eventstudy2: why my command neglects results of Mondays?

[Lynn Lyu]

Hello everyone

I am using the powerful command, eventstudy2 (Thomas Kaspereit, 2020) for a global event study. I was surprised to find that Stata does not generate results for all Mondays but works well for other dates. For every Monday, there is no result for [0,0] and [0,1], but have results for [-1,0]. I have tried many specifications but have not found a good solution. This is my code:

eventstudy2 security_id date using RT_F, ret(security_returns) model(FM) marketfile(RMT) marketreturns(market_returns) idmar(market_id) evwlb(-1)evwub(5) eswlb(-200) eswub(-30) minesw(30) shift(3) car1LB(0) car1UB(0) car2LB(0) car2UB(1) car3LB(0) car3UB(2) car4LB(0) car4UB(3) car5LB(0) car5UB(4) car6LB(0) car6UB(5) car7LB(-1) car7UB(0) car8LB(-1) car8UB(1)

Thanks in advance!

Best wishes,
Lynn

[Thomas Kaspereit]

Dear Lynn,

If you use the sample data that comes with the distribution of eventstudy2 ( net get eventstudy2 ) and type:

gen dow = dow(Date)
tab dow

You will see:

dow | Freq. Percent Cum.
------------+-----------------------------------
1 | 7 7.00 7.00
2 | 20 20.00 27.00
3 | 26 26.00 53.00
4 | 32 32.00 85.00
5 | 15 15.00 100.00
------------+-----------------------------------
Total | 100 100.00

which means that 7 events fall on Mondays.

After executing:

eventstudy2 Security_id Date using Security_returns if Ea > 0.05, ret(Return) car1LB(0) car1UB(0) mod(FM) marketfile(Factor_returns) mar(MKT) idmar(Market_reference) factor1(SMB) factor2(HML) risk(risk_free_rate) replace

on some positive earnings surprises > 0,05, and after loading the crossfile (click on it), and after typing

tab dow if CAR1 != .

you will see:


dow | Freq. Percent Cum.
------------+-----------------------------------
1 | 4 10.53 10.53
2 | 8 21.05 31.58
3 | 10 26.32 57.89
4 | 11 28.95 86.84
5 | 5 13.16 100.00
------------+-----------------------------------
Total | 38 100.00

which means that (C)ARs [0;0] were calculated for all 4 relevant events that fall on Mondays.

To summarize, eventstudy2 has no issue with handling Mondays.

Your issue could have to do with your data structure. Do you have Saturdays and/or Sundays in your dataset with missing returns, either in the security or market file? Deleting these rows could solve the issue.

Best
Thomas

[Lynn Lyu]

Thank you so much Thomas. Sorry for my late reply and just back from holiday. Yes you are right, after processing the values for the weekends, everything works well.

Thanks again for providing such an amazing tool!

Best wishes,
Lynn