|
on Econometric Time Series |
| By: | BOUEZMARNI, Taoufik; ROMBOUTS, Jeroen (UniversitŽ catholique de Louvain (UCL). Center for Operations Research and Econometrics (CORE)); TAAMOUTI, Abderrahim |
| Keywords: | nonparametric tests, conditional independence, Granger non-causality, Bernstein density copula, bootstrap, finance, volatility asymmetry, leverage effect, volatility feedback effect, macroeconomics |
| JEL: | C12 C14 C15 C19 G1 G12 E3 E4 E52 |
| Date: | 2009–06–01 |
| URL: | http://d.repec.org/n?u=RePEc:cor:louvco:2009041&r=ets |
| By: | HEINEN, AndrŽas (Departamento de Estadistica, Universidad Carlos III de Madrid, Spain); VALDESOGO, Alfonso (CREA, University of Luxembourg, Luxembourg) |
| Abstract: | We propose a new dynamic model for volatility and dependence in high dimensions, that allows for departures from the normal distribution, both in the marginals and in the dependence. The dependence is modeled with a dynamic canonical vine copula, which can be decomposed into a cascade of bivariate conditional copulas. Due to this decomposition, the model does not suffer from the curse of dimensionality. The canonical vine autoregressive (CAVA) captures asymmetries in the dependence structure. The model is applied to 95 S&P500 stocks. For the marginal distributions, we use non-Gaussian GARCH models, that are designed to capture skewness and kurtosis. By conditioning on the market index and on sector indexes, the dependence structure is much simplified and the model can be considered as a non-linear version of the CAPM or of a market model with sector effects. The model is shown to deliver good forecasts of Value-at-Risk. |
| Keywords: | asymmetric dependence, high dimension, multivariate copula, multivariate GARCH, Value-at-Risk |
| JEL: | C32 C53 G10 |
| Date: | 2009–11–01 |
| URL: | http://d.repec.org/n?u=RePEc:cor:louvco:2009069&r=ets |
| By: | S. Bordignon; D. Raggi |
| Abstract: | Goal of this paper is to analyze and forecast realized volatility through nonlinear and highly persistent dynamics. In particular, we propose a model that simultaneously captures long memory and nonlinearities in which level and persistence shift through a Markov switching dynamics. We consider an efficient Markov chain Monte Carlo (MCMC) algorithm to estimate parameters, latent process and predictive densities. The insample results show that both long memory and nonlinearities are significant and improve the description of the data. The out-sample results at several forecast horizons, show that introducing these nonlinearities produces superior forecasts over those obtained from nested models. |
| Date: | 2010–02 |
| URL: | http://d.repec.org/n?u=RePEc:bol:bodewp:694&r=ets |
| By: | Timotheos Angelidis; Alexandros Benos; Stavros Degiannakis |
| Abstract: | We evaluate the performance of an extensive family of ARCH models in modelling daily Value-at-Risk (VaR) of perfectly diversified portfolios in five stock indices, using a number of distributional assumptions and sample sizes. We find, first, that leptokurtic distributions are able to produce better one-step-ahead VaR forecasts; second, the choice of sample size is important for the accuracy of the forecast, whereas the specification of the conditional mean is indifferent. Finally, the ARCH structure producing the most accurate forecasts is different for every portfolio and specific to each equity index. |
| Keywords: | Value at Risk, GARCH estimation, Backtesting, Volatility forecasting, Quantile Loss Function. |
| Date: | 2010 |
| URL: | http://d.repec.org/n?u=RePEc:uop:wpaper:0048&r=ets |
| By: | Joanne S. Ercolani |
| Abstract: | This paper considers a fractional noise model in continuous time and examines the asymptotic properties of a feasible frequency domain maximum likelihood estimator of the long memory parameter. The feasible estimator is one that maximises an approximation to the likelihood function (the approximation arises from the fact that the spectral density function involves the finite truncatin of an infinite summation). It is of interest therefore to explore the conditions required of this approximation to ensure the consistency and asymptotic normality of this estimator. It is shown that the truncation parameter has to be a function of the sample size and that the optimal rate is different for stocks and flows and is a function of the long memory parameter itself. The results of a simulation exercise are provided to assess the small sample properties of the estimator. |
| Keywords: | Continuous time models, long memory processes |
| JEL: | C22 |
| Date: | 2010–03 |
| URL: | http://d.repec.org/n?u=RePEc:bir:birmec:10-09&r=ets |