Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Factor-based forecasting in the presence of outliers: Are factors better selected and estimated by the median than by the mean?

Factor-based forecasting in the presence of outliers: Are factors better selected and estimated... Abstract Macroeconomic forecasting using factor models estimated by principal components has become a popular research topic with many both theoretical and applied contributions in the literature. In this paper we attempt to address an often neglected issue in these models: The problem of outliers in the data. Most papers take an ad-hoc approach to this problem and simply screen datasets prior to estimation and remove anomalous observations. We investigate whether forecasting performance can be improved by using the original unscreened dataset and replacing principal components with a robust alternative. We propose to use an estimator based on least absolute deviations (LAD) as this alternative and establish a tractable method for computing the estimator. In addition to this we demonstrate the robustness features of the estimator through a number of Monte Carlo simulation studies. Finally, we apply the estimator in a simulated real-time forecasting exercise to test its merits. We use a newly compiled dataset of US macroeconomic series spanning the period 1971:2–2012:10. Our findings suggest that the chosen treatment of outliers does affect forecasting performance and that in many cases improvements can be made using a robust estimator such as the proposed LAD estimator. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Studies in Nonlinear Dynamics & Econometrics de Gruyter

Factor-based forecasting in the presence of outliers: Are factors better selected and estimated by the median than by the mean?

Download PDF
30 pages

Loading next page...
 
/lp/de-gruyter/factor-based-forecasting-in-the-presence-of-outliers-are-factors-p0sdHpT0b6
Publisher
de Gruyter
Copyright
Copyright © 2014 by the
ISSN
1081-1826
eISSN
1558-3708
DOI
10.1515/snde-2012-0049
Publisher site
See Article on Publisher Site

Abstract

Abstract Macroeconomic forecasting using factor models estimated by principal components has become a popular research topic with many both theoretical and applied contributions in the literature. In this paper we attempt to address an often neglected issue in these models: The problem of outliers in the data. Most papers take an ad-hoc approach to this problem and simply screen datasets prior to estimation and remove anomalous observations. We investigate whether forecasting performance can be improved by using the original unscreened dataset and replacing principal components with a robust alternative. We propose to use an estimator based on least absolute deviations (LAD) as this alternative and establish a tractable method for computing the estimator. In addition to this we demonstrate the robustness features of the estimator through a number of Monte Carlo simulation studies. Finally, we apply the estimator in a simulated real-time forecasting exercise to test its merits. We use a newly compiled dataset of US macroeconomic series spanning the period 1971:2–2012:10. Our findings suggest that the chosen treatment of outliers does affect forecasting performance and that in many cases improvements can be made using a robust estimator such as the proposed LAD estimator.

Journal

Studies in Nonlinear Dynamics & Econometrics – de Gruyter

Published: May 1, 2014

References

  • Forecasting Macroeconomic Variables for the New Member States Central and Eastern European Countries and the European Union edited by University
    Banerjee
  • Robust nd ed New York
    Huber
  • How Successful are Dynamic Factor Models at Forecasting Output and Inflation - Analytic Approach of
    Eickmeier
  • Asymptotics for Least Absolute Deviation Regression Estimators
    Pollard
  • Comparison of Forecasts from Least Absolute Value and Least Squares Regression of
    Dielman
  • Least Absolute Value Regression Recent Contributions of and
    Dielman
  • Stock Forecasting in Dynamic Factor Models Subject to Structural Instability Methodology and Practice of Festschrift in Honour of David edited by University
    Watson
  • Forecasting German GDP using Alternative Factor Models Based on Large of
    Schumacher
  • Asymptotics for Least Absolute Deviation Regression Estimators
    Pollard
  • Stock a Forecasting Using Principal Components from a Large Number of Predictors of the
    Watson
  • Stock a Forecasting Using Principal Components from a Large Number of Predictors of the
    Watson
  • Shortcut to LAD Estimator Asymptotics
    Phillips
  • Comparison of Forecasts from Least Absolute Value and Least Squares Regression of
    Dielman
  • Stock Implications of Dynamic Factor Models for VAR NBER Working Paper National Bureau of Economic Research
    Watson
  • Determining the Number of Factors in Approximate Factor Models
  • Asymptotic Theory of Least Absolute Error Regression of the
    Bassett
  • Least Absolute Error Estimation in the Presence of Serial Correlation of
    Weiss
  • Stock Macroeconomic Forecasting Using Diffusion Indexes of
    Watson
  • Calculation Method for Nonlinear Dynamic Least - Absolute Deviations Estimator of the Japan Statistical
    Hitomi
  • Massachusetts University
    Amemiya
  • Multivariate Statistical th ed Upper Saddle New Jersey
    Johnson
  • Comparing Predictive Accuracy of
    Diebold
  • Stock Diffusion Indexes Working Paper National Bureau of Economic Research
    Watson
  • Forecasting with Many Predictors of Economic Forecasting edited by
    Stock
  • Massachusetts University
    Amemiya
  • Testing for Structural Breaks in Dynamic Factor Models of
    Breitung
  • Estimating Large - Scale Factor Models for Economic Activity in Do They Outperform Simpler Models Jahrbücher für Nationalökonomie und Statistik
    Schumacher
  • Diffusion Index - Based Inflation Forecasts for the Euro Area Working Paper European Central Bank
    Angelini
  • Confidence Intervals for Diffusion index Forecasts and Inference for Factor - Augmented Regressions
  • Performance Measurement with the Arbitrage Pricing Theory New Framework for of
    Connor
  • Diffusion Index - Based Inflation Forecasts for the Euro Area Working Paper European Central Bank
    Angelini
  • Evaluating Factor Models : An Application to Forecasting Inflation in Working Paper Bank of
    Gosselin
  • Multivariate Statistical th ed Upper Saddle New Jersey
    Johnson
  • Stock Forecasting in Dynamic Factor Models Subject to Structural Instability Methodology and Practice of Festschrift in Honour of David edited by University
    Watson
  • Stock New Indexes of Coincident and Leading Economic Indicators edited by MIT
    Watson
  • Stock Dynamic Factor Models InOxford Handbook of Economic Forecasting edited by University
    Watson
  • Estimating Nonlinear Dynamic Models Using Least Absolute Error Estimation
    Weiss
  • Confidence Intervals for Diffusion index Forecasts and Inference for Factor - Augmented Regressions
  • How Successful are Dynamic Factor Models at Forecasting Output and Inflation - Analytic Approach of
    Eickmeier
  • Fitting Multiplicative Models by Robust Alternating Regressions and
    Croux
  • Forecasting German GDP using Alternative Factor Models Based on Large of
    Schumacher
  • Testing for Structural Breaks in Dynamic Factor Models of
    Breitung
  • Inferential Theory for Factor Models of Large Dimensions
  • Stock New Indexes of Coincident and Leading Economic Indicators edited by MIT
    Watson
  • Factor Forecasts for the of
    Artis
  • Forecasting Macroeconomic Variables for the New Member States Central and Eastern European Countries and the European Union edited by University
    Banerjee
  • Performance Measurement with the Arbitrage Pricing Theory New Framework for of
    Connor
  • Sparse and Robust Factor Modelling Discussion Paper TI Institute
    Croux
  • Forecasting Macroeconomic Variables Using Diffusion Indexes in Short Samples with Structural Change InForecasting in the Presence of Structural Breaks and Model Uncertainty of Economics and Globalization edited by Emerald Group Publishing Limited
    Banerjee
  • Stock Dynamic Factor Models InOxford Handbook of Economic Forecasting edited by University
    Watson
  • Forecasting with Many Predictors of Economic Forecasting edited by
    Stock
  • Asymptotic Theory of Least Absolute Error Regression of the
    Bassett
  • Least Absolute Value Regression Recent Contributions of and
    Dielman
  • Robust nd ed New York
    Huber
  • Factor Analysis of Risk of Empirical Series of Textbooks Monographs edited by
    Ludvigson
  • Least Absolute Error Estimation in the Presence of Serial Correlation of
    Weiss
  • Calculation Method for Nonlinear Dynamic Least - Absolute Deviations Estimator of the Japan Statistical
    Hitomi
  • Stock Macroeconomic Forecasting Using Diffusion Indexes of
    Watson
  • Evaluating Factor Models : An Application to Forecasting Inflation in Working Paper Bank of
    Gosselin
  • Arbitrage Factor Structure and Mean - Variance Analysis on Large Asset Markets
    Chamberlain
  • Determining the Number of Factors in Approximate Factor Models
  • Comparing Predictive Accuracy of
    Diebold
  • Shortcut to LAD Estimator Asymptotics
    Phillips
  • Arbitrage Factor Structure and Mean - Variance Analysis on Large Asset Markets
    Chamberlain
  • Estimating Large - Scale Factor Models for Economic Activity in Do They Outperform Simpler Models Jahrbücher für Nationalökonomie und Statistik
    Schumacher
  • Fitting Multiplicative Models by Robust Alternating Regressions and
    Croux
  • Factor Analysis of Risk of Empirical Series of Textbooks Monographs edited by
    Ludvigson
  • Estimating Nonlinear Dynamic Models Using Least Absolute Error Estimation
    Weiss
  • Forecasting Macroeconomic Variables Using Diffusion Indexes in Short Samples with Structural Change InForecasting in the Presence of Structural Breaks and Model Uncertainty of Economics and Globalization edited by Emerald Group Publishing Limited
    Banerjee
  • Sparse and Robust Factor Modelling Discussion Paper TI Institute
    Croux
  • Stock Diffusion Indexes Working Paper National Bureau of Economic Research
    Watson
  • Stock Implications of Dynamic Factor Models for VAR NBER Working Paper National Bureau of Economic Research
    Watson
  • Inferential Theory for Factor Models of Large Dimensions
  • Factor Forecasts for the of
    Artis