You are on page 1of 16

The Suntory and Toyota International Centres for Economics and Related Disciplines

Public Sector Pricing and the Real Exchange Rate

Author(s): John Devereux and Michael Connolly
Source: Economica, New Series, Vol. 60, No. 239 (Aug., 1993), pp. 295-309
Published by: Wiley on behalf of The London School of Economics and Political Science and The
Suntory and Toyota International Centres for Economics and Related Disciplines
Stable URL: .
Accessed: 08/10/2014 04:17

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact

Wiley, The London School of Economics and Political Science, The Suntory and Toyota International Centres
for Economics and Related Disciplines are collaborating with JSTOR to digitize, preserve and extend access to

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions
Economica (1993) 60, 295-309

Public Sector Pricing and the Real Exchange Rate

University of Miami

Final versionreceived9 July 1992.

We examinethe effectsof consumptionsubsidies,publicenterprisepricingandpricecontrols

on the real exchangerate. Underplausibleassumptions,we show that permanentreformin
these areasleads to a permanentreal exchangerate appreciation.An empiricalapplication
to the 1990 FujimoriPlan in Peruis discussed.

Consumption subsidies and public enterprise prices are important policy
instruments in developing economies. Such policies are typically used for
redistribution purposes, sometimes even to buy votes by populist leaders. Since
economic reform usually requires the shutting down of 'loans' from the central
bank to parastatals, reducing consumption subsidies and increasing public
enterprise prices, price reform plays an important role in IMF and World
Bank programmes. Edwards (1989a), for example, found that public enterprise
price increases and reductions in consumption subsidies were involved in 79
and 44 per cent respectively of IMF programmes between 1983 and 1985.
Trade reforms, in contrast, were present in just 35 per cent of cases.
To date, the real exchange rate effects of public sector price policies have
received no formal attention from economists.' In this paper, we show that
lasting public sector price reforms produce, under plausible assumptions, a
permanent real exchange rate appreciation. This finding has important policy
implications and may provide a partial explanation for the puzzling real
appreciations that have taken place in many developing countries following
economic reform. Typically, such real appreciations are believed to arise
because of lags in inflationary expectations or because of credibility problems.
It is not widely known that real exchange rate appreciations can also arise
from public sector reforms and that such real appreciations are permanent if
the reform is permanent. Thus, they cannot be removed by devaluation.
In this paper we examine the effects of public sector pricing policies in
three areas on the real exchange rate. First, there are direct subsidies for certain
consumers or producers; examples include food and fertilizer subsidies. The
second group of policies are implicit governmental subsidies arising from
public sector enterprises such as steel, fertilizers and petrochemicals. State
enterprises, in addition, are often important in the transport and public utility
areas.2 Finally, we analyse the effects of selective price controls on the real
exchange rate.3
The paper is organized as follows. In Section I we develop a model of real
exchange rate determination. Our model is a purely real one which abstracts
from the monetary and intertemporal effects of public sector pricing. Of course,
public sector price reforms can take place in conjunction with macroeconomic
stabilization or with trade liberalization. There is, however, no necessary
? The London School of Economics and Political Science 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

connectionbetweenreformin each area.Publicsectorpricereformsalso occur

in economies with stable prices and with externalbalance.Sections II and III
examinethe real exchangerate effectsof consumptionsubsidies,public enter-
prise prices and price control.
Ourresultscan be summarizedas follows. First,a reductionin a consump-
tion subsidyappreciatesthe real exchangeratein all cases. Second,the impact
on the real exchange rate of public sector price increasesand relaxationsin
price controlsis uncertain.The direct effect of reformin each area is to raise
the domestic price level and hence to appreciatethe real exchangerate. But
the indirectreal exchange rate effects of these policies arisingfrom induced
changes in the relative price of non-tradedgoods cannot, in principle, be
signed.Wearguethatthereis neverthelessa presumptionthatthe realexchange
rate appreciatesin responseto public enterpriseprice increases.
The concludingsection illustratesthe crucialrole played by public sector
pricingpolicies in real exchangerate determinationusing data from Peru.We
provideevidencethatpublicsectorpriceincreasesand reductionsin consump-
tion subsidies associated with the August 1990 public sector price reforms
produceda permanentreal exchangerate appreciation.


The model is a real model in the sense that there is no money or otherassets.4
All markets are assumed competitive. We assume that the economy is a
price-takerproducingone non-tradedgood (services) and a large numberof
traded goods, some of which are produced by governmententerprises.The
assumptionthat public sector output is traded is relaxed later. Withoutloss
of generality,let us choose one of the tradedgoods as our numeraire.Govern-
ment enterprisesare assumedto set marginalcost equal to price.
Let us assumethat the economy consists of a single aggregatehousehold.
This allows us to modelthe demandside of the economyusingthe expenditure
function e(.), given in equation (1). We assume that all goods are normal.
National income is given by the GNP or revenuefunction r(*). In equation
(1), p, is the price of the non-tradedgood, p is a vectorof tradedgoods prices,
m is a vector of imports, U is a scalarrepresentingwelfare, V is a vector of
factorendowments,t is a vectorof tradetaxes and s is a vectorof government
consumptionsubsidies. One element of p, the numeraire,is equal to 1. The
superscriptc denotes prices faced by consumers.Note that the firstderivative
of the expenditureand GNP functionswith respectto the ith price provides
compensateddemandand outputsupplyfunctionsrespectively.Thus,the total
demandand supplyof tradablesis given by ep( - ) and rp(- ), which are vectors
of derivativesof the expenditureand revenuefunctionswith respectto traded
For equilibrium,income mustequal expenditureand the non-tradedgoods
marketmust clear. Equation (1) sets expenditureequal to national income
plus tariff revenue minus consumptionsubsidies. We assume that subsidies
are financedby lump-sumtaxes and that tariff revenue is also redistributed
in a lump-sumfashion.
(1) e(pc pSc; U)o= r(p, pa; V)P+otmi - sep.
@) The London School of Economics and Political Science 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

Net imports are given by equation (2):

(2) m = e, (p', p'; U) - r, (p, ps ; V).
The second condition for equilibrium, that the non-traded goods market
clear, is given by equation (3). In this expression es( *) and rs( *) are derivatives
of the expenditure and GNP functions with respect to the price of non-traded
(3) es(p PS; U) =rs(p. Ps; V).

The relationship between domestic and external prices in terms of the

numeraire is given by equation (4) where p* is the world price of the ith good
and ti is the trade tax in this market. Equation (5) gives the relationship
between producer and consumer prices where sj is the per unit consumption

(4) Pi = P*l+ ti,

(5) pi=P+ Sj.
The GNP function provides a general specification of technology. In
particular, no restrictions are placed on the number of goods and factors in
this economy or on the degree of vertical integration in production.7 As our
final simplification, we assume that substitutability in excess supply holds for
all goods. But it should be noted that this assumption is not as restrictive as
it might appear. In the more general case where complementarities exist, our
results apply as correlations which hold in an average sense.
Substituting (2), (4) and (5) into (1) and (3), we are left with a two-equation
system which can be solved for the non-traded price level and welfare as
functions of the exogenous variables.
The real exchange rate, R, is given by equation (6), where P and P* are
the domestic and foreign price indices expressed in the numeraire (or in any
similar units of account):
(6) R = P*/P.
Equations (7) and (8) specify the price indices for the internal and external
price levels. The domestic price index is given by the consumer price index.8
(7) P= Hi(p),,
(8) p* H(p4*)a*i



In this section, we discuss the real exchange rate effects of consumption

subsidies and public enterprise prices. We start with a consumption subsidy.
In dealing with consumption subsidies, it is important to distinguish between
subsidies on goods that are consumed by households, pure consumption
subsidies, and subsidies on goods that are consumed by households and used
as inputs by firms. The latter we refer to as "mixed consumption subsidies".
? The London School of Economics and Political Science 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

Consumption subsidies on traded goods

Pure consumptionsubsidies A reduction in a consumptionsubsidy on the
jth traded good raises prices faced by consumers.But producerprices are
unchanged because they are traded goods. The direct effect of the subsidy
reductionis thereforeto raise the tradedprice level. However,there are also
indirectprice level effectsworkingthroughinducedchangesin the non-traded
price level.
Differentiatingequation (3) and solving for the non-tradedprice level
yields equation(9), where -ss is the compensatedown elasticityof demand
for services, 7sj is the compensated cross elasticity of demand for services with
respect to the price of the jth traded good, ess is the elasticityof supply of
the non-tradedgood, y is realincomeand qYis the incomeelasticityof demand
for services.A circumflexovera variabledenotesa proportionalrateof change.
Changesin real income are defined as dy = eud U, where eu is the inverseof
the marginalutility of income.

(9) Ps cPJo ++9,

7Lsi >0, _ S >0.
77ss+ ESS 77ss+ rss

The firsttermin (9) is a positive substitutioneffect.The second term is a

real income term.Fromsecond-bestconsiderations,the real income effectsof
the subsidy reductionare uncertain.However, for the special case assumed
here, where all other distortions are zero, as shown by (10), real income
increases.Thus,incomeand substitutioneffectsof a decreasein thejth subsidy
raise the non-tradedgoods price level.

(10) dy (1 + eu) = sjjj dsj -sJes dps-

Equation (11), obtainedby combining(6), (7), (9) and (10), summarizes

the relationshipbetween a consumptionsubsidy and the real exchange rate.
Subsidyreductionsraise the domesticprice level relativeto the foreign price
level, therebyappreciatingthe real exchangerate.
(1 1) R = -(aj+a5j)P -a5y 9<0

Proposition1. The elimination of a pure consumptionsubsidy on a traded

good producesa real exchangerate appreciation.
Mixed consumptionsubsidies Subsidies are often placed on goods that are
inputs, such as fertilizers,or on goods that are consumedby householdsand
are used as inputs by firms,such as petrol or electricity.To examinethe real
exchangerate effectsof a mixed consumptionsubsidy, let us assumethat the
economy has a fixed endowmentof 'oil', 0, which is consumedand used as
an input.9The revenuefunctionfor this economy is given by (12) wherepo is
the price of oil:

(12) r(p, p5, pc pp; V)= r'(p, p5, p; V') + po.

? The London School of Economics and Political Science 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

The real income effects of a reduction in the oil subsidy are given by (13),
which is positive.'0 To obtain this result, all other distortions in the economy
are set to zero.

(13) dy (1 + soeou = so(eoo+ r'o) dso - so(eo0+

\ /~~~~~
r'.) dps.
Solving for the effects on the non-traded price level of a reduction in the oil
subsidy, equation (14) is obtained:
(14) PS (w0+0 O)O+y9,
- Eso
00 - >

?lss + Ess

The oil price increase shifts demand towards non-traded goods while raising
costs in the non-traded sector. However, the oil price rise also changes factor
returns, which has indirect effects on the supply of non-traded goods. The
factor price effect can reduce or can even increase the supply of services.
Consequently, the overall effects of a reduction of a mixed consumption
subsidy on the supply of non-traded goods depends on whether the cost-
increasing effects are outweighed by possible changes in factor rewards. If the
cross elasticity of supply of services with respect to the price of oil (eSO) is
negative, then, as shown by (14), the non-traded price level will increase and
the real exchange rate will appreciate.
(15) R =-[ao+ as(wo+ P)]fO-aY9 <0?
Proposition 2. The removal of a subsidy on a good that is consumed and is
used as an input produces a real exchange rate appreciation if non-traded
goods and the subsidized good are substitutes in supply.

Consumption subsidies to non-traded goods: public sector prices

Subsidies are also placed on non-traded goods. The implicit subsidies present
in the prices of many public enterprises clearly fall into this category. Examples
include electricity, transportation and other public utilities. We examine in
this section the real exchange rate effects of an increase in the price of a
publicly provided non-traded good, 'electricity'. This is assumed to be a pure
consumption good. Let us assume that electricity is produced in the public
sector by a single state-owned enterprise. This enterprise sets marginal cost
equal to price. All losses are covered by lump-sum transfers.
The expenditure-equal-to-output condition is given by (16) where pg is the
electricity price. All other distortions are set equal to zero.
(16) e(p,pc,ps; U)= r(p,pg,ps; V)-sgeg.
Equation (17) gives the market-clearing condition for electricity:
(17) eg(p, pg, Ps; U) = rg(p, pg, Ps; V).
The system given by (3), (5), (16) and (17) can be solved for P,Pc and
U. There are two non-traded goods-services and electricity-in this economy.
Given our assumptions about technology and preferences, it can be shown
? The London School of Economics and Political Science 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

thata reductionin the electricitysubsidyraisesthe consumerpriceof electricity.

Owing to income effects, consumer prices can increase by more than the
reductionin the subsidy, but the latter result requiresan implausibleset of
Equation(18) capturesthe relationshipbetween consumerprices and the
subsidy. The ,u parameteris a complex function of elasticities of demand,
supply and income.
(18) pg = sg,
where ,u< 0.
A reductionin the subsidyon electricityraiseswelfare.This resultdepends
on the assumptionthat the public utility sets price equal to marginalcost. If
the public utility acts as a monopolist, the removal of subsidies can reduce
Solving for the effects of a consumption subsidy on the price level of
services,(19) is obtained:
(19) APs AgCg + og g + A9

From (19), note that the public enterpriseprice increaseaffectsthe service

pricelevel throughthreechannels.First,the electricitypricerise shifts demand
towardsservices.The compensated-demandeffect is given by the firstterm in
(19). Second, the electricityprice increase alters the supply of services.The
sign of the supplyeffectdepends on the relationshipbetweenservicesand the
subsidizedactivityin supply given by the term Esg. By assumption,this term
is negative.11Third,the increasein real income raisesthe demandfor services.
The positive real income effect is given by the final term in (19). Thus, the
effects of public sector price increaseson the service price level is uncertain
as demandand supply effectsworkin differentdirections.The directeffectof
a public enterpriseprice increaseis to appreciatethe real exchangerate. But
the overall effects are uncertainas the non-tradedgoods price level effects
cannot, in principle,be signed.
(20) R =-(ag + aswg)pg -asg pg-a Sa
In practice,increasesin the supply of services arisingfrom reductionsin
public sector subsidies are likely to be minor. First, public sector enterprises
are typically capital-intensivein developing economies. Any increase in
employmentin services is thus likely to be small. Second, public enterprise
output is an importantinput into non-tradedproduction.Public sector price
increasesthereforedirectlyincreasethe non-tradedprice level. Thus, there is
a presumptionthat a public sectorpriceincreaseappreciatesthe real exchange
Proposition 3. A public enterpriseprice increase will usually appreciatethe
real exchangerate.


Under generalizedprice controls, official prices no longer reflect economic

scarcity and black marketsoften appear. The removal of generalizedprice
controlswill, almostcertainly,be accompaniedby an increasein the measured
? The London School of Economics and Political Science 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

price level and hence a real exchange rate appreciation. A more common
situation in developing economies is where price controls are selectively applied
to certain goods or services.13 This section provides a general equilibrium
treatment of the relationship between selective price controls and the real
exchange rate.14
Let us assume that price control is imposed on a non-traded good z
produced by the private sector.15 Define the controlled price as jz. We initially
assume that price control is costlessly imposed and perfectly monitored, that
no black markets emerge and that quality is unchanged. We also assume that
the controlled product is allocated to the highest-value users and that rents
arising from rent control are distributed in lump-sum fashion. Admittedly,
these assumptions associated with traditional textbook treatments of price
control are unrealistic. They do, however, allow us to isolate the economic
forces which determine the impact of price controls on the real exchange rate.
The polar case, where all rents are dissipated, is considered later.

Case 1: no waiting in line

Given our assumptions, the demand side of the economy can be represented
by the restricted expenditure function of equation (21). In the restricted
expenditure function the rationed quantity of the non-traded good, z, appears
as an additional argument. The controlled price pz does not appear at all.16
(21) e(p,p,;2; U)-minpx+p,x +pzz s.t U= U(x,xS,Z),z=z
Virtual prices, as developed by Rotbarth (1940-41) and, especially, Neary
and Roberts (1981), allow us to further simplify the analyses. The virtual price
for z, denoted by -i, is that price which induces consumers to purchase just
the rationed amount.17
There are two non-traded prices in this economy: the virtual price of the
controlled good, and the price of services. The virtual price for z is obtained
from (22), where ez is the derivative of the restricted expenditure function
with respect to the quantity of the rationed good. (See Neary and Roberts
(1981) for a formal proof.) From the properties of the restricted expenditure
function, ez is negative. Not surprisingly, a relaxation in the rationing con-
straint reduces the cost of obtaining a given level of utility.
(22) fi =pz -ez.
Equation (23) modifies the income equal to expenditure condition to allow
for price control. With price controls, consumer expenditure evaluated at
virtual prices is set equal to output:
(23) e(p, ps, PZ; U) = r(p, ps, piz; V) + (pz -p#z)rz.
Equation (24) is the market-clearing condition for the price-controlled
(24) ez(p, PS9PZ; U) = rz(p, P,,Pz; V).

Equation (25) solves for the impact of a relaxation in price control on the
service price level. A relaxation of price control increases the producer price
of z, reduces its virtual price and raises real income.'8 As a result, consumers
substitute from services towards the cheaper z. The pure substitution effect is
? The London School of Economics and Political Science 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

capturedby the firsttermin (25). However,the overalleffectson the demand

for servicesis unclearas realincomeincreases.19On the supplyside, the supply
of services is reduced as the z sector expands. The effect of a relaxationof
price control on the non-tradedprice level, and hence on the real exchange
rate, is thereforeambiguous.20
(25) PS=-zzPz + ozPz+ yy9

Case 2: waiting in line

In practice,rentsarisingfrom price controlare not transferredin a lump-sum
fashion; goods do not always go the highest-valueusers, and resourcesare
used up in black markets and through waiting in line. In this section we
examine how our results are changed when rationedgoods are allocated by
time expenditures.This assumptionallows us to introducethe rent-seeking
behaviourof Krueger(1974) and Bhagwati(1980) into our model.
The income-equal-to-expenditure conditionis given by (26) where L is the
fixed endowmentof labour, Lz is labourused up waitingin line and w is the
wage rate.Waitingin line, in effect,reducesthe effectivelabourforce available
to the economy. In the expression, V the vector of factor endowments,is
defined so as to exclude labour:

(26) e(A,Ps.,Pz; U) = r(p,Psps.h w; V) +(P"Z-,Pz)rz+wL-Z).

The labour marketequilibriumcondition is given by (27), which sets the
demand for labour equal to the supply of labour less the labour used up
waiting in line. Recall that the first derivativeof the revenue function with
respectto the wage rate gives the demandfor labour.
(27) rw( PZ.w; V)=L-Lz.
Because individualsare identical, rents arisingfrom price controls are dissi-
pated. The rent dissipationcondition is given by (28):

(28) wLz= ( -i -jz) rz.

Waitingin line changes our result in two ways. First,the easing of price
controls increasesthe effectivelabour force availableto the economy. Thus,
the real income gains of price decontrolare larger.21Second, the increasein
the effectivelabour force changes the supply of non-tradedgoods. The sign
of this effectis uncertainand dependson the Rybcznskiterm(gsL) for services.
Solving for the non-tradedprice level, the following resultis obtainedwhere
XL, the elasticity of the non-tradedprice level with respect to the effective

labour, is equal to -gsL(L/gs). This term can be positive, negativeor zero.22

(29) Ps = wzPz+ OzPz+ ITLL+ y9.

A relaxationin price control directlyincreasesthe measuredprice level,

as shown by the firsttermin (30), and appreciatesthe real exchangerate.The
overall effects are, however,uncertain.And, by contrastto public enterprise
price increases,thereis no presumptionas to the effectsof price decontrolon
the real exchangerate.
? The London School of Economics and Political Science 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

Proposition 4. The direct effect of a reduction in price control is to appreciate

the real exchange rate but indirect effects may offset the direct effect.
(30) R =-(YaZ +aS0Z)PZ-as(PZJ+ 7LL+ yy)


To sum up, except for an easing of price controls, there is a clear presumption
that public sector price reforms will appreciate the real exchange rate. In this
section we analyze public sector pricing policies associated with the Alan
Garcia and Alberto Fujimori governments in Peru. Our purpose is to illustrate,
for readers unfamiliar with developing economies, the enormous changes in
relative prices that result from such policies. In particular, we focus on the
effects of the large public sector price increases of August 1990 on the Peruvian
real exchange rate.
The Alan Garcia administration took office in August 1985. Its economic
programme called for a partial suspension of payments on international debt,
expansionary demand policies and greater restrictions on international trade.
After an initial growth spurt in 1986-87, Peru ran into severe current account
problems, experienced an acceleration in inflation, and suffered a rapid fall
in output and real wages. We confine our attention to developments in the
areas of interest to the subject matter of our paper: subsidies and public sector
Before the 1990 reforms, Peru had 135 state enterprises. State enterprises
were the sole producers of electricity, gas, oil and telecommunications and
had roughly a one-third share in mining and transport. In 1982 state enterprises
accounted for 32 per cent of GNP (Balassa et al. 1986).
Table 1 presents data on the real prices of public sector products in Peru
for the 1985-92 period. The public sector price index was constructed using


Alan Garci'a AlbertoFujimori

July July July July July July Aug. Aug. Aug.

85 86 87 88 89 90 90 91 92

Petrol (84 octane) 100 82 66 40 29 8 49 36 32

Paraffin 100 53 33 29 29 20 96 56 52
Electricity (residential) 100 60 41 25 9 7 60 33 34
Bus fares 100 86 79 75 58 55 75 91 101
Local telephone calls 100 84 75 51 28 8 45 32 44
Whole milk powder 100 85 66 77 37 53 71 43 43
Drinkable water 100 96 114 56 19 30 53 48 45
Wheat (common bread) 100 79 58 33 57 55 79 30 27
Rice (common) 100 74 63 57 54 51 47 35 38
Public sectorprice indexa 100 76 66 58 32 35 59 40 40

aThe public sector price index is computed in terms of all other goods, i.e. all non-public-sector
Source: Instituto Nacional de Estadistica and Coyuntura Economica, Universidad del Pacifico,
Lima, May 1992.
? The LondonSchool of Economicsand PoliticalScience 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

the prices of 22 importantpublicly provided goods. Altogether,these items

have a weight of 21 per cent in the consumerprice index. The other entries
in the firstpartof the tablewereformedby deflatingeach priceby the consumer
priceindex and then expressingthe resultas a percentageof the corresponding
level for July 1985.24Table 1 also gives the public sectorprice index in terms
of all other goods for the 1985-92 period.
As the table makes clear, from 1985 onwards,public sector prices lagged
consistentlybehindinflation.The Garciaadministrationplacedlargesubsidies
on rice, potatoes, milk and petrol products. By 1987, implicit and explicit
consumptionsubsidiesamountedto at least 7 per cent of GNP (WorldBank
1989,p. 52). Bythen,the relativepriceof publicsectorgoods was only one-third
of its originallevel. As a resultof depressedprices, public enterpriserevenue
as a percentageof GNP decreasedfrom 26 per cent of GNP in 1985to 10 per
cent in 1989.The budgetdeficitalmostdoubledoverthe sameperiod,reaching
11 per cent of GNP in 1989.25
On 8 August1990, 10 days aftersucceedingthe Garciaadministration,the
AlbertoFujimorigovernmentintroducedits economicprogramme.In the same
month, price controls were lifted, the exchange rate was unified and large
increasesin public sector prices were enacted. Altogether,the nominal price
of public sector goods increasedby 654 per cent in August;the relative price
of public sector goods in terms of all other goods, as shown by Table 1,
increasedby 69 per cent. In September,quantitativerestrictionson foreign
trade were removedand a new tariffregimewas initiatedwith tariffrates set
at 15, 25 and 50 per cent.26A temporarysurchargeof 10 per cent was placed
on the two highest rates.
To the surpriseof most observers,the stabilizationprogrammewas accom-
panied by a significantreal appreciation.Figure 1 tracesthe evolution of the
trade-weightedreal exchangeratefor Perufrom January1990to March1992.
We use a simple approachto determinethe effects of public sector price
increases on the real exchange rate. Recall, from equation (6), that real
exchangerate is the ratio of the externalto internalprice levels denominated
in domesticcurrency.Using (6), (7) and (8), we obtainthe followingexpression,
where E is the nominalexchangeratewhich is used to convertexternalprices
into domesticprices:
(31) R = aEg(E pg)+ ars(E-ps)+ aT(E pT) +P*
We aggregatetradedgoods into a compositecommodity,tradables,denoted
by PT. Observethat all the variablesin (31) are measurable.This expression
allows us to decomposechangesin the real exchangerateinto changesarising
from public sector,tradedand non-tradedgoods price movementsrelativeto
the nominal exchange rate. To save space, we concentrateon real exchange
ratemovementsfromJune 1990to October1990,therebycapturingthe effects
of the 8 August 1990 Fujimoriplan.
As a preliminarystep, Table 2 reportsthe percentagechangesin pricesfor
this period for the nominal exchange rate, tradables,non-tradables,public
sector prices and the overalldomesticprice level, P. Publicsector prices were
increasedin Augustby 654 per cent, that is by seven-and-one-halftimes their
July levels! The export and importexchange rates were unified and floated,
and the exchangerate was depreciatedby 378 per cent, or depreciatednearly
? The London School of Economics and Political Science 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

Stabilization plan
8 August 1990

60- _


Jan July jan July Jan

1990 1990 1991 1991 1992
Note: For January-August 1990, the real free rate and the real export rate were averaged. In
August 1990 the exchange rate was unified and floated. Thus, the export rate and the free rate
are the same after August 1990.
Source: For 1990, the real free rate and the real export rate were provided by the Departamento
de Investigaciones Banco Central de Reserva del Peru. For 1991, the real export rate are from
Nota Semanal, no. 49, p. 69, 31 December 1991 (Lima) and Nota Semanal no. 74, 18 May 1992,
p. 50.

FIGURE 1. The real exchange rate in Peru, 1990-92 (July 1985=100)


June 90 62s9 49 0 33s6 47 0 42'6

July 90 859 74.5 41K1 77.5 63J2
Aug. 90
378t1 654 0 272a2 412T7 397 0
Sep. 90 43 5 4*1 32*1 54 13*8
Oct. 906t0 9A6 9s6 11 8

Source: Nota Semanal no. 6, 7 February 1991, p. 57 and no. 18, 18 May 1992, p.
68 and author's calculations.

five times its previousaveragein one month.Inflationwas 397 per cent in the
month of August.Equallynoticeablefromthis table is the fall in the relative
price of non-tradedgoods in August.This is less of a puzzleonce it is realized
that the currencydepreciationand rise in consumerprices reduced the real
money supplyby 40 per cent havinga negativeliquidityeffecton the price of
services.27Finally,the trade liberalizationaccountsfor the relativelylow rate
of increase in the price to tradablesin Septemberand October,when traded
price level fell by 38 per cent in termsof the exchangerate.28
Table 3 gives the decompositionof the sourcesreal exchangerate changes
accordingto equation (31). The reformsof August were accompaniedby a
real appreciationdespite the huge devaluation.The final column of the table
shows that changes in the externalprice level played little role in explaining
m The LondonSchool of Economicsand PoliticalScience1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

(% per month)

R ag(E-_pg) a,(E -_p) aT(E-PT) P

June 90 20-2 2-9 10-2 7-0 0-16
July 90 21-7 2-4 15-7 3-7 016
Aug. 90 -36-1 -57-9 37K1 -15-2 0.19
Sep. 90 29-1 8-3 4-0 16-8 1.91
Oct. 90 -6-9 -0-7 -3-2 -3-0 051
a Whereag= 021, aS= 0 35 and aT= 044.
Forthe purposesof the calculationin Table3, equation(31), whichis an approxi-
mation,is used.
Source: Author's computations from data from the Instituto Nacional de Estadistica,
Coyuntura Economica, Universidad del Pacifico, op. cit., and Nota Semanal, op.
cit. Banco Central de Reserva del Peru Lima.

real exchange rate developments over this period. In fact, the large real
exchange rate appreciation of 36 per cent in August 1990 is explained mainly
by public sector price increases. Finally, the real depreciation of September
1990 is largely accounted for by the trade liberalization.
Augmented Dickey-Fuller tests for the real exchange rate and the relative
price of public sector goods in terms of the exchange rate are displayed in
Table 4. The data are monthly from July 1985 to March 1992. As shown by
the results, both of these variables appear to be non-stationary I(1) variables.
It follows that we may test for the existence of a long-run relationship between
these variables by testing for co-integration.
Following Engel and Granger (1987), we tested for co-integration using
an augmented Dickey-Fuller test (with three lags) on residuals obtained from
regressing the real exchange rate on the relative price of non-traded goods.
The augmented Dickey-Fuller test produced a t-statistic which is significant

JANUARY 1985-APRIL 1992a

Variable ADF

In R 0.46
In (E/pg) 0 07
A ln R 6.02**
A ln (E/pg) 5.80**
Residuals 2.85*
a The variablesare: R = the real exchange
rate, (E/pg) = the ratio of the price of the
dollar to the price of the public sector
output as in equation (31). All variables
are measuredin logarithms.A refers to
first differences.ADF is the augmented
symbols ** and * indicate rejections at the
1 and 5 per cent levels, respectively.
( The LondonSchool of Economicsand PoliticalScience1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

at the 6 per cent level, thus providing further evidence suggesting a long-run
relationship between public sector prices and the real exchange rate.


We have examined the effects of public sector pricing policies on the real
exchange rate. Our principal results are as follows.
1. The removal of a consumption subsidy on a traded good used solely
for consumption purposes will appreciate the real exchange rate, owing to
increased demand for non-traded goods.
2. When a traded good is consumed directly and is also used as an input,
the removal of a subsidy will appreciate the real exchange rate if non-traded
goods and the previously subsidized good are general-equilibrium substitutes
in supply.
3. Increases in the relative price of public sector products tend to increase
the demand for other nontraded goods and thus to appreciate the real exchange
rate. However, the increase in the supply of other non-traded goods arising
from the contraction of the public sector may outweigh the demand effects.
In practice, we expect that the direct effects will outweigh indirect effects.
4. A relaxation of price control increases the measured price of controlled
goods, thus having a direct influence toward real exchange rate appreciation.
However, counter-effects on the supply side, notably through the expansion
of the previously price-controlled sector, may offset the direct effects. Con-
sequently, the net impact on the real exchange rate of a relaxation of price
controls is ambiguous.
5. As an empirical matter, we believe that comprehensive public sector
price reforms tend to appreciate the real exchange rate permanently unless
previous subsidy levels are restored. The public sector reforms of Peru in
August 1990 were followed by a large real appreciation which is still in place.29
The same forces are at work in many other reform packages. Thus, for public
sector programmes which involve the reduction of subsidies and public enter-
prise price increases, we expect real appreciations to be the norm rather than
the exception.30

We are indebted to an anonymousreferee for particularlyhelpful suggestionson a
previousdraftof this paper.

1. Dornbusch et aL (1990) and Buffie (1992) examine some macroeconomic issues arising from
public sector pricing without mentioning the real exchange rate.
2. For many developing countries, state enterprises produce between 10 and 20 per cent of GNP.
They are particularly important in Africa and Latin America. Short (1984), for example, found
that the average share of public enterprises for African economies was 18 per cent while
Balassa et al. (1986) estimated that the average for large Latin American economies was 19
per cent. See also Ayub and Hegstad (1987) and World Bank (1988).
3. We ignore pricing in the education and health areas.
4. Here we follow the approach of earlier papers on the real exchange rate such as Neary (1988)
and Edwards (1989b).
5. For details of the GNP and expenditure functions, see Woodland (1982).
6. In this purely real model, the exchange rate plays the role of a unit of account or 'conversion

? The LondonSchool of Economicsand PoliticalScience 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

7. Using the approach of Woodland (1982) and Neary (1988), the model can be extended to
allow for a vector of non-traded goods. In addition, Neary (1985) allows for unemployment.
These extensions complicate the model without changing our results.
8. The results are unchanged if a producer price index is substituted for the consumer price index.
9. We adopt this assumption for simplicity. The more general case, where the supply of oil is
variable, requires that separate revenue functions be specified for the oil and the non-oil
economy. Our results continue to hold in the more general model.
10. From Shepard's lemma, the first derivative of the revenue function with respect to the price
of oil (r' ) gives the demand for oil as an intermediate input by the non-oil domestic economy.
11. The Esgterm is negative in the specific factor model, but can be positive or negative in the
Heckscher-Ohlin model. In general, there is a presumption that Esg is negative.
12. In practice, public-sector reform means that public enterprises are often sold to the private
sector. If the public-sector enterprise has monopoly power, this changes our results in two
ways. First, the price increases arising from 'privatization' are likely to be greater, and second,
the real income effects are uncertain.
13. Selective price controls have long been a feature of Latin American economies (see Balassa
et al. 1986 for evidence).
14. For partial equilibrium approaches to price controls, see Barzel (1974), Cheung (1974) and
Sah (1987).
15. Our results are readily generalized to cover shortages that arise when public sector prices lag
behind inflation, or to cover the imposition of price controls on traded goods.
16. For details of the restricted expenditure function, see Deaton and Muellebaur (1980) and
Cornes (1992).
17. For a different general equilibrium treatment of price control, see Helpman (1988).
18. These results can be obtained by totally differentiating (3), (24) and (25) and using the
assumption of substitutability in demand and supply.
19. Gould and Henry (1967) showed, in a partial equilibrium model based on Tobin and
Houthakker (1950-51), that the effects of price control on the demand for substitutes is
uncertain; see Neary and Roberts (1981) and Fallis and Smith (1984).
20. If trade ceases, price controls on importables and exportables have identical real exchange
rate effects. If trade continues for an export good, an easing of price controls raises the
non-traded price level.
21. Strictly speaking, this result holds only when all other distortions are set to zero. With more
than one distortion, rent-seeking can increase welfare from second-best considerations.
22. In short-run models this parameter is negative.
23. Much of the following discussion relies on the World Bank (1989). Paus (1991) presents a
more sympathetic account of Peruvian economic policy over this period.
24. The model of the previous section used a traded good as the numeraire. However, in this
context it is more informative to deflate prices by the consumer price index.
25. The Garcia administration intensified trade restrictions. By 1987, all imports required a licence.
The unweighted average tariff for manufactures was estimated at 69 per cent in December
1987 by the World Bank Mission (World Bank 1989). The average effective rate of protection
arising from trade restrictions and multiple exchange rates was estimated at 141 per cent.
26. The 50 per cent rate was abolished in April 1991.
27. For a model of this process, see Connolly and Taylor (1976).
28. The theory of trade liberalization and its effect on the relative price of non traded goods is
dealt with in Connolly and Devereux (1992).
29. Many argue that the real appreciation of the new sole is a result of Dutch Disease, caused
by coca paste exports. This argument, however, if true, applied before the 1990 reforms and
thus could not explain the subsequent real appreciation.
30. The results also hold in intertemporal models. In particular, a permanent reduction in
consumption subsidies or public sector prices will appreciate the real exchange rate. Inter-
temporal models also allow us to analyse the effects of public sector pricing on the current
account. It can be shown that a temporary consumption subsidy along with a public sector
price reduction worsens the current account.

AYUB, MAMOODand HEGSTAD,SVEN (1987). The management of public industrial enterprises.
World Bank Research Observer.Washington, DC: World Bank.
HENRIQUE (1986). Towards Renewed Economic Growth in Latin America. Washington, DC:
Institute for International Economics.
BARZEL, YORAM (1974). A theory of rationing by waiting. Journal of Law and Economics, 17,73-95.

? The LondonSchool of Economicsand PoliticalScience 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions

BHAGWATI, JAGDISH(1980). Lobbying and welfare. Journal of Public Economics, 14, 355-63.
BUFFIE, E1DWARD F. (1992). Stabilization policy and public sector prices. Journal of Monetary
Economics, 29, 253-76.
CHEUNG, STEVEN(1974). A theory of price control. Journal of Law and Economics, 17, 53-71.
CONNOLLY,MICHAELand DEVEREUX,JOHN (1992). Commercial policy, the terms of trade,
and real exchange rates. Oxford Economic Papers, 44, 507-12.
and TAYLOR,DEAN (1976). Adjustment to devaluation with money and non-traded goods.
Journal of International Economics, 6, 289-98.
CORNES, RICHARD (1992). Duality and Modern Economics. New York: Cambridge University
DEATON, ANGUS and MUELLBAUER,JOHN (1980). Economicsand ConsumerBehavior.
Cambridge University Press.
inflation: dynamics and stabilization. Brookings Papers on Economic Activity, 2, 1-84.
EDWARDS,SEBASTIAN(1989a). The International Monetary Fund and the developing countries:
a critical evaluation. Carnegie-Rochester Series on Public Policy, 31 (Amsterdam: North-
(1989b). Real Exchange Rates, Devaluation, and Adjustment: Exchange Rate Policy in
Developing Economies. Cambridge, Mass.: MIT Press.
ENGEL, R. and GRANGER, CLIVE (1987). Co-integration and error correction: representation,
estimation and testing. Econometrica 55, 251-76.
FALLIS,GEORGEand SMITH, LAWRENCE(1984). Uncontrolled prices in a controlled market:
the case of rent controls. American Economic Review, 74, 193-200.
GOULD, JOHN and HENRY, STEVEN (1967). The effects of price control in a related market.
Economica, 54, 411-29.
HELPMAN, ELHANON (1988). Macroeconomic effects of price controls: the role of market
structure. Economic Journal, 98, 341-55.
KRUEGER,ANNE (1974). The political economy of the rent seeking society. American Economic
Review, 64, 291-303.
NEARY, J. PETER(1985). International factor mobility, minimum wages and factor price equaliz-
ation: a synthesis. QuarterlyJournal of Economics, 100, 551-71.
(1988). Determinants of the equilibrium real exchange rate. American Economic Review, 78,
- and ROBERTS,KEVIN (1981). The theory of household behavior under rationing. European
Economic Review, 87, 25-42.
PAUS, EVA (1991). Adjustment and development in Latin America: the failure of peruvian
heterodoxy, 1985-1990. World Development, 19, 411-34.
ROTBARTH,E. (1940-41). The measurement of changes in real income under rationing. Review
of Economic Studies, 8, 100-7.
SAH, RAJ (1987). Queues, rations, and the market: comparisons of outcomes for the poor and
rich. American Economic Review, 77, 69-76.
SHORT,R. PETER(1984). The role of public enterprises: an international statistical comparison.
In R. Peter Short, Robert H. Floyd and Clive S. Gray, Public Enterprisesin Mixed Economies:
Some Macroeconomic Aspects. Washington, DC: International Monetary Fund.
TOBIN, JAMES and HOUTHAKKER,HENRIK (1950-51). The effects of rationing on demand
elasticities. Review of Economic Studies, 18, 140-53.
WORLD BANK (1988). World Development Report. New York: Oxford University Press.
(1989). Peru: Policies to Stop Hyperinflation and Initiate Economic Recovery. Washington,
DC: World Bank.
WOODLAND,ALLAN (1982). International Trade and Resource Allocation. Amsterdam: North-

?) The London School of Economics and Political Science 1993

This content downloaded from on Wed, 8 Oct 2014 04:17:57 AM

All use subject to JSTOR Terms and Conditions