nep-for New Economics Papers
on Forecasting
Issue of 2016‒06‒25
eight papers chosen by
Rob J Hyndman
Monash University

  1. Visualising forecasting Algorithm Performance using Time Series Instance Spaces By Yanfei Kang; Rob J. Hyndman; Kate Smith-Miles
  2. Forecasting implied volatility indices worldwide: A new approach By Degiannakis, Stavros; Filis, George; Hassani, Hossein
  3. Forecasting Employment Growth in Sweden Using a Bayesian VAR Model By Raoufina, Karine
  4. The PRISME model: can disaggregation on the production side help to forecast GDP? By C. Thubin; T. Ferrière; E. Monnet; M. Marx; V. Oung
  5. Taming volatile high frequency data with long lag structure: An optimal filtering approach for forecasting By Dirk Drechsel; Stefan Neuwirth
  6. Bias in Official Fiscal Forecasts: Can Private Forecasts Help? By Jeffrey A. Frankel; Jesse Schreger
  8. A financially stressed Euro area By Kappler, Marcus; Schleer, Frauke

  1. By: Yanfei Kang; Rob J. Hyndman; Kate Smith-Miles
    Abstract: It is common practice to evaluate the strength of forecasting methods using collections of well-studied time series datasets, such as the M3 data. But how diverse are these time series, how challenging, and do they enable us to study the unique strengths and weaknesses of different forecasting methods? In this paper we propose a visualisation method for a collection of time series that enables a time series to be represented as a point in a 2-dimensional instance space. The effectiveness of different forecasting methods can be visualised easily across this space, and the diversity of the time series in an existing collection can be assessed. Noting that the M3 dataset is not as diverse as we would ideally like, this paper also proposes a method for generating new time series with controllable characteristics to fill in and spread out the instance space, making generalisations of forecasting method performance as robust as possible.
    Keywords: M3-Competition, time series visualisation, time series generation, forecasting algorithm comparison
    JEL: C52 C53 C55
    Date: 2016
  2. By: Degiannakis, Stavros; Filis, George; Hassani, Hossein
    Abstract: This study provides a new approach for implied volatility indices forecasting. We assess whether non-parametric techniques provide better predictions of implied volatility compared to standard forecasting models, such as AFRIMA and HAR. A combination of Singular Spectrum Analysis (SSA) and Holt-Winters (HW) model is applied on eight implied volatility indices for the period from February, 2001 to July, 2013. The findings confirm that the SSA-HW provides statistically superior one trading day and ten trading days ahead implied volatility forecasts world widely. Model-averaged forecasts suggest that the forecasting accuracy is further enhanced, for the ten-days ahead, when the SSA-HW is combined with an ARI(1,1) model. Additionally, the trading game reveals that the SSA-HW and the ARI-SSA-HW are able to generate significant average positive net daily returns in the out-of-sample period. The results are important for option pricing, portfolio management, value-at-risk and economic policy.
    Keywords: Implied Volatility, Volatility Forecasting, Singular Spectrum Analysis, ARFIMA, HAR, Holt-Winters, Model Confidence Set, Combined Forecasts.
    JEL: C14 C22 C52 C53 G15
    Date: 2015–09–01
  3. By: Raoufina, Karine (National Institute of Economic Research)
    Abstract: In this paper, Bayesian VAR models are used to forecast employment growth in Sweden. Using quarterly data from 1996 to 2015, we conduct an out-of-sample forecast exercise. Results indicate that the forecasting performance at short horizons can be improved when survey data is included, such as employment expectations in the business sector and forward-looking variables from the trade sector.
    Keywords: Bayesian VAR model; employment forecasting
    JEL: C11 E24
    Date: 2016–06–08
  4. By: C. Thubin; T. Ferrière; E. Monnet; M. Marx; V. Oung
    Abstract: Although a forecasting model has very good statistical properties and the mean of the residuals equals zero, it can produce systematic errors during a short period. In the case of regular publications, forecasters want to prevent such a persistence of errors over several periods. For this reason, a safeguard model can be used to inform the forecaster when there is a risk that the standard model (i.e. the best specified model on average) leads to persistent errors over several months or quarters. This paper explains why and how such a safeguard model has been built in order to improve the forecasts of French GDP at the current quarter horizon (nowcasts), which are officially published by the French central bank. The official benchmark model for GDP nowcasts is an aggregated model that relies exclusively on survey in the manufacturing industry. In the long run, this model still has the best performances. On the contrary, the safeguard model is a disaggregated model which features equations for the valued added of 6 sectors. From this example, we provide general remarks on the advantages of disaggregation as well as how such safeguard models can be used in practice.
    Keywords: GDP nowcasting; Aggregation; Mixed-frequency data.
    JEL: C52 C53 E37
    Date: 2016
  5. By: Dirk Drechsel (KOF Swiss Economic Institute, ETH Zurich, Switzerland); Stefan Neuwirth (KOF Swiss Economic Institute, ETH Zurich, Switzerland)
    Abstract: We propose a Bayesian optimal filtering setup for improving out-of-sample forecasting performance when using volatile high frequency data with long lag structure for forecasting low-frequency data. We test this setup by using real-time Swiss construction investment and construction permit data. We compare our approach to different filtering techniques and show that our proposed filter outperforms various commonly used filtering techniques in terms of extracting the more relevant signal of the indicator series for forecasting.
    Keywords: Forecasting, construction, Switzerland, Bayesian, mixed data frequencies
    Date: 2016–06
  6. By: Jeffrey A. Frankel; Jesse Schreger
    Abstract: Government forecasts of GDP growth and budget balances are generally more over-optimistic than private sector forecasts. When official forecasts are especially optimistic relative to private forecasts ex ante, they are more likely also to be over-optimistic relative to realizations ex post. For example, euro area governments during the period 1999-2007 assiduously and inaccurately avoided forecasting deficit levels that would exceed the 3% Stability and Growth Pact threshold; meanwhile private sector forecasters were not subject to this crude bias. As a result, using private sector forecasts as an input into the government budgeting-making process would probably reduce official forecast errors for budget deficits.
    JEL: E62 H68
    Date: 2016–06
  7. By: Davide De Gaetano
    Abstract: In this paper the problem of instability due to changes in the parameters of some Realized Volatility (RV) models has been addressed. The analysis is based on 5-minute RV of four U.S. stock market indices. Three different representations of the log-RV have been considered and, for each of them, the parameter instability has been detected by using the recursive estimates test. In order to analyse how instabilities in the parameters affect the forecasting performance, an out-of-sample forecasting exercise has been performed. In particular, several forecast combinations, designed to accommodate potential structural breaks, have been considered. All of them are based on different estimation windows, with alternative weighting schemes, and do not take into account explicitly estimated break dates. The model con_dence set has been used to compare the forecasting performances of the proposed approaches. Our analysis gives empirical evidences of the effectiveness of the combinations which make adjustments for accounting the possible most recent break point.
    Keywords: Forecast combinations, Structural breaks, Realized volatility
    JEL: C53 C58 G17
    Date: 2016–06
  8. By: Kappler, Marcus; Schleer, Frauke
    Abstract: The authors analyze 149 newly compiled monthly time series on financial market stress conditions in the euro area. With the aid of a factor model they find different sources of financial stress which are important for selecting and preparing the appropriate policy response. The existence of a "Periphery Banking Crisis" factor, a "Stress" factor and a "Yield Curve" factor seems to explain the bulk of volatility in recent euro area financial sector data. Moreover, by a real-time forecasting exercise, the authors show that including additional factors - that reflect financial sector conditions - improves forecasts of economic activity at short horizons.
    Keywords: financial stress,dynamic factor models,financial crisis,euro area,forecasting
    JEL: C38 G01
    Date: 2016

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.