This paper shows empirical evidence supporting the existence of persistence and nonlinearity in Philippine stock price data. The results reinforce the inadequacies of using Autoregressive Integrated Moving Average (ARIMA) models in explaining the dynamics in Philippine stock price data. The analyses maded use of the Oil Index, Philippine Stock Index (Phisix) and four sectoral stock series namely, PLDT, Ayala, Manila Mining and Petrofields. Autoregressive Fractional Integrated Moving Average (ARFIMA) models and Autoregressive Conditional Heteroskedastic (ARCH) models were used to show mean and variance persistency. A test for time irreversibility known as the Ramsey-Rothman test and the Box-Ljung Q test were employed to test nonlinearity. Finally, a nueral network model (Multilayer Perceptron architecture) was fitted to the sectoral data. The model gave good fit results boosting further the assertion of
nonlinearity.
