nep-for New Economics Papers
on Forecasting
Issue of 2016‒04‒04
twelve papers chosen by
Rob J Hyndman
Monash University

  1. Grouped functional time series forecasting: An application to age-specific mortality rates By Han Lin Shang; Rob J Hyndman
  2. Bayesian Rank Selection in Multivariate Regression By Bin Jiang; Anastasios Panagiotelis; George Athanasopoulos; Rob Hyndman; Farshid Vahid
  3. A Frequency Approach to Bayesian Asymptotics By Tingting Cheng; Jiti Gao; Peter CB Phillips
  4. A generalized exponential time series regression model for electricity prices By Niels Haldrup; Oskar Knapik; Tommaso Proietti
  5. Stock Return Predictability: Evaluation based on Prediction Intervals By Charles, Amelie; Darne, Olivier; Kim, Jae
  6. Box-Jenkins modelling and forecasting of Brent crude oil price By Mensah, Emmanuel Kwasi
  7. Foreign PMIs: A reliable indicator for exports? By Sandra Hanslin; Rolf Scheufele
  8. On the importance of the long-term seasonal component in day-ahead electricity price forecasting By Jakub Nowotarski; Rafal Weron
  9. Adaptive shrinkage in Bayesian vector autoregressive models By Florian Huber; Martin Feldkircher
  10. Long-Term Water Demand Forecasting By Jean-Daniel Rinaudo
  11. When the walk is not random: commodity prices and exchange rates By Emanuel Kohlscheen; Fernando Avalos; Andreas Schrimpf
  12. Forecasting with Large Unbalanced Datasets: The Mixed Frequency Three-Pass Regression Filter By Christian Hepenstrick; Massimiliano Marcellino

  1. By: Han Lin Shang; Rob J Hyndman
    Abstract: Age-specific mortality rates are often disaggregated by different attributes, such as sex, state and ethnicity. Forecasting age-specific mortality rates at the national and sub-national levels plays an important role in developing social policy. However, independent forecasts of age-specific mortality rates at the sub-national levels may not add up to the forecasts at the national level. To address this issue, we consider the problem of reconciling age-specific mortality rate forecasts from the viewpoint of grouped univariate time series forecasting methods (Hyndman, Ahmed, et al., 2011), and extend these methods to functional time series forecasting, where age is considered as a continuum. The grouped functional time series methods are used to produce point forecasts of mortality rates that are aggregated appropriately across different disaggregation factors. For evaluating forecast uncertainty, we propose a bootstrap method for reconciling interval forecasts. Using the regional age-specific mortality rates in Japan, obtained from the Japanese Mortality Database, we investigate the one- to ten-step-ahead point and interval forecast accuracies between the independent and grouped functional time series forecasting methods. The proposed methods are shown to be useful for reconciling forecasts of age-specific mortality rates at the national and sub-national levels, and they also enjoy improved forecast accuracy averaged over different disaggregation factors.
    Keywords: Forecast reconciliation, hierarchical time series forecasting, bottom-up, optimal combination, Japanese Mortality Database
    JEL: C14 C32 J11
    Date: 2016
  2. By: Bin Jiang; Anastasios Panagiotelis; George Athanasopoulos; Rob Hyndman; Farshid Vahid
    Abstract: Estimating the rank of the coefficient matrix is a major challenge in multivariate regression, including vector autoregression (VAR). In this paper, we develop a novel fully Bayesian approach that allows for rank estimation. The key to our approach is reparameterizing the coefficient matrix using its singular value decomposition and conducting Bayesian inference on the decomposed parameters. By implementing a stochastic search variable selection on the singular values of the coefficient matrix, the ultimate selected rank can be identified as the number of nonzero singular values. Our approach is appropriate for small multivariate regressions as well as for higher dimensional models with up to about 40 predictors. In macroeconomic forecasting using VARs, the advantages of shrinkage through proper Bayesian priors are well documented. Consequently, the shrinkage approach proposed here that selects or averages over low rank coefficient matrices is evaluated in a forecasting environment. We show in both simulations and empirical studies that our Bayesian approach provides forecasts that are better than those of the most promising benchmark methods, dynamic factor models and factor augmented VARs.
    Keywords: Singular value decomposition, model selection, vector autoregression, macroeconomic forecasting, dynamic factor models
    JEL: C11 C52 C53
    Date: 2016
  3. By: Tingting Cheng; Jiti Gao; Peter CB Phillips
    Abstract: Ergodic theorem shows that ergodic averages of the posterior draws converge in probability to the posterior mean under the stationarity assumption. The literature also shows that the posterior distribution is asymptotically normal when the sample size of the original data considered goes to infinity. To the best of our knowledge, there is little discussion on the large sample behaviour of the posterior mean. In this paper, we aim to fill this gap. In particular, we extend the posterior mean idea to the conditional mean case, which is conditioning on a given summary statistics of the original data. We stablish a new asymptotic theory for the conditional mean estimator for the case when both the sample size of the original data concerned and the number of Markov chain Monte Carlo iterations go to infinity. Simulation studies show that this conditional mean estimator has very good finite sample performance. In addition, we employ the conditional mean estimator to estimate a GARCH(1,1) model for S&P 500 stock returns and find that the conditional mean estimator performs better than quasi-maximum likelihood estimation in terms of out-of-sample forecasting.
    Keywords: Bayesian average, conditional mean estimation, ergodic theorem, summary statistic
    JEL: C11 C15 C21
    Date: 2016
  4. By: Niels Haldrup (Aarhus University and CREATES); Oskar Knapik (Aarhus University and CREATES); Tommaso Proietti (University of Rome “Tor Vergata” and Creates)
    Abstract: We consider the issue of modeling and forecasting daily electricity spot prices on the Nord Pool Elspot power market. We propose a method that can handle seasonal and non-seasonal persistence by modelling the price series as a generalized exponential process. As the presence of spikes can distort the estimation of the dynamic structure of the series we consider an iterative estimation strategy which, conditional on a set of parameter estimates, clears the spikes using a data cleaning algorithm, and reestimates the parameters using the cleaned data so as to robustify the estimates. Conditional on the estimated model, the best linear predictor is constructed. Our modeling approach provides good fit within sample and outperforms competing benchmark predictors in terms of forecasting accuracy. We also find that building separate models for each hour of the day and averaging the forecasts is a better strategy than forecasting the daily average directly.
    Keywords: Robust estimation, long-memory, seasonality, electricity spot prices, Nord Pool power market, forecasting, robust Kalman lter, generalized exponential model
    JEL: C1 C5 C53 Q4
    Date: 2016–03–18
  5. By: Charles, Amelie; Darne, Olivier; Kim, Jae
    Abstract: This paper evaluates the predictability of monthly stock return using out-of-sample (multi-step ahead and dynamic) prediction intervals. Past studies have exclusively used point forecasts, which are of limited value since they carry no information about the intrinsic predictive uncertainty associated. We compare empirical performances of alternative prediction intervals for stock return generated from a naive model, univariate autoregressive model, and multivariate model (predictive regression and VAR), using the U.S. data from 1926. For evaluation free from data snooping bias, we adopt moving sub-sample windows of different lengths. It is found that the naive model often provides the most informative prediction intervals, outperforming those generated from the univariate model and multivariate models incorporating a range of economic and financial predictors. This strongly suggests that the U.S. stock market has been informationally efficient in the weak-form as well as in the semi-strong form, subject to the information set considered in this study
    Keywords: Autoregressive Model, Bootstrapping, Financial Ratios, Forecasting, Interval Score, Market Efficiency
    JEL: G12 G14
    Date: 2016–03–18
  6. By: Mensah, Emmanuel Kwasi
    Abstract: The volatility in the crude oil price in the international market has risen much interest into the investigation of its price swing. In this project, we examine the dynamics of the monthly Brent oil price for the last two decades using the Box Jenkins ARIMA techniques and show that such model is not able to capture the volatility inherent in the crude oil price for an accurate forecast. We first divided the data into two. The first seventeen years used for the model construction and the last three years validating forecasting accuracy. The data is first differenced for stationarity and autocorrelation and residuals techniques used to select different ARIMA models for analysis. The performance of different models were compared and the result shows that a non-parsimonious ARIMA (1,1,1) model was the best forecasting model amidst the volatilities in the oil price.
    Keywords: Brent crude oil, ARIMA, stationarity, forecasting
    JEL: C22 C51 C52 E37
    Date: 2015–02
  7. By: Sandra Hanslin; Rolf Scheufele
    Abstract: Foreign economic activity is a major determinant of export development. This paper presents an indicator for now- and forecasting exports, which is based on survey data that captures foreign economic perspectives. We construct an indicator by weighting foreign PMIs of main trading partners with their respective export shares. For two very trade exposed countries (Germany and Switzerland) the paper shows that the indicator based on foreign PMIs is strongly correlated with exports (total as well as goods exports). In an out-of-sample forecast comparison we employ MIDAS models to forecast the two different definitions of exports. We document that our export indicator performs very well relative to univariate benchmarks and relative to other major leading indicators using hard and soft data.
    Keywords: Business tendency surveys, mixed frequencies, nowcasting, forecasting, MIDAS, exports
    JEL: F14 F17 C53
    Date: 2016
  8. By: Jakub Nowotarski; Rafal Weron
    Abstract: In day-ahead electricity price forecasting (EPF) the daily and weekly seasonalities are always taken into account, but the long-term seasonal component (LTSC) is believed to add unnecessary complexity to the already parameter-rich models and is generally ignored. Conducting an extensive empirical study involving state-of-the-art time series models we show that (i) decomposing a series of electricity prices into a LTSC and a stochastic component, (ii) modeling them independently and (iii) combining their forecasts can bring - contrary to a common belief - an accuracy gain compared to an approach in which a given time series model is calibrated to the prices themselves.
    Keywords: Electricity spot price; Forecasting; Day-ahead market; Long-term seasonal component
    JEL: C14 C22 C51 C53 Q47
    Date: 2016–03–15
  9. By: Florian Huber (Department of Economics, Vienna University of Economics and Business); Martin Feldkircher (Oesterreichische Nationalbank (OeNB))
    Abstract: Vector autoregressive (VAR) models are frequently used for forecasting and impulse response analysis. For both applications, shrinkage priors can help improving inference. In this paper we derive the shrinkage prior of Griffin et al. (2010) for the VAR case and its relevant conditional posterior distributions. This framework imposes a set of normally distributed priors on the autoregressive coefficients and the covariances of the VAR along with Gamma priors on a set of local and global prior scaling parameters. This prior setup is then generalized by introducing another layer of shrinkage with scaling parameters that push certain regions of the parameter space to zero. A simulation exercise shows that the proposed framework yields more precise estimates of the model parameters and impulse response functions. In addition, a forecasting exercise applied to US data shows that the proposed prior outperforms other specifications in terms of point and density predictions.
    Keywords: Normal-Gamma prior, density predictions, hierarchical modeling
    JEL: C11 C30 C53 E52
    Date: 2016–03
  10. By: Jean-Daniel Rinaudo (BRGM - Bureau de Recherches Géologiques et Minières)
    Abstract: This chapter reviews existing long term water demand forecasting methodologies. Based on an extensive literature review, it shows that the domain has benefited from contributions by economists, engineers and system modelers, producing a wide range of tools, many of which have been tested and adopted by practitioners. It illustrates, via three detailed case studies in the USA, the UK and Australia, how different tools can be used depending on the regulatory context, the water scarcity level, the geographic scale at which they are deployed and the technical background of water utilities and their consultants. The chapter reviews how practitioners address three main challenges, namely the integration of land use planning with demand forecasting; accounting for climate change; and dealing with forecast uncertainty. It concludes with a discussion of research perspectives in that domain.
    Keywords: Water demand
    Date: 2015
  11. By: Emanuel Kohlscheen; Fernando Avalos; Andreas Schrimpf
    Abstract: We show that there is a distinct commodity-related driver of exchange rate movements, even at fairly high frequencies. Commodity prices predict exchange rate movements of 11 commodity-exporting countries in an in-sample panel setting for horizons up to two months. We also find evidence of systematic (pseudo) out-of-sample predictability, overturning the results of Meese and Rogoff (1983): information embedded in our country-specific commodity price indices clearly helps improving upon the predictive accuracy of the random walk in the majority of countries. We further show that the link between commodity prices and exchange rates is not driven by changes in global risk appetite or carry.
    Keywords: commodities, exchange rates, interest rates
    Date: 2016–03
  12. By: Christian Hepenstrick; Massimiliano Marcellino
    Abstract: In this paper, we propose a modification of the three-pass regression filter (3PRF) to make it applicable to large mixed frequency datasets with ragged edges in a forecasting context. The resulting method, labeled MF-3PRF, is very simple but compares well to alternative mixed frequency factor estimation procedures in terms of theoretical properties, finite samle performance in Monte Carlo experiments, and empirical applications to GDP growth nowcasting and forecasting for the USA and a variety of other countries.
    Keywords: Dynamic Factor Models, Mixed Frequency, GDP Nowcasting, Forecasting, Partial Least Squares
    JEL: E37 C32 C53
    Date: 2016

This nep-for issue is ©2016 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.