I. Introduction

There is by now a large literature that has focused on the effects of foreign direct investment (FDI) on growth. The overall consensus appears to be that, provided that the right economic environment is in place, FDI will indeed stimulate growth.¹ Much less attention, if at all, has been paid to the link between capital inflows volatility and output volatility.² In particular, there has been little formal analysis of the idea – often found in policy circles – that FDI should be encouraged because it should lead to lower output volatility. As the logic goes, FDI is more stable than other sources of capital inflows, most notably portfolio and other investments (OTR), and therefore should be encouraged in order to ensure a less volatile level of domestic output.

This paper tackles head on the question of whether more FDI leads to a less volatile level of output.³ To organize the discussion and provide a guide to the empirical analysis, we first develop a simple empirical model of the relationship between output and capital inflows. The model draws on standard portfolio theory in which the volatility of an investor’s portfolio depends on the volatilities of the underlying investments. We show that output volatility depends not only on the volatility of FDI and OTR but also on the correlation between FDI and OTR and the share of FDI in total capital inflows.

The model calls attention to some important caveats that need to be taken into account for some commonly-held beliefs to be true. For example, if would seem intuitively obvious that lower FDI volatility should lead to lower output volatility. This is not, however, necessarily the case. In fact, if the correlation between FDI and OTR is negative, then lower FDI volatility will increase output volatility because FDI cannot provide as much insurance against the volatility of OTR. By the same token, another “obvious” idea – that a higher share of should lead to lower output volatility – is only true in the model if the actual share of FDI in total capital inflows is below the share of FDI that minimizes overall output variability.

We use the model to derive three key testable implications:

• If the correlation between FDI and OTR is zero or positive, output volatility should also depend positively on FDI volatility and OTR volatility.

• Output volatility should be an increasing function of the correlation between FDI and OTR

• Output volatility should be a decreasing function of the share of FDI in total capital inflows, particularly when its initial value is low.

We test the model’s predictions using a sample of 59 countries for the period 1970-2009⁴. For this purpose, we construct five-year non-overlapping series of volatilities and other portfolio and macroeconomic variables. Our empirical findings strongly support our model’s implications.

We control for other possible determinants of output volatility, such as government spending volatility, terms of trade volatility, and country instability. We address endogeneity concerns by using three sets of instruments: (i) five-year non-overlapping lags of portfolio variables, (ii) gravity-based portfolio variables aiming at capturing regional/location effects, and (iii) five-year non-overlapping lags of de jure and de facto measures of restrictions on cross-border financial transactions. Our main findings continue to hold even after controlling for other factors and using instrumental variables.

The paper proceeds as follows. Section 2 develops our simple empirical model. Section 3 discusses the data. Section 4 presents the econometric estimates. Section 5 concludes. An appendix develops a theoretical model that formalizes the tight link between capital inflows and output.

II. Empirical model

To organize ideas and guide our empirics, this section develops a simple empirical model of the relation between output volatility and capital inflows volatility. Consider a small open economy with the following technology:

(1)

where Y is output, A is a positive technological parameter, and K is a (tradable) capital good. Let p be the international price of this capital good.⁵

The capital stock consists of this period’s addition to the existing capital stock: $\text{K}={\text{K}}_{-1}+\Delta K$

Assume that the purchase of this period’s capital good must be fully financed by capital inflows, either in the form of foreign direct investment (FDI) or portfolio and other investments (OTR)⁶. Formally, $\Delta K=\frac{\text{FDI}+\text{OTR}}{p}$

Solving for p and substituting in (1), we obtain

(2)

where

(3)

denotes total capital inflows and $\tilde{\text{A}}\equiv \text{A}/p$. Output is thus a linear function of total capital inflows.⁷

Let $\overline{\mathit{\text{Y}}}$ and $\overline{\text{TF}}$ denote the means of output and total capital flows, respectively. It then follows from (2) that

(4)

Output volatility is thus an increasing function of capital inflows volatility. To proceed further, we need to impose more structure. Specifically, let us assume that the stochastic processes for FDI and OTR take the following multiplicative form:

(5)

(6)

where $\overline{\text{FDI}}$ and $\overline{\text{OTR}}$ are the means of FDI and OTR , respectively, $${\text{ε}}_{\text{FDI}}~\text{N}\left(0,{\sigma }_{\text{FDI}}^2\right)$, ${\text{ε}}_{\text{OTR}}~\text{N}\left(0,{\sigma }_{\text{OTR}}^2\right)$ ,and ${\text{ε}}_{\text{FDI}}$ and ${\text{ε}}_{\text{OTR}}$ are jointly normally distributed. For further reference, let p denote the correlation between ${\text{ε}}_{\text{FDI}}$ and ${\text{ε}}_{\text{OTR}}$.⁸

Let $\mathit{\text{ϕ}}$ denote the share of FDI in total capital inflows; that is,

(7)

(8)

where $\overline{\mathit{\text{TF}}}$ is the mean of TF. Since TF is the sum of FDI and OTR , it will inherit the multiplicative stochastic structure of FDI and OTR. To see this, substitute (5) and (6) into (3), and use (7) and (8), to obtain $\text{TF}=\overline{\text{TF}}\left[\text{ϕ}\right(1+{\text{ε}}_{\text{FDI}})+(1-\varphi \left)\right(1+{\text{ε}}_{\text{OTR}}\left)\right]$

Hence,

(9)

From (2), ${\sigma }_{\text{Y}}^2={\tilde{\text{A}}}^2{\sigma }_{\text{TF}}^2$. Using (9), we can express output volatility as

(10)

This equation thus relates output volatility (${\sigma }_{\text{Y}}^2$) to the volatility of foreign direct investment (${\sigma }_{\text{FDI}}^2$), the volatility of portfolio and other investments (${\sigma }_{\text{OTR}}^2$), the correlation between FDI and OTR ($\rho$), and the share of FDI in total capital inflows ($\mathit{\text{ϕ}}$).

Even though the model is extremely simple, expression (10) already warns us that some commonly-held beliefs regarding the beneficial role of FDI in bringing about lower output volatility in emerging markets are in fact not obvious on closer examination. In particular, we can see that while equation (4) indicates that lower capital inflows volatility does imply lower output volatility, equation (10) tells us that whether lower FDI volatility will actually translate into lower output volatility depends on the correlation between FDI and OTR. As will become clear below, if $\rho <0$, lower FDI volatility could actually lead to higher output volatility! Also, the effect of $\mathit{\text{ϕ}}$ on output volatility is, in principle, ambiguous. In fact, it is possible that a larger share of FDI will lead to higher, rather than lower, output volatility. This suggests that we need to be careful in establishing the conditions under which these ideas may be true and then check in the data if these conditions hold.

To gain insights into expression (10), let us proceed by considering some special cases.

.Case 1: Variances of FDI and OTR are the same and the correlation is on (i.e.,${\sigma }_{\text{FDI}}^2={\sigma }_{\text{OTR}}^2$ and $\rho =1$). Expression (10) then reduces to

${\sigma }_{\text{Y}}^2={\left(\tilde{\text{A}}\overline{\text{TF}}\right)}^2{\sigma }_{\text{FDI}}^2$

Output volatility does not depend on $\mathit{\text{ϕ}}$. Since the variances of FDI and OTR are the same and the correlation is one, there is essentially no difference between FDI and OTR and hence the share of is irrelevant. In this case, higher FDI(or OTR) volatility translates into higher output volatility.

. Case 2: Variances of FDI and OTR are the same, $\mathit{\text{ϕ}}=0.5$, and there is a perfect negative correlation (i.e.,${\sigma }_{\text{FDI}}^2={\sigma }_{\text{OTR}}^2$ and $\rho =1$ ). In this case,${\sigma }_{\text{Y}}^2=0$ . This can be thought of as the “full insurance” case. Due to the perfectly negative correlation, equal variances, and equal share, total capital inflows are constant and hence output volatility is zero.

. Case 3. Variances of FDI and OTR are the same and the correlation is zero (i.e.,${\sigma }_{\text{FDI}}^2={\sigma }_{\text{OTR}}^2$ and $\rho =0$ ). In this case, expression (10) reduces to

(11)

This is the typical benchmark in basic portfolio theory. Think of an investor with two uncertain sources of income (FDI and OTR) that have the same variance but are uncorrelated. What is the share of that would minimize the volatility of the overall portfolio? Set ${\sigma }_{\text{FDI}}^2={\sigma }_{\text{OTR}}^2$ and in (10) and differentiate with respect to $\mathit{\text{ϕ}}$ to obtain,

(12)

This expression is zero for $\mathit{\text{ϕ}}=1/2$ and, as can easily be checked, the second derivative is positive indicating the existence of a minimum. In order words, with two uncorrelated sources of income that have the same variance, splitting the portfolio in half minimizes the overall volatility.

Deviating marginally from $\mathit{\text{ϕ}}=1/2$ has, of course, no first-order effect on output volatility. For values of $\mathit{\text{ϕ}}\ne 1/2$ , however, increasing $\mathit{\text{ϕ}}$ if $\mathit{\text{ϕ}}>1/2$ or reducing $\mathit{\text{ϕ}}$ if $\mathit{\text{ϕ}}<1/2$ will increase output volatility because the shares are getting farther away from the variance-minimizing mix. Formally:

(13)

(14)

Figure 1 illustrates this case by plotting ${\sigma }_{\text{Y}}^2$ as a function of $\mathit{\text{ϕ}}$ for $\rho =0$ and ${\sigma }_{\text{FDI}}={\sigma }_{\text{OTR}}=30$. We can see that, as equation (A.5) indicates, the variance-minimizing value of $\mathit{\text{ϕ}}$ is 0.5 (point A). Given the U-shape of the curve, moving away from point A in either direction increases ${\sigma }_{\text{Y}}^2$. Point B indicates the median value of $\mathit{\text{ϕ}}$ in our sample; 0.31. For any $\mathit{\text{ϕ}}$ between this value and 0.5 , increasing $\mathit{\text{ϕ}}$ will reduce output variability.

Figure 1

Output volatility and share of FDI in total gross capital inflows. (FDI)= (OTR)=30, (FDI,OTR)=0

What happens if we deviate from this benchmark portfolio case in terms of p being different from zero or variances not being the same? Cases 4 and 5 study these deviations from Case 3.

. Case 4. Variances are the same but p is different from zero (i.e., ${\sigma }_{\text{FDI}}^2={\sigma }_{\text{OTR}}^2$ and $\rho \ne 0$)

If the correlation is not zero, then it will still be the case that the value of $\mathit{\text{ϕ}}$ that minimizes output volatility is one-half. Indeed, set ${\sigma }_{\text{FDI}}^2={\sigma }_{\text{OTR}}^2$ in equation (10) and differentiate with respect to $\mathit{\text{ϕ}}$ to obtain

which is zero for $\mathit{\text{ϕ}}=1/2$. Intuitively – and as (10) makes clear – a positive correlation increases overall volatility relative to the p=0 case but does not change the fact that, since FDI and OTR are not perfectly correlated, the variance-minimizing $\mathit{\text{ϕ}}$ is still one-half.

. Case 5. Correlation is zero but variances are different (i.e., p=0 and ${\sigma }_{\text{FDI}}^2\ne {\sigma }_{\text{OTR}}^2$). In this case, the variance-minimizing $\mathit{\text{ϕ}}$ will change. To see this, set p=0 in (10) and differentiate with respect to $\mathit{\text{ϕ}}$ to obtain

Setting this expression to zero, we obtain the variance-minimizing value of $\mathit{\text{ϕ}}$:

(15)

If ${\sigma }_{\text{FDI}}^2<{\sigma }_{\text{OTR}}^2$ , then ${\mathit{\text{ϕ}}}^{min}>1/2$. Intuitively, if FDI is less volatile than OTR, then it would be optimal to hold more than one-half of the TF as FDI. Even though the variance-minimizing $\mathit{\text{ϕ}}$ is larger than one-half, the same intuition developed in Case 3 above holds: deviating from this variance-minimizing value of $\mathit{\text{ϕ}}$ will increase overall volatility.

Needless to say, in practice countries cannot choose the variance-minimizing value of $\mathit{\text{ϕ}}$ ⁹ . But all the intuition developed so far will still help us in thinking about the data. As an illustration, Figure 2 plots equation (10) for the case of Turkey in which ${\sigma }_{\text{FDI}}=58.8$, ${\sigma }_{\text{OTR}}=168.5$ , and $\rho =-0.23$.¹⁰ In this case, the variance-minimizing value of $\mathit{\text{ϕ}}$ is 0.85, given by point A. Given the U-shape of the curve, moving away from point A in either direction increases ${\sigma }_{\text{Y}}^2$. Point B is the actual value of $\mathit{\text{ϕ}}$ for Turkey, $\text{ϕ}=0.14$. Since this value of $\mathit{\text{ϕ}}$ is less than the variance-minimizing $\mathit{\text{ϕ}}$, increasing $\mathit{\text{ϕ}}$ will reduce output volatility. This will be one of the main empirical predictions of our model.

Figure 2:

Output volatility and share of FDI in total gross capital inflows. $\sigma$(FDI)=58.8, $\sigma$(OTR)=168.5, $\rho$ (FDI,OTR)=-0.23

Returning now to the general case captured in equation (10), let us examine how changes in $\rho$ , ${\sigma }_{\text{FDI}}^2$ , and ${\sigma }_{\text{OTR}}^2$ affect output volatility. Taking the corresponding partial derivatives, we obtain

(16)

(17)

(18)

As equation (16) makes clear, a higher $\rho$ always increases output volatility. On the other hand, expressions (17) and (18) indicate that the effects of ${\sigma }_{\text{FDI}}^2$ and ${\sigma }_{\text{OTR}}^2$ are ambiguous. To understand this ambiguity, think of the case in which $\rho =-1$, $\text{ϕ}=0.5$ , and ${\sigma }_{\text{FDI}}<{\sigma }_{\text{OTR}}$. Equation (17) then reduces to $\frac{d{\sigma }_{\text{Y}}^2}{d{\sigma }_{\text{FDI}}}=2\stackrel{˷}{\text{A}}{\overline{\text{TF}}}^2{\text{ϕ}}^2({\sigma }_{\text{FDI}}-{\sigma }_{\text{OTR}})<0$

Here a reduction in FDI volatility would increase output volatility. Intuitively, with perfect negative correlation between foreign direct investment and portfolio and other investments and ${\sigma }_{\text{FDI}}<{\sigma }_{\text{OTR}}$, FDI volatility is actually a good thing because then FDI can offer more insurance against OTR. In other words, if FDI exhibits very low volatility, then it cannot offset the much higher volatility of OTR.

In the data, however, $\rho$ is on average close to zero (sample median is 0.05), in which case an increase in the volatility of either FDI or OTR will increase output volatility. Intuitively, with zero correlation, higher volatility is unambiguously bad because it contributes to output volatility directly without offering any insurance.

To summarize, the main predictions of our empirical model are as follows:

• Output volatility should be an increasing function of the correlation between FDI and OTR.

• Output volatility should be an increasing function of FDI volatility and OTR volatility (under the assumption that p>0).

• Output volatility should be a decreasing function of the share of FDI in total capital inflows (under the assumption that the actual value of is below the variance-minimizing value of ).

III. Data

This study uses a sample of 59 countries: 20 industrial and 39 developing countries for the period 1970-2009.¹¹ Data frequency is annual. Data for real GDP, gross capital inflows, government spending, inflation, and terms of trade data comes from International Financial Statistics (IFS) and World Economic Outlook (WEO), both from the IMF. For capital flows, we use foreign direct investment, portfolio investment, and other investment gross inflows data. As is common practice (see, for instance, BIS (2009)) we group together portfolio and other investments as being more short-term in nature than FDI and denote this aggregate by OTR¹².

The standard deviations and correlations of all variables are computed based on their cyclical components. For this purpose, we use the Hodrick-Prescott filter with a smoothing parameter of 6.5 (Ravn and Uhlig, 2002). Since the cyclical component is expressed in terms of percentage deviations of the actual value from the trend, the corresponding standard deviation is also expressed in those terms. For example, the volatility of FDI mentioned above for Turkey (${\sigma }_{\text{FDI}}=58.8$) means that, on average, the level of FDI is 58.8 percentage points away from its trend. Given that ${\sigma }_{\text{OTR}}=168.5$ for Turkey, this implies that portfolio and other investments are almost three times as volatile as FDI.

Another common practice in the literature (see, for instance, Albuquerque, Loayza, and Serven, 2005) is to normalize capital flows such as FDI by dividing them by GDP. The rationale behind this methodology is to control for country size and avoid nonstationarity problems. While helpful in a different context, we feel that this normalization would not be appropriate in our case because the volatility of such a ratio would capture the volatility of both FDI and output. Since the latter will be our dependent variable, our empirical analysis would suffer from endogeneity problems by construction. Moreover, our focus on the cyclical component of capital inflows avoids nonstationarity problems altogether. Notice also that because we measure the cyclical component in terms of percentage deviations of the actual value from the trend, our volatility measures are independent of the size of the economy or capital inflows. Indeed, using cross-country data, we cannot reject the null hypothesis that the correlation between ${\sigma }_{\text{FDI}}$ and average FDI, as well as the correlation between ${\sigma }_{\text{OTR}}$ and average OTR, are equal to zero at a 5 percent significance level.

We now turn to a broad look at the data. In particular, we focus on volatility and basic statistics discussed in the previous section. Figure 3 shows output volatility.¹³ While output volatility varies substantially across countries, the median is almost twice as large in developing countries as in industrial countries. Figure 4 shows total gross inflows volatility. Not surprisingly, the median of total gross inflows into developing countries is more than one and a half times that of industrial countries.¹⁴

Figure 3

Output volatility.

We now turn to the volatilities of FDI and OTR. Figure 5 shows the ratio of OTR volatility to FDI volatility. The figure is consistent with the idea in the literature that OTR inflows are more volatile than FDI inflows. Indeed, the ratio is larger than one for more than 85 percent of the countries in our sample. The median volatility of OTR is close to 120, compared to less than half (about 44) for FDI. Moreover, the median in developing countries ratios is 76 percent higher than that in industrial economies, reflecting in particular the higher volatility of OTR. In fact, the median FDI volatility is 48 for industrial counties and 41 for developing countries. In sharp contrast, the median OTR volatility is about 30 percent higher in developing countries than in advanced economies (120 for developing countries and 85 for industrial countries).

Figure 4

Total gross inflows volatility

Figure 6 shows that the share of in total gross capital inflows is typically quite low, with the sample median being 0.32. Indeed, for more than 60 percent of the countries, the share is less than 0.5. Furthermore, the median share is three times as high in developing countries as in industrial countries. Finally, Figure 7 depicts the correlation between and . We can see wide variation in this figure across countries, with the sample median being 0.05 and the median for developing countries -0.02.

Figure 5

Ratio of OTR over FDI volatilities.

Figure 6

Median share of gross FDI inflows in total gross capital inflows.

Taking into account the median values of the ratio of OTR volatility to FDI volatility and the correlation between OTR and FDI for industrial and developing countries, we find that the $\mathit{\text{ϕ}}$ (i.e., the share of FDI in total capital inflows) that minimizes output volatility (i.e., expression 10) is 0.7 and 0.8 for industrial and developing countries, respectively. These values are much higher than the actual ones: 0.15 for industrial economies and 0.45 for developing countries. The difference in the optimal shares of FDI reflects the fact that (i) the relative ratio of OTR volatility to FDI volatility is higher in developing countries than in industrial ones (3.7 versus 2.1) and (ii) the correlation between OTR and FDI is positive (0.14) for industrial countries but slightly negative (-0.02) for developing countries. In other words, a higher share of FDI in total capital inflows is more beneficial for developing countries than for industrial economies because (i) it reduces total capital flows volatility directly by substituting a more volatile source of capital (OTR) for one that is less volatile (FDI) and (ii) it provides some insurance given the negative (though rather small) correlation.

Figure 7

Correlation between OTR and FDI gross inflows.

IV. Empirical evidence

In this section we test the main empirical implications derived in Section 2. First, output volatility should depend positively on FDI and OTR volatility. Second, output volatility should be an increasing function of the correlation between FDI and OTR. Third, for low values of the FDI share, output volatility should be a decreasing function of the share of FDI in total capital inflows.¹⁵

We first show our benchmark regressions that link output volatility to the variables highlighted in the empirical model of Section 2. We then control for other variables that, in practice, could affect output variability. We then address endogeneity problems.

IV.1 Basic regressions

Following the empirical growth literature, we use non-overlapping five-year averages. Table 1 reports the basic results using country and five-year fixed effects. Standard errors are robust and we also allow for within-country correlation (i.e., clustered by country). We normalize ${\sigma }_{\mathit{\text{FDI}}}$ and ${\sigma }_{\mathit{\text{OTR}}}$ to be between 0 and 100 to make regression coefficients easier to read.¹⁶ Columns 1-5 test the key implications of our model one variable at a time and column 6 tests them all together.

Results are as predicted by our model. Higher FDI and OTR volatility increase output volatility (columns 1 and 2). A higher correlation between FDI and OTR increases output volatility (column 3). When we include the share of FDI in total capital inflows (column 4), it appears not to matter, contrary to our model’s prediction. However, as captured by (A.5)-(14), the expected relationship between a higher share of FDI in total inflows and lower output volatility tends to occur when the share is small or, to be precise, smaller than optimal. In the particular case of equal variances and zero correlation, an increase in the share of FDI will reduce output volatility when the initial share is smaller than 0.5 (see equation (14). To capture this effect, we interact this term with a dummy variable that equals one when the share is smaller than the sample median share (0.32). Column 5 shows that, indeed, after introducing this distinction, output volatility is a decreasing function of the share of FDI in total capital inflows only when its initial value is low. Finally, when all explanatory variables are included (column 6), the size of the coefficients and significance levels remain essentially unchanged.

Table 1

Basic regression results. Dependent variable is output volatility

IV.2 Controlling for other determinants of output volatility

Having established that output volatility depends on the factors predicted by the portfolio model developed in Section 2, we now proceed to control for other factors that could also affect output volatility. While the basic regressions of Subsection 4.1 control for country and five-year fixed effects, other factors such as idiosyncratic external shocks, fiscal policy volatility, and country instability could also affect output volatility.

Fiscal policy volatility is measured using the standard deviation of the cyclical component of government spending. We proxy external shocks volatility using the standard deviation of the cyclical component of terms of trade. Country instability is measured using the average of internal and external conflicts from the International Country Risk Guide (ICRG). Internal conflict refers to political violence within the country and its actual or potential impact on governance. The risk rating assigned is composed of three subcomponents: civil war/coup threat, terrorism/political violence, and civil disorder. External conflict refers to the risk to the incumbent government from foreign action, ranging from non-violent external pressure (diplomatic pressures, withholding of aid, trade restrictions, territorial disputes, sanctions, and so forth) to violent external pressure (ranging from cross-border conflicts to all-out war). The risk rating assigned is composed of three subcomponents: war, cross-border conflict, and foreign pressures. We normalized this variable so that it varies between 0 and 100, with a low value indicating low risk.

Results are reported in Table 2. Columns 1 to 3 show the effects of the control variables one at a time. The three variables have the expected signs: higher fiscal and terms of trade volatility and more country instability increase output volatility. Surprisingly enough, however, terms of trade volatility is not statistically significant. The reason is that we are also including five-year fixed effects. If such fixed effects are not included, then the coefficient of the terms of trade volatility is positive and significant at the 5 percent level. We thus conclude that, while there is some country idiosyncratic variation over time, an important fraction of terms of trade volatility is common to most countries. This is reflected, for instance in the large terms of trade volatility present in the 1970s and early 1980s (as a result of the oil shocks) and in 2005-2009 (generalized rise in commodity prices) compared to the 1990-2004 period.

When including all controls (column 4), fiscal policy volatility becomes insignificant due to its high correlation with country instability.¹⁷ More importantly for our purposes, Column 6 indicates that the size and significance of our four explanatory variables (FDI volatility, OTR volatility, correlation between FDI and OTR, and interacted share of FDI) remain essentially unchanged relative to column 6 in Table 1.

IV.3 Adressing endogeneity

This section addresses potential endogeneity problems. One could reasonably argue that the positive relationship between output volatility and FDI and OTR volatility may reflect the fact that higher GDP volatility increases the volatility of capital inflows. In other words, the causality may run from output volatility to inflows volatility rather than the other way around. In the same vein, reductions in the share of FDI could reflect the reluctance of foreign firms to invest for the long-term in highly volatile economies.

As is the case in the empirical macro literature that has assessed the influence of FDI on economic growth (see, for instance, Lensink and Morrisey, 2001 and Alfaro, 2003), we lack obvious instruments for ${\sigma }_{\mathit{\text{FDI}}}$, ${\sigma }_{\mathit{\text{OTR}}}$ , $\rho$(FDI, OTR), and FDI share . We then use three sets of instruments. First, we follow the above-mentioned macro literature in using lagged FDI as an instrument for current FDI. In our case, this amounts to using the lagged five-year average of each portfolio variable as an instrument. For example, we use the ${\sigma }_{\mathit{\text{OTR}}}$ for the period 1970-1974 to instrument for the period 1975-1979. The Spearman’s rank correlation between ${\sigma }_{\mathit{\text{FDI}}}$ and its lagged five-year value is 0.38. The corresponding correlation is 0.26 for ${\sigma }_{\mathit{\text{OTR}}}$ and 0.30 for FDI share. In all cases, the correlation is statistically significant at the 5 percent level. In other words, there seems to be a positive association between the volatilities of capital inflows over time, even at the five-year frequency. Unfortunately, the correlation between $\rho$(FDI, OTR) and its lagged five-year value is statistically insignificant.

Table 2

Regressions results with control variables. Dependent variable is output volatility.

Our second set of instruments uses a geographical/gravity approach aimed at capturing the influence of regional effects. Capital inflows respond to economic and political fundamentals which are often shared by different countries within a region (Calvo and Reinhart, 1996; Fernandez-Arias and Montiel, 1995; Alba, Bhattacharya, Claessens, Gosh, and Hernández, 2000; Corbo and Hernández, 2001). During the 1980s, for example, Latin America experienced such political and economic instability that international investors became reluctant to invest for the long-term. Indeed, FDI share was just 0.24 for Latin America during the 1980s, compared to 0.51 during the 1990s, and almost 0.75 during the 2000s. To exploit this geographical dimension, we instrument each portfolio variable using the following expression:${\text{I}}_{it}=\sum _j\frac{1}{{dist}_{ij}}{\text{I}}_{jt},i\ne j,$

where ${\text{I}}_{ij}$ represents the portfolio variable and ${dist}_{ij}$ measures the distance between the capital cities of countries i and j. In other words, we instrument a country’s five-year observation of each portfolio variable with the weighted sum of such variable for other countries. The weight for each other country decreases with its distance. This gravity approach is thus a more generalized version of the idea behind regional effects. The Spearman’s rank correlation between $\rho$(FDI, OTR) and the suggested geographical instrument is 0.35 and statistically significant at the 5 percent level. The gravity approach proves to be a good strategy to predict the patterns of correlation between FDI and OTR. Unfortunately, the correlations for the other portfolio variables are statistically insignificant.

To complement the two previous sets of instruments, we also rely on the literature regarding the determinants of capital flows. In particular we focus on determinants that may help determine portfolio variables, but not have a direct effect on output. Montiel and Reinhart (1999) find that, by imposing capital controls, countries are able to increase the share of FDI. Generally speaking, policies that punish short-term flows should, in principle, induce foreign investors to increase long-term flows. We use three variables to account for this effect. First, we use the Chinn-Ito index (Chinn and Ito, 2006) to measure de jure financial openness. This index, which measures a country’s capital account openness, is based on a binary dummy variable that codifies the tabulation of restrictions on cross-border financial transactions reported in the IMF’s Annual Report on Exchange Arrangements and Exchange Restrictions. A high value of this index is an indication of low de jure financial integration. Second, we use the ratio of total foreign assets and liabilities to GDP from Lane and Milesi-Ferretti (2007) to measure de facto financial integration. A high value of this index indicates a high degree of de facto financial integration. Lastly, we use the investment profile index from the International Country Risk Guide (ICRG). This investment profile assesses factors affecting the risk to investment that are not covered by other political, economic, and financial risk components. The risk rating assigned is composed of three subcomponents: contract viability/expropriation, profits repatriation, and payment delays. We normalized this variable so that it ranges between 0 and 100, with a low (high) value indicating low (high) risk. We use the five-year lag of these three variables to take care of reverse causality concerns (i.e. ,the possibility that output volatility leads to investment risk). The Spearman’s rank correlation between FDI share and the five-year lag of the Chinn-Ito index is -0.14, while the correlation with the five-year lag of investment profile is 0.18. In both cases, the correlation is statistically significant at the 5 percent level. These findings support Montiel and Reinhart’s (1999) arguments. Moreover, in line with the rationale behind some recent policy measures in countries such as Brazil, more capital controls (i.e., lower de jure financial openness) reduce ${\sigma }_{\mathit{\text{OTR}}}$. The Spearman’s rank correlation between ${\sigma }_{\mathit{\text{OTR}}}$ and the five-year lag of the Chinn-Ito index is statistically significant and equal to -0.19.

Having checked that the proposed sets of instruments seem to be good predictors for the variables they are instrumenting for, we proceed to estimate instrumental variables regressions. Table 3 shows the instrumental variable regressions. In all cases we cannot reject the overidentification tests at a 5 percent confidence level. The instruments are valid instruments (i.e., uncorrelated with the error term) and the excluded instruments are correctly excluded from the estimated equation. Moreover, as suggested by the discussion above, instrumental variable regressions confirm that in almost all cases the excluded instruments are not weak instruments (i.e., they are strongly correlated with the endogenous regressors). Column 1 shows that our previous empirical findings hold. The only exception is ${\sigma }_{\mathit{\text{OTR}}}$: while the sign of the coefficient is positive, it is not statistically significant.¹⁸

Table 3

Instrumental variable regression results with control variables. Dependent variable is output volatility

We now add control variables. While terms of trade volatility is typically treated as exogenous, this is certainly not the case of government spending volatility and country instability. Indeed, it seems reasonable to argue that higher output volatility might increase government spending volatility and lead to more instability. To account for this potential reverse causality, we use the five-year lag of government spending volatility and country instability. The Spearman’s rank correlation between ${\sigma }_{gov.spending}$ and its five-year lag is 0.56 and the corresponding correlation for country instability is 0.84. In both cases, the correlation is statistically significant at the 5 percent level. Columns 2 to 4 show the results of including each determinant one at a time. Column 5 includes all portfolio and control variables. The inclusion of these determinants does not change the main results reported in column 1.

V. Conclusions

A commonly-held belief is that a larger share of FDI in total capital inflows will reduce output volatility. There is, however, little, if any, formal evidence on this channel. Based on standard portfolio theory, we first develop a simple econometric model that calls attention to some important caveats. In particular, lower FDI volatility will reduce output volatility only if the correlation between FDI and other flows is positive (which is not always the case in the data). Also, a larger share of FDI will reduce output volatility only if the actual share of FDI is below the variance-minimizing share. Our model thus yields three testable implications: (i) output volatility should depend positively on FDI and OTR volatility; (ii) output volatility should be an increasing function of the correlation between FDI and OTR; and (iii) output volatility should be a decreasing function of the share of FDI in total capital inflows (when the initial share is low). We find strong support in the data for all three implications, even after controlling for other factors that influence output volatility and for possible endogeneity problems.

A. Appendix

This appendix develops a simple theoretical model that provides a theoretical illustration of the key assumption in the empirical model – as captured in equation ([Y2]) – that there exists a tight link between output and capital inflows. In the theoretical model, such a link will arise endogenously as firms choose whether to finance investment with either short-term or long-term external funding¹⁹. In the aggregate, the economy uses both sources of finance and changes in, say, the cost of external funding will lead to changes in output, external finance, and its composition.

Consider a small open economy with a continuum of risk-neutral firms that produce the same final (tradable) good, denoted by q, using the same (tradable) capital, denoted by k. Firms are indexed by their productivity parameter, $\gamma$ ($0<\gamma <1$), which is the only source of heterogeneity. Firms “live” for two periods. Firms buy capital before production and hold it for the entire two periods, after which it depreciates completely.

The production function of a $\gamma$ firm is given by $q_t={ak}^{\gamma },t=1,2,$

where a>0 is a productivity parameter. By construction, output is constant across periods. Firms need to borrow from abroad to finance the purchase of capital. Borrowing can be either short-term or long-term but not a combination of both. Short-term funding (i.e., portfolio investment) requires repayment of principal and interest at the end of the first period. Long-term funding (i.e., foreign direct investment) requires repayment of principal plus interest only after two periods. The one-period short-term and long-term interest rates are, respectively, $r^s$ and $r^l$. We assume that $r^s<r^l$ , reflecting the idea that international lenders may have a preference for a more “liquid” asset.

As an important benchmark, we first solve the firm’s problem under short-term financing and no repayment constraint. We then impose the repayment constraint for short-term financing. We then solve for the case of long-term financing. We then compare profits in the two cases (short-term financing and repayment constraint versus long-term financing) to find out when a firm will chose one or the other. We finally aggregate over all firms to obtain the economy’s aggregate capital stock and output and analyze how the equilibrium changes if the cost of long-term financing changes.

A.1 Short-term financing and no repayment constraint

Denote by the world relative price of q in terms of k and by $b_t$, t=0,1 net foreign assets. Think of period 0 as the period in which the capital stock is purchased. Periods 1 and 2 are the periods in which the firm operates (i.e., produces and sells). The flow budget constraints are thus given by

[A.1]

[A.2]

[A.3]

where ${\pi }_t$ , t=1,2, denotes dividends paid by the firm. Combining these flow constraints, we obtain an intertemporal constraint:

[A.4]

where $\prod (\equiv {\pi }_1/(1+r^s)+{\pi }_2/{(1+r^{\text{S}})}^2)$ is the present discounted value of profits as of time 0.

Firms choose k to maximize (A.4). The first-order condition for capital takes the form:

[A.5]

At an optimum, the firm equates the present discounted value of the value of the marginal productivity to the cost of capital. Solving for the capital stock, we obtain

[A.6]

Substituting this expression into (A.4), we can write profits as:

[A.7]

As expected, profits are positive since, by assumption, $\gamma \in (0,1)$.

In the absence of any additional constraint, all firms would choose short-term financing because, by assumption, it is cheaper than long-term financing. To have a meaningful choice between short-term and long-term financing, we will now introduce a repayment constraint.

A.2 Short-term financing and repayment constraint

Suppose now that a firm can access short-term credit only if it can pay back the loan at the end of the first period. Formally,

[A.8]

Let us check if this repayment constraint binds for the unconstrained problem that we just solved. To this effect, substitute (A.6) into the last expression to obtain $\frac{1+r^s}{2+r^s}\ge \gamma$

Firms whose $\textcolor{red}{}\gamma$ satisfies this condition will thus still be able to choose short-term financing and remain at the first best because the repayment constraint does not bind. Intuitively, low $\textcolor{red}{}\gamma$ firms optimally choose a low level of capital (i.e., units of capital with high marginal productivity) and are thus more likely to satisfy constraint (A.8) given that the repayment cost per unit of capital ($\textcolor{red}{}1+r^s$) does not depend the level of capital.

On the other hand, the unconstrained solution for firms with $\gamma >(1+r^s)/(2+r^s)$ violates condition (A.8). These firms will thus need to choose between “constrained short-term financing” (i.e., choose the optimal level of capital subject to condition (A.8)) or long-term financing. The trade-off is thus between remaining in a first-best equilibrium but facing a higher cost of capital (long-term financing) or choosing a constrained level of capital but at a lower cost (constrained short-term financing).

If constraint (A.8) binds, then the capital stock is given by

[A.9]

If we compare this level of capital with the unconstrained level of capital, given by expression (A.6), for a firm with $\gamma >(1+r^s)/(2+r^s)$, we can see that the stock of capital in the constrained case is lower. In other words, to access short-term financing, the firm needs to have a suboptimally low level of capital to generate enough profits in the first period to repay the loan.

A.3 Maximization under long-term financing

Let us now compute profits under long-term financing. The budget constraints remain the same as in (A.1)-(A.3) with $r^l$ in lieu of $r^s$. Further, since a firm that chooses long-term financing is still operating in a first-best world, the choice of capital will be given by condition (A.6) with $r^l$ in lieu of $r^s$. Profits will thus be given by (A.7) with the corresponding choice of capital.

A.4 Comparison

Firms with $\gamma >(1+r^s)/(2+r^s)$ will choose long-term financing over short-term financing as long as profits are larger: ${\prod |}_{\text{long-term}}\ge {\prod |}_{\text{short-termconstrained}}$

Using equations (A.6) and (A.7), this condition reduces to

Suppose to fix ideas that $r^l=r^s$. In this case, this last expression reduces to $1\ge {\left[\frac{(1+r^l)}{(2+r^s)\gamma }\right]}^{\frac{1}{1-\gamma }}$

Since the choice is only relevant for firms with $\gamma >(1+r^s)/(2+r^s)$, the condition will always hold. In other words, if $r^l=r^s$, then all these firms would choose long-term financing because the cost is the same as short-term financing but they are not subject to the repayment constraint (which, by construction, is binding).

But our maintained assumption is, of course, that $r^l>r^s$. In that case, condition (A.10), holding with equality, defines a threshold value of $\textcolor{red}{}\gamma$, denoted by $\textcolor{red}{}{\gamma }^{*}$, which is given by

[A.11]

We now establish the following result:

Claim 1. Firms with $\gamma \ge {\gamma }^{*}$ ( $\gamma <{\gamma }^{*}$ ) will choose long-term (short-term) financing.

Proof.Consider condition (A.10). Differentiating the right-hand side and evaluating the corresponding expression at $\gamma ={\gamma }^{*}$ , we obtain ${\frac{{d\left[\frac{{(1+r^l)}^2}{(2+r^l)(1+r^s)\gamma }\right]}^{\frac{1}{1-\gamma }}}{d\gamma }|}_{\gamma ={\gamma }^{*}}=\frac{-1}{(1-\gamma )\gamma }<0$

Set $\gamma ={\gamma }^{*}$ in condition (A.10). By construction, it will hold as an equality. An increase in will then reduce the RHS, which means that long-term profits will be higher than constrained short-term profits. The reverse is true for a fall in $\gamma$ .

Intuitively, firms with a large $\gamma$ (i.e., $\gamma >{\gamma }^{*}$ ) are firms that find it more efficient to operate on a larger scale (and thus would be hurt more by the repayment constraint) and hence would be willing to pay the higher cost of long-term financing in order to not be subject to the repayment constraint. In contrast, smaller firms (i.e., firms with $\gamma <{\gamma }^{*}$) would rather not pay the additional cost of financing and choose a second-best level of capital.

From (A.11), we can see that ${\gamma }^{*}$ increases with $r^l$ and decreases with $r^s$. Intuitively, an increase in $r^l$ makes long-term financing more expensive. As a result, marginal firms will choose to switch to short-term financing (i.e., ${\gamma }^{*}$ increases). Conversely, an increase in $r^s$ makes short-term financing more expensive and hence marginal firms will choose to switch to long-term financing (i.e., ${\gamma }^{*}$ decreases).

A.5 Aggregation

As has been established above, there are three types of firms in this economy depending on the value of $\gamma$:

• The range $0<\gamma \le \frac{1+r^s}{2+r^s}$ consists of firms that are operating in a first-best world with short-term financing.

• The range $\frac{1+r^s}{2+r^s}<\gamma \le {\gamma }^{*}$ consists of firms that are operating under constrained short-term financing (i.e., these are firms that would violate the repayment constraint if they chose the first-best level of capital).

• The range ${\gamma }^{*}<\gamma <1$ consists of firms that are operating with long-term financing.

Aggregate capita and output are thus given by, respectively,

[A.12]

[A.13]

where $\tilde{\gamma }\equiv (1+r^s)/(2+r^s)$ .²⁰

Since the first two types of firms buy capital with short-term borrowing (denote it by POR ), while the last type of firm buys it with long term borrowing (denote it by FDI), we can write

[A.14]

[A.15]

We thus have an economy with heterogeneous firms in which the composition of external financing is endogenously determined based on each firm’s productivity and the cost of short-term and long-term financing. This gives us a simple framework to ask how a change in the cost of long-term financing changes the equilibrium.

A.6 Changes in the cost of long-term financing

What are the effects of a change in $r^l$? Specifically, suppose that $r^l$ is lower; how does the equilibrium described above change?²¹

Using Leibniz rule, we can compute the changes in capital and output from equations (A.12) and (A.13), respectively:

Capital and output thus increase. Intuitively, a fall in $r^l$ affects capital and output through two channels:

• From (A.11), we can see that a higher $r^l$ reduces ${\gamma }^{*}$. This means that some marginal firms that were relying on short-term financing will switch to long-term financing. At the margin, however, the capital stock of these firms does not change and thus output is not affected.²²

• The capital stock (and thus output) of firms that rely on long-term financing increases.

What happens to FDI and POR?

In absolute terms, increases FDI and POR falls. The share of FDI also increases because FDI increases by more than total capital inflows (given that POR falls). Intuitively, FDI increases for two reasons. First, firms that relied on long-term financing are now borrowing more. Second, some marginal firms that were relying on POR have now switched to FDI. The change in FDI is thus larger than the change in the capital stock.

It would be easy to accommodate random changes in $r^l$ in our model as long as firms continue to be risk-neutral. In that case, uncertainty regarding changes in $r^l$ (or $r^s$ for that matter) would not change the firms’ behavior derived above (with the expected value of $r^l$ and $r^s$ replacing the actual values). We could imagine that every third period $r^l$ is drawn from some distribution and the above equilibrium materializes. In such a scenario, an increase in the volatility of $r^l$ would lead to higher volatility in output, investment, FDI, POR, and the respective shares. Clearly, being endogenous, the higher volatility of capital inflows or FDI is not “causing” higher output volatility. Rather they are both endogenous responses to the higher volatility in the cost of long-term financing.

References

Alba, Pedro, Amar Bhattacharya, Stijn Claessens, Swati Gosh, and Leonardo Hernández (2000), “Volatility and contagion in a financially integrated world: Lessons from East Asia’s recent experience,” in Asia Pacific Financial Deregulation (Eds) G. de Brouwer and W. Pupphavesa, Routledge, London.

Albuquerque, Rui (2003), “The composition of international capital flows: Risk sharing through foreign direct investment,” Journal of International Economics, Vol. 61, pp. 353-383.

Albuquerque, Rui, Norman Loayza, and Luis Serven (2005), “World market integration through the lens of foreign direct investors,” Journal of International Economics, Vol. 66, pp. 267-295.

Alfaro, Laura (2003), “Foreign direct investment and growth: Does the sector matter?” mimeo, Harvard Business School.

Alfaro, Laura, Chanda Areendam, Sebnem Kalemli-Ozcan, and Selin Sayek (2004), “FDI and economic growth: the role of local financial markets,” Journal of International Economics, Vol. 64, pp. 89-112.

Alfaro, Laura, Sebnem Kalemli-Ozcan, and Vadym Volosovych (2007), “Capital flows in a globalized world: The role of policies and institutions” in Capital Controls and Capital Flows in Emerging Economies: Policies, Practices, and Consequences. (Eds) S. Edwards. National Bureau of Economic Research. University of Chicago Press.

Alfaro, Laura, Areendam Chanda, Sebnem Kalemli-Ozcan, and Selin Sayek (2010), “Does foreign direct investment promote growth? Exploring the role of financial markets on linkages,” Journal of Development Economics, Vol. 91, pp. 242-256.

Alfaro, Laura and Maggie Chen (2012), “Surviving the global financial crisis: Foreign ownership and establishment performance,” American Economic Journal: Economic Policy, Vol 4, pp. 30-55.

Bank for International Settlements (2009), “Capital flows and emerging market economies” CGFS Papers No. 33.

Borensztein, Eduardo, Jose De Gregorio, and Jong-Wha Lee (1998), “How does foreign direct investment affect economic growth?” Journal of International Economics, Vol. 45, pp. 115-135.

Calvo, Guillermo and Carmen Reinhart (1996), “Capital flows to Latin America: Is there evidence of contagion?” in Private Capital Flows to Emerging Markets after the Mexican Crisis (Eds) G. Calvo, M. Goldstein, and E. Hochreiter, Institute for International Economics and Austrian National Bank, Washington DC and Vienna.

Carkovic, Maria, and Ross Levine. (2005), “Does foreign direct investment accelerate economic growth?” in Does Foreign Direct Investment Promote Development? (Eds) T. Moran, E. Graham and M. Blomstrom, Washington, DC, pp. 195-220.

Chinn, Menzie and Hiro Ito (2006), “What matters for financial development? Capital controls, institutions, and interactions,” Journal of Development Economics, Vol. 81, pp. 163-192.

Chuhan, Punam, Stijn Claessens, and Nlandu Mamingi (1998), “Equity and bond flows to Latin America and Asia: The role of global and country factors,” Journal of Development Economics, Vol. 9, pp. 439-463.

Claessens, Stijn, Michael P. Dooley, and Andrew Warner (1995), “Portfolio capital flows: Hot or cold?” World Bank Economic Review, Vol. 9, pp. 153-74.

Fernandez-Arias, Eduardo, and Ricardo Hausmann (2001), “Is foreign direct investment a safer form of financing?” Emerging Markets Review, Vol. 2, pp. 34-49.

Fernandez-Arias, Eduardo, and Peter Montiel (1995), “The surge in capital inflows to developing countries: Prospects and policy response,” World Bank working paper No. 1473.

Hausmann, Ricardo and Eduardo Fernandez-Arias (2000), “Foreign direct investment: Good Cholesterol,” Inter-American Development Bank working paper No. 417.

Kalemli-Ozcan, Sebnem, Bent Sørensen, and Vadym Volosovych (2010), “Deep financial integration and volatility,” NBER working paper No. 15900.

Lane, Philip and Gian Maria Milesi-Ferretti (2007), “The external wealth of nations mark II: Revised and extended estimates of foreign assets and liabilities, 1970–2004,” Journal of International Economics, Vol. 73, pp. 223-250.

Lensink, Robert, and Oliver Morrisey (2001), “Foreign direct investment: Flows, volatility and growth in developing countries,” University of Nottingham SOM Research Report No. E16.

Levchenko, Andrei, and Paolo Mauro (2007), “Do some forms of financial flows protect from sudden stops?” World Bank Economic Review, Vol 21, pp. 389-411.

Montiel, Peter and Carmen Reinhart (1999), “Do capital controls and macroeconomic policies influence the volume and composition of capital flows? Evidence from the 1990’s,” Journal of International Money and Finance, Vol. 18, pp. 619-635.

Ravn, Morten, and Harald Uhlig (2002), “On adjusting the Hodrick-Prescott filter for the frequency of observations,” Review of Economics and Statistics, Vol. 84, pp. 371-376.

Notes

Additional information

Clasificación JEL: F23; F32; F36; F44