|
on Forecasting |
| By: | Thorsten Simon; Peter Fabsic; Georg J. Mayr; Nikolaus Umlauf; Achim Zeileis |
| Abstract: | A probabilistic forecasting method to predict thunderstorms in the European Eastern Alps is developed. A statistical model links lightning occurrence from the ground-based ALDIS detection network to a large set of direct and derived variables from a numerical weather prediction (NWP) system. The NWP system is the high resolution run (HRES) of the European Centre for Medium-Range Weather Forecasts (ECMWF). The statistical model is a generalized additive model (GAM) framework, which is estimated by Markov chain Monte Carlo (MCMC) simulation. Gradient boosting with stability selection serves as a tool for selecting a stable set of potentially nonlinear terms. Three grids from 64×64 km 2 to 16×16 km 2 and 5 forecasts horizons from 5 to 1 day ahead are investigated to predict thunderstorms during afternoons (1200 UTC to 1800 UTC). Frequently selected covariates for the nonlinear terms are variants of convective precipitation, convective potential available energy, relative humidity and temperature in the mid layers of the troposphere, among others. All models, even for a lead time of five days, outperform a forecast based on climatology in an out-of-sample comparison. An example case illustrates that coarse spatial patterns are already successfully forecast five days ahead. |
| Keywords: | lightning detection data, statistical post-processing, generalized additive models, gradient boosting, stability selection, MCMC |
| JEL: | C11 C53 Q54 |
| Date: | 2017–12 |
| URL: | http://d.repec.org/n?u=RePEc:inn:wpaper:2017-25&r=for |
| By: | Cukierman, Alex; Lustenberger, Thomas |
| Abstract: | This paper develops a model of honest rational professional forecasters with different abilities and submits it to empirical verification using data on three and twelve months ahead forecasts of short term interest rates and of long term bond yields for up to 33 countries using data collected by Consensus Economics. The main finding is that, in many countries, less precise forecasters weigh public information more heavily than more precise forecasters who weigh their own private information relatively more heavily. One implication of this result is that less precise forecasters herd after more precise forecasters even in the absence of strategic considerations. The second part of the paper discusses and examines the cross-country relationships between measures of forecast uncertainty, dispersion of forecasts across individual forecasters and the variabilities of short term interest rates and of long term bonds. The main findings are: (i) Forecast uncertainty and dispersion are positively and significantly related across countries for both short rates and yields. (ii) A similar positive, albeit somewhat weaker, association is found between uncertainty and variability. (iii) Dispersion of short term interest rate forecasts and the variability of those rates are also positively associated. The paper also documents differences between the average forecasting errors of more and less able forecasters as well as substantial correlations between the forecast errors of different forecasters. |
| Keywords: | cross-country relation between forecast dispersion and uncertainty.; Forecasting interest rates and bond yields; impact of forecasting ability on forecast formation |
| JEL: | E47 G17 |
| Date: | 2017–12 |
| URL: | http://d.repec.org/n?u=RePEc:cpr:ceprdp:12489&r=for |
| By: | Manuel Gebetsberger; Jakob W. Messner; Georg J. Mayr; Achim Zeileis |
| Abstract: | Non-homogeneous regression models are widely used to statistically post-process numerical ensemble weather prediction models. Such regression models are capable of forecasting full probability distributions and correct for ensemble errors in the mean and variance. To estimate the corresponding regression coefficients, minimization of the continuous ranked probability score (CRPS) has widely been used in meteorological post-processing studies and has often been found to yield more calibrated forecasts compared to maximum likelihood estimation. From a theoretical perspective, both estimators are consistent and should lead to similar results, provided the correct distribution assumption about empirical data. Differences between the estimated values indicate a wrong specification of the regression model. This study compares the two estimators for probabilistic temperature forecasting with non-homogeneous regression, where results show discrepancies for the classical Gaussian assumption. The heavy-tailed logistic and Student-t distributions can improve forecast performance in terms of sharpness and calibration, and lead to only minor differences between the estimators employed. Finally, a simulation study confirms the importance of appropriate distribution assumptions and shows that for a correctly specified model the maximum likelihood estimator is slightly more efficient than the CRPS estimator. |
| Keywords: | ensemble post-processing, maximum likelihood, CRPS minimization, probabilistic forecasting, distributional regression models |
| JEL: | C13 C15 C16 C51 C61 |
| Date: | 2017–11 |
| URL: | http://d.repec.org/n?u=RePEc:inn:wpaper:2017-23&r=for |
| By: | Daniel Kosiorowski; Dominik Mielczarek; Jerzy. P. Rydlewski |
| Abstract: | In economics we often face a system, which intrinsically imposes a structure of hierarchy of its components, i.e., in modelling trade accounts related to foreign exchange or in optimization of regional air protection policy. A problem of reconciliation of forecasts obtained on different levels of hierarchy has been addressed in the statistical and econometric literature for many times and concerns bringing together forecasts obtained independently at different levels of hierarchy. This paper deals with this issue in case of a hierarchical functional time series. We present and critically discuss a state of art and indicate opportunities of an application of these methods to a certain environment protection problem. We critically compare the best predictor known from the literature with our own original proposal. Within the paper we study a macromodel describing a day and night air pollution in Silesia region divided into five subregions. |
| Date: | 2017–11 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:1712.03797&r=for |
| By: | Bucci, Andrea |
| Abstract: | Modeling financial volatility is an important part of empirical finance. This paper provides a literature review of the most relevant volatility models, with a particular focus on forecasting models. We firstly discuss the empirical foundations of different kinds of volatility. The paper, then, analyses the non-parametric measure of volatility, named realized variance, and its empirical applications. A wide range of realized volatility models, both univariate and multivariate, is presented, such as time series models, MIDAS and GARCH-MIDAS models, Realized GARCH, and HEAVY models. We further discuss forecasting evaluation methods specifically suited for volatility models. |
| Keywords: | Realized Volatility; Stochastic Volatility; Volatility Models |
| JEL: | C22 C53 G10 |
| Date: | 2017–12 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:83232&r=for |
| By: | Beckmann, J; Koop, G; Korobilis, D; Schüssler, R |
| Abstract: | This paper considers how an investor in foreign exchange markets might exploit predictive information in macroeconomic fundamentals by allowing for switching between multivariate time series regression models. These models are chosen to reflect a wide array of established empirical and theoretical stylized facts. In an application involving monthly exchange rates for seven countries, we find that an investor using our methods to dynamically allocate assets achieves significant gains relative to benchmark strategies. In particular, we find strong evidence for fast model switching, with most of the time only a small set of macroeconomic fundamentals being relevant for forecasting. |
| Date: | 2017–12–08 |
| URL: | http://d.repec.org/n?u=RePEc:esy:uefcwp:20781&r=for |
| By: | Pandow, Bilal |
| Abstract: | The persistence in manager’s ability to select stocks and to time risk factors is a vital issue for accessing the performance of any asset management company. The fund manager who comes out successful today, whether the same will be able to sustain the performance in the future is a matter of concern to the investors and other stakeholders. More than the stock picking ability of fund managers, one would be interested in knowing whether there is consistency in selectivity and timing performance or not. If a fund manager is able to deliver better performance consistently i.e. quarter-after-quarter or year-after-year, then the managers’ performance in selecting the right type of stocks for the portfolio would be considered satisfactory. This paper has attempted to analyze the persistence in both stock selection and timing performance of mutual fund managers in India through Henriksson & Morton; Jenson, and Fama’s model over a period of five years. It is found that the fund managers present persistence in selection skills, however, the sample funds haven’t shown progressive timing skills in the Indian context. |
| Keywords: | Persistence, selectivity, timing, performance, mutual funds, economy |
| JEL: | G2 |
| Date: | 2017–11–25 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:82975&r=for |