[Limdep Nlogit List] Bivariate Probit model with partial observability

Alex Rutt amr59 at cam.ac.uk
Thu Feb 25 00:46:47 EST 2010


Dear LIMDEP users,

      I am currently trying to run the Abowd and Farber version of the 
bivariate probit model with partial observability, to see what causes 
participation in the IMF and I seem to be running into some problems. It 
is easiest if I show the problem:
I am using the command:
PROBIT; LHS=IMF; RH1=XG; RH2=XI; SELECTION $
+--------+--------------+----------------+--------+--------+----------+
|Variable|          Coefficient  | Standard Error |b/St.Er.|P[|Z|>z]| 
Mean of X|
+--------+--------------+----------------+--------+--------+----------+
---------+Index function for probability
 Constant|           -2.16815707      1.82257143    -1.190   .2342
 RESERVES|      -.24271329       .16623773    -1.460   .1443   3.51233875
 SURPLUS |       -.03111813       .03224380     -.965   .3345  -2.53184732
 TRADE   |         -.00954696       .00626423    -1.524   .1275   76.6414535
 NUMBER  |       .06571627       .04534212     1.449   .1472   62.7368839
 ELECTION|      .29031683       .43239393      .671   .5020    .22575517
 INVESTME|     -.02962498       .02376683    -1.246   .2126   21.1233295
 LIMF    |            3.14136664      2.33501147     1.345   .1785    
.38712242
 LRES    |            .27389091       .17432870     1.571   .1162   
1.30658217
 LSUR    |           -.01134824       .05324245     -.213   .8312  
-1.01219043
 LTRADE  |        .00841883       .00698163     1.206   .2279   28.4096367
 LNUMBER |     -.06133548       .05465241    -1.122   .2617   24.5604134
 LELECT  |        -.74581275       .53669143    -1.390   .1646    .07869634
 LINVEST |        .02877999       .02636821     1.091   .2751   6.73525817
 Constant|           .83231637      1.65477100      .503   .6150
 BOP     |          .00153257       .01117504      .137   .8909      .000000
 NUMBER  |    -.03432942       .02431216    -1.412   .1579      .000000
 REGIME  |     .77239875       .21497799     3.593   .0003      .000000
 LIMF    |         270.476421     .195868D+14      .000  1.0000      .000000
 LBOP    |         -2.56202237     .338401D+12      .000  1.0000      
.000000
 LNUMBER |    2.89944129     .439716D+12      .000  1.0000      .000000
 LREGIME |     -380.076163     .345231D+13      .000  1.0000      .000000


Basically the last four terms are commind out with ver high standard 
errors ad larger than expected results for the coefficients.
I removed these variables and all that happened was that the next four 
variables experienced the same problem.
I was wondering whether anyone knew what was happening here and if it 
has occured else where.

Thanks for any help,

Alex Rutt


More information about the Limdep mailing list