Announcement

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

  • reghdfe and two way clustering

    Dear Statalist

    I am running the regression with reghdfe with two-way clustering. I notice that for some t statistics that are clearly significant, the Stata will output a p-value that is not significant. For example, in my following regression output, I run the following command

    xi: reghdfe abndisx_new treat_post $control ,abs(gvkey fyear) vce(cluster gvkey fyear) keepsingleton

    However, in the stata output, variable mtb has a t statistics of -1.76 and a p-value of 0.117. Is there anyway to reconcile the difference? Thank you


    Robust
    abndisx_new Coef. Std. Err. t P>t [95% Conf. Interval]

    treat_post -.018163 .0088843 -2.04 0.075 -.0386502 .0023243
    mktshare_lag .0391951 .0195459 2.01 0.080 -.0058779 .0842682
    I_noa_sale_lag .0292783 .003381 8.66 0.000 .0214816 .0370749
    I_cycle_lag -.0149608 .0038909 -3.85 0.005 -.0239332 -.0059883
    I_io -.0043657 .0065904 -0.66 0.526 -.0195632 .0108319
    I_big8 .0097768 .0074874 1.31 0.228 -.0074891 .0270427
    I_autenure .0034669 .0050267 0.69 0.510 -.0081247 .0150584
    mtb -.0007963 .0004534 -1.76 0.117 -.0018419 .0002493
    roa .0867488 .0336602 2.58 0.033 .0091283 .1643693
    lnat .0003781 .0087508 0.04 0.967 -.0198013 .0205576
    div .0419227 .060474 0.69 0.508 -.0975305 .1813759
    capx_at -.0293348 .0532592 -0.55 0.597 -.1521509 .0934812
    lev .0150642 .0120801 1.25 0.248 -.0127925 .0429209
    I_debtissue -.0108426 .0023547 -4.60 0.002 -.0162726 -.0054125
    I_equityissue -.002722 .0033298 -0.82 0.437 -.0104006 .0049566

    Attached Files
    Last edited by STERLING HUANG; 22 Aug 2022, 08:20.
Working...
X