1 the role of foreign currency debt in financial crises : 1880-1913 vs. 1972-1997 michael bordo...
TRANSCRIPT
1
The Role of Foreign Currency DebtThe Role of Foreign Currency Debt in in Financial Crises : 1880-1913 vs. Financial Crises : 1880-1913 vs.
119972-199772-1997
Michael Bordo
Rutgers University and NBER
and
Christopher M. Meissner
University of Cambridge
2
IntroductionIntroduction The earlier era of globalization 1870-1914 was rife with frequent
emerging market currency crises, banking crises, and debt crises, with resonance for the recent past.
Many emergers suffered from “original sin” (Eichengreen and Hausmann 1999) - - the external debt that they accumulated was almost strictly denominated in foreign currency or in terms of gold (or had gold clauses).
When the exchange rate depreciates, debt service in gold or foreign currency becomes very difficult leading to default, and the consequent drying up of external funding.
The emerging country experience was in contrast to that of the advanced core countries which were financially mature, had credibility and could issue bonds denominated in terms of their own currency.
There were few crises in these countries.
3
In this paper we ask :
(a) whether these different debt structures might have played a role in explaining the difference in crisis incidence.
(b) whether debt management policies that created or alleviated balance sheet mismatches mattered (Goldstein and Turner 2004).
(c) whether poor reputation and accumulated default experience was a problem (Rienhart, Rogoff and Savastano 2003).
We have developed a database to allow us to identify and distinguish “original sin” and balance sheet crises from more traditional currency and banking crises for roughly 30 countries (both advanced and emerging) from 1880-1914. We also have data for over 40 countries between 1972 and 1997 for the post Bretton Woods period. Our data covers both crisis incidence and determinants.
Our results do not find unambiguous support for the idea that hard currency debt for emerging markets is always associated with more financial turbulence.
4
For the 1880 -1913 period we find evidence that emerging markets with significant amounts of original sin can be divided into two sub-groups:
(a) Argentina, Brazil, Chile, Italy and Portugal each of which suffered a financial catastrophe between 1880 and 1913
(b) Australia, Canada, New Zealand, Norway, and the US, which had relatively little trouble with financial crises.
For the recent period we find that:
(a) Countries like Argentina, Thailand, Indonesia and Brazil look like the former set for 1880-1913.
(b) Greece, Ireland, Israel, Spain, Sweden and Singapore look similar to the latter set
The difference between the two groups of countries we ascribe partially to good debt management and partially to special country characteristics that vary over time.
5
We also argue that there are two other sets of countries:
(a) most peripheral regions that suffer from original sin, but are closed to international capital flows and hence have avoided crises.
(b) most advanced economies without original sin, which have strong financial systems and few currency and debt crises.
We also find that debt management matters even after controlling for the level of original sin.
Many countries matched their hard currency liabilities with hard currency reserves. This helped reduce exposure to currency and banking crises.
Countries with better international repayment records were able to avoid debt crises despite high levels of debt.
6
FrameworkFramework We view banking trouble, currency crises and debt crises as inter-related
phenomena. It draws on Mishkin (2003).
One scenario is a rise in international interest rates which worsens the balance sheets, leads to declining net worth, a rise in performing loans or a decline in lending. A sudden stop or current account reversal may take place leading to further financial stress. Governments may have trouble servicing debt as capital markets become unwilling to roll over debt. The capital flow reversal could lead to abandonment of the peg and a large depreciation.
In the face of original sin, deprecation can led to increases in the real value of debt. The decline in net worth can lead to further disintermediation.
Ceteris Paribus, deterioration in balance sheets would be more severe the greater the amount of fixed interest rate hard currency debt outstanding.
Countries insure themselves against exchange rate movements by
backing up hard currency debt by hard currency assets.
7
Measuring Original SinMeasuring Original Sin
According to Eichengreen and Hausman (1999) foreign currency denominated debt was imposed by international capital markets. Nations with poor reputations, and even nations with good reputations or solid fundamentals, are obliged to issue debt in key international currencies.
The literature so far has done little to ascertain how harmful is original sin. We attempt to do this by examining the empirical association between original sin and financial crises.
Our measure of original sin for the 1880-1913 period is the ratio of hard currency public debt to total public debt. We define hard currency debt as debt that carried a gold clause or was made payable at a fixed rate in a foreign currency.
A key difference between our measure and that used by others is we look at debt issued in domestic and international markets instead of only international issues. The reason is that pre 1914 many domestic issues carried gold clauses.
8
For the 1972-1997 period we use the measure of original sin defined in Eichengreen Hausmann, and Panizza (2005).
The Eichengreen et al data base is only available for the period 1993-1997. To broaden the time span of our sample, we calculate the within country average of the observations and use this average values for all the years in which a country appears in the data set.
The key difference between the two measures we use, the inclusion of domestic original sin in the historical database. Evidence presented by Eichengreen et al suggests that for the most part their measure of original sin is highly correlated with domestic original sin.
0,
icountry by issued Securities
icurrency in issued Securities1maxiOS
9
Figure 1 shows the ratio of hard currency government debt to total government debt by country between 1880 and 1913.
Most countries original sin was constant over time.
Figure 2 shows the evolution of the original sin between 1993 and 1997. We tend to see strong persistence
Figure 3 shows long-run averages of original sin for 1880-1913.
Financial centres have less original sin.
Small peripheral countries have a lot of original sin.
Countries with ostensibly rotten fiscal institutions and poor international track records have intermediate levels of original sin.
Figure 4 shows the 1993-1997 country averages of original sin. It shows that both some very advanced countries like Belgium, Finland, Ireland, and Sweden have high original sin as do really all emerging market countries.
0.5
10
.51
0.5
10
.51
0.5
1
1880 1890 1900 1910 1880 1890 1900 1910
1880 1890 1900 1910 1880 1890 1900 1910 1880 1890 1900 1910 1880 1890 1900 1910
Argentina Australia Austria Belgium Brazil Canada
Chile Denmark Finland France Germany Greece
India Italy Japan Mexico Netherlands New Zealand
Norway Portugal Russia South Africa Spain Sweden
Switzerland United Kingdom United States Uruguay
Rat
io o
f Har
d C
urre
ncy
Deb
t to
Tota
l Gov
ernm
ent D
ebt
YearGraphs by country
Figure 1: Hard Currency Debt as a Percentage of Total Public Debt, 1880-1913
05
01
00
05
01
00
05
01
00
05
01
00
05
01
00
1993 1994 1995 1996 1997 1993 1994 1995 1996 1997 1993 1994 1995 1996 1997 1993 1994 1995 1996 1997 1993 1994 1995 1996 1997
Ar gentina Australia Austria Bangladesh Belg ium
Br azil COLOMBIA COSTA RICA COTE D' IVOIRE Canada
Chile China Denmark ECUADOR EGYPT
Fin land France GHANA Germany Greece
HONG KONG ICELAND INDIA INDONESIA IRELA ND
Ori
gin
al S
in,
bou
nded
at
zero
dateGraphs by country name
Figure 2: The Evolution of Original Sin in Sample Countries 1993-1997
0
50
10
00
50
10
00
50
10
00
50
10
00
50
10
00
50
10
0
1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7 1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7 1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7 1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7
1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7 1 9 9 3 1 9 9 4 1 9 9 5 1 9 9 6 1 9 9 7
IR E L A N D IS R A E L Ita ly J A M A IC A J a p a n K O R E A
M A L A Y S IA M E X IC O N E W Z E A L A N D N I G E R IA N ethe r la nd s N o rw a y
P A K IS T A N P A R A G U A Y P E R U P H IL I P P IN E S P o rt ug a l S E N E G A L
S I N G A P O R E S O U T H A F R IC A S R I LA N K A S p ai n S w e d e n S w it ze r la n d
T A I W A N T H A IL A N D T U R K E Y U R U G U A Y U ni ted K in gd o m U ni ted S ta te s
V E N E ZU E L A Z IM B A B W EOri
gin
al
Sin
, b
ou
nd
ed
at
ze
ro
d a teG ra p hs b y c o u n try na m e
Figure 2: (cont.) The Evolution of Original Sin in Sample Countries 1993-1997
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Per
cent
age
Country
Figure 3: Average Ratio of Hard Currency Public Debt to Total Public Debt, 1880-1913
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1A
vera
ge
Ori
gin
al
Sin
Me
asu
re
Sw
itzer
land
Ger
man
y
Uni
ted
Sta
tes
Japa
n
Italy
Fra
nce
Uni
ted
Kin
gdom
Por
tuga
l
Can
ada
Aus
tral
ia
Spa
in
New
Zea
land
Net
herla
nds
Sou
th A
frica
Hon
g K
ong
Den
mar
k
Bel
gium
Aus
tria
Gre
ece
Sin
gapo
re
Irela
nd
Sw
eden
Indo
nesi
a
Fin
land
Country
Figure 4: Average Level of Original Sin Between 1993 and 1997
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Ave
rage
Ori
gina
l Sin
Mea
sure
Nor
way
Thai
land
Phi
lippi
nes
Mal
aysi
a
Arg
entin
a
Icel
and
Mex
ico
Taiw
an
Kor
ea
Ven
ezue
la
Pak
ista
n
Chi
na
Chi
le
Per
u
Indi
a
Sri
Lank
a
Bra
zil
Col
ombi
a
Gha
na
Zim
babw
e
Ecu
ador
Jam
aica
Isra
el
Turk
ey
Cos
ta R
ica
Uru
guay
Country
Figure 4 (cont.)
16
Currency Mismatches andCurrency Mismatches and Debt IntoleranceDebt Intolerance
Goldstein and Turner (2004) argue that currency “mismatches” are the main problem with foreign currency debt. On the margin, changes in the exchange rate can become a problem the greater the mismatch as local currency assets lose value in terms of the foreign liabilities.
Our measure of mismatches for country i is
where reserves are gold reserves.
The measure above risks combining flow measures (exports) with stock measures. As an alternative measure, we use the total amount of interest payments due in gold or foreign currency.
For the recent period we use the total of external debt given in the World Bank’s Global Development Finance.
17
Debt Intolerance Reinhart, Rogoff and Savastano (2003) suggests that past defaults generate poor sovereign ratings.
For the nineteenth century we account for default history by:
(a) interacting the public debt to government revenue ratio with an indicator variable that equals one if a country had at least one default episode between 1800 and 1880.
(b) interact the debt to revenue ratio with an indicator equal to one if the country is in the periphery.
For the twentieth century we use the ratio of debt to output. We interact this ratio with an indicator that equals one if there was a default in the country between 1824 and 1971.
If the increase in the probability of a financial crisis for a marginal increase in the debt to revenue or debt to GDP ratio is larger for a peripheral country or a past defaulter, we would argue there is evidence in support of the debt intolerance hypothesis.
18
CrisesCrises Figure 5 presents the frequency of various types of crises (banking,
currency, twin, debt, “third generation” crises and all types of crisis together).
We define a crisis as Third generation if there was a debt crises accompanied by either a banking crisis, a currency crisis or both.
The figure shows the frequency of crises – the number of years a country was in crises divided by total possible years of observation.
For the pre 1914 period we see the pattern in Bordo et. al. (2001) that the predominant form of crises before 1914 were banking crises, followed by currency crises, twin and then debt crises.
The more recent period seems much more crisis prone in virtually all dimensions.
Figures 6 and 7, 8 and 9 present scatter plots of the percentage of time a country was in a crisis episode versus our measure of original sin and our mismatch variables.
Figure 5: Crisis Frequency in Percentage Probability per Year, 1880-1913 vs. 1972-1997
0
2
4
6
8
10
12
14
16P
erce
nt
Pro
bab
ilit
y p
er y
ear
BankingCrises
CurrencyCrises
Twin Crises
Debt Crises
"ThirdGen."Crises
All Crises
Type of Crisis
1880-1913
1970-1997
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
-0.2 0 0.2 0.4 0.6 0.8 1
Average ratio of hard currency debt to total debt
perc
enta
ge
Debt Crises
Twin Crises
Currency Crises
Banking Crises
Figure 6: Crisis Frequencies By Country versus the Average Level of Hard Currency Public Debt to Total Public Debt, 1880-1913
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
-10 10 30 50 70 90
Average Original Sin Measure
Pe
rce
nt
Pro
ba
bil
ity Banking Crises
Currency Crises
Debt Crises
Twin Crises
Third Gen. Crises
Alternative Debt Crises
Figure 7: Crisis Frequencies By Country versus the Average Level of Original Sin, 1972-1997
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
Average "Mismatch"
per
cen
tage
Debt Crises
Twin Crises
Currency Crises
Banking Crises
Figure 8: Crisis Frequencies by Country versus the Average Level of the “Mismatch” Measure, 1880-1913
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
-0.5 0 0.5 1 1.5 2 2.5 3 3.5 4
Average Mismatch
Per
cent
pro
babi
lity
Banking Crises
Currency Crises
Debt Crises
Twin Crises
Third Gen. Crises
Alternative Debt Crises
Figure 9: Crisis Frequencies by Country versus the Average Level of the “Mismatch” Measure, 1972-1997
24
Patterns
(a) In the pre-World War I era, there appears to be an inverse U relationship between debt crises and original sin. Countries with intermediate ranges of original sin seem to take longer to resolve their debt crises than those at either end of the spectrum.
(b) Today there appears to be a direct positive relationship between the severity of debt crises and the average level of original sin and similarly for currency and banking crises.
(c) For the mismatch in Figures 8 and 9 there appears to be a positive relationship between mismatches and the length of time spent in a debt crisis.
(d) For the nineteenth century the quadratic seems to pick up the following:
(i) advanced countries have low levels of OS and few crises
25
(ii) The British Dominions, Northern Europe, and the U.S. have high levels of OS and few crises.
(iii) In the intermediate range we see Southern Europe, Eastern Europe and Latin America with medium OS and more crises.
(e) For both centuries the data suggest that at higher per capita income levels original sin is less likely to cause debt or banking crises.
26
Statistical ResultsStatistical Results
We use panel probits to ascertain the role of measures of original sin, currency mismatch and debt intolerance in precipitating the various types of crises.
In addition to the measures of balance sheet strength we control for a number of standard determinants of crises including: total government debt divided by government revenues, growth in the terms of trade, deviation of the exchange rate from the period average, current account to GDP. Yield spread on consols, money growth, gold reserve to note circulation.
27
Basic ResultsBasic Results
(1) We find that more original sin may increase the likelihood of banking and debt crises, but that there is an important interaction with the level of development and other factors.
(2) We find that the positive marginal effect of increased original sin or hard currency debt on debt crises in both periods and banking crises in the twentieth century is smaller at the highest levels of
per capita income. (3) This suggests that original sin may contribute to more financial
crises but that sometimes the damage can be limited as the
development process continues. (4) There is solid evidence that mismatches are positively correlated
with many types of crises.
28
(5) We find that debt crises are associated with the outbreak of a currency crisis in both periods.
(6) To some extent the evidence shows that debt crises are associated with banking crises.
(7) We find that banking crises and currency crises are more positively associated in the nineteenth century than the twentieth century. This evidence suggests that conditional on having a crisis, episodes of instability tended to be more systemic in the past than today.
29
Debt Crises Table 1 shows the results for 1880-1913.
Column 1 shows an inverse U shaped pattern in original sin, in mismatches and there is evidence of debt intolerance.
These are the countries in the areas of recent settlement which had strong financial systems, good fiscal institutions intent and which borrowed largely for productive investments.
Most of the control variables have signs that fit our priors.
We interpret the quadratic in original sin as saying that more original sin is associated with a higher likelihood of a debt crisis, but those observations with very high levels of original sin face a lower likelihood. See figure 10 which shows the shape of the marginal effect of various values of the hard currencies debt to total debt ratio. And figure 11 which presents the predicted probabilities of a debt crisis for various values of the ratio of hard currency debt to total debt.
Table 1: Determinants of Debt Crises
-0.1
-0.05
0
0.05
0.1
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95
Ratio of Hard Currency Debt to Total Debt
Mar
gina
l Effe
ct
c
Figure 10: Marginal Effect of the Ratio of Hard Currency Debt to Total Debt
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Ratio of Hard Currency Debt to Total Debt
Pre
dic
ted
Pro
bab
ility
of
a D
eb
t C
risi
s
Currency Crisis at t &Banking Crisis at t-1
Currency Crisis at t
no currency crisis at t
Figure 11: Predicted Probabilities of a Debt Crisis, 1880-1913
33
Table 2 shows the results for debt crises for the 1972-1997 period.
Column 1 shows there is a positive relationship between our measure of original sin and debt crises.
Column 4 shows that mismatches increase the susceptibility to debt crises
Column 3 shows the interaction between GDP per capita and original sin. Middle income countries, emerging markets are the most prone to have a crisis. See figure 12.
Table 2: Determinants of Debt Crises, 1972-1997
35
Figure 12 suggests that we can break the sample of countries into 3 categories where original sin is high:
a) to the left the poor countries which don’t rely on external finance
b) in the middle income emerging markets which rely on external financing
c) the highly developed countries who either are less exposed or can deal with shocks to the financial system.
Figure 13 classifies countries into 4 groups based on GDP per capita and level of OS.
We then look at the average trade deficit within each group, the average time spent in a debt crisis, the average predicted probability of a debt crisis and the median predicted probability.
We see that the richest countries are lower on all 5 dimensions whereas the middle income are highest.
0
0.05
0.1
0.15
0.2
0.25
real GDP per capita 1989$
Pre
dict
ed P
roba
bilit
y of
a D
ebt C
risi
s
Currency Crisis at t
no currency crisis at t
Figure 12: Predicted Probabilities of a Debt Crisis, 1972-1997.
9
4
0.4
2.91.58
1
8
5
6.111.7
00
0.04
0
0.41-0.77
00
0.4
0
Median Predicted Probability of a DebtCrisis
Avg. Trade Deficit/GDP
Avg. Debt Crisis FrequencyAverage Predicted Probability of a Debt
Crisis
Median mismatch
< $2,900
$2,900 to $8,100
>$8,100 & high OS
>$8,1000 & low OS
Figure 13: “Radar” Chart Showing the Four Part Categorization of Countries
38
Currency Crises
For 1880-1913 with respect to currency crises there is only weak evidence for the role of our balance sheet measures and there is some evidence of a quadratic. See Table 3.
We view most currency crises as a symptom of capital flight from a crumbling financial sector and liquidity problems, and think that indirectly original sin is associated with currency crises.
We see that initial problems in the banking sector (proxied by one-year ahead indicators of debt crises and banking crises) are strongly associated with currency crises.
For the twentieth century, Table 4 shows evidence that original sin leads to a higher chance of a currency crisis as does the mismatch.
Table 3: Determinants of Currency Crises, 1880-1913
Table 4: Determinants of Currency Crises, 1972-1997
41
Banking Crises Banking crises seem to be associated with OS and currency mismatches in both periods but much more so in the nineteenth century.
Between 1880 and 1913 we see a quadratic as we did in Table 1.
Between 1972 and 1997 we see that the impact of more original sin is higher in low and middle income countries than in high income countries.
These results are presented in Table 5 - 6.
Table 5: Determinants of Banking Crises, 1880-1913
Table 6: Determinants of Banking Crises, 1972-1997
44
RobustnessRobustness
We check to see if our results - - that after a certain point more hard currency debt relative to the total seemed to be associated with fewer debt crises and banking crises - - are spurious reflecting characteristics of countries left out of the analysis.
Perhaps those most at risk have ways to deal with crises despite their high levels of OS.
To check this we use a fixed effects linear probability model estimated by OLS in Table 7 which covers the early period.
The results on the coefficients on the original sin mismatch variable are similar to those in the previous table.
The results for the twentieth century in Table 8 were similar but less significant.
45
ConclusionsConclusions
Our key finding is that hard currency debt may not always generate a higher likelihood of a financial crisis.
In both centuries, some countries with high levels of OS averted crises. They tended to be high income and to have backed up their foreign currency debt with hard currency reserves.
We categorize our countries into 4:
(a) financial centers with low original sin and strong financial fundamentals obviously avoid crises.
(b) smaller advanced countries with high OS also avoided crises.
(c) LDC’s which are relatively closed also avoided crises.
(d) middle income emergers with open capital accounts are most prone to crises.
46
Our bottom line is that original sin often makes debt crises more likely and makes avoiding currency and banking crises more difficult.
The lesson from two different periods of globalization is that more careful debt management policies and the development of sound fiscal and financial institutions seem to make it possible for governments and firms to avoid severe financial crises.