I am currently conducting an event study and I want to use SCARs for my univariate regression, however the coefficient I get when running the regression are rather large. I think something goes wrong with calculating the standard deviation (si) to calculate standardized abnormal returns (SAR), because the si seems rather small. The si I get when running the code below is: .0012519. Can someone take a look and see what is wrong with my code?
The data I am using is as follows:
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input int eventtime float(firmid NR_MMM AR_MMM) -171 1 -.02592428 -.0011417216 -170 1 .001208978 -.0021109781 -169 1 .01920737 .00023763232 -168 1 .0019564312 -8.431195e-06 -167 1 .014404905 .000799095 -166 1 .018414522 .0003734784 -165 1 -.0036540066 .00057500665 -164 1 .008071223 .0017457767 -163 1 .008860034 -.0003260339 -162 1 .000696554 .000740446 -161 1 .0012523525 -.0014783525 -160 1 -.006919197 -.0008798032 -159 1 -.00479183 .0004938302 -158 1 .01494759 -.00022159042 -157 1 .004575039 -.0011770388 -156 1 .0002386002 -.0009146002 -155 1 -.001468135 .0002941349 -154 1 -.005910488 -.0021195116 -153 1 .016738048 -.002780048 -152 1 .011091297 -.0015792975 -151 1 -.0019644196 -.0008685804 -150 1 -.002611002 -.001872998 -149 1 .011908351 .0017536485 -148 1 -.00048867875 -.0017213213 -147 1 -.004886649 .0008736486 -146 1 .011942647 .0005193526 -145 1 .0027169974 .0003920026 -144 1 -.003610632 .00021963214 -143 1 -.001176618 -.00022138184 -142 1 -.0003857905 6.790481e-06 -141 1 -.009943306 .0003263055 -140 1 .001489399 -.000199399 -139 1 -.0032908716 -.0008451284 -138 1 -.014146594 -.0003704057 -137 1 -.011339358 .001554358 -136 1 .014998026 -.0012430262 -135 1 -.0013854208 .000017420794 -134 1 .01626799 -.0009789907 -133 1 -.0011675397 .00017253975 -132 1 .0038084204 -.0007964204 -131 1 -.001280515 .000604515 -130 1 .0011978822 .0011741178 -129 1 -.0001648833 .0013298832 -128 1 .00056441315 .00021558686 -127 1 -.004716177 .000860177 -126 1 .010739258 -.0012152576 -125 1 -.01112652 .000397521 -124 1 -.01453394 .00020393883 -123 1 .020669995 -.004214995 -122 1 -.007085634 -.002709366 -121 1 -.00658128 .0004672795 -120 1 -.007841157 .001241157 -119 1 .0022358433 -.00018584334 -118 1 -.01962691 -.00022409014 -117 1 .004762659 -.003165659 -116 1 .010676718 .0006362817 -115 1 .014606647 .0002993528 -114 1 -.015206748 .0005917477 -113 1 -.017986748 .0029187484 -112 1 .007810976 -.0006589764 -111 1 .008859024 -.000020024383 -110 1 .012491385 .0011936155 -109 1 -.001647685 .0008166851 -108 1 -.0022327362 -.0010952638 -107 1 .010687814 -.0006668141 -106 1 -.007314611 -.00006138924 -105 1 -.009528726 .0013267263 -104 1 -.015472038 .0005720377 -103 1 .001994762 -.0005977621 -102 1 -.013263974 -.00055802567 -101 1 -.023941156 -.00001884448 -100 1 -.010967145 .00004214474 -99 1 .0008256687 -.002541669 -98 1 .007836194 -.001758194 -97 1 -.025217174 -.0006418264 -96 1 .016806642 -.0013456416 -95 1 -.02182186 .00012485836 -94 1 .0005018732 -.002919873 -93 1 -.0118306 .001221599 -92 1 .005205482 .0004795182 -91 1 .02042589 .0020711122 -90 1 -.015808947 -.0008870526 -89 1 .014232416 .0017055843 -88 1 -.010992362 .00043836216 -87 1 .00554239 -.0004623905 -86 1 .024940867 -.000371867 -85 1 -.0004816178 .00012261781 -84 1 -.018940987 -.0010710129 -83 1 .005000714 .0007582858 -82 1 .0015055384 .0021454617 -81 1 -.018676706 -.0009412944 -80 1 -.014312023 -.0026449766 -79 1 -.0007045424 -.0023594575 -78 1 -.0002254058 .0004864058 -77 1 -.012442886 .0004638859 -76 1 .019653216 .000492784 -75 1 .016626082 .0010589176 -74 1 .01658876 .0015292395 -73 1 -.004741395 .0005343949 -72 1 -.00006098628 -.0001660137 end
Code:
quietly describe bys eventtime: gen N = _N preserve drop if eventtime>-11 collapse (sd) AR_MMM,by(firmid) rename AR_MMM si keep firmid si save "M:\tmp" restore merge m:1 firmid using "M:\tmp" erase "M:\tmp.dta" drop if eventtime <-2 drop if eventtime >2 gen SAR = AR_MMM/si sort firmid eventtime by firmid: gen CSAR_MMM = sum(SAR) collapse CSAR_MMM, by(eventtime) reg CSAR_MMM
Array
Thank you in advance
0 Response to Standard deviation of SCAR seems to be wrong (event study)
Post a Comment