I am trying Heckman two-step analysis for remittances and agriculture investment based on survey data.
1. the problem is with margins, i get dy/dx = 0 and S.E omitted, why?
2. can i include the select equation variable i.e rem_dom_dummy in my regression equation as well, or it should be only in the select equation?
3. a few comments on the interpretation of the model (Wald, lambda, rho, sigma) will be highly appreciated.
Thanking you in anticipation.
heckman lagri_inv_value $xlist, select (agri_inv_dummy = rem_dom_dummy) twostep
where
lagri_inv_value = log of investment in agriculture
agri_inv_dummy= Dichotomous for agriculture investment
rem_dom_dummy= Domestic remittances dummy (remittances recipient and non recipient household)
$xlist is the list of control variables including, region gender_head age age2 marit_stat hh_size dep_ratio educ_status lincome_hh lexp_hh land_owned total_cultivated irri_land_dummy
. heckman lagri_inv_value $xlist, select (agri_inv_dummy = rem_dom_dummy) twostep
Heckman selection model -- two-step estimates Number of obs = 3,323
(regression model with sample selection) Censored obs = 2,675
Uncensored obs = 648
Wald chi2(13) = 219.38
Prob > chi2 = 0.0000
----------------------------------------------------------------------------------
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-----------------+----------------------------------------------------------------
lagri_inv_value |
region | -.0288873 .2271991 -0.13 0.899 -.4741892 .4164147
gender_head | -.5507128 .6942784 -0.79 0.428 -1.911474 .8100479
age | -.029743 .0310844 -0.96 0.339 -.0906672 .0311813
age2 | .0001303 .0002992 0.44 0.663 -.0004561 .0007166
marit_stat | .2573907 .2887579 0.89 0.373 -.3085644 .8233458
hh_size | -.0325494 .0232412 -1.40 0.161 -.0781013 .0130025
dep_ratio | -.5418882 .3581369 -1.51 0.130 -1.243824 .1600472
educ_status | .2408655 .0950719 2.53 0.011 .054528 .4272031
lincome_hh | -.2368576 .0857027 -2.76 0.006 -.4048318 -.0688834
lexp_hh | 1.035776 .184998 5.60 0.000 .6731866 1.398365
land_owned | -1.413493 .20678 -6.84 0.000 -1.818774 -1.008211
total_cultivated | .0331836 .0050109 6.62 0.000 .0233623 .0430049
irri_land_dummy | -.265573 .2389821 -1.11 0.266 -.7339692 .2028233
_cons | 1.046698 2.53965 0.41 0.680 -3.930924 6.024321
-----------------+----------------------------------------------------------------
agri_inv_dummy |
rem_dom_dummy | -.5159296 .0901819 -5.72 0.000 -.6926829 -.3391763
_cons | -.8084836 .0262016 -30.86 0.000 -.8598379 -.7571294
-----------------+----------------------------------------------------------------
mills |
lambda | .3975703 .7878292 0.50 0.614 -1.146547 1.941687
-----------------+----------------------------------------------------------------
rho | 0.21220
sigma | 1.8735855
----------------------------------------------------------------------------------
. margins, dydx(rem_dom_dummy)
Average marginal effects Number of obs = 3,323
Model VCE : Conventional
Expression : Linear prediction, predict()
dy/dx w.r.t. : rem_dom_dummy
-------------------------------------------------------------------------------
| Delta-method
| dy/dx Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
rem_dom_dummy | 0 (omitted)
-------------------------------------------------------------------------------
1. the problem is with margins, i get dy/dx = 0 and S.E omitted, why?
2. can i include the select equation variable i.e rem_dom_dummy in my regression equation as well, or it should be only in the select equation?
3. a few comments on the interpretation of the model (Wald, lambda, rho, sigma) will be highly appreciated.
Thanking you in anticipation.
heckman lagri_inv_value $xlist, select (agri_inv_dummy = rem_dom_dummy) twostep
where
lagri_inv_value = log of investment in agriculture
agri_inv_dummy= Dichotomous for agriculture investment
rem_dom_dummy= Domestic remittances dummy (remittances recipient and non recipient household)
$xlist is the list of control variables including, region gender_head age age2 marit_stat hh_size dep_ratio educ_status lincome_hh lexp_hh land_owned total_cultivated irri_land_dummy
. heckman lagri_inv_value $xlist, select (agri_inv_dummy = rem_dom_dummy) twostep
Heckman selection model -- two-step estimates Number of obs = 3,323
(regression model with sample selection) Censored obs = 2,675
Uncensored obs = 648
Wald chi2(13) = 219.38
Prob > chi2 = 0.0000
----------------------------------------------------------------------------------
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-----------------+----------------------------------------------------------------
lagri_inv_value |
region | -.0288873 .2271991 -0.13 0.899 -.4741892 .4164147
gender_head | -.5507128 .6942784 -0.79 0.428 -1.911474 .8100479
age | -.029743 .0310844 -0.96 0.339 -.0906672 .0311813
age2 | .0001303 .0002992 0.44 0.663 -.0004561 .0007166
marit_stat | .2573907 .2887579 0.89 0.373 -.3085644 .8233458
hh_size | -.0325494 .0232412 -1.40 0.161 -.0781013 .0130025
dep_ratio | -.5418882 .3581369 -1.51 0.130 -1.243824 .1600472
educ_status | .2408655 .0950719 2.53 0.011 .054528 .4272031
lincome_hh | -.2368576 .0857027 -2.76 0.006 -.4048318 -.0688834
lexp_hh | 1.035776 .184998 5.60 0.000 .6731866 1.398365
land_owned | -1.413493 .20678 -6.84 0.000 -1.818774 -1.008211
total_cultivated | .0331836 .0050109 6.62 0.000 .0233623 .0430049
irri_land_dummy | -.265573 .2389821 -1.11 0.266 -.7339692 .2028233
_cons | 1.046698 2.53965 0.41 0.680 -3.930924 6.024321
-----------------+----------------------------------------------------------------
agri_inv_dummy |
rem_dom_dummy | -.5159296 .0901819 -5.72 0.000 -.6926829 -.3391763
_cons | -.8084836 .0262016 -30.86 0.000 -.8598379 -.7571294
-----------------+----------------------------------------------------------------
mills |
lambda | .3975703 .7878292 0.50 0.614 -1.146547 1.941687
-----------------+----------------------------------------------------------------
rho | 0.21220
sigma | 1.8735855
----------------------------------------------------------------------------------
. margins, dydx(rem_dom_dummy)
Average marginal effects Number of obs = 3,323
Model VCE : Conventional
Expression : Linear prediction, predict()
dy/dx w.r.t. : rem_dom_dummy
-------------------------------------------------------------------------------
| Delta-method
| dy/dx Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
rem_dom_dummy | 0 (omitted)
-------------------------------------------------------------------------------
