nep-for New Economics Papers
on Forecasting
Issue of 2014‒02‒02
fifteen papers chosen by
Rob J Hyndman
Monash University

  1. Does Central Bank Staff Beat Private Forecasters? By Jung, Alexander; El-Shagi, Makram; Giesen, Sebastian
  2. Point and Density Forecasts for the Euro Area Using Many Predictors: Are Large BVARs Really Superior? By Berg, Tim Oliver; Henzel, Steffen
  3. The Empirical (Ir)Relevance of the Interest Rate Assumption for Central Bank Forecasts By Knüppel, Malte; Schultefrankenfeld, Guido
  4. Anticipating business-cycle turning points in real time using density forecasts from a VAR By Schreiber, Sven
  5. Information Rigidities in Economic Growth Forecasts: Evidence from a Large International Panel By Dovern, Jonas; Fritsche, Ulrich; Loungani, Prakash; Tamirisa, Natalia
  6. Value-at-risk Predictions of Precious Metals with Long Memory Volatility Models By Demiralay, Sercan; Ulusoy, Veysel
  7. A latent dynamic factor approach to forecasting multivariate stock market volatility By Gribisch, Bastian
  8. Markov Switching with Endogenous Number of Regimes By Theobald, Thomas
  9. Modelling Stock Return Volatility Dynamics in Selected African Markets By Daniel King and Ferdi Botha
  10. Primary surplus and debt projections based on estimated fiscal reaction functions for euro area countries By Martin Plödt; Claire Reicher
  11. Forecasting business-cycle turning points with (relatively large) linear systems in real time By Schreiber, Sven
  12. Testing for common cycles in non-stationary VARs with varied frecquency data By Hecq A.W.; Urbain J.R.Y.J.; Götz T.B.
  13. Semiparametric Generalized Long Memory Modelling of GCC Stock Market Returns: A Wavelet Approach By Heni Boubaker; Nadia Sghaier
  14. Confidence Bands for Impulse Responses: Bonferroni versus Wald By Helmut Lütkepohl; Anna Staszewska-Bystrova; Peter Winker
  15. Network Risk and Forecasting Power in Phase-Flipping Dynamical Networks By B. Podobnik; A. Majdandzic; C. Curme; Z. Qiao; W. -X. Zhou; H. E. Stanley; B. Li

  1. By: Jung, Alexander; El-Shagi, Makram; Giesen, Sebastian
    Abstract: This paper assesses the relative performance of central bank staff forecasts and of private forecasters for inflation and output. We show that the Federal Reserve (Fed), and less so the European Central Bank (ECB), has a significant information advantage concerning inflation and output forecasts. Using recently developed tests for conditional predictive ability and forecast stability for the US, we find that the driving forces behind the narrowing of the information advantage of Greenbook forecasts have coincided with the Great Moderation. --
    JEL: C53 E37 E52
    Date: 2013
  2. By: Berg, Tim Oliver; Henzel, Steffen
    Abstract: Forecast models with large cross-sections are often subject to overparameterization leading to unstable parameter estimates and hence inaccurate forecasts. Recent articles suggest that a large Bayesian vector autoregression (BVAR) with sufficient prior information dominates competing approaches. In this paper we evaluate the forecast performance of large BVAR in comparison to its most natural competitors, i.e. averaging of small-scale BVARs and factor augmented BVARs with and without shrinkage. We derive point and density forecasts for euro area real GDP growth and HICP inflation conditional on an information set which is appropriate for all approaches and find no consistent outperformance of the large BVAR. While it produces good point forecasts, the performance is poor when density forecasts are used to evaluate predictive ability. Moreover, the ranking of the different approaches depends inter alia on the target variable, the forecast horizon, the state of the business cycle, and on the size of the dataset. Overall, we find that a factor augmented BVAR with shrinkage is competitive in all setups. --
    JEL: C11 C52 E37
    Date: 2013
  3. By: Knüppel, Malte; Schultefrankenfeld, Guido
    Abstract: The interest rate assumptions for macroeconomic forecasts differ considerably among central banks. Common approaches are given by the assumption of constant interest rates, interest rates expected by market participants, or the central bank's own interest rate expectations. From a theoretical point of view, the latter should yield the highest forecast accuracy. The lowest accuracy can be expected from forecasts conditioned on constant interest rates. However, when investigating the predictive accuracy of the forecasts for interest rates, inflation and output growth made by the Bank of England and the Banco do Brasil, we hardly find any significant differences between the forecasts based on different interest assumptions. We conclude that the choice of the interest rate assumption, while being a major concern from a theoretical point of view, appears to be at best of minor relevance empirically. --
    JEL: C12 C53 E58
    Date: 2013
  4. By: Schreiber, Sven
    Abstract: For the timely detection of business-cycle turning points we suggest to use mediumsized linear systems (subset VARs with automated zero restrictions) to forecast the relevant underlying variables, and to derive the probability of the turning point from the forecast density as the probability mass below (or above) a given threshold value. We show how this approach can be used in real time in the presence of data publication lags and how it can capture the part of the data revision process that is systematic. Then we apply the method to US and German monthly data. In an out-of-sample exercise (for 2007-2012/13) the turning points can be signalled before the official data publication confirms them (but not before they happened in reality). --
    Keywords: density forecasts,business-cycle turning points,real-time data,nowcasting,great recession
    JEL: C53 E37
    Date: 2014
  5. By: Dovern, Jonas; Fritsche, Ulrich; Loungani, Prakash; Tamirisa, Natalia
    Abstract: We examine the behavior of forecasts for real GDP growth using a large panel of individual forecasts from 30 advanced and emerging economies during 1989-2010. Our main findings are as follows. First, our evidence does not support the validity of the sticky information model (Mankiw and Reis, 2002) for describing the dynamics of professional growth forecasts. Instead, the empirical evidence is more in line with implications of "noisy" information models (Woodford, 2002; Sims, 2003). Second, we find that information rigidities are more pronounced in emerging economies than advanced economies. Third, there is evidence of nonlinearities in forecast smoothing. It is less pronounced in the tails of the distribution of individual forecast revisions than in the central part of the distribution. --
    Keywords: forecast,economic,information,expectations
    JEL: E27 E37
    Date: 2013
  6. By: Demiralay, Sercan; Ulusoy, Veysel
    Abstract: In this paper, we investigate the value-at-risk predictions of four major precious metals (gold, silver, platinum, and palladium) with long memory volatility models, namely FIGARCH, FIAPARCH and HYGARCH, under normal and student-t innovations’ distributions. For these analyses, we consider both long and short trading positions. Overall, our results reveal that long memory volatility models under student-t distribution perform well in forecasting a one-day-ahead VaR for both long and short positions. In addition, we find that FIAPARCH model with student-t distribution, which jointly captures long memory and asymmetry, as well as fat-tails, outperforms other models in VaR forecasting. Our results have potential implications for portfolio managers, producers, and policy makers.
    Keywords: Long memory, value-at-risk, volatility modeling, precious metals prices
    JEL: C53 C58 G17
    Date: 2014–01–27
  7. By: Gribisch, Bastian
    Abstract: This paper proposes a latent dynamic factor model for low- as well as high-dimensional realized covariance matrices of stock returns. The approach is based on the matrix logarithm and allows for flexible dynamic dependence patterns by combining common latent factors driven by HAR dynamics and idiosyncratic AR(1) factors. The model accounts for symmetry and positive definiteness of covariance matrices without imposing parametric restrictions. Simulated Bayesian parameter estimates as well as positive definite (co)variance forecasts are obtained using Markov Chain Monte Carlo (MCMC) methods. An empirical application to 5-dimensional and 30-dimensional realized covariance matrices of daily New York Stock Exchange (NYSE) stock returns shows that the model outperforms other approaches of the extant literature both in-sample and out-of-sample. --
    JEL: C32 C58 G17
    Date: 2013
  8. By: Theobald, Thomas
    Abstract: This paper uses several macroeconomic and financial indicators within a Markov Switching (MS) framework to predict the turning points of the business cycle. The presented model is applied to monthly German real-time data covering the recession and the recovery after the financial crisis. We show how to take advantage of combining single MSARX forecasts with the adjusting of the number of regimes on the real-time path, which both lead to higher forecast accuracy through the non-linearity of the underlying data-generating process. Adjusting the number of regimes implies distinguishing between recessions which are either normal or extraordinary, i.e. specifically determining as early as possible the point in time from which the recession in the aftermath of the financial crisis structurally exceeded previous ones. In fact it turns out that the Markov Switching model can signal quite early whether a conventional recession will occur or whether an economic downturn will be more pronounced. --
    JEL: C24 C53 E37
    Date: 2013
  9. By: Daniel King and Ferdi Botha
    Abstract: This paper examines whether accounting for structural changes in the conditional variance process, through the use of Markov-switching models, improves estimates and forecasts of stock return volatility over those of the more conventional single-state (G)ARCH models, within and across selected African markets for the period 2002-2012. In the univariate portion of the paper, the performances of various Markov-switching models are tested against a single-state benchmark model through the use of in-sample goodness-of-fit and predictive ability measures. In the multivariate context, the single-state and Markov-switching models are comparatively assessed according to their usefulness in constructing optimal stock portfolios. Accounting for structural breaks in the conditional variance process, conventional GARCH effects remain important in capturing heteroscedasticity. However, those univariate models including a GARCH term perform comparatively poorly when used for forecasting purposes. In the multivariate study, the use of Markov-switching variance-covariance estimates improves risk-adjusted portfolio returns relative to portfolios constructed using the more conventional single-state models. While there is evidence that some Markov-switching models can provide better forecasts and higher risk-adjusted returns than those models which include GARCH effects, the inability of the simpler Markov-switching models to fully capture heteroscedasticity in the data remains problematic.
    Keywords: Stock returns, volatility, GARCH, Africa
    JEL: C52 C58
    Date: 2014
  10. By: Martin Plödt; Claire Reicher
    Abstract: We project the path of the public debt and primary surpluses for a number of countries in the euro area under a fiscal rule based on a set of estimated fiscal policy reaction functions. Our fiscal rule represents a fiscal analogue to a well-known monetary policy rule, and it is calibrated using country-specific as well as euro area-wide parameter estimates. We then forecast the dynamics of the fiscal aggregates under different convergence, growth, and interest rate scenarios and investigate the implications of these scenarios in projecting the future path of fiscal aggregates. We argue that our forecasting methodology may be used to deliver insights into the medium-run effects of different fiscal policy rules and to provide some early warning of future fiscal pressures
    Keywords: fiscal rules, fiscal policy, euro area, forecasting
    JEL: H62 H63 H68
    Date: 2014–01
  11. By: Schreiber, Sven
    Abstract: The detection of business-cycle turning points is usually performed with non-linear discrete-regime models such as binary dependent variable (e.g., probit or logit) or Markov-switching methods. The probit model has the drawback that the continuous underlying target variable is discretized, with a considerable loss of information. The Markov-switching approach in general presupposes a non-linear data-generating process, and the numerical likelihood maximization becomes increasingly dif cult when more covariates are used. To avoid these problems we suggest to rst use standard linear systems (subset VARs with zero restrictions) to forecast the relevant underlying variable(s), and in a second step to derive the probability of a suitably de ned turning point from the forecast probability density function. This approach will never fail numerically. We also discuss and show how this approach can be used in real time in the presence of publication lags and to capture features of the data revision process, and we apply the method to German data; the event of the recent Great Recession is rst signalled in June 2008, several months before the of cial published data con rms it (but due to publication and recognition lags it is found after it already began in reality). --
    JEL: C53 E37 E32
    Date: 2013
  12. By: Hecq A.W.; Urbain J.R.Y.J.; Götz T.B. (GSBE)
    Abstract: This paper proposes a new way for detecting the presence of common cyclical featureswhen several time series are observed/sampled at different frequencies, hence generalizingthe common-frequency approach introduced by Engle and Kozicki 1993 and Vahid andEngle 1993. We start with the mixed-frequency VAR representation investigated in Ghysels2012 for stationary time series. For non-stationary time series in levels, we showthat one has to account for the presence of two sets of long-run relationships. The First setis implied by identities stemming from the fact that the differences of the high-frequencyI1 regressors are stationary. The second set comes from possible additional long-run relationshipsbetween one of the high-frequency series and the low-frequency variables. Ourtransformed VECM representations extend the results of Ghysels 2012 and are very importantfor determining the correct set of variables to be used in a subsequent commoncycle investigation. This has some empirical implications both for the behavior of the teststatistics as well as for forecasting. Empirical analyses with the quarterly real GNP andmonthly industrial production indices for, respectively, the U.S. and Germany illustrate ournew approach. This is also investigated in a Monte Carlo study, where we compare our proposedmixed-frequency models with models stemming from classical temporal aggregationmethods.
    Keywords: Economic History: Transport, Trade, Energy, Technology, and Other Services: Asia including Middle East; Regional and Urban History: General; Microeconomic Analyses of Economic Development;
    JEL: N90 O12 N75
    Date: 2013
  13. By: Heni Boubaker; Nadia Sghaier
    Abstract: This paper proposes a new class of semiparametric generalized long memory model with FIA- PARCH errors (SEMIGARMA-FIAPARCH model) that extends the conventionnel GARMA model to incorporate nonlinear deterministic trend, in the mean equation, and to allow for time varying volatility, in the conditional variance equation. The parameters of this model are estimated in a wavelet domain. We provide an empirical application of this model to examine the dynamic of the stock market returns in six GCC countries. The empirical results show that the model proposed o¤ers an interesting framework to describe the seasonal long range dependence and the nonlinear deterministic trend in the return as well as persistence to shocks in the conditional volatiliy. We also compare its performance predictive to the traditional long memory model with FIAPARCH errors (FARMA-FIAPARCH model). The predictive results indicate that the model proposed out performs the FARMA-FIAPARCH model.
    Keywords: semiparametric generalized long memory process, FIAPARCH errors, wavelet do- main, stock market returns.
    JEL: C13 C22 C32 G15
    Date: 2014–01–06
  14. By: Helmut Lütkepohl; Anna Staszewska-Bystrova; Peter Winker
    Abstract: In impulse response analysis estimation uncertainty is typically displayed by constructing bands around estimated impulse response functions. These bands may be based on frequentist or Bayesian methods. If they are based on the joint distribution in the Bayesian framework or the joint asymptotic distribution possibly constructed with bootstrap methods in the frequentist framework often individual confidence intervals or credibility sets are simply connected to obtain the bands. Such bands are known to be too narrow and have a joint confidence content lower than the desired one. If instead the joint distribution of the impulse response coefficients is taken into account and mapped into the band it is shown that such a band is typically rather conservative. It is argued that a smaller band can often be obtained by using the Bonferroni method. While these considerations are equally important for constructing forecast bands, we focus on the case of impulse responses in this study.
    Keywords: Impulse responses, Bayesian error bands, frequentist confidence bands, Wald statistic, vector autoregressive process
    JEL: C32
    Date: 2014
  15. By: B. Podobnik; A. Majdandzic; C. Curme; Z. Qiao; W. -X. Zhou; H. E. Stanley; B. Li
    Abstract: In order to model volatile real-world network behavior, we analyze phase-flipping dynamical scale-free network in which nodes and links fail and recover. We investigate how stochasticity in a parameter governing the recovery process affects phase-flipping dynamics, and find the probability that no more than q% of nodes and links fail. We derive higher moments of the fractions of active nodes and active links, $f_n(t)$ and $f_{\ell}(t)$, and define two estimators to quantify the level of risk in a network. We find hysteresis in the correlations of $f_n(t)$ due to failures at the node level, and derive conditional probabilities for phase-flipping in networks. We apply our model to economic and traffic networks.
    Date: 2014–01

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.