Can joint-liability microcredit help to share entrepreneurial risks? Insights from Mongolia

Article

Published 11.05.21
Photo credit:
Asian Development Bank/flickr

By allowing risk sharing, joint-liability lending can foster entrepreneurship among microcredit borrowers

Editors’ note: To know more about joint-liability microcredit, read our VoxDevLit on Microfinance.

The past decade has witnessed an intense debate about whether microcredit can lift people out of poverty. Experimental evidence from a variety of countries has fuelled scepticism around the efficacy of microcredit, showing how access to microloans often does not lead to meaningful increases in income or consumption for poor households (Banerjee et al. 2015). One explanation may be that relatively few people take up microcredit when it is made available to them, for example, because they are dissatisfied with some contractual features.

Microcredit contracts differ in various ways. An important question is therefore whether and how microcredit can become a more attractive – and hence possibly more effective – tool to increase entrepreneurship and living standards. Recent evidence suggests that small design changes such as the introduction of grace periods (Field et al. 2013), ex post repayment flexibility (Barboni and Agarwal 2018, Battaglia et al. 2018), or repayments that are ex ante tailored to individual needs (Crépon et al. 2021) may affect how people use microcredit and the impacts it has on various outcomes.

The joint liability model

A quintessential feature of microcredit contracts is their liability structure. In the early days of its existence, much attention was given to microcredit contracts in which borrowers form small groups that are jointly liable for repayment. Group members are treated as being in default when at least one of them fails to repay, effectively denying all members access to subsequent loans. This feature was supposed to improve loan performance and raise repayment rates, hence reducing the lender’s risk. The fact that joint-liability loans can also have a risk-sharing aspect for the borrowers has received less attention. In recent work (Attanasio, Augsburg, and De Haas 2019), we focus on this aspect of risk sharing, as it can possibly increase loan take up in situations where projects are risky and where uncertainty is a salient component of entrepreneurial decisions.

Our main hypothesis is that joint liability might encourage (and provide an institutional setting that allows for) risk sharing among group members, as, de facto, it reduces the amount of risk involved in a given project. Joint liability may therefore lead to an increase in the proportion of borrowers that start a business, compared with individual-liability microcredit. To formalise this intuition, we develop a simple theoretical model where individuals choose whether to take up a loan for a risky project or to pursue a safe project. The model produces two main testable predictions. First, individuals are more likely to take up a loan when offered a joint-liability contract instead of an individual-liability one. Second, although in both contractual frameworks take-up rates go down with the risk of the project, this effect is muted for joint-liability contracts.

Evidence from rural Mongolia

We provide empirical evidence on the relationship between risk and loan take up that is consistent with these theoretical predictions. To do so, we exploit survey data from a randomised controlled trial (RCT) in Mongolia (Attanasio et al. 2015). As part of this RCT, we randomly introduced individual-liability and joint-liability credit across villages (we abstained from introducing credit in a set of control villages). This unique set-up allows for a clean comparison of both types of liability structures while keeping other product features constant. Furthermore, as we discuss below, our data contains unique information on how perceived project risks vary across villages. This information allows us to test the second implication of our model.

Our analysis in Attanasio et al. (2015) focused on measuring the poverty impacts of both types of microcredit and compared their repayment performance. We found a small positive impact on food and total consumption from access to joint-liability loans, but not of individual-liability loans, whereas there were no differences in repayment performance. An open question posed at the end of that paper was why joint-liability loans may have been more effective at raising consumption? Our more recent work answers this question.

Capturing people’s beliefs about entrepreneurial risk

We use data on borrowers’ subjective expected returns that were collected as part of the original RCT. For each respondent, our survey data contain information about the subjective probability distribution of the returns on the investment project that could be financed by the newly available microcredit. Having elicited such subjective probability distributions through a number of novel questions, we computed the riskiness of investment projects at the individual level and, by averaging, the village level. This then allows us to relate differences in loan take up between individual- and joint-liability schemes to the (perceived) riskiness of the projects available to individuals across different villages.

Joint-liability microcredit as insurance

Our results confirm the insurance role of joint-liability contracts. When investment risk is higher – as measured by a high average variance of subjective risk perceptions – the probability of taking up a loan (and presumably engaging in a productive activity) is significantly lower. Potential borrowers that are more uncertain about their future returns thus appear less willing to commit to the fixed repayment schedule of a loan. Importantly, and as our model predicts, this discouraging effect of risk on loan take up is muted in villages where joint-liability credit is offered.

A follow-up question is whether the liability structure of the microcredit contracts also has an impact on the intensive margin? We find that this is the case. Borrowers who face higher investment risk feel comfortable with borrowing more in joint-liability than in individual-liability villages. Moreover, the total amount borrowed for risky projects tends to be higher in joint-liability villages. That is, groups that start relatively risky projects expand their total borrowing more in joint-liability villages. This may be one reason why we find somewhat larger consumption impacts in these villages in the associated RCT.

Importantly, we do not find differences in repayment performance between both types of villages. That is, the faster scaling up of risky projects in joint-liability villages does not come at the expense of higher default rates. Apparently, excessive risk taking is prevented by a credible threat of ex post social sanctions and/or effective ex ante assortative matching. Our interpretation is thus a positive one: joint liability appears to be an effective mechanism to overcome unwarranted risk aversion, at least among some borrowers who have access to risky but potentially profitable (and scalable) investments.

Building on the evidence

While empirical evidence on the impact of microcredit has made significant progress in understanding the impacts of microcredit, many questions remain. We believe one promising research area relevant to both academics and practitioners is the conditions under which microcredit product design can contribute to higher take-up rates, and possibly stronger impacts. Our recent work contributes to this agenda by showing how individuals who are offered a joint-liability loan are more likely to take up credit than individuals who are offered individual-liability credit. Using novel measures of subjective risk perceptions, we find that the probability of loan take up is lower in villages where risk is higher. In line with the insurance role of joint liability, the latter effect is muted (or even fully offset) in villages where joint-liability loans are available.

Policy implications

Our findings confirm that product design can be a key determinant of loan take up and that this may hold particularly true in high-risk environments that are prevalent in many developing countries. In such risky environments, risk-averse borrowers may value the insurance aspect of joint-liability microcredit contracts. Although a continuation of the trend towards liability individualisation may therefore be beneficial to less risk-averse borrowers, this trend may at the same time gradually exclude more risk-averse (and possibly poorer, as indicated by Banerjee and Duflo (2007)) borrowers from the market for formal financial services and preclude the financing of productive activities.

References

Attanasio, O, B Augsburg, R De Haas, E Fitzsimons and H Harmgart (2015), “The Impacts of Microfinance: Evidence from Joint-Liability Lending in Mongolia”, American Economic Journal: Applied Economics 7(1): 90–122.

Attanasio, O, B Augsburg and R De Haas (2019), “Microcredit Contracts, Risk Diversification and Loan Take-up”, Journal of the European Economic Association 17(6): 1797–1842.

Banerjee, A V and E Duflo (2007), “The Economic Lives of the Poor”, Journal of Economic Perspectives 21(1): 141–168.

Banerjee, A, D Karlan and J Zinman (2015), “Six Randomized Evaluations of Microcredit. Introduction and Further Steps”, American Economic Journal: Applied Economics 7(1): 1–21.

Barboni, G and P Agarwal (2018), “Knowing What’s Good for You: Can a Repayment Flexibility Option in Microfinance Contracts Improve Repayment Rates and Business Outcomes?”, working paper.

Battaglia, M, S Gulesci and A Madestam (2018), “Repayment Flexibility and Risk Taking: Experimental Evidence from Credit Contracts”, CEPR Discussion Paper No. 13329.

Crépon, B, R De Haas, F Devoto and W Parienté (2021), “Tailor-made Microcredit in Rural Morocco. Experimental Evidence on Loan Take-up and Poverty Impacts”, AEA RCT Registry No. 0002618.

Field, E, R Pande, J Papp and N Rigol (2013), “Does the Classic Microfinance Model Discourage Entrepreneurship among the Poor? Experimental Evidence from India”, American Economic Review 103(6): 2196–2226.