Eventstudy2 package: Values for variables change in output file

[Marc Pelow]

Dear all,

I am using the eventstudy2 package and as recommended in the help file try to get the examples in the helpfile running.

Downloading the example files:

net get eventstudy2

I tried to perform an event study using the Fama French 3 Factor Model:

eventstudy2 Security_id Date using security_returns, returns(Return) model(FM)
marketfile(factor_returns) marketreturns(MKT) idmarket(Market_reference) factor1(SMB) factor2(HML)
riskfreerate(risk_free_rate) evwlb(-1) evwub(1) eswlb(-250) replace

There weren't any problems when I performed the event study. However, after checking the output file arfile.dta I became curious, since the return of the security in the output file doesn't match with the return of the security in the input file. Same goes for the Market return (MKT) and the factors SMB and HML:




As I looked through the help file I noticed the option logreturns which is the default setting when using the Factor Model:

By default,
eventstudy2 assumes that input returns are discrete, i.e. calculated by the formula
[(p_t+1-p_t)/p_t], where p is the dividend and stock splits adjusted share price, and
transforms those returns to continuously compounded returns, i.e. [ln(p_t+1)-ln(p_t)] if
abnormal_return_model is RAW, MA or FM. bnormal_return_model} is BHAR or BHAR_raw, returns
are kept as discrete returns, or are transformed to discrete returns if option logreturns is
specified.

After considering the logreturns command, the factors SMB and HML do actually match the ones in the input files, however, the security return and the market return still won't match:


Does anyone have a guess what could be the problem? Or maybe I'm wrong with the way I'm interpreting this?
Thanks a lot

[Pete Huckelba (StataCorp)]

No content, bump required while the index gets rebuilt.

-Pete