[Joao Santos Silva]

Dear Deboshmita Brahma,

Please check out the command ivpoisson.

Best wishes,

Joao

[Deboshmita Brahma]

Joao Santos Silva Thank you Sir. Actually, my Y variable does not follow Poisson distribution. To be exact, it follows compound Poisson, sometimes called Poisson Gamma distribution. Here the Var= Mean^p , where 1<p<2. In Poisson distribution, the p=1, but for my Y variable, this is not the case. Will ivpoisson be useful in this case Sir?

[Joao Santos Silva]

Dear Deboshmita Brahma,

As long as you just want to estimate the function (and not compute probabilities of events) that does not matter because the estimator is very robust. Please see the references in the help file.

Best wishes,

Joao

[Michael Zuze]

How are you, I want to run a simultaneous probit model between informal employment and poverty (both binary variables). From literature, there is a bidirectional relationship between informality and poverty, with various factors influencing both. Low earnings from informal employment significantly contribute to household poverty. On the other hand, household poverty can compel the head of the household to accept informal work, as they cannot afford to remain unemployed.


I am aiming to account for the endogeneity between informal employment and poverty, where both are treated as discrete endogenous variables. However, I am encountering difficulties in properly implementing the instrumental variable probit model in Stata, given the complexities involved. My models are

informal employment = f(poverty, other independent variables).......(1)
Poverty = f(informal employment, other independent variables similar to eqn 1)......(2)


I currently face challenges with the IVProbit model in my research especially measuring the relevance and identification:

I have incorporated the instruments into my analysis supporting with literature. Initially, I attempted to use the IVProbit model, but it failed to converge when both endogenous variables were included.
I then specified the model using biprobit as follows:

biprobit (hhinformal = poor hdmale i.hhage i.hheduc hhmarried hdsize tot_informal large_firm urban) (poor = hdmale i.hhage i.hheduc hhmarried hdsize tot_informal urban child_under6 m_hseduc)


instrument child_under6 and m_hseduc

I also specified it alternatively as:


biprobit (poor = hhinformal hdmale i.hhage i.hheduc hhmarried hdsize tot_informal urban child_under6 m_hseduc) (hhinformal = hdmale i.hhage i.hheduc hhmarried hdsize tot_informal large_firm urban)


instrument firm size

Currently, I need to test the relevance of my instruments. However, I am facing issues with the estat overid command in Stata how i can use it in biprobit case.

Could you please provide guidance on how to correctly test the relevance of my instruments or suggest an alternative approach?


The goal is to understand the bidirectional relationship between these variables while addressing potential endogeneity issues. Any advice, references to relevant literature, or examples of similar work do would be invaluable to advancing my analysis.