The endogeneity of the exchange rate as a determinant of FDI: A model of entry and multinational firms

https://doi.org/10.1016/j.jinteco.2006.04.004Get rights and content

Abstract

This paper argues that when the exchange rate and projected sales in the host country are jointly determined by underlying macroeconomic variables, regressions of FDI flows on both exchange rate levels and volatility are subject to bias. The results demonstrate that a multinational firm's response to exchange rate volatility will differ depending on whether the volatility arises from shocks in the firm's native or host country. It is the first study to depart from the representative-firm framework in an analysis of direct investment behavior with money.

Introduction

Foreign direct investment (FDI) has become an increasingly important channel for resource flows across national borders. In 1990, production by overseas branches of multinational firms reached 16% of the world's total manufacturing output (Lipsey, 1998). The share of FDI in total net resource flows to developing countries more than doubled during the 1990s, reaching 82% in 2001 (IMF, 2002). In industrialized countries, the ratio of foreign direct investment to total gross fixed capital formation ranges from zero to almost 50%, depending on the country and the year.1 Thus, the questions of how and why FDI responds to exchange rate fluctuations, first raised in the 1980s, have become acutely relevant to open-economy macroeconomic analysis. This paper addresses the issue by expanding the general equilibrium analysis characteristic of recent optimum currency area theory– where the exchange rate depends on “fundamental” variables that may also impact local demand– to encompass the entry behavior of the multinational firm.

Although empirical and partial-equilibrium analyses suggest that exchange rate uncertainty may be important in a firm's decision to engage in production activity overseas, the literature incorporating the multinational enterprise (MNE) into models of the global economy has remained separate from studies of exchange rate behavior and policy. Previous studies of exchange rate variability and multinational firms have often treated the exchange rate as exogenous.2 The model presented here combines the concept of an exchange rate driven by macroeconomic variables with a sunk cost, which motivates firms' sensitivity to uncertainty in the trade and industrial organization literature. The results indicate that while macroeconomic volatility in the MNE's native country and the country hosting its direct investment venture both increase exchange rate volatility, they can have quite different effects on flows of FDI.

In addition, the model here is unique in that it departs from the representative-firm framework to look at the effect of volatility on entry. The introduction of heterogeneous productivity levels across firms, based on Melitz (2002, 2003), explains why smaller, less productive firms might be deterred from investing overseas by uncertain macroeconomic conditions, while larger, more productive firms are not. A comparison of the zero-cutoff profit conditions for domestic and foreign entrants also allows the decomposition of the impact of monetary volatility into factors affecting all firms and those affecting only entering foreign firms through their exposure to exchange rate fluctuations. The analysis presented in Section 4 illustrates the two effects and provides a theoretical explanation for the observation by Hausmann and Fernandez-Arias (2000) that FDI appears to be cushioned from some types of macroeconomic risk that have been shown to curtail other types of private investment. It also reconciles the conflicting estimates of the direction of the effect of exchange rate volatility on flows of FDI evident in empirical studies based on partial equilibrium models.

The argument presented in this paper rests on two premises. First, it assumes that there is a repeated sunk cost involved in production at home and overseas — some fixed overhead cost such as legal retainers, rental and maintenance contracts, or a property tax that is paid, negotiated, or legislated in advance. In this respect, the model draws on the option value literature sparked by Pindyck (1998), Dixit and Pindyck (1994), and Campa (1993), as well as on trade models incorporating multinational firms with plant-level fixed costs (Horstmann and Markusen, 1992, Brainard, 1997) and sunk costs, which are defined here as fixed costs that must be paid before the realization of a random shock (Grossman and Razin, 1985, Helpman et al., 2004). Because the sunk cost is paid or negotiated in one period under a given exchange rate, but revenues are earned and repatriated at a later date, firms care about fluctuations in the value of the host-country currency.

Second, it is assumed that there are common macroeconomic forces that influence both the exchange rate and the volume of sales by overseas branches. These forces could involve productivity growth or any number of unobservable variables governing the international asset market. However, this study focuses on fluctuations in the growth rate of the money supply as the mechanism influencing both realizations of the exchange rate and, due to sticky prices, the demand for consumption goods in the host country. The exchange rate is a function of the ratio of the home (host-country) and foreign (native-country) money supply. It covaries negatively with the host country's demand for goods, as a positive shock to the home money supply weakens the value of the home currency but simultaneously increases real income—and therefore sales by both domestically owned firms and multinationals operating in the home market. Conversely, a contractionary monetary shock in the host country may generate a more favorable exchange rate at which to convert profits, but will depress local sales. In comparison, a contractionary monetary shock arising in the MNE's native country can adversely affect the value of the home currency with no mitigating influence (or, in the case of complete markets, an exacerbating influence) on sales overseas. Firm owners – consumers – can hedge against these macroeconomic shocks using nominal bonds, but the sunk cost and segmented goods markets introduce an asymmetry which reduces the effectiveness of the hedging when monetary volatility is not equal across countries, so that uncertainty impacts firm entry decisions nonetheless.

Thus, there are two rigidities driving the results of the model. The real rigidity, a sunk cost, motivates firms' sensitivity to uncertainty. The nominal rigidity, sticky prices, causes the uncertainty generated by monetary shocks to influence both the exchange rate and consumer demand in the host market. Because the exchange rate and the demand for goods are impacted differently by monetary shocks, depending on their origin, the analysis below indicates that studies regressing FDI flows on measures of exchange rate volatility may be subject to bias. Considering FDI and exchange rates as variables jointly determined by underlying macroeconomic factors provides an explanation of why the empirical literature analyzing whether exchange rate variability encourages or deters investment by multinational firms remains inconclusive.

The remainder of the text is organized as follows: First, insights from existing theoretical and empirical studies which guided the construction of this model are discussed in 1.1 The theoretical debate, 1.2 Empirical evidence. 2.1 The consumer's problem, 2.2 First-order conditions and the exchange rate describe the consumer's optimization problem and relevant first-order conditions. Section 2.3 defines an expression for expected discounted profits and describes the firm's pricing behavior under uncertainty. Section 2.4 explains the calculation of aggregate productivity and the price level and Section 2.5 explains how aggregate productivity is related to the “threshold” productivity level, or the labor productivity of the least productive entrant. Section 3 presents the key equilibrium conditions governing investment behavior. A special Section 3.1 is devoted to discussing issues of geographic preference and asset-market structure. An analysis of the results is presented in Section 4, where the decision criteria of foreign investors considering a direct investment venture are decomposed into factors affecting all firms operating in the host market and factors rooted in exchange-rate risk, which affect only entrants from overseas. It evaluates the net impact of monetary policy variables on entry by domestic and foreign-owned firms, as well as implications for aggregate prices and consumption in the host market. Section 5 concludes the paper with a discussion of the results and possibilities for future research.

Existing partial equilibrium models provide important insight into the mechanics of the MNE's decision-making behavior, but treat exchange rate fluctuations as exogenous, isolating them from macroeconomic shocks that simultaneously affect demand. Consequently, theoretical arguments based on these models are divided as to whether exchange rate uncertainty will increase or decrease FDI. Authors proposing that exchange rate variations could promote investment abroad assert the long-standing result in trade theory that cross-border investment is a substitute for trade when tariffs or other barriers prevent the free flow of goods (Goldberg and Kolstad, 1995, Cushman, 1985, Cushman, 1988). Mundell (1957) provides the first mathematical proof of this result. Numerous studies provide evidence that exchange rate uncertainty may function as a de facto trade barrier, implying by default that it should increase FDI.3 A related position espouses the “production flexibility” approach — that volatility increases the value of having a plant in both countries, enabling an MNE to decide at any time either to export from home or to produce in its foreign facility, depending on where conditions are most favorable (Sung and Lapan, 2000). Assuming that exchange rate fluctuations are exogenous, multinational firms can take advantage of them by shifting production to the countries where the value of the local currency makes input costs look cheapest, ceteris paribus.4 In earlier work, Itagaki (1981) develops a financial flexibility argument. He posits that an increase in exchange rate risk may incite a firm to invest abroad as a way of hedging against a short position in its balance sheet. A depreciation of the firm's home currency might reduce the value of domestic assets relative to foreign liabilities, but would simultaneously increase the value of assets and revenue streams for its affiliates in foreign countries.

However, theoretical models also exist predicting that exchange rate uncertainty will instead suppress FDI. These arguments assert that unpredictable fluctuations in the exchange rate introduce added uncertainty into both the production costs and future revenues of overseas operations, deterring potential investors. Several studies (Rivoli and Salorio, 1996, Campa, 1993), rooted in the work of Pindyck (1998) and Dixit and Pindyck (1994), declare that currency volatility deters the entry of multinational firms by increasing the “option value” associated with waiting before incurring the sunk costs necessary to produce overseas. They consider that a firm effectively holds an option to invest overseas in any given period. A fixed cost paid in advance (sunk) acts as an exercise price. The return from exercising the option is the expected present discounted value of profits earned from production in the foreign country. Exchange rate risk introduces uncertainty about the size of the return, increasing the value of holding on to the option to wait and motivating the firm to postpone investing until a future period. A salient feature of this literature is that the results hold even for risk-neutral firms, as the key engine is the sunk cost. Without it, there would be no cost to producing when the prevailing exchange rate allows positive returns and exiting when it does not, eliminating any value attached to waiting.

Several models which predict that exchange rate volatility may discourage FDI are driven instead by risk-aversion in firm management. Cushman (1985, 1988) and Goldberg and Kolstad (1995) specify conditions under which exchange rate volatility may reduce the certainty-equivalence value of expected profits from overseas operations — a deterrent to risk-averse prospective investors. The authors show that a deterrent effect arises as long as demand and the elasticity of technical substitution are of a form such that they completely offset the effect of currency fluctuations on profit remittances. More concretely, exchange rate volatility will deter FDI as long as a depreciation of the host-country currency, which would undercut the value of repatriated profits in terms of the firm's native-country currency, is not met by an offsetting increase in host-country demand or reduction in host-country input costs. Like the literature contending that uncertainty promotes FDI, these studies either omit any simultaneous effects of underlying macroeconomic variables on demand and the exchange rate, or consider a correlation between the two but do not explicitly characterize its relationship to the underlying variables.

There are two studies which incorporate multinational firms within a general equilibrium approach, where the exchange rate is endogenous, but they also generate conflicting results. The first study, Aizenman (1992), juxtaposes the production flexibility approach onto the Dixit–Pindyck-type conceptualization of the option value.5 Increases in volatility increase the value of diversification, which pushes firms to shift production to the country where it is cheapest, but also discourages investment by increasing the uncertainty surrounding the return on exercising the Dixit–Pindyck option to invest abroad. Aizenman demonstrates that a floating exchange rate will transmit the effects of country-specific shocks across national borders, which erodes the ability of firms to diversify risk by shifting production across borders. In this sense, the exchange rate volatility associated with a flexible regime can be construed as deterring FDI. In the second study, Devereux and Engel (2001) find that when firms price in the currency of the local market they are serving (pricing-to-market), production by all firms, including the affiliates of multinationals, is higher under a flexible exchange rate than a fixed exchange rate. Hence, exchange rate volatility associated with a flexible regime is loosely linked here with increased production by multinationals. These models are extremely important contributions in that they are the first to incorporate the multinational firm into an open-economy framework with money. Notwithstanding, it is difficult to interpret them in light of the question, “Does exchange rate volatility deter FDI?” because they incorporate a representative firm, implying that either all firms invest abroad or none do.

Existing empirical tests are based on the partial equilibrium models described above or on gravity models and offer two principal conclusions: (1) it is not clear what relationship exists between exchange-rate uncertainty and FDI and (2) local fixed costs make it more likely that FDI will be discouraged by exchange rate volatility. With regard to the first issue, both Cushman (1985, 1988) and Goldberg and Kolstad (1995) find that volatility increases the willingness of U.S. MNEs to locate facilities abroad, in accordance with the early trade theory explaining FDI as a substitute for exports. Zhang (2004) supports their results, finding a positive and significant relationship between exchange rate volatility and FDI flowing into the European Union (EU) from both inside and outside the EU. Nevertheless, there is ample evidence to the contrary. Whereas Galgau and Sekkat (2004) also find a positive link for flows between EU nations, they find that increases in the variance of bilateral exchange rates deter inflows originating outside of the EU. Amuedo-Dorantes and Pozo (2001) report that results may not be robust to the way volatility is measured: there is a positive coefficient associated with volatility of the exchange rate measured as a the standard deviation within a rolling window but a negative coefficient emerges when a GARCH construction is used. Both coefficients are significant. Chakrabarti and Scholnick (2002) find a negative relationship between exchange rate volatility and FDI flows from the U.S. to 20 ECD countries. Using micro-level data, Campa (1993) shows that volatility deters entry by foreign firms contemplating investment in the U.S.

There is less conflict over the influence that fixed costs exert over direct investment behavior. Two papers present evidence that the fixed costs prominent in trade and industrial-organization literature examining multinational firms are important in understanding the response of MNEs to exchange rate uncertainty.6 Campa's (1993) study shows that the negative impact of exchange rate uncertainty on entry by MNEs is more probable and more profound when sunk costs are large. He explains this phenomenon using the options theory reasoning: in the absence of sunk costs, a firm would simply produce overseas in any period when conditions were favorable, “and volatility would have no effect on the entry decision (p.619)”. Although Tomlin (2000) indicates that some of Campa's econometric results may be sensitive to model specification, she clearly demonstrates that local fixed costs, such as advertising expenses, are alone sufficient to quell entry by foreign firms deciding whether to produce in the U.S.

Finally, Goldberg and Kolstad (1995) present empirical evidence explicitly calling into question the practice of considering exchange rate volatility and shocks to demand in the host-country market as separate entities. They emphasize that for the majority of the sample used in their analysis, a depreciation of host-country currency is associated with a simultaneous increase in the host-country's demand for goods (Goldberg and Kolstad, 1995, p.866). They further show that the share of capacity firms choose to locate abroad may be affected by covariance between host country demand and exchange rate movements, but provide mixed evidence as to the direction of the relationship and its robustness across countries. The result invites the construction of a general equilibrium framework tying demand and the bilateral exchange rate to common underlying fundamental variables, depicting the multinational firm's response to the net effect that macroeconomic shocks exert on both the exchange rate and sales abroad. This paper is the first to introduce FDI into a new open-economy macroeconomic model – where exchange rates and local demand are jointly determined – while preserving the fixed-cost component emphasized in analyses of direct investment within the trade and industrial-organization literature and introducing the heterogeneity which governs entry dynamics in recent trade models.7

Section snippets

The model

The model below capitalizes on the insights of partial equilibrium theoretical examinations of MNE behavior and existing empirical evidence, while exploiting the capacity of a general equilibrium framework to connect both demand and the exchange rate to fluctuations in a common underlying variable — money. It builds on the conceptualization of MNEs put forth by Devereux and Engel (2001), the first study that has incorporated multinational firms into an Obstfeld–Rogoff-type general equilibrium

The zero-cutoff profit condition

Two constraints governing firm entry allow one to solve for the threshold productivity levels, φˆH(t) and φˆF (t). Threshold productivity levels are found at the point where a firm is just productive enough that its expected discounted profits equal zero. If expected discounted profits are lower, agents considering engaging in production activities will use funds they might have sunk into the fixed cost for next-period production to invest instead in bonds or to increase their present

The effect of exchange-rate uncertainty on entry by foreign firms

As mentioned above, the relationship between the threshold productivity levels of Home and Foreign firms operating in the Home market reveals the effect of the fundamental variables and the structural parameters governing demand on the relative willingness of Foreign investors to engage in ventures overseas. Another way to look at γ is as a parameter embodying the effect of exchange-rate risk on entry by Foreign firms. Rearranging Eq. (18),ϕˆF(t)=γϕˆH(t)=[(1+ψ1+ψ)(fMNEf)](1μ1)e(1μ1)(σm2σm

Conclusions

The goal of this paper is to explain the conflicting findings of previous empirical work done in a partial equilibrium framework by showing that volatility in the exchange rate may or may not deter foreign direct investment, depending on which underlying variable is the source of the volatility. The result here provides a theoretical account of the link between FDI flows and the correlation between local demand and exchange rate volatility investigated by Goldberg and Kolstad (1995). It bears

Acknowledgement

The author thanks Thomas Lubik and Louis Maccini, as well as Andrew Bernard, Paul Bergin, Jeffrey Campbell, Jimmy Chan, Silvio Contessi, Michael Devereux, Selim Elekdag, Robert Feenstra, Linda Goldberg, Bruce Hamilton, Yingyao Hu, Jane Ihrig, Aylin Isik-Dikmelik, Ali Khan, Michael Krause, John Rogers, Matthew Shum, Matthew Slaughter, Guy Stevens, Jonathan Wright and participants in the 12th European Conference on General Equilibrium at the Bielefeld University IMW, the IF Workshop at the

References (51)

  • B.A. Blonigen

    Firm-specific assets and the link between exchange rates and foreign direct investment

    American Economic Review

    (1997)
  • S.L. Brainard

    An empirical assessment of the proximity–concentration trade-off between multinational sales and trade

    American Economic Review

    (1997)
  • J.M. Campa

    Entry by foreign firms in the United States under exchange rate uncertainty

    Review of Economics and Statistics

    (1993)
  • R. Chakrabarti et al.

    Exchange rate regimes and foreign direct investment flows

    Weltwirtschaftliches Archiv

    (2002)
  • V.V. Chari et al.

    Can sticky price models generate volatile and persistent real exchange rates?

    Review of Economic Studies

    (2002)
  • J.H. Cochrane

    Asset pricing

    (2001)
  • A. Cote

    Exchange rate volatility and trade: a survey

  • D.O. Cushman

    Real exchange rate risk, expectations, and the level of direct investment

    Review of Economics and Statistics

    (1985)
  • D.O. Cushman

    Exchange-rate uncertainty and foreign direct investment in the United States

    Weltwirtschaftliches Archiv

    (1988)
  • A. Deaton

    Understanding consumption

    (1992)
  • G. Dell'Ariccia

    Exchange rate fluctuations and trade flows: evidence from the European Union

    IMF Staff Papers

    (1999)
  • Devereux, M.B., Engel, C., 1999, 2001. The Optimal Choice of Exchange Rate Regime: Price-Setting Rules and...
  • Devereux, M.B., Lane, P., Xu, J., in press. Exchange Rates and Monetary Policy in Emerging Market Economies. Economic...
  • A. Dixit et al.

    Investment under uncertainty

    (1994)
  • J.H. Dunning

    The determinants of international production

    Oxford Economic Papers

    (1973)
  • Cited by (0)

    View full text