nep-for New Economics Papers
on Forecasting
Issue of 2007‒01‒23
eight papers chosen by
Rob J Hyndman
Monash University

  1. Forecast errors and the macroeconomy — a non-linear relationship? By Ulrich Fritsche; Joerg Doepke
  2. Regional employment forecasts with spatial interdependencies By Hampel, Katharina; Kunz, Marcus; Schanne, Norbert; Wapler, Rüdiger; Weyh, Antje
  3. Backtesting VaR Models: An Expected Shortfall Approach By Timotheos Angelidis; Stavros Degiannakis
  4. Periodic Unobserved Cycles in Seasonal Time Series with an Application to US Unemployment By Siem Jan Koopman; Marius Ooms; Irma Hindrayanto
  5. On comparing the accuracy of default predictions in the rating industry By Prof. Dr. Walter Krämer; Andre Güttler
  6. Can investors profit from banks’ stock recommendations? Evidence for the German DAX index By Pierdzioch, Christian; Kempa, Bernd; Hendricks, Torben
  7. Empirical Studies in Consumption, House Prices and the Accuracy of European Growth and Inflation Forecasts By Barot, Bharat
  8. Exchange rates, prices and their speed of adjustment By Fanelli Luca; Paruolo Paolo

  1. By: Ulrich Fritsche (Department for Economics and Politics, University of Hamburg, and DIW Berlin); Joerg Doepke (Fachhochschule Merseburg)
    Abstract: The paper analyses reasons for departures from strong rationality of growth and inflation forecasts based on annual observations from 1963 to 2004. We rely on forecasts from the joint forecast of the so-called "six leading" forecasting institutions in Germany and argue that violations of the rationality hypothesis are due to relatively few large forecast errors. These large errors are shown - based on evidence from probit models - to correlate with macroeconomic fundamentals, especially on monetary factors. We test for a non-linear relation between forecast errors and macroeconomic fundamentals and find evidence for such a non-linearity for inflation forecasts.
    Keywords: forecast error evaluation, non-linearities, business cycles
    JEL: E32 E37 C52 C53
    Date: 2006–02
  2. By: Hampel, Katharina (Institut für Arbeitsmarkt- und Berufsforschung (IAB), Nürnberg [Institute for Employment Research, Nuremberg, Germany]); Kunz, Marcus (Institut für Arbeitsmarkt- und Berufsforschung (IAB), Nürnberg [Institute for Employment Research, Nuremberg, Germany]); Schanne, Norbert (Institut für Arbeitsmarkt- und Berufsforschung (IAB), Nürnberg [Institute for Employment Research, Nuremberg, Germany]); Wapler, Rüdiger (Institut für Arbeitsmarkt- und Berufsforschung (IAB), Nürnberg [Institute for Employment Research, Nuremberg, Germany]); Weyh, Antje (Institut für Arbeitsmarkt- und Berufsforschung (IAB), Nürnberg [Institute for Employment Research, Nuremberg, Germany])
    Abstract: "The labour-market policy-mix in Germany is increasingly being decided on a regional level. This requires additional knowledge about the regional development which (disaggregated) national forecasts cannot provide. Therefore, we separately forecast employment for the 176 German labour- market districts on a monthly basis. We first compare the prediction accuracy of standard time-series methods: autoregressive integrated moving averages (ARIMA), exponentially weighted moving averages (EWMA) and the structural-components approach (SC) in these small spatial units. Second, we augment the SC model by including autoregressive elements (SCAR) in order to incorporate the influence of former periods of the dependent variable on its current value. Due to the importance of spatial interdependencies in small labour-market units, we further augment the basic SC model by lagged values of neighbouring districts in a spatial dynamic panel (SCSAR). The prediction accuracies of the models are compared using the mean absolute percentage forecast error (MAPFE) for the simulated out-of-sample forecast for 2005. Our results show that the SCSAR is superior to the SCAR and basic SC model. ARIMA and EWMA models perform slightly better than SCSAR in many of the German labour-market districts. This reflects that these two moving-average models can better capture the trend reversal beginning in some regions at the end of 2004. All our models have a high forecast quality with an average MAPFE lower than 2.2 percent." (author's abstract, IAB-Doku) ((en))
    Keywords: regionaler Arbeitsmarkt, Beschäftigungsentwicklung, Prognoseverfahren
    JEL: C53 J21 O18
    Date: 2007–01–16
  3. By: Timotheos Angelidis; Stavros Degiannakis
    Abstract: Academics and practitioners have extensively studied Value-at-Risk (VaR) to propose a unique risk management technique that generates accurate VaR estimations for long and short trading positions and for all types of financial assets. However, they have not succeeded yet as the testing frameworks of the proposals developed, have not been widely accepted. A two-stage backtesting procedure is proposed to select a model that not only forecasts VaR but also predicts the losses beyond VaR. Numerous conditional volatility models that capture the main characteristics of asset returns (asymmetric and leptokurtic unconditional distribution of returns, power transformation and fractional integration of the conditional variance) under four distributional assumptions (normal, GED, Student-t, and skewed Student-t) have been estimated to find the best model for three financial markets, long and short trading positions, and two confidence levels. By following this procedure, the risk manager can significantly reduce the number of competing models that accurately predict both the VaR and the Expected Shortfall (ES) measures.
    Keywords: Value-at-Risk, Expected Shortfall, Volatility Forecasting, Arch Models
    JEL: C22 C52 G15
    Date: 2007–01–12
  4. By: Siem Jan Koopman (Vrije Universiteit Amsterdam); Marius Ooms (Vrije Universiteit Amsterdam); Irma Hindrayanto (Vrije Universiteit Amsterdam)
    Abstract: This paper discusses identification, specification, estimation and forecasting for a general class of periodic unobserved components time series models with stochastic trend, seasonal and cycle components. Convenient state space formulations are introduced for exact maximum likelihood estimation, component estimation and forecasting. Identification issues are considered and a novel periodic version of the stochastic cycle component is presented. In the empirical illustration, the model is applied to postwar monthly US unemployment series and we discover a significantly periodic cycle. Furthermore, a comparison is made between the performance of the periodic unobserved components time series model and a periodic seasonal autoregressive integrated moving average model. Moreover, we introduce a new method to estimate the latter model.
    Keywords: Unobserved component models; state space methods; seasonal adjustment; time–varying parameters; forecasting
    JEL: C22 C51 E32 E37
    Date: 2006–11–20
  5. By: Prof. Dr. Walter Krämer (Fachbereich Statistik, Universität Dortmund); Andre Güttler (Universität Frankfurt, Finance Department)
    Abstract: We consider 1927 borrowers from 54 countries who had a credit rating by both Moody's and S&P at the end of 1998, and their subsequent default history up to the end of 2002. Viewing bond ratings as predicted probabilities of default, we consider partial orderings among competing probability forecasters and show that Moody's and S&P cannot be ordered according to any of these. Therefore, the relative performance of the agencies depends crucially on the way in which probability predictions are compared.
    Keywords: credit rating, probability forecasts, calibration
    Date: 2006–10
  6. By: Pierdzioch, Christian; Kempa, Bernd; Hendricks, Torben
    Abstract: We find that banks’ buy and sell recommendations only have a minor effect on the out-of-sample predictability of daily stock returns and the market-timing ability of an investor trading in real time in the German DAX30 stock index. Banks’ stock recommendations may improve the performance of simple trading rules in real time. These improvements, however, are in general small and sensitive to the model-selection criterion being used by an investor to set up a forecasting model for stock returns. Moreover, banks’ stock recommendations are more useful for forecasting one-day-ahead stock returns than one-week-ahead stock returns.
    Keywords: Forecasting stock returns; trading rules; buy and sell recommendations by banks
    JEL: C53
    Date: 2007–01–09
  7. By: Barot, Bharat (National Institute of Economic Research)
    Abstract: BAROT, Bharat, This Ph.D. thesis, Empirical Studies in Consumption, House Prices and the Accuracy of European Growth and Inflation Forecasts contains four self-contained chapters:<p><p> Chapter I gives a brief introduction to the topic of the thesis and summarizes the main results. <p><p> Chapter II an aggregated consumption function based on the life cycle hypothesis using the error correction methodology is estimated for Sweden. Wealth in its disaggregated form (financial and housing wealth) is incorporated in the consumption function, along with basic standard explanatory variables including the unemployment variable. Applying Hendry’s general to specific modelling strategy one final model is deduced. The study finds that each of the primary components of wealth has an equal role for consumer’s expenditure. In addition the study finds significant effects from employment and interest rates.<p><p> Chapter III a stock-flow model serves as the theoretical basis for the fundamental determinants of real estate construction and prices. A housing market model for Sweden has been estimated on semi-annual data for 1970-1998 by separately modelling the demand and the supply sides, specified in error correction form. The supply side is based on Tobin’s q-index. The results indicate that even in a turbulent period, Swedish house prices and housing investment are tracked quite well with this specification. The importance of the simulations and their usefulness to Swedish policy makers is discussed. Both ex post and ex ante forecasts using the model gives reasonable results.<p><p> Chapter IV (with Zan Yang), we estimate quarterly dynamic housing demand and investment supply models for Sweden and the UK for the sample period 1970-1998, using an Error Correction Method (ECM). In order to facilitate comparisons of results between Sweden and the UK we model both countries similarly using comparable exogenous variables. The long run income elasticity for Sweden and the UK are both constrained to be equal to one. The long run semi-elasticity for interest rate is 2.1 for Sweden and 0.9 for the UK. The speed of adjustment on the demand side is 12% and 23% for Sweden and the UK, respectively, while on the supply side it is 6% and 48%. Tobin’s q Granger causes housing investment.<p><p> Chapter V (with Lars-Erik Öller), evaluates the one-year ahead forecasts by the OECD and by national institutes of GDP growth and inflation in 13 European countries. RMSE was large 1.9% for growth and 1.6% for inflation. Six (11) OECD and ten (7) institute growth forecasts records were significantly better than an average growth forecast (the current year forecast). All full record-length inflation forecasts were significantly better than both naive alternatives. There were no significant differences in accuracy between the forecasts of the OECD and the institutes. Two forecasts were found to be biased and one had auto-correlated errors.
    Date: 2007–01–12
  8. By: Fanelli Luca (Department of Statistical Sciences, University of Bologna, Italy); Paruolo Paolo (Department of Economics, University of Insubria, Italy)
    Abstract: This paper addresses the problem of measuring the speed of adjustment of exchange rates and relative prices to purchasing power parity (PPP), in the multivariate context of Vector Autoregressive Processes (VAR). We consider the speed of adjustment of one variable y in response to another variable x, where x, y belong to the VAR. We propose a multivariate measure defined as the forecasting horizon for which the cumulated interim multiplier of x on y surpasses a given fraction p of the corresponding total multiplier. This measure of speed for p = 1/2 coincides with the usual concept of half-life when restricted to univariate processes. We emphasize the importance to separate the concepts of long run e¤ect size and its speed of adjustment, where the latter is unambiguosly defined only when the long run e¤ect is non-zero. We discuss likelihood-based point estimators and confidence sets for this notion of half-life, and reconsider evidence on adjustment to PPP in monthly post-Bretton Woods data for five major industrialized countries against the U.S. dollar. Results show that nominal exchange rates bu¤er the entire adjustment to PPP disequilibrium, wheras relative prices do not adjust either in the short or the long run to PPP deviations. Concluding in such a situation that prices adjust faster than exchange rates is a matter of how one interprets the absence of short run and long run effects.
    Keywords: Invariance, Half-life, purchasing power parity, impact factors, speed of adjustment, vector equilibrium correction, upcrossing and downcrossing.
    JEL: C32 C52 F31
    Date: 2006–09

This nep-for issue is ©2007 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.