preprints.org > arts & humanities > other > doi: 10.20944/preprints201811.0504.v1

Preprint Concept Paper Version 1 This version is not peer-reviewed

Version 1 : Received: 18 November 2018 / Approved: 20 November 2018 / Online: 20 November 2018 (15:29:13 CET)

Modern portfolio theory claims that diversification into non-correlated or negatively correlated activities reduces the overall risk of a portfolio. Considering the total income of a bank as a portfolio of interest income and non-interest income, this paper investigates how the variability of interest income and non- interest income, and covariance between interest income and non-interest income influence the various risk factors of banks. We set out a study in the Indian context. We have extracted data for the period 2005-2017 and employed an extended version of Ridge, Lasso and Elastics Net regression to take care of multi-collinearly in our data. We have considered 10-fold cross-validation techniques to get optimal values of tuning parameters for Ridge, Lasso, and Elastics Net regression (which is a convex combination of ridge and the LASSO). We have compared different regression techniques by comparing RMSE and R2. We observe that non-interest income is positively correlated with interest income in the Indian context, but it does stabilize variance, idiosyncratic risk & market risk (Beta) of Indian Banks.

Income diversification; idiosyncratic risk; emerging economies; Ridge/Lasso/ Elastics net regression