The application of different ARCH/GARCH-models before and after the financial crisis
Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
In this paper the application of different ARCH/GARCH models have been considered on two foreign exchange rates; the Swedish Krona/Euro and the US dollar/Euro. The purpose of the study is to see which volatility model that is the most suitable to use depending on market climate and if there is one that can be recommended to use regardless of the state of the market. The models are estimated on two different time periods, the first period is during a stable market period from the beginning of 2003 to the end of 2007 and the other period is during the more extreme/bear market period under the financial crisis and after from the start of 2008 to the end of 2012. The study is made by first estimating the mean equation of the time series in a Box Jenkins approach and then estimating 4 different volatility models; an ARCH (1)-, a GARCH (1,1)-, an Exponential-GARCH (1,1)- and a GJR-GARCH(1,1)-model. The models are thereafter evaluated by looking at some different evaluation statistics to appraise which model that is the most suitable. From the results there seems to be a difference between the two market periods, during the first period the estimations of the symmetric ARCH (1)-model suits the data best and during the more unstable period of the years 2008 to 2012 the asymmetric GJR-GARCH (1,1)-model is the best. What should be mentioned is that the GJR-GARCH (1,1)-model is also the second best model during the first period and thus seems to be most reliable. As it is hard to know which kind of market period that is present a suggestion from this study is to always use an asymmetric GJR-GARCH (1,1)-model, as it will capture asymmetric impact of the last shock if present and if not, one will simply trade one degree of freedom as an insurance of no leverage effect.
Place, publisher, year, edition, pages
2013. , 24 p.
ARCH-model, GARCH-model, EGARCH-model, GJR-GARCH-model, foreign exchange rates
IdentifiersURN: urn:nbn:se:uu:diva-204236OAI: oai:DiVA.org:uu-204236DiVA: diva2:638166
Subject / course