-estimates table- after -mi estimate- produces error (return code 321)

[Stephen Jenkins]

I have searched the manuals and googled more generally to see if there is a way to avoid the following. Estimates derived using mi estimate are not found by estimates table, presumably because mi estimate saves results to e(V_mi) and e(b_mi), not e(V) and e(b). Is there a way around this? E.g. short of making estimates table, aware of MI estimation, changing the saved result to a recognised form? ... Or adding a post option to mi estimate, analogous to the post option to margins?

.
estimates restore All
(results All are active now)

. estimates

----------------------------------------------------------------------------------------------------------------------------------------------------
Model All
----------------------------------------------------------------------------------------------------------------------------------------------------

Multiple-imputation estimates Imputations = 100
Linear regression Number of obs = 885
Average RVI = 0.0356
Largest FMI = 0.0384
Complete DF = 30
DF adjustment: Small sample DF: min = 27.25
avg = 27.81
max = 28.12
Model F test: Equal FMI F( 3, 28.1) = 2.72
Within VCE type: Robust Prob > F = 0.0633

(Within VCE adjusted for 31 clusters in Country2)
------------------------------------------------------------------------------
gini_net | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
inflation | .0123982 .0072842 1.70 0.100 -.0025414 .0273378
unemp | .1655875 .1795377 0.92 0.364 -.2021075 .5332826
time | .089944 .0323464 2.78 0.010 .0236611 .1562269
_cons | 27.10758 1.911754 14.18 0.000 23.1922 31.02296
------------------------------------------------------------------------------

. estimates table
All estimation results do not have e(b) and e(V)
r(321);

[Yulia Marchenko (StataCorp)]

estimates table does not work after mi estimate. The main reason is that estimates table does not have a logic for using coefficient-specific degrees of freedom when computing p-values which is crucial for mi estimate's results. If you are not interested in displaying p-values, however, you can use mi estimate's option post to repost MI results to e(b) and e(V), so that estimates table could find them.

We will consider extending estimates table to support mi estimate's results in the future.

[Stephen Jenkins]

Apologies, I overlooked mi estimate's post option amongst the long list of options (it's down in the Advanced collection). Problem solved. Thanks, Yulia