Michael L. Morris
DETERMINING COMPARATIVE ADVANTAGE
THROUGH DRC ANALYSIS
guidelines emerging from CIMMYT's experience
CIMMYT is an internationally funded, nonprofit scientific research and training organization. Headquartered in Mexico, the Center is engaged in a worldwide research program for maize, wheat, and triticale, with emphasis on
improving the productivity of agricultural resources in developing countries. It
is one of 13 nonprofit international agricultural research and training centers
supported by the Consultative Group on International Agricultural Research
(CGIAR), which is sponsored by the Food and Agricultural Organization (FAO)
of the United Nations, the International Bank for Reconstruction and Development (World Bank), and the United Nations Development Programme (UNDP).
The CGIAR consists of a combination of 40 donor countries, international and
regional organizations, and private foundations.
CIMMYT receives core support through the CGIAR from a number of sources,
including the international aid agencies of Australia, Austria, Brazil, Canada,
China, Denmark, the Federal Republic of Germany, Finland, France, India,
Ireland, Italy, Japan, Mexico, the Netherlands, Norway, the Philippines, Spain,
Switzerland, the United Kingdom, and the USA, and from the European
Economic Commission, Ford Foundation, Inter-American Development Bank,
OPEC Fund for International Development, UNDP, and World Bank. CIMMYT
also receives non-CGIAR extra-core support from Belgium,the International
Development Research Centre, the Rockefeller Foundation, and many of the
core donors listed above.
Correct Citation: Morris, M.L. 1990. Determining Comparative Advantage Through
DRC Analysis: Guidelines Emerging from CIMMYT's Experience. CIMMYT Economics
Paper NO.1. Mexico, D.F.: CIMMYT.
ISSN - 0188 - 2414
Contents
Abstract
Acknowledgements
Preface
V
VI
vii
1
INTRODUCTION
1
2
A SIMPLE EXAMPLE OF COMPARATIVE ADVANTAGE
3
3
REVIEW OF THE DOMESTIC RESOURCE COST METHODOLOGY 6
6
Step 1: Developing Enterprise Budgets
Step 2: Classifying Inputs and Outputs
7
Step 3: Determining Market Prices and Social Prices
7
Step 4: Calculating Net Social Profitability
12
14
Step 5: Calculating Resource Cost Ratios
Step 6: Conducting Sensitivity Analysis
16
4
USES OF DOMESTIC RESOURCE COST ANALYSIS
Revealing Distorting Effects of Government Policies
Revealing Comparative Advantage Between Enterprises
Revealing Comparative Advantage Between Regions
Revealing Comparative Advantage Between Technologies
Setting Agricultural Research Priorities
5
6
19
19
22
24
25
27
PROBLEMS COMMONLY
ENCOUNTERED IN USING DRC METHODS
Extensive Data Requirements
Determining Social Prices for Primary Factors
Determining Social Prices for Non-traded Tradables
Estimating an Equilibrium Exchange Rate
Predicting General-Equilibrium Effects
Comparative Advantage vs. Competitive Advantage
28
28
29
31
33
LESSONS FROM THE CIMMYT CASE STUDIES
36
References
34
35
41
Tables and Figures
Table 1.
Production conditions in two potential trading partners
3
Table 2.
Net welfare gains achieved through specialization and trade
4
Table 3.
Establishing social prices for fertilizer at the farm level,
Nakuru area, Kenya, 1987
9
Table 4.
Establishing social prices for non-standard tradable inputs
at the farm level, Nakuru area, Kenya, 1987
11
Enterprise budgets for irrigated crops in Zimbabwe,
1986 (Z$lha)
13
Table 6.
Resource cost ratios for irrigated crops in Zimbabwe, 1986
15
Table 7.
Interpretation of resource cost ratios (RCRs)
16
Table 8.
Social profitability of rainfed wheat in Zimbabwe under four
yield levels, 1986
17
Table 9.
Measures of comparative advantage and policy incentives
20
Table 10.
Sources of difference between private and social profitability
of irrigated crops in Zimbabwe, 1987
21
Resource cost ratios for competing enterprises in
Ecuador, 1983
23
Resource cost ratios for two wheat-producing regions in
Mexico, 1985
25
Social profitability of wheat production technologies by size
of field, Kenya, 1987
26
Figure 1.
Conditions giving rise to a non-traded tradable
32
Figure 2.
Establishing social prices for non-traded tradables
32
Figure 3.
Using DRC analysis for research resource allocation
37
Table 5.
Table 11.
Table 12.
Table 13.
Abstract
Knowledge of comparative advantage is important for developing countries, because potential welfare gains from specialization and trade can be used to
foster economic growth. One practical difficulty with using comparative advantage for designing agricultural policies or allocating research resources is that
comparative advantage is difficult to determine empirically. The domestic
resource cost (DRC) framework of analysis offers a way of empirically measuring comparative advantage by generating quantitative indicators of the efficiency of using domestic resources to produce a given commodity, as measured
against the possibilities of trade. These quantitative indicators provide an
empirical measure of comparative advantage. At the same time, the analytical
framework allows the distortionary effects of government policies to be measured. The paper first describes the basic DRC methodology, using examples
drawn from a series of case studies carried out over the past five years by the
CIMMYT Economics Program. This step-by-step description of the DRC methodology is a concise, non-technical guide for use in conducting applied comparative advantage studies and addresses a number of practical and conceptual
problems commonly encountered. The paper then summarizes a number of
lessons learned over the years at CIMMYT and evaluates the usefulness of the
DRC framework as a tool for applied policy analysis and research resource
allocation.
v
Acknowledgements
Acknowledgements
This
Thispaper
paperdescribes
describestheoretical
theoreticaland
andpractical
practicallessons
lessonsemerging
emergingfrom
fromaaseries
seriesofof
comparative
advantagestudies
studiescarried
carriedout
outduring
duringthe
themid-1980s
mid-1980sby
byeconomists
economists
comparativeadvantage
from
theInternational
InternationalMaize
Maizeand
andWheat
WheatImprovement
ImprovementCenter
Center(CIMMYT),
(CIMMYT),inin
fromthe
collaboration
collaborationwith
with colleagues
colleaguesininnational
nationalagricultural
agriculturalresearch
researchprograms.
programs.
Without
Withoutthese
thesestudies,
studies,ititobviously
obviouslywould
wouldhave
havebeen
beenimpossible
impossibletotowrite
writeaa
summary
summarypaper
paperdetailing
detailingshared
sharedexperiences.
experiences.Therefore,
Therefore,the
theother
otherauthors
authorsofof
the
theoriginal
originalcomparative
comparativeadvantage
advantagestudies
studiesalso
alsodeserve
deservecredit
creditfor
forthe
thepresent
present
document:
document:Derek
DerekByerlee,
Byerlee,Jim
JimLongmire,
Longmire,and
andLarry
LarryHarrington
HarringtonofofCIMMYT;
CIMMYT;
Sutat
SutatSat-thaporn
Sat-thapornofofthe
theMinistry
MinistryofofAgriculture,
Agriculture,Thailand;
Thailand;and
andJuma
JumaLugogo
Lugogo
ofofEgerton
EgertonUniversity,
University,Kenya.
Kenya.
InInaddition,
addition,several
severalindividuals
individualscontributed
contributedby
bypreparing
preparingmaterials
materialsfor
foraa
workshop
titled"Research
"ResearchResource
ResourceAllocation
Allocationand
andComparative
ComparativeAdvantage,"
Advantage,"
workshoptitled
sponsored
sponsoredjointly
jointlyby
byCIMMYT
CIMMYTand
andby
bythe
theRegional
RegionalCoordination
CoordinationCentre
Centrefor
for
Research
Researchand
andDevelopment
DevelopmentofofCoarse
CoarseGrains,
Grains,Pulses,
Pulses,Roots,
Roots,and
andTubers
Tubers
(CGPRT),
heldininBogor,
Bogor,Indonesia,
Indonesia,during
duringOctober-November
October-November1987.
1987.These
These
(CGPRT),held
included
JimLongmire,
Longmire,Larry
LarryHarrington,
Harrington,Juan
JuanCarlos
CarlosMartinez,
Martinez,and
andBeatriz
Beatriz
includedJim
Avalos
PearsonofofStanford
Stanford
AvalosofCIMMYT;
ofCIMMYT;Govert
GovertGijsbers
GijsbersofCGPRT;
ofCGPRT;Scott
ScottPearson
University;
University;Mathias
Mathiasvon
vonOppen
Oppenofofthe
theUniversity
UniversityofofHohenheim,
Hohenheim,Germany;
Germany;and
and
Klaus
KlausAltemeier,
Altemeier,consultant.
consultant.
Derek
DerekByerlee,
Byerlee,Robert
RobertTripp,
Tripp,Larry
LarryHarrington,
Harrington,John
JohnBrennan,
Brennan,Mitch
MitchRenkow,
Renkow,
and
andJim
JimLongmire
Longmireofofthe
theCIMMYT
CIMMYTEconomics
EconomicsProgram
Programmade
madedetailed
detailedcomcomments
mentson
onearly
earlydrafts
draftsofofthe
themanuscript.
manuscript.Special
Specialthanks
thanksalso
alsogo
gototoLynn
LynnSalinSalinger
Associatesfor
forInternational
InternationalResources
Resourcesand
andDevelopment
Development(AIRD),
(AIRD),who
who
gerofofAssociates
served
externalreviewer
reviewerand
andprovided
providednumerous
numeroushelpful
helpfulsuggestions.
suggestions.Kelly
Kelly
servedasasexternal
Cassaday
Cassadayand
andMike
MikeListman
Listmancontributed
contributededitorial
editorialservices,
services,Miguel
MiguelMellado
MelladoE.E., ,
Manuel
ManuelFouilloux
Fouillouxand
andEfren
EfrenDiaz
DiazChias
Chiasassisted
assistedwith
withthe
thelayout
layoutand
andfigures,
figures,
and
andRocio
RocioVargas
VargasGranados
Granadosprepared
preparedthe
thetables.
tables.
VI
VI
Preface
This paper is the first in a new series, the CIMMYT Economics Program
Papers. This series is intended to report more substantive research results generated by the staff of the CIMMYT Economics Program and collaborators in
national agricultural research systems. Each CIMMYT Economics Program
Paper is reviewed by one or more independent external referees prior to publication.
This initial paper summarizes some of the lessons learned during the past few
years by CIMMYT economists while carrying out a series of applied comparative advantage studies based on the domestic resource cost (DRC) methodology.
Applied macro-level policy analysis represents something of a change from the
CIMMYT Economics Program's traditional micro-level emphasis on technology
evaluation through on-farm research methods. However, our efforts to examine
comparative advantage issues relating to wheat received an enthusiastic
response from collaborators in the national systems, many of whom have
requested materials to explain and clarify the methodology. This paper attempts to respond to this demand by generating resource materials that can be
used in planning and conducting applied comparative advantage work. We
hope the paper will prove useful to our colleagues in the national agricultural
research systems who wish to design studies to generate the kind of empirical
information which so often is missing from the policy debate.
Derek Byerlee
Director, Economics Program
VII
r
1
Introduction
Knowledge of comparative advantage is important for developing countries, because potential welfare gains from specialization and trade can be used to
foster economic growth. National income often can be increased through
policies encouraging farmers to produce commodities that exploit existing
patterns of comparative advantage. Over the longer run, additional welfare
gains can be assured if research resources are used to strengthen comparative
advantage in the future. This suggests that agricultural policies and research
resource allocation decisions should be based at least in part on comparative
advantage considerations.
Recent developments in the global economy indicate that agricultural policy
makers have indeed started to pay increased attention to comparative advantage. During the 1980s, many countries have implemented policy reforms
designed to reduce state participation in agriculture, increase productivity,
liberalize commodity trade, and free up market prices to playa greater role in
directing economic activity along efficiency lines. At the same time, increased
pressure has been put on research administrators to ensure the cost-effectiveness of agricultural research expenditures. Significantly, these developments
have occurred in industrialized as well as developing countries, both market
economies as well as centrally planned economies.
Unfortunately, one practical difficulty with using comparative advantage for
designing agricultural policies or allocating research resources is that comparative advantage is not easy to determine empirically. Simply comparing production costs between two regions or countries is often inconclusive, because
comparative advantage is not directly related to absolute production costs.
Even if relative production costs are known, frequently these are distorted by
government policies or market failures. Policy makers and research administrators thus require a way to "see through" market distortions in order to
determine true underlying patterns of comparative advantage.
This paper evaluates one method for empirically determining comparative advantage. The domestic resource cost (DRC) framework generates quantitative
indicators of the efficiency of using domestic resources to produce a given
commodity, as measured against the possibilities of trade. These quantitative
indicators provide an empirical measure of comparative advantage. At the
of the distortionsame time, the analytical framework also allows measurement ofthe
ary effects of government policies. The paper first describes the basic DRC
methodology, using examples drawn from a series of case studies carried out
over the past five years by the CIMMYT Economics Program. This step-by-step
description of the DRC methodology provides a concise, non-technical guide for
use in carrying out applied comparative advantage studies and addresses
several practical and conceptual problems commonly encountered. The paper
then summarizes a number oflessons learned over the years at CIMMYT and
evaluates the usefulness of the DRC framework as a tool for applied policy
analysis and research resource allocation.
CIMMYT's
CIMMYT'sinterest
interestinincomparative
comparativeadvantage
advantagework
workoriginally
originallygrew
grewout
outofofthe
the
desire
desiretotodevelop
developanalytical
analyticaltools
toolsfor
foruse
usebybyresearch
researchadministrators
administratorsininnational
national
agricultural
agriculturalresearch
researchsystems
systems(Longmire
(Longmireand
andWinkelmann
Winkelmann1985).
1985).InInananera
eraofof
shrinking
shrinkingresearch
researchbudgets,
budgets,these
theseadministrators
administratorswere
werelooking
lookingfor
forways
waystoto
increase
increasethe
thecost-effectiveness
cost-effectivenessofofresearch
researchexpenditures.
expenditures.Empirical
Empiricalknowledge
knowledge
ofofa acountry's
country'spattern
patternofofcomparative
comparativeadvantage
advantagepresumably
presumablywould
wouldhelp
helpimimprove
provethe
theallocation
allocationofofscarce
scarceresearch
researchresources.
resources.CIMMYT
CIMMYTeconomists,
economists,somesometimes
timesworking
workinginincollaboration
collaborationwith
withlocal
localresearchers,
researchers,carried
carriedout
outa aseries
seriesofof
case
studiesbased
basedtotovarious
variousdegrees
degreesononDRC
DRCmethods.
methods.These
Thesestudies,
studies,most
mostofof
casestudies
which
whichfocused
focusedononthe
theefficiency
efficiencyofofwheat
wheatproduction
productionininnon-traditional
non-traditionalenvironenvironments,
ments,were
werecarried
carriedout
outininThailand
Thailand(Harrington
(Harringtonand
andSat-thaporn
Sat-thaporn1984),
1984),inin
Ecuador
Ecuador(Byerlee
(Byerlee1985),
1985),ininMexico
Mexico(Byerlee
(Byerleeand
andLongmire
Longmire1986),
1986),ininZimbabwe
Zimbabwe
(Morris
(Morris1988a),
1988a),and
andininKenya
Kenya(Longmire
(Longmireand
andLugogo
Lugogo1989).
1989).
22
2
A Simple Example
of Comparative Advantage
The theory of comparative advantage is generally attributed to Ricardo (1817),
who first extended the optimization principle defining efficient choice of output
by firms into the arena of international trade. Ricardo pointed out that a
country can achieve net welfare gains by concentrating productive capacity on
goods and services of which it is a relatively efficient producer and importing
the rest. Significantly, and somewhat counter-intuitively, the country need not
be the lowest cost producer of a given product (Le., the most efficient in absolute terms) in order to have a comparative advantage in its production.
The principle of comparative advantage is perhaps best illustrated through a
simple numerical example (Table 1). Assume the world consists of only two
countries, Alpha and Beta, which produce two commodities, wheat and textiles,
using a single homogeneous factor of production, labor. This means that the
value of each product is determined exclusively by its labor content. Now
assume that one unit oflabor in Alpha produces 6 t of wheat or 2 t of textiles,
while one unit of labor in Beta produces 2 t of wheat or 1 t of textiles. (The
differences in productivity between Alpha and Beta result from differences in
technology and resource endowments.) Clearly, Alpha is absolutely more
efficient in the production of both commodities, since Alpha can produce both
wheat and textiles at lower per-unit costs than Beta.
However, this does not imply that Alpha should produce both wheat and
textiles and Beta should produce nothing. With open economies, both Alpha
and Beta can achieve net welfare gains through specialization and trade. To
see this, it is necessary to consider the relative efficiency of production within
each country. In absolute terms, Alpha may be more efficient than Beta in the
production of both commodities, but in terms of the internal cost ratio (determined in this example by the labor content), Alpha enjoys a relatively greater
advantage in the production of wheat. This becomes evident when we consider
that Alpha must give up 3 t of wheat to free up enough resources (labor) to
Table 1. Production conditions in two potential trading partners
In
One unit of labor
produces
country
Alpha
Beta
6 t wheat
or
2 t textiles
2 t wheat
or
1 t textiles
Absolute production Wheat: Textiles
cost per unit
domestic value ratio
1/6 unit labor
3:1
1/2 unit labor
1/2 unit 'labor
2:1
1 unit labor
3
produce an additional 1 t of textiles, whereas Beta must give up only 2 t of
wheat to free up enough resources (labor) to produce an additional 1 t of
textiles. In other words, the resource costs (or opportunity costs) of the two
commodities differ between countries: 1 t of textiles is worth 3 t of wheat in
Alpha, but only 2 t of wheat in Beta. These domestic resource costs determine
the limits to mutually beneficial trade.
The difference in resource costs is what enables both countries to benefit from
specialization and trade. Consider what will happen if the ratio of the price of
wheat to the price of textiles is 2.5 to 1 in international markets. Alpha can
Table 2. Net welfare gains achieved through specialization and trade
Alpha
Resource endowment
Production cost per unit:
Wheat
Textiles
Beta
100 units labor
100 units labor
1/6 unit labor
1/2 unit labor
1/2 unit labor
1 unit labor
Without trade (assuming 50% allocaton ofresources to each product)
Production and
consumption
300 t wheat
100 ttextiles
100 t wheat
50 t textiles
With trade (assuming 100% allocation ofresources to product with
comparative advantage)
Specialized production
600 t wheat
100 t textiles
Quantity given up in trade
250 t wheat
50 t textiles
Quantity received in trade
100 ttextiles
125 t wheat
Consumption
350 t wheat
100 ttextiles
125 t wheat
50 t textiles
50 t wheat
25 t wheat
Net welfare gains
Note: International exchange ratio is 1 t textiles = 2.5 t wheat.
4
acquire 1 t of textiles by exporting only 2.5 t of wheat, which would be cheaper
than producing the textiles domestically at an opportunity cost of 3 t of wheat.
Conversely, by exporting 1 t oftextiles, Beta can obtain 2.5 t of wheat, more
than the 2 t of wheat it would have been able to produce with the same resources. Thus, in this simple example, both countries can benefit by exploiting
their respective comparative advantage: Alpha would be better off specializing
in wheat and trading for textiles, while Beta would be better off specializing in
textiles and trading for wheat. The net welfare gains from specialization and
trade are apparent in Table 2 in the form of higher levels of consumption in
both countries.
In a more realistic example, the quantity of labor required to produce either
wheat or textiles would not be constant over the entire feasible range of production. In that case, specialization in production would probably be less than
complete, and each country would continue to produce some quantity of both
goods.
5
3
Review of the
Domestic Resource Cost Methodology
Applied comparative advantage analysis essentially seeks to answer the following question: for a given country or region, which among a set of alternative
production activities is relatively most efficient (measured in terms of contribution to national income), ignoring the effects of distortions in the economy
resulting from government policies and market failures? Relative efficiency in
production--and hence comparative advantage--depends on three factors: 1)
technology (which determines production possibilities and influences rates of
product transformation); 2) the resource endowment (which determines the
value of domestic resources, e.g., land, labor, capital, and water); and 3) international prices (which determine the value of all other inputs and outputs).
The DRC method described in this paper generates several measures of the
relative economic efficiency of production alternatives. One of these measures,
net social profitability (NSP), indicates the contribution of each production
alternative to national income, measured in terms of social net returns to land.
A second measure, the resource cost ratio (RCR), indicates the efficiency of each
production alternative in using domestic resources to earn (or save) one unit of
foreign exchange. Since both measures capture the ability of production alternatives to contribute to national income, comparison of social profitability and!
or RCRs provides an empirical measure of the underlying pattern of comparative advantage.
Step 1: Developing Enterprise Budgets
DRC analysis begins with the development of an enterprise budget for each
production alternative being compared. Typically these production alternatives
will consist of different crops which compete for resources. However, they may
also consist of the same crop grown in different locations, for example when
policy makers are trying to determine which of two regions within the same
country enjoys a comparative advantage in the production of a given crop.
Alternatively, they may consist of the same crop grown using different levels of
technology, for example when policy makers are trying to determine whether a
country enjoys a comparative advantage in large-scale mechanized production
ofa crop or small-scale, labor-intensive production. Since calculation of RCRs
depends on accurately accounting for all inputs and outputs, the budgets must
be fairly detailed.
One important use of the enterprise budgets is to permit opportunity costing of
primary factors of production (e.g., land, labor, and capital). Consequently,
budgets must be constructed for all reasonable production alternatives in a
given region. Many applied comparative advantage studies appearing in the
literature focus too narrowly on one or two enterprises of immediate interest.
By ignoring viable production alternatives, these studies may incorrectly
estimate the opportunity costs of primary factors, especially land.
6
Step 2: Classifying Inputs and Outputs
After enterprise budgets have been constructed and verified, all inputs and
outputs must be classified as primary factors or tradables. This distinction is
necessary because RCRs are calculated as the ratio between the total opportunity cost of primary factors and the value added to tradables. Also, shadow
prices are often determined differently for primary factors and tradables. For
these reasons, the two categories of goods must be differentiated.
Primary factors are defined as goods that are not normally traded internationally and include chiefly land, labor, water, and capital. 1 Tradables are defined
as goods that are traded internationally or potentially could be traded. The fact
that some potentially tradable goods may not actually be imported or exported
(because of economic or political reasons) does not affect their status as
tradables; such goods are known as "non-traded tradables" and are classified as
tradables.
Some non-traded production inputs will turn out to be composite goods comprising both a tradable component and a primary factor component. For example, transport costs incurred in delivering fertilizer to the farmer's field
include both the cost of fuel and machinery (tradables), as well as labor costs (a
primary factor). Similarly, the cost of the machinery repairs and maintenance
may include both the cost of spare parts (tradables), as well as labor costs (a
primary factor). If such non-traded composite goods are used in production, it is
necessary to decompose them into the tradable component and the primary
factor component.
Step 3: Determining Market Prices and Social Prices
Once production alternatives have been identified, enterprise budgets developed, and inputs and outputs classified into tradables and primary factors, a
vector of economic (or shadow) prices must be constructed. Referred to here as
"social prices", these prices are intended to reflect the true economic value of
goods and services in the absence of taxes, subsidies, import tariffs, quotas,
price controls, and other government policies. Accurate estimation of social
prices is critically important in DRC analysis, because these prices represent
the opportunity cost to the economy of inputs and outputs. Social prices are
thus extremely influential in determining comparative advantage rankings,
which are based on the ability of alternative enterprises to contribute to national income.
1
Although labor might be considered tradable in cases where seasonal migration
results in remittances offoreign exchange, it is usually treated as a primary factor
in DRC analysis because the international labor market generally is not well
developed.
7
Social prices
prices are
are determined
determined differently
differentlyfor
for primary
primaryfactors
factors and
andtradables.
tradables.
Social
Primaryfactors
factors are
are assigned
assignedsocial
social prices
prices equal
equal to
to their
theiropportunity
opportunitycost
costvalue
value
Primary
(i.e., the
the returns
returns in
in their
their most
mostsocially
sociallyprofitable
profitable alternative
alternative use).
use). While
While the
the
(i.e.,
opportunitycost
costvalue
valueof
ofland
landfrequently
frequently can
can be
be estimated
estimatedfrom
from the
the enterprise
enterprise
opportunity
budgets themselves
themselves (as
(asthe
the residual
residual returns
returns to
toland
landin
inthe
the most
most socially
sociallyprofitprofitbudgets
able cropping
croppingalternative),
alternative), opportunity
opportunitycost
costvalues
values for
forthe
the other
otherprimary
primary
able
factors, especially
especiallylabor
laborand
and capital,
capital, are
are often
oftendifficult
difficultto
to establish.
establish. Estimating
Estimating
factors,
social prices
pricesfor
for labor
laborand
and capital
capitalinvolves
involves adjusting
adjustingmarket
marketprices
prices by
bysome
soIlie
social
compensatoryconversion
conversionfactor
factor selected
selectedto
toreflect
reflect the
the estimated
estimateddegree
degree of
of
compensatory
distortion prevailing
prevailingin
inthe
the economy.
economy.
distortion
Tradables are
are valued
valued at
at their
theirworld
world price
priceequivalent,
equivalent, i.e.,
i.e., the
the price
price at
at which
which
Tradables
theycan
can be
beimported
importedor
orexported,
exported, adjusted
adjusted for
for transport
transport costs
costs and
and exchange
exchange
they
rate anomalies.
anomalies. In
In the
the case
case of
ofimported
importedgoods,
goods, domestic
domestictransportation
transportationand
and
rate
handlingcosts
costs are
are added
added to
to the
the CIF
CIF price
price to
to arrive
arrive at
ataa social
social price
priceequivalent
equivalent
handling
tothe
the import
importparity
parity price;
price; in
in the
the case
case of
ofexported
exportedgoods,
goods, domestic
domestic transportatransportato
tion and
andhandling
handlingcosts
costs are
are subtracted
subtractedfrom
from the
the FOB
FOB price
price to
to arrive
arrive at
at aa social
social
tion
priceequivalent
equivalentto
tothe
theexport
exportparity
parity price.
price. Since
Since domestic
domestictransportation
transportation and
and
price
handlingcosts
costs may
maybe
be considerable,
considerable,ititisisimportant
importantto
to select
selectproduction
productionand
and
handling
consumption reference
reference points
points that
that accurately
accuratelyreflect
reflectthe
the geographical
geographical distribudistribuconsumption
tionof
ofproduction
production and
and consumption
consumption activities.
activities.22
tion
Because decisions
decisions based
basedon
on the
the results
results of
ofDRC
DRC analysis
analysistypically
typically have
have aa longer
longer
Because
term perspective,
perspective, ititmay
may be
be appropriate
appropriate to
to use
use projected
projectedintermediateintermediate- or
orlonglongterm
run CIF
CIF and
and FOB
FOB prices
pricesin
incalculating
calculatingimport
importand
and export
export parity
parityprices
pricesfor
for both
both
run
inputs and
andoutputs.
outputs. Use
Use of
ofprojected
projectedintermediateintermediate- or
orlong-run
long-run prices
prices rather
rather
inputs
than current
current prices
priceshelps
helps overcome
overcomethe
the problem
problem of
ofinternational
internationalmarket
market price
price
than
variability, which
which isis considerable
considerablefor
for many
many commodities.
commodities. Of
Ofcourse,
course, reliable
reliable
variability,
intermediate- or
orlong-run
long-run price
price projections
projectionsare
are not
not always
always available,
available, especially
especially
intermediateinputs. One
Oneconservative
conservative approach
approach for
for dealing
dealingwith
with international
international price
price
for inputs.
for
Thisisisparticularly
particularlyimportant
importantin
inthe
thecase
caseoflow-value
oflow-valuefood
foodcrops,
crops,for
forwhich
whichthe
the
22 This
borderprices
pricesisis usually
usuallyvery
veryhigh.
high. This
Thismeans
means
ratioof
ofdomestic
domestictransport
transportcosts
coststotoborder
ratio
thatititmay
maybe
beefficient
efficientfor
for aacountry
countrytotoproduce
produce aacommodity
commodityin
inremote
remoteregions
regionsfor
for
that
consumption
within
these
regions
and
to
import
the
same
commodity
for
consumption within these regions and to import the same commodity for
consumptionin
inmore
more accessible
accessibleborder
borderzones.
zones.
consumption
88
variability is to start with longer term trend prices and then to conduct sensitivity analysis to determine whether or not comparative advantage rankings
are robust with respect to prices of individual outputs or inputs.
Table 3 presents an example of the calculation of social prices for
imported fertilizers.
Table 3. Establishing social prices for fertilizer at the farm level,
Nakuru area, Kenya, 1987
Diammonium
phosphate
Long term world price, FOB US Gulf Ports (US$/t)
205
plus Freight and handling, US Gulf Ports to East Africa (US$/t) +----±Q
equals CIF Mombasa (US$/t)
245
times Equilibrium exchange rate (Ksh:US$)
equals CIF Mombasa (Ksh/t)
Calcium
ammonium
nitrate
+
184.5
---..A.Q
224.5
18.15
4,075
* 18.15
4,447
plus Port charges and rail freight, Mombasa-Nakuru (Ksh/t) +
452
equals On-rail price, Nakuru (Ksh/t)
4,899
equals On-rail price, Nakuru (Ksh/50 kg bag)
245
+
452
4,527
226
plus Retailing margin, including packaging (25%) (Ksh/bag) +
equals Social retail price, Nakuru (Ksh/bag)
+
---.lili
~
306
plus Transport to farm (Ksh/bag)
equals On-farm social price (Ksh/bag)
Actual price paid on farm (Ksh/bag)
Ratio market price to social price
282
+
-----2
288
234
215
0.75
0.75
Source: Longmire and Lugogo, 1989.
Note:
Transport costs are net of indirect taxes (e.g., subsidy to railroad, sales tax on gasoline).
9
World reference prices (CIF and FOB) may not be available for certain specialized tradables, for example, farm machinery, irrigation equipment, or agricultural chemicals. Social prices for these types of non-standard tradables are
estimated by starting with the domestic market price and removing policyinduced distortions such as tariffs, taxes, and exchange rate anomalies to
arrive at the equivalent of an import or export parity price. As before, the
objective is to arrive at a social price that accurately reflects the economic value
of the good or service in the economy. Table 4 presents an example of the
calculation of social prices for non-standard tradables for which world reference
prices were not available.
In calculating social prices for tradables, it is often necessary to estimate a
shadow exchange rate for converting between domestic currency and international currencies. Exchange rate distortions must be taken into account in DRC
analysis, because they affect the domestic prices of tradables. When the domestic currency is overvalued, imported goods appear cheaper in domestic currency
terms (because they can be purchased with fewer units of the overvalued
domestic currency), while exported goods appear more expensive for foreign
buyers (because more units of the foreign currency are required to pay for
them). When the domestic currency is undervalued, the effects are opposite.
Consequently, if an adjustment is not made to correct for exchange rate distortions, comparative advantage rankings will be biased in favor of import-intensive activities (in the case of overvalued domestic currency) or in favor of
export-intensive activities (in the case of undervalued domestic currency).3
A full set of market price data should be collected and retained when social
prices are estimated so that the effects of government policies and market
failures can later be quantified and disaggregated. Since the calculation of
social prices in any event often begins with market prices, the additional effort
involved is usually minimal.
3 Alternate methods for calculating the shadow exchange rate are discussed in
"Estimating an Equilibrium Exchange Rate," page 33.
10
~
~
343,000
32,000
66,000
95,000
35,660
1,223,000
33,000
45,500
25,200
Tractor, 75 HP (Ksh)
3-disc plow (Ksh)
10-ft harrow (Ksh)
12-ft drill (Ksh)
36-ft sprayer (Ksh)
16-ft harvester (Ksh)
7oft reaper (Ksh)
Large thresher (Ksh)
Small thresher (Ksh)
1
2
3
4
5
137,200
19,200
39,600
38,000
14,264
489,200
19,800
27,300
10,080
6.57
39.54
116.91
4.50
23.88
2.90
2,563
820
1,025
1,025
513
4,100
410
1,025
103
0.26
1.03
1.03
1.03
1.03
0.87
0
0
0
0
0
0
0
0
0
6.78
27.37
81.53
2.84
16.41
0.51
203,238
11,980
25,375
55,975
20,884
729,700
12,790
17,175
15,018
8.29
63.86
190.24
6.63
38.29
1.52
223,561
13,178
27,913
61,573
22,972
802,670
14,069
18,893
16,519
9.12
70.25
209.26
7.30
42.12
1.67
(f=e* 1.1)
(()
(e=a-b-c·d)
(e)
Adjusted
import
price'
(d)
Domestic Estimated Estimated
transport
Import
import
cost2
duty'
price
(c)
(h)
(i)
2,563
820
1,025
1,025
513
4,100
410
1,025
103
0.26
1.03
1.03
1.03
1.03
0.87
150.749
20,997
43,406
41,732
15,656
537,847
21,719
29,876
11,081
4.02
30.55
90.12
3.57
18.49
2.54
376,873
34,995
72,344
104,329
39,141
1,344,617
36,198
49,794
27,703
13.39
101.82
300.41
11.89
61.63
5.09
(i=f+g+h)
Revised
Social
Domestic domestic
price
transport
value
at
cost
added" farm level
(g)
Domestic value added =fixed percentage ofmarket price (varies by item).
Domestic transport cost = weight* domestic freight rate.
Estimated import duty = fIxed percentage of import price (varies by item).
Adjusted import price =estimated import price * 1.1 (reflects estimated 10% overvaluation ofthe Kenyan shilling).
Revised domestic value added =fIxed percentage of adjusted market price (varies by item).
Source: Longmire and Lugogo. 1989.
21.90
131.80
389.70
15.00
79.60
5.80
Domestic
value
added'
Gunnies (Ksh)
Buctril (KshIL)
Roundup (KshIL)
Dipterex 5% (Ksh/kg)
Malathion (Ksh/kg)
Diesel (KshIL)
(b)
(a)
Average
market
price
at farm
level
Table 4. Establishing social prices for non-standard tradable inputs at the farm level, Nakuru area,
Kenya, 1987
0.91
0.91
0.91
0.91
0.91
0.91
0.91
0.91
0.91
1.64
1.29
1.30
1.26
1.29
1.14
(j=a/i)
W
Ratio of
market
price to
social
price
Step 4: Calculating Net Social Profitability
Once social prices have been estimated, the net social profitability (NSP) of
each production alternative can be calculated. This is a straightforward procedure involving use of the enterprise budgets and social prices to calculate net
social returns (usually to land) of each competing enterprise. Table 5 presents a
sample set ofenterprise budgets showing the calculation ofnet social profitability.
Assuming social prices have been estimated accurately, social profitability
rankings provide a preliminary indication of comparative advantage. Social
prices reflect the true economic scarcity value of inputs and outputs, so the
enterprise with the largest positive NSP represents the most profitable production alternative in terms of contributing to national income. Furthermore,
because social prices for primary factors are set equal to their alternative use
values, the NSP figures automatically indicate relative efficiency in production
and thus provide an empirical measure of comparative advantage.
Some of the CIMMYT case studies have used social profitability rankings as
the sole indicator of comparative advantage. However, one shortcoming of the
NSP measure is that it provides no explicit information about the use of
tradables (and hence foreign exchange) in production. Therefore, it is often
helpful additionally to calculate RCRs for each enterprise.The RCRs are simply
a restatement of NSPs, expressed not in terms of domestic currency, but rather
as a unitless ratio (value of primary factors: value-added to tradables). The big
advantage of RCRs is that they provide an explicit indication of the efficiency
with which each production alternative uses domestic resources to generate or
save foreign exchange, which makes them easy to interpret. Also, because they
are not denominated in units of domestic currency, RCRs lend themselves more
readily to cross-country comparison (although care should be exercised in
comparing RCRs between countries, since they reflect productive efficiency
relative to very different patterns of resource endowments).
12
Table 5. Enterprise budgets for irrigated crops in Zimbabwe,
1986 (Z$lha)
Wheat
Maize
Soybeans
Ground·
nuts
Gross social returns
1,979.45
1,686.00
1,020.00
1,627.50
Fixed costs
Irrigation costs
Dam and pump
Irrigation equipment
74.81
74.81
24.94
24.94
24.94
24.94
54.86
54.86
64.84
64.84
39.48
39.48
Farm machinery costs
Tractor depreciation
Tillage equipment depreciation
Cost of capital
56.73
5.84
47.29
82.39
8.48
68.67
43.03
4.43
35.87
64.76
6.66
53.97
53.00
5.45
44.18
102.22
10.52
85.20
Cotton
2,923.38 12,428.19
Tobacco barns and sheds
Variable costs
Machinery operation costs
Tractor fuel
Tractor oils
Tractor repair and maintenance
Tillage equipment repair and
maintenance
Purchased inputs
Seed (and treatment)
Fertilizer and lime
Herbicides
Pesticides
Fungicides
Packing materials
Irrigation costs
Electricity
Repairs and maintenance
Tobacco
163.00
39.73
2.96
53.51
57.70
4.30
77.72
30.14
2.25
40.59
45.35
3.38
61.08
37.12
2.77
49.99
0.71
1.04
0.54
0.81
0.67
72.00
308.38
13.18
4.93
35.75
222.00
45.76
12.96
71.75
115.07
70.80
9.82
15.72
169.89
56.46
163.81
5.00
266.17
525.80
10.89
8.93
2.50
111.25
126.08
108.66
21.77
180.59
8.59
7.75
38.00
244.80
16.03
81.60
5.34
179.52
5.34
212.16
11.76
129.20
13.89
8.46
135.90
55.00
81.90
51.29
390.10
Contract hire services
Aerial pesticide application
Aerial fertilizer application
Combine harvesting
Transport, farm to depot
74.64
60.78
Other costs
Fertilizer transport/handling
Crop insurance
Drying
Levy
27.24
6.93
2.98
9.90
Labor costs
Skilled labor
Unskilled labor
Interest on working capital (6 mo.)
112.75
161.11
13.65
100.91
22.00
6.51
16.76
11.69
11.80
12.57
4.08
2.95
15.30
24.41
40.93
364.00
239.50
167.00
6.59
39.18
9.57
89.16
5.00
26.73
7.52
137.87
6.16
219.14
11.88
618.50
89.58
72.69
55.09
102.52
109.09
278.32
Total fixed costs
Total variable costs
259.48
1,084.94
209.41
866.75
133.21
667.24
235.12
1,060.99
232.31
1,185.30
439.91
3,370.79
Total costs
1,344.42
1,076.17
800.45
1,296.11
1,417.61
3,810.69
635.03
609.83
219.55
331.39
1,505.77
8,617.50
Net social returns
(social profitability)
19.91
9.61
74.76
40.37
Source: Morris, 1988a.
13
Step 5: Calculating Resource Cost Ratios
Using data from the enterprise budgets, RCRs can be calculated according to
the following formula:
RCRc =
(L Wpp
F )
/ ( P CC
T - L P.T.
)
11
where:
RCRc
WP
F
pP
c
Tc
P
T.
I
I
=
=
=
=
=
=
=
resource cost ratio for crop c
social prices (opportunity costs) of primary factors
primary factors of production (usually per ha)
social prices (world price equivalents) of crop c
quantity produced of crop c (usually per ha)
social prices (world price equivalents) of tradable inputs
quantity used of tradable production inputs
(usually per ha)
If the enterprises being compared all compete directly for land and other
resources, only one enterprise will show a positive RCR between 0 and 1. This
is because the net social returns to land used in the most profitable enterprise
represents the opportunity cost (or social price) ofland used in all other enterprises, which by definition drives the RCRs of these other enterprises above 1.
Only if additional assumptions are introduced limiting direct competition
between enterprises is it possible to end up with multiple RCRs between 0 and
1 (e.g., if two crops are grown on the same land but in different seasons).
Calculation of RCRs becomes more complicated if one or more enterprises generates joint products (e.g., groundnut oil and cake; milk, meat, and leather). In
such cases, the different social prices for each final joint product must be
converted into social price equivalents per unit of the initial production output
(e.g., ton of groundnuts; ton oflive cattle). The social price equivalents ofthe
joint products can then be added to obtain the value of production per unit land
area (or net social profitability of the enterprise). In this procedure, domestic
processing costs should also be adjusted, if they are distorted by government
policies or market failures.
Table 6 presents an example of the calculation of RCRs. Following the alternative use value approach, in this example the social price of land for each enterprise has been set equal to net social profitability (NSP) of land in the most
profitable alternative use (as revealed by enterprise budgets calculated previously). Since the most profitable alternative use varies between enterprises, the
social price ofland is not always the same.
Interpretation of RCRs is summarized in Table 7. A positive RCR between 0
and 1 indicates that the value of the domestic resources used in production is
less than the value of the foreign exchange earned or saved; thus, a country
has a comparative advantage in products associated with such a RCR, since the
country earns or saves foreign exchange in their production. A positive RCR
14
Table 6. Resource cost ratios for irrigated crops in Zimbabwe, 1986
Wheat
(Z$/ha)
Maize
(Z$lha)
Soybeans
(Z$lha)
Groundnuts
(Z$lha)
Cotton
(Z$lha)
Tobacco
(Z$lha)
1,979.45
1,886.00
1,020.00
1,627.50
2,923.38
12,428.18
Inputs
Machinery depreciation
212.19
0.75 Repairs and maintenance 52.69
Fuels and oils
42.69
Purchased inputs
394.48
44.01
0.5* Transport
0.5 *Machinery hire charges 37.32
Packing materials
10.89
244.80
Electricity
140.75
63.08
62.00
316.47
60.41
6.83
8.93
81.60
97.34
34.86
32.38
267.43
26.47
37.38
2.50
81.60
181.14
55.24
48.73
548.85
51.95
0.00
8.59
179.52
188.13
48.42
39.88
405.88
34.03
67.95
7.75
212.16
191.71
127.18
112.75
796.97
195.05
27.50
38.00
129.20
Tradables
Outputs
Value of production
Miscellaneous
Drying
Insurance
Levies
2.98
6.93
9.90
9.61
11.80
2.95
3.50
13.12
7.73
28.99
11.69
40.93
239.50
364.00
167.00
Primary factors
Capital
89.58
72.69
55.09
102.52
109.09
278.32
Labor
0.25*Repairs and maintenance
0.5* Transport
0.5 * Machinery hire charges
45.77
17.56
44.01
37.32
98.73
21.03
60.41
6.83
31.73
11.62
26.45
37.38
145.39
18.41
51.95
0.00
225.30
16.14
34.03
67.95
630.33
42.39
195.05
27.50
0.00
1,549.94
1,549.94
1,549.94
678.50
5,137.21
234.24
847.38
1,809.63
852.93
1,712.21
347.04
1,868.21
460.48
1,131.01
1,826.18
6,310.80
10,716.83
0.28
2.12
4.93
4.06
0.62
0.59
Landa
Net cost-primary factors
Value added-tradables
Resource cost ratios
Source: Morris, 1988a.
a Residual returns to land in best competing alternative valued at world price equivalent. Alternatives
assumed as follows: wheat vs. fallow; tobacco vs. rainfed tobacco; maize vs. soybeans vs. cotton VB.
groundnuts. Residual returns (Z$lha) =wheat, 682.31; maize 678.50; soybeans, 255.42; groundnuts,
385,36; cotton, 1,549.94; tobacco, 8,702.70; and rainfed tobacco, 5,137.21.
15
above
above1 indicates
1 indicates
that
that
the
thevalue
value
ofof
the
thedomestic
domesticresources
resourcesused
used
ininproduction
production
exceeds
exceeds
the
thevalue
value
ofof
the
theforeign
foreignexchange
exchange
earned
earned
ororsaved,
saved,and
and
the
thecountry
country
does
doesnot
not
have
havea acomparative
comparative
advantage
advantage
ininproduction.
production.A Anegative
negativeRCR
RCR
indiindicates
catesthat
that
foreign
foreignexchange
exchange
is is
being
being
wasted,
wasted,
i.e.,
i.e.,more
more
foreign
foreignexchange
exchange
is isused
used
inin
production
production
ofof
a commodity
a commodity
than
than
the
thecommodity
commodity
is isworth.
worth.
Step
Step6:6:Conducting
ConductingSensitivity
SensitivityAnalysis
Analysis
One
Oneconvenient
convenient
feature
featureofof
the
theDRC
DRC
framework
framework
is isthat
that
it it
lends
lendsitself
itself
readily
readily
toto
sensitivity
sensitivity
analysis.
analysis.With
Withthe
thehelp
helpofof
any
any
standard
standard
computer
computer
spreadsheet
spreadsheet
package,
package,
it it
is iseasy
easy
totodetermine
determinewhether
whether
comparative
comparativeadvantage
advantage
rankings
rankingsare
are
Table
Table7. 7.Interpretation
Interpretationofof
resource
resourcecost
costratios
ratios(RCRs)
(RCRs)
Value
ValueofRCR
ofRCR
Interpretation
Interpretation
oo<RCR<
<RCR<1 1
Value
Valueofof
domestic
domesticresources
resources
used
used
ininproduction
production
is is
less
lessthan
than
value
valueofof
foreign
foreignexchange
exchange
earned
earned
oror
saved
saved
==
comparative
comparative
advantage.
advantage.
RCR>
RCR>1 1
Value
Value
ofof
domestic
domesticresources
resourcesused
usedininproduction
production
is is
greater
greater
than
than
value
valueofof
foreign
foreignexchange
exchangeearned/saved
earned/saved
==nonocomparative
comparativeadvantage.
advantage.
RCR<O
RCR<O
More
More
foreign
foreignexchange
exchangeused
used
ininthe
theproduction
production
ofof
aa
than
thanthe
thecommodity
commodity
is isworth
worth
commodity
commodity
==
nonocomparative
comparativeadvantage.
advantage.
sensitive
sensitive
totochanges
changes
inin
individual
individualparameters.
parameters.Sensitivity
Sensitivity
analysis
analysisis is
imporimportant,
tant,because
because
technical
technicalcoefficients
coefficients
used
used
ininconstructing
constructing
enterprise
enterprise
budgets
budgets
(e.g.,
(e.g.,
yields,
yields,use
useofof
inputs)
inputs)are
areoften
often
mean
mean
values
valuescalculated
calculated
from
froma arange
rangeofof
observed
observed
values,
values,and
and
because
becauseprices
prices
used
used
inin
calculating
calculating
social
socialprofitability
profitability
the
theshadow
shadow
foreign
foreignexchange
exchange
rate)
rate)are
areoften
oftenestimated
estimated
prices
pricesoror
(including
(including
projected
projected
prices.
prices.Sensitivity
Sensitivity
analysis
analysiscan
can
reveal
reveal
whether
whether
comparative
comparative
advanadvantage
tagerankings
rankingscalculated
calculated
using
using
mean
mean
values,
values,estimated
estimated
values,
values,ororprojected
projected
values
valuesforfortechnical
technical
coefficients
coefficients
and
andsocial
socialprices
pricesare
arelikely
likely
totochange
change
iftechniiftechnicalcal
coefficients
coefficientsand/or
and/or
social
socialprices
prices
eventually
eventually
differ
differfrom
fromexpectations.
expectations.
1616
Which parameters should be tested through sensitivity analysis? While the robustness of DRC results will tend to vary depending on enterprise substitution
possibilities, technology, and relative prices, social profitability levels and RCRs
frequently are sensitive to the following parameters.
World reference prices of outputs
Comparative advantage rankings tend to be highly sensitive to world reference
prices of outputs. Ceteris paribus, a change in the world reference price of an
output will have a greater effect on the social profitability (and the RCR) of an
enterprise than a change of similar magnitude in any other parameter. Since
social prices of outputs are based on long-term trends in world reference prices,
which may not provide an accurate projection of future prices, it is generally a
good idea to determine whether comparative advantage rankings are sensitive
to changes in world reference prices of outputs.
Yields
Comparative advantage rankings tend also to be highly sensitive to the level of
yields assumed for any fixed quantity of inputs. To the extent that improved
management can succeed in raising mean yields above current levels, enterprises that appear socially unprofitable at present could become socially profitable in the future. Table 8 presents data from Zimbabwe showing the expected
social profitability of rainfed wheat under four yield levels, assuming no change
in inputs use and no changes in prices.'
Table 8. Social profitability of rainfed wheat in Zimbabwe under four
;yield levels, 1986
Rainfed wheat at
yield of:
1.5 t/ha
2.0 t/ha
2.5 t/ha
3.0 t/ha
Estimated social net
returns to land:
(Z$/ha)
(47)
125
297
469
Source:~orris, 1988b.
17
Exchange
Exchangerate
rate
InIncases
caseswhere
wherea ashadow
shadowexchange
exchangerate
ratehas
hasbeen
beenestimated
estimatedtotocorrect
correctforfor
apparent
apparentdistortions
distortionsininthe
theofficial
officialexchange
exchangerate,
rate,ititisisgenerally
generallya agood
goodidea
ideatoto
conduct
theexchange
exchangerate.
rate.Because
Becauseshadow
shadowexchange
exchange
conductsensitivity
sensitivityanalysis
analysisononthe
rates
ratesare
arenotoriously
notoriouslydifficult
difficulttotoestimate
estimateempirically,
empirically,conducting
conductingsensitivity
sensitivity
analysis
guardagainst
againstestimation
estimationerrors.
errors.Whether
Whetherorornot
notcomcomanalysisisisone
oneway
waytotoguard
parative
parativeadvantage
advantagerankings
rankingsturn
turnout
outtotobebesensitive
sensitivetotothe
theexchange
exchangerate
ratewill
will
the
depend
theforeign-exchange
foreign-exchangeintensity
intensityofof
theenterprises
enterprisesbeing
being
dependlargely
largelyononthe
compared.
the
tradable
output
compared.IfIf
theuse
useofof
tradableinputs
inputsper
perunit
unitvalue
valueofof
outputisisrelatively
relatively
similar
similaracross
acrossallallenterprises,
enterprises,changes
changesininthe
theshadow
shadowexchange
exchangerate
ratewill
willaffect
affect
social
socialprofitability
profitabilitylevels
levelsmore
moreororless
lessequally,
equally,and
andcomparative
comparativeadvantage
advantage
rankings
the
rankingsare
arelikely
likelytotoremain
remainunaffected.
unaffected.On
Onthe
theother
otherhand,
hand,ifif
theuse
useofof
tradable
tradableinputs
inputsvaries
variessignificantly
significantlybetween
betweenenterprises,
enterprises,changes
changesininthe
theshadow
shadow
exchange
ratewill
willhave
havedifferent
differenteffects
effectsononsocial
socialprofitability
profitabilitylevels,
levels,and
and
exchangerate
comparative
comparativeadvantage
advantagerankings
rankingsare
arelikely
likelytotochange.
change.The
Theexchange
exchangerate
ratemay
may
also
alsobebeimportant
importantwhen
whenthe
theproduction
productionalternatives
alternativesinclude
includenon-traded
non-tradedcrops
crops
(whose
(whosesocial
socialprice
priceisisestablished
establishedininthe
thedomestic
domesticmarket)
market)vs.vs.imported
importedoror
exported
aredetermined
determinedininthe
theworld
worldmarket).
market).
exportedcrops
crops(whose
(whosesocial
socialprices
pricesare
Wage
Wagerate
rate
InIncases
caseswhere
wherea ashadow
shadowwage
wagerate
ratehas
hasbeen
beenestimated
estimatedtotocorrect
correctforforapparent
apparent
distortions
distortionsininthe
thelabor
labormarket,
market,ititmay
maybebeuseful
usefultotoconduct
conductsensitivity
sensitivityanalysis
analysis
the
ononthe
thewage
wagerate.
rate.AsAsininthe
thecase
caseofof
theshadow
shadowexchange
exchangerate,
rate,not
notthe
theleast
least
reason
reasonforforthis
thisisisthat
thatshadow
shadowwage
wagerates
ratesare
arenotoriously
notoriouslydifficult
difficulttotoestimate
estimate
empirically;
empirically;thus,
thus,conducting
conductingsensitivity
sensitivityanalysis
analysisononthe
theshadow
shadowwage
wagerate
rateisis
one
waytotoguard
guardagainst
againstestimation
estimationerrors.
errors.Whether
Whetherorornot
notcomparative
comparative
oneway
advantage
advantagerankings
rankingsturn
turnout
outtotobebesensitive
sensitivetotothe
theshadow
shadowwage
wagerate
ratewill
will
depend
the
the
dependlargely
largelyononthe
thelabor
laborintensity
intensityofof
theenterprises
enterprisesbeing
beingcompared.
compared.IfIf
the
use
labor
output
useofof
laborper
perunit
unitvalue
valueofof
outputisisrelatively
relativelysimilar
similaracross
acrossallallenterprises,
enterprises,
changes
theshadow
shadowwage
wagerate
ratewill
willaffect
affectsocial
socialprofitability
profitabilitylevels
levelsmore
moreoror
changesininthe
less
lessequally,
equally,and
andcomparative
comparativeadvantage
advantagerankings
rankingsare
arelikely
likelytotoremain
remainunafunafthe
labor
fected.
theother
otherhand,
hand,ifif
theuse
useofof
laborvaries
variessignificantly
significantlybetween
between
fected.On
Onthe
enterprises
enterprises(e.g.,
(e.g.,mechanized
mechanizedvs.vs.manual
manualproduction
productiontechnologies),
technologies),changes
changesinin
the
theshadow
shadowwage
wagerate
ratewill
willhave
havedifferent
differenteffects
effectsononsocial
socialprofitability
profitabilitylevels,
levels,
and
andcomparative
comparativeadvantage
advantagerankings
rankingsare
arelikely
likelytotochange.
change.
1818
44
Uses
Usesof
ofDomestic
Domestic
Resource
ResourceCost
CostAnalysis
Analysis
The
Thefollowing
followingexamples
examplesdrawn
drawnfrom
fromCIMMYT
CIMMYTcase
casestudies
studiesdemonstrate
demonstratehow
how
the
theDRC
DRCframework
frameworkcan
canbebeadapted
adaptedtotoaddress
addressa awide
widerange
rangeofofcomparative
comparative
advantage
advantageissues.
issues.
Revealing
RevealingDistorting
DistortingEffects
EffectsofofGovernment
GovernmentPolicies
Policies
Before
BeforeRCRs
RCRsare
areformally
formallycalculated
calculatedand
andanalyzed,
analyzed,private
privateand
andsocial
socialprofitabilprofitability
itycan
canbebecompared
comparedfor
foreach
eachenterprise
enterprisetotodetermine
determineififand
andhow
howgovernment
government
policies
policiesinfluence
influenceproduction
productionincentives.
incentives.The
Theprivate
privateprofitability
profitabilityofofananenterenterprise
priseisissimply
simplythe
thenet
netreturns
returnstotoland
landcalculated
calculatedusing
usingmarket
marketprices;
prices;the
the
social
socialprofitability
profitabilityisisthe
thenet
netreturns
returnscalculated
calculatedusing
usingsocial
socialprices.
prices.Although
Although
technically
technicallynot
notpart
partofofDRC
DRCanalysis
analysisitself,
itself,this
thiscomparison
comparisoncan
canbebeperformed
performedatat
little
littleadditional
additionalcost,
cost,since
sincethe
theprocess
processofofcalculating
calculatingRCRs
RCRsgenerates
generatesallallthe
the
required
requireddata.
data.
The
Thepurpose
purposeofofcomparing
comparingprivate
privateand
andsocial
socialprofitability
profitabilityshould
shouldbebeobvious.
obvious.
Whenever
Wheneverdiscrepancies
discrepanciesexist
existbetween
betweenmarket
marketand
andsocial
socialprices,
prices,the
theinterests
interestsofof
farmers
farmersand
andofofthe
thenation
nationcan
candiverge.
diverge.AAcrop
cropcan
canbebeprofitable
profitabletotofarmers
farmers(e.g.,
(e.g.,
because
becauseofofoutput
outputororinput
inputprice
pricesubsidies),
subsidies),even
eventhough
thoughitsitsproduction
productionmay
maynot
not
represent
representananefficient
efficientuse
useofofresources
resourcesfrom
fromthe
thepoint
pointofofview
viewofofthe
thecountry.
country.
Conversely,
Conversely,a acrop
cropcan
canbebeunprofitable
unprofitabletotofarmers
farmers(e.g.,
(e.g.,because
becauseofofoutput
outputoror
input
inputprice
pricetaxation),
taxation),even
eventhough
thoughitsitsproduction
productionrepresents
representsananefficient
efficientuse
useofof
the
thenation's
nation'sresources.
resources.By
Bycomparing
comparingprivate
privateprofitability
profitabilitywith
withsocial
socialprofitabilprofitabilnotonly
onlycan
canthe
theoverall
overalleffect
effectofofgovernment
governmentpolicies
policiesbebemeasured,
measured,but
butthe
the
ity,
ity,not
influence
influenceofofindividual
individualpolicies
policiescan
canbebequantified
quantifiedbybydisaggregating
disaggregatingthe
theoverall
overall
discrepancy
discrepancyinto
intoitsitsconstituent
constituentparts.
parts.
Policy
Policyeffects
effectsononproducer
producerincentives
incentivesappear
appearasasthe
thedifference
differencebetween
betweenthe
theprice
price
ofofa aparticular
particularproduct
productororinput
inputvalued
valuedatatmarket
marketprices
pricesand
andatatsocial
socialprices.
prices.
These
Thesepolicy
policyeffects
effectsare
aresummarized
summarizedininTable
Table9.9.Thus,
Thus,the
theeffect
effectofofa atariff
tariffonon
imports
importsofofa acommodity
commodityisisindicated
indicatedbybyK,K,the
theeffect
effectofofa asubsidy
subsidyononfertilizer
fertilizerisis
indicated
indicatedbybyL,L,the
theeffect
effectofofa alabor
labormarket
marketdistortion
distortionisisindicated
indicatedbybyM,M,and
andthe
the
indirect
indirecteffects
effectsofofpolicies
policiesononcompeting
competingenterprises
enterprisesthat
thatlead
leadtotodistortions
distortionsinin
the
themarket
marketvalue
valueofofland
landare
areindicated
indicatedbybyNN(assuming
(assumingthat
thatthese
theseeffects
effectsare
are
reflected
reflectedresidually
residuallyininnet
netreturns
returnstotoland).
land).Total
Totalnet
netpolicy
policyeffects
effectsare
areindicated
indicated
bybythe
thedifference
differencebetween
betweenprivate
privateand
andsocial
socialnet
netprofitability
profitability(NPP-NSP),
(NPP-NSP),with
with
a apositive
positivevalue
valueindicating
indicatingthat
thatgovernment
governmentpolicies
policiesononthe
thewhole
wholeincrease
increase
private
privateprofitability,
profitability,and
anda anegative
negativevalue
valueindicating
indicatingthat
thatgovernment
governmentpolicies
policies
ononthe
thewhole
wholedecrease
decreaseprivate
privateprofitability.
profitability.Table
Table9 9also
alsoshows
showshow
howthe
thesame
same
information
informationused
usedininDRC
DRCanalysis
analysiscan
canalso
alsobebeused
usedtotocalculate
calculatemore
moreconvenconventional
tionalmeasures
measuresofofpolicy
policyincentives,
incentives,such
suchasasnominal
nominalprotection
protectioncoefficients
coefficients
andeffective
effectivesubsidy
subsidycoefficients
coefficients
(NPC),
(NPC),effective
effectiveprotection
protectioncoefficients
coefficients(EPC),
(EPC),and
(ESC).
(ESC).
..
1919
Table 9. Measures of comparative advantage and policy incentives
Tradables
Products Inputs
Market prices
World price equivalent
Opportunity cost of resources
Policy effect
Non·tradables
Capital
and labor
Land
Profit
D
NPP=A·B·C·D
A
F
B
G
C
K=A-F
L=G·B
M=H·C
H
Measures of comparative advantage
Resource cost ratio (RCR)
Measures of policy incentives
Nominal protection coefficient (NPC)
Effective protection coefficient (EPC)
Effective subsidy coefficient (ESC)
NSP=F·G·H·I
I
N=I-D O=NPP·NSP
=K+L+M+N
(H+I)/(F·G)
=
AlF
(A·B)/(F·G)
OfF
Source: Adapted from Monke and Pearson, 1989.
A good example of the insights which can be obtained by comparing private and
social profitability measures appears in Morris' (1988) study of wheat in Zimbabwe. Although this study was undertaken primarily to determine whether or
not Zimbabwe has a comparative advantage in wheat production, the DRC
framework of analysis at the same time provided an opportunity to evaluate
the claim made by commercial farmers that government policies strongly
discriminate against the farming sector.
Table 10 presents data from the Zimbabwe study showing the effects of government policies on production incentives for six major commercial crops--wheat,
tobacco, cotton, maize, soybeans, and groundnuts. The difference between the
private and social profitability of each crop, appearing in column 3, represents
the net effect of government policies on producer incentives. This overall policy
effect is disaggregated into effects attributable to specific taxes and subsidies
on producer prices, farm machinery prices, purchased inputs prices, labor
policy, agricultural credit, and other policies. The data show how government
policies affect private profitability, rarely positively (e.g., through subsidies to
agricultural credit programs), and mostly negatively (e.g., through controlled
producer prices, taxes on inputs, and wage policies). The data furthermore
reveal which crops are most affected (tobacco and cotton), and they indicate
which specific policies most influence producer incentives for each crop (producer price policy and labor policy, in most cases):
20
......
......
I\)
I\)
Crop
108.22
108.22
107.82
107.82
115.77
115.77
706.93
706.93
2,698.17
2,698.17
Maize
Maize
Soybeans
Soybeans
Groundnuts
Groundnuts
Cotton
Cotton
Tobacco
Tobacco
a
a
8,617.50
8,617.50
1,505.77
1,505.77
331.39
331.39
219.55
219.55
609.83
609.83
635.03
635.03
Social
Social
profit·
profit·
ability
ability
(Z$lha)
(Z$lha)
0.00
0.00
0.00
0.00
(485.88)
(485.88)
(4,928.19)
(4,928.19)
(111.73)
(111.73)
(215.62)
(215.62)
(798.83)
(798.83)
(5,919.33)
(5,919.33)
(336.00)
(336.00)
(329.45)
(329.45)
(503.97)
(503.97)
(501.61)
(501.61)
Producer
Producer
price
price
policy
policy
(Z$lha)
(Z$lha)
Net
Net
policy
policy
effect
effect
(Z$lha)
(Z$lha)
Includes
of energy,
energy, transport,
Includes effects
effects of
transport, and
and insurance
insurance policies.
policies.
Source:
Morris 1988a.
1988a.
Source: Morris
131.05
131.05
Wheat
Wheat
Crop
-
Private
Private
profit·
profit·
ability
ability
(Z$lha)
(Z$lha)
(67.01)
(67.01)
(76.94)
(76.94)
(39.41)
(39.41)
(55.96)
(55.96)
(52.55)
(52.55)
(54.81)
(54.81)
(91.86)
(91.86)
Purchased
Purchased
input
input
prices
prices
(Z$lha)
(Z$lha)
(29.86)
(29.86)
(25.93)
(25.93)
(28.14)
(28.14)
(22.14)
(22.14)
(41.79)
(41.79)
Farm
Farm
machinery
machinery
prices
prices
(Z$lha)
(Z$lha)
(618.50)
(618.50)
(218.84)
(218.84)
(137.87)
(137.87)
86.51
86.51
29.33
29.33
24.90
24.90
15.22
15.22
20.10
20.10
(89.16)
(89.16)
(26.72)
(26.72)
24.48
24.48
Credit
Credit
policy
policy
(Z$lha)
(Z$lha)
(39.19)
(39.19)
Labor
Labor
policy
policy
(Z$lha)
(Z$lha)
Difference
Difference due
due to:
to:
(342.80)
(342.80)
(26.57)
(26.57)
(20.76)
(20.76)
(19.54)
(19.54)
(19.60)
(19.60)
(26.16)
(26.16)
Other
Other
policies
policies
and
and
market
market 8
distortions
distortionsB
(Z$lha)
(Z$lha)
Table
of difference
profitability of
Table 10.
10. Sources
Sources of
difference between
between private
private and
and social
social profitability
of irrigated
irrigated crops
crops in
in Zimbabwe,
Zimbabwe, 1987
1987
Comparison of private and social profitability thus reveals that agricultural
policies in Zimbabwe provide disincentives for commercial farmers, since
private profitability is less than social profitability for all major commercial
crops. In other words, government policies are taxing away a portion of the
social profits. However, this tax occurs across all commodities with similar
incidence, so the relative ranking among crops in terms of private profitability
is not greatly altered from the ranking in terms of social profitability. This
information on the size and incidence of policy effects can help the government
adjust specific policies to bring about desired changes in the commercial farming sector.
Revealing Comparative Advantage Between Enterprises
Once private and social profitability have been compared, attention turns to indicators of comparative advantage (NSP and RCRs). Probably the most common
use ofDRC analysis is to determine comparative advantage between alternative enterprises, whether cropping activities or other types of agricultural
production activities. A good example of this use of DRC analysis appears in
Byerlee's (1985) study of wheat in Ecuador. This study was motivated by the
Ecuadoran government's concern over the sharp decline in wheat production
which occurred during the late 1970s and early 1980s, at a time when wheat
consumption was increasing rapidly. Policy makers were interested in determining whether or not wheat production represented an efficient use of the
nation's resources.
Table 11 presents data from the Byerlee study showing Ecuador's pattern of
comparative advantage. Wheat in Ecuador competes with three alternative enterprises: barley, potatoes, and dairying. The RCRs appearing in Table 11
(calculated using 1983 data) reveal that potato production represents the most
efficient use of Ecuador's domestic resources, followed by wheat production,
dairying, and finally barley production. 4 This was an important finding at the
time the study was carried out, since the ranking of farmer profitabilities
diverged sharply from this efficiency ranking.
In the Ecuador study, government policies (including a vastly overvalued
exchange rate and differential import tariffs across commodities) were found to
discriminate strongly against wheat, which was the least profitable of all crops
from the farmers' point of view. The strong effects of the exchange rate and of
the import tariffs were particularly noteworthy, since they suggested that the
most realistic way for Ecuador to increase self-sufficiency in wheat was
through macro-economic policy reform, rather than through technological
change within the agricultural sector.
4
22
Note that returns to potato production were not adjusted to reflect capital constraints and
limited marketing opportunities. See page 34 for a discussion of the problem of predicting general
equilibrium effects.
Table 11. Resource cost ratios for competing enterprises in Ecuador,
1983
Inttnii.ve
dairying
Wheat
Tradables
Outputs
Value of production
45,750
35,595
130,130
39,587
5,232
1,379
5,064
1,711
2,379
679
5,232
1,379
4,392
1,556
2,057
634
4,155
1,421
19,280
20,741
8,470
3,234
2,770
948
2,180
3,682
3,951
515
Primary Factors
Capital - machinery
Capital - working
Capital - other investment
1,823
660
1,823
594
1,309
3,990
0
873
0
7,000
Labor
Fertilizer distribution
Machinery distribution
Other inputs distribution
1,547
1,030
425
495
1,547
888
425
450
20,430
3,985
289
6,000
2,190
452
193
1,065
Landa
45,295
45,295
18,850
45,295
Net cost--primary factors
Value added--tradables
51,275
29,306
51,022
20,345
54,853
72,829
56,003
25,541
1.75
2.51
0.75
2.19
Inputs
Machinery depreciation
Fuels and oils
Fertilizer
Other tradable inputs
Seed
Other miscellaneous costs
Resource cost ratio
°
°
Source: Byerlee, 1985.
a
Residual returns to land in best competing alternative valued at world price equivalent. Cost
and returns for dairying adjusted to reflect six-month cropping cycle.
23
Revealing Comparative Advantage Between Regions
Domestic resource cost analysis can also be used to determine comparative advantage rankings between different regions within the same country in producing the same crop. Even if the production technology is similar, regional
differences in RCRs can arise because of differences in local cropping patterns
(which affect the alternate use value of primary factors), or because of differences in transportation and handling costs between the two regions (which
affect the social prices oftradables).
The study carried out by Byerlee and Longmire (1986) focusing on wheat in
Mexico illustrates this particular use of DRC analysis. Wheat is produced in
two widely separated regions of Mexico: the northern irrigated Yaqui Valley
(located far from major consumption points), and the central rainfed high
plateau or altiplano (located adjacent to major consumption points). A DRC
study was undertaken to shed light on the influence of government policies on
producer incentives and to determine Mexico's pattern of comparative advantage in wheat production. 5
Table 12 presents data from the Byerlee and Longmire study relating to the efficiency of wheat production in two different areas of Mexico. The RCR for
wheat in Sonora (Yaqui Valley) is close to 1, indicating that the value of domestic resources invested in wheat production is approximately equal to the net
value-added to tradables. In contrast, the RCR for wheat in Tlaxcala (altiplano)
is well below 1, indicating that the value of domestic resources invested in
wheat production is less than the net value-added to tradables.
These RCRs suggest that wheat production in the altiplano region is slightly
more efficient than wheat production in the Yaqui Valley. This finding came as
a surprise, since the relative efficiency of wheat production in the altiplano is
not widely recognized. While differences in production technologies in part
account for the different RCRs, differential transportation costs involved in
getting wheat to the major consumption point (Mexico City) also significantly
affect the comparative advantage rankings. These results would appear to
justify increased investment in research to develop improved rainfed production technologies for the altiplano. In addition, they would appear to justify a
government initiative to revitalize wheat production in an area where production had been declining.
5
24
In addition to the regional effect, differences in production technology also influenced the
comparative advantage rankings, but these are not discussed here.
Table 12. Resource cost ratios for two wheat-producing regions in
Mexico, 1985
Sonora
(Yaqui Valley)
Tlaxcala
(altiplano)
Tradables
Outputs
Value of production
Inputs
Machinery depreciation
Fuels and oils (net)
Spare parts
Purchased inputs
Other miscellaneous costs
221,990
69,040
8,330
8,540
6,660
50,720
9,210
7,020
3,860
5,570
12,950
9,700
30,590
9,580
2,220
17,160
81,060
20,760
2,060
1,860
0
1,070
140,610
138,530
25,750
29,940
1.02
0.86
Primary Factors
Capital (interest)
Labor
Maintenance
Water
Land
Net cost--primary factors
Value added--tradables
Resource cost ratios
Source: Byerlee and Longmire, 1985.
Revealing Comparative Advantage Between Technologies
Domestic resource cost analysis can also be used to determine comparative advantage between alternative production technologies used for a single crop.
Even though in this case the crop remains the same, social profitability measures and RCRs may differ considerably between production technologies if the
technologies require different types and quantities of inputs. Whenever production takes place across a range of farm sizes and types, it is important to
consider alternative technologies; otherwise, critical issues relating to farm size
and choice of technology are hidden. This is because the primary unit of analysis in DRC analysis is a single cultivated heGtare, which means that unless
farm size and technology level are explicitly taken into account, social profitability measures will be identical for a crop produced on a 1-ha farm and on a
100-ha farm.
25
Longmire's
and
Lugogo's
(1989)
study
wheat
Kenya
provides
a good
Longmire's
and
Lugogo's
(1989)
study
ofof
wheat
inin
Kenya
provides
a good
this
use
of DRC
analysis.
Wheat
produced
Kenya
large-scale
example
ofof
this
use
ofDRC
analysis.
Wheat
is is
produced
in in
Kenya
onon
large-scale
example
commercial
farms
using
high
levels
purchased
inputs
and
machinery.
The
commercial
farms
using
high
levels
ofof
purchased
inputs
and
machinery.
The
Kenyan
government
interested
expanding
wheat
production
into
the
Kenyan
government
is is
interested
inin
expanding
wheat
production
into
the
smallholder
sector,
which
would
necessitate
a shift
less
capital-intensive
smallholder
sector,
which
would
necessitate
a shift
to to
less
capital-intensive
production
technologies
characterized
greater
use
animal
power
and/or
production
technologies
characterized
byby
greater
use
ofof
animal
power
and/or
human
labor.
Domestic
resource
cost
analysis
was
undertaken
attempt
human
labor.
Domestic
resource
cost
analysis
was
undertaken
inin
anan
attempt
to to
assess
the
relative
efficiency
these
proposed
smallholder
production
technoloassess
the
relative
efficiency
ofof
these
proposed
smallholder
production
technologies
under
a range
farm
sizes.
Since
wheat
currently
grown
Kenya
gies
under
a range
ofof
farm
sizes.
Since
nono
wheat
is is
currently
grown
inin
Kenya
using
labor-intensive
technologies,
prototypical
enterprise
budgets
the
using
labor-intensive
technologies,
prototypical
enterprise
budgets
forfor
the
various
small-scale
technologies
were
estimated
based
technical
inputvarious
small-scale
technologies
were
estimated
based
onon
technical
inputoutput
parameters
from
several
Asian
countries
where
wheat
produced
output
parameters
from
several
Asian
countries
where
wheat
is is
produced
byby
smallholders.
smallholders.
Table
presents
data
from
the
Longmire
and
Lugogo
study
relating
the
Table
1313
presents
data
from
the
Longmire
and
Lugogo
study
relating
to to
the
relative
social
profitability
alternative
wheat
production
technologies
under
relative
social
profitability
ofof
alternative
wheat
production
technologies
under
a range
field
sizes.
The
NSP
figures
indicate
that,
the
smallest
sized
a range
ofof
field
sizes.
The
NSP
figures
indicate
that,
forfor
the
smallest
sized
fields
(0.5
and
1 ha),
labor-intensive
production
technology
represents
the
fields
(0.5
haha
and
1 ha),
labor-intensive
production
technology
represents
the
most
socially
profitable
use
domestic
resources,
whereas
the
larger
sized
most
socially
profitable
use
ofof
domestic
resources,
whereas
onon
the
larger
sized
represents
the
most
fields
(>4
ha),
fully
mechanized
production
technology
represents
the
most
fields
(>4
ha),
fully
mechanized
production
technology
socially
profitable
use
resources.
Interestingly,
none
the
intermediatesocially
profitable
use
ofof
resources.
Interestingly,
none
ofof
the
intermediatescale
mechanized
production
technologies
represent
a socially
profitable
use
scale
mechanized
production
technologies
represent
a socially
profitable
use
ofof
resources
fields
any
size.
resources
onon
fields
ofof
any
size.
These
results
suggest
that
labor-intensive
production
technologies
would
These
results
suggest
that
labor-intensive
production
technologies
would
bebe
socially
profitable
Kenya
smallholder
wheat
producers
with
restricted
socially
profitable
inin
Kenya
forfor
smallholder
wheat
producers
with
restricted
access
land.
However,
the
absence
constraints
farm
size,
wheat
access
to to
land.
However,
inin
the
absence
ofof
constraints
onon
farm
size,
wheat
production
would
remain
most
efficient
larger
landholdings
where
high
production
would
remain
most
efficient
onon
larger
landholdings
where
high
Table
Social
profitability
wheat
production
technologies
size
Table
13.13.
Social
profitability
ofof
wheat
production
technologies
byby
size
field,
Kenya,
1987
ofof
field,
Kenya,
1987
Average
field
size
Average
field
size
Wheat
technology
Wheat
technology
0.50.5
haha
1 ha
1 ha
4ha
4ha
10ha
10ha
40ha
40ha
(Ksh/ha)(Ksh/ha)
(Ksh/ha)(Ksh/ha)
(Ksh/ha)(Ksh/ha)
(Ksh/ha) (Ksh/ha)
(Ksh/ha)
(Ksh/ha)
plow
and
labor
intensive
OxOx
plow
and
labor
intensive
Small
motorized
Small
motorized
South Asian
SouthAsian
a a
Large
reaper
and
large
thresher
Large
reaper
and
large
thresher
Fully
mechanized
Fully
mechanized
355
355
(355)
(355)
(63)
(63)
(739)
(739)
(473)
(473)
Source:
Longmire
and
Lugogo,
1989.
Source:
Longmire
and
Lugogo,
1989.
a Tractor
land
preparation,
mechanized
threshing.
a Tractor
land
preparation,
mechanized
threshing.
2626
3939
(568)
(568)
(622)
(622)
(678)
(678)
(39)
(39)
(544)
(683)
(544)
(683)
(1,059) (1,175)
(1,175)
(1,059)
(924)
(854)
(924)
(854)
(878)
(939)
(878)
(939)
544
683
544
683
(781)
(781)
(1,255)
(1,255)
(974)
(974)
(981)
(981)
781781
levels of mechanization are feasible. This information is potentially useful for
Kenyan policy makers, who are currently faced with difficult decisions about
land use and agricultural development. Knowledge of the likely efficiency cost
of breaking up large landholdings for redistribution to smallholder farmers will
help inform the policy debate.
Setting Agricultural Research Priorities
Finally, the DRC framework can be used to help research managers decide how
to allocate scarce resources between commodities, regions, or technologies. By
using sensitivity analysis to vary technical parameters in the enterprise
budgets, research managers can estimate the productivity gains needed to alter
existing comparative advantage rankings. This information can be weighed
against the probability and likely cost of achieving such productivity gains.
An example of this particular use ofDRC analysis appears in Morris' (1988)
study of wheat in Zimbabwe. Under a drought scenario in which water was
assigned a high opportunity cost to reflect its alternative use value in tobacco
production, Zimbabwe was found to lose its comparative advantage in wheat
production. By implication, the introduction of more water-efficient wheat
production technologies might allow the country's comparative advantage in
wheat production to be maintained even in periods when water is scarce.
Sensitivity analysis was used to test the effect on the comparative advantage
rankings of reduced irrigation requirements. Specifically, the variable costs
associated with irrigating wheat were gradually lowered to determine the
irrigation level at which wheat would become competitive with other crops.
Break-even analysis revealed that Zimbabwe's comparative advantage in wheat
production would be maintained even during periods of drought if the crop's
irrigation requirements could be reduced from the present 720 mm to around
420 mm (gross). Research managers can use this information to help decide
whether or not to allocate increased resources to irrigation management
research or to breeding for more drought-resistant varieties.
27
55
ProblemsCommonly
Commonly
Problems
EncounteredininUsing
UsingDRC
DRCMethods
Methods
Encountered
Before
turning
a discussion
of the
strengths
weaknesses
of the
DRC
Before
turning
to atodiscussion
of the
strengths
andand
weaknesses
of the
DRC
framework,
is useful
to cite
a number
of problems
commonly
encountered
framework,
it isit useful
to cite
a number
of problems
commonly
encountered
in in
carrying
applied
comparative
advantage
analysis.
Some
of these
problems
carrying
outout
applied
comparative
advantage
analysis.
Some
of these
problems
practical
in nature,
while
others
relate
more
to conceptual
difficulties
with
areare
practical
in nature,
while
others
relate
more
to conceptual
difficulties
with
theory
underlying
DRC
approach.
of the
problems
have
a bearing
thethe
theory
underlying
thethe
DRC
approach.
AllAll
of the
problems
have
a bearing
on on
likely
accuracy
of empirical
research
results
therefore
should
thethe
likely
accuracy
of empirical
research
results
andand
therefore
should
be be
recognized
analyst
considering
undertaking
a comparative
advantage
recognized
by by
thethe
analyst
considering
undertaking
a comparative
advantage
study.
study.
Extensive
Data
Requirements
Extensive
Data
Requirements
Because
of the
extensive
data
requirements,
applied
comparative
advantage
Because
of the
extensive
data
requirements,
applied
comparative
advantage
analysis
tends
to be
costly
in terms
of time
effort.
Field
visits
must
analysis
tends
to be
costly
in terms
of time
andand
effort.
Field
visits
must
be be
undertaken
early
to identify
enterprise
substitution
possibilities
available
undertaken
early
to identify
thethe
enterprise
substitution
possibilities
available
current
(and
potential)
production
alternatives
must
taken
to farmers,
since
current
(and
potential)
production
alternatives
must
be be
taken
to farmers,
since
account
if the
alternative
values
of primary
factors
to be
estimated
intointo
account
if the
alternative
useuse
values
of primary
factors
areare
to be
estimated
accurately.
Once
production
alternatives
have
been
identified,
a detailed
accurately.
Once
production
alternatives
have
been
identified,
a detailed
enterprise
budget
must
constructed
each
production
alternative,
includenterprise
budget
must
be be
constructed
for for
each
production
alternative,
includcomplete
accounting
fixed
as well
as variable
costs
of production.
Often
inging
complete
accounting
for for
fixed
as well
as variable
costs
of production.
Often
requires
construction
of separate
budgets
capital
inputs
(e.g.,
vehicles,
thisthis
requires
construction
of separate
budgets
for for
capital
inputs
(e.g.,
vehicles,
farm
machinery,
irrigation
systems,
barns
sheds).
many
instances,
farm
machinery,
irrigation
systems,
barns
andand
sheds).
In In
many
instances,
budget
data
already
available
from
secondary
sources,
obviating
budget
data
willwill
already
be be
available
from
secondary
sources,
obviating
thethe
need
primary
data
collection
at the
farm
level.
However,
even
secondary
need
forfor
primary
data
collection
at the
farm
level.
However,
even
secondary
sources
must
carefully
verified
accuracy
representativeness.
This
sources
must
be be
carefully
verified
for for
accuracy
andand
representativeness.
This
a frustrating
exercise,
since
farmers
engage
a wide
range
of produccancan
be be
a frustrating
exercise,
since
farmers
engage
in ainwide
range
of producpractices
typically
a range
of prices
inputs
outputs.
tiontion
practices
andand
typically
faceface
a range
of prices
for for
inputs
andand
outputs.
After
enterprise
budgets
have
been
constructed
inputs
outputs
Mter
thethe
enterprise
budgets
have
been
constructed
andand
inputs
andand
outputs
dis-disaggregated
their
tradable
primary
factor
components,
complete
aggregated
intointo
their
tradable
andand
primary
factor
components,
thethe
complete
vectors
of market
social
prices
must
constructed.
Market
prices
vectors
of market
andand
social
prices
must
be be
constructed.
Market
prices
cancan
usually
obtained
directly
from
agricultural
input
suppliers
other
firms
usually
be be
obtained
directly
from
agricultural
input
suppliers
andand
other
firms
providing
goods
services
to the
agricultural
sector,
generally
andand
services
to the
agricultural
sector,
butbut
thisthis
generally
providing
goods
requires
a costly
time-consuming
data
collection
effort.
Once
market
prices
requires
a costly
andand
time-consuming
data
collection
effort.
Once
market
prices
have
been
collected,
estimation
of social
prices
requires
additional
knowledge
have
been
collected,
estimation
of social
prices
requires
additional
knowledge
of of
world
reference
prices
tradables,
global
transportation
rates
handling
world
reference
prices
for for
tradables,
global
transportation
rates
andand
handling
costs,
government
import
export
policies,
taxes,
subsidies,
costs,
government
import
andand
export
policies,
taxes,
subsidies,
etc.etc.
times
it may
possible
to reduce
data
requirements
estimating
social
At At
times
it may
be be
possible
to reduce
data
requirements
by by
estimating
social
prices
only
some
inputs
outputs.
If market
prices
heavily
distorted
inputs
andand
outputs.
If market
prices
areare
heavily
distorted
prices
for for
only
some
a small
number
of highly
influential
policies
(e.g.,
a vastly
overvalued
by by
a small
number
of highly
influential
policies
(e.g.,
a vastly
overvalued
exchange
rate,
producer
prices
that
way
of line
with
world
equivalent
exchange
rate,
producer
prices
that
areare
way
outout
of line
with
world
equivalent
prices,
large
subsidies
fertilizers),
reasonably
accurate
comparative
advanprices,
large
subsidies
on on
fertilizers),
reasonably
accurate
comparative
advantage
rankings
sometimes
obtained
adjusting
only
those
policies
tage
rankings
cancan
sometimes
be be
obtained
by by
adjusting
for for
only
those
policies
ignoring
other
price
distortions.
Common
sense
help
in identifying
andand
ignoring
all all
other
price
distortions.
Common
sense
cancan
help
in identifying
highly
influential
policies,
a priori
there
is no
way
of knowing
with
thethe
highly
influential
policies,
butbut
a priori
there
is no
way
of knowing
with
28 28
certainty which distortions alter comparative advantage rankings. As the
Zimbabwe case study shows, some government policies affect all enterprises
more or less equally, which may decrease or increase private profitability
relative to social profitability but may not alter comparative advantage
rankings.
In summary, the extensive data requirements of the DRC methodology suggest
that DRC analysis should not be undertaken if the researcher is unable or
unwilling to invest considerable time and effort in data collection activities.
Efforts to cut corners, for example by failing to cost out all major production
alternatives or by guessing at important opportunoty costs, may seriously affect
the quality of the research results.
Determining Social Prices for Primary Factors
A recurring problem in DRC analysis involves the estimation of social prices for
primary factors. As discussed earlier, social prices for primary factors are
supposed to represent the true economic value of these factors in the economy.
This is usually estimated as the opportunity cost value, i.e., as the value of the
factor in its most profitable alternative use. While the concept of alternative
use value is straightforward in principle, in practice each type of primary factor
presents its own unique estimation problems.
The importance of accurately estimating the social price of land cannot be overemphasized, since this price frequently is decisive in determining comparative
advantage rankings. As mentioned previously, a reasonable approximation of
the opportunity cost of land can be obtained from the enterprise budgets
themselves as the net social returns to land in the most socially profitable
cropping alternative. (One advantage of this approach is that it automatically
removes policy distortions on competing crops.) In most of the CIMMYT
studies, this value has been found to differ significantly from the market price
ofland (annualized purchase price or rental value). However, if the enterprise
budgets are to be used for estimating the social price of land, it is important
that all reasonable production alternatives be identified and costed out. As
mentioned earlier, a glaring weakness of many DRC studies is that they focus
on one enterprise of immediate interest and fail to consider alternative enterprises. Social prices of land therefore do not accurately reflect alternate use
values, and comparative advantage findings may be seriously distorted.
Estimating accurate social prices for labor can also be important in DRC
analysis, especially when production alternatives differ significantly in their
labor requirements and there is evidence that the labor market is distorted.
Typically this occurs when government policies (e.g., minimum wage legislation) have raised the market price of agricultural labor above the true marginal
value product. Several methods have been developed for formally calculating
the shadow wage rate of agricultural labor (see McDiarmid 1977; Gittinger
1982). Alternatively, the DRC analyst may simply want to adjust market wage
rates with the help of a conversion factor obtained from published studies,
29
assuming these exist. Sensitivity analysis should then be carried out to determine the importance of the social price assigned to labor. If the final comparative advantage rankings are found to be highly sensitive to the price assigned
to labor, this should be reported.
The problems inherent in estimating the real opportunity cost of capital are
well known and have been discussed elsewhere (for example, see Gittinger
1982). For DRC analysis, it is usually not worth spending a lot of time and
effort estimating the opportunity cost of capital unless the value selected is
likely to have a decisive effect on the comparative advantage rankings. (This
might be the case when potential production alternatives differ drastically in
capital-intensiveness, e.g., a highly mechanized technology vs. a labor-intensive
technology). Therefore, before any calculations are undertaken, it is wise to
carry out sensitivity analysis to determine the importance of the social price
assigned to capital. If comparative advantage rankings are found to be robust
with respect to the opportunity cost of capital, an informed "guesstimate" will
probably be satisfactory.
In some instances a primary factor other than land, labor, or capital will
represent the limiting factor in production, and it will be necessary to estimate
a social price for this factor as well. For example, in the Zimbabwe wheat
study, irrigation water was determined to be the limiting factor of production
during drought years. Since calculation of the marginal value product of
production inputs is usually impractical (especially in the presence of policyinduced distortions), it will often be necessary to resort to less rigorous approximations to estimate an alternative use value. In the Zimbabwe study, an
average value of irrigation water was calculated based on the difference in
profitability between rainfed and irrigated production--not a very exact approximatio~ of the marginal value product, but almost certainly better than the
naive assumption of no difference in the value of water between normal rainfall
years and drought years. (Alternatively, a simple linear programming model of
a representative irrigated farm could have been used to calculate the shadow
price of irrigation water.)
To recapitulate, estimation of social prices for primary factors tends to be
difficult in practice and often requires use of imperfect approximation techniques. Sometimes it will be possible to demonstrate through sensitivity analysis that the comparative advantage rankings are robust with respect to the
social prices assigned to primary factors, but frequently the rankings will be
sensitive to changes in social prices. In such cases, it is important to report the
results of the sensitivity analysis to aid in the interpretation of research
findings.
30
Determining Social Prices for Non-traded Tradables
Problems sometimes arise in establishing social prices for non-traded tradables.
Non-traded tradables typically appear when transportation and handling
charges involved in getting a commodity to and from world markets introduce a
wide gap between the import and export parity prices. Whenever the domestic
price lies within this gap, importing and exporting will both be uneconomical.
For example, in the country depicted in Figure 1, domestic supply of wheat
equals domestic demand at P E and QE' Because of high transport and handling
costs involved in gaining access to world markets, trade in wheat will be
uneconomical; the import parity price is too high to induce consumption of
imports, and the export parity price is too low to encourage additional production for export.
How should social prices be determined for non-traded tradables? If the
market price is an undistorted market-clearing price that effectively equilibrates domestic supply and demand (as in Figure 1), the market price accurately reflects the product's economic value and can be used as an approximation of the social price. But if the market price reflects distortions due to policy
effects or market failure, an adjustment should be made to arrive at the social
price. In Figure 2, government policies have introduced a gap between the
controlled market price (Pc) and the theoretical market-clearing equilibrium
price (PE)' In this example, the controlled market price (Pc) has been set lower
than the theoretical market-clearing equilibrium price (P E)' and demand (QD)
therefore exceeds supply (Q~). The government can react to this excess demand
in two ways: 1) by doing nothing and allowing the excess demand to persist, or
2) by importing the additional quantity necessary to satisfy demand at the
controlled price and selling it to consumers at a loss (represented in Figure 2 by
the shaded area).
Whatever the government's reaction, the difference between P E and Pc represents a policy-induced distortion which should be eliminated for DRC analysis.
The appropriate correction is to estimate a social price equal to the theoretical
market-clearing price (PE)' (Note that even though the government's response
may be to import QD-QS at the import parity price, it is incorrect to set the
social price equal to the import parity price--a common error in the literature,
especially when trade is taking place--since in the absence of price controls
domestic supply would increase to satisfy demand at a price well below the
import parity price.)
31
Price
P import
parity
P
export
parity
Demand
o
QE
Quantity
Figure 1.Conditions giving rise to a non-traded tradable.
Price
P
import
parity
PE 1 - - - - - - -
Pc 1 - - - - - - - - - - : 0
Pexport
t---~--+---+-----+------+----~-~
parity
Demand
o
Quantity
Figure 2. Establishing social prices for non-traded tradables.
32
Estimating an Equilibrium Exchange Rate
Since the exchange rate used to convert between domestic and foreign currencies affects the social prices of all tradables, it is important that exchange rate
distortions be properly identified and corrected for purposes of DRC analysis.
Estimating an accurate real equilibrium exchange rate can present major
problems. Among other things, the issue may be politically sensitive; government officials in many developing countries are reluctant to acknowledge,
much less discuss, exchange rate over- or undervaluation. Even when the issue
can be discussed openly, often there is no easy way to estimate the degree of
exchange rate distortion.
Ifno reliable estimates of the real equilibrium exchange rate are available from
secondary sources, a formal estimation procedure may have to be undertaken.
The purchasing power parity (PPP) method is usually the simplest. The most
basic version of the PPP method involves selecting a base year in which the
official exchange rate is judged to have been valued at its true equilibrium rate
in relation to the currency used by the dominant trading partner (often the US
dollar) and deflating the official exchange rate in subsequent years by the
difference in domestic inflation rates:
where:
E*
E
=
real exchange rate
nominal exchange rate (units of domestic
currency per unit of foreign currency)
price deflator for domestic currency
price deflator for foreign currency
=
=
=
An improvement on this basic version of the PPP method is to calculate a
trade-weighted equilibrium exchange rate, essentially a composite exchange
rate calculated against a basket of foreign currencies, weighted according to
actual trading patterns:
ET*
=
where:
ET*
Ew
=
=
=
=
=
=
real trade-weighted exchange rate
nominal exchange rate (units of domestic
currency per unit of currency in country w)
value of trade with country w
value of total trade
price deflator for domestic currency
price deflator for currency in country w
33
Alone, PPP methods may be inadequate for estimating the equilibrium exchange rate. The existence of trade barriers (e.g., import quotas that drive up
the cost of imported goods, or high tariff protection for domestic industries) can
make selection of a base year problematic unless additional adjustments are
made. Several more sophisticated methods for calculating the equilibrium
exchange rate have been described in the literature (see Dornbusch and Helmers1988; Krueger, Schiff, and Valdes 1988; Ward 1976). No particular method
is advocated here, since much will depend on the availability of data and the
resources for analysis.
Predicting General-Equilibrium Effects
One criticism sometimes directed at the DRC approach concerns the difficulty
of predicting general-equilibrium effects, both on the output side and on the
input side. The criticism is valid, since the RCR for a particular enterprise does
not indicate how much incremental production can be absorbed by the economy
before prices are affected. In certain instances, changes in production patterns
are likely to have significant effects on prices. For example, if potato production
were to increase dramatically in Ecuador, the domestic market would soon be
saturated and potato prices would fall. This would affect the net social profitability of potato production, because potatoes in Ecuador are classified as a nontraded tradable, and the market price is equivalent to the social price. Conventional DRC analysis fails to consider this likely general-equilibrium effect.
Similarly, DRC analysis assumes fixed factor prices, under the implicit assumption that changing patterns of production would have no effect on factor
prices in the future. This assumption may be unrealistic. For example, if
tobacco production were to increase dramatically in Zimbabwe, shortages of
labor could be expected to develop during harvesting, which would result in
upward pressure on wage rates. Once again, conventional DRC analysis fails to
take this likely general-equilibrium effect into account, since it assumes that
wage rates remain constant regardless of the quantity oftobacco produced.
Before attempting to incorporate general-equilibrium effects into the DRC
framework, it is important to recognize the tradeoff between realism and
practicality. While the DRC framework described in this paper admittedly is
based on a partial-equilibrium approach, it is precisely this feature which
makes DRC analysis practical. Realism in the form of general-equilibrium
effects (as in multi-market simulation models) can be introduced only at the
expense of increased complexity, which may make the results unintelligible to
many policy makers.
If important general-equilibrium effects are anticipated, however, these can be
incorporated into DRC analysis in several ways. The simplest approach is to
judge whether or not enterprises which currently are socially profitable offer
viable production alternatives with real possibilities for expansion. If the
answer is clearly no (because of constraints on either the input or the output
34
side), these enterprises should be eliminated from further consideration. Thus,
although potato production was the most socially profitable enterprise in
Ecuador, clearly the market for Ecuadoran potatoes is limited, and potatoes
should not receive high priority in the allocation of scarce research resources. If
the answer is uncertain, sensitivity analysis can help indicate whether a crop's
current ranking would be threatened by possible future changes in input or
output prices. If social profitability levels turn out to be highly sensitive to one
or more prices whose future levels remain uncertain, it may be appropriate to
carry out a formal market study, including price projections.
Comparative Advantage vs. Competitive Advantage
A number of critics have pointed out that the DRC approach does not really
measure comparative advantage, since social profitability levels and RCRs are
calculated using world reference prices which may themselves reflect significant policy-induced distortions. (For example, USDA estimates that the world
price of wheat has been maintained approximately 33% below the level which
would prevail in the absence of the farm support policies and export enhancement programs of the industrialized countries.) Consequently, rather than
providing a true measure of relative efficiency in production, DRC analysis
merely generates a measure of one country's ability to compete with prevailing
world prices. Put another way, this means that DRC does not measure comparative advantage, but rather competitive advantage.
While technically correct, this criticism of the DRC approach does not affect its
usefulness for policy-making. If price distortions in world markets are expected
to continue, the social price oftradables is correctly estimated as the long-term
world reference price, since this represents the effective opportunity cost value
oftradables to the economy. Consequently, policy decisions based oli DRCgenerated comparative advantage rankings retain their validity.
35
6
Lessons From
the CIMMYT Case Studies
During the past decade, concern for agricultural policy reform in many developing countries has generated renewed interest in empirical measurement of
patterns of comparative advantage. One consequence has been a rapid proliferation of DRC studies, since the DRC framework provides a practical approach
for generating quantitative measures of comparative advantage. This paper has
reviewed the DRC methodology and presented examples illustrating different
applications of the DRC approach. In closing, it seems appropriate to reflect
upon the usefulness of applied comparative advantage analysis for CIMMYT
and for our clients in the national agricultural research systems, as indicated
by the reception that the case studies have received.
CIMMYT's original decision to undertake a series of applied comparative
advantage studies was motivated by an interest in developing methods to help
the decision making of research administrators in national agricultural research systems (Longmire and Winkelmann 1985). The idea was that research
administrators would be able to allocate resources more efficiently if they had
knowledge of patterns of comparative advantage--between crops, between
regions, or between alternative production technologies. Thus, DRC analysis
was seen as a way of generating the sort of information which would help the
director of a national research program decide whether to launch a new project
in small grains or in oilseeds, whether to build a research station for wheat in
Region A or in Region B, or whether to hire an irrigation scientist or a rainfed
crop specialist.
To what extent have the CIMMYT DRC studies been able to generate this sort
of information? While most of the studies have generated results which could
help inform research resource allocation, many of the studies apparently have
failed to elicit the anticipated interest among research administrators in the
national systems. One possible reason for this apparent lack of interest may be
that the connection between DRC analysis and research resource allocation is
not always obvious. Many research managers simply may not know what to do
with the efficiency indicators generated by DRC analysis, especially since
research policy is not driven exclusively by efficiency considerations.
Given that non-efficiency considerations frequently enter into research priority
setting, the link between DRC analysis and research resource allocation
becomes quite complex. To see this, consider the three hypothetical enterprises
depicted in Figure 3. Each enterprise is assigned two social profitability levels-one calculated on the basis of existing technology (SPo)' and one estimated on
the basis of improved technology expected to become available after investment
in agricultural research (SP ). For all three enterprises, social profitability
under "improved technology~' is higher than social profitability under "existing
technology." However, the absolute values of the social profitability levels differ
between the three enterprises, as do the changes in social profitability resulting
from investment in agricultural research (DSP, represented by the length of
the arrows). Enterprise 1 is socially profitable under current technology and
becomes even more socially profitable under improved technology, although the
36
Social
profitability
+
o
Enterprise 1
Enterprise 2
Enterprise 3
Figure 3. Using DRC analysis for research resource allocation.
increase in profitability is small. Enterprise 2 is not socially profitable under
current technology, although it becomes socially profitable under improved
technology as the result of a medium increase in profitability. Enterprise 3 is
not socially profitable under current technology or under improved technology,
despite a large increase in profitability (actually a decrease in unprofitability).
Which enterprise merits highest priority with regard to research investment?
The answer depends on a number of considerations, including:
•
the importance of non-efficiency considerations in setting research policy
(e.g., projected distributional impacts oftechnological change across
regions or income classes; food security objectives);
•
the size of the total agricultural research budget (which determines how
many enterprises can be worked on simultaneously);
•
the relative costs of different kinds of research programs (which might
vary considerably among the three enterprises)
37
• • the
theconfidence
confidenceofofresearch
researchadministrators
administratorsthat
thatSP
SP1 1has
hasbeen
beenestimated
estimated
correctly
correctly(which
(whichmay
mayrequire
requirestrong
strongassumptions
assumptionsabout
aboutthe
thefuture
futureimpact
impact
ofoftechnologies
technologiesstill
stilltotobe
bedeveloped);
developed);
• • the
theperceived
perceivedlikelihood
likelihoodthat
thatfarmers
farmerswill
willbe
bewilling
willingand
andable
abletotosubstisubstitute
tuteone
oneenterprise
enterprisefor
foranother
anotheronce
oncethe
theimproved
improvedtechnology
technologyisis
available;
available;
• • the
thelevel
levelofofconfidence
confidenceininprojections
projectionsofoflikely
likelyfuture
futuredevelopments
developmentsininthe
the
global
globaleconomy
economy(e.g.,
(e.g.,supply
supplyand
anddemand
demandofofparticular
particularcommodities;
commodities;
world
worldprices);
prices);
• • the
theprobability
probabilitythat
thatpolicy-induced
policy-induceddistortions
distortionswill
willbe
beremoved
removed(which
(which
may
mayeliminate
eliminatesocially
sociallyunprofitable
unprofitableenterprises
enterprisessuch
suchasasEnterprise
Enterprise3).
3).
Given
Giventhese
theseconsiderations,
considerations,ititisispossible
possibletotoconceive
conceiveofofsituations
situationsininwhich
whichitit
would
wouldbe
bedesirable
desirabletotoinvest
investininresearch
researchon
onanyone
anyoneofofthe
thethree
threeenterprises,
enterprises,oror
totoinvest
investininresearch
researchon
onvarious
variouscombinations
combinationsofofthe
thethree.
three.This
Thismerely
merelyserves
serves
totoemphasize
emphasizethat
thatthe
theefficiency
efficiencyindicators
indicatorsgenerated
generatedby
byDRC
DRCanalysis
analysisdo
donot
not
provide
provideunambiguous
unambiguousinvestment
investmentguidelines
guidelinesfor
forresearch
researchmanagers.
managers.Provided
Provided
they
theyare
areused
usedcarefully,
carefully,DRC
DRCresults
resultscan
canserve
serveasasaabasis
basisfor
forranking
rankingenterenterprises
prisesininterms
termsofofcurrent
currentand
andexpected
expectedfuture
futuresocial
socialprofitability,
profitability,asaswell
wellasasfor
for
segregating
segregatingthose
thoseenterprises
enterprisesthat
thatwaste
wasteforeign
foreignexchange
exchangeorordomestic
domesticcurcurrency.
rency.Through
Throughsensitivity
sensitivityanalysis,
analysis,DRC
DRCresults
resultscan
canalso
alsohelp
helpindicate
indicatethe
thesize
size
ofofchanges
changesininfuture
futureprices
pricesorortechnologies
technologieslikely
likelytotoalter
altercurrent
currentpatterns
patternsofof
comparative
comparativeadvantage.
advantage.However,
However,DRC
DRCanalysis
analysisisisnot
notaa"cookbook"
"cookbook"method
methodfor
for
mechanically
mechanicallycranking
crankingout
outresource
resourceallocation
allocationrules.
rules.Rather,
Rather,the
theDRC
DRCframeframework
workgenerates
generatesone
onetype
typeofofinformation
informationwhich
whichmust
mustbe
becombined
combinedwith
withother
other
types
typesofofinformation
informationtotoinform
informthe
thedecision
decisionmaking
makingofofpolicy
policymakers
makersand
andreresearch
searchmanagers.
managers.The
Thechallenge
challengewill
willbe
betotodevise
devisemethods
methodsfor
fortranslating
translatingthe
the
information
informationgenerated
generatedby
byDRC
DRCanalysis
analysisinto
intoaaform
formininwhich
whichititisismore
moreaccesaccessible
sibletotoresearch
researchmanagers.
managers.
Yet
Yetififthe
theCIMMYT
CIMMYTcomparative
comparativeadvantage
advantagestudies
studieshave
havefailed
failedtotogenerate
generatethe
the
expected
expectedinterest
intereston
onthe
thepart
partofofresearch
researchmanagers,
managers,they
theyhave
havebeen
beenwell
wellrereceived
ceivedby
byothers.
others.The
Theresults
resultsgenerated
generatedby
byindividual
individualcase
casestudies
studies(both
(boththe
the
comparative
comparativeadvantage
advantagerankings
rankingsand
andespecially
especiallythe
theresults
resultsofofcomparative
comparative
profitability
profitabilityanalysis)
analysis)frequently
frequentlyhave
havebeen
beenseized
seizedby
byproducer
producergroups
groupsand/or
and/or
government
governmentofficials
officialstotobe
beintroduced
introducedasasevidence
evidenceininthe
thepolicy
policydialogue.
dialogue.This
This
has
hashappened
happenedeven
eventhough
thoughthe
theauthors
authorshave
havemade
madelittle
littleeffort
efforttotopublicize
publicizethe
the
policy
policyimplications
implicationsofofthe
theresearch
researchfindings,
findings,since
sinceCIMMYT
CIMMYThas
hasnot
notsought
soughttoto
participate
participatedirectly
directlyinindomestic
domesticagricultural
agriculturalpolicy
policydebates.
debates.
38
38
Our experience suggests that comparative advantage studies can help inform
the policy dialogue in two important ways. First, social profitability analysis
reveals how government policies influence agricultural production incentives.
Knowledge of the distribution and strength of policy-induced price distortions
can be extremely valuable in helping policy decision-makers make desired
adjustments, especially knowledge about the relative importance of targeted
taxes and subsidies (e.g., producer price controls, subsidies on inputs) vs.
general economy-wide policies (e.g., labor policy, exchange rate policy). Very
few other analytical methods reveal these effects in such easily intelligible
terms.
Second, social profitability measures and RCRs are quantitative indicators of
comparative advantage which can help decision-makers to assess the cost of
promoting one set of production activities at the expense of another. It is
important to recognize, of course, that comparative advantage rankings are
based on economic efficiency alone and fail to take into account non-efficiency
considerations, which often figure prominently in the policy debate (e.g.,
income distribution, food security, sustainability). In this respect, DRC analysis
clearly must be seen as only one tool for applied policy analysis. Nevertheless,
most agricultural policy decisions start with an assessment of the economic
costs and benefits of alternative courses of action, and the advantage of the
DRC framework is that it provides quantitative measures of the opportunity
cost of pursuing different alternatives. The DRC framework thus represents an
important element--although certainly not the only element--in the policy
analyst's conceptual tool kit.
An additional point that should be made concerning the practical utility of the
DRC approach is the tremendous flexibility introduced by micro-computer
technology. Once enterprise budgets and price vectors have been developed and
recorded in a computerized spreadsheet program, adjustments can easily be
made to reflect changes in technology or prices. Social profitability measures
and RCRs can then be calculated on an annual basis (or more frequently if
desirable) to obtain a dynamic picture of evolving changes in patterns of
comparative advantage. This is obviously very attractive to policy analysts,
since it means that once the initial work of calculating RCRs is completed,
comparative advantage rankings can subsequently be updated at little additional cost.
Final.lY, the CIMMYT comparative advantage studies have provided a practical
lesson in the dynamics of policy reform. It has become clear from the case
studies that the use of comparative advantage results ultimately depends to a
large extent on who is perceived as having carried out the analysis. When
information on comparative advantage is needed to appraise a proposed project
loan, it probably matters little who actually does the calculations, whether a
local economist or an expatriate "expert" brought in from outside. But if DRC
results are to have an effect on domestic policies, in practice it seems to matter
39
a great deal who does the analysis. Comparative advantage studies by nature
tend to be politically controversial, because among other things they reveal the
distributional effects of government policies. Such information can be invaluable in the policy dialogue, but only if it is introduced by someone who is able to
playa forceful role in arguing for policy reform. What this means is that often
it will not be appropriate for comparative advantage studies to be carried out
by expatriate scientists, especially scientists working for international research
centers which have no desire to become directly involved in domestic policy
debates. Scientists in international agricultural research centers can offer
training in DRC methods to local researchers, as well as technical assistance in
carrying out applied DRC studies, but it is the local researchers themselves
who must take the lead in conducting the studies and publicizing the results.
For only if a strong local advocate is willing to take a stand in the policy debate
are the research results likely to make a difference.
40
References
Byerlee, D. 1989. Bread and butter issues in Ecuadoran food policy: A comparative advantage approach. World Development 17(10).
Byerlee, D. 1985. Comparative Advantage and Policy Incentives for Wheat
Production in Ecuador. CIMMYT Economics Program Working Paper
01/85. Mexico, D.F.: CIMMYT.
Byerlee, D., and J. Longmire. 1986. Comparative Advantage and Policy Incentives for Wheat Production in Rainfed and Irrigated Areas of Mexico.
CIMMYT Economics Program Working Paper 01186. Mexico, D.F.:
CIMMYT.
Dornbusch, R., and L. Helmers (eds.) 1988. The Open Economy: Tools for
Policymakers in Developing Countries. New York: Oxford University
Press.
Gittinger, J. 1982. Economic Analysis ofAgricultural Projects. Baltimore:
Johns Hopkins University Press.
Harrington, L., and S. Sat-thaporn. 1984. Wheat in Chiang-Rai, Thailand: A
preliminary look at comparative advantage. In Wheats For More Tropical
Environments: A Proceedings of the International Symposium. Mexico,
D.F.: CIMMYT. 316-320.
Krueger, A., M. Schiff, and A. Valdes. 1988. Agricultural incentives in developing countries: Measuring the effect of sectoral and economywide
measures. The World Bank Economic Review 2(3): 255-271.
Longmire, J., and D, Winkelmann. 1985. Research resource allocation and
comparative advantage. Paper presented at the 19th International
Conference of Agricultural Economists, 26 August-4 September, Malaga,
Spain.
Longmire, J., and J. Lugogo. 1989. The Economics of Small-scale Wheat Production Technologies for Kenya. CIMMYT Economics Program Working
Paper 89/01. Mexico, D.F.: CIMMYT.
McDiarmid, O.J. 1977. Unskilled Labor for Development: Its Economic Cost.
Baltimore: Johns Hopkins/World Bank.
Monke, E., and S. Pearson. 1989. The Policy Analysis Matrix. Baltimore: Johns
Hopkins University Press.
41
Morris,
Morris,M.L.
M.L.1988a.
1988a.Comparative
ComparativeAdvantage
Advantageand
andPolicy
PolicyIncentives
IncentivesFor
ForWheat
Wheat
Production
ProductionininZimbabwe.
Zimbabwe.CIMMYTEconomics
CIMMYTEconomicsProgram
ProgramWorking
WorkingPaper
Paper
02/88.
02/88.Mexico,D.F.:
Mexico,D.F.:CIMMYT.
CIMMYT.
Wheatpolicy
policyoptions
optionsininZimbabwe:
Zimbabwe:AAcomparative
comparativeadvanadvanMorris,
Morris,M.L.
M.L.1988b.
1988b.Wheat
tage
tageapproach.
approach.InInM.M.Rukuni
Rukuniand
andR.R.Bernsten
Bernsten(eds.)
(eds.)Southern
SouthernAfrica:
Africa:
Food
FoodSecurity
SecurityPolicy
PolicyOptions.
Options.Proceedings
Proceedingsofofthe
the3rd
3rdAnnual
AnnualConference
Conference
SouthernMrica.
Mrica.Harare:
Harare:University
Universityofof
ononFood
FoodSecurity
SecurityResearch
ResearchininSouthern
ZimbabwelMichigan
ZimbabwelMichiganState
StateUniversity.
University.
Ricardo,
Ricardo,D.D.1817.
1817.The
ThePrinciples
PrinciplesofPolitical
ofPoliticalEconomy
Economyand
andTaxation.
Taxation.Reprint
Reprint
edition,
edition,1965.
1965.London:
London:J.M.
J.M.Dent
Dentand
andSon.
Son.
Ward,
Ward,W.
W.1976.
1976.Adjusting
Adjustingfor
forOver-Valued
Over-ValuedLocal
LocalCurrency:
Currency:Shadow
ShadowExchange
Exchange
andConversion
ConversionFactors.
Factors.CN-28.
CN-28.Washington:
Washington:Economic
EconomicDevelopDevelopRates
Ratesand
ment
mentInstitute,
Institute,World
WorldBank.
Bank.
4242