Menu Content/Inhalt
Seminars Print
previous year previous month next month next year
See by year See by month See by week See Today Search Jump to month
Jozef Barunik & Lukas Vacha, Charles U. Print
Thursday, 16 February 2012, 10:45 - 13:15

Jozef Barunik & Lukas Vacha, Charles University

Seminar 1: 10:45-12:00 (presenter: L.Vacha)

WAVELET ANALYSIS IN ECONOMIC MODELING

LUKAS VACHA and JOZEF BARUNIK

Institute of Information Theory and Automation, Academy of Sciences of the Czech Republic, Pod

Vodarenskou vezi 4, Prague 8, 182 08, Czech Republic

Institute of Economic Studies, Charles University in Prague, Opletalova 21, 110 00, Prague

A major part of economic time series analysis is done in the time or frequency domain separately. Wavelet analysis combines these two fundamental approaches allowing us study time series in the time-frequency domain. Wavelet methods are free from the assumption of stationarity, thus have significant advantages over many other parametric models. Using wavelets, we are able to analyze structural changes, shocks or transient behavior without violating assumptions of the method. We demonstrate benefits of the wavelet analysis on empirical examples using both univariate and multivariate wavelet tools on daily and high frequency financial data. Namely, we advocate usage of wavelet power spectra, coherence and phase differences.

Seminar2: 12:15-13:30 (presenter: J.Barunik)

REALIZED WAVELET-BASED ESTIMATION OF MULTIVARIATE STOCK MARKET VOLATILITY

JOZEF BARUNIK and LUKAS VACHA

Institute of Economic Studies, Charles University in Prague, Opletalova 21, 110 00, Prague

Institute of Information Theory and Automation, Academy of Sciences of the Czech Republic, Pod

Vodarenskou vezi 4, Prague 8, 182 08, Czech Republic

One of the most fundamental issues in finance is research of the covariance generating process between asset returns. Demand for accurate covariance estimation is becoming more important for risk measurement and portfolio optimization than ever before. We contribute to the current literature and provide a generalized theory for realized measures of covariance which are robust to jumps as well as noise in the underlying process. We introduce wavelet-based realized covariance estimator which brings the covariance estimation into the timefrequency domain for the first time. Moreover, we also present a methodology for detecting multivariate co-jumps using wavelets. The theory is supported by the small sample study of the behavior of estimators. Finally, our estimator is used to decompose the realized covariation of the real-world data into several investment horizon covariations, individual jumps and co-jumps. We utilize this decomposition to construct an ARFIMA-type model for forecasting, and we show that our estimator of realized covariance carries over the highest information content for forecasting of realized covariations and correlations. We also study the impact of co-jumps on the forecasts and find significant bias in the correlation estimates in the presence of cojumps

Location: R42.2.113
Contact: Claude Adan, This e-mail address is being protected from spam bots, you need JavaScript enabled to view it

Back