exchange rate (INCERRER),terms of trade (INCERTOT). Following Aizenman and Marion (1995), Ramey andRamey (1995) or Serven (1997, 1998), uncertainty of a given macroeconomic indicator iscalculated as the standard deviation of the estimated residuals of a first-orderautoregressive process. In order to obtain yearly values this procedure is repeated foreach year, with a time period set at seven years and values of the macroeconomicuncertainty indicator correspond to the standard deviations of the estimated residuals,averaged over three consecutive years—the current year and the preceding two. A random-effects panel estimation is used to run the autoregressive process, instead of country-specific regressions. Although using a panel data model makes the restrictive assumptionthat all countries share the same parameters, it is likely that the indicator generated will fitbetter the structure of the U.S. FDI panel data, by taking into account not only withinvariations but also cross-sectional differences in macroeconomic uncertainty.11 The realexchange rate has been estimated as the nominal exchange rate times the value of the U.S.

9http://www.cepii.fr/francgraph/bdd/distances.htm10These dimensions are (1) government stability, (2) socioeconomic conditions, (3) investment profile, (4)

internal conflict, (5) external conflict, (6) corruption, (7) military in politics, (8) religion in politics, (9) law and

order, (10) ethnic tensions, (11) democratic accountability, (12) bureaucracy quality.11Using the innovations of a cross-country forecasting equation for growth as a measure of uncertainty, Ramey

and Ramey (1995, p. 1145) find that ‘‘the qualitative results are not substantially altered when the standard

http://www.cepii.fr/francgraph/bdd/distances.htm

ARTICLE IN PRESS

Table 2

Summary statistics

Variable Mean Std. dev. Min. Max.

U.S. capital expenditures (U.S.CPX) 33.06 85.15 0 1006

Market size (GDP) 132 308.12 136 384.49 6460 539 000

Host country wealth (GDPPC) 2687.29 1950.35 252.3 9159.12

Agglomeration forces (AGG) 3522.02 4843.51 82.03 26 425.57

Trade openness (TROP) 5.48 2.37 0 10

Bilateral distance (DIST) 9103.49 4115.30 2598.62 15 460.12

Unit wage costs (UWC) 0.27 0.11 0.13 0.65

Corporate taxes (TAX) 0.36 0.08 0.20 0.55

Political risk (POLRISK) 53.72 10.06 34.75 76.33

Output uncertainty (INCERGDPG) 0.04 0.02 0.01 0.09

Inflation uncertainty (INCERINF) 0.17 0.28 0.01 1.39

Exchange rate uncertainty (INCERRER) 0.32 0.84 0.01 4.78

Terms of trade uncertainty (INCERTOT) 0.10 0.06 0.01 0.29


consumer price index divided by the host country consumer price index. Yearly changes inoutput, real exchange rate and consumer price index, have been calculated as the firstdifference of their respective logarithmic values.12 Data come from the World BankGDND and the IMF International Finance Statistics Database.

The confrontation of the RO approach to the SCRM approach will be achieved throughthe introduction of interaction terms. In order to test the RO (SCRM) approach, theuncertainty indicators will be first interacted with a CI (VI) dummy (Table 1) and,afterwards, with the six industry dummies Di. CI industries are chemical, metal andtransport industries and VI industries are electronic, machinery and transport industries(Table 1). If the RO (SCRM) approach is valid, uncertainty should exert a strongernegative impact on the volume of capital expenditures in CI (VI) industries. Hence,acceptance or rejection of either hypothesis will be based on the sign and significance of theinteraction terms’ coefficients a9.

Summary statistics are given in Table 2. Data are available for 19 developing countries,which are the biggest and most ancient recipients of U.S. FDI in the developing world,over the 1984–1996 period.13 Variables are measured in logarithmic values.14 This has twoadvantages: such transformation reduces the influence of large values and coefficients canbe directly interpreted as partial elasticities. Each explanatory variable is also lagged byone year to reduce any endogeneity bias and to take into account that foreign investorsbase their location choice on past information. Industry and time dummies are included tocontrol for unobserved industry-specific and time-specific factors.

(footnote continued)

deviations are estimated using country-specific forecasting equations’’, suggesting that both estimations yield

qualitatively similar results.12The creation of the uncertainty indicators requires unit-free macroeconomic measures: output growth,

inflation, real exchange rate change, correspond to rates of change and the terms of trade are measured relative to

the base year 1995 (index ¼ 100 in 1995 for all countries).13These countries are Argentina, Brazil, Chile, Colombia, Ecuador, Egypt, India, Indonesia, Malaysia, Mexico,

Nigeria, Panama, Peru, Philippines, South Africa, South Korea, Thailand, Turkey, Venezuela.141 is added when the value of a variable equals zero.


Several regression diagnostics have been carried out. The low value of the mean varianceinflation indicates that multicollinearity is not a problem. Influential observations, both interms of leverage and outlierness, have been removed according to a Cook’s D test.Finally, a Breush–Pagan/Cook–Weisberg test (Breusch and Pagan, 1979) and Arellanoand Bond (1991) test, respectively, indicate that residuals are heteroscedastic andautocorrelated up to order 3. In the presence of heteroscedasticity and serial correlation,standard errors of estimators are biased and therefore tests of statistical significance are nomore valid. Standard errors are therefore corrected in order to be robust to bothheteroscedasticity and serial correlation.

4. Empirical results

Estimation results are given in Tables 3 and 4. In regression (1), about 60% of thevariance of the dependent variable is explained. All coefficients have the expected sign andmost are significant at the 5% level. U.S. MNEs tend to invest in large and rich developingcountries, which have been proven to be a favourable location for past U.S. FDI andwhich offer low taxes. The positive and significant coefficient of the trade opennessvariable, combined with a positive but not significant impact of the distance on U.S. FDI,suggest that FDI of the vertical type slightly prevails in the country sample. Unit wagecosts do not appear to be a crucial factor of attractiveness. Finally, coefficient of thevariable of interest, Political risk, is negative and significant at the 5% level. A 1 pointincrease in the political risk rating of a given country would decrease U.S. capitalexpenditures by 8%.15

From regression (1) it is unclear why political risk exerts such a negative impact on thefixed investment of U.S. foreign affiliates. The RO approach suggests that uncertaintydeters FDI when the latter involves sunk costs. Hence, U.S. FDI in CI industries shouldtend to avoid politically risky countries. On the other hand, the SCRM approach putsforward that uncertainty makes difficult the vertical integration of production stageslocated in various countries. According to this hypothesis, U.S. FDI in VI industriesshould also exhibit a preference for politically stable locations. The validity of bothhypotheses can be tested through interaction terms.In regressions (2) and (3), the political risk variable is interacted either with a CI

industries dummy or with a VI industries dummy. If U.S. MNEs behave according to theRO (SCRM) approach, political risk should exert a stronger impact on FDI occurring inCI (VI) industries and coefficient of the interaction term should be negative and significant.Regression (2) tests the RO approach. U.S. MNEs in CI industries do not seem to pay

more attention to political risk than U.S. MNEs in other industries: although thecoefficient of the interaction term has the expected sign, its significance is strongly rejected.Hence, no evidence is found in favour of the RO approach. Regression (3) tests the SCMRapproach. The highly negative and significant coefficient of the interaction term supportsthe hypothesis that a more pertinent typology of the sensitivity of industries to uncertaintycan be achieved by distinguishing between VI and non-VI industries. In fact, the absence ofsignificance of the political risk variable suggests that only U.S. MNEs in VI industriesmay take into consideration business uncertainty in their location decisions as only the

15Semi� elasticityP:Risk ¼ ElasticityP:Risk=P:Risk ¼ �4:2454’ �7:85% where P:Risk is the mean sample value of

the political risk variable.

ARTICLE IN PRESS

Table 3

U.S. capital expenditures and political risk, by industry

Determinants Dep. variable: U.S. capital expenditures

(1) (2) (3) (4)

Market size 1.69a 1.69a 1.70a 1.67a

(0.38) (0.38) (0.31) (0.30)

Host country wealth 1.03b 1.03b 1.05b 1.02b

(0.48) (0.48) (0.44) (0.42)

Agglomeration forces 0.67c 0.67c 0.63c 0.65b

(0.36) (0.36) (0.32) (0.30)

Trade openness 1.15b 1.15b 1.18b 1.11b

(0.57) (0.57) (0.52) (0.46)

Bilateral distance 0.45 0.45 0.44 0.46

(0.74) (0.74) (0.67) (0.62)

Unit wage costs �0.26 �0.26 �0.32 �0.29

(0.74) (0.74) (0.68) (0.64)

Corporate taxes �2.98b �2.97b �2.92b �2.82b

(1.26) (1.26) (1.21) (1.19)

Pol. risk �4.24b �4.05 1.44

(1.96) (2.62) (1.46)

Pol. risk * CI industries dummy �0.38

(2.82)

Pol. risk * VI industries dummy �11.74a

(2.50)

Pol. risk * food industry dummy 2.00

(2.04)

Pol. risk * chemical industry dummy 4.09b

(1.81)

Pol. risk * metal industry dummy �2.52

(2.85)

Pol. risk * machinery industry dummy �12.27a

(4.08)

Pol. risk * electronic industry dummy �2.96

(3.53)

Pol. risk * transport industry dummy �15.94a

(4.13)

Constant �32.13b �32.83c �54.55a �56.08a

(15.42) (17.23) (13.96) (15.29)

R2 0.58 0.58 0.62 0.64

Observations 1131 1131 1131 1131

Number of countries 19 19 19 19

Notes: a, b, and c denote, respectively, significance at the 1%, 5%, and 10% level. Heteroscedasticity-

autocorrelation robust standard errors are in parentheses. Unreported time and industry dummies are included.

Time dummies and industry dummies are jointly significant at the 1% level.


interaction term is significant. A 1-point increase in the political risk rating of a givencountry would generate a staggering 21% decrease in the capital expenditures of U.S.MNEs operating in VI industries, whereas the impact on U.S. MNEs operating in CIindustries is likely to be null.

ARTIC

LEIN

PRES

S

Table 4

U.S. capital expenditures and macroeconomic uncertainty, by industry

Macroeconomic uncertainty Dependent variable: U.S. capital expenditures

Output uncertainty Inflation uncertainty Real exchange rate

uncertainty

Terms of trade

uncertainty

(5) (6) (7) (8)

Coefficient of variable �0.94 0.80 �0.35 0.12 0.01 �0.09 0.40 0.43

(0.60) (0.61) (0.41) (0.37) (0.38) (0.30) (0.67) (0.62)

Interaction with CI industries dummy 1.87c 0.89c 0.66 0.19

(0.98) (0.53) (0.58) (0.85)

Interaction with VI industries dummy �1.64 �0.05 0.91c 0.14

(0.99) (0.56) (0.51) (0.85)

Interaction with food industry dummy 0.78 0.17 �0.19 0.42

(0.72) (0.45) (0.35) (0.91)

Interaction with chemical industry dummy 0.85 �0.05 �0.32 0.56

(0.76) (0.48) (0.31) (0.90)

Interaction with metal industry dummy 0.80 0.33 0.38 0.30

(1.16) (0.59) (0.48) (0.93)

Interaction with machinery industry dummy �2.56c �0.83 0.88 0.44

(1.41) (0.91) (0.72) (1.30)

Interaction with electronic industry dummy �1.05 �0.49 �0.51 0.34

(0.67) (0.46) (0.54) (0.95)

Interaction with transport industry dummy 1.21 1.42 2.00a 0.92

(2.23) (1.09) (0.63) (1.32)

Notes: a, b, and c denote, respectively, significance at the 1%, 5%, and 10% level. Heteroscedasticity-autocorrelation robust standard errors are in parentheses.

Uncertainty variables are included in a specification similar to regression (1).

R.

Desb

ord

es/

Intern

atio

na

lB

usin

essR

eview1

6(

20

07

)7

32

–7

50

742


In regression (4), it is tested in which industries U.S. MNEs are the most sensitive topolitical risk, through the interaction of industry-specific dummies with the political riskvariable. Results once again reject the RO approach and accept the SCRM approach:among the six interaction terms, the highest negative coefficients concern the VI industries,especially the machinery and transport sectors. However, similar to the findings of Wheelerand Mody (1992) and Raff and Srinivasan (1998), U.S. FDI in the electronic industry doesnot seem to be deterred by political risk. Vertical integration with horizontal diversificationmay explain this lack of impact. Concerning non-VI industries, FDI of U.S. MNEs in thefood and metal industries are not statistically influenced by political risk and taking theresults at face value, U.S. MNEs in the chemical industry tend to invest more in riskycountries. The necessity to replicate identical production activities across developingcountries may provide them with a hedge instrument against risk. Results will be furtherdiscussed in the next section.

In Table 4, the impact of macroeconomic uncertainty on industry-specific capitalexpenditures is investigated. Results tend to indicate that U.S. FDI in VI industries is moresensitive to macroeconomic uncertainty than U.S. FDI in CI industries but most of thecoefficients are not significant and some of them have a counter-intuitive sign, likethe positive sign of the interaction term between real exchange rate uncertainty and thetransport industry. Overall, it can be concluded that macroeconomic uncertainty exertslittle impact on U.S. capital expenditures in developing countries, especially if political riskis included into the specification. Lehmann (1999), using the ICRG Economic Risk andFinancial Risk indicators, in addition to the ICRG Political Risk indicator, reaches thesame conclusion.

5. Discussion: uncertainty and global manufacturing presence

The fact that uncertainty has been found, in Section 4, to exert no impact on the FDIchoices of U.S MNEs in non-VI industries suggests that they should invest more evenly indeveloping countries, regardless of country risk, than MNEs in VI industries. The differentexpansion paths of U.S. FDI in VI and non-VI industries can be highlighted through thecalculation of a Herfindahl-like geographic investment index, which measures the degree ofFDI concentration in the developing world, by industry. Following Knetter and Slaughter(1999), it is calculated as

Htic ¼

X USCPX tic

USCPX tiw

� �2

, (4)

where USCPX tic is the volume of U.S. capital expenditures in industry i and country c at

time t and X tiw is the sample ðwÞ total of U.S. capital expenditures in industry i at time t. By

construction, the Herfindhal index ðHticÞ ranges between 0 and 1. It equals 0 if U.S. FDI in

a given industry occurs evenly across all countries of the sample, whereas it equals 1 if U.S.FDI in a given industry is concentrated in one country.

Fig. 1 supports the SCRM hypothesis since U.S. FDI in the non-VI food, chemical andmetal industries are much more dispersed in the developing world than U.S. FDI in the VImachinery and transport industries, especially if the 1991–1996 period is observed;interestingly, as in Section 4, the VI electronic industry remains an outlier with aconcentration index close to those of non-VI industries. In addition, whereas two of the VI

ARTICLE IN PRESSC

on

cen

trati

on

In

dex

0.60

0.50

0.40

0.30

0.20

0.10

0.00

Food Chemical Metal Machinery Electronic Transportation

Industry

Fig. 1. Geographic concentration of U.S. FDI, by industry. Source: BEA and calculation of the author.

R. Desbordes / International Business Review 16 (2007) 732–750744

industries have slightly increased the geographic concentration of their foreign productionover time, the opposite has happened in all non-VI industries, which have stronglyexpanded their worldwide presence in the nineties. For instance, Table 5 shows that thenumber of countries in which significant FDI (more than $3 millions each year) in the foodindustry occurs, has increased from 9 to 16 countries over the 1985–1996 period. It is alsoworth noting that it is in the highly CI chemical industry that U.S. MNEs are significantlypresent in all developing countries of the sample at the end of the nineties.The higher geographic expansion of U.S. MNEs in non-VI industries, even when

confronted to country risk, is related to the specific characteristics of these industries. Thefood, chemical and metal industries cannot take advantage of the international differencesin factor prices. First, production processes of continuous-processes industries, e.g.chemicals, tend to be technically indivisible. Second, industries intensive in capital(Table 1) are less likely to gain from low labour costs. Last, but not least, high weight tovalue ratios defeat the cost-reduction advantages of fragmentation, it greatly increasestransport costs between production sites, and it also makes exporting final goods from thehome market or from export platform countries expensive (Table 5). Hence, MNEs in non-VI industries must adopt a strategy of replication of identical production activities mostlyaimed at serving local markets. This strategy is underlined by greater value added to salesratios than those of VI industries, as a larger fraction of the inputs and outputs thatcomprise total sales are produced in-house, and high local sales to exports or to total salesratios (Table 5).Although MNEs in non-VI industries cannot reap the benefits of differences in

international factor prices, the necessity to replicate identical production activities acrossdeveloping countries, provides them on the other hand with a hedge instrument againstrisk, which both reduces the impact of country-specific uncertainty on their globalactivities and favours their entry on new markets. From an international portfolioperspective, geographic diversification reduces risks related to country-specific uncertaintyas long as correlations between the fluctuations of different foreign markets are less thanperfect (Mathur and Hanagan, 1983; Severn, 1974); Rugman (1976), Aggarwal (1980) andQian (1996) show that international diversification does indeed lower the overall risk ofU.S. MNEs, as measured by volatility of profits or beta coefficients. Furthermore, asbriefly mentioned in Section 2.2, replication of identical production activities in differentcountries provides a firm with operational flexibility, in response to uncertainty, by

ARTICLE IN PRESS

Table 5

Trade costs and geographic expansion, by industry

Industry Average U.S.

import ad valorem

freight costs

(1984–1996) (%)

Average

export unit

values

(1984–1996)

($/kg)

Value

added/

sales ratio

(1994)

(%)

U.S. local

sales/exports

ratio

(1984–1994)

(%)

Local sales/

total sales

(1994/

1984–1996)

(%)

Geographic

expansion

1995–1990 1991–1996

Food 9.5 3.72 28 6.02 79 (89) 9 16

Chemicals 5.5 2.57 31 1.17 66 (94) 16 19

Metals 4.9 1.99 32 1.43 63 (88) 5 9

Industrial

machinery

3.6 9.47 23 0.23 44 (81) 6 7

Electronics 3.4 10.79 30 0.53 59 (66) 7 10

Transportation

equipment

2.8 7.49 24 0.37 47 (83) 5 5

Notes: Data on U.S. import ad valorem freight costs correspond to the markup of the Customs-Insurance-Freight

value of U.S. imports over the a Free-On-Board customs value of U.S import relative to the Free-On-Board

customs value. Data come from Bernard et al. (2006). Data on U.S. unit values are calculated by dividing the

values of exports by the weight of these exports; data come from Nicita and Olarreaga (2006) and cover most

countries in the world. Data on value added and total or local sales come fromMataloni and Fahim-Nader (1996)

and cover the whole universe of MOFAs. For the local sales to total sales ratio, numbers in parentheses

corresponds to values for the country sample. Geographic expansion corresponds to the number of host countries

in the sample in which 3 millions of industry-specific capital expenditures occur each year, on average. Data come

from BEA (various years).


creating arbitrage and leverage opportunities (Aizenman and Marion, 2004; Ghoshal,1987; Kogut, 1985b; Miller, 1992). Faced with a rise in foreign costs, an MNE whichoperates the same production plants in different countries and holds excess capacity cantake advantage of variations in relative international factor price differences, by shiftingabroad production and substituting imports to local production; such flexible networkhypothesis is supported by the empirical works of Miller and Reuer Jeffrey (1998) andRangan (1998), though the latter points out that operational responses are likely to remainmodest. Similarly, threats by the host country government to expropriate the MNE or toblock production if negotiation of a more favourable profit-sharing agreement does notoccur, are less likely to be effective when the MNE produces the same final goods in severalcountries (Aizenman and Marion, 2004; Janeba, 2000; Kogut, 1985b).16 Hence, replicationof production activities ensures that even the full expropriation of the foreign affiliate willnot hinder the international activities of the MNE. A contrario, fragmentation ofproduction increases the exposure of VI industries to political and macroeconomicuncertainty. In line with this reasoning, Aizenman and Marion (2004) find that horizontalFDI is less deterred by uncertainty than vertical FDI.

In presence of uncertainty, MNEs in non-VI industries may in fact find beneficial toexpand their global manufacturing presence. Kogut and Kulatilaka (1994) emphasise the

16For instance, Janeba (2000) interprets the intention of General Motors, in 1997, to build quasi-identical

production plants in Argentina, China, Thailand and Poland, whose joint capacity was likely to exceed demand,

as evidence for a risk-reducing international diversification strategy.


importance of ‘‘investing in opportunities’’ through ‘‘platform investments’’. By investingearly in promising but uncertain markets, some MNEs learn how to do business in anunfamiliar environment and set-up production and distribution facilities, which both givethem the possibility to expand quickly their activities and to gain significant market sharesin the case that high market growth occurs in the future. First-mover advantages may thusdwarf the costs associated with market uncertainty, especially when the potential upside ofa FDI is particularly high and export costs are prohibitive. Put differently, in the realm ofRO approach, it could be argued that FDI occurs because high early mover-advantagesreduce its delayability (Rivoli and Salorio, 1996).Finally, throughout this paper, the location choices of U.S. FDI in the electronic

industry have not fitted really well with the SCRM approach since U.S. MNEs in a VIindustry appear to behave like U.S. MNEs in non-VI industries. Aizenman (2003) providesan answer to this paradox by suggesting that MNEs in VI industries may deal withuncertainty by replicating in several emerging countries the production of the sameintermediate products. However, such strategy can be costly as a new plant must be set-up,economies of scale are lost and the cost of coordinating different stages of the supply chainincreases. Besides, a foreign substitute to the original production site may not be easilyfound, especially if cost benefits are location-specific. Within VI industries, the electronicindustry is the best positioned to gain from this hybrid strategy of vertical integration withhorizontal diversification. Compared to other industries, its low capital intensity makesreplication of production plants financially viable and ensures that sourcing the sameproduct from different plants will only have a moderate impact on average total costs.Some stages of its production process are strongly intensive in unskilled labour and do notrequire significant local technological capabilities, which increase the number of alternativelocations as cost benefits are not location-specific (Lall et al., 2004). Finally, its high valueto weight ratio reduces the impediment of distance. These arguments fit well with the factsthat Moxon (1975) reports that some companies in the electronic industry declared usingmultiple offshore plants to insure against disruptions, that it is the single VI industry tohave significantly increased its global presence over time (Table 5) and that its Herfindahlgeographic index is much lower than those of other VI industries. As summarised byJaneba (2000, p. 1517), the ability of the electronic industry to reduce its exposure tocountry-specific risk through horizontal diversification may explain why Wheeler andMody (1992), Raff and Srinivasan (1998) and this study do not find an impact of politicalrisk on the location choices of FDI in this industry: ‘‘There is also indirect econometricevidence for the hypothesis that multinationals can overcome political risk bysimultaneously holding capacity in different countries. Econometric studies have foundlittle or no support for political risk as determinant of investment. This is particularly truefor footloose industries like electronics where firms typically plants in several countries asexport platform.’’In line with the SCRM approach, these reasons explain why the geographic expansion of

U.S. FDI in non-VI industries, contrary to U.S. FDI in VI industries, has not beensignificantly hindered by uncertainty, especially political uncertainty, as found in Section 4.The specific characteristics of non-VI industries bias the location choices of MNEsoperating in these industries towards a replication of production facilities across countries,mostly aimed at serving local markets. This pattern of industrial expansion abroad allowsthem to build a multinationality-induced hedge instrument against risk, which increasinglyreduces their exposure to country-specific political and macroeconomic uncertainty. On


some new foreign markets, entry in presence of uncertainty may be in fact warrantednecessary in order to build up the first-mover advantages necessary to take advantage ofthe upsides of the market. Finally, although U.S. MNEs in VI industries need generally tolocate their activities in safe countries only to fully benefit from the internationaldifferences in factor prices, the characteristics of the electronic industry give MNEsoperating in this industry the possibility to deal with country-specific uncertainty byadopting a hybrid strategy of vertical integration with horizontal diversification. Thismay explain the absence of econometric impact of uncertainty on U.S. electronic FDI inSection 4.

6. Conclusion and implications

The impact of political and macroeconomic uncertainty on U.S. FDI pattern has beeninvestigated, by trying to discriminate between alternative theoretical explanations. It isfound that U.S. MNEs in VI industries are much more sensitive to uncertainty than U.S.MNEs in other industries, including CI industries. Through their location in relatively safecountries, they attempt to minimise the disruptions that political or macroeconomicuncertainty may trigger in their international supply chain. Conversely to fragmentation,replication of production facilities across countries provides MNEs with a multi-nationality-induced hedge instrument against risk, which reduce their exposure tocountry-specific uncertainty.

This study carries out two implications for managers. First, the fact that market-seeking FDI, i.e. FDI in non-VI industries, is not deterred by political and macroeconomicuncertainty, suggests that geographic diversification is a successful risk-reducing strategy.Even when full fragmentation or concentration of the production process appears, atfirst glance, the cheapest strategy, building several plants and holding excess capacitymay provide arbitrage opportunities which far outweigh the additional costs generated bythis replication strategy (Janeba, 2000; Kogut, 1985b). In other words, the cost-effectiveness analysis needs to be dynamic rather than static, and should acknowledgethat the profitability of geographically distinct foreign affiliates is likely to beinterconnected. However, these arbitrage opportunities can only occur if the MNE isprepared to deal with contingencies. Hence, the second implication is that managersmust ensure that the options generated by multinationality are not drastically reducedby the absence of the organisational structure necessary to the coordination of variousinternational activities (Kogut, 1985b, 1989). For instance, Rangan (1998) finds thatMNEs, facing significant exchange rate changes in the eighties, did not appear to fullyexploit the benefits of operational flexibility, certainly thanks to a lack of planning,inertia, low circulation of information or internal opportunism triggered by inadequatehuman resource management. More worryingly, Pahud de Mortanges and Allers(1996) and Oetzel (2005) report that managers of MNEs, both at the headquarterand subsidiary levels, do not pro-actively manage risk, as they do not engage in riskassessment or management on an ongoing basis and such despite having alreadybeen confronted to economic and political risk. Within the context of rising globalisationof productive activities, managers should then make the necessary organisationaladjustments since the speed of a firm’s response to political or economic events acrossthe world may nowadays strongly determine its ability to preserve or expand its marketposition.


Acknowledgements

The author would like to thank Celine Azemar and anonymous referees for usefulcomments and suggestions. The financial support of the Belgian French-speakingcommunity (ARC 03/08-302) at the Catholique University of Louvain is gratefullyacknowledged.

References

Aggarwal, R. (1980). Investment performance of U.S.-based multinational companies: Comments and a

perspective on international diversification of real assets. Journal of International Business Studies, 11(1),

98–104.

Aizenman, J. (2003). Volatility, employment and the patterns of FDI in emerging markets. Journal of Development

Economics, 72(2), 585–601.

Aizenman, J., & Marion, N. (1995). Volatility, investment and disappointment aversion NBER working paper,

no. 5386.

Aizenman, J., & Marion, N. (2004). The merits of horizontal versus vertical FDI in the presence of uncertainty.

Journal of International Economics, 62(1), 125–148.

Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an

application to employment equations. Review of Economic Studies, 58(2), 277–297.

BEA (various years), U.S. direct investment abroad: operations of U.S. parent companies and their foreign affiliates.

Benassy-Quere, A., Fontagne, L., & Lahreche-Revil, A. (2001). Exchange-rate strategies in the competition for

attracting foreign direct investment. Journal of the Japanese and International Economies, 15(2), 178–198.

Bernard, A. B., Jensen, J. B, & Schott, P. (2006). Trade costs, firms and productivity. Journal of Monetary

Economics, 53(5), 917–937.

Breusch, T. S., & Pagan, A. R. (1979). A simple test for heteroscedasticity and random coefficient variation.

Econometrica, 47(5), 1287–1294.

Campa, J. M. (1994). Multinational investment under uncertainty in the chemical processing industries. Journal of

International Business Studies, 25(3), 557–578.

Dixit, A. K., & Pyndick, R. S. (1994). Investment under uncertainty. Princeton: Princeton University Press.

Fahim-Nader, M. (1994). Capital expenditures by majority-owned foreign affiliates of U.S. companies. Survey of

Current Business, 9, 58–68.

Feenstra, R. C. (1998). Integration of trade and disintegration of production in the global economy. Journal of


Ghoshal, S. (1987). Global strategy: An organizing network. Strategic Management Journal, 8(5), 425–440.

Globerman, S., & Shapiro, D. (2002). Global foreign direct investment flows: The role of governance

infrastructure. World Development, 30(11), 1899–1919.

Globerman, S., & Shapiro, D. (2003). Governance infrastructure and U.S. foreign direct investment. Journal of


Grubert, H., & Mutty, J. (2000). Do taxes influence where U.S. corporations invest? National Tax Journal, 53(1),

825–840.

Gwartney, J., Skipton, C., & Lawson, R. (2001). Trade openness, income levels, and economic growth. In J.

Gwartney, R. Lawson, W. Park, & C. Skipton (Eds.), Economic freedom of the world: 2001 annual report

(pp. 1–5). Vancouver: The Fraser Institute.

Hanson, G. H., Mataloni, R. J. J. & Slaughter, M. J. (2001). Expansion strategies of U.S. multinational firms.

NBER working paper, no. 8433.

Hausmann, R., & Fernandez-Arias, E. (2000). Foreign direct investment: Good cholesterol? Inter-American

Development Bank working papers, no. 416.

Head, K., Ries, J., & Swenson, D. (1995). Agglomeration benefits and location choice: Evidence from Japanese

manufacturing investments in the United States. Journal of International Economics, 38(3), 223–247.

Helpman, E. (1984). A simple theory of international trade with multinational corporations. Journal of Political

Economy, 92(3), 451–471.

Janeba, E. (2000). Tax competition when governments lack commitment: Excess capacity as a countervailing

threat. American Economic Review, 90(5), 1508–1519.


Kinoshita, Y., & Mody, A. (2001). Private information for foreign investment in emerging countries. The

Canadian Journal of Economics, 34(2), 448–464.

Knetter, M. M., & Slaughter, M. J. (1999). Measuring product-market integration. NBER Working Paper,

no. 6969.

Kobrin, S. J. (1991). An empirical analysis of the determinants of global integration. Strategic Management

Journal, 12, 17–31 [Special issue].

Kogut, B. (1983). Foreign direct investment as a sequential process. In C. P. Kindleberger, & D. B. Audretsch

(Eds.), The multinational corporation in the 1980s. Cambridge, MA: The MIT Press.

Kogut, B. (1985a). Designing global strategies: Comparative and competitive vale-added chains. Sloan

Management Review, 26(4), 15–28.

Kogut, B. (1985b). Designing global strategies: Profiting from operational flexibility. Sloan Management Review,

27(1), 27–38.

Kogut, B. (1989). A note on global strategies. Strategic Management Journal, 10(4), 383–389.

Kogut, B., & Kulatilaka, N. (1994). Options thinking and platform investment: Investing in opportunity.

California Management Review, 36(2), 52–71.

Kotabe, M., & Swan, S. K. (1994). Offshore sourcing: Reaction, maturation, and consolidation of U.S.

multinationals. Journal of International Business Studies, 25(1), 115–140.

Lall, S., Albaladejo, M., & Zhang, J. (2004). Mapping fragmentation: Electronics and automobiles in East Asia

and Latin America. Oxford Development Studies, 32(3), 407–432.

Lehmann, A. (1999). Country risks and the investment activity of U.S. multinationals in developing countries. IMF

working paper, no. 99/133.

Levy, D. L. (1995). International sourcing and supply chain stability. Journal of International Business Studies,

26(2), 343–360.

Markusen, J. R. (1990). First mover advantages, blockaded entry, and the economics of uneven development. NBER

working paper, no. 3284.

Mataloni, R., & Fahim-Nader, M. (1996). Operations of U.S. multinational companies: Preliminary results from

the 1994 benchmark survey. Survey of Current Business, December: 11–37.

Mathur, I., & Hanagan, K. (1983). Are multinational corporations superior investment vehicles for achieving

international diversification. Journal of International Business Studies, 14(3), 135–146.

Miller, K. D. (1992). A framework for integrated risk management in international business. Journal of


Miller, K. D., & Reuer Jeffrey, J. (1998). Firm strategy and economic exposure to foreign exchange rate

movements. Journal of International Business Studies, 29(3), 493–513.

Mody, A., & Srinivasan, K. (1998). Japanese and United States firms as foreign investors: Do they march to the

same tune? The Canadian Journal of Economics, 31(4), 778–799.

Moxon, R. (1975). The motivation for investment in offshore plants: The case of the U.S. electronics industry.

Journal of International Business Studies, 6(1), 51–66.

Nicita, A., & Olarreaga, M. (2006). Trade, production and protection 1976– 2004. World Bank project.

Oetzel, J. (2005). Smaller may be beautiful but is it more risky? Assessing and managing political and economic

risk in Costa Rica. International Business Review, 14(6), 765–790.

Pahud de Mortanges, C., & Allers, V. (1996). Political risk assessment: Theory and the experience of dutch firms.

International Business Review, 5(3), 303–318.

Prater, E., Biehl, M., & Smith, M. A. (2001). Tradeoffs between flexibility and uncertainty. International Journal

of Operations & Production Management, 21(5/6), 823–839.

Qian, G. (1996). The effect of multinationality measures upon the risk-return performance of us firms.

International Business Review, 5(3), 247–265.

Raff, H., & Srinivasan, K. (1998). Tax incentives for import-substituting foreign investment: Does signaling play a

role? Journal of Public Economics, 67(2), 167–193.

Ramey, G., & Ramey, V. A. (1995). Cross-country evidence on the link between volatility and growth. American

Economic Review, 85(5), 1138–1151.

Rangan, S. (1998). Do multinationals operate flexibly? Theory and evidence. Journal of International Business

Studies, 29(2), 217–237.

Rivoli, P., & Salorio, E. (1996). Foreign direct investment and investment under uncertainty. Journal of


Rugman, A. M. (1976). Risk reduction by international diversification. Journal of International Business Studies,

7(2), 75–80.


Serven, L. (1997). Irreversibility, uncertainty and private investment: Analytical issues and some lessons for

Africa. Journal of African Economies, 6(3), 229–268.

Serven, L. (1998). Macroeconomic uncertainty and private investment in developing countries: An empirical

investigation. World Bank Working Paper, no. 2035.

Severn, A. K. (1974). Investor evaluation of foreign and domestic risk. The Journal of Finance, 29(2), 545–550.

Stevens, G. V. (2000). Politics, economics and investment: Explaining plant and equipment spending by US direct

investors in Argentina, Brazil and Mexico. Journal of International Money and Finance, 19(2), 153–183.

UNIDO. (1997). Industrial development: Global report. Oxford: Oxford University Press.

Vernon, R. (1971). Sovereignty at bay. New York: Basic Books.

Wei, S.-J. (2000). How taxing is corruption on international investors? The Review of Economics and Statistics,

82(1), 1–11.

Wheeler, D., & Mody, A. (1992). International investment location decisions: The case of US firms. Journal of

International Economics, 33(1), 57–76.

World Bank. (2004). Doing business in 2004: Understanding regulation. Oxford: Oxford University Press.

the sensitivity of u.s. multinational enterprises to political and macroeconomic uncertainty: a...

Documents