New forecasting papers 2011-03-01

In this issue we have Coherent mortality forecasting: the product-ratio method with functional time series models, The value of feedback in forecasting competitions, Does Disagreement amongst Forecasters have Predictive Value?, Heuristic model selection for leading indicators in Russia and Germany, and more.

  1. Coherent mortality forecasting: the product-ratio method with functional time series models
    Date: 2011-02-04
    By: Rob J Hyndman
    Heather Booth
    Farah Yasmeen
    When independence is assumed, forecasts of mortality for subpopulations are almost always divergent in the long term. We propose a method for non-divergent or coherent forecasting of mortality rates for two or more subpopulations, based on functional principal components models of simple and interpretable functions of rates. The product-ratio functional forecasting method models and forecasts the geometric mean of subpopulation rates and the ratio of subpopulation rates to product rates. Coherence is imposed by constraining the forecast ratio function through stationary time series models. The method is applied to sex-specific data for Sweden and state-specific data for Australia. Based on out-of-sample forecasts, the coherent forecasts are at least as accurate in overall terms as comparable independent forecasts, and forecast accuracy is homogenised across subpopulations.
    Keywords: Mortality forecasting, coherent forecasts, functional data, Lee-Carter method, life expectancy, mortality, age pattern of mortality, sex-ratio
    JEL: J11
  2. The value of feedback in forecasting competitions
    Date: 2011-02-09
    By: George Athanasopoulos
    Rob J Hyndman
    In this paper we challenge the traditional design used for forecasting competitions. We implement an online competition with a public leaderboard that provides instant feedback to competitors who are allowed to revise and resubmit forecasts. The results show that feedback significantly improves forecasting accuracy.
    Keywords: Forecasting competition, feedback.
    JEL: C53
  3. Does Disagreement amongst Forecasters have Predictive Value?
    Date: 2010-09-03
    By: Rianne Legerstee (Erasmus University Rotterdam)
    Philip Hans Franses (Erasmus University Rotterdam)
    Forecasts from various experts are often used in macroeconomic forecasting models. Usually the focus is on the mean or median of the survey data. In the present study we adopt a different perspective on the survey data as we examine the predictive power of disagreement amongst forecasters. The premise is that this variable could signal upcoming structural or temporal changes in an economic process or in the predictive power of the survey forecasts. In our empirical work, we examine a variety of macroeconomic variables, and we use different measurements for the degree of disagreement, together with measures for location of the survey data and autoregressive components. Forecasts from simple linear models and forecasts from Markov regime-switching models with constant and with time-varying transition probabilities are constructed in real-time and compared on forecast accuracy. We find that disagreement has predictive power indeed and that this variable can be used to improve forecasts when used in Markov regime-switching models.
    Keywords: model forecasts; expert forecasts; survey forecasts; Markov regime-switching models; disagreement; time series
    JEL: C53
  4. Heuristic model selection for leading indicators in Russia and Germany
    Date: 2011
    By: Ivan Savin (Justus Liebig University Giessen)
    Peter Winker (Justus Liebig University Giessen)
    Business tendency survey indicators are widely recognized as a key instrument for business cycle forecasting. Their leading indicator property is assessed with regard to forecasting industrial production in Russia and Germany. For this purpose, vector autoregressive (VAR) models are specified and estimated to construct forecasts. As the potential number of lags included is large, we compare full–specified VAR models with subset models obtained using a Genetic Algorithm enabling ’holes’ in multivariate lag structures. The problem is complicated by the fact that a structural break and seasonal variation of indicators have to be taken into account. The models allow for a comparison of the dynamic adjustment and the forecasting performance of the leading indicators for both countries revealing marked differences between Russia and Germany.
    Keywords: Leading indicators, business cycle forecasts, VAR, model selection, genetic algorithms
    JEL: C32
  5. Nonlinear Forecasting with Many Predictors using Kernel Ridge Regression
    Date: 2011-01-11
    By: Peter Exterkate (Erasmus University Rotterdam)
    Patrick J.F. Groenen (Erasmus University Rotterdam)
    Christiaan Heij (Erasmus University Rotterdam)
    Dick van Dijk (Erasmus University Rotterdam)
    This paper puts forward kernel ridge regression as an approach for forecasting with many predictors that are related nonlinearly to the target variable. In kernel ridge regression, the observed predictor variables are mapped nonlinearly into a high-dimensional space, where estimation of the predictive regression model is based on a shrinkage estimator to avoid overfitting. We extend the kernel ridge regression methodology to enable its use for economic time-series forecasting, by including lags of the dependent variable or other individual variables as predictors, as is typically desired in macroeconomic and financial applications. Monte Carlo simulations as well as an empirical application to various key measures of real economic activity confirm that kernel ridge regression can produce more accurate forecasts than traditional linear methods for dealing with many predictors based on principal component regression.
    Keywords: High dimensionality; nonlinear forecasting; ridge regression; kernel methods
    JEL: C53
  6. Combining Predictive Densities using Bayesian Filtering with Applications to US Economics Data
    Date: 2011-01-06
    By: Monica Billio (University Ca'Foscari di Venezia)
    Roberto Casarin (University Ca'Foscari di Venezia)
    Francesco Ravazzolo (Norges Bank)
    Herman K. van Dijk (Erasmus University Rotterdam)
    Using a Bayesian framework this paper provides a multivariate combination approach to prediction based on a distributional state space representation of predictive densities from alternative models. In the proposed approach the model set can be incomplete. Several multivariate time-varying combination strategies are introduced. In particular, a weight dynamics driven by the past performance of the predictive densities is considered and the use of learning mechanisms. The approach is assessed using statistical and utility-based performance measures for evaluating density forecasts of US macroeconomic time series and of surveys of stock market prices.
    Keywords: Density Forecast Combination; Survey Forecast; Bayesian Filtering; Sequential Monte Carlo
    JEL: C11
  7. Do Financial Variables Help Predict Macroeconomic Environment? The Case of the Czech Republic
    Date: 2010-12
    By: Tomas Havranek
    Roman Horvath
    Jakub Mateju
    In this paper, we 1) examine the interactions of financial variables and the macroeconomy within the block-restriction vector autoregression model and 2) evaluate to what extent the financial variables improve the forecasts of GDP growth and inflation. For this reason, various financial variables are examined, including those unexplored in previous literature, such as the share of liquid assets in the banking industry and the loan loss provision rate. Our results suggest that financial variables have a systematic and statistically significant effect on macroeconomic fluctuations. In terms of forecast evaluation, financial variables in general seem to improve the forecast of macroeconomic variables, but the predictive performance of individual financial variables varies over time, in particular during the 2008–2009 crisis.
    Keywords: Forecasting, macroeconomic and financial linkages, vector autoregressions.
    JEL: E44
  8. Real-time nowcasting of GDP: Factor model versus professional forecasters
    Date: 2010-12
    By: Liebermann, Joelle
    This paper performs a fully real-time nowcasting (forecasting) exercise of US real gross domestic product (GDP) growth using Giannone, Reichlin and Small (2008) factor model framework which enables one to handle unbalanced datasets as available in real-time. To this end, we have constructed a novel real-time database of vintages from October 2000 to June 2010 for a rich panel of US variables, and can hence reproduce, for any given day in that range, the exact information that was available to a real-time forecaster. We track the daily evolution throughout the current and next quarter of the model nowcasting performance. Analogously to Giannone et al. (2008) pseudo real-time results, we find that the precision of the nowcasts increases with information releases. Furthermore, the Survey of Professional Forecasters (SPF) does not carry additional information with respect to the model best specification, suggesting that the often cited superiority of the SPF, attributable to judgment, is weak over our sample. Then, as one moves forward along the real-time data flow, the continuous updating of the model provides a more precise estimate of current quarter GDP growth and the SPF becomes stale compared to all the model specifications. These results are robust to the recent recession period.
    Keywords: Real-time data; Nowcasting; Forecasting; Factor model
    JEL: C53
  9. Inflation and unemployment in Switzerland: from 1970 to 2050
    Date: 2011-02-14
    By: Kitov, Oleg
    Kitov, Ivan
    An empirical model is presented linking inflation and unemployment rate to the change in the level of labour force in Switzerland. The involved variables are found to be cointegrated and we estimate lagged linear deterministic relationships using the method of cumulative curves, a simplified version of the 1D Boundary Elements Method. The model yields very accurate predictions of the inflation rate on a three year horizon. The results are coherent with the models estimated previously for the US, Japan, France and other developed countries and provide additional validation of our quantitative framework based solely on labour force. Finally, given the importance of inflation forecasts for the Swiss monetary policy, we present a prediction extended into 2050 based on official projections of the labour force level.
    Keywords: Inflation; Unemployment; Labour force; Forecasting; Switzerland
    JEL: J21

Taken from the NEP-FOR mailing list edited by Rob Hyndman.