ï»¿ WPS6386
Policy Research Working Paper 6386
The Effect of Capital Flows Composition
on Output Volatility
Pablo Federico
Carlos A. Vegh
Guillermo Vuletin
The World Bank
Development Research Group
Macroeconomics and Growth Team
March 2013
Policy Research Working Paper 6386
Abstract
A large literature has argued that different types of capital both kinds of inflows. Third, output volatility should
flows have different consequences for macroeconomic be a decreasing function of the share of foreign direct
stability. By distinguishing between foreign direct investment in total capital inflows, for low values of
investment and portfolio and other investments, this that share. The data provide strong support for all three
paper studies the effects of the composition of capital implications, even after controlling for other factors
inflows on output volatility. The paper develops a simple that may influence output volatility, and after dealing
empirical model which, under certain conditions that with potential endogeneity problems. These findings call
hold in the data, yields three key testable implications. attention to the importance of taking into account the
First, output volatility should depend positively on the synchronization and composition of capital flows for
volatilities of both foreign direct investment and portfolio output stabilization purposes, as opposed to just focusing
and other inflows. Second, output volatility should on the volatility of each component of capital flows.
be an increasing function of the correlation between
This paper is a product of the Macroeconomics and Growth Team, Development Research Group. It is part of a larger
effort by the World Bank to provide open access to its research and make a contribution to development policy discussions
around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author
may be contacted at Vegh@econ.umd.edu.
The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development
issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the
names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those
of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and
its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.
Produced by the Research Support Team
The effect of capital flows composition
on output volatility1
Pablo Federico Carlos A. Vegh
BlackRock University of Maryland and NBER
Guillermo Vuletin
Colby College
JEL codes: F23, F32, F36, F44.
Keywords: Foreign direct investment, capital inflows, output volatility.
1
This paper was written while the authors were visiting the World Bank's Office of the Chief Economist
for Latin America and the Caribbean and Vegh and Vuletin were also visiting the Macroeconomics and
Growth Division (DEC) at the World Bank. They are very grateful for the hospitality and stimulating
policy and research environment and would like to thank Laura Alfaro, Constantino Hevia, Sebnem
Kalemli-Ozcan, Daniel Riera-Crichton, Andreas Waldkirch, and, especially, Luis Serven for extremely
helpful comments and suggestions. This research was partly funded by the Knowledge Change Program
(KCP).
1 Introduction
There is by now a large literature that has focused on the eÂ¤ects of foreign direct investment
(F DI ) on growth. The overall consensus appears to be that, provided that the right economic
environment is in place, F DI will indeed stimulate growth.1 Much less attention, if at all, has
been paid to the link between capital inâ€¡ows volatility and output volatility.2 In particular,
there has been little formal analysis of the idea â€“ often found in policy circles â€“ that F DI
should be encouraged because it should lead to lower output volatility. As the logic goes,
F DI is more stable than other sources of capital inâ€¡ows, most notably portfolio and other
investments (OT R), 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 F DI leads to a less volatile level
of output.3 To organize the discussion and provide a guide to the empirical analysis, we â€¦rst
develop a simple empirical model of the relationship between output and capital inâ€¡ows. The
s portfolio
model draws on standard portfolio theory in which the volatility of an investorâ€™
depends on the volatilities of the underlying investments. We show that output volatility
depends not only on the volatility of F DI and OT R but also on the correlation between F DI
and OT R and the share of F DI in total capital inâ€¡ows.
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
1
See, among others, Alfaro et al (2007, 2010), Borenzstein et al (1998), and Carkovic and Levine (2005).
2
A related question â€“the potential beneâ€¦cial eÂ¤ects of FDI in reducing the frequency of crises and/or sudden
stops â€“has been addressed in Fernandez-Arias and Hausmann (2001) and Levchencko and Mauro (2007). For
an early, skeptical look at the notion that long-term â€¡ows may be stabilizing, see Claessens, Dooley, and Warner
(1995). Using â€¦rm level data, Alfaro and Chenz (2012) analyze the role of FDI on establishment performance
before and after the global â€¦nancial crisis of 2008.
3
At a â€¦rm level and for European countries, Kalemli-Ozcan, SÃ¸rensen, and Volosovych (2010) â€¦nd a positive
eÂ¤ect of foreign ownership on volatility of â€¦rmsâ€™outcomes.
2
that lower F DI volatility should lead to lower output volatility. This is not, however, nec-
essarily the case. In fact, if the correlation between F DI and OT R is negative, then lower
F DI volatility will increase output volatility because F DI cannot provide as much insurance
against the volatility of OT R. By the same token, another â€œobviousâ€? idea â€“ that a higher
share of F DI should lead to lower output volatility â€“ is only true in the model if the actual
share of F DI in total capital inâ€¡ows is below the share of F DI that minimizes overall output
variability.
We use the model to derive three key testable implications:
If the correlation between F DI and OT R is zero or positive, output volatility should
also depend positively on F DI volatility and OT R volatility.
Output volatility should be an increasing function of the correlation between F DI and
OT R.
Output volatility should be a decreasing function of the share of F DI in total capital
inâ€¡ows, particularly when its initial value is low.
We test the modelâ€™s predictions using a sample of 59 countries for the period 1970-2009.4
For this purpose, we construct â€¦ve-year non-overlapping series of volatilities and other port-
folio and macroeconomic variables. Our empirical â€¦ndings strongly support our modelâ€™s im-
plications.
We control for other possible determinants of output volatility, such as government spend-
ing volatility, terms of trade volatility, and country instability. We address endogeneity con-
cerns by using three sets of instruments: (i) â€¦ve-year non-overlapping lags of portfolio vari-
4
The number of countries was dictated by data availability on foreign direct investment and other capital
â€¡ows.
3
ables, (ii) gravity-based portfolio variables aiming at capturing regional/location eÂ¤ects, and
(iii) â€¦ve-year non-overlapping lags of de jure and de facto measures of restrictions on cross-
border â€¦nancial transactions. Our main â€¦ndings 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 inâ€¡ows
and output.
2 Empirical model
To organize ideas and guide our empirics, this section develops a simple empirical model of
the relation between output volatility and capital inâ€¡ows volatility. Consider a small open
economy with the following technology:
Y = AK; (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
s addition to the existing capital stock:
The capital stock consists of this periodâ€™
K=K 1 + K:
5
This price could also be interpreted as a rental price. For the purposes of our analysis, we will assume
that this price does not change over time.
4
Assume that the purchase of this periodâ€™s capital good must be fully â€¦nanced by capital in-
â€¡ows, either in the form of foreign direct investment (F DI ) or portfolio and other investments
(OT R).6 Formally,
F DI + OT R
K= :
p
Solving for p and substituting in (1), we obtain
~
Y = AK 1
~ F;
+ AT (2)
where
TF F DI + OT R; (3)
denotes total capital inâ€¡ ~
ows and A A=p. Output is thus a linear function of total capital
inâ€¡ows.7
Let Y and T F denote the means of output and total capital â€¡ows, respectively. It then
follows from (2) that
2
Y
~2
=A 2
TF : (4)
Output volatility is thus an increasing function of capital inâ€¡ows volatility. To proceed
further, we need to impose more structure. Speciâ€¦cally, let us assume that the stochastic
6
We are abstracting, of course, from domestic savings as a source of â€¦nancing in order to focus exclusively
on the eÂ¤ects of volatility of foreign â€¦nancing on domestic output.
7
This very tight link between output and capital â€¡ ows is the key assumption behind our empirical model.
(While helpful to organize the empirical work, what folows below is, formally, a mechanical elaboration of this
main idea in a stochastic setting and involves no implicit theorizing.) To provide some theoretical basis for this
assumption, the appendix develops a simple theoretical framework with heterogenous â€¦rms which delivers an
equilibrium relationship between F DI and OT R, on the one hand, and output on the other. In this context,
the appendix shows how â€¡ uctuations in, for instance, the cost of long-term â€¦nancing leads to â€¡ uctuations in
output, F DI and OT R.
5
processes for F DI and OT R take the following multiplicative form:
F DI = F DI (1 + "F DI ) ; (5)
OT R = OT R (1 + "OT R ) ; (6)
where F DI and OT R are the means of F DI and OT R, respectively, "F DI s N 0; 2 ,
F DI
"OT R s N 0; 2 , and "F DI and "OT R are jointly normally distributed. For further refer-
OT R
ence, let denote the correlation between "F DI and "OT R .8
Let denote the share of F DI in total capital inâ€¡ows; that is,
F DI = TF; (7)
OT R = (1 )TF; (8)
where T F is the mean of T F:Since T F is the sum of F DI and OT R, it will inherit the
multiplicative stochastic structure of F DI and OT R. To see this, substitute (5) and (6) into
(3), and use (7) and (8), to obtain
T F = T F [ (1 + "F DI ) + (1 ) (1 + "OT R )].
Hence,
h i
2
2
TF = TF 2 2
F DI + (1 )2 2
OT R + 2 (1 ) F DI OT R : (9)
8
The normality assumption is not essential for our results to go through.
6
From (2), 2 ~2
=A 2 . Using (9), we can express output volatility as
Y TF
2h i
2
Y
~ F
= AT 2 2
F DI + (1 )2 2
OT R + 2 (1 ) F DI OT R : (10)
This equation thus relates output volatility ( 2) to the volatility of foreign direct invest-
Y
ment ( 2 2
F DI ), the volatility of portfolio and other investments ( OT R ), the correlation between
F DI and OT R ( ), and the share of F DI in total capital inâ€¡ows ( ).
Even though the model is extremely simple, expression (10) already warns us that some
commonly-held beliefs regarding the beneâ€¦cial role of F DI 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 inâ€¡ows volatility does imply
lower output volatility, equation (10) tells us that whether lower F DI volatility will actually
translate into lower output volatility depends on the correlation between F DI and OT R. As
will become clear below, if < 0, lower F DI volatility could actually lead to higher output
volatility! Also, the eÂ¤ect of on output volatility is, in principle, ambiguous. In fact, it is
possible that a larger share of F DI 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 F DI and OT R are the same and the correlation is one (i.e.,
2 = 2 and = 1). Expression (10) then reduces to
F DI OT R
2
2
Y
~ F
= AT 2
F DI .
7
Output volatility does not depend on . Since the variances of F DI and OT R are the
same and the correlation is one, there is essentially no diÂ¤erence between F DI and OT R
and hence the share of F DI is irrelevant. In this case, higher F DI (or OT R) volatility
translates into higher output volatility.
Case 2: Variances of F DI and OT R are the same, = 0:5, and there is a perfect
negative correlation (i.e., 2 = 2 and = 1). In this case, 2 = 0. This can be
F DI OT R Y
thought of as the â€œfull insuranceâ€?case. Due to the perfectly negative correlation, equal
variances, and equal share, total capital inâ€¡ows are constant and hence output volatility
is zero.
Case 3. Variances of F DI and OT R are the same and the correlation is zero (i.e.,
2 = 2 and = 0). In this case, expression (10) reduces to
F DI OT R
2 h i
2
Y
~ F
= AT 2
F DI
2
+ (1 )2 : (11)
This is the typical benchmark in basic portfolio theory. Think of an investor with
two uncertain sources of income (F DI and OT R) that have the same variance but are
uncorrelated. What is the share of F DI that would minimize the volatility of the
overall portfolio? Set 2 = 2 and = 0 in (10) and diÂ¤erentiate with respect to
F DI OT R
to obtain,
d 2 2
Y ~ F
= 2 AT 2
F DI (2 1) 7 0: (12)
d
This expression is zero for = 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
8
sources of income that have the same variance, splitting the portfolio in half minimizes
the overall volatility.
Deviating marginally from = 1=2 has, of course, no â€¦rst-order eÂ¤ect on output volatil-
ity. For values of 6= 1=2, however, increasing if > 1=2 or reducing if < 1=2
will increase output volatility because the shares are getting farther away from the
variance-minimizing mix. Formally:
d 2Y ~ 2
= 2A F DI (2 1) > 0; (13)
d >1 = 2
d 2Y ~ 2
= 2A F DI (2 1) < 0: (14)
d <1 = 2
Figure 1 illustrates this case by plotting 2 as a function of for = 0 and =
Y F DI
OT R = 30. We can see that, as equation (23) indicates, the variance-minimizing value
of is 0.5 (point A). Given the U-shape of the curve, moving away from point A in
either direction increases 2. Point B indicates the median value of in our sample;
Y
0.31. For any between this value and 0:5, increasing will reduce output variability.
What happens if we deviate from this benchmark portfolio case in terms of being
diÂ¤erent 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 is diÂ¤erent from zero (i.e., 2 = 2 and
F DI OT R
6= 0)
If the correlation is not zero, then it will still be the case that the value of that
minimizes output volatility is one-half. Indeed, set 2 = 2 in equation (10) and
F DI OT R
9
diÂ¤erentiate with respect to to obtain
d 2 2
Y ~ F
= 2 AT 2
F DI (2 1) (1 ) 7 0;
d
which is zero for = 1=2. Intuitively â€“and as (10) makes clear â€“a positive correlation
increases overall volatility relative to the = 0 case but does not change the fact that,
since F DI and OT R are not perfectly correlated, the variance-minimizing is still
one-half.
Case 5. Correlation is zero but variances are diÂ¤erent (i.e., = 0 and 2 6= 2
F DI OT R ).
In this case, the variance-minimizing will change. To see this, set = 0 in (10) and
diÂ¤erentiate with respect to to obtain
d 2 2
Y ~ F
= 2 AT ( 2
F DI + 2
OT R )
2
OT R 7 0:
d
Setting this expression to zero, we obtain the variance-minimizing value of :
2
min OT R
= 2 2 . (15)
F DI + OT R
2 2 min
If F DI < OT R , then > 1=2. Intuitively, if F DI is less volatile than OT R, then
it would be optimal to hold more than one-half of the T F as F DI . Even though the
variance-minimizing is larger than one-half, the same intuition developed in Case 3
above holds: deviating from this variance-minimizing value of will increase overall
volatility.
10
Needless to say, in practice countries cannot choose the variance-minimizing value of .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 F DI = 58:8, OT R =
168:5, and = 0:23.10 In this case, the variance-minimizing value of is 0.85, given by point
A. Given the U-shape of the curve, moving away from point A in either direction increases
2. Point B is the actual value of for Turkey, = 0:14. Since this value of is less than
Y
the variance-minimizing , increasing will reduce output volatility. This will be one of the
main empirical predictions of our model.
Returning now to the general case captured in equation (10), let us examine how changes
in , 2 2
F DI , and OT R aÂ¤ect output volatility. Taking the corresponding partial derivatives,
we obtain
d 2Y 2
~2 T F [ (1
= 2A ) F DI OT R ] > 0; (16)
d
d 2 2
Y ~2 T F
= 2A 2
F DI + (1 ) OT R ? 0; (17)
d F DI
d 2 h i
2
Y ~2 T F
= 2A (1 )2 OT R + (1 ) F DI ? 0: (18)
d OT R
As equation (16) makes clear, a higher always increases output volatility. On the other
hand, expressions (17) and (18) indicate that the eÂ¤ects of 2 and 2 are ambiguous.
F DI OT R
To understand this ambiguity, think of the case in which = 1, = 0:5, and F DI < OT R .
9
Even though we could certainly interpret various measures that emerging countries often adopt to encourage
F DI at the expense of other, more volatile, â€¡ ows as an attempt to increase and reduce output volatility.
10
See the data section below for the interpretation of the units in which the standard deviations of F DI and
OT R are expressed.
11
Equation (17) then reduces to
d 2
Y ~ F2 2
= 2AT ( F DI OT R ) < 0:
d F DI
Here a reduction in F DI volatility would increase output volatility. Intuitively, with perfect
negative correlation between foreign direct investment and portfolio and other investments
and F DI < OT R , F DI volatility is actually a good thing because then F DI can oÂ¤er more
insurance against OT R. In other words, if F DI exhibits very low volatility, then it cannot
oÂ¤set the much higher volatility of OT R.
In the data, however, is on average close to zero (sample median is 0.05), in which case
an increase in the volatility of either F DI or OT R will increase output volatility. Intuitively,
with zero correlation, higher volatility is unambiguously bad because it contributes to output
volatility directly without oÂ¤ering 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 F DI and
OT R.
Output volatility should be an increasing function of F DI volatility and OT R volatility
(under the assumption that 0).
Output volatility should be a decreasing function of the share of F DI in total capital
inâ€¡ows (under the assumption that the actual value of is below the variance-minimizing
value of ).
12
3 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 inâ€¡ows,
government spending, inâ€¡ation, and terms of trade data comes from International Financial
Statistics (IFS) and World Economic Outlook (WEO), both from the IMF. For capital â€¡ows,
we use foreign direct investment, portfolio investment, and other investment gross inâ€¡ows
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 F DI and denote this aggregate
by OT R.12
The standard deviations and correlations of all variables are computed based on their
cyclical components. For this purpose, we use the Hodrick-Prescott â€¦lter 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 F DI mentioned above for
Turkey ( F DI = 58:8) means that, on average, the level of F DI is 58.8 percentage points
away from its trend. Given that OT R = 168:5 for Turkey, this implies that portfolio and
other investments are almost three times as volatile as F DI .
Another common practice in the literature (see, for instance, Albuquerque, Loayza, and
11
Industrial countries comprise Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany,
Greece, Ireland, Italy, Japan, Netherlands, New Zealand, Portugal, Spain, Sweden, Switzerland, United King-
dom, and United States. Developing countries comprise Argentina, Bangladesh, Brazil, Cambodia, Cape
Verde, Chile, Colombia, Costa Rica, Czech Rep., Ecuador, El Salvador, Estonia, Georgia, Guatemala, Hong
Kong (SAR China), Hungary, India, Indonesia, Israel, Jordan, Korea, Latvia, Lithuania, Malaysia, Mexico,
Mozambique, Pakistan, Panama, Paraguay, Philippines, Romania, Russia, Singapore, South Africa, Sudan,
Thailand, Turkey, Uruguay, and Venezuela.
12
Speciâ€¦cally, OTR includes portfolio investment (i.e, equity and portfolio debt â€¡ows) as well as loans,
currency, and trade credits.
13
Serven, 2005) is to normalize capital â€¡ows such as F DI by dividing them by GDP. The
rationale behind this methodology is to control for country size and avoid nonstationarity
problems. While helpful in a diÂ¤erent 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 F DI and output. Since the latter will be our dependent variable, our empirical analysis
would suÂ¤er from endogeneity problems by construction. Moreover, our focus on the cyclical
component of capital inâ€¡ows 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 inâ€¡ows. Indeed, using cross-country data, we cannot reject the null hypothesis that
the correlation between F DI and average F DI , as well as the correlation between OT R and
average OT R, are equal to zero at a 5 percent signiâ€¦cance 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 inâ€¡ows volatility.
Not surprisingly, the median of total gross inâ€¡ows into developing countries is more than one
and a half times that of industrial countries.14
We now turn to the volatilities of F DI and OT R. Figure 5 shows the ratio of OT R
volatility to F DI volatility. The â€¦gure is consistent with the idea in the literature that OT R
inâ€¡ows are more volatile than F DI inâ€¡ows. Indeed, the ratio is larger than one for more
than 85 percent of the countries in our sample. The median volatility of OT R is close to 120,
13
In this and following plots, light (yellow) bars denote developing countries while black bars indicate indus-
trial countries.
14
Omitting Sudan and Korea (which have very high total gross inâ€¡ ows volatility) does not aÂ¤ect our results.
14
compared to less than half (about 44) for F DI . Moreover, the median in developing countries
ratios is 76 percent higher than that in industrial economies, reâ€¡ecting in particular the higher
volatility of OT R. In fact, the median F DI volatility is 48 for industrial counties and 41
for developing countries. In sharp contrast, the median OT R volatility is about 30 percent
higher in developing countries than in advanced economies (120 for developing countries and
85 for industrial countries).
Figure 6 shows that the share of F DI in total gross capital inâ€¡ows 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 OT R
and F DI . We can see wide variation in this â€¦gure across countries, with the sample median
being 0.05 and the median for developing countries -0.02.
Taking into account the median values of the ratio of OT R volatility to F DI volatility and
the correlation between OT R and F DI for industrial and developing countries, we â€¦nd that
the (i.e., the share of F DI in total capital inâ€¡ows) 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 diÂ¤erence in the optimal shares of F DI reâ€¡ects the fact that (i) the relative
ratio of OT R volatility to F DI volatility is higher in developing countries than in industrial
ones (3.7 versus 2.1) and (ii) the correlation between OT R and F DI is positive (0.14) for
industrial countries but slightly negative (-0.02) for developing countries. In other words, a
higher share of F DI in total capital inâ€¡ows is more beneâ€¦cial for developing countries than for
industrial economies because (i) it reduces total capital â€¡ows volatility directly by substituting
15
a more volatile source of capital (OT R) for one that is less volatile (F DI ) and (ii) it provides
some insurance given the negative (though rather small) correlation.
4 Empirical evidence
In this section we test the main empirical implications derived in Section 2. First, output
volatility should depend positively on F DI and OT R volatility. Second, output volatility
should be an increasing function of the correlation between F DI and OT R. Third, for low
values of the F DI share, output volatility should be a decreasing function of the share of
F DI in total capital inâ€¡ows.15
We â€¦rst show our benchmark regressions that link output volatility to the variables high-
lighted in the empirical model of Section 2. We then control for other variables that, in
practice, could aÂ¤ect output variability. We then address endogeneity problems.
4.1 Basic regressions
Following the empirical growth literature, we use non-overlapping â€¦ve-year averages. Table 1
reports the basic results using country and â€¦ve-year â€¦xed eÂ¤ects. Standard errors are robust
and we also allow for within-country correlation (i.e., clustered by country). We normalize
F DI and OT R to be between 0 and 100 to make regression coeÂ¢ cients 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 F DI and OT R volatility increase output
15
In principle, one would like to evaluate the interaction eÂ¤ects in a more elaborated way (i.e., by introducing
all neccesary interaction terms). Sample size, however, severely restricts our ability to follow such an approach.
16
After the normalization, F DI and OT R range between 0 and 14.89 and 0.04 and 100, respectively.
16
volatility (columns 1 and 2). A higher correlation between F DI and OT R increases output
volatility (column 3). When we include the share of F DI in total capital inâ€¡ows (column
4), it appears not to matter, contrary to our modelâ€™s prediction. However, as captured by
(23)-(14), the expected relationship between a higher share of F DI in total inâ€¡ows 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 F DI
will reduce output volatility when the initial share is smaller than 0.5 (see equation (14)). To
capture this eÂ¤ect, 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 F DI in
total capital inâ€¡ows only when its initial value is low. Finally, when all explanatory variables
are included (column 6), the size of the coeÂ¢ cients and signiâ€¦cance levels remain essentially
unchanged.
4.2 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
aÂ¤ect output volatility. While the basic regressions of Subsection 4.1 control for country
and â€¦ve-year â€¦xed eÂ¤ects, other factors such as idiosyncratic external shocks, â€¦scal policy
volatility, and country instability could also aÂ¤ect 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
17
of internal and external conâ€¡icts from the International Country Risk Guide (ICRG). Internal
conâ€¡ict 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 conâ€¡ict 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 conâ€¡icts to all-out war).
The risk rating assigned is composed of three subcomponents: war, cross-border conâ€¡ict, 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 eÂ¤ects of the control variables
one at a time. The three variables have the expected signs: higher â€¦scal and terms of
trade volatility and more country instability increase output volatility. Surprisingly enough,
however, terms of trade volatility is not statistically signiâ€¦cant. The reason is that we are also
including â€¦ve-year â€¦xed eÂ¤ects. If such â€¦xed eÂ¤ects are not included, then the coeÂ¢ cient of
the terms of trade volatility is positive and signiâ€¦cant 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 reâ€¡ected, 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), â€¦scal policy volatility becomes insigniâ€¦cant due
18
to its high correlation with country instability.17 More importantly for our purposes, Column
6 indicates that the size and signiâ€¦cance of our four explanatory variables (F DI volatility,
OT R volatility, correlation between F DI and OT R, and interacted share of F DI ) remain
essentially unchanged relative to column 6 in Table 1.
4.3 Addressing endogeneity
This section addresses potential endogeneity problems. One could reasonably argue that the
positive relationship between output volatility and F DI and OT R volatility may reâ€¡ect the
fact that higher GDP volatility increases the volatility of capital inâ€¡ows. In other words,
the causality may run from output volatility to inâ€¡ows volatility rather than the other way
around. In the same vein, reductions in the share of F DI could reâ€¡ect the reluctance of
foreign â€¦rms to invest for the long-term in highly volatile economies.
As is the case in the empirical macro literature that has assessed the inâ€¡uence of F DI on
economic growth (see, for instance, Lensink and Morrisey, 2001 and Alfaro, 2003), we lack
obvious instruments for F DI , OT R , (F DI; OT R), and F DI share. We then use three sets
of instruments. First, we follow the above-mentioned macro literature in using lagged F DI
as an instrument for current F DI . In our case, this amounts to using the lagged â€¦ve-year
average of each portfolio variable as an instrument. For example, we use the OT R for the
s rank correlation
period 1970-1974 to instrument for the period 1975-1979. The Spearmanâ€™
between F DI and its lagged â€¦ve-year value is 0.38. The corresponding correlation is 0.26 for
OT R and 0.30 for F DI share. In all cases, the correlation is statistically signiâ€¦cant at the 5
percent level. In other words, there seems to be a positive association between the volatilities
of capital inâ€¡ows over time, even at the â€¦ve-year frequency. Unfortunately, the correlation
17
The correlation is 0.40 and statistically diÂ¤erent from zero at the one percent level.
19
between (F DI; OT R) and its lagged â€¦ve-year value is statistically insigniâ€¦cant.
Our second set of instruments uses a geographical/gravity approach aimed at capturing
the inâ€¡uence of regional eÂ¤ects. Capital inâ€¡ows respond to economic and political fundamen-
tals which are often shared by diÂ¤erent countries within a region (Calvo and Reinhart, 1996;
Fernandez-Arias and Montiel, 1995; Alba, Bhattacharya, Claessens, Gosh, and HernÃ¡ndez,
2000; Corbo and HernÃ¡ndez, 2001). During the 1980s, for example, Latin America expe-
rienced such political and economic instability that international investors became reluctant
to invest for the long-term. Indeed, F DI 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:
X 1
Iit = Ijt ; i 6= j;
distij
j
where Iij represents the portfolio variable and distij measures the distance between the capital
cities of countries i and j . In other words, we instrument a countryâ€™s â€¦ve-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
s rank correlation
more generalized version of the idea behind regional eÂ¤ects. The Spearmanâ€™
between (F DI; OT R) and the suggested geographical instrument is 0.35 and statistically
signiâ€¦cant at the 5 percent level. The gravity approach proves to be a good strategy to
predict the patterns of correlation between F DI and OT R. Unfortunately, the correlations
for the other portfolio variables are statistically insigniâ€¦cant.
To complement the two previous sets of instruments, we also rely on the literature regard-
20
ing the determinants of capital â€¡ows. In particular we focus on determinants that may help
determine portfolio variables, but not have a direct eÂ¤ect on output. Montiel and Reinhart
(1999) â€¦nd that, by imposing capital controls, countries are able to increase the share of F DI .
Generally speaking, policies that punish short-term â€¡ows should, in principle, induce foreign
investors to increase long-term â€¡ows. We use three variables to account for this eÂ¤ect. First,
we use the Chinn-Ito index (Chinn and Ito, 2006) to measure de jure â€¦nancial openness. This
s capital account openness, is based on a binary dummy
index, which measures a countryâ€™
variable that codiâ€¦es the tabulation of restrictions on cross-border â€¦nancial transactions re-
ported in the IMFâ€™s Annual Report on Exchange Arrangements and Exchange Restrictions.
A high value of this index is an indication of low de jure â€¦nancial integration. Second, we use
the ratio of total foreign assets and liabilities to GDP from Lane and Milesi-Ferretti (2007)
to measure de facto â€¦nancial integration. A high value of this index indicates a high de-
gree of de facto â€¦nancial integration. Lastly, we use the investment proâ€¦le index from the
International Country Risk Guide (ICRG). This investment proâ€¦le assesses factors aÂ¤ecting
the risk to investment that are not covered by other political, economic, and â€¦nancial risk
components. The risk rating assigned is composed of three subcomponents: contract viabil-
ity/expropriation, proâ€¦ts 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 â€¦ve-year lag of these three variables to take care of reverse causality concerns (i.e. ,the
s rank correlation
possibility that output volatility leads to investment risk). The Spearmanâ€™
between F DI share and the â€¦ve-year lag of the Chinn-Ito index is -0.14, while the corre-
lation with the â€¦ve-year lag of investment proâ€¦le is 0.18. In both cases, the correlation is
s
statistically signiâ€¦cant at the 5 percent level. These â€¦ndings support Montiel and Reinhartâ€™
21
(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 â€¦nancial openness) reduce
OT R . s rank correlation between
The Spearmanâ€™ OT R and the â€¦ve-year lag of the Chinn-Ito
index is statistically signiâ€¦cant 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 regres-
sions. Table 3 shows the instrumental variable regressions. In all cases we cannot reject the
overidentiâ€¦cation tests at a 5 percent conâ€¦dence 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 conâ€¦rm 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 â€¦ndings hold. The only exception is OT R : while the sign
of the coeÂ¢ cient is positive, it is not statistically signiâ€¦cant.18
We now add control variables. While terms of trade volatility is typically treated as exoge-
nous, 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 causal-
ity, we use the â€¦ve-year lag of government spending volatility and country instability. The
Spearmanâ€™s rank correlation between government spending and its â€¦ve-year lag is 0.56 and the
corresponding correlation for country instability is 0.84. In both cases, the correlation is
statistically signiâ€¦cant at the 5 percent level. Columns 2 to 4 show the results of including
18
It is worth noting that the sample size of the instrumental variable regression has fallen by almost 45
percent (from 295 and 59 countries in Table 1 to 171 and 38 countries in Table 3).
22
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.
5 Conclusions
A commonly-held belief is that a larger share of F DI in total capital inâ€¡ows will reduce
output volatility. There is, however, little, if any, formal evidence on this channel. Based
on standard portfolio theory, we â€¦rst develop a simple econometric model that calls attention
to some important caveats. In particular, lower F DI volatility will reduce output volatility
only if the correlation between F DI and other â€¡ows is positive (which is not always the case
in the data). Also, a larger share of F DI will reduce output volatility only if the actual
share of F DI is below the variance-minimizing share. Our model thus yields three testable
implications: (i) output volatility should depend positively on F DI and OT R volatility; (ii)
output volatility should be an increasing function of the correlation between F DI and OT R;
and (iii) output volatility should be a decreasing function of the share of F DI in total capital
inâ€¡ows (when the initial share is low). We â€¦nd strong support in the data for all three
implications, even after controlling for other factors that inâ€¡uence output volatility and for
possible endogeneity problems.
6 Appendix
This appendix develops a simple theoretical model that provides a theoretical illustration of
the key assumption in the empirical model â€“ as captured in equation (2) â€“ that there exists
a tight link between output and capital inâ€¡ows. In the theoretical model, such a link will
23
arise endogenously as â€¦rms choose whether to â€¦nance investment with either short-term or
long-term external funding.19 In the aggregate, the economy uses both sources of â€¦nance and
changes in, say, the cost of external funding will lead to changes in output, external â€¦nance,
and its composition.
Consider a small open economy with a continuum of risk-neutral â€¦rms that produce the
same â€¦nal (tradable) good, denoted by q , using the same (tradable) capital, denoted by k .
Firms are indexed by their productivity parameter, (0 < < 1), which is the only source of
heterogeneity. Firms â€œliveâ€? 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 â€¦rm is given by
qt = k ; t = 1; 2;
where > 0 is a productivity parameter. By construction, output is constant across periods.
Firms need to borrow from abroad to â€¦nance 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 â€¦rst 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,
rs and rl . We assume that rs < rl , reâ€¡ecting the idea that international lenders may have a
preference for a more â€œliquidâ€? asset.
As an important benchmark, we â€¦rst solve the â€¦rmâ€™s problem under short-term â€¦nancing
19
At the cost of complicating the model, we could have included domestic saving as well. Our model, however,
can be interpreted as applying to funding needs that go beyond domestic savings, the typical situation for a
developing country.
24
and no repayment constraint. We then impose the repayment constraint for short-term
â€¦nancing. We then solve for the case of long-term â€¦nancing. We then compare proâ€¦ts in
the two cases (short-term â€¦nancing and repayment constraint versus long-term â€¦nancing) to
â€¦nd out when a â€¦rm will chose one or the other. We â€¦nally aggregate over all â€¦rms to obtain
the economyâ€™s aggregate capital stock and output and analyze how the equilibrium changes
if the cost of long-term â€¦nancing changes.
6.1 Short-term â€¦nancing and no repayment constraint
Denote by p the world relative price of q in terms of k and by bt , 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 â€¦rm operates (i.e., produces and sells). The â€¡ow budget constraints
are thus given by
b0 = k; (19)
b1 = (1 + rs )b0 + p k 1; (20)
0 = (1 + rs )b1 + p k 2; (21)
where t, t = 1; 2, denotes dividends paid by the â€¦rm. Combining these â€¡ow constraints, we
obtain an intertemporal constraint:
(2 + rs )
= p k k; (22)
(1 + rS )2
where ( s) + s )2 )
1 =(1 + r 2 =(1 + r is the present discounted value of proâ€¦ts as of time 0.
25
Firms choose k to maximize (22). The â€¦rst-order condition for capital takes the form:
(2 + rs ) 1
p k = 1: (23)
(1 + rs )2
At an optimum, the â€¦rm equates the present discounted value of the value of the marginal
productivity to the cost of capital. Solving for the capital stock, we obtain
1
(2 + rs ) 1
k= p : (24)
(1 + rs )2
Substituting this expression into (22), we can write proâ€¦ts as:
1
=k 1 : (25)
As expected, proâ€¦ts are positive since, by assumption, 2 (0; 1).
In the absence of any additional constraint, all â€¦rms would choose short-term â€¦nancing
because, by assumption, it is cheaper than long-term â€¦nancing. To have a meaningful choice
between short-term and long-term â€¦nancing, we will now introduce a repayment constraint.
6.2 Short-term â€¦nancing and repayment constraint
Suppose now that a â€¦rm can access short-term credit only if it can pay back the loan at the
end of the â€¦rst period. Formally,
p k (1 + rs )k > 0. (26)
26
Let us check if this repayment constraint binds for the unconstrained problem that we just
solved. To this eÂ¤ect, substitute (24) into the last expression to obtain
1 + rs
> :
2 + rs
Firms whose satisâ€¦es this condition will thus still be able to choose short-term â€¦nancing
and remain at the â€¦rst best because the repayment constraint does not bind. Intuitively,
low â€¦rms optimally choose a low level of capital (i.e., units of capital with high marginal
productivity) and are thus more likely to satisfy constraint (26) given that the repayment cost
per unit of capital (1 + rs ) does not depend the level of capital.
On the other hand, the unconstrained solution for â€¦rms with > (1 + rs )=(2 + rs ) vio-
lates condition (26). These â€¦rms will thus need to choose between â€œconstrained short-term
â€¦nancingâ€? (i.e., choose the optimal level of capital subject to condition (26)) or long-term
â€¦nancing. The trade-oÂ¤ is thus between remaining in a â€¦rst-best equilibrium but facing a
higher cost of capital (long-term â€¦nancing) or choosing a constrained level of capital but at a
lower cost (constrained short-term â€¦nancing).
If constraint (26) binds, then the capital stock is given by
1
p 1
k jconstrained short-term = . (27)
1 + rs
If we compare this level of capital with the unconstrained level of capital, given by expression
(24), for a â€¦rm with > (1 + rs )=(2 + rs ), we can see that the stock of capital in the
constrained case is lower. In other words, to access short-term â€¦nancing, the â€¦rm needs to
have a suboptimally low level of capital to generate enough proâ€¦ts in the â€¦rst period to repay
27
the loan.
6.3 Maximization under long-term â€¦nancing
Let us now compute proâ€¦ts under long-term â€¦nancing. The budget constraints remain the
same as in (19)-(21) with rl in lieu of rs . Further, since a â€¦rm that chooses long-term
â€¦nancing is still operating in a â€¦rst-best world, the choice of capital will be given by condition
(24) with rl in lieu of rs . Proâ€¦ts will thus be given by (25) with the corresponding choice of
capital.
6.4 Comparison
Firms with > (1 + rs )=(2 + rs ) will choose long-term â€¦nancing over short-term â€¦nancing as
long as proâ€¦ts are larger:
jlong-term > jshort-term constrained :
Using equations (24) and (25), this condition reduces to
1
(1 + rl )2 1
1> : (28)
(2 + rl ) (1 + rs )
Suppose to â€¦x ideas that rl = rs . In this case, this last expression reduces to
1
(1 + rl ) 1
1> .
(2 + rl )
Since the choice is only relevant for â€¦rms with > (1 + rs ) = (2 + rs ), the condition will
always hold. In other words, if rl = rs , then all these â€¦rms would choose long-term â€¦nancing
28
because the cost is the same as short-term â€¦nancing but they are not subject to the repayment
constraint (which, by construction, is binding).
But our maintained assumption is, of course, that rl > rs . In that case, condition (28),
holding with equality, deâ€¦nes a threshold value of , denoted by , which is given by
(1 + rl )2
= : (29)
(1 + rs ) (2 + rl )
We now establish the following result:
Claim 1 Firms with ( < ) will choose long-term (short-term) â€¦nancing.
Proof. Consider condition (28). DiÂ¤ erentiating the right-hand side and evaluating the
corresponding expression at = , we obtain
1
1
(1+rl )2
d
(2+rl )(1+rs ) 1
= < 0:
d (1 )
=
Set = in condition (28). By construction, it will hold as an equality. An increase in
will then reduce the RHS, which means that long-term proâ€¦ts will be higher than constrained
short-term proâ€¦ts. The reverse is true for a fall in .
Intuitively, â€¦rms with a large (i.e., > ) are â€¦rms that â€¦nd it more eÂ¢ cient 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 â€¦nancing in order to not be subject to the
repayment constraint. In contrast, smaller â€¦rms (i.e., â€¦rms with < ) would rather not
pay the additional cost of â€¦nancing and choose a second-best level of capital.
From (29), we can see that increases with rl and decreases with rs . Intuitively, an
29
increase in rl makes long-term â€¦nancing more expensive. As a result, marginal â€¦rms will
choose to switch to short-term â€¦nancing (i.e., increases). Conversely, an increase in rs
makes short-term â€¦nancing more expensive and hence marginal â€¦rms will choose to switch to
long-term â€¦nancing (i.e., decreases).
6.5 Aggregation
As has been established above, there are three types of â€¦rms in this economy depending on
the value of :
1+r s
The range 0 < 2+rs consists of â€¦rms that are operating in a â€¦rst-best world with
short-term â€¦nancing.
1+r s
The range 2+r s < consists of â€¦rms that are operating under constrained short-
term â€¦nancing (i.e., these are â€¦rms that would violate the repayment constraint if they
chose the â€¦rst-best level of capital).
The range < < 1 consists of â€¦rms that are operating with long-term â€¦nancing.
Aggregate capita and output are thus given by, respectively,
Z ~
1
(2 + rs ) 1
Capital = p d
0 (1 + rs )2
Z 1 Z " # 1
p 1 1 2 + rl 1
+ d + p d ; (30)
~ 1 + rs (1 + rl )2
Z ~
(2 + rs ) 1
Output = p d
0 (1 + rs )2
Z Z " #
p 1 1 2 + rl 1
+ d + p d : (31)
~ 1 + rs (1 + rl )2
30
where ~ (1 + rs )= (2 + rs ).20
Since the â€¦rst two types of â€¦rms buy capital with short-term borrowing (denote it by
P OR), while the last type of â€¦rm buys it with long term borrowing (denote it by F DI ), we
can write
Z ~
1 Z 1
(2 + rs ) 1 p 1
P OR = p d + d ; (32)
0 (1 + rs )2 ~ 1 + rs
Z " # 1
1 2 + rl 1
F DI = p d : (33)
(1 + rl )2
We thus have an economy with heterogeneous â€¦rms in which the composition of external
â€¦nancing is endogenously determined based on each â€¦rmâ€™s productivity and the cost of short-
term and long-term â€¦nancing. This gives us a simple framework to ask how a change in the
cost of long-term â€¦nancing changes the equilibrium.
6.6 Changes in the cost of long-term â€¦nancing
What are the eÂ¤ects of a change in rl ? Speciâ€¦cally, suppose that rl is lower; how does the
equilibrium described above change?21
Using Leibniz rule, we can compute the changes in capital and output from equations (30)
20
In our model, â€¦rms hold no initial capital so the capital stock can be thought of as investment â€¦nanced,
as made clear below, by POR and FDI.
21
Technically, we are solving for the same perfect foresight path for diÂ¤erent values of rl . This can be
interpreted as either two economies with diÂ¤erent values of rl or, more appropriately for our purposes, as an
unanticipated change in rl at the beginning of a third period in which the economy goes through the same
cycle.
31
and (31), respectively:
Z " #
d (capital) 1
p 2 + rl 1
3 + rl
= p d < 0;
drl 1 (1 + rl )2 (1 + rl )3
Z " #2 1
d (Output) 1
p( )2 2 + rl 1
3 + rl
= p d < 0:
drl 1 (1 + rl )2 (1 + rl )3
Capital and output thus increase. Intuitively, a fall in rl aÂ¤ects capital and output through
two channels:
From (29), we can see that a higher rl reduces . This means that some marginal
â€¦rms that were relying on short-term â€¦nancing will switch to long-term â€¦nancing. At
the margin, however, the capital stock of these â€¦rms does not change and thus output
is not aÂ¤ected.22
The capital stock (and thus output) of â€¦rms that rely on long-term â€¦nancing increases.
What happens to F DI and P OR?
" # 1 " # 1 (2+rl )
Z 1 d
dF DI d 2+ rl 1 1
p 2+ rl 1 (1+r l )2
= p + p d < 0,
drl drl (1 + rl )2 1 (1 + rl )2 drl
dP OR d p 1
= > 0.
drl drl 1 + rs
In absolute terms, F DI increases and P OR falls. The share of F DI also increases because
F DI increases by more than total capital inâ€¡ows (given that P OR falls). Intuitively, F DI
increases for two reasons. First, â€¦rms that relied on long-term â€¦nancing are now borrowing
22
To see that capital does not change, notice that expression (24), with rl in lieu of rs and evaluated at
= , is the same as equation (27).
32
more. Second, some marginal â€¦rms that were relying on P OR have now switched to F DI .
The change in F DI is thus larger than the change in the capital stock.
It would be easy to accommodate random changes in rl in our model as long as â€¦rms
continue to be risk-neutral. In that case, uncertainty regarding changes in rl (or rs for that
matter) would not change the â€¦rmsâ€™ behavior derived above (with the expected value of rl
and rs replacing the actual values). We could imagine that every third period rl is drawn
from some distribution and the above equilibrium materializes. In such a scenario, an increase
in the volatility of rl would lead to higher volatility in output, investment, F DI , P OR, and
the respective shares. Clearly, being endogenous, the higher volatility of capital inâ€¡ows or
F DI is not â€œcausingâ€? higher output volatility. Rather they are both endogenous responses
to the higher volatility in the cost of long-term â€¦nancing.
33
7 References
Alba, Pedro, Amar Bhattacharya, Stijn Claessens, Swati Gosh, and Leonardo HernÃ¡ndez
s
(2000), â€œVolatility and contagion in a â€¦nancially integrated world: Lessons from East Asiaâ€™
recent experience,â€? in Asia Paciâ€¦c Financial Deregulation (Eds) G. de Brouwer and W.
Pupphavesa, Routledge, London.
Albuquerque, Rui (2003), â€œThe composition of international capital â€¡ows: Risk sharing
through foreign direct investment,â€?Journal of International Economics, Vol. 61, pp. 353-383.
Albuquerque, Rui, Norman Loayza, and Luis Serven (2005), â€œWorld market integration
through the lens of foreign direct investors,â€? Journal of International Economics, Vol. 66, pp.
267-295.
Alfaro, Laura (2003), â€œForeign direct investment and growth: Does the sector matter?â€?
mimeo, Harvard Business School.
Alfaro, Laura, Chanda Areendam, Sebnem Kalemli-Ozcan, and Selin Sayek (2004), â€œFDI
and economic growth: the role of local â€¦nancial markets,â€?Journal of International Economics,
Vol. 64, pp. 89-112.
Alfaro, Laura, Sebnem Kalemli-Ozcan, and Vadym Volosovych (2007), â€œCapital â€¡ows in a
globalized world: The role of policies and institutionsâ€?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), â€œDoes
foreign direct investment promote growth? Exploring the role of â€¦nancial markets on link-
ages,â€? Journal of Development Economics, Vol. 91, pp. 242-256.
Alfaro, Laura and Maggie Chen (2012), â€œSurviving the global â€¦nancial crisis: Foreign
ownership and establishment performance,â€? American Economic Journal: Economic Policy,
Vol 4, pp. 30-55.
Bank for International Settlements (2009), â€œCapital â€¡ows and emerging market economiesâ€?
CGFS Papers No. 33.
Borensztein, Eduardo, Jose De Gregorio, and Jong-Wha Lee (1998), â€œHow does foreign
direct investment aÂ¤ect economic growth?â€? Journal of International Economics, Vol. 45, pp.
115-135.
Calvo, Guillermo and Carmen Reinhart (1996), â€œCapital â€¡ows to Latin America: Is there
evidence of contagion?â€?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), â€œDoes foreign direct investment accelerate eco-
34
nomic growth?â€? 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), â€œWhat matters for â€¦nancial development? Capital
controls, institutions, and interactions,â€? Journal of Development Economics, Vol. 81, pp.
163-192.
Chuhan, Punam, Stijn Claessens, and Nlandu Mamingi (1998), â€œEquity and bond â€¡ows
to Latin America and Asia: The role of global and country factors,â€? Journal of Development
Economics, Vol. 9, pp. 439-463.
Claessens, Stijn, Michael P. Dooley, and Andrew Warner (1995), â€œPortfolio capital â€¡ows:
Hot or cold?,â€? World Bank Economic Review, Vol. 9, pp. 153-74.
Fernandez-Arias, Eduardo, and Ricardo Hausmann (2001), â€œIs foreign direct investment
a safer form of â€¦nancing?â€? Emerging Markets Review, Vol. 2, pp. 34-49.
Fernandez-Arias, Eduardo, and Peter Montiel (1995), â€œThe surge in capital inâ€¡ows to
developing countries: Prospects and policy response,â€? World Bank working paper No. 1473.
Hausmann, Ricardo and Eduardo Fernandez-Arias (2000), â€œForeign direct investment:
Good Cholesterol,â€? Inter-American Development Bank working paper No. 417.
Kalemli-Ozcan, Sebnem, Bent SÃ¸rensen, and Vadym Volosovych (2010), â€œDeep â€¦nancial
integration and volatility,â€? NBER working paper No. 15900.
Lane, Philip and Gian Maria Milesi-Ferretti (2007), â€œThe external wealth of nations mark
II: Revised and extended estimates of foreign assets and liabilities, 1970â€“2004,â€? Journal of
International Economics, Vol. 73, pp. 223-250.
Lensink, Robert, and Oliver Morrisey (2001), â€œForeign direct investment: Flows, volatility
and growth in developing countries,â€? University of Nottingham SOM Research Report No.
E16.
Levchenko, Andrei, and Paolo Mauro (2007), â€œDo some forms of â€¦nancial â€¡ows protect
from sudden stops?â€? World Bank Economic Review, Vol 21, pp. 389-411.
Montiel, Peter and Carmen Reinhart (1999), â€œDo capital controls and macroeconomic
policies inâ€¡uence the volume and composition of capital â€¡ows? Evidence from the 1990â€™s,â€?
Journal of International Money and Finance, Vol. 18, pp. 619-635.
Ravn, Morten, and Harald Uhlig (2002), â€œOn adjusting the Hodrick-Prescott â€¦lter for the
frequency of observations,â€? Review of Economics and Statistics, Vol. 84, pp. 371-376.
35
Figure 1. Output volatility and share of FDI in total gross capital inflows.
Ïƒ(FDI)= Ïƒ(OTR)=30, Ï?(FDI, OTR)=0
100
90
80
70
Point B
Output volatility
60
Point A
50
40
30
20
10
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Share of FDI in total gross capital inflows
Figure 2. Output volatility and share of FDI in total gross capital inflows.
Case of Turkey. Ïƒ(FDI)= 58.8, Ïƒ(OTR)=168.5, Ï?(FDI, OTR)= -0.23
30000
25000
Point B
20000
Output volatility
15000
10000
Point A
5000
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Share of FDI in total gross capital inflows
0
1
2
3
4
5
6
7
8
France
Belgium
Austria
Australia
Pakistan
Netherlands
Guatemala
Spain
Denmark
Germany
Canada
Italy
Median industrial: 1.4
Japan
Median developing: 2.5
United Kingdom
Sweden
Colombia
United States
New Zealand
Switzerland
India
Greece
South Africa
Russia
Portugal
El Salvador
Finland
Paraguay
Korea
Brazil
Philippines
Ireland
Costa Rica
Indonesia
Thailand
Malaysia
Figure 3. Output volatility
Singapore
Mexico
Cambodia
Hungary
Ecuador
Bangladesh
Cape Verde
Hong Kong
Czech Rep.
Turkey
Georgia
Uruguay
Panama
Sudan
Jordan
Venezuela
Mozambique
Romania
Chile
Argentina
Estonia
Lithuania
Israel
Latvia
0
500
1000
1500
2000
2500
Costa Rica
Czech Rep.
Georgia
Cambodia
El Salvador
India
Romania
Austria
Australia
Hungary
Germany
Estonia
France
Median developing: 82.2
Median industrial: 53.6
Cape Verde
Spain
Canada
United States
Guatemala
Chile
Latvia
Belgium
Mozambique
Singapore
New Zealand
Netherlands
Ireland
Bangladesh
Pakistan
Portugal
Russia
Note: Korea was excluded from this figure due to its extremely high median volatility (4899).
Italy
Israel
Jordan
Colombia
South Africa
Lithuania
Ecuador
Mexico
Figure 4. Total gross inflows volatility
Indonesia
Turkey
Sweden
United Kingdom
Venezuela
Malaysia
Thailand
Panama
Hong Kong
Denmark
Philippines
Brazil
Greece
Paraguay
Switzerland
Argentina
Finland
Japan
Uruguay
Sudan
0
10
20
30
40
50
60
70
80
90
Jordan
Ireland
Germany
Bangladesh
Austria
Australia
South Africa
India
Hungary
Canada
Denmark
Lithuania
Guatemala
Median developing: 3.7
Median industrial: 2.1
Romania
Latvia
United States
Belgium
Czech Rep.
Netherlands
Cape Verde
Italy
Costa Rica
New Zealand
France
Singapore
Turkey
Paraguay
Israel
Russia
Chile
Panama
Thailand
Spain
Georgia
Pakistan
Philippines
El Salvador
Brazil
Switzerland
Korea
Figure 5. Ratio of OTR over FDI volatilities
Indonesia
Colombia
Sweden
Finland
Portugal
Argentina
Cambodia
Estonia
Mozambique
Ecuador
Venezuela
Malaysia
Hong Kong
Mexico
Uruguay
Greece
Japan
Sudan
United Kingdom
0
0.25
0.5
0.75
1
Sudan
Japan
Korea
Finland
Austria
Germany
Ireland
Italy
Belgium
Denmark
Switzerland
Turkey
United Kingdom
South Africa
Median developing: 0.45
Median industrial: 0.15
France
Thailand
Sweden
Portugal
Philippines
Israel
United States
Pakistan
Netherlands
India
Russia
Brazil
Greece
Panama
Lithuania
Latvia
Jordan
Guatemala
Spain
Argentina
Hong Kong
Australia
Canada
Indonesia
Singapore
Ecuador
Romania
New Zealand
Bangladesh
Uruguay
Paraguay
El Salvador
Czech Rep.
Estonia
Figure 6. Median share of gross FDI inflows in total gross capital inflows
Cape Verde
Chile
Venezuela
Mexico
Costa Rica
Hungary
Cambodia
Malaysia
Mozambique
Georgia
Colombia
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
El Salvador
Chile
Lithuania
South Africa
Czech Rep.
Argentina
Cambodia
Japan
Romania
Hungary
Jordan
Turkey
Estonia
Sudan
Median industrial: 0.14
Median developing: -0.02
Korea
India
Guatemala
Israel
Philippines
Georgia
Greece
Portugal
Germany
Mexico
Canada
Brazil
New Zealand
Mozambique
Paraguay
Denmark
Thailand
Cape Verde
Finland
Australia
Belgium
United Kingdom
Bangladesh
Costa Rica
Pakistan
Austria
Ireland
Italy
Russia
Spain
Figure 7. Correlation between OTR and FDI gross inflows
Uruguay
Sweden
United States
Venezuela
France
Indonesia
Panama
Netherlands
Ecuador
Hong Kong
Singapore
Colombia
Switzerland
Malaysia
Latvia
Table 1. Basic regression results.
Dependent variable is output volatility
(1) (2) (3) (4) (5) (6)
Ïƒ(FDI) 0.18*** 0.16***
[3.4] [3.5]
Ïƒ(OTR) 0.01*** 0.01***
[3.8] [3.9]
Ï?(FDI, OTR) 0.24* 0.23*
[1.7] [1.8]
FDI share -0.07 0.03 0.03
[-0.9] [0.9] [0.9]
FDI share Ã— low share dummy -0.56** -0.53**
[-2.2] [-2.1]
RÂ² 0.12 0.10 0.11 0.11 0.15 0.18
Observations 295 295 295 295 295 295
Countries 59 59 59 59 59 59
Note: Regressions include country and five-year fixed effects. t-statistics are reported in brackets. Standard errors are robust and allow for within-
country correlation (i.e., clustered by country). RÂ² in all regressions corresponds to within-country RÂ². Constant and low share dummy coefficients
are not reported. Ã—, *, **, and *** indicate statistically significance at the 15%, 10%, 5%, and 1% levels, respectively.
Table 2. Regression results with control variables.
Dependent variable is output volatility
(1) (2) (3) (4) (5)
Ïƒ(government spending) 0.04** 0.02 0.03
[2.6] [0.6] [0.8]
Ïƒ(terms of trade) 0.02 0.001 0.02
[1.0] [0.03] [0.7]
Country instability 0.02** 0.03** 0.01
[2.1] [2.5] [0.9]
Ïƒ(FDI) 0.17***
[4.0]
Ïƒ(OTR) 0.01***
[2.9]
Ï?(FDI, OTR) 0.24*
[1.9]
FDI share 0.01
[0.4]
FDI share Ã— low share dummy -0.52*
[-2.0]
RÂ² 0.07 0.04 0.10 0.11 0.26
Observations 376 388 321 279 225
Countries 49 49 56 47 47
Note: Regressions include country and five-year fixed effects. t-statistics are reported in brackets. Standard errors are robust and
allow for within-country correlation (i.e., clustered by country). RÂ² in all regressions corresponds to within-country RÂ². Constant and
low share dummy coefficients are not reported. Ã—, *, **, and *** indicate statistically significance at the 15%, 10%, 5%, and 1%
levels, respectively.
Table 3. Instrumental variable regression results with control variables.
Dependent variable is output volatility
(1) (2) (3) (4) (5)
Ïƒ(FDI) 0.21*** 0.21*** 0.18*** 0.18*** 0.15**
[4.0] [3.5] [3.3] [2.7] [2.2]
Ïƒ(OTR) 0.02 0.02 0.02 0.02 0.01
[1.0] [1.0] [1.3] [0.9] [1.0]
Ï?(FDI, OTR) 0.64** 0.61* 0.47* 0.58* 0.46Ã—
[2.0] [1.9] [1.7] [1.8] [1.6]
FDI share -0.02 -0.02 -0.08 -0.03 -0.04
[-0.3] [-0.2] [-1.1] [-0.3] [-0.5]
FDI share Ã— low share dummy -1.18** -1.21** -0.94* -1.21** -1.14**
[-2.4] [-2.2] [-1.8] [-2.4] [-2.0]
Ïƒ(government spending) 0.01 0.01
[0.5] [0.7]
Ïƒ(terms of trade) 0.07 0.04
[1.2] [0.8]
Country instability 0.02*** 0.02***
[2.2] [2.4]
Overidentification test 15.2* 14.7* 14.5* 14.9* 13.3
Weak identification tests
Ïƒ(FDI) 30.1*** 30.4*** 53.0*** 25.1*** 48.9***
Ïƒ(OTR) 1.7Ã— 1.6Ã— 1.5 1.5 1.4
Ï?(FDI, OTR) 7.3*** 7.6*** 7.1*** 7.3*** 7.6***
FDI share 11.1*** 10.4*** 10.5*** 9.6*** 8.6***
FDI share Ã— low share dummy 2.2** 1.8* 2.2** 1.9* 1.5
Observations 171 168 171 171 168
Countries 38 38 38 38 38
Note: Regressions include country and five-year fixed effects. t-statistics are reported in brackets. Standard errors are robust and
allow for within-country correlation (i.e., clustered by country). RÂ² in all regressions corresponds to within-country RÂ². Constant
and low share dummy coefficients are not reported. The over-identification test is the Chi squared Hansen's J statistic; the null
hypothesis is that the instruments are exogenous (i.e., uncorrelated with the error term). The weak-identification test is the first-
stage Angrist-Pischke multivariate F test of excluded instruments; the null hypothesis is that the model is weakly identified (i.e.,
the excluded instruments have a nonzero but small correlation with the endogenous regressors). Ã—, *, **, and *** indicate
statistically significance at the 15%, 10%, 5%, and 1% levels, respectively.