When is it Optimal to Abandon a Fixed Exchange Rate?
ABSTRACT The influential Krugman-Flood-Garber (KFG) model of balance of payment crises assumes that a fixed exchange rate is abandoned if and only if international reserves reach a critical threshold value. From a positive standpoint, the KFG rule is at odds with many episodes in which the central bank has plenty of international reserves at the time of abandonment. We study the optimal exit policy and show that from a normative standpoint, the KFG rule is generally suboptimal. We consider a model in which the fixed exchange rate regime has become unsustainable due to an unexpected increase in government spending. We show that when there are no exit costs, it is optimal to abandon immediately. When there are exit costs, the optimal abandonment time is a decreasing function of the size of the fiscal shock. For large fiscal shocks, immediate abandonment is optimal. Our model is consistent with evidence suggesting that many countries exit fixed exchange rate regimes with still plenty of international reserves in the central bank's vault. Copyright © 2008 The Review of Economic Studies Limited.
- SourceAvailable from: Mei Li[Show abstract] [Hide abstract]
ABSTRACT: This paper examines the optimal appreciation path of an under-valued currency in the presence of speculative capital inflows that are endogenously affected by the appreciation path. A central bank decides the optimal appreciation path based on three factors: (i) Misalignment costs associated with the gap between the actual exchange rate and the fundamental exchange rate, (ii) short-term adjustment costs due to fast appreciation, and (iii) capital losses due to speculative capital inflows. We examine two cases in which speculators do and do not face liquidity shocks. We show that, in the case without liquidity shocks, the central bank should appreciate quickly to discourage speculative capital, and should appreciate more quickly in initial periods than in later periods. In the case with liquidity shocks, the central bank should pre-commit to a slow appreciation path to discourage speculative capital. The central bank should appreciate slowest when the probability of liquidity shocks takes middle values. If the central bank cannot commit and can only take a discretionary policy, appreciation should be faster.Canadian Journal of Economics/Revue Canadienne d`Economique 01/2009; · 0.61 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: This study is aimed at the identification of the fa ctors that influence the conditions of exit from a fixed or intermediate exchange rate regime, to a more flexible arrangement. More specifically, we try to identify the economic varia bles that exercise a significant influence on the probability of an orderly exit. In order to do this, we employ a binary probit estimation procedure, applied to a large number of developed a nd emerging economies who have exited from fixed and intermediate exchange rate arrangements since 1980. We find that the significant variables are those representing the gr owth rate, the evolution of domestic credit, the interest rate, the duration of the initial exch ange rate regime and the incidence of exits over the years preceding and following the year dur ing which the exit episode under consideration takes place. These results point stro ngly towards possible contagion and duration-dependence effects, but also to a strong i nfluence of certain economic fundamentals.
When Is It Optimal to Abandon a Fixed
Sergio Rebelo†and Carlos A. Végh‡
The influential Krugman-Flood-Garber (KFG) model of balance of pay-
ment crises assumes that a fixed exchange rate is abandoned if and only
if international reserves reach a critical threshold value. From a positive
standpoint, the KFG rule is at odds with many episodes in which the cen-
tral bank has plenty of international reserves at the time of abandonment.
We study the optimal exit policy and show that, from a normative stand-
point, the KFG rule is suboptimal. We consider a model in which the
fixed exchange rate regime has become unsustainable due to an unexpected
increase in government spending. We show that, when there are no exit
costs, it is optimal to abandon immediately. When there are exit costs,
the optimal abandonment time is a decreasing function of the size of the
fiscal shock. For large fiscal shocks immediate abandonment is optimal.
Our model is consistent with the evidence that many countries exit fixed
exchange rate regimes with plenty of international reserves in the central
J.E.L. Classification: F31
Keywords: Currency crisis, speculative attacks, optimal policy, fixed
∗We thank Linda Goldberg, Pierre-Oliver Gourinchas, Michael Klein, Assaf Razin, seminar
participants at the NBER IFM Program Meeting, Federal Reserve Bank of New York, Princeton,
and University of Pennsylvania for their comments. We are particularly grateful to Bob Flood,
Fabrizio Zilibotti, and two anonymous referees for their suggestions. Financial support from the
National Science Foundation and UCLA Academic Senate is gratefully acknowledged.
†Northwestern University, NBER and CEPR.
‡University of Maryland, UCLA, and NBER.
Consider an open economy with a fixed exchange rate that suffers an unexpected
fiscal shock. This shock consists of an increase in government expenditures that
has to be financed with seignorage. When, if at all, should the fixed exchange rate
regime be abandoned? Further, suppose that, with some probability, a future fiscal
reform or a financial package from the International Monetary Fund (IMF) can
restore the sustainability of the fixed exchange rate regime. For how long should
policy makers wait for this scenario to materialize?
The decision to exit a fixed exchange rate regime is one of the most important
policy questions in open-economy macroeconomics. This importance was recently
illustrated by Argentina’s abandonment in early 2002 of its 10-year old “Convert-
ibility plan” that had tied the peso to the U.S. dollar at a one-to-one rate since
April 1991. Most analysts agree that fixing the exchange rate was an effective
strategy to eliminate runaway inflation. However, in the mid 1990s, as the fiscal
situation began to deteriorate, the question of whether Argentina should abandon
the fixed exchange rate began to surface with increasing frequency.1The IMF
rescue packages in December 2000 and August 2001 bought Argentina some time.
But, in the end, the fixed exchange rate was abandoned in January 2002.
Economic theory offers surprisingly little guidance as to the optimal time to
exit a fixed exchange rate regime. The dominant paradigm for understanding
this exit is the model proposed by Krugman (1979) and Flood and Garber (1984),
which we refer to as the KFG model.2This model makes two central assumptions.
The first assumption is that the root cause of the eventual abandonment of the
1See Mussa (2002) for a detailed analysis of Argentina’s lax fiscal policy during the mid 1990s.
2The original KFG model does not have microfoundations. However, several authors have
extended the KFG framework to models populated by optimizing representative agents. See,
for example, Obstfeld (1986a), Calvo (1987), Drazen and Helpman (1987), Wijnbergen (1991),
Burnside, Eichenbaum, and Rebelo (2001), and Lahiri and Végh (2003).
fixed exchange rate is an unsustainable fiscal policy. The second assumption is
that the central bank follows an ad-hoc exit rule whereby the fixed exchange rate
regime is abandoned only when the central bank exhausts its foreign exchange
reserves and its ability to borrow.
To study the empirical plausibility of these two hypotheses, we collect in Table
1 data for 51 episodes in which regimes with fixed exchange rates were abandoned.
These abandonments are often called “currency crises.” Our episodes were selected
from an updated version of Kaminsky and Reinhart’s (1999) list of crisis episodes
according to the criteria outlined in Appendix 7.1. Table 1 reports the change
in the exchange rate in the month in which the fixed exchange rate regime was
abandoned, as well as the change in the exchange rate in the 12 months before and
after the abandonment.3Table 1 also reports the rate of change in real government
spending in the three years prior to the crisis and the reserve losses that occurred
in the 12 months prior to the crisis.
We view the fiscal data in Table 1 as lending empirical support to the first KFG
assumption. There were increases in real government spending in the three years
prior to the abandonment of the peg in 80 percent (37 out of 46) of the episodes
for which we have fiscal data. Therefore, fiscal shocks are plausible suspects as
the root cause of the decision to abandon a fixed exchange rate.
We think that the reserve-loss data in Table 1 implies that the second KFG
assumption is empirically implausible. While the KFG model is not explicit about
the critical lower bound for international reserves (is it zero? is it negative?), it
is clearly in the spirit of the model that the monetary authority holds on to
the peg for as long as it can. So we would expect to see central banks exhaust
their international reserves before the fixed exchange rate is abandoned. Figure
3In some of the episodes included in Table 1 the exchange rate was not literally fixed, but
followed a crawling peg or fluctuated within a narrow band.
1 depicts a histogram of the fraction of initial reserves lost during the 12 months
prior to the crisis. In 12 out of 51 episodes countries have non-positive reserve
losses (i.e., they gained reserves). In 38 out of the 51 episodes (or roughly 75
percent), reserve losses were less than 40 percent of initial reserves. While there
were cases in which the monetary authority was willing to lose a large amount
of reserves before devaluing, in most cases the peg was abandoned with plenty
of ammunition left in the central bank’s coffers. In other words, the monetary
authority chooses to devalue as opposed to being forced to devalue by literally
exhausting its reserves and its ability to borrow. We conclude that the KFG exit
rule, a critical component of the KFG model, is inconsistent with the empirical
behavior of reserves in countries that have abandoned fixed exchange rates. In
addition, and given that it assumes an exogenous exit rule, the KFG model is
unsuitable for understanding the decision to abandon a fixed exchange rate regime.
In this paper, we study the optimal exit from a fixed exchange rate regime.4
Our analysis is in the spirit of the literature on optimal monetary and fiscal
policy pioneered by Lucas and Stokey (1983). We argue that the assumption that
central bankers choose the optimal time to abandon the peg generates empirical
implications that are more plausible than those associated with the KFG exit
Ouranalysis is based on a standard cash-in-advance small-open-economy model,
extended to incorporate rational policy makers. We first consider the case where
4Authors such as Buiter (1987), Flood, Garber, and Kramer (1996), Lahiri and Végh (2003),
and Flood and Jeanne (2005) have studied whether it is feasible and/or optimal to delay the
abandonment of the fixed exchange rate regime (i.e., “defend the peg”) by borrowing or by
raising interest rates. While these models give the central bank a more active role than in the
original KFG model, they continue to assume that abandonment of the peg is governed by the
5Second-generation models of speculative attacks introduce an optimizing central banker
(Obstfeld (1986b, 1996)). However, they assume that currency crises do not have a fiscal origin.
Instead, the crises are caused by the incentive to increase output via unexpected inflation in
Barro-Gordon type formulations.
there are no costs of abandoning the peg. In this case it is optimal to abandon the
peg as soon as the fiscal shock occurs and without incurring any reserve losses.
This policy is optimal independently of the level of international reserves and of
whether the central bank faces a borrowing constraint.
We then consider the case in which there are costs of abandoning the peg.
These exit costs can reflect, for instance, output losses or the cost of bailing out
the banking system.6We choose to abstract from the source of these costs and
simply assume that devaluing entails some fiscal and social cost. In this case
there is a certain threshold value for the fiscal shock beyond which it is optimal
to abandon immediately, incurring no reserve losses. For fiscal shocks lower than
this threshold, the optimal exit time is a decreasing function of the size of the
fiscal shock. In other words, the smaller the fiscal shock, the longer is the optimal
Intuitively, the optimal exit time results from the trade off between two factors.
For a given fiscal shock, delaying the abandonment of the peg reduces the present
discounted value of the cost of abandoning. However, a longer delay requires a
permanently higher level of inflation once the peg is abandoned. This increase
in the post-abandonment rate of inflation produces a larger intertemporal distor-
tion in consumption decisions. For large fiscal shocks, the cost of delaying (i.e.,
the larger intertemporal distortion) dominates because the gain from delaying is
bounded by the economy’s resources.
Some back-of-the-envelope calculations — based on our model, the fiscal data
in Table 1, and on empirical estimates of the cost of balance of payment crises —
suggest that an immediate abandonment should be at least as common as delayed
abandonment. Hence, unlike the KFG model, our model is capable of explaining
6See Kaminsky and Reinhart (1999) and Gupta, Mishra, and Sahay (2003) for evidence on
output and banking crises during currency crises.