```
cmd:help movestay}
-------------------------------------------------------------------------------

Title

movestay --    Maximum-likelihood estimation of endogenous switching
regression model

Syntax

movestay (depvar0 [=] varlist0) (depvar1 [=] varlist1) [if] [in] [
weight] , select(depvar_s [=] varlist_s) [options]

Syntax for predict

predict [type] newvar [if] [in] [, statistic]

statistic    Description
-------------------------------------------------------------------------
Main
psel        the probability of being in regime 1
xb0         fitted values for regime 0
xb1         fitted values for regime 1
yc0         fitted values for regime 1
yc1         fitted values for regime 1
mills0      Mills' ratio for regime 0
mills1      Mills' ratio for regime 1
-------------------------------------------------------------------------

options               Description
-------------------------------------------------------------------------
Model
select()             specify selection equation: dependent and
independent variables
collinear            keep collinear variables

SE/Robust
robust              robust estimator of variance
cluster(varname)    adjust standard errors for intragroup correlation

Reporting
level(#)            set confidence level; default is level(95)

Max option
maximize_options    control the maximization process;
-------------------------------------------------------------------------
fweights, iweights, and pweights are allowed;see weight.

Description

movestay uses the maximum likelihood method to estimate the endogenous
switching regression model.  It is implemented using the d2 evaluator to
calculate the overall log likelihood together with its first and second
derivatives.

movestay estimates all of the parameters in the model:

(regression equation for regime 0: y0 is depvar0, x1 is varlist0)
y0 = x0 * b0 + e_0

(regression equation for regime 1: y1 is depvar1, x1 is varlist1)
y1 = x1 * b1 + e_1

(selection equation: Z is varlist_s)
y0 observed if Zg + u <= 0
y1 observed if Zg + u > 0

where:
e_0 ~ N(0, sigma0)
e_1 ~ N(0, sigma1)
u ~ N(0, 1)
corr(e_0, u) = rho_0
corr(e_1, u) = rho_1

Here depvar0, depvar1 and varlist0, varlist1 are the dependent variables
and regressors for the underlying regression models (y0, y1 = xb), and
varlist_s specifies the variables Z thought to determine which regime is
observed.

Options

+-------+
----+ Model +------------------------------------------------------------

select() specifies variables in the selection equation. varlist_s
includes the set of instruments that help identify the model.  This
option is an integral part of the movestay estimation and is
required. The selection equation is estimated based on all exogenous
variables specified in the continuous equations plus instruments.  If
there are no instrumental variables in the model, depvar_s must be
specified.  In that case the model will be identified by
non-linearities and the selection equation will contain all the
independent variables that enter in the continuous equations.

collinear see estimation options.

+-----------+
----+ SE/Robust +--------------------------------------------------------

robust specifies that the Huber/White/sandwich estimator of the variance
is to be used in place of the conventional MLE variance estimator.
robust combined with cluster further allows observations which are
not independent within cluster (although they must be independent
between clusters).

If you specify pweights, robust is implied.  See [U] 23.14 Obtaining
robust variance estimates.

cluster(varname) specifies that the observations are independent across
groups (clusters) but not necessarily within groups.  varname
specifies to which group each observation belongs; e.g.,
cluster(personid) in data with repeated observations on individuals.
cluster() affects the estimated standard errors and
variance-covariance matrix of the estimators (VCE), but not the
estimated coefficients.  cluster() can be used with pweights to
produce estimates for unstratified cluster-sampled data.  Specifying
cluster() implies robust.

+-----------+
----+ Reporting +--------------------------------------------------------

level(#); see estimation options.

+-------------+
----+ Max options +------------------------------------------------------

maximize_options control the maximization process; see maximize.  With
the possible exception of iterate(0) and trace, you should only have
to specify them if the model is unstable.  The maximization uses
option difficult by default. This option need not be specified.

+------------------+
----+  predict options +-------------------------------------------------

psel calculates the probability of being in regime 1.

xb0 calculates the linear prediction for equation 0.

xb1 calculates the linear prediction for equation 1.

yc0 returns the predicted value of the dependent variable(s) in the
regime 0. For example, if earning function is modeled for two sectors
(regimes), then this option predicts the wage rate in sector one for
all individuals in the sample.

yc1 returns the predicted value of the dependent variable(s) in the
regime 1.

mills0 and mills1 calculate corresponding Mills' ratios for two regimes

Examples

To obtain full ML estimates:

Using instruments:

. movestay y1 x1 x2 x3 x4, select(regime1=z1 z2)

. movestay (y1= x1 x2 x3 x4) (y1= x1 x2 x3 x5), select(regime1=z1 z2)

Model is identified through non-linearities:

. movestay (y1= x1 x2 x3 x4) (y1= x1 x2 x3 x5), select(regime1)

To define and use each equation separately:

. global wage_eqn y x1 x2 x3 x4
. global select_eqn regime z1 z2

. movestay (\$wage_eqn), select(\$select_equn)

To use options:

. movestay y= x1 x2 x3 x4 if region=1 [w= hhweight], select(regime=
z1 z2)

. movestay (y= x1 x2 x3 x4) if region=1, select(regime= z1 z2)
tech("dfp")

Prediction:

. movestay y x1 x2 x3 x4, select(regime= z1 z2)

. predict yexpected, xb

. predict mymills1, mills1

Example from the Stata Journal:

. movestay lmo_wage age age2 edu13 edu4 edu5 reg2 reg3 reg4,
select(private =m_s1 job_hold)

Authors

M. Lokshin (DECRG, The World Bank) and Z. Sajaia (Stanford University).

Also see

Online:  help for regress, heckman, ml

```