Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Adjusting the DSS Stata 'Prediction of normal returns formula'

    Dear Statalists,

    in order to be able to calculate non-parametric significance tests when doing an event study following the DSS Princeton explanation
    HTML Code:
    https://dss.princeton.edu/online_help/stats_packages/stata/eventstudy.html
    , i have a question regarding adapting the code. Please also bear in mind i am rather new to STATA as well as statistics, thus might have 'obvious' concept misunderstandings:

    1. The original code (merely adjusted by my variable names):
    Code:
    gen pred_ret10=.
    forvalues i=1(1)754 {
    l id group_id if id==`i' & dif==0
    reg co_ret mkt_ret if id==`i' & est_win1==1
    predict p if id==`i
    replace pred_ret10=p if id==`i' & evt_win10==1
    drop p
    }
    2. Adjusting this by adding a ",robust". What exact statistical/methodological difference will come out adding this "robust" term?

    Code:
    gen pred_ret10=.
    forvalues i=1(1)754 {
    l id group_id if id==`i' & dif==0
    reg co_ret mkt_ret if id==`i' & est_win1==1, robust
    predict p if id==`i
    replace pred_ret10=p if id==`i' & evt_win10==1
    drop p
    }
    3. How would i have to adjust the code so that i can get the standard deviation of estimation errors in order to be able and calculate the e.g.: G RANK T test as shown by eventstudytools.com
    HTML Code:
    https://www.eventstudytools.com/significance-tests
    . scroll down to No. 11.

    Looking forward to your answers.

    Best,
    Max
    Last edited by Max Shepard; 28 Mar 2019, 07:41.
Working...
X