ebook img

Risk Estimation on High Frequency Financial Data: Empirical Analysis of the DAX 30 PDF

78 Pages·2015·1.38 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Risk Estimation on High Frequency Financial Data: Empirical Analysis of the DAX 30

BestMasters Springer awards „BestMasters“ to the best master’s theses which have been com- pleted at renowned universities in Germany, Austria, and Switzerland. Th e studies received highest marks and were recommended for publication by supervisors. Th ey address current issues from various fi elds of research in natural sciences, psychology, technology, and economics. Th e series addresses practitioners as well as scientists and, in particular, off ers guidance for early stage researchers. Florian Jacob Risk Estimation on High Frequency Financial Data Empirical Analysis of the DAX 30 Florian Jacob Karlsruhe, Germany BestMasters ISBN 978-3-658-09388-4 ISBN 978-3-658-09389-1 (eBook) DOI 10.1007/978-3-658-09389-1 Library of Congress Control Number: 2015936902 Springer Spektrum © Springer Fachmedien Wiesbaden 2015 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifi cally the rights of translation, reprinting, reuse of illus- trations, recitation, broadcasting, reproduction on microfi lms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specifi c statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. Printed on acid-free paper Springer Spektrum is a brand of Springer Fachmedien Wiesbaden Springer Fachmedien Wiesbaden is part of Springer Science+Business Media (www.springer.com) Acknowledgements This work would not have been possible without the support, encouragement, and advice of several people. Foremost, I would like to thank Dr. habil. Young Shin (Aaron) Kim for the introduction to the topic during my time as a master student at the Karlsruher Institute of Technology,aswellasfortheguidanceandsupportthroughoutthisthesis. Iwouldalsowantto expressmygratitudetoProf. Dr. Wolf-DieterHellerandProf. Dr. Svetlozar(Zari)Rachevfor giving me the opportunity to pursue my interests in Germany and during my two month stay atStonyBrookUniversity,USA.Thistimeaswellashelpingmetodevelopaprofessionalskill set,letmedevelopasaperson. FlorianJacob Contents 1. Introduction 1 2. Theory of Time Series Modeling and Risk Estimation 5 2.1. FinancialEconometrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.1.1. GARCHmodels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.1.2. FIGARCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.2. ModelChoiceandValidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.2.1. DetectingtheAutocorrelation. . . . . . . . . . . . . . . . . . . . . . . . . 10 2.2.2. TestingforLong-RangeDependency . . . . . . . . . . . . . . . . . . . . . 11 2.3. Theinnovationprocess. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.3.1. L´evyProcesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.3.2. Subordination. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2.3.3. Time-changedStochasticProcess . . . . . . . . . . . . . . . . . . . . . . . 16 2.3.4. Time-changedBrownianMotion . . . . . . . . . . . . . . . . . . . . . . . 16 2.3.5. Theα-stableProcess. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.3.6. TemperedStableDistribution . . . . . . . . . . . . . . . . . . . . . . . . . 20 2.3.7. CTSSubordinator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2.3.8. UnivariateNormalTemperedStableDistribution . . . . . . . . . . . . . . 23 2.3.9. StandardUnivariateNTS . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 2.3.10.MultivariateStandardNormalTemperedStableDistribution . . . . . . . 26 2.4. GoodnessofFit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 2.4.1. Kolmogorov-Smirnov-Test . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 viii Contents 2.4.2. Anderson-DarlingTest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.5. RiskManagement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 2.5.1. Value-at-Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 2.5.2. CoherentRiskMeasures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 2.5.3. AverageValue-at-Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 2.5.4. Computation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 2.5.5. Backtesting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 3. Data and Methodology 37 3.1. DataSelection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 3.2. DataTransformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3.3. AutocorrelationandDependence . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 3.4. EmpiricalEstimationandGoFTests . . . . . . . . . . . . . . . . . . . . . . . . . 44 3.4.1. IndexReturns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 3.4.2. StockReturns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3.4.3. ARMA-GARCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 3.5. ComparisonofRiskMeasuresforARMA-GARCHModels . . . . . . . . . . . . . 57 3.5.1. RiskPrediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 3.5.2. Backtestings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 4. Conclusion 61 A. DAX30Stocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Bibliography 67 List of Figures 2.1. IllustrationofautocorrelationoftheS&P500of01.01.2002-22.10.2012 . . . . . 14 2.2. Illustrationoftheroleofdifferentparametersintheα-stabledistribution. . . . . 19 2.3. CTSSubordinator-Influenceofθ,α . . . . . . . . . . . . . . . . . . . . . . . . . 23 2.4. IllustrationoftheroleofdifferentparametersintheNTSdistribution . . . . . . 25 3.1. IllustrationoftheDAX30IndexduringtheanalyzedtimeperiodbetweenMarch, 22009andDecember,292009 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3.2. Intra-daily volatility profile for the DAX 30 index on 5 min log-returns for the analyzed time period between March, 2 2009 and December, 29 2009. The sum ofsquaredreturnsserveasaproxyfortheunobservedvolatilityinthischart. . 40 3.3. SampleAutocorrelationFunctionoftheabsolutelog-returnsfortheanalyzedtime periodbetweenMarch,22009andDecember,292009. . . . . . . . . . . . . . . 41 3.4. SACFandSPACFforlog-returnsaswellastheSACFforabsoluteandsquared log-returns for DAX 30 time series for 5 min high frequency data for the ana- lyzedtimeperiodbetweenMarch,22009andDecember,292009. Theyexhibit significantautocorrelationaswellassignificantandpersistentdependencyeffects. 43 3.5. R/SStatisticsforlog-returnsaswellasforabsolutelog-returnsforDAX30time seriesfor5minhighfrequencydatafortheanalyzedtimeperiodbetweenMarch, 22009andDecember,292009. . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 x ListofFigures 3.6. KerneldensitiesofstandardizedresidualsofARMA(1,1)-GARCH(1,1)incompar- ison with the fitted standard Normal Distribution and the fitted standard NTS DistributionandQQ-PlotsforDAX30timeseriesfor5minhighfrequencydata fortheanalyzedtimeperiodbetweenMarch,22009andDecember,292009. . . 46 3.7. Kernel densities of standardized residuals of ARMA(1,1)-FIGARCH(1,0.24, 1), compared to fitted standard NTS Distribution and QQ-Plots for DAX 30 time seriesfor5minhighfrequencydatafortheanalyzedtimeperiodbetweenMarch, 22009andDecember,292009. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3.8. Backtesting of MNTS ARMA(1,1)-FIGARCH(1,d,1) and MNTS ARMA(1,1)- GARCH(1,1)forbacktestingwindowof250.. . . . . . . . . . . . . . . . . . . . . 59 Allfigurescanbeaccessedonwww.springer.comundertheauthor’snameandthebooktitle.

Description:
By studying the ability of the Normal Tempered Stable (NTS) model to fit the statistical features of intraday data at a 5 min sampling frequency, Florian Jacobs extends the research on high frequency data as well as the appliance of tempered stable models. He examines the DAX30 returns using ARMA-GA
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.