Increasing access to agricultural credit in rural India is a major policy priority. This column examines whether farmers’ access to formal agricultural loans depends on their caste. It is found that while commercial banks do not discriminate against lower caste farmers in lending, cooperative banks do as they are prone to interest-group capture at the local level.
Enhancing access to financial services such as credit, insurance and savings in rural areas is a major policy priority in India. In particular, agricultural credit has been an important focus of policymakers. The case of agricultural credit is particularly significant because of the crucial role that agriculture plays in the rural economy in terms of employment and income generation. The Reserve Bank of India (RBI), central and state governments therefore encourage bank lending to rural borrowers through a combination of lending targets for banks and softer terms for rural borrowers.
However, rural India is characterised by the divisive features of an entrenched caste system. While the influence of caste on economic transactions might be declining due to the spread of better functioning markets, there is little doubt that caste still plays an important role in shaping both economic and social transactions. The Indian State recognises the regressive role of caste, and has progressively worked to design policies that reduce the disadvantage faced by lower castes. Several credit-related policy guidelines issued by the RBI also reflect these concerns.
Against this backdrop, in a recent study that uses national data1, we examine whether farmers´ access to formal agricultural credit depends on their caste (Kumar 2013).
Loans and caste
If we look at all agricultural production loans irrespective of their source, the data shows that farmers who belong to the lowest caste-group, the Scheduled Castes (SC), receive fewer loans on average than Other Backward Castes (OBCs), who in turn receive fewer loans than higher castes. However, we find that these caste-based differences disappear once we take into account other characteristics of the borrowers, such as the amount of land they own, their wealth and education levels, and various household demographics. Indeed, in an ideal world, a lender such as a bank would evaluate the credit worthiness of a borrower using information that can predict future repayment ability, such as repayment histories and ownership of assets that can be used as security. Lower caste borrowers might well be poorer and have fewer assets, and hence receive loans less often. But if we take loan-relevant characteristics into account explicitly, then caste should not normally be relevant to the lender´s decision.
Bank type matters
Since 95% of all formal agricultural loans come from cooperative and commercial banks, we examine whether caste influences access to loans from either or both types of bank. While commercial banks are large, centralised entities with multiple branches, rural cooperative banks are small, decentralised, relatively independent, and lack direct oversight from the RBI.
We find weak evidence that commercial banks favour SC farmers in lending, in accordance with the RBI´s recommendations. In other words, all else being equal, an SC farmer is somewhat more likely to receive an agricultural loan than his OBC or higher caste counterpart.
In the case of cooperative banks however, we find lending patterns that suggest discrimination against SC farmers. Even after accounting for the various characteristics that determine access to credit, we find that SC farmers are only half as likely to have a loan compared to higher caste farmers (OBC farmers lie roughly in the middle).
Figure 1. Chances of receiving a cooperative bank loan, by caste and asset ownership
Field studies that have observed cooperative banks at close quarters generally corroborate these findings. Various government committees including that headed by Prof. A. Vaidyanathan (2005) point towards key weaknesses in the functioning of cooperative banks. In general, cooperative bank managements have been observed to function poorly, be corrupt, and be actively involved in politics
In India, cooperative banks have a decentralised, three-tier structure spanning village, district and state levels. District entities are expected to oversee the activities of village entities, act as financial clearing-houses for them, and provide guidance on their functioning. The state entities in turn play the same role for district entities.
The structure and observed functioning of cooperative banks suggest a possible explanation for the discrimination against lower caste borrowers. If cooperative bank managements are vulnerable to interest-group capture, then locally-dominant caste groups can shape lending in ways that leave lower caste farmers with reduced access to credit. To test this, we studied the relationship between caste-based loan access and caste-dominance at the district level, where dominance is specified in terms of total land ownership2.
We find that district-level caste dominance is an important part of the explanation for farmers´ access to cooperative-bank credit. Inter-caste disparities in access are prominent only in districts where higher castes are dominant. The pattern visible in Figure 1, wherein higher caste farmers receive loans more often than SC and OBC farmers, is magnified for districts where higher castes are dominant. In such districts, higher castes are more than twice as likely to receive a loan compared to SCs, on average. But in districts where OBCs are dominant, there are few inter-caste disparities. In other words, the entire inter-caste differences in lending from cooperative banks (Figure 1) are due to inter-caste differences in higher caste-dominated districts (Figure 2)3.
Figure 2. Chances of having a cooperative bank loan, by caste and assets ownership, depending on dominant caste in the district
Our analysis suggests that certain cooperative banks discriminate against lower caste borrowers in the provision of agricultural credit. On average, these are cooperative banks located in districts where higher castes are dominant. This finding is indicative of interest-group capture at the level of cooperative bank managements -- locally dominant interests are able to influence, and possibly partially constitute, cooperative bank managements. Meanwhile, our finding that commercial banks do not discriminate against lower caste borrowers and in fact show weak evidence of favouring lower caste borrowers, is consistent with government policy, and implies that large, centralised organisations can sometimes function better than decentralised ones. This is possibly because it is easier to maintain regulatory oversight of their functioning, and their managerial structures are less prone to local interference.
To probe this issue in greater depth, at least two things are needed. First, household surveys, such as the one used here, need to ask questions about borrowers´ experiences of obtaining credit, rather than only asking about actual loan outcomes, as is the case right now. This would help build more complete explanations for borrowers’ decisions to seek or not seek credit, and how they fare in terms of actually receiving it. Second, bank-level surveys could complement these household surveys by providing corroborative evidence. They might also help detail the explanation for why certain management structures perform better than others.
- Round 59, Schedule 18.2 of the National Sample Survey Organisation (NSSO) data.
- The notion of a dominant caste was first discussed by Srinivas (1959), and has been used recently by Anderson (2011).
- Since there are very few SC-dominated districts we do not focus on these. They show few inter-caste differences, like OBC-dominated districts.
- Anderson, S (2011), “Caste as an Impediment to Trade”, American Economic Journal: Applied Economics, 3 (1): 239–263.
- Government of India (2005), ‘Task force on revival of cooperative credit institutions’, Chairman A. Vaidyanathan, Government of India.
- Kumar, Sunil Mitra (2013), “Does Access to Formal Agricultural Credit Depend on Caste?”, World Development, 43: 315–328.
- Srinivas, M N (1959), “The dominant caste in Rampura”, American Anthropologist, 61, 1–16.