ArticlePDF Available

Output and the Real Exchange Rate in Developing Countries: An Application to Mexico

June 1999
SSRN Electronic Journal 61(1):85-109

June 1999
61(1):85-109

DOI:10.1016/S0304-3878(99)00062-0

Source
CiteSeer

Authors:

John H. Rogers

Fudan University

Since Mexico's devaluation in 1994, some observers have called for policies designed to keep the real exchange rate highly competitive in order to promote exports and output growth. However, over the past few decades, devaluations have been associated nearly exclusively with economic contraction, while real appreciations have been followed by expansions. We attempt to disentangle the possible factors underlying this correlation — (1) reverse causation from output to the real exchange rate, (2) spurious correlation with third factors such as capital account shocks, and (3) temporary contractionary effects of devaluation — and determine whether a positive long-run effect of real depreciation on output is in the data. Based on the results of several VAR models, we conclude that even after sources of spurious correlation and reverse causation are controlled for, real devaluation has led to high inflation and economic contraction in Mexico. While changes in Mexico's economic structure and financial situation may qualify the future applicability of this conclusion, our findings point to substantial risks to targeting the exchange rate at too competitive a level.

Granger Causality Tests -Real GDP and Real Exchange Rate

…

Variance Decompositions of Real GDP and Real Exchange Rate from Model 1

…

Variance Decompositions of Real GDP and Real Exchange Rate from Model 3

…

Variance Decompositions of Real GDP and Real Exchange Rate from Model 4

…

Variance Decompositions of Real GDP and Real Exchange Rate from Model 5

…

Figures - uploaded by John H. Rogers

Content may be subject to copyright.

Content uploaded by John H. Rogers

Content may be subject to copyright.

Board of Governors of the Federal Reserve System

International Finance Discussion Papers

Number 580

May 1997

OUTPUT AND THE REAL EXCHANGE RATE IN DEVELOPING COUNTRIES:

AN APPLICATION TO MEXICO

Steven B. Kamin and John H. Rogers

Note: International Finance Discussion Papers are preliminary materials circulated to stimulate

discussion and critical comment. References in publications to International Finance Discussion

Papers (other than an acknowledgment that the writer has access to unpublished material) should

be cleared with the author or authors. Recent IFDPs are available on the Web at

www.bog.frb.fed.us.

OUTPUT AND THE REAL EXCHANGE RATE IN DEVELOPING COUNTRIES:

AN APPLICATION TO MEXICO

Steven B. Kamin and John H. Rogers *

Abstract: Since Mexico's devaluation of the peso in 1994, some observers have called for policies

designed to keep the real exchange rate highly competitive in order to promote exports and output

growth. However, over the past few decades, devaluations of the real exchange rate have been

associated nearly exclusively with economic contraction, while real appreciations have been

followed almost invariably by expansions in economic activity. The purpose of this paper is to

attempt to disentangle the possible factors underlying this correlation--(1) reverse causation from

output to the real exchange rate, (2) spurious correlation with third factors such as capital account

shocks, and (3) temporary contractionary effects of devaluation--and determine whether, once those

factors are accounted for, a positive, long-run effect of real depreciation on output can be identified

in the data. Based on the results of a VAR model designed to explore the linkages between the real

exchange rate and output, we conclude that even after sources of spurious correlation and reverse

causation are controlled for, real devaluation has led to high inflation and economic contraction in

Mexico. While changes in Mexico's economic structure and financial situation may qualify the

future applicability of this conclusion, we view our findings as pointing to substantial risks to

targeting the exchange rate at too competitive a level.

Keywords: contractionary devaluation, Mexico

*The authors are, respectively, Senior Economist and Economist, International Finance Division, Board of

Governors of the Federal Reserve System. We thank Sebastion Edwards for useful comments, and

Bernadette Ryan for superb research assistance. The views expressed in this paper are solely those of the

authors, and do not necessarily reflect those of the Board of Governors or any other members of its staff.

I. Introduction

The economic crisis that followed Mexico's 1994 devaluation of the peso has strengthened

scepticism concerning the benefits of exchange-rate based stabilization in several respects. First,

the crisis highlighted the riskiness of allowing the real exchange rate to appreciate substantially in a

world of high capital mobility. Second, the stark recession in Mexico that followed the devaluation

has refocused attention on the potential output costs of a fixed exchange-rate strategy.

Prior to the devaluation, supporters of Mexico's exchange rate strategy had argued that the

real appreciation of the peso, and the associated widening of current account deficits, had been

signs of economic health, reflecting rising productivity and improved prospects for the Mexican

economy. Concerns that real appreciation was restraining economic growth, such as raised in

Dornbusch and Werner (1994), were largely brushed aside. In the aftermath to Mexico's

devaluation and the severe crisis that followed, however, the use of the exchange rate as a nominal

anchor has been discredited to some degree, while the Dornbusch-Werner view has gained greater

prominence.

In particular, some observers were optimistic that the sharp rise in the real peso value of

the dollar after the December 1994 devaluation, indicated in Chart 1, would launch Mexico on a

path of sustainable, export-led growth. However, much of this adjustment subsequently was

reversed as Mexican inflation continued to exceed foreign levels while the nominal exchange rate--

after its initial devaluation--remained more stable. This reversal alarmed exporters, market

observers, and academic analysts that viewed it as a dangerous throwback to Mexico's pre-

devaluation pattern of overvaluation and widening external imbalances. Dornbusch argued that

further real appreciation had to be prevented--and implicitly, that further real devaluation would be

useful--if Mexico was to achieve its growth potential:

By conventional purchasing-power parity measures, Mexico's competitiveness today is

only about 20 percent better than in 1993. Over the past year and a half, its

competitiveness has been shrinking rapidly. Indeed, it is barely enough to keep the

economy expanding. It definitely is not sufficient to spark high growth and job creation.

(Business Week, November 25, 1996)

Dornbusch's view certainly is consistent with the conventional wisdom that depreciations

of the real exchange rate are expansionary. According to the textbook model, real depreciations

lead to enhanced exports and a substitution from imports to domestically produced goods, thereby

boosting aggregate demand. Moreover, a more depreciated real exchange rate, by encouraging the

growth of the export sector and hence the openness of the economy, is believed to place a country

on a developmental path with greater potential for sustained growth. Finally, a more depreciated

exchange rate would likely prevent a recurrence of the destabilizing financial crises that Mexico

experienced in 1982 and 1995.

While the usefulness of real devaluation in stimulating growth may seem self-evident, this

view is not uniformly supported either by prior theoretical research or by the experience of

countries implementing exchange rate devaluations. First, there is virtual consensus among

economists that the real exchange rate cannot be viably targeted on a sustained basis.1 For limited

periods of time, of course, the monetary authorities can encourage nominal exchange rate

depreciation sufficient to cover the differential between domestic and foreign inflation, thereby

stabilizing the real exchange rate. However, both theoretical considerations and empirical evidence

(Calvo, Reinhart, and Vegh, 1994, and Kamin, 1996) suggest that attempts to keep the real

exchange rate more depreciated than its equilibrium level are likely to lead to a viscious cycle of

rising inflation requiring rising rates of nominal exchange rate depreciation, eventually forcing the

abandonment of the target.

Second, even if a real devaluation could be sustained over time without egregious

inflationary effects, it is by no means clear that such a strategy would lead to higher output growth.

In prior decades, a substantial literature developed which showed that, through a multitude of

different channels, exchange rate devaluations might have contractionary effects on economic

activity. (See Krugman and Taylor, 1978, for one of the first formalizations of this argument, and

Lizondo and Montiel, 1989, for a broad analytical overview.) While the "contractionary

1See Comments by Edwards and Bruno in Dornbusch, Goldfajn, and Valdes, 1995.

devaluation" literature has gone out of fashion more recently, its disappearance does not appear to

reflect convincing refutations either of its assumptions or its predictions.

In fact, actual experience continues to make a strong prima facie case for the

contractionary effect of devaluation. Cooper (1971) found that in many instances, output declined

in the aftermath of devaluation. Updates of Cooper's "before/after" study by Kamin (1988) and

Edwards (1989) confirm that devaluations are associated with declines in output and/or output

growth as well. Edwards (1989) and Morley (1992) also estimate pooled time-series/ cross-

country regressions of output on the real exchange rate and other determinants such as the terms of

trade and still find that devaluations negatively affect output.

Finally, the experience of Mexico itself raises questions about the efficacy of devaluing the

real exchange in order to promote growth. Chart 2 underscores vividly the positive association of

output and the real dollar value of the peso during these years. During 1982-87, when Mexico

maintained a relatively depreciated real exchange rate in response to capital flight and constrained

access to foreign creditors, economic activity stagnated. In 1988-94, Mexico stabilized its nominal

exchange rate, allowing the real exchange rate to appreciate substantially, and growth recovered

markedly. Finally, in the wake of the substantial depreciation of the peso after its float in

December 1994, output fell precipitously.

In fact, there has never been a period in Mexico's recent history when a

competitive real exchange rate and a cyclically high level of economic activity occurred

simultaneously. Chart 3 plots the real exchange rate (now in terms of pesos per dollar, the

standard for the rest of this paper) against the detrended log-level of real GDP.2 There is a

remarkably tight negative correlation between real exchange rates and output. There are no

observations combining highly depreciated real exchange rates with high levels of output, or highly

appreciated exchange rates with very depressed output levels.3

2Detrended GDP is calculated as the residuals from a regression of the log of quarterly real GDP on a

constant and a linear time trend. Quarterly observations of both detrended GDP and the real exchange

rate were then averaged to annual data.

3 Perez-Lopez Elguezabal (1945) presents a similar chart for Mexican industrial production and the real

exchange rate. He finds the two to be cointegrated, but does not analyse the direction of causality.

The mere existence of the correlation indicated in Charts 2 and 3 is far from proof that a

sustained real depreciation would inhibit growth in Mexico. First, and most obviously, it is not

clear from the charts which way the causality runs--the pace of economic activity may influence

the level of the real exchange rate, rather than vice-versa. Second, the tight correlation between

output and the real exchange rate shown in Charts 2 and 3 may be spurious, reflecting the response

of both variables to exogenous shifts in oil export prices or in access to international borrowing.

Finally, it is possible that the direct expansionary effect of devaluation on output may be offset by

contractionary effects in the short run, but that over a longer period, a sustained real devaluation

would promote growth. The observed negative association between devaluation and growth in

Mexico may merely reflect the fact that in the past, real depreciation never was pursued long

enough to allow positive growth effects to become evident.

The purpose of this paper is to disentangle the various factors leading to an observed

negative correlation between exchange rate depreciation and output in Mexico--(1) reverse

causation, (2) spurious correlation with third factors, and/or (3) temporary contractionary effects

of devaluation--and determine whether, once those factors are accounted for, a positive, long-run

effect of real depreciation on output can be identified in the data. Our research is motivated, first,

by the desire to better understand the determinants of economic activity in developing countries,

and in particular, the linkages between exchange rates and growth. By controlling in a highly

comprehensive manner for sources of reverse causation and spurious correlation, we hope to

provide a more compelling test of the contractionary devaluation hypothesis than has been

implemented thus far.

Additionally, we hope to shed light on the options for exchange rate policy currently facing

Mexico and other developing economies. However, we acknowledge in advance that it may not be

possible to draw straightforward policy implications from our results. For example, a finding that

real devaluation raises growth does not necessarily provide the basis for targeting the exchange rate

at a highly competitive level, since such a target may be very inflationary. Conversely, a finding

that real devaluation is contractionary does not mean that the authorities should encourage real

exchange rate appreciation, since that would surely lead to renewed balance-of-payments crisis in

the future.

In Section II, below, we describe in greater detail the various factors that could explain the

correlation observed in the data between the real exchange rate and output in Mexico, and

summarize previous empirical work on this issue. Section III summarizes the results of various

statistical tests aimed at characterizing the bivariate relationship between output and the real

exchange rate. In Section IV, we describe the results of a VAR model designed to explore the

linkages between the real exchange rate and output, once sources of spurious correlation and

reverse causation are controlled for. Section V concludes.

II. Potential Explanations for Output-Real Exchange Rate Linkages

In order to sort out the various factors underlying the observed relationship between real

exchange rates and output in Mexico, we have found the "Salter-Swann" diagram presented in

Chart 4 to be a useful heuristic device. (See Salter (1959) and Swann (1960).) The real exchange

rate--the ratio of tradeables to non-tradeables prices--is plotted against the level of real income in a

small open economy. The internal balance curve IB represents the locus of points in which the

supply of non-tradeable goods is equal to its demand. Increases in income raise the demand for

non-tradeables, thereby raising their price relative to that of tradeable goods and appreciating the

real exchange rate along the IB curve.

The external balance curve EB represents combinations of real exchange rates and income

that equate the trade deficit with the capital account surplus. For a given capital account, increases

in income expand the demand for tradeables and require an offsetting depreciation of the real

exchange rate to keep the trade balance constant. Increases in the capital account surplus allow the

economy to run larger trade deficits and shift the EB curve to the right; conversely, decreases in the

capital account balance shift the EB curve to the left.

Based on this framework, we can put the recent macroeconomic history of Mexico into

perspective. In Chart 5, we superimpose the Salter-Swan diagram shown in Chart 4 on the scatter

plot of real exchange rates and detrended output presented in Chart 3. The IB curve has been

(impressionistically) drawn through the main locus of points stretching from upper left to lower

right; an EB curve for a zero capital account has been (equally impressionistically) drawn through

the 1990 observation, when the current account was approximately in balance, while EB curves for

capital account surpluses and deficits have drawn through appropriate observations as well.

As indicated in the chart, in 1981 Mexico had a cyclically high level of output, a highly

appreciated real exchange rate, and a large capital account surplus. The onset of the debt crisis in

1982, and the associated reversal of capital inflows, led to a sharp real devaluation of the peso and

a marked contraction in output by 1983. Following some recovery of output and reversal of real

depreciation in 1984 and 1985, additional shocks--the Mexico City earthquake and the sharp

decline in oil prices--led to the extreme real depreciation of the peso and depression in economic

activity registered in 1986-87.

After an exchange-rate based stabilization program was initiated in 1988, and Mexico

regained access to international capital markets (in part through the Brady Plan), Mexico appears

to have marched steadily down the IB curve, ending up in 1994 relatively close to where it had

been in 1981, with cyclically high output, a relatively appreciated peso, and a large current account

deficit. Subsequent to the December 1994 devaluation, a new reversal of capital flows

(equivalently, a new shift in the EB curve) forced the economy back to where it had been in the

early 1980s. Since early 1995, however, the process of recovery and real appreciation have begun

anew, as witnessed by the position of the economy in the third quarter of 1996, the latest available

data.

With this history as background, we now turn to various explanations of the observed

negative correlation between devaluation and output in developing countries in general, and Mexico

in particular.

II.1 Spurious Correlation

Devaluations frequently have been implemented in response either to adverse shocks--

declines in terms of trade, increases in international interest rates--or growing domestic imbalances-

-overvalued exchange rates, excessive current account deficits. These problems, as well as the

contractionary economic policies such problems might prompt, could depress economic activity,

even in the absence of devaluation. This view is supported by the fact that various studies,

including Kamin (1988) and Edwards (1989), identify declines in output growth in many countries

even prior to their devaluations. As discussed above, in Mexico, shocks to the capital account

present a particularly important candidate for spurious correlation between output and the real

exchange rate. Both in 1982 and in 1994-95, strong capital outflows forced the authorities both to

devalue the exchange rate and to implement contractionary policies; conversely, the regaining of

access to international capital markets in the 1990s may simultaneously have encouraged real

appreciation and output expansion. Hence, the possibility exists that the strong empirical

association between output loss and devaluation in Mexico merely reflects the correlation of both

of these events with external shocks, especially shifts in the capital account.

II.2 Causality Running from Output to the Real Exchange Rate

As underscored in Charts 2 and 3, not only are real depreciations associated with

recessions in Mexico, but real appreciations are associated with expansions. Based on the

observation that many exchange-rate based stabilizations in developing countries have been

associated both with real appreciation and strong output growth (Kiguel and Liviatan, 1992), some

analysts theorise that the expansion of aggregate demand that typically accompanies disinflation

programs pushes up the price of non-traded goods--while tradable prices are fixed by the pegged

exchange rate--thereby leading the real exchange rate to appreciate.

Different analysts have proposed different explanations for why demand expands after the

nominal exchange rate is stabilized, including the lack of credibility of the stabilization program

(Calvo and Vegh, 1993), the impact of reduced inflation on perceptions of permanent income

(Uribe, 1995, Roldos, 1995, and Uribe and Mendoza, 1996), or the impact on expected income of

associated improvements in the fiscal balance (Rebelo, 1994). In all cases, however, the direction

of causality is portrayed as running from domestic demand to the real exchange rate rather than

vice-versa.

The role of increasing aggregate demand in appreciating the real exchange rate is plausible

and, in the case of Mexico, has some empirical support as well (Kamin, 1996). The progression of

real exchange rate/output combinations, shown in Chart 5, from 1988 through 1994 certainly is

suggestive of an equilibrium mechanism of this sort. However, because it takes time for aggregate

demand to affect domestic prices and hence the real exchange rate, this hypothesis probably is best

suited to explain the longer term co-movements of the exchange rate and output, particularly when

the nominal exchange rate is fixed. The nearly simultaneous, high-frequency correlation of output

and the real exchange rate evident in Chart 2 is less likely to reflect the impact of demand on

domestic prices. Most obviously, the sharp real depreciations in 1982 and 1995 were caused by

devaluations of the nominal exchange rate, not by economic contractions that lowered domestic

prices.

II.3 Causality Running from the Real Exchange Rate to Output

Even when devaluations are prompted by external shocks, as likely occurred in Mexico in

1982 and 1994, they may exert effects on economic activity that are independent of any direct

impact on output of the shocks themselves. Devaluation usually is hypothesized to raise the

production of tradable goods. However, this positive effect on output may be slow and may be

offset, particularly in the short run, by contractionary impacts on the demand for non-tradeables.

In discussing the effects of devaluation on demand, it is important to distinguish between

dynamics and long-run effects. Theories of contractionary devaluation generally have focused on

the case of a single nominal devaluation--that is, a single, permanent increase in the domestic

currency value of foreign currency. However, because in the long run, nominal devaluations are

believed to elicit proportionate increases in prices that leave real exchange rates and economic

activity unchanged, these analyses usually have confined themselves to the impact effects of

devaluation on output. (See Lizondo and Montiel, 1989.)

Below, we first review a number of such impact effects, commenting on their prima facie

applicability to Mexico. We then discuss the possible effects of a steady-state devaluation of the

exchange rate, maintained by continuous devaluation of the nominal exchange rate in order to

stabilize the real exchange rate at its new level.

Rigid nominal levels Devaluations exert supply-side effects on the price level by raising

the domestic currency value both of traded goods and of imported intermediate goods (Lizondo and

Montiel, 1989). To the extent that other nominal variables, such as wages, lag the rise in the

general price level, this causes changes in the real value of those variables, and may give rise to

real effects on economic activity:

•In Diaz-Alejandro (1963), Krugman and Taylor (1978), and others, sticky wage rates adjust

only slowly to increases in prices, thereby reducing real wages, shifting income from low-

saving workers to high-saving capitalists, and hence reducing consumption. As indicated in

Chart 6, real wages typically fall after devaluations in Mexico, although they subsequently

remain depressed for longer than is explainable by nominal rigidities alone.

•If the level of the nominal money supply is rigid, devaluation-induced inflation may erode the

real money supply and depress activity (Krugman and Taylor, 1978). Chart 7 attests that

Mexican real balances typically have fallen after devaluation. However, it is not clear why

nominal money should be rigid--developing country governments typically target the interest

rate or the exchange rate, leaving the money supply to be determined endogenously by changes

in money demand.4

•If the trade balance is in deficit, a real devaluation raises the real domestic currency value of

that deficit, thereby reducing aggregate demand (Krugman and Taylor, 1978). As may be seen

in Chart 8, the Mexican trade balance certainly was in deficit before each of its major

devaluations, although it turned to surplus almost immediately thereafter.

Increased rates of change Because the price level does not leap instantly to its new

steady-state following devaluation, the rate of inflation rises temporarily. Increases in inflation,

along with the heightened prospects for future devaluation that inevitably attend an initial

devaluation, raise the expected rate of exchange rate depreciation. Finally, higher inflation and

depreciation prospects will tend to raise nominal interest rates.

4Edwards (1989) estimates an equation for real output that includes both the exchange rate and measures

of the nominal money supply (both actual and shocks) as explanatory variables. Not surprisingly, with the

nominal money supply held constant, depreciations of the exchange rate lower output. However, insofar

as monetary growth usually rises, along with inflation, after a devaluation (see Edwards and Montiel,

1989), it is not clear whether, on net, devaluation exerts a contractionary effect through the monetary

channel.

•Increased inflation rates may directly lower business and consumer confidence, thereby

depressing spending. As indicated in Chart 9, devaluations clearly have been associated with

additional inflation in Mexico.

•Increases in interest rates may raise debt-service burdens for borrowers and hence contract

spending.5 This effect is believed to have been particularly important in Mexico after the 1994

devaluation, when nominal interest rates soared, as shown in Chart 10.

•Higher inflation and interest rates may depress the demand for bank deposits, causing banks to

restrict the supply of credit and thereby inducing declines in spending. (See Copelman and

Werner, 1996.) As shown in Chart 11, the stock of real bank credit plunged in Mexico after

the 1994 devaluation, although it remains unclear whether this reflected predominantly declines

in the demand for credit or its supply.

Associated economic policies Devaluations generally are coupled with other

contractionary economic policies, often to contain the adverse inflationary and expectational effects

of the devaluation itself. Hence, post-devaluation recessions may in part reflect the contractionary

economic policies prompted or made necessary by devaluation.

•Chart 12 indicates that fiscal policy in Mexico clearly tightened after each major devaluation

episode, and hence must be considered an important candidate to explain each subsequent

recession.

•The pattern of monetary policy after devaluation, as reflected in the level of ex post real

interest rates shown in Chart 13, is less obvious.

Balance-sheet effects When individuals are net debtors in foreign currency, a real

devaluation raises the real, domestic currency value of their liabilities (and of interest payments on

these liabilities), inducing balance-sheet adjustments that may lead to reduced domestic demand.

(See Lizondo and Montiel, 1989.) These effects undoubtedly have been present in Mexico, where

levels of both external debt (net of official reserves) and dollar-denominated domestic debt have

been high. Post-devaluation balance sheet problems may have been even higher in 1995 than in

1982, since in the prior episode, much of the external debt was held by the Mexican public sector,

while the private sector held large amounts of unrecorded foreign-currency flight capital.

5See Bernanke and Gertler (1995) for a detailed discussion of the impact of interest rate changes on

borrower cash-flows and spending decisions.

Access to international capital markets In the aftermath to both the 1982 and 1994

devaluations, Mexico temporarily lost access to borrowing abroad. While to some extent, it was a

reversal of net capital inflows that precipitated these devaluations in the first place, it seems clear

that the loss of credit access accelerated in the aftermath to devaluation. This may have reflected

the role of the devaluation in "waking up" previously myopic foreign investors, as well as the

additional information conveyed by devaluation: that economic imbalances were greater, and the

government less committed to address them through contractionary policies, than had been believed

prior to devaluation.

The loss of external credit access, in turn, might have had contractionary effects by (1)

depriving domestic banks of credit to on-lend to the domestic non-financial sector, (2) restricting

the ability of firms to finance imports of intermediate inputs for production, and (3) inducing a

more contractionary stance of economic policy in order to meet balance-of-payments constraints

and reassure foreign investors.

Impact versus long-run effects of sustained real devaluation As noted above, in the long

run, a single nominal devaluation is expected to have no effect on the level of output. In this

context, the question arises as to what would be the effects on output of a sustained real

devaluation, that is, of (possibly) continuous depreciation of the nominal exchange rate in order to

keep the real exchange rate persistently more depreciated than its initial level. While the effects of

real exchange rate targeting on inflation have received some attention in the literature, its effects on

output have not been analysed closely.

On the one hand, sustaining the real depreciation of the exchange rate may maximize the

probability that output expands, since more time is provided, compared with the case of a single

nominal devaluation, for tradable production to respond fully to the real devaluation. Moreover, to

the extent that the private sector believes that real devaluation will be sustained, this will encourage

an increase in investment in the tradable sector.

On the other hand, considerable evidence suggests that targeting the real exchange rate at

too depreciated a level is inflationary. Kamin (1996) shows that the level of the real exchange rate

was a primary determinant of the rate of inflation in Mexico during the 1980s and 1990s, while

Calvo, Reinhart, and Vegh (1994) identify correlations between the temporary components of

inflation and the real exchange rate in Brazil, Chile, and Colombia. Looking at a larger sample of

countries in which crawling peg policies were instituted, Edwards (1989) determines that in many

of them, real depreciation was achieved at the cost of considerable additional inflation.

To the extent that targeting the real exchange rate at a depreciated level is inflationary, this

means that continuous nominal devaluation will be required. If a single nominal devaluation

initially is contractionary, sustained nominal depreciations might be so as well. Hence, on

theoretical grounds, the prospective impact of a sustained real depreciation on growth is uncertain.

III.4 Results of Previous Empirical Work

Direct tests of the contractionary devaluation hypothesis The emergence in the 1970s and 1980s

of a large theoretical literature on contractionary devaluation encouraged relatively little "follow

up" empirical work, in part because this topic was eclipsed by growing interest in other topics such

as the effects of exchange-rate based stabilization. In a pooled time series/cross-country sample,

Edwards (1989) regresses real GDP on measures of the nominal and real exchange rate,

government spending, the terms of trade, and measures of money growth. In general, he finds that

even holding those other factors constant, devaluations tended to reduce output. While

devaluations may have been correlated with other external shocks omitted from the analysis--

particularly increases in international real interest rates or country-specific reversals of capital

flows--Edwards' estimates provide a prima facie case for devaluations' contractionary effect.

Agenor (1991) regresses output growth in a time-series/cross-country sample on

contemporaneous and lagged levels of the real exchange rate and on deviations of actual from

expected changes in the real exchange rate, government spending, the money supply, and foreign

income. He finds that surprises in real exchange rate depreciation actually boost output growth,

but that increases in the level of the real exchange rate exerts a contractionary effect. Finally,

Morley (1992) analyses the effect of real exchange rates on output, also in a time-series/cross-

country regression framework, holding constant a broad array of factors that could simultaneously

induce devaluation and reduce output: the terms of trade, import growth, the money supply, and the

fiscal balance. He also finds depreciations of the level of the real exchange rate to reduce output.

Findings from VAR models The expanding use of VAR models to study macroeconomic behavior

in developing countries, while not focused on the effects of exchange rates on output per se, has

yielded much relevant information. Rogers and Wang (1995) estimate a five-variable VAR--

output, government spending, inflation, the real exchange rate, and money growth--in an attempt to

decompose the movements of Mexican output in the 1977-90 period. They find that most of the

variation in Mexican output was attributable to shocks to "own shocks" to output itself, but in their

model, positive shocks to the real exchange rate (that is, devaluations) do lead to declines in output.

Rodriguez and Diaz (1995) get similar results for the Peruvian economy. In a VAR with six

variables--output growth, real wage growth, exchange rate depreciation, inflation, monetary

growth, and the Solow residual--output growth mainly is explained by its own shocks, but is

negatively effected by increases in exchange rate depreciation as well.

Copelman and Werner (1996) analyse some related five-variable VARs--output, the real

exchange rate, the rate of depreciation of the nominal exchange rate, the real interest rate, and a

measure of real credit or real money balances--also for Mexican data. They find that positive

shocks to the rate of exchange rate depreciation significantly reduce credit availability and depress

the level of output. Interestingly, shocks to the level of the real exchange rate have no effects on

output, suggesting that the contractionary effects of devaluation are associated more with the rate

of change of the nominal exchange rate than with the changed level of the real exchange rate. Also

interestingly, shocks to real credit itself have no effect on output, suggesting (contrary to the

authors' interpretation) that the rate of depreciation depresses output through some means other

than reducing credit availability.

Finally, various analyses of the effects of exchange-rate based disinflation on output have

provided indirect evidence concerning the impact of devaluations on output. In a two-variable

VAR for Mexico--output and the rate of nominal exchange rate depreciation--Santaella and Vela

(1996) find that a reduction in the rate of exchange rate depreciation causes an initial rise in output

which was reversed after about 12 quarters; this suggests that, conversely, a rise in the rate of

depreciation would have prompted an initial decline in output. Hoffmaister and Vegh (1996)

perform a similar experiment with a five-variable VAR--output, inflation, nominal exchange rate

depreciation, money growth, and Argentine output--applied to Uruguay. They find that a

permanent reduction in exchange rate depreciation leads to a long-lasting increase in output, much

as in Copelman and Werner (1996), whereas a negative shock to money growth strongly depresses

output (unlike Copelman and Werner's results).

In sum, econometric analyses indicate almost uniformly that devaluations--either increases

in the level of the real exchange rate or in the rate of depreciation--are associated with reductions in

output. Moreover, these studies provide little evidence that the initial contractionary effects of

devaluation are subsequently reversed. On the other hand, few of the VAR studies described above

control adequately for the full range of exogenous shocks, particularly those to the capital account.

This is crucial, since devaluation usually represents a response to adverse economic conditions. If

indicators of external shocks are not controlled for adequately, their effects may be mistakenly

attributed to the exchange rate.

III. Bivariate Data Analysis

Before describing the multivariate analysis of Mexican data in Section IV, below, we first

evaluate the bivariate correlations between output and the real exchange rate in Mexico. The data

are quarterly, from 1980.1 through 1996.2. The measure of output is real seasonally-adjusted

GDP; the measure of the real exchange rate is the CPI-adjusted peso price of the U.S. dollar, so

that increases indicate real depreciation.6

6This measure moves relatively closely with broader measures such as the IMF's real, effective

multilateral exchange rate, and is available on a more timely basis.

Tables 1 to 3 analyses the correlations between various transformations of the real

exchange rate and of output at various leads and lags. The lag number indicates the number of

quarters by which the real exchange rate is lagged relative to GDP. The data are consistent with

what is observed in Chart 2: in nearly all cases, the real exchange rate is negatively associated with

GDP, that is, devaluations are associated with downturns in output, and appreciations with output

growth. The particular filter used to transform the data does not seem to matter much, although

the HP filter appears to reduce the absolute magnitude of the correlation between the lagged real

exchange rate and GDP, while increasing the magnitude of the correlation between lagged GDP

and the real exchange rate.

For the other filters, correlations appear to be somewhat stronger running from lagged real

exchange rates to GDP than vice-versa, but the evidence is by no means clear-cut. To address the

issue of timing more directly, the first two rows of Table 4 present the results of bivariate Granger

causality tests. Applied over the full sample, the tests indicate unambiguously that lags of real

GDP do not help to explain movements in real exchange rate (over and above the effect of the

lagged dependent variable), but lagged real exchange rates unambiguously do help to explain real

GDP. Hence, the full-sample estimates provide prima facie support for the view that the tight

correlation between real exchange rates and output shown in Chart 2 reflects causality running

from exchange rate to output, rather than vice-versa.

Notwithstanding the strength of the full-sample results, it would be logical to assume that

the bivariate relationship between the real exchange rate and output would be affected by the

exchange rate regime. Over the period in question, four alternating regimes can be identified:

1980-81, when the nominal exchange rate was fixed; 1982-87, when the nominal exchange rate

was depreciated frequently in order to vary the level of the real exchange rate; 1988-94, when the

nominal exchange rate was fixed and then closely managed; and 1995-present, when the real

exchange rate has largely floated. We re-did our analysis, focusing on the two longest periods,

1982-87 and 1988-94.

During the 1982-87 period, the real exchange rate was highly variable, in large part

reflecting sharp policy-induced movements in the nominal exchange rate. Under these

circumstances, we might expect to see the real exchange rate causing output (with a negative sign).

The actual results, however, are quite mixed. Table 2 indicates that for most filters, real exchange

rates were negatively correlated with future output, but the HP-filtered data represent a prominent

exception. Moreover, the Granger-causality results presented in the second two rows of Table 4

suggest, at best, two-way causation between output and the real exchange rate.

During the 1988-94 period, the nominal exchange rate was largely fixed, so that variations

in the real exchange rate mainly reflected Mexican/foreign inflation differentials. Under these

circumstances, causality might be expected to run from output to the real exchange rate, again with

a negative sign, as discussed in Section II.2. The correlation results, shown in Table 3, are mixed,

with increases in output associated with appreciations of the real exchange rate only when the data

are log- or HP-filtered. On the other hand, the Granger-causality results, shown in the last two

rows of Table 4, certainly confirm that over this sub-period, causality ran from output to real

exchange rates rather than vice-versa.

The somewhat murky nature of our sub-sample results, compared with the strong full-

sample results, highlights two issues. First, in any large, complicated system, bivariate

correlations between two endogenous variables are likely not to remain constant over time.

Second, to the extent that a particular relationship between two endogenous variables does prevail,

identifying that relationship requires a data sample sufficiently large to incorporate significant

variation in the variables. In the Mexican case, the business cycles associated with three

significant episodes of real depreciation--1982, 1985, and 1995--probably helped to define the full-

sample results. Shortening the sample, by cutting out significant episodes, reduces the clarity of

the results accordingly.

IV. VAR Analysis

The analysis described above suggests that, for the sample as a whole, causality has run

from the real exchange rate to output rather than vice-versa. However, the bivariate results do not

exclude the possible of spurious correlation, nor do they provide information regarding the

channels by which the real exchange rate might affect output. In this section, we derive and

estimate a model to address these issues.

IV. 1 Derivation of the VAR Model

In our work, we hope to build on the results of previous research described above, and

develop a VAR model that (1) more clearly identifies the sources of shocks, and (2) more fully

controls for important external shocks affecting economic performance. We identify the variables

to be included in the VAR, as well as the recursive ordering of their contemporaneous shocks, with

the following illustrative framework.7 Because it incorporates all the potential channels through

which the exchange rate might affect output, this framework implies too large a number of

variables to be estimated in a single VAR model. Therefore, after describing the illustrative model,

we discuss its reduction to the more parsimonious models actually estimated.

The illustrative model The illustrative model consists of 12 endogenous variables and three

exogenous variables--the U.S. interest rate, U.S. inflation, and the dollar price of Mexican oil.

Equation 1 merely divides total GDP, Y, into domestic demand DD and net exports NX:

1) Y = DD + NX

Equation 2 relates net exports positively to the real exchange rate RER, defined as real the

peso value of foreign currency (up means depreciate). Increases in output Y raise the demand for

imports and hence lower net exports.

2) NX = a21RER - a22Y

The literature on contractionary devaluation, as well as the subsequent literature on

exchange-rate based stabilization programs, highlights a long list of factors influencing domestic

7Because the estimation method requires only restrictions on the contemporaneous correlations between

variables, for brevity we do not write out the lagged terms, although these are included in estimating the

structural equations.

demand--equation 3--including the real interest rate r, the fiscal deficit FISCDEF, the stock of real

bank credit RCREDIT, the nominal interest rate i, the rate of inflation π, the real exchange rate

RER, and the level of the real wage RW. In most cases, the conventional, expected sign of the

coefficient is written--in the case of RER, since its positive effect on net exports is included in

equation 2, it is assumed that any additional effects on aggregate demand are negative:

3) DD = -a31r + a32FISCDEF + a33RCREDIT - a34i - a35π - a36RER + a37RW

Equation 4 relates the supply of bank credit to the banking system's main sources of

loanable funds, real domestic money holds RM and borrowing from abroad, proxied by net capital

inflows KA:

4) RCREDIT = a41RM + A42KA

Equation 5 is a standard money demand equation. Kamin and Rogers (1996) found that

once interest rates and income were included in money demand equations for Mexico, other

variables such as inflation or the rate of exchange rate depreciation were not significant.

5) RM = a51Y - a52i

The nominal interest rate i is determined by a central bank reaction function, based on

inflation, output, and the level of capital inflows. This departs from the more standard uncovered

interest parity condition, reflecting our view that Mexican and foreign assets are only imperfectly

substitutable, providing the monetary authority with some scope for setting interest rates.

Increases in inflation and economic activity cause the central bank to raise interest rates, while

increases in net capital inflows cause the central bank to allow interest rates to decline. (See

similar reaction functions for Mexico estimated by Kamin and Rogers, 1996, and Kamin and

Wood, 1996.)

6) i = a61π + a62Y - a63KA

The CPI inflation rate is determined as in Kamin (1996). Increases in the level of the real

exchange rate RER shift demand toward non-traded goods, thereby raising overall prices for a

given level of the nominal exchange rate. Increases in output Y raise inflation through traditional

demand effects, while increases in the rate of nominal exchange rate depreciation E' raise inflation

by increasing the cost of imported goods.

7) π = a71RER + a72Y + a73E'

We assume net capital inflows KA are determined by the standard elements of the interest

parity condition, although we do not impose restrictions on their coefficients. iUS represents a

U.S. interest rate of comparable maturity to the Mexican rate.

8) KA = a81i - a82E' - a83iUS

Based on Mexico's experience in the mid-1980s, when the government appeared to be

targeting the real exchange rate, we assume the government sets a real exchange rate target first,

and then adjusts the nominal exchange rate in response to domestic and foreign inflation in order to

meet that target. Hence, nominal depreciation E' reflects Mexican inflation, US inflation, and the

real exchange rate.

9) E' = a91π - a92πUS + a93RER

The government is assumed to set the real exchange rate in response to balance-of-

payments pressures. Hence, improvements in net exports cause it to let the real exchange rate

appreciate, as do increases in net capital inflows. Because real net exports are, to a first

approximation, not affected by changes in the terms of trade, the price of oil is included separately-

-increases in oil prices improve the dollar valuation of the balance of payments and hence

encourage real exchange rate appreciation.

10) RER = -a101NX - a102KA - a103poil

The fiscal deficit is assumed automatically to decline in response to increases in output,

reflecting improvements in tax revenues. Increases in net capital inflows are assumed to raise the

fiscal deficit, both because improved access to foreign credit allows the government to borrow

more abroad, and because improved investor attitudes allow the government to pursue less austere

policies. Conversely, increases in inflation are assumed to prompt the government to tighten fiscal

policies; this depends on our use of the primary (non-interest) fiscal deficit as our measure of the

budget balance--increases in inflation that feed through to interest rates tend to raise the overall

budget deficit in Mexico. Finally, increases in oil prices, a major source of revenues for the

government, tend to lower fiscal deficits.

11) FISCDEF = -a111Y + a112KA - a113π - a114poil

Our last key variable, the level of real wages RW, is assumed to depend positively on

output, but, if the contractionary devaluation theorists are correct, negatively on the rate of

inflation, at least in the short run.

12) RW = a121Y - a122π

The core model The full system described above would be ideal to analyse the impact of

devaluation on output. It incorporates nearly all of the prominent channels of causality linking

exchange rates and output that have been raised in the literature. It also incorporates the most

obvious possible sources of spurious correlation--shocks to U.S. interest rates, oil prices, and the

capital account.

Unfortunately, given data limitations--quarterly observations over a relatively short 1 1/2

decade period--estimating the full system described above would not be practical. As an

alternative, we progressively substitute out most of the variables in the full model, leaving a core

model containing what we judge to be the four most important variables. These are comprised of

three endogenous variables--output, the real exchange rate, and inflation--and one exogenous

variable--the U.S. interest rate. This core model is then used in two ways. First, we estimate it to

see what information it can provide, by itself, on the relation between output and the real exchange

rate. Second, by adding one or two variables at a time to the core mode and re-estimating, we can

(1) determine the robustness of our model results, and (2) determine which channels of causation

between exchange rates and output figure most prominently.

By progressively substituting out those endogenous variables not included in the core

model--and by setting coefficients on the excluded exogenous variables to zero--the full 12-

equation system described above is reduced to the 4-equation system (not counting the equation for

exogenous iUS) shown below. The coefficients are complicated combinations of the coefficients in

the full model--in cases where the signs of the coefficients in the core model were ambiguous, we

wrote down our prior view as to which effect would dominate.

1') Y = -a'11π + a'12RER - a'13iUS

2') π = a'21RER + a'22Y

3') RER = a'31iUS + a'32π+ a'33Y

The model was estimated using the familiar two-stage approach to VAR estimation. In the

first stage, the variables were regressed on lags of all the variables in the system. In the second

stage, the Cholesky decomposition technique used by Sims (1980) was used to orthogonalize the

residuals.8 Based on the theoretical considerations raised in the illustrative model and its boiled

down version, the "core" model, the following order was employed: the U.S. interest rate, the real

exchange rate, the inflation rate, and real GDP. The rationale for placing the U.S. interest rate first

is obvious--it is unlikely to be affected by contemporaneous shocks to the Mexican economy. The

ordering for the remaining variables reflects the view underlying the illustrative model. Shocks to

the capital account--such as might be caused by shocks to the U.S. interest rate--induce adjustment

to the real exchange rate that, by altering nominal price levels, lead to changes in GDP.

Alternative models We consider the "core model" to be useful as a relatively parsimonious

starting point. The model is inadequate in two respects. First, it provides little sense of which

channels link the real exchange rate to output--the inflation variable is the only "intermediate"

variable in between the real exchange rate and output, and it could be proxying for many things.

Second, the core model does not account for all of the shocks that could simultaneously affect the

real exchange rate and output, thereby inducing spurious correlation between the two variables.

Therefore, in addition to the core model, we estimate a number of additional models in

which the core model is augmented by one additional variable. This provides an indication of the

robustness of our results, while at the same time keeping the size of the VAR model to a

8We initially attempted to estimate the model using the Bernanke (1986) technique, which would allow us

to give a more structural interpretation to the estimated innovations, but ran into convergence problems

with the maximum likelihood routine used in the estimation.

manageable level. The models and their ordering are listed below (the "core model" is shown as

Model 1):

Model 1: U.S. interest rate, real exchange rate, inflation rate, real GDP

Model 2: U.S. interest rate, gov’t size, real exchange rate, inflation rate, real GDP

Model 3: U.S. interest rate, real exchange rate, real M2, inflation rate, real GDP9

Model 4: U.S. interest rate, capital account, real exchange rate, inflation rate,

real GDP

Model 5: Capital account, gov’t size, real exchange rate, inflation rate, real GDP

Model 6: Oil price, capital account, real exchange rate, inflation rate, real GDP

IV.2 Estimation Results

Data Quarterly values of the following data were obtained from the Federal Reserve Board’s

INTL database and the Bank of Mexico’s Indicadores Economicos. As noted above, the data

begin in 1980.1.

U.S. interest rate: Nominal three-month Treasury Bill rate.

Real exchange rate: Log CPI-deflated peso value of the dollar.

Inflation rate: Log-change in level of Mexican CPI.

Real GDP: Seasonally adjusted log-level of Mexican real GDP.

Gov’t Size: Ratio of real (national income accounts) government spending

(consumption plus investment) to real GDP.

Real M2: CPI-adjusted level of M2.

Capital account: Ratio of capital account to GDP.

Oil price: Log of the price of crude oil and petroleum products that the U.S.

imports from Mexico (in U.S. $).

Preliminary analysis of the above series (not reported) suggests near unit root behavior in each

series, and hence suggests that a VAR in first-differences is appropriate. [All of the following

VARs were estimated using first-differences of the variables listed above. We have not yet tested

for the number of cointegration relationships among the variables in the VAR; the results of such

9 We also examined a version of model 3 switching the ordering of M2 and inflation. The results are

unaffected by this change, and so we do not report them.

tests may alter our estimation strategy.]. Given four lags in each VAR equation, and taking into

account differencing, the VARs were estimated over the period 1981.3 to 1995.4.10

Variance-decomposition results Tables 5 through 10 present the variance decompositions of

GDP, the real exchange rate, and inflation. These give the fraction of the forecast error variance

for each variable that is attributable to its own innovations and to innovations in the other variables

in the system. Each table provides the variance decompositions estimated from one of the 6 models

described above; results are reported for forecast horizons 1, 2, 4, and 12 (quarters). Each cell

represents the fraction of the forecast error variance attributable to innovations in the associated

column variable. For example, in Table 5, at the 1-quarter forecast horizon, 8.92 percent of the

variance in real GDP is attributable to shocks to U.S. interest rates.

The salient results from the variance decompositions are as follows. First, mirroring

results from previous studies, in all models, the predominant source of variation in GDP forecast

errors are "own shocks" to GDP--these generally account for 40-60 percent of the forecast error

variance at medium-term horizons, and upwards of 60-90 percent at horizons of 1 or 2 quarters.

Second, for those models that do not include the capital account, the second most

important source of variation in GDP is the real exchange rate, whose shocks account for 20-30

percent of the error variance at medium horizons. None of the shocks to the other variables in the

model--U.S. interest rates, government spending, inflation, real M2, or oil prices--explain more

than 15 percent of real GDP's forecast errors, and usually much less.11 On the other hand, in

models that include the capital account (models 4-6), shocks to the capital account are second only

to "own shocks" in accounting for the variance of GDP forecast errors, usually explaining 20-30

percent of it. In those models, real exchange rate shocks become the third most important source

of variation in GDP errors, accounting for about 10-20 percent. The contributions of capital

account and real exchange rate shocks are slightly larger when we estimate over the sample

10 We also estimated the models up to 1994.4, to see if the results were unduly influenced by the four post-

crisis observations we have available to us. We discuss below any noteworthy differences in results that

are due to the change in sample period.

11 We are in the process of computing standard errors for the variance decompositions and impulse

response functions.

excluding the post 1994.4 observations. For example, in model 4 the contributions of shocks to the

capital account, real exchange rate, and GDP are 27.9, 20.0, and 32.1, respectively, at the 20-

quarter horizon in the shorter sample, as compared to 24.4, 15.9, and 44.4 in the full sample.

Third, in models that do not include the capital account, after accounting for the effects of

"own shocks" to the real exchange rate, no other variable consistently accounts for a significant

fraction of the forecast errors in the real exchange rate. However, in models that include the

capital account, capital account shocks explain 10-20 percent of real exchange rate forecast errors.

This result is quite robust to the exclusion of the 1995 observations.

In conclusion, these results suggest that when capital account movements are not taken

into account, the real exchange rate emerges as a key determinant of macroeconomic developments

in Mexico. In particular, reverse causality from real GDP to the real exchange rate does not

appear to have been important, nor spurious correlation of devaluations and output with oil shocks

or US interest rates. On the other hand, it is evident that shocks to the capital account have been

important both to movements in the real exchange rate and to movements in output. What is not

clear is whether capital account shocks affect output only through their impact on the real

exchange rate, or whether they affect output through some independent channel as well. This

leaves open the possibility that the observed contractionary response of output to devaluation may

really reflect the movements of both the real exchange rate and output to capital account shocks.

This possibility will be explored further in the next section.

Impulse-response functions Figures 1, 2, and 3 contain the impulse-response functions of GDP,

inflation, and the real exchange rate, respectively. In each figure, the responses of a particular

variable to a one-time shock in each of the variables in the system are displayed; each panel in the

figure shows the impulse-responses calculated from one of the 6 different models. Each figure thus

allows us to determine whether the estimated response of a variable to a particular shock is robust

with respect to changes in model specification.

In interpreting the impulse-response functions, recall that the VAR models are estimated

using first-differences of the variables described above. Therefore, a one-time shock to the first-

difference in a variable is a permanent shock to the level of that variable. This allows us to address

directly certain questions of interest--in particular, what is the effect on output of a sustained

depreciation of the level of the real exchange rate? To aid in the interpretation of the results, the

responses of the levels of GDP, inflation, and the real exchange rate are displayed. These, of

course, are simply the accumulated responses of the first-differences.

The following conclusions emerge from examination of the impulse-response functions.

First, as indicated in Figure 1, a permanent, positive (depreciation) shock to the level of the real

exchange rate leads to a sustained reduction in the level of real GDP. The permanence of this

effect is consistent across all of the models estimated. In particular, it is not diminished by

inclusion of variables in the system that might simultaneously induce devaluation and reduce

output: U.S. interest rates, oil prices, or the capital account. Furthermore, the magnitude of the

effect of real exchange rate shocks is only slightly diminished by including the capital account in

the model.

Second, this contractionary effect of real devaluation on output comes out of models where

almost all the other estimated impulse-response functions are sensible and consistent with the

illustrative framework outlined earlier. As we would expect, increases in U.S. interest rates

usually reduce output, while increases in oil prices, the capital account, government spending, and

real M2 tend to raise real GDP.

An important exception is the role of inflation shocks. As discussed above, inflation

shocks might be expected to reduce output by (1) reducing business and consumer confidence, (2)

raising nominal interest rates and hence domestic debt-service burdens, and/or (3) reducing the

demand for money and hence the supply of loanable funds. However, as indicated in Figure 1,

positive shocks to inflation are associated with increases--not decreases--in output. One possible

explanation for this is that increases in inflation are not particularly contractionary, and that the

depressive effect of devaluation on output operates through some other channel not explicitly

accounted for in the models. The other possibility is that we are not identifying inflationary

"supply" shocks adequately, and that the estimated inflation shocks actually reflect positive shocks

to aggregate demand.

To address the latter possibility, the recursive ordering of the contemporaneous shocks in

the model was reversed, so that inflation always follows GDP (this implies that inflation shocks

have no contemporaneous effect on GDP). By reordering the variables in this way, we hope to be

able to interpret the resultant, estimated inflation shocks as arising from factors other than demand

shocks (since demand shocks already should be accounted for by GDP). However, in results not

reported, we find that even with the variables re-ordered, inflation shocks continue to lead to higher

GDP (after a very short initial period), casting doubt on the "mis-identification" explanation.

A third conclusion from the impulse-response functions is that the effect of a persistent

real devaluation on increasing inflation is as sustained as its effect in reducing output. This is

consistent with Kamin's (1996) findings, using a single-equation model, that the level of the real

exchange rate, as well as the rate of change of the nominal exchange rate, affects the rate of

inflation.

Finally, the impulse-response functions provide some, albeit limited, clues to the specific

mechanisms by which real devaluation affects output. As noted above, devaluation appears not to

affect output through its impact in raising inflation, although that result remains (as far as we are

concerned) in question. On the other hand, more consistently with our prior intuition, devaluation

works consistently to reduce government spending, to depress M2, and to lower the capital

account. These results suggest the possibility that the effects of devaluation in Mexico may be so

pervasive that no single, predominant channel of causation linking the real exchange rate and

output can be identified.

V. Conclusion

The results of the econometric work we have completed to date very clearly point to the

conclusion that in Mexico, sustained real devaluations have been associated with persistent high

inflation and contraction in economic activity. There is no evidence that this effect merely is

reflective of a spurious correlation of output and devaluation with other shocks. Moreover, our

results are consistent with virtually all prior analysis of devaluations and economic activity, both in

Mexico and elsewhere. What, then, are the implications of these findings for Mexican exchange

rate policy at present?

First, we do not draw from our results the conclusion that Mexico should encourage real

appreciation in order to disinflate quickly and encourage economic recovery. This is the strategy

that got Mexico into trouble in 1994, and given time, the same strategy would certainly lead to a

new financial crisis in the future.

Second, it is not clear that, just because devaluation and recession were associated with

each other in the past, a sustained real devaluation necessarily would be contractionary in the

future. As noted earlier, in Mexico in recent history, real devaluations have been sustained only

during periods of adversity, not during periods where internal and external circumstances have

been relatively favorable. It is possible that a strategy of real devaluation at this point--with

financial markets largely stabilized, re-access to international capital markets under way, and the

economy recovering--might yield greater benefits with less inflation than in the past. Additionally,

the many structural reforms that have been implemented in Mexico in the past decade--

privatization, dismantling of import barriers, reduction of fiscal deficits, etc.--may cause Mexican

GDP to be more (positively) responsive to devaluation, and inflation less responsive, than has been

the case in the past.

Third, the above qualifications notwithstanding, we view our results as pointing to

substantial risks--both in terms of inflation and economic activity--to targeting the real exchange

rate at too competitive a level. In this regard, we believe the level at which the real exchange rate

is targeted to be crucial. It is plausible that the more competitive the level at which the real peso is

targeted, the more likely it will be that the unfavorable outcomes predicted by our model will

eventualize. In this sense, a pro-growth policy based on a "super-competitive" exchange rate

would seem to be unduly risky.

On the other hand, there should be much less risk inherent in adopting policies that

maintain the real exchange rate at an equilibrium level--by encouraging a sufficiently rapid

depreciation of the nominal exchange rate--and that prevent further real appreciation.12 The

difficulty, of course, is in identifying that equilibrium level. Targeting the real exchange rate at too

depreciated a level could lead to the adverse combination of low growth and high inflation

described in Section IV. Conversely, allowing the real exchange rate to become too appreciated

undoubtedly would lay the foundation for renewed financial and economic crisis.

Properly identifying the equilibrium or sustainable real exchange rate for Mexico is beyond

the scope of this paper. It requires taking into account a broad array of variables--such as the

terms of trade, the level of tariffs, the stance of fiscal policy, and the level of world interest rates, to

name a few--and estimating the economy's response to their movements in general equilibrium.

However, as a "quick and dirty" alternative, eyeballing Chart 5, a case could be made for adjusting

monetary and/or exchange rate policy so as to keep the real exchange rate in the neighborhood of

its level in 1990. This was a period when (1) output was close to its trend level, (2) the trade

balance was close to being in deficit, (3) inflation was about 35 percent, high but down from

previous years’ levels, and (4) the real exchange rate was close to its historical average. All these

factors suggest that such a rate might be sustainable, allowing Mexico to get off the

macroeconomic rollercoaster ride it has been on for the past two decades.

12Logic also suggests that if the exchange rate were overvalued to start with, a sustained real devaluation

that brought the exchange rate into line with its equilibrium level might have more positive effects. In

recent papers advocating a policy of keeping the real exchange rate competitive, Dornbusch (1994, 1995)

has argued that during exchange-rate based stabilization programs, inflationary inertia causes the real

exchange rate to appreciate relative to its equilibrium value, and therefore that in those circumstances, a

nominal devaluation would have a good chance of leading to a sustained real devaluation with beneficial

economic effects. Research by Kamin (1996) casts doubt on the extent to which inflationary inertia can

explain the real appreciation of the Mexican peso during the 1988-94 period. However, he shows that

even equilibrium forces can cause real appreciations that enhance the risk of balance-of-payments crises,

thereby providing a role for policies geared toward avoiding overvaluation.

REFERENCES

Agenor, Pierre-Richard (1991), "Output, Devaluation, and the Real Exchange Rate in

Developing Countries," Weltwirtschaftliches Archiv, 127: 18-41.

Bernanke, Ben S. (1986), "Alternative Explanations of the Money-Income Correlation,"

Carnegie-Rochester Series on Public Policy, 25: 49-100.

_____ and Mark Gertler (1995), "Inside the Black Box: The Credit Channel of Monetary

Policy Transmission," Journal of Economic Perspectives, 9: 27-48.

Calvo, Guillermo and Carlos A. Vegh (1993), "Exchange Rate Based Stabilization under

Imperfect Credibility," in Helmut Frisch and Andreas Worgotter, Open Economy

Macroeconomics, London: MacMillan, 3-28.

Calvo, Guillermo, Carmen Reinhart and Carlos A. Vegh (1994), "Targeting the Real

Exchange Rate: Theory and Evidence," IMF Working Paper /94/22.

Cooper, Richard N. (1971), "Currency Devaluation in Developing Countries," in G. Ranis

(ed.), Government and Economic Development, New Haven: Yale University

Press.

Copelman, Martina and Alejandro M. Werner, "The Monetary Transmission Mechanism in

Mexico," working paper, Federal Reserve Board.

Diaz-Alejandro, Carlos F. (1963), A Note on the Impact of Devaluation and the

Redistributative Effects, " Journal of Political Economy, 71: 577-580.

Dornbusch, Rudiger and Alejandro Werner (1994), "Mexico: Stabilization, Reform, and No

Growth," Brookings Papers on Economic Activity, 1: 253-315.

_____ , Ilan Goldfajn, and Rodrigo O. Valdes (1995), "Currency Crises and Collapses,"

Brookings Papers on Economic Activity, 2: 219-294.

Edwards, Sebastian (1989), Real Exchange Rates, Devaluation, and Adjustment, Cambridge,

Massachusetts: MIT Press.

_____ and Peter J. Montiel (1989), "Devaluation Crises and Macroeconomic Consequences

of Postponed Adjustment in Developing Countries," IMF Staff Papers, 36: 875-903.

Hoffmaister, Alexander W. and Carlos Vegh (1996), "Disinflation and the Recession-Now-

Versus-Recession-Later Hypothesis: Evidence from Uruguay," IMF Staff Papers, 43:

355-394.

Kamin, Steven B. (1988), "Devaluation, External Balance, and Macroeconomic Performance

in Developing Countries: A Look at the Numbers," Princeton Essays in International

Finance No. 62.

_____ (1996), "Exchange Rates and Inflation in Exchange-Rate Based Stabilizations: An

Empirical Examination," International Finance Discussion Paper No. 554, Board of

Governors of the Federal Reserve System, Washington, D.C.

Kamin, Steven B., and John H. Rogers (1996), “Monetary Policy in the End-Game to

Exchange-Rate Based Stabilizations: The Case of Mexico,” Journal of International

Economics, forthcoming.

Kiguel, Miguel (1992), "Exchange Rate Policy, the Real Exchange Rate, and Inflation: Lessons

from Latin America," World Bank, WPS 880.

_____ and Nissan Liviatan (1992), "The Business Cycle Associated with Exchange Rate

Based Stabilization," World Bank Economic Review, 6: 279-305.

Krugman, Paul and L. Taylor (1978), "Contractionary Effects of Devaluation," Journal of'

International Economics, 8: 445-456.

Lizondo, Saul and Peter J. Montiel (1989), "Contractionary Devaluation in Developing

Countries: An Analytical Overview," IMF Staff Papers, 36: 182-227.

Mendoza, Enrique G., and Martin Uribe (1996), "The Syndrome of Exchange-Rate-Based

Stabilizations and the Uncertain Duration of Currency Pegs," International

Finance Discussion Paper No. 548, Board of Governors of the Federal Reserve

System, Washington, D.C.

Morley, Samuel A. (1992), "On the Effect of Devaluation During Stabilization Programs in

LDCs," Review of Economics and Statistics, LXXIV: 21-27.

Perez-Lopez Elguezabal, Alejandro (1995), "Un Model de Cointegracion Para Prognosticar

el PIB de Mexico," Banco de Mexico, Documento de Investigacion No. 71.

Rebelo, Sergio (1994), "What Happens When Countries Peg Their Exchange Rates? (The

Real Side of Monetary Reforms)," working paper, University of Rochester.

_____ and Carlos A. Vegh (1995), "Real Effects of Exchange-Rate-Based Stabilizations: An

Analysis of Competing Theories," NBER Macro Annual, National Bureau of

Economic Research, Cambridge, MA.

Rodriguez, Gabriel H. and Guillermo Diaz Gazani (1995), "Fluctuaciones Macroeconomicas

en la Economia Peruana," working paper, Banco Central de Reserva del Peru.

Rogers, John H. and Ping Wang (1995), "Output, Inflation, and Stabilization in a Small Open

Economy: Evidence from Mexico," Journal of Development Economics, 46:

271-293.

Roldos, Jorge E. (1995), "Supply-Side Effects of Disinflation Programs," IMF Staff Papers,

42: 158-183.

Santaella, Julio, and Abraham E. Vela (1996), "The 1987 Mexican Disinflation Program: An

Exchange-Rate-Based Stabilization?" IMF Working Paper /96/24.

Salter, W.E. (1959), "Internal and External Balance: The Role of Price and Expenditure

Effects," Economic Record, 35: 226-38.

Sims, Christopher (1980), "Macroeconomics and Reality," Econometrica, 48: 1-48.

Swan, T.(1960), "Economic Control in a Dependent Economy," Economic Record, 36: 51-66.

Uribe, Martin (1995), "Exchange-Rate-Based Inflation Stabilization: The Initial Real Effects

of Credible Plans," International Finance Discussion Paper No. 503, Board of

Governors of the Federal Reserve System, Washington, D.C.

Table 1 - Full Sample

Cross-correlations of real GDP and the real exchange rate with the same tranformation applied to

both series.

Full Sample

Lag

Log First

Difference Deviation

from Linear

Trend

Deviation

from

Quadratic

Trend

Deviation

from Cubic

Trend

HP-Filtered

4 -0.247 -0.052 -0.676 -0.513 -0.225 -0.062

3 -0.266 -0.041 -0.743 -0.608 -0.393 -0.067

2 -0.263 -0.154 -0.789 -0.697 -0.581 -0.065

1 -0.243 -0.485 -0.783 -0.730 -0.711 -0.058

0 -0.177 -0.049 -0.655 -0.606 -0.601 -0.045

-1 -0.156 -0.053 -0.527 -0.445 -0.381 -0.100

-2 -0134 -0.638 -0.389 -0.278 -0.153 -0.150

-3 -0111 0.084 -0.235 -0.089 0.078 -0.194

-4 -0.986 0.201 -0.102 0.089 0.241 -0.233

Note: "Lag" refers to the number of periods by which the corresponding detrended real exchange

rate lags real GDP.

Table 2 - 1980:1 - 1986:4

Cross-correlations of real GDP and the real exchange rate with the same tranformation applied to

both series.

Sub-period

1980:1-

1986:4

Lag

Log First

Difference Deviation

from Linear

Trend

Deviation

from

Quadratic

Trend

Deviation

from Cubic

Trend

HP-Filtered

4 -0.261 -0.144 -0.292 -0.415 -0.254 0.515

3 -0.283 -0.226 -0.441 -0.618 -0.534 0.625

2 -0.219 -0.264 -0.509 -0.735 -0.731 0.741

1 -0.070 -0.338 -0.485 -0.738 -0.808 0.863

0 0.152 -0.039 -0.316 -0.551 -0.656 0.988

-1 0.312 -0.288 -0.089 -0.283 -0.437 0.872

-2 0.514 0.051 0.2412 0.119 -0.074 0.761

-3 0.630 0.181 0.508 0.440 0.232 0.652

-4 0.709 0.189 0.677 0.682 0.458 0.547

Note: "Lag" refers to the number of periods by which the corresponding detrended real exchange

rate lags real GDP.

Table 3 - 1987:1 - 1996:2

Cross-correlations of real GDP and the real exchange rate with the same tranformation applied to

both series.

Sub-period

1987:1 -

1996:2

Lag

Log First

Difference Deviation

from Linear

Trend

Deviation

from

Quadratic

Trend

Deviation

from Cubic

Trend

HP-Filtered

4 -0.651 0.042 -0.358 0.115 0.281 -0.786

3 -0.671 0.144 -0.455 -0.027 0.163 -0.810

2 -0.706 -0.060 -0.633 -0.314 -0.117 -0.822

1 -0.741 -0.665 -0.779 -0.600 -0.446 -0.824

0 -0.706 -0.037 -0.718 -0.371 -0.224 -0.815

-1 -0.559 0.204 -0.603 -0.141 -0.022 -0.705

-2 -0.451 -0.186 -0.521 -0.114 0.037 -0.610

-3 -0.326 -0.031 -0.395 0.038 0.246 -0.511

-4 -0.204 0.184 -0.255 0.150 0.397 -0.410

Note: "Lag" refers to the number of periods by which the corresponding detrended real exchange

rate lags real GDP.

Table 4

Granger Causality Tests - Real GDP and Real Exchange Rate

Full Sample

Log First

Difference Deviation

from the

Linear

Trend

Deviation

from the

Quadratic

Trend

Deviation

from the

Cubic

Trend

HP - Filtered

Real GDP 0.118

(.976) 1.005

(.413) 0.262

(.901) 0.268

(.897)

0.274

(.893)

----

Real Exchange

Rate 7.02

(.001) 5.695

(.001) 9.021

(.000) 7.673

(.000)

5.888

(.001)

9.314

(.000)

Sub-period

80:1-86:4

Real GDP 3.246

(.042) 1.765

(.192) 1.428

(.273) 2.429

(.093)

2.693

(.071)

----

Real Exchange

Rate 0.431

(.784) 1.120

(.3862) 0.171

(.950) 0.992

(.442)

1.427

(.273)

0.004

(.999)

Sub-period

87:1-96:2

Real GDP 0.535

(.711) 0.937

(.456) 0.547

(.703) 0.693

(.602)

3.012

(.033)

----

Real Exchange

Rate 8.8543

(.000) 9.161

(.000) 9.530

(.000) 4.720

(.005)

5.366

(.002)

1.107E-4

(.999)

Notes: Each cell reports the F-statistic for the null hypothesis that lags of GDP of the exchange

rate are zero. P-values are in parentheses.

Table 5

Variance Decompositions of Real GDP and Real Exchange Rate from Model 1

A. VDC of Real GDP

Horizon/Shock U.S.

Interest

Rate

Real

Exchange

Rate

Inflation

Real GDP

1 8.92 0.56 0.73

89.80

2 6.89 26.67 0.54

65.90

4 6.61 26.39 1.84

65.16

20 7.94 26.21 2.57

63.28

B. VDC of Real Exchange Rate

Horizon/Shock U.S.

Interest

Rate

Real

Exchange

Rate

Inflation Real GDP

1 5.25 94.75 0.00 0.00

2 5.05 82.27 4.02 8.66

4 7.39 77.75 3.83 11.03

20 7.94 26.21 2.57 63.28

C. VDC of Inflation Rate

Horizon/Shock U.S.

Interest

Rate

Real

Exchange

Rate

Inflation Real GDP

1 0.01 6.26 93.73 0.00

2 0.06 17.95 73.69 8.29

4 0.27 17.28 72.15 10.30

20 0.59 16.50 73.43 9.48

Table 6

Variance Decompositions of Real GDP and Real Exchange Rate from Model 2

A. VDC of Real GDP

Horizon/Shock U.S.

Interest

Rate

Fiscal Real

Exchange

Rate

Inflation Real

GDP

1 5.92 1.94 0.12 0.01 92.01

2 3.79 4.99 28.43 0.01 62.78

4 4.05 9.82 25.67 1.78 58.68

20 4.69 11.02 27.67 2.10 54.51

B. VDC of Real Exchange Rate

Horizon/Shock U.S.

Interest

Rate

Fiscal Real

Exchange

Rate

Inflation

Real GDP

1 0.60 0.66 98.74

0.00

2 3.24 19.72 69.19

2.45

5.40

4 3.45 22.60 65.24

2.37

6.34

20 5.49 21.05 57.82

3.56

12.07

C. VDC of Inflation Rate

Horizon/Shock U.S.

Interest

Rate

Fiscal Real

Exchange

Rate

Inflation

Real GDP

1 0.18 6.43 3.83

89.56

0.00

2 14.46 9.46 6.55

64.64

4.88

4 14.47 17.54 8.07

54.12

5.79

20 14.54 22.21 11.80

45.12

6.34

Table 7

Variance Decompositions of Real GDP and Real Exchange Rate from Model 3

A. VDC of Real GDP

Horizon/Shock U.S.

Interest

Rate

Real

Exchange

Rate

Real M2 Inflation

Rate Real

GDP

1 1.31 0.37 0.68 0.00 97.63

2 1.30 32.34 4.89 0.34 61.12

4 1.68 30.65 5.19 1.40 61.08

20 2.37 28.45 8.41 3.38 57.39

B. VDC of Real Exchange Rate

Horizon/Shock U.S.

Interest

Rate

Real

Exchange

Rate

Real M2 Inflation

Rate Real GDP

1 0.69 99.31 0.00 0.00 0.00

2 6.49 80.35 0.11 5.10 7.95

4 8.90 73.60 4.62 4.56 8.32

20 8.89 68.81 5.74 6.14 10.42

C. VDC of Inflation Rate

Horizon/Shock U.S.

Interest

Rate

Real

Exchange

Rate

Real M2

Inflation

Rate

Real GDP

1 1.26 2.73 0.47

95.54

0.00

2 0.77 14.41 0.70

79.58

4.54

4 3.20 11.71 9.34

69.33

6.43

20 3.41 12.16 9.78

67.96

6.69

Table 8

Variance Decompositions of Real GDP and Real Exchange Rate from Model 4

A. VDC of Real GDP

Horizon/Shock U.S.

Interest

Rate

Capital

Account Real

Exchange

Rate

Inflation

Rate Real GDP

1 4.09 1.01 0.54 4.55 89.81

2 3.54 1.46 19.97 7.46 67.57

4 2.73 16.87 15.41 10.54 54.46

20 5.52 24.37 15.93 9.81 44.37

B. VDC of Real Exchange Rate

Horizon/Shock U.S.

Interest

Rate

Capital

Account Real

Exchange

Rate

Inflation

Rate Real GDP

1 10.72 0.49 88.80 0.00 0.00

2 10.89 10.17 67.27 6.85 4.82

4 10.63 9.24 64.50 7.73 7.90

20 10.08 9.59 58.30 10.87 11.16

C. VDC of Inflation Rate

Horizon/Shock U.S.

Interest

Rate

Capital

Account Real

Exchange

Rate

Inflation

Rate Real GDP

1 0.49 0.33 13.59 85.59 0.00

2 1.26 5.72 17.85 70.77 4.40

4 3.62 11.17 14.69 63.87 6.65

20 5.00 14.42 15.45 59.01 6.12

Table 9

Variance Decompositions of Real GDP and Real Exchange Rate from Model 5

A. VDC of Real GDP

Horizon/Shock Capital

Account Fiscal Real

Exchange

Rate

Inflation

Rate Real GDP

1 0.38 0.00 3.42 5.78 90.41

2 2.64 5.70 16.55 13.55 61.56

4 23.82 5.02 12.66 11.80 46.70

20 40.18 5.77 11.12 7.93 35.00

B. VDC of Real Exchange Rate

Horizon/Shock Capital

Account Fiscal Real

Exchange

Rate

Inflation

Rate

Real GDP

1 1.45 0.09 98.47

0.00

2 25.58 1.00 70.24

2.62

0.55

4 20.56 11.40 62.10

3.09

2.84

20 18.95 16.95 51.93

5.23

6.93

B. VDC of Real Exchange Rate

Horizon/Shock Capital

Account Fiscal Real

Exchange

Rate

Inflation

Rate

Real GDP

1 3.65 0.16 4.67

91.52

0.00

2 9.22 2.82 7.60

74.93

5.43

4 15.79 12.29 7.80

58.71

5.42

20 20.24 17.17 9.25

47.36

5.96

Table 10

Variance Decompositions of Real GDP and Real Exchange Rate from Model 6

A. VDC of Real GDP

Horizon/Shock Oil Price Capital

Account Real

Exchange

Rate

Inflation

Rate Real GDP

1 3.07 0.01 1.50 2.46 92.97

2 2.25 0.93 18.97 5.81 72.03

4 1.85 18.29 13.91 11.18 54.77

20 8.99 24.16 14.70 9.41 42.73

B. VDC of Real Exchange Rate

Horizon/Shock U.S.

Interest

Rate

Capital

Account Real

Exchange

Rate

Inflation

Rate

Real GDP

1 0.17 1.18 98.65

0.00

2 0.14 13.10 80.28

4.99

1.49

4 2.93 12.46 75.26

5.50

3.83

20 5.70 13.34 67.10

7.82

6.04

B. VDC of Inflation Rate

Horizon/Shock U.S.

Interest

Rate

Capital

Account Real

Exchange

Rate

Inflation

Rate

Real GDP

1 0.00 0.41 9.63

89.96

0.00

2 1.36 5.28 20.41

69.67

3.28

4 2.89 12.92 18.56

61.38

4.25

20 4.68 16.24 16.65

57.87

4.55

Chart 1.

Real Exchange Rate

log of CPI-adjusted peso-value of $

1990 1991 1992 1993 1994 1995 1996

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

Average for period Jan80 to Nov96

Chart 2.

*CPI-deflated U.S. dollar value of peso

Real Exch Rate

Real GDP

Real Exchange Rate* ($/peso) and Seasonally Adjusted Real GDP

1980 1982 1984 1986 1988 1990 1992 1994 1996

-0.27

-0.26

-0.25

-0.24

-0.23

-0.22

13.75

13.80

13.85

13.90

13.95

14.00

14.05

14.10

Chart 3.

Real Exchange Rate vs. Detrended Real GDP

-0.06 -0.04 -0.02 0.00 0.02 0.04 0.06

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

96Q3

Real Exchange Rate (pesos/$)

Deviation of Real GDP from a Linear Trend

IB: DN=SN

EB: DT-ST=KA

(KA = 0)

EB (KA < 0)

EB (KA > 0)

RER

=PT/PN

Chart 4.

Chart 5.

Real Exchange Rate vs. Detrended Real GDP

-0.06 -0.04 -0.02 0.00 0.02 0.04 0.06

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

96Q3

EB (KA = 0)

EB (KA < 0)

EB (KA > 0)

Real Exchange Rate (pesos/$)

Deviation of Real GDP from a Linear Trend

Chart 6.

*Seasonally Adjusted Real Manufacturing Wage

Real Wage* Real Exch Rate

Real Manufacturing Wage and the Real Exchange Rate

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996

4.2

4.3

4.4

4.5

4.6

4.7

4.8

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

Real Wage* Real GDP

Real Manufacturing Wage and Seasonally Adjusted Real GDP

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996

4.2

4.3

4.4

4.5

4.6

4.7

4.8

13.75

13.80

13.85

13.90

13.95

14.00

14.05

14.10

Chart 7.

Real Money Supply and the Real Exchange Rate

1980 1982 1984 1986 1988 1990 1992 1994 1996

1.50

1.75

2.00

2.25

2.50

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

Real M2

Real Exch Rate

Real Money Supply and Seasonally Adjusted Real GDP

1980 1982 1984 1986 1988 1990 1992 1994 1996

1.50

1.75

2.00

2.25

2.50

13.75

13.80

13.85

13.90

13.95

14.00

14.05

14.10

Real M2

Real GDP

Chart 8.

*Nonseasonally Adjusted Merchandise Trade Balance (U.S. dollars)

Trade Balance* Real Exch Rate

Merchandise Trade Balance and the Real Exchange Rate

1980 1982 1984 1986 1988 1990 1992 1994 1996

-2000

-1500

-1000

-500

500

1000

1500

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

Trade Balance* Real GDP

Merchandise Trade Balance and Seasonally Adjusted Real GDP

1980 1982 1984 1986 1988 1990 1992 1994 1996

-2000

-1500

-1000

-500

500

1000

1500

13.75

13.80

13.85

13.90

13.95

14.00

14.05

14.10

Chart 9.

*Monthly rate of CPI inflation, compounded to an annual rate

Inflation and the Real Exchange Rate

1980 1982 1984 1986 1988 1990 1992 1994 1996

100

120

140

160

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

Inflation*

Real Exch Rate

Inflation and Real GDP

1980 1982 1984 1986 1988 1990 1992 1994 1996

100

120

140

160

13.75

13.80

13.85

13.90

13.95

14.00

14.05

14.10

Inflation*

Real GDP

Chart 9.

*Monthly rate of CPI inflation, compounded to an annual rate

Inflation and the Real Exchange Rate

1980 1982 1984 1986 1988 1990 1992 1994 1996

100

120

140

160

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

Inflation*

Real Exch Rate

Inflation and Real GDP

1980 1982 1984 1986 1988 1990 1992 1994 1996

100

120

140

160

13.75

13.80

13.85

13.90

13.95

14.00

14.05

14.10

Inflation*

Real GDP

Chart 10.

*Log of 91-Day Cetes Rate

Nominal 91-Day Cetes Interest Rate and the Real Exchange Rate

1980 1982 1984 1986 1988 1990 1992 1994 1996

2.0

2.5

3.0

3.5

4.0

4.5

5.0

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

Nom. Int. Rate*

Real Exch Rate

Nominal 91-Day Cetes Interest Rate and Seasonally Adjusted Real GDP

1980 1982 1984 1986 1988 1990 1992 1994 1996

2.0

2.5

3.0

3.5

4.0

4.5

5.0

13.75

13.80

13.85

13.90

13.95

14.00

14.05

14.10

Nom. Int. Rate*

Real GDP

Chart 11.

*Domestic Credit Granted by the Banking System to the Private Sector

Stock of Real Bank Credit and the Real Exchange Rate

1980 1982 1984 1986 1988 1990 1992 1994 1996

0.75

1.00

1.25

1.50

1.75

2.00

2.25

2.50

2.75

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

Real Bank Credit

Real Exch Rate

Stock of Real Bank Credit and Seasonally Adjusted Real GDP

1980 1982 1984 1986 1988 1990 1992 1994 1996

0.75

1.00

1.25

1.50

1.75

2.00

2.25

2.50

2.75

13.75

13.80

13.85

13.90

13.95

14.00

14.05

14.10

Real Bank Credit

Real GDP

Chart 12.

Govt Spending Real Exch Rate

Government Spending and the Real Exchange Rate

1980 1982 1984 1986 1988 1990 1992 1994 1996

11.94

12.00

12.06

12.12

12.18

12.24

12.30

12.36

12.42

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

Govt Spending Real GDP

Government Spending and Seasonally Adjusted Real GDP

1980 1982 1984 1986 1988 1990 1992 1994 1996

11.94

12.00

12.06

12.12

12.18

12.24

12.30

12.36

12.42

13.75

13.80

13.85

13.90

13.95

14.00

14.05

14.10

Chart 13.

*Real 91-Day Cetes Rate, compounded annually

Real 91-Day Cetes Interest Rate and the Real Exchange Rate

1980 1982 1984 1986 1988 1990 1992 1994 1996

-27

-18

-9

-4.5

-4.4

-4.3

-4.2

-4.1

-4.0

-3.9

-3.8

-3.7

Real Int. Rate*

Real Exch Rate

Real 91-Day Cetes Interest Rate and Seasonally Adjusted Real GDP

1980 1982 1984 1986 1988 1990 1992 1994 1996

-27

-18

-9

13.75

13.80

13.85

13.90

13.95

14.00

14.05

14.10

Real Int. Rate*

Real GDP

Figure 1: Impulse Responses of Real GDP

Model 1

1 4 7 10131619

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015 US interest rate

Real exch. rate

Inflation

Real GDP

Model 2

1 4 7 10131619

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015 US interest rate

Fiscal

Real exch. rate

Inflation

Real GDP

Model 3

1 4 7 10131619

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015 US interest rate

Real exch. rate

Inflation

Real GDP

Model 4

1 4 7 10131619

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015 US interest rate

Capital account

Real exch. rate

Inflation

Real GDP

Model 5

1 4 7 10131619

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020 Capital account

Fiscal

Real exch. rate

Inflation

Real GDP

Model 6

1 4 7 10131619

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015 Oil price

Capital account

Real exch. rate

Inflation

Real GDP

Figure 2: Impulse Responses of Inflation

Model 1

1 4 7 10131619

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035 US interest rate

Real exch. rate

Inflation

Real GDP

Model 2

1 4 7 10131619

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

0.030 US interest rate

Fiscal

Real exch. rate

Inflation

Real GDP

Model 3

1 4 7 10131619

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

0.030 US interest rate

Real exch. rate

Inflation

Real GDP

Model 4

1 4 7 10131619

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035 US interest rate

Capital account

Real exch. rate

Inflation

Real GDP

Model 5

1 4 7 10131619

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

0.030 Capital account

Fiscal

Real exch. rate

Inflation

Real GDP

Model 6

1 4 7 10131619

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

0.030

0.035 Oil price

Capital account

Real exch. rate

Inflation

Real GDP

Figure 3: Impulse Responses of Real Exchange Rate

Model 1

1 4 7 10131619

-0.025

0.000

0.025

0.050

0.075

0.100

0.125 US interest rate

Real exch. rate

Inflation

Real GDP

Model 2

1 4 7 10131619

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10 US interest rate

Fiscal

Real exch. rate

Inflation

Real GDP

Model 3

1 4 7 10131619

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10 US interest rate

Real exch. rate

Inflation

Real GDP

Model 4

1 4 7 10131619

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100 US interest rate

Capital account

Real exch. rate

Inflation

Real GDP

Model 5

1 4 7 10131619

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100 Capital account

Fiscal

Real exch. rate

Inflation

Real GDP

Model 6

1 4 7 10 13 16 19

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10 Oil price

Capital account

Real exch. rate

Inflation

Real GDP

Figure A-1

Impulse Response Functions with Standard Errors, Model 1

Response of:

Shock to:

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

-0.012

-0.006

0.000

0.006

0.012

0.018

0.024

0.030

0.036

0.042

-0.012

-0.006

0.000

0.006

0.012

0.018

0.024

0.030

0.036

0.042

-0.012

-0.006

0.000

0.006

0.012

0.018

0.024

0.030

0.036

0.042

-0.012

-0.006

0.000

0.006

0.012

0.018

0.024

0.030

0.036

0.042

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

Figure A-2

Impulse Response Functions with Standard Errors, Model 2

Response of:

Shock to:

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

1.25

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

1.25

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

1.25

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

1.25

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

1.25

-0.006

-0.004

-0.002

0.000

0.002

0.004

0.006

-0.006

-0.004

-0.002

0.000

0.002

0.004

0.006

-0.006

-0.004

-0.002

0.000

0.002

0.004

0.006

-0.006

-0.004

-0.002

0.000

0.002

0.004

0.006

-0.006

-0.004

-0.002

0.000

0.002

0.004

0.006

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.025

0.000

0.025

0.050

-0.025

0.000

0.025

0.050

-0.025

0.000

0.025

0.050

-0.025

0.000

0.025

0.050

-0.025

0.000

0.025

0.050

-0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

Figure A-3

Impulse Response Functions with Standard Errors, Model 3

Response of:

Shock to:

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

-0.014

-0.007

0.000

0.007

0.014

0.021

0.028

0.035

-0.014

-0.007

0.000

0.007

0.014

0.021

0.028

0.035

-0.014

-0.007

0.000

0.007

0.014

0.021

0.028

0.035

-0.014

-0.007

0.000

0.007

0.014

0.021

0.028

0.035

-0.014

-0.007

0.000

0.007

0.014

0.021

0.028

0.035

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

Figure A-4

Impulse Response Functions with Standard Errors, Model 4

Response of:

Shock to:

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

0.05

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

0.05

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

0.05

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

0.05

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

0.05

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

Figure A-5

Impulse Response Functions with Standard Errors, Model 5

Response of:

Shock to:

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

-0.0050

-0.0025

0.0000

0.0025

0.0050

0.0075

-0.0050

-0.0025

0.0000

0.0025

0.0050

0.0075

-0.0050

-0.0025

0.0000

0.0025

0.0050

0.0075

-0.0050

-0.0025

0.0000

0.0025

0.0050

0.0075

-0.0050

-0.0025

0.0000

0.0025

0.0050

0.0075

-0.075

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.075

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.075

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.075

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.075

-0.050

-0.025

0.000

0.025

0.050

0.075

0.100

0.125

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

-0.02

-0.01

0.00

0.01

0.02

0.03

0.04

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

Figure A-6

Impulse Response Functions with Standard Errors, Model 6

Response of:

Shock to:

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

-0.06

-0.04

-0.02

0.00

0.02

0.04

0.06

0.08

0.10

0.12

-0.018

-0.009

0.000

0.009

0.018

0.027

0.036

0.045

-0.018

-0.009

0.000

0.009

0.018

0.027

0.036

0.045

-0.018

-0.009

0.000

0.009

0.018

0.027

0.036

0.045

-0.018

-0.009

0.000

0.009

0.018

0.027

0.036

0.045

-0.018

-0.009

0.000

0.009

0.018

0.027

0.036

0.045

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

0.015

0.020

When are devaluations more contractionary? A Quantile VAR estimation for Argentina

Article

Full-text available

Mar 2022

This paper presents empirical evidence on the short- and medium-run contractionary effects of exchange rate shocks and currency devaluations for bimonetary (i. e., highly dollarized) countries. In particular, for Argentina for the period January 2004 December 2018. Using a VAR representation with quantile heterogeneity, it implements a multivariate model with four macroeconomic variables: exchange rate variations, inflation, economic activity and nominal wage growth. The empirical results show a 30% price pass-through effects and a bimodal effect on output, with both positive and negative effects. Wages ad-just less than prices with the consequent effect that real wages have a negative elasticity of 0.23 with respect to exchange rate shocks. Further analysis on the multivariate responses show that the negative effect on output is associated with a decline in real wages: a 1% fall in real wages after a currency devaluation produces a 2.3% decline in output.

Effects of real effective exchange rate volatility on export earnings in Ethiopia: Symmetric and asymmetric effect analysis

Article

Full-text available

Dec 2023

Obsa Urgessa

The exchange rate is one of the key monetary policy instruments used to dictate the economy. However, the effect of exchange rate volatility on macroeconomic outcomes has been a subject of debate in previous research. Thus, this study examined the effects of real effective exchange rate (REER) volatility on Ethiopia's export earnings using quarterly data covering 2007 to 2021. The GARCH model is employed to estimate the volatility of the real effective exchange rate, and its effect is examined at the aggregate and disaggregate levels. Disaggregation is made at commodity and partner country levels. Besides these, the study also examined the symmetric and asymmetric effects of exchange rate volatility on the three categories of export earnings. To estimate the effects, both the linear autoregressive distributed lag (ARDL) and nonlinear ARDL (NARDL) models are employed. The short-run ARDL result shows that the real effective exchange rate and its volatility affect export earnings only in a few cases. Real effective exchange rate appreciation is significant and reduces total export earnings and export earnings from vegetables, meat, and oilseed-related products. Similarly, exchange rate volatility reduces total export earnings and earnings from coffee, vegetables, and meat-related products. At the aggregate level, the NARDL result shows that exchange rate volatility has an asymmetric effect on total export earnings. Positive volatility reduces total export earnings, while negative volatility is not significant. At the commodity level, volatility also has an asymmetric effect on earnings from coffee and meat exports in the short run. However, in the long run, there is no asymmetric effect of exchange rate volatility on total and commodity-level export earnings. In terms of the country of destination, there is no strong evidence of the effect of real effective exchange rate appreciation and its volatility on export earnings. Thus, policies that stabilize the volatility of the exchange rate are needed in Ethiopia in order to enhance export earnings.

Inverse Balassa–Samuelson effect in Mexico: the role of the oil sector

Article

Full-text available

May 2023
EMPIR ECON

In contrast to the oil boom in the USA, Mexico’s oil production has been declining continuously for the last 15 years. Lower output in the sector has affected various aspects of Mexico’s economy, one of which we found to be the real exchange rate: the decrease in productivity in the oil sector compared to the non-tradable sector, which in turn has caused a depreciation of the real exchange rate. This is an inverse Balassa–Samuelson effect, and the experience of Mexico serves as a wake-up call to countries whose economies are highly dependent on non-renewable commodities. To perform our analysis, we begin by calculating productivity by sector as total factor productivity. We then estimate the determinants of the real exchange rate by including the productivity of Mexico’s oil- and non-oil tradable sectors in order to disentangle their contributions.

Trade Integration in ECOWAS: Any Evidence of Market Size Effects and Knowledge Transfer?

Article

Full-text available

Mar 2022

Chukwuemeka Onyebuchi Onyimadu

This paper examined the degree of trade integration among selected ECOWAS countries vis-a-vis market size effects and knowledge transfer. In specific, the study examined if smaller countries in ECOWAS have benefitted from improved access to larger markets and tested the proposition that trade integration is capable of inducing knowledge transfers from more advanced to less advanced countries. Data from 2008–2018 sourced from the United Nation’s UNCOMTRAD database and a sample of 10 ECOWAS countries were used to estimate a GMM panel regression. The estimated GMM panel regression revealed evidence of a causal relationship between annual growth and trade integration, country size, initial income, macroeconomic control variables, and interacting terms between trade integration, country size, and initial incomes. Specifically, the empirical findings indicated that size matters for economic performance, with smaller countries benefiting more from increased trade integration relative to larger countries. In addition, the study found that through increased trade integration, technologically more advanced countries transfer knowledge to less technologically advanced countries. These findings are consistent with the extant literature. Based on the findings, it is therefore recommended that for the ECOWAS subregion to maximize the benefits from increased market access and experience knowledge transfer that are associated with trade integration, the composition of its bilateral trade flows must be focused on manufacturing rather than mineral fuels and oils.

Devaluaciones contractivas en América Latina. Un análisis VAR de los casos de Argentina, Chile y Colombia

Conference Paper

Full-text available

Nov 2021

Nicolas Bertholet

Resumen Utilizando metodología VAR bajo un esquema recursivo, el trabajo estima los efectos del tipo de cambio nominal sobre la inversión, el producto y la balanza comercial en Argentina, Chile y Colombia para el periodo 1950-2010. Con el objetivo de analizar la sensibilidad de los resultados obtenidos con respecto al periodo y frecuencia utilizada, también se estimaron VARs trimestrales para un periodo más reciente. Los resultados obtenidos mediante las funciones impulso-respuesta y descomposición de varianza muestran que los shocks cambiarios tienen: a) un efecto positivo sobre la balanza comercial tanto en el corto como en el mediano plazo. b) un efecto contractivo de corto plazo en el producto que está liderado por la inversión y que en el mediano plazo dicho efecto se revierte parcial o totalmente. Mediante el análisis de descomposición de varianza también encontramos que en la mayoría de los casos los shocks cambiarios explican una parte importante de la varianza de la inversión, el producto y la balanza comercial.

DO EXCHANGE RATES MATTER FOR ETHIOPIA'S EXPORT?

Thesis

Full-text available

Jan 2019

Samuel Elias Kayamo

This study estimated the long and short-run effects of the real exchange rate on Ethiopian export earnings by using the ARDL panel approach. The study used balanced annual panel data for the period 1997-2016. The results of the study showed that the real exchange rate has a significant impact on export earnings in the long-run, while it has no effect in the short-run. This implies that the depreciation of the real exchange rate improves export earnings of the country.

Economic Growth Driven by the Mexican Fourth Transformation Policy

Article

Feb 2024
Rev Polit Econ

The economic policy of the fourth transformation (4T) of Mexico - as the actual president Andrés Manuel López Obrador defines his mandate - is characterized by being cautious and responsible for public spending, making investment decisions more effective, and redistributing wealth to sustain domestic consumption. The internal demand has been promoted by increasing poor people's disposable income through subsidies and transfers and several policies aimed at raising minimum wages. As a framework for these direct policies to boost economic growth, the fight against corruption helped achieve these results more effectively. This work aims to analyze whether the 4T policy increased economic growth in Mexico. We applied two econometric techniques to test our hypothesis: interrupted time series analysis, controlling for robustness of the results, and breakpoint regression analysis. In both cases, we robustly showed that the 4T economic policy has positively impacted economic growth in Mexico.

Nexus between Exchange Rate Spill-over and Macroeconomic Determinants in South Asian Countries

Article

Full-text available

May 2023

This paper examines the relationship between the Exchange Rate spill-over and Macro Economic Determinants in selected South Asian Counties (Afghanistan, India and Pakistan) employing 10 years of data from 2011 to 2021. We use macroeconomic variables-Total Investment (TIN), Inflation (INF), Imports (IMP), Exports (EXP), Government Total Expenditure (GTE), Government Net Lending (GNL) and Exchange Rate (ER). Result evidenced that selected macroeconomic variables changes do not have any connection in the changes of exchange rate of India, Pakistan and Afghanistan. The exchange rate changes are highly autocorrelated and act highly independence in the last ten years. Moreover, the depreciation of currency values replicates economic instability and drastically straight fall. Asian countries operating their currency based on a managed floating exchange rate regime. It is marking a market-determined exchange rate regime of the currency with provision for timely intervention by the central bank. This study would be useful to the retail investors to speculate and arbitrage and (economic policy) policymakers to monitor, forecasting and simulation the exchange rate movements.

Revisiting the Dynamic MacroeconomicEffects of Real Exchange Rate and Economic growth: Application of Structural Vector Autoregressive (SVAR) Analysis

Article

Full-text available

Jan 2021

Resource Dependence, Commodity Shocks and the Role of the Exchange Rate: An Empirical Study of Papua New Guinea

Article

Jan 2023

Rubayat Chowdhury

Targeting the Real Exchange Rate: Theory and Evidence

Article

Full-text available

Jan 1994

Targeting the Real Exchange Rate: Theory and Evidence

Article

Full-text available

Jan 1994

Exchange-Rate-Based Stabilization under Imperfect Credibility

Article

Full-text available

Jan 1991

In the late 1970s, the Southern-Cone countries comprising Argentina, Chile and Uruguay launched stabilisation programmes based on a pre-announced path for the exchange rate that exhibited a declining rate of devaluation — the Tablitas.2 Policymakers expected a stabilisation plan based on the exchange rate as the nominal anchor to act directly on inflationary expectations — which were viewed as a central determinant of short-run inflation — thus increasing the chances of lowering inflation at no significant real cost.

The Syndrome of Exchange-Rate-Based Stabilization and the Uncertain Duration of Currency Pegs

Article

Apr 1996

This paper conducts a quantitative examination of the hypothesis that uncertain duration of currency pegs causes the sharp real appreciations and business cycles that affect chronically countries using fixed exchange rates as an instrument to stop high inflation. Numerical solutions of equilibrium dynamics of a two-sector small open economy with incomplete markets show that uncertain duration rationalizes the syndrome of exchange-rate-based stabilizations without price or wage rigidities. Three elements of the model are critical for these results: (a) a strictly-convex hazard rate function describing time-dependent devaluation probabilities, (b) the wealth effects introduced by incomplete insurance arkets, and (c) the supply-side effects introduced via capital accumulation and elastic labor supply. Uncertain duration also entails large welfare costs, compared to the perfect-foresight credibility framework, although temporary disinflations are welfare-improving. The model's potential empirical relevance is examined further by reviewing Mexico's post-war experience with the collapse of six currency pegs.

Exchange-Rate Based Inflation Stabilization : The Initial Real Effects of Credible Plans

Article

Mar 1995

Martin Uribe

This paper presents a dynamic general equilibrium model of a small, open, monetary economy in order to analyze the short-run effects of credible stabilization plans that fix the nominal exchange rate in a regime of free convertibility. In this model inflation acts as a tax on domestic market transactions. In particular, it generates a wedge between the rate of return on investment in domestic capital and the rate of return on investment in foreign assets. The model stresses the importance of adjustment costs (including gestation lags) in explaining the precise character of the initial dynamics. The main stylized facts of this type of programs namely an initial phase characterized by several months of real exchange rate appreciation, trade balance deterioration and expansion in aggregate demand and production, followed by a deflationary slowdown in real activity, are replicated without resorting to credibility problems, sticky prices, adaptive expectations, or gradual disinflation schemes. Finally, the model is calibrated using long-run relations from the Argentinean economy, and its quantitative predictions are compared to the initial effects of that country's Convertibility Plan of April 1991.

Output and the Real Exchange Rate in Developing Countries: An Application to Mexico

Article

Jan 1997

Since Mexico's devaluation of the peso in 1994, some observers have called for policies designed to keep the real exchange rate highly competitive in order to promote exports and output growth. However, over the past few decades, devaluations of the real exchange rate have been associated nearly exclusively with economic contraction, while real appreciations have been followed almost invariably by expansions in economic activity. The purpose of this paper is to attempt to disentangle the possible factors underlying this correlation--(1) reverse causation from output to the real exchange rate, (2) spurious correlation with third factors such as capital account shocks, and (3) temporary contractionary effects of devaluation--and determine whether, once those factors are accounted for, a positive, long-run effect of real depreciation on output can be identified in the data. Based on the results of a VAR model designed to explore the linkages between the real exchange rate and output, we conclude that even after sources of spurious correlation and reverse causation are controlled for, real devaluation has led to high inflation and economic contraction in Mexico. While changes in Mexico's economic structure and financial situation may qualify the future applicability of this conclusion, we view our findings as pointing to substantial risks to targeting the exchange rate at too competitive a level.

Economic Control in a Dependent Economy

Article

Jan 1960

T. Swan

UN MODELO DE COINTEGRACIÓN PARA PRONOSTICAR EL PIB DE MÉXICO

Article

Currency Devaluation in Developing Countries

Article

Jan 1971

Richard Cooper

This chapter describes the currency devaluation in developing countries. In analyzing devaluation, the exact nature of the initial disequilibrium is important and much analysis misleads by its focus on economies that are assumed to be in equilibrium at the moment of devaluation. The elasticities approach focuses on the substitution among commodities, both in consumption and in production induced by the relative price changes wrought by the devaluation. The principal relative-price change is between goods, whether imported or exported, whose price is strongly influenced by conditions in the world market, and those home goods and services that are not readily traded. For a small country, it can be assumed that the prices in domestic currency of foreign-trade goods will rise by the amount of devaluation. This rise will divert purchases out of the existing income to nontraded goods and services, thereby reducing domestic demand for imports and for export goods, releasing the latter for sale abroad.

A Note on the Impact of Devaluation and the Redistributative Effects

Article

Jan 1963

Carlos F. Diaz-alejandro

Output and the Real Exchange Rate in Developing Countries: An Application to Mexico

Abstract and Figures

Recommended publications

EXCHANGE RATE PASS-THROUGH IN TURKEY: NONLINEAR PROPAGATION OF SHOCKS APPROACH(TÜRKİYE’DE DÖVİZ KURU...

Dynamics Of Indonesia's International Trade a VAR Approach

Una Revisión de la Transmisión Monetaria y el Pass-Through en Chile

Exchange rate pass-through into inflation in Vietnam: evidence from VAR model