The Impact on Uganda of Agricultural Trade Liberalisation

Journal of Agricultural Economics frac34 Volume 53 Number 2 frac34 July 2002 frac34 Pages 365-381

Oacute Agricultural Economics Society

The Impact on Uganda of Agricultural TradeLiberalisation

Adam Blake Andrew McKay and Oliver Morrissey1

(Original received September 2000 revision received November 2001 accepted January2002)

This paper evaluates the impact on Uganda of the liberalisation of world trade especiallyin agricultural commodities as proposed in the Uruguay Round and of unilateralliberalisation of Ugandarsquos own import tariffs A CGE model of Uganda is used to modelthe effects of trade liberalisation taking into account the distribution effects by householdtypes We can draw three broad conclusions First the impact of multilateralliberalisation on a low-income country such as Uganda appears to be quite slight albeitpositive largely because there is only a slight impact on the world prices of theagricultural commodities it exports Second the principal gains actually arise from tradereforms that are essentially unilateral in nature Third the impact is likely to be pro-poorAlthough the largest proportional gains are to the urban self-employed there aresignificant gains in agriculture that benefit almost all categories of rural household

1 IntroductionUganda is in many respects a typical sub-Saharan African country it is low-income dependent on aid tends to have a persistent trade deficit is dominatedby agricultural production and processing with a small manufacturing sector andmerchandise exports are principally of a few primary commodities (notably inthis case coffee) Furthermore it suffers internal security problems and iscaught up in regional conflicts therefore faces the threat of political instabilityIt is untypical in that it has exhibited high growth rates throughout most of the1990s is self-sufficient in food has made considerable progress in liberalising itstrade regime and was the first country to qualify for debt relief under the Highly-Indebted Poor Countries initiative From the context of evaluating the likelyimpact of multilateral trade liberalisation Uganda is a valid case to represent themost salient features of better performing sub-Saharan African economies

1Adam Blake is Research Fellow Christal DeHaan Tourism and Travel Research Institute Nottingham

University Business School E-mail AdamBlakenottinghamacuk for correspondence Andrew McKay isSenior Lecturer and Internal CREDIT Fellow and Oliver Morrissey is Reader and Director of CREDIT both inthe School of Economics University of Nottingham Useful comments were received from anonymous refereesof this journal and participants at the DESG Annual Conference 2000 University of Nottingham 27-29 March2000 Adam Blakersquos work was supported by a DfID grant (CNTR 96 0494A) when he was a CREDIT ResearchFellow

366 Adam Blake Andrew McKay and Oliver Morrissey

In the context of the potential impact on a low-income country such asUganda the major Uruguay Round (UR) commitments as embodied in GATT1994 were those for liberalisation of trade in agricultural commodities Inpractice little real agricultural liberalisation has been implemented to dateNevertheless we estimate the impact on Uganda assuming UR commitments areimplemented by 2002 using a base model for 1997 (projected from a 1992 database) The reforms incorporated in the simulations are changes in world pricesas a result of all countries implementing the UR and unilateral liberalisation(Uganda reduces its own tariffs)

Section 2 discusses the commodity composition of Ugandan trade and outlinesthe potential effects of multilateral trade liberalisation (MTL) using a simplepartial analysis This section is intended to provide background information onthe Ugandan economy and the partial equilibrium estimates offer a context andprecursor to the CGE estimates We outline the structure of the computablegeneral equilibrium (CGE) model briefly in section 3 while section 4 presentsthe results of using this to estimate the impact of MTL on Uganda The mainresults of the paper are summarised in Section 5

2 Commodity Composition of Trade and Effects of LiberalisationThe Museveni regime that came to power in Uganda in 1987 was quick to beginquite radical economic reforms with major phases of macroeconomic reformand liberalisation (associated with World Bank adjustment programmes) startingin the early 1990s In regard to trade export diversification was advocated t oinsulate the economy from adverse terms of trade and instability in exportearnings associated with commodity concentration (notably in coffee) The aimhas been to move away from traditional (coffee tea cotton and tobacco)towards non-traditional exports mainly composed of agricultural commoditiessuch as sesame seeds maize beans horticulture and fish The Ugandan economyis dominated by agricultural activities as is the commodity composition ofexports The share of agricultural products declined from about 90 per cent ofexports in the late 1980s to about 70 per cent in 1992 but recovered to over 75per cent in the mid-90s this is largely a reflection of the variability in coffeeproduction and earnings

With the increase in coffee prices and growth of non-traditional exports sincethe mid-90s the trade balance has improved (although the composition ofexports is still dominated by agriculture) There has been a significant rise inexport revenue from non-traditional exports from $71 million in 1993representing 35 per cent of total export earnings to $151 million in 1995although then only 27 per cent of total export earnings (Sharer et al 1995)Traditional exports especially coffee still constitute the largest share offoreign exchange earnings (and are likely to remain so for the foreseeablefuture) accounting for over 70 per cent of commodity exports for most of theperiod between 1989 and 1995 Consequently export earnings are very

The Impact on Uganda of Agricultural Trade Liberalisation 367

sensitive to trends in world prices for these traditional commodities (Morrisseyand Rudaheranwa 1998)

Significant trade reforms that are unilateral in nature have been implemented inUganda since 1992 This was intended amongst other things to reverse andeven eliminate the trade deficit through increasing export earnings Incentivesgeared towards export-oriented sectors combined with market-determinedexchange rate policies are expected to encourage both traditional and non-traditional exports Although merchandise exports continued to decline whilstimports remained steady throughout the period from 1988 to 1992 thereafterthe value of exports improved markedly (Morrissey and Rudaheranwa 1998)Nevertheless exports remained at around ten per cent of GDP in the 1990swhereas imports were 20 per cent or more The persistent trade deficit reflectsthe composition of Ugandarsquos export basket (primarily coffee and other cashcrops) and import basket (manufactures equipment and machinery) and theimpact of deteriorating terms of trade

Multilateral implementation of UR commitments will have little impact on theprices of Ugandan imports as these are primarily manufactures and fuels (theexceptions being agricultural products which are less than ten per cent of totalimports and textiles which are less than five per cent) Uganda itself does notinfluence world prices for primary commodities Increasing exports ofagricultural crops is not simply a trade policy issue It is also an issue ofagricultural policy such as providing education and extension services to smalldispersed farmers and encouraging adoption of the most appropriatetechnology to ensure high quality output Trade policy contributes to improvingprice incentives but other policy interventions are necessary to relaxconstraints and ensure supply response (McKay et al 1997) Some of these willrelate to agricultural policy including improved provision of inputs Improvedinfrastructure and institutional support such as for coffee transport andmarketing are also important components of export promotion While manyreforms of this nature have been implemented in Uganda in the 1990s andthere is some evidence that efficiency in agriculture has increased these are notincorporated into our model (as it is based on pre-reform 1992 data)

Impact of Changes in World PricesThe easiest way of judging the possible impact of MTL on Uganda is to considerpossible effects on the price of exports and imports and evaluate the potentialresponse This is a simple partial equilibrium approach but at least providesguidance There are three stages First we identify the products likely to beaffected (ie to experience changes in world prices due to MTL) and select afigure for the volume of trade in those products for a base year in our case1995 We use the average volume traded for the four principal exportcommodities over 1994-96 (to smooth the series) as our base year estimateSuch data are not available for textiles or cereals (the only import commodities


likely to be affected) However for an average over 1994-96 we do know thatcereal and cereal preparations imports were worth $422m and textile importswere worth $316m average total imports were $10095m so these productsrepresent 42 per cent and 32 per cent respectively (Uganda Statisti calAbstract 1998 p 77)

Second we require an estimate of the price change where available we use therange from estimates reported in other studies otherwise we use our ownestimates Third we require an estimate of potential supply response There areno available estimates for Uganda Relevant information is provided in McKayet al (1999) who estimated aggregate agricultural supply response for Tanzaniadistinguishing export and food crops They found that food crop production wasresponsive to changes in the relative price of food to export crops with ashort-run elasticity of 039 and a long-run elasticity of 092 However theywere unable to estimate supply response for aggregate export crops the(downward) trend in output was almost entirely explained (statistically ) by atime trend As the time trend captured a secular decline in world prices this doesnot imply that producers of export crops are not responsive (rather the data donot permit one to estimate the response) Nevertheless the resultsdemonstrated that farmers are in principle quite responsive to prices and theaggregate agricultural output response was 035 We adopt this as our estimateof agricultural supply response for Ugandan producers

Combining all of these we can estimate the overall impact of MTL on theprincipal products affected for Uganda The results are summarised in Table 1for the principal export commodities Perhaps the most important point t onote is the very wide range of estimates of the impact of MTL on coffee prices(although estimates were not available a similar wide range should be expectedto apply to the other commodities) Obviously the actual impact on Ugandawill depend on how world prices change and MTL is only one (and not the mostimportant) factor affecting world commodity prices and also on the ability ofUgandan producers to respond As coffee is such a dominant crop for Ugandaand there is a tenfold difference in the estimated effects of MTL on coffeeprices the overall impact depends heavily on what happens to coffee prices(the impact associated with other crops is relatively small) As a result of MTLUgandan coffee exports could increase by between $05m and $3m relative to a1995 base Total exports (of the four commodities) could increase by between$06m and $32m Even at the maximum this is no more than 05 per cent oftotal exports This effect is quite small albeit positive

We have insufficient data to estimate the corresponding effect of MTL onimports using the above approach for exports However estimates show that ifagricultural liberalisation in developed countries is implemented the priceeffects on grain and cereals could be quite significant (see sources for pricechanges listed under Table 1) The lowest estimated increase is two per cent


(under partial liberalisation) ranging up to 26 per cent for full liberalisation (avery long-run scenario) On the not unreasonable assumption that Ugandanfarmers are as responsive as Tanzanian farmers the increased price of cerealimports should encourage increased production in Uganda (there is potential t oexpand production of millet maize and sorghum) This is especially true as foodcrop prices are likely to increase by more than export crop prices Ifappropriate agricultural policies are in place to provide incentives to localfarmers and facilitate increased production an increase in world cereal pricesneed not disadvantage Uganda as it has the potential to displace more expensiveimports with increased domestic production Although Uganda may experiencesome terms of trade losses there is unlikely to be a net adverse effect In factgiven that the impact is likely to be beneficial to agriculture and that is the coreof the Ugandan economy MTL could be expected to benefit Uganda

Table 1 Partial Impact of World Price Changes on UgandaExport Crop Coffee Tea Cotton Tobacco

Base volume (tonnes) 213965 12088 6331 3225Base price ($Kg) 182 102 101 212Price change () 400 290 233 -088

041 234Volume change () 140 102 082 -031

014 082Impact ($m) 300 013 012 -006

055 010 exports 077 106 188 088

014 081Notes Base volume is 1994-96 average

Price change - max and min estimates from Zietz and Valdes (1986 1990) and UNCTAD (1990) theestimate for cotton is from Brandatildeo and Martin (1993) estimate for tobacco from Blake et al (1999)Volume change - assumes supply response elasticity of 035 for all crops calculated as 035 times theprice changeImpact - the value of the volume change at the new price ie (base volume times volume change) times(base price times price change) using maximum and minimum estimates of price changes whereavailable Exports - impact ($m) expressed as a percentage of the value ($m) of base year exports of thecommodity

3 The CGE Model for UgandaThe primary data sources were the 1992 input-output (IO) table of Uganda andhousehold data from the 1992-93 Ugandan Integrated Household Survey TheIO table gives data at a 50-sector level of aggregation and includes separate IOmatrices for both domestic and imported goods While the householdexpenditure survey contains data at a more detailed level it has been aggregatedto match the 50-sector IO table In doing this ten household groupings havebeen identified with six categories of labour

Table 2 lists the factors of production giving the percentage of factor returnscontributed by each of five broad sector classifications Labour is definedaccording to literacy levels (low medium and high) and according to wage andnon-wage income Overall factor returns are principally from ldquoother primaryrdquo(mainly food crops) and services Wage labour is predominantly in services


although a significant share is in processing and manufacturing There are twomajor types of non-wage employment in Uganda subsistence agriculture whichleads to a concentration of non-wage low and medium literacy labour in ldquootherprimaryrdquo and informal sector or self-employment (mostly in services) Wageincome particularly of labour with high literacy is concentrated in the servicessector

Table 2 Factors of production identified in the modelPercentage of factor returns from (sectors)

Cashcrops

Otherprimary

Processing ofagricultural products

Manufacturing Services

1 Wage earning low literacy 9 25 13 2 51

2 Wage earning medium literacy 4 10 7 8 71

3 Wage earning high literacy 0 1 3 5 91

4 Non-wage low literacy 5 87 3 1 4

5 Non-wage medium literacy 4 81 2 2 11

6 Non-wage high literacy 2 44 2 2 49

7 Capital (includes land) 2 36 5 3 55

All Factors 3 47 3 3 44

Table 3 shows the definitions for the ten household groups defined in the modelThe total income of each group as a proportion of total household income(column 1) is disaggregated into the proportion of household income earnedfrom wage labour (column 2) and from highly literate labour (column 3) As agreater proportion of the population live in rural than in urban areas moreruralagricultural households are defined than urban ones The first twohousehold groups are defined as wage earners from urban and rural areas Forthese groups wage income accounts for 89 per cent and 69 per cent respectivelyof income and this is predominantly from highly literate labour Agriculturalhouseholds are classified for each of the four regions The Northern region isthe poorest this region is semi-arid is the main area for cotton production(which performed poorly until the late 1990s) and is unstable (subject to rebelactivities) Non-working households are a miscellaneous category this mayinclude some of the poorest (such as households headed by widows or the old)but also includes households living on remittances and those that do not ldquoneedrdquoto work (ie they have non-earned income)


Table 3 Households identified in the modelShare of household

type in total householdincomes

Wage income as aproportion of household

income

Income from highly literatelabour as a proportion of

household income1 Urban wage earners 12 89 79

2 Rural wage earners 10 69 57

3 Agricultural central 14 4 14

4 Agricultural eastern 14 4 16

5 Agricultural western 15 3 11

6 Agricultural northern 9 3 20

7 Urban non-farm self employed 12 1 49

8 Rural non-farm self employed 9 0 39

9 Urban non-working 1 0 4

10 Rural non-working 3 0 5

All households 100 20 33

The representation of production in the model follows fairly standard structuresin the CGE modelling literature (see Blake et al 1999 Goldin and Knudsen1990 Martin and Winters 1996) We assume constant returns to scaletechnology and perfect competition Output is modelled as a Leontiefcombination of intermediate inputs and aggregate value added and aggregatevalue added as a CES combination of different factors The Armingtonassumption is invoked to differentiate imports from domestically producedgoods and also to differentiate exports from goods for domestic use In bothimport and export markets Uganda is modelled as a small country that cannotinfluence world prices

Factor markets are modelled as displaying imperfect mobility Capital isconsidered to be specific to the sector it is used in while the six categories oflabour are each characterised by a transformation frontier between sectors ofuse2 The government collects tax revenue from production taxes importtariffs and income taxes and uses this revenue on public expenditure to financethe trade deficit3 and to provide transfers to households

2 In fact this is a constant elasticity of transformation (CET) frontier The elasticities of transformation varyaccording to level of literacy and wagenon-wage labour The elasticities that we use are 025 05 and 075 forlow medium and high literacy non-wage labour and 05 1 and 15 for the corresponding wage labourcategories These values imply significant rigidity in the non-wage and low-literacy wage labour markets3 A standard way of explaining the trade deficit in CGE models is that the trade deficit equals net foreign savings(X-M = S-I) Normally the lsquorepresentativersquo household would finance this As we have ten households and asmost foreign savings (notably aid receipts) in Uganda are through the government we model the public sector asfinancing this deficit Furthermore the current account deficit is a constant (for model closure) implying thatexports can only change if matched by imports


Each of the ten household groups receives income from the seven factors ofproduction (of which three are wage labour three non-wage labour and one iscapital) transfers from other household groups and transfers from thegovernment Each household group pays income tax and exhausts its remainingincome on either savings transfers to other households and expenditure ongoods and services This consumption is represented by a LES functioncalibrated to income elasticities The Armington assumption again differentiatesimports from domestically produced goods

Table 4 Sector data and average growth rates 1992-1997--------from 1992 Data-------- --------from 1997 Data-------- Baseline

simulation resultExports

totalexports

Imports total imports

Valueadded

GDP

GDPgrowth92-97

GDP pcgrowth92-97

Export price92-97

TFP growth 92-97

(1) (2) (3) (4)

Coffee 347 00 17 107 76 66 10

Other cash crops 97 11 21 138 107 34 66

Food 141 47 484 33 01 00 -10

Other primary 55 13 39 62 31 00 -02

Manufacturing 09 778 60 157 126 00 86

Services 352 152 379 98 67 00 31

Total 100 100 100 70 39 26 14

Notes 1 Average annual real GDP growth rate between 1992 and 19972 Average annual real GDP per capita growth rate 1992-19973 Average annual real export price growth rate 1992-19974 Average annual total factor productivity growth rate 1992-1997 These are calculated in the 1997

baseline scenario

1997 Baseline SimulationThe IO table and SAM are based on 1992 data so our first step is to project thisforward to represent the Ugandan economy in 1997 (based on the structuralfeatures of the 1992 IO table) the latest year for which adequate data areavailable To do this we take data on a limited number of indicators for 1992and 1997 and apply movements in these indicators to the model Theseindicators are output growth export import and GDP shares (by value added) ofthe principal sectors as indicated in Table 4 As productivity rates areunavailable they are made endogenous4 and are produced by the simulatedprojection so that the representation meets the 1997 level of GDP with thesame structure of production as is evident in the economy in 1997 Anaggregated version of the growth rates that this entails is given in Table 4

4 While other simulations that we perform use some of the same exogenous shocks introduced here this is theonly simulation where productivity is endogenous


Population figures are taken for urban and rural areas5 and the householdsrsquofactor endowments and population are expanded in these proportions Inaddition an extra 1 per cent per annum increase in factor supply is assumed forall factors over and above that implied by population growth This additionalgrowth in factor endowments is to account for skill and quality improvementsover time World price changes for Ugandarsquos major export goods (coffee teacotton and tobacco) are incorporated in the model these only affect coffee andother cash crops The implied total factor productivity (TFP) growth rates areconsistent with an economy that was performing well during this period ofanalysis

Historical DecompositionThe results from the 1997 baseline simulation are decomposed according to thedifferent ldquoshocksrdquo that we apply in the baseline simulation We perform foursimulations for the historical decomposition (i) a productivity simulationwhere the productivity rates derived in the 1997 baseline simulation are applied(at exogenous rates) with no other changes from the 1992 base (ii) a worldprice simulation where the only change from the 1992 base is to introduce1997 world prices (iii) a population simulation where population (andproportionately factor endowments) is scaled up to the 1997 level at differentrates for rural and urban households (iv) a factor simulation where theadditional factor endowment growth of one per cent per annum is introduced

Table 5 shows the results for these simulations and the 1997 baselinesimulation in terms of equivalent variations (EVs) as a percentage of 1992income for each household group and for a welfare function that is a Cobb-Douglas function of the householdsrsquo utility The baseline simulation has totalwelfare growing by 39 per cent per annum which has differential impacts onhouseholds Productivity growth and factor growth have the largest effects onwelfare and household EVs with export prices having small positive effects forall households except urban and rural non-working and population growthhaving purely redistributive effects As population growth is here accompaniedby proportionate growth in factor endowments the overall effect must be zeroon per capita welfare The exogenous population trend is that urban householdsgrow at a higher rate than rural households so factors supplied by urbanhouseholds (particularly medium and high literacy wage labour) are oversuppliedrelative to other factors of production Relative returns to these factorstherefore fall leaving urban households and particularly urban wage earnerssignificantly worse off Rural wage earners also have a negative EV as they arealso supplying the same types of labour as the urban wage earners

5 These show that rural population grew by an annual average of 26 between 1992-1997 while urbanpopulation grew by 70 pa


Table 5 Equivalent variations for households and sector output underthe 1997 baseline simulation and historical decomposition Annualisedpercentages of 1992 income

BaselineSimulation

Productivitygrowth

Exportprices

Populationgrowth

Factorgrowth

Equivalent Variation (per capita) (per capita)

1 Urban wage earners 22 21 01 -23 17

2 Rural wage earners 22 15 01 -16 18

3 Agricultural central 35 12 02 01 17

4 Agricultural eastern 36 12 02 01 17

5 Agricultural western 34 10 02 02 17

6 Agricultural northern 34 12 02 00 17

7 Urban non-farm self employed 50 29 04 -07 16

8 Rural non-farm self employed 42 21 04 -04 16

9 Urban non-working 43 22 00 -05 20

10 Rural non-working 31 10 -01 -05 23

Total (Cobb-Douglas composite) 39 16 02 00 14

Output

Coffee 75 25 18 11 26

Other cash crops 113 72 06 14 22

Food 07 -10 -01 -01 19

Other primary 22 02 -01 04 17

Manufacturing 105 72 -03 15 21

Services 60 31 00 10 19

Productivity growth has the largest effect on sectors where GDP growth hasbeen highest between 1992 and 1997 The ldquoother cash cropsrdquo andmanufacturing sectors have experienced high growth rates so have highproductivity rates in order to reach these GDP targets in the baseline simulationExport prices alone have substantial effects only on the coffee and ldquoother cashcropsrdquo sectors with small output reductions in other sectors that are competingwith cash crop sectors for factor inputs This pattern is reflected in householdEVs where those households (rural and urban self-employed) with factors(medium and high literacy non-wage labour) employed in cash crop productionhave the largest EVs

4 CGE Estimates for UgandaWe choose 1997 as the baseline because this was the most recent year for whichadequate data on the structure of the economy was available to update the 1992IO and SAM structures The 1992-97 projections represent a period in whichthe economy was performing relatively well This strong performance hascontinued at least until the end of 2000 (data for 2001 are not available) As i tis useful to simulate the effects of trade liberalisation over a period when theperformance of the economy was similar to that of 1992-97 we assume tradepolicy reforms are implemented by 2002 (and thus both sets of projections


cover periods of five years) To examine the effects of liberalisation under theUruguay Round we first project the model forward to 2002 and assume fullimplementation of the UR commitments analysed We include projections ofpopulation growth to 2002 with factor endowments increasing by the rate ofpopulation growth plus 1 per cent (to account for skill and qualityimprovements) We also model productivity changes between 1997 and 2002 atthe same growth rates as occurred in the projection from 1992 to 1997

From this projected base for 2002 we introduce changes to world prices forUgandarsquos imports and exports to represent multilateral liberalisation Inaddition we introduce across the board reductions of 24 per cent in Ugandantariffs (the UR commitment) to represent trade liberalisation that is unilateralin nature Although Uganda did reduce tariffs in the mid-90s and continued t odo so after 1997 the available information on tariff reductions is not of a formthat can be incorporated into the CGE model (most information relates t oscheduled tariffs on import commodities rather than tariffs corresponding t oproduction sectors) Our simulation is intended to capture how reforms that areunilateral in nature would have affected the economy given the structureprevailing in 19976 In fact in the late 1990s tariff reductions tended to beacross the board in contrast to the earlier reforms that reduced the highesttariff rates Thus the implementation of the UR commitment to reduce tariffsby 24 per cent corresponds in principle to what was actually executed (ifanything Uganda may have exceeded this) The total UR effect is decomposedinto the effects of (i) import price changes (ii) export price changes and (iii)Ugandan tariff reductions The results are shown in Tables 6 and 7

Considering first the overall impact the estimates show that GDP wouldincrease by 04 per cent in real terms (measured as EV per capita) Although thismay appear very slight it should be interpreted in the context of relativelyminor impacts on prices of principal exports and imports which in turn arerelatively low shares of GDP In particular exports represent about ten per centof GDP Most of this gain arises from unilateral liberalisation (reduction intariffs across the board) a 026 per cent increase in EV followed by the gainfrom lower import prices 010 per cent This is consistent with the fact thatimports are over 20 per cent of GDP In these estimates we assumed coffeeprices increased by only 04 per cent hence it is not too surprising that therewas no welfare gain from export price changes (although there was a gain in theldquoagriculture westernrdquo households the principal producers of coffee)

6 Dirty tariffication is unlikely to be an issue as the tariff rates in the model for 1997 are if anything anoverestimate of the level of protection prevailing at that time


Table 6 Percentage change in EV per capita Uruguay Round andcomponents

UruguayRound exportprice changes

Uruguay Roundimport price

changes

Uruguay RoundUgandan tariff

changes

Uruguay Round(all factors)

1 Urban wage earners 0000 0158 -0079 0079

2 Rural wage earners 0000 0079 0079 0236

3 Agricultural central 0000 0000 0358 0502

4 Agricultural eastern 0000 0000 0286 0357

5 Agricultural western 0073 0073 0436 0509

6 Agricultural northern 0000 0073 0218 0291

7 Urban non-farm self employed 0126 0252 0882 1197

8 Rural non-farm self employed 0067 0201 0669 0870

9 Urban non-working -0067 -0067 -0467 -0534

10 Rural non-working -0074 -0074 -0445 -0593

Total (Cobb-Douglas composite) 0000 0104 0259 0415

Note We conducted four sets of sensitivity analysis doubling the Armington import and export elasticitiesdoubling the elasticity of transformation of labour between sectors and doubling the elasticities ofsubstitution between factors within sectors None of the results reported in this table were found to besensitive (details available on request) signs and relative magnitudes did not change and absolutemagnitudes were altered only slightly

In terms of the distributional effects (Table 6) the welfare gains are greatest forurban and rural non-farm self-employed this is likely to reflect gains to tradersand owners of cash crop processing facilities and significant gains due tocheaper imports The only groups to suffer a welfare loss are the urban and ruralnon-working (in the Ugandan context such households are not necessarily thepoorest non-working can often be interpreted as not needing to work ratherthan being unemployed) The benefits to agriculture are greater in the main foodand coffee growing central and western regions because the factors that benefitare more prevalent in those regions Table 7 shows that the largest output gainsare in cash crops especially those other than coffee Food output is estimatedto fall slightly as factors are diverted into cash crops (but note that overallagriculture incomes rise) Manufacturing output is estimated to fall slightly asimport-competing producers face more competition Perhaps the single mostimportant conclusion is that the welfare gains arise predominantly from reformsthat are essentially unilateral trade liberalisation


Table 7 Exogenous price changes due to the Uruguay Round andpercentage changes in output for the Uruguay Round and components

Exogenous Percentage change in output per annum

Change inexport

prices dueto the UR

Change inimport

prices dueto the UR


UruguayRound importprice changes


changes


Coffee 04 04 005 -007 067 062

Other cash crops 18 27 055 -002 052 103

Food 09 68 001 002 -010 -007

Other primary -03 -06 -006 -001 012 004

Manufacturing -02 -10 -012 -024 018 -020

Services -02 -07 001 007 -016 -008

Average 04 -05

Coffee Price IncreasesIn the previous simulations we used the Brandatildeo and Martin (1993) estimate forthe effect of the UR on world prices for coffee This is rather low so wesimulate the effects of alternative coffee price increases Table 8 shows theresults from these simulations where equivalent variation (as a percentage of2002 post-UR income) is given for each household and output changes fromthe full Uruguay Round scenario are given These results indicate a low supplyresponse for coffee (a 10 per cent price increase generates an output increase of232 per cent) and that returns from such price increases go predominantly t othe agricultural households and to non-farm self employed This is lower thanthe supply response of 035 assumed in Table 1 However that figure wasderived from food crops in Tanzania whereas this one is generated from themodel while the Tanzanian study suggests that supply response for cash crops islower than that for food

One may feel that coffee supply response in Uganda should be higher thanestimated and indeed it may be greater in the now liberalised market Our resultscan be interpreted as conservative estimates given the assumption that factormobility is limited (and land is constrained) The quality of the coffee tree stockin Uganda is quite low and planting new trees is both a slow process and wouldnot be reflected in yields for some five years The coffee sector in Uganda hasbeen fully liberalised in the early 1990s coffee production was controlled by thestate-owned Coffee Marketing Board whereas by the late 1990s some 90 percent of trade was in the hands of five multinationals Coffee production roughlydoubled between 1992 and 1997 as did the share of the world price received byproducers (Morrissey and Rudaheranwa 1998) This gives rise to cautiousoptimism for the future of the sector (although the decline in world coffeeprices in 2001 is ominous)


Table 8 The effects of additional coffee export price increasesAdditional coffee export price increase 2 4 6 8 10

Equivalent Variation

1 Urban wage earners 000 000 000 000 000

2 Rural wage earners 008 008 008 008 008





7 Urban non-farm self employed 019 037 056 075 087

8 Rural non-farm self employed 013 033 046 060 080

9 Urban non-working 000 000 000 000 000

10 Rural non-working 000 000 -007 -007 -007


Output

Coffee 050 097 145 190 232

Other cash crops -017 -032 -048 -063 -077

Food -001 -006 -006 -007 -012

Other primary -006 -015 -021 -025 -030

Manufacturing -015 -032 -048 -066 -081

Services 002 003 005 007 010

Note Signs and relative magnitudes of the results reported in this table did not change noticeably in thesensitivity analysis (see Table 6) As would be expected the supply response of coffee output wasgreater if labour markets were more flexible or of transformation elasticities were higher The generalresults and distributional effects were not altered

5 ConclusionsThe results presented here are based on simulations from a CGE model ofUganda It is interesting to note that the low impact estimates of the CGEmodel are comparable to the simple estimates based on an examination of thecomposition of trade We can make three broad conclusions First the impactof multilateral trade liberalisation on a low-income country such as Ugandaappears to be quite slight although it is positive This should not be surprisingbecause although the country is predominantly agriculture-based theliberalisation of world trade in agricultural commodities has little impact on theworld prices of the products Uganda exports This general result would alsoapply to other low-income countries especially those in sub-Saharan Africa butgenerally those exporting tropical agricultural commodities Furthermore low-income countries face constrained supply response and many other domesticpolicy reforms would be required to avail them of the increased opportunitiesassociated with multilateral liberalisation of trade in agricultural commoditieswhile benefits of enhanced market access would not be captured in the model (asthey are not reflected in changes in world prices)

Second the greatest share of gains in welfare actually arises from reforms thatare essentially unilateral in nature The results can be interpreted as indicating


the potential gains from reducing tariffs against the rest of the world even ifother countries do not reciprocate During the 1990s Uganda did simplify andreduce tariffs quite significantly certainly by more than the 24 per centreduction simulated here This liberalisation did improve the relative incentivesfor agricultural producers and would have benefited manufacturers usingimported inputs Given the results of our model the trade policy reformsimplemented by Uganda in the 1990s are likely to have had positive welfareeffects

Third there are distributional consequences of the impact of trade liberalisationAlthough the largest proportional gains are to the urban self-employed thereare significant gains in agriculture The benefits to agriculture are greater in themain food and coffee growing central and western regions because the factorsthat benefit are more prevalent in those regions This is consistent with theevidence that agricultural growth and poverty reduction in Uganda in the1990s was concentrated in these areas (whereas the Northern region fared leastwell) Appleton (1998) reports a decline in the poverty headcount index from56 per cent in 1992 to 46 per cent in 1996 due largely to growth especially incash crop production (coffee) However there is no evidence of a decline inincome inequality and the poorest quintile have experienced falling livingstandards especially households with a non-working head (AIDS is an importantfactor here) The Northern regions have benefited least It is encouraging thatour results are consistent with this evidence even to the extent that we identifynon-working households as the major losers (although not all of thesehouseholds are the poorest)

Our estimates of the effect of the Uruguay Round on Uganda suggest that theoverall effect will be small but on balance positive As there are major benefitsto agriculturerural households the impact is likely to be pro-poor Althoughmultinationals now dominate coffee marketing in Uganda there is competition(at least five multinationals have a significant presence) and prices paid t ofarmers have been significantly increased The effect for specific products orsectors depends on how the relevant world prices change and the flexibility ofthe Ugandan economy in allowing and encouraging (via the structure ofincentives) producers to respond to relative price changes Uganda is likely t obenefit because prices for the cash crops that it exports will rise and increases infood prices can stimulate domestic farmers to increase production (imports falldomestic production rises and possibly even exports increase) Howeverproducers respond not so much to the prices they face for individualcommodities but the relative prices faced for substitute commodities The CGEmodels address the effects associated with relative price changes As shown theoutcome depends on the magnitude of relative price changes and the ability ofproducers to respond (as represented by factor mobility) In general Uganda willbenefit It will benefit by more the greater the increase in relative prices of thegoods it exports It will benefit even more the greater the degree of factor


mobility between sectors in the economy Extended analysis based on a refinedmodel is important to evaluate these conclusions and test their robustness

The model also helps to identify appropriate non-trade policies In the leastdeveloped countries agricultural policy reform is required to increase theincentives facing domestic producers and to remove the constraints (such asaccess to inputs credit and technology) that peasant farmers tend to be subjectto Unilateral trade liberalisation within poor countries is desirable in that i tpermits a more efficient allocation of resources and allows world prices to betransmitted to domestic producers While there will be short-run adjustmentcosts as the losses to contracting sectors tend to materialise faster than thegains to expanding sectors the long-run effect on welfare should be positiveOur results suggest that even the short-run impact will be positive There arehowever likely to be substantial redistribution effects although in Uganda theseare likely to be mostly within agriculture as farmers shift from one crop toanother Policies to facilitate internal factor mobility are shown to beimportant and more so the greater the increase in relative price incentives Itshould not be forgotten that perhaps the most important policies are those indeveloped countries Agricultural trade liberalisation in the EU for example willhave a greater impact on world prices and even on low-income countries thanpolicy reforms in low-income countries themselves The developed countryreforms alter the world trade environment reforms in developing countries canenhance their ability to benefit from a more open global trading environment

ReferencesAppleton S (1998) Changes in Poverty in Uganda 1992-96 University of Oxford CSAE

Working Paper 98-15Blake A Rayner AJ and Reed GV (1999) A Computable General Equilibrium

Analysis Of Agricultural Liberalisation The Uruguay Round And CommonAgricultural Policy Reform Journal of Agricultural Economics 50 3 400-424

Brandatildeo and Martin (1993) Implications of Agricultural Trade Liberalisation for theDeveloping Countries mimeo The World Bank

Goldin I and Knudsen O (eds) (1990) Agricultural Trade Liberalization Implicationsfor Developing Countries Paris OECD

Martin W and Winters A (1996) The Uruguay Round and the Developing CountriesCambridge Cambridge University Press

McKay A Morrissey O and Vaillant C (1997) Trade Liberalisation and AgriculturalSupply Response Issues and Some Lessons European Journal of DevelopmentResearch 92 129-147

McKay A Morrissey O and Vaillant C (1999) Aggregate Supply Response inTanzanian Agriculture Journal of International Trade and Economic Development81 107-123

Morrissey O and Rudaheranwa N (1998) Ugandan Trade Policy and Export Performancein the 1990s University of Nottingham CREDIT Research Paper 9812

Sharer R De Zoysa R and McDonald H (1995) Uganda Adjustment with Growth1987-94 Washington DC International Monetary Fund Occasional Paper 121

UNCTAD (1990) Agricultural Trade Liberalization in the Uruguay Round Implicationsfor Developing Countries UNCTADUNDWIDER United Nations Publications


Zietz J and Valdes A (1986) The Costs of Protectionism to Developing Countries anAnalysis for Selected Agricultural Products World Bank Staff Working Papers No769

Zietz J and Valdes A (1990) International Interactions in Food and Agricultural PoliciesEffects of Alternative Policies in Goldin and Knudsen (eds)


In the context of the potential impact on a low-income country such asUganda the major Uruguay Round (UR) commitments as embodied in GATT1994 were those for liberalisation of trade in agricultural commodities Inpractice little real agricultural liberalisation has been implemented to dateNevertheless we estimate the impact on Uganda assuming UR commitments areimplemented by 2002 using a base model for 1997 (projected from a 1992 database) The reforms incorporated in the simulations are changes in world pricesas a result of all countries implementing the UR and unilateral liberalisation(Uganda reduces its own tariffs)

Section 2 discusses the commodity composition of Ugandan trade and outlinesthe potential effects of multilateral trade liberalisation (MTL) using a simplepartial analysis This section is intended to provide background information onthe Ugandan economy and the partial equilibrium estimates offer a context andprecursor to the CGE estimates We outline the structure of the computablegeneral equilibrium (CGE) model briefly in section 3 while section 4 presentsthe results of using this to estimate the impact of MTL on Uganda The mainresults of the paper are summarised in Section 5

2 Commodity Composition of Trade and Effects of LiberalisationThe Museveni regime that came to power in Uganda in 1987 was quick to beginquite radical economic reforms with major phases of macroeconomic reformand liberalisation (associated with World Bank adjustment programmes) startingin the early 1990s In regard to trade export diversification was advocated t oinsulate the economy from adverse terms of trade and instability in exportearnings associated with commodity concentration (notably in coffee) The aimhas been to move away from traditional (coffee tea cotton and tobacco)towards non-traditional exports mainly composed of agricultural commoditiessuch as sesame seeds maize beans horticulture and fish The Ugandan economyis dominated by agricultural activities as is the commodity composition ofexports The share of agricultural products declined from about 90 per cent ofexports in the late 1980s to about 70 per cent in 1992 but recovered to over 75per cent in the mid-90s this is largely a reflection of the variability in coffeeproduction and earnings

With the increase in coffee prices and growth of non-traditional exports sincethe mid-90s the trade balance has improved (although the composition ofexports is still dominated by agriculture) There has been a significant rise inexport revenue from non-traditional exports from $71 million in 1993representing 35 per cent of total export earnings to $151 million in 1995although then only 27 per cent of total export earnings (Sharer et al 1995)Traditional exports especially coffee still constitute the largest share offoreign exchange earnings (and are likely to remain so for the foreseeablefuture) accounting for over 70 per cent of commodity exports for most of theperiod between 1989 and 1995 Consequently export earnings are very















041 234Volume change () 140 102 082 -031

014 082Impact ($m) 300 013 012 -006

055 010 exports 077 106 188 088








Cashcrops

Otherprimary
















income




















totalexports


Valueadded

GDP

GDPgrowth92-97

GDP pcgrowth92-97

Export price92-97

TFP growth 92-97

(1) (2) (3) (4)

Coffee 347 00 17 107 76 66 10

Other cash crops 97 11 21 138 107 34 66

Food 141 47 484 33 01 00 -10

Other primary 55 13 39 62 31 00 -02

Manufacturing 09 778 60 157 126 00 86

Services 352 152 379 98 67 00 31

Total 100 100 100 70 39 26 14


baseline scenario










BaselineSimulation

Productivitygrowth

Exportprices

Populationgrowth

Factorgrowth













Output

Coffee 75 25 18 11 26


Food 07 -10 -01 -01 19



Services 60 31 00 10 19












changes


changes


















Change inexport

prices dueto the UR

Change inimport

prices dueto the UR




changes


Coffee 04 04 005 -007 067 062

Other cash crops 18 27 055 -002 052 103

Food 09 68 001 002 -010 -007

Other primary -03 -06 -006 -001 012 004

Manufacturing -02 -10 -012 -024 018 -020

Services -02 -07 001 007 -016 -008

Average 04 -05

















Output

Coffee 050 097 145 190 232

Other cash crops -017 -032 -048 -063 -077

Food -001 -006 -006 -007 -012

Other primary -006 -015 -021 -025 -030

Manufacturing -015 -032 -048 -066 -081

Services 002 003 005 007 010







































041 234Volume change () 140 102 082 -031

014 082Impact ($m) 300 013 012 -006

055 010 exports 077 106 188 088








Cashcrops

Otherprimary
















income




















totalexports


Valueadded

GDP

GDPgrowth92-97

GDP pcgrowth92-97

Export price92-97

TFP growth 92-97

(1) (2) (3) (4)

Coffee 347 00 17 107 76 66 10

Other cash crops 97 11 21 138 107 34 66

Food 141 47 484 33 01 00 -10

Other primary 55 13 39 62 31 00 -02

Manufacturing 09 778 60 157 126 00 86

Services 352 152 379 98 67 00 31

Total 100 100 100 70 39 26 14


baseline scenario










BaselineSimulation

Productivitygrowth

Exportprices

Populationgrowth

Factorgrowth













Output

Coffee 75 25 18 11 26


Food 07 -10 -01 -01 19



Services 60 31 00 10 19












changes


changes


















Change inexport

prices dueto the UR

Change inimport

prices dueto the UR




changes


Coffee 04 04 005 -007 067 062

Other cash crops 18 27 055 -002 052 103

Food 09 68 001 002 -010 -007

Other primary -03 -06 -006 -001 012 004

Manufacturing -02 -10 -012 -024 018 -020

Services -02 -07 001 007 -016 -008

Average 04 -05

















Output

Coffee 050 097 145 190 232

Other cash crops -017 -032 -048 -063 -077

Food -001 -006 -006 -007 -012

Other primary -006 -015 -021 -025 -030

Manufacturing -015 -032 -048 -066 -081

Services 002 003 005 007 010




























Cashcrops

Otherprimary
















income




















totalexports


Valueadded

GDP

GDPgrowth92-97

GDP pcgrowth92-97

Export price92-97

TFP growth 92-97

(1) (2) (3) (4)

Coffee 347 00 17 107 76 66 10

Other cash crops 97 11 21 138 107 34 66

Food 141 47 484 33 01 00 -10

Other primary 55 13 39 62 31 00 -02

Manufacturing 09 778 60 157 126 00 86

Services 352 152 379 98 67 00 31

Total 100 100 100 70 39 26 14


baseline scenario










BaselineSimulation

Productivitygrowth

Exportprices

Populationgrowth

Factorgrowth













Output

Coffee 75 25 18 11 26


Food 07 -10 -01 -01 19



Services 60 31 00 10 19












changes


changes


















Change inexport

prices dueto the UR

Change inimport

prices dueto the UR




changes


Coffee 04 04 005 -007 067 062

Other cash crops 18 27 055 -002 052 103

Food 09 68 001 002 -010 -007

Other primary -03 -06 -006 -001 012 004

Manufacturing -02 -10 -012 -024 018 -020

Services -02 -07 001 007 -016 -008

Average 04 -05

















Output

Coffee 050 097 145 190 232

Other cash crops -017 -032 -048 -063 -077

Food -001 -006 -006 -007 -012

Other primary -006 -015 -021 -025 -030

Manufacturing -015 -032 -048 -066 -081

Services 002 003 005 007 010





























income




















totalexports


Valueadded

GDP

GDPgrowth92-97

GDP pcgrowth92-97

Export price92-97

TFP growth 92-97

(1) (2) (3) (4)

Coffee 347 00 17 107 76 66 10

Other cash crops 97 11 21 138 107 34 66

Food 141 47 484 33 01 00 -10

Other primary 55 13 39 62 31 00 -02

Manufacturing 09 778 60 157 126 00 86

Services 352 152 379 98 67 00 31

Total 100 100 100 70 39 26 14


baseline scenario










BaselineSimulation

Productivitygrowth

Exportprices

Populationgrowth

Factorgrowth













Output

Coffee 75 25 18 11 26


Food 07 -10 -01 -01 19



Services 60 31 00 10 19












changes


changes


















Change inexport

prices dueto the UR

Change inimport

prices dueto the UR




changes


Coffee 04 04 005 -007 067 062

Other cash crops 18 27 055 -002 052 103

Food 09 68 001 002 -010 -007

Other primary -03 -06 -006 -001 012 004

Manufacturing -02 -10 -012 -024 018 -020

Services -02 -07 001 007 -016 -008

Average 04 -05

















Output

Coffee 050 097 145 190 232

Other cash crops -017 -032 -048 -063 -077

Food -001 -006 -006 -007 -012

Other primary -006 -015 -021 -025 -030

Manufacturing -015 -032 -048 -066 -081

Services 002 003 005 007 010





























totalexports


Valueadded

GDP

GDPgrowth92-97

GDP pcgrowth92-97

Export price92-97

TFP growth 92-97

(1) (2) (3) (4)

Coffee 347 00 17 107 76 66 10

Other cash crops 97 11 21 138 107 34 66

Food 141 47 484 33 01 00 -10

Other primary 55 13 39 62 31 00 -02

Manufacturing 09 778 60 157 126 00 86

Services 352 152 379 98 67 00 31

Total 100 100 100 70 39 26 14


baseline scenario










BaselineSimulation

Productivitygrowth

Exportprices

Populationgrowth

Factorgrowth













Output

Coffee 75 25 18 11 26


Food 07 -10 -01 -01 19



Services 60 31 00 10 19












changes


changes


















Change inexport

prices dueto the UR

Change inimport

prices dueto the UR




changes


Coffee 04 04 005 -007 067 062

Other cash crops 18 27 055 -002 052 103

Food 09 68 001 002 -010 -007

Other primary -03 -06 -006 -001 012 004

Manufacturing -02 -10 -012 -024 018 -020

Services -02 -07 001 007 -016 -008

Average 04 -05

















Output

Coffee 050 097 145 190 232

Other cash crops -017 -032 -048 -063 -077

Food -001 -006 -006 -007 -012

Other primary -006 -015 -021 -025 -030

Manufacturing -015 -032 -048 -066 -081

Services 002 003 005 007 010
































BaselineSimulation

Productivitygrowth

Exportprices

Populationgrowth

Factorgrowth













Output

Coffee 75 25 18 11 26


Food 07 -10 -01 -01 19



Services 60 31 00 10 19












changes


changes


















Change inexport

prices dueto the UR

Change inimport

prices dueto the UR




changes


Coffee 04 04 005 -007 067 062

Other cash crops 18 27 055 -002 052 103

Food 09 68 001 002 -010 -007

Other primary -03 -06 -006 -001 012 004

Manufacturing -02 -10 -012 -024 018 -020

Services -02 -07 001 007 -016 -008

Average 04 -05

















Output

Coffee 050 097 145 190 232

Other cash crops -017 -032 -048 -063 -077

Food -001 -006 -006 -007 -012

Other primary -006 -015 -021 -025 -030

Manufacturing -015 -032 -048 -066 -081

Services 002 003 005 007 010


































changes


changes


















Change inexport

prices dueto the UR

Change inimport

prices dueto the UR




changes


Coffee 04 04 005 -007 067 062

Other cash crops 18 27 055 -002 052 103

Food 09 68 001 002 -010 -007

Other primary -03 -06 -006 -001 012 004

Manufacturing -02 -10 -012 -024 018 -020

Services -02 -07 001 007 -016 -008

Average 04 -05

















Output

Coffee 050 097 145 190 232

Other cash crops -017 -032 -048 -063 -077

Food -001 -006 -006 -007 -012

Other primary -006 -015 -021 -025 -030

Manufacturing -015 -032 -048 -066 -081

Services 002 003 005 007 010




























Change inexport

prices dueto the UR

Change inimport

prices dueto the UR




changes


Coffee 04 04 005 -007 067 062

Other cash crops 18 27 055 -002 052 103

Food 09 68 001 002 -010 -007

Other primary -03 -06 -006 -001 012 004

Manufacturing -02 -10 -012 -024 018 -020

Services -02 -07 001 007 -016 -008

Average 04 -05

















Output

Coffee 050 097 145 190 232

Other cash crops -017 -032 -048 -063 -077

Food -001 -006 -006 -007 -012

Other primary -006 -015 -021 -025 -030

Manufacturing -015 -032 -048 -066 -081

Services 002 003 005 007 010







































Output

Coffee 050 097 145 190 232

Other cash crops -017 -032 -048 -063 -077

Food -001 -006 -006 -007 -012

Other primary -006 -015 -021 -025 -030

Manufacturing -015 -032 -048 -066 -081

Services 002 003 005 007 010

























The Impact on Uganda of Agricultural Trade Liberalisation

Documents

Transcript of The Impact on Uganda of Agricultural Trade Liberalisation