Dear Statalisters,
I am researching poverty dynamics following the methods of Cappellari & Jenkins (2004; Modelling Low Income Transitions; http://www.jstor.org/stable/25146308) and Villa & Nino-Zarazua (2018; Poverty dynamics and graduation from CCT; https://doi.org/10.1007/s10888-018-9399-5). Their method requires estimation of trivariate probit model, which I did, using a -cmp- command by David Roodman. However, I am stuck when it gets to the specification tests reported in both papers, in particular, tests of:
rho_12 = rho_13=0
rho_12 = rho_23=0
rho_12 = rho_13=rho_23
Also, they do the LR test, including and excluding some variables from the 1st and 2nd equations of the trivariate probit. Is there any command to do such specification tests? I tried to find solutions to the issues faced, but I have not been able to find any answers. I would appreciate your assistance.
I am researching poverty dynamics following the methods of Cappellari & Jenkins (2004; Modelling Low Income Transitions; http://www.jstor.org/stable/25146308) and Villa & Nino-Zarazua (2018; Poverty dynamics and graduation from CCT; https://doi.org/10.1007/s10888-018-9399-5). Their method requires estimation of trivariate probit model, which I did, using a -cmp- command by David Roodman. However, I am stuck when it gets to the specification tests reported in both papers, in particular, tests of:
rho_12 = rho_13=0
rho_12 = rho_23=0
rho_12 = rho_13=rho_23
Also, they do the LR test, including and excluding some variables from the 1st and 2nd equations of the trivariate probit. Is there any command to do such specification tests? I tried to find solutions to the issues faced, but I have not been able to find any answers. I would appreciate your assistance.
