Transcript Exchange rate, output and employment: revisiting the

Exchange rate, output and
employment:
revisiting the contractionary
devaluation hypothesis
Saúl Keifman
Universidad de Buenos Aires
I. Introduction
• Purpose: discuss the relationship between the
exchange rate and employment.
• Motivation: understand some recent behavioral
changes in the Argentinean economy to intervene in the
current macropolicy debate.
• Debate: what to do with the exchange under a current
account and balance of payments surplus.
Two opposite views:
(a) Heteredox economists support Kirchner’s
competitive (undervalued) real exchange rate policy to
sustain output and employment expansion, relying on
sterilization and some capital controls.
(b) Orthodox economists propose a clean float and
inflation targeting, which would lead to significant real
appreciation.
• Novelty: in the past, heteredox Argentinean economists,
criticized IMF adjustment programs based on
structuralist models that proved that devaluations were
contractionary, contrary to their current view.
• Approach: this paper attempts to revisit the current
relevance of the theory of contractionary devaluations
which arose from structuralist macroeconomics in the
seventies and that has been recently discarded by many
of their former proponents. Though the article refers to
the Argentinean case, we believe this revisionist exercise
presents a more general interest for two reasons:
(1) The contractionary devaluation hypothesis was very
popular in the developing world.
(2) From a methodological viewpoint this paper presents
macroeconomic models that take into account the
structural characteristics of a specific country.
II. Devaluation and employment.
Structuralism from the seventies to the new
millennium
• Structuralism: comprehensive vision of developing
countries problems in opposition to orthodoxy.
• Contractionary orthodox adjustment and stabilization
programs applied in Latin America inspired alternative
macroeconomic models. Two outstanding examples:
(a) Structural inflation theory
(b) Contractionary devaluation theory.
• Background: in a Keynesian open economy model,
achievement of internal balance (full employment and
price stability) and external balance (sustainable current
account balance) required two kinds of policies:
expenditure-switching policies, to reallocate expenditure
between tradable and nontradable goods, and
expenditure-reducing policies, to control aggregate
expenditure or absorption. If Marshall-Lerner conditions
held, devaluations would improve trade balance and also
raise output. Devaluations would be expansionary.
• Right policy mix depended on initial imbalances.
• But orthodox recipe for balance of payments deficits
always recommended a currency devaluation combined
with tight monetary and fiscal policies (e.g. IMF during
developing countries foreign debt crisis in the early ´80s
or in “emerging market” crises during the ´90s). Note that
even if devaluations were expansionary, the treatment of
a balance of payments crisis under conditions of idle
capacity and high unemployment, should not include
expenditure-reducing policies, that is to say, a fiscal and
monetary contraction, since the latter would hinder the
achievement of internal balance.
• The theory of contractionary devaluations explain the
recessionary bias of orthodox adjustment programs:
Ferrer (1963), Díaz-Alejandro (1963, 1965) and Braun
and Joy (1968). It is based on some typical features of
the Argentinean economy in the sixties and fifties:
(1) The export good is the wage-good: food;
(2) Domestic food supply is inelastic but foreign food
demand is perfectly elastic;
(3) The price elasticity of the domestic demand for food
is less than one;
(4) Manufacturing production is oriented to the domestic
market and its supply curve is very elastic;
(5) Imported goods are not substitutes for, but rather
complementary to domestic production (nonproduced
manufacturing inputs);
(6) Workers have a marginal propensity to spend on
domestic goods greater than urban and rural property
owners.
• Structuralist explanation emphasizes the redistributive
mechanism. A devaluation makes food more expensive
and reduces real wages. This has a larger impact on the
domestic demand for other goods (manufactures)
because of the low price elasticity of domestic food
demand. Given the inelasticity of the exportable good
supply, its price increase does not induce higher output
or employment in the food sector. The fall in workers’
manufactures demand is not offset by a higher demand
from higher profits or rents in the exportable good sector.
This can be due to the post-Keynesian assumption that
workers have a marginal propensity to consume higher
than capitalists, or because higher savings are not
translated into a higher demand for domestic capital
goods (negative accelerator effect in manufacturing or
the increase in the interest rate), or because higher
profits or rents in the exportable sector are channeled to
import goods demand.
• The contractionary view of devaluation spread outside
Latin America Krugman and Taylor (1978), and became
popular in the ’80s (Hanson, 1983; Katseli, 1983)
providing a robust critique to the excessive adjustment
external caused the orthodox recipe and known as
“overkill”.
• Krueger (1983) replied the critique by blaming the
contractionary effect of devaluations on the “distortions
caused by the import substitution strategy”. In her view,
opening up the economy would give domestic industry
enough competitiveness as to be able to substitute
imports efficiently in response to a devaluation.
• Before the ´90s, most Argentinean macroeconomists
supported the contractionary view of devaluations
regardless of their agreement with structuralism in toto.
The macroeconomic experience from the ´60s to the
´80s had been consistent with the former.
• The severe consequences of the currency board
regime, locally known as Convertibility which set a one
to one parity between the Argentinean peso and the US
Dollar, from April 1991 to December 2001, had a deep
impact on heterodox and structuralist economists views.
• The appearance of two-digit unemployment rates since
1993 in a context of strong real appreciation of the peso
and the strong negative impact of currency depreciation
in Brazil (Argentina’s main trade partner) on Argentinean
manufacturing output in the late nineties, weakened the
contractionary devaluation view.
• The fast recovery of output and employment which
started in the second quarter of 2002 just after the
devaluation had a devastating effect on the former.
• As a result, some structuralist and heterodox authors
such as Roberto Frenkel and Aldo Ferrer emphasized
the paramount importance of relative prices and the
need of a “competitive [undervalued] exchange rate”
to foster output and employment expansion.
• Curious fact: structuralist and orthodox economists
traded places in terms of their attitudes towards the
relevance of real exchange rates!
• In the ´60s, in a world of fixed exchange rates and low
international capital mobility, orthodox economists
recommended devaluations to “get prices right”, in order
to stimulate the production of exportable agriculture.
Structuralist economists dissented not only for their
contractionary effects but also because they distrusted
price incentives as they emphasized role of “structural
features” such as land tenure and the technological and
financial support from the state.
• Since the ´90s, in a world of financial globalization,
more liberal trade and flexible exchange rates, the new
monetarist orthodoxy associated with the “equilibrium
approach to exchange rates” considers that the real
exchange rate and the current account balance are
irrelevant.
• Given the historically conditioned character of
structuralist models, this turn does not necessarily go
against the methodological principles of the
structuralist approach.
• Insofar as the structure of the Argentinean economy has
changed, it should not be discarded that its behavioral
relations have changed too. However, in contrast to the
hypotheses developed in the ´60s and ´70s that were
supported by formal models and detailed discussions
about the features of the Argentinean economy, the
recent turn has not offered comparable foundations so
far.
III. Summary of three macroeconomic
models
• We built three structuralist models intended for Argentina
that examine the relationship between the real exchange
rate (given by the ratio of the exchange rate to the
nominal wage rate) and employment (see Appendix).
We summarize the main assumptions and qualitative
results.
• The models attempt to capture some structural changes
in the economy to assess its theoretical impact on the
relationship between the exchange rate and
employment:
(a) Production of food, the exportable good is price
sensitive (a stylized fact).
(b) We explicitly take into account the impact of trade
liberalization (semi-closed economy versus open
economy) and the presence of domestic production of
importable manufactures.
• In all models we obtain an aggregate labor demand that
we identify with an employment function under the
simplifying assumption that labor supply is greater than
labor demand at the prevailing market wage and
inelastic to wages.
• The effect of real devaluation (an increase in the
exchange rate/wage ratio) on employment is positive at
the minimum level of the exchange rate/wage ratio.
For some values of the model parameters, devaluation
will always be expansionary but for not for others. The
employment function could have interior maximum, so
that devaluation would be expansionary for low levels of
the exchange rate and contractionary for high levels of
the exchange rate. In other words relationship could
have and inverted U shape (see Figure, the 1).
FIGURE 1
Employment
Exchange rate/wage rate
• An important result is that the possibility of an inverted
U-shape employment function is independent of the
tariff level, which refutes Krueger’s conjecture that
devaluation cannot be contractionary in an open
economy.
• A decrease in tariffs shifts the employment function
downwards and to the right (Figure 1), so that:
(1) Employment will be lower in a more open economy
(given the nominal exchange rate and wage rate).
(2) The exchange rate to wage ratio that maximizes
employment will be higher in a more open economy.
• The shift of the employment function caused by tariff
reduction implies that there is an interval of the
exchange rate/wage ratio for which devaluation would go
from being contractionary before the tariff reduction, to
expansionary, after the tariff reduction, although with
lower levels of employment.
IV. An interpretation of the Argentinean
experience
• The traditional structuralist interpretation placed
emphasis on a monotonically decreasing relationship
between the exchange rate and employment, that was
consistent with the Argentinean experience until the
nineties.
• The recent structuralist interpretation, that has
captured the experience of the last decade, proposes a
reversion of the relationship posed before, and stresses
a monotonically increasing relationship between the
exchange rate and employment. The change in stance
is made without discussing the probable determinants of
the reversion.
• Our point is that employment function could be an
inverted U.
• Two central features of the “Convertibility” policy
experiment were:
(a) Strong real appreciation of the peso, which implied
an important reduction of the exchange rate/wage ratio.
(b) fast, across the board, and deep trade liberalization.
• The significant fall in the exchange rate/wage ratio could
have led the economy from the decreasing interval to the
increasing interval of the employment function.
• On the other hand, the shift of the employment curve
caused by tariff reductions probably strengthened the
former effect because the slope of the employment
function turned from negative to positive in some interval
of the exchange rate/wage ratio.
• The recent structuralist interpretation and ours have
some similar local predictions. But, they might have
different policy implications. The recent structuralist
view assumes that a higher real exchange rate will
always be beneficial for employment. Of course, there
would still be a conflict between the real exchange
rate and real wages, something very important for
income distribution and poverty.
• According to our interpretation, there could be a level of
the real exchange rate that maximizes (ceteris paribus)
the level of employment. If the current real exchange
rate was greater than the employment maximizing
level, there would be room to increase both real
wages and employment, because the economy would
lie in the decreasing interval of the employment curve.
• From our point of view, the moderate real appreciation
that followed the initial overshooting of the nominal
and real exchange rate after the collapse of the
currency board was not necessarily damaging to
employment creation (let alone, income distribution).
On the other hand, if an attempt was made to raise the
real exchange rate to reverse the moderate appreciation
with the purpose of raising employment, the effects could
be contrary to those expected.
V. Conclusions
• This paper has two goals. On the one hand, to stimulate
the discussion of macroeconomic models intended to
capture each country’s specific traits, in our case,
Argentina, thus resuming the structuralist
methodological tradition.
• On the other hand, we are interested in pointing out that
macroeconomic policymaking should not be based
on an empiricism of trends and variables associations
which could be shortlived, and emphasize the need to
do more research on these issues in order to make
policies that help best to overcome our severe social
crisis. For example, such empiricism led to underestimate the job creation ability of output growth in the
last decade, by extrapolating the nineties low
employment-output elasticity. Post-convertibility strong
recovery in output and employment buried this
hypothesis. To be cautious, we believe that we should
not limit ourselves to the extrapolation of recent trends in
employment and exchange rates. Instead, we should
attempt to understand them in a comprehensive
analytical framework.