Unit Root Dynamics in Financial Time Series: A Comparative Study of Parametric and Non-Parametric Testing Frameworks

The study explores unit heterogeneity implication in financial time series analysis and investment decision guide, using some parametric and non-parametric tests that includes Augmented Dickey-Fuller (ADF), Phillips-Perron (PP), and Kwiatkowski-Phillips-Schmidt-Shin (KPSS), i.e one parametric unit root test (ADF) against two non-parametric unit root test (PP and KPSS). The study selects 5 US financial and macroeconomic variables spanning from January 2020 to December 2023 (monthly series) that includes US Dollar Index, S&P 500 Index, NASDAQ, Bond Yield, and Consumer Price Index for all urban consumers. The findings underscore the heterogeneous nature of unit roots in financial time series, with differences in persistence, integration order, and mean-reverting properties across variables, which reinforces the need for a multi-test approach combining parametric and non-parametric techniques to achieve robust stationarity inference.