Modeling Liquidity Adjusted Value at Risk Using Quantile Regression Analysis

Abstract

The master's thesis deals with modeling Value at Risk model adjusted by liquid- ity. For this purpose we use quantile regression analysis and liquidity proxies. We find out that Garman-Klass volatility estimator can be very useful in pe- riod 2000-2008 for the small and mid-size semiconductor companies but not in period 2008-2015. The NASDAQ composite Garman-Klass volatility is useful for all semiconductor companies for period 2008-2015. We might conclude that from the outbreak of the crisis returns of all semiconductor companies might depend on movement of NASDAQ composite index. We use Amihud and Roll measures as the liquidity proxies but the results are not persuasive regardless or size of companies and period we analyzed. JEL Classification G11, G14, G17, G18, G32 Keywords liquidity, value at risk, quantile regression Author's e-mail michalnd@gmail.com Supervisor's e-mail barunik@utia.cas.cz Abstrakt Diplomová práce se zabývá modelováním hodnoty v risku upravenou o likvid- itu. Pro tuto analýzu jsme použili kvantilovou regresi a proměnné indikující likviditu. Došli jsme k závěru, že Garman-Klass volatility estimator je velmi užitečný pro malé a středně velké firmy operující na trhu s polovodiči a to v ob- dobí 2000-2007, nikoliv však období 2008-2015. NASDAQ composite...

The master's thesis deals with modeling Value at Risk model adjusted by liquid- ity. For this purpose we use quantile regression analysis and liquidity proxies. We find out that Garman-Klass volatility estimator can be very useful in pe- riod 2000-2008 for the small and mid-size semiconductor companies but not in period 2008-2015. The NASDAQ composite Garman-Klass volatility is useful for all semiconductor companies for period 2008-2015. We might conclude that from the outbreak of the crisis returns of all semiconductor companies might depend on movement of NASDAQ composite index. We use Amihud and Roll measures as the liquidity proxies but the results are not persuasive regardless or size of companies and period we analyzed. JEL Classification G11, G14, G17, G18, G32 Keywords liquidity, value at risk, quantile regression Author's e-mail michalnd@gmail.com Supervisor's e-mail barunik@utia.cas.cz Abstrakt Diplomová práce se zabývá modelováním hodnoty v risku upravenou o likvid- itu. Pro tuto analýzu jsme použili kvantilovou regresi a proměnné indikující likviditu. Došli jsme k závěru, že Garman-Klass volatility estimator je velmi užitečný pro malé a středně velké firmy operující na trhu s polovodiči a to v ob- dobí 2000-2007, nikoliv však období 2008-2015. NASDAQ composite...