An Inquiry into the Financial Literacy and Cognitive Ability of Farmers: Evidence from Rural India

Abstract

Poor understanding of financial products and an inability to process financial information prevent millions of rural households in the developing world from making informed financial decisions. This article assesses the financial literacy and cognitive ability of farmers using data from a unique field experiment in the Indian state of Gujarat. Using ordered response models, the effect of farmers' education on cognitive ability and financial literacy is estimated on the one hand, and the relationship between cognitive ability and financial literacy is analysed on the other. Farmers' education and financial experience are shown to be significantly correlated with achievements in customized tests for ability in mathematics and probability, which are taken as the two components of cognitive ability. Cognitive ability, in turn, predicts financial aptitude and debt literacy, the two components of financial literacy. By focusing on farmers in a developing country, the findings contribute to an improved understanding of financial literacy in such settings and can inform the design of inclusive financial systems that are sensitive to the cognitive and informational limitations of rural households.

Notes

The authors acknowledge financial support from the ILO Microinsurance Innovation Facility for a research grant to conduct a field experiment in the Indian state of Gujarat, which is the source of data for this study. Sarthak Gaurav was a principal investigator along with Shawn Cole of Harvard Business School for the field experiment. The authors thank Sachin Oza of the Development Support Centre, Ahmedabad, and Natu Macwana of Sajjata Sangh, Ahmedabad, for organizational assistance and logistics. Mayur Kumar Rathore provided excellent research assistance. The authors also thank three anonymous referees, an associate editor and the managing editor of Oxford Development Studies for suggestions and insightful comments on earlier versions of this paper. The paper has benefited from Sarthak Gaurav's discussions with Nilesh Fernando, Jeremy Tobacman, Shawn Cole and Sripad Motiram.

¹ See Behrman et al. (2010) and van Rooij et al. (2011, 2012) for the treatment of endogeneity.

² See Cunha et al. (2010) for a model of cognitive and non-cognitive skill formation.

³ Wu (1977) suggests that, in a densely populated agricultural system where production is typically carried out by small farmers, a minimum level of education (about 6 years of schooling) is necessary for farmers to learn to make effective use of modern agricultural inputs.

⁴ A non-profit and non-governmental organization, Development Support Centre, and its farmer groups' network, Sajjata Sangh, which are well known to farmers in the area, introduced rainfall insurance to 600 study households.

⁵ We also conducted a Wald test for joint significance of coefficients in the models. The Wald test provided sufficient evidence for rejecting the null of joint insignificance of the coefficients in the models, thus justifying the inclusion of the independent variables in the models. The results of the Brant test (omnibus likelihood ratio test) and Wald test can be provided on request.

⁶ As calculated from the coefficients of the estimated model. However, the coefficients and predicted probabilities are not reported in the tables in this article due to space constraint, but can be provided on request.

⁷ In this context, the literature on the role of family background (for example, parental education) in the formation of cognitive skills can also be explored (Cunha et al., 2005; Heckman, 2007).

⁸ As calculated from the coefficients of the estimated model. The coefficients and predicted probabilities are not reported in the tables due to space constraint, but can be provided on request.

⁹ See Moscardy & de Janvry (1977) and Binswanger (1980, 1981) for detailed discussion on risk attitudes and farmers in the developing world.

¹⁰ Though this is a limited indicator of farmers' risk aversion, we use it in the absence of a meaningful alternative measure in the survey data.

¹¹ See Johnson et al. (1993) for related work on framing effects in probability comprehension.

¹² The socio-economic characteristics of the SCs in our sample may not represent the overall socio-economic situation of the SCs in India.

¹³ We do not discuss the importance of returns to education or the resulting policy implications, as much has already been written on this theme.