nep-for New Economics Papers
on Forecasting
Issue of 2016‒03‒06
nineteen papers chosen by
Rob J Hyndman
Monash University

  1. Complex Exponential Smoothing By Svetunkov, Ivan; Kourentzes, Nikolaos
  2. Forecasting oil price realized volatility: A new approach By Degiannakis, Stavros; Filis, George
  3. Inflation Dynamics and the Hybrid Neo Keynesian Phillips Curve: The Case of Chile By Carlos Medel
  4. New Housing Registrations as a Leading Indicator of the BC Economy By Calista Cheung; Dmitry Granovsky
  5. Forecasting with VAR Models: Fat Tails and Stochastic Volatility By Ching-Wai (Jeremy) Chiu; Haroon Mumtaz; Gabor Pinter
  6. Model Averaging in Markov-Switching Models: Predicting National Recessions with Regional Data By Pierre Guérin; Danilo Leiva-Leon
  7. Forecasting euro area recessions in real-time By Pirschel, Inske
  8. Experts, firms, consumers or even hard data? Forecasting employment in Germany By Lehmann, Robert; Wohlrabe, Klaus
  9. Does Realized Volatility Help Bond Yield Density Prediction? By Shin, Minchul; Zhong, Molin
  10. Nowcasting Business Cycles: a Bayesian Approach to Dynamic Heterogeneous Factor Models By D'Agostino, Antonello; Giannone, Domenico; Lenza, Michele; Modugno, Michele
  11. Evaluating Forecasts of a Vector of Variables: A German Forecasting Competition By Tara M. Sinclair; Hans Christian Müller-Dröge; Herman Stekler
  12. A daily indicator of economic growth for the euro area By Valentina Aprigliano; Claudia Foroni; Massimiliano Marcellino; Gianluigi Mazzi; Fabrizio Venditti
  13. Merger options and risk arbitrage By Van Tassel, Peter
  14. Nowcasting Indonesia By Luciani, Matteo; Pundit, Madhavi; Ramayandi, Arief; Veronese, Giovanni
  15. Eliciting GDP Forecasts from the FOMC’s Minutes Around the Financial Crisis By Ericsson, Neil R.
  16. The Impact of Jumps and Leverage in Forecasting Co-Volatility By Asai, M.; McAleer, M.J.
  17. Regime-Switching Models for Estimating Inflation Uncertainty By Nalewaik, Jeremy J.
  18. Plausibility of big shocks within a linear state space setting with skewness By Koloch, Grzegorz
  19. Estimation of DSGE Models under Diffuse Priors and Data-Driven Identification Constraints By Markku Lanne; Jani Luoto

  1. By: Svetunkov, Ivan; Kourentzes, Nikolaos
    Abstract: Exponential smoothing has been one of the most popular forecasting methods for business and industry. Its simplicity and transparency have made it very attractive. Nonetheless, modelling and identifying trends has been met with mixed success, resulting in the development of various modifications of trend models. We present a new approach to time series modelling, using the notion of ``information potential" and the theory of functions of complex variables. A new exponential smoothing method that uses this approach, ``Complex exponential smoothing" (CES), is proposed. It has an underlying statistical model described here and has several advantages over the conventional exponential smoothing models: it allows modelling and forecasting both trended and level time series, effectively sidestepping the model selection problem. CES is evaluated on real data demonstrating better performance than established benchmarks and other exponential smoothing methods.
    Keywords: Forecasting, exponential smoothing, ETS, model selection, information potential, complex variables
    JEL: C5 C53
    Date: 2015–05–01
  2. By: Degiannakis, Stavros; Filis, George
    Abstract: This paper adds to the extremely limited strand of the literature focusing on the oil price realized volatility forecasting. More specifically, we evaluate the information content of four different asset classes’ volatilities when forecasting the oil price realized volatility for 1-day until 66-day ahead. To do so, we concentrate on the Brent crude oil and fourteen other assets, which are grouped into four different asset classes, based on Heterogeneous AutoRegressive (HAR) framework. Our out-of-sample forecasting results can be summarised as follows. (i) The use of exogenous volatilities statistically significant improves the forecasting accuracy at all forecasting horizons. (ii) The HAR model that combines volatilities from multiple asset classes is the best performing model. (iii) The Direction of Change suggests that all HAR models are highly accurate in predicting future movements of oil price volatility. (iv) The forecasting accuracy of the models is better gauged using the Median Absolute Error and the Median Squared Error. (v) The findings are robust even during turbulent economic periods. Hence, different asset classes’ volatilities contain important information which can be used to improve the forecasting accuracy of oil price volatility.
    Keywords: Volatility forecasting, realized volatility, crude oil futures, Brent crude oil, HAR, MCS.
    JEL: C22 C53 G13 Q02 Q47
    Date: 2016–01–29
  3. By: Carlos Medel
    Abstract: It is recognised that the understanding and accurate forecasts of key macroeconomic variables are fundamental for the success of any economic policy. In the case of monetary policy, many efforts have been made towards understanding the relationship between past and expected values of inflation, resulting in the so-called Hybrid Neo-Keynesian Phillips Curve (HNKPC). In this article I investigate to which extent the HNKPC help to explain inflation dynamics as well as its out-ofsample forecast for the case of the Chilean economy. The results show that the forward-looking component is significant and accounts from 1.58 to 0.40 times the lagged inflation coefficient. Also, I find predictive gains close to 45% (respect to a backward-looking specification) and up to 80% (respect to the random walk) when forecasting at 12-months ahead. The output gap building process plays a key role delivering better results than similar benchmark. None of the two openness measures used—neither real exchange rate nor oil price—are significant in the reduced form. A final estimation using the annual variation of a monthly indicator of GDP deliver reasonable forecast accuracy but not as good as the preferred forecast-implied output gap measure.
    Date: 2015–09
  4. By: Calista Cheung; Dmitry Granovsky
    Abstract: Housing starts and building permits data are commonly used as leading indicators of economic activity. In British Columbia, all new homes must be registered with the Homeowner Protection Office, a branch of BC Housing, before the issuance of building permits and the start of construction. Data on new housing registrations (NHR) could thus potentially be used as an even earlier leading indicator of economic activity. This study assesses whether NHR data have significant predictive power for economic activity in British Columbia. The authors find that quarterly increases in new registrations for single detached homes have statistically significant predictive content for growth in real GDP over the next one to three quarters, and provide stronger signals compared to housing starts and building permits over this forecast horizon. These signals remain significant for growth in real GDP over the next two quarters even in the presence of other leading indicators in the equations. However, forecasts using quarterly NHR data with other leading indicators are not able to outperform simple benchmark forecasts in an out-of-sample forecasting exercise. Nonetheless, adding the NHR variable to an AR(1) equation does produce forecasts that are superior to a simple AR(1) and that at one quarter ahead also outperform an AR(1) augmented with building permits.
    Keywords: Business fluctuations and cycles, Housing, Regional economic developments
    JEL: C13 C53 E32 E37
    Date: 2016
  5. By: Ching-Wai (Jeremy) Chiu (Bank of England); Haroon Mumtaz (Queen Mary University of London); Gabor Pinter (Bank of England)
    Abstract: In this paper, we provide evidence that fat tails and stochastic volatility can be important in improving in-sample fit and out-of-sample forecasting performance. Specifically, we construct a VAR model where the orthogonalised shocks feature Student’s t distribution and time-varying variance. We estimate this model using US data on output growth, inflation, interest rates and stock returns. In terms of in-sample fit, the VAR model that features both stochastic volatility and Student’s t-distributed disturbances outperforms restricted alternatives that feature either attributes. The VAR model with Student’s t disturbances results in density forecasts for industrial production and stock returns that are superior to alternatives that assume Gaussianity, and this difference appears to be especially stark over the recent Great Recession. Further international evidence confirms that accounting for both stochastic volatility and Student’s t-distributed disturbances may lead to improved forecast accuracy.
    Keywords: Financial Frictions, Predictive Densities, Great Recession, Threshold VAR
    JEL: C11 C32 C52
    Date: 2015–02
  6. By: Pierre Guérin; Danilo Leiva-Leon
    Abstract: This paper introduces new weighting schemes for model averaging when one is interested in combining discrete forecasts from competing Markov-switching models. In particular, we extend two existing classes of combination schemes – Bayesian (static) model averaging and dynamic model averaging – so as to explicitly reflect the objective of forecasting a discrete outcome. Both simulation and empirical exercises show that our new combination schemes outperform competing combination schemes in terms of forecasting accuracy. In the empirical application, we estimate and forecast U.S. business cycle turning points with state-level employment data. We find that forecasts obtained with our best combination scheme provide timely updates of the U.S. business cycles.
    Keywords: Business fluctuations and cycles, Econometric and statistical methods
    JEL: C53 E32 E37
    Date: 2015
  7. By: Pirschel, Inske
    Abstract: I present evidence that the linear mixed-frequency Bayesian VAR provides very sharp and well calibrated monthly real-time recession probabilities for the euro area for the period from 2004 until 2013. The model outperforms not only the univariate regime-switching models for a number of hard and soft economic indicators and their optimal linear combinations, but also a real-time recession index obtained with Google Trends data. This result holds irrespective of whether the joint predictive distribution of several economic indicators or the marginal distribution of real GDP growth is evaluated to extract the real-time recession probabilities of the mixed-frequency Bayesian VAR. The inclusion of the confidence index in industry turns out to be crucial for the performance of the model.
    Keywords: Density nowcasting,Real-time recession forecasting,Mixed-frequency data,Bayesian VAR,Regime-switching models,Linear opinion pool,Google Trends
    JEL: C53 E32 E37
    Date: 2016
  8. By: Lehmann, Robert; Wohlrabe, Klaus
    Abstract: In this paper, we forecast employment growth for Germany with data for the period from November 2008 to November 2015. Hutter and Weber (2015) introduced an innovative unemployment indicator and evaluate the performance of several leading indicators, including the Ifo Employment Barometer, to predict unemployment changes. Since the Ifo Employment Barometer focuses on employment growth instead of unemployment developments, we mirror the study by Hutter and Weber (2015). It turns out that in our case, and in contrast to their article, the Ifo Employment Barometer outperforms their newly developed indicator. Additionally, consumers’ unemployment expectations and hard data such as new orders exhibit a high forecasting accuracy.
    Keywords: survey data; employment forecasts; model confidence set
    JEL: C52 C53 E24 E27 J00
    Date: 2016–02–19
  9. By: Shin, Minchul (University of Illinois); Zhong, Molin (Board of Governors of the Federal Reserve System (U.S.))
    Abstract: We suggest using "realized volatility" as a volatility proxy to aid in model-based multivariate bond yield density forecasting. To do so, we develop a general estimation approach to incorporate volatility proxy information into dynamic factor models with stochastic volatility. The resulting model parameter estimates are highly efficient, which one hopes would translate into superior predictive performance. We explore this conjecture in the context of density prediction of U.S. bond yields by incorporating realized volatility into a dynamic Nelson-Siegel (DNS) model with stochastic volatility. The results clearly indicate that using realized volatility improves density forecasts relative to popular specifications in the DNS literature that neglect realized volatility.
    Keywords: Dynamic factor model; forecasting; stochastic volatility; term structure of interest rates; dynamic Nelson-Siegel model
    JEL: C5 E4 G1
    Date: 2015–12–18
  10. By: D'Agostino, Antonello (European Stability Mechanism); Giannone, Domenico (Federal Reserve Bank of New York); Lenza, Michele (European Central Bank); Modugno, Michele (Board of Governors of the Federal Reserve System (U.S.))
    Abstract: We develop a framework for measuring and monitoring business cycles in real time. Following a long tradition in macroeconometrics, inference is based on a variety of indicators of economic activity, treated as imperfect measures of an underlying index of business cycle conditions. We extend existing approaches by permitting for heterogenous lead-lag patterns of the various indicators along the business cycles. The framework is well suited for high-frequency monitoring of current economic conditions in real time - nowcasting - since inference can be conducted in presence of mixed frequency data and irregular patterns of data availability. Our assessment of the underlying index of business cycle conditions is accurate and more timely than popular alternatives, including the Chicago Fed National Activity Index (CFNAI). A formal real-time forecasting evaluation shows that the framework produces well-calibrated probability nowcasts that resemble the consensus assessment of t he Survey of Professional Forecasters.
    Keywords: Current Economic Conditions; Dynamic Factor Models; Dynamic Heterogeneity; Business Cycles; Real Time; Nowcasting.
    JEL: C11 C32 C38 E32
    Date: 2015–08–06
  11. By: Tara M. Sinclair (Department of Economics/Institute for International Economic Policy, George Washington University); Hans Christian Müller-Dröge (Handelsblatt Newspaper); Herman Stekler (Department of Economics, George Washington University)
    Abstract: In this paper we present an evaluation of forecasts of a vector of variables of the German economy made by different institutions. Our method permits one to evaluate the forecasts for each year and then if one is interested to combine the years. We use our method to determine an overall winner for a forecasting competition across twenty-five different institutions for a single time period using a vector of eight key economic variables. Typically forecasting competitions are judged on a variable-by-variable basis, but our methodology allows us to determine how each competitor performed overall. We find that the Bundesbank was the overall winner for 2013.
    Keywords: Mahalanobis Distance, forecasting competition, GDP components, German macroeconomic data
    JEL: C5 E2 E3
    Date: 2014–07
  12. By: Valentina Aprigliano; Claudia Foroni; Massimiliano Marcellino; Gianluigi Mazzi; Fabrizio Venditti
    Abstract: In this paper we study alternative methods to construct a daily indicator of growth for the euro area. We aim for an indicator that (i) provides reliable predictions, (ii) can be easily updated at the daily frequency, (iii) gives interpretable signals, and (iv) it is linear. Using a large panel of daily and monthly data for the euro area we explore the performance of two classes of models: bridge and U-MIDAS models, and di¤erent forecast combination strategies. Forecasts obtained from U-MIDAS models, combined with the inverse MSE weights, best satisfy the required criteria. JEL classi…cation: C51, C53, E27. Keywords: Nowcasting, mixed-frequency data.
    Date: 2016
  13. By: Van Tassel, Peter (Federal Reserve Bank of New York)
    Abstract: Option prices embed predictive content for the outcomes of pending mergers and acquisitions. This is particularly important in merger arbitrage, where deal failure is a key risk. In this paper, I propose a dynamic asset pricing model that exploits the joint information in target stock and option prices to forecast deal outcomes. By analyzing how deal announcements affect the level and higher moments of target stock prices, the model yields better forecasts than existing methods. In addition, the model accurately predicts that merger arbitrage exhibits low volatility and a large Sharpe ratio when deals are likely to succeed.
    Keywords: financial economics; option pricing; mergers and acquisitions
    JEL: G00 G12 G34
    Date: 2016–01–01
  14. By: Luciani, Matteo (Board of Governors of the Federal Reserve System (U.S.)); Pundit, Madhavi (Asian Development Bank); Ramayandi, Arief (Asian Development Bank); Veronese, Giovanni (Banca d'Italia)
    Abstract: We produce predictions of the current state of the Indonesian economy by estimating a dynamic factor model on a dataset of eleven indicators (also followed closely by market operators) over the time period 2002 to 2014. Besides the standard difficulties associated with constructing timely indicators of current economic conditions, Indonesia presents additional challenges typical to emerging market economies where data are often scant and unreliable. By means of a pseudo-real-time forecasting exercise we show that our model outperforms univariate benchmarks, and it does comparably with predictions of market operators. Finally, we show that when quality of data is low, a careful selection of indicators is crucial for better forecast performance.
    Keywords: Dynamic Factor Models; Emerging Market Economies; Nowcasting
    Date: 2015–11–09
  15. By: Ericsson, Neil R. (Board of Governors of the Federal Reserve System (U.S.))
    Abstract: Stekler and Symington (2016) construct indexes that quantify the Federal Open Market Committee's views about the U.S. economy, as expressed in the minutes of the FOMC's meetings. These indexes provide insights on the FOMC's deliberations, especially at the onset of the Great Recession. The current paper complements Stekler and Symington's analysis by showing that their indexes reveal relatively minor bias in the FOMC's views when the indexes are reinterpreted as forecasts. Additionally, these indexes provide a proximate mechanism for inferring the Fed staff's Greenbook forecasts of the U.S. real GDP growth rate, years before the Greenbook's public release.
    Keywords: Autometrics; bias; Fed; financial crisis; FOMC; forecasts; GDP; Great Recession; Greenbook; impulse indicator saturation; projections; Tealbook; United States
    JEL: C53 E58
    Date: 2015–11–17
  16. By: Asai, M.; McAleer, M.J.
    Abstract: __Abstract__ The paper investigates the impact of jumps in forecasting co-volatility, accommodating leverage effects. We modify the jump-robust two time scale covariance estimator of Boudt and Zhang (2013) such that the estimated matrix is positive definite. Using this approach we can disentangle the estimates of the integrated co-volatility matrix and jump variations from the quadratic covariation matrix. Empirical results for three stocks traded on the New York Stock Exchange indicate that the co-jumps of two assets have a significant impact on future co-volatility, but that the impact is negligible for forecasting weekly and monthly horizons.
    Keywords: Co-Volatility, Forecasting, Jump, Leverage Effects, Realized Covariance, Threshold, Estimation.
    JEL: C32 C53 C58 G17
    Date: 2015–02–01
  17. By: Nalewaik, Jeremy J. (Board of Governors of the Federal Reserve System (U.S.))
    Abstract: This paper constructs regime-switching models for estimating the probability of inflation returning to its relatively high levels of variability and persistence in the 1970s and 1980s. Forecasts and probabilities of extreme events from the models are evaluated against comparable estimates from other statistical models, from surveys, and from financial markets. The paper then uses the models to construct prediction intervals around Federal Reserve Board staff forecasts of PCE price inflation, combining the recent non-parametric forecast error distribution with parametric information from the model. The outer tails of the prediction intervals depend importantly on the probability inflation is in its high-variance, high-persistence regime.
    Keywords: Inflation; Markov-Switching; Uncertainty
    JEL: E30
    Date: 2015–09–01
  18. By: Koloch, Grzegorz
    Abstract: In this paper we provide formulae for likelihood function, filtration densities and prediction densities of a linear state space model in which shocks are allowed to be skewed. In particular we work with the closed skew normal distribution, see González-Farías et al. (2004), which nests a normal distribution as a special case. Closure of the csn distribution with respect to all necessary transformations in the state space setting is guaranteed by a simple state dimension reduction procedure which does not influence the value of the likelihood function. Presented formulae allow for estimation, filtration and prediction of vector autoregressions and first order perturbations of DSGE models with skewed shocks. This allows to assess asymmetries in shocks, observed data, impulse responses and forecasts confidence intervals. Some of the advantages of using the outlined approach may involve capturing asymmetric inflation risks in central banks forecasts or producing more plausible probabilities of deep but rare recessionary episodes with DSGE/VAR filtration. Exemplary estimation results are provided which show that within a linear setting with skewness frequency of big shocks can be rather plausibly identifed.
    Keywords: Maximum likelihood estimation, state space models, closed skew-normal distribution, DSGE, VAR
    JEL: C13 C51 E32
    Date: 2016–01–25
  19. By: Markku Lanne (University of Helsinki and CREATES); Jani Luoto (University of Helsinki)
    Abstract: We propose a sequential Monte Carlo (SMC) method augmented with an importance sampling step for estimation of DSGE models. In addition to being theoretically well motivated, the new method facilitates the assessment of estimation accuracy. Furthermore, in order to alleviate the problem of multimodal posterior distributions due to poor identification of DSGE models when uninformative prior distributions are assumed, we recommend imposing data-driven identification constraints and devise a procedure for finding them. An empirical application to the Smets-Wouters (2007) model demonstrates the properties of the estimation method, and shows how the problem of multimodal posterior distributions caused by parameter redundancy is eliminated by identification constraints. Out-of-sample forecast comparisons as well as Bayes factors lend support to the constrained model.
    Keywords: Particle filter, importance sampling, Bayesian identification
    JEL: D58 C11 C32 C52
    Date: 2015–08–18

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.