nep-for New Economics Papers
on Forecasting
Issue of 2014‒06‒28
eleven papers chosen by
Rob J Hyndman
Monash University

  1. Forecast rationality tests in the presence of instabilities, with applications to Federal Reserve and survey forecasts By Barbara Rossi; Tatevik Sekhposyan
  2. Forecasting Short-Term Real GDP Growth in the Euro Area and Japan Using Unrestricted MIDAS Regressions By Maxime Leboeuf; Louis Morel
  3. The role in index jumps and cojumps in forecasting stock index volatility: Evidence from the Dow Jones index By Adam Clements; Yin Liao
  4. Forecasting with the Standardized Self-Perturbed Kalman Filter By Stefano Grassi; Nima Nonejad; Paolo Santucci de Magistris
  5. A fear index to predict oil futures returns By Julien Chevallier; Benoit Sevi
  6. Dynamic State-Space Models By Karapanagiotidis, Paul
  7. Coping with area price risk in electricity markets: Forecasting Contracts for Difference in the Nordic power market By Egil Ferkingstad; Anders L{\o}land
  8. Exact fit of simple finite mixture models By Dirk Tasche
  9. Generalized Forecast Error Variance Decomposition for Linear and Nonlinear Multivariate Models By Markku Lanne; Henri Nyberg
  10. Open-economy Distribution Forecast Targeting, Macroeconomic Volatility and Financial Implication By Alessandro Flamini; Iftekhar Hasan; Costas Milas
  11. Empirical evidence for nonlinearity and irreversibility of commodity futures prices By Karapanagiotidis, Paul

  1. By: Barbara Rossi; Tatevik Sekhposyan
    Abstract: This paper proposes a framework to implement regression-based tests of predictive ability in unstable environments, including, in particular, forecast unbiasedness and efficiency tests, commonly referred to as tests of forecast rationality. Our framework is general: it can be applied to model-based forecasts obtained either with recursive or rolling window estimation schemes, as well as to forecasts that are model-free. The proposed tests provide more evidence against forecast rationality than previously found in the Federal Reserve's Greenbook forecasts as well as survey-based private forecasts. It confirms, however, that the Federal Reserve has additional information about current and future states of the economy relative to market participants.
    Keywords: Forecasting, forecast rationality, regression-based tests of forecasting ability, Greenbook forecasts, survey forecasts, real-time data
    JEL: C22 C52 C53
    Date: 2014–06
  2. By: Maxime Leboeuf; Louis Morel
    Abstract: In this paper, the authors develop a new tool to improve the short-term forecasting of real GDP growth in the euro area and Japan. This new tool, which uses unrestricted mixed-data sampling (U-MIDAS) regressions, allows an evaluation of the usefulness of a wide range of indicators in predicting short-term real GDP growth. In line with previous Bank studies, the results suggest that the purchasing managers’ index (PMI) is among the best-performing indicators to forecast real GDP growth in the euro area, while consumption indicators and business surveys (the PMI and the Economy Watchers Survey) have the most predictive power for Japan. Moreover, the results indicate that combining the predictions from a number of indicators improves forecast accuracy and can be an effective way to mitigate the volatility associated with monthly indicators. Overall, our preferred U-MIDAS model specification performs well relative to various benchmark models and forecasters.
    Keywords: Econometric and statistical methods, International topics
    JEL: C C5 C50 C53 E E3 E37 E4 E47
    Date: 2014
  3. By: Adam Clements (QUT); Yin Liao (QUT)
    Abstract: Modeling and forecasting realized volatility is of paramount importance. Previous studies have examined the role of both the continuous and jump components of volatility in forecasting. This paper considers how to use index level jumps and cojumps across index constituents for forecasting index level volatility. In combination with the magnitude of past index jumps, the intensity of both index jumps and cojumps are examined. Estimated jump intensity from a point process model is used within a forecasting regression framework. Even in the presence of the diffusive part of total volatility, and past jump size, intensity of both index and cojumps are found to significantly improve forecast accuracy. An important contribution is that information relating to the behaviour of underlying constituent stocks is useful for forecasting index level behaviour. Improvements in forecast performance are particularly apparent on the days when jumps or cojumps occur, or when markets are turbulent.
    Keywords: Realized volatility; diffusion; jumps; point process; Hawkes process; forecasting
    JEL: C22 G00
    Date: 2014–06–17
  4. By: Stefano Grassi; Nima Nonejad; Paolo Santucci de Magistris
    Abstract: A modification of the self-perturbed Kalman filter of Park and Jun (1992) is proposed for the on-line estimation of models subject to parameter instability. The perturbation term in the updating equation of the state covariance matrix is weighted by the measurement error variance, thus avoiding the calibration of a design parameter. The standardization leads to a better tracking of the dynamics of the parameters compared to other on-line methods, especially as the level of noise increases. The proposed estimation method, coupled with dynamic model averaging and selection, is adopted to forecast S&P 500 realized volatility series with a time-varying parameters HAR model with exogenous variables.
    Keywords: TVP models; Self-Perturbed Kalman Filter; Dynamic Model Averaging; Dynamic Model Selection; Forecasting; Realized Variance
    JEL: C10 C11 C22 C80
    Date: 2014–02
  5. By: Julien Chevallier; Benoit Sevi
    Abstract: This paper evaluates the predictability of WTI light sweet crude oil futures by us- ing the variance risk premium, i.e. the difference between model-free measures of implied and realized volatilities. Additional regressors known for their ability to ex- plain crude oil futures prices are also considered, capturing macroeconomic, finan- cial and oil-specific influences. The results indicate that the explanatory power of the (negative) variance risk premium on oil excess returns is particularly strong (up to 25% for the adjusted R-squared across our regressions). It complements other fi- nancial (e.g. default spread) and oil-specific (e.g. US oil stocks) factors highlighted in previous literature.
    Keywords: Oil Futures, Variance Risk Premium, Forecasting
    JEL: C32 G17 Q47
    Date: 2014–06–16
  6. By: Karapanagiotidis, Paul
    Abstract: A review of the general state-space modeling framework. The discussion focuses heavily on the three prediction problems of forecasting, filtering, and smoothing within the state- space context. Numerous examples are provided detailing special cases of the state-space model and its use in solving a number of modeling issues. Independent sections are also devoted to both the topics of Factor models and Harvey’s Unobserved Components framework.
    Keywords: state-space models, signal extraction, unobserved components
    JEL: C10 C32 C51 C53 C58
    Date: 2014–06–03
  7. By: Egil Ferkingstad; Anders L{\o}land
    Abstract: Contracts for Difference (CfDs) are forwards on the spread between an area price and the system price. Together with the system price forwards, these products are used to hedge the area price risk in the Nordic electricity market. The CfDs are typically available for the next two months, three quarters and three years. This is fine, except that CfDs are not traded at NASDAQ OMX Commodities for every Nord Pool Spot price area. We therefore ask the hypothetical question: What would the CfD market price have been, say in the price area NO2, if it had been traded? We build regression models for each observable price area, and use Bayesian elicitation techniques to obtain prior information on how similar the different price areas are to forecast the price in an area where CfDs are not traded.
    Date: 2014–06
  8. By: Dirk Tasche
    Abstract: How to forecast next year's portfolio-wide credit default rate based on last year's default observations and the current score distribution? A classical approach to this problem consists of fitting a mixture of the conditional score distributions observed last year to the current score distribution. This is a special (simple) case of a finite mixture model where the mixture components are fixed and only the weights of the components are estimated. The optimum weights provide a forecast of next year's portfolio-wide default rate. We point out that the maximum-likelihood (ML) approach to fitting the mixture distribution not only gives an optimum but even an exact fit if we allow the mixture components to vary but keep their density ratio fix. From this observation we can conclude that the standard default rate forecast based on last year's conditional default rates will always be located between last year's portfolio-wide default rate and the ML forecast for next year. We also discuss how the mixture model based estimation methods can be used to forecast total loss. This involves the reinterpretation of an individual classification problem as a collective quantification problem.
    Date: 2014–06
  9. By: Markku Lanne (University of Helsinki and CREATES); Henri Nyberg (University of Helsinki)
    Abstract: We propose a new generalized forecast error variance decomposition with the property that the proportions of the impact accounted for by innovations in each variable sum to unity. Our decomposition is based on the well-established concept of the generalized impulse response function. The use of the new decomposition is illustrated with an empirical application to U.S. output growth and interest rate spread data.
    Keywords: Forecast error variance decomposition, generalized impulse response function, output growth, term spread
    JEL: C13 C32 C53
    Date: 2014–05–19
  10. By: Alessandro Flamini (Department of Economics and Management, University of Pavia); Iftekhar Hasan (Fordham University and Bank of Finland); Costas Milas (University of Liverpool)
    Abstract: In an open-economy faced with model uncertainty, this paper uses distribution forecasts to investigate the impact of alternative infl?ation targeting policies on macroeconomic volatility and their potential implications on ?financial stability. Theoretically, Domestic Infl?ation Targeting (DIT) leads to less volatility than Consumer Price index In?flation Targeting (CPIIT) for several macroeconomic variables and, in particular, for the interest rate. Empirically, a positive relationship between interest rate volatility and fi?nancial instability emerges for the US, UK and Sweden since the early 1990s. Bridging theory and empirical evidence, we conclude that the choice of the in?flation targeting regime has an important impact on macroeconomic volatility and potential implications for fi?nancial stability.
    Keywords: Macroeconomic volatility; fi?nancial stability; interest rate volatility; multiplicative uncertainty; Markov jump linear quadratic systems; open-economy; optimal montary policy; infl?ation index.
    JEL: E52 E58 F41
    Date: 2014–06
  11. By: Karapanagiotidis, Paul
    Abstract: Theory suggests that commodity futures price levels and returns data may exhibit both nonlinear and nonreversible features. This paper attempts to provide a thorough empiri- cally investigation of these claims. The data set is composed of 25 individual continuous contract commodity futures series which fall within a number of industry sectors including softs, precious metals, energy, and livestock. Employing both time-domain and frequency- domain tests examining the higher order cumulant properties of these series, it is shown that they exhibit both nonlinearities and irreversibility differing across industry sector. Furthermore, in modeling these series I estimate a number of parametric models able to capture irreversibility such as the linear mixed causal/noncausal autoregressive model and various purely causal nonlinear models, since there is a close connection between these two classes of models. It is shown that the linear causal ARMA model is unable to adequately account for the features of the data and while the mixed causal/noncausal model improves model fit significantly by capturing latent irreversibility, the vast majority of the nonlinearity these series exhibit is of the “nonlinear in variance” type. Finally, out of sample forecasts and an evaluation of the estimated unconditional distribution of the mixed causal/noncausal models suggest that there may still exist model misspecification.
    Keywords: mixed causal/noncausal autoregressions, nonlinear models, commodity futures, speculative price bubbles.
    JEL: C22 C50 C51 C52 C58
    Date: 2013–08–21

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.