Available for an HonestDiD package for Sun&Abraham in Stata 18? #27
Description
Hi, I met some difficuties when applying honestdid in Sun and Abraham estimation. I could not generate a sensitivity coefplot. Here is the working example for your reference.
Sun and Abraham using command
eventstudyinteract
forvalues l = 0/4 {
gen L`l'event = K==`l'
}
forvalues l = 1/5 {
gen F`l'event = K==-`l'
}
drop F1event
eventstudyinteract ln_sum L*event F*event lnGDP lnpop lnstudent, vce(cluster province) absorb(province year) cohort(treat_year) control_cohort(lastcohort)
event_plot e(b_iw)#e(V_iw), default_look graph_opt(xtitle("Periods since the event") ///
ytitle("Average causal effect") xlabel(-5(1)4) title("Sun and Abraham (2020)") ///
name(SA, replace)) stub_lag(L#event) stub_lead(F#event) together
matrix sat_b = e(b_iw) // storing the estimates for later
matrix sat_v = e(V_iw)
estimates store eventstudyinteract
estimates replay eventstudyinteract
honestdid, post(1/5) pre(6/9) mvec(0(0.5)2) coefplot
However , honestdid generates a coefplot without estimates. And the reported results from eventstudyinteract are different from estimates dir eventstudyinteract:
eventstudyinteractcommand:
------------------------------------------------------------------------------
| Robust
ln_sum | Coefficient std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
L0event | .0106074 .0068322 1.55 0.131 -.0033661 .0245809
L1event | .03547 .0110083 3.22 0.003 .0129556 .0579844
L2event | .0365463 .0151308 2.42 0.022 .0056004 .0674923
L3event | .0442416 .0198347 2.23 0.034 .0036751 .0848082
L4event | .0769768 .0258086 2.98 0.006 .0241923 .1297614
F2event | -.0036461 .0072195 -0.51 0.617 -.0184117 .0111195
F3event | .0007 .0064344 0.11 0.914 -.0124599 .0138599
F4event | -.0073783 .0173413 -0.43 0.674 -.0428452 .0280886
F5event | -.0104371 .0230692 -0.45 0.654 -.057619 .0367447
lnGDP | -.0005063 .0084592 -0.06 0.953 -.0178073 .0167946
lnpop | -.0448455 .0414904 -1.08 0.289 -.1297028 .0400119
lnstudent | -.0427345 .0511465 -0.84 0.410 -.1473409 .0618719
------------------------------------------------------------------------------
estimates replay eventstudyinteractcommand:
------------------------------------------------------------------------------
| Robust
ln_sum | Coefficient std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
__00000U | 0 (omitted)
__00000V | .0174089 .0091639 1.90 0.067 -.0013333 .0361511
__00000W | -.0014576 .0083436 -0.17 0.863 -.0185222 .015607
__00000X | .0175572 .0068407 2.57 0.016 .0035663 .031548
__00000Y | 0 (omitted)
__00000Z | .0411389 .012254 3.36 0.002 .0160766 .0662011
__000010 | .0330809 .0157284 2.10 0.044 .0009126 .0652492
__000011 | 0 (omitted)
__000012 | 0 (omitted)
__000013 | .0367688 .0152167 2.42 0.022 .0056472 .0678904
__000014 | 0 (omitted)
__000015 | 0 (omitted)
__000016 | 0 (omitted)
__000017 | .0445295 .0199571 2.23 0.034 .0037126 .0853463
__000018 | 0 (omitted)
__000019 | 0 (omitted)
__00001A | 0 (omitted)
__00001B | .077482 .0259628 2.98 0.006 .024382 .1305819
__00001C | 0 (omitted)
__00001D | 0 (omitted)
__00001E | 0 (omitted)
__00001F | -.0164215 .0104065 -1.58 0.125 -.0377053 .0048622
__00001G | .0001596 .0085733 0.02 0.985 -.0173749 .017694
__00001H | .0059214 .0064786 0.91 0.368 -.0073288 .0191715
__00001I | 0 (omitted)
__00001J | 0 (omitted)
__00001K | .0014988 .0110473 0.14 0.893 -.0210954 .0240931
__00001L | -2.86e-06 .0086694 -0.00 1.000 -.0177337 .017728
__00001M | 0 (omitted)
__00001N | 0 (omitted)
__00001O | -.0307817 .0152167 -2.02 0.052 -.0619032 .0003399
__00001P | .0158392 .0101073 1.57 0.128 -.0048326 .036511
__00001Q | 0 (omitted)
__00001R | 0 (omitted)
__00001S | -.0419752 .0203078 -2.07 0.048 -.0835094 -.0004409
__00001T | .0211257 .0084485 2.50 0.018 .0038465 .0384049
__00001U | .1556738 .1001824 1.55 0.131 -.0492221 .3605698
__00001V | 0 (omitted)
__00001W | 0 (omitted)
__00001X | .3038947 .0934702 3.25 0.003 .1127266 .4950628
__00001Y | -.6078498 .4810137 -1.26 0.216 -1.591633 .3759337
__00001Z | 0 (omitted)
__000020 | 0 (omitted)
__000021 | 0 (omitted)
__000022 | .2683345 .3078596 0.87 0.391 -.3613091 .8979781
__000023 | 0 (omitted)
__000024 | 0 (omitted)
__000025 | 0 (omitted)
_cons | 13.68915 1.82066 7.52 0.000 9.96548 17.41282
------------------------------------------------------------------------------
And then I run the command honestdid, and return me a very weird result:
_honestARPConditionalTest(): LP for eta did not converge properly. Not rejecting
_honestARPConditionalTest(): LP for eta did not converge properly. Not rejecting (repeated for many times)
| M | lb | ub |
| ------- | ------ | ------ |
| . | 0.000 | 0.000 | (Original)
| 0.0000 | 0.000 | 0.000 | (*)
| 0.5000 | -0.017 | 0.017 | (*)
| 1.0000 | -0.033 | 0.033 | (*)
| 1.5000 | -0.050 | 0.050 | (*)
| 2.0000 | -0.066 | 0.066 | (*)
(method = C-LF, Delta = DeltaRM, alpha = 0.050)
(*) CI is open at both endpoints; CI length may not be accurate.
Try expanding the grid using the grid_lb() and grid_ub() options
Could you try to help me fix that? Many thanks!