Forecasting stock prices of five major commercial banks in India and stress testing: A multivariate approach

Abstract

This study explains a novel glide path to model the stock prices of five major commercial banks in India. The purpose of this work is concerned with identifying the macroeconomic drivers of stock prices of five core banks listed in NIFTY50 index, namely, SBI, HDFC, ICICI, INDUSIND and KOTAK MAHINDRA, and to test their stability during periods of fiscal tension. The study utilized a Principal Component Analysis (PCA) of the respective stock prices and has proceeded with the multivariate method to make the required predictions. The choice of statistical technique primarily consists of decomposing the respective stock prices of the five banks using PCA and then selecting one optimal Principal Component (PC1). A multivariate regression model is employed to predict the optimal PC of such stock prices in which PC1 is the dependent variable and NIFTY50 and INR‐USD is the independent variables. Ultimately, the optimal PC is returned to the stock prices of the corresponding levels. The model is set up to be performed comfortably and the hardiness of the model is supported from various selection criteria; particular, the model assumption testing, goodness of fit, sample stressed period (using Rolling Average method on Indian VIX: 2008:Q1 to 2009:Q2) testing and out of sample performance. In addition to this, the estimated results are consistent with Monte‐Carlo simulation results. The envisioned model satisfies the model objective, while sufficiently capturing the historical dynamics of the stock prices as reflected by various performance indicators (RMSE, MAPE, RMSE over standard deviation) presented in this paper.