The Poverty Discourse and the Poor in Sri Lanka

8/3/2019 The Poverty Discourse and the Poor in Sri Lanka

1/22

The Poverty Discourse and the Poor in Sri LankaAuthor(s): Lakshman YapaSource: Transactions of the Institute of British Geographers, New Series, Vol. 23, No. 1 (1998),pp. 95-115Published by: Blackwell Publishing on behalf of The Royal Geographical Society (with the Institute of

British Geographers)Stable URL: http://www.jstor.org/stable/623159 .

Accessed: 21/09/2011 02:20

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

Blackwell Publishing and The Royal Geographical Society (with the Institute of British Geographers) are

collaborating with JSTOR to digitize, preserve and extend access to Transactions of the Institute of BritishGeographers.
http://www.jstor.org/action/showPublisher?publisherCode=blackhttp://www.jstor.org/action/showPublisher?publisherCode=rgshttp://www.jstor.org/action/showPublisher?publisherCode=rgshttp://www.jstor.org/stable/623159?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/623159?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=rgshttp://www.jstor.org/action/showPublisher?publisherCode=rgshttp://www.jstor.org/action/showPublisher?publisherCode=black


2/22

95T h e p o v e r t y discourse a n d t h e p o o r

in S r i L a n k a

Lakshman YapaSri Lanka is cited as an exemplary case of direct poverty alleviation because ofa long history of social welfare and high values in quality of life indices.Notwithstanding, anti-poverty measures in Sri Lanka founded on the internationaldiscourse of poverty and development do not serve the interests of poor people.This discourse begins by locating poor people in a distinct poverty sector andproceeds to examine its characteristics. Several attributes of that discourse make itintellectually incapable of seeing how poverty is socially constructed in a diffusednexus of production relations that extends far beyond the so-called 'poverty sector'.An alternative 'substantive approach to poverty' is presented. The arguments areillustrated using the theme of food production in Sri Lanka.key words Sri Lanka poverty discourse theory nexus of production relationsconstructed scarcity substantive approach to poverty food productionDepartment of Geography, Pennsylvania State University, University Park, Pennsylvania 16802, USAe-mail: [email protected] manuscript received 29 May 1997

Poverty exists when people are unable to satisfytheir basic needs for food, clothing, shelter andhealth. It is widely believed that poverty is causedby economic underdevelopment. Within thediscourse of development and underdevelopment,the most common approach to the study of povertybegins by defining a 'poverty sector'. Researchersthen try to understand causes of poverty by study-ing the characteristics of this sector and its people.It is my view that this approach is not a helpfulway to address problems of poor people. In thispaper I argue for an alternative 'substantiveapproach' to poverty. My argument is founded onseveral claims:* the causes of poverty cannot be understood bymaking poor people the object of our study* the causes of poverty are embedded in a nexusof production relations diffused throughout thelarger society that extends far beyond theso-called 'poverty sector'* material deprivations experienced by the poorare socially constructed at every node of thenexus of production relations

* the existing academic discourse on poverty con-tributes to that scarcity and conceals the socialorigin of scarcity* the issue of poverty may more helpfully beaddressed by moving away from the presentreified representation of poverty to a substantiveapproach - why do particular groups in specificplaces experience hunger, malnutrition, lack ofshelter, etc?These theoretical claims are explained at somelength in Yapa (1993, 1996a, 1996b) and in Yapaand Wisner (1995). In this paper I illustrate theargument by focusing on the theme of food in SriLanka. Important as it is, my concern is not foodperse; I present the empirical narrative on food asan example of an alternative way of speakingabout poverty.A preliminary word on what this paper is not. Itcontains no 'proof' that the substantive approach Ipropose is superior to competing explanations ofpoverty. Sri Lanka has a well-established andextensive poverty discourse that includes liberal,conservative, marxist and populist approaches to

TransInst Br GeogrNS 23 95-115 1998ISSN 0020-2754 ? Royal Geographical Society (with the Institute of British Geographers) 1998


3/22

Table I Sri Lankan GNP per capita and quality of life indiceswithin the world economy, 1990GNP per Infantcapita Life mortality PercentageUS$ expectancy (/000) literacy

Sri Lanka 470 71 19 88High-income economies 19 590 77 8 94Middle-income economies 2220 66 48 78Low-income economies 320 55 92 55Source:World Bank (1992)

the problem. Despite profound philosophical andtactical differences in their approaches, they all seeeconomic development - an expansion of produc-tion forces and increasing income - as the solutionto poverty (Yapa 1996b). It is this axiomatic beliefthat I contest. In this paper I have neither describedthe competing discourses on poverty, nor subjectedeach to critical inquiry. Instead I present my ownideas as an invitation to a conversation and adebate about the axiomatic belief that developmentwill eradicate poverty.Sri Lanka is cited as an exemplary case of directpoverty alleviation because of a long history ofsocial welfare and high values in quality of lifeindices. The first part of the paper presents a briefreview and critique of these measures. The secondpart engages various aspects of the conventionaldiscourse on poverty before the concept of sociallyconstructed scarcity is presented in part three as asubstantive approach to poverty. The fourth partemploys the scheme of the nexus of productionrelations to illustrate the social construction of foodscarcity in Sri Lanka.

Poverty alleviation in Sri LankaDespite a very low per capita income, Sri Lankahas achieved remarkably high levels in quality oflife indices such as life expectancy, infant mortalityand literacy. This record is impressive as thesemeasures are closer to those of countries whoseGNP (gross national product) per capita is nearly50 times as great (Table I). Many scholars believesuch statistics reflect the benefits of sustained gov-ernment intervention in social welfare.1 Indeed, SriLanka is often cited as a 'test case' of the efficacyofdirect public intervention in poverty alleviation(Anand and Kanbur 1995, 228).

A study of poverty in Sri Lanka is particularlyinstructive because the country has a sustainedhistory of anti-poverty programmes from colonialtimes. The state, which has consistently presenteditself as the patron of common people, has carriedout a series of costly social welfare programmesproviding free health care, free education and evensubsidized food. And yet, according to the CentralBank of Sri Lanka (Gunaratne 1987), a fifth of allhouseholds do not consume the required mini-mum in caloric food energy. Whilst it is certain thatthe poor in Sri Lanka would have been worse offwithout such programmes, providing welfare isnot the same as 'solving the poverty problem'.Anti-poverty measures are (in)formed by the waypolicy-makers understand poverty and that under-standing depends in turn on the way academicshave represented poverty. My argument is thatanti-poverty measures cannot solve the povertyproblem in Sri Lanka because they are founded ona discourse that perpetuates the very problem theyare designed to solve.Food subsidies and rationing, first introducedin 1942 by the British as part of a wartime reliefmeasure, continued into the postwar period,becoming a cornerstone of Sri Lanka's post-colonial welfare services (De Silva 1981). Afterindependence in 1948, programmes of social wel-fareexpanded in scope with expenditure absorbingover 56 per cent of the annual government budget(ibid.).Expenditure on public health, hospitals, freedrugs, free education and subsidies for importedessential foods continued at expanded levels,despite a deteriorating foreign account balance(Fig. 1). In 1977, the food subsidy alone absorbedaround 17 per cent of the government budget(Bandaranaikeand de Alwis 1987).Despite severalchanges in political leadership between 1956 and1977, the broad outline of poverty policy remained

96 LakshmanYapa


4/22

Thepovertydiscourseand the poor in SriLanka1Social Expenditure as a Pecentage of GNP

AtLz-So

C196-1977: Period of closed werst economy

4 1960-1977: Perod of closed wearlst economy14 1 1 1 1 1 1 1 1 I l l l l l l l l l l l l l l l i I I I53 55 57 59 61 63 65 67 69 71 73 75 77 79 81Year

Figure 1 Social expenditure as a percentage ofGNP, 1951-82

intact with social expenditure on education, healthcare and food subsidies taking up almost 10 percent of the GNP each year (Anand and Kanbur1995). But the economy grew so slowly that SriLanka was cited in the development literature as acase of 'equity-with-no-growth' (Chenery 1989, 21).The open economy (1977-94) ushered in a newperiod. It brought to an end a distinct phase ofsocial welfare legislation in a closed economy withstate control of exports, imports, plantations,industries, transport and foreign exchange. Thenew policy dismantled import controls, exchangeregulations and price controls, and several state-owned enterprises were privatized. Expenses onsocial welfare were reduced. The governmentadopted the World Bank's notion of providing asafety net for the poorest of the poor by 'targeting'food subsidies and social welfare benefits. In 1979,the food subsidy scheme was eliminated andreplaced with an American-style food stamp pro-gramme. Besides the traditional welfare measuresin food, education and health care, there weretwo new programmes of poverty alleviation: theNational Housing Program (Siriwardena 1994)andthe Janasaviya people empowerment) Programme.Janasaviya eceived funds from international devel-opment agencies because its logic was compatiblewith the new development philosophy of freemarkets, investment in human capital, enterprisedevelopment and safety nets for the poor (Stokke1995).There are several aspects of the open economythat had a direct bearing upon poverty alleviationmeasures. The economy grew at a very rapid rate(Rajapathirana1988) but the growth was fuelled bya massive influx of foreign capital combined with

97deficit financing (Sahn 1987). Despite the free-market philosophy, the scope of the governmentexpanded during this time. Unprecedented levelsof government expenditure and expanded creditfuelled a very high rate of inflation: the cost ofliving index (1952=100) rose from 203 in 1977 to1408 by 1993 (Shastri 1995). Rapid inflation had adeleterious impact on the poor through a decline inreal wages and the value of food stamps (Sahn1987).Therewere also large outlays of public fundson internal security and the military following theoutbreak of ethnic violence in 1983 and insurgencyamong Sinhalese youth (Warnapala1993).A principal debate of the poverty discourse inSri Lanka concerns the efficacy of direct (basicneeds) as distinct from indirect (economic growth)measures in promoting social welfare. Measures ofpoverty alleviation undertaken during the period1960-77 were direct, while post-1977 policies areseen as indirect. Although Sri Lanka is frequentlycited as a country which has successfully imple-mented the direct approach, Bhalla and Glewwe(1986) have argued that the improvement in livingconditions during the years 1960-77 was not par-ticularly remarkable when compared to othercountries and that Sri Lanka's achievements inquality of life should not be attributed to the largesocial expenditures made by the government dur-ing these years. Several scholars have questionedthis line of argument on the grounds that the highquality of life in Sri Lanka should be seen as apart of a long and sustained history of socialinvestments and not just that of the 1960s.2Interesting as it is academically, this debateon the relative merits of direct and indirectapproaches to social welfare is not very helpful inconsidering poverty alleviation in Sri Lanka. Itseems obvious that if a government spends moneyon providing cheap food, education and healthservices, the quality of life will be better than if themoney were not spent for that purpose. In 1977,even after 30 years of post-independence welfareprogrammes, nearly half the population was eli-gible to receive food stamps. Furthermore, thedebate about the superiority of direct or indirectmeasures is not one that can be resolved in amethodological sense. It is not possible to take astandard of living and apportion changes in valuesof its various components to this or that cause. Forexample, suppose that in a given year there was a1 per cent drop in the rate of infant mortality, wecannot know whether this was caused by increased

14


5/22

98expenditure on the health budget or by expenses ineducation which helped parents to become bettercarers,or by better nutrition purchased with higherincomes, or by some other cause.This brief review of the history of poverty allevi-ation in SriLankais designed partly to focus on therole of the state. All shades of political leadershipbelieve poverty to be an economic problem whichcan be eradicated through economic developmentand social welfare, and the state is central to bothtasks. During the period 1956-77 the state cameto control the economy and thus became the chiefjob provider. This, coupled with a slow growtheconomy and the contraction of the private sector,led to a greater dependence of people on thestate for their economic well-being, leading tonepotism and graft. Dependent as they were ongovernment jobs, the Tamil middle class founditself at an increasing disadvantage with thegrowth in the state power of the Sinhala Buddhists(Nithiyanandan 1987; Piyadasa 1988; Tambiah1986).After 1977,although there were more privatesector jobs available, the state continued to grow inpower at the expense of civil society. The longinvolvement of the state with poverty alleviationchannelled imagination, discourse and initiativeinto the narrow space of its political economy.Simultaneously, this had the effect of concealingthe multiple sites at which the problem of povertycould have been addressed creatively. This is thewider argument of the paper: poverty policies inSri Lanka are grounded in a discourse that isintellectually incapable of seeing how scarcity issocially constructed in a largernexus of productionrelations.

The poverty sector approach and reificationof povertyThe primary object of my critique is the official -i.e. state - discourse on poverty in Sri Lanka. Theofficial methodology is identical to that advocatedby the World Bank. This discourse begins with adescription of the poor in the poverty sector - theobject of study. The poverty sector is viewed as adistinct entity with stable internal characteristicswhose study will reveal the causes of poverty(Yapa1996a).The point of such a study is to deviseoperational methods that will help 'solve' the prob-lem. This is what I mean by reification of poverty:the lack of basic needs by large numbers of people

LakshmanYapa|ThePovertyProfileT I

Poverty indicators Distinguishingcharacteristics of the poorObs. No. Y1 Y2 X1 X2 X3 Xk12345

N

Figure 2 The poverty profile as a databasehas been transformed into a quantifiable povertyproblem existing in a distinct and coherent sectorwith stable inner characteristics,the study of whichwill reveal the causes of poverty and thus help usto find solutions. As reasonable as it seems, thisapproach is not helpful.The first step in the poverty sector approach is toconstruct a poverty profile (Fig. 2); visualize this asa data table with n rows and k columns. Then rows represent individuals, households or places(census tracts, counties, villages, districts and soon). The k columns contain two types of infor-mation about the observations: poverty indicatorsand 'distinguishing characteristics of the poor'.Poverty indicators include such things as incomeand status on a poverty line. Distinguishing char-acteristics are variables such as location, urban/rural, gender, race, ethnicity, family size, maritalstatus, employment, occupation, education, assetsand access to markets. In most analyses the distin-guishing characteristics of the poor are the inde-pendent variables that 'explain' magnitudes of thepoverty indicators. There is no implication herethat all poverty studies follow a statistical regres-sion format, although many do. Even a simple twoby two table - for example one that shows povertyby gender - is constructed with the implicationthat gender, among other factors, 'explains' pov-erty. The independent variables in poverty studiesare subdivided into two groups - those amenableto change such as education and employment, andthose that are not, such as gender and race. Thelatter may help to pinpoint beneficiary targetgroups of poverty programmes.According to the World Bank's Povertyreductionhandbook 1993, 13-17), 'The poverty profile por-trays the extent and nature of poverty and thedistinguishing characteristics of the poor ... [it] is


6/22

Thepovertydiscourse and the poor in SriLankaa snap-shot of the poor'. To solve the problem, wemust first find out who the poor are, what theextent of their poverty is, where they live and whatcauses their poverty. This is based on the belief thata study of the poor will reveal why they are poor.In the words of a Sri Lankan sociologist well-known for his poverty studies (Ratnapala1989, 20):'A pre-condition for understanding poverty in SriLanka is the understanding of the target com-munity whose poverty we are attempting todefine'. This is why surveys are the most popularinstrument of poverty research.The main source ofinformation on Sri Lankan poverty is the island-wide Reporton consumer inancesand socioeconomicsurvey (1985) conducted by the Central Bank of SriLanka.Dualistic thinking pervades the entire notionof a poverty sector which is viewed as a distinct,measurable, bounded entity, that part of theeconomy where the poor reside - the locus of thepoverty problem; those who are not poor reside inthe realm of the non-problem. The poverty sectorhas little capital and no resources. Presumably thatis why it is poor. Capital, technology and resourcesmust be infused from outside. The sector of thenon-poor is the seat of intellect, resources andsolutions - the knowing subject reflecting on theproblems of the needy object, an idea well-captured in the term 'poor as target group'. Aninteresting example of the subject/object binary isthe work of Ratnapala (1985), who lived amonga group of beggars after disguising himself asone. He described what he called the 'living-in-experience' methodology in his book Thebeggar nSri Lanka.Ratnapala's objective was not to providean ethnographic description of the life of beggarsin Sri Lanka; he wanted to know the reasons whybeggars become beggars. Ratnapala's writing istypical of a wider Sri Lankan discourse also heardin the sermons of the clergy and civic leaders. Ittakes the stance of a knowing, compassionate sub-ject reflecting on the poor as the needy object: theother who needs to be educated, morally uplifted,materially advanced and culturally modernized.Poverty sector thinking is an example of thelogic of 'internalism' which assumes that causes ofpoverty lie 'internally' within the characteristicsof the poor. Conceptually, each observation in thepoverty table - that is, each row - is treated as anintegral unit: the direction of causation runs acrossthe row from the distinguishing characteristics ofthe poor to the poverty indicator (Fig. 2). The

99degree to which values of the independent vari-ables 'co-vary' with the poverty indicator willdetermine how much of total variation has been'explained', and which variables contribute to thatexplanation. Although I have used the language ofstatistical regression to communicate this notion, itis clear that the logic of 'internalism' is present inany contingency table of poverty statistics. In 1987,the Central Bankof Sri Lankaused Rs69 per monthper person in 1978/9 prices as the poverty line, thisbeing the income required for a minimum caloriediet. Overall, 23.6 per cent of the total populationand 20-1 per cent of households were in poverty(Gunaratne 1987). In an effort to find explanatorycauses, the poor were broken down by numerouscategories such as socio-economic group, status ofemployment, education, age, gender, householdsize and geographic location (ibid.). I argue thatthe causes of poverty cannot be discovered bystudying such tables.Another characteristic of the poverty sectorapproach is 'operationalism'. I refer to the beliefthat 'the problem can be solved' through appropri-ate policy and target group strategies. Thisthinking provides the very raison d'etre of insti-tutions such as the World Bank. According to theBank's president (World Bank 1993, Foreword),

povertyreduction s the overarching bjectiveof theWorld Bank. It is the bench-markby which ourperformanceas a development institutionwill bemeasured.The Bank's approach is nicely captured in thetitle of their widely distributed authoritative pub-lication, the Povertyreductionhandbookibid.),which'summarizes the Bank procedures and guidelinesfor operations in poverty reduction'. The WorldBank's concept of the poverty gap - the incometransfer needed to lift everybody above the pov-erty line - is a good example of what I have called

operationalism. Using identical logic, the CentralBank of Sri Lanka calculated the poverty gap at3.14, meaning that poverty can be eliminated bytransferring 3.14 per cent of total income from thenon-poor to the poor (Gunaratne 1987). Povertyin Sri Lanka (or elsewhere) has no operationalsolution, at least not in the sense offered by theofficialdiscourse of the WorldBank and the CentralBank of Sri Lanka.The poverty sector approach is embedded ina wider discourse whose characteristics includewhat may be called 'economism' and 'political


7/22

100economism'. 'Economism' is the idea that lack offood, shelter and health care is an economic prob-lem that calls for economic solutions. According tothe governor of the Central Bank of Sri Lanka(1987, 4),

There s no doubt that successin povertyalleviationprogrammesargelyhingeson the abilityto maintainsustainedeconomicgrowth .. [to]enablethe poortoget morejobs,higher ncomes,and higher aborforceparticipation.By the term 'political economism', I refer to thedominant role of state power in affairs of theeconomy and civil society. In this sense politicaleconomism is evident throughout the povertydiscourse: since poverty is believed to represent amalfunctioning of the economy (or a lack of

economic development), the state is expected toplay an active role through corrective measures(liberal political economy) or direct intervention(radical political economy). And, as Gramsci(1971, 160) has argued, even a policy of non-intervention like laissez-faire is established bystate regulation and maintained 'through legis-lative and coercive means'. Sri Lanka is anexcellent example of political economism in thepoverty discourse. Programmes for poverty alle-viation, social welfare and creation of employ-ment figure prominently in the electoral politicsof Sri Lanka (Warnapala 1993). Throughoutpostcolonial history, the official discourse ofgovernment publications, policy documents andparty manifestos shows that poverty was viewedas an economic problem and the primary re-sponsibility for poverty alleviation lay with thestate.A summary of the argument so far.In Sri Lanka,poverty is viewed as residing in a concrete sector ofthe economy. Those who are not poor and thosewho study the poor see themselves as distinct andseparate from those in the poverty sector. Theofficial analysis of poverty begins with the con-struction of a poverty profile - a description of whothe poor are,what the extent of their poverty is anda list of characteristics that explains why they arepoor. The causes of material deprivation are soughtfrom within these distinguishing characteristics.Both academics and the public in Sri Lanka seepoverty as an economic problem. There are twoaspects to this economism. The first is the notionthat poor people have characteristics that preventthem from participating fully in the economy and

LakshmanYapathe second that the economy itself needs develop-ment because it is now underdeveloped. Thosewho emphasize the first seek 'direct' solutions topoverty and others look to development for 'indi-rect' solutions. The state in Sri Lanka was a keyagent in the poverty discourse. This is because it iswidely believed that the poverty problem has asystemic deep structure so that only a powerfulagent like the state has the capacity to interveneand rectify these flaws. The manner in which statepower was exercised had at least three unfortunateconsequences. First, anti-poverty initiatives cameto be seen increasingly as a matterof implementinggovernment programmes of social welfare. Sec-ondly, poverty and development acted as instru-ments which facilitated the rise of state power atthe expense of civil society. And finally, the beliefthat the state possessed 'operational' solutions topoverty took attention away from numerous othersites at which agents outside the state could havefunctioned. The net result of the discursive ele-ments of poverty acting in concert was to producea discourse that is intellectually incapable of seeinghow material deprivation is socially constructedwithin a larger nexus that extends far beyond theso-called 'poverty sector'.

A substantive approach to povertyInstead of viewing poverty in the abstract as aneconomic problem, I wish to treat it in a concrete,substantive manner. To the question of why sub-stantial numbers in Sri Lanka lack adequate basicgoods such as food and housing, I am not satisfiedwith the answer that 'they do not make enoughmoney'. By situating each basic good within thenexus of production relations, we can 'uncover'how scarcity is socially constructed at each site ornode of a network of relations - technical, social,cultural, political, ecological and academic - dif-fused throughout the larger society. Hence eachsite is also a locus of 'opportunity' creatively toengage the very forces that create scarcity (Fig. 3).Poverty has no root causes because scarcity iscreated everywhere. There can be no grand projectof economic development or policy solution topoverty (Yapa 1996a).The Central Bank of Sri Lanka (1987) definedpoverty as the lack of income to buy the basicminimum in food caloric energy. However, insteadof dealing with poverty as an abstract economic


8/22

Thepovertydiscourseand the poor in Sri Lanka

SOCIAL -' P- POLITI(Figure 3 The nexus of production 3poverty

issue of inadequate income, I shall asdirect substantive question: why do 2CSri Lanka's 16 million people not ha,mum calories they need? In seeking ansquestion, we shall see how food scarcilconstructed at numerous sites, manyreach of normal poverty alleviation pBy looking at the circumstances surroubasic needs, such as housing, transporcare, there is a much larger story to tesocial construction of scarcity in Sri L


9/22

102initiated to serve the paddy and subsistence needsof the peasantry.Moreover,colonial science viewedpeasant agriculture as primitive, unproductive andbackward. While many colonial officers admiredthe industriousness of the neatly laid-out paddymonocultures, they were not convinced of thevalue of home gardens and chenaswhich appearedhiggledy-piggledy, jungle-like and uncultivated.These impressions formed the basis of formaltheorizing in the dual economy model that concep-tualized the plantations as modern, dynamic andproductive, and peasant agriculture as traditional,static and unproductive. The backwardness of thepeasants' economy was attributed to their cultureof primitive values and traditions.The fourth aspect is the influence of the colonialdiscourse on local elites. In 1912, the Departmentof Agriculture was established in response todemands for institutional support for 'peasantagriculture'.However, as Pain (1981,4) has pointedout, 'the early activities and focus of theDepartment reflected the economic crop bias of theBotanicalGardens'. Colonial assessment of peasantagriculture was accepted, adopted and elaboratedby educated Sri Lankan elites who administeredthe Ministries of Agriculture, Land Developmentand Irrigation and taught at the PeradeniyaUniversity Faculty of Agriculture (Brohier 1975;Goonatilake 1984). The educated Sri Lankan eliteswere closer in outlook, values and philosophy totheir British counterparts than they were to SriLankan peasant farmers. These attitudes were toplay a majorrole in the formulation of agriculturalpolicy in post-1948 Sri Lanka.

Nexus of food relations in Sri LankaDespite the heavy emphasis on agriculture duringthe last three decades, the poor in Sri Lanka facepersistent shortage of food. The average intake offood per day is about 2200 calories and it hasgrown at a modest 0-2 per cent per year (FAO1993). The modest average annual growth in cal-orie intake has not been steady; there was a seriousfall in the national average during the years 1970-5and another decline between the years 1985-90. Itmust be remembered that the poorest groups con-sume well below this modest national average(Census and Statistics 1993; Central Bank of SriLanka 1987). Throughout the 1960s and 1970s thecost of living index increased at a modest rate butbetween 1980 and 1991 it increased dramatically

LakshmanYapa

EEc

61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91Year

Figure4 Index of food prices, 1961-91(1980=100)

690, .

X x ? r-: v/ . ,,

70A Totaloodproduction

60 ---| Food productionpercapita

61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93Year

Figure 5 Index of total and per capita foodproduction, 1961-91(1979-81 = 100)

from a value of 100 to 375, much of this increasebeing fuelled by the price of food (Fig. 4). Suchprice increases are highly regressive because theaverage household spends about 71 per cent of itsincome on food items while the poorest groupsspend nearly 95 per cent (Census and Statistics1993; Central Bank of Sri Lanka 1987). Accordingto the United Nations' Food and AgricultureOrganization (FAO 1993), food production in SriLanka increased by 2-1 per cent per year between1961 and 1991 but per capita food productionincreased by only 0.4 per cent. Total food pro-duction stagnated throughout the 1980s whileper capita food production declined dramatically(Fig. 5). The shortfall in food production wasrectified by importing wheat. Over the last 30 yearsthe wheat tonnage has grown at a rate of 3.4 percent per year and the dollar outlay on the importshas grown by 6-2 per cent per year (ibid.).


10/22

Thepovertydiscourseand the poorin SriLanka

8tDEmc0'a

300Index fRiceProduction

250

200

150

100

50

61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93Year

Figure 6 Index of rice production, 1961-91(1961=100)

ThecontemporaryiceeconomyThe rice economy in Sri Lanka has flourishedover the last 30 years, despite the stagnation inoverall food production. Between 1961 and 1991total rice production increased by over two and ahalf times (Fig. 6) with much of this increasecoming from the widespread adoption of high-yielding varieties. Surprising as it may seem, thesuccessful diffusion of high-yielding rice narrowedthe range of options for production and reduceddiversity in diet. I shall illustrate this by placingthis rice within a nexus of production relations andshow how scarcity has been constructed at eachnode of this nexus.Successive governments in postcolonial SriLanka have made rice a centrepiece of statepolicy. But why the emphasis on rice? First, it isthe staple diet of Sri Lankans. Secondly, import-substitution of rice was seen as a way of savingvaluable foreign exchange spent on rice imports.Thirdly, food subsidies have existed in Sri Lankasince the Second World War and their costsbecame excessive. In 1977, the rice subsidy costaround 17 per cent of government expenditure.However, the reasons for the 'rice bias' are morecomplex.The ruling classes in post-independence SriLanka have gone to great lengths to preserve thepeasant economy. Despite party philosophy, suc-cessive administrations channelled revenue fromthe export sector into large-scale irrigation to servepeasant paddy cultivators (Stokke 1992). The rea-sons for the state interest in the peasantry are manyand, as Moore (1989, 188) has argued, it is strongly

103tied to the elite's interpretation of Sinhalesehistory:5

the leading actors see themselves as fulfilling anhistoric mission which involves the recreationofan authentic, raditionalvillage, family-farming ndrice-based inhalese ociety.The three main symbols of Sinhala Buddhistnationalism are the temple (pansala), he irrigationtank (wewa) and the paddy field (kumbura).Riceproduction in Sri Lanka is not about food alone; itis also a powerful symbol of culture, restorationand history. According to Moore (ibid.,206),publicpolicyhas been unusually and, most peoplewould argue, ineffectively- focussed on rice, thesymbolically-potentrop,at the expenseof otheragri-culturalproducts.Thewhole focus of the workof theDepartment f Agriculturehas been on rice,and the'rice bias' can be seen in the spheresof extension,research, credit, irrigated land development andsubsidy.The postcolonial emphasis on irrigation andfood crops was a definite break with the colonialpast. However, as Moore pointed out, the focus onrice came at the expense of other agriculturalproducts. The Department of Agriculture was pre-occupied with rice as a commodity and paid lessattention to the sociology and polycultures of peas-ant farming systems. Improved rice was a symbolof economic development and modernization ofagriculture: monocultural rice farming, large-scaleirrigation, material inputs, extension and banking.But it was accompanied by a parallel logic of socialconstruction of scarcity: expensive rice, loss ofrural livelihood, reduced variety in foods and agrowing dependence on imported wheat.The following sections of the paper describe insome detail what these sites are. The logic here isnot one that lends itself to a neat cost-benefit

analysis in an econometric framework. By neces-sity I have chosen a narrative style to make thisargument. Visiting each site of the nexus begins tomake it possible to understand the mechanismsthat are at work and to reveal how the mechanismsat one site are constituted by those at other sites ofthe nexus. The remainder of the paper elaboratesthis claim.

Technical relations in high-yielding riceDespite the common belief in the backwardnessof traditional farmers, the Sri Lankan elites


11/22

104implemented a series of irrigation and land devel-opment policies in the Dry Zone designed topreserve and strengthen the peasantry.6The agri-cultural sciences responded to the political initia-tive by working on a programme to improve riceyields. Yet,despite impressive gains in rice produc-tion, the new discourse/practice of agriculturefocused on expensive commercial inputs and mar-ginalized less expensive 'low-tech' ways of raisingfood. Technology became a site at which scarcitywas socially constructed: it created new demandsfor expensive inputs and reduced existing sup-plies by marginalizing a range of alternativetechniques.7By the end of the 1960s Sri Lankan agriculturalscientists had produced an agro-ecological map ofthe island that showed the intense place-to-placevariation of optimal growing conditions of differ-ent crops; in particular, the map provided anappreciation of the range of paddy ecologies thatwere present in the island. Sri Lankan crop scien-tists had developed a line of locally adaptedimproved rice called the H varieties (and later BG)independently of the plant breeding efforts at theInternational Rice Research Institute (IRRI) n LosBaios, Philippines. This experience enabled themsuccessfully to resist the efforts of IRRIscientists topromote IR8 in Sri Lanka, a genetically uniformhighly vulnerable variety (Anderson et al. 1991).Despite the more enlightened efforts at agro-ecologically sensitive plant breeding, the SriLankan scientists continued to work within theIRRIparadigm which placed the central focus onbiological manipulation of plant characteristics toprovide high yields in response to intensiveapplication of technical inputs - improved seeds,chemical fertilizer, pesticides and irrigation.Before the spread of improved seeds, farmerskept healthy seed from each harvest as 'seedpaddy' for the next planting season. Now farmerspurchase seed from 'seed stores' because harvestsdo not provide seeds that are stable enough forplanting. It is true that the cost of seed paddy in SriLankais modest compared to costs of other inputs.However, 'store-bought seeds' make farmers morevulnerable because they have lost control over seedquality. The reproduction of seeds has moved fromfarmers' fields into the realm of formal science,experimental plots of research institutes, commer-cial seed suppliers and bureaucratic processes ofseed certification (Kloppenburg 1988; Mooney1979).

LakshmanYapa900800 | .Indexof FertilizerImports | ,700600

--Quant l 1 A400 Value300

61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91Year

Figure 7 Index of chemical fertilizer imports byquantity and value, 1961-91(961=100)

That these new seeds require large quantities ofchemical fertilizer and pesticides is an excellentexample of my argument about constructed scar-city. Most commercial fertilizers consist of somecombination of the chemicals nitrogen (N), phos-phorous (P) and potassium (K).Since the 1960s theMinistry of Agriculture, on the advice of theFaculty of Agriculture at Peradeniya University,undertook a massive campaign to promote the useof chemical fertilizers. Between 1961 and 1991 thequantity of chemical fertilizer imported to SriLankagrew at a steady rate of 2.2 per cent per yearwhile the cost of those imports grew at 5-9per cent.The total outlay on fertilizer imports during thatperiod was US$947 million (Fig. 7). The use ofchemical fertilizer was actively promoted usingstate subsidies. The average subsidy per year wasabout Rs1000 million - 1/60 of total governmentexpenditure in 1985. In previous years the propor-tion was 1/32 (FADINAP 1987; Palm and Sandell1989). Between 1989 and 1990 the price of a metricton of fertilizer rose from Rs3500 to about Rs9700when the government discontinued the subsidy,causing great hardship among small-scale paddyfarmers (National Fertilizer Secretariat1992).8There are several alternative sources of plantnutrients, including agricultural waste; greenmanure; the cultivation of leguminous crops andleguminous permanent trees; animal manure;human waste; compost; intercropping; companionplanting and crop rotation;biological nitrogen fix-ation, as in blue-green algae living in paddy fields;and slurry from the anaerobic decomposition oforganic waste in methane digesters.9 For example,the anaerobic decomposition of animal and human


12/22

Thepovertydiscourseand the poor in Sri Lanka1100o000 Indexof Pesticide ImportsbyValuei900 1'.800700 ^600 , P .'500 ,/ X400 I V,

\s300'200 / ,^ _^ . . . |--.- Dollarvalue |00 p 0I0I100 ..

o I I I I I ii I I I i ii i i I61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91

Year

Figure 8 Index of pesticide imports by value,1961-91(1961=100)

waste produces a clean combustible fuel (methane)in which the residual sludge forms an excellentsource of fertilizer. The technology is widelyknown in India and China (Lichtman 1983;McGarry and Stainforth 1979). Despite the wealthof information available, there has been very littlesupport for biogas technology in Sri Lanka. Duringthe mid-1980s there were only a few hundredfunctioning units in the country (Wijesinghe andChandrasiri 1986).10The use of pesticides in Sri Lanka for combatingcrop disease is another example of constructedscarcity. Again, there are several effective alterna-tive methods of pest control: cultural, genetic andbiological.11 Integrated pest management includesthe combined use of all or some of these methods,including chemical pesticides. Of these methodsof pest control, chemical pesticides receive themost support both from the government and theprivate sector in Sri Lanka. Originally confinedto the plantation sector, since the 1960s pesticideshave come increasingly to be used in the foodcrop sector. The change is primarily a result ofthe wide diffusion of improved rice varietiesduring this period (Abeysekera 1988). Over theperiod 1961-91 pesticide imports have grown atthe rate of 8.3 per year and cost nearly US$150million (Fig. 8).In such ways, scarcity has been socially con-structed at the technical site of the nexus of agri-cultural relations by creating new demands forexpensive inputs while simultaneously marginaliz-ing alternative, less expensive techniques. Thus thealleviation of poverty among farmers is not amatter of measuring their poverty, identifying tar-

105get groups and increasing income but should beginwith a systematic understanding of why their costsof production are high and why the new methodsare not sustainable. But consider the followingobjection to that argument: improved seeds, chemi-cal fertilizer and chemical pesticides are a part of arational system of agriculture that increased riceproduction in Sri Lanka. Surely 'alternative tech-niques' would carry higher opportunity costs interms of lower production and continued hunger.My claim about constructed scarcity is not about'rice production' itself. Constructed scarcity has todo with high costs of production, the marginaliz-ation of alternative and less expensive techniques,and the neglect of a range of other food strategies.Moreover, the productivist logic of the agriculturalstrategy in Sri Lanka does not permit us to see thediverse number of sites at which scarcity is sociallyconstructed.Ecological relationsAt the ecological node of the nexus, the new ricetechnology contributed to scarcity in two primaryways. First, the cluster of genetically uniformseeds, chemical fertilizer, pesticides and irrigationled to the degradation of the very conditions ofproduction which demanded higher levels of inputto maintain existing levels of output, inviting com-parison to a treadmill.12 Secondly, very heavyopportunity costs were incurred by ignoring arange of alternative technologies related touse-values created in nature and to people'sknowledge of local ecology, thus replacing the'reproductive capacity' of nature with the 'produc-tive capacity' of industrial inputs.Among the estimated 12 000 rice cultivars in theworld, Sri Lanka alone has recorded 2800 varieties,a product of thousands of years of selection, tra-ditional farming methods and the island's ecologi-cal diversity. The varieties are adapted to a range ofecological conditions. Some upland varieties areadapted to drought and low temperatures, andsome coastal varieties are adapted to submergence,flash floods and saline soils; some varietieshave medicinal properties and others are usedfor religious rituals. But modern rice farmingrelies on a few imported seed types with anarrow genetic base of about five to ten varieties,displacing the diversity of traditional rice. Lessthan 5 per cent of the total area under paddygrows the traditional indigenous cultivars(NARESA 1991).

11

Exc)


13/22

106Genetically uniform varieties of rice, wheat andcorn grown in monocultural stands are more vul-nerable to pests and pathogens than older varieties- which have coevolved with the local environment- thereby necessitating the use of pesticides and

so, as suggested above, contributing to the socialconstruction of scarcity (Bull 1982). The heavyconsumption of nitrogen by high-yielding varietiesproduces succulent plants that favour insects anddisease-causing pathogens. The increasing resist-ance to pesticides of such insects as the paddy bug(Leptocorrhizaariconis)has become a serious prob-lem (Abeysekera 1988). Insecticides also destroydelicate prey-predator relationships, leading togreater and secondary outbreaks of the target pest(Vanden Bosch 1978).In the 1970s, the brown planthopper (Nilaparvatalugens) became a major ricepest throughout Asia. It is endemic to SriLanka buthad never been a serious problem before the intro-duction of diazanon. This pesticide was commonlyused for combating the brown plant hopper despiteevidence available to IRRIscientists at Los Baniosthat rice fields treated with diazanon had higherhopper infestation than untreated ones (Andersonet al. 1991). The problem was that diazanon alsokilled the natural enemies of the plant hopper, themost important being spiders. When outbreaks ofplant disease occur, pesticide salesmen respond byrecommending the use of larger doses, puttingfarmers on a pesticide treadmill.13 A survey ofvegetable farmers in Sri Lanka revealed that themost important source of information regardingpesticides was private traders (Abeysekera 1988).Itmust be stressed that pesticide traders in rural SriLanka have received only a few years of formaleducation and they lack specialized technical infor-mation concerning the health hazards of the pesti-cides they sell. Thus improved rice in Sri Lanka isnot simply a technique of increasing food produc-tion but represents the emergence of a mode ofproduction that is destroying the productive baseof subsistence.We have seen how scarcityis socially constructedat the ecologicalnode of the nexus of the modern ricetechnology by not only degrading the productivebase of subsistence but also by marginalizing low-cost techniques which allow the poor to harness'value' created by nature. To return to my basicargument, alleviating poverty is not merely an econ-omic matter of increasing income but, among otherthings, one of engaged ecological relations of theagriculturaldiscourse/practices of Sri Lanka.

LakshmanYapaSocial relationsThe primary social relations of agriculture inSri Lanka involve questions of landownership.Contemporary land-use patterns of Sri Lankashowlittle change from the framework established incolonial times. According to the 1982 Sri Lankaagricultural census, agricultural land covers justover 31 per cent of the total land area, only slightlyhigher than the figure for 1962 (Survey Departmentof Sri Lanka 1988). Of the total area under agricul-tural holdings, about 40 per cent is devoted to tea,rubber and coconut, the major plantation crops.Paddy is grown on about 28 per cent of the areabutthe actual area cultivated in any one year is higherbecause of double-cropping. About 9 per cent ofthe land is under permanent crops, a categorywhich includes land under home gardens, andaround 10 per cent of the agricultural land is undertemporary crops other than paddy, a category thatincludes the cultivation of cereals, legumes, yams,potatoes and vegetables (Department of Censusand Statistics 1988).

Numerically, Sri Lanka is a nation of smallfarmers (Department of Census and Statistics1986).Over 73 per cent of the total agriculturalareais in small holdings and the rest are in largeholdings called estates. The total number of agri-cultural holdings in 1982 was 1 800 238, of which99-5per cent was under small holdings and 0.5 percent was in estates (TableII).Holdings under 1 acrein extent made up 5-8 per cent of the land and42 per cent of the total number of holdings; hold-ings under 2 acres comprised 14.8 per cent of theland and 63-5per cent of the holdings. Over 10 percent of the agricultural operators held no land at allwhile 38-4 per cent owned only home gardens.Thus nearly half of the agricultural operatorsbelonged to a class that either owned a single plot(under 1/8 acre) or no land at all (Table III).The first major land reform in independent SriLanka was the Paddy Lands Act of 1958, whichwas designed to give security of tenure, particu-larly to share-croppers of paddy. Sponsored by thethen marxist minister of agriculture in a coalitiongovernment with strong conservative land-owning interests, the Act received little supportduring implementation. The next round of landreforms took place in the 1970s, under the left-of-centre government of Sirima Bandaranaike, as aresponse to the insurrection of leftish youth in 1971(Samaraweera1982). The reforms took place in twostages: in the first stage, holdings of over 25 acres


14/22

107he poverty discourse and the poor in Sri LankaTable II Distribution of land holdings by size in Sri LankaSize class Extent Percentage Percentage(acres) (acres) area No. of holdings holdingsUnder 1/2 91 091 2.0 445 641 24.81/2-


15/22

108sphere, the continued belief in the backwardness oftraditional agriculture prevented society from fullyutilizing the food production potential of homegardens, highland plots, multiple cropping, indig-enous knowledge and regenerative agriculture. Inthe consumption sphere, a strong preference forrice, and now an increasing reliance on importedwheat flour, are accompanied by neglect of a rangeof readily cultivable traditional foods.The unacknowledged history of constructedscarcity paralleling the celebrated 'success story' ofthe modernization of paddy rice cultivation in SriLanka has a subplot - a sort of 'non-history' of thebenign neglect of traditional farming technology.As mentioned earlier, the traditional peasant farmin Sri Lanka consisted of three production units:the paddy field, the home garden of permanenttree crops and vegetables, and a highland plotwith food crops such as finger millet and manioc(Manihot esculentus). Several varieties of paddywere grown, usually twice a year (Yalaand Mahaseasons). After the paddy harvest, fields were leftfallow or cash crops such as betel leaves or vege-tables were grown. Animals (buffaloes) were usedin ploughing and threshing, and their dung wasreturned to the fields. The small plots of the homegardens were micro-simulations of the ecologicalprinciples of the tropical forest. A large numberof food, medicinal and ornamental plants weregrown. The gardens also yielded fodder for theanimals, fuel for cooking and materials for build-ing. Additional food was provided by backyardchickens, a cow and an occasional pig or goat. Theanimals gave valuable manure that was returned tothe soil (Everett 1987). The highland plots whichgrew temporary crops also had a high crop mix(though not as high as in the home garden). In theless densely populated part of the island, thehighland plot was part of a cycle of shiftingcultivation.As elsewhere in the third world, the officialagricultural discourse did not hold traditionalfarming technology in high regard, an attitudeconsistent with the representation of peasants inthe dominant academic model of the dualeconomy. The peasant economy in Sri Lanka suf-fered at the hands of academics in at least twoways. First, beginning with colonial science, theplantation crops received a great deal of attentionand resources;peasant crops received little patron-age. Contemporary perceptions of these two sys-tems - one as modern and the other as backward -

LakshmanYapaignore how resources and research were allocatedto them in the past. Secondly, the presumed back-wardness of peasant farming methods was notbased on evidence; it was an expression of a valuesystem that judged the worth of agriculture by thesize of the marketable surplus produced.The governing principle of traditional peasantfarming is the use of 'internal resources' providedby the regenerative power of nature and culture ata place (Harwood 1983; Rodale 1983; Shiva 1991).Multiple cropping allows an optimal use of space,time, sunlight, water, soil nutrients, family labourand management of risk. Francis (1986) hasdescribed the following advantages of interplant-ing: genetic diversity; diversity in insects associ-ated with complex prey-predator relationships;internal supply of plant nutrients;protection of theorganic content of the soil by continuous vege-tation cover which reduces soil erosion and con-serves rainwater whilst plant root systems tap intodifferent layers of the soil profile for nutrients andwater; good use of light and water throughout theyear; and low risk of complete crop loss in a givenyear.These characteristics of multiple-cropping areenhanced by the incorporation of farm animalssuch as chickens, ducks, cattle, goats and pigs.Agricultural waste is an important source of nutri-ents and fodder in multiple cropping which alsoreduces the incidence of pests and weeds (Altieriand Liebman 1986).Total yields per hectare of multiple croppingsystems are often higher than monocrop yields(Bray 1994). This attribute, called 'overyielding', ismeasured by the land equivalent ratio (LER)whichexpresses the monoculture land area required toproduce the same amount as 1 ha of multiplecrops using the same plant population. If the LERis greater than one, the polyculture overyields(Altieri 1987). Traditionalfarming systems are alsomore energy efficientthan fossil fuel based modernsystems, having energy output/input ratios in therange of 10 to 15, whereas modern systems typi-cally exhibit ratios of 1 to 3 (Pimentel and Pimentel1979). The energy efficiencies of a rice growingsystem in the village of Wangala, south India,before and after the green revolution were 7.14 and4-7, respectively (Bayliss-Smith 1984).The food balance sheets of Sri Lanka containstatistics on the amount of calories derived fromdifferent commodities consumed. Between 1961and 1991 the average calorie consumptionincreased at a modest rate of 0-2 per cent. In


16/22

The poverty discourse and the poor in Sri Lankageneral, over three-quarters of the total caloriescome from four commodities: rice, wheat flour,sugar and coconut (FAO 1993). Rice occupies adominant place in the Sri Lankan diet: a littleunder 40 per cent of all daily calories is obtainedfrom this source. It is not uncommon for those whocan afford it to eat three meals of rice every day.The heavy emphasis on, and investment in, grow-ing new varieties of rice were driven by a desire forthe country to become self-sufficient in rice. But thelack of self-sufficiency is not simply a matter ofinsufficient supply; it is also related to demand.The matter of self-sufficiency in rice can be posedas a question of why the demand is what it is.Brohier (1975, 163) has argued that

Surely hen,to putfirst hings irst, t does seem certainthat a campaign fostering the slogan Eat Less Rice,should find a place concurrentlywith a campaign oGrowMoreFood,and that as a preliminary step in thatdirectionsomethingshould be done to adjust thatwhich haspromotedmbalancen thepresent-day ietof the people.Only about 0-5 per cent of total calories comefrom cereals other than rice. The production ofmaize in 1990 was double that in 1970 but the total

quantity is small. The production of finger millet(kurakkan)n 1990 was less than half that of 1970.Anutritious crop such as kurakkan,which is drought-resistant and widely grown in chenaplots, is con-sidered by middle and upper class Sri Lankans tobe an inferior food. A little over 4 per cent of thecalories comes from the consumption of roots andtubers, primarily potatoes, manioc (cassava) andsweet potatoes (Table IV). As income grows, fam-ilies increase their consumption of potatoes whilethat of manioc and sweet potatoes is reduced(Central Bank of Sri Lanka 1985). This is consistentwith the cultural perception of manioc and sweetpotato as poor people's food. In the period 1961-91, calories derived from the consumption of rootsand yams declined at the rate of 0-5 per cent (FAO1993). Jak (Artocarpusheterophyllus) nd breadfruit(Artocarpusaltilis), important sources of carbo-hydrates, are commonly grown in home gardensbut the annual food balance sheets compiled by theDepartment of Census and Statistics do not evenmention these items. The leaves of the jak tree arealso excellent sources of fodder for animals. Per-ceived as lower class, the consumption of thesefoods declines with increasing income (CentralBank of Sri Lanka 1985). There is no organized

109Table IV Sources of calorie intake of the averagediet in Sri Lanka, 1987Commodity Caloriesper day Percentage

Rice 903 39-8Wheat flour 279 12.3Other cereals 13 0.5Roots/tubers 99 4.3Sugar 272 11-9Coconuts 315 13.9Pulses 53 2.3Fruits and vegetables 101 4.4Meat, fish and eggs 69 3.0Milk 62 2.7Oils and fats 91 4.0Total 2267 100.0Source:Department of Census and Statistics (1988)

61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93Year

Figure 9 Wheat imports by quantity and value,1961-91effort to encourage the production or consumptionof jak fruit, jak seed and breadfruit.Only a little over 2 per cent of the total foodcalories is derived from pulses (Table IV). This isquite unfortunate because pulses can be an import-ant source of inexpensive protein. Pulses are quitepopular among all classes of Sri Lankan society butthe total quantity consumed is small and the cal-ories derived from pulses has been declining at therate 1-1 per cent per year. On the other hand, over12 per cent of calories is obtained from importedwheat flour consumed in the form of bakery bread.The calorie intake from this source has beenincreasing at 2 per cent per year, the fastest rate ofgrowth for any item in the Sri Lankan diet (Figs 9and 10).According to FAO(1993) statistics between1961 and 1991, Sri Lanka spent over US$2-7billionon imported wheat. Unfortunately, the largest


17/22

110450

l 1400 -350-300 -

eE0s

A Index of Per CapitaKilocalorieConsumption----Total r---- From wheat r-/ \- ?....... From Rice . F.aL./ ~ ~ n~'250

200150 /100 n '].-.......... .l.--.*t . E .. ........"-- * A .R?..?..!. .

61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91Year

Figure 10 Index of kilocalories from wheat andrice in the average diet, 1961-91(1961 =100)consumers of imported wheat flour and breadwere among the poorest in Sri Lanka, many livingin rural areas (Central Bank of Sri Lanka 1985).In summary, the alleviation of poverty is not amatter of increasing income alone. The deprivationthat people experience in food has been culturallyconstructed by modern attitudes towards 'tra-ditional agriculture' and class-bias in the patternsof food preferences. Thus the cultural node in thenexus provides a site for the creative engagementof the social forces that create such scarcity.Political relationsBy 'political relations of production', I referprimar-ily to interactions between the state and society inthe organization of economic activity. In Sri Lanka,the postcolonial state has come to play an all-pervasive role in civil society through its commandover the development project, resource allocation,jobs and patronage. We have already seen themajor role of the state in the discourse/practice ofmodern agriculture. There are several ways inwhich the state and political relations of modernagriculture are implicated in the creation of scar-city. First, the research, technological and commer-cial relations of modern rice farming in Sri Lankathat we have examined are primarily a product ofstate policy. Secondly, the state has assumed therole of, and encouraged the public to perceive it as,the principal agent in the solution of the povertyproblem. Thirdly, by essentializing poverty as aneconomic problem that can be solved throughpolicy prescriptions, the state has helped to concealthe numerous sites at which scarcity is created andthus disabled the agency of members of civil

LakshmanYapasociety from exercising their substantive power atthese sites. If it is true that development is deeplyimplicated in the social construction of scarcity, itfollows that the state as an agent of developmentis, by virtue of that role, at the same time acausative agent in the creation of scarcity. So it isironic that the poor have come to look to the stateas a way out of their poverty.Drawing on Foucault's (1980, 1990) critique,conventional notions of power - which focus onthe state, repressive institutions, class and power-ful individuals - are not helpful to the resolution ofthe poverty problem. Agents of change act notthrough a general exercise of power but by exert-ing their will in substantive networks related toissues of food, nutrition, housing, education, trans-port, culture, geographical location and so on. Ihave identified several sites at which power can becreatively exercised in the discourse/practice offood and agriculture in Sri Lanka. There are manypoints of engagement and it is not helpful toreduce that plurality to a single abstract struggleagainst the state. Moreover,each substantive themein society is informed by powerful discourses suchas 'good nutrition' and 'scientific agriculture'.Power that will serve the resolution of the povertyproblem must have the ability to counter thepower of scarcity-constructing discourses that cir-culate throughout society; a sovereign notion ofpower could not serve that end.15

How academic relations construct scarcityThe task of understanding the links betweenacademic relations and scarcity is somewhat diffi-cult because the poverty discourse conceals its ownrole in the social construction of scarcity. I shallmake this argument by recognizing two kinds ofacademic relations: internal and external.Internal relations arise from the self-image sci-ence has of itself; an image that has been passed onto the discourse on poverty and development.Science views itself as value-free, neutral and non-political (Proctor 1991). Science has the power toname and define a problem. It gathers data, testshypotheses and helps to frame policy to solve theproblem. Scientists believe that science is locatedoutside the problem; it is an observer, a facilitatorand a problem-solver. The concept of the povertysector is a classic example of this subject/objectdualism of social science. Social science speaks of apoverty problem - a problem that exists among


18/22

Thepovertydiscourse and the poor in SriLankapoor people. Criteria for measuring poverty aregiven, data are collected, the nature of poverty isanalysed and, finally, development proposes policyto solve the problem (Central Bank of Sri Lanka1987;World Bank 1990).According to this logic, theproblem resides in the external material world andpoverty policy can frame the solution. From thatperspective, the notion that the subject is part ofthe problem becomes puzzling and inconsistentwith science's self-image.External academic relations refer to discoursesthat are produced at substantive sites - technology,political economy, culture, politics and nature.These give rise to what I have called the technical,social, cultural, political and ecological relations ofproduction. For example, plant breeders at IRRIin Sri Lanka chose to develop a technologicaltrajectory that included hybrid seeds, inorganicfertilizer, chemical pesticides and irrigation. Theychose not to focus on organic fertilizer, non-chemical methods of pest control, strengtheningrain-fed crops, multiple cropping systems andhome gardens. How did the plant breeders makethese decisions and why? Here is a part of theanswer: the plant breeder's science functions ina series of enabling texts that formulates socialtheory of the economy, culture and nature - theexternal academic relations. For example, plantbreeders responded to a given set task - that ofincreasing yields. The plant breeder has beenacademically socialized by the economist to believethat labour-intensive is backward and capital-intensive is modern; that progress is measured bythe size of market transactions; and large scale ismore efficient (Levins 1986). Enabling texts ineconomics also helped to exclude questions ofaccess to land or the development of technologythat considered land relations. Cultural texts pro-vided the view of peasants as traditional society;consequently, multiple cropping and folk knowl-edge were seen as backward. Theories of natureprovided additional authority; since the yield ofplants was the primary objective, all other aspectsof nature - water, soil, plants and insects - must bemanipulated, dominated and controlled to gain thedesired objective (Merchant 1983).There is a third aspect to academic relationsinvolving both internal and external relations thathave helped in the social construction of scarcity.By reifying poverty as an economic problem ame-nable to policy solution, and by equating powerwith conventional politics, social science robbed

111members of civil society of their power of agencyto overcome scarcity by acting at numeroussubstantive sites diffused throughout society.

ConclusionsFirst, a brief comment on looking for root causes.Poverty does have multiple causes but is it not thevery point of social science to separate the import-ant from the less important? The nexus of relationsoffers no means by which 'causes' can be arrangedin some logical order of importance. In my viewthe social science preoccupation with finding rootcauses is part of the problem. If some causes aremore important than others, we need to ask'important to whom?' For example, consider thecase where scholars privilege social relations ofproduction in their analysis. Accordingly, they mayargue that in a poor agricultural country like SriLanka, land reform is more important than otherfactors. If that is so, then the state, as the solesource of power to carry out land reform, auto-matically becomes the most important agent in thepoverty problem.Next, consider a nutritionist who wishes tomake a contribution to alleviating hunger in SriLanka by promoting the consumption of inexpen-sive, indigenous foods. By what logic do we judgeland reform to be more important than changes inhabits of food consumption? Is there a logic of'causation' that is more important than the logic of'agency'? I think not. Assume for a moment thatthe nutritionist in our example has no power ofagency over land reformbut does have knowledge,interest and power in the area of food consump-tion. The nutritionist does not need to wait untilland reform is completed to carry out his or herwork. This is because there is no logical basis to saythat land reform is more important than changes inhabits of food consumption. These two examplescome from incommensurable processes. There isno standard or metanarrative logic of poverty thatcan help us arbitrate on the relative importance ofthis or that cause, or help us find root causes.The general discourse on poverty, particularlythose of the Central Bank in Sri Lanka and theWorld Bank in Washington, posit the existence of apoverty sector in the economy. The economisticapproach defines this sector using an incomecriterion and proposes economic growth, jobsand higher income as a solution to the problem.


19/22

112Following that approach, the Central Bank in SriLanka defined absolute poverty as not havingenough income to buy the minimum required inintake of food calories. However, instead of seeingthe problem as one of inadequate income, I asked asubstantive question in this paper: why are thepoor in Sri Lanka unable to obtain the minimumlevel of required food? What we found was that thescarcity of food was socially constructed at eachnode of a nexus of production relations diffusedthrough the larger society.Such a view has several advantages: it helps ussee the connection between development andsocially constructed scarcity; it shows how dis-course has served to conceal the history of thisscarcity; going beyond income, it reveals myriadways - technical, social, cultural, ecological, politi-cal and academic - of addressing issues of scarcity;identifying numerous sites of action takes usbeyond poverty experts to marshall different kindsof agents; and, by moving away from a single greatlocus of action, we strengthen civil society atnumerous substantive sites and de-emphasize therole of sovereign power. These arguments are notjust unique to food; a resolution of the povertyquestion in Sri Lanka requires sustained criticalengagement 'on the ground' in other areas such asenergy, housing, transport and health care, but thattakes us back to discourse. The task of buildingsocial theory to aid the poor must necessarily beginwith a critical engagement of the existing discoursethat is part of the problem.Notes

1 See Anand and Kanbur (1995); Gunatilleke et al.(1992);Isenman (1980);Sahn (1987);and Sen (1981).2 See Anand and Kanbur (1995); Isenman (1986);Pyatt (1986);and Sen (1995).3 Even today over 80 per cent of the land in Sri Lankais under some form of state control (Land Commis-sion 1990).4 Though both governors Ward and Gregory paidattention to the restoration of ancient irrigationworks to serve the paddy economy, colonial policyas a whole had little interest in continuing suchefforts (Bandarage 1983;De Silva 1981).5 A discussion of the reasons for these policies wouldtake us beyond the scope of this paper. A goodexplanation appears in Moore (1989).6 While these policies served the ideological interestsof upper class Sinhalese and the material interestsof some Sinhalese peasant families, they did muchto alienate large segments of the Tamil population.

Lakshman Yapa7 Between 1961 and 1991 the value of agriculturalinputs (fertilizer, pesticides and machinery) grewat a rate of 6 per cent; total agricultural outputgrew at a modest 1-5 per cent and agriculturaloutput per capita declined at a rate of -0-3 per cent.The corresponding statistics for total food outputwas 2.1 per cent and food output per capita was0-4 per cent (FAO 1993).8 A related story is the cost of the state-owned ferti-lizer factory at Sapugaskanda about 5 miles fromColombo designed to use naphtha (a byproduct ofthe Petroleum Corporation's oil refinery) as feed-stock for the manufacture of urea. The plant wascommissioned in the late 1980s at an estimatedinitial capital cost of Rs2800 million. Plagued withcost overruns and mechanical breakdowns, the

plant was commonly referred to as 'Sri Lanka'sbiggest white elephant'. With the high price ofnaphtha and falling prices of urea on the inter-national market, the plant was closed in 1985after recurrent annual losses and finally sold forscrap.9 See FAO (1977); National Research Council (1989);Palm and Sandell (1989); Ulluwishewa (1991); andWolf (1977).10 The National Engineering Research and Develop-ment Centre in Sri Lanka is a small underfundedresearch institute located a few miles north ofColombo. Among other projects, the engineers atthe centre have developed a biogas technologycalled the DryBatch system which overcomes someof the disadvantages of the Chinese and Indianversions of biogas systems. According to their lit-erature, a single acre of paddy produces enoughstraw which, when digested by the DryBatch sys-tem, is sufficient to fertilize an acre of paddy. Inaddition, the straw will produce biogas that canbe used as a fuel for heating and lighting. Lack ofinterest in such alternatives is so pervasive that Iwas unable to obtain any information about thesystem from the National Fertilizer Secretariat inColombo, a state agency which was establishedin 1979 with FAO and German aid to coordinatefertilizer policy in Sri Lanka. A perusal of theannual reports and databases maintained by theSecretariat makes it clear that their primary mis-sion is the promotion of chemical fertilizer in SriLanka.11 See Bull (1982); Dover (1985); Miller (1992);National Research Council (1989); Ulluwishewa(1992);Van den Bosch (1978);and Wolf (1977).12 See Abeysekera (1988); Cochrane (1993); Merrill(1976);and Shiva (1991).13 Despite a ten-fold increase in the use of insecticidesin the US, since the 1940s crop losses to insects havenearly doubled from 7-1 to 13 per cent. The esti-mated environmental and health costs range from


20/22

The poverty discourse and the poor in Sri LankaUS$4 to 10 billion a year (Miller 1992;Pimentel andLevitan 1986). I have not found similar calculationsmade for Sri Lanka.14 For a similar history of green revolution researchin Mexico, see Hewitt de Alcantara (1976) andJennings (1988). For a discussion of the institutionalhistory of IRRI in the context of the RockefellerFoundation and the Ford Foundation, see Andersonet al. (1991).15 A much longer version of this argument appears inYapa (1996a).

ReferencesAbeysekera W A T 1988 Pesticide use in the foodproduction sector of Sri Lanka in Use of pesticidesandhealth hazards in the plantation sector Friedrich-Ebert-Stiftung, Colombo, Sri Lanka 12-37Altieri M A ed. 1987 Agroecology: he scientific basis ofalternativeagricultureWestview Press, Boulder, COAltieri M A and Liebman M 1986 Insect, weed and plantdisease management in multiple cropping systems inFrancis C A ed. Multiple croppingsystems Macmillan,New York 183-218Anand S and Kanbur S M R 1995 Public policy and basicneeds provision: intervention and achievements in SriLankain Dreze J Sen A and Hussain A eds Thepoliticaleconomyof hungerClarendon Press, Oxford 298-331Anderson S Levy D and Morrison B M 1991 Rice scienceand developmentpoliticsOxford University Press, NewYorkBandarage A 1983 Colonialism n Sri Lanka: he politicaleconomy of the KandyanHighlands, 1833-1886 MoutonPublishers, New YorkBandaranaike A D and de Alwis S 1987 Health andnutrition in Thealleviationof poverty n Sri LankaCentralBank of Sri Lanka, Colombo, Sri Lanka 165-85Bayliss-Smith T P 1984 Energy flows and agrarianchanges in Karnataka: the green revolution at micro-scale in Bayliss-Smith T P and Wanmali S edsUnderstanding green revolutions: agrarian change anddevelopmentplanningin SouthAsia Cambridge Univer-

sity Press, Cambridge 153-72Bhalla S S and Glewwe P 1986 Growth and equity indeveloping countries: a reinterpretation of the SriLankan experience The WorldBank EconomicReview 135-63Bray F 1994 Agriculture for developing nations ScientificAmerican270 30-7Brockway L H 1979 Science and colonial expansionAcademic Press, New YorkBrohier R L 1975 Food and the people Lake House,Colombo, Sri LankaBull D 1982 A growingproblem: esticides nd the thirdworldpoorOxfam, Oxford

113Central Bank of Sri Lanka 1985 Report on consumer

finances and socio economic survey 1981/82 Sri LankaCentral Bank of Sri Lanka, Colombo, Sri LankaCentral Bank of Sri Lanka 1987 Thealleviationof povertyin Sri LankaCentral Bank of Sri Lanka, Colombo, SriLankaChenery H B 1989 Growth-equity tradeoffs in Meier G Med. Leading issues in economic development OxfordUniversity Press, Oxford 21-5Cochrane W W 1993 Thedevelopment f Americanagricul-ture:a historicalanalysis University of Minnesota Press,Minneapolis, MNDepartment of Census and Statistics 1986 Sri Lankacensus of agriculture1982 Department of Census andStatistics, Colombo, Sri LankaDepartment of Census and Statistics 1988 Agriculturalstatistics of Sri Lanka1988 Department of Census andStatistics, Colombo, Sri LankaDepartment of Census and Statistics 1993 Householdandincome expenditure survey 1990/91 Department ofCensus and Statistics, Colombo, Sri LankaDe Silva K M 1981 A history of Sri LankaUniversity ofCalifornia Press, Berkeley,CADover M J 1985 A better mousetrap: improving pestmanagement or agricultureWorld Resources Institute,Washington, DCEverett Y F 1987 Seeking principles of sustainability:a forest model applied to forest gardens in SriLanka Unpubl. MA thesis, Wildlife Resource Science,University of California, Berkeley,CAFADINAP (Fertilizer Advisory, Development and Infor-

mation Network for Asia and the Pacific) 1987 Supply,marketing,distributionand use of fertilizer in Sri LankaFADINAP,Bangkok, ThailandFAO (Food and Agriculture Organization) 1977 Organicmaterials and soil productivity Food and AgricultureOrganization of the United Nations, RomeFAO (Food and Agriculture Organization) 1993Agrostat-PCFood and Agriculture Organization of theUnited Nations, RomeFoucault M 1980 POWER/KNOWLEDGE:elected inter-views and otherwritings Pantheon, New YorkFoucault M 1990 The history of sexuality Vintage Books,New YorkFrancis C A 1986 Perspectives for the future of multiplecropping in Francis C A ed. Multiple croppingsystemsMacmillan, New York 351-70Goonatilake S 1984 Crippledminds: an explorationintocolonialculture Lake House, Colombo, Sri LankaGramsci A 1972 Prison notebooksnternational Publishers,New YorkGunaratne L 1987 The poorest of the poor in Sri LankainThealleviationof poverty n Sri LankaCentral Bank of SriLanka, Colombo, Sri Lanka 243-75Gunatilleke G Perera M Wanigaratne R A M CFernando R E Lakshman W D Chandrasiri J K M D


21/22

114and Wanigaratne R D 1992 Ruralpoverty in Sri Lanka:priority issues and policy measures Asian DevelopmentReview10 164-98Harwood R R 1983 International overview of regener-ative agriculture in Resource-efficientarmingmethodsforTanzaniaRodale Press, Emmaus, PA 24-35Hewitt de Alcantara C 1976 ModernizingMexicanagricul-ture: socioeconomicimplicationsof technologicalchange1940-1970 United Nations Research Institute for SocialDevelopment, GenevaIsenman P 1980 Basic needs: the case of Sri Lanka WorldDevelopment8 237-58Isenman P 1986 Comment on 'Growth and equity indeveloping countries: a reinterpretation of the SriLankan experience by Bhalla and Glewwe' The WorldBankEconomicReview 1 521-31

Jennings B H 1988 Foundationsof internationalagriculturalresearchWestview Press, Boulder, COKloppenburg J R 1988Firstthe seed: hepoliticaleconomyofplant biotechnology1492-2000 Cambridge UniversityPress, New YorkLand Commission 1990 Reportof the Land Commission1987 Government Publications Bureau, Colombo, SriLankaLevins R 1986 Science and progress: seven developmen-talist myths in agriculture MonthlyReview 38 13-20Lichtman R J 1983 Biogas systems in IndiaVolunteers inTechnicalAssistance, Arlington, VAMcGarry M G and Stainforth J eds 1979 Compost, erti-lizer,andbiogasproductionromhumanandfarmwastes inthePeople'sRepublic fChinaInternationalDevelopment

Research Centre, Ottawa, CanadaMerchant C 1983 Thedeathof nature:women,ecologyandthe scientificrevolutionHarper & Row, San Francisco,CAMerrill R ed. 1976 RadicalagricultureHarper & Row, NewYorkMiller G T 1992 Living in the environmentWadsworth,Belmont, CAMooney P R 1979 Seedsof the earthInstitute for Food andDevelopment Policy, San Francisco, CAMoore M 1989 The ideological history of the Sri Lankan'peasantry' ModernAsian Studies 23 179-207NARESA (National Resources, Energy and ScienceAuthority) 1991 Natural resourcesof Sri LankaNARESAand USAID, Colombo, Sri LankaNational Fertilizer Secretariat 1992 The reviewof fertilizer- 1992 National Fertilizer Secretariat, Colombo, SriLankaNational Research Council 1989 Alternative agricultureNational Academy Press, Washington, DCNithiyanandan V 1987 An analysis of economic factorsbehind the origin and development of Tamil National-ism in Sri Lanka in Abeysekera C and GunasingeN eds Facets of ethnicity in Sri LankaSocial ScienceAssociation, Colombo, Sri Lanka

Lakshman YapaPain A 1981 Agricultural research: 1906-1981 TropicalAgriculturalist137 3-12Pain A 1986Agricultural researchin Sri Lanka: an histori-cal account ModernAsian Studies 20 755-78Palm O A and Sandell K 1989 Sustainable agricul-ture and nitrogen supply in Sri Lanka: farmers' andscientists' perspective Ambio18 442-8Pimentel D and Levitan L 1986 Pesticides: amounts

applied and amounts reaching pests BioScience 3686-91Pimentel D and Pimentel M 1979 Food,energyand societyJohn Wiley, New YorkPiyadasa L 1988 Sri Lanka: he unfinished quest for peaceMarram Books, LondonProctor R N 1991 Value-free cience? Purity and powerin modern knowledge Harvard University Press,Cambridge, MAPyatt G 1986 Comment on 'Growth and equity in devel-

oping countries: a reinterpretation of the Sri Lankanexperience by Bhalla and Glewwe' The WorldBankEconomicReview1 515-20Rajapathirana S 1988 Foreign trade and economic devel-opment: Sri Lanka's experience WorldDevelopment161143-57Ratnapala N 1985 The beggar in Sri Lanka D NPalihawadana, Moratuwa, Sri LankaRatnapala N 1989 Rural poverty in Sri LankaDeepanee,Nugegoda, Sri LankaRodale R 1983 Importance of resource regeneration inResource-efficientarming methodsfor TanzaniaRodalePress, Emmaus, PA 20-3Sahn D E 1987 Changes in the living standards of thepoor in Sri Lanka during a period of macroeconomicrestructuring WorldDevelopment15 809-30Samaraweera V 1982 Land reform in Sri Lanka ThirdWorldLegalStudies1982 104-22Sen A 1981 Public action and the quality of life indeveloping countries OxfordBulletinof EconomicsandStatistics 43 287-319Sen A 1995 Food, economics, and entitlements in Dreze JSen A and Hussain A eds The political economy ofhungerClarendon Press, Oxford 50-68Shastri A 1995 Transitions o a free market: conomic iber-alization in Sri LankaPaper read at the fifth Sri LankaConference held at the University of New Hampshire,Durham, NHShiva V 1991 The violence of the green revolution: thirdworld agriculture, ecology and politics Zed Books,LondonSiriwardena S 1994 An analysis ofurbanpovertyalleviationthrough community-based programmes in Sri Lankamimeo, Colombo, Sri LankaStokke S 1992 Dynamic growth or pauperization? Ruralsmall-scale industries in Sri LankaUnpubl. PhD thesis,Department of Geography, Pennsylvania State Univer-sity, University Park, PA


22/22

The poverty discourse and the poor in Sri LankaStokke K 1995 Poverty as politics: the Janasaviya PovertyAlleviation Programme in Sri Lanka Norsk GeografiskTidsskrift 9 123-35Survey Department of Sri Lanka 1988 ThenationalatlasofSri LankaSurvey Department of Sri Lanka, Colombo,

Sri LankaSwan B 1987 SriLankamosaic: nvironment,man,continuityand changeMarga Institute, Colombo, Sri LankaTambiah S J 1986 Sri Lanka:ethnic fratricide and thedismantling of democracy Oxford University Press,New Delhi, IndiaUlluwishewa R 1991 Soil fertility management ofpaddy fields by traditional farmers in the Dry Zone ofSri LankaJournalof SustainableAgriculture1 95-106Ulluwishewa R 1992 Indigenous knowledge systems forsustainable development: the case of pest control bytraditional paddy farmers in Sri Lanka Journal ofSustainableAgriculture3 51-62Van den Bosch R 1978 ThepesticideconspiracyDoubleday,Garden City, NYWarnapala W A W 1993 The Sri Lankanpolitical sceneNavrang, New Delhi, India

115Wijesinghe L C A D S and Chandrasiri J A 1986Operating experience with biogas plants in Sri LankaNaturalResourcesForum10 221-9Wolf R ed. 1977 Organic arming:yesterday's ndtomorrow'sagricultureRodale Press, Emmaus, PAWorld Bank 1990 Worlddevelopment eport1990 OxfordUniversity Press, New YorkWorld Bank 1992 Worlddevelopment eport1992 OxfordUniversity Press, New YorkWorld Bank 1993 Poverty reductionhandbookThe WorldBank, Washington, DCYapa L 1993 What are improved seeds? An epistemologyof the green revolution EconomicGeography 9 254-73Yapa L 1996a What causes poverty? A postmodern viewAnnals of the Association of American Geographers86707-28Yapa L 1996b Innovation diffusion and paradigms ofdevelopment in Earle C Matthewson K and Kenzer

M S eds Concepts in human geography Rowman &Littlefield, Lanham, MD 231-70Yapa L and Wisner B 1995 Building a case againsteconomic development GeoJournal5 105-18

The Poverty Discourse and the Poor in Sri Lanka

Documents