New forecasting papers 2008-06-02

In this issue we have: Predicting Stock Market Returns by Combining Forecasts ; Assessing probabilistic forecasts about particular situations ; General to specific modelling of exchange rate volatility ; Stock Picking via Nonsymmetrically Pruned Binary Decision Trees ; Some New Approaches to Forecasting the Price of Electricity, and more.


  • Predicting Stock Market Returns by Combining Forecasts
    Date: 2008-03
    By: Laurence Fung (Research Department, Hong Kong Monetary Authority)
    Ip-wing Yu (Research Department, Hong Kong Monetary Authority)
    The predictability of stock market returns has been a challenge to market practitioners and financial economists. This is also important to central banks responsible for monitoring financial market stability. A number of variables have been found as predictors of future stock market returns with impressive in-sample results. Nonetheless, the predictive power of these variables has often performed poorly for out-of-sample forecasts. This study utilises a new method known as "Aggregate Forecasting Through Exponential Re-weighting (AFTER)" to combine forecasts from different models and achieve better out-of-sample forecast performance from these variables. Empirical results suggest that, for longer forecast horizons, combining forecasts based on AFTER provides better out-of-sample predictions than the historical average return and also forecasts from models based on commonly used model selection criteria.
    Keywords: Forecasting, Model combination, Model uncertainty
    JEL: G11 G12 C13

  • Assessing probabilistic forecasts about particular situations
    Date: 2008-04-25
    By: Green, Kesten C.
    How useful are probabilistic forecasts of the outcomes of particular situations? Potentially, they contain more information than unequivocal forecasts and, as they allow a more realistic representation of the relative likelihood of different outcomes, they might be more accurate and therefore more useful to decision makers. To test this proposition, I first compared a Squared-Error Skill Score (SESS) based on the Brier score with an Absolute-Error Skill Score (AESS), and found the latter more closely coincided with decision-makers' interests. I then analysed data obtained in researching the problem of forecasting the decisions people make in conflict situations. In that research, participants were given lists of decisions that might be made and were asked to make a prediction either by choosing one of the decisions or by allocating percentages or relative frequencies to more than one of them. For this study I tr! ansformed the percentage and relative frequencies data into probabilistic forecasts. In most cases the participants chose a single decision. To obtain more data, I used a rule to derive probabilistic forecasts from structured analogies data, and transformed multiple singular forecasts for each combination of forecasting method and conflict into probabilistic forecasts. When compared using the AESS, probabilistic forecasts were not more skilful than unequivocal forecasts.
    Keywords: accuracy; error measures; evaluation; forecasting methods; prediction
    JEL: D74 C0 F51

  • General to specific modelling of exchange rate volatility : a forecast evaluation
    Date: 2008-04
    By: Luc Bauwens
    Genaro Sucarrat
    The general-to-specific (GETS) methodology is widely employed in the modelling of economic series, but less so in financial volatility modelling due to computational complexity when many explanatory variables are involved. This study proposes a simple way of avoiding this problem when the conditional mean can appropriately be restricted to zero, and undertakes an out-of-sample forecast evaluation of the methodology applied to the modelling of weekly exchange rate volatility. Our findings suggest that GETS specifications perform comparatively well in both ex post and ex ante forecasting as long as sufficient care is taken with respect to functional form and with respect to how the conditioning information is used. Also, our forecast comparison provides an example of a discrete time explanatory model being more accurate than realised volatility ex post in 1 step forecasting.
    Keywords: Exchange rate volatility, General to specific, Forecasting
    JEL: C53 F31

  • The Intranational Business Cycle: Evidence from Japan
    Date: 2008
    By: Pedro M.D.C.B. Gouveia
    Denise R. Osborn
    Paulo M.M. Rodrigues
    Forecast combination methodologies exploit complementary relations between different types of econometric models and often deliver more accurate forecasts than the individual models on which they are based. This paper examines forecasts of seasonally unadjusted monthly industrial production data for 17 countries and the Euro Area, comparing individual model forecasts and forecast combination methods in order to examine whether the latter are able to take advantage of the properties of different seasonal specifications. In addition to linear models (with deterministic seasonality and with nonstationary stochastic seasonality), more complex models that capture nonlinearity or seasonally varying coefficients (periodic models) are also examined. Although parsimonous periodic models perform well for some countries, forecast combinations provide the best overall performance at short horizons, implying that utilizing the! characteristics captured by different models can contribute to improved forecast accuracy.

  • The role of country-specific trade and survey data in forecasting euro area manufacturing production. Perspective from Large Panel factor models.
    Date: 2008-05
    By: Laurent Maurin (European Central Bank, Kaiserstrasse 29, 60311 Frankfurt am Main, Germany.)
    Matthieu Darracq Pariès (European Central Bank, Kaiserstrasse 29, 60311 Frankfurt am Main, Germany.)
    Several factor-based models are estimated to investigate the role of country-specific trade and survey data in forecasting euro area manufacturing production. Following Boivin and Ng (2006), the emphasis is put on the role of dataset selection on the empirical performance of factor models. First, spectral analysis is used to assess the information content for euro area manufacturing production of external trade and surveys data of the three largest economies as well as two medium-sized highly opened economies. Second, common factors are estimated on four datasets, following twomethodologies, Stock andWatson (2002a, 2002b) and Forni et al. (2005). Third, a rolling out of sample forecast comparison exercise is carried out on ninemodels. Compared to univariate benchmarks, our results are supportive of factor-basedmodels up to two quarters. They show that incorporating survey and external trade information improves th! e forecast of manufacturing production. They also confirm the findings of Marcellino, Stock and Watson (2003) that, using country information, it is possible to improve forecasts for the euro area. Interesting, the medium-sized highly opened economies provide valuable information to monitor area wide developments, beyond their weight in the aggregate. Conversely, the large countries do not add much to the monitoring of the aggregate, when considered separately. JEL Classification: E37, C3, C53.
    Keywords: Factor models, Dataset, Forecasting.

  • Stock Picking via Nonsymmetrically Pruned Binary Decision Trees
    Date: 2008-05
    By: Anton Andriyashin
    Stock picking is the field of financial analysis that is of particular interest for many professional investors and researchers. In this study stock picking is implemented via binary classification trees. Optimal tree size is believed to be the crucial factor in forecasting performance of the trees. While there exists a standard method of tree pruning, which is based on the cost-complexity tradeoff and used in the majority of studies employing binary decision trees, this paper introduces a novel methodology of nonsymmetric tree pruning called Best Node Strategy (BNS). An important property of BNS is proven that provides an easy way to implement the search of the optimal tree size in practice. BNS is compared with the traditional pruning approach by composing two recursive portfolios out of XETRA DAX stocks. Performance forecasts for each of the stocks are provided by constructed decision trees. It is shown that BN! S clearly outperforms the traditional approach according to the backtesting results and the Diebold-Mariano test for statistical significance of the performance difference between two forecasting methods.
    Keywords: decision tree, stock picking, pruning, earnings forecasting, data mining
    JEL: C14 C15 C44 C63 G12

  • Some New Approaches to Forecasting the Price of Electricity: A Study of Californian Market
    Date: 2008-01-01
    By: Eduardo Mendes
    Les Oxley (University of Canterbury)
    Marco Reale
    In this paper we consider the forecasting performance of a range of semi- and non- parametric methods applied to high frequency electricity price data. Electricity price time-series data tend to be highly seasonal, mean reverting with price jumps/spikes and time- and price-dependent volatility. The typical approach in this area has been to use a range of tools that have proven popular in the financial econometrics literature, where volatility clustering is common. However, electricity time series tend to exhibit higher volatility on a daily basis, but within a mean reverting framework, albeit with occasional large 'spikes'. In this paper we compare the existing forecasting performance of some popular parametric methods, notably GARCH AR-MAX, with approaches that are new to this area of applied econometrics, in particular, Artificial Neural Networks (ANN); Linear Regression Trees, Local Regressions and Generali! sed Additive Models. Section 2 presents the properties and definitions of the models to be compared and Section 3 the characteristics of the data used which in this case are spot electricity prices from the Californian market 07/1999-12/2000. This period includes the 'crisis' months of May-August 2000 where extreme volatility was observed. Section 4 presents the results and ranking of methods on the basis of forecasting performance. Section 5 concludes.
    Keywords: Electricty Time Series; Forecasting Performance; Semi- and Non- Parametric Methods
    JEL: C14 C45 C53

  • The usefulness of infra-annual government cash budgetary data for fiscal forecasting in the euro area.
    Date: 2008-05
    By: Luca Onorante (DG Economics, European Central Bank, Kaiserstrasse 29, 60311 Frankfurt am Main, Germany.)
    Diego J. Pedregal (ETSI Industriales, Edificio Politécnico, Universidad de Castilla-la-Mancha, campus universitario s/n, 13071 Ciudad Real, Spain.)
    Javier J. Pérez (European Central Bank, Kaiserstrasse 29, 60311 Frankfurt am Main, Germany.)
    Sara Signorini (Global Economics & FI/FX Research, HVB Milan, Via Tommaso Grossi 10, 20121 Milan, Italy.)
    Short-term fiscal indicators based on public accounts data are often used by European policy makers. They represent one of the main sources of publicly available intra-annual fiscal information. Nevertheless, these indicators have received limited attention from the academic literature analysing fiscal forecasting in Europe. Some recent literature suggests the validity of public accounts data to forecast government deficits in the euro area. We extend this literature on two fronts: (i) we shift the focus from indicators of government deficits to look at indicators for government total revenue and total expenditure; (ii) we use a mixed-frequency state-space model to integrate readily available monthly/quarterly cash-based fiscal data with annual general government series (National Accounts). By doing so, we are able to maintain the focus on forecasting and monitoring annual outcomes, while making use of infra-annua! l fiscal information, available within the current year. The paper makes a case for the use of monthly cash indicators for multilateral fiscal surveillance at the European level. JEL Classification: C53, E6, H6.
    Keywords: Leading indicators, Fiscal forecasting and monitoring, Euro area.

  • Scenarios of Carbon Dioxide Emissions from Aviation
    Date: 2008-05
    By: Karen Mayor (Economic and Social Research Institute (ESRI))
    Richard S. J. Tol (Economic and Social Research Institute (ESRI))
    We use a model of international and domestic tourist numbers and flows to forecast tourist numbers and emissions from international tourism out to 2100. We find that between 2005 and 2100 international tourism grows by a factor of 12. Not only do people take more trips but these also increase in length. We find that the growth in tourism is mainly fuelled by an increase in trips from Asian countries. Emissions follow this growth pattern until 2060 when emissions per passenger-kilometre start to fall due to improvements in fuel efficiency. Forecasted emissions are also presented for the four SRES scenarios and maintain the same growth pattern but the levels of emissions differ substantially. We find that the forecasts are sensitive to the period to which the model is calibrated, the assumed rate of improvement in fuel efficiency and the imposed climate policy scenario.
    Keywords: Carbon dioxide emissions, international tourism, long-term forecasting, aviation
  • Taken from the NEP-FOR mailing list edited by Rob Hyndman.