Public Disclosure Authorized
Public Disclosure Authorized
Public Disclosure Authorized
I
_ l
_
and ExternalAffairs
Research,
LPOICYy
|
4_
)
p___'
/
C'
'
WORKING PAPERS
TradeandFinance
TechnicalDepartment
Africa RegionalOffice.
TheWorld Bank
March1991
WPS640
The Impact of Policy
in African Agriculture
An EmpiricalInvestigation
0~~
-
Public Disclosure Authorized
WilliamJaeger
-Policyin Sub-SaharanAfrican countries,is linked with the
region's,agric uqalperformance.Exchangerate policie",high
taxeson agriculture,andgovernmentcontrolofexportmarketing
are associatedwith the deteriorationin agriculturalexport perof the late 1980s
formancein 1970-87. And the policyr,eformns
linked with increased
sustained and effective
agriculturalproductivity.
-where
-are
work in prooross anl
The Policy, Research, and External Affairs Comiplexdistribates MREWorking Papers to d. Iremnate the findingsof
cenissues.nThes papers catry the names of
in develo
Lo encourage the cxchangc of ideas anlong Bank staff and all others intefcsted
and conclusions are the
the authors. reflect only theif views, and should be used and citted accordingy. TDe findings, interpretations,
any of its metber countries.
or
management,
its
Directors,
of
Iloard
its
Blank.
World
t~,thc
attributed
be
not
shiould
They
own.
authorse
~~
640
WPS
This paperisa productof the Tradeand FinanceDivision,TechnicalDepartment,AfricaRegi9nalOffice.
20433. Pleasecontanq
Copiesare availablefreefromtheWorldBank,1818H StreetNW,W'a'shingtorn,`DC
Azeb Yideru, room J3-(P80,extension 34663 (69 pages, with figures and tables).
Jaegerexaminesthe relationshipbetween
governmentpolicy and agriculttal performance
in Sub-SaharanAfricabetween1970and 1987.
Usingnewly compileddata enablinga wider
rangeof empiricalanalyses,the study assesses,
per capitaproductionreflectprimarilya productionproblem. Econometricresultsindicatethat
most of the rise in Africa's food importsIs
associatedwith shiftingdemandtowardimported
foods,ratherthan a failure of supply, The main
the impact of policy distortions on productivity
factors causing the shift in demand are increas-
over tune and acrosscountries. It assesses
exportagricultureand foodproductionsFparately.
eThe analysisconfirns thatthe deterioration
oAfrica's agricultual exportsduring the 1970s
sndearly 1980swas associatedwith
agriculture'shigh levels of direct taxationand of
indirecttaxationthrtugh govermnentcontrols
and overvaluedcurrencies. Goverment controls
in the marketingand pricingof exportcrops have
contributedto thc detetiorationin exportperfirmance. But the large indirectdistortionsand
disincentivescausedby exchangeratepolicies
are what havedistinguishedAfricanpolicy
environmentsfrom thosein non-Africandeveloping countries. Econometrc reryltsshow that
the responsivenessof agriculturalexports to
change in incenttivesis moderatein the short run
for countriesexportingtree crops but more
elasticin countriesexportingannualcrops.
The authoralso investigatesAfrica's chronic
food crises and questionsthe conventional
wisdomtiha rising foodimportsand declining
ing urbanizatiopi,higherimportcapacity,and
exchangeratedistortionstAatmake imported
food relativelycheap. Whenthe variationof
these factorshas been taken into account,the
remainingunexplainedtrend is only I percenta
year,causedin part by,declininginterational
pricesfor wheatand rice.
r
Jaegerestablishesa link betweenpolicy
reformsand the iniprovementsobservedin
agriculturalperfortance in the late 1980s.
Countrieswith favorablepolicyenvironments
haveperformedbetter in the 1980s,on average,
than those with unfavorablepolicye9vironments. This has been true both in agricultureand
in overall economicgrowth. Andin countries
wherepolicyreformprogams resultedin
sinificant and sustainedimprovementsin
'Incentves (for example,Ghana and Togno),
productivityhas improvedsubstantially.But in
countrieswherereforis havenot led to improved incentivesor where the imProvements,
were short-lived(for example,Tanzaniaand
Zaire),little responsewas observable.
and Extema
ujnderwayi thoexBarnpolic, Resarch
omeaeinaetheshrtrungofwr
PRE WolnSaper Seies dm
Thanes
oeuctuicl,veifprhasenaionsrelesusthan tially. Butlishd
bul togrteefdig
Anobedvo these ies
Afoarscounplesexpo
policy.
not
necsledersetorcilBn
paperedoomshv
in thesrie
andcncuaicons.
Telasticingscounterreaieexorlns,
by the PREJisseminationCoiver
Produced
TABLE OF CONTENTS
Executive Sunmary
I.
Introduction 1
Objectiveof the study 1
Background I
Conceptualframework 2
I.
The Impact of Policyon Export Agriculture 3
Trends in prices, policydistortions,and export 3
Export supplyresponse 12
m. The Impact of Policy on the Food Crop Sector 20
Cross-priceeffects betweenexportprices and food 21
Food policy 22
Real food price trends 23
Food importsand slow food productiongrowth 25
Policy and rural-urbanmigration 31
IV. Diversity within Africa: PolicyEnvironments and Performance 33
Comparingperformancein favorableversus unfavorablepolicyenvironments 33
Country diversity:policyand response 39
V. Concluding Comments 43
Summaryof empiricalfindings 43
Forward lookingissues 45
AnnexesA. Data and methodology46
B. The extent of chronichungerin Africa 48
C. Classificationof countriesby policy environment50
D. Data tables 53
Bibliography 66
Text Tables
1 RegressionEquationfor AgriculturalSupplyin Sub-SaharanAfrica 17
2 RegressionEquationsfor AgriculturalSupplyin Sub-SaharanAfrica: Crop models 18
3 RegressionEquationsfor AgriculturalSupplyin Sub-SaharanAfrica: Country models19
4 Regressionequationsfor cross-priceeffecton foodproduction 22
5 Comparisonof real consumerfood prices in Africa 25
6 Regressionequationfor food imports 30
7 Rural-urbanmigrationin Ghanasincepolicy reform 32
8 Comparisonof performanceand policy amongAfricancountrygroups 34
AnnexTables
Cl Classificationof countriesby policyenvironment52
Dl Averagenominalprotectioncoefficientsfor majorexport commodities53
D2 Nominalprotectioncoefficientfor principalexportcommodities 54
D3 Real protectioncoefficientsfor principalexport commodity55
D4 Averagereal producerprice for major exportcommodities56
D5 Averagenominalprotectioncoefficientsfor ti.ded food crops 57
D6 Real producerpricesfor majorfood crops 58
D7 Comparisonof estimatedtransportcosts 59
D8 Ratio of implicitfinancialoutflowfrom agriculaire 60
D9 Weathervariable(deviationfrom trendof cerealyields, in logarithms) 61
D1O "Disasters"data: numberof peopleaffectedby major disasters,in thousands 62
D1I Food crop price and policyeffects 63
012 Comparisonof domesticstaple foodprices to importedrice and wheatprices 64
D13 Variousspecificationsof regressionequationfor agriculturalsupply 65
Figures
1 Policy, incentives,and agriculturalexports in Sub-SaharanAfrica 4
2 Nominaland real protectioncoefficientsfor major agriculturalexports 6
3 NPCs and RPCs in CFA countriesfor major exportcrops 8
4 AdjustedRPCs for Africa's major exportcrops 10
5 Decompositionof changesin RPCs for Africanexport crops 11
6 The impactof weatheron Africanagriculture 14
7 Real producerprices for major crops in Sub-SaharanAfrica 24
8 Ratio of export/foodproducerprices in Sub-SaharanAfrica 24
9 Africa's food imports 28
10 Africa's import capacity 28
11 Africa's agriculturalexports 36
12 Total agriculturalproduction 36
13 Total food production 37
14 Africa's GDP 37
15 Real effectiveexchangerates 38
16 Weatherpatternsin Africa 38
17 Ghana's price incentivesand agriculturalexports 40
18 Zaire's price incentivesand agricutural exports 40
19 Tanzania'sprice incentives 42
20 Togo's price incentivesad agriculturalexports 42
ACKNOWLEDGEMENTS
This paper was largelypreparedwhilethe authorwas an economistat the WorldBank. Data
for this researchwas compiledunder UNDPproject RAP 861058. The authoris gratefialto
CharlesHumphreys,MichaelSarris, and GilbertUwujarenfor commentson earier draft. The
researchassistanceof KatharinaKatterbach,JamesTeffi, and PeterRosendorff,and the secrearial
supportof Azeb Yideru,are gratefuilyacknowledged.
- iii FOREWARD
The Lst two decadeshave witnesseddeclinein Africa's agriculturalexportsand a sharp rise
in the region's food imporis. With over 70 percentof Africa's peopledependenton agriculture
for their livelihood,gettingagriculturemovingin Africais one of the most importantand
formidablechallengesfazingAfrica in the 1990s. There is now wide recognitionthat
governmentpolicieshave contributedto Africa's crisis and that policyreforms can help foster
restoredgrowth.
"The Impactof Policyon AfricanAgriculture:An EmpiricalInvestigation'examinesthe
link betweengovernmentpolicyand agriculturalperformanceacross countriesand over most of
the past two decades. The analysisconfirmsthat governments'agriculturalpricing and
marketingarrangementsand their macroeconomicpolicieshave slowedagriculturalgrowth. In
particular,overvaluedcurrencieshave often causedthe largestdistortions. Recenttrends,
however, offer evidencethat policyreformshave helpedimproveagriculturalgrowth where
these reforms have sustainedagriculturalprocedures.
The authoralso offers evidencethat governmentpolicieshave been substantiallyresponsible
for Africa's growingfood imports,by encouragingurban migrationand by loweringthe prices
for importedfoodsrelativeto domestically-grown
foods.
This study contributesto a better understandingof the causesof Africa's agriculturalcrisis
and can help forge a consensuson how to restorevitalityin Africanagriculture. This work was
carried out as part of the World Bank's ongoingmonitoringof Africa's economicgrowth and
policy reforms.
Ismail Serageldin
Director
TechnicalDepartment
Africa Region
EXECUTIVE SUMMARY
Africa's loss of world market sharesfor its major agriculturalexportsand the declinein food
productionper capitaare the most tellingsignsof stagnationand declinein Africasince 1970.
Africa's internationalmarket share for its 12 majoragriculturalcommoditiestell by half between
1970and 1983. Over the sameperiod, per capitafood productionfell about 1 percentper year.
Since the mid-1980s,however,there have been improvementsin both food and export agriculture,a
reversal whichhas coincidedwith the implementationof policyreformprogramsby a subsiantial
numberof Africancountries.
This paper examinesthe impactof policyon agriculturalperformancein Sub-SaharanAfricain
the 1970sand 1980s. Based on recentlyavailabledata enablinga wider rangeof detailedempirical
analysis,the study assessesthe distortingeffects of governmentpolicieson incentives,the impactof
those distortionson agriculturalperformance,and the changesover time and across countriesin both
policies and agriculturalperformance Exportagricultureand food producdonare assessed
separately.
Africa's agriculturalexportperformancedeterioratedsubstandallyfrom 1970-84. The pattern of
decline- and partialrecovery in the late-19SOs- followscloselythe high levels of distortionin the
agriculturalsector. Thesedistortionswere the result of both direct governmentcontrolson producer
prices for Africa's principalexportcrops and the indirecteffectsof exchangerate policies. As a
result of these policies,real producerprices feil by one-fourthbetweenthe early 1970sand early
1980s. The share of these crops' value retainedbv farmers (as measuredby the real protection
coefficient)droppedfrom about90 percent to about 50 percent, on average, from 1971to 1983.
Governmentcontrolsand involvementin the marketingand pricingof export crops has been
responsiblefor a considerableshare of the deteriorationof producerincentives,althoughthe
magnitudeof these distortionshas not been larger, on average, than in other non-Africandeveloping
countries. However,large indirectdistortionsand disincentiveshave been causedin Africaby the
prevalenceof highlyovervaluedcurrencies,especiallyfrom the mid-1970¢to mid-1980s. The large
non-African
size and prevalenceof theseindirectdistortionscontrastwith conditionsiL
developingcountries. Pegged exchangerate regimes,coupledwith poor fl6. managementand high
domesticinflation,resultedin highlyovervaluedcurrencies. Those countrieswhichmaintained
grosdlyovervaluedcurrenciesover extendedperiods experiencedsevere deteriorationof export
agriculture. CFA-zonecountrieshave been more successfulin avoidinglarge exchangerate
distortionsover the period, and thus have been less affectedby the deleteriouseffects on agricultural
growth. In the late 1980s,however,overvaluadonof the CFA franc has becomemore of a problem.
There has been considerabledebate in the literatureon the responsivenessof agricultureto price
incentivesin Africa. The issueis addressedhere in an econometricframeworkcombiningcrosssectionand time-seriesdata. In the short-run,tree crops exportsate shown to be ordy moderately
responsiveto changesin price incentives. Elasticitiesare on the order of 0.1. However,for
countriesexportingannualcrops the short-runresponseis moreelastic, above0.9. Exportsare also
shown to be responsiveto changesin the real exchangerate, reflectingthe effectson price incentives
and on efficiencycausedby changesin exchangeand trade restrictionsthat generallyaccompany
overvaluedcurrencies. Elasticitiesof exportsupplywith respectto the real exchangerate varied
from 0.1 to 0.3 for tree crop exporters,and 0.6 for annualcrop exporters. Theseestimatesmay
not, however,reflect long-runresponses,which are influencedby expectedprices.
-
vi -
The possibilitythat export agriculturemay crowdout food productionis examined. Growthin
exportagriculturedoes not appear to come at the expenseof food production. Food producdonis
positivelycorrelatedwith exportcrop pricesand real exchangerates suggestingcomplementarity.
However,lack of data meansthat the impactof policy on food productionin Africacannotbe
analyzedin the sameway as exportagriculture. Food producdondata are oftenquestionable.
Marketprice data for food crops are unavailablefor most Africancountries. BecausemanyAfrican
governmentsfix officialprices for food, only those prices are generallyreportedand oftendo not
correspondto actualmarket prices. However,the impactof policyon Africa's food problemscan
be examinedindirectly.
Averagefood supplyper capitahas not declinedin Africasince the 1960s. Food supplyper
capita - accordingto FAO data - has risen slightly,on average, s' estingthat only the
compositionof consumptionhas shiftedfrom domesically-grownfoods to importedfoods(whichare
often preferred,especiallyamongurban and wealthierconsumers). However,Africa's chronicfood
crisis is commnonly
assumedto be both widespreadand worsening,as evidencedby decliningfood
productionper capitaand rising food imports. It is generallyassumedthat weak domestic
productioncreates a need for more imports. The causaldirectionbetweenimportsand domestic
productionmaybe the reverse: that is, a shift in demandtowardimportedfoods mayhave resulted
in increasedconsumptionof imports,and thereforereduceddemandfor domesticfoods. This weak
demandthen led to weak domesticoutput. Demandfor importedfoodsmay rise due to urbanization,
import capacity,and the relativeprice of importedto domesticfoods(whichwill be affectedby real
exchangerates). Indeed, the doublingof Africa's food importsin the late-1970swas largely the
resultof eventsin Nigeria, whichaccournedfur 70 percent of the increase. This occurredduring
Nigeria'soil boom whenthere were large shiftsin importcapacity(a quadrupling),the real
exchangerate, and urban migration.
To t_st the hypothesisthat both rising food importsand decliningdomesticproductionis demand
driven rather than the resultof a productionconstraint,an econometricmodel is estimated. The
results confirmthat muchof the rise in Africa's food importscan be attributedto urbanization,
import capacity,and exchangerate distortions. Whenthese factorshave been taken into account, the
remainingtrend is only 1 percentper year, and this can probablybe explainedby the declining
internationalprices for wheatand rice over the period.
Rural-urbanmigration,which appearsto be a causeof rising food imports,is a resultof
governmentpolicies, imciudingexchangerate policiesthat have alteredthe internalterms of trade in
favor of urban areas. Evidenceof the extent to whichurban migrationis influencedby policy - and
the potentialfor policyreform to alter existingmigratorypatterns- is demonstratedwith recent data
from Ghana. Surveydata confirmwhathas been observedanecdotallyin severalAfricancountries:
significantreversemigrationis occurringfrom non-agriculturalto agriculturaloccupationssince the
introductionof Ghana's reformprogram.
The improvementsin aggregatepatternsof productionand trade in Africasince the mid-1980s
coincidewith the implementationof structuraladjustmentprogramsin a large numberof African
countries. To assess the evidenceof a causalrelationship,economicperfob.ce in countrieswhere
policydistortionshave been relativelysmall, is comparedto that in countrieswhere policy
distortionshave resultedin an unfavorablepolicyenvironment. In aggregate,countrieswith
favorablepolicyenvironmentshave performedbetter in the late-1980sthanthe countrieswith
unfavorablepolicyenvironmentsin virtuallyall measuresof economicperformance,including
agriculturalproduction,agriculturalexports, and overalleconomicgrowth. The experiencein
-
vii -
individualcountrieshas been quite disparate,but in countrieswhere policyreforms have resultedin
significantand sustainedimprovementsin incentives(e.g., Ghana, Togo) productivityhas improved
substantially. In countrieswhere thesereformsdid not improveprice incentives,or where the
improvementswere shortlived(e.g., Tanzania,Zaire), no significantresponsewas observed.
Overall,the evidencefrom this study suggeststhat poor policieshave had a major role in the
decr-neof Africanagriculture. Similarly,there is evidencethat policyreforms in the 1980shave
cor.vibutedto some modestimprovement.
There has been considerabledebateaboutthe scope for policyresponsein Africa, at times
framed in termsof whetherprice or nonpricefactorsconstraingrowthmost. The debate maybe
misleading,however. First, the distinedonbetweenprice and nonpricefactorsis sometimesblurred;
nonpriceconstraintson agriculturemay be seen as beingprice related (e.g., the lack of roads implies
high transportcosts; the absenceof extensionservicesraises informationcosts of new technology).
Second,the relativeimportanceof price and nonpricefactorswill vary from countryto country;
thereforethe debate of their relativeimportancemust take accountof specificcountrycontexts. And
third, the classificadonof constraintsinto price-and nonprice-relatedis to a large extent analogous
to the dichotomybetweenshort-runand long-runresponseto prices. For example,bad roads, lack
of irrigaion, poor institutionalinfrastructure,and other "nonprice"constraintsmay be the result of
neglectedinvestmentscausedby policy distortionswhichhave made sociallyprofitableinvestments
privatelyunprofitable. Investmentsthat wouldremovenonpriceconstraintsmaybecomeattractive
when policydistortionsalteringrelativeprio s are lessened.
I. INTRODUCTION
Objectiveof the study
The responsivenessof Sub-SaharanAfrican1/ agricultureto changesin farm prices and macropolicieshas emergedas a critical issue for understandingAfrica's poor agriculturalperformanceas wellas
for prescribingcorrectiveactions. Viewson the importanceof pricingpoliciesvary widely,and a debate
has emergedrecentlyabout the relativeimportanceof price and "nonprice' factorsboth in explaining
Africa's poor performanceand as the main impedimentto recovery. Surprisinglylittle empiricalworkhas
been done, however, on aggregateagriculturalsupplyresponsein Africa. This has been due at leastin part
to data limitationsthat have preventedextensivecross-sectionaland time-seriesanalys,.
Recently,however, suitabledata have been compiledto enablea wider range of more detailed
empiricalanalyses. Based on thesedata, the objectivesof this paper are: a) to examinethe trends and
magnitudeof policydistortionsin Africaand their effectson agriculturalincentives;b) to estimatethe
responsivenessof Africanagricultureto thesepolicychanges;c) to assessthe relativeimportanceof these
policiesin explainingAfrican's poor agriculturalperformanceboth fcr food and export crops since 1970;and
d) to evaluatethe evidenceof a link betweenAfrica's policyreformprogramsand recent evidenceof
recoveryin Africa's agriculturalgrowth.
Withinthe agriculturalsector, exportedcrops and food crops have itatures that make it difficultto
addressthemjointly. Analysisof foodpolicyand growthgives rise to a variety of distinctivecharacteristics
and problemsdue to the nature of governmentpolicies,the importanceof domesticdemand,and especially
data limitationson producerfood prices. Giventhese complexities,exportand food crop agricultureare
examinedseparatelybelow.
Agricultureis Africa's most importantsector, accountingfor about80 percent of employmentand 50
to 90 percent of exports. The sectorhas performedpoorly since 1970: foodproductionhas failed to keep
pace with populationgrowthand exportagricuture has experenced decliningmarketshares. This slow
growth and loss of competitivenessin internationalmarketscoincidedwith increasinglydistorted
macroeconomicpoliciesand intensifiedgovernmentcontrolsand restrictions(WorldBank 1989a).
Furthermore,in the late 1970s,exceptionallyhigh commodityprices for Africa's major exports(oil, tropical
beverages,phosphates)led to unrealisticexpectations,overextendedborrowing,and an unmanageabledebt
burden. The substantialimbalanceson externaland domesticpublicaccountscreatedduring this period
evenually compelledmany Africangovernmentsto adoptmacroeconomicpolicyreformprograms.
Decliningrevenuesfrom agriculturalexportsand rising food importsled to the inclusionof agriculural
sectorreforms in many of thesereform programs.
Mostof thesereformprogramswere initiatedduring the first half of the 1980s. Andbeginningin
1985,agriculturalperformanceand agriculturalexportmarket sharesbeganto show signsof recovery, with
agriculturalproductiongrowingat 4 percentper year from 1985-88- faster thanpopulationgrowthfor the
first extendedperiod since 1970(WorldBarl; 1989a). Thesepatternshave heightenedinterest- as well as
the debate - on the merits and limits of poliayreformas a basis for recoveryin Africa.
1 Throughoutthe paper "Africa"is used to donateSub-SaharanAfrica.
-2-
ConceptualFramework
It is generallyagreedthat agriculturalgrowth is lnked to farm profits - where farm profits are
affectedby a rangeof factorsincludinggovernmentpolicy. Output prices,input costs,and
exchangerates are central to most discussionsof the responsivenessof agricultureto policy,but
since farm profits are affectedby wages,interest rates, market imperfections,information,etc.,
these factorsneed to be taken into account as well (Binswanger).
GovermnentpolcHycan affect farm profitabilitythrough a) control over output and Input
prices,b) taxationor subsidiesthat affectthose prices,c) controlson wagesand interest rutes, d)
Institutionalarrangements(eg., accessto credit, inputs, information),and e) actions that affect
profitabiity and productivityin other sectors.
Direct governmentpoicies, includingprice fixingfor productsor inputs, or the taxationof
their trade, affect the profitabilityof farmingdirectly,and result i the shiftingof resources
betweencrops, or in movingresourcesout of agricultureinto other sectors.
MacroeconomicpoUciesaffect farm profits in severalindirectbut critical ways. Nominal
exchangerates set an upper bound on the price paid to farmersfor exportedcommodities(less
transport and processingcosts, and net of subsidies). In the same way,exchangerates (together
with import taxes and other restrictions)set pricesof inputs and agriculturalimportswhich
competewith domesticproduction.
Overvaluationof the exchangerate can result in severewelfareand efficiencycosts, both
directlyfrom the misallocationof productiveresources,and indirectlyas a result of the exchange
and trade controls that usuallyaccompanyovervaluation.These indirect effectsfrom trade and
exchangecontrolscan be the biggestcosts associatedwith overvaluation(Edwards1989a). in
addition,such controlsgenerateunproductiveresourceuse from rent seekingactivities.
An appreciationof the real exchangerate (or raisingthe relativeprices of nontradablesto
tradables)raises the cost of producingtradables in terms of nontradables,and therefore reduces
directlythe profitabilityof producingthese goods. Moreover,the shiftingof the internal terms of
trade against agriculturewill encouragemigrationfrom rral to urban areas due to the loss of
competitivenessof agricultureand growth in demandfor nontradables,as well as from protection
of urban-basedmanufacturing.Migrationmay occur as a result of other policiestoo, such as the
provisionof cheap food and other servicesfor urban populations.
In addition to pricingpoliciesand macroeconomicpolicies,governmentexpenditureand
i'vestment In the agriculturalsector can have importanteffectson fAm profits and are critical to
long-termcompetitivenessand agriculturalgrowth. By and large these pollcymeasuresare aimed
at red-..-ng costs of productionin order to raise profits and stimulategrowth:constructionof
transjrt infrastructurewill lower transport costs thus reducinginput prices and raisingoutput
prices at the famgate; extensionservicescan be seen as reducingthe costs of information(that are
otherwiseunavailableat almostany reasonableprice); rural credt institutionsmake credit available
at a lower cost to farmers;and researchstrivesto raise proflts by way of technologicalchange.
Factors affectingfarm proflts of this type are sometimesreferred to as "non-pricefators!
The macoeconomicpolicyenvironmentthat has led to real exchangerate appreciationis
generallycharcterized by intemnalas well as externalimbalances. High inflation,negativereal
interest rates, and restrictedaccessto credit add to the harmful effectson investmentsin
*3agriculturalas wellas non-agriculturalsectors. ro the extent that agricultureis not a protected
industry,the adverse,effec-t on investmentwillbe even larger.
Governmentinterventionin marketingof farm products and inputs affectsfarm profits when
these institutionalarrangementsresuit in inefficiencies,delays,fragmentedmarkets,or inflatedcosts
that depress producerprices. And governmentsintervenein other areas that affect farm profits as
well,such as in land tenure policiesor the allocationof newlyproduct'velands (eg. irrigated).
In addition to policymeasuresaimed at promotingagricvuturalgrowth,governmentsintervene
in the interestsof consumers(price levelsand stabilization),and to raise revenues(taxationof
agriculturalexports is an importantsource of revenuein most Africancountries). These two typW
of policieshave distributionalconsequencesbetweenrural and urban peoples,and betweenexport
crop and food crop producers.
Most Africangovernmentsexercisesome control in the marketingand pricingof exportcrops,
fixingprices in most cases,althoughso. ' have recentlyliberalizedthese markets. Major food
commodities,both domestcallyproducedand imported,have come under governmentcontrol for
reasons of price stabilization,keepingprices low for consumers,and to prohibit excessprofits and
control of marketsby private traders. Cheap food policies,although commonin Africa,have been
difficultto enforce in most countries,givingrise to large parallelmarkets,and makingofficially
announcedpricesoften of little relevanceto producers.
EXPORTS
H. TEE IMPACTOF POIYCYON AGRICULTURAL
Trends in Prices.PolicyDistortione,and Exports
Africa'sloss of world market shares for its major agriculturalexports is the most telling
evidenceof its loss of competitiveness.Africa'sshare of those markets for its 12 major agricultural
exports fell by half between 1970and 1983. And the volumeof agriculturalexportactually
declinedas well,falling from 14 millionmetrictons to about 11 millionover the same period. _y
As figure 1 confirms,the averagepattern across26 countrieswhere completedata are available
suggestthat the loss of competitivenessin internationalmarketshas followedcloselythe pattern WL
overvaluedexchangerates (real effectiveexchangerate indexis invertedfor better visual
presentation),and lower real producerprices. For these 26 countries,only after exchangerate
distortionsbegan to declineand real pricespaid to farmersbegan to rise in 1984,did export
volumesimprove.
The declineand recent recoveryof agriculturalexportsappears to followcloselythe pattern of
real effectiveexchangerates over the period, exceptfor 1987. Real producerprices havegenerally
followeda similar pattern, althoughdecliningmore slowlyin the mid-1970sdue the boom in
internationalcoffeeand cocoaprices. The declinein exportvolumesin 1987is primarilythe result
y FAO weighted 'verageexportvolumesfor total agriculturalcommodities.Much of the analysis
is focusedon exl. . volumesrather than agriculturalproduction. Agriculturalproductiondata is
of questionablequalitygiventhe rough estimationjrocedures relied on for most ccuntries. Since
for most export crops domesticconsumptionis small relativeto the total exported(cocoa,coffee,
tea, rubber, tobacco,cotton), exportvolumescan be takes as a proxyfor total productionfor most
of these crops.
4-
Figure 1.Policy, incentives, and agricultural exports in Sub-Saharan Africa
Index 1970
=
100
120 110 -
90
80
1970
-
1972
REER
1974
1976
1978
1980
1982
arithmetic mean for 26 countries
agricultural
exports
1984
1986
real producer price
-5of lower outputamonigmajor coffeegrowers. Exportsrose in 1988accordingto preliminarydata available
at the time of this writing.
Direct pricing policydistortions. A varietyof governmentacdons can affect producerprices directly(fixing
prices, export taxes, marketingarrangements,etc). The net effect of which is oftencomplex. To assess the
magnitudeof the collectivedistortingeffectsof thesepolicies,some kind of measureis needed to capturethe
net effectof these complexinterventions. The most commonlyused measureof the distortingeffectsof
policy is the nominalprotectioncoefficient(NPC) - or ratio of producerprice to border price adjustedfor
marketingcosts - which compareswhata farmer receivesto the maximumhe couldreceiveshort of
subsidies(an NPC less thanone indicatesa tax on producers,an NPC greater than one wouldreflecta
subsidy). NPCs have been widelyused to assess the effect of policy on agriculturalincentives,as a
relativelysimplemeansof assessingthe divergenceof producerprices from what they wouldbe in the
absenceof governmentpolicies.I/
Figure 2 indicatesthat NPCs for Africa's major exportshavefallen and risen twice since 1970,and
are now at a level of about 1.0, on average,but having fallenbelow 0.5 in 1976. The range of NPCs varies
widelyamongcountriesand over time (see AnnexD tables 1 and 2), from as low as 0.16 for Ghanaian
cocoa in 1976to as high as 2.6 for Senegalesegroundnutsin 1987. In most countriesthe NPC varies
erraticallyfrom year to year becauseproducerprices are usualy set at the beginningof the growingseason,
long before the internationalprice that wiUlbe receivedfor their productionin the folowing year is
known.4/ Only in the few countrieswhere the producerprice is based on what is eventuallyreceivedfor
their product (usinga rebate system)as in Kenyanand Ethiopiancoffee,and Malawiantobacco,are the
NPCs more stableacrossyears.
The degree of taxationimplicitin theseNPCs is in manycases substantial. However,it does not
appear to be significantlyhigherduring this period than in otherdevelopingregions whencomparedto those
estimates(Binswangerand Scandizzo;Krueger, Schiff, Valdes).
I e NPC as a measureof policydistortionsuffers,however, from at
Thons.
Macconomic policy distor
least two importantdrawbacks. First, the NPC does not take accountof exchangerate misalignmentor
the implicittaxationthat it can represent,and thus will understatethe degreeof agricultural
taxationwhenexchangerates are overvalued. Andsecond, changesin the NPC over time arise
from three sources,but the reladve importanceof any one factor relativeto the net
I/ A broad range of agriculturaland macroeconomicdata for 1970to 1988were used to examinepolicies,
incentives,and performancefor up to 40 Africancountries(see AnnexA for details). The time period being
consideredand the numberof countriesinvolvedlends itself to an analysisof cross-sectionaland time-series
data. NPCs for each country's majorexportand food crops havebeen computedas the ratio of producer
price to border price net of all processingand marketingcosts. Both numeratorand denominatorare
adjustedto reflect the comparisonof producerto border pricesat the "lastjoint marketingpoint' (Westlake);
the border in the case of exports,and the major consumptioncenter for importsubstitutingfood crops. See
Scandizzoand Bruce (1980)for computationaldetails.
41/The wide swings in NPCsfrom year to year explains,in part, conflictingfindingsin the literatureas to
the degree - and direction- of policydistortingeffects in a particularcountry. Unlesssuchstudiesare
comparingproducerprices and internationalprices for the sameyear, significantdifferencesare likely to
occur in the conclusions. Given thesewide variationsover time, multi-yearanalysesare advisable.
-6-
Figure 2. Nominal and real protection
coefficients for major agric. exports
1.2
0.8
0.6
0.41
1970
1972
1974
1976
-RPC
1978
1980
-NPC
1982
1984
1986
-7effects is ambiguous. Taking accountof theseeffectsis important,especiaUybecauseexchangerates
across Africabecamehighlyovervaluedduring the 1970s,and have becomeless overvaluedin the
1980s.
Both limitationsof the NPC are addressedhere, first by computing"real protectioncoefficients"
(RPC) to take into accountchangesin the degree of currencyovervaluationusing the real exchange
rate, and secondby "decomposing"the annualchangesin theseRPC's in order to reveal the source
of the changesbecauseof changesin producerprice, internationalprice, or exchangerate.
The RPC is computedby adjustingthe NPC for changesin the real exchangerate, using 1970as
a base year. 5/ The RPC will divergefrom the NPC to the extent that the implicittaxationfrom
exchangerate policieshas changed.{/ Given the extent of exchangerate misalignmentthat has
occurredamongAfricancountries,the RPC offers a better meansfor assessinghow changesin
policyhave affectedagriculture(See AnnexTable 3). Figure 2 depict both the NPC and RPC from
1970to 1987. The two are equalby definitionin 1970,but divergesignificantlybetween 1978and
1984indicatingan increasein the degreeof exchangerate misalignment.The differenceis quite
large, accountingon averagefor 20 percent of the producer's fair valueof his production,and
maldngclear that the rise in the NPC in 1980-81was really a "false improvement"wherebya higher
share of the border price (convertedat the officialexchangerate) was being receivedby producers,
but at the sametime the indirecteffectsof increasily overvalueddomesticcurrenciesreduced
agriutural competitivenessin a less direct but potentiallymore distortingway.
The differencesbetweenthe NPC and the RPC is quitelarge in those countrieswhere exchange
rate misalignmenthad becomesevere: an NPC of 2.8 comparesto an RPC of 0.51 for Ghanaian
cocoa m 1981;in Ugandathe NMCfor coffeeis 1.51, whilethe RPC is 0.17 in 1980. For CFA
coumtrieswhich historicallyhave keptexchangerates in better alignment,the differencesare small
(less than 0.1), with some divergenceduring 197741 and 198546 (figure3). The divergencein
1985-86appearsto have diminishedin 1987,the resultof a declinein the average REERin 1988for
CFA countriesdue to the strengtheningof the frenchfrancvis-a-visto U.S. dollar.
5/ Both the NPC and RPC are computedfor splitcrop years (i.e. 1970/71),so that the producer
prices for 1970crop year are comphredto internatonl pricesfor the correspondingmarketing
year, whichis 1971. For that reasonl1971rather than 1970is used as a base for the real
effectiveexchangerate in computingthe RPC for 1970.
{/ The RPC will take accountof changesin the degreeof exchangerate misalignment,but not
in the absolutelevel. Only if there had been no misalignmentin the base year, 1971,would the
RPC accuratelyshowthe level of both direct and implicittamtion of agriculture. The average
level of exchangerate distron appearsto have been relaively smallin 1971based on the ratio
of officialto parallelmarket exchangerates, which averaged1.2 for countrieswith available
data. Choosinga base year prior to the first oil price shockand subsquent volatilityhas
obviousadvantages. Still, exchangerates were serously overvaluedat that tme in several
countries,reflectedin high rados of parallelto officialexchangerates: Ethiopia(1.17), Ghana
(1.48), Kenya (1.35), Malawi(1.37), Sudan(1.77), Tanzania(1.62), Uganda(1.48), Zaire
(1.45), and Zambia(1.49). In Nigeria the parallelmarketrate in 1971was below the official
rate.
Figure 3. NPCs and RPCs in CFA countries
for major export
crops
1.3.
_
0.9
0.7
0.5-
0.3
1970
1972
1974
1976
1978
NPC
1980
-RPC
1982
1984
1986
The actual levels of direct and implicit taxation (not just relative to a base year) of agricultural
producers is difficult to assess short of elaborate and computationally difficult methods such as
those recently developed by Krueger, Schiff, and Valdez, but as an approximation the RPC can be
adjusted proportionally to the base year ratio of parallel market exchange rate to official exchange
rate, and termed "adjusted RPC". Using this rough estimation procedure, the average adjusted
RPC is compared for CFA-zone countries (where no significant parallel market distortion existed in
1971) and non-CFA countries (where the average ratio of parallel market to official exchange rate
was 1.46 in 1971). The result (figure 4) suggests that: a) the degree of direct and implicit taxation
of agriculture has been substantially higher in non-CFA countries throughout most of the past two
decades, averaging 0.5 from 1970-1985,b) the adjusted RPC has fluctuated more in CFA countries,
with averages ranging from 0.4 to 1.25, and c) for both sets of countries the level of direct and
implicit taxation have moved in the direction of increased incentives to producers since 1984.2/
The underlying cav-sesof the large fluctuations in the RPC can be revealed by "decomposing"
the annual change into its component effects. §/ The three principal effects causing changes in the
RPC are the nominal producer price, the real exchange rate, and the international price. 2/ In
figure 5 each set of three bars will sum to the annual change in the overall RPC. Each bar
represents the influence of each of the three major factors on the overall change in the RPC from
the previous year; the sum of the three bars will reflect the change in the RPC from the previous
year. For example, in 1972, the large negative effect of international price changes (rising
international commodity prices) outweighs the smaller positive change due to nominal producer
price changes and exchange rate changes, so that the net effect is a decline in the NPC between
1971 and 1972. By contrast, in 1986 the rise in nominal producer prices outweighed the small
downward influence of changes in the exchange rate, resulting in a net effect raising the average
RPC in 1986 over 1985.
Figure 5 shows that the decline in RPCs in the early 1970s was due to rising international
prices unaccompanied by higher farm prices (which causes the RPC to fall), and smaller real
exchange rate depreciations in 1973 and 1975. Producer prices were raised as indicated in the
figure, by substantial amounts in 1973, 1975, and 1976,although only a portion of the international
price rise was passed on to producers.
Beginning in 1980, large nominal devaluations in exchange rates drove the RPC down. The
average RPC rose ag,ain in 1981 due to a drop in international prices (which were not fully passed
on to producers), but in 1982 and 1983, the large devaluations lowered the average RPC
substantially. Not until 1984, with the increased scope for raising nominal producer prices
following devaluation, were the RPCs raised as a result of large increases in producer prices in
1984-87. In 1987 increased producer prices were offset by exchange rate devaluation (occurring in
1988), resulting in a small decline in the RPC.
2/ These results are reiatively close to those estimated in Krueger, Valdez, Schiff for Cote d'Ivoire
and Zambia. But for Ghana, they estimate a lower level of total taxation of cocoa during the
1975-79 period than estimated here.
/ See Jaeger and Humphreys for the derivation.
2/ If only producer prices rise, the RPC will rise; if only international prices rise, the RPC will
fall; if only the real exchange rate rises, the RPC will fall.
-10
Figure 4. Adjusted RPOs for Africa's
major export crops
1.3
1.1
0.9
/
0.7-
0.3
1970
1972
1974
-
1976
CFA countries
1978
-
1980
1982
1984
non-CFA countries
1986
-
11 .
Figure 5. Decomposition of changes in
RPCs for African export crops
averoge RPC
average annuol chonge
/1.28
0.2
-0.2 -0.8
-0.4 -0.6
-0.61971
0.4
1973
1975
_
prod'er price effect
FII
inter'l price effect
1977
1979
1981
1983
exchangerate effect
-
RPC(right scale)
1985
1987
-
12-
As of 1986,the averageRPC was greater than 1.0. This reflectsseveralfactors. First, several
CFA countriesrecentlybegan subsidizingproducers(rather than taking the politicallyunpopular
decisionto lower producerprices).IQ/ In addition, severalother countriesnow have REERs
which are "undervalued"relativeto 1971, makingtheir RPCs greater than 1.0. Taking account,
however, of the exchangerate distortionsat that time - accomplishedto a large extent by the
"adjusted"RPC -- leaves the net effect as a taxationon agriculture(figure4).
Real producerprice trends. The net effect of both direct and indirectpolicy changes,as well as
changesin internationalprices, are felt most directlyby producerin terms the real prices received.
Real producerprices for Africa's major exportcommoditieshave witnessedwide fluctuationssince
1970(figure 1). Throughoutmuchof the 1970stheseprices rose, on average, reflectingthe rise in
internationalcommodityprices in the mid-to late-1970s,primarilyfor tropicalbeverages. The rise
in producerprices for these commodities,however, wasfar less than the increaseon internadonal
marketswhere real prices for primarilycommoditiesmore thandoubled,whereasreal producer
prices rose, on average, less than 15 percent.
Between1977and 1982, the real prices receivedby farmersfor these exportsfell 25 percent
resuldng in part from a return to lower prices followingthe "boom' in tropicalbeverageprices in
the mid 1970s,but also due to high domesticinflationand the reluctanceof governmentsin Africa to
respondby adjustingtheir exchangerates. Onlyafter exchangerates were devalued,and real
producerprices raised, did exportperformancebeginto improve.
Exprt supDDlresnonse
The responseof agriculturalproductionto changesin policiesand incentivesis a centralissue
with importantimplicationsfor governmentsand donors. Viewsrange widelyon the responsiveness
of agriculturalproductionto price and this has given rise, to muchdebate (Chhibber,Binswanger,
Cleaver 1988).
One source of thesedifferingviews is the distinctionbetweenthe supplyresponseof individual
crops - when factorsof productioncan be quicklyshiftedfrom one crop to another - and aggregate
supplyresponsewhich, in the long-runrequireseither additionalresourcesto movefrom other
sectors, higher investmenu,or technologicalchangeto bring abouthigher productivity. The fact
that farmersrespondstronglyto changesin the relativepricesof individualcrops is well documented
in empiricalstudies. This, however, tells us little aboutthe aggregateresponseto changesin
agriculual pricesoverall.
For Africancountries,relativelyfew supplyresponsestudieshave tried to assess the relative
importanceof price, non-pricefactors, and exogenousshocksin explainingagriculturalgrowdt
(Wheeleris a notableexception). And none has exploitedthe potentialstrengthof using pooled
cross-sectionaltime-seriesanalysisto explain thesereladonships.
IQ/ This situationarose as the CFA franc, which is ded to the French franc, has appreciated
makingthe US$ denominatedinternationalcommoditypricesappear muchlower in local
currencyterms.
-
13 -
In addition,while muchattentionhas beenplaced on the direct price incentives(generally
producerprices), little work has includedthe indirecteffectsof macroeconomicpolicy distortionson
agriculturalgrowth, especiallyexports(balassa, 1988,has used the real exchangerate as a proxy for
iLt.entives).Real exchangerates affect the competitivenessof agriculturalexports by alteringthe
relativeprices of inputs and products. But inclusionof the real exchangerate - in additionto real
prices - serves as a proxy for the costs -id efficiencylossesassociatedwith trade and exchange
restrictionsthat generallyoccur wheh exchangerates are highlyovervalued.
The analysispresentedhere utilizestime-seriesdata for the 1970-87periodand for 21 countries
to estimatethe responseof agricultureto changesin real producerprices, macioeconomicpolicies
(specificaUythe real exchangerate), the effect of weather,and major shocksin the form of
"disasters"(includingwar, civil strife, famine, floods,or other disruptions).L1/
The modelis as follows:
AGRICULTURALEXPORTS= b0 + b, REAL PRODUCERPRICE +b2 REAL
EFFECTIVEEXCHANGERATE + b3 WEATHER+
b4 DISASTERS+ e
The analysisfocuseson exportsof agriculturalcommodities(cocoa, coffee,tea, etc.). For most
of these commodities,nearlyall productionis exportedso that exportsare a reasonable
approximationof productionexceptin cases where domesticconsumptionis large relativeto exports
(as in the case of palm oil), or where smugglingresults in a significantdivergenceof recorded
exports from actualproduction.12/ The food crop sectoris, for the moment,left out of the
analysis. To estimatesimilarsupplyequationsfor food crops wouldbe problematicdue to the lack
of adequateproducerprice data. This results in part from the prevalenceof officiallyannounced
prices for food crops which in many countrieshave little relationto what farmersactuallyreceive.
These officialprices are oftenthe only availableprice series. The interactionbetweenfood
productionand policyvariablesis addressedbelow.
Weatherhas an importantimpacton agriculturalproductionin Africa,but quantifyingthat effect
is problematic. Rainfalldata is spottyand difficultto interpretcorrectlyshort of complexsimulation
modelsof soil moisture,daily rainfall, and evapotranspiration
data, and differentrainfallpatterns
will effect individualcrops differently. An approximationfor the generaleffect of weatheron
agricultureis derivedhere by estimatinga regressiontrend line for cerealyields, and takingthe
iL/ Specificationof variablesand data sourcescan be foundin AnnexA.
12/ Severalcountrieshave been excludedfrom the anm
1ysis becauseof obvious"bordereffects"
and smuggling. TheseincludeBeninand Zambiawheredramaticswingsin agriculturalexports
appear to have resultedfrom policychangesand border closingsin neighboringcountries. Also,
Rwandaand Burundiare excludeddue to apparentsmuggling- exports exceededproductionby
substantialamountsin severalyears. Ghanais oftenthoughtto have substantialsmugglingbut it
is unclearthat this accountsfor a substantialshare of total production.
-
'4 -
Figure 6. The impact of weather on
African agriculture
(deviations from trend in cereal yield)
0.08 0.06
0.04 0.020-0.02-0.04-0.06
-0.08
-0.1AI
I
I
1970
1972
1974
1976
i~~~~~~~~~~~~~
I
1978
I
1980
1982
1984
1986
-
15 -
residualsof that esimated trend to be a proxy for the effect of weatheron agriculturein each
year. Il/ The results are consistentwith expectadonsoverall(figure6) and for individual
countriesand years where drought(Ghanaand Senegal,1983)or abundantharvests(Burkina1985,
1986)are knownto have occurreJ.
Major shockssuch as war, civil strife, and naturaldisasterssuch as floodsor cycloneshave also
had devastatingeffects on specificAfricancountries. The potendalimpactof thesephenomenonare
taken into accountin the modelby includingthe percentageof the populatdonaffectedby 'disasters"
as recordedin "MajorDisastersWorldwide"(USAID).
The real effectiveexchangerate (REER)is includedin the model(an increasein the REER
indicatesappreciation). As a measureof the eompetitivenessof agricultureit incorporatessome of
the incentivesaccountedfor by the real producerprice. However,REERsalso providea more
generalmeasureof distortionsin productand faitor markets,as well as serving as a proxy for the
indirecteffectsof the exchangeand trade controlsthat usuallyaccompanyovervaluation.
The results of the pooled cross-sectiontime-seriesmodelJAIare presentedin table 1 (and
Annextable 14). They includea numberof variationsof laggedand movingaveragesin the
independentvariables,and two subgroupsof countries; a) those exportingprimarilytree crops, and
b) those exportingprimarilyannualcrops. (Crop specificand countrymodelsare presentedbelow.)
The weathervariableis stronglysignificantin all of the equadons. The coefficientranges from
0. 15to 0.46 and is highestfor the annualcrop producers. The coefficientscan be interpretedas a
10 percentdrop from trend for cerealyieldsassociatedwith a 1.5 to 4.6 percentdeclinein exports.
The "disasters"variablehas the expectedsign and is significantin most cases.
The estimatedprice elasticitiesof export supplyfor all countriesrangefrom 0.1 to 0.3, and the
elasticitieswith respectto REE rangefrom -0.1 to -.25. For countriesexportingprimarilytree
crops (wherelongerlaggedresponsesare expected)the REERis the most consistentexplanatory
variablewith elasticitiesrangingfrom -0.14 to -0.25; the price elasticitiesfor tree crop exportersare
only significantwhen the REERis excluded(with an elasticityof 0. 115)or in one case significant
with the wrong sign (an elasticityof -0.06).
Higher price elasticitiescan be expectedfrom countriesexportingannualcrops such as tobacco
and cotton,as farmersare more able to respondquicklyto changesin relativeprice incentivesby
,3/ Cerealsare the most susceptiblecrop to moisturestress,and for most countriesvariationin
averageyieldsof cereals will resultprimarilyfrom variationsin weather. While exportedcrops
may responddifferentlyto specificweatherpatterns,the deviationsfrom trend can be expected
to be of the same sign. Year to year variationsin cerealyields couldarise due to fluctuationsin
ferdlizeravallabilityor other policyrelatedfactors,but with few excepdonsferdlizeris not
widelyused on cereals and other factorsare not likely to dominatethe effectsof rainfall.
J1/ The poolingof cross-sectiontme-series data can be accomplishedseveralways depending
on the characteristicsof the data. Sincethesedata are time-wiseautoregressiveand crosssecdonallycorrelated,a hree-stageprocedureis requiredto produceconsistent,unbiased
estimatesfoliowingKmenta(pp. 512-514),and using the SAS Park estimationprocedurefrom
SAS SUGISupplementalsoftware.
- 16 alteringtheir croppingpattern. The estimatedelasticitiesfor thesecountriesare indeedmuchhigher,
rangingfrom 0.56 to 0.94, suggestinghighlyresponsiveadjustmentsto changesin real prices. For
this subgroup,however,the real exchangerate variableis only weaklysignificantin one case, and
has the wrong sign. II/
Overall, these resultsindicatethat agricultureis only moderatelyresponsiveto changesin
pricingand exchangerate policiesin the short-term. Producersof annualcrops such as cottonor
tobaccorespondquicklyand with an elasticitynear 1.0. The estimatedelasticitiesfor tree crop
exportersare low, reflectingthe limitationson quickadjustmentsfor these crops. Even a three to
five year laggedprice variablewill be inadequatefor estimatinglong-runsupplyresponsesin these
cases, both becauseadjustmentscan take longerthan that, and becausefarmersrespondto changesin
expectedprices, not short-termbooms or buststhat will not be expectedto persist. Shorter-run
responsescan come only from rehabilitatingexistingplantations.
These results indicatethat for tree crops the indirectrelationshipwith respectto the exchange
rate.is stronger than the relationshipwith producerprices. This might reflectindirecteffectscaused
by the constraintsand distortingtrade and exchangecontrolsthat commonlyoccur when exchange
rates are overvalued. Or the estimatesmay be biasedupwardas a resultof the omissionof
unrecordedproductionand smuggledexportsthat tend to occur whenexchangerates are seriously
overvalued.
Crop specificmodelsof agriculturalexportswere esimated for severalmajor export
commodities(table 2). The results are similarto those in table 1. The short-runresponsivenessto
price is about 0.23 for coffee, aboutthe samefor cocoabut not significant. For cotton, an annual
crop, the responsivenessof exports to price is muchhigher with an elasticityof 0.67 both for the
producerprice and the real exchangerate. For tea the results were not significant,apparentlydue to
the small sample.
J1/ The significanceof the overall equationshas been testedusing a covariancemodeland is
highly significantat the 99% level. The three-stagepooledesdmationproceduredoes not permit
computationof F-values.
AfricaL
Table 1. Rearession EAuation for Agricultural Supplv in Sub-Saharan
All countries
Model
Producer Price
Moving average (t and t-1)
Real effective exchange rate
Moving average (t and t-1)
(1)
0.202
(12.97)
(2)
0.240
(204.2)
Tree Crop Exporters
(3)
0.017
(1.13)
(4)
0.115
(5.04)
-0.0010
(49.7)
Disasters variable
-0.0006
(-2.15)
Weather variable
0.373
(21.64)
0382
(299)
Intercept
4.12
(26.87)
3.44
(625)
Degrees of freedom
331
332
(5)
0.940
(6.26)
(6)
0.920
(6.08)
0330
(2.09)
-0.253
(-13.86)
-0.104
(4.15)
Annual Crop Eyporters
-0.0003
(-0.43)
-0.0016
(-1.69)
0.0013
(0.92)
0.0009
(0.67)
0.153
(5.18)
0.219
(5.54)
0.440
(6.29)
0.435
(5.7)
5.62
(41.5)
4.01
(36.8)
-1.29
(-1.46)
033
(0.47)
219
220
91
92
been expressed in logarithms. Their coefricianctmay be
Note. Exports, producer prices, exchange rate, and the weather variable have
interpreted as elasticities.
Figure in parentheses are t-values.
-
18
-
Table 2. RegresIo Eaudtons for Agdricutural-SuDply
In Sub-SabaranA-frca: CrM models
Depndant v
Producerprice a/
ble: total agrcultural exports
Cocoa
Coffee
0.22*
(1.75)
0.23**
(8.07)
Cotton
Tea
0.67** -0.04
(4.02)
(-0.50)
Real effecdve exchangerate a/
4.35**
(-3.91)
Disastersvariable
-0.0067 -0.0037** 0.0026 -0.0192**
(-1.49) (-2.86)
(1.64)
(-3.14)
Weathervariable
-0.20
(-1.65)
Intercept
Degreesof freedom
Countries
0.055* -0.68**
(1.94)
(-5.04)
-0. 163** 0.226*
(-2.82)
(2.4)
5.16** 3.30**
(5.8)
(15.2)
4.43**
(3.9)
0.126
(0.65)
0.033
(0.21)
4.24**
(4.02)
107
219
171
59
7
14
11
4
Notes: Exports, producerprice, exchangerate, and the rainfallvariablehave been expressedin
logarithms. Their coefficiantsmaybe interpretedas elasticities.
Years refer to split crop years (e.g. 1980/81). Sincemarketingoccurs in the secondof the two
calandaryears, exportdata for the 1981calandaryear correspondto prices, rainfal, etc. for 1980.
The pooled cross-ection time series procedurewas used.
Figures in parenthesesare t-values. Significancelevels are * (955) and ** (99%).
a/ Two year movingaverage (t and t-1)
Table 3. Regression Eguations for Agricultural SuRpIv in Sub-Saharan Africa: Country models.
Dependant variable: total agricultural exports
Ethiopia
Ghana
Kenya
Producer Price a/
-0.84*
(-2.3)
1.17*
(3.8)
-0.48
(-2.0)
Real effective exchange rate a/
-1.64*
(-2.55)
0.32
(1.6)
Malawi Nigeria
Senegal
Tanzania Togo
Zimbabwe
-0.07
(-0.14)
0.22
(0.56)
1.52
(1.83)
0.42
(0.58)
2.1**
(3.3)
0.86*
(2.48)
0.35
(0.27)
2.13*
(2.99)
-1.03**
(-4.4)
1.65
(0.22)
-1.07*
(-2.6)
0.24
(0.29)
0.29
(1.75)
Disasters variable
-0.0016 -0.019* -0.046
(-0.15) (-2.68) (-0.75)
-0.005
(-0.12)
4.18
(1.52)
0.0008 -0.007
(0.126) (-0.24)
Weather variable
-0.029
(-0.069)
-0.40
(1.84J
0.96*
(2.23)
-0.40
-0.023
(-0.049) (-0.66)
Intercept
15.9
(3 79)
-2.15
(-2.17)
5.2
(0.80)
Degree of freedom
12
12
12
R square
0.66
0.85
0.72
0.068* -9.03
(-0.66)
(2.51)
1.36
(2.02)
-0.65
(-0.66)
-0.146
(-0.38)
0.56**
(4.08)
-4.90
(-3.78)
&16** -10.07
(-1.45)
(3.1)
7.54
(1.62)
-6.17
(-1.01)
-0.73
(-.41)
12
12
12
12
12
0.85
0.81
12
0.64
0.75
0.78
0.62
Note: Exports, producer price, exchange rate, and the rainfall variable have been expressed in logarithms. Their coefficiants may
be interpreted as elasticities.
Years refer to split crop years (eg. 1980181). Since marketing occurs in the second of the two calendar years, export data
for the 1981 calendar year correspond to prices, rainfall, etc for 1980.
*
indicates statistically significant at the 95% leveL
indicates statistically significant at the 99%OleveL
**
Two-year moving average.
a/
-
20
-
Country-specific models were also estimated (table 3). There are striking differences between
these models and those estimated pooling cross-section and time-series data. The estimated
elasticities differ widely between countries. For a number of countries the elasticity with respect to
price, or with respect to the exchange rate, is quite high (Ghana, Nigeria, Tanzania, Togo), often
around 1.0, and in one case exceeding 2.0. / These wide differences in estimated elasticities may
reflect the small number of observations or other data problems, or the diversity among African
countries that gives rise to real differences in the responsiveness of farmers to price incentives (see
Section IV below).
Estimates of aggregate supply response in the literature range from 0.1 to 0.5 (Binswanger,
Bond, Chhibber). But procedures used to estimate long-run elasticities such as the Nerlove
techniques (which uses lagged dependent variables to derive intertemporal adjustment coefficients)
provide estimates that are generally believed to be too low and not good estimates of the response
of crops to a permanent change in the price regime of agriculture (Binswanger). In addition, there
are problems with the interpretation of distributed lag models, and simultaneity problems that can
arise in the data.
In summary, current statistical methods appear inadequate to estimate reliably long-run supply
response. Since the year-to-year price fluctuations in the data normally reflect short-lived
commodity booms rather than permanent changes from a low price to a high price regimen, any
estimate based on these data will reflect short-run adjustments rather than long-run responses to
permanent changes in price levels (Binswanger). Farmers will respond to expcted prices, which
are generally not observable. Given the preponderance of tree crop exports in Africa, it is not
surprising that these short-run elasticities are low. The long-run aggregate supply response will
include the effect of reallocation of productive resources, labor and capital, among sectors in the
economy overall. It may also include changes in government expenditures and public capital
investments for infrastructure, research, human capital and institutional support which may be more
forthcoming in the context of higher incentives.
m. THE IMPACT OF POLICY ON THE FOOD CROP SECTOR
Analysis of the food crop sector in African countries is more problematic than export crops
due to poor quality and limited availability of the production and price data -- in addition to other
differences. Many African countries set official food prices, and thus only the officially announced
price figures are available. For a majority African countries, however, official prices are not
effectivelyenforced, and most producers receive a market determined price for their food sales.
16/ In three cases the estimated price elasticity is negative -- Ethiopia, Kenya, and Malawi -although only significant for Ethiopia. These three countries differ from the others in that
producer prices are determined ex post at international auctions. Farmers will not respond to
current or lagged prices unless they reflect changes in expected price. Because prices received by
farmers in these countries fluctuate widely from year to year, very high prices paid in one year may
do little to change their price expectations unless there is some perceived change in the long-term
price level for their products in international markes. Hence, the lack of a signiflcant relationship
is not surprising, although in the case of Ethiopia the negative sign is puzzling, and may reflect
complications related to coffee quotas, or the influence of Ethiopian exports on price.
- 21 -
Cross-Driceeffectsbetweenexportprices and food
The elasticitiesesdmatedabove for exportcrops will overstatethe responsivenessof aggregate
agricultureif higherexports are offset by a faUin food production(in the sameway singlecrop
esdmationstend to overstate.ggregate supplyresponse). If the exportresponsehas comefrom a
shift of resourcesaway from food cropsand toward exportcrops, then the net increasein total
agriculturaloutputwill be lower.
Direct esdmationof food crop modelsin the aboveanalysisis not possibledue to a lack of price
data. However,by replacingthe dependentvariablein the aboveformulationwith food producdon,
it is possibleto test the hypothesisthat exportincentiveslead to offsettingeffectsin food production.
Similarly,the relationshipbetweenfoodproductionand the real exchangerate can be esdmated.
The results of these modelsare shown in table 4. Three differentdependentvariableshave been
used; total agriculturalproduction,total food producdon,and staplefood production. The estimated
elasticitiesindicatethat foodproductionrespondspositivelyto changesin exportcrop producer
prices. Total agriculturalproduction,total food productionand staplefood producdonrespond
positivelyto changesin exportcrop prices and real exchangerates. The resultwith respectto the
exchangerate is expected,as both tradablefood crops and exportcrops are affected in similarways
by exchangerate distortions.17i
There is no evidenceof offsettingchangesbetweenexport and food crops. Even for countries
where the principalexportedcrops are annualswith high supplyelasticitiesof about 0.9, the crossprice elasticitieson total food or staplefoodsprovideno statisticallysignificantevidenceof a traderesult is for a positivecross-priceelasticityfor staple
off. Indeed,the only statisticallysignificaint
foods.
There are severalpossibleexplanationsfor the positivecross-priceeffjct of export crop price on
food production: a) food and exportcrops maybe complementsin prodaction(fertilizerbenefits
both crops whengrown togetheror in succession),b) returns on exportproductionpermit higher
inputuse or investmentin food crop production,or c) a third, omitted,factor - such as government
policy that generallyfavors agrdiulture- affectsboth exportand food crops in the sameway. The
results presentedhere are consistentwith other empiricalfindingsthat have founda positive
correlationbetweenexpansionof exportcrops and growthin foodproduction. One such study
concludedthat "conditionsfavoringagriculturalgrowth, includingappropriateeconomicpolicies,
encourageboth cash crop and basicstaple food crop production"(von Braunand Kennedy).
17/ To the extent that the REER maybe a proxy for agriculturalterms of trade overall (for both
exportand food crop incentives)the aggregateresponsivenessimpliedhere for total agricultural
productionis an elasticityof 0.045, which is very closeto the 0.05 estimatefor aggregateoutput
in India by Bapna(reportedin Binswanger).
-22-
Table 4. Regerssion equations for cross-price effect on food production
Independent variables a/
Dependent variable
All Countries
Total agricultural production
Food production
Staple foods
Annual crop Droducinifcountries
Total agricultural production
Food production
Staple foods
Export crop producer
price
(2 year moving avg.!
0.033**
0.046**
0.065**
0.028
-0.016
0.183*
REER
(2 ear moving avg.)
-0.045**
-0.013*0
-0.018**
-0.110
0.085
0.401
Statistically significant at the 95 percent level.
Statistically significant at the 99 percent level.
S*
Note: "staple foods" is an aggregate of total production of cereals, roots and tubers, pulses, oil
crops, bananas, and plantains in grain equivalent units converted on the basis of their relative
caloric content.
a/ Full model is same as model 1, table 1.
*
The importance of food production and policy goes beyond its relationship with export
agriculture. Food policy in Africa differs from export agricultural policy In many respects. For
this reason, and due to the more limiting data, food policy and productivity are assessed separately
below.
Food Policv
In only about half the 35 African countries in the sample do governments intervene
substantially in the marketing and pricing of food crops (eleven fewer than in the 1970s). In some
of those that do, or did, intervene, government intervention was restricted to marketing of
imported foods, mainly rice or wheat to urban consumers, having indirect effects on domestic
producers.
The effectiveness of government food marketing and pricing policies is difficult to assess
precisely given the paucity of information on the size and scope of parallel markets in many
African countries. Among those countries that do intervene to set prices and control marketing,
however, evidence suggests that few do it effectively. Where details are available significant parallel
markets appear to exist. Zambia appears .o be an exception, where the very high subsidy to
official marketing has prevented much private trade from developing (Harvey), although there are
substantial cross-border food movements to neighboring countries. In Zimbabwe, Kenya, and
Botswana, controls are at least partially successful,but parallel markets exist. Nevertheless, accounts
differ on the importance of parallel markets in many countries. For example, in Mali prior to
liberalization, one account reports that the existence of parallel markets implied free price
-
23 -
formation(Lecaillonand Morrisson).Andin Ethiopiagovernmentconfiscationand restrictionson
tnsportation of food affect only certainregions. I8/
Where data are availablecomparisonsof food crop producerprices to border prices, or NPCs,
show wide swingsfrom year to year and across countries(AnnexD table 11). In many countries
producerprices are higher thanborder prices (the NPCs are greater than 1.0) becauseof protection
from imports- due either to governmentpoliciesor high transportcosts which createnatural
barriers. While most of thesecountriesimportfood, in goodyears they mayalso export food to
neighboringcountries. In some countriesthe net effect on producerprices is an implicitsubsidydue
to restrictiveimport quotas(Nigeria, Ghana)or rationingof foreignexchange.
The high transportcosts (annexD table 7) and year-to-yearswingsin domesticproductionmakes
establishingthe level of protectiondifficult,since the food price may fluctuatewithina wide range
betweenexport and importparity prices. As long as the price remainswithinthat band, the
influenceof border prices on domesticmarketprices is ambiguous,and it is unclear whetherthe
producerprice shouldbe comparedto exportor importpar.ty at all for assessingdistortions. In
situationslike this the NPC provideslitte useful information.12/
Real food price trends
Real producerprices for food crops have fluctuatedmuchless thanfor exportcrops, exceptfor a
rise in prices foUowingthe world food shortageof 1974(figures7 and 8), aside from which there is
no perceivedtrend on average. The ratio of exportto food crop p;ices, however,has changed
significantlyover the period, falling from 1970to 1984(exceptduring the commodityboom in 197579). OveraUlthe price ratio declinedby about20 percentover the period, Since1984, however,the
trendhas seen a large reversal, with export crops regainingtheir 1970parity in just two years,
followedby a small declinein 1986.
These patternshave been due to a combinationof exchangerate, trade, and pricingpolicies.
Exchangerate policieshave discriminatedagainsttradablefoods,drivingdown the real prces of
exportedand import competng commodities.In some cases, however, foreignexchangeconstraints
and import controlslimit importsand thus may giverise to somewhathigher foodprices. The net
effect of thesepolicies is thus ambiguous. To the extent that domesticallyproducedfoodsare
nontradedcrops (cassavaand yams) that are imperfectsubstitutes,the appreciationof the real
exchangerate may raise their price relativeto tradedfood and exportcrops.
Trends in consumerfood pricesshow that relativeto border prices at officialexchangerates,
domesticallyproducedfood cropsbecamemoreexpensiveduring the 1970s(AnnexD table 12)
jJ/ Even with more detailedinformationon the paraUelmarketsin food crops, it wouldstill be
difficultto assess the impactof governmentefforts to restrict private trade on producers
incentives. Enforcementis rare, but where it does occur, and where transportersare restricted
by law from carryingfood crops, low officialpriceswill have some disincendiveeffects.
12/With high transportcosts from producingregionto border, an NPC (computedas an import
competingcommodity)less than one has an inconclusiveinterpretationof the effectof policy. It
may simplybe that the price has not falen sufficientlylow for the commodityto becomean
exportedgood.
-24 -
Figure 7. Real producer prices for major
crops in Sub-Saharan Africa
120
Index 1970-75 = 100
110
100 X
90
80
\
70
1970
1972
1974
-
1976 1978 1980 1982
18 country average
Food crops
-
Figure 8. Ratio of export/food
prices in Sub-Saharan Africa
110
Index 1970-75
=
1984
1986
1984
1986
Export crops
producer
100
105
100
9590
\
851970
1972
1974
1976 1978 1980
18 country sample
1982
- 25 prices and
relative to imported rice and wheat. Cheaperimportsresultedfrom lower internatdonal
overvaluedexchangerates. Trade restrictionsand imperfectsubstitutionbetweenfood'sappearto
have keptprices of domesticallyproducedfoodssomewhathigher.
When traded foods are distinguishedfrom nontradedroots and tubers, the trends in the 1980s
illustratethe effectsof exchangerate distortionsand changesin domesticpricingpolicies(table 5).
As distortionsin real exchangerates have declined-- in additionto food market liberalizationin
some countries- prices of tradedfoods have risen, whilenontradedfoodprices have fallen. This
should have the effect of shiftingdemandtowarddomesticallyproducedfoodsand away from
imports(discussedin moredetailbelow).
Tale 5. Comarison of real consumerfood prices in Africa
1977-79
Indexes
1980-82
1985-87
Traded food crops (n= 14)
100
103
112
Nontradedfood crops (n=9)
100
96
89
Overall, food priceshave been affectedless thanexport cropsby governmentpolicies(with a
few notableexceptionswherepricing and markedngare effectivelyenforced). This has been
because: a) in manycountriesdirect policyinterventionshavebeen ineffective,and b) exchangerate
policieshave effecteddomesticfoodprices less in part becauseof trade restrictionsthat have limited
importsof cheapfood.
Food imnort. and slow food mroductiongrow
Food productionin Africahas grownmore slowlythan populationsince 197020/ leading
manyobserversto concludethat this 'deterioration"is evidenceof the generalfailure of African
agricultural:
"Many Africanstateshave slowlylost the capacityto feed their people"(Eicher 1986).
"[Africa's] ppoorperformancehas led to a remarkableincreasein food imports.... The
availabilityof food importshas greatly offset the impactof poor agidulture performanceon food
security,but not enoughin aggregateto ensure an adequatediet to the average African
(Serageldin)."
20/ Per capitafood productionhas fllen at a rate of about 2 percentper year since 1970- the
only regionin the world with decliningper capitafood production. At the sametme Africahas
becomea net importer of food, doublingfood importsduringthe late-1970swhile food exporu
stagnated. Compoundingtheseoverall trends are the periodicand transitoryshortagesand food
insecuritythat affect large mnmbersof peoplein Ethiopia,Sudan,Mozambique,and the Sahelian
countriesmost prone to food shortagesover the past 15 years due to droughtor war or both.
- 26 -
Consumptionof food, however,has not declined. Sincethe 1960saveragecaloricconsumptionhas
been relativelystableor rising slightlyin Africa. There has, however,been a shift in the compositionof
consumptiontowardan highershare of importedfoods, especiallyrice and wheat.
But while importshave clearlyb-en substitutedfor domesticproduction,this leaves open the
questionof the casualdirectionof the shift betweenimportsand domesticsupply. Has deteriorationof
productivecapacityforcedAfricansto importmore food? Or has a shift in demandin favor of imports
resultedin decliningdemandfor domesticallygrownfood (and thus resultingin slowergrowth in
production)?
Demandfor food importsmay rise for severalreasonsindependentof the adequacyof domestic
supplies. Importedfoods(generallyrice and wheat)are imperfectsubstitutesfor other, domestically-grown
staples(millet, sorghum,maize,cassava,and yams). Importedrice and wheattend to be preferredfoodsin
Africafor reasonsof tastes, socialstatus, and ease of preparation,and they tend to exhibita high income
elasdcityof demand(Delgado). Preferencefor importedfoods is most oftenmanifestamongurban
populationsin Africa.
As a result, changesin importdemandshouldresult from changesin the followingvariables:a) the
relativeprice of importedfoodsto domesticfoods, b) income, especiallyforeignexchangeearnings,and c)
demographicchangessuch as urbanizationwhich wouldincreasethe share and numbersof the population
with a high propensityto consumeimportedfoods. Beforetestingthesehypotheseseconometrically,each is
elaboratedfurtherbelow.
The relativeprice of importedto domesticfoods willbe affectedby changesin internationalprices,
exchangerate policies, changesin barriersto trade such as importrestrictions,and officialdomesticpricing
policies. Exchangerate policieswiUalter the price of domesticfood relativeto importedfoods to the extent
that domesticfood pricesare not fully affectedby internationalmarketprices. High transportationcosts,
other barriersto trade, or the lack of internationalmarketsfor staplessuch as yams, will make these
productsessentiallynontradablegoodsand thereforetheir prces wil rise relativeto importedfood prices as
currenciesbecomeovervalued. Additionally,when currenciesare overvaluedgovernmentsgeneraUymust
rationforeignexchangeand place restrictionson importsin order to avoidexhaustingreserves. These
restrictionsmay lead to inefficienciesin domesticfood productionand marketingto the extent that it is
dependenton importedinputs such as fertilizer, fuel, or fbrm machinery. If, however,import restrictions
involvefood importsthemselves(eg. the banningof food importsin Nigeria)then the net effect of exchange
rate distortionscouldbe to reducefood imports,but ornya few countrieshave implementedeffectiveimport
restrictions. Domesticpricing policieswhichdiscriminateagainstproducerswill limit the extent to which
domesticproductionis stimulatedby growthin demand,givingrise to a policy-inducedfood gap. In
addition, consumeroriented"cheapfood' policieswill, to the extent that they applyto importedfoods such
as rice and wheat,further shift demandtowardimportsand away from domesticallyproducedfood.
Higher incomeswiUraise consumption,includingimports. A numberof studieshave observeda
strong positiverelationshipbetweenhigh agriculturalgrowth and rising food importsin developingcountries
generaly. Thesestudies cite the incomeeffect of strong agriculturalgrowthas leadingto increaseddemand
for importedfoods (Bautista). Whilefew Africancountriesfit the categoryof havingstrong agricultural
growth, a numberhave seen periodsof rising incomesfrom other sourcesover an extendedlengthof time
(eg. oil exporters). In particular,a country's importcapacity,or the total of exportearnings,foreig
investment,borrowing,and grant aid, wiUaffect demandfor importedfoods overall.
Growthin urban populationis likely to lead to higherfood importsif the propensityto
- 27 consume imported food is higher among urban dwellers. This appears to be the case for reasons
of higher urban incomes, status, and easier and quicker time of preparation (Delgado). Urban
growth due to rural-to-urban migration is influenced by government policy. Macroeconomic
policies, pricing policies, and the provision of social services all contribute to shifting the
ruraVurban terms of trade in favor of urban dwellers and thereby making migration from rural to
urban areas more attractive. This is likely to increase food impets due to higher demand, but
also there may be a supply-side effect if urban migration depletes the rural labor force leading to a
decline in domestic production.
Agerereatepattern of food imports. The aggregate patterns of Africa's food imports are presented
in figure 9. Food imports varied little between 1970 and 1976, ranging between 7-8 million metric
tons. Over the next four years, however, food imports doubled to more than 14 million tons, and
remained at these high levels until 1986 when they began to fall gradually. Surprisingly, the period
during which this sharp increase occurred was a very favorable period in terms of the weather
pattern and its effect on cereal yields (except in 1979). And food aid accounts for less than 10
percent of the increase, although food aid shipments continued to grow in the 1980s raising its
share in the total considerably, especially following the 1982 and 1983 droughts (imports net of
food aid fell to under 10 million tons by the mid-1980s).
On closer examination of these data, however, one notes that the doubling of Africa's food
imports during this period resulted in large part from events in one country, Nigeria (figure 9).
Indeed, seventy percent of the increase in Africa's food imports -- or about 5 million tons - is
accounted for by changes in Nigeria's imports during that country's oil boom. Given the events in
Nigeria during that period, a number of contributing factors suggest themselves. First, oil revenues
resulted in a sharp rise in the nation's income and capacity to import. Annual import capacity,
defined as export earnings, net borrowing, direct foreign investment, and aid, nearly doubled,
increasing by $14 billion during the same period (figure 10). Second, the disruptive economic
impact and "dutch disease" effects of the oil boom on macroeconomic variables resulted in rising
prices and wages, an increasinglyovervalued currency, a shift of labor from traditional agriculture
to industrial and service sectors, and rural-to-urban migration. All of these factors contributed to
both rising demand for food imports and a decline in domestic production relative to population.
The changes which were brought about were of such magnitude that for three of Nigeria's four
major agricultural exports -- peanuts, cotton, and palm oil -- exports ceased, and in all three cases
Nigeria became a net importer of these commodities by the 1980s.
Thesevery dramatic changes that occurred in Nigeria -- Africa's largest economy -- are likely
to be indicative of phenomenon at work in other countries, albeit to a lesser degree. How
generalizable these phenomenon are in explaining Africa's food crisis, is the question to which we
now turn.
Empirical mode. To test the significance and relative importance of these variables in explaining
changes in Africa's food imports an econometric model is estimated for a sample of 31 countries
with data from 1971 to 1987. For a sample with these characteristics, pooled cross-section, timeseries analysis was chosed as potentially the most powerful method. Food imports are regressed on
-28 -
Figure 9. Africa's food imports
million metric tons
16
14 13 -
.....
.
.
121
10
1 1 .~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. .....
......
. ......
,-
-
Y
. ......... .............
,,
..........
------~,,
7-
2. .... .. s;f._A
... _Y_.__
. ____
__~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
... ....
1 4
.
...................
....
...............
... ..... ... .. ......
.
. ._..
...-.......
.. ....
_
_ _
2
1970
1972
-
1974
1976
Total
6
1978
1980
Total less Nigeria
1982
1984
1986
1988
Food aid
-4-
Figure 10, Africa's import capacity
(exports, borrowing, aid, investment)
70
billions of $US
65
..
..
Total-- -- ---
- ----
l
60
Nigeria
..........
. ........................
....\...........
50 -\
<
45
. .....-.
40
1970
L_
.............
.......
\.
----- ...
.........
1972
1974
-
1976
Total
1978
1980
1982
°Total less Nigeria
19B4 1986
-29 urban population,the real effectiveexchangerate, importcapacity,a "weather"variable,and trend in the
followingway: 21W
FOODIMPORTS = bo + bjURBANPOPULATION+ b2 REER+
b3EMPORTCAPACITY+ B4 WEATHER+ BsTREND+ e
Food importsare total volumefigures, net of food aid. 22/ URBANPOPULATIONis an index
based on the total numbersof urban residents. The REERis the real effectiveexchangerate as definedby
the IMF. It is an index which measuresthe evolutionof a country's prices relativeto those of its trading
partners, adjustmentfor nominalexchangerate changes. The IMPORTCAPACITYis the sum of export
earnings, net borrowing,foreigninvestment,and grant aid deflatedby Africa's importprice index.
The two variables, WEATHERand TREND, reflectthe hypothesesfor a supplyside explanationof
variationin food imports, whichare unrelatedto prices, policyvariables,or urbanization. A secularly
increasingTRENDin food importswouldsupportthe hypothesisthat there has been a deteriorationin
Africa's productivecapacity. Andpoor WEATHER(rainfallas well as tornadosor floods)that affect
domesticproductionwill necessitatehigherimports. Clear cases of weather-relatedcauses occurredduring
the periodconsidered,includingfrequentseveredroughtbetween 1973and 1984. Since 1984Africaappears
to have seen more normal rainfallpatternson average. The WEATHERvariableis a proxy derivedfrom
annualdata on cereal yieldsper hectarefor each country. It is esimated as the percentagedeviationfrom
trend in cerealyields over the period.
Empiricalresults. Table 6 presents the results of the regressiondescribedabove. All estimatesof the
coefficientsin the equationagree in sign with a priori expectations. All the explanatoryvariablesare
statisticallysignificantat the 1 percentlevel, exceptfor the REERwhich is significantat the 5 percentlevel.
The estimatedcoefficientand significancefor urban populationindicatesthat urban growthis
stronglyassociatedwith increasedfood imports. This variableheld the strongeststatisticalrelationshipin
terms of the additionto the r-squaredwhen the variablewas addedto the equation(usingordinaryleast
squares estimation). The additionto the r-squaredwas0.237, as comparedto 0.07 for import capacity,0.05
for the weathervariable, and 0.0017 for the trend.
It shouldbe recognizedwith respectto the urban populationvariablethat the estimatedrelationship
may concealan underlying,omittedvariablewhich is correlatedto both. In particular,governmentpolicies
which alter the rural-urbantermsof trade in favor of urban consumers- such as cheap foodpolicies- will
simultaneouslyencourageurban consumptionof the subsidizedfoods(generallythese programsinclude
importedrice and wheat)and urban migration,which will likewise
21/ Data are from World Bankand FAO. Data for food imports,urban population,and import
capacityare normalzedas indexesto abstractfrom countrysize (1970-87average = 100).
L2/ Food aid is excludedfrom food importsbecauseit is a componentof grant aid and therefore
it would createan endogeneityproblemif includedin the right-hand-sidevariable, "import
capacity." Also, while emergencyfood aid shipmentsrespondto domesticshortfallsin
production,increasinglyin the 1980sthe relationshipbetween"structuralfood aid" shipments
and patternsof demandand supplyin Africawere not obvious.
-
30 -
lead to rising food consumptionand imports. Indeed, Zambia,where cheapfood policies
have been in place for 40 years, has the most urbanizedpopulaion in Africa.
Table 6. Regressioneuation for Sub-SaharanAfrica's foodimports
variable
Dependentvariable: food imports
Coefficient
Constant
Urbanpopulation
Real exchangerate
Importcapacity
Weather
Trend
13.3
0.29
0.06
0.40
-59.3
1.09
t-value
( 3.3)**
(10.7)**
( 2. 1)*
(21.0)**
(-23.7)**
( 7.4)**
Degree of freedom = 487
Source: WorldBank and FAO.
* indicatessignificanceat the 95 percent level.
** indicatessignificanceat the 99 percent level.
Note: pooledcross-sectiontime-seriesGLSmethodwas used.
An additionalambiguityiiathis result is the extent to whichurbanizadonhas affecteddomestic
food productionas a result of a reducedagriculturallabor force. Outmigrationof agriculturallabor
may resultin slowergrowthin productionrelativeto overallpopulationgrowth. The upward
pressure of prices for domesticfoods that might resultwould furthershift demandtowardimports.
If, however, producerfood prices are controlled,or if cheap food importsor food aid keepproducer
prices low, the price signalsnecessaryto stimulategrowthin domesticfood productionfor an
increasinglyurban populationmay not ensue. To the extent that this is occurring,some of the
increasein food importsmaybe seen as a supplyproblem,but one whoseorigins lie in rural to
urban migrationand the policies whichhave encouragedit.
The estimatedcoefficientfor the real exchangerate impliesthat exchangerate policieshave
contributedto the growthin Africa's food imports. There are severalways in which overvalued
exchangerate will contributeto food imports. The most direct way is through its effect on the
relativeprices of importedversusdomesticallyproducedfoods,manyof which are nontradedor
have significantnatural barriersbetweendomesticand internationalprices. Thus whenexchange
rates becomeovervaluedimportedfoodprices will fall relativeto domesticprices leadingto a shift
in consumptiontowardimportedfood. In addition,overvaluedcurrenciesare usuallyaccompanied
by trade and exchangerestrictions(in order to stem the excessdemandfor foreignexchange
reserves). This may have the addedeffect of reducingproductivityin domesticfood productionin
cases where food productionrelieson importssuch as fertilizer, fuel and farm machinery. In some
cases, howeve , the accompanyingimport restrictionsmay also applyto some importedfoods. If
effectivethis wouldhave the oppositeeffect of reducingfood imports. The existenceof such
contradictoryeffects mayexplainthe lower significancelevel for this variable.
The coefficientfor import capacityis positiveand significantindicatingan incomeeffect, oa that
changesin aggregatepurchasingpower are positivelyassociatedwith changesin food imports. As
describedabovethe vcry large increasesin Nigeria's importcapacitycoincidedwith the doublingof
-31Africa's food imports,70 percentof which went to Nigeria. Manyother cour. ies were affectedto
a lesser degree by commoditybooms in the late 1970sfor cocoa, coffee,and phosphates,as well as
the four other countriesexportingoil. And growthin aid and borrowingof recycledOPECrevenues
raised tae capacityto importof manyAfricancountriesduring the late 1970s. By contrast,the debt
crisis in the 1980sreducedsubstantiallythe importcapacityof manyof these countries.
The weathervariable(laggedone year)is significantand has the expectedsign: cereal yields
below trend in year t-l are associatedwith increasedfood importsin year t. Despitethe exclusion
of food aid from the dependentvariable,commercialfood importsare responsiveto shortfallsin
domesticproductioncausedby weather-relatedfactors. Deterioratingweather,however, does not
explain the secularrise in food importssincethis variableis in terms of residuals,or deviationsfrom
trend, and the trend for most Africancountries,showedincreasingyieldsfor cerealsover the period
The trendvariableis significantand has the expectedsign. The value of the coefficientis small,
however, suggestinga rise in food importsof only about 1 percentper year. 23/ In addition,the
significanceof the variablein termsof its additionto the r-squaredwas very smallas reported
above. Nonetheless,a likelyexplanationfor the significanceof a slight trend is the declinein
internationalrice and wheatpricesover the period, whichwouldbe expectedto raise imports.
Wheatprices have falien in real termsby a third, and rice prices by half, between 1970and 1986.
Thus, in additionto the impactof exchangerate policies on the relativepricesof domesticversus
importedfoods,border pricesfor importedfoodshave fallen over the period, encouragingadditional
substitutiontoward importsin consumption.
Overall, the regressionresults provideimportantevidencethat Africa's food import growthhas
been, in part, demanddrivenrather thanresultingfrom failed domesticproduction- althoughurban
migrationand trade and exchangerestrictionsare likely to havehad an effect on supplyas well.
Demandhas been influencedby go urnmentpolicies,urbanization,and importcapacity.
Urbanizationis the strongestexplanatoryfactor. But governmentpolicies,to the extent they account
for differencesin rates of urban migration,under the strongrelationshipateributedto urbanization.
Evidenceof that link is presentedbelow.
Policyand rural-urbanmigration
The link betweenpoliciesand the rate of rural-urbanmigrationhas only been imprecisely
identifiedthus far. Governmentsaffect the rate of rural-urbanmigrationthroughmacroeconomic
and pricing policiesthat alter the domestic,or rural-urbanterms of trade. Theseincludeexchange
rate policies and cheapfood policies, both of whichgenerallydiscriminateagainstthe rural sector.
In addition, govermnentprogramssuchas the provisionof schools,health service,and water
supplies,may favor urban locationstherebyencouragingmigration. But determiningthe extent of
these influencesis problematicdue to poor censusdata and the gradualnature of thesedemographic
shifts. Anecdotalevidenceprovidessome supportto the importanceof thesepolicieson
urbanization. For example,Zambia,whichhas had highly subsidizedcheapfood policiesfor more
than40 years, has the highestshare of urbanpopulationin Sub-SaharanAfrica.
2D Althoughnot expressedpreciselyas a growthrate, the dependentvariableis an indexwith
an average value of 100for the entire period. Thus a coefficientof 1 can be takenas
approximatinga 1 percent per year trend.
-
32 -
Only with large changesin policyvariableswouldone expectobservableshifts in the rate (or
even the direction)or rural-urbanmigration. Someobserverscontendthat in several countriessuch
as Ghana, Nigeria,and Tanzania,whereprofoundpolicy changesalteredthe domestictermsof trade
in favor of the rural sector, urban dwellersare returningto the farm. Becauseof poor demograplhic
data, however,these patternshave been difficultto substantiate.
In the case of Ghana,however, recent surveydata from the World Bank's GhanaLiving
StandardsMeasurementSurveyoffers an opportunityto authenticatethe pattern more concretely. In
a surveyof over 8,000 individualsin Ghana,informationon currentand prior employmentwas
collected. From this data, respondentswere broken downby principaloccupationinto agricultural
and non-agricultural,and similarlyfor previousoccupation. For those whochangedoccupations
betweenmid-1984and 1988- the timeof the survey-- prior and currentoccupationwas tabulated
as shownin table 7.
Table 7. Rural-urbanmigrationin Ghana sincestructuraladiustment
(principaloccupation1984-88,in percent)
CurrentOccupation
Agriculture Non-agriculture
Peius
Total
ocMuation
Agriculture
Non-agriculture
Total
63
2
65
4
31
35
67
33
100
Source: Ghana Uving StandardsSurvey (WorldBank).
Note: samplesize = 5570
Thesedata reveal a significant evre migrationfrom urban (or nonagricultural)to rural
(agricultural)occupationssincethe reformprogramwas initiatedin the early 1980s. Among
individualsthat have changedoccupationsduringthe period, those movingfrom nonagriculturalinto
agricultureoutnumberthose movingin the oppositedirectiontwo-to-one. Given the proportionsof
the rural and urban populationsas a whole (and assumingit is roughlythe samefor the agriculturalto-nonagriculturalbreakdown)these data indicatea two-percentnet increasein the share of the
populationearninga living in agriculturesince 1984. This contrastswith aggregatedata prior to
1984indicatingmigrationin the other directionat about 1 percentper year. The impactof this
reversalon productivityof both food and exportagricultureappearsto have been considerable: In
the late 1980sGhanasubstantiallyimprovedfood productionand dramaticallyincreasedcecoa
exports(figure 17 below).
These data confirmwhat many observershave suspectedor observedanecdotally;that
governmentpolicyhas a profoundeffecton migrationfrom rural to urban areas. As a corollary,
thesepolicieshave also had an importanteffect, both directlyand indirectly,on the growth in
Africa's food imports.
33 IV. DIVERSITYWITHIN AFRICA:
POLICY ENVIRONMENTAND AGRICULTURALPERFORMANCE
The pattern for individual countries will vary from the average relationships described above.
And the responsiveness of agriculture to changes in price and nonprice variables will certainly
differ across countries. To some extent this will depend on the type of policies pursued, whether
farmers anticipate policy changes to be permanent or temporary, as well as structural factors such
as the state of transport and other infrastructures. Recent debate on the effectiveness of policy
reforms and adjustment programs has focused both on what evidence there is to indicate that these
programs are working, and on to what extent can one expect adjustment programs to be equally
effective in different countries.
To try to address these two questions, and to provide evidence complementary to the patterns
and results above, this section will examine the patterns for countries grouped according to their
policy environment, and it will examine a number of individual countries as well. The emphasis on
the comparison of country groupings is to discern evidence of divergent trends between those
countries pursuing policy reform programs and those that do not.
Comparing performance in favorable versus unfavorable policy environment
One approach to addressing the importance of the policy environrent empirically is to
compare the performance of countries that have pursued policies favorable to agriculture, with that
of countries whose policies have been less favorable to agriculture. Although the set of domestic
policies which impinge on the agricultural sector differs by country, key policy variables can be
used to create this typology. The approach taken here is to compare countries which have
maintained or adopted a policy environment favorable to the agricultural sector. The classification
is based primarily on two variables, exchange rate policies and producer pricing policies. Thus, by
examining 1) the extent of exchange rate distortion - revealed in the real exchange rate and the
ratio of parallel market exchange rate to official rates -- and 2) the trends in real prices paid to
produceis as well as the nominal protection coefficient, the countries in Sub-Saharan Africa have
been classified as those with a Favorable Policy Environment (FPE), and those with Unfavorable
Policy Environments (UPE) which hinder or discriminate against agriculture. The classification and
criteria used are given in Annex C.
Differential performance between groups. The comparisons in table 8 indicate that since the early
1980s the countries with favorable policies have performed better than those with unfavorable
policies. This has been true in agriculture, and in terms of overall economic growth as weLL
Between 1982 and 1988, agricultural exports and agricultural value-added rose in FPE countries
4.15 percent and 3.50 percent per year, respectively,while both of these indicators declined in UPE
countries. Total agricultural production grew nearly three times as fast in FPE countries, and food
production grew nearly 4 times as fast, as was the case in the UPE countries. Given the
importance of agriculture and its linkages to other sectors, it is not surprising that overall GDP
also grew faster in FPE countries -- at 3.25 percent per year -- over the 1982-88 period. Since
Table &8 Comparison of gerformance and policy amone African county erouns.
Countries with . favorable policies
growth rate
average
average
198248
198548
198082
Economic Performance
Volume of agricultural exports a/cf
Total agricultural production a/
Food production
Agricultural Value-added a/c/
Gross domestic product a/
Polc Performance
Real effective exchange rate b/
Real producer prices for exports a/cl
Nominal protection coefficient
(export commodities) c/
Real protection coefficient
(export commodities) c/
Food Sector
Real producer food price cd
Nominal protection coefficient
(tradeable foods) d/
Real protection coefficient
(tadeable foods) d/
Food imports cl
ogenous Factors
Index of export prices a/
Weather effect
a/
b/
c/
d/
Countries with unfavorable policies
growth rate
average
average
198248
1985488
1980482
99.1
104.2
103.9
101.5
103.9
121.5
118.3
116.5
119.7
125.6
4.15
2.55
2.33
3.50
3.25
101.8
102.9
103.3
101.7
104.1
102.8
109.1
107.7
99.4
118.4
-1.80
0.89
0.62
-0.91
2.14
133.0
94.3
95.6
104.9
-9.4
4.37
168.2
95.8
112.4
96.5
-5.3
0.85
0.84
1.14
0.96
0.99
0.65
1.12
0.75
0.97
98.4
100.6
2.9
1A
1.18
106
1.97
133
99.1
-0.06
87.1
0.05
104
5.5
108
0.74
1.27
0.63
112
1.03
160
97.8
0.02
89.8
-0.05
Index, 1979-81 = 100.
Index, 1970-75 = 100.
Period is 1985-87 only, no 1988 data are available.
Periods are 1981-83,and 1987 only. Reduced samples are 6 and 4 for FPE and UPE, respectively.
4.5
- 35 -
1984, growthin the FPE countrieshas exceededpopulationgrowthior all the performancevariables
mentionedabove. The divergenttrends are illustratedin figures 11-14. 24/
The policyvariablesin table 8 mirror the way the countrieswere classified: PPE countrieshave, on
average, reducedreal effectiveexchangerates rapidlysince 1982, nearly twiceas fast as the UPE countries,
and to a level lower than during the early 1970swhereasfor the UPE countries' the REERremains 12
percenthigher thanin the early 1970s(figure 15). 2S/
With respectto direct policiesaffectingagriculturalincentives,the averagenominalprotection
coefficientfor major exportcrops in FPE countriesrose to 1.14 by 1985-87from a level of 0.84 in the early
1980s. In UPP.countriesthe NPC rose as well, but by a lesser amount,to 0.99. Sincethe NPC does not
accountfor the indirecteffects of exchangerate disturtions,the real protectioncoefficients(RPCs)are
comparedbetweenthe two groups. They indicatethat in the early 1980s,producer. in FPE countrieswere
facedhigher rates of implicittaxationthan did those in UPE countries. Sincethat time the pattern has
shiftedin favor of producersin FPE countriesas the RPCs for that group rose 70 percent. 2AI
For the FPE group, the level of RPC in 1985-87indicatessomedegree of subsidizationto
producers. Indeed, amongCFA countries,governmentshavebeen unwillingto reduce producerprices in
the face of the declinein the French franc denominatedpricesof their major commodityexports. This has
for exampleresultedin an RPC of 2.56 for Senegal,and 1.15 for Cote D'Ivoire.
Exogenousfactorsand shocks. Exogenousfactorssuch as better weatherpatternsor higher demandfor a
country's exports couldpotentiallyexplainsome or all of the differentialperformanceobservedhere. To
assess that possibilityboth a weathervariable,and exportprices are comparedfor the two groupsof
countries. In the case of exportprices, the data suggestthat exportprices facingboth groupsof countries
have fallen by similar magnitudes.
To assess the impactof rainfall, the weathervariabledefinedin Section2 is used: deviationfrom
trendin cerealyields. Those residuals(in logarithms)are comparedin table 1 for the two groupsof
countries. They indicatethat, on average, FPE countriesdid indeedreceivemore favorableweatherduring
1985-87,but faired somewhatworse in the early 1980s(figure 16).
Can the differencesin performancebetweenthe two groupsbe explainedby the better weather
experiencedby the FPE countries? This quesdoncanbe answeredin a precise way. Sincethe
relationshipbetweenthe weathervariableand exportswas esimated in table 1, it was also
esimated with respectto total productionand food. Basedon the esdmatedcoefficientsof the
24/ Performancevaried witin the two groups. Not all FPE countriesperformedbetter than all
UPE countries,but in the case of agriculturalvalue added,for example,of the 12 countrieswith
growthrates above2 percent, 9 were FPE countries;of the 11 countrieswith growthrates below
2 percent9 were UPE countries. For agriculturalexports the dispersionoverlappedmore; in
fact Senegal,a FPE country,had the worst performancewith -12 percent growthfor 1982-87.
2JI This improvementfor FPE countriesoccurredin spiteof an increasein REERsby CFA
countriesof about 1 percentper year from 1982-87.
2I The actuallevel of the "real" protectioncoefficientmaybe misleadingif the real exchange
rate was distortedin the base year. Attention,therefore, shouldbe focusedon the changeover
time and differencebetweenthe NPC and the "real"protectioncoefficient.
-36-
Figure 11,Africa's agricultural exports
Comparing countries grouped by favorable
and unfavorable policy environments
Index 1979-81 = 100
1 70
150
130
110
90
70 1 '
I '
l
I
1970 1972 1974 1976
1978
-FPE
1980
1982
1984
1986
-UPE
Figure 12. Total agricultural production
Comparing countries grouped by favorable
and unfavorable policy environments
125
Index 1979-81 = 100
115-
105 -
95
1970
1972
1974
1976 1978 1980 1982 1984
FAOproduction index numbers
FPE
-UPE
1986
1988
-
37 -
Figure 13. Total food production
Comparing countries grouped by favorable
and unfavorable policy environments
Index 1979-81D
120
100
110p
100
90
1970
1972
1974
1976 1978 1980 1982 1984
FAO production index numbers
FPE
1986
UPE
Figure 14. Africa's GDP
Comparing countries grouped by favorable
and unfavorable policy environments
Index 1979-81 = 100
130
120 110 100
90
80
70
1970
1972
1974
1976
-FPE
1978
1980
-UPE
1982
1984
1986
1988
-38-
Figure 15. Real effective exchange rates
Comparing countries grouped by favorable
and unfavorable policy environments
Index 1970-75 100
200
180
160
140120 100
80
,o
1970
II
1972
I
1974
1976
1978
- FPE
1980
1982
1984
1986
1988
-UPE
Figure 16. Weather patterns in Africa
Comparing countries grouped by favorable
and unfavorable policy environments
0.15 0.1
F
0.05
-0.05 -0. 1
-0.15-0.2
-0.25-
.
1980
.
I
.
1981 1982 1983 1984 1985 1986
deviations from trend in cereal yield
FPE
3UPE
1987
-
39 -
regressionsin table 1, only aboutone-thirdof the differencesin exportgrowthbetweenthe two groups or
countriescan be attributedto differencesin rainfall.27/
The comparisonsin table 9 also indicatethat both food and total agriculturalproductiongrew faster
since 1982in countrieswith favorableagriculturalpolicies- with food productiongrowingnearly four times
as fast in FPE countriesas in UPE countries. Again,based on the coefficientsestimatedfor their response
to better weather,only aboutone-thirdof the differencecan be attributedto the differentweatherpatterns.
It is noteworthythat for those countrieswhere producerfood price data are available(excluding
officialprices where they are not enforced)therehas been very little changein the real prices paid to
producersfor food crops. Again the earlier results are confirmed:improvedprices for exportcrops tend to
benefitfood productionas well.
County diversity:policy and response
While the patternsin many individualcountriessupport the aggregateresults and grouped
comparisonsabove, performanceand responsivenessvariesacross countrieswithinthe two groups. Aside
from differentiationbetweenannualand tree crops, two major reasonscan accountfor differing
responsivenessto price signals: first, if price changesare viewedas temporaryfluctuationsrather than
permanentshiftsin the price policyregime, farmerswill not respond; second,nonpriceconstraintsmay
obstructproducersfrom respondingto incentives.
The experienceof severalcountriesprovideseviderce of both responsiveand nonresponsive
policies can
agriculture. Ghana is a strikingexampleof both how deleteriousthe effectsof bad economnic
effect for agricultureand for the economy
be, and how changingthose poor policiescan have a dramau^-.
overall. GDP has grownby 6 percentper year since 1983,led by rhe agriculural sector. Agricultural
exports(essentiallycocoa), whichhad been fallingsince the 1960s,have finallyrebounded,3-4 years after
real producerprices were nearly doubled(figure 17).
There are many reasonsfor the inpressive responseof Ghana's economy. Includedamongtheseare
the return of expelledworkersfrom Nigeriaat the time of the reformsbringingadditionallabor to the rural
sector, high levels of internationalassistancewhichenabledthe governmentto perseverein the difficult
adjustmentprograms, and the ostensibleperceptionthat the reforms representeda permanentchangein
Ghana's policyregime. A surveycarried out in early 1988showeda substantialamountof new cocoa
plantingwas takingplace, an investmentwhosereturn will dependon producerprices at least four years
hence. The credibilityof the changein price level had been perceived(accuratelythus far) to be a
permanentchangeand thus led to revisionof the prices exocte by farmersin the future, not just current
prices.
In addition,though, despitea long periodof deterioratingeconomicconditions,the physicaland
institutionalinfrastructureneededto servicethe cocoasectorhad remainedreasonablyintact.
22/ Usingthe estimateof the exportsupplyresponseto changesin the weathervariable,only 4
percentagepoints of the 22 percent growthover the period can be explainedby more favorable
weather. Similarly,for the UPE group whichexperiencepoor weatherduring part of the
period, their exportscouldhave been 3 percentagepointsbetter than the 1 percentrise between
the two periodsif weatherhad remainedfavorablein the secondperiod. Therefore,7 of the 21
percentagepoint differencein growthbetweenthe two groupsis accountedfor by different
weather. (Similarresults are obtainedby comparingthe differences:n total productionand food
production.)
-40 .
Figure 17. Ghana's price incentives and
agricultural exports
130
Index 1970 = 100
110
90
7050300I
1970
1972
,
1974
I
1976
I
1978
- producer price
1980
-
1982
1984
1986
1988
export volume
Figure 18. Zaire's price incentives and
agricultural exports
Index 1970 = 100
17014512095
70-
4520
1970
1974
1972
-
1976
1980
1978
producer price
-
1982
1984
export volume
1988
- 41 Some roads are in need of repair, but for the mostpart the capacityexistedto supportthe response
in the sector. Thesenonpricefactorswere, in Ghana's case, not as limitingto agriculturalresponse
as may be the case elsewhere. Togoalso providesan examplewhere agriculturehas been responsive
to incentives(figure20).
By contrast, Zaire is a case where policyreform led to strong policychanges,but had little effect
on production. The improvementin incentives,however, was shortlived(figure 18), and to the
extent that producers(correctly)viewedthe changetransitoryrather thana credibleshift in policy
regime, they were unresponsiveto the change. The potendalresponsivenessof agriculturein Zaire
is less clear for other reasons. The interior transportnetwork,as wellas supportservicesto
agriculturewere never as well developedas in Ghana, and theyhave deterioratedconsiderably. The
loss of this infrastructure,combinedwith the policiesthat weighedheavily on taxationof agriculture,
resultedin a degenerationto essentiallysubsistenceagriculture. Recentanalysisof the situationin
Zaire refers to a labor shortagein the rural areas due to low returns to agriculture,lack of consumer
goods, and cheap food policiesfavoringthe urban areas: quite simplyagriculturewill not respondto
price incentivesif few peoplelive in rural areas!
Similarto Zaire in many respects,Tanzaniaprovidesanotherexampleof disappointingresponse
to policy reforms. While the policydistortionsover the past 20 years havebeen very severe
(Lofchie),recent policyreformsin Tanzaniahave resultedin little responsefrom agriculture. There
appearto be two main reasonsfor this. First, the reformshaveyet to result in sustained
improvementsin the incentivesfacingproducers. Exchangerate distortionsremain, and real
producerprices have not improved(figure 19). For investmentto occurin agriculture,prices
mMr&k4in the futuremust be revised. If even crrnt priceshave not risen, it is unlikelyto change
behavior. Secondly,the physicalinfrastructurehad decayedsubstantiallyin Tanzania,leavingmany
producers, or potentialproducers,withouteasy accessto the marketsthat are intendedto foster
incentives.
Many observershave remarkedat the strikingdifferencesin performancebetweenKenyaand
Tanzania. The differencesare clearly relatedto differencesin policy. Kenya's coffeeand tea
growershave enjoyedhigh percentagesof their crops internationalprice (NPCshave averaged0.9
for both), althoughthis leavesfarmersvulnerableto wideyear-to-yearswingsin th"prices they
receive. Nevertheless,there has been steadygrowthin exportvolumesof both crops.
Landpolicy in Kenya, and the Ujaamapolicyin Tanzaniaare just two additionalfactorsthat
enter into understandingthe sourcesof the two countriesdifferences.The complexitiesof the
differencesand similaritiesof these two countriesare, however.impossibleto summarize
briefly. 21/ One key difference,though, is the identificationof key constituenciesfor policy
making:Kenya's farmer elites, and Tanzania'surban elite. In Kenya, the elitehas financially
investedin rural areas, oftenderivingan importantportionof its incomefrom agriculture. Their
influencehas undoubtedlybeen a factor in the consistentpolicyenvironmentthat has favored
agriculture.
By contrast,the Tanzaniangovernmentnationalizedthe land, makingit virtuallyimpossiblefor
agriculturalsystemsbased on private ownershipto develop,and the government'sleadershipcode
21/ See MichaelLofchie, 1989.The PolicyFactor: Agricultml Performancein Kenya and
Tanzaniafor a recentanalysisof the economicand politicalissues.
- 42 -
Figure 19. Tanzania's price incentives
and agricultural exports
Index 1970
100
=
120
100
80
4020
0
o
1970
1I
1972
I
1 I
I
I
I
I
1974
1976
1978
1980
1982
1984
producer price
-
I
1986
export volume
Figure 20. Togo's price incentives an,
agricultural exports
Index 1970
100
=
120-
100
80
60
40
1970
1972
1974
-
1976
1978
producer price
1980
-
1982
export volume
1984
1986
-43
-
"framed and implemen-e'S by politicians whose political and economic base was exclusively
urban. Tanzania's agn -Itural policies thus provide one of Africa's clearest examples of a
system designed to trans ;r economic resources from the countryside to the city (Lofchie, p.
191).
Indeed, many of the African countries where agriculture has been favored and as a result
performed well, are countries where influential elites have vested interests in the agricultural sector:
Kenya, Malawi, and Zimbabwe, are all countries with large-scale farms owned by influential
citizens. By contrast, in the majority of African countries where the sector is made up almost
entirely of smallholder farmers with little or no political voice, agriculture has been disadvantaged
in favor of the influential urban class.
This adds a confounding factor to the arguments in favor of a broadly-based smallholder
development strategy versus a "bimodal" strategy including larger, estate type farms. While the
economic arguments for smallholder-based strategies are persuasive (Johnston 1982), one sees in
recent experience that they have tended to be politically disadvantaged and suffered as a result. By
contrast, where there is an elite with vested interest in agriculture there appears to have been a
powerful voice supporting the interest of farmers. Does the importance of assuring a political
lobby outweigh the economic arguments for unimodal development? Have the large-scale
landholders lobbied for their interests only, at the expense of smaUholders? In some cases, there
has been a discriminatory influence that has protected the interests of the large-scale farmers,
sometimes at the expense of the smallholder, as in the cases of Malawi and Zimbabwe.
V. CONCLUDING COMMENTS
Summary of Empirical Findings
The analysis above supports the following conclusions:
1.
The decline in Africa's agricultural exports production between 1970 and the early1980s coincided with substantial and widespread macroeconomic policy distortions and
deteriorating real prices paid to agricultural producers. Detailed econometric analysis indicates
that agricultural exports have, overall, been depressed by direct and indirect price distortions.
The effects have been substantial; aggregate patterns indicate that as both real producer prices
and real exchtangerates deteriorated by about 20 percent on average during the 1970s,
agricultural exports declined by a similar percentage.
2.
Beginning in the early 1980s, policy reforms and macroeconomic adjustments have
reduced the degree of direct and indirect policy distortion giving rise to improvements in
overall agricultural export performance. The estimates of supply responsiveness confirm the
relationship between improved policies and better growth. Moreover, countries that have
adopted or maintained favorable policy environments since 1984 have seen higher agricultural
production, exports, and overall economic growth than countries with less favorable policies.
3.
The degree of policy distortion in Africa differs from other developing countries
primarily with respect to exchange rate distortions rather than in direct pricing policy. The
degree of direct pricing policy intervention, as measured by the rates of nominal protection,
have been significant in many African countries, but they have not been significantly larger than
in other developing countries. The degree of ex.changerate overvaluation, however, has been
- 44 larger, resulting in a loss of competitiveness and leading to inefficiencies,including those arising
from the exchange and trade controls that usually accompany overvaluathon.
4.
The response of export agriculture to policy changes has not come by way of shifting
resources out of food production. Econometric analysis indicate that food production, like
exports, responds positively to improvements in real exchange rates. Moreover, food production
correlates positively with higher producer prices for export crops, suggesting either that they are
complements in production, that policies favorable to export agriculture also favor food
producers, or that a third, omitted, variable is correlfr- 4 with both improved policy and output.
5.
Urban migration has been both a symptom of policy distortions and a cause of poor
agricultural performance. In many African countries, macroeconomic imbalances, declining real
producer prices, "cheap food policies", and other policies favoring urban dwellers have shifted
the rural-urban terms of trade significantlyagainst the rural sector. This has led to a shift of
productive resources out of agriculture and has resulted in high rates of urban migration.
Policy reforms that realign the internal terms of trade can slow, or reverse, these resource
flows. This has been documented in Ghana wh-re substantial reverse migration to the
agricultural secto. has been occurring since the implementation of their recovery program.
6.
Africa's food imports doubled during the late 1970s, making Africa a net food impolter
for the first time. In the aggregate, however, this 7 million metric ton increase in annual
imports is almost entirely accounted for by two factors: a) seventy percent of the increase is
due to a quadrupling of Nigeria's food imports during the oil boom when its "capacity to
import" tripled; and b) food aid to Africa increased from annual levels of 1 million metric tons
per year to around 4 million metric tons, accounting for most of the remaining increase.
7.
At the country level, rising food imports have not resulted from a failure of African
agriculture (yields per hectare have, in general, been rising). Slow growth in domestic food
production in Africa is primarily the result of urban migration, the availability of cheap food
imports, and a tendency among urban dwellers to consume imported rice and wheat. This has
resulted in a reduction in the demand for domestically produced foods, which, along with the
outmigratiun of agricultural labor, has led to supply and demand for domestic production
growing more slowly than population. Econometric analysis indicates that the most important
factors explaining rising food imports are: a) chiangesin the strcture of demand (due to
demographic and income changes related to urbanization), b) changes in the ability to pay
(foreign exchange earnings or "capacity to import"), and c) relative prices (due to the
overvaluation of c'irrencies which have made imported foods relatively cheap). Most of the
variation in food imports can be explained by these factors. The remaining trend in food
imports is only about a 1 percent annual increase, which is likely the result of the substantial
drop in international rice and wheat prices between 1970-86.
Taken together, these conclusions place much of the blame for the deterioration of African
agriculture on government policies which have shifted the internal terms of trade strongly against
agriculture and created market distortions that reduced efficiency. The result was a shift of
resources -- especially labor -- out of the sector, and a decline in both private and public
investment.
-
45 -
Forward Looking Issues
The prescription that follows from these findings should not be one of limiting the focus of
attention to price policy. But rather it should be clear that the elimination of policy distortions
sets the stage for getting on with strengthening the real sources of long-term agricultural growth:
productive investment and technological change. In addition to investments by private agents,
public investments in physical, human and institutional capital are essential such as transport
infrastructure, support services, and agricultural research. All of these seek to raise productivity
and reduce the cost of production to make agriculture more competitive.
In specific countries, prices may appear to be less important than nonprice factors; where price
policies have not led to major distortions, attention should rightly be placed on promoting
productive private and public investments. Where price distortions have been large and as a result
agriculturai investments neglected, both price and nonprice constraints are likely to be binding, but
romoving price distortiens should be seen as a prerequisite to encouraging appropriate investments
in the sector.
The debate over the relative importance of price and nonprice factors is to some degree
misguided, or at least misleading. There are three reasons for this. First, nonprice constraints on
the responsiveness of agriculture may be interpretable as being price-related (i.e., the lack of roads
is a transport cost which will reduce the farmgate price; the absence of extension services raises
information costs making technology prohibitively expensive to most farmers). Second, the relative
importance of these two categories clearly differs across countries. Therefore, the debate should
rightly take place only within a specific country context.
And third, nonprice constraints in many cases may be a reflection of the very slow speed of
adjustment for investments, maintenance, migration, and the attention of government to provide
supporting services, in response to potentiaUy profitable agricultural activities. In the absence of
price incentives, it will do little good to relieve nonprice constraints. In fact, it should be difficult
to identify them without proper price signals (in the extreme, it does little good to build roads to
rural areas if nobody lives there anymore).
In many cases constraints exist which inhibit a swift response of farmers to changes in prices.
But bad roads, lack of irrigated land, and nonexistent input suppliers can be seen as the result of
neglected investments in the face of unprofitable choices. The point here mirrors the dicotomy
between short-run and (very) long-run supply response. In the same way that planting cocoa trees
is a long-run investment decision by a farmer in response to expectations of future profitability,
public funds for roads, storage facilities, or establishing institutional support reflect public sector
responses to future returns on investment. In this way, the very long-run supply response includes
"induced institutional change." Indeed, it should be clear that the highly developed physical and
institutional infrastructure of Kenya's coffee sector would not have developed if pricing and
exchange rate policies had been strongly adverse. The iasults presented here makc a case for the
importance of reducing price policy distortions. Viewed in this way, it should be seen as a
necessary but not sufficient condition for agricultural growth.
-46ANNEX A - DATAAND METHODOLOGY
Time-series data for official and market prices were compiled from a variety of sources
including World Bank and IMF documents, consultant's reports, the FAO, the USDA, the
International Federation of Agricultural Producers, and direct contacts with World Bank Resident
Missions and country economists. 22/
Producer and consumer price data have been compiled for the most important agricultural
commodities in each of 35 countries; for up to four majot export commodities, up to three traded
food crops, and up to two nontraded staples.
Data for agricultural production and trade are from FAO. National accounts data are from the
World Bank data files and IMP. Transport costs per kilometer-ton have been estimated based on
the large number of existing detailed studies, and are applied to distances from the major
producing and consuming region for each commodity via the most commonly used mode.
Historical transport cost data, where unavailable, are extrapolated on the basis of the domestic CPL
Processing costs have been estimated similarly. Ocean freight costs are estimated based on World
Bank figures.
The critical distinction between official prices and open market or parallel market prices has
been maintained throughout. Where the available information indicates that most farmers receive
the official price for their product, that price is also assumed to ba the market price. '.Vhere data
on open or parallel market prices are available these prices are used as being more representative
of what producers receive. Averages across countries for export or food crops are taken as
weighted averages (by production value) within countries, and simple averages across country
groups. 2Q/
For comparative analysis between country groups, group averages are simple means. Weighted
averages would be inappropriate since the objective is to illuminate a central tendency among
countries in the group, rather than to assess their aggregate magnitude.
The producer price data assembled are deflated by the domestic CPI for up to four major
export commodities in each country. An average index of export producer prices is computed
weighing each commodity's price by its relative value of total production.
The following potential data problems should be borne in mind when interpreting the results of
this study:
1.
Production data and, to a lesser extent, export data, may be inaccurate. Export data be
29/ In general, the producer price data reflect the average annual price received by farmers in the
most important producing regions, although due to the wide regional and seasonal variations and
the lack of specifics from some sources, this level of precision cannot be assured. Data series from
different sources have sometimes been spliced when they seemed reasonably consistent.
L/
Detailed data on producer prices was compiled for 35 countries, excluding very small countries
(Comoros, Seychelles,Cape Verde, Sao Tome, Lesotho, Swaziland), ones where no data were
available (Angola), or where agriculture is of minor importance in the overall e.nmomy (Gabon,
Djibouti).
l!
iI|
1
-
47 -
biased when shifts in the share of unreported or smuggled exports changed. This could have
resulted in an positive bias in the results above.
The producer price data have been carefully compiled to distinguish between official
2.
prices and actual producer prices where divergences were apparent. Nevertheless, 'he prices
used are for most countries officially reported prices and in some cases will not accurately
reflect what farmers are being paid.
The index of producer prices computed is based on from 1 to 4 export crops per
3.
country. For countries where the composition of exports includes significant other crops, the
price incentives are omitted from the analysis, while their export volumes are included.
Changes in input prices have been ignored. To the extent that producers are
4.
responding to changes in farm profits, changes in fertilizer and other prices should be taken
into account. Typically,higher output prices and devaluation have been associated with raising
the prices for imported inputs such as fertilizer. To the extent that this phenomenon has been
omitted from the above analysis, the elasticities with respect to price and exchange rate have
been biased downward.
There are simultaneity problems for countries where their exports account for a sizable
5.
hare of inteational nmarketsanid face less than perfectly elastic demand (Ghana, Cote d'Ivoire,
Nigeria, Kenya, Cameroon). This problem could bias the results downward. For countries
where the farm price reflects the highly variable international price (coffee in Kenya and
Ethiopia, tobacco in Malawi), occasional sharp increases in price will not affect current
production decisions, and will be viewed and temporary in nature. This may partly explain the
low, or negative, results for these three countries.
-48 ANNEX B - THE EXTENT OF CHRONIC HUNGER IN AFRICA
At the heart of any analysis of food policy in Africa is a concern over the welfare of African
people who live in a harsh environment, where rainfall is erratic and where several regions have
been prone to drought, especially in the past 15 years. Africa's famines in drought or war prone
areas such as Ethiopia, Sudan, Mozambique, and the Sahel, require emergency respoases and better
global preparedness for the future. But these emergency situations, now referred to as "transitory"
food insecurity (World Bank 1986a) are different from the issue of chronic hunger or food
insecurity, where people suffer from a continuously inadequate diet caused by inability to acquire
food.
There are, however, relatively few reliable statistics with which to asses the prevalence and
magnitude of chronic hunger in Africa. In Africa and in other developing countries food
availabilities are estimated using food balance sheet techniques which estimate availability as a
residual, introducing errors which are almost invariably in the direction of understatement. In a
thorough analysis of the statistics and methods quantifying nutritional situations, Poleman
concluded that
"With food availability estimates .hat understated set against requirement figures that
overstated, the cards were so stacked that almost all LDCs could be classified as "diet
deficit." Redone with truly accurate information, it is probable that few countries would he
so classified. Much as the protein gap proved a statistical illusion, the list of diet-deficient
countries would be whittled away." (Poleman, p.15)
Given the weaknesses in the data, and their likely biases, the case being made about Africa's
chronic food problem, that "averagedaily caloric intake was 96 percent of requirements in the
1980s"seems a remarkably weak basis on which to assert that the "inadequacy of food manifests
itself both at a national level and at the household and individual lever (Serageldin).
The principal evidence that may indicate widespread undernutrition comes from dietary and
anthropometric studies showing that a large share of children in Sub-Saharan Africa are stunted.
Stunting, however, can result from several causes, including diseases which lead to reduced intake
of food and impair the metatolism and absorptive capacity of the individual. And there is a
growing consensus among nutritionists that both disease and lidernutrition can result in stunting,
as weU as child morbidity and mortality. In fact, over the past 20 years the infant and child
mortality rates have not dropped more in African countries where aggregate food supplies have
increased, than in the countries where food supplies have declined (Svedberg).
In a systematic assessment of available evidence Svedberg (1987) concludes that "there is no
firm evidence of widespread and severe undernutrition in the African population at large (p. 87).
But certain patterns observable in Africa seems at odds with tlue oonclusion that "about a
quarter of Africa's population ... do not consume enough food to alowvan active working life
(World Bank 1988a). The high rate and continuation of rural-urban migration, the growing
importation of rice and wheat that in many countries is more costly that domestic cereals, and
depressed market prices following abundant harvests suggesting market saturation, are difficult to
reconcile with the assertion that in many of these countries 30 to 40 percent of the population are
chronicaly "food insecure."
However, even with adequate "average" food availabilities, chronic hunger may persist if poverty,
income distribution, and other factor prevent certain groups from obtaining enough food. These
- 49 -
are issues that need more careful unalysis and attention. The evaluations by Poleman and Svedberg
suggest that available data overstate the extent of chronic undernutrition, and that reports including
strong statementt about the severity of the problem have no reliable empirical basis. Clearly, what
is needed is better data and more focused analysis on the extent of the problem and its causes.
-
50
-
ANNEX C - CLASSIFICATIONOF COUNTRIESBY POLICY ENVIRONMENT
For comparative purposes countries in Sub-Saharan Africa have been classified into two groups
based on the policy environment that existed in the mid- to late-1980s. The principal
considerations for classification of countries are the direct policies that affect agricultural incentives
(the real producer prices, levels of direct taxation i.e. NPC), and indirect policy measures which
affect the competitiveness of the agricultural sector vis-a-vis other sectors in the economy and in
international markets (real exchange rate).
On the basis of criteria for these key policy variables the countries were classified as shown in
the table below as having a Favorable Policy Environment (FPE) or as having an Unfavorable
Policy Environment (UPE). This classification differs in several respects from that used recently in
(Africa's Adjustment and Growth in the 1980s (World Bank 1989) to compare "strong reforming"
and "weak reforming" countries. Policy reform programs differ among countries and, given the
kinds of policy changes undertaken they may or may not be expected to have any relatively shortrun impact on economic performance overall, or agricultural performance in particular. For
example, reforms in public enterprises, govemment revenue collection, or reduction of government
payrolls are all measures that at least partly share the goal of alleviating fiscal imbalances, to
reduce deficit financing through money creation that is inflationary, and thereby lead to lower
inflation, positive real interest rates, and better allocation of investments to promote efficient
growth. These measures have longer-term objectives and therefore cannot be expected to affect
productivity in the economy or in the agricultural sector over a five year period.
In addition, the categories of "strong" and "weak"reforming countries rely on formal agreements
with Bretton Woods institutions as the criteria for being considered "strong reforming", rather than
by having these reform programs manifested in objective policy performance indicators such as the
real exchange rate or level of agricultural taxation. As a result of this different approach to
classification, several differences arise in the grouping of countries.
First, the PPE category includes countries which have adopted or mantaned a favorable policy
environment, thereby including countries which have not undergone a stmctured reform program,
but where on the basis of these policy performance measures, the need for such a program - or
the detrimental effect of their policies on productivity - is less apparent (Benin, Burkina, and
Cameroon fit this group).
Second, several countries considered to be "strong reformers", on the basis of the policy
performance variables examined here, cannot be considered to provide a favorable policy
environment, because exchange rate misalignment persists. As a result, Mauritania and Tanzania
are considered to provide unfavorable policy environmtnts due to their exchange rate policy which
continue to impose heavy indirect taxation on the agricultural sectors, and to make them less
competitive with respect to other sectors and in international markets.
Several countries classified as "strong reforming" countries are here also considered to have
favorable policy environments, but because the improvements in their policy performance variables
are so recent (in 1986 or 1987), they are excluded from the comparisons of performance, since it is
too soon to expect any significant response, especially in countries where agricultural exportables
are tree crops where production response can take four or five years. These countries are Burundi,
Ghana, Guinea, Madagascar, and Nigeria. Malawi is also a recent reformer, but because its
agricultural exports are annual crops, a more immediate response can be expected.
-
51
Several countries are excluded from considerations due to lack of data, small size, or volatile
swings in agricultural exports related to policies and cross-border smuggling of neighboring
countries (Benin, Zambia). The resulting classification of FPE and UPE includes 12 countries in
each group. 31/
/f Angola, Cape Verde, Chad, Comoros, Djibouti, Gabon, Guinea-Bissau, Lesotho, Mozambique,
Sao Tome, Seychelles, and Swaziland have been excluded due to their small size or lack of data.
The countries excluded from comparisons of performance for other reasons are Burundi, Ghana,
Guinea, Madagascar, and Nigeria.
Annex Tabb Cl.
Clelficaton
of Countries by Poliy Environment.
Pa
O01a
Caflcbatlcn
Averag
198486
1987
NOmn potweon coafficmt
for majo szprtad crow
Ratio of pfeel ma,kat to
hemga ta
otffIl
afectlw exchange rts
11970.73) * 100)
Crdtaron
(below 120
Inid-l9108a
Citwion
tbelow 1.3)
Averge
19848
1987
1.49
1.02
1.29
1.01
UPE
WE
1.11
1.01
1.01
WE
WE
Averag
198446
1987
0.78d
1.18
Criterion
Wio2 o.n
mPP
Control of
markating bv
W
BEoin
FPW
Eo_wan
Bunkih
UPE
WPE
NA
82
NA
78
Euruni
Canroon
CA.R.
FPE'
WE
UPS
141
114
102
II
139
107
FPE'
WE
1.31
1.02
1.02
Chnd
Congo
Coa dlve r
FPE\b
WE
P
WE
93
105
95
127
FWE
WE
1.1
1.02
1.02
1.02
1.01
1.01
FPE
FPE
WE
0.79
1ie
FWE
yW
Etopi
Gambe
Ghan
UPE
FPE
FWE'
162
102
112
11l
89
S0
UPE
WE
WPE'
2.04
1.17
2A1
221
1.09
1.37
UPE
WE
UPE
0.70
1.11
0.96
0.77
1.92
0.52
FPE
WE
FWE
per
paW
VW
uinze
Klnvy
Lbera
FPE'
FWE
UPE
90
124
74
10
FWE
UPE
9.26
1.08
1.Seelnota)
1.04
1.06
FE'
WE
1.79
0.82
OIS
NA
0.70
1.03
WE'
FPE
FPE
no
no
VW
Medagasa
Melawi
Mol
FPE'
FWE'
UPE
90
96
122
el
S2
11
FWE
WE
UPE
1.26
1.37
1.02
0.79
1.13
1.01
WE'
UPE
WE
0.58
0.63
0.83
0.70
0.68
OS
WE*
WE'
UPS
PE
pWl
Pser
Munbtan
Maurtls
Mozaq6u
UPWE
FWE
UExt
10
95
92
94
WE
FPE
2.32
2.30
UPw
12.0
UPE
Nigt
Nigi
Rwanda
FPE
FPE'
UPWE
93
249
66
78
E0
155
WE
WE'
UPE
1.02
3.94
1.A
1.01
3.32
1.27
WE
UPE
UPE
1.67
1.03
.77
4.02
0.74
0.97
WE
FPE
WE
no
no
VW
Snegal
Siera Leons
Snoml
FPE
UPES"
UPE
10
139
198
108
91
95
FPE
UPE"
UPW
1.02
1.62
2.07
1.01
8.13
1.43
FPE
UPE
UPE
1.12
1.14
1.0
2.62
2.07
1.386
FPE
FPE
WE
We
no
pati
Sudan
TUnlass
Teo
UPE
UPE
FPE
97
187
91
E1B
74
91
WE
FPE'
FPE
1.97
4.15
1.02
1.84
2.72
l.01
UPE
UPE
FPE
1J0
0.94
0.62
1.36
EP
1.16
FPE
FPE
FPE
pati
VW
pal
Ugand
Zaid
Zambia
Zinbtbwe
UPE
FPE
UPE I
UPE
140
79
62
72
190
66
37
83
UPE
FPE
FPE
FPE
1.99
1.04
1.61
1.70
1.8
UPE
WE
UPE
UPE
0.5i
0.57
0.54
1.5
1.09
UPSE
UPE
WE
FWE
V
W
we
no
WPE
mE
42.6
0.80
1. 9
1.59
UPE * unfavorbl Polcty evimnt.
Note: FPE - favoabe poklicy nvIrnmn
ccmpaIun becae of volatie eoxpt relatd to neighboFn commW polce.
a ESkdad fom perfoman
dt n.
%b E=kWed due to lok of dta and adve politica
, no ppl on
e
pt
e theit
In
adopting fevoabl polciery
Indiotecornla
in ti
a
tbe
copp a involved (Mal
Wnua
Whem
epoot.
agrclra
conlPlng
of
Thu.. tOa hey. bean *luded fot pumpooe
in 198was suspndd in 1988.
mb
tn_ p
L- n st
Sir
'
mE'
a be oeo
for
sp
0.66
0.92
FPE
no
W
0.69
0.90
0.78
0.79
1.38
0.99
WE
WE
FPE
we
w"
prtl
VW
-
W
-
O06
0.89
1.11
in dt det
s anases.
t
t
poiy ch
.
uJ
.53 -
|~
~~~~a
_
ad d
_d
d
d
d
-
d
d
d
ddf
d
d
d
d
dd
d
d
C S6d
a
d
a
-
d
od
d
d
d
d
dd d
d
d
l fi
_
l
-
d
dddddddddd_
di
d
ad
ddd
r
f0_0td
d
d
d
d_dz
d
d
d
5f
d
dd~~~~4
d
d
d
d
d
5d
d
-d
2-^
d
d
d
d
-
-
-
d
-~~~~~~~~~~~~~~~~~
ddIidIii d doii
iiiI
ed° iiIdd
iddIdidd
-
-
-
-
-
-…-
-
-d
-
.
0
-
CI
pmwteon coefficienit
for principalexpoft commodite.,
AnnexD Table 2. NomninG
Oqmy
iuznFPso
op
1970
1971
1972
1973
1974
1975
1978
1977
1978
1979
19s8
1961
1982
1983
1984
1985
1988
1087
Coro
0.65
0.64
0.49
0.39
0.71
0.43
.40
0.87
0.84
0.59
0.52
0.55
O4A
o0.38
059
124
0.88
0.92
oand
Cofee
.71
0.i1
0.52
0.57
0.89
024
0.2
0.53
0.51
Qs2
0.79
0.73
0.70
0.54
Q58
0.45
098
0.79
Cam
Coff
1.28
129
1.29
1.13
1.34
9ss
0.39
087
0.91
1.13
1.52
1.10
0.12
0S1
0.98
098
g1.80
1.87
CAR
Colon
064
0.Q8
0.51
0.3
085
0.50
0.52
0.88
0.84
0.50
0.6
0.0
0.
0.48
0.71
1.55
1.01
Os9
COo
Ccoa
0.73
.65
0.40
030
.Q44
024
0.17
023
027
040
041
0.44
0.32
0.25
029
0.43
0.63
0.73
COWdvc8
CclIe
1.11
1.09
1.08
1.0S
1.48
0.87
0.23
0.84
0.77
1.09
1.44
1.09
0.80
0,72
0.8
0.88
1.05
1.98
0.5
s8
0.77
2.00
1.92
30
0.52
EiDlmpl
Cdbe
0.73
0.89
0.78
0.77
a8
.79
0.40
0.81
0.57
0.59
0.73
0.71
Q05
0.59
G381*
Gmundn
0.o8
0.90
0.88
0.49
0.4
0.89
0.79
010
0.80
1.18
0.91
1.38
08o
0.52
78
ohw
Como
0.84
0.48
0.33
022
0.42
029
0.18
022
030
0.55
01
2.81
1.98
027
0.27
33
KW"
co..e
Q0.7
0.98
Q93
0.91
0.01
1.03
.77
089
0.52
.89
0.78
0.79
0.38
0.82
0.88
0.77
0.8
a70
LUbI
Fkbbw
0.65
0.88
Q32
083
09e
0.52
0.82
053
048
0.58
0.78
0.67
0.51
0.71
080
0.74
0.03
074
Madag_m
Cob.
1.32
131
1.3
1.12
1.58
0.77
030
.58
0s.
0.73
Q97
Q79
as4
0.47
0.54
0.48
087
083
maw
Tob
1.34
1.39
1.98
1.3
0.47
12
1.17
1.45
1.17
1.02
0.82
0.89
1.12
1.02
O0.0
0.8
1.05
1.08
kwaD
Colon
O.
0.51
0.40
030
0.8
0.40
.3s
05
0ss
0s3
0.55
Qi38
045
0.42
a55
1.2D
0.84
ass
lgmi
CMoc
0.84
0.95
0.48
03
9.5
0.61
928
0.49
0R3
0.Q5
1.16
1.27
1.03
0.80
0.13
0A7
1.41
0.74
Fmatd
Corme
074
0.7
082
U54
0.73
0.3
022
0.58
0.Q4
0.82
0.78
072
0.77
0.67
0.85
0s.
1.22
0.97
Groundn
0.65
0.9
0.72
03
0.90
0.911
073
0.85
0.77
08
.e8
1.11
0.72
0.48
058
l.9
247
z71
Code
1.7
1S.
0.94
1.08
1.19
045
0.31
1.19
0.98
1.34
1.70
1.15
075
8
1.15
1.
0.84
129
.07
eaul
Siam leone
e.7
13
Sanuma
u
121
1.0
1.18
1.38
77
0Q04
09
0.94
0.81
ass
074
1.01
0.2
0.95
0.75
0.7
1.08
1.0
Sudan
Co6tn
1.51
14
0.95
1.5
1.52
1.19
1.36
1.69
1.58
1.40
1.57
1.9
121
1S8
124
2.40
1.71
13
Tuanla
Cole
0.85
0.78
0.Q58
0.57
048
0.41 R
43
0.52
3
055
Toga
coa
Igd
C*
z2
er
0d.
Zmdha
Tobuc
ztnbabe
Tobaco
0.78
0.72 Q
43
029
036
Q40
38
0.38
0.71
0.85
.3
0.3
0.74
044
42
0.47
0.49
0.35
0.32
035
0.57
0.75
1.51
05
0.41
28
0.0
0.23
0l.
0.8
027
15
019
0.2s R
.511
038
Q0.5
Q40
0.83
0.63
044
a.44
a.s
0S4
0.9
Qes
1.41
0.31
01
0.383
0.1
054
1.3
0.42
0.22
0A.
0.41
0.55
04
1.10
1.22
1.48
1.48
1.02
1.22
0.93
1.16
1.00
1.07
97
1.31
1.16
1.00
073
0.44
0.48
1.
7
1.04
1.0
1.02
1.13
1.0
0.8
0.2
0.8
0.73
0.65
124
0.983
1.90
061
1.04
1.22
0.
s0.1
Anne D Table 3.
eFa Drtcon
coefft
for oincoi
.
eo ot
(1971-100 fo ree ecobangeratse)
Bwn
Fewo
Buwuid
1972
1973
1974
1975
1978
1977
1978
1979
1990
1981
1962
1903
1984
1985
1968
1987
068
.51
Q043
0.72
0.54
043
0.73
0.o6
0.60
0.56
063
0.49
0.48
0.72
1.58
1.16
1.09
0.67
0S.8
0.o6
0.72
02B
0o.
O.6
0A7
0.58
2
0.52
0.49
0.38
040
3
0.69
0.65
0.75
1.27
121
1.40
0.s2
0.95
0.75
1.50
1970
1871
Cdon
o68
Coffee
0.71
crop
Coqa
Cameroon
Cdbe
128
1.23
1.17
1.03
1.15
0.56
0.34
0.72
0.76
o0.9
1.40
1.10
0.64
072
0.84
CAR
Colton
0.64
0.64
05.0
0.38
Q.S6
0.48
0.53
0.87
0.61
0.54
0.57
Q57
0.49
0.47
0.71
0.3
0.33
.44
0.25
o.7
0.22
0.27
0.42
0.43
0A7
0.35
0.28
30
.44
0.64
1.13
1.07
1.04
1.37
0684
030
0.51
58
078
1.19
09e
0.78
0.71
084
0.72
127
97
0.84
0.95
Qe0
Q75
0.42
0.52
0.47
0.52
0.Q9
0.63
0.48
.38
35
049
0.75
.e6
1.58
2.19
1.98
.55
Q6s
1.16
1.15
CongD
Co
073
.62
Coat dlore
Coffee
1.11
Ebl
Coffee
0.73
Gambia
GUim
QS6
0.90
Goudnd
Cocoa
0.4
0.s
079
0.37
045
0.22
0.74
038
0.2
0S
e.3
0o.
0.59
0.12
0.75
097
0.78
1.17
0.77
0.48
0.64
0.19
0.28
0.1
Q.51
0.5
0.19
0.25
Keanya
e
CoffC
0.3
1.11
1.11
1.09
1.
1.25
0.99
0.98
1.05
1.02
0.62
0.92
1.09
o09
1.00
1.02
1.32
LUbei
Rubber
0.8
0.71
033
0.60
o0.9
Q48
0.60
0.84
047
0.54
0.68
0s.
0.41
0.55
0e6
0.87
o.60
0.71
Madagec
Code
1.32
1.29
1.34
1.05
1.44
0.75
0.41
0.2
0.59
0.
0.7
0.67
0.53
0.46
0.55
0.50
1.09
13
1.13
1.3
133
Maleaw
Tbeco
1.34
1.38
223
1.87
0.62
1.38
1.33
1.5t
1.23
1.o3
0.63
1.03
1.16
1.07
0.9a
Ma
Colon
00
0.4
0.35
0.27
0.2
0.38
34
0.40
0.47
047
0.42
0.51
038
0.35
043
Nigerlia
CcNoa
0.4
0s9
050
0.37
0.57
Rands
Cathe
0.74
0.72
0.3
02
o06
Senega
Gsoundn
0.6Q
Qs3
0.o8
0.34
SIM'-_n
Somi
Cadte
AMNINr
1.07
1.21
1.07
1.18
1.03
1.30
.es
Qs9
0.69
0.62
ZS9
1.48
0.8
020
0.34
0.35
Q53
Q.68
0.oe
0.48
0.2
0.31
Q46
029
0.17
4S
044
0.52
0.9
Q45
045
03
0.37
.3s
0.75
0.80
Q64
0.70
G94
0.72
1.15
0.74
0.48
Q5s2
1.68
2.2
2.58
0.89
0.78
0.7S
IA
2.04
0.32
0.48
0.O
1.18
0.91
2.49
1.90
1.37
.44
05
121
1.43
0a.
0.43
1.48
128
1.73
1.90
1.04
0.55
1.38
075
0.78
074
OS.2
Qs
0.37
0.35
0.QO
0.33
Sudan
Coaon
1.51
1.39
Q0.4
0.2
1.21
0S8
1.0
1.53
1.45
1.43
1.52
1.87
1As
19
128
TaNnia
Coffee
0.65
0.7
0.63
059
0A7
0.42
0.48
0.83
0.42
0.52
0.52
0.41
0.35
0.34
0.35
029
0.61
0o.
TOP
cocoa
0.75
0.42
Q30
0.40
.25
12
0.16
0.23
038
0.43
0.47
0.34
0.34
040
0.59
0.61
0.99
021
0.23
0.18
0.09
0.20
028
1.15
0.53
0.84
0.78
1.93
075
Uganda
Coafe
Z*de
Cdbfee
0.36
.e9
0.45
039
025
0.35
020
a.S
0.10
0.12
0.07
0.16
023
021
0.82
091
075
1.11
0.22
0.09
0.14
O.
0.29
0.71
0.50
020
Z2mbi
Tobacco
1.10
1.29
1.58
1.58
1.17
1.35
1.05
1.35
1.28
127
1.13
1.37
1.33
1.32
1.04
1.30
127
Zbnaw
Tab.w.
087
1.04
1.18
1.68
12
1.07
0.93
1.18
1.10
0.94
0.78
1.38
1.16
1.25
1.13
1.60
1.97
mal roducer pdrcefor malorexDortcmole
Annex D Wbbe4 Avte
#rfd-
low0
1972
1s71
1970
COF"
tG00
*
serdn
BudaFaS
10
sumr
94
Camro
135
1tn
13a
ts7a
ta7
1974
W3
1s77
t978
1979
tsao
tost
1a82
1ssa
ta94
100
115
10s
Wr
107
109
S4
es
7a
119
130
131
115
10S
104
137
144
121
132
140
129
112
142
108
5
as
93
S2
a7
S5
S
71
4
70
79
tO5
121
CAJ
ttO
17
S5
eS
132
119
Can
dno
EFd',
G&Yb
hom
so
85
7a
so
82
so
09
SS
10s
137
1eR
123
125
148
1as
tss
t20
282
09
tot
104
112
t3s
13D
221
20a
2U4
28
28
Gut
70
tOS
IC"
ss3
el
tt5
lie
185
2a
123
So
107
sO
89
85
79
75
71
119
117
tot
87
So
103
tos
S9
94
n
aS
Be
IGos
79
80
1aa
tsoD
t13
aoD
115
100
ss
M3
10o
tto
10o
so
102
so
75
eS5
14S
9t
157
SO
2ts
2aa
Is?
15D
IOD
113
20a
20a
a5
el
10w
100
n
78
0
SO
as
aS9
72
444
91
3t
33
125
10D
109
123
7s
tes
135
el
2B4
125
tos
174
147
115
109
III
ts
tco
e5
oo
so
as
sD
52
e3
t04
10D
ss
eS
ea
ao
aS9
BB
tot
161
12D
lsO
ss
990
1a2
13t
el
eO
tts
114
S2
143
tas
tss
145
log
1sa
122
bil
147
13s
115
15a
145
134
124
99
109
100
177
84
so
st
173
1E9
117
123
t10
so
129
NW
N - b
05
sa
Rwnd
121
t20
117
lag
te5
8d""
1G0
121
114
112
IN8
st
9t
14s
113
III
49
as
72
2Es
215
110
1w
Oa
120
sD
2Z7
17a
S2
of
119
t7s
119
141
la?
ts4
10a
78
74
73
III
10D
102
73
so
1n7
123
125
172
1S3
1e4
S3
9C
es
so
tOO
135
2d
12s
122
III
110
141
1#2
tsX
131
12D
117
12a
Zmnbh
tt3
114
13s
1aD
ISO
t34
12s
140
1s3
132
10a
III
10s
so
To
?4
el
el
104
go
So
too
9t
tOO
Be
ea
go
103
lie
123
123
12s
173
248
tst
1m
104
235
t3s
14t
7n
139
78
el
104
78
S
7a
54
20
78
122
so
34
143
e4
84
94
45
a
79
52
144
102
74
104
104
o
80
tO
122
tOO
t
n7
7o
t4
109
7n
128
84
121
UpndI
ISO
224
tss
ta1
t
122
20t
52
123
10O
132
195
S2
log
112
so
152
67
137
T"p
as9
1s5
7n
so
124
S2
131
log
l2a
os
el
el
Tminh
so
sOD
124
67
1 a
103
94
10?
too
tOD
10D
108
125
eS9
rOD
SwJ0
go
tos
107
10w
2S3
too
132
102
102
242
113
119
s7
8o0dlh
94
t0s
tOO
147
13
83
271
107
102
es
113
ZkFbx_
84
as
as
65
13
a7
89
lie
2
73
as
t2s
0
83
94
10a
GI
ea
813
12s
too
130
s1
as
10D
to7
aw
12S
71
tos
eD
MWad
124
75
tos
13S
magm
12D
10w
a2
n7
Ubzt
el
ioo
104
10s
S5
as
Go
85
Cads
C4n90
105
10O
too
tsa7
tsas
t8e
too
1es
100
so
n2
as
III
soo
110
ttO
as
S5
10s
or
t0S
tOO
la?
14s
132
1
141
1us
e4
Oss
Arnnx 0 tabb . Aveas nominalpatAclon coefficentsfor tradedfood o
CWY
atR
A
1970
1971
1072
1973
1974
1975
1970
1977
1978
1979
1.01
0.93
074
0.82
Q0.5
58
074
e.
0.74
0.70
1910
198Z
1983
1984
1985
1988
0g9s7
06
0.97
1.09
1.31
1.01
0.75
0.89
07
1081
090.99
Sundn Faso
1.38
1.57
1.69
2-31
0.73
09D
0.95
0.93
2.74
2.70
Z15
0.85
0.92
1.08
1t2
-0.30
-0.Z
5.00
0.83
t0.
0.80
0.74
074
0.9
1.42
2.07
1.97
1.64
1.03
1.02
0.93
0.4
0.78
1.22
1.5e
97
472
2ee
.6
8.75
14838
5.57
z77
1.03
9.08
Ghan
Q04
0.75
0.50
043
0.Q8
0.08
084
0.88
070
0.79
0.e
0.85
0.88
0.7
Kenya
1.07
0.94
0.57
0.50
0.75
IAI
1.55
1.19
1.29
1.04
1.53
2.10
2.2
2Z7
UbeaB
0.60
Malee
Ss
064
035
0.2
Q53
0.74
0.89
0.8e
0.74
25.81
1.09
072
1.31
1.89
4.9
6.89
.74
0.88
0.52
0.67
0.73
0.78
1Z5
12.18
-0.32
.0.47
.339
0.59
0.85
Mai
1.88
1.3
074
0.96
Cob d%alm
Medoca
1987
442
0.94
0.69
0.48
1.70
NIge
1.85
Nda&
Renda
Q02
c04
0.51
0.37
51
080
0.88
078
0.4
O.9
1.50
077
228
0.69
2.03
0.7
1.40
0.55
1.19
&78
2.03
0.78
0.87
1.47
1.87
4newaw
078
0.77
1.50
1.13
1.48
1.45
smn"
aand
123
1.18
0.70
0.8
1.88
1.19
1.35
1.10
0.98
Tanhs
Togo
0.1
1.05
G09
094
1.08
2.51
9.71
.44
2.2
4.09
320
1.97
1.19
12D
2.04
Z21
1.11
1.29
078
1.01
1.0
1.72
2.30
1.40
1.S
1.01
2.05
1.50
1.40
121
0.69
0.68
070
0.80
0.Q0
0.94
0.73
0.9
0.81
0.52
064
0.7
Q40
0.35
0.39
045
080
0.43
0.39
0.48
Q47
0.92
Z7mba
0.85
0.78
08D
048
0s.
0.o
ZbnbabWe
042
.40
0.30
Q.27
0.35
0.32
m37
38
hiod on
Pe,= fo folmalw
el
AnnexD lbibl Q
ts=o-
p
tS7D
CA
10ol
197t
1973
1S2
Boni
1974
1975
1s7B
1977
ts7
ts7
1sE0
105
107
114
IC"
1or
105
IOD
9t
ES
75
go
sobo_
67
so
u&mFeso
auuEw
Cams"wn
r-AtR
s7
Si
109
75
112
112
116
113
104
esB
S4
03
89
91
92
102
107
103
117
so
la?
Chad
117
12B-
122
194
20B
227
172
214
221
210
t0s
es
GhU
78
79
so
30
3t
Gu_
10t
Kea
~43
109
113
tco
103
102
1c
as
79
10
s
0
10
E
209
O
128
s
115
s100
o
s78
7B
175
1e4
1e7
109
100
120
123
io7
10D
123
148
10t
30
114
III
10S9
112
10B
119
151
4
144
12S
7
eo
70
73
Es
78
sD
7a
n
108
S~~~~~72
8
103
117
Md
10a3
S5
es
09S
10
MO&A
107
Bes
104
Oa
103
85
E
n
97
1a
r2
94
102
101
104
94
85
70
S1
75
72
lie
17S
1B
0
87
e7
79
t15
ttSTo
75
73
el
92
S3
117
120
117
110
114
114
105
tos
so
10
10o
84
793
s7
104
103
la?
71
e4
82
9t
73
lie
S3
es
tto
10o
89
100
112
a
E
0
90
87
100
112
14
O?
s
10D
EOD
114
S
98
97
to
S
72
so
178
o
1SB7
7
10
0
8
93
195
141
101
110
10B
115
107
Es
102
102
75
17
S3
93
8S
70
83
104
119
108
t2B
137
eas
33
S9
134
14?
bs
43
_
140
122
114
10D
7E
Ubeda
i
123
s7
9
109
1t
lie
112
0
tSHB
93
83
eSB
t9E
109
129
107
1SB4
10
112
119
Ehbpha
74
97
1SEs
1962
1oo
117
203
78
ss
OD
S2
owp
Cdet dWbeM
Gsnbh
104
aSB
1991
72
73
aSB
esB
175
57
231
s
8
WBs
103
00
Maushd
sB
NxprM"
25
.
~~127
F"h
127
Rwnzp
ew
Si
73
E
15
153
144
as
a
11
141
137
OD
U
3
13
143
113
109
93
t311711
70
93
8
to
a
121
1
1
Ila
a3
too
92
91
97
97
181
133
n
109
1a0
lie
1ot
S5
102
so
too
142
112
asB
100
So
79
ST
10
85
esB
B4
5t
44
95
95
s7
120
125
114
113
184
145
1a
122
92
135
13E
123
47
as
54
as
77
ss
71
w
79
El
Sudn
Tazf
TOgt*
75
83
BOD
10t
so
121
130
131
143
147
139
128
100
164
U8e
137
13B
174
212
115
so
100
les
185
549
307
225
t50
100
540
370
100
140
Ise
155
14B
127
136
la
146
104
117
112
IGO
150
1a
95
105
123
113
S4
el
10B
S3
77
6se
6B
Z77
107
119
149
65
EB
1
138
ZWn&
82
es
100
149
so
10t
83
S4
1sB
7s
74
So
70
103
181
Zambkl
172
e7
as
ISe
7ss
149
83
10s
80meh
7ss
128
89
eS9
103
go
UgDnda
142
100
lo
85
er
el
72
so
Be
7t4
eo
so
77
82
s7
1ao
e4
e7
e2
78
IOo
84
59
d
t
I
tI
dd
i ll
1
d
d
6
dd
dd
ddd
d
d
o
d
000
-
d
d d
d,d
d
dd
dd
dd
ddd
d
ddd
d
dod
~11
d
o t §
I
aJ
l
I
do
d
dd
d
dd
d d
dd
d
od d
d d d
d
d
dd
I~1ZRifJtlXJgIff{0
Ils
{XZoi{§}Z}l||zg
d
d
d
d
dd
d
d
-60Annex D Table 8. Ratio of implicit financial outflow from agriculture
(1980-85 average, in percent)
Country
Benin
Burkina Faso
Burundi
Cameroon
CAR
Congo
Cote d' Ivoi.e
Ethiopia
Gambia
Ghana
Guinea
Kenya
Liberia
Madagascar
Malawi
Mali
Mozambique
Niger
Nigeria
Rwanda
Senegal
Sierra Leone
Somalia
Sudan
Tanzania
Togo
Uganda
Zaire
Zambia
Zimbabwe
To total
government
expenditure
in agriculture
49.58
86.65
68.79 a
To total
government
revenue
0.55
2.41
15.69
4.67
15.12
74.19
204.05
25.00
37.33
0.32
17.86
3.36
3.14
47.07
-0.39
1.78
6.29
22.38
4.70
3.94
-0.01
2.57
135.87
4.57 a
93.94
14.89
-25.05 a
56.35
101.35 a
19.80
124.38 a
1.19
-0.47
0.00
0.16
11.08
0.42
13.12
0.89
-3.28
6.83
8.18
1.20
6.41
G.25
-0.05
450.05
22.68
13.95
420.54
To total
government
tax revenue
2.82
17.03
5.18
22.20
4.41
3.50
55.23
2.01
6.56
28.63
5.50
4.60
0.02
5.12
0.02
15.12
-1.28
61.62
0.97
1.84
0.02
-0.09
To
agricultural
GDP
0.18
0.92
4.06
4.21
4.04
1.37
17.80
1.65
2.16
7.39
-0.26
1.39
4.06
8.80
2.90
0.94
0.00
0.08
3.?7
0.43
3.54
0.13
-1.29
. 47
8.20
0.14
3.02
0.41
-0.13
Note: Implicit financial outflows are the estimated unit transfers based on the nominal
protection coefficients weighted by total export quantities.
a. Estimated Government expenditures in agriculture were available for entire
1980-85 period.
from trend of cereal vieds. In logarWnM
Amnex0 able 9. Weather valable (dewAatlon
197o
CoUb
1971
4.0147
1972
-0.120
1973
1974
1S75
1978
1977
1978
1979
1s8o
1981
1982
0.078
0.Q01
0.115
0.079
0.105
0.073
0.032
0.013
40.094
4Q118
0.346
1984
1985
1988
1987
.0211
0.058
0.098
0.092
4D092
1983
Senn
4-052
Boft A
-1.628
0173
0.986
0.228
0.736
0.342
0.750
40.0s8
4sesG
.0.116
0.099
0.154
0.103
-0.503
40084
-0.15S
0.177
0.138
0.037
-0.025
-0.104
Q001
0.092
40.044
4Q011
0.080
0.032
.0.010
-0.023
-0.082
4-.188
-0.075
0.085
0.145
0.033
0.010
0.003
0.088
4Q.004
0.008
-0.04e
MOSS
0.052
-0.030
0.093
-0.005
-. o0
-0.120
.0.015
0O28
0.000
-0.015
0.004
Suidna FeS
Bunid
Cewnemon
0.031
-0.123
-0.013
-0.017
-0.009
40.083
0.075
0.211
0.184
4QOB
40104
0.23
1.34
.0.037
0.334
0.385
0.304
-0221
-1.275
0.141
0.539
0.473
Chad
Q179
0.079
40.1B9
-0.122
0.115
0.012
Q0.107
.0.076
40.05e
0.031
40015
Con0
0.122
0211
0.166
-0.087
40.28
4.0.9
0074
400389
-0383
0.023
0.161
.015
0.197
0.010
0.058
-0.025
-0.125
40.125
-0.132
-0.119
-0.117
-0.115
0.031
0.107
0.080
-0.098
0.107
-0.014
Cape
Cow
Vede
d7eore
CAR
Q.128
0.149
-0.004
0.128
Q114
40.185
EtIopia
4.010
-0.024
-0.074
-0.114
-0.095
Gabon
-0.030
-0.032
.0.001
0.013
4015
GaMbia
Ghana
Guinea
Gtu
Kea
Lesotho
0.094
a027
0.027
-0.011
40.001
-0.249
-0.088
-0.343
022
0.439
0.20
40.13
-0.322
0.117
0,113
Q.007
-0.093
-0.070
-0.050
0.027
0.201
Q221
0.188
4.235
0cas
-109
0.037
-0.144
-0.07
4021
-0.024
-0.041
0.150
0.121
0.070
0.189
0.037
0.178
-0.084
-0.022
40.007
0.102
0.069
0.031
0.142
0.132
0.057
0.058
0.071
-0.098
-Q151
40.317
0.021
-0.159
0.053
0.100
0.038
-. 229
.012
0.019
0.03 i
Q031
0.138
Q004
0.013
0.010
0.020
0.032
40.128
0.059
-0.199
.0.780
0.005
0.013
0.018
0.024
0.024
0.034
0.029
0.053
0.081
0.088
0.094
0117
-04208
A140
-173
-0.140
0.04
-0.095
0.082
dnasa
4G025
0.083
.0.150
4.012
4-.2
-0.084
0.093
-0.133
-0.041
-0.083
-22
0.127
-0.124
4.105
40125
-0.091
056
-0.183
40.097
Q.018
0.022
0.028
40.010
.089
0.050
0.023
-0.079
-0.18o
0.217
0.007
227
0.012
-0.4Q2
023
0.179
-0.083
4Q181
0.671
Qs94
038
0.208
0058
4-170
4152
4203
0.094
0.001
-0.028
40.008
0.080
0o.1
0.004
-0.001
Q002
-0.037
40005
Q027
4058
4072
0014
0.134
-240
.0.012
015
-0217
0021
0.101
0.090
0200
as20
-0.101
a167
o097
-0.102
-0.008
Madagasow
0.040
0.009
0.014
4014
0.028
0.003
4C.027
-0o08
-0.028
4025
0.088
0.094
.110
-0.435
0.02Q
0.12
-0 '38
-0.240
0.315
40.148
0.052
0.0
0.028
4oss09
4-04
0148
0.071
0.083
40.073
40o04
4.319
0.052
0281
0482
0.13
4B14
40308
4047
4074
4200
4205
40.28
4M
4027
4120
0.237
0279
4092
4028
4Q007
0.019
0.042
0.283
0271
MWamMu
0.088
Q323
mamnbu
0.097
0c.62
4o41
O.Q03
Q.048
40o48
NIger
N%p%
Fhnda
Sena
Sene Leone
soma9a
Q0.05
0.016
4098
.202
0S
4047
m0oo
0.011
0.056
G051
4305
Q1e2
4.010
-0.254
Q128
.039
-4051
4333
0.02s
4188
-0217
4Q110
m171
0.014
4Q189
0.001
4010
0.012
0.011
4C002
4X14
4144
49ss
4126
Q013
ambeb
4.333
0.081
Q153
Q164
Q316
4033
0.128
0.oe
0.088
4050
0.028
0.081
4L041
01ls
0.240
4295
0.083
0587
4007
4LO42
Zambia
0.s1
41ao
0258
0.002
0.011
0.010
4-140
4873
Zabe
4Q129
4QD98
Tarami
TogP
Uganda
4ne3
4-020
Sw,d
0.24
0137
0.038
.0390
4018
4123
4143
.10S
4003
0209
4009
OA023
4092
0.190
413S
MOe07
4050
413
0.009s
4001
0.053
4183
0231
4293
4o
0.151
.075
0.123
Q034
0.08
Sud-n
4.090
0.015
0.194
0.097
Q038
0.045
Q012
023
.089
0.030
4s05s
Q011
0.021
.083
4.015
0.097
0.224
0.019
0234
0.04
0219
MOO0
00.82
40o14
Qo87
4c039
4182
4-114
4107
4109
os0
0.001
Q027
4071
a168o
4144
nc28
4048
.082
4004
.030
Q027
0.091
40.089
402
4091
4153
4QDe2
0.139
0O046
4077
0.021
4204
0.13s9
4973
0.007
4094
4092
0252
41i0S
0.077
4087
0.178
0223
0.25
Qos.1
Q
071
0.100
I189
Mi110
4089
4620
0087
4048
4103
4L114
4087
4189
104
4088
0.109
0.034
4137
40128
0c.9
0.191
4608
0.011
0.034
Q072
Q020
06143
Q8OS1
0.007
0.154
-4380
4072
m0ss
4075
Q222
095
0.189
4401
4514
n24s
4048
4Q020
4040
.0o0
0.153
0.127
4369
o053
4024
0235
MOD0
4.132
0.057
0230
0.088
.183
Q002
4300
0.158
0O093
498o8
.32
40009
4130
078
QOOE
02s9
40125
40017
0.044
4.038
M110
4L004
0C.23
4Q122
Q032
0.019
4029
.17S
-. 168
Q003
n128
4.018
0.195
4147
0.040
0.127
4017
.049
0131
.085
009
n144
0.068
0L051
Q013
0.098
Q12s
4.054
0233
mawal
4197
4-0.2
0.15s
-0.11
0212
0.003
MaWawnl
0.071
0.022
0.204
0.003
Mabu
0122
0.148
0.003
.015
0.028
-0.042
123
Ubeft
0.085
0.033
4019
-0.059
0.307
4013
0.121
4482
0.007
0.008
0.089
0Q057
:w
-Q013
4-433
0.048
0.018
0230
40m18
QDQ7
4153
0.075
405
0.24
0.0S
0.027
0o0s
0.008
4L013
0.004
0204
4032
4988
4092
4L043
0.072
4314
0370
4L049
40.24
4398
0.489
0300
4301
0.029
0.013
Annex0 tabie 10. 1lasts
data: rnumber
of pople affectedby majordisastes, In thousands.
Bonin
FAwanda
IwidnFaso
3urunm
Canewon
Cape Verte
CAR
Chad
Conorc
Congo
Cote dlvof
EquaL Guinea
1979
1960
1981
1962
1983
1984
1985
1968
1987
0
0
0
F81
0
0
0
500
0
1
0
0
0
0
0
0
0
1900
35
0
4
0
442
2
0
0
0
0
0
0
0
0
2
0
0
0
500
475
0
0
30
0
0
1037
2502
0
0
0
0
1503
0
0
350
s80
0
0
0
a
0
0
35
0
375
848
0
0
a
0
0
0
0
0
0
671
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50
0
am
0
1417
250
0
0
25
0
0
0
0
0
2000
0
500
0
8051
0
0
0
7330
1973
1974
0
0
0
1is6
8O
0
8S
0
0
0
0
0
0
0
0
0
0
0
0
400
0
C
10
0
0
115
1
0
5O0
0
0
0
0
0
0
0
0
1300
0
0
0
13
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
5
0
0
0
0
0
1700
EtNopa
1978
0
1i72
0
BOO
0
1580
1975
1977
1971
0
1970
Angohp
0
1580
1976
0
0
0
0
0
0
0
1500
0
0
0
0
4
0
0
0
1500
0
0
500
410
7
0
0
2
0
1700
0
0
250
410
0
0
0
0
0
i00
Gabon
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
GmbIa
Ghana
0
0
150
12
0
0
0
0
0
0
0
0
0
0
0
7
85
0
0
0
500
0
301
0
0
0
0
1320O
0
0
0
2
0
0
0
0
GuInea
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2D
0
0
0
0
0
0
0
0
4
0
0
0
10
0
0
0
0
0
3
82
0
150
0
0
0
0
0
0
0
501
0
0
0
0
1
0
0
0
0
0
0
13
8
0
9o5
0
0
0
10
0
0
0
87
8
0
9O0
18
0
0
500
0
0
0
0
0
0
0
0
0
0
30
0
0
710
0
400
0
0
1700
800
500
0
114
0
1503
1853
0
5010
3500
0
0
0
90
0
0
0
1503
8wo0
0
3068
0
1
420
o
0
84
0
0
0
0
0
0
0
0
0
0
0
28
715
0
0
0
6580
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
30
0
0
0
200
0
0
0
0
0
0
0
0
0
0
0
150
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
124
0
0
0
0
0
0
0
0
0
0
5
125
75
0
0
0
0
0
0
0
0
0
5
0
70
0
0
0
0
1
0
0
0
0
0
173
0
0
20
0
0
8
0
0
0
0
0
a
4200
0
1o9
0
0
0
0
0
3
0
3
8
20Dr
C
0
0
10o0
0
0
0
0
0
0
50i
0
0
0
0
331
0
0
0
Guuna-Blsa
K-na
Lesotho
Ubela
Madaga
MalawI
mani
mawIlhl
MuAws
Dn-bkq
Mo
NIger
Nwla
Fkrda
SWoTorn&a,Ffdpe
SeNa
9eychel
S8M Le"
Salak
Suda
tand
Tanzania
TogP
Ugan
zaIo
Zambia
Zabnb
SOUR0
US0
May 196& Ue"M
MW.
0
0
0
2500
0
0
0
25
0
0
0
0
0
0
0
0
0
1900
1300
0
0
160o
0
3
0
1400
0
0
0
60
0
0
0
0
0
0
0
Clbe d U.S ForeignOtrsv
828
Assanoe
ao
0
0
1
0
0
0
710
0
200
0
0
1
0
0
204
18
20
0
27
0
so
0
0
1
0
0
0
0
0
3715
0
0
20
1W
0
0
0
0
0
0
0
95D
0
0
0
0
0
0
0
500
18o
11
0
0
0
0
20
0
0
0
0
0
0
12
0
0
0
0
0
0
0
40
a
0
0
0
1
16
80
0
0
6S
30
0
0
400
500
1
0
40
0
2
0
0
0
0
100
0
0
0
0
0
0
0
1000
0
5
0
0
6006
0
0
0
4
0
0
118
0
0
0
32
0
3
0
0
13
0
0
0
0
0
80
0
400o
67
0
0
0
605
0
0
0
140
1
0
2D
15D
0
0
0
983
600
0
0
0
5202
632
0
0
0
300
0
0
o
N
-
63
Annex D Table 11. Food Crop Price and Policy Effects
Real producer prices
(index 1979-81 = 100)
Benin
Botswana
Burkina Paso
Burundi
Cameroon
CAR.
Chad
Congo
Cote d'Ivoire
Ethiopia
Gambia
G;hana
Guinea
Guinea-Bissau
Kenya
Liberia
Madagascar
Malawi
Mali
Mauritania
Mozambique
Niger
Nigeria
Rwanda
Senegal
Sierra Leone
Somalia
Sudan
Tanzania
Togo
Uganda
Zaire
Zambia
Zimbabwe
Nominal protection coefficients
Average
1970-72
Average
1975-77
66
123
89
107
108
87
89
104
98
115
120
219
71
84
30
90
207
181
117
138
114
96
93
85
74
93
105
59
80
95
98
134
96
113
101
96
104
80
172
94
109
71
140
140
103
71
181
66
72
299
91
101
102
106
91
142
88
130
139
202
83
76
84
84
Average
1985-87
Average
1970-72
Average
1975-77
Average
1985-87
136
85
0.89
0.66
1.2
0.73
1.04
0.93
3.32
1.89
0.63
0.86
0.53
10.92
0.79
1.38
0.7
4.58
0.93
2.43
1.06
1.49
0.58
0.59
0.66
1.27
1.03
1.24
92
92
88
1.59
109
88
139
94
0.68
0.37
1.7
1.32
1.34
1.58
1.62
0.61
0.36
0.68
0.62
-64 Arinex D Table 12. Comparison of domestic staple food prices to imported rice and wheat prices
As a fraction of international
rice price
Country
Crop
Burkina Faso
1985-87
As of fraction of international
wheat price
1970-72
1980-82
1985-87
1970-72
1980-82
sorghum
maize
0.58
0.58
0.79
0.79
1.31
1.31
2.08
2.08
Cameroon
mai7e
cassava
plantains
0.99
0.36
0.53
1.08
0.33
0.69
2.23
0.82
1.19
2.86
0.88
1.83
Cote d'Ivoire
maize
cassava
yams
1.25
0.93
0.92
1.29
1.25
1.19
2.80
2.09
2.07
3.43
3.32
3.15
Ghana
maizz
cassava
1.02
0.42
6.88
3.43
2.28
0.95
18.26
9.09
Kenya
maize
0.56
0.70
1.27
1.85
Malawi
maize
0.32
0.27
0.41
0.72
0.71
0.75
Mali
sorghum
0.75
0.86
1.22
1.69
2.29
2.24
Niger
sorghum
0.75
1.32
1.69
3.50
Togo
maize
sorghum
0.72
0.80
0.80
0.64
1.61
1.79
2.12
1.69
Senegal
sorghum
1.00
0.74
2.25
1.97
Zambia
maize
0.60
0.60
1.35
1.60
Zimbabwe
maize
0.47
0.43
1.06
1.15
1.92
1.10
1.54
0.65
3.52
2.01
2.83
1.18
Am=e 0 Tebb 1. Vkaus Sp
_eafiofRd
Pebn
Equelon for Aadcjftura SuPnh.
Oeendt
vabble - egdcuur
ADcotants
Producer Pfw
aop yw (8
a8s
Pj!3"
agge me
(-2 to t)
FRe yer
(t to t'4
a-Sp
Logged
awe
01s
We,offer
Anua crop expoibs
4.OD1
0.560
(400
(4.00)
40tS2
(46
0.091
(204
Ias rate
0.002
p09Ed
7-a65M)
"
..4oie
173
(431)
to
lbbV
4QDDtO
f264
bb
0.372
(16.9)
dnt*eedom
FIw
mOOtt
(t3
0.303
p30)
4.33
(11.95)
Now E8or
QW5
PM)
150
(427)
bawoegI
De_
t
Tree empempost
9mo arage
R.W
Qep yew(8
e
puu
In
4.10s
68
pdce
-Ig.
rawe,ard fe rbt
4.18
(.45)
0250
(1.79)
4Q000
(454
40011
(-1.67
40042
(.1127)
40011
(.1.19)
4.008
(40.)
0.0022
(1.M
Q405
0.404
(10.1)
0.184
(14.3)
0.187
P.08)
0Q242
(39)
Q48
(5.6)
.33
(81.1)
5.40
(3.5
3.10
(18.)
217
e A We tioues
4140
(.04)
40210
(-10.07)
1w
lbo:bh been
...
3as
(174
233
In logeaur_
llet
... hanta may be Ibtorpee
6.S1
p.9t)
97
as ebstCkbLs
- 66 BIBILOGRAPHY
Ademola T. Oyejide (1986). "The Effects of Trade and ExchangeRate Policies on Agriculture
in Nigeria." Research Report 55. WashingtonD.C.: International Food PolicyResearch
Institute.
Balassa Bela (1988). "IncentivePoliciesand AgriculturalPerformance in Sub-SaharanAfrica."
PPR Working Paper No. 77. WashingtonD.C.: World Bank.
Bateman, Merrill, David M. Newbery,WilliamAsenso Okyere and Gerald T. O'Mara (1988).
"Ghana Cocoa Pricing PolicyPaper." World Bank consultantsreport (processed).
Bates, Robert H. (1981). Markets and States in Trogical Africa - The PoliticalBasis of
Aricultural Policies. Berkeley: Universityof CaliforniaPress.
Bautista, Romeo M. (1988?undated). "Agriculturalgrowth and food imports in developing
countries: a reexamination,"Processed. Washington,D.C.: International Food Policy
Research Institute.
Berg, Robert J. and Jennifer Seymour Whitaker,eds. (1986). Strateeies for African
Development. Berkeley: Universityof CaliforniaPress.
Binswanger,Hans (1989). 'The PolicyResponse of Agriculture,"Proceedingsof the World Bank
Annual Conference on DeveloRmentEconomics 1989. Washington,D.C.: World Bank.
Binswanger,Hans P. and Pasquale L. Scandizzo(1983). "Patterns of AgriculturalProtection."
ARU DiscussionPaper No. 15. Processed. World Bank.
Bond, Marian E. (1983). "AgricultureResponses to Prices in Sub-SaharanAEican Countries.
"International MonetaryFund Staff Papers, Vol. 20 No. 4 pp. 703-736. WashingtonD.C.
ResRonsein Agriculture. Forthcomingin S.
Chhibber, Ajay (1988). The AmereeateSUDPIY
Commander (ed) Structural Adjustment in Agriculture: Theory and Practice. London:
James Curry Publishers.
Cleaver, Kevin M. (1985). "The Impact of Price and ExchangeRate Policieson Agriculture in
Sub-SaharanAfrica." World Bank Staff WorkingPapers Number 728. WashingtonD.C.:
World Bank.
Cleaver, Kevin (1988). "AgriculturalPolicyReform and Structural Adjustment in Sub-Saharan
Africa: Results to date. "World Bank Africa Department I. Washington,D.C. Processed.
Delgado, Christopher L. (1987). "The Role of Prices in the Shift to Rice and
Wheat Consumption in Francophone West Africa." Paper prepared for the Conference on
the Dynamicsof Cereals Consumptionand ProductionPatterns in West AfricaJuly 15-17,
1987, Dakar, Senegal: International Food PolicyResearch Institute.
-
67
-
Devarajan, Shantayananand Jaime de Melo (May 1987). "Adjustmentwith a Fixed Exchange
Rate: Cameroon,Cote d'Ivoire, and Senegal." The WorldBank EconomicReiew, Vol. 1, No.
3: 447487.
'eerch
Edwards, Sebastian(1989a). "ExchangeRateMisalignmentin DevelopingCountries."
Bank.
Qbsrver,Vol. 4, No. 1. World
Edwards, Sebastdan(1989b). "Openness,OutwardOrientation,Trade Liberalization,and
EconomicPerformancein DevelopingCountries". PPR WorkingPaper No. 191. Washington
D.C.: World Bank.
Eicher, Carl K. (1986). "StrategicIssues in CombatingHungerand Poverty in Africa" in Berg
and whitaker.
Gitfinger,Price J., JoanneLeslie and CarolineHoisington(1987). Food Policy:Integrat
Supply.Distribution,and Consumption.Baltimore: The Johns HopkinsUniversityPress.
Haggblade,Steven,Peter B. Hazeland James Brown(1987). "Farm/NonfarmLinkagesin Rural
Sub-SaharanAfricaEmpiricalEvidenceand PolicyImplications.'BGRAPEconomicDiscussion
Paper No. 2. WashingtonD.C.: WorldBank.
Harvey, Charlesed. (1988). AgriculturalPricingPolicyin Africa. Four Country Case Studies,
Hayami,Yujiro and VernonW. Ruttan(1971). AgriculturalDavelopment.Baltimore: Ihe
Johns HopkinsUniversityPress.
Imran, Mudassarand DuncanRon (1988). "OpfimalExportTaxesfor Exportersof Pernnial
Crops." PPR WorkingPaper No. 10. WashingonD.C.: World Bank.
Jaeger, Williamand CharlesHumphreys(1988). "The Effectof PolicyReformson Agricultural
Incentivesin Sub-SaharanAfrica." AmericanJournalof Agriculural EconomicsVol. 70 No. 5,
December 1988.
Jansen, Doris (1988). Trade ExchangeRate. and AgrictlturalPricingPoliciesin Zambia.
WashingtonD.C.: WorldBank.
Johnston,BruceF. and WilliamC. Clark(1982). RedesiningRural Development- A Strategic
Baltimore: The Johns HopkdnsUniversityPress.
Perctive.
Johnston,BruceF. (1986). "GovernmentStrategiesfor AgriculturalDevelopment'in Berg and
Whitaker.
Kmenta,Jan (1971). Elementsof Econometrics.New York: MacmillanPublishingCo., Inc.
Krueger, Anne 0. (1974)"The PoliticalEconomyof Rent-Seeking"AmericanEconomic
Review64, No. 3: 291-303.
-68 Krueger, AnneO., MauriceSchiff, and AlbertoValdes(1988). "AgriculturalIncentivesin
DevelopingCountries: Measuringthe Effectof Sectoraland EconomywidePolicies." Mm
WoQdBank EconomicRevie, Vol. 2, Number3.
Lecaillon,Jacques and ChdstianMorrisson(1986a). EconomicPicies arndAgicultural
Performance: The Case of BurldnaFaso. Paris: DevelopmentCenter of the Organizadonfor
EconomicCo-Operationand Development.
Lecaillon,Jacquesand ChristianMorrisson(1986b). EconomicPQliciesand Agricull
Pejq=mance: The Case of Mali, 196041983.Paris: DevelopmentCenterof the Organizadon
for EconomicCo-Operationand Development.
Lele, Uma (1989). "Sourcesof Growthin East AfricanAgriculture." The World Do&
EconomicReview.Vol. 3, No. 1: 119-144.
Upton, Michael(1987). "Limitsof Price Policyfor Agriculture: Which Wayfor the World
Bank?" DevelopmentPolicyReview. Vol. 5, 197-215.
Lofchie, MichaelF. (1989). The PolicyFactor: AericulturalPerformancein Kenyaand
Tanzgnia.Boulder: LynneRiennerPublishers.
Mellor, John W. and AhmedRaisuddin(1988). AgriclturalPrice Policy for Developing
Countries, Baltimore: The Johns HopkdnsUniversityPress.
Mellor JohnW. and BruceF. Johnston(1984). "The WorldFood Equation: Interrelations
AmongDevelopment,Employment,and Food Consumption."Journalof EconomicLiterature.
Vol. XXI (June 1984),pp. 531-574.
Poleman,ThomasT. (1981). "Quantifyingthe Nutrtion Situadonin DevelopingCountries."
Food ResearchInsItute Studies.Vol. XVIII, No. 1, 1981.
Rassas,Bechir, CharlesRabenarivoand LawrenceMeserve(1988). "Evaluationof the Food for
ProgressRice Program in Madagascar." WashingtonD.C.: InternationalScienceand
TechnologyInstitute,Inc. (Processed).
Roberts, Coileen(1988). Trade. Aid. and PolicyReform.Proceedingsof the EighthAgriculture
SectorSymposium. WashingtonD.C.: World Bank.
Serageldin,I. (1989)Povert. Adiustmentand Growthin Africa. Washington,D.C.: World
Bank.
Sharply,Jennifer(1986). EconomicPoliciesand Agicultural Performance. The Case of Kenya.
Paris: De.velopmentCenterof the Organizationfor EconomicCo-Operationand Development.
Scandizzo,PasqualeL. and Colh Bruce(1980). Methgdologiesfor MAeasuring
Agricultural
Price InterventionEffects,World Bank StaffWorkingPaper No. 394. Washington,D.C.:
World Bank.
-69 Stryker, Dick, J. (1988). "Ghana: A ComparativeStudyof the PoliticalEconomyof
AgriaulturalPricingPolicies."Processed. WashingtonD.C.: World Bank.
Svedberg,Peter (1987). "Undernutritionin Sub-SaharanAfrica: A criticalAssessmentof the
Evidence." WIDER workingpaper 15. Helsinid: WorldInstitutefor DevelopmentEconomics
Research, and Institutefor InternationalEconomicStudiesUniversityof Stockholm,Stockholm.
USAID.(1988). DisasterHistory: SignificantData on Major DisastersWorldwide.
1900pre . Washington,D.C.
UnitedNations (1989). Updateon the NutritionSituation.RecentTrends in Nutritionin 33
countries. Geneva: AdministrativeCommitteeon Coordinadon- Sub-Committeeon Nutridon.
Valdes, Albertoand AmmarSiamwalia(1988). "ForeignTradeRegime, ExchangeRate Policy,
and the Structureof Incentives."in Gittingeret al.
von Braun, Joachimand EileenKennedy(1987). "CashCropsversusSubsistenceCrops:
Incomeand NutritionalEffectsin DevelopingCountries."In Gittingeret al.
Westlake,M.J. "The Measurementof AgriculturalPrice Distordonsin DevelopingCountries."
Journal of DevelogmentStudies. 23(1987):367-81.
Wheeler,David (1984). "Sourcesof Stagnationin Sub-SaharanAfrica". WorldDevelopment
Vol. 12, No. 1, pp. 1-23.
WorldBank (1986a). Povertyand Hunger. Issues and Opions for Food Secumity
in
DevelopingCountries. WashingtonD.C.: WorldBank.
World Bank (1986b). WorldDevelopmentReport 1986. NewYork: OxfordUniversityPress.
World Bank (1988a). The Challengeof Hungerin Africa. WashingtonD.C.: WorldBank.
World Bank (1988b). CottonDeyelopmentProgramsin BurkinaEaso. Cote d'Ivoire. and Togo.
WashingtonD.C.: OperationsEvaluadonDepartment.
WorldBank (1988c). The NigerianStructuralAdjustmentProgram: Policies,Impact, and
Prospects. WashingtonD.C.
and Growthin the 1980s. WashingtonD.C.: World
WorldBank (1989a). Africa's Adjustmnent
Bank.
World Bank (1989b). Ghana:StructuralAdjustnentfor Growth. WashingtonD.C.: World
Bank.
World Bank (1989c). SuccessfilDevelopmentin Africa. WashingtonD.C.: World Bank.
World Bank (1989d). World DevelopmentReport 1989. NewYork: OxfordUniversityPress.
I
ia
PREWorkingPaperSeries
Contact
gape
Auftho
DirectInvestmentin
WPS619 F6.ceign
Sub-SaharanAfrica
LaurenceCockoroft
RogerC. Riddell
March1991
S. King-Watson
33730
WPS620 Hovetommercial BanksIgnored
SuleOzler
March1991
S. King-Watsrn
33730
History?
forHighly
WPS621 SensibleDebtBuybacks
IndebtedCountries
EnricaDetragiache
March1991
S. King-Watson
33730
WPS622 HowFactorsin CreditorCountries
MarketPrices
AffectSecondary
CountryDebt
for Developing
SuleOzler
HarryHuizinga
March1991
S. King-Watson
33730
Adjustment ViftoriqCotbo
WPS623 WorldBank-Supported
2 PatricioRojas
Performance
Programs:Country
March1991
A. Oropesa
39075
and Effectiveness
-.
WPS624 ChoosingPolicyIntrumentsfor
t
A Review
Pollution.Control:
Gunnar Eskeland
Jimenez
Emmanuel
March1991
A.Bhalla
37699
WPS625 HowTradeandMacroeconomic
PoliciesAffectEconomicGrowth
In.
andCapitalAccumulation
Counfties
Developing
RamonLopez
March1991
K. Cabana
37946
of thePublic
WPS626 TheMacroeconomics
SectorDeficit:TheCaseof
Colombia
WilliamEasterly
March1991
R. Luz
39059
InPoverty
of Institutions
WPS627 The>Role
Reduction:A Focusonthe
ProductiveSectors
SharonL. Holt
March1991
WDROffice
31393
DovChernichovsky
FamilyPlanning
WPS28 TheIndonesian
Program:An EconomicAwspective HenryPardoko
DavidDeLeeuw
Rahardjo
3
¢ DPudjo
14
CharlesLerman
March1,991
0. Nadora
31091
ot SuccOs PatrickConway
WPSO2 AnAtheoreticEvaluation
Adjustment
in Structural
March1991
Zou
WPS630 TheSpiritr'tCapitnlisrandLong-run Heng-fu
Growth
Marc 191t
, A. Shalla
37699
March1991
R. Lux,
34303
of thePublic
WPS631 TheMacroeconomics
of
Case
The
SectorDeficit:
Morocco
RlccardFalni
,
D.Ballantyne
37947
N
PRE WorkingPaperSerieas
Contact
~~~~~
Iilli
~~Author
for paaeer
12aQ
of the Public
WPS632 TheMacroeconomics
SectorDeficit:TheCaseof
Argentina
CarlosAlredoRodrigtez March1991
R. Luz
34303
of the Public
WPS633 TheMacroeconomics
SectorDeficit:TheCaseof
Thailand
Virabongse
RamangkuraMarch1991
Bhanupongse
Nidhiprabha
R. Luz
34303
Country
WPS634 Trendsin Developing
Expot's,1963-88
BelaBalassa
March1991
WDROffice
31393
WPS635 Exchange
RatesandForeignTrade BelaBalassa
in Korea
March1991
C. CuskellyYoung
39413
in Eastern
WPS636 EconomicIntegration
Europe
BelaBalassa '
March1991
C. CuskellyYoung
39413
WPS637 Povertyin Poland,1978-88
BrankoMilanovic
March1991
WDROffice'
31393
theTrade-ProductivityJamesTybout
WPS638 Researching
Link: NewDirections
March1991
D. Ballantyne
37947
WPS639 The'highCostof Protecting
Uruguay'sAutpmotive
Industry
WendyE.Takacs
:
'9rch 1991
D. B3allantyne
37947
WPS640 The Impactof Policyin African
An Empirical
Agriculture;,
Investigation
WiliamJaeger 'YMarch
C
,,>
ideru
34663
19
a
S~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
H~~~~~
.'
:V\
..-.
,
(~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~J
0
a~~~~~~~~~~~~~~~~~~~~~~~~
, -,
.X0