Version 1.06 User Manual
Contents
1. cccccsscccccceseceeeceeeeeeeeceeeeeeeeseeeececeseeaeeeesseaaeeeessageeesessagss 30 72 W Weighted Least Squares Regression ccceecccceccceseeeseeeceeeeseeeeseeeeeeeseaeaeeeeeeeeseeaeeeeeeeeseaas 72 White Robust Standard Errors 72 NORE EE 52. Regression Dataset The software automatically denotes missing values by 20 Export Data To use the data of your Event Study Metrics Workfile in any other program you can simply copy amp paste the data from any arbitrary spreadsheet You can change the sort order of most spreadsheets by simply clicking on the desired column heading Edit Data Center Help Simula a Cut Copy Paste Settings Event Date Check Se Price gt Return Converter gt Furthermore you can transpose your dataset while exporting by choosing the Transpose option on the O Settings form This can be helpful in case you want to open a large dataset in a spreadsheet program which only has a limited amount of columns Additionally the Export feature from the File menu provides a simple way to export a complete spreadsheet of your results to a comma separated value file csv File Edit DataCenter Help Simulation New Workfile Open Workfile Save Workfile Save Workfile As Bens Model Event List Print gt Dataset Quit Return Model CARs Individual CARs ARs CAR Realized Returns CAR BHARs Matched Returns BHAR 21 The CAR Method A common method to analyze abnormal performance around a single event is the application of the cumulative abnormal return measure Event Study Metrics offers a powerful and simple implementation of this method All necessary controls are located in the CAR menu For a detailed di
3. Cluster by V Advanced gen varl expression conditions 53 Advanced Mode Creating Variables Examples gen var new varl var2 a var_new is generated as V Advanced gen var_new varl var2 the sum of varl and var2 Event List Dataset Return Model Regression Dataset var new 12 01178 11 02327 gen var new NA var new contains only NA values You are able to edit each cell in the Editor by Event List Dataset Been Model Regression Dataset double clicking the value var new N N v Advanced gen var new NA gen var new vari var2 var2 lt 10 var_new is generated as the sum of varl and var2 var2 lt 10 values for var2 Event List Dataset Return Model Regression Dataset which are gre ater than Or en equal to 10 are not 9 82774 NA considered V Advanced gen var new varl var2 var2 lt 10 54 Advanced Mode Deleting Variables Description You can delete variables using the del command Syntax del vor CAR BHAR CTIME REGRESSION Cluster by Robust Advanced del varl Example del var new var new IS deleted 55 Advanced Mode Replacing Variables Description You can replace the contents of an existing variable by using the rep command Event Study Metrics offers the possibility to apply mathematical operations within expressions An overview of the provided mathematical operations is available in section Mat
4. D 53332 Bornheim info eventstudymetrics com www eventstudymetrics com May 30 2014 Getting Started Installing Event Study Metrics cccccccccccceeseesecccescnsusecececeesssueesseecesssssssuseeeesssaeaeseeecessssaaaees 1 Registering Event Study Metrics rrrrrrrrrrnnnrnrrrrrrrrrnnnnrrrrrrrrrnnnnnrrrrrrnrrnnnsnnrrrnnnnnnennrrennnnnnnn 2 Updating Event Study Metrics cccccccccccccccseeceseeecccsecsseeceecccssessuueseeeeeesssaueeceeesssssuneeeseeeesseas 3 Data Management The Event Study Metrics WOorkfile cccccccccccssseeeeecceccescsseeueeeccessesueeseeeecesssaaeeeeessssssaesseeeses 5 EE 6 BE EEE EE 11 FOTIA N O ERE deste 16 Regression Dataset Event Study Metrics Plus 20 EDON RE 21 Analysis The CAR Method 22 TEEN 26 The Calendar Time Portfolio Method wiiiesiesvnsexessscaseievnbetortecidandaseutiedeensieciesserinisinatectsannedests 29 Event Study Metrics Plus The Regression Analysis rrrrrnrnrnrnnnrrnrnnrerrnnnrernnnnrrnrnnnrnrnanrrnrnnnrnnnnsernrnanrrnnnernnnasnrnrnnsserenn 32 Ordinary Least Squares OLS Hegoreseion 33 Eeer 38 PROV ACCC 100 SSE EEE EEE EEE EE 40 Advanced Mode OLS Regression ssscccccssseececesneeecessaueeeessaueeeesssaueeesseaaeeeesssaseeessssanes 41 Advanced Mode Logistic Hegreseion 45 Advanced Mode Summary Statistics cccccccccccccccccssseeseecccecsssseeececessssseeeeeeessssaaeeseeeseeseas 48 Advanced Mode Plotting Varables 50 Advanced Mode C
5. METRICS Version 1 06 User Manual Event Study Metrics Copyright O 2011 Event Study Metrics UG haftungsbeschr nkt This software product including program code and manual is copyrighted and all rights are reserved by Event Study Metrics UG haftungsbeschrankt Your rights to the software are governed by the accompanying software license agreement Under the copyright law this manual may not be copied in whole or in part without the prior written permission of Event Study Metrics UG haftungsbeschrankt Disclaimer Event Study Metrics UG haftungsbeschrankt assumes no responsibility for any errors that may appear in this manual or the Event Study Metrics software The user assumes all responsibilities for the selection of the program to achieve intended result and for the installation use and results obtained from the Event Study Metrics software Trademarks Event Study Metrics is a registered trademark of Event Study Metrics UG haftungsbeschrankt Windows and Excel are registered trademarks of Microsoft Corporation Other company and product names mentioned herein are trademarks of their respective companies Mention of third party products is for informational purposes only and constitutes neither an endorsement nor a recommendation Event Study Metrics UG haftungsbeschrankt assumes no responsibility with regard to the performance or use of these products Event Study Metrics UG haftungsbeschrankt Schornsberg 21
6. cccccccccccccccceseseseeceeeeeeseeeeseeeeeeesseaeseceeeeessaeaeeeeeeseaaageseess 61 D BES EEE EEE EEE EE ENN 11 E EsimaHorn VV INC 0 EE 63 Fel NN 6 7 Event Study Metrics Workile ssrisseimincsssisaniineni nnn aiai See Workfile Feide 63 ENN 21 Fama French 3 Factor Model 69 G Generalized Sign Test 80 Eg SEE EE EE 24 27 TN 7 Installing Event Study Metrics ccccccccesseeeeeeeeeeeeeeeeeeeeeeeeceeeeeeceeeeesseeeeeeeseessaeeaeeeeeeeeessaas 1 L Large Sample Mode naannnnnnnnnnennnnnnnnnennnnnnnennnnnnnrensrnnrrrrnnssnnrrrrennsnnrrrrennsnnnrennrnnnnnreennenen nnne 13 Logisti REN 45 M Markt Model EE 66 67 Market ROT EEE EEE EEE ENE 65 Matched Films P OMMONOS E 70 Minimum System Heourements 1 Multifactor dee 69 N Normal Return Model 63 O Ordinary Least Squares Hegoreseion 33 41 P PT ve See Standardized Residual Test Registering Event Study Metrics ccccccccccccceeeeeeeeeeeeeeeeeeeeeeeeeceeeeeesseeeseeeeeeseaeeeeeeeeeeeeeaaas 2 ge Me EE EE EE reenn 32 Fen Clive 60 ROI ell ME 16 S ScholesWiliams Approach WEE 66 SU EEE EE ENE A E NE EE A AE E 1 Simple Net Hetum ssori ei ERER E E OE EEEE AER ESSE 60 Skewness Adjusted KEE 81 82 83 Standardized Cross Sectional Test 77 78 Standardized Residual Test 75 T Me vr 73 TANS DO gt EEE EE EEE OE 21 U Bel gt EE 3 4 GEBE 9 EE EE ENE EEE cepa ae err aner mea eet sere ae 30 72 V Value Weighted Portfolios
7. round expression Round function abs expression Returns the absolute value of a number pos expression Transforms all values in positive values neg expression Transforms all values in negative values 59 Return Calculation Event Study Metrics automatically calculates returns if your Dataset contains prices By default Event Study Metrics calculates simple net returns P LT Riz l ef If you want to apply a short term event study based on the cumulative abnormal return measure Event Study Metrics optionally calculates continuously compounded returns log returns Tic In 1 Riz op In Pi r 1 You can specify your preferred calculation method by selecting the corresponding option in the Return Calculation Method section For further details on the properties of simple and continuously compounded return you may refer to section 1 4 of Campbell Lo and MacKinlay 1997 60 Abnormal Returns Abnormal Returns are the crucial measure to assess the impact of an event The general idea of this measure is to isolate the effect of the event from other general movements of the market The abnormal return of firm i and event date t is defined as the difference of the realized return and the expected return given the absence of the event AR E Riz a E Ri cu The expected return henceforth referred to as normal return is unconditional on the event but conditional on a separate information set Dependent on the
8. 83 Barber Brad M and John D Lyon Detecting Long Run Abnormal Stock Returns The Empirical Power And Specification of Test Statistics Journal of Financial Economics 1997 43 3 341 372 Boehmer Ekkehart Jim Musumeci and Anette B Poulson Event Study Methodology under Conditions of Event Induced Variance Journal of Financial Economics 1991 30 2 253 272 Bessembinder Hendrik Kahle Kathleen M Maxwell William F and Danielle Xu Measuring Abnormal Bond Performance Review of Financial Studies 2009 22 10 4219 4258 Blume Marshall E Betas and Their Regression Tendencies The Journal of Finance 1975 30 3 785 795 Brown Stephen J and Jerold B Warner Measuring Security Price Performance Journal of Financial Economics 1980 8 3 205 258 Brown Stephen J and Jerold B Warner Using Daily Stock Returns The Case of Event Studies Journal of Financial Economics 1985 14 1 3 31 Campell John Y Andrew W Lo and A Craig MacKinaly The econometrics of financial markets Princeton University Press 1997 Princeton New Jersey Carhart Mark M On Persistence in Mutual Fund Performance The Journal of Finance 1997 52 1 57 82 Corrado Charles J A Nonparametric Test for Abnormal Security Price Performance in Event Studies Journal of Financial Economics 1989 23 2 385 396 84 Corrado Charles J and T L Zivney The Specification and Power
9. 06 Events csv Dataset The Dataset spreadsheet contains the asset pricing data of all sample firms Event List Dataset Return Model Regression Dataset Date DE0006969603 DEOOO7164600 DEOONADHN404 DE0005165906 D gt 80 58 1901 36 96 67 58 45 BE 02 01 1997 81 81 1892 29 94 44 57 58 BE 03 01 1997 84 13 1901 36 95 58 62 8E 06 01 1997 84 13 1305 9 94 44 61 07 87 07 01 1997 83 36 1948 56 94 72 63 16 BE 08 01 1997 83 36 1937 66 95 64 38 BE 09 01 1997 85 67 19422 95 65 08 8 10 01 1997 81 5 1998 47 67 35 BE 13 01 1997 79 03 1996 66 69 44 e 14 01 1997 81 04 1969 43 69 44 ou 15 01 1997 83 36 2042 04 70 84 ge 10 rm 1007 01 01 MAA 5 7N 40 OF b The Import gt Dataset feature from the File menu provides a simple way to import a sample from a comma separated value csv file created by any spreadsheet or database program Edit Data Center Help Simulation New Workfile B Open Workfile Save Workfile Save Workfile As Event List Return Model Regression Dataset Your comma separated value csv file can contain the asset pricing data of all sample firms Each column should exhibit a time series for an asset over the full sample period You must denote missing values by or 11 Dataset The first column should contain the trading dates of the sample period You do not need to specify the treatment of missing values when importing your data but you may do so before conducting your study by select
10. 10 1997 815 1998 47 Number Format Format 1 decimal separator dot Format 2 decimal separator comma Number 1 000 00 1 000 00 CSV Settings Delimiter 14 Dataset If at least one of the comma separated value file csv preferences differs from the preferences selected in the import dialogue the software will display an error message that will provide you with the specific error description An additional explanation of the error codes is provided in appendix of this document An error message will also occur if an invalid preferences combination e g the Date Format 1 the Number Format 2 and a comma as the Delimiter is chosen in the import dialogue A detailed manual on how to set up a Dataset based on a comma separated value file csv can be found here http eventstudymetrics com wp content uploads 2014 07 ESM IMPORT Dataset CSV manual pdf A sample file can be found here http eventstudymetrics com wp content uploads 2014 06 Dataset csv 15 Return Model Depending on the type of study you plan to conduct you can choose between different options for modeling the normal i e expected return If you want to implement the market model market adjusted return the CAPM or a multifactor model you need to add the corresponding data to the Return Model spreadsheet The Import gt Return Model feature from the File menu provides a simple way to import a sample from a comma separated value fil
11. By standardizing abnormal returns before forming portfolios the standardized residuals test assigns a lower weight to abnormal returns of securities with large variances than a simple time series t test Boehmer Musumeci and Poulson 1991 show that under the absence of an event induced variance increase the standardized residuals test is well specified and has appropriate power If the variance of stock returns increases around the event date the standardized residuals test rejects the null hypothesis too often 76 Standardized Cross Sectional Test Boehmer Musumeci and Poulson 1991 combine the standardized residuals test with an empirical variance estimate based on the cross section of event window abnormal returns to construct a test that is robust to event induced variance increases of stock returns Initially abnormal returns are standardized as described in the previous section Then the cross sectional average of CSAR t T2 IS calculated N 1 CSAR T4 T2 st CSAR T1 T2 t 1 The standard deviation of CSAR t T2 is estimated from the cross section of event window abnormal returns N X CSAR 4 72 CSAR t4 T2 2 1 The standardized cross sectional test statistic for the null hypothesis that the cumulative average abnormal return is equal to zero is CSAR T1 T2 TBoehmer et al S CSAR Event Study Metrics allows you to use an adjusted version of the standardized cross sectio
12. Firms e The first line contains the column headings security identifiers e The date format should correspond to one of the formats available in the import dialogue within the Event Study Metrics software You are able to select among four most common date formats MM DD YYYY format 1 DD MM YYYY format 2 YYYYMMDD format 3 or YYYY MM DD format 4 e A valid number is either of the form 1 000 00 format 1 or 1 000 00 format 2 that can be chosen in the import dialogue within the Event Study Metrics software 13 Dataset The Preview window Source in the import dialogue shows you how your original data IS formatted This allows you to check whether your data and the chosen settings in the import dialogue are correct Once the data has been imported the Event Study Metrics software will convert all datasets into a unitary format the Date Format 1 and the Number Format 1 The Preview window Target in the import dialogue shows you how your data is ultimately formatted and used with the Event Study Metrics software DE0006969603 DE0007164600 DE0006969603 DE0007164600 80 58 1901 36 80 58 1901 36 02 01 1997 81 81 1892 29 01 02 1997 81 81 1892 29 03 01 1997 84 13 1901 36 01 03 1997 84 13 1901 36 06 01 1997 84 13 1905 9 01 06 1997 84 13 1905 9 07 01 1997 83 36 1948 56 01 07 1997 83 36 1948 56 08 01 1997 83 36 1937 66 01 08 1997 83 36 1937 66 09 01 1997 85 67 19422 01 09 1997 85 67 19422 10 01 1997 81 5 1998 47 01
13. You are able to reject your selection by picking a variable from the Independent Variables list and clicking the button 2 Select your Dependent Variable by clicking the Dependent Variable box and picking the preferred variable from the drop down list 33 Ordinary Least Squares OLS Regression Regression Variables All Variables Dependent Variable 3 By selecting Cluster by or Robust you are able to add one of the two options to your estimation These additional options are described below Standard Error VI Cluster by Robust Reports Huber White Sandwich robust standard errors Cluster by Reports clustered standard errors allowing for intragroup correlation Groups are defined by varX You can use any numeric or non numeric variable 34 Event Study Metrics Plus Ordinary Least Squares OLS Regression 4 Click the button to conduct the regression analysis 35 Ordinary Least Squares OLS Regression Example 1 BHAR CTIME REGRESSION Regression Variables All Variables Independent Variables x D 2 43 vi x4 y Dependent Variable Standard Error V Cluster by Robust by Advanced Dependent Variable y Std Err t Statistic Prob 95 Conf Intervall constant 5 5608 0 1208 46 0173 0 0000 5 7986 x 0 7158 0 0091 79 0089 0 0000 0 6980 x2 7 6649 0 1186 646094 0 0000 7 4314 Observations F 2 488 45706516 P P
14. definition of the information set e g past asset returns and the functional form there exist various models for the normal return Those models are extensively discussed in the following section Event Study Metrics offers two different measures of aggregated abnormal returns that are commonly used in event study analyses Cumulating abnormal returns across time yields the cumulative abnormal return measure T2 CAR T1 T2 H AR t T1 The second measure the buy and hold abnormal return is defined as the difference between the realized buy and hold return and the normal buy and hold return T2 T2 BHAR t1 t2 FIG Rit Ja E Rit Qie t t1 t t1 61 Abnormal Returns Statistical test of abnormal returns are commonly based on the cross average of each measure For cumulative abnormal returns the cross sectional average is N 1 CAAR t T2 gt CAR T1 T2 i 1 Whereas the mean buy and hold abnormal return is N 1 BHAR t T gt gt BHAR t4 T gt i 1 For a detailed discussion of the difference between the two measures you may consult Barber and Lyon 1997 or Ritter 1991 62 Normal Return Models A substantial feature of an event study is the choice of an appropriate normal return model Some models contain parameters that need to be estimated constant mean return model market model CAPM and multifactor models The time period over which parameters are estimated
15. excess return of past winning over past losing stocks For a detailed description of the momentum factor and the construction of the underlying portfolios refer to Carhat 1997 69 Matched Firms Portfolios Lyon Barber and Tsai 1999 propose the use of a portfolio matched by size and market to book ratio as measure of normal returns for each event They claim that this measure is free of the new listing and rebalancing bias and propose to draw statistical evidence applying a bootstrapped version of the skewness adjusted t test However the authors cast doubt whether this approach yields well specified test statistic in non random samples To apply the matched firms portfolio approach you need to specify a reference firm portfolio to each event The Event List contains additional fields for each event that allow you to match a firm portfolio Event Study Metric treats the asset pricing data of matched firms portfolios equal to the data of event firms Therefore you can simply add the asset pricing data of your matched firms to the common Dataset Event Study Metrics will calculate abnormal returns by subtracting the contemporaneous return of the individually matched firm portfolio ARir Rit up where Ryp IS the contemporaneous return of the individually matched firm portfolio 70 Bonds Matched Portfolios Bessembinder et al 2008 discuss different methods to measure abnormal bond performance Since some fir
16. is based on the ratio of positive cumulative abnormal returns p over the event window Under the null hypothesis this ratio should not systematically deviate from the ratio of positive cumulative abnormal returns over the estimation window pn Since the ratio of positive cumulative abnormal returns is a binominal random variable the follow test statistic is used tr Po Pest Pose 1 Pest N Under the null hypothesis that the average cumulative abnormal return is not statistically different from zero the test statistic approximately follows a normal distribution 80 Skewness Adjusted t Test Buy and hold abnormal returns are positively skewed e g Barber and Lyon 1997 The skewness adjusted t test originally developed by Johnson 1978 is a transformed version of the usual t test to eliminate the skewness bias The test statistic for the null hypothesis that the mean buy and hold abnormal return is equal to zero is 1 1 T skewness Adjusted VN e er where ga BHAR T1 12 and je Vi BHAR T1 72 BHAR Tt1 72 3 A 3 OBHAR NOBHAR Since VNS is the usual t statistic the estimated standard deviation is defined by N 1 EE OBHAR BHAR t1 72 BHAR 1 72 i 1 Lyon Barber and Tsai 1999 recommend the use of a bootstrapped version of the skewness adjusted t test that yields well specified test Statistics You can enable the bootstrap in the Bootstrap section of the or Setting
17. is commonly denoted as the estimation window Since the normal return is the expected return in absence of the event overlapping event and estimation windows should be avoided Otherwise normal return model parameters are estimated from returns affected by the event Event Study Metrics applies the common approach by restricting the estimation window to the time period prior to the event window estimation event post event window window IC window I Ti 0 1 gt 13 L L By choosing the option Skip near singular events on the Settings form all events are deleted automaticaly in case no regression parameters are obtained for the normal return model 63 Constant Mean Return Assume that expected asset returns can differ by company but are constant over time Then the constant mean return model Is Riz Hi t ir With EI 0 and VARIG d The parameter u is estimated by the arithmetic average of estimation window returns T S 1 Mi m H Riz i To 1 where M is the number of non missing returns over the estimation window Please note that M lt Ly Even though the constant mean return model is simple and highly restrictive compared to other models Brown and Warner 1980 1985 show that results based on this model do not systematically deviated from results based on more sophisticated models Please note that Brown and Warner 1980 1985 only analyze short term event studies The selection of the bench
18. or use the Export feature from the File menu 25 The BHAR Method A common method to analyze the long term abnormal performance around a single event is the application of the buy and hold abnormal return measure Event Study Metrics offers a powerful and simple implementation of this method All necessary controls are located at the BHAR CTIME menu For a detailed discussion of this method you may refer to Lyon Barber and Tsai 1999 and the appendix of this document CAR BHAR CTIME REGRESSION Normal Return Model Method Market Return v Buy Hold Calendar Time Portfolio Event Window Main Event Window Sub Event Window Sub amp Event Window Sub amp Event Window Sub Event Window Sub Event Window Sub amp Event Window Sub To start with your analysis you might select the Buy Hold option from the Method menu The Event Window Main defines the period over which a possible influence of the event on the asset involved should be examined Applying a typical event study it is common to take account of multiple sub periods prior during and following the event Event Study Metrics offers the simultaneous consideration of different sub periods that can be defined in Event Window Sub The start and the end of each window are defined relative to the event date event time 26 The BHAR Method To specify the start and the end of each window you can directly type i
19. 976 14 2 246 274 Ritter Jay R The long run performance of initial public offerings The Journal of Finance 1991 46 3 27 Scholes Myron and Joseph T Wiliams Estimating Betas from Nonsynchronous Data Journal of Financial Economics 1977 5 3 309 327 White Halbert A heteroscedasticity consistent covariance matrix estimator and a direct test for heteroscedasticity Econometrica 1980 48 817 838 86 PUIG Fre WU EEE EE 61 Auto Update ee EE 3 B BHAR TO EE 26 Blume Adjustment ccceeeecccceeeeeeecceeeeeecceaeeseeeecseuseceeseeaseeeesseaaeeeesseuaecesseaeeesssaaaeeeessaeaees 66 Boehmer et l T Est uuanemmssme degt See Standardized Cross Sectional Test Bonds Matched Portfolios ccccccsseseccceseseceecceeececcceuescececeeaeeeeseeaecessesaeeesseeaneeeesessages 71 9 6 EE EE 81 Buy And Hold Abnormal Return cccscccccsseccecseeececsseeeceueeeeseeeeessueeeseueeeeseneesseeeesaaees 61 C Calendar Time Portfolio Method 29 Calendar Time Portfolio Hegreseions 72 Pi 68 CAR 1000 EEE EEE EEE 22 arhar Momentum g ee JE EE 29 69 72 Constant Mean Return arunnrnrnrnnnnnnnnvnnnnnnnnennnnrnnnnnnerurennevnnnrnsnnnernsnsusnnnerusennerusnnnenunevusenusenn 64 Continuously Compounded Return cccccccseseeecceeeeeceeeeeeeeeceeeeeeeeeeeeeeesseeeeeeeeeeesaeeeeeeeees 60 Gla EE 21 25 28 31 ROO Rank EEE ED a 79 Gje pe el e 16 ENK EN 74 Cumulating Abnormal Return
20. Estimation Method menu Instead of ordinary least squares Event Study Metrics will then apply the method proposed by Scholes and Williams 1977 to account for non synchronous trading Event Study Metrics calculates the market model parameters applying the Scholes Williams approach by Bi ER d H sy ilag a Llea and 1 2PM sw 2 DEL Bisw JERN where Big Dr Bieaa are the OLS estimates from the regression of Rur 1 Ruz and Ry r 1 ON Riz and Py IS the first order autocorrelation of uc Some financial databases use adjusted betas following Blume 1975 By choosing the option Blume Adjustment on the Settings form betas are adjusted according to Blume 1975 Gi plume 0 33 0 678 66 Market Model Event Study Metrics offers a multi country version of the market model To apply this approach you need to specify a reference portfolio index to each event The Event List contains additional fields for each event that allow you to match a portfolio or index Departing from the ordinary market model you can simply add the asset pricing data of your reference portfolios indices to the Dataset 67 CAPM According the capital asset pricing model the expected excess return of asset is given by E R re i BIR r Cit where Tr is the risk free return Event Study Metrics estimates the model parameters of the capital asset pricing model by a time series regression based on realized returns Rix S
21. STER NET 06 28 1999 DE0005079909 ASCREATION 07 22 1999 Format 2 DD MM YYYY Number Format Format 1 decimal separator dot Number 1 000 00 CSV Settings Delimiter Compamy ID BAL EDOB ABWICKL KOEGEL FAHRZ ENBW ENGE BA DE0006300734 DE0005220008 DE0005151005 DE0006335037 DE0006219934 DE0005082903 DE0005079909 Format 4 YYYY MM DD Company Name BASF SE KRONES AG JUNGHEINRICH LOBSTER NET AS CREATION Format 2 decimal separator comma 1 000 00 Date 10 26 1998 12 21 1998 12 28 1998 01 13 1999 01 18 1999 06 25 1999 06 28 1999 07 22 1999 Event List 10 If at least one of the comma separated value file csv preferences differs from the preferences selected in the import dialogue the software will display an error message that will provide you with the specific error description An additional explanation of the error codes is provided in the appendix of this document An error message will also occur if an invalid preferences combination e g the Date Format 1 the Number Format 2 and a comma as the Delimiter is chosen in the import dialogue A detailed manual on how to set up an Event List based on a comma separated value file csv can be found here http eventstudymetrics com wp content uploads 2014 02 ESM IMPORT Event List CSV manual 1 06 pdf A sample file can be found here http eventstudymetrics com wp content uploads 2014
22. alue file csv preferences differs from the preferences selected in the import dialogue the software will display an error message that will provide you with the specific error description An additional explanation of the error codes is provided in appendix of this document An error message will also occur if an invalid preferences combination e g the Date Format 1 the Number Format 2 and a comma as the Delimiter is chosen in the import dialogue e A detailed manual on how to set up a Return Model based on a comma separated value file csv can be found here http eventstudymetrics com wp content uploads 2014 02 ESM IMPORT Return Model CSV manual 1 06 pdf 19 Regression Dataset Event Study Metrics Plus The Regression Dataset spreadsheet contains the observations for different variables in your dataset Event List Dataset Return Model Regression Dataset varl vard gt 1 01178 11 1 02327 1 3804 1 24767 1 13902 0 921679 1 0857 1 07408 1 16762 1 12629 0 861166 0 82774 10 13 12 10 10 10 NS SoS od e SO The Import gt Regression Dataset feature from the File menu provides a simple way to import a sample from a comma separated value csv file created by any spreadsheet or database program H H gt E New Workfile Open Workfile Save Workfile Save Workfile As Import Export Print Quit Model Event List Dataset Return Model
23. among four most common date formats MM DD YYYY format 1 DD MM YYYY format 2 YYYYMMDD format 3 or YYYY MM DD format 4 e A valid number is either of the form 1 000 00 format 1 or 1 000 00 format 2 that can be chosen in the import dialogue within the Event Study Metrics software 17 Return Model 18 The Preview window Source in the import dialogue shows you how your original data IS formatted This allows you to check whether your data and the chosen settings in the import dialogue are correct Once the data has been imported the Event Study Metrics software will convert all datasets into a unitary format the Date Format 1 and the Number Format 1 The Preview window Target in the import dialogue shows you how your data is ultimately formatted and used with the Event Study Metrics software 1997 01 02 260 91 1997 01 03 261 96 1997 01 06 263 94 1997 01 07 264 34 1997 01 08 266 15 1997 01 09 265 38 1997 01 10 268 38 1997 01 13 270 21 1997 01 14 269 87 Format 1 O Format 2 MM DD YYYY DD MM YYYY Number Format Format 1 decimal separator dot Number 1 000 00 CSV Settings Delimiter Format 3 YYYYMMDD 01 02 1997 260 91 01 03 1997 01 06 1997 01 07 1997 01 08 1997 01 09 1997 01 10 1997 01 13 1997 01 14 1997 261 96 263 94 264 34 266 15 265 38 268 38 270 21 269 87 Format 2 decimal separator comma 1 000 00 Return Model If at least one of the comma separated v
24. are defined by varX You can use any numeric or non numeric variable Estimates the model without a constant Examples logit varY varx1 varx2 varx3 Logistic Regression Dependent Variable vary Std Err z Statistic Prob 95 Conf Intervall constant 19 9612 var 1 5576 van2 4 0183 varx3 3 3343 Observations 301 Log likelih LR chi2 3 118 3988 Prob chi2 Pseudo R 0 7519 The table shows the regression results for the dependent variable varY with values 0 or 1 and the independent variables varX1 varX2 var X3 46 Advanced Mode Logistic Regression logit varY var X1 var X2 var X3 robust Logistic Regression Dependent Variable vary Robust Std Err constant var van2 var Observations Wald chi2 3 Pseudo R 19 9612 1 5576 4 0183 3 3343 301 13 8957 0 7519 Std Err z Statistic Log likelih Prob chi2 Prob 95 Conf Intervall 30 8555 0 7279 6 6796 1 5557 The table shows the results for the dependent variable varY with values 0 or 1 as well as the independent variables varX1 varX2 varX3 and reports Huber White Sandwich robust standard errors 47 Advanced Mode Summary Statistics Description The mean median standard deviation minimum maximum and the number of observations for each variable are reported by using the sum command Syntax sum vars conditions BHAR CTIME REGRESSION Cluster by Robust V A
25. chnique from the Estimation Method menu By default Event Study Metric forms equal weighted portfolios You might select the Use Weights option to form value weighted portfolios Raw Data Dataset and Return Model Price Return Use Weights Once all selections have been made you can execute the event study analysis by clicking on the Run button Depending on the sample size the execution can take up to several minutes The window might freeze while the program is in execution mode but will be accessible after all calculations have been made Results Realized Returns CT Portfolio Matched Returns Graph t Statistic White 0 0006 0 6641 0 5067 0 5606 Value t Statistic Prob 0 5196 8 4600 0 0000 5 3356 0 1235 1 8335 0 0667 2 3833 0 0192 0 2979 0 7658 0 2469 2631 0000 0 0264 The regression results intermediate calculation and sorting steps are shown in the Results section 30 The Calendar Time Portfolio Method The Print feature from the File menu provides a simple way to print a summary report containing a table of summary statistics and a graph showing buy and hold abnormal returns over the event window To use the results in any other program you can simply copy amp paste the data from any arbitrary spreadsheet or use the Export feature from the File menu 31 The Regression Analysis The Regression Section allows you to conduct cross sectional regression analyses of abnor
26. dels and estimation techniques are available in Event Study Metrics You might select a particular model from the Normal Return Model menu Where applicable you can select an appropriate regression technique from the Estimation Method menu Once all selections have been made you can execute the event study analysis by clicking on the Run button Depending on the sample size the execution can take up to several minutes The window might freeze while the program is in execution mode but will be accessible after all calculations have been made 23 The CAR Method CAAR 25 25 The summarized results intermediate calculation and sorting steps and a graph are shown in the Results section Results Realized Returns Matched Returns Model Parameter Date CAAR Pos Neg ea Prob 40 40 0 0036 204 234 0 2611 0 7940 GESA 0 0263 168 270 4 0929 0 0000 10 2 0 0167 170 268 3 6722 0 0002 GLE 0 0357 317 121 13 6029 0 0000 E 0 0351 314 124 11 5814 0 0000 2 2 0 0349 301 137 10 2978 0 0000 2 20 0 0042 223 215 0 6413 0 5213 0 0 0 0323 343 95 21 3039 0 0000 24 The CAR Method The Print feature from the File menu provides a simple way to print a summary report containing a table of summary statistics and a graph showing cumulative abnormal returns over the event window To use the results in any other program you can simply copy amp paste the data from any arbitrary spreadsheet
27. dvanced sum vars conditions 48 Advanced Mode Summary Statistics Examples sum varl var var3 var4 Summary Statistics Median Std Dev 1 27355 8 85547 5 98047 0 67425 sum varl var2 var3 10 Summary Statistics Mean Median Std Dev Min 0 94543 0 86285 0 14644 0 73737 1 16621 11 71429 1050000 1 70434 10 00000 14 00000 The summary statistics for vari and var2are reported only for the observations satisfying the condition that var3 equals 10 49 Advanced Mode Plotting Variables Description A plot visualizes the observations for a single variable or the relationship between two variables in your Dataset Event Study Metrics offers the most common types of plots By default the chart is shown as a point plot You may select your favored type by adding the options below In addition setting distinct conditions allows for subsample analyses Syntax plot vari var2 options conditions CAR BHAR CTIME REGRESSION Cluster by VI Advanced plot vart var2 options conditions 50 Advanced Mode Plotting Variables Options point In the case of plotting a single variable the chart is shown as an index plot where the values of the variable on the y axis are plotted against the corresponding observation numbers on the x axis In the case of plotting two variables the chart shows the distribution of points each having the corresponding values of one variable on th
28. e csv created by any spreadsheet or database program File Edit DataCenter Help Simulation New Workfile i gt Open Workfile kd Save Workfile Kl Save Workfile As gt Import gt Event List gt Dataset gt Return Model Regression Dataset At least your comma separated value file csv must contain the time series of your market index The time series can either consist of index values or it can consist of return data You must specify the type of data in the Raw Data section see Dataset If your model is defined by excess returns the second column must contain the corresponding interest rate You must not use annualized rates but rather rates that coincide with your individual data frequency 16 Return Model The following columns can carry up to four factors If you plan to implement a factor model all data must be defined in terms of returns instead of prices Additionally you must select the return option in the Raw Data section Any comma separated value file csv must satisfy the following requirements e Each line contains observations of a single point in time and the values are separated by a unique delimiter that can be chosen in the import dialogue within the Event Study Metrics software e The first line contains the column headings e The date format should correspond to one of the formats available in the import dialogue within the Event Study Metrics software You are able to select
29. e x axis and of the other variable on the y axis line Shows the graph where the data points are linked by lines This kind of plot is usually used to present the frequency of data on a number line Shows a chart with horizontal bars For instance a bar graph can be used to compare the values x axis of different categories y axis Shows how proportions of data shown as pie shaped pieces contribute to the data as a whole 51 Advanced Mode Plotting Variables Examples plot vari var2 Sa Sees Scatter Plot var1 values on the x axis var2 values on the y axis plot vari var2 bar vari lt 1 Plot of varl x axis and var2 y axis Ba r Ch a r t 0 903475 0 853475 var1 values on the x axis 0 803475 0 753475 vare values on the y axis 0 703475 0 653475 0 602475 varl lt 1 only values less ones than or equal to 1 for var1 0 503475 o are shown 52 Advanced Mode Creating Variables Description To create a new variable you can use the gen command The values of the new variable are specified by an expression Event Study Metrics offers the possibility to apply mathematical operations within expressions An overview of the provided mathematical operations is available in section Mathematical Expressions You can also define custom values In addition conditions can be used Syntax gen varl expression conditions BHAR CTIME REGRESSION
30. g as the ID of the Event List corresponds to the security identifiers of your Dataset You can specify a Matched Firm This allows you to apply the matched firms portfolio approach to estimate CARs and BHARs Event List You can add a Weight to each event to apply an individual weighting scheme to your event study You can match each event to a specific Group This allows you to create subsamples out of your overall sample e g if you want to analyze CARs for different groups The Import gt Event List feature from the File menu provides a simple way to import a sample from a comma separated value file csv Most spreadsheet or database programs allow you to store your data as a comma separated value file csv File Edit DataCenter Help Simulation New Workfile gt Open Workfile kd Save Workfile Kl Save Workfile As gt Dataset gt Return Model Regression Dataset Any comma separated value file csv must satisfy the following requirements e There is one event per line and the values are separated by a unique delimiter that can be chosen in the import dialogue within the Event Study Metrics software e The first line contains the column headings Event List The column order of the comma separated value file csv should be identical with the column order of Event List 1st ID 2nd company name 3rd date 4th ID matched firm 5th company name matched firm 6th Weight 7th Group Example Your eve
31. he regression results for the dependent variable varY and the independent variables varX1 and varX2 var X3 group variable specifies to which group each of the observations for varY varX1 and varX2 belongs varX3 5 determines that the estimation contains only the observations for group 5 ols varY varX1 varX2 robust OLS Regression Dependent Variable varY Std Err t Statistic Prob 95 Conf Intervall constant 0 0142 0 7057 varX1 0 0022 y 0 0891 varX2 0 0031 0 1108 Observations F 2 488 R Prob F Adj R Log likelih 655 0107 AIC SIC 2 6302 The table shows the results for the dependent variable varY as well as the independent variables varX1 and varX2 and reports Huber White Sandwich robust standard errors 44 Advanced Mode Logistic Regression Description The logistic regression estimates the likelihood of an event arising The dependent variable takes only two values e g 0 and 1 You are able to conduct the logistic regression by using the logit command The additional options are described below Syntax logit depvar indepvars options conditions CAR BHAR CTIME REGRESSION Cluster by Robust V Advanced logit depvar indepvars options conditions 45 Advanced Mode Logistic Regression Options Reports Huber White Sandwich robust standard errors cl varX Reports clustered standard errors allowing for intragroup correlation Groups
32. hematical Expressions You can also define custom values In addition conditions can be used Syntax rep varl expression conditions BHAR CTIME REGRESSION Cluster by V Advanced rep varl expression conditions 56 Advanced Mode Replacing Variables Examples rep vari var2 2 2 V Advanced rep varl var2 2 2 The values of varl are replaced by the values of var2 squared and multiplied Event List Dataset Retum Model Regression Dataset by 2 rep varl var2 2 2 vari lt 1 The values of var1 which are greater than or equal to 1 are replaced by the values Event List Retum Model Regression Dataset of var 2 squared and 1 02327 multiplied by 2 The other est values are retained 1 24767 1 13902 V Advanced rep varl var2 2 2 varl lt 1 57 Advanced Mode Renaming Variables Description Each variable may be renamed by using the rename command Syntax rename vari vari new CAR BHAR CTIME REGRESSION Cluster by Robust Advanced rename varl vart new Example rename vari var new orl is renamed var new Yy S ri e yy var 58 Advanced Mode Mathematical Expressions Arccos inverse cosine function Actan inverse tangent function max expressioni expression2 Returns the maximum of expression1 and expression2 expression2 ion exp expression Exponential function
33. ing an option from the Missing Data section Missing Data Dataset and Retum Model Stop Ignore These options have the following consequences in case necessary return data i e observations within the estimation or event window is missing choosing e Stop will abort the current estimation and an error message is displayed e Ignore will keep the asset but the missing data point s i e single or multi day return s will be ignored Estimates will be based on the remaining return data Test statistics will be adjusted accordingly You can either import a dataset based on price data or return data You must specify the type of data in the Raw Data section Raw Data Dataset and Retum Model Price C Retum Use Weights 12 Dataset If your dataset is relatively large more than 50 MB you might want to use the option large sample mode on the Settings form In this mode the dataset is not displayed in the respective spreadsheet to obtain results much faster and bind less memory resources Any comma separated value csv file must satisfy the following requirements e Each line contains observations of a single point in time and the values are separated by a unique delimiter that can be chosen in the import dialogue within the Event Study Metrics software If you want to apply the matched firms portfolios approach your dataset must contain the asset pricing data of all Matched
34. is that the cumulative average abnormal return is equal to zero Under the assumption that abnormal returns are uncorrelated and variance is constant over time each abnormal return is standardized by its estimated standard deviation AR S AR SAR 1 The standard deviation is estimated from the time series of abnormal returns of the estimation window 1 EStmax A 2 Oar d gt AR t Estmin where M is the number of non missing returns and d the degrees of freedom e g market model d 2 To account for the fact that the event window abnormal returns are an out of sample prediction the standard error is adjusted by the forecast error 1 Rmr R 2 S AR GAR 1 M se mest gt l Este Rint gt Rm Est As simple abnormal returns the standardized version can be cumulated over time T AR CSAR T T2 S AR l t T1 75 Standardized Residual Test Under the null hypothesis the distribution of SAR is a Student s t distribution with M d degrees of freedom for a further discussion see Campbell Lo and MacKinlay 1997 pp 160 It directly follows that the expected value of CSAR is zero and the standard deviation is M d S CSAR Iren 1 KI The test statistic for the null hypothesis that the cumulative average abnormal return is equal to zero is N r 1 CSAR T1 T2 i The standardized residual test is robust to heteroscedastic event window abnormal returns
35. mal returns subsequent to an event study Furthermore this section can be used as a stand alone statistic program in order to conduct an arbitrary ordinary least squares OLS or logistic regression In addition summary statistics or graphical illustrations of your data are available within this section The Regression Menu is a user interface which allows you to conduct an OLS regression and edit your data with just a few clicks CAR BHAR CTIME REGRESSION Regression Variables All Variables Independent Variables Dependent Variable Standard Error Cluster by Robust Advanced The Advanced Mode which can be activated by selecting the Advanced option allows to conduct an OLS estimation complemented by distinct conditions Furthermore the following operations are available in in this mode Logistic regression Plotting variables Summarize and Edit your data 32 Ordinary Least Squares OLS Regression Description You are able to conduct an ordinary least squares OLS regression by selecting the variables for your estimation from the Regression Menu or you may use the ols command in the Advanced Mode The Regression Menu allows a fast application of the OLS method with a few clicks 1 Select your Independent Variables by picking a variable from the All Variables list and clicking the button BHAR CTIME REGRESSION Regression Variables All Variables Independent Variables vi x1 P
36. mark models is crucial when performing a long term event study 64 Market Return Abnormal returns are calculated by subtracting the contemporaneous return of a market index Ah ES Rit Rur where Hu is the return of a market index e g S amp P 500 This model is can be viewed as a restricted market model with alpha equal to zero and beta equal to one for each stock see MacKinlay 1997 Since the parameters are predefined a separate estimation window is not necessary Thus Event Study Metrics will ignore any settings specifying the estimation window when you select Market Return as normal return model However some of the reported test statistics require an estimation window Therefore Event Study Metrics also allows you to apply the market return model with an estimation window The estimation window has no influence on the normal return measure itself but is solely used to calculate test statistics To apply this approach you need to select the Market Return Est option from the Normal Return Model menu 65 Market Model The market model is based on the assumption of a constant and linear relation between individual asset returns and the return of a market index Rir i BiRm r Eit with Els 0 and VAR Ee Og Event Study Metrics estimates the model parameters by ordinary least squares regressions based on estimation window observations Alternatively you may choose the Scholes Williams option from the
37. ms might have multiple bonds outstanding they propose to conduct a bond event study on firm level portfolios To apply this approach you need to select the Bonds Matched Portfolios option from the Normal Return Model menu Event Study Metrics will then automatically create firm level portfolios Each portfolio consists of all assets that share the same event date and company name based on the entries of the Event List Event Study Metrics allows you to utilize rating equivalent reference portfolios as measure of normal returns for each event For a detailed discussion of feasible reference portfolios you may refer to Bessembinder et al 2008 Event Study Metrics will calculate abnormal returns by subtracting the contemporaneous return of the individually matched firm portfolio ARir Rit up where Ryp IS the contemporaneous return of the individually matched firm portfolio Some of the reported test statistics require an estimation window Event Study Metrics allows you to conduct the aforementioned matching approach with an estimation window The estimation window has no influence on the normal return measure itself but is solely used to calculate test statistics To apply this approach you need to select the Bonds Matched Portfolios Est option from the Normal Return Model menu 71 Calendar Time Portfolio Regressions The Calendar Time Portfolio method allows you to assess if event firms persistently earn abnormal return
38. n these values or use the up down buttons The two most common approaches matched firms portfolios market return to estimate the normal return of a given asset are available in Event Study Metrics You may select a particular model from the Normal Return Model menu By default the cross sectional average of buy and hold abnormal returns is calculated based on an equal weighting scheme You may employ a value weighted average by checking the Use Weights option in the Raw Data section Event Study Metrics then calculates a value weighted average based on your own weights as specified in the Event List Raw Data Dataset and Return Model Price Return Use Weights Once all selections have been made you can execute the event study analysis by clicking on the Run button Depending on the sample size the execution can take up to several minutes The window might freeze while the program is in execution mode but will be accessible after all calculations have been made The summarized results intermediate calculation and sorting steps and a graph are shown in the Results section The Print feature from the File menu provides a simple way to print a summary report containing a table of summary statistics and a graph showing buy and hold abnormal returns over the event window 27 The BHAR Method To use the results in any other program you can simply copy amp paste the data from any arbitrary spreadsheet or use the Exp
39. n order to enter the Editor double click on any value within the Regression Dataset Event List Dataset Retur Model Regression Dataset 1 01178 1 02327 38 Editor 2 Removing checkmarks in the Variables list allows you to hide each variable x4 1 01178 1 02327 1 3804 3 The Y button allows you to enter distinct conditions in order to filter your data 4 The Ei button allows you to enter the Edit mode In the Edit mode you are able to copy paste delete or modify the observations within your dataset x1 x2 x3 x4 y A gt II 0 5455 4 1 01178 6 10 0 5 3 1 02327 5 5 In order to adopt the changes click the button 39 Advanced Mode To enter the Advanced Mode select the Advanced option In the Advanced Mode the output can be modified by setting distinct conditions CAR BHAR CTIME REGRESSION Cluster by Robust VI Advanced 40 Advanced Mode OLS Regression Description You are able to conduct an ordinary least squares OLS regression by selecting the variables for your estimation from the Regression Menu or you may use the ols command in the Advanced Mode Syntax ols depvar indepvars options conditions BHAR CTIME REGRESSION Cluster by Robust V Advanced ols depvar indepvars options conditions 41 Advanced Mode OLS Regression Options Reports Huber White Sandwich robust standard errors cl
40. nal test following Kolari and Pynn nen 2010 77 Standardized Cross Sectional Test The adjusted standardized cross sectional test statistic for the null hypothesis that the cumulative average abnormal return is equal to zero iS 1 p TBoenmer et al adj TBoenmer et al 1 n D where p denotes the average cross correlation among abnormal returns 78 Corrado Rank Test The non parametric rank test proposed by Corrado 1985 tests the null hypothesis that the average abnormal return is equal to zero Initially abnormal returns are transformed into ranks This is done asset by asset for the joint time period consisting of the estimation window and the event window Kit rank ARj Tied ranks are treated by the method of midranks see Corrado 1985 footnote 5 Corrado and Zivney 1992 propose a uniform transformation of ranks to adjust for missing values 1 M where M is the number of non missing returns for each asset The single day test statistic is defined as N 1 Tcorrado gt Ui SZ 0 5 S U VN i 1 The estimated standard deviation is defined as TX I LY the SC VN S U where N is the number of non missing returns cross section at t t A multiday version can be achieved by taking the average of single day Statistics multiplied by the inverse of the square root of the period s length 79 Generalized Sign Test The generalized sign test proposed by Cowan 1992
41. nt list should only contain items 1 to 3 and the Group item In that case your comma separated value file csv needs to be organized as follows 1st column ID 2nd company name 3rd date 4th to 6th column blank 7 Group The date format should correspond to one of the formats available in the import dialogue within the Event Study Metrics software You are able to select among four most common date formats MM DD YYYY format 1 DD MM YYYY format 2 YYYYMMDD format 3 or YYYY MM DD format 4 A valid number is either of the form 1 000 00 format 1 or 1 000 00 format 2 that can be chosen in the import dialogue within the Event Study Metrics software Event List The Preview window Source in the import dialogue shows you how your original data IS formatted This allows you to check whether your data and the chosen settings in the import dialogue are correct Once the data has been imported the Event Study Metrics software will convert all datasets into a unitary format the Date Format 1 and the Number Format 1 The Preview window Target in the import dialogue shows you how your data is ultimately formatted and used with the Event Study Metrics software ISIN COMPANY NAME EVENT DATE gt EDOB ABWICKL 10 26 1998 DE0006300734 KOEGEL FAHRZ 12 21 1998 DE0005220008 ENBWENGE BA 12 28 1998 DE0005151005 BASF SE 01 13 1999 DE0006335037 KRONESAG 01 18 1999 DE0006219934 JUNGHEINRICH 06 25 1999 DE0005082903 LOB
42. of the Sign Test in Event Study Hypothesis Tests Using Daily Stock Returns Journal of Financial and Quantitative Analysis 1992 465 478 Cowan Arnold R Nonparametric Event Study Tests Review of Quantitative Finance and Accounting 1992 2 Dec 343 358 Fama Eugene F and Kenneth R French Common Risk Factors in the Returns of Stocks and Bonds Journal of Financial Economics 1993 33 1 3 56 Fama Eugene F Market Efficiency long term returns and behavioral finance Journal of Financial Economics 1998 49 283 307 Hall Peter On the Removal of Skewness by Transformation Journal of the Royal Statistical Society Series B 1992 54 1 21 228 Johnson Norman J Modified t tests and confidence intervals for asymmetrical populations Journal of the American Statistical Association 1978 73 363 536 547 Kolari J W and Pynn nen S Event Study Testing with Cross sectional Correlation of Abnormal Returns Review of Financial Studies 2010 23 11 3996 4025 Lyon John D Brad M Barber and Chih Ling Tsai Improved Methods for Tests of Long Run Abnormal Stock Returns The Journal of Finance 1999 54 1 165 201 MacKinlay A Craig Event Studies in Econometrics and Finance Journal of Econometric Literature 1997 35 1 13 39 85 Patell James M Corporate Forecasts of Earnings Per Share and Stock Price Behavior Empirical Tests Journal of Accounting Research 1
43. or message is generated when the Event Window exceeds the length of the Dataset Reduce the length of the Event Window or expand the length of the Dataset This error message is generated when the date format of your raw data does not correspond to the date format of your Workfile Change the date format of your raw data or Workfile and try again Event window ends after the last entry Invalid date format 82 Error Codes N This error message is generated when the ear singular independent variables of a regression are perfectly collinear List is empty Add some events and try again specified specified Add an D to the related event and try again This error message is generated when your Return No model data Model spreadsheet is empty Import your model data and try again Nomodel This error message is generated when no Normal select d Return Model is selected Select a model from the Normal Return Model menu and try again enter a value that is not a valid date Adjust your entry This error message is generated when you click on Please select the Edit button without selecting an event to edit entry to edit Select an event from the Event List and click on Edit again This error message is generated when then date Date format format of your raw data does not correspond to the must be date format of your Workfile Change the date format dd mm yyyy of your raw data or Workfile and try again
44. ort feature from the File menu 28 The Calendar Time Portfolio Method The Calendar Time Portfolio method allows you to assess if event firms persistently earn abnormal returns The general idea is to form a portfolio of event firms and to test if this portfolio exhibits any abnormal return not captured by common risk factors Event Study Metrics offers a powerful and simple implementation of this method All necessary controls are located at the BHAR CTIME menu For a detailed discussion of this method you may refer to Lyon Barber and Tsai 1999 and the appendix of this document CAR BHAR CTIME REGRESSION Normal Return Model Method 3 Factor Model v Buy Hold Estimation Method Calendar Time Portfolio OLS v 0 Inclusion Date 20 Exclusion Date To start with your analysis select the Calendar Time Portfolio option from the Method menu The date each asset is added to the portfolio is defined by the Inclusion Date Each asset remains in the portfolio until the Exclusion Date Both dates are defined in days months as defined by the user relative to the event date event time 29 The Calendar Time Portfolio Method You may employ the Fama French three factor model or a factor model that additionally contains the Carhart 1997 momentum factor to analyze returns of calendar time portfolios from the Normal Return Model menu Furthermore you can select an appropriate regression te
45. py of your Event Study Metrics Workfile on your hard disk drive Alternatively you can select the Save Workfile Save Workfile As feature from the File menu Event List The Event List is the single tool to manage the sample of your research project independent of the type of study you may plan to conduct Ata minimum you must enter an asset identifier ID a company name and an event date Event List Dataset Retum Model Regression Dataset Event Data ID ID MATCHED FIRM OPTIONAL WEIGHT OPTIONAL NAME MATCHED FIRM OPTIONAL GROUP OPTIONAL ID Company Name Date 2014 DE0005692107 EDOB ABWICKLUNGS AG 26101 TORT DE0006300734 KOEGEL FAHRZEUGWERK 21 121 DE0005220008 ENBW ENGE BADENWUR 28121 i Mi Do Fr Sa DE0005151005 BASF SE 13 01 1 28 29 30 1 2 3 DE0006335037 KRONES AG 18 01 1 7 8 9 DE0006219934 JUNGHEINRICH AG 25 061 DE0005082903 LOBSTER NET STORAGE AG 28 06 1 28 2 30 a DE0005079909 AS CREATION TAPETEN AG 22 07 1 gt 3 4 5 6 DE0007249104 SOFTM SFTW amp BERATUN 27 071 m t Delete Edit The Date is the event date on which the event study is centered Each Date should be a trading day that occurs in your Dataset Otherwise Event Study Metrics will display an error message You may use the Event Date Check function from the Edit menu to check the validity of your event dates The ID is used to match a unique security to each event You are free to choose any format e g ISIN CUSIP as lon
46. r the null hypothesis the cumulative average abnormal return is equal to zero The statistic follows asymptotically the normal distribution The variance estimator of this statistic is based on the time series of abnormal returns from the estimation window 2 1 EStmax 1 EStmax A2 OAAR md H ar ay H AAR t Estmin EStmin where M is the number of non missing returns and d the degrees of freedom e g market model d 2 To account for the fact that the event window abnormal returns are an out of sample prediction the standard error is adjusted by the forecast error In the market model the adjustment is 1 1 Rint _ Rmest M EStmax gt 2 Dr Rm WW Ra ger 73 Cross Sectional t Test The cross sectional t test is defined as CAAR t1 T gt OCAAR t t CTOSS Under the null hypothesis the cumulative average abnormal return is equal to zero The variance estimator of this statistic is based on the cross section of abnormal returns N 1 GCAAR t 1 gt Naa ZAR tp AAR t2 j Brown and Warner 1980 show that the cross sectional t test is robust to an event induced variance increase However Boehmer Musumeci and Poulson 1991 provide evidence that their standardized cross sectional test requiring an estimation window exhibits a comparable size but is more powerful 74 Standardized Residual Test The standardized residual test developed by Patell 1976 tests the null hypothes
47. reating Variables 0000nnnan0aaannnnnannnnnnannnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnrnnnnnn 53 Advanced Mode Deleting Variables 1cccccccsscccccsssseeeecssseseeesansseeessauseeesssaaeeeessenseesssanes 55 Advanced Mode Replacing Varatfies 56 Advanced Mode Renaming Variables cccccscccccssseeeccscaneeeecssaneeeeecssasseeessasaeseessneeeeessaaes 58 Advanced Mode Mathematical Expression 59 Appendix FN Eer EE 60 Vee 61 NTN 63 Constant Mean EEN 64 VNR 65 VENN 66 Pl 68 M CO MOO CIS EEE EE ENER 69 Matched Rettel EE 70 Bonds Matched POMONOS seess E ATE 71 Calendar Time Portfolio Hegoreseions 72 Time Series EEE 73 Crosse Sectional EL EEE EEE EG ED 74 Slandardized Residual EE EEE EE 75 Standardized Cross Sectional Test 77 COM GOO Fank OSE EEE 79 Generalized SION ToS EE 80 Gkeuwness Adustedt Test 81 10 ee SEE EEE 82 References EE 78 Index Installing Event Study Metrics Event Study Metrics is either distributed on a single DVD ROM or as a download version Before you start the installation process close all other applications To install Event Study Metrics from a DVD ROM simply insert the disk into the drive If the setup does not begin automatically you will need to run Setup exe from the disk s root directory For either method you might need administrator privileges The installation of the software is straightforward You are advised to read the terms of the license before proceeding with
48. rob F 0 0000 Adj P Log likelih 372 1250 AIC SIC 1 5536 The table shows the regression results for the dependent variable Y and the independent variables X1 and X2 36 Ordinary Least Squares OLS Regression Example 2 BHAR CTIME REGRESSION Regression Yariables All Variables Independent Variables x1 e 4 L Dependent Variable Standard Error 7 Cluster by E Robust x3 Advanced Dependent Variable y Clustered Std Err by x3 Std Er tStatistic Prob 95 Conf Intervall constant 5 5608 0 0237 234 8766 0 0000 5 6074 xi 0 7158 0 0031 2345041 0 0000 0 7098 x2 7 6649 0 0548 139 9857 0 0000 7 5571 Observations F 2 488 4570 6516 P Prob F 0 0000 Adj P2 Log likelih 372 1250 AIC SIC 1 5536 The table shows the regression results for the dependent variable Y and the independent variables X1 and X2 X3 grouping variable assigns each observation to a specific cluster 37 Editor Description The Editor allows you to browse and to edit your data with a few clicks In the Edit mode you are able to copy cut and modify each observation in your dataset In addition the Editor allows you to delete variables and provides a variable filter x4 1 01178 1 02327 11 3804 1 24767 1 13902 0 921679 1 0857 1 07408 1 16762 1 12629 0 861166 0 82774 9949488 0 910678 0 865334 ninan oa noo oO ns wo Using the Editor 1 I
49. s The general idea is to form a portfolio of event firms and to test if this portfolio exhibits any abnormal return not captured by common risk factors Suppose you want to asses abnormal returns over a 3 year period For each month in calendar time the portfolio is constructed by all firms that had an event in the three year prior to the calendar month By default Event Study Metrics forms equal weighted portfolio You may select the Use Weights option to form value weighted portfolios You can employ the Fama French three factor model or a factor model that additionally contains the Carhart 1997 momentum factor to analyze returns of calendar time portfolios Ric Tfr A Bim Ruz Tfr BismeSMB biym HM Ei Under the null hypothesis of no abnormal return the estimate of a should be not statistically different from zero Lyon Barber and Tsai 1999 suggest the error term in calendar time portfolio regressions may be heteroscedastic since the number of securities varies over time They propose to employ a weighted least Squares regression where the weighting factor is based on the number of assets in the portfolio You may follow their proposal by selecting the WLS option from the Estimation Method menu Additionally Event Study Metrics reports t statistics based on White 1980 robust standard errors 72 Time Series t Test The time series t test is defined as CAAR Ttime 1 T2 T 1 2 Chare Unde
50. s Tr Ur T Pi Ruz Tfr Cit with Els 0 and VAR Eiz Ge Please make sure that the time series of risk free returns is not annualized but instead matches your data frequency 68 Multifactor Models Event Study Metrics offers the possibility to apply a multifactor model to measure normal returns You can choose a multifactor model based on three or four factors The best known approach is the three factor model developed by Fama and French 1993 Based on their empirical findings they add two additional factors to the CAPM that should increase explanatory power of the model Ric Tfr a amp i Bim Ruz Tfr BismeSMB biym HM Ei where SMB stands for small minus big and HML stands for high minus low The SMB factor should capture the excess return of small over big stocks measured by market cap The HML factor should capture the excess return of stock with a high market to book ratio over stocks with a low market to book ratio For a detailed description of the factors and the construction of the underlying portfolios you might refer to Fama and French 1993 You may obtain time series data from Kenneth French s website A common extension of the three factor model is the four factor model that additionally contains the momentum factor MOM as introduced by Carhat 1997 Ric Tfr A Bim Ruz Tfr BismeSMB Bj nm HM Pi nm UMD Ei where MOM is a factor that should capture the
51. s form Additionally you can specify the number and size of resamples Event Study Metrics will report bootstrapped p values and critical values for the 5 significance level two sided 81 Error Codes This error message is generated when Event Study Metrics can t create a new file Ensure that enough disk space is available and you have the permission to write and create files in the desired direction This error message is generated when the folder or File used by files within the folder are locked because they are another being used by Windows or another program running in program Windows Close any program that is using the file and try again Bani ds This error message is generated when an ID of the TEE Event List is not contained in the Dataset Delete the l event or add the related asset to your Dataset Can t find date This error message is generated when the event date ndice is not contained in the Dataset Delete the event or l shift the event date to the next trading date This error message is generated when a necessary Can t create file Can t find value datapoint is not contained in the Dataset Delete the event or select the Ignore option from the Missing Value section Estimation This error message is generated when the Estimation window starts Window exceeds the length of the Dataset Reduce before the first the length of the Estimation Window or expand the entry length of the Dataset This err
52. scussion of this method you may refer to Campbell Lo and MacKinlay 1997 and the appendix of this document CAR BHAR CTIME REGRESSION Normal Return Model Return Calculation Option Market Model a Log Return Estimation Method Simple Return OLS gt Estimation Window Event Window Main amp Event Window Sub Event Window Sub amp Event Window Sub Event Window Sub Event Window Sub Event Window Sub Event Window Sub The Event Window Main defines the period over which a possible influence of the event on the asset involved should be examined When applying a typical event study it is common to take account of multiple sub periods prior during and following the event Event Study Metrics offers the simultaneous consideration of different sub periods that can be defined in Event Window Sub The Estimation Window defines the period used to estimate the Normal Return Model s parameters where applicable 22 The CAR Method The start and the end of each time window are defined relative to the event date event time To specify the start and the end of each window you can directly type in these values or use the up down buttons Event Study Metrics prohibits any overlap of the Estimation Window and the Event Window Numerous approaches to estimate the normal return of a given asset have evolved since researchers have started conducting event studies The most common mo
53. the installation Thereafter you can specify the directory you wish to install your copy of Event Study Metrics By default Event Study Metrics will install to program files event study metrics Once the installation process is complete you can launch Event Study Metrics with a double click on the Event Study Metrics icon on the start menu 2 GHz processor dual core or more recommended Minimum 1024 MB RAM 50 MB free hard drive space System Windows XP SP2 or above Windows Vista Windows 7 8 Requirements NET Framework 3 5 SP1 or higher 1280 x 800 pixel display Registering Event Study Metrics The first time you run Event Study Metrics on your computer you must register the program using the serial number printed on your license form The registration is a one time process of assigning a serial number to a specific computer and validating the license If the copy of Event Study Metrics is not registered the program will automatically display a registration form Company Serial Number 00C025 22DA84 CX67AC Computer ID sum mai J oa You must fill in your name and the serial number giving a company name is optional If you are connected to the Internet you can automatically validate your license by clicking on Submit The option to register your product manually is recommended if you do not have an internet connection You can display the information for manual registration by clicking on Mail Upda
54. ting Event Study Metrics Event Study Metrics provides an automatic update feature that checks for new updates Whenever a new update is available the program will display a message and offer the opportunity to install the latest version You can enable disable the automatic updates feature in the Auto Update Function section of the Settings form Settings Auto Update Function Enable automatic updates Bootstrap _ Enable bootstraping ples 100 193746 gt Test Statistics _ Kolari Pynnonen Cross Correlation Adjustment Estimation Settings _ Blume Adjustment _ Skip near singular events Automatic event date check Import Settings _ Large sample mode Export Settings Transpose You can either select the Settings Form feature from the Edit menu or you can simply click on the Settings Form icon Updating Event Study Metrics You may manually check for updates by selecting Update from the Help menu The Event Study Metrics Workfile The Event Study Metrics Workfile offers you the possibility to store and load the raw data settings and results of your analysis into a single file To create and setup up a new Event Study Metrics Workfile you can either select the New Workfile feature from the File menu or you can simply click on the New Workfile icon Edit Data Center Save Workfile Save Workfile As Import Export Print Quit Click on the tal Save icon to save a co
55. varX Reports clustered standard errors allowing for intragroup correlation Groups are defined by varX You can use any numeric or non numeric variable Estimates the model without a constant Examples ols varY varX1 varX2 OLS Regression Dependent Variable varY Std Err tStatistic Prob 95 Conf Intervall constant 0 7156 varX1 y 0 0879 varX2 0 1133 Observations F 2 488 2136 0799 R Prob F 0 0000 Adj R Log likelih 655 0107 AIC SIC 2 6302 The table shows the regression results for the dependent variable varY and the independent variables varX1 and varX2 42 Advanced Mode OLS Regression ols varY varX1 varX2 cl varx3 Dependent Variable varY lustered Std Err by varX3 Coef Std Err t Statistic Prob 95 Conf Intervall constant varX1 varX2 Observations F 2 488 R Prob F Adj R Log likelih AIC SIC The table shows the regression results for the dependent variable varY and the independent variables var X1 and varX2 var X3 cluster variable assigns each observation to a specific cluster ols varY varX1 varX2 cl varX3 varX3 5 OLS Regression Dependent Variable varY lustered Std Err by varX3 Coef Std Err t Statistic Prob 95 Conf Intervall constant 22 4655 varX1 27 9246 varX2 19 1070 Observations F 2 50 407 7310 R Prob F 0 0000 Adj R Log likelih 99 9664 AIC SIC 3 5476 43 Advanced Mode OLS Regression The table shows t
Download Pdf Manuals
Related Search