Realized Jump GARCH model: Can decomposition of volatility improve its forecasting?

• Název práce v češtině: Realized Jump GARCH model: pomůže dekompozice volatility vylepšit predikční schopnosti modelu?
• Název v anglickém jazyce: Realized Jump GARCH model: Can decomposition of volatility improve its forecasting?
• Klíčová slova: Neparametrické odhady realizované volatility, skoky, vysokofrekvenční data, GARCH, Realized (Jump) GARCH, HAR
• Klíčová slova anglicky: Realized measures of volatility, Jumps, High-frequency data, GARCH, Realized (Jump) GARCH, HAR
• Akademický rok vypsání: 2012/2013
• Typ práce: diplomová práce
• Jazyk práce: angličtina
• Ústav: Institut ekonomických studií (23-IES)
• Vedoucí / školitel: doc. PhDr. Jozef Baruník, Ph.D.
• Řešitel: skrytý - zadáno vedoucím/školitelem
• Datum přihlášení: 08.02.2013
• Datum zadání: 08.02.2013
• Datum a čas obhajoby: 24.09.2014 09:00
• Místo konání obhajoby: IES
• Datum odevzdání elektronické podoby: 26.06.2014
• Datum proběhlé obhajoby: 24.09.2014
• Oponenti: Mgr. Barbara Pertold-Gebicka, M.A., Ph.D.

Předběžná náplň práce v anglickém jazyce

Volatility of financial time-series forms a significant part of current econometric research as volatility is a key component of modern finance. Understanding of volatility is crucial to large span of fields such as risk management, hedging or asset pricing which closely rely on the volatility estimates. In other words, estimation of volatility and its accuracy is very important because this measure is relevant for all market participants and for their decision-making. The base stone of the volatility estimation forms the GARCH methodology of Engle (1982), which opened new avenues to the financial engineering. In recent years, the availability of high frequency data brought new measures of volatility and initiated the discussion about the accuracy and goodness of forecasts given by GARCH models.

Realized measures of volatility offered a possibility to gain a consistent estimate of quadratic returns variation out of high frequency data. This measure was used in a variety of models, for example Heterogeneous Auto-Regressive model. HAR models showed promising results that exceeded those gained by the GARCH model.

Promising area of research is to combine parametric models with high frequency measures. One of the most current GARCH extensions, Realized GARCH (Hansen, 2011), offers the inclusion of realized measures of volatility into variance equation and yet shows promising results. Realized measures of volatility also enable us to decompose quadratic variation into the integrated and jump variation. Current research shows that the role of jumps in the analysis of financial series is very important. The goal of this thesis is to incorporate the jumps component into the Realized GARCH model, find if there is a significant effect of jumps on volatility forecast and provide comparison with HAR and ordinary GARCH model.