Foreign Direct Investment, Exchange Rate Volatility and Political Risk Chiara Del Bo Università degli Studi, Milano [email protected]
This version: August 2009
Abstract This paper investigates the effect of exchange rate and institutional instability on the level of Foreign Direct Investment flows between developed and developing countries by presenting an empirical investigation on a panel of countries over two decades, both with cross country and cross sector data, justified by a partial equilibrium model of foreign entry. The issue is first presented with a partial equilibrium model of FDI in an oligopolistic industry , where n identical foreign firms have to decide whether to enter a host market characterized by exchange rate volatility and political risk. The results are that both exchange rate variability and political risk have a dampening effect on FDI flows, and that the interaction term is negative, indicating that the two effects reinforce each other. The econometric analysis confirms and verifies these results. The sectoral evidence points in the direction of specific industry effects, especially with respect to the role of interest rates and wages. The general conclusion regarding the role of exchange rate instability and institutional risk are confirmed, with some qualifications for the primary, financial, depository, trade and service sectors. Keywords: FDI, Exchange rate volatility, Political risk, Gravity equation. JEL Codes: F21, F31, C31
1. Introduction Foreign Direct Investments (FDI) are a significant source of flows of resources across borders, accounting for a large share of net flows, especially from developed to developing countries2 and are becoming increasingly more relevant over time. According to the World Investment Report (2008) global FDI inflows rose in 2007 by 30% to reach $1,833 billion, with $500 billion of flows directed towards less developed economies, with the three largest recipients, among developing and transition economies, being China, Hong Kong and the Russian Federation. As an illustration of the destination of FDI, the following figure constructed from UNCTAD data on FDI inflows provides some evidence on the trends of inward FDI both in developed and developing countries from 1970 to 2006.
This paper is originates from my PhD dissertation, and was written in part during my research stay at the Economics Department, Boston College. I would like to thank Fabio Ghironi, Marzio Galeotti and Giovanni Facchini for useful comments and discussions. The usual disclaimer applies. 2 There is also evidence of significant flows among OECD countries and a strand of literature examining resource flows towards developed countries, e.g. Lucas, 1990, but is not the focus of the present analysis.
1 600 000 1 400 000 1 200 000 1 000 000
600 000 400 000 200 000 2006
Figure 1: Inward FDI 1970-2006 (UNCTAD)
FDI represent one of the modes of accessing a foreign market and of internationalizing activities. A firm that decides to become international faces a number of risks that must be taken into account and analysed. These risks are associated with the overall uncertainty regarding the success of the venture and the economic, political and regulatory environment in the host country. Country specific factors that might influence a multinational's firm success can be summarized as country risk, which can be economic, commercial and political. Economic risk is related mainly to macroeconomic conditions of the host country such as the real interest rate structure and movements, and exchange rate volatility. Commercial risk is related to enforceability of contracts with local partners and suppliers. Finally, political risk can be a very important factor and is related to the possibility of changes in the regulatory and operational framework, such as changes in price controls, instability of the legal system and its enforcement and, last but not least, the risk of expropriation. The latter refers to the possibility that the local government decides to expropriate the foreign capital completely or forces the owner to give local partners ownership shares, below face value.3 As an illustration, we report results of a survey conducted on European parent companies investing in Southern Africa4 illustrating the percentage of respondents identifying particular risk factors associated with the host economy. The two most common risk factors that emerge from this survey are unstable macroeconomic environments, characterised by exchange rate volatility, and regulatory uncertainty. These considerations lead to the question of understanding the effect of exchange rate variability and political risk, seen as the possibility of expropriation or changes in the regulatory and taxation framework by the host government, on the firm's decision to access foreign markets, focusing specifically on Horizontal FDI,5 which are defined by the IMF as ``a category of international investment that reflects the objective of a resident in one economy (the direct investor) obtaining a lasting interest in an enterprise resident in another economy (the direct investment enterprise)”. The lasting interest implies the existence of a long-term relationship between the direct investor and the direct investment enterprise, and significant degree of influence by the investor on the management of the enterprise. A direct investment relationship is established when the direct investor has 3 Nordal, 2001. 4 Source: Jenkins and Thomas, 2002. 5 There is also the interesting aspect of Vertical FDI, which involves the fragmentation of stages of production, but will not be addressed here.
acquired 10 percent or more of the ordinary shares or voting power of an enterprise abroad. Given that a large portion of FDI flows are towards developing countries, the role of exchange rate variability is also important. Therefore, an interesting issue to be addressed, which is the focus of the present paper, is the combined effect of exchange rate volatility and political risk on the amount and direction of FDI. The paper is structured as follows: Section 2 briefly summarizes the relevant literature, while Section 3 sets out the research questions. The partial equilibrium oligopoly model is presented in Section 4, as a basis and justification for the country level empirical analysis in Section 5, with a specific focus on the interaction effect of ER volatility and institutional risk. Section 6 presents the sector-level data and econometric analysis. Section 7 summarizes and concludes, highlighting possible policy implications.
2. Related Literature The effect of political risk, in the form of expropriation, changes in law and regulations, and the institutional framework in the host country of FDI has been assessed through several empirical studies, using both cross country comparisons and panel data sets. A detailed literature review can be found in Busse, 2004. The main findings from cross country studies have highlighted the importance of the institutional framework of the receiving country on the flow of private investment. Brunetti and Weder, 1998, find a significant and negative effect of institutional uncertainty and FDI, while Lee and Mansfield, 1996, find a positive effect of intellectual property rights protection and private investment flows. More recently, Wei, 2000, reports a negative impact of corruption on inward FDI. Loree and Guisinger, 1995, find mixed results on the effect of political stability on FDI. Other recent contributions take into consideration also the time series dimension, by exploiting panel data sets. Jun and Singh, 1996, and Gastanaga et al., 1998, find a significant and negative statistical effect of measures of political risk and other political variables (such as corruption, nationalization risk and imperfect contract enforceability) on flows of private investment. The role of democracy is also explored, with studies by Busse, 2004, Harms and Ursprung, 2002 and Jensen, 2003, showing that countries with democratic regimes are more likely to attract FDI inflows. However, Li and Resnick, 2003 and Sethi et al., 2003, fail to find a significant effect of political stability on FDI flows. Li, 2006, also analyzes the effect of different measures of political violence, both on the location choice and on the amount of FDI, finding significant results and highlighting that political violence and instability affect the location choice and the investment amount differently. In a recent contribution, Busse and Heffeker, 2007, explore the relations between political risks, institutions and FDI flows using a sample of 83 developing countries over a 20 year period. Their findings suggest that several indicators for political risk and institutional quality are significant determinants of FDI inflows. The theoretical literature has also contributed to the understanding of the link between political risk, especially in the form of expropriation, and direct investment. The design of optimal contracts between the host government and the MNE is first analyzed in a static setting (Eaton, Gersovitz, 1984), then extended to a fully dynamic environment (Thomas and Worrall, 1994). Eaton and Gersovitz propose a theory of expropriation based on the optimizing behaviour of host country government and foreign investors. They highlight the role of the threat of expropriation, rather than the act itself, and show how this policy can be harmful for the host country's welfare. The authors 3
also examine the role of risky projects (with risks other than that of expropriation) and assume risk neutral firms and risk averse host countries. Thomas and Worrall stress the importance of limited contract enforceability between multinational enterprises (MNE) and local governments, and show that the short term incentive to expropriate is actually outweighed by the long term incentive to attract more FDI in the future. The analysis is then extended to determine the form of the optimal self enforcing dynamic contract. More recent literature endogenizes the decision of the host government to expropriate by introducing knowledge and technological spillovers from the MNE to local firms, in the context of imperfect contract enforceability. Two interesting contributions are Albuquerque, forthcoming, and Maliar et al., 2005. The first paper endogenizes inalienability through spillovers in the form of a human capital externality and organizational capital that flow from foreign-owned production establishments to the host country, building from the endogenous growth literature. Maliar et al. predict that when spillovers are large a developing host country will not expropriate foreign capital and will behave as under perfect enforcement of foreigners' property rights. Otherwise, when spillovers are insignificant, there will be expropriation and the long term outcome will be financial autarky, since no FDI flows will take place. The second strand of literature is related to the effect of exchange rate (ER) volatility on FDI. A recent theoretical contribution is Russ, 2007, which analyzes the effect of exchange rate volatility on Horizontal FDI flows and multinational firms' decision to enter foreign markets in the context of a general equilibrium model, such that the exchange rate is endogenous, in which heterogeneous firms face repeated sunk costs in production at home and over seas. The main findings are that macroeconomic volatility (here due to monetary shocks) in the firm's native and host country both increase ER volatility. However, the firm's response to ER volatility depends on whether the underlying shocks originate in the home or host country. The effects also depend on the firm's specific productivity level. Smaller and less productive firms may be discouraged from investing overseas by macroeconomic instability, while larger and more productive firms are not. The origin of monetary volatility is important when analyzing the entry and investment decisions of foreign and native firms. Exchange-rate variability can mitigate the effects of uncertainty in the hostcountry money supply on FDI encouraging FDI. Monetary volatility in a firm's home market introduces exchange rate risk without offsetting effects on sales, deterring FDI. Other theoretical contributions to this topic include Aizenman, 1993, and Goldberg and Kolstad, 1995. The empirical evidence on the effect of ER variability on FDI is somewhat mixed. While Campa, 1993, finds a negative effect of volatility for FDI, mainly due to the presence of fixed costs, other authors, such as Cushman, 1985 and 1988, and Goldberg and Kolstad ,1995, find that ER volatility increases FDI by US firms, while Zhang, 2001 supports these results, for FDI within the EU. A paper that combines these aspects together and examines the role of exchange rate risk, contract enforceability and trade is Sayinta, 2001. Given the mixed empirical results on the sign of the effect of ER risk on trade, the author sets up a model with risk neutral agents6 and imperfect contract enforceability and shows that ER volatility depresses trade when contractual insecurity is high, then the effect is reversed when insecurity decreases, and the effect dies out in the extreme case of perfect enforcement. The original contributions of this paper are the modelling of imperfect contract enforceability through a parameter, θ , which is a parametric probability that a defaulter is forced to 6 Most of the previous literature in trade effects of exchange rate variability assumed risk-aversion. Relaxing this assumption and taking into account the possibility if hedging is another contribution of this paper.
execute his contract, and its interaction with exchange rate risk in determining the overall effects on trade.
3. Motivation and Research Question This paper aims at examining the role of two determinants of FDI inflows, reflecting macroeconomic conditions of instability in the host country, namely exchange rate volatility and institutional or political risk. The analysis is conducted in three progressive stages. First, the macro level is considered, both from a theoretical and empirical standpoint, by examining the separate and combined effect of the two risk measures on foreign capital inflows. This analysis confirms the dampening effect of both variables, and a negative marginal effect of institutional risk when exchange rate volatility increases, indicating a negative complementarity between these risk indicators. Moving on to a meso-level, sectoral data are considered, in order to investigate/gauge the magnitude and relevance of ER and political risk on disaggregated foreign investments. The dampening effect of both variables on FDI inflows is retained, and sector-specific patterns emerge. In this case data is stacked by sectors, and the relevant interaction effects are those obtained by interacting industry dummies with the variables of interest. Finally, as an additional consistency check, other relevant FDI determinants (namely wage and interest rate variables) are included in the analysis, shedding light on the overall mechanisms underlying foreign capital movements. For a multinational enterprise, FDI can be considered as an internationalization strategy for different purposes. It can be used as a mode of entry in the host country market and is therefore a substitute for exports, with the multinational facing the choice between investing in that country or exporting directly into it. FDI can also take place to take advantage of specific characteristics of the host country in terms of input availability and costs or taxation regimes, and the output of production is then exported back to the home or a third consuming country. In the latter case, the multinational firm has a choice between several possible host countries; therefore, the number of foreign firms in a specific host country is endogenous. The focus of this paper is the analysis of the effect of exchange rate variability and political risk on the FDI flows. Therefore, we don't consider the case of FDI as a substitute for exports in the host country. Following Lahiri and Ono, 1998, 7 we make use of the concept of a small open economy in FDI in which there is free entry and exit of foreign firms. The outside option for the multinationals is to invest in another host country, and this defines the free entry condition. In the theoretical model presented in the next Section, we consider an oligopolistic environment, with multinational firms competing strategically in a Cournot fashion. This choice is motivated by previous literature, which has stressed how FDI is indeed associated with imperfect competition, strategic choices and the presence of economics rents.8 Several papers that explore optimal tax policy in the presence of economic rents earned by multinational enterprises that rely on imperfect competition, following the seminal paper by Brander and Spencer, 1985. 9 While the separate effects of various measures and indicators of political risk and ER volatility on the flows of FDI to developing countries have been object of several empirical and theoretical 7 In several papers, summarized in Lahiri, Ono, 2004, the focus of the authors is the analysis of optimal host country policies to attract FDI. 8 Justification for the oligopolistic setting can be found in Devereux, Hubbard, 2003, and Buch et al. 2005. 9 Levinshon, Slemrod,1993, and Janeba, 1996.
studies, the issue of the interaction of these two forms of risks facing the multinational firm has been analyzed mainly with respect to trade and export flows, and not to FDI. This paper aims at filling this gap by analysing the combined effexr of ER volatility (i.e. ER risk) and political risk on FDI outflows. There also seems to be scarce evidence and analysis on the differential impact of these variables in different sectors, which could be related to different capital intensity, technology content and other sector-specific characteristics. We therefore extend the analysis to take into account cross-sector differences, by analyzing sector specific FDI inflows, with political risk and institutional measures as control variables. Results of this empirical investigation are presented in Section 6.
4. Theoretical Background As a theoretical framework for the empirical analysis, we consider a two country, partial equilibrium model of an oligopolistic industry in which n identical foreign multinational firms seek to enter the host country through FDI, set up production facilities, using local inputs, and then sell the output back in the home country. We assume that foreign firms do not face domestic competition since there are no local firms producing the good in the host country. The n foreign firms' headquarters are located in country F (the foreign home country), while their production facilities are located in country D (the domestic host country). Output produced in country D is then sold back into country F in home country prices. Host country inputs are paid by the foreign firms in domestic currency and there is a form of political risk associated with the host country, modelled as the possibility of partial expropriation of output.10 The final produced good x serves both for consumption and investment purposes, justifying how expropriation risk enters the model. The multinational firms face an inverse demand function in the home country F given by: (1.) PF = α − β D where D = nxF is the market clearing condition for the good, where x F is the quantity produced by each foreign firm. Therefore: (2.) PF = α − β xF We assume that firms face linear production costs, so that marginal costs coincide with average ) variable costs. Marginal costs C F are paid in the local currency and are therefore subject to ) ) ) exchange rate variability. Specifically: C F = ε CD , where ε is the exchange rate between the domestic and foreign country. The exchange rate is considered an exogenous random variable. Firms also face a political risk in the host country, modelled as the possibility of expropriation of output, captured by the parameter θ ∈ [ 0,1[ . Multinational's profits, expressed in terms of home country currency are: ) ) ) (3.) π = PF xF − C F xF − θ xF = PF xF − ε CD xF − θ xF = [α − β nxF ] xF − ε CD xF − θ xF If we allow for the exchange rate to be distributed as a Normal random variable, we can apply the concept of certainty equivalence of profits, as has been done by Cushman, 1985 and Lahiri, Mesa, 2006. 10 The backbone for this model is the work in Lahiri, Ono, 1998 and Lahiri, Mesa, 2006.
This assumption has been frequently used both in the empirical and theoretical literature.11 Assuming that the bilateral exchange rate ε follows a log-normal distribution: ln ε ≈ N (µ ε , σ ε2 ) , µ + 1σ 2
then µε = e
and σ ε2 = e 2 µ +σ (eσ − 1) = e
2( µ + 12σ 2 )
(eσ − 1) = e
( µ + 12σ 2 )2
(eσ − 1) = µε2 (eσ − 1) . 2
So: σ ε = µε (eσ − 1)1/2 . In this case, the certainty equivalence of profit for each multinational firm is: ) ) (4.) π CE = E PF − C F − θ xF − γ F StD C F xF which can be re-written as: (5.) π CE = PF xF − θ xF − aF CD xF = [ PF − θ − aF CD ] xF , where aF = µε + γ F σ ε . The firm's profit maximization problem can be expressed as: max π CE . xF
The first order conditions, making use of Cournot conjecture, are: ∂π CE (6.) =0 ∂xF
∂π CE =0 ∂xF From the first order conditions, using equation (2.) and the definition of costs, we derive the expression for optimal output: α − θ − aF C D (7.) xF = β [ n + 1] The number n of foreign multinational firms investing in the host country is endogenous, and is defined by the following entry condition: (8.) π = π R where π R are the reservation profits that the firm can obtain by investing elsewhere. This entry condition is based on the idea that the host country is a small open economy for FDI, and that multinational firms have the outside option of investing elsewhere, reaping the reservation profits πR. Substituting the equilibrium value of xF in the expression of profits, and then taking into account the free entry condition (8.), we can derive an expression for the number of firms entering the market as a function of the relevant parameters, to be seen as a proxy for entry of Foreign capital in the Domestic country. Formally: (α − θ − a F C D ) (9.) [ n + 1] =
In order to evaluate the effect of exchange rate volatility and political risk on the number of foreign firms (and consequently of the volume of FDI inflows), we take the total differential of the equation (9.) and obtain: 2(α − θ − aF CD ) 2(α − θ − aF CD ) (10.) 2 ( n + 1) dn = − dθ − daF R R
11 This specification for the distribution of the exchange rate is typical in the literature. See Cushman, 1985, Goldberg, Kolstad,1995, Lahiri, Mesa, 2006.
From this expression the direct effect of a change in θ (political risk) on the number of Foreign firms entering the Domestic market n, and consequently on the amount of FDI flows, can be found by holding σε (exchange rate volatility) fixed: dn α − θ − aF C D (11.) =− 0 , used in Equation (12.): ∂σ 2
Since we defined aF = µε + γ F σ ε , we can re-write it as:
aF = µε + γ F µε (eσ − 1)1/2 = µε 1 + γ F (eσ − 1)1/2 > 0 2
We are also interested in the sign of the derivative of αF with respect to σε, so we compute
∂aF ∂σ 2
which is positive. Formally: 1
2 µ+ σ aF = µε 1 + γ F (eσ − 1)1/2 = e 2
∂aF ∂σ 2
= µε 1 + γ F
2eσ −1 1 2 ( eσ −1) 2
1 + γ (eσ 2 − 1)1/2 F
Appendix B: Data sources and description Data sources and frequency The sources of data that match the desired characteristics are described below and summarized in Table B.8.
Variable FDI outflows
Bilateral Exchange Rates Institutional Risk Measure Control Variables
Source Bureau of Economic Analysis, Direct Investment Position OECD Federal Reserve of St. Louis International Country Risk Guide, Political Risk Service Group World Development Indicators CEPII
Frequency yearly, 1982-1997 yearly, 1997-2005 monthly, 1982-2005 monthly, 1982-2005 yearly, 1982-2005 yearly, 1982-2005
Table B.8: Data Sources and Frequencies
PRS Risk measures Corruption: A measure of corruption within the political system that is a threat to foreign investment by distorting the economic and financial environment, reducing the efficiency of government and business and introducing instability into the political process. Scale: 0-6 Economic Risk Rating: A means of assessing a country’s current economic strengths and weaknesses. Comparability between countries is ensured since risk components are based on accepted ratios between measured data and the ratios are compared, not the original data. Scale: 050 Financial Risk Rating: A means of assessing a country’s ability to pay its way by financing its official, commercial and trade debt obligations. Comparability between countries is ensured since risk components are based on accepted ratios between measured data and the ratios are compared, not the original data. Scale: 0-50 Investment Profile: A measure of the government’s attitude toward inward investment as determined by four components: the risk to operations, taxation, repatriation and labor costs. Scale: 0-12 Political Risk Rating: A means of assessing the political stability of a country on a comparable basis with other countries by assessing risk points for the component factors of government stability, socioeconomic conditions, investment profile, internal conflict, external conflict, corruption, military in politics, religious tensions, law and order, ethnic tensions, democratic accountability and bureaucracy quality. Scale: 0:100 Socioeconomic Conditions: an estimate of the general public’s satisfaction or dissatisfaction with the government’s economic policies, covering a broad spectrum of factors, ranging from infant mortality to housing and interest rates. Different weights are applied in different societies, depending upon the relative political impact. Scale: 0-12