An empirical investigation of direct and iterated multistep conditional forecasts

Abstract

Summary When constructing unconditional point forecasts, both direct and iterated multistep (DMS and IMS) approaches are common. However, in the context of producing conditional forecasts, IMS approaches based on vector autoregressions are far more common than simpler DMS models. This is despite the fact that there are theoretical reasons to believe that DMS models are more robust to misspecification than are IMS models. In the context of unconditional forecasts, Marcellino et al. (Journal of Econometrics, 2006, 135, 499–526) investigate the empirical relevance of these theories. In this paper, we extend that work to conditional forecasts. We do so based on linear bivariate and trivariate models estimated using a large dataset of macroeconomic time series. Over comparable samples, our results reinforce those in Marcellino et al.: the IMS approach is typically a bit better than DMS with significant improvements only at longer horizons. In contrast, when we focus on the Great Moderation sample we find a marked improvement in the DMS approach relative to IMS. The distinction is particularly clear when we forecast nominal rather than real variables where the relative gains can be substantial.

Deep Mobile Path Prediction With Shift-and-Join and Carry-Ahead

“…Early studies on multi-step prediction using direct approach are done with auto-regression model for forecasting and estimating indicators in economic area [17]. In various studies, results of direct approach are comparatively analyzed against recursive approach using nonlinear data [18], and it is concluded that the direct approach shows better results for long target sequences in terms of prediction error, parameter estimation error, and efficient finite data samples [19]. A study for vehicular route prediction defines mobility patterns as consecutive series of road segment selections.…”

Predicting Water Availability in Water Bodies under the Influence of Precipitation and Water Management Actions Using VAR/VECM/LSTM

“…They can be customized to produce a forecast for multiple intervals ahead. McCraken et al [27] talk about VAR models in detail, focusing on their utilization in direct and recursive forecasting. Further in his paper, McCraken demonstrated a comparison of two methods for conditional forecasting and concluded that the recursive method acts better than the direct method [27].…”

Multi-Day-Ahead Electricity Price Forecasting: A Comparison of Fundamental, Econometric and Hybrid Models

“…Marcellino et al (2006), Taieb and Atiya (2016)), empirical investigations in macroeconometrics have uncovered the opposite for both unconditional and conditional forecasts (e.g. Marcellino et al (2006), McCracken and McGillicuddy (2019)). To the best of our knowledge, no previous study has compared recursive and direct forecasting models in the context of EPF.…”