nep-for New Economics Papers
on Forecasting
Issue of 2017‒06‒04
six papers chosen by
Rob J Hyndman
Monash University

  1. Linear and Nonlinear Predictability in Investment Style Factors: Multivariate Evidence* By Massimo Guidolin; Francesco Chincoli
  2. Density Forecasts of Polish Industrial Production: a Probabilistic Perspective on Business Cycle Fluctuations By Blazej Mazur
  3. Forecasting GDP Growth with NIPA Aggregates By Knotek, Edward S.; Garciga, Christian
  5. Are Multifractal Processes Suited to Forecasting Electricity Price Volatility? Evidence from Australian Intraday Data By Mawuli Segnon; Chi Keung Lau; Bernd Wilfling; Rangan Gupta
  6. Prediction Bands for Functional Data Based on Depth Measures By Elías Fernández, Antonio; Jiménez Recaredo, Raúl José

  1. By: Massimo Guidolin; Francesco Chincoli
    Abstract: This paper studies the predictive performance of multivariate models at forecasting the (excess) returns of portfolios mimicking the Market, Size, Value, Momentum, and Low Volatility factors isolated in asset pricing research. We evaluate the accuracy of the point forecasts of a number of linear and regime switching models in recursive, out-of-sample forecasting experiments. We assess the accuracy of the models using several measures of unbiasedness and predictive accuracy, and, using Diebold and Mariano’s approach to test whether differences in expected losses from all possible pairs of forecast models are statistically significant. We fail to find evidence that complex statistical models are uniformly more accurate than a naïve constant expected return model for factor-mimicking portfolio (excess) returns. However, we show that it is possible to build simple portfolio strategies that profit from the higher out-of-sample predictive accuracy of forecasting models with Markov switching in conditional mean coefficients. These results appear to be independent of the forecasting horizon and robust to changes in the loss function that captures the investors’ objectives.
    Keywords: Factor mimicking portfolios, forecasting, Markov regime switching models, equal predictive accuracy tests.
    JEL: G11 G12 C32
    Date: 2017
  2. By: Blazej Mazur (Cracow University of Economics, Poland)
    Abstract: Current approaches used in empirical macroeconomic analyses use the probabilistic setup and focus on evaluation of uncertainties and risks, also with respect to future business cycle fluctuations. Therefore, forecast-based business conditions indicators should be constructed using not just point forecasts, but rather density forecasts. The latter represent whole predictive distribution and provide relevant description of forecast uncertainty.We discuss a problem of model-based probabilistic inference on business cycle conditions in Poland. In particular we consider a model choice problem for density forecasts of Polish monthly industrial production index and its selected sub-indices. Based on the results we develop indicators of future economic conditions constructed using probabilistic information on future values of the index. In order to develop a relevant model class we make use of univariate Dynamic Conditional Score models with Bayesian inference methods. We assume that the conditional distribution is of the generalized t form in order to allow for heavy tails. Another group of models under consideration relies on the idea of business cycle modelling using the Flexible Fourier Form. We compare performance of alternative models based on ex-post evaluation of density forecasting accuracy using such criteria as Log-Predictive Score (LPS) and Continuous Ranked Probability Score (CRPS). The assessment of density forecasting performance for Polish industrial production index turns out to be difficult since it depends on the choice of verification window. The pre-2013 data supports the deterministic cycle model whereas more recent observations can be explained by a very simple mean-reverting Gaussian AR(4) process. This provides an indirect evidence indicating the change of pattern of Polish business cycle fluctuations after 2013. A probabilistic indicator of business conditions is also sensitive to details of its construction. The results suggest application of forecast pooling strategies as a goal for further research.
    Keywords: density forecasts; Bayesian inference; business cycle; Dynamic Conditional Score models; Generalized t distribution.
    JEL: E37 C53
    Date: 2017–05
  3. By: Knotek, Edward S. (Federal Reserve Bank of Cleveland); Garciga, Christian (Federal Reserve Bank of Cleveland)
    Abstract: Beyond GDP, which is measured using expenditure data, the U.S. national income and product accounts (NIPAs) provide an income-based measure of the economy (gross domestic income, or GDI), a measure that averages GDP and GDI, and various aggregates that include combinations of GDP components. This paper compiles real-time data on a variety of NIPA aggregates and uses these in simple time-series models to construct out-of-sample forecasts for GDP growth. Over short forecast horizons, NIPA aggregates—particularly consumption and GDP less inventories and trade—together with these simple time-series models have historically generated more accurate forecasts than a canonical AR(2) benchmark. This has been especially true during recessions, although we document modest gains during expansions as well.
    Keywords: forecasting; GDP; GDI; real-time data; consumption;
    JEL: C32 C53 E01
    Date: 2017–05–19
  4. By: Davide De Gaetano
    Abstract: This paper proposes some weighting schemes to average forecasts across different estimation windows to account for structural changes in the unconditional variance of a GARCH (1,1) model. Each combination is obtained by averaging forecasts generated by recursively increasing an initial estimation window of a fixed number of observations v. Three different choices of the combination weights are proposed. In the first scheme, the forecast combination is obtained by using equal weights to average the individual forecasts; the second weighting method assigns heavier weights to forecasts that use more recent information; the third is a trimmed version of the forecast combination with equal weights where a fixed fraction of forecasts with the worst performance are discarded. Simulation results show that forecast combinations with high values of v are able to perform better than alternative schemes proposed in the literature. An application to real data confirms the simulation results
    Keywords: Forecast combinations, Structural breaks, GARCH models.
    JEL: C53 C58 G17
    Date: 2017–05
  5. By: Mawuli Segnon (Westfälische Wilhelms-Universität Münster, Department of Economics (CQE), Germany and Mark E AG, Germany); Chi Keung Lau (Newcastle Business School, Department of Economics and Finance, UK); Bernd Wilfling (Westfälische Wilhelms-Universität Münster, Department of Economics (CQE), Germany); Rangan Gupta (Department of Economics, University of Pretoria, Pretoria, South Africa)
    Abstract: We analyze Australian electricity price returns and find that they exhibit multifractal structures. Consequently, we let the return mean equation follow a long memory smooth transition autoregressive (STAR) process and specify volatility dynamics as a Markov-switching multifractal (MSM) process. We compare the out-of-sample volatility forecasting performance of the STAR-MSM model with that of other STAR mean processes, combined with various conventional GARCH-type volatility equations (for example, STAR-GARCH(1,1)). We find that the STAR-MSM model competes with conventional STAR-GARCH specifications with respect to volatility forecasting, but does not (systematically) outperform them.
    Keywords: Electricity price volatility; multifractal modeling; GARCH processes; volatility forecasting
    JEL: C22 C52 C53
    Date: 2017–05
  6. By: Elías Fernández, Antonio; Jiménez Recaredo, Raúl José
    Abstract: We propose a new methodology for predicting a partially observed curve from a functional data sample. The novelty of our approach relies on the selection of sample curves which form tight bands that preserve the shape of the curve to predict, making this a deep datum. The involved subsampling problem is dealt by algorithms specially designed to be used in conjunction with two different tools for computing central regions for functional data. From this merge, we obtain prediction bands for the unobserved part of the curve in question. We test our algorithms by forecasting the Spanish electricity demand and imputing missing daily temperatures. The results are consistent with our simulation that show that we can predict at the far horizon.
    Keywords: daily temperatures; electricity demand; central regions; depth measures
    Date: 2017–05

This nep-for issue is ©2017 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.