Socrates,
My guess is the key is your Question 3. These variables are all
prefixed with "d". Are they dummies? If so, it's easy to see how
first-differencing would wipe a lot of them out. Solve this and
you'll probably solve Q2 as well. Regarding Q1, lots of instruments
relative to the number of observations can generate severe finite-
sample bias problems. It's nice that -xtabond2- alerts you to this;
it's a potentially serious problem that many researchers aren't aware
of.
--Mark
From: Socrates Mokkas <[email protected]>
Date sent: Wed, 13 Oct 2004 13:23:25 +0100 (BST)
To: [email protected]
Subject: st: xtabond2 problem
Send reply to: [email protected]
> Hi there,
>
> I am trying to estimate a model with xtabond2 and I keep coming up with serious problems.
> Below is the output of my effort.
>
> Question No1: Why is it warning me that instruments are large relative to number of observations and how can I reduce the instruments. The only variable that I want to instrument is the lag of the dependent variable.
>
> Question No2: Why is the covariance matrix of moment conditions singular? How can I get that right?
>
> Question No3: Why is it dropping almost all the variables? I know that they are not multicollinear!
>
> Thanks very much for your help,
>
> Socrates
>
> . xtabond2 ca L.ca dsolpr dsol dsolaft dwolpr dwol dwolaft dwcpr dwc dwcaft d1-d131, gmmstyle(L.ca, lag(1 2)) twostep robust
> d1 dropped because of collinearity.
> Building GMM instruments..
> Warning: Number of instruments may be large relative to number of observations.
> Estimating.
> Warning: Two-step estimated covariance matrix of moment conditions is singular.
> Number of instruments may be large relative to number of groups.
> Using a generalized inverse to calculate optimal weighting matrix for two-step estimatio
> > n.
> Computing Windmeijer finite-sample correction...........
> Performing specification tests.
>
> Arellano-Bond dynamic panel-data estimation, two-step system GMM results
> ------------------------------------------------------------------------------
> Group variable: cntr Number of obs = 1169
> Time variable : t Number of groups = 10
> Number of instruments = 390 Obs per group: min = 87
> F(140, 9) = 0.08 avg = 116.90
> Prob > F = 1.000 max = 131
> ------------------------------------------------------------------------------
> | Corrected
> | Coef. Std. Err. t P>|t| [95% Conf. Interval]
> -------------+----------------------------------------------------------------
> ca |
> L1 | (dropped)
> dsolpr | .0456957 .1109821 0.41 0.690 -.2053633 .2967546
> dsol | (dropped)
> dsolaft | (dropped)
> dwolpr | (dropped)
> dwol | (dropped)
> dwolaft | .0516802 .6528166 0.08 0.939 -1.425094 1.528454
> dwcpr | (dropped)
> dwc | (dropped)
> dwcaft | -.3606597 1.910496 -0.19 0.854 -4.682502 3.961182
> d2 | (dropped)
> d3 | (dropped)
> d4 | (dropped)
> d5 | (dropped)
> d6 | (dropped)
> d7 | (dropped)
> d8 | (dropped)
> d9 | (dropped)
> d10 | (dropped)
> d11 | (dropped)
> d12 | (dropped)
> d13 | (dropped)
> d14 | (dropped)
> d15 | (dropped)
> d16 | (dropped)
> d17 | (dropped)
> d18 | (dropped)
> d19 | (dropped)
> d20 | (dropped)
> d21 | (dropped)
> d22 | .004763 .0807997 0.06 0.954 -.1780188 .1875447
> d23 | .0389608 .0370386 1.05 0.320 -.0448263 .1227479
> d24 | (dropped)
> d25 | (dropped)
> d26 | (dropped)
> d27 | (dropped)
> d28 | (dropped)
> d29 | (dropped)
> d30 | (dropped)
> d31 | (dropped)
> d32 | .0040198 .0365386 0.11 0.915 -.0786363 .086676
> d33 | (dropped)
> d34 | (dropped)
> d35 | (dropped)
> d36 | (dropped)
> d37 | (dropped)
> d38 | (dropped)
> d39 | (dropped)
> d40 | (dropped)
> d41 | (dropped)
> d42 | (dropped)
> d43 | (dropped)
> d44 | (dropped)
> d45 | (dropped)
> d46 | (dropped)
> d47 | (dropped)
> d48 | (dropped)
> d49 | (dropped)
> d50 | (dropped)
> d51 | (dropped)
> d52 | (dropped)
> d53 | (dropped)
> d54 | (dropped)
> d55 | (dropped)
> d56 | (dropped)
> d57 | (dropped)
> d58 | (dropped)
> d59 | (dropped)
> d60 | (dropped)
> d61 | (dropped)
> d62 | (dropped)
> d63 | (dropped)
> d64 | (dropped)
> d65 | (dropped)
> d66 | (dropped)
> d67 | (dropped)
> d68 | (dropped)
> d69 | (dropped)
> d70 | (dropped)
> d71 | (dropped)
> d72 | (dropped)
> d73 | (dropped)
> d74 | (dropped)
> d75 | (dropped)
> d76 | (dropped)
> d77 | (dropped)
> d78 | (dropped)
> d79 | (dropped)
> d80 | (dropped)
> d81 | (dropped)
> d82 | (dropped)
> d83 | (dropped)
> d84 | (dropped)
> d85 | (dropped)
> d86 | (dropped)
> d87 | (dropped)
> d88 | (dropped)
> d89 | (dropped)
> d90 | (dropped)
> d91 | (dropped)
> d92 | (dropped)
> d93 | (dropped)
> d94 | (dropped)
> d95 | (dropped)
> d96 | (dropped)
> d97 | (dropped)
> d98 | (dropped)
> d99 | (dropped)
> d100 | (dropped)
> d101 | (dropped)
> d102 | (dropped)
> d103 | (dropped)
> d104 | (dropped)
> d105 | (dropped)
> d106 | (dropped)
> d107 | (dropped)
> d108 | (dropped)
> d109 | (dropped)
> d110 | (dropped)
> d111 | .9831839 3.246474 0.30 0.769 -6.360851 8.327219
> d112 | .7838466 3.484041 0.22 0.827 -7.097601 8.665295
> d113 | (dropped)
> d114 | -.0019199 .3928971 -0.00 0.996 -.8907149 .886875
> d115 | (dropped)
> d116 | (dropped)
> d117 | (dropped)
> d118 | (dropped)
> d119 | (dropped)
> d120 | (dropped)
> d121 | (dropped)
> d122 | (dropped)
> d123 | (dropped)
> d124 | (dropped)
> d125 | (dropped)
> d126 | (dropped)
> d127 | (dropped)
> d128 | (dropped)
> d129 | (dropped)
> d130 | (dropped)
> d131 | (dropped)
> _cons | .0095233 .1761511 0.05 0.958 -.3889581 .4080047
> ------------------------------------------------------------------------------
> Hansen test of overid. restrictions: chi2(249) = 0.00 Prob > chi2 = 1.000
>
> Arellano-Bond test for AR(1) in first differences: z = -0.06 Pr > z = 0.949
> Arellano-Bond test for AR(2) in first differences: z = -0.23 Pr > z = 0.814
> ------------------------------------------------------------------------------
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
Prof. Mark E. Schaffer
Director
Centre for Economic Reform and Transformation
Department of Economics
School of Management & Languages
Heriot-Watt University, Edinburgh EH14 4AS UK
44-131-451-3494 direct
44-131-451-3008 fax
44-131-451-3485 CERT administrator
http://www.som.hw.ac.uk/cert
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
