Eventstudy using either the estudy or eventstudy2 command

[Christian Balle]

Hi,

I am currently trying to conduct an event study for my master thesis. However, I do have some trouble with the two commands listed in the title of this thread.

estudy:
For some reason I can't get the grouped CAAR to make sense. From the data shown in the table I would assume that the grouped CAAR would be -4.61% and not -4.35% as shown in the picture below. This is the issue with all the grouped CAARs. Does anyone know why this happens?
Furthermore, does anyone know if it possible to get the t-scores in the output instead of the command only showing p-values?

eventstudy2:
With this command I simply can't tell if I am doing something wrong. I am using the following command. However, I don't know if the command means that I am using the market model. I assume it does but I am not sure. For input I use discrete returns as I understood from the help-file that the code converts the returns to ln by itself. The output from eventstudy2 differs substantially from the output from the estudy command. If comparing the outputs in the event window [0,5] the CAAR from the estudy command is -27.28% while it for the eventstudy2 command is -16.36%. This just doesn't make much sense. I have tried using the exact same estimation windows, exact same event windows and so on. Hopefully someone can tell me what I am doing wrong here - maybe I misunderstood something in the eventstudy2 code as it is quite complicated to say the least.

eventstudy2 security_id date using price_file, ret(security_return) mod(MA) marketfile(market_file) mar(return_mkt) idmar(mkt_ref) car1LB(0) car1UB(5) car2LB(-5) car2UB(-5) car3LB(0) car3UB(10) car4LB(-10) car4UB(10) car5LB(0) car5UB(15) car6LB(-15) car6UB(15) car7LB(0) car7UB(0) replace

I look forward to hearing from you guys. I appreciate your time

[Jared Greathouse]

Read the FAQ please on how to ask questions. Using Statalist is just like making wild mushroom risotto. If I asked you to make my favorite version of risotto, you'd likely need to know what the SPECIFIC ingredients are and the steps you do to make said dish. So far you've given improperly formatted code and no dataset to work work- you're asking us for help to make a dish, and you've not given the recipe (the code) or ingredients (the data).

Precisely, you must give your dataset using dataex and your code that you've tried (formatted accordingly). No dataset, no code=no real assistance.

I would assume that the grouped CAAR would be -4.61% and not -4.35% as shown in the picture below.

why?


You need to read the help files and articles describing these commands. They likely have different assumptions and do certain things under the hood which would explain these differences

[Max Kerk]

Hello Christian,

Saw you responding on my post, I'm having the same problem. I recalculated the AR and CARs, these seem to be correct. However, I think there is something wrong in the average calculation.
For the T-scores, you can add the command: "outputfile(filename)" in the estudy line. In the Excel output there is a tab with all T-scores

Edit: MA is is market adjusted model, FM with one market index is the MM I guess.

[Christian Balle]

This is my eventdate file - all events have the same date as I am examining the impact of WHO's announcement that COVID-19 officially became a pandemic. In this file I have the eventdate and a security ID:

* Example generated by -dataex-. For more info, type help dataex
clear
input float(date sec_id)
21986  1
21986  2
21986  3
21986  4
21986  5
21986  6
21986  7
21986  8
21986  9
21986 10
21986 11
21986 12
21986 13
21986 14
21986 15
21986 16
21986 17
21986 18
21986 19
21986 20
21986 21
21986 22
21986 23
21986 24
21986 25
21986 26
21986 27
21986 28
21986 29
21986 30
21986 31
21986 32
21986 33
21986 34
21986 35
21986 36
21986 37
21986 38
21986 39
21986 40
21986 41
21986 42
21986 43
21986 44
21986 45
end
format %d date

Secondly I have my market-file - This file contains dates, price, return, log return and market reference:

* Example generated by -dataex-. For more info, type help dataex
clear
input float date double return_mkt byte mkt_ref double price float log_mkt
21686                     . 1 2116.505             .
21689   -.00560405007311598 1 2104.644   -.005619812
21690    .00618489397731886 1 2117.661    .006165846
21691   -.00277523172972442 1 2111.784    -.00277909
21692    -.0115196440545057 1 2087.457    -.01158651
21693    .00362546390177149 1 2095.025   .0036189076
21696   .000670159067314231 1 2096.429   .0006699346
21697   -.00565580804310567 1 2084.572   -.005671863
21698   -.00948492064558093 1   2064.8   -.009530189
21699   .000814606741572931 1 2066.482   .0008142751
21700   -.00979248790940346 1 2046.246    -.00984075
21703  -.000577643157274353 1 2045.064 -.00057781005
21704     .0159388654829383 1  2077.66    .015813176
21705    .00874637813694253 1 2095.832     .00870835
21706    .00398409796205035 1 2104.182   .0039761825
21707     .0107880401980438 1 2126.882    .010730265
21710     .0040547618532668 1 2135.506   .0040465635
21711    .00198407309555678 1 2139.743   .0019821075
21712    -.0024881492777404 1 2134.419    -.00249125
21713     .0011914249264086 1 2136.962   .0011907157
21714   -.00281240377695053 1 2130.952   -.002816366
21717 -.0000417653705951127 1 2130.863 -.00004176624
21718    .00940886392039281 1 2150.912    .009364876
21719    .00496673039157343 1 2161.595    .004954437
21720     .0100675658483667 1 2183.357    .010017226
21721   -.00238165357291539 1 2178.157   -.002384494
21724             4.130e+10 1 2178.166  4.131925e-06
21725   -.00662805314195537 1 2163.729   -.006650116
21726   -.00245825609399327 1  2158.41  -.0024612825
21727    .00373978993796365 1 2166.482    .003732814
21728    .00547662062274242 1 2178.347    .005461678
21731    .00637685364177499 1 2192.238    .006356608
21732    .00287788096000534 1 2198.547    .002873748
21733    .00673126387564149 1 2213.346     .00670871
21734   .000823188060068384 1 2215.168   .0008228495
21735   -.00433375707846985 1 2205.568   -.004343175
21738   -.00496561429980859 1 2194.616   -.004977984
21739  -.000442446423428975 1 2193.645  -.0004425443
21740    .00341577602574707 1 2201.138   .0034099554
21741    .00191173838260031 1 2205.346   .0019099134
21742    .00330968473881195 1 2212.645     .00330422
21745   .000766051490410872 1  2214.34   .0007657582
21746   -.00282341465176981 1 2208.088   -.002827408
21747   -.00459764284756769 1 2197.936  -.0046082446
21748   .000210652175495437 1 2198.399     .00021063
21749   -.00172261723190368 1 2194.612  -.0017241027
21752    .00139432391693837 1 2197.672   .0013933528
21753    .00542164617831959 1 2209.587    .005407002
21754    .00369480812477636 1 2217.751    .003687999
21755   -.00488287458781449 1 2206.922   -.004894835
21756    .00398111034282139 1 2215.708   .0039732065
21759   -.00144603891848573 1 2212.504  -.0014470854
21760   -.00411389086754187 1 2203.402   -.004122376
21761   -.00718797568487284 1 2187.564   -.007213934
21762   -.00657672186962288 1 2173.177   -.006598444
21763    -.0106682520567815 1 2149.993   -.010725566
21766    -.0244996146499082 1 2097.319   -.024804724
21767    .00566532797347474 1 2109.201     .00564934
21768    .00230039716461345 1 2114.053   .0022977553
21769     .0155407645882105 1 2146.907    .015421243
21770   -.00538635348433831 1 2135.343   -.005400912
21773   -.00828719320502604 1 2117.647   -.008321723
21774    .00797583355488421 1 2134.537    .007944195
21775    -.0226166142821604 1 2086.261    -.02287629
21776   -.00164073430889029 1 2082.838  -.0016420818
21777     .0123216495954077 1 2108.502    .012246356
21780     .0110542935221309 1  2131.81    .010993642
21781   -.00573972352132689 1 2119.574   -.005756259
21782    .00752556881713024 1 2135.525    .007497393
21783   -.00180189883049833 1 2131.677  -.0018035242
21784    -.0173778672847717 1 2094.633   -.017530635
21787    .00505291380399349 1 2105.217     .00504019
21788  -.000176703874232584 1 2104.845  -.0001767195
21789    .00335986735365317 1 2111.917   .0033542356
21790     .0100193331461418 1 2133.077    .009969472
21791    .00255218166057763 1 2138.521   .0025489305
21794  -.000906701407187634 1 2136.582  -.0009071127
21795   -.00451983588741256 1 2126.925   -.004530081
21796     .0107535526640572 1 2149.797    .010696145
21797     .0113820049055795 1 2174.266    .011317717
21798    .00212991418713251 1 2178.897    .002127649
21801  -.000225802321082518 1 2178.405 -.00022582783
21802   .000468232491203418 1 2179.425   .0004681229
21803    .00688209045963949 1 2194.424    .006858517
21804    .00336762631105013 1 2201.814   .0033619686
21805     .0018144130248969 1 2205.809    .001812769
21808   -.00450129634977466 1  2195.88   -.004511458
21809    .00268092974115148 1 2201.767   .0026773424
21810  -.000337002053350665 1 2201.025 -.00033705885
21811    .00165740961597423 1 2204.673   .0016560376
21812   -.00323267895057451 1 2197.546  -.0032379155
21815   -.00173193189130059 1  2193.74  -.0017334335
21816   -.00531877068385482 1 2182.072   -.005332966
21817    .00114111725002648 1 2184.562   .0011404667
21818  -.000506279977404975 1 2183.456 -.00050640816
21819   -.00362452918675724 1 2175.542   -.003631114
21822    .00205741833529305 1 2180.018   .0020553048
21823   -.00988340463243873 1 2158.472    -.00993257
21824    -.0171426824160797 1  2121.47   -.017291319
21825    .00466044770842867 1 2131.357   .0046496214
end
format %d date

Lastly I have my return-file - In this file I have the date, return of x security, security id, market reference, price and log returns.

21686        0 1 1 7.12         0
21689 -.002809 1 1  7.1 -.0028129
21690  .022535 1 1 7.26  .0222851
21691 -.002755 1 1 7.24 -.0027587
21692 -.009669 1 1 7.17 -.0097155
21693  .009763 1 1 7.24  .0097155
21696  .041436 1 1 7.54   .040601
21697  .022546 1 1 7.71   .022296
21698        0 1 1 7.71         0
21699   .01297 1 1 7.81  .0128868
21700        0 1 1 7.81         0
21703 -.053777 1 1 7.39 -.0552772
21704 -.005413 1 1 7.35 -.0054274
21705 -.008163 1 1 7.29 -.0081968
21706        0 1 1 7.29         0
21707  .006859 1 1 7.34  .0068353
21710  .020436 1 1 7.49  .0202299
21711 -.008011 1 1 7.43 -.0080429
21712 -.008075 1 1 7.37 -.0081081
21713  .005427 1 1 7.41  .0054127
21714   .00135 1 1 7.42  .0013486
21717        0 1 1 7.42         0
21718 -.010782 1 1 7.34 -.0108402
21719  .001362 1 1 7.35  .0013614
21720  .004082 1 1 7.38  .0040734
21721  .017615 1 1 7.51  .0174618
21724 -.014647 1 1  7.4 -.0147555
21725 -.006757 1 1 7.35 -.0067797
21726   .02449 1 1 7.53  .0241948
21727 -.017264 1 1  7.4 -.0174151
21728        0 1 1  7.4         0
21731  .001351 1 1 7.41  .0013504
21732        0 1 1 7.41         0
21733  .021592 1 1 7.57  .0213627
21734  .002642 1 1 7.59  .0026385
21735  .002635 1 1 7.61  .0026316
21738 -.013141 1 1 7.51 -.0132277
21739 -.021305 1 1 7.35 -.0215352
21740  .004082 1 1 7.38  .0040734
21741   .01355 1 1 7.48  .0134591
21742  .001337 1 1 7.49   .001336
21745 -.012016 1 1  7.4 -.0120888
21746  .001351 1 1 7.41  .0013504
21747        0 1 1 7.41         0
21748        0 1 1 7.41         0
21749 -.006748 1 1 7.36 -.0067705
21752 -.001359 1 1 7.35 -.0013597
21753  .020408 1 1  7.5  .0202027
21754  .001333 1 1 7.51  .0013325
21755 -.002663 1 1 7.49 -.0026667
21756 -.010681 1 1 7.41 -.0107383
21759        0 1 1 7.41         0
21760 -.005398 1 1 7.37 -.0054127
21761  .005427 1 1 7.41  .0054127
21762  .002699 1 1 7.43  .0026954
21763 -.001346 1 1 7.42 -.0013468
21766 -.013477 1 1 7.32 -.0135687
21767  .006831 1 1 7.37  .0068073
21768 -.006784 1 1 7.32 -.0068073
21769        0 1 1 7.32         0
21770 -.005464 1 1 7.28 -.0054795
21773        0 1 1 7.28         0
21774 -.034341 1 1 7.03 -.0349442
21775 -.029872 1 1 6.82 -.0303272
21776        0 1 1 6.82         0
21777  .024927 1 1 6.99   .024621
21780  .052933 1 1 7.36  .0515794
21781 -.001359 1 1 7.35 -.0013597
21782 -.010884 1 1 7.27  -.010944
21783  .019257 1 1 7.41  .0190741
21784 -.020243 1 1 7.26 -.0204506
21787 -.012397 1 1 7.17 -.0124742
21788  .012552 1 1 7.26  .0124742
21789 -.016529 1 1 7.14 -.0166671
21790 -.004202 1 1 7.11 -.0042105
21791  .001406 1 1 7.12  .0014054
21794 -.005618 1 1 7.08 -.0056338
21795 -.016949 1 1 6.96 -.0170944
21796  .027299 1 1 7.15  .0269329
21797  .002797 1 1 7.17  .0027933
21798 -.006974 1 1 7.12  -.006998
21801 -.011236 1 1 7.04 -.0112995
21802        0 1 1 7.04         0
21803  .024148 1 1 7.21  .0238608
21804   -.0319 1 1 6.98   -.03242
21805  .034384 1 1 7.22   .033806
21808   .00277 1 1 7.24  .0027663
21809 -.015193 1 1 7.13 -.0153099
21810  .033661 1 1 7.37  .0331064
21811  .001357 1 1 7.38   .001356
21812   .00813 1 1 7.44  .0080972
21815        0 1 1 7.44         0
21816 -.024194 1 1 7.26  -.024491
21817  .001377 1 1 7.27  .0013764
21818  .004127 1 1  7.3  .0041181
21819   .00411 1 1 7.33  .0041011
21822  .002729 1 1 7.35  .0027248
21823  .010884 1 1 7.43  .0108255
21824  -.02288 1 1 7.26  -.023146
21825  .001377 1 1 7.27  .0013764
end
format %d date

Finally the code I am using:

eventstudy2 sec_id date using price_file, ret(sec_ret) mod(FM) marketfile(market_file) mar(return_mkt) idmar(mkt_ref) car1LB(0) car1UB(5) car2LB(-5) car2UB(-5) car3LB(0) car3UB(10) car4LB(-10) car4UB(10) car5LB(0) car5UB(15) car6LB(-15) car6UB(15) car7LB(0) car7UB(0) replace

[Thomas Kaspereit]

Hello Christian,

Saw you responding on my post, I'm having the same problem. I recalculated the AR and CARs, these seem to be correct. However, I think there is something wrong in the average calculation.
For the T-scores, you can add the command: "outputfile(filename)" in the estudy line. In the Excel output there is a tab with all T-scores

Edit: MA is is market adjusted model, FM with one market index is the MM I guess.

Max is correct; FM with one factor is the MM. MA is the market adjusted model. See helpfile.

[Thomas Kaspereit]

Dear Christian,

Jared is correct; there are things going on "under the hood", in particular the Maynes & Rumsey algorithm to handle missing returns and defining appropirate datelines, see helpfile of eventstudy2.

For a comparison of commands, please see my 2021 Stata Journal article.

There is also another post on Statalist where I reply to a gentleman with the name "Glen". There, I write about why it is not easy to replicate and reconcile eventstudy2's results with one own*s Excel-based calculations.

Best
Thomas

[Adam Thomas]

Hi there,

I am currently attempting an event study using the eventstudy2 command. I have actually done the event study manually but want to use this command as it automatically completes the robustness tests. I was following a guide here https://thedatahall.com/event-study-...mand-in-stata/, however whenever I attempt the command, I receive this error when state is ''preparing security_return'': ''variables date __000005 do not uniquely identify observations in the using data r(459)''.

I have spent countless hours trying to figure out why this is, my data and formatting appear to be the exact same as the website and tutorial video.

Any help or advice on this matter would be greatly appreciated!
Thanks, Adam

[Thomas Kaspereit]

Dear Sir,

Thank you for using eventstudy2.

1) Please use the sample files that come with the installation (findit eventstudy2) and make the command run on them; this is not related to your problem but ensures that there are no other issues.
2) It seems that there are duplicates in your files where there should be not be any. Use the command "isid [your firm id] [your date variable]" on all your files, and see that you do not get any error there.

Best
Thomas Kaspereit

[Adam Thomas]

Hi Thomas

Thanks a lot for getting back to me.
I ran the sample files for the command and there was no issues. When I ran the isid command for my variables it gave me '' firm_id does not uniquely identify observations in the using data r(459)'', when I ran a duplicates report I got

Duplicates in terms of returndate

--------------------------------------
Copies | Observations Surplus
----------+---------------------------
1 | 1 0
62 | 62 61
64 | 38784 38178
127 | 127 126
--------------------------------------

As far as I can see, the duplicates in the stock return are coming from the return happening to be the same on different days, as I have rounded them to two decimal places. is there a way to mitigate that?

Thanks, Adam

[Thomas Kaspereit]

Dear Sir,

Same returns values are not the problem, but same firmid-date-combinations are the problem. Or same marketid-date combinations in the market return file.

Best
Thomas

[Adam Thomas]

Hi Thomas,

Thanks for the help, I eventually sorted the problem. However if it's not too much trouble I have another question. When I run it successfully, Stata runs an infinite loop of events despite me having only 3 in my event data file. Would you have any idea why this could be happening?