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

Learn More →

Nonlinear and nonparametric modeling approaches for probabilistic forecasting of the US gross national product

Nonlinear and nonparametric modeling approaches for probabilistic forecasting of the US gross... Abstract Numerous time series models are available for forecasting economic output. Autoregressive models were initially applied to US gross national product (GNP), and have been extended to nonlinear structures, such as the self-exciting threshold autoregressive (SETAR) and Markov-switching autoregressive (MS-AR) models. We show that while these nonlinear models fit the in-sample data better than linear models, the out-of-sample forecast performance of extremely simple methods, such as the unconditional mean is competitive compared with both previously published linear and nonlinear models for GNP time series. Motivated by Occam′s razor and the forecasting competitiveness of the unconditional mean, we seek parsimonious models which are based on simple assumptions, incorporate few parameters, and can generate accurate forecasts. We propose nonlinear and nonparametric models based on nearest neighbor and kernel regression by forming a state-space of time delayed GNP observations. The rationale of the proposed methodology lies in identification of local regions of state-space known as nearest neighbor sub-spaces, whereby we utilize future trajectories of the neighboring states and weight them using a double kernel estimator to generate density forecasts. The models proposed in this work incorporate only one or two parameters, and the model estimation framework relies on optimizing the performance of in-sample density forecasts. The proposed modeling approach does not require prior assumptions regarding the form of the error distribution, and can provide transition between regimes of growth and contraction. We compare the forecasts from proposed models with classical time series models for GNP and a range of benchmarks, and validate our results on two post-war GNP time series using different performance scores, such as the root mean squared error (RMSE), mean absolute error (MAE), and the continuous ranked probability score (CRPS). We find that proposed models consistently outperformed previously published models for point and density forecasts of GNP at varying horizons. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Studies in Nonlinear Dynamics & Econometrics de Gruyter

Nonlinear and nonparametric modeling approaches for probabilistic forecasting of the US gross national product

Download PDF
26 pages

Loading next page...
 
/lp/de-gruyter/nonlinear-and-nonparametric-modeling-approaches-for-probabilistic-pILSuLF5yT
Publisher
de Gruyter
Copyright
Copyright © 2013 by the
ISSN
1081-1826
eISSN
1558-3708
DOI
10.1515/snde-2012-0029
Publisher site
See Article on Publisher Site

Abstract

Abstract Numerous time series models are available for forecasting economic output. Autoregressive models were initially applied to US gross national product (GNP), and have been extended to nonlinear structures, such as the self-exciting threshold autoregressive (SETAR) and Markov-switching autoregressive (MS-AR) models. We show that while these nonlinear models fit the in-sample data better than linear models, the out-of-sample forecast performance of extremely simple methods, such as the unconditional mean is competitive compared with both previously published linear and nonlinear models for GNP time series. Motivated by Occam′s razor and the forecasting competitiveness of the unconditional mean, we seek parsimonious models which are based on simple assumptions, incorporate few parameters, and can generate accurate forecasts. We propose nonlinear and nonparametric models based on nearest neighbor and kernel regression by forming a state-space of time delayed GNP observations. The rationale of the proposed methodology lies in identification of local regions of state-space known as nearest neighbor sub-spaces, whereby we utilize future trajectories of the neighboring states and weight them using a double kernel estimator to generate density forecasts. The models proposed in this work incorporate only one or two parameters, and the model estimation framework relies on optimizing the performance of in-sample density forecasts. The proposed modeling approach does not require prior assumptions regarding the form of the error distribution, and can provide transition between regimes of growth and contraction. We compare the forecasts from proposed models with classical time series models for GNP and a range of benchmarks, and validate our results on two post-war GNP time series using different performance scores, such as the root mean squared error (RMSE), mean absolute error (MAE), and the continuous ranked probability score (CRPS). We find that proposed models consistently outperformed previously published models for point and density forecasts of GNP at varying horizons.

Journal

Studies in Nonlinear Dynamics & Econometrics – de Gruyter

Published: Sep 1, 2013

References

  • How Accurate are Private Sector Forecasts ? Cross - Country Evidence from Consensus Forecasts of Output Growth of
    Loungani
  • Predicting Inflation Professional Experts Versus No - Change Forecasts arXiv
    Gneiting
  • Predicting Inflation Professional Experts Versus No - Change Forecasts arXiv
    Gneiting
  • What Causes the Forecasting Failure of Markov - switching Models Monte Carlo Study in Nonlinear Dynamics and Article
    Bessec
  • Inference When a Nuisance Parameter is Not Identified under the Null Hypothesis
    Hansen
  • Atmospheric Predictability as Revealed by Naturally Occuring Analogues of the
    Lorenz
  • Can Markov Switching Models Predict Excess Foreign Exchange Returns of
    Dueker
  • Nearest - neighbor Forecasts of Interest Rates Working in Number
    Barkoulas
  • Forecast Combination and the Bank of England s Suite of Statistical Forecasting Models Bank of England Working Paper no
    Kapetanios
  • Generalised linear Models for Site - specific Density Forecasting of UK Daily Rainfall
    Little
  • Short - term Forecasting of GDP Using Large Monthly Datasets Pseudo Real - time Forecast Evaluation Exercise European Central Bank Occasional Paper Series No
    Barhoumi
  • a Erratum The Likelihood Ratio Test Under Nonstandard Conditions : Testing the Markov Switching Model of of
    Hansen
  • Evaluating Long horizon Forecasts University of Missouri Columbia Manuscript
    Clark
  • Averages of Forecasts Some Empirical Results
    Makridakis
  • The Importance of Nonlinearity in Reproducing Business Cycle Features Federal Reserve Bank of St Working Paper
    Morley
  • Nonlinear Models for Macroeconomic Time Series Studies in Nonlinear
    Öcal
  • Using Monthly Indicators to Predict Bank of Working Paper No
    Zheng
  • On the Estimation of the Discrepancy Between Empirical Curves of Distribution for Two Independent Samples of University
    Smirnov
  • Inference When a Nuisance Parameter is Not Identified under the Null Hypothesis
    Hansen
  • Short - term Forecasting of GDP Using Large Monthly Datasets Pseudo Real - time Forecast Evaluation Exercise European Central Bank Occasional Paper Series No
    Barhoumi
  • Comparison of the Forecast Performance of Markov - switching and Threshold Autoregressive Models of US
    Clements
  • Forecasts Incremental Relative to Naive of
    Clemen
  • Forecasting Processes of
    Clements
  • Combining Predictive Distributions arXiv
    Gneiting
  • On the Estimation of the Discrepancy Between Empirical Curves of Distribution for Two Independent Samples of University
    Smirnov
  • Is The Cycle Characterized by Deterministic of
    Brock
  • Comparison of Direct and Iterated Multistep AR Methods for Forecasting Macroeconomic Time Series of
    Marcellino
  • Atmospheric Predictability as Revealed by Naturally Occuring Analogues of the
    Lorenz
  • Comparison of Linear and Nonlinear Univariate Models for Forecasting Macroeconomic Time Series InCh in Cointegration Causality and Forecasting Festschrift for Clive edited by University
    Stock
  • The Importance of Nonlinearity in Reproducing Business Cycle Features Federal Reserve Bank of St Working Paper
    Morley
  • Local Random Analogue Prediction of Nonlinear Processes
    Paparella
  • Nonparametric Regression University
    Härdle
  • Forecast - based Model Selection in the Presence of Structural Breaks Federal Reserve Bank of Kansas City
    Clark
  • What Causes the Forecasting Failure of Markov - switching Models Monte Carlo Study in Nonlinear Dynamics and Article
    Bessec
  • On Non linearity Forecasting of
    Clements
  • Some Methods for Assessing the Need for Non - linear Models in Business Cycle of
    Engel
  • - Cycle Analysis with a Markov - Switching Model of
    Goodwin
  • Can Markov Switching Models Predict Excess Foreign Exchange Returns of
    Dueker
  • Tests of Equal Forecast Accuracy and Encompassing for Nested Models of
    Clark
  • Floor and Ceiling Model of US Output of Economic Dynamics and
    Pesaran
  • Asymptotic Inference about Predictive Ability
    West
  • Comparison of Direct and Iterated Multistep AR Methods for Forecasting Macroeconomic Time Series of
    Marcellino
  • Some Methods for Assessing the Need for Non - linear Models in Business Cycle of
    Engel
  • Averages of Forecasts Some Empirical Results
    Makridakis
  • Forecasting Economic and Financial Time - Series with Non - linear models of
    Clements
  • Combining Predictive Distributions arXiv
    Gneiting
  • Nearest - neighbor Forecasts of Interest Rates Working in Number
    Barkoulas
  • Nonparametric Exchange Rate Prediction of
    Diebold
  • a Erratum The Likelihood Ratio Test Under Nonstandard Conditions : Testing the Markov Switching Model of of
    Hansen
  • Asymptotics for Out - of - sample Tests of Causality of
    McCracken
  • Systems Overcoming the Pitfalls of Forecasting Models via a Multidisciplinary Approach of
    Orrell
  • Local Linear Regression of the Statistical
    Yu
  • The of Statistical Learning Data Mining Inference Prediction nd ed Springer Series in
    Hastie
  • Forecasting Economic and Financial Time - Series with Non - linear models of
    Clements
  • Theory and an Extension of the Maximum Likelihood Principle International on Information Theory edited by Budapest
    Akaike
  • Combining Forecasts of
    Clemen
  • Box Time Series Forecasting Control San
    Jenkins
  • Econometric Modelling of Markov - switching Vector Autoregressions using MS VAR for Ox Discussion paper University of Oxford
    Krolzig
  • Local Random Analogue Prediction of Nonlinear Processes
    Paparella
  • Cycle Measurement in the Presence of Structural Change : Evidence of
    Krolzig
  • Using Conditional Kernel Density Estimation for Wind Power Density Forecasting of the
    Jeon
  • Generalised linear Models for Site - specific Density Forecasting of UK Daily Rainfall
    Little
  • Model Selection Approach to Real - time Macroeconomic Forecasting Using Linear Models and Artificial of and
    Swanson
  • Macroeconomic Forecasting in the Euro Area : Country Specific versus Area - wide
    Marcellino
  • a Comparing Density Forecasts Using Threshold and Quantile - Weighted Scoring Rules of
    Gneiting
  • Model Selection Approach to Real - time Macroeconomic Forecasting Using Linear Models and Artificial of and
    Swanson
  • - Cycle Analysis with a Markov - Switching Model of
    Goodwin
  • Nonlinear Modelling of Autoregressive Structural Breaks in a US Diffusion Index Dataset Bank of England Working Paper no
    Kapetanios
  • Nonparametric Exchange Rate Prediction of
    Diebold
  • Content for Univariate Time - series Forecasts of
    Galbraith
  • Probabilistic Forecasts Calibration Sharpness of the
    Gneiting
  • On Non linearity Forecasting of
    Clements
  • Modelling Nonlinear Relationships University
    Granger
  • Nonlinear Modelling of Autoregressive Structural Breaks in a US Diffusion Index Dataset Bank of England Working Paper no
    Kapetanios
  • Tests for Forecast of
    Harvey
  • The Future of Macroeconomic Forecasting : Understanding the Forecasting Process of
    Stekler
  • The Likelihood Ratio Test under Nonstandard Conditions : Testing the Markov Switching Model of of
    Hansen
  • Modelling Nonlinear Relationships University
    Granger
  • Kernel Density Estimation for Time Series Data of
    Harvey
  • Nonparametric Regression University
    Härdle
  • Regression by Local Fitting Properties Computational of
    Cleveland
  • Characterizing Nonlinearities in Business Cycles Using Smooth Transition AutoreGressive Models of
    Teräsvirta
  • The New Econometrics of Structural Change Dating Breaks in Labor Productivity The of
    Hansen
  • Exchange - rate Forecasts with Simultaneous Nearest - neighbor Evidence from the EMS of
    Fernandez
  • Cycle Measurement in the Presence of Structural Change : Evidence of
    Krolzig
  • Systems Overcoming the Pitfalls of Forecasting Models via a Multidisciplinary Approach of
    Orrell
  • Probabilistic Forecasts Calibration Sharpness of the
    Gneiting
  • Asymptotic Inference about Predictive Ability
    West
  • The Future of Macroeconomic Forecasting : Understanding the Forecasting Process of
    Stekler
  • Using Monthly Indicators to Predict Bank of Working Paper No
    Zheng
  • Econometric Modelling of Markov - switching Vector Autoregressions using MS VAR for Ox Discussion paper University of Oxford
    Krolzig
  • Forecasts Incremental Relative to Naive of
    Clemen
  • a Comparing Density Forecasts Using Threshold and Quantile - Weighted Scoring Rules of
    Gneiting
  • Modelling in National Product Empirical
    Teräsvirta
  • Kernel Density Estimation for Time Series Data of
    Harvey
  • Macro Variables and International Stock Return Predictability of
    Rapach
  • On the Empirical Determination of a Distribution Function Giornale dell Instituto Italiano degli
    Kolmogorov
  • New Approach to the Economic Analysis of Nonstationary Time Series and the Cycle
    Hamilton
  • Content for Univariate Time - series Forecasts of
    Galbraith
  • Strictly Scoring Rules Prediction Estimation of the
    Gneiting
  • Macro Variables and International Stock Return Predictability of
    Rapach
  • Modelling in National Product Empirical
    Teräsvirta
  • Theory and an Extension of the Maximum Likelihood Principle International on Information Theory edited by Budapest
    Akaike
  • Macroeconomic Variables and South African Stock Return Predictability University of Pretoria Working Paper
    Gupta
  • The of Statistical Learning Data Mining Inference Prediction nd ed Springer Series in
    Hastie
  • What Does the Yield Curve Tell us About GDP Growth of
    Ang
  • On the Empirical Determination of a Distribution Function Giornale dell Instituto Italiano degli
    Kolmogorov
  • Box Time Series Forecasting Control San
    Jenkins
  • Macroeconomic Forecasting in the Euro Area : Country Specific versus Area - wide
    Marcellino
  • Regression by Local Fitting Properties Computational of
    Cleveland
  • Comparing Predictive Accuracy of and
    Diebold
  • Characterizing Nonlinearities in Business Cycles Using Smooth Transition AutoreGressive Models of
    Teräsvirta
  • What Does the Yield Curve Tell us About GDP Growth of
    Ang
  • Exchange - rate Forecasts with Simultaneous Nearest - neighbor Evidence from the EMS of
    Fernandez
  • Nonlinear Models for Macroeconomic Time Series Studies in Nonlinear
    Öcal
  • The Likelihood Ratio Test under Nonstandard Conditions : Testing the Markov Switching Model of of
    Hansen
  • Monte Carlo Study of the Forecasting Performance of Empirical SETAR Models of
    Clements
  • New Approach to the Economic Analysis of Nonstationary Time Series and the Cycle
    Hamilton
  • Some Advances in Non - linear and Adaptive Modelling in Time - series of
    Tiao
  • Comparison of the Forecast Performance of Markov - switching and Threshold Autoregressive Models of US
    Clements
  • Monte Carlo Study of the Forecasting Performance of Empirical SETAR Models of
    Clements
  • Model Selection Approach to Assessing the Information in the Term Structure Using Linear Models and Artificial of and
    Swanson
  • The New Econometrics of Structural Change Dating Breaks in Labor Productivity The of
    Hansen
  • Combining Forecasts of
    Clemen
  • Evaluating Long horizon Forecasts University of Missouri Columbia Manuscript
    Clark
  • How Accurate are Private Sector Forecasts ? Cross - Country Evidence from Consensus Forecasts of Output Growth of
    Loungani
  • Forecasting Processes of
    Clements
  • Visual Goodness - of - Fit Test for Econometric Models Studies in Nonlinear
    Gençay
  • Comparing Predictive Accuracy of and
    Diebold
  • Is The Cycle Characterized by Deterministic of
    Brock
  • Macroeconomic Variables and South African Stock Return Predictability University of Pretoria Working Paper
    Gupta
  • Strictly Scoring Rules Prediction Estimation of the
    Gneiting
  • Some Advances in Non - linear and Adaptive Modelling in Time - series of
    Tiao
  • Model Selection Approach to Assessing the Information in the Term Structure Using Linear Models and Artificial of and
    Swanson
  • Tests for Forecast of
    Harvey
  • Asymptotics for Out - of - sample Tests of Causality of
    McCracken
  • Forecast - based Model Selection in the Presence of Structural Breaks Federal Reserve Bank of Kansas City
    Clark
  • Visual Goodness - of - Fit Test for Econometric Models Studies in Nonlinear
    Gençay
  • Local Linear Regression of the Statistical
    Yu
  • Using Conditional Kernel Density Estimation for Wind Power Density Forecasting of the
    Jeon
  • Floor and Ceiling Model of US Output of Economic Dynamics and
    Pesaran
  • Forecast Combination and the Bank of England s Suite of Statistical Forecasting Models Bank of England Working Paper no
    Kapetanios
  • Comparison of Linear and Nonlinear Univariate Models for Forecasting Macroeconomic Time Series InCh in Cointegration Causality and Forecasting Festschrift for Clive edited by University
    Stock
  • Tests of Equal Forecast Accuracy and Encompassing for Nested Models of
    Clark