The Role of Household-Level Characteristics in Predicting the Unbanked in Rural India : A Comparison of Eastern and Western Regions

In India, despite numerous measures adopted by Reserve Bank of India (RBI) to ensure adequate supply of financial services in rural areas, the banking participation by households in terms of owning a bank account is considerably low. The present paper employs Discriminant Analysis (DA) and Logistic Regression Analysis (Logit) technique to determine the unbanked households in rural India using information on various socio-demographic characteristics of the households from National Family Health Survey (NFHS) – 2005-06 data. The unique feature of the paper is that it combines the results from both the methods to improve the predictive accuracy and compares the household-level characteristics that affect the banking participation in the two regions with highest and lowest unbanked households in rural India. Combining both the methods, the paper correctly predicts 95.1% and 83.3% of unbanked households in Eastern and Western region respectively. It is also observed that the number of adults along with their education level increases the probability of being banked, while families belonging to two major minority religions (Muslim and Christian) reduce the probability of the same. While female headship and belonging to SC or ST category reduces the banking participation in Western region, female headship is found to increase banking participation in Eastern region