Does financial globalization promote risk sharing?

Abstract

In theory, one of the main benefits of financial globalization is that it should allow for more efficient international risk sharing. In this paper, we provide an empirical evaluation of the patterns of risk sharing among different groups of countries and examine how international financial integration has affected the evolution of these patterns. Using a variety of empirical techniques, we conclude that there is at best a modest degree of international risk sharing, and certainly nowhere near the levels predicted by theory. In addition, only industrial countries have attained better risk sharing outcomes during the recent period of globalization. Developing countries have, by and large, been shut out of this benefit. Even emerging market economies, many of which have reduced capital controls and all of which have witnessed large increases in cross-border capital flows, have seen little change in their ability to share risk. We find that the composition of flows may help explain why emerging markets have not been able to realize this presumed benefit of financial globalization. In particular, our results suggest that portfolio debt, which had dominated the external liability stocks of most emerging markets until recently, is not conducive to risk sharing.

Full text

Page 1

Does financial globalization promote risk sharing?☆
M. Ayhan Kosea,⁎, Eswar S. Prasadb, Marco E. Terronesa
aResearch Department, International Monetary Fund, United States
bCornell University and Brookings Institution, United States
a b s t r a c ta r t i c l ei n f o
Article history:
Received 5 June 2007
Received in revised form 29 August 2008
Accepted 2 September 2008
Available online xxxx
JEL classification:
F02
F21
F36
F4
Keywords:
Financial globalization
Consumption risk sharing
Emerging markets
In theory, one of the main benefits of financial globalization is that it should allow for more efficient
international risk sharing. In this paper, we provide an empirical evaluation of the patterns of risk sharing
among different groups of countries and examine how international financial integration has affected the
evolution of these patterns. Using a variety of empirical techniques, we conclude that there is at best a
modest degree of international risk sharing, and certainly nowhere near the levels predicted by theory. In
addition, only industrial countries have attained better risk sharing outcomes during the recent period of
globalization. Developing countries have, by and large, been shut out of this benefit. Even emerging market
economies, many of which have reduced capital controls and all of which have witnessed large increases in
cross-border capital flows, have seen little change in their ability to share risk. We find that the composition
of flows may help explain why emerging markets have not been able to realize this presumed benefit of
financial globalization. In particular, our results suggest that portfolio debt, which had dominated the
external liability stocks of most emerging markets until recently, is not conducive to risk sharing.
© 2008 Elsevier B.V. All rights reserved.
1. Introduction
In theory, one of the main benefits of financial globalization is that
it provides better opportunities for countries to smooth consumption
growth in the face of country-specific fluctuations in output growth.
With well-developed domestic financial markets, economic agents
within a country can share risk amongst themselves. However,
insuring against country-wide shocks requires openness to financial
flows that would allowagents in different countries to pool their risks.
Thus, financial globalization should generate welfare gains by
reducing thevolatilityof aggregateconsumptionandalsobydelinking
fluctuations in national consumption and output.
There is a substantial literature examining patterns of risk sharing
among advanced industrial economies (some of the notable contribu-
tions include Obstfeld, 1994, 1995; Lewis, 1996, 1999; Sørensen and
Yosha,1998). The main conclusion of this literature is that the degree of
risksharingisratherlimitedevenamongadvancedindustrialeconomies,
leaving a considerable amount of potential welfare gains unexploited.
Recent work examining the evolution of risk sharing among these
economies presents conflicting results. While some studies suggest that
it has increased during the recent period of globalization (e.g., Sørensen
et al., 2007; Artis and Hoffman, 2006a,b; Giannone and Reichlin, 2006),
others have found little evidence of better risk sharing among industrial
economies (see Moser et al., 2004; Bai and Zhang, 2005).
In contrast, the literature on risk sharing patterns for non-industrial
economiesisrelativelysparse.Obstfeld(1994)andLewis(1996,1997)do
include some of these countries in their analysis, but their samples
(which end in 1988 and 1992, respectively) do not cover much of the
recent wave of financial globalization that enveloped the emerging
marketeconomiesstartinginthemid-1980s.Giventherelativelyhigher
volatility of consumption fluctuations in these economies, and the
higher potential welfare gains of stabilizing these fluctuations, under-
standing these economies' risk sharing patterns is of considerable
interest.1
The objective of this paper is to study the impact of financial
globalization on the degree of international consumption risk sharing
for a large set of industrial and developing countries. In particular, we
makethreecontributionstotheempiricalliteratureoninternationalrisk
Journal of Development Economics xxx (2008) xxx–xxx
☆ Earlier versions of this paper were presented at the 2006 IMF Annual Research
Conference, the January 2007 AEA meetings, the 2007 IMF-Cornell conference on “New
Perspectives on Financial Globalization,” workshops at the Bank of England, Inter-
American Development Bank and the ECB-Bundesbank Joint Seminar Series. We would
like to thank the editors, Gordon Hanson and Enrique Mendoza, and two anonymous
referees for helpful comments that significantly improved the paper. We are grateful to
our discussants, Jonathan Heathcote and Bent Sørensen, for their helpful suggestions.
We also thank Karen Lewis, Fabrizio Perri and seminar participants for helpful
comments. Dionysios Kaltis and Yusuke Tateno provided able research assistance. The
views expressed in this paper are those of the authors and do not necessarily reflect the
views of the IMF or IMF policy.
⁎ Corresponding author.
E-mail addresses: akose@imf.org (M.A. Kose), eswar.prasad@cornell.edu
(E.S. Prasad), mterrones@imf.org (M.E. Terrones).
1Quantitative estimates suggest that the potential welfare gains for developing
countries can be very large (Prasad et al., 2003; Imbs and Mauro, 2007).
DEVEC-01408; No of Pages 13
0304-3878/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jdeveco.2008.09.001
Contents lists available at ScienceDirect
Journal of Development Economics
journal homepage: www.elsevier.com/locate/econbase
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 2

sharing. First, we extend the analysis to a large group of emerging
markets and other developing economies, and investigate the extent of
risk sharing by these economies in a unified framework. Second, we
examinechanges over timeinthe degree of risksharingacrossdifferent
groupsof countries and attempt to relatethosetochanges in thedegree
of financial openness. Third, we provide a careful evaluation of
alternative measures of risk sharing, drawn from different empirical
approaches. In principle, many of these approaches are equivalent, but
there are subtle differences that affect the results. Thus, our compre-
hensiveevaluationofrisksharingpatternsbasedonarangeofmeasures
provides a benchmark set of results that should be useful for further
theoretical and empirical work in this area. Our analysis focuses on
different de jure measures of financial integration; these measures
capture the restrictions a country imposes on cross-border capital
account transactions. We also examine the effects of changes in the
magnitude and composition of actual financial flows.
Our main conclusion is that, notwithstanding the prediction of
conventional theoretical models that financial globalization should
foster increased international risk sharing, there is no evidence that
this is true for developing countries. Even for the group of emerging
market economies – which have opened up their capital accounts and
becomefar moreintegrated intoglobalmarkets thanotherdeveloping
countries – financial globalization has not helped improve the degree
of risk sharing. In contrast, for industrial economies, there is some
evidence that risk sharing has improved in the last decade and a half, a
period during which there was a substantial increase in the volume of
cross-border financial flows relative to the previous two decades.
Why are non-industrial countries unable to share risk more
efficiently despite their increasing integration into global financial
markets? One possibility is that these countries rely more on less
stable capital such as bank loans and other forms of debt that may not
allow for efficient risk sharing. Indeed, when we break up stocks of
external assets and liabilities into different categories – FDI, portfolio
equity, portfolio debtetc. – we find that the underlying composition of
capital flows influences the ability of developing countries to share
risk. In particular, external debt appears to hinder the ability of
emerging market economies to share their consumption risk.
In Section 2, we present the main features of ourdataset. In Section
3, we first derive a basic regression equation and then examine how
the degree of risk sharing has changed over time using various
approaches. In Section 4, we extend the regression model to evaluate
the direct impact of financial globalization on the degree of risk
sharing. In Section 5, we examine if the composition of flows could
explain the inability of emerging markets to attain the risk sharing
benefits of financial globalization. We conclude with a brief summary
of our findings in Section 6.
2. Dataset
We examine patterns of international consumption risk sharing
using a large dataset that includes industrial and developing
countries. The basic data are from the Penn World Tables 6.2 (Heston
et al., 2006) and the World Bank's World Development Indicators. Per
capita real GDP, real private consumption, and real public consump-
tion constitute the measures of national output, private consumption
and government consumption, respectively. All data are in constant
(2000) international prices.
We use different measures of de jure capital account openness to
evaluate the impact of financial integration on risk sharing. Our
benchmark de jure measure is the widely used one based on
information from the IMF's Annual Report on Exchange Arrangements
and Exchange Restrictions (AREAER). We also use three other de jure
measures to check the sensitivity of our results. The first is based on
the dates of equity market liberalizations (Bekaert et al., 2005). Both
theIMFandBHLmeasures arebinary(zeroorone)variables.The other
two de jure measures are taken from the work of Chinn and Ito (2006)
Fig. 1. Risk sharing-cross-section regressions.
2
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 3

and Edwards (2005), respectively. These are continuous measures but
are also largely derived from the IMF's AREAER publications.2We also
examine the robustness of our results to using measures of de facto
financial openness – gross stocks of external assets and liabilities as
ratios to GDP – taken from the External Wealth of Nations Database
(Lane and Milesi-Ferretti, 2006). These measures capture the out-
comes of financial globalization rather than exogenous changes in the
regime of capital controls. However, since the de jure measures are
also imperfect indicators of the true degree of capital mobility, it is
worth showing that our results do not hinge on the measure of
financial integration.
Our dataset comprises annual data over the period 1960–2004 for
69 countries (see the Appendix for a list of countries). The size and
country composition of the sample are dictated by data availability,
with historical data on the measures of financial integration being a
key constraint. The sample consists of 21 industrial and 48 developing
countries. We divide developing countries into two coarse groups —
21 emerging market economies (EMEs) and 27 other developing
countries.
For analyzing the effects of globalization on risk sharing, it is
important to consider the modern era of globalization (1987–2004).
There have been dramatic increases in the volumes of cross-border
trade and financial flows during this period. In particular, private
capital flows from industrialized economies to developing economies
have increased dramatically since the mid-1980s, with most of these
flows going to emerging market economies (see Kose et al., 2006a,b).
This increase in trade and financial flows has been fueled bya series of
trade and financial liberalization programs undertaken by these
economies since the mid-1980s. Roughly 30% of the countries in our
sample had liberalized their trade regimes in 1986; by 2004, this share
had risen to almost 85%. The share of countries with de jure open
financial accounts rose from 20% to about 55% over this period. In light
of these facts, we separately report results pertaining to the
globalization period, 1987–2004, in addition to the full sample.
3. Evolution of consumption risk sharing
Conventional theoretical models in open economy macroeco-
nomics and international finance yield clear predictions about the
impact of financial integration on risk sharing. These predictions are
mostly based on the dynamics of correlations between domestic
consumption and output or between domestic consumption and
world output/consumption. In Kose et al. (2007), we review these
predictions in detail and examine whether they are supported by
empirical evidence. For example, theoretical models with complete
markets predict that the correlation of a country's consumption
growth with the growth of world output (or, equivalently, world
consumption) should be higher than its correlation with that
country's output growth. However, we find that for most countries
the correlation between domestic consumption and output is higher
than the correlation between domestic consumption and world
output. The gap between the two measures is much larger for
emerging markets and other developing countries than for industrial
countries.
Industrial countries in general appear to have higher correlations
of consumption and output with the corresponding world aggregates;
these correlations are typically much lower for developing countries.
A particularly interesting result is that, for emerging market
economies, these correlations with world aggregates have,if anything,
declined slightly during the globalization period. This seems at odds
with the notion that financial integration should have helped these
2The Edwards measure includes additional country-specific information and also
draws upon information from other indexes of financial liberalization. Cross-
correlations across these four de jure measures are very high (see Schindler,
forthcoming).
Fig. 2. Risk sharing-time-series regressions.
3
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 4

economies, which have removed controls on international asset trade
and received the bulk of capital flows to developing countries, to
better share risk with the rest of the world.
Although the unconditional correlations presented in Kose et al.
(2007) are useful in obtaining a preliminary assessment of the
relevance of certain theoretical predictions about risk sharing, they
haveobvious limitations. Wenowturnto a more formalanalysisof the
roles played by factors such as common shocks and the increase in
trade andfinancial linkages in explaining the extentof comovementof
macroeconomic aggregates. In particular, we undertake a more
rigorous test of the risk sharing implications of models with complete
markets. In these types of models, the growth rates of discounted
marginal utility between the periods t and (t+1) are equal across
countries as dictated by the first order conditions with respect to
consumption:
U0 citþ1
U0 cit
where U′ denotes the derivative of the temporal utility function with
respect to (per capita) consumption (c) in country i or j. λ is the
respective Lagrange multiplier or the shadow price of consumption.3
This equation implies that the difference between the growth of
marginal utility of two countries (i and j) should not depend on any
country-specific variables. A number of studies in the literature use
this equationtoarrive at a basic risk sharingregression. Assumingthat
ðÞ
ðÞ
¼U0 cjtþ1
U0 cjt
??
?? ¼λtþ1
λt
ð1Þ
the functional form of the utility function is isoelastic, one can rewrite
this equation as:
E Δlogcit−ΔlogCtjZit
where Z represents a vector of factors specific to country i and C is the
world (per capita) consumption. This yields the following regression
specification:
ð Þ ¼ 0
ð2Þ
Δlogcit−ΔlogCt¼ bZitþ ɛit
If there is perfect risk sharing, the difference between the
consumption growth rates on the left hand side should be equal to
zero, implying that the regression should yield a zero coefficient.
Building on this implication of the complete markets model yields
our basic risk sharing equation, which is similar to others in this
literature4:
ð3Þ
Δlogcit−ΔlogCt¼ constant þ βtΔlogyit−ΔlogYt
where cit(yit) denotes per capita consumption (GDP) of country i in
year t, Ct(Yt) is world per capita consumption (GDP). Growth rates of
Ctand Ytare, respectively, measures of aggregate (common) fluctua-
tions in consumption and output. Since it is not possible to share the
risk associated with common fluctuations, the common component of
each variable is subtracted from the corresponding national variable.
The difference between the national and common world component
ðÞ þ ɛit
ð4Þ
3This equation also has implications for cross-country correlations of consumption.
The theoretical predictions mentioned earlier – that cross-country correlations of
consumption should be equal to unity (or be very high) and cross-country correlations
of consumption should be much higher than those of output – are derived from
models utilizing similar first order conditions.
4For extended discussions of the derivation of this equation, see Obstfeld and Rogoff
(2004, Chapter 5), Asdrubali et al. (1996), Sørensen and Yosha (1998) and Artis and
Hoffman (2006a).
Fig. 3. Risk sharing-panel regressions.
4
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 5

Fig. 4. A. Variance of output growth. B. Variance of residuals.
5
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 6

of each variable captures the idiosyncratic (country-specific) fluctua-
tions in that variable (see Sørensen et al., 2007). The error term, εit, is
assumed to follow a stationary process and captures errors in
measuring consumption (see Obstfeld, 1994).
In a model with complete international financial markets and
perfect risk sharing, the coefficient βt, which captures the average
degree of synchronization between countries' idiosyncratic consump-
tion growth and GDP growth at time t, is equal to zero. Asdrubali et al.
(1996) argue that this coefficient can be used to measure the degree of
risk sharing. The smaller the extent of idiosyncratic comovement, βt,
the greater the degree of international risk sharing.
We analyze how the extent of international consumption risk
sharing has evolved over time using three different approaches in
order to fully exploit the cross-section and time-series dimensions of
the data. Our first approach follows that of Sørensen et al. (2007) and
involves year-by-year estimation of a cross-section regression of the
country-specific component of consumption growth (measured as a
deviation of domestic consumption growth from world consumption
growth) on the country-specific component of output growth. The
second approach is similar to the first one but, relying on the idea
advanced by Obstfeld (1995), involves running the same regression
equationforeachcountryovera giventimeperiod.Thethird approach
is a combination of the first two as it involves estimation of the same
underlying model in a panel framework.5
3.1. Cross-section regressions (year-by-year)
We estimate the basic risk sharing regression above for each year
over the period 1960–2004 and trace the evolution of estimates of
(1−βt) in order to evaluate the changes in the extent of risk sharing
over time. This variable should range from 0 (no risk sharing) to 1
(perfect risk sharing). Since the estimates of (1−βt) fluctuate from
year to year, we smooth them by computing their average over a 9-
year rolling window.
5We also experimented with regressions using levels rather than the growth rates
of consumption and output (Artis and Hoffman, 2006b). The general message about
the extent of risk sharing and its evolution was no different from that of the other
approaches.
Table 1a
Risk sharing and financial integration (de facto and de jure measures, full period, 1960–2004)
De jureDe factoDe jure — IMFDe jure — Edwards
IMFBHL Chinn–ItoEdwards AssetsLiabilities AssetsLiabilities AssetsLiabilities
I. All countries
Output 0.679⁎⁎⁎
[0.033]
0.676⁎⁎⁎
[0.040]
−0.014
[0.074]
0.656⁎⁎⁎
[0.033]
0.160⁎⁎⁎
[0.053]
0.661⁎⁎⁎
[0.038]
−0.014
[0.026]
0.591⁎⁎⁎
[0.087]
0.151
[0.158]
0.709⁎⁎⁎
[0.038]
0.760⁎⁎⁎
[0.043]
0.709⁎⁎⁎
[0.040]
−0.040
[0.089]
0.084
[0.076]
0.528
2228
0.757⁎⁎⁎
[0.044]
−0.014
[0.078]
−0.038
[0.048]
0.530
2228
0.672⁎⁎⁎
[0.084]
0.062
[0.165]
0.036
[0.052]
0.529
1984
0.702⁎⁎⁎
[0.084]
0.105
[0.152]
−0.043
[0.054]
0.537
1987
Output×interactiona
Output×interactionb
0.057
[0.063]
0.528
2261
−0.043
[0.045]
0.533
2262
R2-adjusted
N
0.473
2886
0.471
2493
0.477
2618
0.471
2226
0.478
2020
II. Industrial countries
Output0.652⁎⁎⁎
[0.027]
0.638⁎⁎⁎
[0.043]
0.026
[0.047]
0.633⁎⁎⁎
[0.032]
0.043
[0.041]
0.616⁎⁎⁎
[0.052]
0.036
[0.025]
0.564⁎⁎⁎
[0.125]
0.106
[0.147]
0.651⁎⁎⁎
[0.038]
0.646⁎⁎⁎
[0.040]
0.638⁎⁎⁎
[0.048]
0.058
[0.042]
−0.004
[0.015]
0.618
673
0.639⁎⁎⁎
[0.049]
0.059
[0.042]
−0.005
[0.017]
0.618
673
0.558⁎⁎⁎
[0.091]
0.133
[0.102]
−0.001
[0.014]
0.636
613
0.569⁎⁎⁎
[0.094]
0.101
[0.113]
0.014
[0.015]
0.636
613
Output×interactiona
Output×interactionb
−0.001
[0.016]
0.617
697
0.004
[0.019]
0.617
697
R2-adjusted
N
0.620
882
0.607
758
0.632
798
0.595
673
0.608
624
III. Developing countries
Output 0.693⁎⁎⁎
[0.041]
0.681⁎⁎⁎
[0.046]
−0.005
[0.102]
0.664⁎⁎⁎
[0.040]
0.214⁎⁎⁎
[0.068]
0.654⁎⁎⁎
[0.052]
−0.026
[0.040]
0.597⁎⁎⁎
[0.098]
0.120
[0.202]
0.590⁎⁎⁎
[0.055]
0.775⁎⁎⁎
[0.063]
0.586⁎⁎⁎
[0.054]
−0.105
[0.080]
0.617⁎⁎⁎
[0.129]
0.528
1555
0.772⁎⁎⁎
[0.065]
−0.048
[0.113]
−0.024
[0.076]
0.523
1557
0.583⁎⁎⁎
[0.104]
−0.008
[0.173]
0.582⁎⁎⁎
[0.164]
0.525
1371
0.745⁎⁎⁎
[0.112]
0.055
[0.192]
−0.048
[0.077]
0.524
1373
Output×interaction
Output×interactionb
0.551⁎⁎⁎
[0.134]
0.527
1565
−0.034
[0.074]
0.525
1568
R2-adjusted
N
0.463
2004
0.455
1739
0.464
1817
0.453
1557
0.452
1397
IV. Emerging market economiesc
Output0.826⁎⁎⁎
[0.055]
0.836⁎⁎⁎
[0.057]
−0.118
[0.108]
0.792⁎⁎⁎
[0.062]
0.117
[0.077]
0.775⁎⁎⁎
[0.058]
−0.079⁎⁎
[0.028]
0.838⁎⁎⁎
[0.093]
−0.069
[0.192]
0.727⁎⁎⁎
[0.083]
0.680⁎⁎⁎
[0.076]
0.743⁎⁎⁎
[0.083]
−0.132
[0.090]
0.397⁎⁎
[0.141]
0.688
682
0.697⁎⁎⁎
[0.073]
−0.113
[0.099]
0.214⁎⁎⁎
[0.068]
0.682
680
0.805⁎⁎⁎
[0.114]
−0.213
[0.187]
0.424⁎⁎
[0.185]
0.685
598
0.723⁎⁎⁎
[0.115]
−0.085
[0.165]
0.191⁎⁎
[0.077]
0.678
596
Output×interactiona
Output×interactionb
0.379⁎⁎⁎
[0.132]
0.686
682
0.210⁎⁎⁎
[0.068]
0.680
680
R2-adjusted
N
0.627
874
0.625
766
0.626
791
0.649
695
0.639
616
Note: This table shows the results of panel regressions with yearly data. For details of the regression specification, see Section 4. The standard errors robust to heteroscedasticity and
within-country serial correlation are in brackets. Regressions also include country fixed effects and year dummies. The symbols ⁎, ⁎⁎, and ⁎⁎⁎ indicate statistical significance at the
10%, 5%, and 1% levels, respectively. “Assets” and “Liabilities” refer to gross foreign assets and gross foreign liabilities relative to GDP.
aDe jure measure — either IMF, Bekaert–Harvey–Lundblad, Chinn–Ito or Edwards.
bDe facto measure — either assets or liabilities.
cEmerging Market Economies are a part of the group of Developing Countries.
6
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 7

Fig. 1 shows that, for the full sample, the extent of risk sharing
appears to increase in the globalization period, but it is lower than the
levels observed during the late 1970s. The degree of risk sharing is
often higher among industrial countries than other country groups.
Moreover, it rises modestly for the group of industrial countries
during the late 1990s, but to a levelthat is not much abovethat seen in
the 1970s. There is little evidence that the period of globalization has
seen a marked increase in risk sharing capabilities of emerging
markets and other developing countries.
3.2. Time-series regressions (for each country)
Next, we turn again to Eq. (4) but, rather than estimating it for each
year, we estimate it for each country over rolling nine-year periods
starting from 1960. This is similar to the regressions used by Obstfeld
(1995), who estimates his models for each country for different
periods and analyzes the changes in the relevant coefficients over
time. In his set up, perfect risk sharing implies that, in a regression of
the growth rate of domestic consumption on world consumption
growth and national output growth, the coefficient on world
consumption should be one and that on national output should be
equal to zero. To be consistent with the basic risk sharing regression
above, we focus only on the coefficient associated with consumption.6
After runningthe regressionforeach country, wecompute themedian
of βtover the country sample for each period.
Fig. 2 presents the plots of the extent of consumption risk sharing,
measured by the median of (1−βt), for the full sample and for each
country group based on the time-series regressions. In other words,
the extent of risk sharing in 1969 in each panel refers to the median of
(1−βt) of the respective country group and βtis the regression for
country i over the period 1961–1969. For industrial countries, there
is a steady and substantial increase in the degree of risk sharing
during the globalization period. By contrast, both emerging market
Table 1b
Risk sharing and financial integration (de facto and de jure measures, globalization period, 1987–2004)
De jureDe factoDe jure — IMF De jure — Edwards
IMFBHLChinn–ItoEdwardsAssets LiabilitiesAssetsLiabilitiesAssets Liabilities
I. All countries
Output0.860⁎⁎⁎
[0.046]
0.879⁎⁎⁎
[0.043]
−0.074
[0.099]
0.848⁎⁎⁎
[0.051]
0.008
[0.075]
0.856⁎⁎⁎
[0.048]
−0.028
[0.031]
0.924⁎⁎⁎
[0.076]
−0.143
[0.148]
0.859⁎⁎⁎
[0.048]
0.928⁎⁎⁎
[0.054]
0.871⁎⁎⁎
[0.047]
−0.086
[0.100]
0.027
[0.045]
0.607
1186
0.936⁎⁎⁎
[0.050]
−0.062
[0.107]
−0.080
[0.052]
0.598
1180
0.923⁎⁎⁎
[0.076]
−0.146
[0.153]
0.007
[0.034]
0.620
924
0.980⁎⁎⁎
[0.072]
−0.118
[0.149]
−0.083
[0.057]
0.613
920
Output×interactiona
Output×interactionb
0.003
[0.043]
0.605
1191
−0.090⁎
[0.047]
0.598
1185
R2-adjusted
N
0.606
1193
0.607
1188
0.619
925
0.608
1184
0.620
924
II. Industrial countries
Output 0.661⁎⁎⁎
[0.026]
0.721⁎⁎⁎
[0.073]
−0.064
[0.075]
0.662⁎⁎⁎
[0.030]
…
…
0.713⁎⁎⁎
[0.090]
−0.022
[0.042]
0.964⁎⁎⁎
[0.284]
−0.341
[0.317]
0.704⁎⁎⁎
[0.029]
0.716⁎⁎⁎
[0.031]
0.783⁎⁎⁎
[0.081]
−0.101
[0.073]
−0.028⁎⁎
[0.011]
0.626
354
0.794⁎⁎⁎
[0.081]
−0.099
[0.074]
−0.033⁎⁎⁎
[0.011]
0.626
354
1.088⁎⁎⁎
[0.296]
−0.456
[0.328]
−0.019
[0.012]
0.643
279
1.107⁎⁎⁎
[0.300]
−0.474
[0.334]
−0.018
[0.012]
0.643
279
Output×interactiona
Output×interactionb
−0.037⁎⁎⁎
[0.012]
0.618
359
−0.041⁎⁎⁎
[0.012]
0.617
359
R2-adjusted
N
0.617
357
0.625
352
0.643
277
0.624
348
0.643
277
III. Developing countries
Output0.873⁎⁎⁎
[0.051]
0.878⁎⁎⁎
[0.046]
−0.027
[0.123]
0.836⁎⁎⁎
[0.054]
0.062
[0.082]
0.869⁎⁎⁎
[0.057]
−0.010
[0.039]
0.890⁎⁎⁎
[0.088]
−0.072
[0.192]
0.761⁎⁎⁎
[0.066]
0.979⁎⁎⁎
[0.064]
0.763⁎⁎⁎
[0.065]
−0.087
[0.102]
0.357⁎⁎
[0.137]
0.588
825
0.979⁎⁎⁎
[0.060]
−0.005
[0.127]
−0.119⁎
[0.062]
0.584
824
0.836⁎⁎⁎
[0.108]
−0.104
[0.179]
0.229
[0.185]
0.597
640
1.030⁎⁎⁎
[0.107]
−0.082
[0.186]
−0.152⁎⁎
[0.067]
0.596
639
Output×interactiona
Output×interactionb
0.312⁎⁎
[0.132]
0.588
825
−0.120⁎
[0.060]
0.585
824
R2-adjusted
N
0.584
828
0.584
828
0.597
641
0.584
828
0.596
640
IV. Emerging market economiesc
Output0.966⁎⁎⁎
[0.048]
0.984⁎⁎⁎
[0.050]
−0.136
[0.093]
1.008⁎⁎⁎
[0.045]
−0.062
[0.077]
0.938⁎⁎⁎
[0.047]
−0.059⁎⁎
[0.025]
1.028⁎⁎⁎
[0.058]
−0.119
[0.114]
0.890⁎⁎⁎
[0.069]
0.908⁎⁎⁎
[0.094]
0.895⁎⁎⁎
[0.068]
−0.063
[0.116]
0.210
[0.156]
0.804
356
0.921⁎⁎⁎
[0.094]
−0.141
[0.100]
0.071
[0.070]
0.799
357
0.987⁎⁎⁎
[0.086]
−0.188
[0.120]
0.178
[0.216]
0.816
273
0.967⁎⁎⁎
[0.125]
−0.084
[0.130]
0.049
[0.077]
0.812
275
Output×interactiona
Output×interactionb
0.196
[0.143]
0.804
356
0.065
[0.073]
0.797
357
R2-adjusted
N
0.797
357
0.799
357
0.813
276
0.800
357
0.812
275
Note: This table shows the results of panel regressions with yearly data. For details of the regression specification, see Section 4. The standard errors robust to heteroscedasticity and
within-country serial correlation are in brackets. Regressions also include country fixed effects and year dummies. The symbols ⁎, ⁎⁎, and ⁎⁎⁎ indicate statistical significance at the
10%, 5%, and 1% levels, respectively. “Assets” and “Liabilities” refer to gross foreign assets and gross foreign liabilities relative to GDP.
aDe jure measure — either IMF, Bekaert–Harvey–Lundblad, Chinn–Ito or Edwards.
bDe facto measure — either assets or liabilities.
cEmerging Market Economies are a part of the group of Developing Countries.
6When countries have different rates of time preference, this would be reflected in
the constant term in the regression equation, but it has no impact on the main
prediction of the complete markets model that cross-country correlations of
fluctuations in the growth rates of consumption should be equal to one. Since the
regression is estimated separately for each country, differences in average growth rates
would also be picked up by the regression constant (intercept). Risk sharing is a
relevant concept only in the context of short-term fluctuations, not permanent shocks
or differences in trend growth.
7
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 8

Table 2
Risk sharing and financial integration (dissaggregated de facto measures and de jure measure)
A
IMF measureBHL measure
Full period Globalization periodFull periodGlobalization period
FDIEquityDebtFDI+
equity
FDI+
equity
FDIEquityDebt FDI
equity
FDI+
debt
FDIEquity Debt FDI+
equity
FDI+
equity
FDI EquityDebtFDI+
equity
FDI debt
I. Industrial countries
Output0.622⁎⁎⁎
[0.058]
0.031
[0.102]
0.641⁎⁎⁎
[0.049]
−0.032⁎
[0.016]
0.638⁎⁎⁎
[0.049]
−0.002
[0.014]
0.642⁎⁎⁎
[0.049]
−0.013
[0.016]
0.626⁎⁎⁎
[0.047]
−0.074
[0.082]
0.039
[0.051]
0.058
[0.042]
0.618
673
0.770⁎⁎⁎
[0.084]
−0.074
[0.094]
0.776⁎⁎⁎
[0.078]
−0.072⁎⁎⁎
[0.016]
0.765⁎⁎⁎
[0.082]
−0.017
[0.011]
0.785⁎⁎⁎
[0.079]
−0.047⁎⁎⁎
[0.014]
0.779⁎⁎⁎
[0.076]
−0.174⁎
[0.099]
0.070
[0.059]
−0.104
[0.071]
0.631
354
0.592⁎⁎⁎
[0.052]
0.058
[0.109]
0.612⁎⁎⁎
[0.039]
−0.026
[0.038]
0.608⁎⁎⁎
[0.042]
0.004
[0.014]
0.612⁎⁎⁎
[0.040]
−0.011
[0.027]
0.601⁎⁎⁎
[0.043]
−0.052
[0.090]
0.031
[0.050]
0.078⁎⁎⁎
[0.027]
0.637
613
0.698⁎⁎⁎
[0.042]
−0.090
[0.092]
0.692⁎⁎⁎
[0.033]
−0.074⁎⁎
[0.035]
0.694⁎⁎⁎
[0.035]
−0.019⁎
[0.010]
0.702⁎⁎⁎
[0.035]
−0.051⁎
[0.025]
0.708⁎⁎⁎
[0.039]
−0.191⁎
[0.108]
0.063
[0.057]
…
…
0.644
279
Output×
interactiona
Output×debt
Output×de
jure measure
R2-adjusted
N
0.058
[0.047]
0.613
674
0.048
[0.044]
0.612
675
0.057
[0.043]
0.618
673
0.044
[0.042]
0.612
675
−0.087
[0.073]
0.625
353
−0.096
[0.075]
0.628
354
−0.083
[0.076]
0.626
353
−0.098
[0.073]
0.627
354
0.080⁎⁎
[0.033]
0.632
614
0.077⁎⁎
[0.035]
0.631
615
0.078⁎⁎
[0.027]
0.637
613
0.075⁎⁎
[0.035]
0.631
615
…
…
0.642
278
…
…
0.642
279
…
…
0.642
278
…
…
0.642
279
II. Emerging market economies
Output 0.727⁎⁎⁎
[0.062]
0.813⁎⁎⁎
[0.185]
0.840⁎⁎⁎
[0.057]
1.086
[0.873]
0.726⁎⁎⁎
[0.070]
0.169⁎⁎⁎
[0.052]
0.746⁎⁎⁎
[0.060]
0.630⁎⁎⁎
[0.162]
0.661⁎⁎⁎
[0.076]
0.380⁎
[0.186]
0.164⁎⁎⁎
[0.046]
−0.137
[0.101]
0.678
681
0.927⁎⁎⁎
[0.069]
0.180
[0.279]
1.000⁎⁎⁎
[0.053]
−0.433
[0.677]
0.912⁎⁎⁎
[0.085]
0.074
[0.052]
0.961⁎⁎⁎
[0.064]
0.022
[0.227]
0.823⁎⁎⁎
[0.081]
0.032
[0.203]
0.135⁎⁎⁎
[0.040]
−0.048
[0.121]
0.799
356
0.693⁎⁎⁎
[0.067]
0.861⁎⁎⁎
[0.234]
0.800⁎⁎⁎
[0.066]
0.576
[1.147]
0.694⁎⁎⁎
[0.079]
0.140⁎⁎⁎
[0.047]
0.706⁎⁎⁎
[0.069]
0.551⁎⁎
[0.224]
0.639⁎⁎⁎
[0.094]
0.315
[0.313]
0.151⁎⁎⁎
[0.050]
0.077
[0.074]
0.673
598
0.936⁎⁎⁎
[0.082]
0.243
[0.372]
1.006⁎⁎⁎
[0.050]
0.157
[0.997]
0.954⁎⁎⁎
[0.088]
0.044
[0.054]
0.988⁎⁎⁎
[0.067]
0.069
[0.329]
0.906⁎⁎⁎
[0.116]
0.069
[0.335]
0.067
[0.058]
− 0.056
[0.081]
0.802
275
Output×
interactiona
Output×debt
Output×de
jure measure
R2-adjusted
N
−0.074
[0.090]
0.677
678
−0.124
[0.100]
0.684
681
−0.115
[0.100]
0.683
682
−0.144
[0.106]
0.676
681
−0.049
[0.117]
0.778
356
−0.118
[0.099]
0.802
356
−0.131
[0.099]
0.799
356
−0.039
[0.119]
0.794
356
0.069
[0.078]
0.677
595
0.099
[0.086]
0.683
598
0.113
[0.069]
0.681
599
0.105
[0.083]
0.674
598
−0.058
[0.087]
0.786
274
−0.065
[0.084]
0.817
275
−0.044
[0.081]
0.812
276
−0.083
[0.079]
0.801
275
8
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 9

B
Chinn–Ito measureEdwards measure
Full periodGlobalization periodFull period Globalization period
FDIEquity DebtFDI+
equity
FDI+
equity
FDIEquity DebtFDI+
equity
FDI+
debt
FDIEquityDebtFDI+
equity
FDI+
equity
FDIEquityDebt FDI+
equity
FDI+
equity
I. Industrial countries
Output0.619⁎⁎⁎
[0.057]
0.010
[0.099]
0.636⁎⁎⁎
[0.049]
−0.039⁎
[0.020]
0.634⁎⁎⁎
[0.048]
− 0.005
[0.013]
0.637⁎⁎⁎
[0.049]
−0.019
[0.015]
0.618⁎⁎⁎
[0.049]
− 0.088
[0.085]
0.043
[0.052]
0.093⁎
[0.052]
0.619
662
0.707⁎⁎⁎
[0.037]
−0.089
[0.096]
0.703⁎⁎⁎
[0.026]
− 0.081⁎⁎⁎
[0.018]
0.704⁎⁎⁎
[0.031]
− 0.020⁎
[0.011]
0.712⁎⁎⁎
[0.028]
−0.054⁎⁎⁎
[0.016]
0.704⁎⁎⁎
[0.040]
− 0.173⁎
[0.097]
0.065
[0.057]
…
…
0.629
350
0.592⁎⁎⁎
[0.057]
0.065
[0.093]
0.617⁎⁎⁎
[0.046]
− 0.020
[0.036]
0.608⁎⁎⁎
[0.047]
0.005
[0.011]
0.617⁎⁎⁎
[0.047]
−0.007
[0.024]
0.600⁎⁎⁎
[0.049]
− 0.055
[0.092]
0.034
[0.049]
0.067
[0.040]
0.636
613
0.725⁎⁎⁎
[0.046]
− 0.022
[0.090]
0.729⁎⁎⁎
[0.037]
−0.042
[0.036]
0.723⁎⁎⁎
[0.039]
−0.004
[0.012]
0.734⁎⁎⁎
[0.039]
−0.027
[0.026]
0.729⁎⁎⁎
[0.043]
− 0.181⁎
[0.103]
0.070
[0.055]
− 0.095
[0.067]
0.645
279
Output×
interactiona
Output×debt
Output×de
jure measure
R2-adjusted
N
0.085
[0.055]
0.613
663
0.073
[0.059]
0.612
664
0.084
[0.053]
0.618
662
0.069
[0.056]
0.612
664
…
…
0.624
349
…
…
0.627
350
…
…
0.625
349
…
…
0.626
350
0.067
[0.044]
0.631
614
0.057
[0.050]
0.630
615
0.065
[0.039]
0.636
613
0.055
[0.048]
0.630
615
− 0.127⁎
[0.066]
0.645
278
−0.120
[0.070]
0.645
279
−0.125⁎
[0.072]
0.645
278
−0.124⁎
[0.070]
0.644
279
II. Emerging market economies
Output 0.739⁎⁎⁎
[0.069]
0.813⁎⁎⁎
[0.181]
0.894⁎⁎⁎
[0.072]
1.143
[0.909]
0.773⁎⁎⁎
[0.078]
0.166⁎⁎⁎
[0.047]
0.762⁎⁎⁎
[0.069]
0.602⁎⁎⁎
[0.153]
0.719⁎⁎⁎
[0.080]
0.357⁎
[0.175]
0.165⁎⁎⁎
[0.043]
− 0.106
[0.064]
0.678
677
0.903⁎⁎⁎
[0.124]
0.130
[0.270]
1.019⁎⁎⁎
[0.076]
− 0.470
[0.684]
0.928⁎⁎⁎
[0.105]
0.075
[0.057]
1.003⁎⁎⁎
[0.083]
0.033
[0.231]
0.859⁎⁎⁎
[0.096]
0.031
[0.188]
0.131⁎⁎⁎
[0.043]
−0.049
[0.082]
0.799
356
0.650⁎⁎⁎
[0.094]
0.850⁎⁎⁎
[0.249]
0.779⁎⁎⁎
[0.099]
1.052
[1.112]
0.673⁎⁎⁎
[0.109]
0.140⁎⁎
[0.059]
0.664⁎⁎⁎
[0.093]
0.608⁎⁎
[0.217]
0.623⁎⁎⁎
[0.114]
0.381
[0.296]
0.149⁎⁎
[0.058]
0.038
[0.091]
0.672
597
0.941⁎⁎⁎
[0.093]
0.167
[0.367]
1.001⁎⁎⁎
[0.069]
0.012
[0.887]
0.940⁎⁎⁎
[0.122]
0.050
[0.064]
1.000⁎⁎⁎
[0.074]
−0.006
[0.301]
0.915⁎⁎⁎
[0.134]
0.018
[0.313]
0.066
[0.066]
−0.057
[0.090]
0.801
274
Output×
interactiona
Output×debt
Output×de
jure measure
R2-adjusted
N
−0.029
[0.062]
0.679
673
−0.099
[0.076]
0.685
676
−0.087
[0.073]
0.683
678
−0.044
[0.068]
0.676
676
0.027
[0.125]
0.778
356
−0.040
[0.087]
0.800
356
−0.041
[0.084]
0.798
357
−0.061
[0.087]
0.795
356
0.084
[0.086]
0.677
594
0.051
[0.101]
0.682
597
0.069
[0.092]
0.679
598
0.083
[0.091]
0.673
597
−0.050
[0.094]
0.786
273
−0.049
[0.081]
0.816
274
−0.028
[0.091]
0.812
275
−0.084
[0.083]
0.801
274
Note: Thistable shows the results of panel regressions withyearlydata. Fordetails of the regression specification, see Section 4. The standarderrors robust to heteroscedasticityand within-country serial correlation are in brackets. Regressions
also include country fixed effects and year dummies. The symbols ⁎, ⁎⁎, and ⁎⁎⁎ indicate statistical significance at the 10%, 5%, and 1% levels, respectively.
aThe interaction term is the de facto financial openness measure listed in each column.
9
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 10

economies and the group of developing countries experience a slight
trend decline in the degree of risk sharing during the period of
globalization.
3.3. Panel regressions
Our next approach combines the first two by estimating the
standard regression model in a panel framework. In particular, we run
the same regression based on Eq. (4) but estimate it over nine-year
rolling panels. This allows us to utilize all the time-series and cross-
sectional information available in the data. Fig. 3 presents plots of the
extent of consumption risk sharing based on the panel regressions.
The extent of risk sharing in 1969 is again equal to (1−βt) and βtis the
coefficient from the panel regression covering the period 1961–1969.
The patterns in these figures are broadly consistent with our earlier
results. While industrial countries attain slightly better risk sharing
outcomes during the period of globalization, neither emerging
markets nor developing countries exhibit the same pattern.
Werana batteryof tests tocheck if therearestatisticallysignificant
changes in the risk sharing coefficientsacross countrygroups andover
time. The results indicate that the extent of risk sharing is significantly
higher for industrial countries than for either emerging market or
other developing economies over the past two decades. In addition,
there is a significant increase in the extent of risk sharing from the
period 1985–1994 to 1995–2004 for industrial countries alone.7
Another question related to the types of regressions we employ
here is whether the results are driven by changes in idiosyncratic
country-specific noise.8To explore this possibility, we check the
evolution of the variance of idiosyncratic output fluctuations, Δlogyit−
ΔlogYt, and the variance of the residual, εit, in our panel regressions.
As presented in Fig. 4A and B, for the group of industrial countries,
these variances appear to be quite stable. In contrast, for the other
country groups, the variances are not as stable. These findings lend
further support to our claim that the results about changes in the
degree of risk sharing are significant only in the case of industrialized
countries.9
There are two messages from the results in this section. First,
industrial countries have attained improvements in risk sharing
during the period of globalization. These results are consistent with
thefindingsin Sørensen etal. (2007); ourworkextends theirresults to
a larger set of countries and a longer temporal span of the data. We
also employ a larger menu of regressions to analyze the robustness of
our findings to different approaches. While the extent of risk sharing
among industrial countries does improve during the period of
globalization, we find that this is only a modest improvement relative
to the 1970s.10Second, we find that emerging market economies and
other developing economies have not registered any major changes
during the period of globalization in terms of their ability to share
idiosyncratic income risk.
Could measurement error in the consumption data for emerging
markets be driving our puzzling results for this group of countries? It
is unlikely that the temporal evolution of risk sharing patterns can be
simply explained by measurement error. Indeed, if consumption is
becoming better measured over time, it should be getting mechani-
cally delinked from output, which would naturally drive up estimates
of risk sharing. Another approach to dealing with measurement error
is to look at income smoothing, on the logic that income (measured as
GNP or GNI) is better measured than consumption and reflects risk
sharing via international financial flows. Sørensen et al. (2007) find
that international financial flows have helped industrial countries to
smooth their income since the early 1990s. Ultimately, however, it is
smoothing of consumption rather than income that matters for
welfare.
The temporal patterns we document in this section are suggestive,
but do not directly address the question of whether financial
globalization has played an important role in the evolution of the
degree of risk sharing displayed by different country groups. So we
now turn to a regression model that augments the standard risk
sharing regression with an interaction variable in order to explicitly
capture the effects of financial globalization.
4. Financial globalization and risk sharing
We now use panel regressions to directly examine the impact of
financial globalization on the degree of risk sharing. In particular, we
interact the idiosyncratic component of output with various measures
of financial integration. That is, we estimate the following regression
using panel data:
Δlogcit−ΔlogCt¼ constant þ μtΔlogyit−ΔlogYt
þ γ0tFOitΔlogyit−ΔlogYt
where FOitis the set of measures of the degree of financial openness of
country i. Parallel to the analysis in the previous section, the degree of
risk sharing attained by country i is equal to (1−µt−γt′FOit).11When an
element of the coefficient vector of interaction terms, γ′, is negative, it
indicates that thegreaterthe degree of financial integration associated
with the respective measure of financial openness, the higher the
amount of risk sharing achieved by a country. The panel regressions
we employ include both country fixed effects and time effects.
In Tables 1a and 1b, we report the results for different country
groups and also separately for the full sample (1960–2004) and the
period of globalization (1987–2004). We focus on four de jure
measures of financial openness (IMF, BHL, Chinn–Ito, and Edwards)
andtwode facto ones (gross stocks of assets and liabilities, both scaled
by GDP). We experiment with each of these measures in turn and then
consider various combinations of de jure and de facto measures in our
regressions.
The first column of Table 1a shows the results of panel regressions
without interaction terms associated with financial integration. The
findings are broadly consistent with the results reported in the
previous section. The extent of risk sharing appears to be slightly
higherin industrial countries than in developing countries;it is lowest
for the group of emerging market economies.
In the remaining columns of Table 1a, we first report results from
regressions that include interaction terms with each of the de facto
and de jure measures. In the final fourcolumns, wereport results from
regressions that include interactions with one de facto and one de jure
measure at a time. To keep the volume of these results manageable,
we report results using the two de facto measures and only two of the
de jure measures — IMF and Edwards. The results were similar when
we used the other two de jure measures.
For the full sample period, virtually none of the interaction
coefficients is significant with the expected sign for all countries and
the group of industrial economies. The sole exception is the negative
coefficient on the interaction with the Chinn–Ito measure for
emerging markets, but this is not a robust result as it is limited to
this one indicator. In fact, many of the coefficients on the interactions
with the de facto measures are positive for emerging markets —
ðÞ
ðÞ þ ɛit
ð5Þ
7Detailed results from these tests are available from the authors.
8We would like to thank one of the referees for raising this point.
9These results suggest that changes over time in the variances of idiosyncratic
output and the error term can potentially bias our estimates in the case of developing
countries and emerging markets. Since we are making no claim that there are
statistically significant changes in the degree of risk sharing for these groups of
countries, we do not pursue this further.
10Sørensen (2006) argues that there may be more measurement error in the
consumption data in the earlier period, which could explain this result.
11See Sørensen et al. (2007) for a similar model. We also estimated models allowing
for time trends associated with the measures of financial integration, but the trends
were not statistically significant.
10
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 11

implying a deterioration in risk sharing. In short, it is difficult to claim
that financial globalization has had a beneficial effect on the amount
of risk sharing around the world over the period 1960–2004.
The results look slightly more promising when we restrict our
analysis to the period of globalization. The results in Table 1b indicate
that higher levels of gross external assets and liabilities improve risk
sharing for industrial countries and it is only for that group that the
coefficients on the interaction terms with both of the de factofinancial
integration variables are generally significantly negative. The de jure
measures are not associated with significant changes in the extent of
risk sharing during the globalization period.12The statistically
significant negative sign on the interaction with the stock of liabilities
is preserved in regressions involving the full group of developing
economies. However, in the case of emerging markets, financial
integration appears to have no significant impact in the globalization
period.Theseresults areconsistent withthestylizedfacts inKose et al.
(2007) that the risk sharing benefits of financial integration have
accrued mainly to industrial countries over this period.
5. Why is there so little risk sharing?
The results that we have presented thus far suggest that the degree
of international risk sharing is limited. Furthermore, only industrial
countries seem to have attained clear benefits from financial
integration in terms of improved risk sharing. Why is it that even
emerging markets have not benefited much in this dimension, despite
having liberalized their capital accounts and attained much higher
levels of financial integration than other developing economies? In
this section, we investigate the possibility that different types of
capital flows may be more or less conducive to risk sharing, and
emerging markets may just not be getting the “right” types of flows.13
To address the issue of whether specific types of financial flows
(and corresponding stocks) are more conducive to attaining the risk
sharing benefits of globalization, we now consider disaggregated
measures of gross external assets and liabilities relative to GDP. In
particular, we focus on stock measures (assets+liabilities) of foreign
direct investment (FDI), Equity (portfolio equity), Debt (portfolio debt)
and (FDI+Equity). We also examine the effects of (FDI+Equity) and
Debt when we include them simultaneously. In all of our regressions,
we control for de jure measures of financial integration as well. To
make the coefficients comparable across regressions, we convert the
Chinn–Ito and Edwards measures to 0–1 variables. We do this based
on whether a country is below (0) or above (1) the median level of the
respective de jure measure (within its countrygroup) over a particular
period. Our results in the previous section indicated sharp differences
in risk sharing outcomes for industrial countries and emerging
markets. Since these are the two groups of countries that have
attained significant levels of de facto financial integration, we present
results only for these two groups in order to conserve space.
The first panel of Table 2 present the results of panel regressions
with these measures of the structure of external assets and liabilities
and our benchmark IMF de jure measure for the full sample and
globalization periods. The results add to our previous findings by
showing that particular forms of capital flows have not contributed to
increased risk sharing during the full sample period. The only
exception to this finding is equity in the case of industrial countries.
In the case of emerging markets, the interaction coefficients
associated with most forms of external capital are positive, suggesting
that higher levels of de facto financial openness actually led to worse
risk sharing outcomes for these economies. Of course, their levels of
de facto integration were low during the pre-globalization period, so
one should not make too much of these results.
The results for the globalization period are quite different. The
interaction coefficients associated with Equity, (FDI+Equity) and (FDI+
debt) are significantly negative for industrial countries, implying that
higher de facto financial openness improves risk sharing outcomes.
Interestingly, for emerging markets, neither stocks of FDI nor equity
seem to help in sharing risk while debt stocks by themselves reduce
the level of risk sharing during the globalization period.14Note that,
despite these effects on risk sharing identified for particular forms of
external capital stocks, the benchmark de jure measure of financial
openness is not statistically significant in any of the specifications.
The remaining three panels of Table 2 repeat the same set of
regressions with the BHL, Chinn–Ito and Edwards measures of de jure
capital account openness, respectively. These results largely confirm
our baseline findings. Different forms of capital flows do not appear to
have a beneficial effect on the extent of risk sharing during the full
sample period. In the globalization period, virtually all types of flows
improve risk sharing outcomes for industrial countries. The de jure
openness variables are rarely significant, except whenwe consider the
Edwards measure. Based on this measure, it is de jure openness that
has improved the degree of risk sharing attained by industrial
countries. The Edwards measure is the most comprehensive of the
de jure measures and includes country-specific information other
than just formal capital account restrictions as reported to the IMF,
which makes it a closer proxy for overall financial openness than the
other de jure measures.
For emerging markets, the coefficients on the interaction with
external debt stocks are positive, indicating an adverse effect on risk
sharing, in all four regressions and statistically significant when de
jure openness is proxied by the IMF or Chinn–Ito measures. Putting
this result together with the fact that, until recently, debt liabilities
dominated the external liability positions of emerging markets could
explain why these economies have not attained the risk sharing
benefits of financial globalization. However, it still leaves open the
question of why the rising importance of FDI and portfolio equity
stocks has not yet resulted in a marked improvement in risk sharing
achieved by these economies.15
6. Conclusion
In this paper, we examined the implications of increased financial
integration for the patterns of international risk sharing among
different groups of countries using a variety of empirical approaches.
First, we examined the evolution of the extent of risk sharing as
measured by changes in the comovement between idiosyncratic
components of the growth rates of consumption and output. The
12Since most industrial countries undertook equity market liberalizations before the
start of the globalization period, the BHL measure is unchanged over this period and
drops out of the panel regression for this group of countries.
13Other theoretical explanations for the low degree of risk sharing include: the
importance of non- traded goods and the prevalence of large preference shocks; the
dearth of financial instruments for efficiently sharing macroeconomic risk (Heathcote
and Perri, 2002); and large transaction costs associated with international trade of
goods and assets (Obstfeld and Rogoff, 2001). Lewis (1996, 1999) finds that neither
nonseparabilities between consumption and leisure, nor the inclusion of capital
controls, nontradables and/or durable goods, accounts for imperfect risk sharing but
she also reports that when she considers both nonseparabilities and certain types of
controls, risk sharing cannot be rejected. Consistent with our findings, these studies
suggest that, even after controlling for various factors, the risk sharing prediction of
the standard models has often been rejected. For a survey of these explanations, see
Kose et al. (2007).
14These results may be related to the empirical link between exposure to short-term
debt and the likelihood of financial crises (see Rodrik and Velasco, 2000; Berg et al.,
2004).
15Kose et al. (2006) report that the share of debt in gross stocks of foreign assets and
liabilities has declined from 75% in 1980–84 to 59% in 2000–2004 in a group of 71
developed and developing countries. For emerging markets, the share of FDI and
portfolio equity has risen from a total of 13% in 1980–84 to 37% in 2000–2004, while
the share of debt has declined from 78% to 47% over the same period. In the process of
accumulating massive foreign exchange reserves, emerging markets have recently
been buying large quantities of industrial country government bonds; these may not
be conducive to efficient international risk sharing.
11
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 12

results suggest that industrial countries share more of their idiosyn-
cratic consumption risk through their integration with global capital
markets than do emerging market economies. Moreover, industrial
countries have been able to increase the extent of risk sharing in the
globalization period while emerging market economies have not been
able to attain any such benefits. Second, we directly analyzed the
impact of integration with global financial markets on the extent of
risk sharing a country is able to attain. We find only limited evidence
suggesting that financial integration has helped improve risk sharing
outcomes in industrial countries, and no evidence that it has done so
for emerging markets.
These congruent results point to an interesting puzzle. Theory
predicts that financial integration should allow countries to improve
the extent of international risk sharing by diversifying their idiosyn-
cratic income risk. Contrary to theoretical predictions, however,
emergingmarketeconomies – which haveeliminated various controls
on capital account transactions and experienced a significant increase
in international financial flows during the past two decades – have
been unable to enjoy the risk sharing benefits of financial globaliza-
tion. We examined whether the composition of external capital stocks
could explain this puzzle, and found that, in general, FDI and portfolio
equity stocks seem to improve risk sharing outcomes while debt
stocks have the opposite effect. Only the latter result comes through
strongly for emerging markets, however. De jure measures of financial
integration generally have little effect on the degree of risk sharing
during the period of globalization.
Our results suggest three avenues to be explored in future work.
The puzzle we have identified in this paper might be related to a
threshold effect in terms of how financial globalization improves risk
sharing — only countries that are substantially integrated into global
markets (in de facto terms) appear to attain these benefits. Indeed,
Kose et al. (2003) document that the volatility of consumption growth
relative to that of income growth, a crude alternative proxy for risk
sharing, tends to increase at intermediate levels of financial integra-
tion, and then declines at higher levels of integration. This suggests
that, in order to reap the risk sharing benefits of financial globaliza-
tion, emerging markets and other developing countries need to
become more integrated into global financial markets.
Another possibility is that, despite increased financial integration,
there are other characteristics of certain countries – either structural
features or policies – that preclude them from attaining improved risk
sharing through financial integration. Our preliminary results indicate
that this explanation does not work well (see Kose et al., 2007), but
this hypothesis warrants a more careful investigation.
A third reason for the inability of emerging market economies to
attain the presumed risk sharing benefits of financial integration is
that capital flows to emerging market economies tend to be
procyclical — they increase in good times and fall in bad times (see
Kaminsky et al., 2004). The very feature of procyclicality might be
preventing emerging markets from utilizing these flows to smooth
their consumption fluctuations. Since flows to these economies are
shifting away from debt – which tends to be more procyclical – to FDI
and portfolio equity flows – which tend to be more stable – it is
possible that the risk sharing benefits of international financial
integration will become more apparent for emerging markets in the
future.
Appendix. List of countries
The sample comprises 69 countries — 21 industrial and 48
developing.16
Industrial countries
Australia (AUS), Austria (AUT), Belgium (BEL), Canada (CAN),
Denmark (DNK), Finland (FIN), France (FRA), Germany (DEU), Greece
(GRC), Ireland (IRL), Italy (ITA), Japan (JPN), Netherlands (NLD), New
Zealand (NZL), Norway (NOR), Portugal (PRT), Spain (ESP), Sweden
(SWE), Switzerland (CHE), United Kingdom (GBR), and United States
(USA).
Developing countries17
These countries are grouped intoEmerging Markets (21) and Other
Developing Countries (27).
Emerging Market Economies
Argentina (ARG), Brazil (BRA), Chile (CHL), China (CHN), Colombia
(COL), Egypt (EGY), India (IND), Indonesia (IDN), Israel (ISR), Jordan
(JOR), Korea (KOR), Malaysia (MYS), Mexico (MEX), Morocco (MAR),
Pakistan (PAK), Peru (PER), Philippines (PHL), South Africa (ZAF),
Thailand (THA), Turkey (TUR), and Venezuela (VEN).
Other Developing Countries
Algeria (DZA), Bolivia (BOL), Cameron (CMR), Costa Rica (CRI), Cote
d'Ivoire (CIV), Dominican Republic (DOM), Ecuador (ECU), El Salvador
(SLV), Fiji (FJI), Gabon (GAB), Ghana (GHA), Guatemala (GTM), Haiti
(HTI), Honduras (HND), Iran (IRN), Jamaica (JAM), Mauritius (MUS),
Nicaragua (NIC), Papua New Guinea (PNG), Paraguay (PRY), Senegal
(SEN), Sri Lanka (LKA), Togo (TGO), Trinidad and Tobago (TTO), Tunisia
(TUN), Uruguay (URY), and Zimbabwe (ZWE).
References
Artis, Michael J., Hoffmann, Mathias, 2006a. Declining Home Bias and the Increase in
International Risk Sharing: Lessons from European Integration. Working Paper.
University of Manchester, Manchester.
Artis, Michael J., Hoffmann, Mathias, 2006b. The home bias and capital income flows
between countries and regions. CEPR Discussion Paper No. 5691. Centre for
Economic Policy Research, London.
Asdrubali, Pierfederico, Sørensen, Bent E., Yosha, Oved,1996. Channels of interstate risk
sharing: United States 1963–90. Quarterly Journal of Economics 111 (4),1081–1110.
Bai, Yan, Zhang, Jing, 2005. Financial Integration and International Risk Sharing.
University of Michigan. Working Paper.
Bekaert, Geert, Harvey, Campbell R., Lundblad, Christian, 2005. Does financial liberal-
ization spur growth? Journal of Financial Economics 77 (1), 3–55.
Berg, Andrew, Borensztein, Eduardo, Pattillo, Catherine, 2004. Assessing Early Warning
Systems: How Have They Worked in Practice? IMF. Working Paper 04/52.
Chinn, Menzie, Ito, Hiro, 2006. What matters for financial development? Capital
controls, institutions, and interactions. Journal of Development Economics 61 (1),
163–192.
Edwards, Sebastian, 2005. Capital Controls, Sudden Stops and Current Account
Reversals. NBER. Working Paper No. 11170.
Giannone, Domenico, Reichlin, Lucrezia, 2006. Trends and Cycles in the Euro Area: How
Much Heterogeneity and Should We Worry About It? European Central Bank,
London. ECB Working Paper No. 595.
Heathcote, Jonathan, Perri, Fabrizio, 2002. Financial autarky and international business
cycles. Journal of Monetary Economics 49 (3), 601–628.
Heston, Alan, Summers, Robert, Aten, Bettina, 2006. Penn World Tables Version 6.2,
Center for International Comparisons, Income and Prices at the University of
Pennsylvania.
Imbs, Jean, Mauro, Paolo, 2007. Pooling Risk Among Countries. IMF. Working Paper No.
07/132.
Kaminsky, Graciela L., Reinhart, Carmen, Végh, Carlos A., 2004. When it rains, it pours:
procyclical capital flows and policies. In: Gertler, Mark, Kenneth S., Rogoff (Eds.),
NBER Macroeconomics Annual. MIT Press, Cambridge, Mass, pp. 11–53.
Kose, M. Ayhan, Prasad, Eswar S., Terrones, Marco E., 2003. Financial integration and
macroeconomic volatility. IMF Staff Papers 50 (1), 119–142.
Kose, M. Ayhan, Prasad, Eswar S., Rogoff, Kenneth, Wei, Shang-Jin, 2006a. Financial
Globalization: A Reappraisal. IMF. Working Paper 06/189.
16We excluded from the analysis small countries (those with population below
1 million), transition economies, major oil producers, and other countries with
incomplete or clearly unreliable data.
17Hong Kong, Panama, and Singapore were excluded from the analysis because these
countries are financial hubs and are therefore outliers in terms of standard measures of
de facto financial integration.
12
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001

Page 13

Kose, M. Ayhan, Prasad, Eswar S., Terrones, Marco E., 2006b. How do trade and financial
integration affect the relationship between growth and volatility? Journal of
International Economics 69 (1), 176–202.
Kose, M. Ayhan, Prasad, Eswar S., Terrones, Marco E., 2007. How Does Financial
Globalization Affect Risk Sharing? Patterns and Channels. IMF. Working Paper No:
07/238.
Lane, Philip, Milesi-Ferretti, Gian Maria, 2006. The External Wealth of Nations Mark II:
Revised and Extended Estimates of Foreign Assets and Liabilities, 1970–2004. IMF.
Working Paper 06/69.
Lewis, Karen K.,1996. What can explain the apparent lack of international consumption
risksharing? Journal of Political Economy 104 (2), 267–297.
Lewis, Karen K., 1997. Are countries with official international restrictions liquidity
constrained? European Economic Review 41 (6), 1079–1109.
Lewis, Karen K.,1999. Trying to explain home bias in equities and consumption. Journal
of Economic Literature 37, 571–608 June.
Moser, Gabriel, Pointner, Wolfgang, Scharler, Johann, 2004. International risk sharing in
Europe: has anything changed? In: Liebscher, K., Christl, J., Mooslechner, P.,
Ritzberger-Grünwald, D. (Eds.), The Economic Potential of a Larger Europe. Edward
Elgar Publishing, Northampton, pp. 260–275.
Obstfeld, Maurice,1994. Are industrial-country consumption risks globally diversified?
In: Leiderman, L., Razin, A. (Eds.), Capital Mobility: The Impact of Consumption,
Investment and Growth. Cambridge University Press, New York, pp. 13–47.
Obstfeld, Maurice, 1995. International capital mobility in the 1990s. In: Kenen, P.B. (Ed.),
Understanding Interdependence. Princeton University Press, Princeton, pp. 201–261.
Obstfeld, Maurice and Rogoff, Kenneth, 2001, “The Six Major Puzzles in International
MacroeconomicsSolved,”NBERMacroeconomicsAnnual2000,Vol.15,No.1,pp.339–90.
Obstfeld, Maurice, Rogoff, Kenneth, 2004. Foundations of International Macroeco-
nomics. MIT Press.
Prasad, Eswar S., Rogoff, Kenneth, Wei, Shang-Jin, Kose, M. Ayhan, 2003. Effects of
Financial Globalization on Developing Countries: Some Empirical Evidence.
International Monetary Fund, Washington. IMF Occasional Paper No. 220.
Rodrik, Dani, Velasco, Andres, 2000. Short-term capital flows. Annual World Bank
Conference on Development Economics 1999. World Bank, Washington, pp. 59–90.
Schindler, Martin, forthcoming. Measuring Financial Integration: A New Dataset. IMF
Staff Papers.
Sørensen, Bent E., 2006, Comments on “Trends and Cycles in the Euro Area: How Much
Heterogeneity and Should We Worry About it?” European Central Bank Working
Paper Series No: 595.
Sørensen, Bent E., Yosha, Oved,1998. International risk sharing and European monetary
unification. Journal of International Economics 45, 211–238.
Sørensen, Bent E., Yosha, Oved, Wu, Yi-Tsung, Zhu, Yu, 2007. Home bias and
international risk sharing: twin puzzles separated at birth. Journal of International
Money and Finance 26 (4), 587–605.
13
M.A. Kose et al. / Journal of Development Economics xxx (2008) xxx–xxx
ARTICLE IN PRESS
Please cite this article as: Kose, M.A., et al., Does financial globalization promote risk sharing? Journal of Development Economics (2008),
doi:10.1016/j.jdeveco.2008.09.001