A Comparison of Direct and Iterated Multistep Ar Methods for Forecasting Macroeconomic Time Series

By Massimiliano Marcellino, James H. Stock, Mark W. Watson. 2006

794 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Management Science and Operations Research

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Economics and Econometrics

In Management Science and Operations Research

Citations in the corpus

Time-series forecasting with deep learning: a survey (2020, methodology): “re inputs (e.g. date information, such as the day-of-week or month) across the entire horizon, and xtare inputs that can only be observed historically. In line with traditional econometric approaches [57, 58], deep learning architectures for multi-horizon forecasting can be divided into iterative an”
Temporal Fusion Transformers for Interpretable Multi-horizon Time Series Forecasting (2019, methodology): “In line with traditional methods for multi-horizon forecasting [29, 35], recent deep learning methods can be categorized into iterated approaches using autoregressive models [15, 25, 31] or direct methods using sequence-to-sequence models [13, 39].”
Real-Time Inflation Forecasting in a Changing World (2009, neutral): “Whether an iterative procedure provides more accurate forecasts than a direct approach is a matter of ongoing debate, see the discussion in Marcellino et al. (2006). (10)That is, if one defines the out-of-sample forecast error of a model for yt+h as 2̂t+h then RMSE = √√√√ 1 T − t0 − h T−h ∑”
Realising the future: forecasting with high frequency based volatility (HEAVY) models (2010, methodology): “Marcellino et al. (2006) provide an extensive discussion of the literature.”
Macroeconomic forecast accuracy in a data‐rich environment (2019, result): “Marcellino et al. (2006) conclude that the direct approach provides slightly better results but does not dominate uniformly across time and series.”
Forecast Combination and Model Averaging Using Predictive Measures (2007, neutral): “While the literature on forecast combination is extensive, see Clemen (1989) and Timmermann (2006) for reviews, and Hendry and Clements (2004), and Elliott and Timmermann (2004) for recent theoretical contributions, relatively little attention has been given to the use of predictive measures of fit ”
Multi-Period Forecasts of Volatility: Direct, Iterated, and Mixed-Data Approaches (2009, background): “…is gratefully acknowledged. yDepartment of Finance, Kenan-Flagler Business School and Department of Economics, University of North Carolina, Gardner Hall CB 3305, Chapel Hill, NC 27599-3305, e-mail: eghysels@unc.edu. zUniversity of Alicante, Campus de San Vicente, CP 03080, Alicante, Spain.”
Non-Parametric Direct Multi-step Estimation for Forecasting Economic Processes (2005, background): “However, as seen in Sargan (1982) and Hendry (1991), the non-existence of the moments does not preclude sensible Monte Carlo results, which can be seen as analogous to Nagar (1959) approximations.”
Credit Market Shocks and Economic Fluctuations: Evidence from Corporate Bond and Stock Markets (2009, result): “The results from forecasting exercises that rely on credit spreads in our EDF-based bond portfolios indicate that most of the predictive power of these default-risk indicators comes from the middle of the credit-quality spectrum, a result consistent with that of Mueller [2007] who examines the predi”
Bayesian VARs : specification choices and forecast accuracy (2011, positive): “For a discussion and a comparison of these alternative methods in a classical context see, for example, Marcellino et al. (2006) and Pesaran et al. (2011).”
Unrestricted mixed data sampling (MIDAS): MIDAS regressions with unrestricted lag polynomials (2015, methodology): “Following for example Marcellino et al. (2006), a simpler approach is to use a form of direct estimation and to construct the forecast as ỹTk+k|Tk = c̃.”
Local Projections vs. VARs: Lessons From Thousands of DGPs (2021, methodology): “This observation suggests the presence of a bias-variance trade-off whenever the true DGP is not a finite-order VAR, perfectly analogous to the choice between “direct” and “iterated” predictions in multi-step forecasting (Marcellino et al., 2006).”