Professional Documents
Culture Documents
17621783, 2012
2012 Elsevier Ltd. All rights reserved
0305-750X/$ - see front matter
www.elsevier.com/locate/worlddev
http://dx.doi.org/10.1016/j.worlddev.2012.04.024
1. INTRODUCTION
1763
1764
WORLD DEVELOPMENT
Dt
;
Yt
DP t et DF t F t Det
gt pt dt ;
Dt
Dt
Dt
pbt it P t iFC
F t Det
t et F t S t
gt pt dt :
Yt
Yt
et
Det
b t p pit
pt pt )
R ER
t
et
et
Yt
In deriving this expression, we have also used the fact that the
nominal GDP growth rate is equal to real growth (gt) and
ination (pt). rt is the real interest rate on domestic-curthe real interest rate on forrency-denominated debt and rFC
t
eign-currency-denominated debt. The last term on the righthand side is the correction factor that arises from the existence of foreign-currency-denominated debt, RERt is the real
exchange rate and the hat denotes percentage change.
Notice that ignoring this correction factor will result in a
miscalculation of the primary decit consistent with a stable
debt-to-GDP ratio. In particular, if the country experiences
a real devaluationthat is, if (RERt/RERt) > 0, 14 ignoring
the correction factor will result in an underestimated primary balance consistent with sustainability.
The case where there is a BalassaSamuelson eect in which
the currency exhibits a real appreciation over time can be
thought of as leading to a larger sustainable primary decit
(see Eqn. (6)), since this implies that the value of the foreign
debt as a percentage of GDP is declining over time due to this
eect. This, however, does not invalidate our concern regarding the dramatic valuation eects that can take place if there is
a devaluation of considerable magnitude.
Assuming that d* is the sustainable public-debt-to-GDP ratio in the sense that it is consistent with the demand for that
debt, then if dt < d*, the country will be able to sustain smaller
primary balances than those of a steady state for a while. In
Eqn. (6) Ddt is positive (hence the primary balance smaller),
that is, the debt-to-GDP ratio can increase and still be consistent with debt sustainability. The inverse occurs if dt > d*.
We assume that the uncovered real interest rate parity holds:
b t qt rFC ;
rt rt ER ER
t
7
rt
F t DEt
:
Yt
10
Expression (9) tells that as t approaches t, the level of accepted public debt approaches the steady-state level d*. On
one extreme, if the market gives the country just one period
to adjust to d*, then d
t d . This means that the primary
balance will have to increase as much as necessary so that
the ratio of public debt to GDP remains unchanged at level
d*. On the other extreme, if the market gives the country
innite time to adjust (t 1), then
dt d0 . This means that
the new equilibrium debt-to-GDP ratio becomes d0 rather
than d*.
Finally, we assume that when public debt is above its sustainable level, GDP growth is below its potential, the reasons
being the higher interest rate discussed above and the increased uncertainty derived from the fact that taxes or seigniorage might be increased to nance the larger primary
decit. 17 The larger the gap between sustainable and actual
debt, the greater the impact on growth. Additionally, to explore potential nonlinearities of the relation between public
debt and GPD growth, we follow earlier studies by Reinhart
and Rogo (2010) and Kumar and Woo (2010) and work with
a threshold of initial public debt beyond which the debt begins
to have an adverse eect on growth. This relationship can be
expressed as:
if dt 6 d
gt
;
11
g
gt bdt d if dt > d
where g* is the potential GDP growth rate and d* is the debt
threshold beyond which debt will have adverse eect on
growth.
4. ESTIMATIONS
In this section, we are interested in estimating the eect of
public debt on the interest rate of that debt (Eqns. (7) and
(8)), the dynamics of the debt when there is a shock to the real
exchange rate (Eqns. (9) and (10)), and the impact of debt on
growth (Eqn. (11)).
Our estimations considered ve Latin American countries:
Brazil, Chile, Colombia, Mexico, and Peru. 18 We used quarterly data for the period 1999:I2007:IV. The variables included country risk as measured by the quarterly average of
the EMBI Plus index qt . 19 The data on external public debt
as a percentage of GDP were obtained from the World Bank.
For the real exchange rate, we used the traditional measure:
1765
Et P t
RERt
;
Pt
where P t is the dollar-denominated CPI of the main trading
partners, E is the nominal exchange rate (local currency units
per dollar), and P is the domestic CPI. All these data plus the
GDP growth data were obtained from the central banks of the
countries considered in the study. 20
(a) Public debt and interest rates
We now proceed to estimate the eect of a change in public
debt on the interest rate on that debt. Because there is no total
public debt quarterly data available for all the countries considered, we do our estimations using only foreign public
debt. 21 Our dependent variable is the risk premium (see Eqns.
(7) and (8)), that is, we want to estimate the eect of a change
in the public debt on the risk premium paid on that debt.
We estimate the following panel:
qit a1 a2 qit 1 a3
li nit ;
EPD
GDP
a4 RERit a5 git kt
it
12
where subscript i denotes the country (i = 1, 2, . . . , 5) and subscript t denotes the quarter:
EPD
GDP
it is the ratio of public foreign debt to GDP. As we
mentioned above, it is expected that the greater the public
debt, the higher the risk premium.
RERit is the real exchange rate index. Like in Edwards
(1986) and Min (1998), we analyze whether a less competitive real exchange rate (appreciation) can adversely aect
the risk premium. According to Cline (1983), real appreciations in LDCs played a major role in the overborrowing
process.
git is the rate of growth of GDP. We expect that a higher
GDP growth rate will reduce country risk.
Finally, kt is a temporary xed eect, li is a country-specic xed eect, and nit is an iid 0; r2n distributed-error
term.
We rst estimated Eqn. (12) using a static GLS xed-eect
method. The Hausmann test rejects the null hypothesis of no
correlations between the xed eects and the explanatory variables, thus validating the estimation through xed eects.
However, considering both the dynamic structure of the model
and the potential endogeneity of the explanatory variables
([EPD/GDP]iRERit and git ) we proceeded to estimate using
System GMM proposed by Arellano and Bover (1995) and
Blundell and Bond (1998).
The results are presented in Table 1. The rst regression
shows the results of the xed-eect estimation. The eect of
the debt on the risk premium (and hence on the interest rate)
is positive and statistically signicant. The coecient shows
that a one percentage-point increase in the debt-to-GDP ratio
(say, from 30% to 31%) produces an increase in the interest
rate of 21 basis points (say, from 5% to 5.21%). Regressions
24 show the results of the dynamic estimation using System
GMM methodology. The results are consistent with those
found in regression 1 in the sense that the coecient of the
debt variable is positive and statistically signicant, and the
magnitude of the eect is 0.11%. However, we are able, with
the dynamic specication, to obtain long-term eects. Note
that the coecient of the lagged dependent variable is positive
and statistically signicant, indicating persistence in the eect
on the risk premium. 22
1766
WORLD DEVELOPMENT
Table 1. Estimates using xed eects model and System GMM
Eqn. (1)
qt1
(External public debt/GDPt)
RERt
GDP growth ratet
Time eect
2
R
Observations
Number of id
Serial correlation order 1
Serial correlation order 2
Sargan p value
0.213***
(0.0265)
0.0504***
(0.00860)
0.0812
(0.0555)
Yes
0.877
155
5
Eqn. (2)
Eqn. (3)
Eqn. (4)
0.691***
(0.0729)
0.0969**
(0.0269)
0.537**
(0.139)
0.113***
(0.0259)
0.0473
(0.0276)
Yes
Yes
0.659***
(0.156)
0.113*
(0.0462)
0.0232
(0.0224)
0.288
(0.184)
Yes
153
5
0.0696
0.219
0.107
153
5
0.0878
0.202
0.118
153
5
0.0384
0.220
0.108
The Sargan test statistic for overidentifying restrictions conrms the exogeneity of the instruments validating our GMM
estimations. The second-order residual serial correlation tests
conrm the fact that the errors are serially uncorrelated.
Given that including Brazil in the estimations could be troublesome because of the persistently high interest rates, we perform a sensitivity analysis with Brazil excluded from our
sample. When excluding Brazil from the sample, the coecient
of public debt to GDP declines to 0.05, meaning that a onepercentage-point increase in the debt-to-GDP ratio produces
an increase in the interest rate of 5 basis points (see Appendix
Table 6). This coecient is about half of what we obtained
when Brazil was included in the sample.
(b) The dynamics of public debt
The second step in our analysis is to study the dynamics of
public debt when the exchange rate varies (Eqns. (9) and (10)).
To address this issue, we use vector autoregressions (VAR) to
analyze the dynamic impact on public debt in each country of
a random disturbance in the real exchange rate.
Let Xt be the vector of endogenous variables. It would be
written in reduced form as follows for the VAR system:
X t ALX t1 U t ;
h
i
EPD
; RERt ; qt ; gt and Ut is a reduced residuals
where X t GDPt
vector dened as:
h EPD
i
; uRER
; uqt ; ugt :
U t uGDP
t
t
The equations include a lag, established using Akaikes criteria.
The reduced residuals (Ut) and, more specically, the resid
EPD
,
uals of the ratio between foreign public debt and GDP uGDP
t
can also be determined as: (1) a linear combination of the response of the foreign-public-debt-to-GDP ratio when there are
unexpected shocks in the other variables, (2) the discretionary
response of the policy-maker to changes in the variables 23 and
(3) the random shocks of foreign debt. 24
As mentioned earlier, our analysis concentrates on the
response of the foreign-public-debt-to-GDP ratio to an
unexpected shock in the real exchange rate. We are specically
looking to estimate the time (t) that a certain country takes to
adapt to its sustainable public debt level. We therefore estimated the eect of an unexpected shock (one-time only) on
the RER, assuming that the RER will not react contemporaneously to changes in the other model variables and that the
other variables (EPD/GDP, q, g) do not react contemporaneously to shocks in the other variables, except for changes in the
RER. Moreover, the path of EPD/GDP is estimated separately from the shocks to the rest of the variables (i.e., the only
shock that is received throughout the period of analysis hits
the RER).
The results are presented in Figure 1, panels AE. This gure shows the response of the foreign-public-debt-to-GDP ratio to a one standard deviation shock in the RER using the
Cholesky decomposition with the small sample degrees of freedom correction to orthogonalize the impulses. Asymptotic
standard errors are used to construct the condence bands.
The Cholesky scheme implies that RER shocks have no contemporaneous eect on the interest rate and GDP growth rate.
The RER shock aects only the debt within the same quarter.
The results show the responses over a horizon of 15 years. The
gure shows that the foreign public debt-to-GDP ratio increases and that this eect lasts for approximately two to three
years in the cases of Brazil, Chile, Colombia, and Mexico. In
the case of Peru we found no signicant eect. 25
Hence, for the simulations in our model, we will assume that
the time in which the country has to return to an acceptable
level of public debt (in the sense discussed in the previous section) is three years, but we also use a range going from 1 to
10 years.
(c) Public debt and growth
Finally, we turn to the question about the eect of public
debt on growth (Eqn. (11)). To nd the answer, we estimate
the following dynamic regression:
External Public Debt
git b1 b2 git1 b3
GDP
it
External Public Debt
Dum 30 b4
GDP
it
Dum 30 60 b5 RERit b6 rit kt ui vit :
13
1767
Figure 1. Impulse response functions response of external debt-to-GDP ratio to Cholesky one S.D. RER innovation.
As in Eqn. (12), we address the possible endogeneity problem of explanatory variables (external public debt, RER,
and spread) using System GMM. Table 2 presents the results.
The coecient of external public debt when there is a low initial debt is negative and statistically insignicant in all of the
equations. In contrast, the coecient of external public debt
when debt is above 30% is negative and statistically signicant,
indicating that growth declines as the debt increases. The coefcient indicates that a one percentage point increase in the
public-debt-to-GDP ratio reduces the growth rate by 0.09%
points in the short term.
The Sargan test and the second-order residual serial correlation tests conrm both the fact that the errors are serially
uncorrelated and the validity of the instrumental variables.
As in our estimates of interest rates we performed a sensitivity analysis excluding Brazil from our sample. Results are
shown in Appendix Table 7. The results show that a change
of one percentage point in the debt ratio would reduce growth
by 0.06% at levels above 30%, The impact of a change in debt
on growth rate is about two thirds of what we nd in the full
sample regressions. Reinhart (2010) nd that growth deteriorates markedly at debt levels over 60%. 26
5. SIMULATIONS
In this section, we simulate the sustainable primary balance
and debt dynamics of a country facing an initial real deprecia-
1768
WORLD DEVELOPMENT
Table 2. Estimates using System GMM
Eqn. (1)
Eqn. (2)
Eqn. (3)
0.535***
(0.115)
0.127*
(0.0533)
0.219
(0.143)
0.535***
(0.106)
0.0914*
(0.0358)
0.0925
(0.0541)
0.0412
(0.116)
Yes
Yes
0.536***
(0.107)
0.0912**
(0.0348)
0.0930
(0.0515)
0.0448
(0.151)
0.000711
(0.0117)
Yes
155
5
0.0979
0.552
0.406
154
5
0.0398
0.187
0.239
154
5
0.0388
0.189
0.224
Symbol
(e0F0/D0)
r0
5.0%
a2
0.659
a3
0.113
g0
4%
b2
0.536
b3
0.09
t
S/Y
Assumed value
35%
5570%
Between 1 and 10
0.4%
It is assumed to be equal to
r* + q = rFC
SGMM estimator in
autoregressive panel data model
SGMM estimator in
autoregressive panel data model
Simulation can be made using
dierent rates of growth
SGMM estimator in
autoregressive panel data model
SGMM estimator in
autoregressive panel data model
According to impulse response
functions
Consistent with a monetary base
of 6% of GDP and a nominal
GDP growth rate of 7%
1769
Table 4. Real devaluation of 25% initial foreign currency-denominated debt as a percentage of total debt: 55%
Panel A
Primary balance
Panel B
Public debt (% of GDP)
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
0.05
1.60
0.08
0.03
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.97
0.95
0.66
0.02
0.04
0.05
0.05
0.05
0.05
0.05
0.05
0.85
0.95
0.78
0.58
0.38
0.01
0.04
0.05
0.05
0.05
0.05
0.79
0.95
0.84
0.69
0.54
0.40
0.26
0.02
0.05
0.05
0.05
0.75
0.95
0.88
0.78
0.67
0.57
0.46
0.36
0.26
0.16
0
1
2
3
4
5
6
7
8
9
10
39.8
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
39.8
38.2
36.6
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
39.8
38.9
37.9
36.9
36.0
35.0
35.0
35.0
35.0
35.0
35.0
39.8
39.1
38.4
37.8
37.1
36.4
35.7
35.0
35.0
35.0
35.0
39.8
39.3
38.9
38.4
37.9
37.4
36.9
36.4
36.0
35.5
35.0
Panel C
Real interest rate
Panel D
Rate of real GDP growth
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
5.00
5.65
5.12
5.02
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
6.08
5.85
5.38
5.07
5.01
5.00
5.00
5.00
5.00
5.00
5.00
6.16
6.13
5.86
5.55
5.23
5.04
5.01
5.00
5.00
5.00
5.00
6.20
6.24
6.07
5.85
5.62
5.39
5.17
5.03
5.01
5.00
5.00
6.23
6.33
6.22
6.07
5.91
5.75
5.60
5.44
5.28
5.12
0
1
2
3
4
5
6
7
8
9
10
4.0
3.6
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.3
3.5
3.8
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.2
3.3
3.5
3.7
3.9
4.0
4.0
4.0
4.0
4.0
4.0
3.2
3.3
3.4
3.5
3.7
3.8
3.9
4.0
4.0
4.0
4.0
3.2
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.9
Now we assume a real devaluation of 25% at the end of period zero. As mentioned earlier, it is assumed that there are no
further expected changes in the real exchange rate. 29 Table 4
shows the results of our simulations for dierent values of t
and for 10 years after the devaluation occurs, when 55% of
public debt is denominated in foreign currency. Panel A shows
the primary balance sustainable path. If the country has three
years to adjust to the equilibrium debt-to-GDP ratio, it has to
have primary surpluses of 0.97%, 0.95%, and 0.66% respectively in the rst three years. This means a primary balance
adjustment of between 1.02% and 0.71% of GDP per year
for three years as compared to the nondevaluation situation.
This is clearly not a minor adjustment for a country that
was supposed to be scally sustainable. Panel B shows the
path of public debt as a percentage of GDP. For the same
t 3, it goes up to 39.8% of GDP right after the devaluation.
At the end of the rst year, it is down to 38.2% of GDP, and it
is back to its equilibrium level (35%) at the end of period 3.
Interest rates (panel C) go up to 6.08% at the end of year 1
and back to 5.0% in year 5. 30 Growth declines to 3.3% in
the rst year and returns to 4% after the fourth year.
If the country has to adjust in just one year, debt goes back
to its initial level at the end of year 1. But the required adjustment in the primary balance in that year is 1.6% of GDP.
In Table 5 we simulate the case where 70% of public debt is
denominated in foreign currency. If the country has three
years to return to the equilibrium debt-to-GDP ratio, it has
to adjust its primary balance between 1.27% and 0.86% of
GDP per year for three years as compared to the nondevaluation situation. Public debt goes up to 41.1% of GDP immediately after the devaluation. At the end of the rst year, it is
1770
WORLD DEVELOPMENT
Table 5. Real devaluation of 25% initial foreign currency-denominated debt as a percentage of total debt: 70%
Panel A
Primary balance
Panel B
Public debt (% of GDP)
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
0.05
2.06
0.11
0.02
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
1.27
1.24
0.86
0.04
0.04
0.05
0.05
0.05
0.05
0.05
0.05
1.11
1.24
1.02
0.75
0.50
0.00
0.04
0.05
0.05
0.05
0.05
1.05
1.25
1.10
0.90
0.71
0.52
0.34
0.01
0.04
0.05
0.05
1.00
1.25
1.16
1.03
0.88
0.74
0.61
0.48
0.35
0.22
0
1
2
3
4
5
6
7
8
9
10
41.1
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
41.1
39.1
37.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
41.1
39.9
38.7
37.5
36.2
35.0
35.0
35.0
35.0
35.0
35.0
41.1
40.3
39.4
38.5
37.6
36.8
35.9
35.0
35.0
35.0
35.0
41.1
40.5
39.9
39.3
38.7
38.1
37.5
36.8
36.2
35.6
35.0
Panel C
Real interest rate
Panel D
Rate of real GDP growth
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
5.00
5.82
5.16
5.03
5.01
5.00
5.00
5.00
5.00
5.00
5.00
5.00
6.37
6.08
5.48
5.09
5.02
5.00
5.00
5.00
5.00
5.00
5.00
6.48
6.43
6.09
5.70
5.30
5.06
5.01
5.00
5.00
5.00
5.00
6.53
6.58
6.36
6.08
5.79
5.50
5.21
5.04
5.01
5.00
5.00
6.56
6.70
6.55
6.36
6.16
5.96
5.76
5.55
5.35
5.15
0
1
2
3
4
5
6
7
8
9
10
4.0
3.5
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.1
3.4
3.8
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.0
3.2
3.4
3.6
3.9
4.0
4.0
4.0
4.0
4.0
4.0
3.0
3.1
3.2
3.4
3.6
3.7
3.9
4.0
4.0
4.0
4.0
3.0
3.0
3.1
3.2
3.3
3.5
3.6
3.7
3.8
3.9
nated public debt, but also considering the adverse eects that
external debt could have on the overall economic performance, more specically on the countrys interest rate and
GDP growth.
Assuming that the country has three years to return to the
equilibrium public debt-to-GDP ratio, the required primary
balance adjustment after the devaluation of 25% is about 1%
of GDP for each of those three years. Moreover, not taking
into account the dynamics of the interest rate and GDP
growth rate, would have led us to the mistaken conclusion that
the required scal adjustment was about half of the actual required adjustment.
The longer the adjustment period, the larger the underestimation of the adjustment in the primary balance required to
return to the equilibrium level of public debt, since as the
adjustment period increases, both interest rates and growth
are aected for a longer period. By ignoring the impact on
interest rate and growth assuming them to be exogenous, previous studies underestimate the scal costs of a real exchange
rate devaluation.
1771
NOTES
1. Naturally, the debt ratio may be calculated relative to alternative
benchmarks. On sustainability analyses see, for example, Milesi-Ferretti
and Razin (1996, 2000) and Edwards (2002).
2. The primary balance is dened as the nominal balance, excluding
interest payments.
3. See, for example, World Bank and IMF (2002, 2004) and Lachler
(2001).
4. See Cuddington (1997) and Arnone, Bandiera, and Presbitero (2008).
5. See IMF and World Bank (2010).
6. It could be argued that a devaluation from a position of an
unsustainably appreciated currency can make room for reductions rather
than increases in interest rates. Although this is a plausible case, it is usual
to see that devaluations initially bring turmoil to the market hence
increasing interest rates. In our model and simulations we assume that this
is the case (via an increase in public debt) as it has been so in most cases of
signicant depreciations in Latin American countries. In addition, our
empirical estimates indicate that interest rates rise after a depreciation (the
channel is an increase in public debt). On the other hand, it could also be
argued that devaluations are expansionary rather than contractionary,
hence GDP growth would increase rather than decrease. It is generally
observed, though, that this is not the case in the short term. Initially
devaluations are traumatic and reduce economic growth. In our simulations we assume that growth returns to its long-term value after the
country returns to a sustainable level of public debt. In addition our
empirical estimates indicate that growth declines after a depreciation.
7. See also Eichengreen, Hausmann and Panizza (2003).
8. We consider this sample is representative of Latin American economies. Although for more general conclusions for the whole set of emerging
economies it would be necessary a broader sample.
9. See Edwards and Vergara (2002) for a detailed discussion of some of
these parameters.
10. See for example Wessel (2007) and Dizard (2008).
17. Barro (1979) and Dotsey (1994) argue that high debt can aect
growth via higher future distortionary taxation.
18. The sample is limited by the availability of country risk data for
Latin American countries. In a sensitivity analysis we excluded Brazil from
our sample. Estimates without Brazil are shown in Appendix tables.
19. The EMBI Plus measures the interest rate dierential between the
dollar bonds issued by governments and the US Treasury bonds. The
source is www.valorfuturo.cl.
1772
WORLD DEVELOPMENT
20. The unit root tests of our variables reject the null hypothesis that
variables are order-1 integrated.
21. For the ve countries considered in this paper, foreign public debt
represents about 50% of total public debt. For less developed countries,
the share of foreign debt is even higher.
26. Reinhart (2010) nd that average GDP growth falls 3% for external
debt levels above 60%. As they point out, In light of this, it is more
understandable that over one half of all defaults on external debt in
emerging markets since 1970 occurred at level of debt that would have met
the Maastricht criteria of 60% or less.
27. See Edwards and Vergara (2002) and Edwards (2002).
22. We also tested specications with nonlinear terms, but we did not
obtain dierent results.
23. Like in Blanchard and Perotti (2002), the discretionary responses are
assumed to take more than one quarter to appear and, therefore, they do
not capture the quarterly data from the series used.
24. See Perotti (2005) for an in-depth discussion of this subject.
25. In addition, we included, in Appendix gures, a set of rolling IRFs of
External Debt to Cholesky One S.D. RER innovation, in order to examine
more closely the sensitivity of external debt responses to sample variation.
In the recursive regressions, the initial sample is xed (rst 12 quarters)
and then observations are added one at a time. Appendix Figures A1A5
show snapshots of the rolling regression for dierent time periods (one,
three, ve, seven and 10 years). As the gure shows, results for most of the
samples point to an increase in public debt-to-GDP ratio that lasts for two
to three years.
b 0.
28. Remember that we assume that ER ER
29. This is clearly a simplifying assumption since the scal adjustment
can be expected to produce further changes in the exchange rate.
Nonetheless, although the scal adjustment obtained in this exercise is
large as compared to scal expenditure, it is relatively small as compared
to the overall aggregate demand of the economy. Hence, it is reasonable to
assume that the further eects over the real exchange rate will be rather
small.
30. Notice that neither growth nor the interest rate return to their
previous levels immediately after t years due to the presence of persistence
in these variables.
31. Even if the companies are not state-owned, there is an eect through
tax revenues obtained from private companies producing and exporting
commodities.
REFERENCES
Akitoby, B., & Stratmann, T. (2006). Fiscal policy and nancial markets.
International Monetary Fund working paper 06/16. <http://www.imf.org/external/pubs/ft/wp/2006/wp0616.pdf> Retrieved 16.12.11.
Arellano, M., & Bover, O. (1995). Another look at the instrumental
variable estimation of error-components models. Journal of Econometrics, 68, 2951.
Arnone, M., Bandiera, L., & Presbitero, A. (2008). Debt sustainability
framework in Hipcs: A critical assessment and suggested improvements.
<http://ssrn.com/abstract=871171> Retrieved 16.12.11.
Baldacci, E., & Kumar, M. (2010). Fiscal decits, public debt and sovereign
bond yields. International Monetary Fund working paper 10/184.
<http://www.imf.org/external/pubs/ft/wp/2010/wp10184.pdf>
Retrieved 16.12.11.
Barro, R. (1979). On the determinants of the public debt. Journal of
Political Economy, 85(5), 940971.
Blanchard, O., & Perotti, R. (2002). An empirical characterization of the
dynamic eects of changes in government spending and taxes on
output. The Quarterly Journal of Economics, 117(4), 13291368.
Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions
in dynamic panel data models. Journal of Econometrics, 87, 115143.
Bohn, H. (1998). The behavior of U.S. public debt and decits. Quarterly
Journal of Economics, 113(3), 949963.
Bohn, H. (2005). The sustainability of scal policy in the United States.
<http://www.econ.ucsb.edu/~bohn/papers/DebtUS.pdf>
Retrieved
16.12.11.
Bordo, M., & Meissner, C. (2006). The role of foreign currency debt in
nancial crises: 18801913 vs. 19721997. Journal of Banking Finance,
30(12), 32993329.
Calvo, G., & Reinhart, C. (2002). Fear of oating. Quarterly Journal of
Economics, 117(2), 379408.
Calvo, G., Izquierdo, A., & Talvi, E. (2003). Sudden stops, the real
exchange rate and scal sustainability: Argentinas lessons. NBER
working paper 9828. <http://www.nber.org/papers/w9828> Retrieved
16.12.11.
Cuddington, J. (1997). Analyzing the sustainability of scal decits in
developing countries. World Bank policy research working paper 1784.
<http://papers.ssrn.com/sol3/papers.cfm?abstract_id=597231>
Retrieved 16.12.11.
Cline, W. (1983). Interest and debt: Systematic risk and policy response.
Washington, DC: Institute for International Economics, MIT Press.
Dizard, J. (2008, October 21). Now the suering spreads to emerging
market banks. Financial Times. <http://www.ft.com/cms/s/0/
1773
1774
WORLD DEVELOPMENT
APPENDIX A
See Figures A1A5 and Tables 613.
Figure A1. Brazil rolling IRFs of External Debt to Cholesky One S.D. RER innovation at one, three, ve, seven, and 10 years.
Figure A2. Chile rolling IRFs of External Debt to Cholesky One S.D. RER innovation at one, three, ve, seven, and 10 years.
1775
1776
WORLD DEVELOPMENT
Figure A3. Colombia rolling IRFs of External Debt to Cholesky One S.D. RER innovation at 1 quarter, one, three, ve, seven, and 10 years.
Figure A4. Mexico rolling IRFs of External Debt to Cholesky One S.D. RER innovation at one, three, ve, seven, and 10 years.
1777
1778
WORLD DEVELOPMENT
Figure A5. Peru rolling IRFs of External Debt to Cholesky One S.D. RER innovation at 1 quarter, one, three, ve, seven, and 10 years.
1779
Table 6. Estimates using xed eects model and System GMM Excluding Brazil
Dependent variable: interest rate spread (qt)
Eqn. (1)
qt1
(External public debt/GDPt)
RERt
GDP growth ratet
Time eect
Observations
R2
Number of id
Serial correlation order 1
Serial correlation order 2
Sargan p
0.219***
(0.0252)
0.00259
(0.0110)
0.0955**
(0.0430)
Yes
122
0.910
4
Eqn. (2)
Eqn. (3)
Eqn. (4)
0.782***
(0.0942)
0.0642**
(0.0191)
0.813***
(0.0753)
0.0311**
(0.00895)
0.0133
(0.00703)
Yes
Yes
0.693***
(0.0476)
0.0548***
(0.0116)
0.00144
(0.00781)
0.0892**
(0.0252)
Yes
120
120
120
4
0.134
0.348
0.958
4
0.0775
0.189
0.123
4
0.0818
0.217
0.132
Eqn. (1)
Eqn. (2)
***
***
Eqn. (3)
0.573
(0.0373)
0.0636*
(0.0245)
0.0516
(0.125)
0.00879
(0.0806)
0.479
(0.237)
Yes
0.710
(0.0760)
0.0672***
(0.00777)
0.202
(0.0996)
0.0396
(0.0337)
0.737***
(0.0669)
0.0628**
(0.0169)
0.180
(0.103)
Yes
Yes
121
4
0.0534
0.423
0.109
122
4
0.118
0.895
0.168
122
4
0.0997
0.878
0.147
1780
WORLD DEVELOPMENT
Table 8. Parameter values used in the scal sustainability analysis
Parameter
Symbol
Assumed value
d*
(e0F0/D0)
r0
5.0%
a2
0.693
a3
0.0548
g0
4%
b2
0.573
b3
0.0636
35%
t
Seigniorage
S/Y
5570%
Between 1 and 10
0.4%
Table 9. Simulations using parameters estimated excluding Brazil Real devaluation of 25% initial foreign currency-denominated debt as a percentage of total
debt: 55%
Panel A
Primary balance
Panel B
Public debt (% of GDP)
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
0.05
1.44
0.05
0.03
0.04
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.75
0.77
0.59
0.02
0.04
0.05
0.05
0.05
0.05
0.05
0.05
0.61
0.72
0.62
0.48
0.34
0.01
0.04
0.05
0.05
0.05
0.05
0.56
0.70
0.64
0.54
0.43
0.33
0.23
0.02
0.04
0.05
0.05
0.51
0.68
0.65
0.58
0.51
0.43
0.36
0.29
0.21
0.15
0
1
2
3
4
5
6
7
8
9
10
39.8
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
39.8
38.2
36.6
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
39.8
38.9
37.9
36.9
36.0
35.0
35.0
35.0
35.0
35.0
35.0
39.8
39.1
38.4
37.8
37.1
36.4
35.7
35.0
35.0
35.0
35.0
39.8
39.3
38.9
38.4
37.9
37.4
36.9
36.4
36.0
35.5
35.0
Panel C
Real interest rate
Panel D
Rate of real GDP growth
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
5.00
5.41
5.12
5.03
5.01
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.64
5.59
5.31
5.08
5.02
5.00
5.00
5.00
5.00
5.00
5.00
5.69
5.75
5.60
5.41
5.20
5.05
5.01
5.00
5.00
5.00
5.00
5.71
5.82
5.73
5.59
5.45
5.29
5.14
5.04
5.01
5.00
5.00
5.72
5.87
5.83
5.73
5.63
5.53
5.42
5.31
5.21
5.10
0
1
2
3
4
5
6
7
8
9
10
4.0
3.7
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.5
3.6
3.9
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.4
3.5
3.6
3.8
3.9
4.0
4.0
4.0
4.0
4.0
4.0
3.4
3.4
3.5
3.6
3.7
3.8
3.9
4.0
4.0
4.0
4.0
3.4
3.4
3.5
3.5
3.6
3.7
3.7
3.8
3.9
4.0
1781
Table 10. Simulations using parameters estimated excluding Brazil real devaluation of 25% initial foreign currency-denominated debt as a percentage of total
debt: 70%
Panel A
Primary balance
Panel B
Public debt (% of GDP)
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
0.05
1.85
0.08
0.02
0.04
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.98
1.01
0.76
0.03
0.03
0.05
0.05
0.05
0.05
0.05
0.05
0.81
0.94
0.81
0.63
0.45
0.00
0.04
0.05
0.05
0.05
0.05
0.74
0.92
0.84
0.71
0.57
0.44
0.31
0.01
0.04
0.05
0.05
0.68
0.90
0.86
0.77
0.67
0.57
0.47
0.38
0.29
0.20
0
1
2
3
4
5
6
7
8
9
10
41.1
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
41.1
39.1
37.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
41.1
39.9
38.7
37.5
36.2
35.0
35.0
35.0
35.0
35.0
35.0
41.1
40.3
39.4
38.5
37.6
36.8
35.9
35.0
35.0
35.0
35.0
41.1
40.5
39.9
39.3
38.7
38.1
37.5
36.8
36.2
35.6
35.0
Panel C
Real interest rate
Panel D
Rate of real GDP growth
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
5.00
5.53
5.16
5.04
5.01
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.82
5.74
5.39
5.11
5.03
5.01
5.00
5.00
5.00
5.00
5.00
5.87
5.95
5.77
5.52
5.25
5.07
5.02
5.00
5.00
5.00
5.00
5.90
6.04
5.93
5.76
5.57
5.37
5.18
5.05
5.01
5.00
5.00
5.92
6.10
6.05
5.94
5.80
5.67
5.53
5.40
5.26
5.13
0
1
2
3
4
5
6
7
8
9
10
4.0
3.6
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.3
3.5
3.8
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.3
3.4
3.5
3.7
3.9
4.0
4.0
4.0
4.0
4.0
4.0
3.2
3.3
3.4
3.5
3.7
3.8
3.9
4.0
4.0
4.0
4.0
3.2
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.9
Symbol
(e0F0/D0)
r0
a3
5.0%
0.053
g0
4%
b3
0.04
t
S/Y
Assumed value
35%
5570%
Between 1 and 10
0.4%
1782
WORLD DEVELOPMENT
Table 12. Simulations using parameters from BaldacciKumar and KumarWoo Real devaluation of 25% initial foreign currency-denominated debt as a
percentage of total debt: 55%
Panel A
Primary balance
Panel B
Public debt (% of GDP)
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
0.05
1.28
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.54
0.46
0.39
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.39
0.35
0.30
0.26
0.21
0.05
0.05
0.05
0.05
0.05
0.05
0.33
0.30
0.26
0.23
0.20
0.17
0.14
0.05
0.05
0.05
0.05
0.29
0.26
0.24
0.21
0.19
0.17
0.15
0.12
0.10
0.08
0
1
2
3
4
5
6
7
8
9
10
39.8
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
39.8
38.2
36.6
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
39.8
38.9
37.9
36.9
36.0
35.0
35.0
35.0
35.0
35.0
35.0
39.8
39.1
38.4
37.8
37.1
36.4
35.7
35.0
35.0
35.0
35.0
39.8
39.3
38.9
38.4
37.9
37.4
36.9
36.4
36.0
35.5
35.0
Panel C
Real interest rate
Panel D
Rate of real GDP growth
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
5.00
5.06
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.19
5.11
5.02
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.22
5.17
5.11
5.06
5.01
5.00
5.00
5.00
5.00
5.00
5.00
5.23
5.19
5.15
5.12
5.08
5.05
5.01
5.00
5.00
5.00
5.00
5.24
5.21
5.18
5.16
5.13
5.11
5.08
5.06
5.03
5.01
0
1
2
3
4
5
6
7
8
9
10
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.9
3.9
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.8
3.9
3.9
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.8
3.9
3.9
3.9
3.9
4.0
4.0
4.0
4.0
4.0
4.0
3.8
3.8
3.9
3.9
3.9
3.9
3.9
4.0
4.0
4.0
1783
Table 13. Simulations using parameters from BaldacciKumar and KumarWoo Real devaluation of 25% initial foreign currency-denominated debt as a
percentage of total debt: 70%
Panel A
Primary balance
Panel B
Public debt (% of GDP)
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
0.05
1.64
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.71
0.61
0.51
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.52
0.46
0.40
0.34
0.29
0.05
0.05
0.05
0.05
0.05
0.05
0.44
0.40
0.35
0.31
0.27
0.23
0.19
0.05
0.05
0.05
0.05
0.38
0.35
0.32
0.29
0.26
0.23
0.20
0.17
0.14
0.12
0
1
2
3
4
5
6
7
8
9
10
41.1
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
41.1
39.1
37.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
35.0
41.1
39.9
38.7
37.5
36.2
35.0
35.0
35.0
35.0
35.0
35.0
41.1
40.3
39.4
38.5
37.6
36.8
35.9
35.0
35.0
35.0
35.0
41.1
40.5
39.9
39.3
38.7
38.1
37.5
36.8
36.2
35.6
35.0
Panel C
Real interest rate
Panel D
Rate of real GDP growth
Period
t
3
10
Period
t
3
10
0
1
2
3
4
5
6
7
8
9
10
5.00
5.08
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.24
5.13
5.03
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.27
5.21
5.15
5.08
5.02
5.00
5.00
5.00
5.00
5.00
5.00
5.29
5.24
5.20
5.15
5.10
5.06
5.01
5.00
5.00
5.00
5.00
5.30
5.27
5.23
5.20
5.17
5.14
5.11
5.07
5.04
5.01
0
1
2
3
4
5
6
7
8
9
10
4.0
3.9
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.8
3.9
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.8
3.8
3.9
3.9
4.0
4.0
4.0
4.0
4.0
4.0
4.0
3.8
3.8
3.9
3.9
3.9
4.0
4.0
4.0
4.0
4.0
4.0
3.8
3.8
3.8
3.8
3.9
3.9
3.9
3.9
4.0
4.0