Thank you very much, Nicola. I will try it and see what happens.
Branko
Development Research, World Bank
Email: [email protected] or branko_mi@yahoo.
tel: 202-473-6968
World Bank, Room MC 3-559
1818 H Street NW
Washington D.C. 20433
For "Worlds Apart" see
http://www.pupress.princeton.edu/titles/7946.html
Website:
http://econ.worldbank.org/projects/inequality
For papers see also:
http://econpapers.hhs.se/
http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=149002
nicola.baldini2
Sent by:
owner-statalist@hsp To
hsun2.harvard.edu [email protected]
cc
[email protected]
01/07/2008 12:14 PM Subject
Re: Re: st: Re:Two hurdle models
Please respond to
statalist@hsphsun2.
harvard.edu
I like collecting software and I downloaded -dhurdle-, too. I never used it, but
I wrote an help (just in case I am going to use it!) which I report below. I
mixed information from the webpages from which I downloaded it and information
from other commands for most used options (e.g. robust).
Obviously, since I never used the command, and I don't even have much of the
statistical knowledge underlying it, you can use it as a guidance - at YOUR OWN
RISK!!! I don't know what option -bracket- is and how it works.
You'd be better off contacting Fennema direclty for an extra help/a review of my
work. The command is quite old:
. which dhurdle
c:\ado\plus\d\dhurdle.ado
*! version 1.0 12 Feb 2003
Nicola
Double-hurdle model (model with selection and censoring)
--------------------------------------------------------
dhurdle varlist [weight] [if exp] [in range] [, select(d = varlist_s)
independent
cluster(varname) nocoef noconstant first from(init_values) level(integer)
noheader
nolog mlmethod(method) mlopts(option) offset(varname) robust score(string)
noskip
technique(algorithm_spec) ]
pweights, aweights, fweights, and iweights are allowed; see help weights.
Description
Censoring of the dependent variable is traditionally dealt with using the
Tobit
model. Cragg (1971) proposed the extension that the probability of a zero
realisation, is not directly to the density for a continuous realisation,
but
instead govern me other process. The original model made the assumption that
the
two error terms were jointly normal and uncorrelated. Similar to that
demonstrated
by McDowell (2003) for count models, the separability of the likelihood
function
permits the use bination of commands to estimate this model: truncreg and
probit.
This assumption has been relaxed in later work, e.g. Jones (1992). Both
parts of
this likelihood function must, however, be maximised simultaneously; there
is no
two-step equivalent. This is available using the dhurdle command. Note, of
course, the difficulties of these procedures in achieving convergence. See
Flood
and Gr�sj�, 2001, "A Monte Carlo Simulation of Tobit models", Applied
Economics
Letters, 8, pp.581-584 for an assessment of the problems of
misspecification.
Options
select(d = varlist_s) specifies selection equation: dependent and
independent
variables. d is a dummy taking the value of 1 if the dependent variable
is
greater than zero, 0 otherwise. This option is required.
independent estimates a double hurdle model with independent errors, the
original
Cragg formulation.
bracket �
cluster(varname) specifies that the observations are independent across
groups
(clusters) but not necessarily within groups. varname specifies to
which
group each observation belongs. See [U] 23.14 Obtaining robust variance
estimates.
nocoef does not display the coefficient table; seldom used.
noconstant supresses constant term.
first reports first-step estimates.
from(init_values) specifies initial values for the coefficients. You can
specify
the initial values in one of three ways: by specifying the name of a
vector
containing the initial values (e.g., from(b0) where b0 is a properly
labeled
vector); by specifying coefficient names with the values (e.g.,
from(age=2.1
/sigma=7.4)); or by specifying a list of values (e.g., from(2.1 7.4,
copy)).
from() is intended for use when you are doing bootstraps (see bootstrap)
and
in other special situations (e.g., with iterate(0)). Even when the
values
specified in from() are close to the values that maximize the
likelihood, only
a few iterations may be saved. Poor values in from() may lead to
convergence
problems.
level(integer) sets confidence level; default is level(95).
noheader suppresses display of the header reporting the estimation method
and the
table of equation summary statistics.
nolog prevents the iteration log from being shown.
mlmethod(method) where method is { lf | d0 | d1 | d1debug | d2 | d2debug }.
See
help mlmethod.
mlopts(option) specifies ml options.
nowarning suppresses "convergence not achieved" message of iterate(0).
novce substitutes the zero matrix for the variance matrix.
score(newvars) new variables containing the contribution to the score.
offset(varname) includes varname in model with coefficient constrained to 1.
robust specifies that the Huber/White/sandwich estimator of variance is to
be used
in place of the traditional calculation; see [U] 23.14 Obtaining robust
variance estimates.
score(newvars) calculates the equation-level score; the derivative of the
log
likelihood with respect to the linear prediction.
noskip specifies that a full maximum-likelihood model with only a constant
for the
regression equation be fitted. This model is not displayed but is used
as the
base model to compute a likelihood-ratio test for the model test
statistic
displayed in the estimation header. By default, the overall model test
statistic is an asymptotically equivalent Wald test of all the
parameters in
the regression equation being zero (except the constant). For many
models,
this option can substantially increase estimation time.
technique(algorithm_spec) specifies how the likelihood function is to be
maximized. The following algorithms are currently available.
technique(nr) specifies Stata's modified Newton-Raphson (NR) algorithm.
technique(bhhh) specifies the Berndt-Hall-Hall-Hausman (BHHH) algorithm.
technique(dfp) specifies Davidon-Fletcher-Powell (DFP) algorithm.
technique(bfgs) specifies the Broyden-Fletcher-Goldfarb-Shanno (BFGS)
algorithm.
The default is technique(nr).
You can switch between algorithms by specifying more than one in the
technique() option. By default, dhurdle will use an algorithm for five
iterations before switching to the next algorithm. To specify a
different
number of iter nclude the number after the technique in the option. For
example, specifying technique(bhhh 10 nr 1000) requests that dhurdle
perform
10 iterations using the BHHH algorithm perform 1000 iterations using the
NR
algorithm, and then swi to BHHH for 10 iterations, and so on. The
process
continues until convergence or until the maximum number of iterations is
reached.
At 02.33 05/01/2008 -0500, you wrote:
>Thanks a lot, Wilner. This is most helpful. I thought that the two-hurdles
>models have to be estimated jointly, but with independent estimation, Stata
>commands should suffice. Thanks again!
>
>Best regards,
>Branko
*
* 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/
*
* 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/
