Why "weakiv" doesn't show the complete test?

[Grace Cai Lee]

Dear all,

I am using weakiv to generate the weak-instrument robust inference.

According to weakiv's document, there should be several tests. But I find that when there is single instrument, it only shows the AR test.

For example, use weakiv's first example:

clear
use http://www.stata.com/data/jwooldridge/eacsap/mroz.dta
gen byte poshours=(hours>0)
ivregress 2sls lwage exper expersq (educ = fatheduc)
weakiv

The result only shows AR test without the CLR, K, or J test. I was wondering why it is the case.

Any advice is greatly appreciated. Thanks!

[Fei Wang]

Grace, a quote from the help file (help weakiv) of weakiv (from SSC) would exactly answer your question.

In an exactly-identified model where the number of instruments equals the number of endogenous
regressors, it reports the Anderson-Rubin (AR) test statistic. When the IV model contains more
instruments than endogenous regressors (the model is overidentified), weakiv also conducts the
conditional likelihood ratio (CLR) test, the Lagrange multiplier K test, the J overidentification test,
and a combination of the K and overidentification tests (K-J).

[Grace Cai Lee]

Thanks a lot, Fei! Sorry for the overlook.

But do you know why it is the case? As I understand it, the AR test is the joint test for weak instruments and the exclusion condition. Is there anything I can do if I just want the test for weak instruments? Thanks a lot!

[Joro Kolev]

Try the user written -weakivtest-.

sysuse auto

. ivregress 2sls price (mpg = headroom) trunk

Instrumental variables 2SLS regression            Number of obs   =         74
                                                  Wald chi2(2)    =       0.87
                                                  Prob > chi2     =     0.6479
                                                  R-squared       =          .
                                                  Root MSE        =     9148.9

------------------------------------------------------------------------------
       price | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         mpg |   1657.354   4818.731     0.34   0.731    -7787.185    11101.89
       trunk |   1520.483   3798.846     0.40   0.689    -5925.119    8966.085
       _cons |  -50048.81   154814.1    -0.32   0.746    -353478.9    253381.3
------------------------------------------------------------------------------
Instrumented: mpg
 Instruments: trunk headroom

. weakivtest
(obs=74)

Montiel-Pflueger robust weak instrument test
--------------------------------------------
Effective F statistic:        0.159
Confidence level alpha:          5%
--------------------------------------------

--------------------------------------------
Critical Values             TSLS      LIML
--------------------------------------------
% of Worst Case Bias
tau=5%                    37.418    37.418
tau=10%                   23.109    23.109
tau=20%                   15.062    15.062
tau=30%                   12.039    12.039
--------------------------------------------

.

[Grace Cai Lee]

Thanks a lot for your reply, Joro Kolev! But weakivtest doesn't provide the weak-instrument robust inference, right? Thanks a lot!

Best,
Yajie