nep-for New Economics Papers
on Forecasting
Issue of 2015‒06‒13
eleven papers chosen by
Rob J Hyndman
Monash University

  1. Overcoming the Forecast Combination Puzzle: Lessons from the Time-Varying Effciency of Phillips Curve Forecasts of U.S. Inflation By Christopher G. Gibbs
  2. Crowdsourcing of economic forecast: Combination of forecasts using Bayesian model averaging By Kim, Dongkoo; Rhee, Tae-hwan; Ryu, Keunkwan; Shin, Changmock
  3. Weather, the forgotten factor in business cycle analyses By Döhrn, Roland; an de Meulen, Philipp
  4. Efficient estimation of Bayesian VARMAs with time-varying coefficients By Joshua C.C. Chan; Eric Eisenstat
  5. With or without you: Do financial data help to forecast industrial production? By Kitlinski, Tobias
  6. Gold price forecasts in a dynamic model averaging framework: Have the determinants changed over time? By Baur, Dirk G.; Beckmann, Joscha; Czudaj, Robert
  7. Multiple hypothesis testing of market risk forecasting models By esposito, francesco paolo; cummins, mark
  8. The role of targeted predictors for nowcasting GDP with bridge models: Application to the Euro area By Kitlinski, Tobias; an de Meulen, Philipp
  9. Leading indicators of financial stress: New evidence By Borek Vašícek; Diana Žigraiová; Marco Hoeberichts; Robert Vermeulen; Katerina Šmídková; Jakob de Haan
  10. The demand for euro banknotes in Germany: Structural modelling and forecasting By Bartzsch, Nikolaus; Seitz, Franz; Setzer, Ralph
  11. Do macroeconomic shocks affect intuitive inflation forecasting? An experimental investigation By Deversi, Marvin

  1. By: Christopher G. Gibbs (School of Economics, UNSW Business School, UNSW)
    Abstract: This paper proposes a new dynamic forecast combination strategy for forecasting inflation. The procedure draws on explanations of why the forecast combination puzzle exists and the stylized fact that Phillips curve forecasts of inflation exhibit significant time-variation in forecast accuracy. The forecast combination puzzle is the empirical observation that a simple average of point forecasts is often the best forecasting strategy. The forecast combination puzzle exists because many dynamic weighting strategies tend to shift weights toward Phillips curve forecasts after they exhibit a significant period of relative forecast improvement, which is often when their forecast accuracy begins to deteriorate. The proposed strategy in this paper weights forecasts according to their expected performance rather than their past performance to anticipate these changes in forecast accuracy. The forward-looking approach is shown to robustly beat equal weights combined and benchmark univariate forecasts of inflation in real-time out-of-sample exercises on U.S. and New Zealand inflation data.
    Keywords: Forecast combination, inflation, forecast pooling, forecast combination puzzle, Phillips curve
    JEL: E17 E47 C53
    Date: 2015–04
  2. By: Kim, Dongkoo; Rhee, Tae-hwan; Ryu, Keunkwan; Shin, Changmock
    Abstract: Economic forecasts are quite essential in our daily lives, which is why many research institutions periodically make and publish forecasts of main economic indicators. We ask (1) whether we can consistently have a better prediction when we combine multiple forecasts of the same variable and (2) if we can, what will be the optimal method of combination. We linearly combine multiple linear combinations of existing forecasts to form a new forecast ('combination of combinations'), and the weights are given by Bayesian model averaging. In the case of forecasts on Germany's real GDP growth rate, this new forecast dominates any single forecast in terms of root-mean-square prediction errors.
    Keywords: Combination of forecasts,Bayesian model averaging
    JEL: E32 E37
    Date: 2015
  3. By: Döhrn, Roland; an de Meulen, Philipp
    Abstract: In periods of unusual weather, forecasters face a problem of interpreting economic data: Which part goes back to the underlying economic trend and which part arises from a special weather effect? In this paper, we discuss ways to disentangle weather-related from business cycle-related influences on economic indicators. We find a significant influence of weather variables at least on a number of monthly indicators. Controlling for weather effects within these indicators should thus create opportunities to increase the accuracy of indicator-based forecasts. Focusing on quarterly GDP growth in Germany, we find that the accuracy of the RWI short term forecasting model improves but advances are small and not significant.
    Keywords: weather,short term forecasting,bridge equations,forecast accuracy
    JEL: C53 E37
    Date: 2015
  4. By: Joshua C.C. Chan; Eric Eisenstat
    Abstract: Empirical work in macroeconometrics has been mostly restricted to using VARs, even though there are strong theoretical reasons to consider general VARMAs. A number of articles in the last two decades have conjectured that this is because estimation of VARMAs is perceived to be challenging and proposed various ways to simplify it. Nevertheless, VARMAs continue to be largely dominated by VARs, particularly in terms of developing useful extensions. We address these computational challenges with a Bayesian approach. Specifically, we develop a Gibbs sampler for the basic VARMA, and demonstrate how it can be extended to models with time-varying VMA coefficients and stochastic volatility. We illustrate the methodology through a macroeconomic forecasting exercise. We show that in a class of models with stochastic volatility, VARMAs produce better density forecasts than VARs, particularly for short forecast horizons.
    Keywords: state space, stochastic volatility, factor model, macroeconomic forecasting, density forecast
    JEL: C11 C32 C53
    Date: 2015–06
  5. By: Kitlinski, Tobias
    Abstract: This paper analyzes the forecasting performance of financial market data in comparison to other indicator groups to forecast industrial production for Germany and the US. We focus on single-indicator models and various weighting schemes and evaluate the forecasting performance using a significance test. In addition, we investigate the stability of forecasting models before and during the recent financial crisis. This paper shows that financial market indicators are useful for short-term forecasting, especially for the US and longer forecast horizons. Nevertheless, the results indicate that the Great Recession was not foreseeable even if financial market indicators were taking into account. Furthermore, the reliability of pooled forecasts is higher than most of the forecasts obtained from single-indicator models.
    Abstract: In diesem Papier wird die Fähigkeit von Finanzmarktindikatoren im Vergleich zu anderen Kategorien von Indikatoren für die Prognose der Industrieproduktion in Deutschland und der USA verglichen. Dafür werden einzelne Gleichungen, in die die Indikatoren einfließen, und verschiedene Gewichtungsschemen herangezogen, um die Prognoseleistung zu evaluieren. Darüber hinaus wird die Stabilität der Prognosegüte untersucht, indem sie vor und während der Finanzmarktkrise verglichen wird. Es zeigt sich, dass Finanzmarktindikatoren durchaus nützlich für Kurzfristprognosen sind, insbesondere für die USA und wenn der Prognosezeitraum mehrere Monate umfasst. Nichtsdestotrotz lässt sich festhalten, dass auch unter Berücksichtigung von Finanzmarktindikatoren die Große Rezession nicht hätte vorhergesehen werden können. Zudem zeigt sich, dass Prognosen, die auf Gewichtungsschemen beruhen, stabiler sind als die von einzelnen Indikatorenmodellen.
    Keywords: forecasting,financial market data,single-indicator model,pooling of forecasts
    JEL: C53 E37
    Date: 2015
  6. By: Baur, Dirk G.; Beckmann, Joscha; Czudaj, Robert
    Abstract: The price of gold is influenced by a wide range of local and global factors such as commodity prices, interest rates, inflation expectations, exchange rate changes and stock market volatility among others. Hence, forecasting the price of gold is a notoriously difficult task and the main problem a researcher faces is to select the relevant regressors at each point in time. This combination of model and parameter uncertainty is explicitly accounted for by Dynamic Model Averaging which allows both the forecasting model and the coefficients to change over time. Based on this framework, we systematically evaluate a large set of possible gold price determinants and use both the predictive likelihood and the mean squared error as a measure of the forecasting performance. We carefully assess which predictors are relevant for forecasting at different points in time through the posterior probability. Our findings show that (1) DMA improves forecasts compared to other frameworks and (2) provides clear evidence for the time-variation of gold price predictors.
    Abstract: Der Goldpreis wird durch eine breite Palette von lokalen und globalen Faktoren wie Rohstoffpreisen, Zinssätzen, Inflationserwartungen, Wechselkursänderungen und Aktienmarktvolatilität beeinflusst. Daher ist es eine schwierige Aufgabe den Goldpreis zu prognostizieren. Das Hauptproblem besteht darin, die relevanten Regressoren zu jedem Zeitpunkt auszuwählen. Diese Kombination aus Modell- und Parameterunsicherheit wird in unserem Dynamic Modell Averaging (DMA) Ansatz berücksichtigt, in dem sich sowohl das Prognosemodell als auch deren Koeffizienten im Laufe der Zeit ändern können. Basierend darauf untersuchen wir eine große Bandbreite möglicher Goldpreisdeterminanten hinsichtlich ihrer Vorhersagekraft und nutzen dabei sowohl die Predictive Likelihood als auch den mittleren quadratischen Fehler als Maß für die Vorhersageleistung. Wir analysieren anhand der a posteriori-Wahrscheinlichkeit, welche Prädiktoren zu verschiedenen Zeitpunkten für die Prognose relevant waren. Unsere Ergebnisse zeigen, dass (1) DMA die Prognose im Vergleich zu anderen Verfahren verbessert und (2) dass die Goldpreisprädiktoren sich über die Zeit verändert haben.
    Keywords: Bayesian econometrics,dynamic model averaging,forecasting,gold
    JEL: C32 G10 G15 F37
    Date: 2014
  7. By: esposito, francesco paolo; cummins, mark
    Abstract: Extending previous risk model backtesting literature, we construct multiple hypothesis testing (MHT) with the stationary bootstrap. We conduct multiple tests which control for the generalized confidence level and employ the bootstrap MHT to design multiple comparison testing. We consider absolute and relative predictive ability to test a range of competing risk models, focusing on Value-at-Risk (VaR) and Expected Shortfall (ExS). In devising the test for the absolute predictive ability, we take the route of recent literature and construct balanced simultaneous confidence sets that control for the generalized family-wise error rate, which is the joint probability of rejecting true hypotheses. We implement a step-down method which increases the power of the MHT in isolating false discoveries. In testing for the ExS model predictive ability, we design a new simple test to draw inference about recursive model forecasting capability. In the second suite of statistical testing, we develop a novel device for measuring the relative predictive ability in the bootstrap MHT framework. The device, we coin multiple comparison mapping, provides a statistically robust instrument designed to answer the question: ''which model is the best model?''.
    Keywords: value-at-risk, expected shortfall, bootstrap multiple hypothesis testing, generalized familywise error rate, multiple comparison map
    JEL: C12
    Date: 2015–03–01
  8. By: Kitlinski, Tobias; an de Meulen, Philipp
    Abstract: Using factor models, it has recently been shown that a pre-selection of indicators improves GDP forecasts in the very short-term. The aim of this paper is to adopt this research to the methodology of bridge models in combination with pooling approaches. Focusing on Euro Area GDP between 2005 and 2013, we find that a selection of targeted predictors by means of soft- and hard-threshold algorithms improves the forecasting performance, especially during periods of economic crisis. While a critical number of indicators are needed to include all relevant information, adding additional indicators has a negative effect on forecasting performance, all the more, if the set of indicators becomes unbalanced.
    Abstract: In der Vergangenheit konnte gezeigt werden, dass eine Vorauswahl von Indikatoren die Prognoseleistung von Kurzfristprognosen, die auf Faktormodellen beruhen, deutlich verbessert. In diesem Papier wird untersucht, ob sich dieses Ergebnis auf Brückengleichungen in Kombination mit verschiedenen 'pooling' Ansätzen übertragen lässt. Dabei werden Prognosen für das Bruttoinlandsprodukts (BIP) des Euroraums im Zeitraum 2005 bis 2013 evaluiert. Es zeigt sich, dass eine Auswahl von Indikatoren durch 'soft- threshold ' und 'hard-threshold' Ansätze die Prognoseleistung deutlich verbessert. Dies gilt insbesondere in Zeiten von Wirtschaftskrisen. So wird zwar eine bestimmte Zahl von Indikatoren benötigt, die die notwendigen Informationen für eine möglichst genaue Prognose enthalten. Aber die Hinzunahme weiterer Indikatoren führt zu einer schlechteren Prognoseleistung. Dies gilt insbesondere dann, wenn zu viele Indikatoren aus einer bestimmten Kategorie berücksichtigt werden.
    Keywords: Forecasting,bridge equations,pooling of forecasts
    JEL: C53 E37
    Date: 2015
  9. By: Borek Vašícek; Diana Žigraiová; Marco Hoeberichts; Robert Vermeulen; Katerina Šmídková; Jakob de Haan
    Abstract: This paper examines which variables have predictive power for financial stress in a sample of 25 OECD countries, using a recently constructed Financial Stress Index (FSI). First, we employ Bayesian model averaging to identify leading indicators of our FSI. Next, we use those indicators as explanatory variables in a panel model for all our countries and in models at the individual country level. It turns out that panel models can hardly explain FSI dynamics. Although better results are achieved in models estimated at the country level, our findings suggest that (increases in) financial stress is (are) hard to predict out-of-sample.
    Keywords: financial stress index; Bayesian model averaging; early warning indicators
    JEL: E5 G10
    Date: 2015–06
  10. By: Bartzsch, Nikolaus; Seitz, Franz; Setzer, Ralph
    Abstract: This paper explains and forecasts the demand for banknotes issued in Germany. For small and large denomination notes we estimate vector error correction models (VECM). The results suggest that the long-run demand for German small denomination notes is mainly driven by domestic transactions and demand from outside the euro area. The transaction motive in the rest of the euro area is part of the short-term dynamics. The long-run demand for German large denomination notes is mainly driven by foreign demand both from the rest of the euro area and outside the EMU. The global financial crisis led to a one-time increase in the (real) demand for these notes. Our results are in line with estimates according to which the level and dynamics of banknote demand are mainly determined by foreign demand. Additionally, we present RegARIMA models for the medium denominations as it was not possible to build a VECM for these denominations.
    Keywords: banknotes, vector error correction, RegARIMA, forecasts
    JEL: C22 C32 E41
    Date: 2015–06–10
  11. By: Deversi, Marvin
    Abstract: In an experimental setting impulse-response behaviour in intuitive inflation forecasting is analysed. Participants were asked to forecast future values of inflation for a fictitious economy after receiving charts and lists of past values of inflation and output gap. Thirty periods were forecasted stepwise and feedback on performance was provided after each period. In a between subjects design, participants experienced a negative or positive supply shock. The results suggest that participants barely report rational forecasts. Instead, simple backward-looking rules describe stated forecast series. Forecasting is heterogeneous across agents and over time. Before the shock, most participants can be described by natural expectations. Due to the shocks 69% of participants are found to switch their forecasting rule. After the negative supply shock, subjects increase efficiency of forecasts. But, after a positive supply shock efficiency drops down to zero; this is evidence for a negativity bias. As a main result, macroeconomic shocks do alter the way experimental participants form intuitive inflation forecasts, however, to what extent depends on the shocks' characteristics.
    Abstract: Die vorliegende Studie untersucht die Effekte makroökonomischer Extremsituationen auf das intuitive Vorhersageverhalten von Individuen in einer experimentellen Umgebung. Die Probanden haben die Aufgabe, zukünftige Werte der Inflation einer fiktiven Volkswirtschaft vorherzusagen. Dies geschieht, nachdem sie die historische Entwicklung des Bruttoinlandsprodukts und der Inflation entsprechender Volkswirtschaft in grafischer und tabellarischer Form erhalten haben. In 30 Perioden werden jeweils Vorhersagen über die Inflation in der nächsten Periode verlangt, wobei die Probanden entsprechend der Präzision ihrer Vorhersage bezahlt werden. Nach jeder Prognose erhalten sie ein Feedback. Während der Vorhersageperioden erfahren die Probanden einen ein-periodischen, nicht-vorhersehbaren Schock; die Simulation einer Extremsituation. In einem between-subjects-design wird ein positiver oder negativer Angebotsschock induziert. Die Ergebnisse zeigen, dass nur sehr wenige Probanden rationale Vorhersagen abgeben. Eher lassen sich die Prognosereihen anhand einfacher statistischer Regeln, sog. Heuristiken, beschreiben. 69 Prozent der Probanden wechseln ihre Vorhersageheuristik durch den Schock. Nach dem negativen Angebotsschock steigt die statistische Effizienz der Vorhersagen stark an. Nach dem positiven Schock ist dies allerdings nicht zu beobachten. Dies ist Evidenz für einen sog. negativity bias. Es zeigt sich also, dass makroökonomische Extremsituationen das Vorhersageverhalten von Experimentteilnehmern beeinflussen. Allerdings variieren die Effekte entsprechend der Charakteristika der Extremsituation.
    Keywords: macroeconomic experiment,inflation expectations,intuitive forecasting,shocks,heterogeneity
    JEL: C91 D84
    Date: 2014

This nep-for issue is ©2015 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.