Access to Credit in Eastern India

Access to appropriate credit reduces poverty and increases the income of agricultural households (Binswanger and Khandker 1995; Carter 1989; Carter and Wiebe 1990; Feder et al 1990; Pitt and Khandker 1996, 1998; Khandker and Faruqee 2003; Guirkinger 2008; Awotide et al 2015; Narayanan 2016; Luan and Bauer 2016; Kumar et al 2017).Credit helps farmers buy necessary inputs, such as seeds, fertilisers, pesticides, animal feed, and irrigation services, and invest in long-term productive assets for agriculture and livestock. However, many agricultural households have limited access to credit. Recognising the importance of the agricultural sector in the national economy, the Government of India (GoI) has undertaken a number of initiatives to strengthen the agricultural credit system. These include the nationalisation of commercial banks in 1969 and 1980; establishing regional rural banks (RRBs) in 1975 and the National Bank for Agricultural and Rural Development (NABARD) in 1982; setting up special agricultural credit plans (SACP) in 1994–95 and the Kisan Credit Card (KCC) Scheme in 1998–99; doubling the SACP within three years (2004); establishing the agricultural debt waiver and debt relief scheme in 2008, the interest subvention scheme in 2010–11, and the Pradhan Mantri Jan Dhan Yojana (PMJDY) in 2014; and extending KCC facilities to livestock and fish farmers in 2018. These initiatives have had a positive impact on the flow of agricultural credit (Ghosh 2005; Golait 2007; Kumar et al 2010; Mohan 2006; Hoda and Terway 2015; Kumar et al 2015), and the ratio of agricultural credit to agricultural gross domestic product (GDP) has increased from 10% in 1999–2000 to about 43% in 2016–17 (GoI 2018). However, about half of agricultural households still have no access to credit services (Kumar et al 2017). Limited access to credit squeezes investment in agriculture and other productive activities (Udry 1994). Lack of credit is regarded as one of the crucial reasons for poor households remaining poor (Collins et al 2009).

While inadequate access to credit is a major concern in India in general, the situation is worse in the eastern region of the country. Previous studies have identified inadequate access to credit as one of the primary impediments to agricultural development in eastern India (Joshi and Kumar 2017). However, the source of credit is equally important as some of it may be offered at an exploitative rate of interest. It is well-documented that the rural credit market in India is characterised by the coexistence of formal and informal credit agencies. Formal credit agencies include public and private banks, RRBs, post offices, and cooperative banks, while informal agencies comprise moneylenders, loans from friends or relatives, agricultural traders, and commission agents. Formal and informal sources have different implications for agricultural households’ welfare, but little empirical evidence has been derived from comparative analyses of the impacts of different sources of credit. Against this background, and with the help of a large field survey conducted in the eastern states of India during 2018, this study aims to contribute to the literature on the comparative impact of different sources of credit. The study focuses on two specific objectives: first, the factors associated with access to credit from different sources (formal and informal) are examined, and second, the impact of different sources of agricultural credit are analysed.

Data and Descriptive Analysis

This study uses observational data from a 2018 field survey of 1,940 agricultural households from Bihar, Jharkhand, and eastern Uttar Pradesh, in eastern India. Of these, 890 households were from Bihar (45.88%), 698 from eastern Uttar Pradesh (35.98%), and the remaining 352 were from Jharkhand (18.14%). The number of sample households in a state was allocated in proportion to the rural population in that state, with Bihar having the highest rural population among the surveyed states. We randomly selected 10 districts from Bihar, eight districts from eastern Uttar Pradesh, and four from Jharkhand (Figure 1 shows the location of selected districts). We then randomly selected two blocks from each district, and from each block, again, randomly selected two villages. Finally, we randomly selected 30 households to be surveyed from each village. The survey instruments collected information on resource endowments (household, agricultural, business, and financial) as well as on access to, and use of, a wide variety of formal and informal financial institutions, such as commercial banks, cooperatives, self-help groups (SHGs), microfinance institutions (MFIs), moneylenders, friends, and relatives. The data also provided detailed information on household demographics, education, and other characteristics. These data provided rich and detailed information about households and financial intermediaries and thus are particularly well-suited for our analysis. We now turn to a brief description of some of the salient features of the data.

Table 1 reports the distribution of sample farmers according to different credit sources (formal and informal). About 49% of agricultural households did not use credit (F₀I₀), and among those who did use it, the majority used credit from only one source. About 24% of agricultural households accessed credit from formal sources, while 21% accessed it from informal sources. Only 6% of agricultural households accessed credit from both formal and informal sources (F₁I₁).

The study sample consisted of 1,940 agricultural households, 954 (49%) of which had not accessed any credit during the previous year, while the remaining 986 households (51%) had done so. Among agricultural households that had accessed credit, the average amount that had been borrowed during the year previous to the survey was `17,448. Of this, 62% was borrowed from formal sources and 38% from informal sources (Table 2). Among formal sources, public sector commercial banks were the dominant players, providing 63.5% of the formal credit. They were followed by RRBs (13.7%), SHGs (10.2%), and MFIs (6.1%). Private sector commercial banks, private sector finance companies, and cooperatives provided the remaining 6.5% of formal credit to agricultural households in eastern India (Appendix A1, p 53). Moneylenders (56.5%) were the largest source of informal credit, while friends and relatives (who usually do not charge interest or charge lower interest rates) provided 41.3% of informal credits in the study area. Agricultural traders and commission agents accounted for a negligible share of the informal credit accessed in the eastern states of India. Interest rates charged by formal and informal sources showed a considerable variation. The average annual interest rates charged by formal and informal agencies were 12% and 27%, respectively. There was also a significant variation in the interest rates charged within the formal and informal sources, ranging from 5.8% charged by cooperatives to as high as 24% charged by private commercial banks and private finance companies. Interest rates charged by SHGs and MFIs hovered at around 20%. Among the informal sources, the highest annual interest rates were charged by moneylenders (37%), followed by commission agents (36%), and friends and relatives (7.8%).

In general, agricultural households used credit for multiple purposes, such as farming and non-farming investment, household consumption expenditures, education, medical treatment, and housing. The pattern of use of formal and informal credit differed significantly. About 70% of formal credit was used for farming activities, while only 28% of informal credit was used for farming. The highest proportion of informal credit (25%) was used for medical treatment (Table 3).

Table 4 presents the descriptive statistics of the key variables of interest. The average household size was about seven people and the average age of heads of households was about 52. Agricultural households had an average operational landholding of 0.96 hectares (ha). About 97% of households were headed by males, and the majority of respondents were literate (62%) and had about five years of education. Other Backward Castes (OBCs) accounted for 58% of the agricultural households, followed by Scheduled Castes (SCs) and Scheduled Tribes (STs) who constituted 21% of the sample households. The remaining 20.6% were from a variety of other castes. About 82% of households possessed ration cards and almost all agricultural households had bank accounts. About 51% of agricultural households had heard of the Pradhan Mantri Fasal Bima Yojana (PMFBY) but the use of the crop insurance was quite low (only 5%). Twenty-eight percent of the sample households received remittances. The awareness level of rural agricultural households was quite high, with about 76% being aware of the direct benefit transfer scheme, about 93% being aware of the Mahatma Gandhi National Rural Employment Guarantee Act (MGNREGA), about 83% being aware of the loan waiver scheme, and 7% being actively associated with a political party. The agricultural households had an annual per hectare farm income of `2,69,649, and yields of rice and wheat crops were 36.4 and 29.5 quintals per hectare (q/ha) respectively.

Table 4 further compares the means of selected variables across different categories of borrowing and non-borrowing households. The difference in household characteristics of these groups are of interest as they help explain the variations in access to credit. However, the results in Table 4 cannot be used to make inferences about the impact of credit on farm income and productivity without controlling for other confounding factors.

Methodology

The main aim of this study is to assess which sources are most effective in improving farmers’ welfare. To this end, we estimate the impact on the economic welfare of agricultural households of various choices of formal and informal credit.

Multinomial endogenous switching regression (MESR): We have taken income and agricultural productivity as indicators of the economic welfare of agricultural households. Income and productivity have been widely used as a proxy for household welfare in a number of previous studies (Kumar et al 2017; Li et al 2011; Arouri et al 2015; Wetterberg 2007). However, the identification of the cause and effect relationship between credit sources, and potential outcome indicators is complex due to an endogeneity bias, as we cannot observe the counterfactual. As mentioned earlier, farmers can avail credit from multiple sources and the selection of any source is based on the farmer’s expected net return subject to the constraints. Access to credit is therefore based on an individual’s choice and may be correlated with unobservable characteristics that would also affect his performance in farming. The precise estimation of impacts therefore necessitates controlling for both observable and unobservable characteristics through random selection of individuals or households for treatment. Several methods have been proposed and used to deal with such issues and are documented in the literature, ranging from instrumental variable methods to experimental and quasi-experimental methods. We employ an MESR framework to estimate the parameters. This framework has the advantage of evaluating individual as well as alternative combinations of practices. It also captures both self-selection bias and the interaction between choices of alternative practices (Mansur et al 2008; Wu and Babcock 1998). In the first stage, the impact of each combination of credit sources is modelled using a multinomial logit model, while recognising the interrelationships among the credit-source choices. In the second stage, the impacts on outcome variables of each combination of the credit sources are evaluated using ordinary least squares (OLS) regression with a selectivity correction term from the first stage. For identification, we use the distance of the bank from the village as an instrument variable. We checked the validity of the instrument and conducted an admissibility test (Di Falco at al 2011; Di Falco and Veronesi 2013) to confirm that this variable jointly affects credit sources and, thus, does not affect our outcome variables.

We assume that agricultural households aim to maximise their income and productivity (Y_i) by comparing the positive return provided by m alternative credit sources. The requirement for agricultural household i to choose credit source j over any alternative source m is that Y_ij> Y_im ≠ j, or equivalently, ΔY_im = Y_ij – Y_im> 0 m ≠ j. The expected outcome variable Y*_ij that the agricultural households derive from selection of credit source j is a latent variable determined by observed characteristics (X_i) and unobserved characteristics (Î_ij),

... (1)

where X_i is a vector of observed exogenous variables (socio­demographic characteristics, social and physical capital, social safety net, and location characteristics). Let l be an index that denotes the agricultural household’s selection of credit source, such that:

... (2)

where η_ij = max_{m ¹ j}(Y^*_im – Y^*_ij) <0 (Bourguignon et al 2007). Equation 2 implies that the i_th agricultural household will select credit source j to maximise the expected positive Y if credit source j provides greater expected positive Y than any other credit source m ¹ j, that is, if η_ij = max_{m ¹ j}(Y^*_ij – Y^*_im )> 0. Assuming that ϵ are identically and independently Gumbel distributed, the probability that agricultural household i with characteristics X_i will select credit source j can be specified by a multinomial logit model (McFadden 1973):

... (3)

The parameters of the latent variable model can be estimated by a maximum likelihood function. In the second stage of the MESR,¹ the relationship between the outcome variables and a set of exogenous variables Z (household characteristics such as age, education, social group, assets, and livestock) is estimated for the selected credit source. In our set of possible credit sources (Appendix Table A1), the base category which “does not borrow credit from any source” is denoted as j = 1. In the remaining set of possible credit sources (j = 2, 3, and 4), at least one credit source is selected by the agricultural household. The outcome equation for each possible regime j is given as:

... (4)

where Q_ij refers to the outcome variables of the i_th agricultural households in Regime j, and the error terms (μ) are distributed with and E(u_ij|X,Z) = 0 and var(u_ij|X,Z) = σ_j². Q_ij is observed if, and only if, credit source j is used, which occurs when . If ϵ and u are not independent, OLS estimates obtained from Equation (4) will be biased. A consistent estimation of α_j requires inclusion of the selection bias correction terms of the alternative credit source choices in Equation (4). The Durbin and McFadden (DM) model makes the following linearity assumption:

... (5)

with Σj_m = 1r_j = 0 (by construction, the correlation between the u‘s and ϵ‘s sums to zero).

Using this assumption, the equation of the MESR in Equation (7) is specified as:

... (6)

where σ_j is the covariance between the u‘s and ϵ‘s. Whereas ω‘s are the error terms with an expected value of zero and λ_j is the inverse Mills ratio computed from the estimated probabilities in Equation (6) as follows:

... (7)

ρ is the correlation coefficient of the u‘s and ϵ‘s . In the multinomial choice setting, there are J − 1 selection bias correction terms, one for each alternative credit source. The standard
errors in Equation (7) are bootstrapped to account for the heteroscedasticity arising from the generated regressor (λ_j).

Estimation of Average Treatment Effects

The MESR framework is used to examine the average treatment effects on the treated (ATT) by comparing the expected outcomes of each alternative credit source. The challenge of impact evaluation using observational data is to estimate the counterfactual outcome, which is the outcome households could have achieved had they not chosen the one they did. Following Carter and Milon (2005) and Di Falco and Veronesi (2013), we compute the ATT in the actual and counterfactual scenarios as follows:

Adopters with adoption (actual adoption observed in the sample):

 

... (8)

... (9)

Adopters, had they decided not to adopt (counterfactual):

... (10)

 

... (11)

 

These expected values are used to derive unbiased estimates of the ATT. The ATT is defined as the difference between Equations (8a) and (10a) or Equations (8b) and (10b). For instance, the difference between Equations (8a) and (10a) is given as:

... (12)

The first term on the right-hand side of Equation (12) represents the expected change in the mean outcome attributed to a credit source if an associated agricultural household with a credit source characteristic had the same outcome variable as that of a non-associated agricultural household with a corresponding credit source. The second term (λj) is the selection term that captures all the potential effects of differences in unobserved variables. On the other hand, the average treatment effect on the untreated (ATU) is the difference between Equations (9a) and (11a) and can be specified as:

... (13)

Determinants of Access to Credit

Table 5 presents the coefficient and marginal effects of the multinomial regression estimated with 1,940 observations. The model is significant at the 1% level.² The estimated coefficients differ significantly across alternative sources of credit. Size of landholding had a positive significant effect on access to credit. Households with larger landholdings were more likely to take credit from both formal and informal sources. Households from general castes were more likely to take credit from formal sources than were SC and ST households. Farmers who were more educated preferred to take credit from formal sources. The awareness of the households also affected their decision to access credit and the selection of credit outlets. For instance, agricultural households that had heard about loan waiver schemes preferred to take credit from formal sources.

The promotion of loan waiver schemes by political parties motivated farming households to take credit from formal sources. Similarly, awareness of the direct benefit transfer scheme also had a significantly positive effect on taking credit from both formal and informal sources. On the other hand, agricultural households which were dependent on remittances preferred to get credit from informal sources. Households that had opened Jan Dhan Yojana (JDY) bank accounts after the PMJDY was launched in 2014 were more likely to borrow from informal sources. Households that had sought information from any source had a higher propensity to use both formal and informal sources of credit. Agricultural households that had more livestock preferred to take credit from formal sources.

Impacts of Source of Credit

Income effects: The impact of sources of credit on agricultural households’ net farm income and on the productivity of major crops (rice and wheat) is examined next. This net farm income and the productivity of rice and wheat is used as a measure of agricultural household welfare. The estimated average net farm income from different sources of credit is calculated from the MESR model. We calculated ATT and ATU effects (Table 6). The findings in Table 6 should be viewed as two scenarios: (i) agricultural households preferring a single source of credit (formal or informal), and (ii) agricultural households prefer both formal and informal sources simultaneously.

The second-last column in Table 6 reports the treatment (ATT) and counterfactual (ATU) effects. Interestingly, in all three combinations of source of credit, the ATT and ATU effects are positive, suggesting that agricultural households that accessed credit realised higher annual net farm income than non-borrowing households, regardless of the source of credit chosen. However, agricultural households that accessed credit from both formal and informal sources simultaneously were more likely to experience enhanced annual net farm income.

Yield effects: Table 7 reports the impact of source of credit on the productivity of rice and wheat. The ATT of informal channels for outcome indicators (productivity of rice and wheat crops) have less value than formal sources. This suggests that taking credit from formal sources can result in higher productivity for agricultural households. In all counterfactual (ATU) cases, agricultural households that had taken credit from formal sources would have had a higher productivity had they not borrowed money. For example, agricultural households that had not taken credit from formal channels would have increased the productivity of their rice and wheat crops by about 1.7 q/ha and 1.8 q/ha, repectively, had they chosen to take credit from formal channels. This similarly positive impact was found for both formal and informal channels.

Heterogeneity Effects

The previous results for the ATT of credit access on outcome indicators depicted the important role played by credit. The estimates reported in Tables 6 and 7 assume a heterogeneous impact of credit access on all farmers; however, the estimated ATT of credit access on welfare outcome indicators can vary among different sets of farm households. Capturing the differential impacts of the Kisan Vigyan Kendra (KVK) access is therefore important for targeting individual farm households as well as designing the best-fit approach instead of a “one-size, one-institution and one-method-fits-all” approach. In this section, we present the heterogeneous treatment effect of access to credit. Following Verhofstadt and Maertens (2015) and Wossen et al (2017), we use the ATT of individual outcome indicators as a dependent variable in an OLS regression and then examine how the estimated ATT varies with the socio-economic characteristics of farmers. The estimated results, as shown in Table 8, indicate that credit has hetero­geneous effects on farm households. We find a statistically significant differential impact of KVK access with respect to age, household size, gender, education, occupation, and farmer’s awareness. These results emphasise that households headed by a male, and particularly when he is more educated, benefit most from KVK services. However, the impact of access to credit seems to be neutral to scale, which implies that once the marginal and small-farm households overcome the barriers of accessing formal credit, the likelihood of benefitting from the credit use remains the same as for large-farm households.

Conclusions

Despite consistent growth in the national economy, agricultural development—especially in the eastern region—faces a number of challenges. Farm households’ lack of access to appropriate and adequate credit is one of the most important concerns in eastern India. This study explores the impediments to credit access experienced by rural households and the impact of credit on household incomes.

Three states in eastern India where the incidence of poverty is the highest were selected as the research area for this study. Most of the rural population in the selected region derives its primary income from agriculture and has limited access to credit schemes. There is increasing concern that eastern India is at a disadvantage in terms of poverty-reduction measures as compared to other regions of the country.

The results show that credit access is strongly associated with the socio-economic and demographic characteristics of agricultural households. Access to credit has increased their household income substantially and has significantly raised yields of major staple crops. The effects of credit access have an observable heterogeneous impact across different groups of households based on education and social group, implying that credit policies should be made adaptable to different types of farm households. However, ceteris paribus, credit access depicts homogeneous effects on marginal, small, medium, and large farm households, suggesting that the impact of credit access is neutral to scale.

This study is subject to certain limitations and also provides insights for further research. Since the study is based on cross-sectional data, it was not possible to analyse the dynamics of household resources and credit issues over time. More longitudinal studies are needed to assess the long-term effects of credit on agricultural households’ welfare.

Notes

1 The second-stage parameter estimates of the outcome variables are given in the Appendix Table A2 (p 54).

2 Note that the model included block fixed effects as control variables.

References

Arouri, M, C Nguyen and A Ben Youssef (2015): “Natural Disasters, Household Welfare, and Resilience: Evidence from Rural Vietnam,” World Development, Vol 70, pp 59–77.

Awotide, B A, T Abdoulaye, A Alene and V M Manyong (2015): “Impact of Access to Credit on Agricultural Productivity: Evidence from Smallholder Cassava Farmers in Nigeria,” conference paper presented at the International Association of Agricultural Economists, Milan, Italy, 9–14 August.

Binswanger, H P and S Khandker (1995): “The Impact of Formal Finance on the Rural Economy of India,” Journal of Development Studies, Vol 32, No 2, pp 234–262.

Bourguignon, F, F Ferreira and M Walton (2007): “Equity, Efficiency and Inequality Traps: A Research Agenda,” Journal of Economic Inequality, Vol 5, pp 235–56.

Carter, M R (1989): “The Impact of Credit on Peasant Productivity and Differentiation in Nicaragua,” Journal of Development Economics, Vol 31, No 1, pp 13–36.

Carter, D W and J W Milon (2005): “Price Knowledge in Household Demand for Utility Services,” Land Economics, Vol 81, No 2, pp 265–83.

Carter, M R and K D Wiebe (1990): “Access to Capital and Its Impact on Agrarian Structure and Productivity in Kenya,” American Journal of Agricultural Economics, Vol 72, No 5, pp 1146–50.

Collins, D, J Morduch, S Rutherford and O Ruthven (2009): Portfolios of the Poor: How the World’s Poor Live on $2 a Day, Princeton, NJ: Princeton University Press.

Di Falco, S and M Veronesi (2013): “How Can African Agriculture Adapt to Climate Change? A Counterfactual Analysis from Ethiopia,” Land Economics, Vol 89, No 4, pp 743–66.

Di Falco, S, M Veronesi and M Yesuf (2011): “Does Adaptation to Climate Change Provide Food Security? A Micro-perspective from Ethiopia,” American Journal of Agricultural Economics, Vol 93, No 3, pp 829–46.

Feder, G, L J Lau, J Lin and X P Luo (1990): “The Relationship between Credit and Productivity in Chinese Agriculture: A Microeconomic Model of Disequilibrium,” American Journal of Agricultural Economics, Vol 72, No 5, pp 1151–57.

Ghosh, D N (2005): “A Policy Approach for Agricultural Lending,” Economic & Political Weekly, Vol 40, No 2, pp 93–96.

GoI (2018): Economic Survey, 2018–19, Ministry of Finance, Government of India, http://mofapp.nic.in:8080/economicsurvey/

Golait, R (2007) “Current Issues in Agriculture Credit in India: An Assessment,” Reserve Bank of India Occasional Papers, Vol 28, No 1, pp 79–100.

Guirkinger, C (2008): “Understanding the Co-existence of Formal and Informal Credit Markets in Piura, Peru,” World Development, Vol 36, No 8, pp 1436–52.

Hoda, A and P Terway (2015): “Credit Policy for Agriculture in India: An Evaluation, Working Paper 302, New Delhi: Indian Council for Research on International Economic Relations.

Joshi, P K and A Kumar (2017): “Transforming Agriculture in Eastern India: Challenges and Opportunities,” Vicissitudes of Agriculture in the Fast-Growing Indian Economy: Challenges, Strategies and the Way Forward, C Ramasamy and K R Ashok (eds), New Delhi: Indian Society of Agricultural Economics, pp 125–49.

Khandker, S R and R R Faruqee (2003): “The Impact of Farm Credit in Pakistan,” Agricultural Economics, Vol 28, No 3, pp 197–213.

Kumar, A, A Mishra, S Saroj and P K Joshi (2017): “Institutional versus Non-institutional Credit to Agricultural Households in India: Evidence on Impact from a National Farmers’ Survey,” Economic Systems, Vol 41, No 3, pp 420–32.

Kumar, A, R K P Singh, Shiv Jee, S Chand, G Tripathi and S Saroj (2015): “Dynamics of Access to Rural Credit in India: Patterns and Determinants,” Agricultural Economics Research Review, Vol 28 (Conf), pp 151–66.

Kumar, A, K M Singh and S Sinha (2010): “Institutional Credit to Agriculture Sector in India: Status, Performance and Determinants,” Agricultural Economics Research Review, Vol 23, No 2, pp 253–64.

Li, X, C Gan and B Hu (2011): “The Welfare Impact of Microcredit on Rural Households in China,” Journal of Behavioural and Experimental Economics, No 40, Vol 4, pp 404–11.

Luan, D X and S Bauer (2016): “Does Credit Access Affect Household Income Homogeneously across Different Groups of Credit Recipients? Evidence from Rural Vietnam.” Journal of Rural Studies, Vol 47, Part A, pp 186–203.

Mansur, E T, R Mendelsohn and W Morrison (2008): “Climate Change Adaptation: A Study of Fuel Choice and Consumption in the US Energy Sector, Journal of Environmental Economics and Management, Vol 55, No 2, pp 175–93.

McFadden, D (1973): “Conditional Logit Analysis of Qualitative Choice Be,” Frontiers in Econometrics, P Zarembka (ed), New York: Academic Press, pp 105–42.

Mohan, R (2006): “Agricultural Credit in India: Status, Issues and Future Agenda,” Economic & Political Weekly, No 41, No 11, pp 1013–21.

Narayanan, S (2016): “The Productivity of Agricultural Credit in India,” Agricultural Economics, No 47, No 4, pp 399–409.

Pitt, M M and S R Khandker (1996): “Household and Intrahousehold Impact of the Grameen Bank and Similar Targeted Credit Programs in Bangladesh,” World Bank Discussion Paper, No 320, World Bank, Washington, DC.

— (1998): “The Impact of Group-Based Credit Programs on Poor Households in Bangladesh: Does the Gender of Participants Matter?” Journal of Political Economy, Vol 106, No 5, pp 958–96.

Udry, Christopher (1994): “Risk and Insurance in a Rural Credit Market: An Empirical Investigation in Northern Nigeria,” Review of Economic Studies, Vol 61, No 3, pp 495–526.

Verhofstadt, E and M Maertens (2015): “Can Agricultural Cooperatives Reduce Poverty? Heterogeneous Impact of Cooperative Membership on Farmers’ Welfare in Rwanda,” Applied Economic Perspectives and Policy, Vol 37, No 1, pp 86–106.

Wetterberg, A (2007): “Crisis, Connections, and Class: How Social Ties Affect Household Welfare,” World Development, Vol 35, No 4, pp 585–606, DOI: 10.1016/j.worlddev.2006.06.005.

Wossen, T T Abdoulaye, A Alene, G M Haile, S Feleke, A Olanrewaju and V Manyong (2017): “Impacts of Extension Access and Cooperative Membership on Technology Adoption and Household Welfare,” Journal of Rural Studies, Vol 54, pp 223–33.

Wu, J J and B A Babcock (1998): “The Choice of Tillage, Rotation, and Soil Testing Practices: Economic and Environmental Implications,” American Journal of Agricultural Economics, Vol 80, No 3, pp 494–511.