Real Effective Exchange Rates and Foreign Direct Investment Inflows: Empirical Evidence from India’s Sub-National Economies

This paper investigates the impact of real effective exchange rates (REER), both in terms of levels and volatility, on foreign direct investment (FDI) inflows for a panel of 35 Indian sub-national economies over the period 2000-2013. In light of the asymmetric distribution of FDI inflows within India, we focus on examining the nexus between FDI inflows at the sub-national level and India’s competitiveness captured by REER. Our empirical analysis reveals that movements in REER have a significant and negative impact on FDI inflows, while REER volatility is found to be inducing FDI. Our results are suggestive that FDI inflows into India are largely domestic market oriented in nature. Purpose: In light of the asymmetric distribution of FDI inflows within India, we focus on examining the nexus between foreign direct investment (FDI) inflows at the sub-national level and India’s competitiveness captured by real effective exchange rates (REER). This paper investigates the impact of REER, both in terms of levels and volatility, on FDI inflows to 35 Indian sub-national economies over the period 2000-2013. Research Methodology: To examine the impact of REER on FDI inflows, we compile a panel dataset for 35 sub-national economies covering the time period 2000 to 2013. We employ panel fixed effects www.scholink.org/ojs/index.php/jepf Journal of Economics and Public Finance Vol. 6, No. 2, 2020 79 Published by SCHOLINK INC. models to explore our relationship of interest between REER and FDI, controlling for other characteristics specific to a sub-national economy. Findings: Our empirical analysis reveals that movements in REER have a significant and negative impact on FDI inflows, while REER volatility is found to be inducing FDI. Our results are suggestive that FDI inflows into India are largely domestic market-oriented in nature. Originality/Value: Considering that India’s FDI inflows exhibit significant concentration patterns among selected regions, we exploit this heterogeneity at the sub-national level to empirically understand the determinants of FDI, with a particular focus on cost competitiveness as captured by REER. The extant literature has not explicitly focused on testing the impact of REER both in terms of its levels and volatility on FDI inflows to India at the sub-national level, especially not at the sub-national level. While admittedly the exchange rate varies only at the national level, the value-addition comes from understanding its interaction with state-varying macroeconomic indicators.


Figure 1. FDI Inflows to India (US$ Billion) and as a Percent of GDP and World FDI, 1990-2014
Source: Reserve Bank of India and World Bank.
There is a well-established academic literature that points to a large number of factors that determine FDI inflows especially into emerging and developing economies (See Blonigen, 2005 for a comprehensive review of this literature). Studies such as Sahoo (2012) and Sahiti et al. (2018) have identified market size, labour cost, trade openness, infrastructure, economic reforms and labour quality as determining factors for FDI inflows. One of the many determinants on FDI relates to the movement of exchange rates both in terms of level and volatility. At a very basic level, when an economy experiences a depreciation of its currency for example, viz. the value of its currency declines relative to another currency or a basket of currencies, it potentially improves the attractiveness of that country as a destination for FDI inflows as the country gains a "locational advantage" as a result of a possible reduction in its wages and costs of production, ceteris paribus (Froot & Stein, 1991;Klein & Rosengren, 1994;Goldberg, 2009). Although there are various other confounding factors such as future expectations of exchange rates that matter in order to empirically determine the extent to which exchange rate movements affect FDI inflows, the broader point to note is that cost competitiveness remains one of the crucial determining variables affecting FDI inflows. Popovici and Calin (2015) in their study examine the impact of enhancing competitiveness on FDI inflows for Central and Eastern European countries. Their findings reveal that FDI inflows can be increased by improving competitiveness variables. For most emerging and developing economies like India, remaining cost competitive has become a pre-requisite to continue being an attractive destination for global FDI inflows. In light of this background, this paper examines the impact of cost competitiveness, broadly proxied by Real Effective Exchange Rates (REER) on FDI inflows in India.
While there is some literature to date that studies the relationship between exchange rates and FDI in the context of India, most of the literature investigates this relationship at the aggregate level.
Considering that India's FDI inflows exhibit significant concentration patterns among selected regions, we exploit this heterogeneity at the sub-national level to empirically understand the determinants of FDI, with a particular focus on cost competitiveness as captured by REER. Further, consistent with the related literature that points to varying levels of competitiveness (Note 3) and governance structures observed in India, undertaking an empirical analysis at the sub-national level is warranted.
The remainder of the paper is structured as follows. Section 2 will provide an overview of the FDI trends and patterns in India at the sub-national level. Section 3 will discuss the theoretical and empirical literature on determinants of FDI, specifically focusing on the nexus between exchange rates and FDI. Section 4 will outline the empirical model employed in the paper, along with details on the data sources and definitions. Section 5 with furnish the empirical results as well as the robustness checks, while Section 6 summarizes the paper highlighting some policy implications.   Maharashtra's share of FDI at par with the national average ( Figure 4).

Figure 4. FDI (Percent of GSDP) of Top 7 Indian Sub-national Economies
Source: Authors.
In the relatively bigger sub-national economies (in terms of GSDP) such as Delhi, Karnataka, and Tamil Nadu, the share of FDI as a proportion of their output has more than doubled during the corresponding period. Evidently, Delhi has shown the largest increase in the FDI as a share of GSDP, representing an increase from 4.7 per cent in 2001 to 9.7 per cent in 2013. However, the main surprises came from Gujarat, Chhattisgarh and Madhya Pradesh. In Gujarat, FDI inflows as a proportion of GSDP, increased from about 0.1 per cent in 2001 to 0.7 per cent in 2013 this translates to an approximate increase of 7 times.
To sum up, our discussion on the trends and patterns of FDI inflows to India at the sub-national makes it apparent that FDI inflows to India are highly skewed towards selected regions. This warrants the need of examining FDI patterns at a disaggregated level. In Section 3, we offer a discussion of the theoretical and empirical literature examining the impact of exchange rates on FDI before proceeding with our empirical analysis.

Literature Review
The following discussion proceeds in two parts. The first part provides an overview of the theoretical and empirical literature addressing the nexus between exchange rate movements and FDI inflows. The second part of the section focuses specifically on the relevant literature for India.

Overview of Literature
How do movements in REER, both in terms of their levels and volatility, affect FDI inflows? The theoretical and empirical literature seems ambiguous at best. However, notwithstanding the ambiguity, the literature points to some directions as to what to expect from the nexus between exchange rate movements both in terms of their levels and volatility and FDI inflows.
The literature posits that the exchange rate effects operate broadly through the valuation channel which can affect FDI inflows through three specific ways. A positive relationship between host country's currency depreciation and its FDI inflows can come about in three specific ways through the valuation channel (Froot & Stein, 1991;Goldberg & Klein, 1998;Blonigen, 1997 reasonable degree will insulate it against the risk of uncertainties imposed by exchange rate volatility. Empirically, some notable papers such as Cushman (1988), Stokman and Vlar (1996), De MÈnil (1999), Pain and Welsum (2003) find a significantly positive relationship between REER volatility and FDI inflows in the host country.
On the other hand, a body of literature also suggests that higher exchange rate volatility can deter firms from moving to the host country because of risk aversion reasons. To be sure, a firm planning to undertake an investment in a country that is prone to greater exchange rate fluctuations might imply a riskier stream of profits. This coupled with the sunk costs involved in the investment activity would encourage the firm to place on hold its investment rather than undertake it (See for instance the discussion in Goldberg (2009) and Foad (2005)). In other words, as summarised by Goldberg (2009) Similar to the case of movements of REER (in levels), one of the key points to be noted from the related literature is that the relationship between REER volatility and FDI inflows depends on the nature of FDI inflows. If FDI is horizontal or domestic-market oriented in nature, an increase in REER volatility could induce FDI because costs of exporting becomes high which leads firms to serve the domestic market by establishing a base in the host country. The intuition, as alluded to earlier, is that the firm attempts to establish an early base in the country in order to avoid dealing with exchange rate risks since they already know that they are going to serve the specific host country market. However, if the nature of FDI is vertical or export-oriented, we expect to see a negative relationship between REER volatility and FDI, viz. an increase in REER volatility is likely to deter FDI, for risk aversion reasons elaborated earlier (Note 6).
Overall, the impact of REER both in terms of levels and volatility may have an ambiguous impact on FDI reflecting the complex nature of the relationship governing the variables of interest.

Literature on India
Unlike the vast literature on the determinants of FDI which exist for advanced economies, studies Dua and Garg (2015)  A small but growing set of studies have departed from the aggregate focus on India as a whole and factored in the regional inequality in distribution of FDI inflows into India. Mukherjee (2011) for example focuses on the regional inequality in the FDI flows to India and finds a positive association of FDI inflows to a particular region with the region's market size, agglomeration effects and size of its manufacturing and services base. A similar conclusion has been drawn by Tsuchiya (2015), who performs a region and sector wise analysis of India's FDI inflows using yearly data from 2008 to 2013.
Clearly, to the best of our knowledge, the extant literature has not explicitly focused on testing the impact of REER both in terms of its levels and volatility on FDI inflows to India at the sub-national level, exploiting the variation using disaggregated data. While admittedly the exchange rate varies only at the national level, the value-addition comes from understanding its interaction with state-varying macroeconomic indicators.

Data and Empirical Model
To examine the impact of exchange rate on FDI inflows, we compile a panel dataset for 36 sub-national economies using annual data from 2000 to 2013. We employ panel fixed effects models to explore our relationship of interest between REER and FDI.
The basic estimating equation will be to understand the impact of REER on FDI inflows to India's sub-national economies, controlling for other characteristics specific to a sub-national economy. The baseline equation takes the form: Where, refers to FDI inflows to a sub-national economy i at time t; represents the time-varying independent variable captured by REER index; represents a matrix of control variables measured at time t for a sub-national economy i.
represents a set of control variables in sub-national economy i capturing entity fixed effects is the idiosyncratic error term.
In the equation (1), are the parameters to be estimated.
For our study we expect to see a positive relationship between the depreciation of the host country's currency and its FDI inflows. Our control variables are informed by the related literature (Blonigen, 2005;  Paved Roads in Length: assesses the extent of infrastructure development in a sub-national economy; quality physical infrastructure helps reduce costs of production for firms which induces FDI. A priori, we expect to see a positive impact of improved infrastructure on FDI flows. We first estimate equation (1) to capture the impact of REEER movements in levels on FDI inflows before controlling for volatility of REER to understand its impacts on FDI inflows. In the baseline model, we use a simple measure of REER volatility as calculated by the standard deviation of the monthly REER index, while we use coefficient of variation of the REER series as a robustness check.
Two primary econometric problems can potentially produce biased estimates in the specified empirical model. One is the classic issue of simultaneity bias or reverse causality which remains an unresolved issue in the decades old general exchange rate-FDI literature. The other is that of endogeneity that arises from omitted variable bias in specifying the model. Our panel data estimation can handle the concern of omitted variable bias to a reasonable extent by incorporating entity fixed effects.
It has been well established that such estimation allows us to control for unobserved entity-specific fixed characteristics that might affect the impact of REER on FDI. We expect the estimates of the  Source: Authors.
The second part of our empirical analysis is to capture the possible effects of an exchange rate expectations on FDI inflows. Specifically, we use accumulation of foreign exchange reserves and interact it with REER to capture expectations of sustained appreciation in the country's exchange rate.
When a country intervenes in the foreign exchange market and builds foreign exchange reserves, thereby attempting to prevent its currency from appreciating through sustained reserve accumulation, there is a likelihood of market expectations of further appreciation in the future. This is reflected in equation (2) given below.
= 0 + 1 + 2 + + + 3 + 4 * + + If reserve changes are a good proxy for sustained REER appreciation, then a priori we should observe a positive relationship between REER and FDI implying that a REER appreciation is likely to induce domestic-market oriented FDI.

Empirical Results
We start with our fixed effects estimation outlined in equation (1), the results of which are summarized in Table 2. We build our model by assessing the importance of several macroeconomic, institutional and financial determinants of FDI (Columns 1 to 3) before testing for the specific impact of REER and REER volatility (Columns 4 and 5). Note. Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1; State FE included.
The results offer some interesting insights. First, with regard to the key variables of interest, we find REER to be consistently statistically significant and negatively associated with FDI inflows, across all specifications. This implies that a host country's currency appreciation measured by an increase in REER deters FDI inflows in the host country.
Recall (see section 3.1) that regardless of whether FDI inflows are export-promoting (vertical) or domestic-oriented (horizontal) in nature, a REER appreciation deters FDI, with the only difference being that in the case of the latter, the channel of impact operates through wealth effects. Consistent with our discussion, our results suggest that a REER appreciation reduces FDI inflows, viz. a 10 percentage point increase in REER is associated with a reduction in FDI inflows as a share of GDP by .0025 percentage points.
It must be noted that the coefficient of REER is only weakly significant at the 10 percent level.
However, when we factor in REER volatility, we find that the statistical significance of REER improves with the variable being significant at the 5 percent level, while still consistently exerting a negative impact on FDI inflows to sub-national economies.
In contrast to the results for REER movements in levels, the effect of REER volatility on FDI inflows to India is positive. That is, a 10 percentage point increase in REER (appreciation) is associated with an increase in FDI inflows to sub-national economies as a share of GDP by .0173 percentage points, which is economically quite significant.
As explained earlier in the paper, if FDI is domestic-market oriented in nature, an increase in REER volatility is likely to induce FDI as firms that have decided to serve the domestic market establish a base in the host country (in light of higher costs of exporting) to avoid dealing with exchange rate risks since they already know that they are going to serve the specific host country market. positive sign implying that sub-national economies with a larger market size will attract more FDI.
Higher inflation has a significant and negative impact on FDI inflows across all specifications, suggesting that sub-national economies with higher inflation tend to deter FDI, which also conforms to the priors. Other significant determinants of FDI include the positive and significant role of human capital in attracting FDI and a positive yet weak statistical significance for credit creation, proxying for the level of financial sector development in sub-national economies.
Considering the possibility that there could be reverse causality between FDI and financial sector development, we drop this variable from our regression and re-run our model with the same set of control variables used before.  Finally, we run our augmented model to test the effects of exchange rate expectations on FDI inflows.
In the augmented regression specification, as noted earlier, we introduce the changes in foreign exchange reserves as an additional control variable and also interact it with REER to capture the effect of expected exchange rate appreciation on FDI. Our main results continue to remain robust and the coefficient of the interaction term carries the appropriate positive sign though it is not statistically significant.

Robustness Checks
We perform two kinds of robustness checks to ascertain the consistency of the findings of our baseline fixed effects estimates and that of the augmented model. The first type of robustness check is to use an alternative definition of REER volatility and the second type is to re-run our empirics using alternative series of REER.
The results using coefficient of variation as a measure of volatility is given in Table 4. The results indicate that exchange rate volatility, measured by coefficient of variation, is positively and significantly associated with FDI inflows, and has consistent signs with the baseline model.
Interestingly, exchange rate is negative but becomes significant at the 10 percent level when we use coefficient of variation as a measure of volatility. Finally, the interaction term of reserves and REER remains positive and insignificant, akin to our results in Table 3.  Table 5 also suggest that most series are highly correlated. Some are perfectly correlated like BIS and FRED, while some like Bruegel and BIS are highly correlated but not complete (0.87).
In light of the above, we re-estimate our model using Bruegel REER series. Table 6 shows that exchange rate volatility continues to be positive and significant at 5 percent level even when we use an alternative REER series. REER is negative but insignificant, consistent with our main findings in Table   3. Finally, it is worth noting that the signs and significance levels of the control variables such as population, inflation, and student-teacher ratio also concur with our main findings. Source: Authors.
The findings from both the robustness checks, using an alternative volatility measure and REER series, corroborates the main findings of this paper thus validating the robustness of our model. We undertake additional robustness checks (Note 8) using Bruegel REER series with an alternative measure of volatility (coefficient of variation) and using alternative REER RBI (TWB) and RBI (EWB) series.  The use of national level REER to evaluate its impact on the FDI inflows at the sub-national level may be viewed as a limitation of our study. However, building a sub-national level REER series is beyond the scope of our study. Yan et al. (2016) have constructed provincial-level REER indices to study the effect of REER on regional economic growth in China. It would be interesting to construct a similar state-level index for India for our future research on the impact of REER on FDI inflows to India.