nep-for New Economics Papers
on Forecasting
Issue of 2014‒06‒14
eleven papers chosen by
Rob J Hyndman
Monash University

  1. Fast computation of reconciled forecasts for hierarchical and grouped time series By Rob J Hyndman; Alan Lee; Earo Wang
  2. Low-dimensional decomposition, smoothing and forecasting of sparse functional data By Alexander Dokumentov; Rob J Hyndman
  3. Forecasting US Real Private Residential Fixed Investment Using a Large Number of Predictors By Goodness C. Aye; Rangan Gupta; Stephen M. Miller; Mehmet Balcilar
  4. Forecasting financial market activity using a semiparametric fractionally integrated Log-ACD By Yuanhua Feng; Chen Zhou
  5. Interpreting Financial Market Crashes as Earthquakes: A New early Warning System for Medium Term Crashes By Francine Gresnigt; Erik Kole; Philip Hans Franses
  6. On Forecast Evaluation By Wilmer Osvaldo Martínez-Rivera; Manuel Dario Hernández-Bejarano; Juan Manuel Julio-Román
  7. Forecasting Multivariate Time Series under Present-Value-Model Short- and Long-run Co-movement Restrictions By Guillén, Osmani; Hecq, Alain; Issler, João Victor; Saraiva, Diogo
  8. The Naive Extrapolation Hypothesis and the Rosy-Gloomy Forecasts By Vasileios Barmpoutis
  9. Proportional loss functions for debris flow events By Christoph Rheinberger; Hans E. Romang; Michael Bründl
  10. Learning from the past: statistical performance measures for avalanche warning services By Christoph Rheinberger
  11. On the Fundamental Relation Between Equity Returns and Interest Rates By Jaewon Choi; Matthew P. Richardson; Robert F. Whitelaw

  1. By: Rob J Hyndman; Alan Lee; Earo Wang
    Abstract: We describe some fast algorithms for reconciling large collections of time series forecasts with aggregation constraints. The constraints arise due to the need for forecasts of collections of time series with hierarchical or grouped structures to add up in the same manner as the observed time series. We show that the least squares approach to reconciling hierarchical forecasts can be extended to more general non-hierarchical groups of time series, and that the computations can be handled efficiently by exploiting the structure of the associated design matrix. Our algorithms will reconcile hierarchical forecasts with hierarchies of unlimited size, making forecast reconciliation feasible in business applications involving very large numbers of time series.
    Keywords: combining forecasts, grouped time series, hierarchical time series, reconciling forecasts, weighted least squares.
    JEL: C32 C53 C63
    Date: 2014
  2. By: Alexander Dokumentov; Rob J Hyndman
    Abstract: We propose a new generic method ROPES (Regularized Optimization for Prediction and Estimation with Sparse data) for decomposing, smoothing and forecasting two-dimensional sparse data. In some ways, ROPES is similar to Ridge Regression, the LASSO, Principal Component Analysis (PCA) and Maximum-Margin Matrix Factorisation (MMMF). Using this new approach, we propose a practical method of forecasting mortality rates, as well as a new method for interpolating and extrapolating sparse longitudinal data. We also show how to calculate prediction intervals for the resulting estimates.
    Keywords: Tikhonov regularisation, Smoothing, Forecasting, Ridge regression, PCA, LASSO, Maximum-margin matrix factorisation, Mortality rates, Sparse longitudinal data
    JEL: C10 C14 C33
    Date: 2014
  3. By: Goodness C. Aye (University of Pretoria); Rangan Gupta (University of Pretoria); Stephen M. Miller (University of Nevada, Las Vegas and University of Connecticut); Mehmet Balcilar (Eastern Mediterranean University)
    Abstract: This paper employs classical bivariate, factor augmented (FA), slab-and-spike variable selection (SSVS)-based, and Bayesian semi-parametric shrinkage (BSS)-based predictive regression models to forecast US real private residential fixed investment over an out-of-sample period from 1983:Q1 to 2011:Q2, based on an in-sample estimates for 1963:Q1 to 1982:Q4. Both large-scale (188 macroeconomic series) and small-scale (20 macroeconomic series) FA, SSVS, and BSS predictive regressions, as well as 20 bivariate regression models, capture the influence of fundamentals in forecasting residential investment. We evaluate the ex-post out-of-sample forecast performance of the 26 models using the relative average Mean Square Error for one-, two-, four-, and eight-quarters-ahead forecasts and test their significance based on the McCracken (2004, 2007) MSE-F statistic. We find that, on average, the SSVS-Large model provides the best forecasts amongst all the models. We also find that one of the individual regression models, using house for sale (H4SALE) as a predictor, performs best at the four- and eight-quarters-ahead horizons. Finally, we use these two models to predict the relevant turning points of the residential investment, via an ex-ante forecast exercise from 2011:Q3 to 2012:Q4. The SSVS-Large model forecasts the turning points more accurately, although the H4SALE model does better toward the end of the sample. Our results suggest that economy-wide factors, in addition to specific housing market variables, prove important when forecasting in the real estate market.
    Keywords: Private residential investment, predictive regressions, factor-augmented models, Bayesian shrinkage, forecasting
    JEL: C32 E22 E27
    Date: 2014–05
  4. By: Yuanhua Feng (University of Paderborn); Chen Zhou (University of Paderborn)
    Abstract: This paper discusses forecasting of long memory and a nonparametric scale function in nonnegative financial processes based on a fractionally integrated Log-ACD (FI-Log-ACD) and its semiparametric extension (Semi-FI-Log-ACD). Necessary and sufficient conditions for the existence of a stationary solution of the FI-Log-ACD are obtained. Properties of this model under log-normal assumption are summarized. A linear predictor based on the truncated AR(oo) form of the logarithmic process is proposed. It is shown that this proposal is an approximately best linear predictor. Approximate variances of the prediction errors for an individual observation and for the conditional mean are obtained. Forecasting intervals for these quantities in the log- and the original processes are calculated under log-normal assumption. The proposals are applied to forecasting daily trading volumes and daily trading numbers in financial market.
    Keywords: Approximately best linear predictor, FI-Log-ACD, financial forecasting, long memory time series, nonparametric methods, Semi-FI-Log-ACD
    Date: 2013–04
  5. By: Francine Gresnigt (Erasmus University Rotterdam); Erik Kole (Erasmus University Rotterdam); Philip Hans Franses (Erasmus University Rotterdam)
    Abstract: We propose a modeling framework which allows for creating probability predictions on a future market crash in the medium term, like sometime in the next five days. Our framework draws upon noticeable similarities between stock returns around a financial market crash and seismic activity around earthquakes. Our model is incorporated in an Early Warning System for future crash days. Testing our EWS on S&P 500 data during the recent financial crisis, we find positive Hanssen-Kuiper Skill Scores. Furthermore our modeling framework is capable of exploiting information in the returns series not captured by well known and commonly used volatility models. EWS based on our models outperform EWS based on the volatility models forecasting extreme price movements, while forecasting is much less time-consuming.
    Keywords: Financial crashes; Hawkes process; self-exciting process; Early Warning System
    JEL: C13 C15 C53 G17
    Date: 2014–06–03
  6. By: Wilmer Osvaldo Martínez-Rivera; Manuel Dario Hernández-Bejarano; Juan Manuel Julio-Román
    Abstract: We propose to assess the performance of k forecast procedures by exploring the distributions of forecast errors and error losses. We argue that non systematic forecast errors minimize when their distributions are symmetric and unimodal, and that forecast accuracy should be assessed through stochastic loss order rather than expected loss order, which is the way it is customarily performed in previous work. Moreover, since forecast performance evaluation can be understood as a one way analysis of variance, we propose to explore loss distributions under two circumstances; when a strict (but unknown) joint stochastic order exists among the losses of all forecast alternatives, and when such order happens among subsets of alternative procedures. In spite of the fact that loss stochastic order is stronger than loss moment order, our proposals are at least as powerful as competing tests, and are robust to the correlation, autocorrelation and heteroskedasticity settings they consider. In addition, since our proposals do not require samples of the same size, their scope is also wider, and provided that they test the whole loss distribution instead of just loss moments, they can also be used to study forecast distributions as well. We illustrate the usefulness of our proposals by evaluating a set of real world forecasts.
    Keywords: Forecast evaluation, Stochastic order, Multiple comparison.
    JEL: C53 C12 C14
    Date: 2014–06–06
  7. By: Guillén, Osmani; Hecq, Alain; Issler, João Victor; Saraiva, Diogo
    Abstract: This paper has two original contributions. First, we show that the presentvalue model (PVM hereafter), which has a wide application in macroeconomicsand fi nance, entails common cyclical feature restrictions in the dynamics of thevector error-correction representation (Vahid and Engle, 1993); something thathas been already investigated in that VECM context by Johansen and Swensen (1999, 2011) but has not been discussed before with this new emphasis. Wealso provide the present value reduced rank constraints to be tested within thelog-linear model. Our second contribution relates to forecasting time seriesthat are subject to those long and short-run reduced rank restrictions. Thereason why appropriate common cyclical feature restrictions might improveforecasting is because it finds natural exclusion restrictions preventing theestimation of useless parameters, which would otherwise contribute to theincrease of forecast variance with no expected reduction in bias. We applied the techniques discussed in this paper to data known to besubject to present value restrictions, i.e. the online series maintained and up-dated by Shiller. We focus on three different data sets. The fi rst includes thelevels of interest rates with long and short maturities, the second includes thelevel of real price and dividend for the S&P composite index, and the thirdincludes the logarithmic transformation of prices and dividends. Our exhaustive investigation of several different multivariate models reveals that betterforecasts can be achieved when restrictions are applied to them. Moreover,imposing short-run restrictions produce forecast winners 70% of the time fortarget variables of PVMs and 63.33% of the time when all variables in thesystem are considered.
    Date: 2014–06–02
  8. By: Vasileios Barmpoutis
    Abstract: I study the behavior and the performance of the long-term forecasts issued by financial analysts with respect to the Extrapolation Hypothesis. That hypothesis states that investors, extrapolating from the firms' recent performances, are too optimistic about growth and large firms and too pessimistic about value and small firms. I find that the forecasting errors are higher for the growth firms and large firms, thus providing support for the Extrapolation Hypothesis. However, in addition to the rosy picture of the growth and large firms, the forecasts of the value and small firms are not so gloomy in many cases. My analysis also reveals that expectations move together for all categories of book-to-market and all sizes of firms. I proceed by investigating some common factors that may influence analysts' long-term forecasts, including co-movement and excessive optimism. I find that macro factors beyond a firm's recent performance may influence the formation of expectations.
    Date: 2014–06
  9. By: Christoph Rheinberger (Economie des Ressources Naturelles, INRA); Hans E. Romang (Swiss Federal Office of Meteorology and Climatology); Michael Bründl (WSL Institute for Snow and Avalanche Research SLF)
    Abstract: Quantitative risk assessments of debris flows and other hydrogeological hazards require the analyst to predict damage potentials. A common way to do so is by use of proportional loss functions. In this paper, we analyze a uniquely rich dataset of 132 buildings that were damaged in one of five large debris flow events in Switzerland. Using the double generalized linear model, we estimate proportional loss functions thatmay be used for various prediction purposes including hazard mapping, landscape planning, and insurance pricing. Unlike earlier analyses, we control for confounding effects of building characteristics, site specifics, and process intensities as well as for overdispersion in the data. Our results suggest that process intensity parameters are the most meaningful predictors of proportional loss sizes. Cross-validation tests suggest that the mean absolute prediction errors of our models are in the range of 11 %, underpinning the accurateness of the approach.
    Keywords: risk-assessment, landslide risk, vulnerability, damage, management, regression, hazards
    Date: 2013
  10. By: Christoph Rheinberger (Economie des Ressources Naturelles, INRA)
    Abstract: Avalanche warning services (AWS) are operated to protect communities and traffic lines in avalanche-prone regions of the Alps and other mountain ranges. In times of high avalanche danger, these services may decide to close roads or to evacuate settlements. Closing decisions are based on field observations, avalanche release statistics, and snow forecasts issued by weather services. Because of the spatial variability in the snowpack and the insufficient understanding of avalanche triggering mechanisms, closing decisions are characterized by large uncertainties and the information based on which AWS have to decide is always incomplete. In this paper, we illustrate how signal detection theory can be applied to make better use of the information at hand. The proposed framework allows the evaluation of past road closures and points to how the decision performance of AWS could be improved. To illustrate the proposed framework, we evaluate the decision performance of two AWS in Switzerland and discuss the advantages of such a formalized decisionmaking approach.
    Keywords: statistical performance measures, discriminating ability, signal detection theory, avalanche warning services, Performance statistiquesavalancheservice d'alerte
    Date: 2013
  11. By: Jaewon Choi; Matthew P. Richardson; Robert F. Whitelaw
    Abstract: This paper uses contingent claim asset pricing and exploits capital structure priority to better understand the relation between corporate security returns and interest rate changes (i.e., duration). We show theoretically and, using a novel dataset, confirm empirically that lower priority securities in the capital structure, such as subordinated or distressed debt and equity, have low or even negative durations because these securities are effectively short higher priority, high duration fixed rate debt. This finding has important implications for interpreting existing results on (i) the time-varying correlation between the aggregate stock market and government bonds, (ii) the use of bond factors for multifactor asset pricing models and forecasting bond and stock returns, (iii) the Fisher effect and inflation, and (iv) the betas of corporate bonds.
    JEL: G12 G13 G32
    Date: 2014–06

This nep-for issue is ©2014 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.