29-03-2012, 02:18 PM
Currency Crises and Capital Controls: A Selective Survey
Currency Crises and Capital Controls A Selective Survey.pdf (Size: 247.28 KB / Downloads: 68)
Abstract:
This survey discusses theoretical models of speculative attack and currency
crises, and reviews the empirical evidence. The paper outlines the correspondence of the
models to different cases of crisis (e.g. Latin American crises, the ERM breakdown, and
the recent Asian crisis), and points to gaps in the theoretical literature for explaining the
Asian crisis. The large economic costs resulting from the severe depreciation of Asian
currencies and general problems with macroeconomic management in the presence of
large capital flows has recently led to proposals for limiting capital flows. The paper
reviews the arguments and models for and against capital controls.
1. Introduction
The recent history of the international financial markets is characterized by numerous
currency crises. Various countries around the world have come under pressure or faced a
crisis at different points in time. The recent counts were the crises in Mexico in 1976 and
Argentina, Brazil, Peru and Mexico in the early and mid-80s, the crises in Chile and
Argentina in 1980s and ERM in 1992, then the one in Mexico in 1995. Now, in 1997 and
1998, a major part of Asia is under a financial crisis.
Economists, who are doing some catching-up work, are trying to provide analysis of
these crises. So far, most of the work focuses on three different but related areas:
(a) Theoretical analysis of the causes and effects of currency crises
Undoubtedly, most of the work on currency crisis focuses on explaining the causes and
effects of currency crisis. Earlier work by Salant and Henderson (1978), Krugman
(1979), and Flood and Garber (1984), which often are dubbed as the first-generation
models, painstakingly points out how persistent government budget deficits may lead to
capital flight and currency crisis. The crises in Chile and Argentina in the 1980s and
ERM in 1992 led to the development of second-generation models, which emphasize the
existence of multiple equilibria in the foreign exchange markets and the possibility of
having crises as self-fulfilling outcomes.
The current crisis in Asia, however, has some features that either were not present or
were not so obvious in previous crises. For example, these countries had responsible
fiscal policies of the governments, and the economies showed solid fundamentals.
Another feature of these economies is that they had been growing with impressive rates
for a long period of time. Prior to the crisis, these countries were regarded by many as the
model of growth for many developing countries. All of a sudden, they faced new
problems in the financial markets that were not expected. It is, therefore, interesting to
investigate any possible links between growth and crisis.
(b) Empirical Studies of Crises
On the empirical side, people have tried to determine whether crises can be predicted. In
particular, there has been interest on finding the relationship between a crisis and certain
variables of the economy, and whether there exist good leading indicators of a crisis.
© Policy Recommendation
As the analysis of various crises is being developed, a question that easily pops up in
people’s mind is what the government should do to avoid a crisis. Two approaches can be
suggested. First, one can focus on a particular crisis, determine its causes, and try to see
whether some of these causes can be eliminated through a change in some government
policies. For example, if one looks at the crises in Latin America, one can simply suggest
that lowering the government deficits could avoid a crisis down the road. Second, one can
2
examine the similarities between various crises and determine whether there are some
government policies that could diminish the chances of a crisis in the future.
In terms of the second approach, one can note that while all crises can be distinguished in
terms of their causes and effects, they do share two common features: (i) a fixed
exchange rate regime, and (ii) capital flight and speculative attacks. Corresponding to
these two features, two proposal have been suggested: (i) give up the fixed exchange rate
regime; (ii) capital control, i.e., strict constraints on the inflow and/or outflow of capital
across the borders of a country.
In this paper, we survey some of the more important issues related to currency crises. Our
discussion will be based the three areas described above: the analysis of causes and
effects of crises, empirical studies, and policy recommendation.
In analyzing the causes and effects of crises, we begin with the two main areas
emphasized in earlier work: the existence of persistent fiscal deficits, and existence of
multiple equilibria and self-fulfilling crises. These are discussed respectively in Sections
2 and 3.
We then turn to more recent work that examines other areas: (a) simultaneous existence
of a banking crisis and a currency crisis, so called the twin crises (Section 4); (b) herd
behavior and its relationship to capital flight and speculative attacks (Section 5); and ©
moral hazard and currency crises (Section 6).
Next, in Section 7, we turn to the work that examines the predictability of crises.
In Sections 8 to 13, we choose to discuss one policy recommendation to avoid a currency
crisis that has been proposed before: capital control. We will not mention about the other
policy suggestion: the floating of a local currency. It is because a fixed exchange rate
versus a flexible exchange rate has long been an important issue in the literature, and we
decided not to cover this area.
In Section 8, we provide some basic material about capital account control, including a
brief history of capital control. In Section 9, we present some traditional arguments for
and against capital control, while Section 10 focuses on more recent arguments. Section
11 explains an alternative recommendation: Instead of liberalizing the capital account in a
single step, as some of Asian countries did prior to the crisis, it has been suggested the
country the account should be liberalized in several steps. We will present some of the
arguments. Section 12 discusses several measures to discourage capital inflow and
outflow, so that an economy may not be so risky under the threat of capital flight. Section
13 presents some empirical studies related to capital control, while the last section
concludes.
2. Domestic Credit Creation and Currency Crises
3
For an economy under a fixed exchange rate regime, a currency crisis usually refers to a
situation in which the economy is under the pressure to give up either the prevailing
exchange rate or the regime. For the former, the economy in crisis in most cases is
required to devalue its currency by a substantial amount,1 and the exchange rate then
moves to a new, but at least temporarily fixed, level. For the latter, the alternative regime
is a flexible exchange rate one.
In the literature, many models and theories have been suggested to explain the causes and
occurrence of a currency crisis. In this section, we focus on the impacts of domestic
credit creation on the exchange rate pegged by the government. Models that use these
impacts to explain the existence of a currency crisis are sometimes called the firstgeneration
models.
2.1 A Simple Model of Exchange Rate Determination
To introduce the main features of the first-generation models, let us first lay down a
simple model of exchange rate determination. Consider a one-sector, small, open
economy. To focus on the monetary side of the economy, we assume that its real side is
characterized by full employment, with constant factor endowment and technology. The
following equilibrium conditions are used to describe the monetary side of the economy:
(1) Mt / Pt a bit
(2) Mt StRt Dt
(3) Pt Pt*St
(4) it it* St
where Mt , Pt , it , are the quantity of (high-powered) money, the general price level, and
the interest rate, respectively. Equation (1) is the money demand equation, with the
output level always at a fixed level. The two coefficients, a and b, are positive numbers.
Equation (2) gives the money supply, which consists of foreign reserves (in foreign
currency) held by the government/central bank, Rt , plus domestic credit, Dt . Variable St
is the spot exchange rate, defined as the domestic currency price of foreign currency.
Equation (3) is the purchasing power parity, where an asterisk represents a foreign
variable, while equation (4) is the interest parity condition, where a dot above a variable
represents the rate of change of that variable with respect to time. Assuming perfect
foresight, expected rate of depreciation is equal to the actual rate of depreciation. Since
we are considering a small economy, foreign variables are treated as given exogenously.
This allows us to normalize Pt* 1 and i*t 0 .
In this subsection, we assume that the spot exchange rate always adjusts to its equilibrium
level instantaneously and costlessly. With no government intervention, the amount of
1 Devaluation and appreciation of a currency are usually treated asymmetrically. While a forced devaluation
is considered as a crisis, a forced appreciation is not.
4
foreign reserve is fixed and is denoted by R0 . The domestic credit is assumed to be
increasing at an exogenously given rate of 0, i.e., Dt .
The increase in domestic credit is the main feature of most currency crisis models.
Several reasons can be used to explain why it increases, but the most common one is that
the government is running continuous deficits, and that these deficits are financed by
printing money (increase in domestic credit).2
Combining equations (1), (3) and (4) together, we have
(5) Mt aSt bSt
.
The solution to equation (5) can be found to be
(6) St Mt ,
where b / a2 and 1/ a .3 Recall that the high-powered money consists of
foreign reserves and domestic credit, (2) can be substituted into (6) to give
(7) St Dt ,
where 1/(1 R0 ) . Equation (7) is represented by line ABCE in Figure 1. The
vertical intercept of the line, A, represents /(1 R0 ) .4 Suppose that initially at time
t t0 , the quantity of domestic credit is D0 . This is represented by point B. As the
domestic credit increases, the currency is devalued and the exchange rate moves up along
the line BCE. Note that because of the constant rate of increase in Dt , Figure 1 can be
interpreted as a diagram showing the change in St with respect to time, with t
represented by the horizontal axis.
Because Dt is changing at a rate of , the value of Dt at time t is Dt D0 ( t t0 ) .
Using this equation, the exchange rate will change according to the following equation:
(8) St D0 ( t t0 ) .
2 This is what was observed in many Latin American countries, which experienced currency crises of
various degrees in the seventies and eighties.
3 To derive the solution, we conjecture that the solution of the following form: t St M .
Differentiating both sides gives St Mt
. Substitute this condition into (5), and comparing this with the
assumed solution form gives 1/ a and b / a2 .
4 To have a positive exchange rate in equation (7), we assume that 1 0 R0 .
5
Note that the exchange rate given in (8) depends on the initial value of foreign reserve.
We can consider an alternative rate that corresponds to another reserve level, conveniently
chosen to be zero, R0 0. This implies that 1. Denote the corresponding rate
by St
~
, which is described by
(9) S D ( t t )
~
t 0 0 .
The exchange rate in (9) is represented by line GHJ in Figure 1.
Suppose now that at time t t0 , in an unanticipated move, the government chooses to
raise the exchange rate from the prevailing level to a higher level. In Figure 1, this policy
can be represented by a jump of the exchange rate from point B to F instantaneously.
After the jump, the exchange rate is then fixed, as represented by horizontal line S FC.
When the exchange rate is fixed, St 0 . Equation (4) implies that i = i* = 0, so that
equation (1) reduces to
(9) Mt aS ,
i.e., the equilibrium stock of money is proportional to the given exchange rate. Therefore
when the government pegs the exchange rate, the economy accumulates foreign reserve
and thus money by running a balance of payment surplus.
2.2 Domestic Credit Creation and Currency Crisis
Mexico in 1976, and Argentina, Brazil, Peru and Mexico in the early and mid-80s
experienced various degrees of currency crises. These countries chose to peg their
currencies against foreign ones. At some points they observed capital flight and
speculative attacks on their currencies, which resulted in enormous pressure on the
central bank to devalue their currencies.
Continuous government fiscal deficits had been attributed as one major factor of these
currency crises. These deficits were financed mainly by printing money, i.e., through an
increase in the domestic credit held by the central banks. As Salant and Henderson
(1978), Krugman (1979), and Flood and Garber (1984) pointed out, there is an
inconsistency between deficit financing policy and the policy of a fixed exchange rate. In
the model described above, the money supply that equilibrates the money market is given
exogenously.5 An increase in the central bank’s domestic credit will be matched by a
drop in foreign reserve. Because the amount of foreign reserve held by the central bank is
finite, the government cannot maintain a fixed exchange rate regime indefinitely.
5 This feature is an important one in the Mundell-Fleming models of an small open economy with a fixed
exchange rate and perfect capital mobility.
6
Krugman went on to argue that a crisis occurs when the central bank’s foreign reserve
reaches a minimum level. At this point, the government will have to devalue its currency
or give up its fixed exchange rate policy.
Let us make use of the simple model of exchange rate introduced earlier to explain such
an inconsistency.6 Let us assume that at some point, the economy is represented by point
F in Figure 1, with the exchange rate fixed at S , while domestic credit is increasing at a
rate of . The increase in domestic credit means that point F shifts to the right along the
horizontal line. However, by equation (9), the equilibrium money stock is fixed, meaning
that as the economy is creating domestic credit, it is losing foreign reserve by running
balance of payment deficits. In other words, because Mt 0 , we have
(10) Rt Dt / S / S 0
.
Equation (10) suggests that the amount of foreign reserve at any time t is given by
(11) Rt R0 ( t t0 ) ,
where / S . The change in the foreign reserve over time is illustrated by schedule
ABC in Figure 2.
If this situation continues, the central bank will run out of foreign reserve at time t tx ,
where tx is obtained from (11) by setting Rt 0 :
(12) tx t0 R0 / .
At t tx , the value of domestic credit is
(13) Dx D0 ( tx t0 ) D0 SR0 .
As more domestic credit is created, both flexible exchange rates St (with positive foreign
reserve) and St
~
(with no foreign reserve) are rising. Specifically, St S when t t y ,
where from (8),
(14) t y t0 ( S D0 ) .
Similarly, S S
~
t when t tz t y , where (with 1)
(15) tz t0 ( S D0 ) .
6 The following model is based on Floor and Garber (1984).
7
It is usually assumed that t y tx .7 The value of domestic credit at t t y is equal to
(15) Dy D0 ( t y t0 ) (1 )D0 ( S ) ,
and that at t tz is equal to
(16) Dz D0 (tz t0 ) (1 )D0 (S ) .
Obviously, because the amount of foreign reserve held by the central bank is limited, the
present situation cannot continue indefinitely. Suppose that the government decides that
this situation terminates when foreign reserve hits a minimum level such as zero, and that
it gives up its fixed exchange rate policy and lets the currency float.8 When the constraint
is removed, the exchange rate becomes the same as the shadow rate.
Such a reversal of the exchange rate policy is a crisis for the economy under
consideration because the government is forced to abandon its originally set exchange
rate policy, after losing a considerable amount of foreign reserve.
Krugman (1979) points out that the fixed exchange rate regime can no longer survive
when the amount of foreign reserve held by the central bank reaches a minimum level.
This implies that the crisis occurs at t tx . So the change in foreign reserve follows
schedule ABC in Figure 2. Flood and Garber (1984), however, argue that if people know
in advance of the devaluation, speculation will occur and could force the devaluation to
happen earlier. For example, shortly before tx , speculators can purchase from the central
bank the remaining foreign reserve, denoted as R' , with an amount of domestic currency
of SR' . When the exchange rate is changed to S' S , where S' is the corresponding
shadow exchange rate, the speculators can sell the foreign reserve back to the central
bank, and get a profit of ( S'S )R' . This profit may not be much if R' is small, but the
profit rate is very high considering the short duration of time. So devaluation will not
occur at t tx , but earlier. However, the same argument can be used to say that
devaluation will not occur just earlier than tx , but even earlier. As a result, at any time
when the shadow exchange rate is higher than the pegged rate, one can argue that
devaluation will occur slightly earlier. This means that the crisis will occur when t tz ,
and the change in central bank’s foreign reserve over time is described by ABE in Figure
2. Furthermore, the above analysis implies that there is no discrete jump in the exchange
rate.
7 This is the case considered in most, if not all, papers analyzing the present issue.
8 The same analysis applies even if the minimum level is positive, but it is usually assumed that this amount
is less than Ry , meaning that the government will not consider giving up the regime before t t z .
8
However, the path ABE of foreign reserve requires perfect information, costless
exchange transactions, unlimited resources owned by the speculators, and a clear policy
in terms of when the fixed exchange rate is given up. It should also be pointed out that
between t y and tz , the currency is overvalued (relatively to the existing stock of foreign
reserve held by the central bank). Therefore, if the determination of the government in
defending the pegged exchange rate is unknown to the public and people do not know the
minimum amount of foreign reserve below which the government will give up the
exchange rate regime, then it is possible that some investors may start moving the money
out of the country at or shortly after t y . These may be people who are most pessimistic
about the economy or the resolve of the government. This leads to loss of foreign reserve
of the government, making devaluation more likely. Very soon, more people may take
similar action, and speculative attack occurs. Finally, the capital flight and speculation
become so overwhelming for the government to defend the exchange rate anymore. So a
possible adjustment path of foreign reserve is described by schedule AFG in Figure 2.