The Development of the Rural Non-Farm Economy in ...projects.nri.org/rnfe/pub/papers/keyissues.pdf · Developing Countries and Transition Economies: Key Emerging and Conceptual Issues

The Development of the

Rural Non-Farm Economy

in Developing Countries

and Transition Economies:

Key Emerging and Conceptual Issues

Junior R. Davis and Dirk J. Bezemer

Enterprise development, trade, finance, and empowerment are central to the improvement of

people’s livelihoods in developing and transition countries. The Enterprise, Trade and Finance

Group together with the Livelihoods and Institutions Group at the Natural Resources Institute

apply practical solutions to rural and economic development through research, consultancy and

training activities. Our teams have a substantial experience in agricultural and development

economics, marketing and market research, commodity and international trade, financial

service development, and social and institutional development. Core areas of expertise include:

• Microfinance, Enterprise Development and Poverty Reduction

• Ethical Trade and Corporate Social Responsibility

• Improving the Performance of Agricultural Markets

• International Marketing and Trade

• Community Participation in Policy Formulation

• Social Impact Assessment

• Institutional Analysis and Capacity Building

The Institute is also able to arrange the following services: programmed study tours; training

courses; seminars; communication for development; and CD Rom design and production.

For further information please contact:

Enterprise, Trade and Finance Group

Natural Resources Institute

Chatham Maritime

Kent, United Kingdom

ME4 4TB

Email: [email protected]

Internet: http://www.nri.org/rnfe/

Tel: + 44 1634 883199

Fax: + 44 1634 883706

ISBN: 0 85954 557 1

The Development of the RuralNon-Farm Economy in

Developing Countries andTransition Economies:

Key Emerging andConceptual Issues

Junior R. Davis1 and Dirk Bezemer2

1Natural Resources Institute

2Imperial College at Wye

ii

© The University of Greenwich 2004

The Natural Resources Institute (NRI) of the University of Greenwich is an internationally recognizedcentre of expertise in research and consultancy in the environment and natural resources sector. TheInstitute carries out research and development and training to promote efficient management and useof renewable natural resources in support of sustainable livelihoods.

Short extracts of this publication may be reproduced in any non-advertising, non-profit-making contextprovided that the source is acknowledged as follows:

DAVIS, J. R. and BEZEMER, D. (2004) The Development of the Rural Non-Farm Economy inDeveloping Countries and Transition Economies: Key Emerging and Conceptual Issues. Chatham,UK: Natural Resources Institute.

Permission for commercial reproduction should be sought from the Publications Manager, Universityof Greenwich at Medway, Central Avenue, Chatham Maritime, Kent ME4 4TB, United Kingdom.

This publication is an output from a research project funded by the United Kingdom Department forInternational Development (DFID) under the DFID/World Bank Collaborative Programme for RuralDevelopment. The views expressed are solely those of the authors and not necessarily those of DFIDor the World Bank.

Natural Resources InstituteISBN: 0 85954 557 1

University of Greenwich, a registered charity and company limited by guarantee, registered in England (Reg. No. 986729). Registered Office:Old Royal Naval College, Park Row, Greenwich, London SE10 9LS.

Contents

1 Relevance of Rural Diversity and the Rural Non-Farm Economy 1

2 Definitions and Measurement Problems 3The Rural Non-Farm Economy 3Incomes and Livelihoods 3Activities, Assets and Diversification 5Concepts, Assets, Activities and Livelihoods 7Diversification Typologies 9

Distress-push and demand-pull diversification 9Other diversification typologies 10

3 Economic Diversity and Growth 15Factors Enabling Household and Enterprise Diversification 15Motivation for Diversification 17Factors Enabling Growth of the RNFE 17Socio-cultural Aspects of RNFE Growth 19Linkages Between the Farm and Non-farm Economy 20The RNFE, Poverty and Inequality 21

4 The Rural Non-Farm Economy Policy Processes 25

5 Conclusions 29

References 31

iii

For most rural people in developing andtransitional economies, rural non-farm activitiesare part of a set of livelihood activities thatincludes farming, that is, they are part of adiversified livelihood portfolio. The ruralpopulation in developing countries derives animportant share of its income from rural non-farm activities. Ellis (2000a) states that 30–50%is common in sub-Saharan Africa, and Reardonet al. (1998)1 gives a mean figure of 42% for sub-Saharan Africa. In Asia, and Latin America, FAOestimates the figures to be 32% and 40%,respectively. Ellis (2000a) gives appreciablyhigher estimates for South Asia.

There has been an increasing recognitionrecently that the rural economy is not confined tothe agricultural sector, but embraces the broadspectrum of needs of all rural people includingsocial service provision, economic activities,infrastructure and natural resources (Csaki andLerman, 2000). Since the 1970s, a large numberof studies have investigated the role of non-agricultural economic activities for ruraldevelopment. Evidence from the developingworld suggests that economic diversity in thecountryside has the potential to foster localeconomic growth and alleviate the rural-urbanincome gap and rural poverty.

These findings are also relevant to the post-socialist transition countries, where typically alarge part of the population lives in rural areas,and economic growth and the reduction ofpoverty are significant challenges. This isparticularly true for those transition countries

outside Central Europe. Analysis of the transitionprocess in general and of transition in theagricultural sector has generated a large volumeof literature, but less has been specificallydevoted to the wider rural non-farm economy(RNFE). However, studies in this field are nowbeing undertaken, since it is recognized that inthe longer term the development of the rural non-farm sector is a critical factor in providing ruralemployment and income (see Bright et al., 2000;Davis and Pearce, 2001).

The RNFE is of interest to governments, bilateraland multilateral donor agencies, NGOs anddevelopment practitioners because of itsincreasing prevalence in both developing andtransition economies. In many parts of the world,the number of poor people in rural areas exceedsthe capacity of agriculture to provide sustainablelivelihood opportunities. Even with a decline infertility rates and a slowing of population growth,this situation will not change significantly. Out-migration is not possible for all types of people,and urban centres cannot (or should not, foreconomic and social reasons) be assumedcapable of providing adequate livelihoodopportunities for all those unable to make aliving in agriculture. For these reasons, a healthyRNFE holds out the prospect of improvedlivelihoods for rural people. This set ofcircumstances puts the spotlight on the RNFE asa potential vehicle for poverty reduction in ruralareas. The RNFE may:

• absorb rural surplus labour

• help farm-based households spread risks

Relevance of Rural Diversityand the Rural Non-Farm

Economy11Chapter

1 The FAO study summarizes data from over 100 studies undertaken over three decades (1970s to the 1990s), focusing mainly on farmhouseholds (as opposed to rural town residents).

• offer more remunerative activities tosupplement or replace agricultural income

• provide a means for the rural poor to cope orsurvive when farming fails

• exploit rural comparative advantages(resources, location, labour costs)

• foster rural growth

• improve the overall quality of life, goodsand services in rural areas.

From a policy perspective, it is important tounderstand participation in the RNFE,particularly participation of the poor. Why doindividuals enter the rural non-farm economy?What RNFE opportunities are accessible tothem? Does participation in the RNFE contributeto an individual’s ‘upward’ or ‘downward’livelihood trajectory (a concept elaborated bySwift, 1998)2. Further, just as it is important tounderstand entry into the RNFE, it is equallyimportant to understand exit – whether peopleremain in the RNFE or leave, either throughchoice or circumstance.

Answers to these questions can inform thedevelopment of policies that:

• support the efforts of the rural poor

• protect them from deleterious livelihoodtrajectories

• improve access to sustainable andremunerative non-farm livelihoods.

Participation in the RNFE requires bothmotivation to enter the RNFE, and ability toaccess sustainable and remunerative livelihoodsfrom it. These two aspects – motivation andability – are important because the reasons whypeople enter the RNFE may have an implicationfor the types of access barriers faced, forexample, a person who is forced to diversify into

non-farm activities because of lack of access tocredit to purchase seeds will, in all probability,also face problems in accessing credit to start upa new business. This might be termed distress-push diversification3. Conversely, those enteringthe RNFE for demand-pull, that is, in response toan observed market gap or entrepreneurialreasons, are more likely to have access to higherentry barrier activities that allow accumulation(we discuss these two concepts in greater detailbelow). Setting aside the problem of how todefine ‘low’ and ‘high’ asset endowments, wemay for convenience simplify this to representtwo extreme types. In one, the social unit(characteristically an individual or household)has low endowments of all types of capital asset,human, natural, physical, financial, social, whilstin the other, the social unit has high endowments.In practice of course, particular social units willhave particular endowments of different types ofasset, thus there would be a multi-strandedspectrum of asset endowment, each strandcorresponding to a type of asset. Particularconfigurations of contextual factors, such as thepolicy environment, institutions and thevulnerability context, combined with particularconfigurations of asset endowments will result indiffering RNFE entry motivations, accesscapabilities and livelihood trajectories. Theselivelihood strategies may be better understood byseeing them as being on a spectrum between‘demand-pull’ and ‘distress-push’ diversity.

This paper outlines key emerging and conceptualissues in the development of the RNFE in lessdeveloped countries and transition economies. Itis based on a conceptual framework where theRNFE is discussed as being either part of agrowth strategy for the economy (demand-pull),or as a ‘defensive’ survival strategy for the ruralpoor (distress-push).

2

Relevance of Rural Diversity and the Rural Non-Farm Economy

2 When thinking about the rural non-farm economy, it is important to realize that different social units are constantly engaged in a dynamicprocess of livelihood adaptation. Taking the unit of the household: households operating within a particular livelihood system may be on anyof a number of different livelihood trajectories. These may be ‘downward’, in the sense that there is a process of disaccumulation of assets;‘upward’ in which case there will be asset accumulation; or more or less constant in the sense that the household asset base is neitherexpanding nor contracting. In each of these different scenarios, the role and importance of off-farm strategies takes on a different meaning.3 Reardon et al. (1998) suggest that when relative returns are higher to the RNFE than to farming, and returns to farming are relatively morerisky, ‘pull’ factors are at work. Conversely, when farm output is inadequate and opportunities for consumption smoothing, such as credit andcrop insurance, are missing, or when input markets are absent or fail and the household needs cash to pay for farm inputs, ‘push’ factors areat work. As evidence of distress-push, wages or incomes are likely to be lower in the RNFE.

The Rural Non-FarmEconomy

In this study the rural non-farm economy(RNFE) is defined as being all those incomegenerating activities (including income in-kind)that are not agricultural but are located in ruralareas (Lanjouw and Lanjouw, 1997). A key termin this definition of the RNFE is ‘rural’. TheOxford English Dictionary (1996) classifiespredominantly rural areas as those where morethan 50% of the population live in ruralcommunities, and significantly rural areas asthose where between 15% and 50% live in ruralcommunities; but different studies includedifferent definitions of ‘rural’. ‘Agriculture’ ishere taken to mean all primary production offood, flowers and fibres, thus it includes growingcrops, rearing livestock, horticulture (flowers,fruit and vegetables), forestry and fisheries. Itexcludes any food processing (although this maytake place on-farm), agricultural services(whether technical or commercial) and otherprimary sectors, such as mining or quarrying.

This definition is not solely activity based(waged work or self-employment), as it alsoincludes non-earned income (e.g. remittances) aswell as the rural institutional framework (roads,schools, hospitals, etc.), which are an integralpart of the rural economy. This also includessocial payments (pensions, social insurance,etc.), which are often a significant source ofunearned household income, but for which noactivity is undertaken by household members (incontrast with remittances or asset income).

However, social payments have a clear impact onthe RNFE as they reduce poverty levels,influence household work-leisure decisions, andmay create opportunities for investment.

Thus, the RNFE might include agro-processing,the setting up of a small business, or the receiptof transfer payments such as interest, dividendsor remittances from temporary, seasonal orpermanent migration. The RNFE incorporatesjobs which range from those requiring significantaccess to assets, such as education or credit, toself-employed activities such as the roadside‘hawking’ of commodities which have lowbarriers to entry and low asset requirements(Davis and Pearce, 2001). As regards the conceptitself, it could be argued that the term ‘RNFE’,although in common usage is technicallyincorrect, as non-agricultural activities mayactually take place on farms. Thus, although therural non-agricultural economy would be a moreaccurate definition, the terminology in this paperconforms to usage in the literature, where thefocus is often on ‘farm’ versus ‘non-farm’ or ‘on-farm’ versus ‘off-farm’ activities.

Incomes and Livelihoods

Davis and Pearce (2001), in a review of the levelof RNFE diversification, assert that it isimportant to consider the potential sources ofincome available to each farm or rural household.These are shown in Figure 1 for a farm-basedhousehold. The traditional main component hereis income from agricultural core activities.

Definitions and MeasurementProblems

2Chapter

4

Figure 1: Potential sources of income Source: Davis and Pearce (2001)

In defining RNFE income, diversification andother economic activities of farmers and ruraldwellers, two central problems emerge: (i)capturing the appropriate unit for incomeanalysis; and (ii) recognizing and ordering themultiple nature of income sources.

The question of what the appropriate unit mightbe for income analysis is important, and partlyrelates to social and cultural factors. The mostobvious units would be either the ‘individual’ orthe ‘household’. The definition of the former isnot in doubt. The latter – the household – is notso straightforward, since the co-resident unit isnot always the only economically relevant unit interms of production or consumption. There maybe smaller units, which are relevant in anextended or joint family situation, or there maybe larger units where closely related householdscollaborate and co-operate in activities that areeconomically significant. It might be argued thatthe most appropriate unit should be identified inthe specific cultural and social context, however,this presents additional empirical problemsbecause it is difficult to draw out parallels anddifferences between different countries, and even

between different regions of the same country.One undisputed characteristic of households isthat its members share income to some extent.

A disadvantage with this focus on householdlivelihoods and incomes is that it provides noinsight into enterprise behaviour. Its mainadvantage is that it is better suited formeasurement purposes: all rural income fromwhatever source sooner or later ends up inhousehold wallets. However, micro- and small-scale rural enterprises have often beeninvestigated as potential motors of localeconomic growth. The dynamics, drivers andbarriers for these businesses are not fullycaptured by the livelihoods approach. They areusually captured as family businesses, but notonce they go beyond the size of a micro-enterprise.

By defining rural economic diversification as allrural income generation other than foodproduction, a great heterogeneity in the activitiesundertaken by, or sources of income of, ruralhouseholds and enterprises is implied (Start,2001: 496). This ‘bewildering diversity’

Definitions and Measurement Problems

FarmHouseholdIncome

On Farm Income

Off Farm Income

Unearned Income

Agricultural CoreActivities

DiversifiedEnterprises

Non FarmEnterprises

Non-AgriculturalEmployment

Non Home FarmAgriculturalEmployment

Pensions,Dividends,Interest etc.

5

(Haggblade et al., 2002) presents problems ofconcepts and definitions relating to both the unitof measurement and the definition of incomesand activities (Barrett et al., 2001b; Reardon etal., 1998). In response, many dichotomies orcategorizations have been used in empiricalresearch to address the above problems ofdefining and measuring the RNFE, such as off/on-farm, business/wage income, local/urbanactivities, earned/non-earned income, tradable/non-tradable, activity-based/income-based, etc.

Davis and Pearce (2001) suggest that oneapproach is to study the components of potentialsources of income (see Figure 1). In the case offarms, on-farm income can come both fromagricultural core activities and non-agriculturalactivities. Potential sources of non-agriculturalincome can be divided into three components:income from non-agricultural employment; non-farm enterprises; and unearned income. As such,one can distinguish between enterprise andincome diversification. Enterprise diversificationactivity embraces both on- and off-farm businesscreation outside agricultural core activities.Income diversification will embrace these twocomponents plus any movement towards non-farm employment (whether agriculturally basedor not). Finally, unearned income, such asremittances, pensions, dividends and interest,which while usually ignored, can be substantial,and decisions made in this sphere may have animportant bearing on such crucial choices as timeof retirement and intensity of farming.

Thus, potential sources of income are disparate,likely to vary substantially in importancebetween rural households, and exhibit widevariations in their attractiveness as sources offinancial gain. These variations betweencomponents of income are, therefore, likely tohave a major effect on the decision-making ofrural households and individuals and there is a

need to understand the importance of each, ratherthan subsuming them all into binaryclassifications such as the ‘part/full-time’dichotomy4. Moreover, there is no reason whyRNFE income diversification has to be eitherabout setting-up new enterprises or even be farmbased at all. For many, other intermediate optionsmay prove more fruitful or promising (Pearceand Davis, 2000). This heterogeneity is one ofthe reasons why the concept of livelihoods,introduced earlier, is more appropriate as itencompasses all income sources.

Activities, Assets andDiversification

Rural non-farm activities may be defined in anumber of different ways. One simple distinctionis between waged and self-employment. This is afunctional distinction. In addition, activities maybe classified according to sector (e.g. primarysector vs. secondary sector) and/or space. SeeBarrett and Reardon (2000) who explored thesedistinctions in great detail. RNFE activities mayfall anywhere within the shaded part of Table 1below.

The situation is, however, even more complexthan that depicted in Table 1, because in additionto the three way classification, it is alsoimportant to make a distinction between the assetimplications of particular activities. That is, it isimportant to distinguish between activities thataccumulate, spread or denude assets. Activitiesand assets are difficult to measure for some of thesame reasons that incomes are, that is,complexity, oversight and respondent/subjectrecall problems, dubious legality of certainactivities and certain (financial) assets. Inaddition, there are further valuation difficulties inrelation to assets (Barrett and Reardon, 2000:27), and certain types of activity:


4 See Mishra and Goodwin (1997) who address farm income variability and how this affects the supply of off-farm labour. They also attemptto test whether spouses make joint decisions in their off-farm employment activity. Of course, their paper assumes that markets functionefficiently, again not always the case in the rural economies of developing countries. However, utilizing an econometric approach, the authorsfound that the off-farm labour supply of farmers is positively correlated with the riskiness of farm incomes; that farmers and their spouseswith more farming experience are less likely to work off-farm; and that off-farm labour supply is correlated with off-farm experience.

6

• the quantity of the asset may not be knownaccurately (this is common with land, forinstance)

• it may be difficult to value assets for whichno local market exists

• some assets are held in common with otherhouseholds; describing a ‘share’ is difficult

• some of the most important assets –especially components of human capital(e.g. skills, health) and social capital (e.g.capacity to make claims on others) – aredifficult to observe accurately

• it is difficult to observe (in a survey) qualitydifferentiation (e.g. soil quality, animalhealth)

• there may be multiple activities undertakenby several household members over severalseasons of the year

• some activities are illegal or ‘informal’ andare, therefore, not readily reported

• the activities are often undertaken part-timeand mixed with other activities (such asoperating a small-scale retail commercebusiness mixed with household chores andfarm labour in a given season)

• there are also several income valuationproblems given the diversity of incomesources, non-monetary income (e.g. barter),levels of remuneration, and empiricalproblems in accurately measuring andcollecting reliable income data in lessdeveloped countries.

To illustrate the difficulties, we contrast themethods used by Fafchamps and Minten (1998)with those of Minde and Nakhumwa (1997).Both pairs of authors were trying to understandthe activities of small traders. Fafchamps andMinten attempted to quantify social capitalamongst agricultural traders and their clients inMadagascar using a questionnaire-based samplesurvey for data collection and econometrictechniques for analysis of data7.

Explicit consideration of social capital inquantitative analysis is a useful and relativelynew development in research of this kind. Theauthors were aware, however, of problems ofunder-reporting. For example, Fafchamps andMinten (1998) noted that they were unable to

5 Ellis (2000b: 12) makes the point that there are no hard and fast rules governing income classifications (and the same can be said for activityclassifications). ‘Agriculture’ may be taken as a rough shorthand for renewable natural resources, so that gathering/cultivation of forestproducts and fishing are also included. This is the definition used in this paper reflecting a perceived need to correct for past oversight infocusing largely on agriculture, forestry and fisheries in rural areas. Non-farm activity includes agro-processing and trading activities, neitherof which is primary production, even if conducted on-farm. 6 Migratory activity and incomes are difficult subjects. Rural non-farm activity cannot include the activities of permanent migrants. The samecannot necessarily be said, however, for rural non-farm incomes, as under some definitions, remittances from former members of thehousehold who have permanently moved away would be regarded as unearned rural non-farm income. In this paper, unearned income fromsuch sources is included in our definition of the RNFE.7 It is important to measure social capital because this can have important implications for the operation of labour markets and barriers toentering the RNFE. For example, certain employment opportunities may not require a great deal of capital, experience or skill, but a friendshipor kinship relationship might be an important determinant of access. Fafchamps and Minten (1998) suggest that social capital can“…substitute for perfect markets and enable agents to economize on transactions costs”.


Table 1: The heterogeneity of rural non-farm activities

Primary Secondary Tertiary

Agriculture5 Mining/other Manufacture Servicesextractive

Waged L M L M L M L Memployment

Self- L M L M L M L Memployment

Source: Barrett and Reardon (2000: 40)

L = local; M = migratory6

7

trace traders who are the “least formal and havethe least permanent form of operation”. This wasbecause they used a standard two round samplesurvey and were thus unable to trace the moreitinerant traders interviewed in the first round.The importance of social capital to this type oftrader and vice versa is, therefore, not explored.Secondly, the authors also noted that “Indo-Pakistani traders, who constitute a small minorityof traders, tended to refuse participation to thesurvey”. The reasons for or implications of thisare not discussed. However, experience suggeststhat traders can be very suspicious of researcherswho record responses on forms or in writing.More generally, accessing information on‘shady’ or illegal activities remains problematicwith a standard sample survey methodology. Thefact that such activities are known to bewidespread and important sources of incomediversification in many rural areas hasimplications for the accuracy of, amongst otherthings, attempts to quantify social capital.

It is probably the case that gathering accurateinformation on certain types of activities requiresa high degree of trust, and this can only bedeveloped over time scales which are typicallymuch longer than those set aside for relationshipbuilding in most sample surveys. The need todevelop relationships over time is highlighted ina study of informal cross-border trade betweenMalawi and neighbouring countries (Minde andNakhumwa, 1997). This activity is known to be akey source of livelihood diversification amongstrural people around Malawi’s borders(Whiteside, 1998). Like Fafchamps and Minten,the authors were studying trading activities, incontrast, however, they used more qualitativemethods and monitored over a full calendar year.Minde and Nakhumwa (1997: 13) noted that:“…there was considerable suspicion about themonitors during the first two months of thesurvey because they were mistaken for eitherpolicemen or customs personnel”. This problemwas tackled by the monitors actively distancingthemselves from police and customs personneland through a process of getting to knowindividual traders informally. “Because the

traders tended to work along fixed routes, fixingthe monitors allowed a rapport to developbetween them, thereby facilitating collection ofvalid and accurate information” (Minde andNakhumwa, 1997: 13).

Concepts, Assets, Activitiesand Livelihoods

This report is structured around the concepts oflivelihood and diversity. “A livelihood comprisesthe assets (natural, physical, human, financialand social capital), the activities, and the accessgained to these … that together determine theliving gained by … the household” (Ellis, 2000b:10).

Assets form a household’s endowment ofresources with which to gain a living. In thisdefinition, the conventional meaning of assets isexpanded to include, besides material andfinancial resources, household members’ skillsand experience (human capital) and theirrelations within wider communities (socialcapital). This inclusive definition, as well as theuse of the term ‘capital’ in these senses, is notuncontroversial (Fine, 1999), but it serves tohighlight several unifying features of diverseresources. They require investment, in terms oftime or money, to be obtained or formed. Theycan (but need not) be used in an economicallyproductive way and in so doing, they are(imperfectly) substitutable and complementhousehold labour.

Activities comprise all the ways in whichhousehold members utilize their non-leisure timeto support their livelihoods. This broad definitionincludes work and care, employment andentrepreneurship, agricultural production andtrade, and a range of other dichotomies, some ofthem depicted in Figure 1. Engagement inactivities both requires assets and may increasehouseholds’ stock of assets. Households’endowment of assets and involvement inactivities jointly support their level of well-being(Figure 2).


8

The second central term used in this paper isdiversity, which follows naturally from the ideaof livelihood. Diversity in a household’sactivities and income (which is one measure fora household’s living standard) “refers to theexistence, at a point in time, of … differenthousehold income sources…” (Ellis, 2000b: 14).Given heterogeneity in assets, diversity inincome is almost implied. Indeed, bothindividual and household income normallyderives from more than one source: incomediversification is the norm, specialization theexception (Barrett et al., 2001b).

Typically, household income diversity isespecially great in rural areas. Rural householdsare more often producers as well as consumers,which implies the presence of profit (from soldoutput) or in-kind income (if output is consumed)as income components in addition to, forinstance, wages. Also, the relatively lowerremuneration of capital and labour and the morelimited market development that often

characterizes rural areas make it less likely thatany single source of income is sufficient to meetrural household needs.

For similar reasons, income diversity isparticularly relevant to developing economies. Inthis context, it is unsurprising that theorizingabout, and empirical study of, the economics oflivelihoods and diversity mainly draw onevidence from the developing world (Reardon etal., 1998; Start, 2001; Haggblade et al., 2002)and, more recently, also from the transitioneconomies in Central and Eastern Europe and theformer Soviet Union. Of particular interest is theevolution of diversity, i.e. the process ofdiversification of activities and incomes. This ishow households respond to changes in theireconomic environment, drawing on their variousassets to preserve or improve their livelihoods.As reflected in the definition of livelihoods, theyare enabled or restricted in doing so by theirsocial and economic environment.


Figure 2: A framework for livelihoods analysisSource: Adapted from Ellis (2000b: 30)

AssetsNatural capital: soil type, sunshine,precipitationPhysical capital: land, animals, machines, andbuildingsHuman capital: education, skills, experienceFinancial capital: savings, creditSocial capital: ethnic ties, membership oforganisation

Activities

Waged employment / independententerprise

On-farm / off-farmAgricultural / non-agriculturalCommercial / for subsistence, gift orbarter

Well-being

Reflected in: income, consumption assets, andexpenditures

LawsPoliciesIncentivesInstitutionsPROCESSES

STRUCTURELevels ofGovtPrivate Sector

TRANSFORMINGSTRUCTURES &

PROCESSES

Livelihood

9

Diversification Typologies

Distress-push and demand-pulldiversificationIn studying households’ diversificationstrategies, it is important to account for the factthat the motivations, means and outcomes ofdiversifying are heterogeneous. The twoextremes in this respect are, on the one hand, thetraditional belief that diversity of activitiessignifies a lack of economic development.Diversity is then juxtaposed with specializationof labour and efficiency gains (Lewis, 1954). Onthe other hand, more recently rural diversity inactivities and incomes has been identified as apotential motor for rural economic growththrough additional income generation andproduction and consumption linkages betweenagriculture, industry and services (Reardon et al.,1998; Start, 2001; Haggblade et al., 2002).Although such effects have been observed, ruraldiversity per se is clearly not a panacea for ruraldevelopment (Lanjouw and Lanjouw, 1995;Piesse and Thirtle, 2001; Deininger and Olinte,2001).

An approach that is more sensitive to thedifferent potentialities of rural diversity issuggested by a distinction in the literaturebetween ‘demand-pull’ and ‘distress-push’diversification (e.g. Efstratoglou-Todoulo, 1990;Reardon, 1999; Ellis, 2000b; Barrett et al.,2001a,b; Pearce and Davis, 2000; Haggblade etal., 2002). Distress-push diversification typicallyoccurs in an environment of risk, marketimperfections, and hidden agricultural un-employment, and is typically triggered byeconomic adversity, which sets the household ona downward income trajectory. It impliesengaging in economic activities that are lessproductive than agricultural production could beon a full-employment basis, and is motivated bythe need to avoid further income decreases.

Demand-pull diversification, on the other hand,is characterized as a response to evolving marketor technological opportunities, which offer thepotential for increasing labour productivity andhousehold incomes8. This distinction suggests anumber of specific inferences in terms of therelationship between diversification strategies,household characteristics and the socio-economic environment.

Regionally, distress-push diversification willdominate in rural areas which have one or moreof the following characteristics: geographicalisolation, low-quality physical infrastructure,low human capital, underdeveloped markets,scarcity of resources, or recent shocks to thenatural environment, economic system oragricultural sector. Demand-pull diversificationis possible in the presence of expandingtechnological innovations (whether within oroutside agriculture), market development orintensifying links with markets outside the localeconomy (Davis and Pearce, 2001).

Within any rural area, distress-pushdiversification attracts households in a ruralpopulation, which are either less well-endowedor have lower incomes. These households willenter non-agricultural activities that are, onaverage, less rewarding (e.g. in terms of labourproductivity) than demand-pull diversificationactivities, since the higher-return activitiestypically require higher investment that only thericher households can afford. For instance,poorer households will obtain a larger share oftheir non-agricultural income from wageemployment, while richer households have betteropportunities to enter non-agricultural activitiesin their own independent enterprise. Sinceincome inequality is typically such that there aremore relatively poor than relatively richhouseholds, distress-push diversification will bemore prevalent than demand-pull diversification.Distress-push and demand-pull diversification


8 Several studies in Asia (cited in Islam, 1997) show that the poorest groups (the landless and small farmers) diversify into activities wherewages are no higher than those in the agricultural sector, whilst higher income groups (larger farmers) also diversify, but into better paidsectors.

10

activities will be more clearly separatelyobserved as inequality is larger.

One implication of this approach is that thedistribution of diversification activities overhouseholds would follow a bimodal distributionover household incomes in the presence of bothdemand-pull and distress push diversification.There would be two clusters of low-return andhigh-return activities, engaged in by poor andaffluent households, respectively9. Moreover, ifdistress-push diversification dominates, wewould expect that poorer households are moreinvolved in diversification than others. In thecase of predominantly demand-pull diver-sification, we would expect that higher incomehouseholds engage more in non-agriculturaldiversification than the poorest households. Thisrelationship between returns to diversificationactivities and income levels of householdsengaged in them is reflected in some empiricalfindings on rural diversity (see Seppala, 1996;Carter and May, 1999).

This framework is one way of accounting forvaried evidence on the diversification-incomerelationship from different geographical areas,signifying different rural development patterns(Deininger and Olinte, 2001; Start, 2001; Imbsand Wacziarg, 2000; Piesse et al., 1999). Thedistinction between demand-pull and distress-push diversification is also useful for evaluatingthe economic significance of the RNFE. In manydeveloping countries, particularly in South Asia,demand-pull diversification occurrs, signifyingrural economic growth in the sense of increasingefficiency (Haggblade et al., 2002). In contrast,in many transition economies, it appears thathousehold diversification has often occurredduring a downhill trajectory of householdincomes, which would be distress-pushdiversification, and in conditions of a general

‘primitivization’ of the economy, that is, adecrease in the value-addition in the economy(Hedlund and Sundstrom, 1996; Ellman, 2000).

Other diversification typologies10

Based on the literature about peasant economics,we could in theory identify two principalcomponents when analysing the process of non-farm diversification – income and activity11. Theincome-driven, non-farm diversificationhypothesis assumes diversifiers are profit-maximizers, while the second, activity-driven,non-farm diversification, points to the differentcomparative advantage of household members asunderlying incentives for non-farm diver-sification (Ellis, 1993: 65–81, 146–123). Thus,two types of non-farm diversification may bedefined as follows: the first, income-drivendiversification, coincides with a period of capitalaccumulation (including financial and socialcapital, and information), while the second type,activity-driven diversification, often occurs later,when the afore-mentioned capital accumulationhas already taken place. Income diversificationdoes not necessarily exclude activitydiversification; we see it as a mixed and dynamicprocess, with income and activity diversification(depending on the household) often overlappingor occurring at the same time. Thus, for manyrural poor households, capital accumulation isthe consequence of income diversification, notthe aim of income diversification.

We would argue that there are two stages, whichare the components of a process that is notnecessarily sequential, but cyclical. First, theincome-dominant phase is more linked to the aimof covering households’ basic needs. This phasewill be dominant so long as meeting basic needsis the households’ main priority, as reflected inlow levels of income. When incomes are securelyabove a particular threshold, a certain amount of


9 What about those that are neither rich nor poor? Although numerous analyses of the RNFE and diversification tend to distinguish betweenthese two extremes, often the situation on the ground is not so clear-cut, therefore, it is important that diversification typologies are notoversimplified.10 This section is very much a ‘work in progress’, in terms of thinking through the process of livelihood diversification, identifying broadconcepts and patterns of diversification.11 For an analysis of peasant economics see Ellis (1993).

11

capital (whether financial, education, physical,land, etc.) may be accumulated. This is aconsequence of the income-diversificationstage). This enables the activity diversificationmotive to become more important, allowinghousehold members to pursue their comparativeadvantages in selecting particular activities, freedfrom the necessity of catering for basic needs bywhatever means available to them.

It bears repetition that, although we havedescribed this as a sequential process, it shouldbe thought of as a dynamic (possibly cyclical)process, with the dominance of one or anothertype varying from one stage to another (as a newincome allows the addressing of needs but alsoother potential investment opportunities).Income-driven diversification places stress onobtaining the necessary income to cover basicneeds while activity-driven diversification makesuse of surplus resources once the main incomesource(s) is (are) assured and thus encourages amore active entrepreneurial behaviour, that is,demand-pull diversification12.

To identify which of the two non-farmdiversification drivers are most prevalent at thevillage level (NUTS-5)13, Davis and Cristoiu(2002) in Romania and Davis (2002a) inArmenia, constructed two ratio-based incomeand activity diversification indices. Davis(2002a) applied them to village/rural municipallevel activity and employment data14. They foundthat a key weakness of these ratios was that theydid not consider ‘agriculture’ itself as a possible

second income generating activity. A moredetailed approach to diversification patternsshould consider pure non-farming ruraldiversification and hybrid non-farm ruraldiversification. The former takes into accountonly those individuals having a secondary non-farming activity while the latter accounts for bothfarming and non-farming activities (i.e. the ratioof the active population with a secondaryoccupation in farming or non-farming withrespect to the total active population).Nonetheless, diversification of the economymight well imply specialization of households,so these indices reveal little about household-level diversification strategies (see Davis andCristoiu (2002) for a full explanation of theapproach).

There are many interesting ways of thinkingabout livelihood diversification, some of whichwe have touched on above and other ‘ideas inprogress’ which we will briefly outline below.When considering the type of activities in whichan active population is involved, three differentdiversification patterns may occur: (i) inside-; (ii)ebb- (or distress-push); and (iii) flow- (ordemand-pull)15 diversifiers.

Inside-diversifiers are those who choose a secondjob in the same domain (either agricultural ornon-agricultural sector) as their primary activity(e.g. a farmer with a secondary activity of off-own farm work for cash). This would be mostcommon in the case of low capital endowments(financial or human), or among those rural


12 It should be noted that it is possible to look at two different aspects of the same problem. The theoretical observations offered above havebeen framed in terms of groups of households. However, if we consider each individual, it is as likely that the first impulse for diversificationis income (primary diversification) and then activity driven (secondary diversification). To a certain extent, this could also be applied to asingle household, assuming that all its members have the same goal and comparative advantage. Thus, at an individual household level,diversification could be considered a sequential process. On the other hand, as we have argued above, different households are usually atdifferent stages of development (or capital accumulation) so for some of them the income-driven stage will predominate while for othersactivity-driven diversification would be more important.13 The NUTS nomenclature (Nomenclature of Territorial Units for Statistics) is a five-level hierarchical classification (three regional levelsand two local levels) drawn up by Eurostat to provide a single uniform breakdown of territorial units for the production of EC regionalstatistics, for socio-economic analyses of the regions and for the framing of EC regional policies.14 The data used comprised two components: (i) regional employment data (in agriculture, public sector, non-agriculture, etc.); and (ii) regionaldata based on types of activities (as handicrafts, trade/commerce, services, etc.). Thus, the estimated income ratios included those having asource of income (from employment) while activity ratios included those into non-farming activities (as recorded in government databases).This differentiation (income and activity) was drawn by the way the data were collected for the database. See Davis and Cristoiu (2002), fora fuller explanation of the approach.15 Ebb and flow diversifier notions are more illustrative in the context of developing and transition economies, emphasizing the dynamiccharacter of diversification in an unstable economic environment. This suggests that diversification does not have a permanent character.

inhabitants who are not prepared to risk enteringinto a different activity domain16. However, thereare some exceptions to this case, such as roadsidehawkers and ‘higglers’.

Ebb-diversifiers are those whose primary activityis in the non-farm domain and who choose asecond activity in the agricultural sector. Apredominance of ebb-diversifiers indicates asituation where either non-farm income does notcover subsistence needs, forcing people back intoagriculture, or where there are distortedagricultural prices (either high due to low levelsof agricultural productivity and efficiency, or lowdue to state policies protecting low incomeconsumers in urban areas but with a concomitantde-capitalizing impact in farming communities).

Flow-diversifiers are those with a primaryactivity in agriculture and a second activity in thenon-farm economy. These are the demand-driven, risk-taking diversifiers, often having abetter financial and/or human capital

endowment, hence better equipped to takeadvantage of market opportunities, and thus ableto diversify. It may also be the case that theseflow-diversifiers cannot find opportunities fordiversification within agriculture and, therefore,try to re-orient their activities (and/or sources ofincome) to non-agricultural activities. Figure 3summarizes the different diversification patternsoutlined above.

To illustrate further, a farmer running aprocessing plant is considered an inside-diversifier if the plant processes agriculturaloutputs (i.e. a bakery) and a flow-diversifier ifwhat is processed is non-agricultural (i.e. TV setproduction). In using the ebb and flow terms, weare trying to suggest a possibly unstable(fluctuating) labour market so that people willuse farming as a temporary buffer or safety netduring unemployment periods or temporary lackof opportunities in their main expertise domain.Thus they may return to their main job when theyidentify an opportunity to do so (e.g. an

12


16 It should also be noted, that sometimes diversifying within the same activity domain can in fact increase risk, not lessen it as incomefluctuations tend to co-vary.

Farming

Activity

Primary Secondary

Non-FarmingFarmingNon-Farming

Inside-diversifier

Flow-diversifier Ebb-diversifier

Inside-diversifier

Figure 3: Diversification patternsSource: Davis and Cristoiu (2002)

13

unemployed factory worker will temporarilymove into agriculture to cover his/her basicneeds but on identifying an opportunity to returnto their job at a factory, will flow-out ofagriculture (unless the agricultural income ishigher than the income they would obtain at thefactory). We view this as an ebb-flow non-farmdiversification, that is, temporary movementsinside-outside agriculture.

Davis and Cristoiu (2002) in Romania and Davis(2002a) in Armenia identify which of the twonon-farm diversification drivers are mostprevalent at the village level, and found that mostpeople in both countries had had more than onejob based on a process of income diversification.In Romania, Davis and Cristoiu (2002) foundthat with more than half of the population livingin poverty, the main priority of rural inhabitantswas to cover their basic needs. Davis’ (2002a)study of Armenia shows that there appears to bea relatively low level of non-farm diversification.In Armenia 63.5% of the surveyed ruralpopulation is primarily employed in farming,25% in the purely non-farm group and theremaining 11.3% in the hybrid diversifiers group.Davis and Cristoiu (2002) found that the level ofnon-farm diversification in Romania was higherwith around 46% of the population primarilyemployed in agriculture and 37% in non-farm

employment with 17% hybrid diversifiers. Theyfound in both communities that inside-diversifiers, that is, those who select theirsecondary activity from the sphere of their primeactivity, diversify within their primary branchmainly because of a shortage of capital (mainlyfinancial) and/or their reluctance to take risks. Inall the surveyed communities, there were moreflow-diversifiers (main activity farming,secondary activity non-farm), than ebb-diversifiers (main activity non-farm, secondaryactivity farming). Thus, farming is the mostfrequent primary activity in rural areas. Thedifferences are greatest in the most affluentregions (in terms of per capita GDP), where thepopulation has a better financial situation and canafford to diversify for demand-pull reasons. Theshare of ebb-diversifiers was only larger in a fewcommunities, typically those which had well-developed tourism/day-tripper services andfacilities, as well as a vibrant natural resourcebased industry, for example, fishing or forestry(see Davis, 2002b; Davis and Cristoiu, 2002).Although secondary employment is probablyunder-reported in official statistics for bothArmenia and Romania, appropriate policies andprogrammes need to be put in place whichconsider both local endowments and humancapital characteristics (in terms of gender,education, age, etc.) to promote the RNFE.


Taking a livelihoods view on rural economicdiversity implies a holistic perspective that goesbeyond defining and measuring the size of theRNFE. The distress-push/demand-pull distinc-tion introduced in chapter 1 suggests that thereare different prerequisites, constraints,motivations and outcomes for householdsengaging in the RNFE (see Ellis (2000b), Barrettet al. (2001a) and Haggblade et al. (2002), formore detailed surveys).

Factors Enabling Householdand Enterprise Diversification

One approach in exploring these factors is torealize that development of the RNFE is oneform of local economic growth. There are anumber of factors that are important for(regional) economic growth in general, includinggrowth of the RNFE through diversification offarm activities or the operation of non-farmenterprises. Such economic growth throughincreased diversification may be apparent in bothincreased diversification of farm activities, inincreasingly diversified household incomesources, and in changes in the distribution ofincome, leading to larger or smaller rural incomeinequality.

Not surprisingly, therefore, the literature on ruralhousehold specialization and diversification islargely based on evidence from the developingworld; and the literature on the factorsunderlying regional economic growth largelyoverlap. We may distinguish betweenhousehold/enterprise-level factors and group-level (village, region) factors that affect the

distribution of household labour over incomesources, and thus diversification. Household/enterprise-level factors include the following.

• Asset endowments (such as land, livestock,real estate) and savings, i.e. wealth, as wellas income levels increase the opportunity toinvest in education, contacts or productiveassets that generate income either throughentrepreneurship or wage labour. It could beargued that asset endowment is moreimportant because many developing countrymarkets, particularly credit markets, eitherfunction poorly or are non-existent.Endowments and the level of income tend toencourage specialization in the mostproductive activity.

• Access to markets. Markets may begenerally absent or malfunctioning in aregion (i.e. land markets, credit markets), orthey may be inaccessible for people(typically the rural poor) with low social,financial or human capital. Market access isalso determined by factors such as distanceto markets, access to transport, infrastructureand telecommunications, access to marketinformation, the quality of goods andservices produced, volumes produced, etc.

• Human capital attributes (age, skills,education) broaden the set of employmentand entrepreneurial options for individuals.Household age composition (usuallyassessed in the form of dependency ratios)and education levels are an often-citedmeasure of human capital used empiricallyto explain the degree of participation across

Economic Diversity and Growth 3Chapter

16

a wide range of income groups in the RNFE.For example, Abdulai and Delgado (1999)found that the probability of participation innon-farm work increases with age up to 33for men and 30 for women and is, thereafter,inversely related to age. A higher level ofeducation is positively correlated with ahigher probability of participation for bothhusbands and wives in the RNFE, and ishigher for wives than husbands. However, ahigher level of educational attainment for awife lessens the probability of the husbandparticipating in the RNFE. Women’sparticipation in non-farm work was moresensitive to a lack of household cash thantheir husbands’ participation. Yet despite theintuitive appeal of these findings, therelationship between incomes and educationis not that clear-cut. Lanjouw (1999)suggests that educational credentials may beused to ration access to scarce regular non-farm employment opportunities. A generalincrease in education levels may ratchet upthe educational requirement (regardless ofits practical use) or result in a shift to otherselection criteria, still tending to exclude thepoor.

Reardon et al. (2000) argue that whereaccess to education is fairly equallydistributed, the effect will be to equalize theoverall size distribution of income.Moreover, where there are more non-farmemployment opportunities with loweducation requirements, rural non-farmincome inequality should be less. Reardon etal. (1998) attributes the poor distributionalconsequences of RNFE participation inAfrica to a scarcity of labour-intensiveactivities that have low entry barriers.

• Social capital . Participation in socialnetworks also broadens the set ofemployment and entrepreneurial options forindividuals. The concept of social capital hasseveral different interpretations. Fafchampsand Minten (1998) provide two definitions

from an economist’s perspective: “The firstmeaning sees social capital as a ‘stock’ oftrust and an emotional attachment to a groupor society at large that facilitates theprovision of public goods … The secondmeaning sees social capital as an individualasset that benefits a single individual or firm;this meaning is sometimes referred to associal network capital to emphasize thatagents derive benefits from knowing otherswith whom they form networks ofinterconnected agents.”

From a livelihoods perspective, the seconddefinition is perhaps most pertinent for thispaper. If social relationships are not taken intoaccount, the significance of barriers to enteringthe RNFE may be seriously under or over-estimated. For example, certain employmentopportunities may not require a great deal ofcapital, experience or skill, but a friendship orkinship relationship might be an importantdeterminant of access (Davis, 2002b; Bleahu2002). Fafchamps and Minten (1998) suggestthat social capital can “substitute for perfectmarkets and enable agents to economize ontransactions costs”. It is difficult to capture fullythe significance of social capital using a formalquestionnaire approach. However, such anapproach has been used by Fafchamps andMinten (1998), and Lanjouw (1998a,b) amongothers, who attempted to measure quantitativelythe impact of social capital. Using regressionanalyses, Fafchamps and Minten (1998),demonstrated that social network capital raisestotal sales and gross margins. Similarly, Lanjouw(1998b) in his study of the non-farm economy inMexico’s ejidos, used a social capital index andfound that ejidos with a higher score weresignificantly less likely to be poor.

Economic Diversity and Growth

17


Motivation forDiversification

• Risk may induce people to diversify income.The risk inherent in agricultural productionmay cause single-source income tofluctuate, which can be mitigated bydiversifying the portfolio of activities(Reardon et al., 1998). Economic theoryindicates that risk-neutral farmers willdivide their labour supply between on-farmand non-farm employment opportunitiessuch that the expected marginal returns to anextra hour of effort/work are equal. Iffarmers are risk-averse, either less time willbe allocated to the more risky jobs if theexpected returns to each sector are the same,or alternatively, the farmer will be willing toaccept lower wages in the less-riskyenvironment (Mishra and Goodwin, 1997).Non-farm labour can be used by farmers toreduce the total variance of their income,that is, the overall risk, or to increase thetotal returns to labour. However, this doesnot necessarily mean that risks associatedwith non-farm opportunities are lower than,independent of, or inversely related to on-farm risks – it is more the case that on-farmopportunities are often limited (Davis andPearce, 2001). While a combination of theabove demand/supply and labouravailability conditions must still hold inorder for RNFE activities to be viable, priceor income shocks may have constituted anadditional, or a major reason for individualsto consider diversifying into the RNFE.Price increases (indeed hyperinflation attimes), delayed payment of wages, and thecollapse of much of the socialist transportand outlet system (implying higher retailtransaction costs) are among the real income

shocks that rural people have experiencedduring transition. This would imply that theyare willing to pay a risk premium, in whichcase non-farm rural production could be lessproductive than food production and stillexpand.

• Seasonality . Seasonal labour and assetemployment of agricultural production maybe another reason for the growth of theRNFE. Using idle labour or machinery andempty buildings for non-agriculturalactivities may supplement incomes withoutcapital investments and at low opportunitycosts. As the demands of agriculturalproduction on labour and capital are usuallyseasonal, this motive would imply a strongcompetitive position for rural non-farmproducers, since revenue and profits arepractically equal as the additional costs ofexisting assets are fairly small. It wouldrestrict non-farm activities to those that arefarm-asset-based or capital-intensive. Itwould also interact with the risk motive as itstabilizes income over time.

Factors Enabling Growthof the RNFE

At the group (village, regional) level, the factorsstimulating diversification are mainly the sameas those that stimulate economic growth, whichimplies that we can consider diversification fromthe viewpoint of (endogenous) growth theory17

Relevant factors include the following.

• An important factor in growth prospects islocal natural/physical resources. However,although production is obviously predicatedon productive resources, resourcesendowment is not necessarily an important

17 The new growth economics approach assumes the presence of important externalities when combining labour and capital with an increasingamount of knowledge. The gains from education, for example, are not simply determined by how much a person's productivity is raised bytheir investment in education, but is also the result of interaction with many well-educated people and this constitutes positive externalities.An effective enabling institutional environment and good governance at all levels of administration operates in a similar way. Therefore,overall, models of economic growth make clear that productivity growth as well as capital formation (real and human capital) are both criticalto achieve high rates of growth in per capital GNP and thus sustainable livelihoods (see Breitschopf and Schreider, 1999).

18

factor in understanding growth patterns, fortwo reasons. First, given a set of factorendowments, there are many possible uses,each of which may or may not generategrowth. How resources are used is crucial togrowth, but it is not determined by theirmere presence. An illustration of this is astudy by Bernstein and Weinstein (1998),who statistically examined the relationshipbetween factor endowments and productionpatterns using international and Japaneseregional data. They found evidence ofsubstantial production indeterminacy, aresult which implies that regressions of tradeor output on endowments have weaktheoretical foundations. An additional factor,which complicates the link betweenresources and growth, is the possibility of a‘resource curse’ or ‘Dutch disease’ (Auty,2000). Countries that are richly endowedwith primary (e.g. mineral) resources tend tospecialize in the production and export ofthese resources, more so than othercountries. Since the economic returns to thisare low relative to alternative developmentpaths, such as expanding the trade orservices sectors, resource abundance maycrowd out high-productivity alternatives andso impede growth.

• Quality of local government (corruption, aidprogrammes, consistency of policies andtheir implementation) expressed, forinstance, by levels of corruption,government stability, policy volatility(mostly measured as monetary impulses),the annual number of coups and revolutions,or, sometimes, the level of democracy. SeeMoers (2000) for an overview of theliterature and Moers (1999) for an empiricalanalysis in transition countries. In this area,one would look for a rural-urban differencein corruption levels, bureaucratic quality, orthe nature of civic society that can helpexplain differences in economic activities.Governments are typically insensitive toRNFE development and the agricultural

paradigm usually dominates. In case ofintervention, actions by local administrativebodies seem most appropriate (Haggblade etal., 2002: 20). They have probably lessurban bias, are more knowledgeable aboutrural needs, and operate more efficiently atthe local level.

• Local physical infrastructure includingdensity of the road and telephone networksand household services is an importantaspect of the RNFE and growth (seeLanjouw and Feder, 2001). Jalan andRavaillon (1998) found that road density isone of the determinants of household-levelprospects of escaping poverty in rural China.Komives et al. (2001) found that there is adivision in access to household facilities(electricity, water, sewer and telephoneservices) between urban and ruralhouseholds, with the exception of EasternEurope and Central Asia. Here, ruralhouseholds have relatively higher levels ofcoverage.

• Proximity to towns, linkages with urbanareas. Rural growth often depends on linkswith urban areas, either through theacquisition of manufactured or consumptiongoods, or commuting incomes, or themarketing of rural produce in towns. Ruraltowns are also important for the followingreasons: public service provision,information, credit services, economies ofscale and agglomeration, etc. Also, ruraltowns can function as ports through whichrural producers can transport their goods tonational or global markets (World Bank,2000a,b).

• Trade and regional growth. Someeconomists stress that trade is important forgrowth; others maintain that productionlinkages, implying a barrier to entry forprospective trading enterprises, constitute analternative path to growth. Rauch (1997)shows that both views may be reconciled.


19

Another alternative to trade-led growth isimport substitution. Srinivasan andBhagwati (2001) gave an overview ofarguments in the trade versus importsubstitution debate, and argued that the costsof import substitution are greater than itsbenefits. They cited studies showing theweak theoretical foundations and the use ofincorrect methods deployed in support ofimport substitution; and that trade is still thebest growth option. Does openness, apartfrom influencing growth, also affectinequality, and by, implication, would thetransition affect rural/urban inequality?Theoretically, this is ambiguous. It dependson factor endowments, the type of openness(trade flows, capital flows or labour flows),and on complementarity and substitutabilityof factors of production, and on thedistribution of endowments over individualson different income levels. Empirically, theopenness-inequality connection researchedin regression analyses has suggested impactsin both directions in different countries(O’Rourke, 2001). There is also empiricalevidence that labour mobility may be afactor in decreasing inequality and thus ruralpoverty. Razin and Yuen (1996) showed thatlabour mobility is capable of generatingincome level equalization across regions inthe presence of knowledge spillovers, whilerestrictions on labour flows tend to makeindividual region/country per capita incomemore divergent.

Socio-cultural Aspects ofRNFE Growth

• Ethnicity and ethnic heterogeneity. Where anumber of different ethnic groups live in thesame area, whether in the same village ornot, anecdotal evidence suggests that oftenthere is occupational specialization alongethnic or ethno-religious lines, which affectsboth current participation in differentoccupations and the potential whichindividuals and households have for

diversification (see Fafchamps and Minten,1998). For example, Smith et al. (2001) intheir study of two districts in Uganda foundthe diverse social ethnic structure played acritical role in governing access to resourcesand thus patterns of rural livelihooddiversification. Bleahu and Janowski (2001)maintained in the case of Romania that someethno-religious groups have internal socialcharacteristics, which make it relativelyeasier for them to diversify out ofsubsistence agriculture. This is likely toenable members of that group to collaboratein, for example, marketing agriculturalproducts. It also enables members to succeedin entering non-agricultural activities; whereone member is already involved in a certainoccupation niche, he or she will facilitate theinvolvement of other members of the ethnicgroup. This is what has happened inTransylvania, where ethnic Germans(‘Saxons’) appear to be much better athelping each other gain entry to non-farmretail trading activities than their ethnicRomanian neighbours within the samevillage (Bleahu, 2002). Consequently thisgroup has had an involvement in trade andcommerce which the Romanian ethnic grouphas not had because they operate moreeffectively through networks of kin andneighbours. Bleahu and Janowski (2001)also noted the impact of discriminatingagainst particular ethnic groups. In theirstudy, Roma gypsies were essentiallyexcluded from accessing most forms of localemployment. Ethnicity is one dimension ofsocial capital.

• Gender and cultural aspects of RNFEaccess. Religion and a variety of culturalfactors may mean that there is a preferencefor involvement in certain types of non-farmlivelihood activity on the part of all membersof a community or some section of it. Thereare often activities which are seen asundesirable by members of certaincastes/classes or certain ethnic groups. There


20

are also activities which are seen asinappropriate for certain categories ofindividuals, because of their gender or age,for example. Access barriers may also berelated to caste or class divisions, ethnicity,language or other cultural factors (aspects ofsocial capital). High status groups of allkinds, including high castes and highstatus/majority ethnic groups, may gainaccess more easily into remunerative non-farm activities. Individuals and householdsbelonging to low status groups, on the otherhand, find it difficult to diversify into better-paid sectors, and tend to be forced intocertain less remunerative non-farmactivities. In many developing countries,women play a key role in farming and non-farm ancillary services. They are oftenresponsible for selling produce and forsubsistence production (see Canagarajah etal., 2001). Women have a strong influence inthe family in both providing and promotingthe education of children. Therefore, theymay warrant targeted programmes in bothextension and education, for example, fornon-farm and on-farm financial recordkeeping and organization. Women also oftenplay a key role in activities such as agro-tourism, weaving, light manufacturing work,etc., which often require credit and otheraids for small business assistance. Finally,Smith et al. (2001) found that in Ugandaculturally proscribed gender roles can shiftwhen the household is under pressure tobring in sufficient food and income tosurvive. Conflict, economic deteriorationand the impact of AIDS have intensified theburden on women, with an increase infemale-headed households that have inmany cases been forced into livelihooddiversification. In the case of Rakai, inUganda, women have received targetedsupport from NGOs in the form of technicaland financial assistance to handicraftsgroups and, to a lesser extent, baking groups.

• Regional human and social capital ismeasured by the level of education of thepopulation, the level of trust and theintensity of civic society (i.e. number ofassociations and clubs). Note that this is agroup characteristic and, as such, distinctfrom the educational level, skills andconnection that a single individual/household/farmer/farm manager may have.Putnam (1999) in his study of Southern Italyshowed that social capital (e.g. networks,norms, trust) co-ordinates actions and is anasset in economic growth; see also Temple(1998) for an overview of macro-economicperformance and social capital.

Linkages Between the Farmand Non-farm Economy

Haggblade et al (2002) stated that one reasonwhy the RNFE should be actively encouraged isbecause, when agriculture grows, the ruraleconomy benefits from powerful income andemployment multipliers. In many developingcountries, discrimination against small rural non-farm firms constrains the effects of thesemultipliers. As previously noted, the prevailingconception is that RNFE activities have closelinks with the agricultural sector; Heidhues et al.(1998) and Davis and Gaburici (1999) provideevidence of this in Romania. Although thissection focuses on the linkages between the farmand rural non-farm economy, these must also beviewed within the context of broader links. TheWorld Bank (Csaki and Lerman, 2000)emphasizes the links between the rural sector andall other sectors of the economy – not only thosebetween the rural sector and the agriculturalsector. In contrast to the distinction used inTable 1, they argue for a cross-sectoral context torural development due to “the ‘connectedness’ ofrural residents to many economic sectors, onlyone of which is agriculture”. For example, ruralindustry has strong links with the urban sector,due both to the market provided by the urbanarea and the links between firms, which may be


21

either competitive or complementary: ruralenterprises may provide components for urbanfirms, or may assemble or finish their products(Islam, 1997).

The farm and non-farm economy may be linkeddirectly through production activities, orindirectly through incomes or by investment(Reardon et al., 1998). Production linkages maybe either upstream or downstream: upstreamlinkages occur either when the farming sectorgrows and induces growth upstream in the supplyof inputs and services, or when growth of localmanufacturing and services reduces the price andincreases the availability of inputs upstream;downstream linkages take place when activities,such as agro-processing and distribution, thatrely on farm inputs, are increased and thusincrease the demand for farm products. Incomelinkages occur when income earned in one sectoris spent on the outputs of the other, andinvestment linkages take place when profits fromone sector are invested in the other. All theselinkages are important in the development ofnon-farm enterprises in developing countries andtransition economies. However, linkages may beweak and the strength of different linkages iscontext specific and depends on a number offactors. On production linkages, for example,constraints downstream in the RNFE may raiseprocessing and distribution costs and so inhibitfarm sector development, or upstream RNFEconstraints may raise input and services costs(Reardon et al., 1998). Increased opportunitiesfor rural non-farm employment would absorb theexcess labour found in agriculture and tend toresult in increased labour productivity(Christensen and Lacroix, 1997).

The RNFE, Poverty andInequality

We now turn from the constraints and conditionsof RNFE growth to its possible consequences.One of the reasons that diversification of therural economy is now a subject of interest is itspotential to reduce income inequality and thus

rural poverty. Again, this concern is also centralto the literature on economic growth, especiallyas applied to the developing world. We focus onthis and related literature on growth anddiversification. In treating diversification as aparticular form of economic growth andexploring the relation between diversificationand income inequality, a conceptual difficultyshould be noted. There is evidence that incomeinequality is often associated with take-offeconomic growth or with economic shocks ingeneral. Thus, economic growth may imply anincrease in inequality. As such, it is often asymptom of economic development rather than aproblem in itself.

Jian et al. (1996) found divergence betweencoastal and inland regions in China due to the‘economic zones’ policy. Ferreira (1997)investigated the distributional consequences ofpolicies and developments associated with thetransition from central planning to a marketsystem. The model suggests that even an efficientprivatization designed to be egalitarian may leadto increases in inequality (and possibly poverty),both during the transition and in the new steadystate. Another reason why growth generatesinequality is that it is usually accompanied by theformation of a centre of economic activity(Hanson, 1996), where incomes are probablyhigher.

If income levels and growth rates that arerelatively low in some (often rural) regions seemoften associated with economic growth, is therestill a case for aiming for a reduction ininequality? There are two rationales for such apolicy aim. First, extreme inequality and povertymay be problematic for social reasons, whilethere is also evidence that it actually impedesgrowth. Second, it may be that only certain typesof growth increase urban-rural incomeinequality.

There is not always a growth-inequality trade-off. Barro (1991, 1997) found a tendencytowards convergence among the US states,


22

among Japanese prefectures, and among regionswithin Western Europe. Also for poorercountries, empirical work in Bangladesh showsthat growth, reducing poverty in both urban andrural areas, is associated with rising inequalityonly in urban areas (Woodon, 1999). This findingreflects a large strand of literature on theconditions for growth to be pro-poor, focusedparticularly on rural areas (e.g. Ravallion andDatt, 1999).

Just as inequality, urbanization is often defined aspart of the rural development problem, while it isalso often a symptom of economic growth at thenational level. The degree of urbanization is anindicator of economic development at low levelsof per capita income. There is evidence that thereis a best degree of national urban primacy, whichincreases sharply up to a per capita income ofabout US$ 5000 (PPP 1987 income), beforedeclining modestly. The best degree of primacydeclines with country scale (Henderson, 2000).This would imply that urbanization, and thusdwindling rural populations, as such should notbe viewed as problematic in developing andtransition countries, where per capita incomesare generally lower. Indeed, in an exploration ofurban and regional dynamics in Poland after thetransition, Deichmann and Henderson (2000)found that the degree of urbanization andprimacy remains low in Poland. The largest citiesare not growing at the rate that would beexpected if post-transition adjustments wereoperating freely. As a result, Poland is not fullyrealizing external economies from urbanagglomeration, probably due to housingshortages and low labour mobility.

Having this in mind, we now consider if and howrural diversification, through the generation ofnon-farm income, can reduce rural poverty bydecreasing inequality. We have already noted thatthe small share of full-income farms in Centraland Eastern Europe imply a potentially majorrole for the non-farm economy in rural areas. Inaddition, extensive evidence of the role of non-farm income generation in other developing

country settings is a reason to investigate itspotential in the context of Central and EasternEurope. For example, Lanjouw (2001)maintained that the non-agricultural rural sector(in his definition comprising non-farmbusinesses) represents a potentially importantroute out of poverty in Ecuador. Poverty declinesas the share of income from non-agriculturalsources rises. Non-agricultural employment andearnings are positively associated with bettereducation and infrastructure access. Povertycould be expected to fall substantially withexpansion in the non-farm sectors ofconstruction, transport, commerce and services.Lanjouw (2001) also analysed a recent householdsurvey for Ecuador to assess the impact of thenon-agricultural rural economy in reducingpoverty. That sector accounts for roughly 40% ofrural incomes in Ecuador, 75% of which comesfrom non-agricultural enterprises as opposed towage labour. The sector provides employmentfor nearly 40% of economically active men and50% of women. This survey shows that, all otherthings equal, the greatest fall in poverty could beexpected from expanding employmentopportunities in transport, commerce-relatedactivities and services such as administration andthe hotel and restaurant trade. Although theevidence is from developing countries (as inAdams (2001, 2002) and Islam (1997)), it isworth considering whether it is also relevant torural areas in transition countries. And if so,which factors stimulate equity enhancingdiversification?

Reardon et al. (1998) identified a number ofconditions for the development of the RNFE tobe more equality enhancing which includeproximity to urban markets, physical and marketinfrastructure, resource endowments and thedistribution of productive resources within ruralareas. Piesse and Thirtle (2001) showed thataccess to markets increases the poverty reductionpotential of the RNFE in Zimbabwe. Deiningerand Olinte (2001), studying data from Colombia,found that specialization, in farm or non-farmactivities, increases linearly in wealth and in


23

income levels. The relationship betweendiversification and wealth/income is a U-shapedcurve. This suggests that there are two types ofdiversification: a low-return refuge from poverty,and a high-return innovative diversificationbased on high levels of asset endowment andhuman capital and a well-developed ruralinfrastructure, including access to credit markets.A similar U-shaped relationship is reported forfarms in the developed world. Theseobservations support the distinction between‘demand-pull’ and ‘distress-push’ factors in therural diversification process posed in chapter 1.They also explain the recent shift in attitudestowards the RNFE from viewing it as a symptomof backwardness towards a potential motor of therural economy (Lanjouw and Lanjouw, 1997).Rather than being contradictory, the older view

fits the low-income/wealth, ‘poverty refuge’,distress-pushed type of diversification, while themore recent view connects to the highincome/wealth, innovative, demand-pull type ofdiversification.

The implications for the study of the impact ofthe RNFE on poverty, or income inequality,reduction are the following. First, the increase indiversification in recent years (assuming thisexists) may be a positive sign of (renewed)economic viability, or a negative sign ofincreasing poverty, in line with the aboveclassification. Which type it is depends on thefactors identified above, which may well beincomplete given the ongoing explorations of thesubject (Deininger and Olinte, 2001).


This study has explored some of the scale anddefinitional complexities underlying the RNFE,and showed how susceptible it is to a widevariety of trends, shocks and processes. Amongthese wide-ranging influences, we need to beable to determine the extent to which the RNFEcan be treated as a distinct entity and, therefore,assess how much it might specifically be liable tobeing influenced by policy initiatives. Can theRNFE be identified in such a way that it issusceptible to separate analysis and, therefore,amenable to distinct policy interventions?

The difficulty is not only in the heterogeneity ofthe activities but also that in many cases wherethere is an expansion of non-farm activities, itmay be a consequence of a wide range ofinfluences. These may include changes in theagriculture sector, as with the Green Revolutionin parts of Asia, which acted as a driving forcefor small-scale industrial expansion in rural areasboth to supply inputs and process outputs. (Muchof the rapid expansion of rural industry andcommerce in India’s provinces of Punjab andHaryana can be ascribed to this process.) It mightbe a product of processes like those in China inthe early 1980s, with a combination of macro-economic shifts, local governmentdecentralization and decollectivization offarming coupled with higher procurement prices.In combination these allowed a rapid expansionof agricultural output, a resultant rise in incomes(driving consumer demand, especially for house-building), and surplus capital for investmentretained by the localities, all of which producedan enormous expansion in small-scale industryand commerce. The most successful places forthis were also in coastal provinces, and many

also benefited from the newly permitted influx offoreign investment.

The issues relating to potential expansion of theRNFE range from the overall impacts ofeconomic growth in the wider economy (whichmay or may not have a positive effect on theRNFE), through the (supply-side) increase ininvestment, the availability of labour foremployment, and the seeking of opportunities inother types of livelihood. It is important torecognize that the last – seeking other livelihoods– includes both those taken up out of preferenceand choice, and many others which are adoptedas a result of crisis or collapse of other activities,that is, they are distress-driven and survivalstrategies. These various components form aspectrum of ‘driving forces’ or political-economic environments that affect the potentialfor growth (or decline) of the RNFE and whichhave an impact by design or default on itsgrowth.

Governance, if defined as “the manner in whichpower is exercised in the management of acountry’s economic and social resources fordevelopment”, indicates that it is different fromgovernment, and that power may be exercisedwhich is not incorporated into administrative oreconomic structures that are formally constitutedor subject to democratic or other forms of socialmediation. Such power may clearly be exercisedat the local level, by both local and higher-levelplayers comprising not only constitutionalauthority (government), but also the privatesector, local elites and expressions of civilsociety (including co-operatives, NGOs andproducer organizations). Decentralization

The Rural Non-FarmEconomy Policy Processes

4Chapter

26

processes impact directly on local governance.The degree to which decentralization affects theRNFE positively or negatively will dependessentially on the degree to which it strengthensthe efficiency, effectiveness and relevance of theinstitutional environment. The institutionalenvironment influences the livelihood strategiesof non-farm rural households, throughdetermining their access to, and utilization of, arange of livelihood resources (natural, economic,human and social capitals). The relationshipbetween local governance and the RNFE is,therefore, a complex one, with institutionalchange and the processes driving it (such asdecentralization and economic reform andrestructuring) continually modifying andchanging the nature of the relationship.

Clearly the enormous range of issues outlined inthis paper demonstrates not only the complexityof the RNFE, but also the fact that policyinterventions at one level may also impact others(positively or negatively). Moreover, in terms ofpoverty-reduction strategies, the need is torecognize that there should be a movement awayfrom distressed or constrained livelihoodstowards chosen or superior activities (includingthe need to avoid wage-labour or self-employment in dangerous or low-pay activitiesthat may represent poor substitutes for existinglivelihoods).

Recognition of the complexity of the RNFE andits interrelationships with wider social andeconomic processes should enable the reductionof negative knock-on effects elsewhere,sectorally or spatially. Policy design should beshown to operate with positive generalequilibrium effects and without displacednegative impacts on the poor (e.g. throughharmful income or asset redistribution), orworsening gender inequality, or degradation ofthe natural environment. Subject to theseconstraints, and relating to the issues discussed inthis paper, key policy questions emerge:

• Economic growth – Where relevant, whatmacro-level policies will foster economicgrowth in the RNFE and which willundermine its development?

• Investment – (i) How can reinvestmentlocally be fostered to enhance locallivelihoods (and without significantinefficiencies in opportunity costs of thatinvestment?); (ii) How can investment fromoutside which promotes sustainablelivelihoods be encouraged?

• Employment – How can waged employmentbe generated as a significant form oflivelihood enhancement?

• Chosen livelihoods – What can be done toenhance the opportunities for people,especially the poor, to gain access to morediverse livelihoods?

• Constrained livelihoods – What circum-stances lead to people having to take updistress or coping strategies, and if they areincreasing in number what policies can treatthe problem? Is the expansion in copingactivities in part driven by negativeconsequences of other policies? How canpoor people be enabled to widen their choiceof livelihood alternatives such that they arenot reliant on endangering natural capital,low-pay and exploitative alternatives to farmpoverty?

We should also consider problems from theperspective of rural people themselves, and thiscan be done in respect to the way livelihoodopportunities are constructed through differentialaccess to the five types of capital (natural,financial, physical, social and human) that areincreasingly familiar in the analysis ofsustainable livelihoods (Ellis, 1998). Some of thekey issues that emerge from this approachinclude a focus on: property rights, markets,governments, trends and demographics, shocksand stresses, which are designed to abstract someof the key considerations that a priori areconsidered to be significant in influencing thecapital portfolios of people in the promotion ofthe RNFE. To be effective policy-makers,

The Rural Non-Farm Economy Policy Processes

governments and researchers will need todemonstrate how policy interventions thatinfluence RNFE diversification ‘drivers’ orlivelihood frameworks can positively affect theasset holdings of the rural poor. To evaluate theimpacts of RNFE opportunities for the rural poor,the following will need to be considered.

• Property rights systems mean theconglomeration of processes and structuresthat determine the relations between peopleand groups of people (including socialrelations such as class, gender, ethnicity,age) and their differential access to incomesand assets. It incorporates the de factodistribution of assets and the property rightsthat govern this pattern of control, togetherwith the process of appropriation anddistribution of surplus within the society.These systems by and large determine themanner in which power is held and operateswithin societies.

• Markets or commercial environments maynot always be ‘proper/fully functioning’markets. This is because although theyreflect the way property rights and politicalsystems operate (e.g. property rights thatdetermine economic influence also affectmarket structures significantly).

• Government is distinguished for tworeasons: firstly, in most societies it claims tobe independent of property rights (whileproviding them with a legal framework) andto be neutral in the way political institutionsoperate; and secondly it usually claims to becompetent in the implementation of policiesdesigned to impact on the economy, forwhich reason it is seen as the sovereignagent for receipt of, and use of, most officialdevelopment assistance, including policiesfor the RNFE that may emerge from thiswork.

• Trends and demographics provide the basisfor particular analysis of recognizablepatterns of change (in economic systems at

any level, or technologies), and bothquantitative and qualitative processesaffecting the population. This allows scopefor discussing the direction of change in arange of factors that may have a seriousimpact (positive or negative) on the RNFE,for instance, structural adjustment packages.Shocks and stresses are distinct events orparticular processes that need to beseparately identified from the broadertrends, in particular the impacts of disastersor epidemics and pandemics. Policies for theRNFE that neglect peoples’ vulnerability tohazards or the impact of HIV/AIDS areunlikely to be very robust.

• Migration and the associated remittancesconstitute another route to improvedincomes in rural areas, especially whereeconomic opportunities are lacking orlimited. Important conditions for smoothmigration flows include access toinformation about job opportunities indifferent areas, a functioning house marketin destination areas, availability of credit tocover initial migration costs, and support incase migrants fail and need to return home.Individuals who have little or no earnedincome and cannot or are reluctant tomigrate must rely on income support fromothers or the state. Although modest andincapable of sustaining a thriving ruraleconomy, state subventions to both poorindividuals and the job creation impact oflocating state services in rural areas (such ashospitals, schools, etc.) generate much-needed local income and provide a source ofdemand for locally produced goods andservices.

27

The Rural Non-Farm Economy Policy Processes

We began this paper by arguing that the RNFEwould be discussed as part of a growth strategyfor the economy and also as a defensive survivalstrategy. Macro-economic factors have a majoreffect on the RNFE, as they affect generalemployment opportunities and the institutionalframework within which the RNFE functions, inparticular, the education system, financialinstitutions and credit market, factors whichinfluence the development of micro- and small-medium enterprises and the land market and farmstructure. Reforms within the agriculture sectoralso have a major impact on the RNFE becauseof the linkages, both positive and negative,between the two sectors. In general terms,growth in the farming sector has a positiveinfluence on the RNFE and vice versa, but it isvital that the RNFE is expanded in order toimprove rural livelihoods in the long-run whenemployment in the farming sector is expected tocontract.

As the natural resource base has declined inmany developing countries, so livelihoods havebecome less secure and sources of income morevaried. However, their stocks of assets of variouskinds condition the ability of individuals andhouseholds to access non-erosive andaccumulative off-farm strategies. At theindividual level, health status and education asdeterminants of human capital are important,particularly education. At the household-level,gender and age profiles will to some extentdetermine the pattern of non-farm activitieswhich can be undertaken. Financial capital,physical capital (infrastructure, especially roads)and the quality of social networks can all becrucial, depending on livelihood circumstances.

In several societies, being female represents animportant ‘barrier to entry’ to non-farmemployment. Diversification in rural incomesmay also affect gender relations (women maybecome more marginalized if they are moreconstrained than men in their access to non-farmopportunities, or they may be empowered by newopportunities to earn income, and develop skillsand networks). Generational considerations canalso be important. Characteristically, it is theyoung men who migrate, either seasonally orpermanently to urban centres for work. The oldand women in most societies are less likely toundertake this type of strategy.

The role of off-farm income generation activitiesis seen as increasingly important for the viabilityand development of rural areas, and for ruralhouseholds to diversify income sources andenhance livelihood opportunities. The on-farmsector exerts considerable influence on the ruraleconomy, both directly and indirectly, throughancillary and associated industries, and throughemployment and income multiplier effects.Private sector activity unrelated to the on-farmsector, and generally centred in rural towns, isviewed as a key area for development of the ruraleconomy, and also a means of more closelylinking rural areas with urban and economiccentres. Micro and small-medium enterprisesplay an increasingly central role in public sectorinitiatives aimed at promoting off-farm economicactivity and the commercialization of the ruraleconomy (see Davis and Gaburici, 2001).

There remains a question as to whether theRNFE should be left to itself – with nationalgovernments and their agencies merely ensuring

Conclusions 5Chapter

30

that the institutional and other reforms continueto progress – or whether it requires positivesupport/intervention. We would argue that thelatter would be helpful, possibly even essential,but intervention needs to be informed by a clearconception of what the rural sector is likely tolook like 10 to 20 years into the future.Experience from less developed countries

suggests that broad-based economic growth can,under the right conditions, reduce the overallpoverty level. However, even with growth, therewill remain groups, which descend further intopoverty. Therefore, the national governments anddonor community need to identify clearadditional interventions that are demonstrablybeneficial alongside general economic policies.

Conclusions

Abdulai, A. and CroleRees, A. (2001)

Determinants of income diversification amongstrural households in southern Mali. Food Policy,

26 (4).

Abdulai, A. and Delgado, C. L. (1999)

Determinants of nonfarm earnings of farm-basedhusbands and wives in Northern Ghana.

American Journal of Agricultural Economics, 81(1): 117–130.

Adams, R. (2001) Non-farm Income, Inequality

and Poverty in Rural Egypt and Jordan.Washington DC: World Bank. (unpublished)

Adams, R. H. (2001) Nonfarm income,

inequality and poverty in rural Egypt and Jordan.World Bank Policy Research Paper, No. 2572.

Adams, R. H. (2002) Nonfarm income,

inequality and land in rural Egypt, EconomicDevelopment and Cultural Change, 50 (2).

Antle, J. (1983) Infrastructure and aggregate

agricultural productivity: International evidence.Economic Development and Cultural Change, 31

(3): 609–619.

Auty, R. (2000) How natural resources affect

economic development. Development Policy, 18(4): 347–364.

Babb, F. (1989) Between Field And Cooking Pot:

The Political Economy Of Marketwomen InPeru. Texas Press Sourcebooks in Anthropology,

No. 15. Austin: University of Texas Press.

Barrett, C. and Reardon, T. (2000) Asset, Activityand Income Diversification Among AfricanAgriculturists: Some Practical Issues.(unpublished)

Barrett, C., Bezuneh, M. and Abdillahi, A.(2001a) Income diversification, poverty trapsand policy shocks in Cote d’Ivoire and Kenya.Food Policy, 26 (4): 367–384.

Barrett, C., Reardon, T. and Webb, P. (2001b)Nonfarm income diversification and householdlivelihood strategies in rural Africa: Concepts,issues, and policy implications. Food Policy, 26(4): 315–331.

Barro, R. J. (1991) Economic growth in a cross-section of countries. Quarterly Journal ofEconomics, 106: 407–443.

Barro, R. J. (1997) Determinants of EconomicGrowth: A Cross Country Empirical Study.Cambridge Massachusetts: The MIT Press.

Bebbington, A. and Perrault, T. (1999) Socialcapital, development, and access to resources inHighland Ecuador. Economic Geography, 75 (4):395–418.

Bernstein, J. and Weinstein, D. (1998) Doendowments predict the location of production?Evidence from national and international data.NBER Working Paper, No. W6815.

Bezemer, D. and Davis, J. (2003) The rural non-farm economy in Armenia. In: Transition,Institutions, and the Rural Sector. Spoor, M.(ed.). In series: Rural Economies in Transition.

References

32

Lerman et al. (eds). Lanham, Maryland:Lexington Books.

Bhaduri, A. (1989) Employment and livelihood:The rural labour process and the formulation ofdevelopment policy. International LabourReview, 128 (6): 685–700.

Binswanger, H., Khanker, S. and Rosenzweig,M. (1993) How infrastructure and financialinstitutions affect agricultural output andinvestment in India. Journal of DevelopmentEconomics, 41 (2): 337–366.

Bleahu, A. (2002) Social factors affectinginvolvement in non-farm activities in Romania.Paper presented at the Workshop on the RuralEconomy in Central and Eastern Europe and theCommonwealth of Independent States, NationalResources Institute, University of Greenwich,6–7 March 2002.

Bleahu, A. and Janowski, M. (2002) Rural Non-farm Livelihood Activities in Romania: A reporton Qualitative Fieldwork in Two Communities.Chatham Maritime, UK: Natural ResourcesInstitute. (http://www.nri.org/rnfe/papers.htm)

Block, S. and Webb, P. (2001) The dynamics oflivelihood diversification in post-famineEthiopia. Food Policy, 26 (4): 333–350.

Breitschopf, B. and Schreider, G. (1999) Ruraldevelopment in transition economies – the casesof two counties in Romania. Paper presented atthe 48th International Atlantic EconomicConference 7–10 October, Montreal.

Bright, H., Davis, J., Janowski, M., Low, A. andPearce, D. (2000) Rural Non-Farm Livelihoodsin Central and Eastern Europe and Central Asiaand the Reform Process: A Literature Review.NRI Report, No. 2633. Chatham Maritime, UK:Natural Resources Institute.http://www.nri.org/rnfe/papers.htm).

Buchenreider, G. and Knuepfer, J. (2002)Household determinants of access to rural non-

farm employment in the Balkans. Paperpresented at the Workshop on the Rural Economyin Central and Eastern Europe and theCommonwealth of Independent States. NationalResources Institute, University of Greenwich,6–7 March 2002.

Canagarajah, S., Newman, C. and Bhattamishra,R. (2001) Non-farm income, gender, andinequality: Evidence from rural Ghana andUganda. Food Policy, 26 (4).

Carter, M. and May, J. (1999) Poverty, livelihoodand class in rural South Africa. WorldDevelopment, 27 (1): 1–20.

Chirca, C. and Tesliuc, E. (eds) (1999) FromRural Poverty to Rural Development.Washington DC: World Bank and NationalCommission for Statistics, Romania.

Christensen, G. and Lacroix, R. (1997)Competitiveness and employment: A frameworkfor rural development in Poland. World BankDiscussion Paper, No. 383. Washington DC:World Bank.

Csaki, C. and Lerman, Z. (2000) AgriculturalTransition Revisited: Issues of Land Reform andFarm Restructuring in East Central Europe andthe Former USSR. Washington DC: World Bank.

Davis, D. and Weinstein, D. (1996) Economicgeography and regional production structure: Anempirical investigation. American EconomicReview, 86 (5): 1266–1278.

Davis, J. (2001a) Conceptual Issues in Analysingthe Rural Non-Farm Economy in TransitionEconomies. NRI Report, No. 2635. ChathamMaritime, UK: Natural Resources Institute.http://www.nri.org/rnfe/papers.htm).

Davis, J. (2001b) Sustainable non-farm rurallivelihoods in Transition Economies:Understanding the access and capacityconstraints. Eastern European Countryside, 7.

References

33

Davis, J. (2002a) Constraints and Potential to theDevelopment of Rural Non-Farm Activities inArmenia. Joint Endeavour EuropeanCommission and UN-FAO/SEUR GCP/ INT/758/ EC-ARM. (unpublished)

Davis, J. (2002b) Diversification of Income andEmployment Opportunities: RNFE FeasibilityStudy in Selected Areas in Bulgaria andRomania. UN-FAO/SEUR. (unpublished)

Davis, J. and Cristoiu, A. (2002) Patterns ofRural Non-farm Diversification and Employmentin Romania: A County Level Analysis. NRIReport, No. 2639. Chatham Maritime, UK:Natural Resources Institute.http://www.nri.org/rnfe/papers.htm).

Davis, J. and Gaburici, A. (2001) Non-FarmEmployment in Small-Scale Enterprises inRomania: Policy and Development Issues. NRIReport, No. 2637. Chatham Maritime, UK:Natural Resources Institute.http://www.nri.org/rnfe/papers.htm).

Davis, J. and Pearce, D. (2001) The rural non-farm economy in Central and Eastern Europe. In:The Challenge of Rural Development in the EUAccession Process. Lerman, Z and Csaki, C.(eds). World Bank Technical Publication.Washington DC: World Bank.

Davis, J. and Robinson, E. (2000) Participationof the Poor in the Rural Non-Farm Economy:Conceptual and Practical Considerations, NRIReport. Chatham Maritime, UK: NaturalResources Institute.http://www.nri.org/rnfe/papers.htm).

Davis, J. R. and Gaburici, A. (1999) Theeconomic activity of private farms in Romaniaduring transition. Europe-Asia Studies, 51 (5):843–869.

Deichmann, U. and Henderson, V. (2000) Urbanand regional dynamics in Poland. World BankPolicy Research Working Paper, No. 2457.

Deininger, K. and Olinte, P. (2001) Ruralnonfarm employment and income diversificationin Colombia. World Development , 29 (3):455–465.

Efstratoglou-Todoulo (1990) Pluriactivity indifferent socio-economic contexts: a test of thepush-pull hypothesis in Greek farming. Journalof Rural Studies, 6 (4): 407–413.

Ellis, F. (1993). Peasant Economics. FarmHousehold and Agrarian Development. Secondedition. Cambridge: Cambridge UniversityPress.

Ellis, F. (1998) Survey article: Householdstrategies and rural livelihood diversification indeveloping countries. Journal of AgriculturalEconomics, 51 (2): 289–301.

Ellis, F. (2000a) The determinants of rurallivelihood diversification in developingcountries. Journal of Agricultural Economics, 51(2): 289–302.

Ellis, F. (2000b) Rural Livelihoods and Diversityin Developing Countries. Oxford: OxfordUniversity Press.

Ellman, M. J. (2000) The Russian economyunder El’tsin. Europe-Asia Studies, 52 (8): 1424.

Fafchamps, M. and Minten, B. (1998) Returns tosocial capital amongst traders. Markets andStructural Studies Division Discussion Paper,No. 23. Washington DC: International FoodPolicy Research Institute.

Feder, G. (1986) Growth in semi-industrialcountries: A statistical analysis. In:Industrialization and Economic Growth.Chenery, H., Robinson, S. and Syrquin, M. (eds).New York: Oxford Economic Press.

Ferreira, F. (1997) Economic transition and thedistributions of income and wealth. World BankPolicy Resarch Paper, No. 1808.

References

34

Ferreira, F. and Lanjouw, P. (2001) Ruralnonfarm activities and poverty in the BrazilianNortheast. World Development, 29 (3): 509–528.

Fine, B. (1999) The developmental state is dead,long live social capital? Development andChange, 30 (1): 1–19.

Gordon, A. (1999) Non-farm Rural Livelihoods.Policy Series 4. Chatham, UK: NaturalResources Institute.

Gordon, A. and Craig, C. (2001) Rural Non-farmActivities and Poverty Alleviation in Sub-Saharan Africa. Policy Series 14. Chatham, UK:Natural Resources Institute.

Grootaert, C. (2000) Social capital, householdwelfare, and poverty in Indonesia. World BankPolicy Resarch Paper, No. 2148.

Grooteart, C. and Narayan (2001) Localinstitutions, poverty, and household welfare inBolivia. World Bank SDFESSDN Working Paper.

Haggblade, S., Hazell, P. and Brown, J. (1989)Farm-nonfarm linkages in rural Sub-SaharanAfrica. World Development, 17 (8): 1173–1201.

Haggblade, S., Hazell, P. and Reardon, T. (2002)Strategies for stimulating poverty-alleviatinggrowth in the rural nonfarm economy indeveloping countries. International Food PolicyResearch Paper. Washington DC: InternationalFood Policy Research Institute.

Hanson, G. (1996) Localization economies,vertical organization, and trade. AmericanEconomic Review, 86 (5): 1266–1278.

Harper, D. and StLouis, L. (1999) Regionaleconomic diversification and efficiency:Baumol’s likely lower confidence limit measureof risk. Review of Regional Studies, 29 (2):197–211.

Haug, R. (1999) From integrated ruraldevelopment to sustainable livelihoods: What is

the role of food and agriculture? Forum forDevelopment Studies, 10 (2): 181–201.

Hedlund, E. and Sundstrom, E. (1996) TheRussian economy after systemic change. Europe-Asia Studies, 48 (6): 900–901.

Heidhues, F., Davis, J. R. and Schrieder, G.(1998) Agricultural transformation andimplications for designing rural financial policiesin Romania. European Review of AgriculturalEconomics, 25 (3).

Helliwil, J. (1996) Do borders matter for socialcapital? Economic growth and civic culture inUS States and Canadian Provinces. NBERWorking Paper, No. W5863.

Henderson, V. (2000) How urban concentrationaffects economic growth. World Bank PolicyResearch Working Papers, No. 2412.

Hill, B. (1999) Farm household incomes:Perceptions and statistics. Journal of RuralStudies, 15: 345–358.

Imbs, J. and Wacziarg, R. (2000) Stages ofdiversification. NBER Working Paper, No. 2642.

Islam, N. (1997) The non-farm sector and ruraldevelopment – review of issues and evidence.Food, Agriculture and the EnvironmentDiscussion Paper , 22. Washington DC:International Food Policy Research Institute.

Jalan, J. and Ravaillon, M. (1998) Geographicpoverty traps? Institute for EconomicDevelopment, Discussion Paper, No. 86. Boston:Institute for Economic Development, BostonUniversity.

Jamal, V. and Weeks, J. (1993) AfricaMisunderstood or Whatever Happened to theRural-urban Gap? ILO Studies. London:Macmillan Press.

Janowski, M. (2002) Social factors affectinghousehold level involvement in the RNFE in

References

35

Armenia, Georgia, and Romania: Lessonslearned. Paper presented at the Workshop on theRural Economy in Central and Eastern Europeand the Commonwealth of Independent States,National Resources Institute, University ofGreenwich, 6–7 March 2002.

Jian, T., Sachs, J. and Warner, A. (1996) Trendsin regional inequality in China. NBER WorkingPaper, No. W5412.

Knight, J. and Song, L. (1997) Chinese peasantchoices: Farming, rural industry, or migration.University of Oxford Applied EconomicsDiscussion Paper Series , No. 188. Oxford:University of Oxford.

Komives, K., Whittington, D. and Wu, X. (2001)Infrastructure coverage and the poor: A globalperspective. World Bank Policy Resarch Paper,No. 2551.

Lanjouw, J. and Lanjouw, P. (1995) Ruralnonfarm employment: A survey. World BankPolicy Research Paper, No. 1463.

Lanjouw, J. and Lanjouw, P. (1997) The ruralnon-farm sector: an update. Paper presented atthe XXIII International Conference ofAgricultural Economists (IAAE) on FoodSecurity, Diversification and ResourceManagement: Refocusing the Role ofAgriculture, Sacramento, USA, 10–16 August.

Lanjouw, P. (1998a) Ecuador’s rural nonfarmsector as a route out of poverty. World BankPolicy Resarch Paper, No. 1904.

Lanjouw, P. (1998b) Poverty and the Non-FarmEconomy in Mexico’s Ejidos: 1994–1997.Amsterdam: Free University of Amsterdam andDERCG, World Bank. (unpublished)

Lanjouw, P. (1999) Rural nonagriculturalemployment and poverty in Equador. EconomicDevelopment and Cultural Change, 48 (1):91–122.

Lanjouw, P. (2001) Nonfarm employment andpoverty in rural El Salvador. World Development,29 (3): 529–547.

Lanjouw, P. and Feder, G. (2001) Rural Non-farmActivities and Rural Development: FromExperience Towards Strategy. Washington DC:World Bank. (unpublished)

Lanjouw, P. and Ferreira, F. (2000) Rural non-agricultural activities and poverty in theBrazilian Northeast. Economics Department ofPUC-Rio Discussion Paper, No. 428.

Lanjouw, P., Quizon, J. and Sparrow, R. (2001)Non-agricultural earnings in periurban areas ofTanzania: Evidence from household survey data.Food Policy, 26 (4): 385–403.

Leibrandt, M., Woolard, C. and Woolard, I.(2000) The contribution of income componentsto income inequality in the rural formerhomelands of South Africa: A decomposable ginianalysis. Journal of African Economies, 9 (1):79–99.

Lewis, A. W. (1954) Economic developmentwith unlimited supplies of labour. ManchesterSchool of Economics and Social Studies, 22:139–191.

Minde and Nakhumwa (1997) Informal Cross-Border Trade Between Malawi and HerNeighbouring Countries. Report for TechnoserveInc., Nairobi, Kenya, March 1997. Funded byUSAID Regional Economic DevelopmentSupport Office for East and Southern Africa(REDSO/ESA) and executed by AgriculturalPolicy Research Unit (APRU), Bunda College ofAgriculture, University of Malawi.

Mishra, A. K. and Goodwin, B. K. (1997) Farmincome variability and the supply of off-farmlabour. American Journal of AgriculturalEconomy, 79: 880–887.

Moers, L. (1999) Growth empirics internationalinstitutional measures and its application to

References

36

transition countries: a survey. Tinburgen InstituteDiscussion Paper, No. 98-126/2.

Moers, L. (2000) How important are institutionsfor growth in transition countries? TinburgenInstitute Discussion Paper, No. 99-004/2.

Morrison, B. (1980) Rural household livelihoodstrategies in a Sri Lankan village. The Journal ofDevelopment Studies, 16 (4): 443–462.

O'Rourke, K. (2001) Globalization andinequality: historical trends. pp. 39–67. In:Annual World Bank Conference on DevelopmentEconomics. Washington DC: World Bank.Reprinted 2002 in Aussenwirtschaft, 57 (1):65–101.

Pearce, D. and Davis, J. R. (2000) The role of thenon-farm rural sector in the reconstruction of theBalkans. MOST-MOCT: Economic Policy inTransition Economies, 10 (2): 207–228.

Piesse, J. and Thirtle, C. (2001) Leveller forSome? Non-Farm Income and Equality inZimbawe. London: Birkbeck College, Universityof London. (unpublished)

Poirson, H. (2000) Factor reallocation andgrowth in developing countries. IMF WorkingPaper WP/00/94.

Putnam, R. (1999) Transforming social practice:An action science perspective. ManagementLearning, 30 (2).

Ravallion, M. and Datt, G. (1999) When isgrowth pro-poor? Evidence from the diverseexperience of India’s states. World Bank PolicyResearch Working Paper Series.

Rauch, J. (1997) Balanced and unbalancedgrowth. Journal of Development Economics, 54:493–518.

Razin, A. and Chi-Wa Yuen (1996) Labormobility and fiscal coordination. NBER WorkingPaper, No. W5433.

Reardon, T. (1997) Using evidence of householdincome diversification to inform study of therural non-farm labour market in Africa. WorldDevelopment, 25 (5): 735–747.

Reardon, T. (1999) Rural Non-Farm income InDeveloping Countries. Rome: Food andAgricultural Organization of the United Nations.(unpublished)

Reardon, T., Crawford, E. and Kelly, V. (1994)Links between nonfarm income and farminvestment in African households: Adding thecapital market perspective. American Journal ofAgricultural Economics, 76 (5): 1172–1176.

Reardon, T., Stamoulis, K., Cruz, M-E.,Balisacan, A., Berdegue, J. and Banks, B. (1998)Rural Non-Farm Income in DevelopingCountries. The State of Food and Agriculture1998: Part III. Rome: Food and AgriculturalOrganization of the United Nations.

Reardon, T., Taylor, J., Stamoulis, K., Lanjouw,P. and Balisacan, A. (2000) Effects of nonfarmemployment on rural income inequality indeveloping countries: An investment perspective.Journal of Agricultural Economics, 51 (2).

Reinert, K. (1998) Rural non-farm development:A trade-theoretic view. Journal of InternationalTrade and Economic Development, 7 (4):425–437.

Roberts, D. (1998) Rural-urban interdepen-dencies: Analysis using an inter-regional SAMmodel. European Review of AgriculturalEconomics, 25: 506–527.

Sarris, A., Doucha, T. and Matijs, E. (1999)Agricultural restructuring in Central and EasternEurope: Implications for competitiveness andrural development. European Review ofAgricultural Economics, 26: 305–329.

Schultz, T. W. (1954) Transforming TraditionalAgriculture. New Haven: Yale University Press.

References

Seppala, P. (1996) The politics of economicdiversification: Reconceptualizing the ruralinformal sector in South-east Tanzania.Development and Change, 27 (3): 557–578.

Shorrocks, A. (1982) Inequality decompositionby factor components. Econometrica, 193–211.

Smith, D. R., Gordon, A., Meadows, K. andZwick, K. (2001) Livelihood diversification inUganda: Patterns and determinants of changeacross two rural districts. Food Policy, 26 (4):421–435.

Srinivasan, T. N. and Bhagwati, J. (2001)Outward orientation and development: Arerevisionists right? In: Trade, Development andPolitical Economy: Essays in Honour of AnneKruegar. London: Palgrave.

Start, D. (2001) The rise and fall of the rural non-farm economy: Poverty impacts and policyoptions. Development Policy Review, 19 (4):491–505.

Swain, N. (1999) Rural development and socialchange in the post socialist central Europeancountryside. Zemedelska Ekonomika, 45 (2):79–84.

Swift, J. (1998) Factors Influencing theDynamics of Livelihood Diversification andRural Non-farm Employment in Time and Space.(unpublished)

Swinnen, J. and Mathijs, E. (1998) Theeconomics of agricultural decollectivization inEast Central Europe and the Former SovietUnion. Economic Development and CulturalChange, 47 (1): 1–26.

Temple (1998) Initial conditions, social capitaland growth in Africa. Journal of AfricanEconomies, 7 (3).

Tykkyläinen, M. (ed.) (1996) Local and RegionalDevelopment During the 1990s Transition inEastern Europe. Aldershot: Avebury.

Van de Walle, D. (1999) Assessing the Impact ofRural Road Projects. Washington DC: WorldBank, Development Research Group.

Visser, O. (2003) Farm Restructuring, PropertyRights, and Household strategies in Rural Russia.In: Transition, Institutions, and the Rural Sector.Spoor, M. (ed.). In series: Rural Economies inTransition. Lerman et al. (eds). Lanham,Maryland: Lexington Books.

Vogel, S. (1994) Structural changes inagriculture: Production linkages and agriculturaldemand-led industrialisation. Oxford EconomicPapers, 46: 136–156.

Wei, S-J. and Wu, Y. (2001) Globalization andinequality: Evidence from within China. NBERWorking Paper, No. W8611.

Whiteside (1998) When the Whole is More Thanthe Sum of The Parts: The Effect of Cross-BorderInteractions on Livelihood Security in SouthernMalawi and Northern Mozambique. Oxford:Oxfam UK.

Woldehanna, T. (2000) Economic Analysis andPolicy Implications of Farm and Off-farmEmployment: A Case Study in the Tigray Regionof Northern Ethiopia. Ph.D. thesis, WageningenUniversity, the Netherlands.

Woodon, Q. T. (1999) Growth, poverty, andinequality: a regional panel for Bangladesh.World Bank Policy Research Working Paper, No.2072. Washington DC: World Bank.

World Bank (2000a) The Rural Non-FarmEconomy: Report on Presentations andDiscussions at World Bank.http://www.worldbank.org/research/rural/workshop/May2000/RNFEmay.pdf

World Bank (2000b) Rural-Urban Linkages andInteractions: Synthesis of Issues, Conclusions,and Priority Emerging from the March 9Workshop.http://www.worldbank.org/research/rural/workshop.htm

37

The Development of the

Rural Non-Farm Economy

in Developing Countries

and Transition Economies:

Key Emerging and Conceptual Issues

Junior R. Davis and Dirk J. Bezemer

Enterprise development, trade, finance, and empowerment are central to the improvement of

people’s livelihoods in developing and transition countries. The Enterprise, Trade and Finance

Group together with the Livelihoods and Institutions Group at the Natural Resources Institute

apply practical solutions to rural and economic development through research, consultancy and

training activities. Our teams have a substantial experience in agricultural and development

economics, marketing and market research, commodity and international trade, financial

service development, and social and institutional development. Core areas of expertise include:

• Microfinance, Enterprise Development and Poverty Reduction

• Ethical Trade and Corporate Social Responsibility

• Improving the Performance of Agricultural Markets

• International Marketing and Trade

• Community Participation in Policy Formulation

• Social Impact Assessment

• Institutional Analysis and Capacity Building

The Institute is also able to arrange the following services: programmed study tours; training

courses; seminars; communication for development; and CD Rom design and production.

For further information please contact:

Enterprise, Trade and Finance Group

Natural Resources Institute

Chatham Maritime

Kent, United Kingdom

ME4 4TB

Email: [email protected]

Internet: http://www.nri.org/rnfe/

Tel: + 44 1634 883199

Fax: + 44 1634 883706

ISBN: 0 85954 557 1

Documents

The Development of the Rural Non-Farm Economy in ...projects.nri.org/rnfe/pub/papers/keyissues.pdf · Developing Countries and Transition Economies: Key Emerging and Conceptual Issues