Testing mispricing-augmented factor models in an emerging market: A quest for parsimony

Abstract

This study is the first to test a financing-based misvaluation factor (UMO, undervalued-minus-overvalued), first proposed by Hirshleifer and Jiang (2010), for the Pakistani stock market. I find that the UMO factor, long underpriced (repurchase) stocks and short overpriced (new issue) stocks, earns significant mean and risk-adjusted returns. Further, I jointly examine the performance of UMO-augmented factor models – the Capital Asset Pricing Model, Carhart's four-factor model, and Fama and French's three-, five- and six-factor models – to find out which of these models or their subsets is most pertinent in the Pakistani stock market. A battery of tests – factor spanning regressions, Barillas and Shanken's (2017) maximum squared Sharpe ratio tests, and examination of two-way and one-way sorted portfolios using Gibbons-Ross-Shanken and Fama and French (2015, 2018) performance metrics over the 2003–2018 period – reveals that the UMO factor carries distinctive information that cannot be described by other factors under study. Finally, this study proposes a parsimonious four-factor model that combines the market, UMO, size, and profitability factors and outperforms the other models in Pakistan.