Determinants and Prediction Accuracy of Price Multiples for South East Asia: Conventional and Machine Learning Analysis

Abstract

The present study evaluates determinants of price multiples and their prediction accuracy using ordinary least square (OLS) regression and machine learning-based shrinkage methods for the South East Asian markets. Price multiples examined in the research are price to earnings (P/Es), price to book (P/B), and price to sales (P/S). Data has been collected from Thomson Reuters Eikon. The study recommends that the P/B ratio is the best price multiple for developing a price-based valuation model. Beside fundamental determinants of the multiple, various firm-level control variables, namely, firm size, cash holding, strategic holding, stock price volatility, firms’ engagement in Environment, Social, and Governance (ESG) activities, dividend yield, and net profit margin impact firm’s P/B multiple. Positive coefficients of consumer non-cyclical and healthcare dummies indicate a preference for defensive stocks by the investors. Application of machine learning-based shrinkage methods ensures the accuracy of prediction even with out-of-sample forecasting.