nep-for New Economics Papers
on Forecasting
Issue of 2013‒09‒28
six papers chosen by
Rob J Hyndman
Monash University

  1. On the Usefulness of Constant Gain Least Squares when Forecasting the Unemployment Rate By Antipin, Jan-Erik; Boumediene, Farid Jimmy; Österholm, Pär
  2. Short-term forecasts of French GDP: A dynamic factor model with targeted predictors. By Bessec, Marie
  3. Credible Granger-Causality Inference with Modest Sample Lengths: A Cross-Sample Validation Approach By Richard A. Ashley; Kwok Ping Tsang
  4. On Measuring Time Preferences By James Andreoni; Michael A. Kuhn; Charles Sprenger
  5. Which Parametric Model for Conditional Skewness? By Bruno Feunou; Mohammad R. Jahan-Parvar; Roméo Tedongap
  6. Promptness and Academic Performance By Novarese, Marco; Di Giovinazzo, Viviana

  1. By: Antipin, Jan-Erik (Finnish Tax Administration); Boumediene, Farid Jimmy (Confederation of Swedish Enterprise); Österholm, Pär (National Institute of Economic Research)
    Abstract: In this paper, we assess the usefulness of constant gain least squares (CGLS) when forecasting the unemployment rate. Using quarterly data from 1970 to 2009, we conduct an out-of-sample forecast exercise in which univariate autoregressive models for the unemployment rate in Australia, Sweden, the United Kingdom and the United States are em-ployed. Results show that CGLS very rarely outperforms OLS. At horizons of six to eight quarters, OLS is always associated with higher forecast precision, regardless of model size or gain employed for Australia, Sweden and the United States. Our findings suggest that while CGLS has been shown valuable when forecasting certain mac-roeconomic time series, it has shortcomings when forecasting the unemployment rate. One problematic feature is found to be an increased tendency for the autoregressive model to have explosive dynamics when estimated with CGLS.
    Keywords: Out-of-sample; forecasts
    JEL: E24 E27
    Date: 2013–09–10
  2. By: Bessec, Marie
    Abstract: In recent years, factor models have received increasing attention from both econometricians and practitioners in the forecasting of macroeconomic variables. In this context, Bai and Ng (2008) find an improvement in selecting indicators according to the forecast variable prior to factor estimation (targeted predictors). In particular, they propose using the LARS-EN algorithm to remove irrelevant predictors. In this paper, we adapt the Bai and Ng procedure to a setup in which data releases are delayed and staggered. In the pre-selection step, we replace actual data with estimates obtained on the basis of past information, where the structure of the available information replicates the one a forecaster would face in real time. We estimate on the reduced dataset the dynamic factor model of Giannone, Reichlin and Small (2008) and Doz, Giannone and Reichlin (2011), which is particularly suitable for the very short-term forecast of GDP. A pseudo real-time evaluation on French data shows the potential of our approach.
    Keywords: Factor model; GDP forecasting; Large dataset; Targeted predictors; Variable selection;
    JEL: C22 E32 E37
    Date: 2013–09
  3. By: Richard A. Ashley; Kwok Ping Tsang
    Abstract: Credible Granger-causality analysis appears to require post-sample inference, as it is well-known that in-sample fit can be a poor guide to actual forecasting effectiveness. But post-sample model testing requires an often-consequential a priori partitioning of the data into an 'in-sample' period - purportedly utilized only for model specifi- cation/estimation - and a 'post-sample' period, purportedly utilized (only at the end of the analysis) for model validation/testing purposes. This partitioning is usually infeasible, however, with samples of modest length – e.g., T less than 100 - as is common in both quarterly data sets and/or in monthly data sets where institutional arrange- ments vary over time, simply because there is in such cases insufficient data available to credibly accomplish both purposes separately. A cross-sample validation (CSV) testing procedure is proposed below which substantially ameliorates this predicament - preserving most of the power of in-sample testing (by utilizing all of the sample data in the test), while also retaining most of the credibility of post-sample testing (by al- ways basing model forecasts on data not utilized in estimating that particular model's coefficients). Simulations show that the price paid, in terms of power relative to the in-sample Granger-causality F test, is manageable. An illustrative application is given, to a re-analysis of the Engel and West (2005) study of the causal relationship between macroeconomic fundamentals and the exchange rate.
    Keywords: Time Series, Granger-causality, causality, post-sample testing, exchange rates.
    Date: 2013
  4. By: James Andreoni; Michael A. Kuhn; Charles Sprenger
    Abstract: Eliciting time preferences has become an important component of both laboratory and field experiments, yet there is no consensus as how to best measure discounting. We examine the predictive validity of two recent, simple, easily administered, and individually successful elicitation tools: Convex Time Budgets (CTB) and Double Multiple Price Lists (DMPL). Using similar methods, the CTB and DMPL are compared using within- and out-of-sample predictions. While each perform equally well within sample, the CTB significantly outperforms the DMPL on out-of-sample measures.
    JEL: D03 D14 G02
    Date: 2013–08
  5. By: Bruno Feunou; Mohammad R. Jahan-Parvar; Roméo Tedongap
    Abstract: This paper addresses an existing gap in the developing literature on conditional skewness. We develop a simple procedure to evaluate parametric conditional skewness models. This procedure is based on regressing the realized skewness measures on model-implied conditional skewness values. We find that an asymmetric GARCH-type specification on shape parameters with a skewed generalized error distribution provides the best in-sample fit for the data, as well as reasonable predictions of the realized skewness measure. Our empirical findings imply significant asymmetry with respect to positive and negative news in both conditional asymmetry and kurtosis processes.
    Keywords: Econometric and statistical methods
    JEL: C22 C51 G12 G15
    Date: 2013
  6. By: Novarese, Marco; Di Giovinazzo, Viviana
    Abstract: This article uses university administration data to investigate the relation between student behavior (rapid response in finalizing enrolment procedures) and academic performance. It shows how student promptness in enrolling, or lack of it, can prove a useful forecast of academic success. Several explanations can be given, including simply the greater or lesser tendency to procrastinate.
    Keywords: procrastination, academic performance, motivation
    JEL: D83 D99 I21
    Date: 2013–09

This nep-for issue is ©2013 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.