forked from fbelotti/Stata
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathxsmle.pkg
87 lines (82 loc) · 3.16 KB
/
xsmle.pkg
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
v 3
d xsmle. Spatial Panel Data Models (version 1.4.5 5jun2017).
d
d
d {cmd:xsmle} fits fixed or random effects spatial models for balanced panel data. See the
d {help mi} prefix command in order to use {cmd:xsmle} in the unbalanced case. Consider the
d following general specification for the spatial panel data model:
d
d y_it = tau*y_it-1 + psi*W*y_it-1 + rho*W*y_it + beta*X_it + D*Z_it*theta + a_i + gamma_t + v_it
d v_it = lambda*E*v_it + u_it
d
d where u_it is a normally distributed error term, W is the spatial matrix for the autoregressive
d component, D the spatial matrix for the spatially lagged independent variables, E the spatial
d matrix for the idiosyncratic error component. a_i is the individual fixed or random effect
d and gamma_t is the time effect.
d
d {cmd:xsmle} fits the following nested models:
d
d {cmd:i)} the SAR model with lagged dependent variable (theta=lambda=psi=0)
d
d y_it = tau*y_it-1 + rho*W*y_it + beta*X_it + a_i + gamma_t + u_it
d
d {cmd:ii)} the SAR model with time and space lagged dependent variable (theta=lambda=tau=0)
d
d y_it = psi*W*y_it-1 + rho*W*y_it + beta*X_it + a_i + gamma_t + u_it
d
d {cmd:iii)} the full dynamic SAR model (theta=lambda=0)
d
d y_it = tau*y_it-1 + psi*W*y_it-1 + rho*W*y_it + beta*X_it + a_i + gamma_t + u_it
d
d {cmd:iv)} the classical SAR model (theta=lambda=psi=tau=0)
d
d y_it = rho*W*y_it + beta*X_it + a_i + gamma_t + u_it
d
d {cmd:v)} the SDM model with lagged dependent variable (lambda=psi=0)
d
d y_it = tau*y_it-1 + rho*W*y_it + beta*X_it + D*Z_it*theta + a_i + gamma_t + u_it
d
d {cmd:vi)} the SDM model with time and space lagged dependent variable (lambda=tau=0)
d
d y_it = tau*y_it-1 + rho*W*y_it + beta*X_it + D*Z_it*theta + a_i + gamma_t + u_it
d
d {cmd:vii)} the full dynamic SDM model (lambda=0)
d
d y_it = tau*y_it-1 + psi*W*y_it-1 + rho*W*y_it + beta*X_it + D*Z_it*theta + a_i + gamma_t + u_it
d
d {cmd:viii)} the classical SDM model (lambda=tau=psi=0)
d
d y_it = rho*W*y_it + beta*X_it + D*Z_it*theta + a_i + gamma_t + u_it
d
d {cmd:xsmle} allows to use a different weighting matrix for the spatially lagged dependent variable (W) and the spatially lagged regressors (D) together with a different sets of explanatory (X_it) and spatially lagged regressors (Z_it).
d The default is to use W=D and X_it=Z_it.
d
d {cmd:ix)} the SAC model (theta=tau=psi=0)
d
d y_it = rho*W*y_it + beta*X_it + a_i + gamma_t + v_it
d v_it = lambda*E*v_it + u_it
d
d for which {cmd:xsmle} allows to use a different weighting matrix for the spatially lagged dependent variable (W) and the error term (E). {* The default is to use W=E.}
d
d {cmd:x)} the SEM model (rho=theta=tau=psi=0)
d
d y_it = beta*X_it + a_i + gamma_t + v_it
d v_it = lambda*E*v_it + u_it
d
d {cmd:xi)} the GSPRE model (rho=theta=tau=psi=0)
d
d y_it = beta*X_it + a_i + v_it
d a_i = phi*W*a_i + mu_i
d v_it = lambda*E*v_it + u_it
d
d where also the random effects have a spatial autoregressive form.
d
d
d Authors: Federico Belotti, Gordon Hughes, Andrea Piano Mortari.
d Distribution-Date: 20140312
* uploaded on 20 Dec 2016
f xsmle/xsmle.ado
f xsmle/xsmle.sthlp
f xsmle/xsmle_p.ado
f xsmle/xsmle_postestimation.sthlp
F xsmle/lxsmle.mlib