ERR77.pdf

7/29/2019 ERR77.pdf

1/42

United StatesDepartment oAgriculture

Economic

ResearchService

EconomicResearchReportNumber 77

July 2009

Trade and Development When

Exports Lack Diversifcation

A Case Study rom Malawi

Suresh Chand Persaud

Birgit Gisela Saager Meade

7/29/2019 ERR77.pdf

2/42

www

.ers

.usd

a.govVisit Our Website To Learn More!

National Agricultural Library

Cataloging Record:

The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs

and activities on the basis of race, color, national origin, age, disability, and, where

applicable, sex, marital status, familial status, parental status, religion, sexualorientation, genetic information, political beliefs, reprisal, or because all or a part of an

individual's income is derived from any public assistance program. (Not all prohibited

bases apply to all programs.) Persons with disabilities who require alternative means

for communication of program information (Braille, large print, audiotape, etc.) shouldcontact USDA's TARGET Center at (202) 720-2600 (voice and TDD).

To file a complaint of discrimination write to USDA, Director, Office of Civil Rights, 1400Independence Avenue, S.W., Washington, D.C. 20250-9410 or call (800) 795-3272

(voice) or (202) 720-6382 (TDD). USDA is an equal opportunity provider and employer.

Persaud, Suresh Chand, 1969-Trade and development when exports lack diversification.(Economic research report (United States. Dept. of Agriculture.Economic Research Service) ; no. 77)

1. Tobacco industryMalawi. 2. Economic developmentMalawiCase studies.3. MalawiEconomic conditions. I. Meade, Birgit Gisela SaagerII. United States. Dept. of Agriculture. Economic Research Service.

III. Title.HD9146.M32

Photo credits: Howard F. Schwartz, Colorado State University,Bugwood.org (tobacco field); Claudio Angelini (truck and bales);and Jupiterimages (tobacco leaf).

7/29/2019 ERR77.pdf

3/42

United States

Department

of Agriculture

www.ers.usda.gov

A Report from the Economic Research Service

Abstract

Developing countries, particularly those that depend heavily on a small number o agri-cultural exports, are vulnerable to domestic and international shocks. These countriesoten have diculty achieving sustained economic growth. This analysis uses Malawi,a country that earns most o its oreign exchange rom tobacco, as a case study o exportconcentration and heavy exposure to volatility. The econometric results suggest that the

decline in Malawis gross domestic product (GDP) when tobacco exports are allingis almost three times greater than the increase in GDP when exports are rising. Model-based simulations indicate that variability in tobacco exports leads to slower economicgrowth because GDP alls by a relatively large amount in response to a given decreasein exports, while recovering little during an upswing in exports. Gains in tobacco yieldand improvements in marketing eciency, however, can help buer Malawis GDP romvariability in export revenues.

Keywords: Malawi, tobacco, export-led growth, asymmetry, volatility, productivity,trade, development, marketing eciency, price transmission

Acknowledgments: The authors extend their appreciation to sta members o the ERSMarket and Trade Economics Division Suchada Langley, Chie o the Agricultural Policyand Models branch, Mary Anne Normile, Chie o the International Demand & Tradebranch, and economists Steven Zahniser, John Dyck, and Matt Shane or their thor-ough reviews and helpul comments on earlier drats. We would also like to thank ourthree anonymous reviewers or their valuable suggestions. Special thanks are extendedto Courtney Knauth, Cynthia Ray, and Thomas McDonald or editorial and designassistance.

Suresh Chand Persaud

Birgit Gisela Saager Meade

Trade and Development When

Exports Lack Diversifcation

A Case Study rom Malawi

Economic

Research

Report

Number 77

July 2009

7/29/2019 ERR77.pdf

4/42

iiTrade and Development When Exports Lack Diversifcation: A Case Study rom Malawi / ERR-77

Economic Research Service/USDA

Recommended citation format for this publication:

Persaud, Suresh Chand, and Birgit Gisela Saager Meade. Trade and

Development When Exports Lack Diversifcation: A Case Study rom

Malawi. ERR-77. U.S. Dept. o Agriculture, Econ. Res. Serv. July 2009.

7/29/2019 ERR77.pdf

5/42

iii

Trade and Development When Exports Lack Diversifcation: A Case Study rom Malawi / ERR-77Economic Research Service/USDA

Contents

Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Export Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Asymmetric Impacts of Exports on GDP . . . . . . . . . . . . . . . . . . . . . . . . 5

Productivity Gains Potentially Offset Export Variability . . . . . . . . . . . 9

Marketing Eciency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Agricultural Productivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Exports and Economic Growth in Malawi . . . . . . . . . . . . . . . . . . . . . . 12

Symmetry: Reerence Case. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Asymmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Simulation Model Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Overview o Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Implications for Other Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Futures Markets and Developing Countries . . . . . . . . . . . . . . . . . . . . . 18

Price Transmission and Marketing Ineciency . . . . . . . . . . . . . . . . . . 19

Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Appendix. Data Sources, Model Parameters,

and Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

7/29/2019 ERR77.pdf

6/42

ivTrade and Development When Exports Lack Diversifcation: A Case Study rom Malawi / ERR-77


Summary

In developing countries, export concentration is a critical obstacle tosustained economic growth. A number o countries in Arica, Latin America,and the Caribbean depend heavily on a limited number o cash crop exportsand are vulnerable to domestic and international shocks. Malawi stands outamong these countries. Malawis agricultural exports averaged 20 percent ogross domestic product (GDP) between 2000 and 2004, and one commodity,tobacco, provided over hal o Malawis export earnings.

What Is the Issue?

The contribution o agricultural exports to economic growth in developingcountries has been neglected in the literature on export-led growth. In addi-tion, past studies have eectively assumed that an increase in exports wouldaect GDP to the same extent as an equivalent decrease. This assumptionmay not be correct. For example, i export revenues are ineciently utilized(or instance, diverted toward low-return enterprises that are not consistentwith a countrys comparative advantage), rising exports will do little to

increase GDP, and the growth-enhancing eects o export expansion may belargely lost. Moreover, inecient utilization o export revenues may leavethe countrys economy unprepared or unavorable shits in market condi-tions that lead to alling exports. The economic benets o export expansionmay then be muted, and the economic losses o export contraction may beaccordingly larger.

This analysis uses Malawi as a case study o export concentration and heavyexposure to volatility, a topic with broad relevance or other developingcountries that have diculty in drawing sustained economic growth rom anarrow portolio o cash crop exports. The study investigates the relationshipbetween Malawis tobacco and nontobacco exports and its GDP, particularly

the impact o rising compared with alling exports. I the economic impactso alling exports exceed those o rising exports, how does variability inoreign sales infuence the pace o Malawis GDP growth, compared with ascenario in which Malawis economy is aected equally by increasing anddecreasing exports?

In the absence o osetting improvements in productivity, a country thatdepends heavily on commodity exports is less likely to experience persis-tent economic growth because its economic ortunes would be closely tiedto booms and busts in world commodity markets. In developing countriessuch as Malawi, considerable potential exists to enhance the productivity othe agricultural export sectors by raising arm productivity and marketing

eciency. I tobacco exports are indeed an engine o Malawis GDP growth,might eciency gains in the tobacco sector reduce the potentially adverseeects o export volatility on Malawis economy?

What Did the Study Find?

The statistical ndings o the study bear out anecdotal evidence thatMalawis tobacco exports are positively related to its GDP. The analysis,which disaggregated Malawis total exports, showed no evidence that non-tobacco exports drive the countrys economic growth.

7/29/2019 ERR77.pdf

7/42

1


Variability in Malawis tobacco exports leads to slower economic growth,because GDP alls by a relatively large amount in response to a givendecrease in exports, while recovering little during an upswing in exports; theanalysis showed that the impact o tobacco exports on GDP is almost threetimes greater when exports are alling than when they are rising (asymmetry).This result or Malawi provides a cautionary tale or other countries withsimilar economic structures: Ineective management o export revenues,along with production and marketing ineciencies, may diminish the posi-

tive GDP eects o rising exports without tempering the negative eects oalling exports. Such inherent weaknesses may, in act, exacerbate the nega-tive impacts o export variability on a countrys GDP.

Model results show that Malawis GDP would be least vulnerable to volatilityin tobacco export earnings i gains in yield and marketing eciency werecombined with an export-GDP relationship that was symmetric, that is, withrising and alling exports having GDP impacts o equal magnitude. Growth inarm productivity would require improvements in the availability and quality oresources and inputs, as well as in human capital. Greater marketing eciencycould be achieved by reducing internal costs o transportation and distribution,which could partially insulate Malawis GDP rom alling export prices. Lowerexport prices would not be ully transmitted to the arm price i increasedeciency along the marketing chain led to increases in the armers share othe export price. Consequently, as marketing margins contracted, Malawiseconomy would be shielded to a degree rom lower international prices: themodel results indicate that a decrease in the export price would only slightlyreduce arm prices, domestic tobacco production, exports, and hence GDP.

On the other hand, i export prices rose, increased eciency along themarketing chain could ampliy the benets accruing to growersan increasein the armers share would mean that armers received a greater portion oa higher export price. The analysis shows that this would lead to relatively

large increases in arm prices, production, and exports, and thereore in GDP.As export variability inevitably occurs, the combination o margin compres-sionwhich reduces the gap between arm prices and export pricesandrising tobacco yields can partially oset the negative impact o alling exportprices on Malawis GDP, while ampliying the benets o rising exportprices, generating an upward ratcheting eect.

How Was the Study Conducted?

The ramework or empirically investigating the relationship between exportsand economic growth is based on an extension o an aggregate productionunction model. The value o exports is partitioned into rising and alling

components, leading to a more fexible model specication that allowsor dierential impacts o increasing vs. decreasing exports. Although theeconometric results clearly suggest an asymmetric relationship between realtobacco exports and real GDP, this study does not predict or orecast thecontinuation o relationships ound in historical data. Rather, it comparesdierent scenarios involving asymmetric vs. symmetric export-GDP link-ages. Econometric estimates o the export-GDP relationship or Malawi areembedded in a simulation model that also incorporates relationships or thearm supply response o tobacco and the transmission (to varying degrees) oexport prices to producer prices.

7/29/2019 ERR77.pdf

8/42

2Trade and Development When Exports Lack Diversifcation: A Case Study rom Malawi / ERR-77


Introduction

In large parts o Arica, Latin America, and the Caribbean, economic growthand stability depend heavily on a limited number o cash crop exports.However, these export commodities are prone to periods o volatility. Tradeshocks infuencing exports o commodities such as tobacco, coee, tea,cocoa, sugar, and cotton oten lead to macroeconomic instability and pooreconomic perormance in developing countries, reducing U.S. agriculturalexport potential to these regions.

Among countries with a narrow range o exports, Malawi stands out. Malawiranks among the poorest countries, with per capita gross domestic product(GDP) in purchasing power parity terms o $760 in 2007 (World Bank,2009). The country also has one o the most highly concentrated export port-olios one cash crop, tobacco, accounts or approximately 60 percent oits total merchandise export earnings. Malawi is a small landlocked countryin Arica with a population o approximately 12.9 million. Tobacco accountsor 13 percent o Malawis GDP and 23 percent o its total tax base (Jaee,

Figure 1

Coffee exports as share of total merchandiseexports, 1994-2004 average

Burundi Ethiopia Guatemala Honduras Rwanda Uganda0

10

20

30

40

50

60

70

80

Source: FAOSTAT database.

Percent

Figure 2

Sugar exports as share of total merchandiseexports, 1994-2004 average

Percent

0

5

10

15

20

25

30

35

40


Belize Cuba FijiIslands

Guyana Mauritius Swaziland

Figure 3

Banana exports as share of total merchandise

exports, 1994-2004 average


Percent

CostaRica

Dominica Ecuador Panama SaintLucia

SaintVincent/

Grenadines

0

10

20

30

40

50

60

Figure 4

Key agricultural exports as share of total

merchandise exports, 1994-2004 average


Percent

Cote dIvoire Ghana Madagascar Malawi0

10

20

30

40

50

60

70

Cocoa Cocoa Vanilla Tobacco

7/29/2019 ERR77.pdf

9/42

3


2003). This study uses Malawi as a case study o export concentration andheavy exposure to volatility, an issue with broad relevance or other devel-oping countries that experience diculties in drawing sustained economicgrowth rom a narrow portolio o cash crop exports.

The study investigates the relationship between Malawis exports and itsGDP. Although the export-led growth hypothesis has gained wide acceptance(Krueger, 1990; World Bank, 1991, 1993), the contribution o agricultural

exports to economic growth in developing countries has been neglected in theliterature (Dawson, 2005). In addition, past studies have eectively assumedthat a given increase in exports has the same magnitude o impact on GDPas an equivalent decrease in exports. This assumption may not be correct. Iexport revenues are diverted to subsidize low-return enterprises that are notconsistent with a countrys comparative advantage, the positive GDP eectso export expansion could be largely lost. Moreover, inecient utilizationo export revenues would leave the countrys economy unprepared or una-vorable shits in market conditions that lead to alling exports. A distinctiveeature o this study is that it recognizes that GDP may all by a relativelylarge amount in response to a given decrease in exports, while recoveringlittle during an upswing in exports, a phenomenon called asymmetry.

7/29/2019 ERR77.pdf

10/42



Export Concentration

Many countries in Sub-Saharan Arica, Latin America, and the Caribbeanearn a large share o their export revenues rom a small number o agricul-tural commodities (FAO, 2004). Commodity concentration is particularlyextreme in Burundi and Malawi, countries that earn approximately 70and 60 percent o their total merchandise exports rom sales o coee andtobacco, respectively (gs. 1 and 4). Among sugar exporters, Cuba andGuyana stand out with the highest levels o export concentration (g. 2),and banana exports account or between 30 and 60 percent o merchandiseexports in several Caribbean countries (g. 3). A common characteristic oheavily indebted poor countries is that they obtain more than hal o theirmerchandise export earnings rom primary (i.e., unprocessed) commodities(World Bank, 1999). Relative to industrial economies, developing countries(excluding those in East Asia) tend to experience larger shocks in terms otrade growth (Loayza et al., 2007), implying that price instability in interna-tional markets is a major concern or developing economies.

When supply and demand are highly inelastic, even small changes in demand

or supply can be a source o considerable price variability. In response totight market conditions, characterized by low stocks and high prices, armerscan increase their planting, but they cannot reduce the time it takes or cropsto ripen to harvest, a period that can extend to years in the case o peren-nial crops such as coee or cocoa. When higher crop production occurs,prices all as supplies quickly outgrow demand in importing countries,given that demand does not grow signicantly in response to lower prices.Consequently, price fuctuations are prolonged and deepened, potentiallyleading to a pattern o short-lived booms ollowed by lingering slumps(FAO, 2004). Nevertheless, the demand or market-based risk managementservices tends to be unmet in developing countries (World Bank, 1999). Pricevariability may induce countries to orgo the benets o specialization in

accordance with comparative advantage, or to enact costly price stabilizationschemes, while also contributing to unwillingness to liberalize markets.

Although variability is a concern in developing countries, empirical studiesare not unanimous in their support or the hypothesis that volatility adverselyaects economic growth. Indeed, a positive relationship between exportinstability and economic growth is possible. For example, under the assump-tion o risk-averse behavior, instability in export earnings may lead todecreased consumption, higher savings, increased investment, and hencehigher economic growth (MacBean, 1966; Knudsen and Parnes, 1975;Yotopoulos and Nugent, 1976). In addition, i higher risk enterprises areassociated with higher returns, countries that on average experience aster

GDP growth will also experience greater volatility. In contrast, it can beargued that export instability restrains economic growth, i risk-averse inves-tors reduce their investment or i export instability induces countries to holdlarge oreign exchange reserves that may incur substantial opportunity costs.Empirical studies, such as those o Glezakos (1973), Voivodas (1974), Ozlerand Harrigan (1988), and Gyimah-Brempong (1991), nd a negative relation-ship between export instability and economic growth.

1A number o studies attempt to

separate volatility into predictable vs.

unpredictable components (Dehn, 2000;

Dehn et al., 2005; Ramey and Ramey,

1995), which is also problematic. I

theory provides little guidance or sup-

port regarding measures o volatility,

then attempts to separate volatility into

predictable vs. unpredictable compo-

nents will also lack rm oundations.

7/29/2019 ERR77.pdf

11/42

5


Collier and Gunning (1996) and Schuknecht (1998) nd that when countriesexperience positive shocks rom windall gains in export revenues, slowergrowth results rom reduced eciency o public investment projects duringthe boom phase, ollowed by higher scal decits at the end o the boom.Similarly, Dehns (2000) empirical results show no evidence that extremepositive commodity price shocks have an impact on economic growth,while extreme negative shocks do adversely aect growth. Note that Dehnsresults rely on various decompositions o price volatility into predictable vs.

unpredictable components, where the latter was not signicantly associatedwith economic growth. Dehn et al. (2005) maintain that the literature has notestablished either a quantitatively important or statistically signicant linkbetween the variability o commodity export prices and economic growth.On the other hand, FAO (2004) maintains that declines and fuctuations inexport earnings have adversely aected income, investment, and employ-ment in developing countries where exports are not diversied.

7/29/2019 ERR77.pdf

12/42



The lack o consensus in the literature cited here may partly refect uncer-tainty as to how to measure volatility. A key challenge o estimating theimpacts o volatility, whether in exports, exchange rates, or prices, is thechoice o a particular variable to measure volatility. Theory provides littleguidance1 or selecting an appropriate measure, and the empirical results maybe sensitive to the measure chosen (Goodwin, 2002). To avoid the uncertain-ties associated with attempts to measure variability, we ocus this study ona more undamental issue: an increase in export revenues (as distinct rompositive commodity price shocks) may not have the same impact on GDP asan equivalent decrease in export revenues. In other words, the export-GDPrelationship may be asymmetric.

Figure 5

Negative supply-side shocks and asymmetry

Exports S2 S1

GDP response to an increasein exports

E2

E3 GDP response with a subsequent

decrease in exports

E1

Q3 GDPQ1 Q2

7/29/2019 ERR77.pdf

13/42

7


Asymmetric Impacts o Exports on GDP

Even when developing countries rely on a narrow range o exports, theireconomic diculties are not entirely a consequence o commodity concen-tration. Other actors also tend to reduce the positive GDP eects o risingexports. For example, ineective management o export revenues, combinedwith policies that do not sterilize large infows o oreign currency (i.e.,that do not limit exchange rate appreciation by reducing domestic currencyliquidity), can allow destabilizing side eects (Lee, 1997) and misalignmentso relative prices (Varangis et al., 1995). Consequently, asymmetry mayarise, i.e., the economic benets o export expansion may be muted, whilethe economic losses o decreasing exports would be relatively larger.

When a country experiences a period o rapidly rising exports, correlatedchanges in policies and macroeconomic conditions are possible, such asincreased government spending and indebtedness and currency appreciation,which partially oset the positive GDP eects o export expansion (Varangiset al., 1995). The exchange rate can be slow to adjust as a country enters aphase o deteriorating export perormance, which is exemplied by Malawi

(Pryor, 1990). Although Malawis exchange rate regime has been relativelyfexible since 1994 (FAO, 2003) and is characterized as a managed foat, poli-cymakers have alternated between periods o liberal and restrictive exchangerate management (IMF, 2007). Factors eroding the protability o tobacco inMalawi include deteriorating price, quality, and productivity, in addition to themovement (or management) o the exchange rate (Jaee, 2003). The exchangerate issue has broad relevance to other developing countries. Currency appreci-ation reduces the positive GDP eects o export booms, and, since the currencytends to be slow to depreciate during periods o alling international pricesand/or export revenues, the result is to ampliy the adverse GDP eects duringperiods o deteriorating export markets (Varangis et al., 1995).

A number o other actors may truncate a countrys upward economic growthpotential during avorable periods o rising exports, while not amelioratingthe downside during periods o alling exports. For example, windall gainsare oten channeled into low-return projects (Collier and Gunning, 1996).Governments may utilize revenues rom cash crop exports to nance invest-

Figure 6

Asymmetric export-GDP relationship

E3 E5 Y1 Y3

E1 Y5

Y2Y4

E2 E4

Period Period

1 2 3 4 5 1 2 3 4 5

Exports GDP

7/29/2019 ERR77.pdf

14/42



ments or subsidies or low-return enterprises that are not consistent with thecountrys comparative advantage, tending to blunt the positive GDP eectso rising exports. Note that tobacco is Malawis primary source o politicalpatronage (Jaee, 2003). Less-than-optimal use o export earnings may alsobe due to weak institutions (Rodrik, 1998) and rent-seeking behavior (i.e., theinvestment o resources in eorts such as lobbying to create monopolies orto restrict competition). All these actors would reduce the benets o risingexports, but there is no reason to conclude that they would shield an economyrom alling exports.

7/29/2019 ERR77.pdf

15/42

9


We illustrate a possible orm o asymmetry based on the notion o an export-GDP schedule, shown in gure 5 as S1. Exports are initially at E1 and GDPis at Q1. In this hypothetical example, windall gains in export revenues arenot invested in productivity improvements, and in this instance, GDP onlyrises along a static export-GDP curve rom Q1 to Q2 when export revenuesrise rom E1 to E2. Note that fuctuations in a countrys agricultural exportrevenues may result rom actors such as variability in weather, leading tochanges in the quantities o arm output and exports, or rom international

price movements. Although the example presented here ocuses on pricemovements as a source o instability in export revenues, similar reasoningapplies when fuctuations in export quantity cause variability in oreignexchange earnings.

When a boom in export revenues ends, due, or instance, to weather-inducedcontractions in arm output or alling international prices, a country may bein a heightened state o susceptibility to losses in export earnings. Windallgains in export revenues leading to currency appreciation would tend tosharply reduce the competitiveness o export industries that have not experi-enced the boom. As a result, a countrys export portolio will become evenmore skewed toward the main commodity with its rapidly rising internationalprices, implying larger economic losses when either the price or arm outputo that commodity enters a alling phase. Indeed, when the boom ends, thecountrys economy may be worse o than beore the onset o the boom(Yabuki and Akiyama, 1996). The currency can be slow to adjust, remainingovervalued even ater the boom ends and prices enter a declining phase.Consequently, the protability o the once-booming export sector suers dueto the combined eects o deteriorating international commodity prices andan overvalued currency (Varangis et al., 1995).

Moreover, asymmetries may be induced by limited integration with globalcapital markets in combination with thin domestic capital markets (Aizenman

and Pinto, 2005). Malawis nancial sector is underdeveloped, character-ized by measures o nancial depth that are low even by the weak regionalstandards o Sub-Saharan Arica (IMF, 2007). Large negative fuctuationsand the tightening o binding investment constraints adversely aect outputgrowth, particularly in countries that are poor, nancially and institution-ally underdeveloped, or unable to conduct countercyclical scal policies.Empirical studies suggest that large adverse shocks contract investment,make liquidity constraints binding, and eventually lead to asset destruction(Loayza et al., 2007).

The preceding discussion suggests that deteriorating commodity prices maylead to a negative supply-side shock. Accordingly, a decrease in exports may

have two eects. First, GDP may decrease along the static curve S1 in gure5, and second, there may be a shit to a new export-GDP curve rom S1 toS2. Consequently, as exports decrease rom E2 to E3, GDP contracts alongthe diagonal line between S1 and S2, ultimately reaching Q3, implying alarge decrease in GDP relative to its previous level o Q2. Forecasts based ona static concept o an export-GDP curve would tend to understate the adverseeconomic impacts o a negative shock. In summary, gure 5 illustrates a casewhere the GDP response to a decrease in exports exceeds the response to anincrease in exports. A scenario in which the GDP response elasticity is higher

7/29/2019 ERR77.pdf

16/42



or a decrease in exports than or an increase would, all else being equal,refect greater vulnerability to instability in export markets.

I the export-GDP relationship is asymmetric in the manner just described,fuctuations in exports would tend to generate a downward trend in GDP.Figure 6 claries this point. In the let panel, exports are on the y-axis. In theright panel, GDP is on the y-axis. In both panels, the x-axis indicates the timeperiod. Exports are initially at E1 in period 1 (let panel) and GDP is at Y1

(right panel). Exports all to E2 in period 2, leading to a contraction in GDPto Y2. When exports recover in the subsequent period to E3, GDP does notully recover and only reaches a level o Y3alling exports have greaterimpacts than rising exports. The process continues to repeat itsel: a dropin exports to E4 leads to a relatively large drop in the size o the economyto Y4. In period 5, note that exports are at their initial level since E5 = E1,while, in contrast, GDP is below its starting point, since Y5 < Y1. Althoughexports are fuctuating without any trend, GDP is experiencing a downwardtrend as a consequence o a orm o asymmetry in which the impact o allingexports exceeds the impact o rising exports.

The downward ratcheting eect illustrated in gure 6 is not an inevitableoutcome. Booms can be economically benecial when windalls are wellmanaged, but damaging when poorly managed. Moreover, alling worldprices do not inevitably reduce the gains rom primary commodity exports.Productivity gains in the export sector may allow countries to improve theirreturns rom commodity exports, even in the ace o deteriorating prices.

7/29/2019 ERR77.pdf

17/42

11


Productivity Gains Potentially OsetExport Variability

In the absence o osetting improvements in productivity, a country thatdepends heavily on commodity exports is less likely to experience persistenteconomic growth because its economic ortunes would be closely tied to pricebooms and busts in world commodity markets. Developing country exports

contribute directly to the growth o their economies because exports are acomponent o GDP. An increase in the value o exports tends to increaseGDP, while a decrease in exports has the opposite eect, all else being equal.The economic well-being o a society will vary with international commoditycycles, unless export revenues are used to nance investment and productivitygrowth in the export sector and other segments o the economy.

Furthermore, exports may accelerate GDP growth i trade leads to scaleeconomies and increased capacity utilization, as well as to externalitiesor spillovers that positively aect growth in the nonexport sectors (Feder,1983; Bhagwati and Srinivasan, 1978; Krueger, 1980). Trade can infuencegrowth through investment (actor accumulation) (Levine and Renelt, 1992;

Wacziarg, 2001), as well as through human capital accumulation (Frankeland Romer, 1999). Open economies may experience aster productivitygrowth, and exports may be a source o productivity gains (Bernard andJensen, 1999). In cases where a country has a comparative advantage in agri-culture, export-led growth rom agriculture allows or better use o limitedresources based on their true opportunity costs, while also providing oreignexchange (Johnston and Mellor, 1961) to upgrade technology.2 These indi-rectcontributions o exports to GDP growth (involving productivity gains)are distinct rom the direct eects o exports on GDP noted earlier, and it isprecisely these indirect contributions that may protect an economy rom thevicissitudes o primary-commodity markets.

Marketing Efciency

2The literature on the nexus between

trade and growth has grown to vast pro-

portions over the past our decades. For

an excellent review o empirical studies,

see Lewer and Van den Berg (2003).

Figure 7

Tobacco yields


1990 92 94 96 98 2000 02 04 06

0

500

1,000

1,500

2,000

2,500

3,000

Malawi

United States of America

Africa

Latin America & Caribbean

World developing countries

kg per ha

7/29/2019 ERR77.pdf

18/42



Steady improvements in marketing eciency may shield exporters o cashcrops rom volatility that involves deteriorating international prices. Tradecosts, including internal costs o transportation and distribution, are impor-tant barriers to exports,3 and these costs can be much larger than trade policybarriers (Anderson and van Wincoop, 2004). Greater eciency in marketingand reduced trade cost barriers may oset the impacts o alling exportprices, allowing a country to continue to benet rom cash crop exportseven during unavorable market conditions. On the other hand, ineciency,bottlenecks, corruption, and excess prots due to market power along themarketing chain help to divert the proceeds rom tobacco sales/exports awayrom productivity-enhancing investments.

3In addition, empirical evidence rom

Dennis and Shepherd (2007) indicates

that reductions in trade costs are associ-

ated with gains in export diversication.

Figure 8

Real value of Malawis tobacco exports, 1970-2003


Exports (1,000 dollars, real)

1970 74 78 82 86 90 94 98 02

0

50,000

100,000

150,000

200,000

250,000

300,000

350,000400,000

450,000

Figure 9

Tobacco exports and GDP in Malawi1, 1970-2003 (R2 = 0.71)

1Logarithmic scale on the Y and X axes, i.e., natural logs of tobacco exports (1,000 U.S. dollars)and GDP (1,000 U.S. dollars).

Sources: World Bank; FAOSTAT database.

Real GDP

B

BB B

BB

B B

BBB

B BB

BB B B

BBB

BB

B

B

BB

BBB

BBBB

13.0

13.2

13.4

13.6

13.8

14.0

14.2

14.4

14.6

11.2 11.7 12.2 12.7 13.2

7/29/2019 ERR77.pdf

19/42

13


The Government o Malawi has traditionally regulated tobacco arming.Prior to 1990, legal restrictions prevented smallholders rom growingtobacco, while estate arms were allocated production quotas with the right tomarket their quota on the auction foors. Beginning in the 1990/91 croppingseason, smallholders were granted quotas, allowing them to legally growtobacco. However, smallholders were obligated by law to sell their output toMalawis state marketing board (ADMARC), which paid lower prices thanthe auction foors. The Government o Malawi rapidly increased the small-

holders quota until 1995/96, and it abolished the quota system in 1996/97(Zeller et al., 1998).

As alternatives to ADMARC, Malawis reorms permitted smallholders toorganize themselves into armer clubs that had the right to directly markettobacco to auction foors. In addition, under the reorms intermediate buyerswere licensed to buy tobacco rom smallholder or estate arms to be soldon the auction foor (Zeller et al., 1998). Intermediate buyers unction as animportant logistical link between tobacco armers and the distant auctionfoors, allowing armers to market their crop even i they are unable to bringtheir tobacco to the auctions. Intermediate buyers also pay armers immedi-ately in cash, although the amount is typically hal o the auction foor price(Jaee, 2003).

Malawis production and marketing reorms were successul at achievinggreater smallholder participationsmallholders account or 70 percent oMalawis tobacco output. However, the mass entry into production o small-holder armers is associated with high and rising marketing and transactioncosts (Jaee, 2003). Transportation is the most important cost along themarketing chain. These costs are increased by regulations that restrict exportsto a single auction house, thereby creating bottlenecks and transport conges-tion in ront o the auction foorstransport costs or moving tobacco to thefoor are infated by as much as 300 percent, resulting in reduced arm prices.

Moreover, these bottlenecks provide incentives or corruption and bribes(Koester et al., 2004). Auction Holdings Ltd. (AHL) is the sole operatoro Malawis three auction foors, a monopoly in which the Government oMalawi has a controlling stake through ADMARC. Given the legal require-ment that tobacco must be sold through auction foors, AHL passes throughto armers the expenses arising rom corruption, extravagant corporate ocecosts, and excessive sta numbers at senior levels (Maleta, 2004).

In summary, ineciencies and corruption along the marketing chain, aswell as monopoly or monopsony rents accruing to intermediaries, resultin reduced payments or Malawis tobacco armers (Koester et al., 2004).Indeed, high local marketing costs may depress arm incomes by discour-

aging growers rom cultivating export crops (Balat et al., 2007). In contrast,steady improvements in marketing eciency that lead to higher arm pricesand, in the intermediate to long term, to expanding agricultural productionand exports, may contribute to aster, more sustained economic growth.Factors that reduce unit costs in the marketing sector would allow rms tooperate protably at smaller margins, involving lower and more competi-tive export prices. That is to say, marketing eciency can be a key actorin coping with volatile periods in international markets when alling exportprices lead to margin compression. Advances in arm productivity also

7/29/2019 ERR77.pdf

20/42



enhance a countrys ability to cope with volatility by allowing armers tocover their costs o production at lower prices.

Agricultural Productivity

With increases in total actor productivity (TFP)dened as the outputper unit o all inputs combinedthe agricultural sector can continue toprot, despite declining real prices, i productivity eects compensate orprice decreases. Considerable potential exists or raising productivity inSub-Saharan Arica (Wiebe et al., 1998), as well as in other developingregions. Continued growth in agricultural production hinges on producingmore output per unit o land, apartial measure o productivity known as

yield, rather than on bringing more land under cultivation.

Malawis tobacco armers have experienced a decreasing trend in produc-tivity, with yields dropping at an annual rate o approximately 6 percentrom 1990 through 2006. In 2006, Malawi tobacco yields were 25 percentlower than the Arican average, while yields in the United States were sixtimes higher (g. 7). Possible explanations include decreasing ertilizeruse as well as a more general decline in soil ertility. Additional actorsadversely aecting productivity include decreases in larger estate arms andthe growing preponderance o relatively unskilled smallholder armers, wide-spread use o low-quality, own-saved seed, and high and rising marketingand transaction costs (Jaee, 2003). Moreover, despite potentially abundantsupplies o water rom Lake Malawi and several other large lakes (Tobin andKnausenberger, 1998), almost all o Malawis tobacco is grown under rain-ed conditions (USDA, 2000) and hence is vulnerable to weather-inducedshocks. Future yield trends will be infuenced by the degree to which Malawiand other developing countries address decits in the availability and qualityo resources and inputs, as well as decits in human capital.

Growth in arm productivity and improved marketing perormance are likelyto be largely concurrent. In countries where inrastructure is underdeveloped,

Figure 10

Simplified depiction of simulation model

Export price

Export revenues

GDP

Tobacco

production

Farm price

7/29/2019 ERR77.pdf

21/42

15


large increases in agricultural productivity may be possible rom investmentsin rural roads and utilities (Wiebe et al., 1998). Improvements in the quantityand/or quality o transportation and distribution systems would give armersbetter access to productivity-enhancing agricultural inputs at lower costs.Diculties in marketing larger quantities o production outputs would tendto discourage the adoption o output-increasing technology, while improveddistribution systems would have the opposite eect. Growth in agriculturalproductivity, combined with parallel improvements in marketing, imply thatoutward supply shits may be accompanied by alling margins.

Exports and Economic Growth in Malawi

Tobacco export earnings are a key driver o Malawis economy (Koester etal., 2004; USDA, 2000), and those earnings have experienced substantialfuctuations (g. 8), particularly since 1989. Figure 9 illustrates the positiverelationship between real GDP and the real value o tobacco exports.

On a more ormal level, we quantiy the linkage between these two variablesvia an aggregate production unction model (see appendix or a descriptiono the model). The econometric estimates o the export-GDP relationship aredeveloped in two steps: (1) to provide a reerence scenario, we estimate themodel assuming that a given increase in exports has an impact on GDP iden-tical to an equivalent decrease in exports, an assumption known as symmetry,and (2) we relax the preceding assumption to allow dierential impacts on

GDP o rising and alling exports, in a phenomenon known as asymmetry.Comparisons o symmetry and asymmetry yield useul insights.4

Symmetry: Reerence Case

The model results indicate that a 10-percent increase in the real value otobacco exports is associated with a 1.8-percent increase in real GDP.Similarly, because the assumption o symmetry is imposed, a 10-percentdecrease in the real value o tobacco exports is also associated with a1.8-percent decrease in real GDP. These results, despite their useulness in

4Also o interest is the impact o

nontobacco exports. Accordingly, when

estimating the model, total exports

are split into tobacco vs. nontobacco

exports, to obtain the separate impacts

on GDP o these export variables. The

ndings indicate no evidence that nonto-

bacco exports, whether rising or alling,

infuence Malawis economic growth in

any o the model specications.

Figure 11

Assumed growth in export price of tobacco

Source: Model assumption.

2003 2005 2007 2009 2011 2013 2015 20172,000

2,500

3,000

3,500

4,000

4,500

Export price ($ per metric ton)

7/29/2019 ERR77.pdf

22/42



providing a benchmark or model-based simulations later in this report, donot account or the possibility that the impact o a change in exports maydepend on the direction o the change.

Asymmetry

Statistical evidence suggests that both increases and decreases in tobaccoexports aect Malawis economy, although not to the same degree (app. table1). The results indicate a rejection o the null hypothesis o symmetry, andaccording to the point estimates, the impacts on GDP o decreasing tobaccoexports ar exceed those o rising exports. Specically, a 10-percent increasein the value o tobacco exports is associated with an increase in GDP o 0.90percent, while a 10-percent decrease reduces GDP by 2.51 percent. Thisdierence is signicant, both statistically and economically. The latter isbrought into sharp relie in the next section, where we build multiyear GDPprojections under alternate scenarios.

Simulation Model Characteristics

Given empirical evidence that oreign exchange earnings rom agricul-ture have macroeconomic eects, variables that aect a countrys exportrevenuessuch as domestic arm production, export prices, and behind-the-border ineciencieslikely infuence its national income. We illustratethese relationships with a simple economic model or Malawi. In addition,the model ramework is used to evaluate the likely role o asymmetry in theexport-GDP relationship in constraining economic growth.

As depicted in gure 10, the export price and the predetermined quantityo tobacco production infuence the level o export revenues, which in turncontribute to GDP. A change in the export price has contemporaneousimpacts on export revenues and GDP, as well as on the producer price o

tobacco.

Figure 12

Malawis real GDP under alternative scenarios

Source: Model results.

GDP (U.S. $1,000, real)

2005 2007 2009 2011 2013 2015 20171,500,000

1,700,000

1,900,000

2,100,000

2,300,000

2,500,000

2,700,000

2,900,000

Scenario I

Scenario III

Scenario II

7/29/2019 ERR77.pdf

23/42

17


Changes in the producer price (due to export price movements) alterMalawis arm production o tobacco in subsequent time periods, implyinglagged eects on GDP, in addition to the contemporaneous impacts that werepreviously mentioned. As indicated by gure 10, the ramework incorporatesrelationships or the arm production o tobacco and the transmission (tovarying degrees) o export prices to producer prices. The ull model structurealso captures the previously mentioned GDP-export linkages (symmetric andasymmetric).

Price Transmission

Export prices o tobacco, which are considered to be determined outside othe model,5 fuctuate around an increasing trend (g. 11), and grow at anannual rate o 2.5 percent, by assumption in all the various scenarios.6 Theexperiments involving partial price transmission are based on statisticalevidence suggesting an estimated elasticity o 0.55 (see appendix).

Tobacco Production and Exports

Tobacco acreage responses are computed using area elasticities o 0.329and 0.951 in the short run and the long run, respectively, and are based onChembezi (1991). Farm production o tobacco is the product o croppedarea and yield. For simplicity, domestic consumption, imports, and stocks otobacco are assumed to be negligible, implying that the quantity o tobaccoexported equals arm production.

Overview o Scenarios

Every scenario considered in the model is similar in that (1) the exportprice ollows the same pattern o fuctuating around an upward trend, and(2) we hold constant actors such as nontobacco exports, labor, and capital

throughout the multiyear model projections.

However, there are also key dierences across scenarios. In the Reerencescenario (Scenario I), the export-GDP relationship is symmetric. In addition,tobacco yield is held constant. Changes in the export price have (1) imme-diate impacts on export revenues, GDP, and arm prices via partial pricetransmission eects, and (2) lagged impacts on arm production o tobacco(via changes in cropped area), with implications or export revenues, andhence GDP.

Scenario II diers rom the Reerence case in that the export-GDP relation-ship is asymmetric, but again, yield is held constant throughout the projection

period, and export prices are partially transmitted to armers. Scenario IIIreturns to the symmetric export-GDP relationship used in the Reerence case,while also adding insight by evaluating the impacts o rising yield. ScenarioIII also incorporates steady improvements in marketing eciency that eec-tively overturn the price transmission relationship used in the Reerencescenario and Scenario II.

In evaluating the projections, it is not the base-year levels themselves that arecritical, but rather changes relative to the base yearas well as the relativedierences between the various scenariosthat are relevant. The scenarios,

5Empirical results rom Koester et al.

(2004) indicating price-taking behavior

suggest that Malawi acts as a small

country.

6The assumed price growth, along

with the other components o the

simulation model, leads to a 3.4-per-

cent growth rate in the value o tobacco

exports, which matches the 1970-2005

growth rate.

7/29/2019 ERR77.pdf

24/42



which are neither orecasts nor predictions, are meant to instruct, with thestarting point and slope not crucial to the argument.

Scenario I: Symmetry, With No Improvements in Yield

and Marketing Efciency

The export unit value fuctuates around an increasing trend (g. 11), growingat 2.5 percent per annum (app. table 2) throughout the projection period (by

assumption). With a passthrough elasticity o 0.55, rising export prices arenot completely passed through to armers as increasing margins dampenthe price transmission eectsproducer prices rise relatively slowly, at anannual rate o 1.6 percent in this experiment (app. table 2). Margins expandat an annual rate o 2.9 percent.

Although tobacco yields remain constant at 0.73 metric tons (mt) per hectare(ha) throughout the projection period (by assumption), area planted totobacco expands at 0.9 percent per year, in response to growth in arm prices.At 3.4 percent per year, export revenues grow aster than the unit value, dueto increases in the quantity o production/exports and rising export prices.Given that the GDP-export relationship is symmetric in this experiment,economic growth tracks export revenues. Despite exhibiting year-to-yearfuctuations, GDP grows at an average annual rate o 0.6 percent (g. 12,app. table 2).

Scenario II: Asymmetry, With No Improvements in Yield and Marketing

Efciency

The export price continues to fuctuate around an upward trend and growsat an average annual rate o 2.5 percent, as beore. Yield remains xed at0.73 mt/ha, and price transmission is incomplete as beore. An importantdierence now is that the relationship between exports and national incomeis asymmetricthe impacts on GDP are almost three times greater whenoreign sales arealling than when they are rising. The implication is thatupward and downward fuctuations in exports, or variability, tend to reducethe pace o economic expansion. Other things equal, GDP alls during aperiod o decreasing export revenues. However, suppose, or example, thatduring a period o rising exports that is assumed to ollow, GDP grows rela-tively little. In that case, GDP does not experience a recovery when exports

Futures Contracts

A utures contract, which is an agreement priced and entered on an

exchange to trade a commodity (with standardized characteristics) at aspecied uture time, allows growers to reduce risks. For example, a armproducer may use short hedging as protection against decreases in outputprice, a practice that entails selling utures contracts at the beginning oor during the growing period. The grower holds the short utures posi-tion until the product is ready to be sold, and during this holding period,losses (gains) in the value o the output due to unexpected price changestend to be oset by gains (losses) in the value o the utures position(Harwood et al., 1999).

7/29/2019 ERR77.pdf

25/42

19


return to (or exceed) their previous level, implying that a sustained period oexport variability reduces economic growth. Scenario II illustrates just such acaseGDP now contracts at an annual rate o 0.4 percent.7

Scenario III: Symmetry, With Improvements in Yield

and Farmers Share

As in Scenario I, the GDP-export relationship is symmetric and the export

price fuctuates around an increasing trend, growing at an annualized rate o2.5 percent. An important dierence is that we now relax the previously heldassumptions regarding marketing eciency and yield.

In the scenarios thus ar, tobacco yields were xed throughout the projectionperiod, implying that any growth in arm output originated rom area expan-sion rather than productivity improvements. However, at 0.73 mt/ha in 2005,tobacco yields in Malawi are below those o most other major producers.In addition, the armers share o the export price in Malawi is relativelylow (34 percent). The analysis now assumes that the armers share o theexport price rises to the 2003-2005 average level in Zimbabwe (57 percent).Reductions in the internal costs o transportation and distribution, as wellas greater competition in the marketing sector, narrow the gap or marginbetween arm and export prices. Malawis tobacco yield also grows to the2003-2005 average level in Zimbabwe, i.e., 1.67 mt/ha.

In the scenario, armers steadily increase their share o the export pricethroughout the projection period, and margins decrease by 0.8 percentper year. Consequently, the arm price grows at 6.5 percent, substantiallyaster than in the previous two scenarios. Moreover, the nature o the pricetransmission diers markedly rom the previous scenarios: it is no longeran econometrically estimated parameter imposed on the model; rather, it isdetermined within the model. The simulatedtransmission elasticity o exportprices to arm prices is now 82.6 percent larger during years o rising vs.alling export prices (1.26 vs. 0.69).

Lower export prices are not ully transmitted to the arm price i increasedeciency along the marketing chain leads to increases in the armers shareo the export price. Consequently, as marketing margins contract, Malawiseconomy is to a degree shielded rom lower international prices: A decreasein the export price only slightly reduces arm prices, domestic tobaccoproduction, exports, and hence GDP. On the other hand, when export pricesrise, increased eciency along the marketing chain amplies the benetsaccruing to growersmeaning that armers receive a greater portion o ahigher export price. This leads to relatively large increases in arm prices,

production, and exports, and thereore in GDP. As export variability inevi-tably occurs, the combination o margin compressionwhich reducesthe gap between arm prices and export pricesand rising tobacco yieldspartially osets the negative impact o alling export prices on MalawisGDP, while ampliying the benets o rising export prices, generating anupward-ratcheting eect.

Propelled by rising arm yields and stronger price incentives to expand thearea under tobacco cultivation, under Scenario III tobacco production andexport revenues grow at average rates o 12.1 and 14.8 percent per year,

7These projections are not predictions

or orecasts, but rather scenarios that un-

derscore the opposing economic eects

o trend growth in export prices vs. fuc-

tuations around the trend. Experiments

using varying magnitudes o the growth

trend reveal that aster growth in export

prices relative to the degree o pricevariability has a positive eect on GDP

growth. In addition, sensitivity analysis

reveals that the opposing eect o price

growth and variability is reduced when

the dierential between the elasticities

o rising and alling export revenues is

narrowed in the GDP model.

7/29/2019 ERR77.pdf

26/42



respectively, leading to the astest rate o economic growth (2.5 percent peryear) among the three scenarios.

7/29/2019 ERR77.pdf

27/42

21


Implications or Other Countries

Trade shocks infuencing developing country exports o commodities such astobacco, coee, tea, cocoa, sugar, and cotton oten lead to macroeconomicinstability and poor economic perormance, which reduce U.S. agriculturalexport potential to these regions. The state o the U.S. agricultural economyis closely tied to international trade, and developing countries, spurred bychanging demographics and, particularly, by growing purchasing power,are becoming increasingly important outlets or U.S. agricultural exports.Continued expansion o developing country imports depends on the abilityo those countries to achieve sustained increases in incomes and oreignexchange earnings to pay or imports.

This study presents its ndings conditionally: I the GDP impacts o allingexports exceed those o rising exports in a given country, the adverseconsequences o export volatility are exacerbated, all else equal. We do notattempt to generalize the statistical evidence or Malawi that suggests asym-metry. However, the issue may have relevance or other developing countrieswhere economic growth and stability depend heavily on a narrow range o

agricultural exports, particularly those with limited means to cope with riskexposure.

Futures Markets and Developing Countries

Although developing countries are heavily exposed to volatility associatedwith commodity markets, the demand or risk management services tends tobe unmet. The ve largest producers among developing countries accountor more than 75 percent o global production o cocoa, tea, rice, palm oil,coconut oil, groundnut meal, groundnut oil, rubber, and tin. Yet, developingcountries in total account or less than 2 percent o the volume o utures andoptions instruments (see box, Futures Contracts), a remarkably small rac-

tion given that these countries produce a large share o the physical output ocommodities.

Developing country participation in commodity utures markets is lacking,partly because, or certain commodities that are important to developingcountries (such as rice and tea), no liquid exchanges exist. In addition,the available instruments in risk management markets are not necessarilyconsistent with developing country risk management needs, which otentend to be or smaller contract sizes, longer maturities, or a dierent qualityo commodities, or or contracts that are not actively traded in internationalmarkets (World Bank, 1999).

Despite the act that utures markets oer substantial leveraging potential,traders require extremely liquid nancial reserves in order to maintain uturespositions. This obstacle is compounded by the act that traders in the devel-oping country cannot use domestic currency to meet margin calls, and insteadmustor instanceutilize dollars or interest-bearing U.S. Governmentsecurities. A urther complication stems rom risks associated with exchangerate fuctuationswhen traders in a developing country hold U.S. uturescontracts, they are, in eect, speculating on the value o their domesticcurrency relative to dollars (Thompson, 1985).

7/29/2019 ERR77.pdf

28/42



The trader may attempt to manage exchange rate risk by concurrentlyhedging in oreign exchange markets,8 at the cost o requiring added liquidityon the part o the developing country trader (Thompson, 1985). The need tohold liquid nancial reserves may exacerbate problems o oreign exchangescarcity. Moreover, there is also an opportunity cost o tying up oreignexchange or hedging purposes, in that it likely precludes its use in acquiringproductivity-enhancing technology, which, as discussed earlier, may in itselhelp a country cope with volatility.

Price Transmission and Marketing Inefciency

The empirical ndings or Malawi that illustrate the role o productivity inbuering the impacts o export variability have broad application, givenwidespread ineciencies in developing country marketing sectors. Useulinormation on the market environment in developing countries may beobtained rom proxy indicators (Townsend, 1999)or example, the trans-mission o changes in exchange rates to producer prices. The use o proxyindicators is a pragmatic approach, since attempts to characterize the marketenvironment in developing countries are plagued by measurement problems.

Cross-country dierences in transmission elasticities indicate dierencesin market structure, product characteristics, competition, and trade barriers(Menon, 1995). Empirical evidence rom a wide cross-section o countriesindicates that the transmission to producer prices o changes in internationalprices and exchange rates, while not altogether absent, tends to be slow andincomplete in developing countries (Lieert and Persaud, 2009).

Barriers to Price Transmission

Historically, policy interventions were considered key barriers to price trans-mission. However, economic reorms, along with shits toward substantially

reduced government intervention in developing country agricultural markets,have enhanced the relative importance o other possible causes o incompletetransmission. These causes include, in particular, poor inrastructure andother market imperections (Lieert and Persaud, 2009).

Although there are dierences across countries, marketing o most majorexport and ood commodities tended to be highly regulated by developingcountry governments throughout the 1960s, 1970s, and part o the 1980s(Barrett and Mutambatsere, 2008). In response to price variability, govern-ments in developing countries created institutional arrangements thatattempted to insulate producers and consumers rom market price fuctua-tions through price controls or subsidies, or to replace the price discovery by

markets with a planned and regulated system o prices (World Bank, 1999).

In the long run, price stabilization schemes are expensive to maintain: sharpfuctuations in currency values or other economic events may require largetransers o resources in consecutive years o low prices in order to meetadministratively set benchmarks. Consequently, programs that attempted toseparate domestic commodity prices rom international prices were otenunsustainable in the ace o depleted stabilization reserves, limited borrowingcapacity, and generally unhedged exposure to price risks. Lack o sustain-ability was a key actor in the unilateral abandonment o ood marketing

8This assumes that there is an active

orward market or the particular devel-

oping countrys currency. I no orward

market exists, the trader will have to

resort to a (possibly imperect) cross-

hedge in a dierent currency, which may

be o limited value in reducing exchange

rate risk.

7/29/2019 ERR77.pdf

29/42

23


agencies and the once-ubiquitous marketing boards or export crops, such ascoee and cocoa (World Bank, 1999). Economic reorms and shits towardsubstantially reduced government intervention in developing country agri-cultural markets in the 1980s and 1990s led to reduced rates o indirect anddirect agricultural taxation. Nevertheless, a disaggregated view o net taxa-tion by exportable and import-competing products shows that exportables arestill heavily taxed in many developing countries (World Bank, 2007).

Reorm eorts were complicated by the creation o voids resulting rom thewithdrawal o government rom agricultural markets. Farmers experiencedreduced access to input nancing and increased input prices, as the privatesector oten ailed to replace parastatals in core input marketing activitiesdue to underdeveloped communications, power, and transport inrastruc-ture, along with credit constraints. Dismantling state-run monopolies andmonopsonies did not automatically lead to ecient, workably competi-tive marketing industries. Even ater removing legal and policy barriers,the private sector has played only a limited role in meeting the demand ormotorized transportation in rural markets as a consequence o high sunkcosts, as well as economies o scale (Barrett and Mutambatsere, 2008). In

addition, most orms o inrastructure have characteristics o a public good.For example, given the likelihood o nonpaying trac on roads, it is dicultto recoup investments in road building. I let solely to market orces, therewould be suboptimal provision o inrastructure, implying a role or govern-ment in encouraging and unding investments in inrastructure (Coyle, 2005).

Price Instability and Incomplete Price Transmission

Slow or incomplete transmission o international prices to domesticmarkets, whether a consequence o government intervention or decienciesin marketing sectors, may worsen price instability. For example, i lowerinternational prices are not ully transmitted to domestic prices, decreases

in world supply and increases in world demandwhich would otherwiseoccurwill not take place, making international price reductions more acuteand prolonged (Quiroz and Soto, 1995). Incomplete or slow price trans-mission in a small country would not aect world prices, but a group oexporting countries with a similar characteristic o low transmission couldhave global impacts. Inecient marketing sectors may also ampliy inter-national price volatility, in much the same way as government interventionsaimed at separating domestic commodity prices rom international prices.

Industry Consolidation and Inrastructure

Vertical integration o supply chain activities, a hallmark o ecient ood

supply chains in developed economies, may improve marketing eciencyin low-income countries by reducing the number o middlemen along themarketing chain and collapsing margins (Landes et al., 2004). By operatingon a higher scale o production, a smaller number o larger marketing rmscould spread out lumpy, indivisible xed costs and overhead over a largervolume o output, leading to lower unit costs. These changes in market struc-ture would tend to place greater demands on developing country inrastruc-ture, since it may be necessary to move nal products and inputs over greaterdistances. Thus, improved inrastructure plays a role in the emergence omore ecient market structures.

7/29/2019 ERR77.pdf

30/42



Sharp reductions in marketing margins potentially generate a combina-tion o higher prices or armers and lower prices or end users, even whena relatively small number o rms play a large role in setting prices due toshits toward greater industry concentration (Landes et al., 2004). Moreover,in the event o industry consolidation, agribusiness rms may be willing topay higher arm prices i those higher prices enable them to acquire largersupplies o raw agricultural materials, achieve higher rates o capacity utili-zation, and lower costs overall (Persaud and Landes, 2006; Persaud and

Tweeten, 2002). Improved transportation inrastructure within a countryacilitates competition and improves resource allocation (Coyle, 2005),encouraging the movement o ood products and services rom regions withlower prices to more remote, higher-priced areas.

Improvements in marketing eciency, leading to a decreasing trend inmarketing margins, markedly alters the nature o the price transmissioneects rom incomplete passthrough to a avorable orm o asymmetrictransmission. As demonstrated in the case o Malawi, a decreasing trend inthe margin that separates arm and export prices allows armers to steadilyincrease their share o the export price, thus buering Malawi rom allingexport prices while ampliying the benets o rising export prices. Countriesthat experience productivity gains are more likely to preserve the protabilityo their export industries in the ace o adverse price shocks, while those thatdo not achieve productivity growth tend to experience losses in market shareand greater diculty in coping with volatility.

7/29/2019 ERR77.pdf

31/42

25


Conclusions

Developing countries oten depend on a limited number o cash crop exports.Trade shocks and associated macroeconomic fuctuations may constrainincome growth in such countries. Adverse eects o price volatility are, insome respects, a consequence o underdevelopment and low productivity.In the absence o osetting improvements in productivity, a country thatdepends heavily on commodity exports is unlikely to experience persistenteconomic growth because its economy would be closely tied to booms andbusts in world commodity markets. Future trends in economic growth arelikely to be infuenced by the ability to cope with volatility through produc-tivity gains and improved management o export revenues. Malawi, an arche-typal country with a highly concentrated export portolio, illustrates theseissues surrounding trade and economic growth.

Statistical evidence suggests that both increases and decreases in tobaccoexports aect Malawis economy, though not to the same degree.Specically, the negative impacts on GDP o decreasing tobacco exports arexceed the positive eects o rising exports. Model-based simulations reveal

the role o this asymmetry in reducing Malawis economic growth. There isan inverse relationship between export variability and GDP because Malawibears the ull brunt o contractions in its exports, while recovering littleduring a subsequent period o rising exports. In other words, fuctuations inexports have a downward ratcheting eect on Malawis GDP.

Although statistical evidence or Malawi clearly indicates asymmetry, it isinstructive to model and compare dierent scenarios involving asymmetricvs. symmetric GDP-export linkages, rather than predicting or orecastingthe continuation o relationships ound in historical data. Alternate scenariosdemonstrate that improvements in arm yields and marketing eciency canmitigate the adverse eects o export variability. Growth in yields would

require improvements in the availability and quality o resources and inputs,as well as in human capital. Greater marketing eciency may be achieved byreducing internal costs o transportation and distribution. The results showthat as export variability occurs, the combination o margin compressionwhich reduces the gap between arm prices and export pricesand risingyields partially osets the negative impact o alling export prices on thecountrys GDP, while ampliying the benets o rising export prices, gener-ating an upward-ratcheting eect.

7/29/2019 ERR77.pdf

32/42



Reerences

Aizenman, J., and B. Pinto (2005). Overview, in J. Aizenman and B. Pinto,eds.,Managing Economic Volatility and Crises. Cambridge UniversityPress, UK.

Al-Yousi, Y.K. (1997). Exports and Economic Growth: Some EmpiricalEvidence rom the Arab Gul Countries,Applied Economics 29:693-97.

Anderson, James E., and Eric van Wincoop (2004). Trade Costs,Journalo Economic Literature 42(3):691-751. September.

Balat, Jorge, Irene Brambilla, and Guido Porto (2007). Realizing theGains rom Trade: Export Crops, Marketing Costs, and Poverty. NBERWorking Paper No. 13395. National Bureau o Economic Research,Cambridge, MA.

Barrett, Christopher B., and Emelly Mutambatsere (2008). AgriculturalMarkets in Developing Countries. The New Palgrave Dictionary oEconomics, Second Edition. Steven N. Durlau and Lawrence E. Blume,eds. Palgrave Macmillan.

Bernard, J., and J.B. Jensen (1999). Exceptional Exporter Perormance:Cause, Eect, or Both?Journal o International Economics 47:1-25.

Bhagwati, J.N., and T.N. Srinivasan (1978). Trade Policy andDevelopment, in R. Dornbusch and J.A. Frenkel, eds.,InternationalEconomic Policy: Theory and Evidence. John Hopkins University Press,Baltimore, MD.

Chembezi, Duncan M. (1991). Acreage Response to Price and Risk:The Case o Tobacco in Malawi,European Review o AgriculturalEconomics 18:85-102.

Collier, P., and J.W. Gunning (1996). Policy Towards Commodity Shocksin Developing Countries, IMF Working Paper (WP/96 84). InternationalMonetary Fund, Washington, DC.

Coyle, William (2005). Asia-Pacic Transportation Inrastructure: LinkingFood Sources to Urban Centers.Amber Waves 3(4), September. U.S.Department o Agriculture, Economic Research Service.

Dawson, P.J. (2005). Agricultural Exports and Economic Growth in Less-Developed Countries,Agricultural Economics 33:145-52.

Dehn, J., C.L. Gilbert, and P. Varangis (2005). Agricultural CommodityPrice Volatility, in J. Aizenman and B. Pinto, eds.,Managing EconomicVolatility and Crises. Cambridge University Press, UK.

Dehn, J. (2000). The Eects on Growth o Commodity Price Uncertaintyand Shocks, Policy Research Working Paper No. 2455, World Bank,Washington, DC.

7/29/2019 ERR77.pdf

33/42

27


Dennis, Allen, and Ben Shepherd (2007). Trade Costs, Barriers to Entry,and Export Diversication in Developing Countries, Policy ResearchWorking Paper No. 4368, World Bank, Washington, DC.

Engle, R.F., D.F. Hendry, and J.F. Richard (1983). Exogeneity,Econometrica 51:277-304.

Feder, G. (1983). On Exports and Economic Growth,Journal o

Development Economics 12:59-73.

Fosu, A. K. (1990). Export Composition and the Impact o Exports onEconomic Growth o Developing Economies,Economic Letters 34:67-71.

Frankel, J.A., and D. Romer (1999). Does Trade Cause Growth?AmericanEconomic Review 89:379-99.

Glezakos, Constantine (1973). Export Instability and Economic Growth:A Statistical Verication,Economic Development and Cultural Change21:670-78.

Goodwin, Barry (2002). Exchange Rate Risk and Agricultural Trade:Discussion, in Suchada Langley, Samarendu Mohanty, Marcelo Giugale,and William Meyers, eds.,Exchange Rate Volatility and InternationalAgricultural Trade. Captus Press Inc., Ontario, Canada.

Gujarati, Damodar (1995).Basic Econometrics, 3rd Edition. McGraw-Hill,New York.

Gyimah-Brempong, Kwabena (1991). Export Instability and EconomicGrowth in Sub-Saharan Arica,Economic Development and CulturalChange 39(4):815-28.

Harwood, Joy, Richard Heiner, Keith Coble, Janet Perry, and AgapiSomwaru (1999).Managing Risk in Farming: Concepts, Research, andAnalysis. Agricultural Economic Report No. 774. U.S. Department oAgriculture, Economic Research Service. March.

Houck, P. J. (1977). An Approach to Speciying and EstimatingNonreversible Functions,American Journal o Agricultural Economics59:570-72.

International Monetary Fund (IMF) (2007).Malawi: 2006 Article IVConsultation and Third Review Under the Three Year Arrangement

Under the Poverty Reduction and Growth Facility. Country Report No.07/147. IMF, Washington, DC. May.

Jaee, Steven (2003). Malawis Tobacco Sector: Standing on One StrongLeg Is Better Than on None. The World Bank Group, Arica RegionWorking Paper Series No. 55, June.

Johnston, B. F., and J. W. Mellor (1961). The Role o Agriculture inEconomic Development,American Economic Review 51:566-93.

7/29/2019 ERR77.pdf

34/42



Knudsen, O., and A. Parnes (1975). Trade Instability and EconomicDevelopment. Lexington Books, Lexington, MA.

Koester, Ulrich, Gavin Olney, Charles Mataya, and Tamanda Chidzanja(2004). Status and Prospects o Malawis Tobacco Industry: A ValueChain Analysis.Ministry o Agriculture, Malawi. November.

Krueger, A. (1980). Trade Policy as an Input to Development,American

Economic Review 70:288-92.

Krueger, A. (1990). Perspectives on Trade and Development. University oChicago Press.

Landes, Maurice, Suresh Persaud, and John Dyck (2004).Indias PoultrySector: Development and Prospects, WRS-04-03. U.S. Department oAgriculture, Economic Research Service. January.

Lee, Jang-Yung (1997). Sterilizing Capital Infows,Economic Issues 7.International Monetary Fund, Washington, DC.

Lieert, William, and Suresh Persaud (2009). Transmission o Changes inExchange Rates to Agricultural Prices, Economic Research Report No.76. Economic Research Service, U.S. Department o Agriculture. July.

Levine, R., and D. Renelt (1992). A Sensitivity Analysis o Cross-CountryGrowth,American Economic Review 82:942-63.

Lewer, Joshua J., and Hendrik Van den Berg (2003). How Large isInternational Trades Eect on Economic Growth? Journal o EconomicSurveys 17(3):363-96.

Loayza, Norman V., Romain Ranciere, Luis Seren, and Jaume Ventura

(2007). Macroeconomic Volatility and Welare in Developing Countries:An Introduction, The World Bank Economic Review 21(3):343-57.

MacBean, Alasdair I. (1966).Export Instability and Economic Development.Harvard University Press, Cambridge, MA.

Maleta, Marc (2004). Malawi Tobacco Sector Perormance Audit: Reviewo the Perormance o Sector Institutions. Ministry o Agriculture,Malawi. November.

Menon, J. (1995). Exchange Rate Pass-Through, Journal o EconomicSurveys 9:197-231.

Ozler, Sule, and James Harrigan (1988). Export Instability and Growth.Department o Economics Working Paper No. 486, University oCaliornia, Los Angeles.

Persaud, Suresh, and Maurice Landes (2006). The Role o Policy andIndustry Structure in Indias Oilseed Markets. Economic Research ReportNo. 17, U.S. Department o Agriculture, Economic Research Service.April.

7/29/2019 ERR77.pdf

35/42

29


Persaud, Suresh, and Luther Tweeten (2002). Impact o AgribusinessMarket Power on Farmers, inAgricultural Policy or the 21st Century,Luther Tweeten and S. Thompson, eds. Iowa State University Press,Ames, Iowa.

Pryor, Frederic L. (1990). Malawi and Madagascar, in The PoliticalEconomy o Poverty, Equity, and Growth, Deepak Lal and Hla Myint, eds.Oxord University Press, UK.

Quiroz, Jorge, and Raimundo Soto (1995). International Price Signals inAgricultural Prices: Do Governments Care? GERENS and ILADES/Georgetown University, Washington, DC.

Ramey, Garey, and Valerie Ramey (1995). Cross-Country Evidence onthe Link Between Volatility and Growth,American Economic Review85(5):1138-51.

Rodrik, Dani (1998). Where Did All the Growth Go? External Shocks,Social Confict, and Growth Collapses, NBER Working Paper Series(6350). National Bureau o Economic Research, Cambridge, MA.

Roy, Atrayee, and Hendrik Van den Berg (2000). Are Petroleum Exportsan Engine or Growth? Time-Series Evidence or Five Oil Exporters,Journal o Energy and Development26:55-69.

Schuknecht, Ludger (1998). Tying Governments Hands in CommodityTaxation. Sta Working Paper ERAD-97-06, World Trade Organization,Geneva, Switzerland. May.

Thompson, Sarahelen (1985). Use o Futures Markets or Exports by Less-Developed Countries,American Journal o Agricultural Economics67(5):986-91, Proceedings Issue. December.

Tobin, Richard J., and Walter I. Knausenberger (1998). Dilemmas oDevelopment: Burley Tobacco, the Environment and Economic Growthin Malawi, Journal o Southern Arican Studies 24(2):405-24.

Townsend, Robert E. (1999).Agricultural Incentives in Sub-Saharan Arica,World Bank Technical Paper No. 444. Washington, DC.

Ukpolo, Victor (1994). Export Composition and Growth o SelectedLow-Income Arican Countries: Evidence rom Time-Series Data,Applied Economics 26:445-49.

United Nations Food and Agriculture Organization (FAO) (2007).FAOSTAT database. Rome.

United Nations Food and Agriculture Organization (FAO) (2003). WTOAgreement on Agriculture: The Implementation Experience - Developing

Country Case Studies. Rome.

United Nations Food and Agriculture Organization (FAO) (2004). The Stateo Agricultural Commodity Markets. Rome.

7/29/2019 ERR77.pdf

36/42



U.S. Department o Agriculture, Foreign Agricultural Service (2000).Malawi Tobacco and Products Annual 2000, GAIN Report #MI0001.July 20.

Varangis, Panos, Akiyama Takamasa, and Donald Mitchell (1995).Managing Commodity Boomsand Busts. World Bank, Washington, DC.

Vohra, Rubina (2001). Export and Economic Growth: Further Time Series

Evidence rom Less-Developed Countries,International Advances inEconomic Research 7(3):345-50.

Voivodas, Constantine S. (1974). The Eect o Foreign ExchangeInstability on Economic Growth,Review o Economics and Statistics56:410-12.

von Cramon-Taubadel, Stephan, and Jens-Peter Loy (1996). PriceAsymmetry in the International Wheat Market: Comment, CanadianJournal o Agricultural Economics 44:311-17.

Wacziarg, R. (2001). Measuring the Dynamic Gains rom Trade, World

Bank Economic Review 15:393-429.

White, H. (1980). A Heteroskedasticity-Consistent Covariance MatrixEstimator and Direct Test or Heteroskedasticity,Econometrica 48:817-38.

Wiebe, Keith D., Meredith J. Soule, and David E. Schimmelpennig (1998).Agricultural Productivity and Food Security in Sub-Saharan Arica.Food Security AssessmentGFA-10. Economic Research Service, U.S.Department o Agriculture. December.

Wolram, R. (1971). Positivistic Measures o Aggregate Supply

Elasticities: Some New Approaches - Some Critical Notes,AmericanJournal o Agricultural Economics 53:356-59.

World Bank (1999). Dealing with Commodity Price Volatility inDeveloping Countries: A Proposal or a Market-Based Approach.Discussion Paper or the Roundtable on Commodity Risk Management inDeveloping Countries, Washington, DC. September.

World Bank (2007). World Development Report 2008: Agriculture orDevelopment. Washington, DC.

World Bank (2009). World Development Indicators (WDI) database.

Washington, DC.

World Bank (1993). The East Asian Miracle: Economic Growth and PublicPolicy. Washington, DC.

World Bank (1991). World Development Report 1991. Washington, DC.

Yabuki, Nanae, and Takamasa Akiyama (1996). Is Commodity-DependencePessimism Justied? World Bank Policy Research Working Paper 1600.Washington, DC.

7/29/2019 ERR77.pdf

37/42

31


Yotopoulos, Pan, and Jerey B. Nugent (1976).Economics o Development.Harper and Row, New York.

Zeller, Manred, Aliou Diagne, and Charles Mataya (1998). Market Accessby Smallholder Farmers in Malawi: Implications or TechnologicalAdoption, Agricultural Productivity and Crop Income,AgriculturalEconomics 19:219-29.

7/29/2019 ERR77.pdf

38/42



AppendixData Sources, Model Parameters, andSimulation Results

The underlying conceptual ramework or empirically investigating the rela-tionship between exports and economic growth is based upon an aggregateproduction unction (APF) model. The APF model posits that, along with

conventional inputs (capital and labor), unconventional inputs (exports andother variables) may be included in the model to capture their contribu-tions to economic growth. This approach has been used by Dawson (2005),Al-Yousi (1997), Feder (1983), Fosu (1990), and Ukpolo (1994).

The APF can be expressed as ollows:

(A1) Yt = AtKa

tLb

t ,

where Yt is real aggregate output, Kt and Lt are capital and labor inputs,respectively, and At is the level o total actor productivity. Exports canincrease total actor productivity by expanding knowledge transers, by

improving access to better technologies, inputs, and intermediate goods,by permitting a country to borrow external capital at more avorable terms,and by allowing an economy to realize economies o scale and scope.Accordingly, At can be expressed as a unction o exports, Xt, and otherexogenous actors, C:

(A2) At = g(Xt, C) = Xr

tC .

Next, equation A2 is combined with equation A1 to obtain

(A3) Yt = CKa

tLb

tXr

t ,

where a, b, and r are the elasticities o output with respect to capital, labor,and exports, respectively.

Taking natural logs, ln, o both sides o equation A3 results in the ollowingunction, which is linear in the parameters:

(A4) lnYt = c + alnKt + blnLt + rlnXt + vt ,

in which all coecients are constant elasticities, c is a constant parameter,and vt is the usual error term, which refects the infuence o all other actors.We take the analysis a step urther by disaggregating the export value vari-able. Specically, we decompose exports into tobacco (TX) vs. nontobacco

(NTX) components, yielding,

(A5) lnYt = c + alnKt + blnLt + dlnTXt + flnNTXt + vt ,

where d and f are elasticities o output with respect to tobacco and nonto-bacco exports, respectively. Although tobacco comprises the majority oMalawis

ERR77.pdf

Documents