nep-for New Economics Papers
on Forecasting
Issue of 2022‒01‒03
three papers chosen by
Rob J Hyndman
Monash University

  1. Forecasting Regional GDPs: a Comparison with Spatial Dynamic Panel Data Models By Anna Gloria Billé; Alessio Tomelleri; Francesco Ravazzolo
  2. A Bayesian Spatio-temporal model for predicting passengers' occupancy at Beijing Metro By Cabras, Stefano; Sunhe, Flor
  3. Dividend Momentum and Stock Return Predictability: A Bayesian Approach By Juan Antolin-Diaz; Ivan Petrella; Juan F. Rubio-Ramirez

  1. By: Anna Gloria Billé; Alessio Tomelleri; Francesco Ravazzolo
    Abstract: The monitoring of the regional (provincial) economic situation is of particular importance due to the high level of heterogeneity and interdependences among different territories. Although econometric models allow for spatial and serial correlation of various kinds, the limited availability of territorial data restricts the set of relevant predictors at a more disaggregated level, especially for GDPs. This paper evaluates the predictive performance of a spatial dynamic panel data model with individual fixed effects and some relevant exogenous regressors by using data on total GVA for 103 Italian provinces (NUTS-3 level) over the period 2000-2016. A comparison with nested panel sub-specifications as well as pure temporal autoregressive specifications has also been included. The main finding is that the spatial dynamic specification increases forecast accuracy more than its competitors throughout the out-of-sample, recognizing an important role played by both space and time. However, when temporal cointegration is detected, the random walk specification is still to be preferred in some cases even in the presence of short panels.
    Keywords: Prediction,, Spatial Correlation, Panel Data, Regional GVA forecasting
    JEL: C33 C52 C53 E37 R11
    Date: 2021–12
  2. By: Cabras, Stefano; Sunhe, Flor
    Abstract: This work focuses on predicting metro passenger flow at Beijing Metro stations and assessing uncertainty using a Bayesian Spatio-temporal model. Forecasting is essential for Metro operation management, such as automatically adjusting train operation diagrams or crowd regulation planning measures. Different from another approach, the proposed model can provide prediction uncertainty conditionally on available data, a critical feature that makes this algorithm different from usual machine learning prediction algorithms. The Bayesian Spatio-temporal model for areal Poisson counts includes random effects for stations and days. The fitted model on a test set provides a prediction accuracy that meets the standards of the Beijing Metro enterprise.
    Keywords: Bayesian Modelling; Integrated Nested Laplace Approximation; Spatio-Temporal Modelling; Poisson Counts
    Date: 2021–12–16
  3. By: Juan Antolin-Diaz; Ivan Petrella; Juan F. Rubio-Ramirez
    Abstract: A long tradition in macro finance studies the joint dynamics of aggregate stock returns and dividends using vector autoregressions (VARs), imposing the cross-equation restrictions implied by the Campbell-Shiller (CS) identity to sharpen inference. We take a Bayesian perspective and develop methods to draw from any posterior distribution of a VAR that encodes a priori skepticism about large amounts of return predictability while imposing the CS restrictions. In doing so, we show how a common empirical practice of omitting dividend growth from the system amounts to imposing the extra restriction that dividend growth is not persistent. We highlight that persistence in dividend growth induces a previously overlooked channel for return predictability, which we label "dividend momentum." Compared to estimation based on ordinary least squares, our restricted informative prior leads to a much more moderate, but still significant, degree of return predictability, with forecasts that are helpful out of sample and realistic asset allocation prescriptions with Sharpe ratios that outperform common benchmarks.
    Keywords: CS restrictions; Bayesian VAR; optimal allocation
    JEL: C32 C53 G11 G12 E47
    Date: 2021–11–10

This nep-for issue is ©2022 by Rob J Hyndman. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.