Tango with the Gringo:
The hard peg and real misalignment in Argentina
Enrique Alberola, Humberto López and Luis Servén
Abstract
Between 1990 and 2001 the Argentine peso appreciated by 80 percent in real terms,
and its overvaluation has been singled out as one of the main suspects in the debate
on the causes of the Argentina collapse of late 2001. This paper assesses the degree
of real misalignment in Argentina over the Convertibility period using a model in
which the equilibrium real exchange rate is defined as the value consistent with (i) a
balance of payments position where any current account imbalance is financed by a
sustainable flow of international capital (external equilibrium), and (ii) traded /
nontraded sector productivity differentials (internal equilibrium). Empirical
implementation of the model suggests that the initial real appreciation of the peso,
between 1990 and 1993, was consistent with the productivity increases that
Argentina enjoyed following the stabilization of the economy after the
hyperinflation of the late 1980s. But after 1996 a widening gap opened between the
observed real exchange rate and that consistent with a sustainable net foreign asset
position. Our estimates indicate that in 2001 the peso was overvalued by over 50
percent. The model allows us to assess how much of the overvaluation resulted from
Argentina's inadequate choice of anchor currency and how much from a divergence
of fundamentals between the U.S. and Argentina, ultimately due to the maintenance
of policies inconsistent with the peg. We find that both factors played a role in the
overvaluation accumulated between 1977 and 2001 that preceded the collapse of the
Convertibility regime.
JEL codes: F31, F41
World Bank Policy Research Working Paper 3322, June 2004
The Policy Research Working Paper Series disseminates the findings of work in progress to
encourage the exchange of ideas about development issues. An objective of the series is to get the
findings out quickly, even if the presentations are less than fully polished. The papers carry the
names of the authors and should be cited accordingly. The findings, interpretations, and conclusions
expressed in this paper are entirely those of the authors. They do not necessarily represent the view
of the World Bank, its Executive Directors, or the countries they represent. Policy Research Working
Papers are available online at http://econ.worldbank.org.
Bank of Spain, World Bank (LCSPE) and World Bank (LCRCE) respectively. For
their helpful comments we thank, without implication, Bill Cline, Ali Leila, an
anonymous referee, and seminar participants at the World Bank, the Bank of Spain
and the LACEA 2003 Annual Meetings. Patricia Macchi and Maria Salgado
provided able research assistance.
I. Introduction
In late 2001, after many months of recession and mounting financial turmoil, Argentina
was forced to abandon its currency board. The demise of the decade-old dollar peg was
accompanied by a financial crisis of unprecedented severity, and followed by a dramatic
devaluation of the peso. From the 1 peso = 1 U.S. dollar exchange rate of the
Convertibility decade, the Argentine currency depreciated to nearly four pesos per dollar
by mid 2002, before recovering toward three pesos per dollar at the end of that year and
throughout 2003.
The causes of the financial collapse, as well as its implications for macroeconomic and,
especially, exchange rate policy, have attracted considerable attention. A number of
observers have underscored the role of fiscal imbalances in precipitating the crisis, while
others blame self-fulfilling default expectations in international financial markets.1 In this
debate, the trajectory of the real exchange rate of the peso has played a prominent role.
Between 1990 and 2001, Argentina's real effective exchange rate appreciated by almost
80 percent. In spite of the productivity gains that Argentina probably achieved over the
1990s a result of the stabilization and reform of the economy after the hyperinflation of
the late 1980s -- a majority of observers concur that the real exchange rate had become
overvalued by 2001, especially after the devaluation of the Brazilian real in 1999 and
coinciding with the persistent appreciation of the U.S. dollar at the end of the decade.
However, the extent of the peso overvaluation, as well as its significance in triggering off
the crisis, remain disputed.2 Furthermore, assessments of peso misalignment advanced by
various observers are typically based on rough comparisons between the actual and past
values of some measure of the real exchange rate that is, they implicitly rely on a PPP-
like assumption that the equilibrium real exchange rate is constant over time.
In this paper we go beyond that crude approach and develop a formal analytical model
for the study of the equilibrium real exchange rate. We implement the model empirically
to assess the misalignment of the Argentine peso. Thus, the paper is most closely related
to the literature on real misalignment and equilibrium exchange rates.3
1 Fiscal imbalances are particularly stressed by Mussa (2002) and Teijeiro (2002). On the
role of investor expectations see Stiglitz (2002) and Sachs (2002).
2 The view that by 2000-2001 the peso was significantly overvalued has been stated,
among others, by Calvo et. al (2002), Corden (2002), Hausmann and Velasco (2002),
Roubini (2001), Rodrik (2002), Sachs (2002) and Stiglitz (2002). Most of these authors
do not provide a specific estimate of the magnitude of the overvaluation, although Sachs
(2002), for example, places it in the 30-40 percent range. A lone dissenter is Schuler
(2002).
3 See Williamson (1994), Wren-Lewis and Driver (1998), Baffes et al. (1999),
MacDonald and Stein (1999), or MacDonald (2000) for recent contributions on
equilibrium exchange rates.
Our approach extends the literature along several dimensions. First, the model we
construct is capable of encompassing two leading approaches to real exchange rate
determination.4 On the one hand, it assigns a role to productivity differentials, along the
lines of the Balassa-Samuelson hypothesis. In our framework productivity differentials
explain the evolution of the price of non-tradables relative to tradables. Since such price
determines the allocation of resources within the economy, this captures what we may
label the `internal equilibrium' of the economy. On the other hand, the model also assigns
a role to current account sustainability (as in Frenkel and Mussa 1985), which can be
viewed as the `external equilibrium' of the economy.
Second, to assess empirically the degree of real misalignment, we rely on an orthogonal
decomposition of the observed exchange rate series into a permanent component, which
captures long-run movements, and is therefore identified with the equilibrium real
exchange rate, and a transitory component, which captures equilibrium-reverting
movements and thus corresponds to the short-run deviation from equilibrium or
misalignment.5
Third, our framework allows us to go beyond measuring misalignment and assess the
factors behind the observed real exchange rate trajectory. On the one hand, we gauge the
roles of those domestic factors related to the internal and external equilibrium
respectively. This permits us to assess the contribution of structural reforms and spending
decisions to the real exchange rate path. On the other hand, we also examine how the peg
to the U.S. dollar contributed to the misalignment of the peso. The fact that trade with the
U.S. accounted for only a small fraction (around 15 percent) of Argentina's total trade,
and an even smaller share of its GNP (less than 3 percent) has led some observers to
conclude that the U.S. dollar was the wrong anchor currency for Argentina, and that this
inadequate choice is largely to blame for the misalignment of the peso at the end of the
1990s.6 Our empirical calculations allow us to evaluate this assertion.
The rest of the paper is organized as follows. Section II offers a brief overview of the
main existing approaches to the equilibrium real exchange rate. Section III lays out the
analytical model, and section IV describes the empirical strategy for its implementation.
Section V reports the empirical results, and assesses the estimated equilibrium real
exchange rate as well as its main determinants. Section VI goes one step beyond and
examines the impact of the recent evolution of the US dollar in the overall misalignment
of the peso. Finally, section VII concludes.
4 The model follows along the lines of Alberola et al (2002).
5 Huizinga (1987) and Clarida and Gali (1994) also interpret the permanent component of
the real exchange rate as a measure of equilibrium.
6 See Alesina, Barro and Tenreyro (2002), whose analysis shows that on the basis of
observed macroeconomic comovement the Euro would have been a less-inadequate
choice of anchor for Argentina's hard peg.
2
II. Exchange rate appreciation and misalignment
After decades of monetary mismanagement, the hard peg to the dollar that Argentina
adopted under the Convertibility regime in 1991 quickly led to nominal stability and
financial deepening.7 In the process, however, the peso experienced a large real
appreciation. Figure 1 shows that throughout the Convertibility decade the real effective
exchange rate (REER) of the peso remained consistently above8 its historical average
over the period 1960-2000, by a margin that ranged between 15 and 50 percent. This
persistent appreciation has led many observers to conclude that the peso was overvalued,
particularly at the end of the decade.
There is broad consensus among economists that large real overvaluations have negative
macroeconomic consequences. Their correction requires at best the adoption of painful
adjustment programs, and may lead at worst to an abrupt currency collapse and a major
financial crisis. In addition to the Argentine episode, there are numerous other examples
of abrupt currency collapses in recent memory, including Mexico (1994), East Asia
(1997), and Brazil (1999). In each of these episodes, a few observers had warned of the
impending collapse in view of what they perceived as an unsustainable overvaluation of
the currency.9
More generally, analyzing the evidence from a sample of 93 countries over 1960-94,
Goldfajn and Valdes (1999) conclude that once a currency appreciates significantly, a
smooth return to equilibrium becomes highly unlikely. In their sample, 85 percent of the
cases in which currencies reached a misalignment of 25 percent or more ended abruptly
with a collapse of the nominal exchange rate. This is particularly relevant for the
Argentine case, in which much of the policy discussion in the run up to the crisis, and
even after it, has centered on the question of whether the authorities might have
succeeded in salvaging the Convertibility regime, achieving whatever real depreciation
was necessary through nominal deflation.10 The answer to this question therefore
depends, among other things, on the actual degree of overvaluation of the peso in the
final years of Convertibility.
Assessing the degree of misalignment, however, is not straightforward. The most popular
method relies on the purchasing power parity (PPP) doctrine. In its absolute form, it
7 Financial deepening under Convertibility was accompanied by an increasing degree of
financial dollarization which, according to many observers, made the financial system
vulnerable and helped precipitate the financial crash. See for example De la Torre and
Schmukler (2003).
8Throughout we define the real exchange rate so that an increase is an appreciation.
9 Dornbusch and Werner (1994) and Werner (1997) examine the case of Mexico, and
Chinn (1998) that of East Asia.
10 In this vein, it has been argued that a course of action preferable to the one that
eventually emerged would have involved full dollarization, letting nominal price
adjustment realign subsequently the real exchange rate. The feasibility of such strategy
has been widely debated; see Roubini (2001) and De la Torre and Schmukler (2003).
3
maintains that after conversion into a common numeraire a basket of goods should cost
the same at home and abroad; i.e. the equilibrium real exchange rate should equal 1. A
weaker version of PPP, sometimes referred to as relative PPP, holds that changes in
nominal effective exchange rates must compensate for the inflation differential between a
country and its trading partners, implying that the equilibrium real exchange rate is
constant over time (although it may not equal 1).
Based on the relative PPP notion of the equilibrium real exchange rate, which in Figure 1
is measured by its long-term average, the peso would have been undervalued in 1990, but
would have become increasingly overvalued after the introduction of the Convertibility
Law and the currency board in 1991. The PPP-based overvaluation peaked initially in
1993 at about 40 percent, declined later through 1996 to about 14 percent, and rose again
to exceed its average historical level by almost 50 percent in 2001. Thus the PPP
approach would point towards a persistent and significant overvaluation over the whole
decade.
However, the PPP approach to exchange rate misalignment can also be misleading.
Rogoff (1996) presents evidence supporting a positive relationship between income and
price levels. That is, a given a sum of dollars converted into a less developed economy
currency at the nominal exchange rate will buy a larger basket of commodities and
services than can be bought in the United States. Also, a considerable literature finds that
deviations of exchange rates from their PPP equilibrium (both absolute and relative) are
large, fairly persistent and with little tendency to revert towards a fixed long run
equilibrium level.11 There are two well-documented factors that can contribute to the
empirical failure of PPP. The first one is the existence of non-traded goods. The second is
imperfect substitutability between traded goods produced in different countries. We shall
take them in turn.
If the exchange rate is viewed as the relative price of traded goods, then the existence of
non-traded goods is irrelevant. But if the exchange rate is viewed as an asset price (i.e.,
the price of one national money in terms of another) then the appropriate price index has
to be broader, given that the value of money is given by the reciprocal of the general
price level, which includes both traded and non-traded goods and services.
The introduction of non-traded goods poses a challenge for the PPP hypothesis, if for
example there are productivity biases as noted by Balassa (1964) and Samuelson (1964).
The Balassa-Samuelson hypothesis postulates that under fairly general assumptions non-
traded goods prices will be lower in the country with lower productivity in the traded
goods sector (usually the poorer country). As a result, a basket of traded and non-traded
goods will be cheaper in the lower wage (i.e. the poorer) country. By the same token,
increases in relative productivity will be associated with an increase in relative prices,
that is, an appreciation of the real exchange rate.
11 On the persistence of deviations from PPP, see for example Krugman (1978), Frenkel
(1981), Rogoff (1996) and Cashin and McDermott (2003).
4
The same result can arise without international productivity differences if there are
instead demand asymmetries or endowment biases. For example, Genberg (1978) argues
that if the income elasticity of demand for non-traded goods is greater than unity, under
the assumption of unbiased productivity growth the relative price of non-traded goods
will rise with income. Bhagwati (1984) proposes instead a model which predicts lower
price levels in poorer countries if they are labor-abundant relative to rich countries. Under
the assumption of similar consumption tastes across countries (which rules out preference
biases towards non-tradable goods that could raise wages enough to offset the
endowment disparity), wages and the price of labor-intensive non-tradable goods will be
relatively lower in poorer countries.
Thus, in addition to productivity asymmetries, demand asymmetries and factor
endowments may also drive the relative price of non-tradables. In reality, the empirical
failure of PPP may be due, at least in part, to differences in endowments (poorer
countries are usually more labor abundant than rich countries) and demand biases
(services usually represent a larger share of the consumption basket in richer economies).
But productivity biases still seem an integral part of the model: richer countries would be
so because of higher labor productivity. Empirical evidence pointing in this direction is
reported by De Gregorio et al. (1994) or Alberola and Tyrväinen (1998), who find a
strong link between changes in the price of non-tradables and changes in productivity
differentials between tradables and non-tradables in OECD countries, as predicted by the
Balassa-Samuelson hypothesis.
For the Argentine case, Figure 2 plots the real exchange rate and two series that capture
international productivity differentials over 1990-2000.12 The first is overall labor
productivity in Argentina relative to its main trading partners. The second is the ratio of
non-tradable prices to tradable prices, again measured relative to Argentina's main
trading partners. According to the Balassa-Samuelson argument, this relative price
differential between countries should be driven by their respective tradable-nontradable
productivity differentials.
Inspection of the figure reveals a remarkably similar pattern for both measures of
productivity differentials. Regardless of which one is used, casual empiricism would
seem to support the Balassa-Samuelson hypothesis, as Figure 2 suggests a very close
correlation between the productivity series and the real exchange rate. During the first
years of the currency board, following the stabilization of the economy after the
hyperinflation of the late 1980s, the graph shows that Argentina enjoyed significant
productivity increases relative to her trading partners, which may help explain the sharp
real appreciation of the peso between 1990 and 1993. Relative productivity stalled after
1993 and this process was accompanied by a parallel correction in the REER until 1996.
After 1996 productivity starts recovering (according to relative prices until 2000 and
according to relative labor productivity until 1998) and the real exchange rate appreciates
again, although during this period the appreciation outpaces the relative productivity
12In the figure, the two relative productivity series have been rescaled (to zero mean) for
ease of comparison.
5
gains. All in all, however, the real exchange rate seems to have evolved roughly in line
with productivity differentials, at least until 1995.
Aside from non-traded goods, the other main factor that can lead to the breakdown of
PPP is the lack of perfect substitution between domestic and foreign traded goods. The
Balassa-Samuelson hypothesis is based on the assumption of perfect substitutability
between traded goods produced at home and abroad, but in reality traded goods produced
in different countries are not identical, and the relative price of foreign and domestic
tradables (i.e. the terms of trade) shows significant time variation. This in turn may result
in substantial changes in the real exchange rate over time.
To account for this variability of exchange rates, Mussa (1984) proposed a model of
exchange rate determination, which can be summarized as follows. On the one hand,
changes in the exchange rate affect the trade balance, the current account balance and its
stock counterpart, the country's net foreign asset position.13 Under this approach to
exchange rate determination, the exchange rate must be consistent with an external
position where any current account imbalance is financed by a sustainable flow of
international capital, which in turn is determined by the desired stock of foreign asset and
liabilities among nations. A country may run current account deficits (surpluses) and
therefore decumulate (accumulate) assets in the adjustment process towards its desired
stock. Such imbalances would be due to cross-country differences in the propensities to
save and invest, which are assumed independent from exchange market developments. In
the long-run, however, when assets are at their desired levels, the current account and the
exchange rate should be consistent with a stable net foreign assets-output ratio (see the
empirical applications in Broner et al. (1998), Alberola et al. (1999), and Alberola and
Lopez (2000)).
Figure 3 reviews the evolution of the stock of Net Foreign Assets (NFA) relative to GNP
for Argentina over 1990-2001.14 Inspection of Figure 3 reveals that during the 1990s
Argentina witnessed a steady increase in its foreign liabilities: the NFA stock fell from
about -15 percent of GNP in the early 1990s to close to -40 percent of GNP at the end of
the decade. Figure 3 also indicates that the decline in the country's net asset position
accelerated after the Asian crisis of 1997. Before 1997 Argentina was adding liabilities at
a rate of about 1 percent of GNP per year, but after 1997 it started adding about 4 percent
of GNP per year. If the real exchange rate had adjusted to this reduction in wealth, it
should have depreciated, leading to a trade balance surplus to compensate for the interest
payments associated with the increasing liabilities. Such adjustment did obviously not
take place, and therefore the disparate evolution of net foreign asset holdings and the
observed real exchange rate in the late 1990s suggests that a gap opened between the
actual exchange rate and the one that would have been consistent with a sustainable NFA
/ GNP ratio.
13Changes in the net foreign asset position, and therefore in the net wealth of the country,
also affect the real exchange rate, through the wealth effect on consumption.
14 See Section IV for a description of the procedure used in the computation of the net
foreign asset position.
6
To examine these issues more rigorously, the next section develops a theoretical
framework that incorporates the two ingredients reviewed above, productivity
differentials and asset equilibrium. Importantly, the approach combines both theoretical
strands without restricting the exchange rate to be determined by either one of them
alone.
III. The economic model
Consider two economies, each producing two goods: one tradable (denoted by the
subscript T in what follows) and one non-tradable (N). The (log) real exchange rate (q) is
defined as the relative price of two consumption baskets at home and abroad:
q = p-(s + p*), (1)
where s is the (log) nominal exchange rate, defined as the price of foreign currency in
terms of domestic currency, and p and p* are the (log) domestic and foreign price indices
respectively. Throughout we use asterisks to denote foreign variables. An increase in q
represents an appreciation of the real exchange rate.
The consumer price index (CPI) for each country is a weighted-average of the exportable,
non-tradable, and importable prices, all expressed in the currency of the respective
country:
p=(1-N- T)pT+ Np N + T (s+p T*) (2)
p*=(1-*N-* T)p*T+* Np* N +* T (p T -s),
where the s are the weights of the respective goods in the consumer basket. Substituting
these expressions in (1), assuming that N=* N, and rearranging terms we obtain
q = (1- T -* T)q X+ N q I, (3)
where qX=pT-(s+p*T) is the relative price of domestic tradables in terms of foreign
tradables, and qI=(pN-pT)-(p*N-p*T) is the price of non-tradables relative to tradables
across countries. Here qX captures the competitiveness of the economy, and as we shall
explain shortly it is related to the evolution of the foreign asset position. Since sustainable
capital flows eventually lead to the desired stocks of assets and liabilities across
countries, the equilibrium level of qX is associated with the external equilibrium of the
economy. On the other hand, the cross-country differential in relative tradable-
nontradable prices qI is related to productivity differentials. Since these prices determine
the allocation of resources across sectors in a given country, the equilibrium level of qI
can be associated to the internal equilibrium of the economy.
The equilibrium exchange rate is attained when both qX and qI are at their equilibrium
values, and thus follows from internal and external equilibrium:
7
q = (1-T -T )qX +NqI
* (4)
with the bars denoting equilibrium values. We next characterize the internal and external
equilibrium of the economy.
III.1 Internal equilibrium
The differential behaviour of sectoral relative prices between countries determines the
evolution of the internal real exchange rate. Sectoral prices are in turn related to the
evolution of sectoral productivity. These notions can be illustrated using a simple model
with two production factors, labor (L) and capital (K). Output in each sector is
determined by a Cobb-Douglas production technology:
YN=ANLN KN 1- (5)
YT=ATLT KT ,
1-
where 0<,<1 represent the intensity of labor in each sector. Labor is perfectly mobile
between sectors (but not across countries), implying nominal wage equalization:
WT = WN = W. (6)
Labor is paid the value of its marginal product Yi/Li=W/Pi. Under Cobb-Douglas
technology the ratio of marginal productivities is proportional to the ratio of average
productivities:
YT /LT = YT /LT . (7)
YN /LN YN /LN
From (7) it follows that the (log) sectoral price differential is equal to the labor
productivity differentials plus a drift capturing the relative intensity of labor. Expressing
with lower case the natural logarithms of sectoral labor productivities, (7) reduces to
pN - pT =log(/)+(yT-yN). (8)
Neglecting constant terms and denoting n = (yT-yN)- (y*T - y*N), the internal equilibrium
exchange rate is just:
qI = n , (9)
III.2 External equilibrium
Portfolio models of real exchange rate determination (Mussa 1984) focus on asset
equilibrium, as defined by the attainment of agents' desired foreign asset stock. Over
time, the accumulation of net foreign assets (F) is given by the current account balance
8
(CA), which equals the trade balance (XN), plus the net income that residents receive (or
pay) on F:
F=CA=XN+i*F (10)
where i* is the international interest rate, which is assumed given. It will be more
convenient to focus on the trajectory of the foreign asset stock relative to GNP, which can
be written
f=ca=xn+(i*-g) f (11)
where f and xn denote the ratios to GNP of the respective uppercase variables, and g is
the rate of GNP growth. If the Marshall-Lerner condition holds, an increase in the relative
price of domestic tradables qX shifts consumption toward foreign tradables and worsens
the trade balance. Consistent with this interpretation, it is plausible to assume that the
trade balance as a percentage of GNP (xn) is given by:
xn=-qx, >0. (12)
The capital account deficit reflects the desired rate of accumulation of net foreign assets
by the home country, which is assumed to depend on the divergence between the current
level of assets as a percentage of GNP (f) and the desired equilibrium level ( f ), itself
determined by exogenous factors such as saving preferences and demographics which
will not be modelled here
f = a( f - f ) a>0 (13)
Equation (13) indicates that if the actual net foreign asset position is below its desired
level, agents will accumulate assets to reach the target; conversely, if f is greater than f
agents will be reduce their asset holdings until they reach f .
Equating (13) and (11) after using (12), and solving for qx we get:
qx =[a/] (f- f )+ [(i*-g) / ] f (14)
Equation (14) shows that the external real exchange rate depends on (i) the divergence
between current and equilibrium asset holdings; and (ii) the current stock of net foreign
assets f. Defining the equilibrium external real exchange rate qX as that consistent with f
= f (i.e. the exchange rate consistent with asset holdings at their equilibrium level) it
follows that
qX = [(i*-g)/] f , (15)
9
III.3 The equilibrium real exchange rate
Substituting (9) and (15) into (3), we get the final expression for the equilibrium
exchange rate:
q =[(1-T-* T)(i*-g)/] f + N n . (16)
Observe that in principle both (1-T-* )(i*-g)/ and N should be positive, and thus the
T
equilibrium real exchange rate appreciates in response to both a higher long-run asset
stock and a higher relative productivity differential.15
IV. Empirical issues
The theoretical model outlined above identifies two fundamentals for the evolution of the
real exchange rate: the level of net foreign assets f and relative sectoral productivity (n).
In this framework, a suitable empirical model for estimation would be
q=0 + F f + N n + u (17)
At this stage, one would be tempted to think that if a long-run cointegration relationship
between the real exchange rate and its fundamentals can be found, it will automatically
yield an estimate of the equilibrium rate. However, for this to be true, one would have to
observe the equilibrium levels of the fundamentals, and then apply a cointegration
analysis to them. Unfortunately, we can observe only the actual values of the variables,
and therefore some further econometric manipulation is needed to estimate the
equilibrium real exchange rate.
Intuitively, the observed exchange rate can be decomposed into two components: the first
one, when the fundamentals are at their steady state levels, would be the equilibrium
exchange rate
q =0+F f +N n (18)
The second component, when the fundamentals are away from their respective steady
states, would correspond to the deviation of the exchange rate from its equilibrium level.
q^t =0+F f^t +N n^t (19)
where f^t and n^t denote deviations of the fundamentals from their equilibrium values.
Our strategy for estimation of the equilibrium real exchange rate is based on the
econometric decomposition of the observed real exchange rate into a transitory and a
15The assumption i*>g amounts to ruling out dynamic inefficiency.
10
permanent component. The estimated equilibrium exchange rate is taken to be the
permanent component, while the transitory component reflects deviations from
equilibrium. In what follows, we first relate the equilibrium exchange rate with the
concept of cointegration, and then show how cointegration allows for the extraction of
the two unobserved components from the observed exchange rate and fundamental series.
In order to understand the link between equilibrium and cointegration, it is useful to
depart from the PPP view, which implies a constant value for the equilibrium real
exchange rate q or, in econometric terms, that qt is integrated of order zero (I(0)). Failure
of PPP to hold does not necessarily imply that no equilibrium exists, but rather that the
equilibrium may be time-varying. In our case, if q, f, and n are cointegrated, then u in
(17) will be I(0), and an equilibrium real exchange rate will exist. In other words, q will
fluctuate around a time-varying equilibrium characterized by the long-run cointegrating
vector [1, -F, -N].
As noted above, the time-varying equilibrium exchange rate cannot be inferred by simply
imposing the cointegration vector on the observed values of the explanatory variables. In
this regard, cointegration among a set of variables presents a very desirable property: it
allows for the decomposition of the relationship among the variables into two
components: a permanent or secular I(1) component, which describes the long-run
properties of the relationship, and can be identified with a time-varying equilibrium path;
and a transitory I(0) component, which corresponds to deviations from the permanent
component and represents departures of the fundamentals from their steady state values.
The decomposition of the observed series into their permanent and transitory components
requires identification of the basic properties of the latter. We follow Gonzalo and
Granger (1995) and derive a decomposition where the transitory component does not
Granger-cause the permanent component in the long run, and where the permanent
component is a linear combination of contemporaneous observed variables. In other
words, the first restriction implies that a change in the transitory component will not have
an effect on the long-run values of the variables.16 The second restriction makes the
permanent component observable and implies that all the information necessary to extract
it is contained in the contemporaneous observations of the variables.
Let us consider the 3x1 vector xt = [qt, ft nt]' which under the null hypothesis of one
cointegration vector admits the following representation:
xt = D1xt +...+ Dp xt
-1 -1 -p+1+ xt -p + et, (20)
where et is a vector white noise process with zero mean and variance and is 3 x 3
matrix with rank 1. Given that is not full rank, it can be written as the product of two
rectangular matrices and of order 3 x 1 such that ='. In turn, is the
cointegration vector and is the factor-loading vector. Next, we can define the
orthogonal complements and as the eigenvectors associated with the unit
16 In essence, this decomposition rules out hysteresis in the real exchange rate.
11
eigenvalues of the matrices (I- (' )-1 ') and (I- (' )-1 '), respectively. Notice
that ' = 0 and ' = 0. With this notation we can write
xt = (' )-1xt +(')-1'xt , (21)
where (' )-1 xt captures the permanent component and (' )-1 ' xt the
transitory component. Gonzalo and Granger (1995) show that the transitory component
defined in this way does not have any effect on the long-run values of the variables,
which are captured by the permanent component.
The identification of the permanent component with equilibrium implies that
xt = (' )-1xt and x^t = (')-1'xt ,
from where an estimate of the equilibrium exchange rate and its deviation follows
directly.
A final issue to consider in this section is the extent to which the statistical model can
yield additional knowledge regarding the theoretical model. As noted above the
theoretical model encompasses the Balassa-Samuelson hypothesis and the balance of
payments approach. It is therefore natural to enquire if the data provide support for both
views, for one of them only, or for neither. This can be assessed by constructing test
statistics for the following hypotheses on the cointegration vector:
Balassa-Samuelson approach: F =0
Under the null F=0, the real exchange rate would be determined only by the Balassa-
Samuelson model, and would evolve according to relative productivity differentials.
Under the null, the test statistic, which we denote BASA, is distributed as a 2(1).
Balance of payments approach: N = 0
The case N = 0 corresponds to a situation in which either n does not enter the model or
N =0 (i.e., the economy is fully tradable). The Balassa-Samuelson effect then plays no
role as determinant of the real exchange rate. The test statistic, that we denote BOP, is
distributed as a 2(1) under the null.
PPP: F = N = 0
If F = N = 0, the real exchange rate is stationary, as posited by the PPP approach.
Econometrically we should find a cointegration vector of the form [1, 0, 0]. Under the
null, the test statistic, denoted PP, is distributed as a 2(2). Notice also that it would be
possible to discriminate further between absolute and relative PPP, depending on whether
the cointegration vector is centered around 0 (absolute PPP) or not (relative PPP).
12
V. Misalignment and its sources
In this section we estimate the equilibrium real exchange rate for the Argentine peso. The
sample covers the period 1960-2001 and the data are annual. We first describe the
construction of the series and then present the results.
V.1 Data
Empirical estimation of our model requires data on three variables: the real effective
exchange rate q, the stock of net foreign assets as a percentage of GNP f, and relative
sectoral productivity n. Following the majority of the literature, we use a CPI-based index
of the real effective exchange rate (henceforth REER), whose weights are based on
Argentina's bilateral trade flows averaged over the period 1998-2000. We construct
weights for 40 trading partners which combined account for 90 percent of Argentina's
trade. Figure 4 reports the weights corresponding to a higher level of aggregation
distinguishing only five trading partners: Europe, United States, Asia, Mercosur and the
rest of the World (ROW). Mercosur is Argentina's main trading partner, accounting for
almost 40 percent of her total trade. Within Mercosur, Brazil is the leading trading
partner, with about 30 percent of Argentina's trade. Next comes Europe, with 25 percent
(of which 23 percent corresponds to the Euro area). The U.S. follows with 17 percent,
and finally the rest of the world accounts for about 5 percent of Argentina's trade.
The stock of net foreign assets F can be computed by accumulating past current account
balances:
t
Ft = CAt-j + F0. (22)
j=0
Of course, to implement this procedure we still need an initial value F0. We take it from
Broner et al. (1998), who estimate F1965 at US$-3,853 million; f is then computed as the
ratio of F to the U.S. dollar-denominated GNP.
Finally, data on sectoral productivity or even for overall productivity are not available for
Argentina and many of the partner countries for the sample period. However, we take
advantage of the already robust evidence of a long-run relation between sectoral
productivity and sectoral prices (see, among others, Canzoneri et al. (1999), Alberola and
Tyrväinen (1999)) to use an index of relative sectoral prices as a proxy for sectoral
productivities. More precisely, we use the comparative index of the relative price of non-
tradable versus tradable goods devised by Kakkar and Ogaki (1999) for estimation
purposes: we proxy n with the (log) ratio of Argentina's consumer price index (CPI) to
the wholesale price index (WPI), relative to the corresponding ratio for Argentina's
trading partners. The trade weights used in the computation are those used to construct
the real effective exchange rate.
V.2 Econometric Results
Table I reports the results of Johansen cointegration tests for Argentina using a VAR of
order 3. Inspection of the table indicates that on the basis of the maximal eigenvalue test
13
one can reject at the 5 percent level the null of no-cointegration in favor of the existence
of one cointegration vector.
Table I. Cointegration
Ho: r Eigenvalue Trace -max Critical Values 5%
Trace -max
2 0.0077 0.2619 0.2619 8.1800 8.1800
1 0.1769 6.8829 6.6211 17.9500 14.9000
0 0.4672 28.2870 21.4041a 31.5200 21.0700
Cointegration vector: q = 0 +1.82 f + 0.69n
Loading Matrix: = [-1.14 (s.e. .25) -.01 (s.e. .06) -.18 (s.e. .11)]'
PPP = 24.43a BASA=16.91a BOP=23.99a
a. Indicates rejection of the null at the 5 percent.
The table also shows the cointegration vector and the loading matrix (whose standard
errors are shown in parentheses). As predicted by the analytical model, the cointegration
vector uncovers a positive relationship between the real exchange rate and the stock of
net foreign assets and relative productivity growth. Also, the magnitude of the parameter
estimates is comparable to that reported by Alberola et al. (1999), who apply a similar
specification to 12 OECD currencies.
As for the VAR loading factors shown in the table, we find that that the only loading
factor significantly different from zero at the 5 percent level is the one for the real
exchange rate equation. In other words, the error-correction term does not enter the
equations for productivity and net foreign assets.
Table I also reports the BASA, BOP and PPP statistics described above. Recall that
BASA is the test statistic for the null hypothesis that the real exchange rate follows the
Balassa-Samuelson hypothesis alone, while BOP corresponds to the hypothesis that the
Balance of Payments approach suffices to explain the time path of the real exchange rate.
The 5- percent critical value for both statistics is 3.84. In turn, the PP statistic tests the
null hypothesis of PPP, and its 5-percent critical value is 5.99.
As shown in the table, all three statistics provide strong rejections of their respective null
hypotheses. We take this as evidence that both the balance of payments approach and the
Balassa-Samuelson hypothesis play a role in the determination of the real exchange rate,
as predicted by the analytical model.
Using the estimated cointegrating vector and the loading factors, we can decompose the
real exchange rate into its permanent and transitory components, following the
methodology described in the previous section. As already noted, the permanent and
transitory components are the empirical counterparts of the equilibrium real exchange
rate and its deviation from equilibrium, respectively.
14
Figures 5 and 6 present the equilibrium exchange rate for the Argentine peso and the
evolution of its misalignment (with 95 percent standard error bands17) since the inception
of the currency board. In Figure 6, positive values imply an overvaluation of the
multilateral rate, and negative values mean undervaluation.
The results reveal a large overvaluation of the peso at the end of the sample, reaching 53
percent18 in 2001. This figure is somewhat higher than the overvaluation derived from a
simple-minded PPP approach, which equals 43 percent in 2001. Likewise, the time
profile of estimated misalignment derived from our model is quite different from that
derived from the PPP approach. The PPP benchmark suggests that the real exchange rate
was significantly overvalued throughout the period of analysis. In contrast, our
estimations suggest that the peso was undervalued at the inception of the currency board
(by about 20 percent), remained close to its equilibrium value from 1993 to 1997, and
became grossly overvalued over 1998-2001.
The main ingredient behind this rising overvaluation in the final years of the sample is
the depreciation of the equilibrium exchange rate after 1993, shown in Figure 5. In the
model, the time path of the equilibrium real rate reflects the evolution of equilibrium
foreign assets and productivity differentials. From Figures 2 and 3, we can conclude that
the depreciation of the equilibrium real exchange rate after 1993 was mostly driven by
Argentina's declining net foreign asset position relative to GNP, given that relative
productivity, after a sharp increase in 1990-91, experienced only modest changes over the
rest of the decade.
In principle, the steady fall in the ratio of net foreign assets to GNP could have been due
to substantial current account deficits, sluggish growth, or both. In fact, Argentina ran
persistent current account deficits over the 1990s, averaging 3 percent of GNP, although
their magnitude increased in the second half of the decade. As for growth, it averaged a
high 6 percent per year in the first half of the decade, but in the second half (1996-2001)
it declined to an average of 2 percent per year. Further, growth turned negative at the end
of the period (1999-2001).
To disentangle the relative roles of sluggish growth and current account deficits in the
observed evolution of the equilibrium real exchange rate, we performed two simulation
experiments to assess the impact on the net foreign asset ratio of (i) maintaining a zero
current account balance from 1991 to 2001; and (ii) 6 percent average growth over 1996-
2001 (while keeping the original current account balance unchanged). In the first case,
we find that the "virtual" net foreign asset position would have been consistent with a 3
percent overvaluation in 2001 or, in other words, the real exchange rate would have been
17 See Appendix I in Alberola et al. (1999) for the derivation of the bands.
18 More precisely, this is the difference between the logs of the actual REER and its
equilibrium counterpart. The other figures quoted in this section also refer to log
deviations. We use these deviations rather than percentages to allow straightforward
additive decompositions of overall misalignment into its various components.
15
nearly in equilibrium at the end of the period. In the second simulation we find a more
modest effect on the equilibrium real exchange rate: we are able to wipe out just 7
percent of the 53 percent estimated overvaluation. This suggests that overspending, rather
than sluggish growth, was the main driving force behind the steady depreciation of the
equilibrium REER.
Finally, it is also interesting to view the degree of overvaluation of the peso that our
estimates reveal for the final years of Convertibility in the light of the results obtained by
Godfajn and Valdés (1999) from analysis of a large number of episodes of real
overvaluation. As noted earlier, the evidence they report suggests the existence of a
threshold of overvaluation beyond which correction of the misalignment without nominal
devaluation becomes very unlikely. Indeed, in their sample few overvaluations exceeding
25 percent, and none above 35 percent, were undone without a collapse of the nominal
exchange rate. Our estimates place the overvaluation of the peso in 2000 and 2001 well
above those thresholds, suggesting that at such point the collapse of the Convertibility
regime had become virtually unavoidable.
VI. The currency board: wrong peg or diverging fundamentals ?
In the previous section we have estimated the misalignment of the real effective exchange
rate of the peso and explained its trajectory on the basis of the fundamentals implied by
our model. In this section we take a different perspective in order to assess the role of
changes in the real exchange rates of other currencies in the misalignment of the peso. In
this regard, a number of observers have attributed the bulk of the misalignment to the
appreciation of the U.S. dollar against Argentina's major trading partners or, to put it
differently, to the depreciation of the Euro and the Brazilian real (the currencies of
Argentina's top two trading partners) at the end of the1990s. In contrast, other analysts
assign most of the blame for the collapse of Convertibility and the ensuing crisis to the
adoption by the authorities of economic (particularly fiscal) policies inconsistent with the
hard peg.19
To assess these views, consider the following identity:
q^A = (1$ - qA) + inadequate+peg
q42 (23)
43 (1
qA4- q$) q^$
4244 3
diverging
fundamentals
where we use the subscripts A and $ to denote the Argentine peso and the U.S. dollar,
respectively. This is just a decomposition of peso misalignment into three terms. The first
one captures the divergence between the equilibrium REERs of the dollar and the peso.
In the context of Argentina's peg to the dollar, a nonzero value for this term implies a
long-term divergence between the fundamentals of the anchor and client countries. To the
extent that in the model presented earlier the fundamental determinants of the real
19See for example Mussa (2002).
16
exchange rate the foreign asset position and relative productivity growth can be
affected by macroeconomic policies and structural reforms, we can relate this component
of peso misalignment with persistent policy divergences between Argentina and the U.S.,
which in the long run are inconsistent with Argentina's dollar peg.
The other two terms in the right hand side of (23) capture the peso misalignment
occurring even in the absence of policies inconsistent with the peg. Thus, we may view
such misalignment as reflecting the inadequacy of the peg itself. It combines two items.
One is the divergence between the actual REERs of the peso and the U.S. dollar, captured
by the second term on the right-hand side of (23). Since each REER can be expressed as a
weighted sum of the bilateral real exchange rates of trading partners (where the weights
are their respective trade shares), this term basically arises from differences between the
trade structures of the U.S. and Argentina. In practice, the key difference in their trade
structures concerns Brazil, who is a major trading partner for Argentina but not for the
U.S.20 Hence, the time path of (qA - q$) is dominated by the real exchange rate of the
Brazilian real, and we should expect this term to exhibit a significant increase in 1999
due to the abrupt devaluation of the latter currency.
The rest of the misalignment attributable to inadequacy of the peg, captured by the last
term in the right-hand side of (23), is just the misalignment of the REER of the U.S.
dollar, which the peso inherits through the peg. The logic of this term is quite simple:
absent policy divergences and asymmetries in trade structure (already captured by the
first two terms in the right-hand side of (23) above), pegging to a misaligned anchor
currency necessarily leads to misalignment. In practice, existing empirical studies21
suggest that the dollar was overvalued after 1996-97, and thus we would expect this last
term to add to the overvaluation of the peso in the final years of Convertibility.
To implement empirically the decomposition in (23) we need an estimate of the
misalignment of the U.S. dollar real effective exchange rate in order to capture the last
term in the right-hand side of the expression. To construct such estimate, we build from
the work of Alberola et al. (1999), who report multilateral misalignments for the period
1980-98 for twelve OECD currencies, by extending their sample to the period 1999-01.
Figure 7 reports the estimated real misalignment of the dollar that results from this
procedure. Inspection of the figure indicates that during the first half of the 1990s the
dollar was slightly undervalued, but in 1996 there was a change in trend which lead to an
increasing overvaluation. By 1999 the overvaluation was about 10 percent, and it rose to
15 percent in 2000 and over 20 percent in 2001.22
Using Figure 7, we can calculate the decomposition in equation (23). Its results are
shown in Figure 8, which reports the sources of the annual change in peso misalignment
20 Brazil accounts for 30 percent of Argentina's trade, but only for 2 percent of U.S.
trade.
21See Alberola et al (1999).
22 Subsequently the dollar depreciated by about 10 percent on a trade-weighted basis
between mid 2001 and mid 2002, and further over 2003.
17
over 1993-2001 (with the changes measured by the first difference of the overall peso
misalignment shown in Figure 6).
The figure shows some salient facts. First, except for 1993, diverging fundamentals made
a positive contribution every year; that is, given other things, they invariably added to
overvaluation. Thus, Argentina's economic fundamentals in terms of productivity
trends and spending relative to income diverged persistently from those that would
have been required to sustain the currency board. Ultimately, this suggests that the stance
of macroeconomic policies and structural reforms was not tuned to the maintenance of
Convertibility in the long run.
Second, the role of peg inadequacy in the overvaluation of the peso was more erratic. Its
contribution was positive in 1993, became negative in 1994-95 thus pushing the peso
towards undervaluation -- and then returned to positive values throughout the rest of the
sample period. Thus, the `wrong peg' did add considerably to overvaluation in the final
years of Convertibility.
Further, the two components of this term trade asymmetries and dollar overvaluation
also exhibited changing behavior over the period. Between 1994 and 1998 trade
asymmetries made a negative contribution to the overvaluation of the peso i.e., they
tended to render the peso undervalued. This is consistent with the appreciation of the
Brazilian real over those years. However, its abrupt depreciation in 1999 added a
substantial push towards peso overvaluation (about 11 percent, as already noted). After
that year, the contribution of trade asymmetries was fairly modest. Finally, the
misalignment of the dollar also made a contribution of changing sign to peso
misalignment. But from 1996 on, as the overvaluation of the dollar developed, the
contribution became positive and also added to peso overvaluation.
Figure 9 offers a different perspective on the same results. As we already saw, over 1993-
97 the real exchange rate of the peso was roughly in equilibrium, and its overvaluation
developed after the latter date. Figure 9 shows the contribution of the three ingredients
under consideration to the cumulative change in the misalignment of the peso over 1993-
97 and 1997-2001. The contrast between the two periods is revealing. During the first
one, the misalignment of the peso barely changed. The figure shows that this was the
result of two opposing forces: inconsistent fundamentals, that tended to render the peso
overvalued (at a rate of 6.3 percent per year), and trade asymmetries, that worked in the
opposite direction (by 6.1 percent per year). As a result, the net effect was almost
negligible. Importantly, the contribution of dollar misalignment over this period was
small as well.
In contrast, in 1997-2001 all three ingredients worked in the same direction towards
overvaluation of the peso. Diverging fundamentals continued to contribute close to 6
percent per year to the overvaluation. But the inadequacy of the peg had also a positive,
and even bigger, impact, primarily through the increasing overvaluation of the dollar,
which by itself added almost 5 percent per year to overall peso overvaluation, and also
18
through the reversal in the effect of trade asymmetries, as the appreciation of the
Brazilian real ended with its collapse in 1999.
VII. Conclusions
Argentina's Convertibility regime quickly led to nominal stability and financial
deepening. However, it also allowed a large real appreciation to develop. Overvaluation
of the peso has been underscored by many observers as one of the key ingredients in the
Argentine crisis. This paper has assessed the degree of real exchange rate misalignment
in Argentina over the Convertibility decade and has explored the main factors behind it.
The paper develops a model in which the equilibrium real exchange rate is that consistent
with both a sustainable balance of payments position (external equilibrium) and the
efficient use of domestic resources (internal equilibrium). Thus the model encompasses
two leading views of real exchange rate determination -- the balance of payments and the
Balassa-Samuelson approaches. In the model, the fundamental determinants of the
equilibrium exchange rate are the stock of net foreign assets and the relative productivity
in tradable and nontradable goods vis-à-vis trading partners.
Empirical implementation of the model to the Argentine peso reveals that its real
exchange rate exhibits a unit root, which constitutes evidence against the PPP hypothesis.
Further, we find evidence supporting the presence of a cointegration relationship between
the real exchange rate and its fundamental determinants, as predicted by our model.
Finally, the data allow us to reject the hypotheses that either the balance of payments
approach or the Balassa-Samuelson approach alone suffice to explain the evolution of the
real exchange rate.
The results show that during the first years of the currency board, and coinciding with the
stabilization of the economy after the hyperinflation episode of the late 1980s, Argentina
enjoyed significant productivity increases relative to her trading partners, which may help
explain the sharp real appreciation of the peso between 1990 and 1993. After the latter
year, productivity stalled. We also note, however, the parallel process of foreign asset
loss, which brought the stock of net foreign liabilities from 15 percent of GNP in the
early 1990s to about 40 percent of GNP in 2001. In the model this implies a significant
correction in the equilibrium real exchange rate.
According to our results, after 1997 the peso became increasingly overvalued. By 2001,
the overvaluation exceeded 50 percent. Our framework allows us to assess whether this
reflected primarily the pursuance of macroeconomic policies inconsistent with the dollar
peg, or the inadequacy of the peg itself for the Argentine economy either because the
dollar was the wrong anchor given the Argentine trade structure, or because the anchor
was overvalued. Over 1997-2001, the period in which the overvaluation developed, we
find that both factors share the blame.
19
The paper's results also cast some light on the retrospective policy debate about whether
Argentina's nominal peg could have been salvaged by letting deflation do its job to
realign the real exchange rate. Given the international experience on how appreciations
are reversed, the large magnitude of the overvaluation that the paper finds in the final
years of Convertibility suggests that such course of action would have been very unlikely
to succeed.
20
Bibliography
ALBEROLA, E., S. GARCÍA, H. LÓPEZ, and A. UBIDE (1999). Global Equilibrium
Exchange Rates: Euro, Dollar, "Ins," "Outs," and Other Major Currencies in a Panel
Cointegration Framework, IMF Working Paper n.º 99/175.
-- (2002). «Quo Vadis Euro», European Journal of Finance, 8, 352-370.
ALBEROLA, E., and H. LÓPEZ (2001). ««Internal and External Exchange Rate Equilibrium
in a Cointegration Framework: An Application to the Spanish Peseta», Spanish Economic
Review, Vol.. 3, pp. 23-40.
ALBEROLA, E., T. TYRVÄINEN (1999). «¿Hay Margen para los Diferenciales de Inflación
en la UEM?», Moneda y Crédito, Vol. 208, pp. 65-120.
ALESINA, A., R. BARRO, and S. TENREYRO (2002). Optimal Currency Areas, Harvard
Institute of Economic Research Discussion Paper n.º 1958.
BAFFES, J., I. ELBADAWI, and S.A. O'CONNEL (1999). «Single Equation Estimation of
Equilibrium Real Exchange Rates», Chapter 10 in Estimating Equilibrium Exchange
Rates in Developing Countries, L.E. Hinkle and P.J. Montiel (eds.), The World Bank,
Washington.
BALASSA, B. (1964). «The Purchasing Power Parity Doctrine: A Reappraisal», The
Journal of Political Economy, Vol. 72, pp. 584-596.
BAYOUMI, T. and R. MACDONALD (1999). «Deviations of Exchange Rates from
Purchasing Power Parity: A Story Featuring Two Monetary Unions», Staff Papers,
International Monetary Fund, Vol. 46 n.º1, pp. 89102.
BREUER, J. (1994). «An Assessment of the Evidence of Purchasing Power Parity»,
Estimating Equilibrium Exchange Rates, J. Williamson (ed.), Institute for International
Economics, Washington.
BRONER, F., N. LOAYZA, and H. LÓPEZ (1998). «Desalineación y Variables
Fundamentales: Tasa de Cambio de Equilibrio en Siete Países Latinoamericanos»,
Coyuntura Económica, XXVII, n.º 4, pp. 101-124.
CALVO, G., A. IZQUIERDO, and E. TALVI (2002). «Sudden Stops, the Exchange Rate and
Fiscal Sustainability: Argentina's Lessons», IADB, mimeo.
CANZONERI, M., R. CUMBY, and B. DIBA (1999). «Relative Labor Productivity and the
Real Exchange Rate in the Long Run: Evidence from a Panel of OECD Countries»,
Journal of International Economics, Vol. 47, pp. 245266.
CASHIN, P., and C. J. McDERMOTT (2003). «An unbiased appraisal of Purchasing Power
Parity », IMF Staff Papers Vol. 50, pp. 321 --351.
CHINN, M. D. (1998). «Before the Fall: Were East Asian Currencies Overvalued?»,
NBER Working Paper n.º 6491.
21
CLARIDA, R. and J. GALI (1995). «Sources of Real Exchange Rate Fluctuations: How
Important are Nominal Shocks?», in Carnegie-Rochester Conference Series on Public
Policy, Vol. 41, pp. 1-56.
CORDEN, W. M. (2002). Too Sensational: On the Choice of Exchange Rate Regimes, MIT
Press.
DE GREGORIO, J., A. GIOVANNINI, and H. WOLF (1994). «International Evidence on
Tradables and Non-Tradables Inflation», European Economic Review, Vol. 38, pp.1225-
1244.
DE LA TORRE, A.,E. LEVY YEYATI and, S. L. SCHMUKLER (2003). Living and Dying with
Hard Pegs: The Rise and Fall of Argentina's Currency Board, The World Bank, mimeo.
DORNBUSCH, R. and A. WERNER (1994). «Mexico: Stabilization, Reform, and No
Growth», Brookings Papers on Economic Activity 1994, Vol.1 (Spring 1994).
FRENKEL, J.A. (1981). «The Collapse of Purchasing Power Parity During the 1970s»,
European Economic Review, Vol.16, pp.145-165.
FRENKEL, J. and M. MUSSA (1985). «Asset Markets, Exchange Rates and the Balance of
Payments», Handbook of International Economics, R.W. Jones and P. B. Kenen , North-
Holland (eds.), Amsterdam, Chapter 14.
GAY, A. and S. PELLEGRINI (2003). Tipo de cambio real y crisis cambiaria en Argentina,
paper presented at LACEA 2003, Puebla (Mexico).
GENBERG, H. (1978). «Purchasing Power Parity under Fixed and Flexible Exchange
Rates», Journal of International Economics, Vol. 8, pp.247-276.
GOLDFAJN, I. and R. VALDÉS (1999). «The Aftermath of Appreciations», The Quarterly
Journal of Economics, February, pp. 229-260.
GONZALO, J. and C. GRANGER (1995). «Estimation of Common Long-Memory
Components in Cointegrated Systems», Journal of Business and Economic Statistics,
Vol. 13, pp. 2735.
HAUSMANN, R. and A. VELASCO (2002). «The Argentine Collapse: Hard Money's Soft
Underbelly», Kennedy School of Government - Harvard University, mimeo.
HUIZINGA, J. (1987). «An Empirical Investigation of the Long-Rin Behaviour of Real
Exchange Rates», in Carnegie-Rochester Conference Series on Public Policy, Vol. 27,
pp.149-214.
ISARD, P. (1977). «How Far Can We Push the Law of One Price?», American Economic
Review, n.º 67, pp.942-948.
KRUGMAN, P. (1978). «Purchasing Power Parity and Exchange Rates: Another Look at
the Evidence», Journal of International Economics, n.º 8, pp. 397-407.
22
LANE, P. and G. M. MILESI-FERRETTI (2001). «The External Wealth of Nations:
Measures of Foreign Assets and Liabilities for Industrial and Developing Countries»,
Journal of International Economics, Vol.55 n.º2, pp. 263-294.
MACDONALD, R. (2000), Exchange Rate Economics: Theories and Evidence, 2nd Edition,
London: Routledge Keegan Paul.
MACDONALD, R. and J.L. STEIN (1999). Equilibrium Exchange Rates, Kluwer Academic
Publishers, London, UK.
MUSSA, M. (2002). Argentina and the Fund: From Triumph to Tragedy, Institute for
International Economics, mimeo.
NURSKE, R. (1944). «Conditions of International Monetary Equilibrium», reprinted in
The International Monetary System: Highlights from Fifty Years of Princeton essays in
International Finance, Peter Kenen (eds.), pp. 124.
RODRIK, D. (2002). Argentina: Globalization Gone Too Far or not Far Enough,
Kennedy School of Government Harvard University, mimeo.
ROGOFF, K. (1996). «The Purchasing Power Parity Puzzle», Journal of Economic
Literature, Vol. XXXIV, pp. 647-668.
ROUBINI, N. (2001), Should Argentina Dollarize or Float? The Pros and Cons of
Alternative Exchange Rate Regimes and their Implications for Domestic and Foreign
Debt Restructuring/Reduction, Stern School of Business, New York University, mimeo.
SACHS, J. (2002). Understanding and Responding to Argentina's Economic Crisis, Center
for International Development Harvard University, mimeo.
SAMUELSON, P. (1964). «Theoretical Notes on Trade Problems», Review of Economics
and Statistics, Vol. 46, pp.145-154.
SCHULER, K. (2002). Fixing Argentina, Cato Institute, mimeo.
STIGLITZ, J. (2002). «Argentina short-changed», Washington Post May 27, 2002.
TEIJEIRO, M. (2001). Una Vez Más, La Política Fiscal ... , Centro de Estudios Públicos,
mimeo.
WARNER, A. (1997). Mexico's 1994 Exchange Rate Crisis Interpreted in Light of the
Non-Traded Model, NBER Working Paper n.º 6165.
WILLIAMSON, J. (ed.) (1994). Estimating Equilibrium Exchange Rates, Institute for
International Economics, Washington.
WREN-LEWIS, S. and R. DRIVER (1998), Real Exchange Rates for the Year 2000, Institute
for International Economics, Washington.
23
Figure 1. Real Effective Exchange Rate of the Peso
Average REER
1960-2001=1
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Figure 2. Relative Productivity, Prices, and the Real Exchange Rate
2 1.75
1
1.5
REER
prices 0
(Average
and
-1 1.25 1960-2001=1
productivity
-2
)
Relative
Relative productivity 1
-3 Relative prices
Real exchange rate
-4 0.75
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
24
Figure 3. Net Foreign Assets
(Percent of GDP)
-10%
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
-15%
-20%
-25%
-30%
-35%
-40%
Figure 4. Argentina's Trade Structure, 1998 - 2000
(Percent of total trade)
Mercosur
Europe
USA
Asia
ROW
25
Figure 5. Equilibrium Real Exchange Rate
Average REER
1960-2001=1
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
26
Figure 6. Estimated Real Misalignment of the Peso
(Percentages)
80%
60%
40%
20%
0%
PPP-based
-20% Model-based
2 S.E. bands
-40%
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Figure 7. Estimated Real Misalignment of the U.S. Dollar
(Percentages)
35%
30%
25%
20%
15%
10%
5%
0%
-5%
-10%
-15%
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
27
Figure 8. Source of Peso Misalignment
(Annual contribution of each factor)
25%
20% 19.6%
15%
14.0%
10% 9.6%
8.3%
6.1%
5% 5.2%
4.2%
0%
-2.6% -3.1%
-5% Diverging fundamentals
Trade structure
-10% Dollar overvaluation
Total
-15%
1993 1994 1995 1996 1997 1998 1999 2000 2001
Figure 9. Sources of Cumulative Peso Overvaluation
(Percentages)
15%
10%
5.6%
2.0%
5% 6.3%
4.7%
1.5%
0%
-6.1%
Diverging fundamentals
-5%
Trade structure
Dollar overvaluation
-10%
1993-1997 1997-2001
28