An empirical investigation of the usefulness of ARFIMA models for predicting macroeconomic and financial time series

This paper addresses the notion that many fractional I(d) processes may fall into the "empty box" category, as discussed in Granger (1999). We present ex ante forecasting evidence based on an updated version of the absolute returns series examined by Ding, Granger and Engle (1993) that suggests that ARFIMA models estimated using a variety of standard estimation procedures yield “approximations” to the true unknown underlying DGPs that sometimes provide significantly better out-of-sample predictions than AR, MA, ARMA, GARCH, and related models, with very few models being “better” than ARFIMA models, based on analysis of point mean square forecast errors (MSFEs), and based on the use of Diebold and Mariano (1995) and Clark and McCracken (2001) predictive accuracy tests. Results are presented for a variety of forecast horizons and for recursive and rolling estimation schemes. The strongest evidence in favor of ARFIMA models arises when various transformations of 5 major stock index returns are examined. For these data, ARFIMA models are frequently found to significantly outperform linear alternatives around one third of the time, and in the case of 1-month ahead predictions of daily returns based on recursively estimated models, this number increases to one half of the time. Overall, it is found that ARFIMA models perform better for greater forecast horizons, while this is clearly not the case for non-ARFIMA models. We provide further support for our findings via examination of the large (215 variable) dataset used in Stock and Watson (2002), and via discussion of a series of Monte Carlo experiments that examine the predictive performance of ARFIMA model.

[1]  J. Breitung,et al.  Inference on the cointegration rank in fractionally integrated processes , 2002 .

[2]  Philip Hans Franses,et al.  Inflation, forecast intervals and long memory regression models , 2002 .

[3]  J. Stock,et al.  Macroeconomic Forecasting Using Diffusion Indexes , 2002 .

[4]  Glenn D. Rudebusch,et al.  Long Memory and Persistence in Aggregate Output , 1989, Business Cycles.

[5]  P. Robinson.  Gaussian Semiparametric Estimation of Long Range Dependence , 1995 .

[6]  K. West,et al.  Asymptotic Inference about Predictive Ability , 1996 .

[7]  Norman R. Swanson,et al.  Predictive Density Accuracy Tests , 2004 .

[8]  Norman R. Swanson,et al.  Predictive Density Evaluation , 2005 .

[9]  Clive W. J. Granger,et al.  Properties of Nonlinear Transformations of Fractionally Integrated Processes , 2000 .

[10]  Norman R. Swanson,et al.  OUT-OF-SAMPLE TESTS FOR GRANGER CAUSALITY , 2001, Macroeconomic Dynamics.

[11]  C. Granger,et al.  Varieties of long memory models , 1996 .

[12]  Norman R. Swanson,et al.  Predictive ability with cointegrated variables , 2001 .

[13]  Peter Reinhard Hansen,et al.  Model Confidence Sets for Forecasting Models , 2005 .

[14]  W. Andrew.  LO, . Long-Term Memory in Stock Market Prices, Econometrica, , . , 1991 .

[15]  Philip Hans Franses,et al.  A nonlinear long memory model, with an application to US unemployment ☆ , 2002 .

[16]  F. Diebold,et al.  Comparing Predictive Accuracy , 1994, Business Cycles.

[17]  A. Lo.  Long-Term Memory in Stock Market Prices , 1989 .

[18]  Anthony S. Tay,et al.  Multivariate Density Forecast Evaluation and Calibration In Financial Risk Management: High-Frequency Returns on Foreign Exchange , 1999, Review of Economics and Statistics.

[19]  Glenn D. Rudebusch,et al.  On the Power of Dickey-Fuller Tests against Fractional Alternatives , 1991, Business Cycles.

[20]  Ying-Wong Cheung,et al.  TESTS FOR FRACTIONAL INTEGRATION: A MONTE CARLO INVESTIGATION , 1993 .

[21]  Peter C. B. Phillips,et al.  Exact Local Whittle Estimation of Fractional Integration , 2002 .

[22]  Clive W. J. Granger,et al.  Spectral Analysis for Economic Time Series , 1964 .

[23]  Murad S. Taqqu,et al.  Robustness of whittle-type estimators for time series with long-range dependence , 1997 .

[24]  F. Diebold,et al.  On maximum likelihood estimation of the differencing parameter of fractionally-integrated noise with unknown mean , 1994 .

[25]  Norman R. Swanson,et al.  An Out of Sample Test for Granger Causality , 2000 .

[26]  Norman R. Swanson,et al.  A Consistent Test for Nonlinear Out of Sample Predictive Accuracy , 2002 .

[27]  Richard T. Baillie,et al.  Long memory processes and fractional integration in econometrics , 1996 .

[28]  Anthony S. Tay,et al.  Evaluating Density Forecasts with Applications to Financial Risk Management , 1998 .

[29]  F. Diebold,et al.  Long Memory and Regime Switching , 2000 .

[30]  Michael P. Clements,et al.  Evaluating multivariate forecast densities: a comparison of two approaches , 2002 .

[31]  Michael W. McCracken.  Asymptotics for out of sample tests of Granger causality , 2007 .

[32]  H. E. Hurst,et al.  Long-Term Storage Capacity of Reservoirs , 1951 .

[33]  P. Robinson.  Log-Periodogram Regression of Time Series with Long Range Dependence , 1995 .

[34]  Todd E. Clark,et al.  Tests of Equal Forecast Accuracy and Encompassing for Nested Models , 1999 .

[35]  J. Geweke,et al.  THE ESTIMATION AND APPLICATION OF LONG MEMORY TIME SERIES MODELS , 1983 .

[36]  C. Granger,et al.  A long memory property of stock market returns and a new model , 1993 .

[37]  F. Diebold,et al.  Optimal Prediction Under Asymmetric Loss , 1994, Econometric Theory.

[38]  Marius Ooms,et al.  Computational aspects of maximum likelihood estimation of autoregressive fractionally integrated moving average models , 2003, Comput. Stat. Data Anal..

[39]  Robert F. Engle,et al.  The Reviewof Economicsand Statistics , 1999 .

[40]  C. Granger.  Investigating causal relations by econometric models and cross-spectral methods , 1969 .

[41]  Ben Goertzel,et al.  Long-Term Memory , 1993 .

[42]  Yixiao Sun,et al.  Adaptive Local Polynomial Whittle Estimation of Long-Range Dependence , 2002 .

[43]  Jan Beran,et al.  Maximum Likelihood Estimation of the Differencing Parameter for Invertible Short and Long Memory Autoregressive Integrated Moving Average Models , 1995 .

[44]  P. Newbold,et al.  BIAS IN AN ESTIMATOR OF THE FRACTIONAL DIFFERENCE PARAMETER , 1993 .

[45]  Francis X. Diebold,et al.  The Review of Economics and Statistics VOL . LXXIII FEBRUARY 1991 NUMBER 1 IS CONSUMPTION TOO SMOOTH ? LONG MEMORY AND THE DEATON PARADOX , 2008 .

[46]  Lutz Kilian,et al.  On the Selection of Forecasting Models , 2003, SSRN Electronic Journal.

[47]  Peter C. B. Phillips,et al.  Time Series Regression With a Unit Root and Infinite-Variance Errors , 1990, Econometric Theory.

[48]  Zeng Zhen-yu.  The Empirical Analysis of Long-term Memory in Stock Market , 2003 .

[49]  Clive W. J. Granger,et al.  Spectral Analysis for Economic Time Series , 1964 .

[50]  Fallaw Sowell.  Maximum likelihood estimation of stationary univariate fractionally integrated time series models , 1992 .

[51]  Eric Zivot,et al.  Long Memory and Structural Changes in the Forward Discount: An Empirical Investigation , 2007 .

[52]  C. Granger,et al.  Spectral Analysis for Economic Time Series , 1964 .

[53]  Fallaw Sowell.  Modeling long-run behavior with the fractional ARIMA model , 1992 .

[54]  W. Newey,et al.  A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelationconsistent Covariance Matrix , 1986 .

[55]  Barbara Rossi,et al.  Testing Long-Horizon Predictive Ability with High Persistence, and the Meese-Rogoff Puzzle , 2005 .

[56]  Michael W. McCracken.  Asymptotics for Out of Sample Tests of Causality , 1998 .

[57]  P. Newbold,et al.  Tests for Forecast Encompassing , 1998 .

[58]  C. Granger,et al.  Forecasting Economic Time Series. , 1988 .

[59]  Norman R. Swanson,et al.  A Model-Selection Approach to Assessing the Information in the Term Structure Using Linear Models and Artificial Neural Networks , 1995 .

[60]  Michael P. Clements,et al.  Evaluating The Forecast of Densities of Linear and Non-Linear Models: Applications to Output Growth and Unemployment , 2000 .

[61]  C. Granger.  Long memory relationships and the aggregation of dynamic models , 1980 .

[62]  C. Granger,et al.  An introduction to bilinear time series models , 1979 .

[63]  Norman R. Swanson,et al.  The Block Bootstrap for Parameter Estimation Error In Recursive Estimation Schemes, With Applications to Predictive Evaluation , 2003 .

[64]  H. White,et al.  A Reality Check for Data Snooping , 2000 .

[65]  P. Franses,et al.  Structural breaks and long memory in US inflation rates: Do they matter for forecasting? , 2001 .

[66]  P. Robinson.  Time Series with Long Memory , 2003 .

[67]  W. Newey,et al.  A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelationconsistent Covariance Matrix , 1986 .

[68]  C. Granger,et al.  Properties of Nonlinear Transformations of Fractionally Integrated Processes , 2000 .

[69]  Norman R. Swanson,et al.  A Model Selection Approach to Real-Time Macroeconomic Forecasting Using Linear Models and Artificial Neural Networks , 1997, Review of Economics and Statistics.

[70]  H. R. Kuensch.  Statistical Aspects of Self-Similar Processes , 1986 .

[71]  Uwe Hassler,et al.  Long Memory in Inflation Rates: International Evidence , 1995 .

[72]  C. Granger,et al.  AN INTRODUCTION TO LONG‐MEMORY TIME SERIES MODELS AND FRACTIONAL DIFFERENCING , 1980 .

[73]  P. Schmidt,et al.  On the Power of the KPSS Test of Stationarity Against Fractionally-Integrated Alternatives , 1996 .

[74]  Clive W. J. Granger,et al.  Occasional Structural Breaks and Long Memory , 1999 .