Rural Livelihood Diversification in Sub-Saharan Africa: A Literature Review

Abstract

This article provides a comprehensive review of the literature on the nature and evolution of rural livelihood diversification in sub-Saharan Africa, and the situation regarding smallholders. It reveals mixed findings about the causes and consequences of livelihood diversification on rural smallholders adopting this strategy. A lot of evidence from the literature suggests that it is relatively better-off smallholders with sufficient assets who achieve successful livelihood diversification, mainly by exploiting opportunities and synergies between farm and nonfarm activities. Because of asset constraints, increase in incomes and wealth based on livelihood diversification has not yet benefitted the large majority of smallholders.

View correction statement:

Corrigendum

1. Introduction

Livelihood diversification has received much attention from researchers and policy-makers in the past decades, with high hopes that promoting it can offer a pathway for poverty reduction and economic growth in sub-Saharan Africa (SSA) (World Bank, 2007). The term ‘diversification’ refers to processes taking place at different levels of the economy, which are usually, but not always directly linked (Start, 2001). Firstly, ‘diversification of the rural economy’ refers to a sectoral shift of rural activities away from farm to non-farm activities, associated with the expansion of the rural non-farm economy (Start, 2001); normally as part of a broader process of structural transformation (Timmer, 2009). Secondly, ‘individual or household diversification’ refers to income strategies of rural individuals or households in which they increase their number of activities, regardless of the sector or location. Livelihood diversification is an active social process of individual or household diversification, involving the maintenance and continuous adaptation of a highly diverse portfolio of activities over time in order to secure survival and improve standards of living (Ellis, 2000b). The components of rural livelihood diversification are commonly classified by sector (farm or non-farm), by function (wage employment or self-employment) or by location (on-farm or off-farm) as summarised in Table 1.

Table 1. Classification of the components of rural livelihood diversification

Classification	Category	Definition
By sector	Farm (agricultural)	Production of unprocessed crops, livestock, forest, or fish products from natural resources. This category also includes farm wage labour, sale of farm output and consumption-in-kind of own farm output.
	Nonfarm (non-agricultural)	Includes all income-generating activities other than the production of primary agricultural commodities. Examples include mining, manufacturing, utilities, construction, commerce, transport, government services, among others. It also includes agro-processing, transport or trading of unprocessed crop, livestock, forest and fish products.
By function	Wage employment	Involves an employer–employee relationship, where the employee sells labour services to the employer in exchange for a wage or salary.
	Self-employment	Involves the sale of labour services to oneself, rather than working for an employer that pays a wage or salary. Income is earned through activities operated directly by the owner.
By location	On-farm	Income activity takes place on the farm, for example, crop and livestock production, hunting, fishing or gathering from natural resources.
	Off-farm	Income activity takes place away from the farm (rural or urban, local or foreign). Typically includes all in the nonfarm category. Also refers to wage or exchange labour on other farms, and labour payments in kind such as harvest sharing and other non-wage labour contracts.

Notes: The classification by sector follows standard national accounting systems. Classification by function depends on how labour is compensated, while that by location depends on where the activity takes place.

Source: Adapted from Barrett et al. (2001); Ellis (1998); Haggblade et al. (2010); Losch et al. (2012).

In SSA, many rural smallholder farmers have increasingly diversified their livelihoods through nonfarm activities and migration (Barrett, Reardon, & Webb, 2001; Losch, Freguin-Gresh, & White, 2012; Reardon, 1997). These diversified livelihoods are facilitated by infrastructural development, emergence of rural towns and improving accessibility to urban areas (Losch, Magrin, & Imbernon, 2013). Whether diversification will provide impetus for improving standards of living in SSA is still a subject of much debate, however. A contrasting perspective views livelihood diversification in rural SSA as a long-term deagrarianisation process of adjustment and reorientation of livelihoods in distress; one in which smallholders are invariably moving away from farming (Bryceson, 1999, 2002). Since most empirical studies on this subject in SSA have been based on cross-sectional data, the medium- to long-term impacts of livelihood diversification on smallholders and its links to the process of structural transformation have not yet been well understood.

This review article broadly examines recent empirical studies on SSA relating to the nature and evolution of rural livelihood diversification, its causes and consequences for rural smallholders, and the overall process of structural transformation. The aim is not to be exhaustive, but to point at some issues for reflection and for further research. The next five sections examine various literature on the subject and conclude with a discussion of main issues arising from the review.

2. Diversification of the Smallholder Rural Economy in Sub-Saharan Africa

Historical lessons from structural transformation in Europe and North America indicate that rising agricultural productivity, together with industrialisation and urbanisation, has been the stimuli for economic development (Timmer, 2009). In Asia, agricultural transformation occurred through the Green Revolution in which productivity was raised by growing high-yielding grain varieties – a process which was driven by the state, mediated by markets and based on smallholders (Djurfeldt, Holmen, Jirstrom, & Larsson, 2005). The structural transformation process at the macro level was characterised by a declining share of agriculture in GDP and employment, rural–urban migration leading to urbanisation, the development of a modern industrial and service economy, and a demographic transition (Winters, Essam, Zezza, Davis, & Carletto, 2010). Although agriculture became less important relative to other sectors, it continued to grow in absolute terms (Timmer, 2009). At the micro level, rural household participation in farm activities declined relative to nonfarm activities (Winters et al., 2010). In the early stages of the process, most rural households were subsistence farmers who produced most of the farm and nonfarm goods and services they required (Timmer, 2009). Because agriculture was mainly for subsistence, trade and commerce remained marginal. With better functioning markets and improved transport and communications infrastructure in rural areas, farm households diversified to include nonfarm activities as a way to increase their incomes. In the later stages, with rising incomes and higher standards of living, they either specialised in farming on larger consolidated farms or moved into high-return nonfarm sectors (Timmer, 2009).

Evidence suggests that SSA deviates in many ways from this expected path of structural transformation and economic development. Firstly, instead of farms becoming consolidated as it happened in Europe and North America, farm sizes in SSA are generally becoming smaller (Andersson Djurfeldt & Jirström, 2013; Jayne et al., 2003; Jirström, Andersson, & Djurfeldt, 2010). Recent studies on land issues in SSA (Headey & Jayne, 2014; Jayne, Chamberlin, & Headey, 2014; Muyanga & Jayne, 2014) have mainly attributed the declining farm sizes especially in land constrained areas in SSA to high population growth resulting from high fertility rates. According to these studies, while rural populations in Asia and Latin America are expected to decline by 2050, in SSA they are expected to increase further. The already declining farm sizes coupled with the high population growth could have a potentially negative impact on rural welfare and food security in SSA. The increasing population density has already encouraged more intensive use of land in high density areas of SSA, albeit in the absence of modern input use (fertiliser or irrigation), indicating unsustainable intensification. Increase in food production in SSA has so far been mainly based on the expansion of cultivated areas (Jirström et al., 2010; World Bank, 2013), which is now limited by declining farm sizes and the expansion of urban areas (Andersson Djurfeldt, 2015; Losch et al., 2012). Shrinking farm sizes and growing landlessness are by default pushing unskilled farm labour into mainly low-return nonfarm sectors (Haggblade, Hazell, & Reardon, 2007; Headey & Jayne, 2014).

Secondly, urbanisation in SSA is taking place without industrialisation (Andersson Djurfeldt, 2015; Losch et al., 2012), in contrast to green revolution Asia where urbanisation and emerging industries gradually allowed rural people to leave agriculture and enter nonfarm employment (Haggblade et al., 2007), and rewarded investments in education and migration (Jayne et al., 2014). In the absence of manufacturing industries and high-return service sectors to provide skilled nonfarm opportunities, prospects for increased employment and rising incomes in urban areas of SSA remain limited. This leaves smallholder farming as the primary option for gainful employment for SSA’s growing young labour force (Losch et al., 2012). However, rapid growth in nonfarm sectors fuelled by improvements in education and infrastructure can potentially alter this situation (Haggblade, Hazell, & Reardon, 2010).

Thirdly, persistent low agricultural productivity coupled with chronic food insecurity and severe poverty characterises the smallholder rural economy in SSA (Reardon & Timmer, 2007). As opposed to green-revolution Asia where modern inputs such as fertiliser and irrigation were important in raising agricultural productivity (Djurfeldt et al., 2005), in SSA low agricultural productivity is mainly linked to low fertiliser use, low responsiveness to fertiliser use due to overexploitation of land leading to nutrient mining and loss of organic matter, low use of irrigation, insecure land tenure, environmental degradation and underinvestment in crop research (Dethier & Effenberger, 2012; Headey & Jayne, 2014; Tittonell & Giller, 2013). Therefore, poverty gaps are increasing, with yield gaps resulting from such factors, particularly in regions with low agricultural potential (Dzanku, Jirström, & Marstorp, 2015). As a consequence of poverty and food insecurity, a large proportion of smallholders remain deeply engaged in subsistence staple crop production, but at the same time seasonally rely on the market for their staple food needs (Jirström et al., 2010; Losch et al., 2012). However, panel studies following agricultural transformation in nine SSA countries between 2002 and 2010 (Djurfeldt, Aryeetey, & Isinika, 2011; Djurfeldt et al., 2005; Djurfeldt, Larsson, Holmquist, Jirström, & Andersson, 2008), attribute increased agricultural productivity among smallholders in some regions to participation in agricultural markets and the nonfarm sector, and to the use of modern inputs and technology. Amidst the new opportunities and threats for smallholders linked to market liberalisation and globalisation (Reardon & Timmer, 2007), there is hope that with more public expenditure on infrastructure, modern technologies, promoting agricultural marketing and agribusiness, and pro-poor nonfarm growth, smallholder agriculture in SSA might be transformed (Haggblade et al., 2007).

3. Smallholder Livelihood Diversification in Sub-Saharan Africa

Recent studies indicate that asset, activity and income diversification characterise the livelihood strategies of rural smallholders in SSA (Barrett et al., 2001; Ellis, 2000b). Incomes from nonfarm sources have grown in importance; accounting for about 35 per cent of rural household incomes in SSA and 50 per cent in Asia and Latin America (Haggblade et al., 2010). Diversification at household level is viewed as an outcome of dynamic livelihood adaptation to various constraints and opportunities faced by smallholders (Ellis, 2000b). Diversification is therefore associated with both livelihood survival and distress under deteriorating conditions, as well as with livelihood security under improving economic conditions (Niehof, 2004). It is aimed at securing better living standards by reducing risk, vulnerability and poverty, increasing income, enhancing security and increasing wealth (Yaro, 2006). In order to use livelihood diversification to secure better living standards, rural households have to be able to generate cash, build assets and diversify across farm and nonfarm activities (Ellis & Freeman, 2004). It is a cumulative process that requires investment in improved farm practices or in nonfarm assets, or a combination of both, according to the options available for risk reduction and income generation. Where there are no feasible opportunities to diversify income activities, migration and remittances between rural and urban areas may be important in sustaining rural livelihoods (World Bank, 2007). There is substantial evidence showing that some rural households are sustained by multi-spatial livelihood activities (Andersson Djurfeldt, 2014; Ellis, 2000a; Losch et al., 2012) or food transfers (Andersson, 2011; Andersson Djurfeldt, 2012; Andersson Djurfeldt & Wambugu, 2011). Agricultural entrepreneurship, a vitalised rural labour market and migration are thus often complementary (World Bank, 2007). While farm income may provide capital for rural nonfarm employment and migration, nonfarm income plays a key role in strengthening the potential of smallholder farming as a pathway out of poverty.

Given the prevalence of risk in the rural SSA smallholder context, diversification may often be a strategy for survival or coping with risk, especially where agriculture fails to offer sufficient means of livelihood (Bryceson, 2002; Larsson, 2005; Reardon, 1997). In situations of high-risk agriculture and poverty, poorer smallholders without the necessary assets may be pushed to seek alternative incomes by engaging in low-return and sometimes risky nonfarm activities (Barrett, Bezuneh, Clay, & Reardon, 2001). However, it is mainly among richer households or in regions with favourable agricultural conditions that livelihood diversification driven by motives to raise incomes or accumulate wealth prevails (Haggblade et al., 2007). Although diversification is a common livelihood strategy, not all households enjoy equal access to high-return opportunities (Barrett et al., 2001; Lay, Mahmoud, & M’Mukaria, 2008), and for many rural households there are limited possibilities for remunerative nonfarm work (Jayne, Mather, & Mghenyi, 2010; Jirström et al., 2010; Otsuka & Yamano, 2006). The constraints and opportunities are unevenly distributed socially and geographically, and households with better asset endowments are more likely to access better opportunities for diversification (Barrett, Bezuneh, & Aboud, 2001; Barrett et al., 2001). The usual pattern is for the range of activities that can lead to increase in incomes and wealth to rise with income level (Ellis, 1999; Oya, 2007), and for such activities to be more common in areas with favourable agro-ecology and good market access (Losch et al., 2012; Reardon, 1997). Even in rural areas with favourable endowments or opportunities, some households are better off in terms of welfare, while others remain trapped in structural poverty (Losch et al., 2012).

4. Empirical Approaches to Studying Livelihood Diversification

Two main approaches are commonly used in the economic literature to study livelihood diversification behaviour: ‘the household economic model’ (Singh, Squire, & Strauss, 1986; Taylor & Adelman, 2003) and ‘the livelihood approach’ (Ashley & Carney, 1999; Chambers & Conway, 1992; Scoones, 2009). The household economic model considers farm households as production units that maximise utility by combining time and other inputs to produce output, subject to price and resource constraints (Becker, 1965). Diversification is seen as a function of returns to labour from farm activities compared to off-farm activities (Singh et al., 1986). Given an asset base, the farm household makes choices by comparing the returns from farm labour time and time spent on off-farm activities (Yaro, 2006). The assumption is that increases in off-farm incomes provide incentives for farm households to diversify their activities. In the SSA context, the household model has been used to investigate household production and off-farm labour allocation decisions (Barrett et al., 2001; Reardon, 1997; Reardon, Delgado, & Matlon, 1992), farm/nonfarm interactions (Davis, Winters, Reardon, & Stamoulis, 2009; Haggblade, Hazell, & Brown, 1989), participation, patterns and drivers of diversification at household level (Abdulai & CroleRees, 2001; Barrett et al., 2001, 2001; Bezu & Barrett, 2012; Bezu, Barrett, & Holden, 2012; Canagarajah, Newman, & Bhattamishra, 2001; Lay et al., 2008; Lay, Narloch, & Mahmoud, 2009; Winters et al., 2009, 2010).

The household models have been criticised for not taking the inter-temporal dimensions of livelihoods into account and for failing to capture survival strategies of livelihoods under stress (Ellis, 2000a, 2000b). They are also criticised for not considering the social relationships between household members, which in many cases have strong influence on household choices (Ellis, 1998). Furthermore they simplify reality by assuming that incomes and preferences are shared between household members (Taylor & Adelman, 2003). In reality, division of responsibilities and tasks between men and women in the household affects their production decisions and income distribution (Ellis, 1993). The models further assume that markets are perfectly functioning; whereas in developing countries, households are frequently exposed to incomplete or imperfect markets that limit their choices and thus affect their behaviour (De Janvry & Sadoulet, 2006; Ellis, 1993).

The livelihood approach, on the other hand, takes a more people-centred view on the study of rural livelihoods in different contexts, even under stress. The approach has been widely used in empirical studies of livelihood strategies and adaptation (Ellis, 2000a; Orr & Mwale, 2001; Yaro, 2006), livelihoods, risk and poverty (Ansoms & McKay, 2010; Bebbington, 1999; Bird & Shepherd, 2003; Ellis & Bahiigwa, 2003), and livelihood diversification (Ellis, 2000a, 2000b; Smith, Gordon, Meadows, & Zwick, 2001). The livelihood approach has also been adopted by many development and non-government organisations (NGOs) as a tool for monitoring livelihoods and their transformation (Ashley & Carney, 1999). The approach commonly employs the ‘sustainable livelihoods framework’ (SLF) to assess people’s livelihood assets and how the external environment of social relations, institutions, organisations, policies, seasonality, trends and shocks modify access to and ability to convert livelihood assets into livelihood outcomes (Ansoms & McKay, 2010; Vedeld, Jumane, Wapalila, & Songorwa, 2012). The approach has its strength in recognising the multiple and diverse character of livelihoods (Ellis, 1998, 2000a; Ellis & Biggs, 2001) and has proved useful in examining the diversity of farming systems (Sourisseau et al., 2012). Furthermore it accounts for the influence of institutions on livelihoods (Ellis & Freeman, 2004) and the social and economic character of livelihood strategies (Ellis, 2000b). The SLF has also been used to understand the costs and benefits of different livelihood decisions and strategies (Ashley & Carney, 1999). At the same time, it has been criticised because many of its components are difficult to measure and often require the use of proxy indicators, which are sometimes difficult to find. The approach also fails to account for prices and wages, which is necessary when comparing the costs and benefits of different livelihood outcomes (Barrett & Reardon, 2000).

Most of the studies on livelihood diversification in SSA using the above analytical approaches have been based on cross-sectional data from individual countries or from sample regions within countries. In some cases, studies compare two or more country or regional situations (Barrett et al., 2001; Canagarajah et al., 2001; Dercon & Krishnan, 1996; Losch et al., 2012; Winters et al., 2009, 2010), and few studies have used panel data from one or more countries to add a time dimension to their analyses (Abdulai & CroleRees, 2001; Bezu & Barrett, 2012; Bezu et al., 2012; Block & Webb, 2001; Dercon, 2004; Djurfeldt et al., 2011; Kijima, Matsumoto, & Yamano, 2006; Lay et al., 2009; Porter, 2012). Despite the need for empirical evidence from panel data to capture changes over time, there is a lack of financial and skilled human resources in SSA to collect and analyse data of sufficient quality and scope to inform policy. Where panel surveys depend on irregular financing by donors, it becomes difficult to plan ahead, with negative repercussions for the collection of panel data (Carletto, Jolliffe, & Banerjee, 2013). Hence, the wide heterogeneity of the rural economy and funding constraints have limited most empirical studies to one-time shots, with limited scope for making comparisons and generalisations.

5. Patterns and Determinants of Diversification in Sub-Saharan Africa

Individuals and households may diversify their assets, incomes and activities in response to incentives that may be classified as push and pull factors (Ellis, 2000b; Reardon, Berdegué, Barrett, & Stamoulis, 2006). However, the processes and outcomes of push and pull factors are different in dynamic and in marginalised or stagnant regions (Haggblade et al., 2007).

5.1 Push Factors

Push factors are negative factors that may force farm households to seek additional livelihood activities within or outside the farm. Push factors tend to dominate in high-risk and low-potential agricultural environments, subject to drought, flooding and environmental degradation (Haggblade et al., 2007). When agricultural activities are seasonal and environments are full of uncertainty, like in many parts of SSA, rural households tend to reduce risk by diversifying into activities with lower covariate risk in order to make consumption and incomes less volatile (Barrett et al., 2001; Dercon, 2002; Ellis, 2000b; Matlon, 1991). The most common push factors are related to different forms of risk, such as seasonality and climatic uncertainty (Ellis, 1998, 2000b). Others include land constraints driven by population pressure and fragmented land holdings, missing or incomplete factor markets, and market access problems due to poor infrastructure and high transaction costs (Barrett et al., 2001).

Diversification may be used as a strategy for coping or risk management (Dercon, 2002; Ellis, 1998; Matlon, 1991; Start & Johnson, 2004). Risk management is an ex-ante deliberate strategy where a household anticipates failures in their income streams and thereby maintain a range of income activities to safeguard against it, while coping is a response to disaster or unanticipated failure in major sources of survival. In SSA, the general lack of social insurance or safety nets from government transfers, NGOs, community or family members may push households into diversification for risk management (Barrett et al., 2001).

Regarding seasonality, many nonfarm income activities tend to peak during the dry seasons when there is a decline in farm activities (Reardon, 1997). During the dry season, especially in semi-arid regions, some rural households depend on incomes from selling farm products and from nonfarm activities, including migration remittances (Ellis, 1998; Losch et al., 2012; Reardon, 1997). This is the case in the Sahelian agricultural systems, where farmers turn to nonfarm sources to supplement farm incomes when harvests fail (Bryceson, 2002; Grawert, 1998). Diversification is also driven by differences in relative returns in different agro-climatic zones (Reardon, 1997).

Social factors such as social positions, networks, associations, religion and culture are important drivers of diversification (Ellis, 1998). Labour market opportunities may be restricted by gender, class or social inequalities (Oya, 2007; Start & Johnson, 2004). In terms of gender, rural women are often constrained in accessing land and other productive assets (Gladwin, Thomson, Peterson, & Anderson, 2001). Therefore, they often adopt multiple livelihood strategies (Andersson Djurfeldt, Djurfeldt, & Lodin, 2013). However, nonfarm income may contribute more to inequality among female-headed households, where self-employment is important and nonfarm opportunities more constrained (Canagarajah et al., 2001). Institutional factors also play a significant role in creating opportunities or constraints to the improvement of rural livelihoods. In some regions, institutional factors such as regressive tax systems at local level tend to discourage rather than foster livelihood diversification (Ellis & Freeman, 2004).

5.2 Pull Factors

Pull factors are positive and these may attract farm households to pursue additional livelihood activities to improve their living standards. These factors provide incentives for people to expand their range of income activities outside farming by increasing the returns from nonfarm activities. Such factors tend to dominate in less risky, more dynamic agricultural environments (Haggblade et al., 2007). Diversification becomes a deliberate strategy for an individual or household in order to generate assets for accumulation and reinvestment (Ellis, 1998, 2000b). Pull factors include the commercialisation of agriculture and the emergence of improved nonfarm labour market opportunities linked to better market access, improved infrastructure, and proximity to urban areas (Losch et al., 2012; Reardon et al., 2006; Winters et al., 2009). Other pull drivers of diversification are supply factors, such as improved technology, expansion of education, increased demand for non-food goods and services driven by higher per capita incomes (Reardon, 1997).

5.3 Survival-led or Opportunity-led Diversification

Diversification resulting from push or pull factors have been categorised as either ‘survival-led’ or ‘opportunity-led’ respectively (Ellis, 2000b; Lay et al., 2008; Reardon et al., 2006). Survival-led diversification, mainly driven by push factors, occurs when poorer rural households engage in low-return nonfarm activities by necessity to ensure survival, to reduce vulnerability or to avoid falling deeper into poverty. They are pushed towards diversifying their income sources to manage risks or cope with shocks, such as declines or stagnation in agriculture, differentiated labour markets, credit market imperfections, demographic pressures and land constraints (Barrett et al., 2001; Lay et al., 2008; Reardon et al., 2006). They are pushed into low-return nonfarm activities because they have low endowments of assets such as land, capital, livestock and credit, making them more vulnerable to seasonal and other risk factors (Barrett et al., 2001; Ellis, 1998; Lay et al., 2008; Reardon & Taylor, 1996). Many poor households also tend to lack formal education and skills, which act as entry barriers preventing them from engaging into high-return activities like nonfarm waged and skilled employment (Abdulai & CroleRees, 2001; Barrett et al., 2001; Ellis, 1998; Reardon, 1997). The poor are confined to low-income, labour-intensive nonfarm activities that leave them trapped in structural poverty, while richer households tend to specialise in high-return farm or nonfarm activities (Haggblade, Hazell, & Reardon, 2005; Losch et al., 2012). The poor tend to be food insecure all year round, and depend on selling their labour or on safety net supports (Ellis & Freeman, 2004). Sometimes they are unable to sustain their subsistence needs and may be forced to engage in activities with returns below those in the agricultural sector (Lay et al., 2008).

Opportunity-led diversification is mainly driven by pull factors. It occurs when wealthier rural households engage in high-return nonfarm activities, with accumulation objectives, in order to increase household income by maximising returns from their assets. They are able to diversify their income activities in more favourable labour markets or take advantage of off-farm opportunities created by technological advances, new market possibilities, proximity to urban centres or improved infrastructure (Lay et al., 2008; Losch et al., 2012). High returns to nonfarm activities may emerge from increased demand for nonfarm goods and services or off-farm opportunities created by growth motors in different rural sectors such as agriculture, mining or tourism (Reardon et al., 2006). Better-off households are those with high endowments of assets such as land, livestock and buildings (Ellis & Freeman, 2004), and are more likely to engage in diverse high-return nonfarm activities, some of which have similar or higher returns than farming (Barrett et al., 2001; Lay et al., 2008). In this way some better-off households are capable of accumulating capital by combining commercial farming and nonfarm activities while still relying more on commercial agriculture (Andersson Djurfeldt, 2013; Barrett et al., 2001; Ellis & Freeman, 2004; Oya, 2007).

6. The Welfare Impacts of Diversification in Sub-Saharan Africa

The literature on diversification in rural Africa generally shows a positive relationship between nonfarm income and household welfare indicators such as income, wealth (estimated through size of land holdings or livestock), consumption and nutrition (Barrett et al., 2001; Ellis, 1998, 2005; FAO, 1998; Reardon, 1997). Panel and longitudinal data evidence from Ethiopia suggest that bigger nonfarm income results in a more rapid growth in income and consumption, especially among wealthier farm households (Bezu et al., 2012; Block & Webb, 2001). A reason for this is that substantial entry barriers (Abdulai & CroleRees, 2001; Barrett et al., 2001; Davis et al., 2009) limit access to high-return rural nonfarm income to relatively better-off households, while the poor are mainly confined to low-return activities (Barrett et al., 2001; Bezu et al., 2012). High-return nonfarm opportunities are often found in formal sector employment and activities which are skilled, capitalised or protected from competition, while the low-return opportunities generally have little requirement for skill or capital, for example, unskilled factory or porter jobs, traditional cottage activities, and micro-enterprise like petty-trade, handicrafts, sand mining, brick making, burning charcoal or collecting firewood (Start & Johnson, 2004).

A number of studies also find that nonfarm income diversification has a positive impact on farm productivity and food security. For instance, in Burkina Faso, some households that lacked credit used nonfarm incomes to invest in farm assets such as animal traction (Savadogo, Reardon, & Pietola, 1998). In Senegal, nonfarm incomes enabled some households to access farm inputs like groundnut seeds, fertilisers and livestock (Kelley, Diagana, Reardon, Gaye, & Crawford, 1996). In Tanzania and Ethiopia, Dercon and Krishnan (1996) found that households engaged in off-farm activities with high entry barriers such as trade or business, had higher levels of assets, income and consumption. Ellis and Mdoe (2003) found that in Tanzania richer households tended to diversify into high-return nonfarm activities and had higher agricultural productivity compared to the poor households. Whilst in Ethiopia, farm households with more diversified income sources had higher agricultural productivity and that off-farm income was complementary to farm income when farm households lacked credit (Woldehanna, 2000). Evidence from Kenya shows that involvement in high-return nonfarm activities such as salaried employment has positive effects on agricultural productivity (Lay et al., 2008; Marenya, Oluoch-Kosura, Place, & Barrett, 2003). In Western Kenya, Andersson Djurfeldt (2012) finds that wealthier farm households with access to nonfarm incomes were able to profit from seasonality of agricultural markets through trade-based or barter exchanges for agricultural produce. In contrast, poorer farm households that lacked nonfarm incomes were more vulnerable and their food security was worsened by seasonal changes in food prices and in the agricultural production cycle.

The overall effect of nonfarm activities on rural income distribution in SSA generally remains mixed (Barrett et al., 2001; FAO, 1998, Haggblade et al., 2005; Reardon, 1997; Reardon & Taylor, 1996; Reardon, Taylor, Stamoulis, Lanjouw, & Balisacan, 2000). In some cases, nonfarm activities reduce overall income inequality (Adams, 2002; Van Den Berg & Kumbi, 2006), while in others they tend to increase inequality (Block & Webb, 2001; Canagarajah et al., 2001; Reardon & Taylor, 1996). When relatively poor households are able to engage in nonfarm activities, it reduces total income inequality, if incomes are large enough and accessible to the poor (Reardon & Taylor, 1996; Van Den Berg & Kumbi, 2006). Where high-return nonfarm activities are unequally distributed in favour of relatively richer households, it tends to reinforce total income inequality, even when incomes are generally increasing across income strata (Canagarajah et al., 2001; FAO, 1998; Reardon & Taylor, 1996). There are differences in the nature and returns to labour in different nonfarm activities undertaken by rural SSA households according to their income strata, due to the presence of asset entry barriers (Lay et al., 2009; Reardon & Taylor, 1996; Woldenhanna & Oskam, 2001). However, it seems that households with less diversified income sources struggle hard to diversify more over time (Barrett et al., 2001). In Ethiopia, panel evidence (Bezu & Barrett, 2012; Bezu et al., 2012) shows that poor households who were able to accumulate capital through low-return nonfarm activities could subsequently access high-return nonfarm activities. In other words, participation in the rural nonfarm economy provided a pathway for upward mobility. This suggests that even if opportunity-led diversification in SSA is biased in favour of the wealthier households, survival-led diversification has more potential than just being an important safety net for poorer households.

The effect of nonfarm activities on income inequality is commonly analysed by considering the relationship between diversification (share of nonfarm income in total household income or absolute level of nonfarm income), and total household income (or the size of landholdings). There is generally conflicting empirical evidence on the patterns and on whether nonfarm income contributes more to the income of the relatively poor or richer rural households; with an apparent contradiction in which several patterns of the relationships emerge in different regions (FAO, 1998; Losch et al., 2012; Reardon, 1997; Reardon & Taylor, 1996; Reardon et al., 2000). Roughly five main patterns of the relationship between diversification and total household income emerge from this literature: strongly negative and linear; strongly positive and linear; the U-shaped pattern; the inverted U-shaped pattern; or otherwise with no clear relationship. The patterns depend on whether diversification is measured using the share of nonfarm income in total household income or absolute level of nonfarm income. There is evidence in many cases, that the ratio of the absolute levels of nonfarm incomes between the highest and lowest income strata is much higher than the ratio of the shares (FAO, 1998). In many cases there is also a high correlation between total household income and the size of landholdings (Reardon et al., 2000).

In the strongly negative and linear pattern, the share of nonfarm income declines sharply as total household income increases, following the conventional wisdom. This means that the relatively poor households are highly diversified compared to the relatively rich households. For the strongly positive and linear pattern, the share of nonfarm income increases sharply as total household income increases, contradicting the conventional wisdom. The relatively rich households are highly diversified as opposed to the relatively poor households. In general, a positive pattern of the relationship between diversification and total household income or size of landholdings is reported in much of Africa, while the negative pattern is reported mostly in Latin America, and mixed patterns found in Asia (FAO, 1998; Reardon, 1997; Reardon et al., 2000). This pattern in SSA is attributed to high entry barriers to nonfarm opportunities for the poor, because farming is mainly subsistence, land distribution is relatively equal, and infrastructure, rural town economies and capital markets are relatively undeveloped; hence, the scarcity of labour intensive activities with low entry barriers and prevalence of high entry barriers in capital intensive activities (FAO, 1998; Reardon & Taylor, 1996; Reardon et al., 2000).

While the U-shaped pattern means that both the relatively poor and the relatively rich households have a higher share of nonfarm income (highly diversified), while the middle-income households are less diversified. Although the poorest households have higher shares of nonfarm income compared to the middle-income households, their absolute level of nonfarm income is considerably lower. Asset-poor households may spend a large share of their time on nonfarm activities but receive low returns, while richer households with more assets may spend the same or less time on nonfarm activities and get higher returns (FAO, 1998). The U-shaped pattern has been found most frequently in Asia and Latin America (less in Africa) because there is greater availability of labour intensive activities with low entry barriers for the poor, and richer households with more assets are able to diversify into capital intensive activities (FAO, 1998; Reardon & Taylor, 1996; Reardon et al., 2000).

In the inverted U-shaped pattern, the middle-income households have a higher share of nonfarm income compared to the relatively poor and the relatively rich households with a lower share of nonfarm income (FAO, 1998; Losch et al., 2012; Reardon et al., 2000). A comparison of diversification and rural change at household level was done across seven countries in Africa and Latin America at different stages of structural change (Losch et al., 2012): The findings show a strong positive relationship between income and the process of structural change towards a more diversified rural economy in these countries, which include four from SSA (Kenya, Senegal, Mali and Madagascar). At very low income levels rural households focused on survival strategies, while as incomes grew they began to diversify their activities in order to cope with risk and find additional incomes. At higher income levels households started to specialise into farm or off-farm activities. In the SSA countries, most households seemed to be trapped in structural poverty and were neither able to earn sufficient income through diversification to become secure in their livelihoods nor able to reach the point of specialisation. This inverted U-shaped pattern was mainly attributed to poverty and prevalence of high entry barriers in the nonfarm sector.

7. Conclusion

This article discusses some recent studies on structural and agricultural transformation, and rural livelihood diversification in SSA, with a special focus on the situation of smallholders. The literature review reveals some important issues for reflection and further research: Firstly, because of persistent low agricultural productivity and declining farm sizes coupled with rising population, SSA’s structural and agricultural transformation appears to move very slowly. In addition, the transformation path clearly differs from the one taken by developed economies in Europe, America or Asia, where urbanisation and industrialisation accompanied the rural transformations. Although this leaves farming as the main employment option for the majority, there is an important role for the nonfarm sector in providing employment for those smallholders that are forced to straddle between farm and nonfarm activities or to completely exit farming.

Secondly, it is clear that rural farm and nonfarm livelihood diversification is of increasing importance for economic growth, poverty reduction, food security and creation of employment. Evidence from studies in rural SSA indicates their positive welfare impacts on income, wealth, consumption, nutrition, agricultural productivity and food security. However, increases of income and accumulation of wealth as a result of livelihood diversification is not yet happening on a large enough scale to affect a majority of smallholders in rural SSA. The process is biased in favour of relatively better-off farmers with sufficient assets, while the poor tend to be hindered by entry barriers. The relatively better-off smallholders who exploit opportunities and synergies between farm and nonfarm activities are able to use livelihood diversification to expand their incomes and accumulate wealth. Thus, growth in the rural nonfarm economy in SSA is currently neither inclusive nor redistributive. Although the benefits of livelihood diversification mainly favour the better-off, it still provides a safety net for the rural poor and sometimes offers a means for upward mobility. There is therefore good reason for governments and development partners to promote livelihood diversification among smallholders in SSA, and to make it more inclusive through policies and programmes that lower entry barriers for the poor.

Thirdly, because of wide heterogeneity in the rural economy and of data limitations, the medium- to long-term impact of livelihood diversification on smallholders and their environments, and its role in the broader process of structural and agricultural transformation in SSA, remain to be fully understood. Most studies have so far been based on cross-sectional data rather than panel or longitudinal data. This suggests that more is revealed about rural diversity in different contexts and at different points in time than about livelihood diversification as a dynamic process. Thus, there is urgent need for more longitudinal research projects focusing on livelihood diversification and transformation in SSA, and for existing panel studies to get the financial support needed to continue. Panel studies in SSA can be encouraged by providing consistent funding to such survey efforts and technical support to build their analytical capacity.

Disclosure statement

No potential conflict of interest was reported by the author.

Acknowledgements

The author is highly indebted to Magnus Jirström, Mikael Hammarskjöld and Agnes Andersson Djurfeldt of the Department of Human Geography, Lund University (Sweden) for the major contributions they made to the final manuscript and earlier drafts. Many thanks to my thesis committee at CIRAD/Montpellier (France): Pierre-Marie Bosc, Bruno Losch, Michel Benoit-Cattin, Céline Bignebat, Jean-Michel Sourriseau, Jean-François Bélières for comments on the earlier versions. The author is also grateful to the reviewers at the Journal of Development Studies for their insightful comments that enriched this article. The article was written with support from CIRAD and the Afrint Project, Lund University. The views expressed here are those of the author.