Applicability of the fractal market hypothesis (FMH) to the Philippine stock market

• Main Author: Vital, Villamor De Guzman
• Format: text
• Language: English
• Published: Animo Repository 2004
• Subjects: Investments > Philippines; Stocks > Philippines; Fractals; Hypothesis; Stock exchanges; Business Administration, Management, and Operations
• Online Access: https://animorepository.dlsu.edu.ph/etd_doctoral/42; https://animorepository.dlsu.edu.ph/context/etd_doctoral/article/1041/viewcontent/CDTG003707_P.pdf
• Institution: De La Salle University

This dissertation tests the applicability of the fractal market hypothesis (FMH) to the Philippine stock market. Peters (1994) first used the FMH to prove the persistence of the US market. FMH uses Rescaled Range (R/S) analysis to compute the Hurst exponent (H). If H less than or equal to .50, then the series is either white noise of antipersistent, and if H is greater than or equal to .50, then it is persistent. If the test shows persistence, ARFIMA forcasting is suggested because Hosking (1981), and Binsol and Bonzo (1996), found ARFIMA more appropriate in modeling persistent time series otherwise, use ARIMA. The results of tests in this dissertation cannot reject the null hypothesis (Ho) of white noise and antipersistence for the 18 Philippine stock market return series (1958-1997). Two stocks are white noise in line with classic efficient market hypothesis (EMH) over a short horizon and antipersistent (still EMH) over a longer horizon of returns. Such findings are explained by the unique characteristics of an emerging market. ARIMA forecasting is feasible for the PSE Composite Index (PHISIX) 20-day and 60-day returns, Commercial-Industrial Index (CI) 60-day returns, Mining Index (MINE) 20-day and 60-day returns, Lepanto Consolidated (LC) 20-day returns, San Miguel Corporation (SMC) 20-day and 60-day returns, and PLDT (Tel) 1-day and 60-day returns. Since Ho cannot be rejected, the data fit the EMH and Markowitz (1952) mean-variance efficient is the applicable portfolio model. Still, there is need to: (1) distinguish an efficient portfolio in normal times and in turbulent times as expounded by Chow, Jacquier, Kritzman and Lowry (1999) and (2) complement the Markowitz model with active asset management and contrarian timing to take advantage of the antipersistent returns of certain stocks.