Academy of Economic Studies Bucharest Doctoral School of
Academy of Economic Studies - Bucharest Doctoral School of Finance and Banking DOFIN Long Memory in Stock Returns: Research over Markets Supervisor: Professor Dr. Moisă Altăr Bucharest, July 2008 MSc Student: Silvia Bardoş
Contents • Long memory & Motivation • Literature review • Steps & data used: • Testing stationarity and long memory • ADF & KPSS • Hurst exponent through R/S test & Hurst exponent through wavelet estimator • Determining long memory by estimating fractional differencing parameter • Geweke and Porter-Hudak test & Maximum Likelihood Estimate for an ARFIMA process • Conclusions
Long Memory & Motivation • Long memory has important implications in financial markets because if it is discovered it can be used to construct trading strategies. • Long memory or long range dependence means that the information from “today” is not immediately absorbed by the prices in the market and investors react with delay to any such information. So: • A long memory process is a process where a past event has a decaying effect on future events AND • Memory is the series property to depend on its own past realizations
Long Memory & Motivation • Mathematic view: Long memory processes relates to autocorrelation • If a time series of data exhibits autocorrelation, a value from the data set x s at time ts is correlated with another value xs+z at time ts+z. For a long memory process autocorrelation decays over time and the decay is slower than in a stationary process ( I(0) process) • So, if a long memory process exhibits an autocorrelation function that is not consistent for a I(1) process (a process integrated of order 1) nor for an I(0) process (a pure stationary process) we can consider a long memory process as being the layer separating the non-stationary process from the stationary ones – namely a fractionally integrated process.
Literature review • Evidence of long memory was first brought up by E. Hurst in 1951 when, testing the behavior of water levels in the Nile river, he observed that the flow of the river was not random, but patterned • Mandelbrot (1971) was among the first to consider the possibility of long range dependence in asset returns • Wright, J. (1999) is detecting evidence of long memory in emerging markets stock returns (Korea, Philippines, Greece, Chile and Colombia) • Caporale and Gil-Alana (2002), studying S&P 500 daily returns found results indicating that the degree of dependence remains relatively constant over time, with the order of integration of stock returns fluctuating slightly above or below zero • Henry Olan (2002) makes a survey for finding long memory in stock returns from an international perspective. Evidence of long memory is found in the German, Japanese, South Korean and Taiwanese markets against UK, USA, Hong Kong, Singapore and Australia where no sign of long memory appears.
Steps – Modeling long memory • A series x t follows an ARFIMA (p, d, q) process if: where Φ(L), θ(L) are the autoregressive and moving average polynomials, L is the lag, d is the fractional differencing parameter, εt is white noise. • For d within (0, 0. 5), the ARFIMA process is said to exhibit long memory or long range positive dependence • For d within (-0. 5, 0), the process exhibits intermediate memory or long range negative dependence • For d within [0. 5, 1) the process is mean reverting and there is no long run impact to future values of the process • The process is short memory for d=0 corresponding to a standard ARMA process
Testing stationarity • Memory is closely related to the order of integration • In the context of non-fractionally integration is equivalent to establish whether the series is I(0) or I(1) and the commonly used tests are ADF and KPSS ADF Null hypothesis: H 0: d = 1 (returns series are containing a unit root) Hassler and Wolter (1994) find that this test of unit root is not consistent against fractional alternatives so the ADF can be inappropriate if we are trying to decide whether a set of data is fractionally integrated or not. KPSS Null hypothesis: H 0: d = 0 (return series are stationary) Lee and Schmidt (1996) find that KPSS test can be used to distinguish short memory and long memory stationary processes
Testing stationarity KPSS Consider xt ( t = 1, 2, …, N), as the observed return series for which we wish to test stationarity The test decomposes the series into the sum of a random walk, a deterministic trend a stationary error with the following linear regression model: The KPSS statistics: and is the residual from regressing the series against a constant or a constant and a trend Under the null hypothesis of trend stationary, the residuals et (t regression of x on an intercept and time trend. ) are from the = 1, 2, …, N Under the null hypothesis of level stationarity, the residuals et are from a regression of x on intercept only. Rejection of ADF and KPSS indicates that the process is described by neither I(0) and I(1) processes and that is probable better described by the fractional integrated alternative (d is a non-integer).
Estimating long memory using R/S test • Mandelbrot & Wallis (1969) method allows computing parameter H, which measures the intensity of long range dependence in a time series Return time series of length T is divided into n sub-series of length m. For each sub-series m = 1, . . . , n, we: a) find the mean (Em) and standard deviation (Sm); b) we subtract the sample mean Zi, m = Xi, m − Em for i = 1, . . , m; c) produce a time series taking form of Wi, m = j, m where i = 1, …, m d) find the range Rm = max{W 1, m, …. , Wn, m} – min{ W 1, m, …. , Wn, m} e) rescale the range Rm by How does this procedure relates to the Hurst exponent?
Estimating long memory using R/S test • Einstein discovered that the distance covered by a random variable is close related to the square root of time (Brownian motion) , where R is the distance covered by the variable, k is a constant and T is the length of the time. • Using R/S analysis, Hurst suggested that: , where R/S is the rescaled range, m is the number of observations, k is the constant and H is the Hurst exponent, can be applied to a bigger class of time series (generalized Brownian motion) • The Hurst exponent can be than found as: H value Return time series log(R/S)m= log k + H log m = 0. 5 follow a random walk and are independent (0, 0. 5) are anti-persistent, process covers only a small distance than in the random walk case (0. 5, 1) are persistent series, process covers a bigger distance than a random walk (long memory)
Hurst exponent using wavelet spectral density • For computing the Hurst Exponent, the R wavelet estimator uses a discrete wavelet transform then: • averages the squares of the coefficients of the transform, • performs a linear regression on the logarithm of the average, versus the log of the parameter of the transform The result provides an estimate for the Hurst exponent. Wavelet transform behaves as a microscope that decomposes our return series into components of different frequency so this is why we tend to consider that results obtained for H through the wavelet estimator are being more accurate.
The GPH test (1983) • Semi-parametric approach to obtain an estimate of the fractional differencing parameter d based on the slope of the spectral density function around the frequency ξ=0 Periodogram (estimator of the spectral density) of x at a frequency ξ I (ξ) = • Geweke, J. and S. Porter-Hudak(1983) proposed as an estimate of the OLS estimator of d from the regression: , λ= 1, …. . , v the bandwidth v is chosen such that for but Geweke and Porter-Hudak consider that the power of T has to be within (0. 5, 0. 6). In our test we have considered: V=
Maximum likelihood estimates for ARFIMA model • In the present paper we have used the MLE implemented based on the approximate maximum likelihood algorithm of Haslett and Raftery (1989) in R. If the estimated d is significantly greater than zero, we consider it an evidence of the presence of long-memory.
Data used • For testing the existence of long memory we have selected indexes around the world trying to compare return series in mature markets (US, UK, Germany, France, Japan) with emerging markets (Romania, Poland the BRIC countries) • For the data series (1997 – 2008) we have first established the length as being 2 n (for the wavelet transform performed by the soft) and then we have transformed it in return series through: • For testing and comparing we have selected mainly, daily returns • Stationarity test were run in Eviews and long memory tests and estimation procedures were run in R
Is there evidence of long memory in the return time series? S&P 500 daily return series ADF KPSS Null Hypothesis: SP 500 DAY has a unit root Null Hypothesis: SP 500 DAY is stationary Exogenous: Constant, Linear Trend Exogenous: Constant Lag Length: 0 (Automatic based on SIC, MAXLAG=25) Bandwidth: 22 (Newey-West using Bartlett kernel) t-Statistic Prob. * Augmented Dickey-Fuller test statistic -47. 60859 0. 0000 Test critical values: 1% level -3. 962531 5% level -3. 412005 10% level -3. 127909 LM-Stat. Kwiatkowski-Phillips-Schmidt-Shin test statistic 0. 289248 Asymptotic critical values*: 1% level 0. 739000 5% level 0. 463000 10% level 0. 347000 R/S Hurst Exponent Diagnostic: 0. 4834943 Wavelet estimator for H: 0. 4108623 GPH 0. 45 0. 55 0. 75 0. 8 d 0. 032 -0. 0906 0. 082143 -0. 05057 -0. 03568 tstat sd (d=0) 0. 258 -0. 97417 0. 973893 -1. 25906 -1. 09493 tstat asd (d=0) 0. 231 -0. 82632 0. 935704 -1. 31442 -1. 12114 ARFIMA (0, d, 0) mle d Value 4. 583 E-05
FTSE 100 daily return series ADF KPSS Null Hypothesis: FTSE 100 DAY has a unit root Null Hypothesis: FTSE 100 DAY is stationary Exogenous: Constant, Linear Trend Exogenous: Constant Lag Length: 2 (Automatic based on SIC, MAXLAG=25) t-Statistic Prob. * Augmented Dickey-Fuller test statistic -29. 58352 0. 0000 Test critical values: -3. 962535 1% level Bandwidth: 17 (Newey-West using Bartlett kernel) LM-Stat. Kwiatkowski-Phillips-Schmidt-Shin test statistic 0. 279658 Asymptotic critical values*: 1% level 0. 739000 5% level -3. 412007 5% level 0. 463000 10% level -3. 127911 10% level 0. 347000 R/S Hurst Exponent Diagnostic: 0. 5587119 Wavelet estimator for H: 0. 3972144 GPHd=0 0. 45 0. 55 0. 75 0. 8 d 0. 0449066 -0. 024836 -0. 01847 -0. 06293 -0. 03258 tstat sd 0. 4092198 -0. 2552364 -0. 20357 -1. 47644 -0. 91725 tstat asd 0. 3206718 -0. 2265169 -0. 21044 -1. 63558 -1. 02361 ARFIMA (0, d, 0) mle d Value 4. 583 E-05
BET-FI daily return series ADF KPSS Null Hypothesis: BETFIDAYP has a unit root Null Hypothesis: BETFIDAYP is stationary Exogenous: Constant, Linear Trend Exogenous: Constant Lag Length: 0 (Automatic based on SIC, MAXLAG=21) Bandwidth: 7 (Newey-West using Bartlett kernel) t-Statistic Prob. * Augmented Dickey-Fuller test statistic -29. 22895 0. 0000 Test critical values: 1% level -3. 967044 5% level -3. 414212 10% level -3. 129218 LM-Stat. Kwiatkowski-Phillips-Schmidt-Shin test statistic 0. 443275 Asymptotic critical values*: 1% level 0. 739000 5% level 0. 463000 10% level 0. 347000 R/S Hurst Exponent Diagnostic: 0. 6177791 Wavelet estimator for H: 0. 6394731 GPHd=0 0. 45 0. 55 0. 75 0. 8 d 0. 1112765 0. 188012 0. 26837 0. 105296 0. 11638 tstat sd 0. 9531621 1. 830524 2. 329716 2. 023603 2. 708848 tstat asd 0. 6536144 1. 39642 2. 44605 2. 069451 2. 717905 ARFIMA (0, d, 0) mle Value d 0. 07864
BOVESPA daily return series ADF KPSS Null Hypothesis: BOVESPADAY has a unit root Null Hypothesis: BOVESPADAY is stationary Exogenous: Constant, Linear Trend Exogenous: Constant Lag Length: 0 (Automatic based on SIC, MAXLAG=25) Bandwidth: 16 (Newey-West using Bartlett kernel) t-Statistic Prob. * Augmented Dickey-Fuller test statistic -44. 17009 0. 0000 Test critical values: 1% level -3. 962531 5% level -3. 412005 10% level -3. 127909 LM-Stat. Kwiatkowski-Phillips-Schmidt-Shin test statistic 0. 414950 Asymptotic critical values*: 1% level 0. 739000 5% level 0. 463000 10% level 0. 347000 R/S Hurst Exponent Diagnostic: 0. 5681442 Wavelet estimator for H: 0. 5579485 GPHd=0 0. 45 0. 55 0. 75 0. 8 0. 055495 0. 124207 0. 138253 -0. 0504179 -0. 03275 tstat sd 0. 4703172 1. 199346 1. 650381 -1. 505468 -1. 09286 tstat asd 0. 3962822 1. 132831 1. 574859 -1. 310459 -1. 02899 d ARFIMA (0, d, 0) mle d Value 0. 0003773
RTS daily return series ADF KPSS Null Hypothesis: RTSDAY has a unit root Null Hypothesis: RTSDAY is stationary Exogenous: Constant, Linear Trend Exogenous: Constant Lag Length: 0 (Automatic based on SIC, MAXLAG=25) Bandwidth: 1 (Newey-West using Bartlett kernel) t-Statistic Prob. * Augmented Dickey-Fuller test statistic -42. 96666 0. 0000 Test critical values: 1% level -3. 962531 5% level -3. 412005 10% level -3. 127909 LM-Stat. Kwiatkowski-Phillips-Schmidt-Shin test statistic 0. 083395 Asymptotic critical values*: 1% level 0. 739000 5% level 0. 463000 10% level 0. 347000 R/S Hurst Exponent Diagnostic: 0. 543887 Wavelet estimator for H: 0. 531688 GPHd=0 0. 45 0. 55 0. 75 0. 8 -0. 1386964 -0. 11365 -0. 05167 -0. 0030198 0. 028219 tstat sd -1. 185368 -1. 29636 -0. 67683 -0. 0828427 0. 909179 tstat asd -0. 990412 -1. 03654 -0. 58861 -0. 0784892 0. 886579 d ARFIMA (0, d, 0) mled= 0 Value d 0. 03032
SENSEX daily return series ADF KPSS Null Hypothesis: SENSEXDAY has a unit root Null Hypothesis: SENSEXDAY is stationary Exogenous: Constant, Linear Trend Exogenous: Constant Lag Length: 0 (Automatic based on SIC, MAXLAG=25) Bandwidth: 10 (Newey-West using Bartlett kernel) t-Statistic Prob. * Augmented Dickey-Fuller test statistic -41. 41383 0. 0000 Test critical values: 1% level -3. 962531 5% level 10% level LM-Stat. Kwiatkowski-Phillips-Schmidt-Shin test statistic 0. 509953 Asymptotic critical values*: 1% level 0. 739000 -3. 412005 5% level 0. 463000 -3. 127909 10% level 0. 347000 R/S Hurst Exponent Diagnostic: 0. 568345 Wavelet estimator for H: 0. 525448 GPHd=0 0. 45 0. 55 0. 75 0. 8 d 0. 0759082 0. 009728 0. 031772 -0. 0129993 0. 013071 tstat sd 0. 5596446 0. 078911 0. 34705 -0. 3247642 0. 402457 tstat asd 0. 5420498 0. 088728 0. 361916 -0. 3378765 0. 410681 ARFIMA (0, d, 0) mled=0 d Value 0. 04682
Hang Seng daily return series ADF KPSS Null Hypothesis: HANGSENGDAY has a unit root Null Hypothesis: HANGSENGDAY is stationary Exogenous: Constant, Linear Trend Exogenous: Constant Lag Length: 0 (Automatic based on SIC, MAXLAG=25) Bandwidth: 5 (Newey-West using Bartlett kernel) t-Statistic Prob. * Augmented Dickey-Fuller test statistic -45. 92523 0. 0000 Test critical values: 1% level -3. 962531 5% level 10% level LM-Stat. Kwiatkowski-Phillips-Schmidt-Shin test statistic 0. 499391 Asymptotic critical values*: 1% level 0. 739000 -3. 412005 5% level 0. 463000 -3. 127909 10% level 0. 347000 R/S Hurst Exponent Diagnostic: 0. 528084 Wavelet estimator for H: 0. 495059 GPHd=0 0. 45 0. 55 0. 75 0. 8 d 0. 131589 0. 051335 0. 039096 0. 0278982 0. 019198 tstat sd 0. 896207 0. 469499 0. 48214 0. 7254383 0. 60606 tstat asd 0. 939662 0. 468203 0. 445347 0. 7251295 0. 603159 ARFIMA (0, d, 0) mle d Value 4. 583 E-05
Comparison between indices - Hurst Index daily H value via R/S S&P 500 0. 48 FTSE 100 Index daily H value via Wavelet estimator S&P 500 0. 4109 0. 5587 FTSE 100 0. 3072 CAC 40 0. 4738 CAC 40 0. 4161 DAX 0. 5189 DAX 0. 4957 NIKKEI 225 0. 5045 NIKKEI 225 0. 4927 WIG 0. 593 WIG 0. 4786 BET 0. 4232 BET 0. 5337 BET C 0. 6306 BET C 0. 5894 BET FI 0. 6187 BET FI 0. 6395 BOVESPA 0. 5681 BOVESPA 0. 5579 RTS 0. 5439 RTS 0. 5317 SENSEX 0. 5683 SENSEX 0. 5254 HANG SENG 0. 5281 HANG SENG 0. 4951 BRIC
Conclusions • Using a range of test and estimation procedures we have investigated whether stock returns exhibit long memory • Our results come to increase a bit the idea that emerging markets have a weak form of long memory as resulted in case of Russia and India or a stronger form like discovered in case of Romania (BET-FI), China and Brazil. Mature markets, in which we include US & UK among Germany, France show mixed evidence • We have tested for long memory the return series for BRIC countries indices Why? Because it is important to see is there is some kind of correlation between distant observations in these markets as emerging markets are of great interest to potential investors first taking into account their returns and second because they can be used in case of portfolio diversification as emerging market countries have low correlation with mature markets.
- Slides: 23