Respected Sir,
I am Nadeem Ahmad, research scholar at Jawaharlal Nehru University. I am using following model in my research work, can anyone please help me in how to Jointly Estimate these 3 equation in stata, when 2 equation are Tobit and 1 is OLS. (it means can you please tell me the stata command/code for this model).
Model is:
Taking into account the fact that we cannot observe health spending for some individuals because they are not ill, we used a system of Tobit regressions that involved censoring through a conditional mixed process estimator procedure developed by Roodman(2007).
The procedure estimates multi-equations where the dependent variable of each equation may have a different format. The model is recursive with three structural equations and the estimation is a full information maximum likelihood (FIML).
The methodological approach is justified by our attempt to sort out the issues and concepts related to the econometric analysis of data such as health expenditure that contains a large number of zero expenditure observations, which can lead to a number of estimations bias when using inappropriate estimators.
A Seemingly Unrelated Equations (SUR) system of two Tobit models for household out-of-pocket health expenditure and the catastrophic health expenditure gap combined with a linear equation explaining household health status (measured as an ill-health score by counting the reported diseases within the household) is implemented. This accounts for correlations among unobservable factors affecting our endogenous dependent variables in addition to dealing with censoring issues.
The empirical model can be written as follows:
Where H∗i represents a censored variable measuring the distance from the catastrophic expenditure threshold for the householdi , observed only for a household whose health expenditure is above the threshold. The endogenous I∗i
corresponds to the total out-of-pocket household health expenditures. S reflects the (number of diseases in the household) measure of health status, inversely related to the aggregate household health stock. Xi, Mi and Ei are sets of exogenous household socioeconomic, environmental and control variables, such as location and household size, that enter into consideration in the specification of the endogenous variables. εHi εIi and εSi are random disturbances of the equations. The empirical model is jointly estimated.
Thanking YOU
with warm regards
I am Nadeem Ahmad, research scholar at Jawaharlal Nehru University. I am using following model in my research work, can anyone please help me in how to Jointly Estimate these 3 equation in stata, when 2 equation are Tobit and 1 is OLS. (it means can you please tell me the stata command/code for this model).
Model is:
Taking into account the fact that we cannot observe health spending for some individuals because they are not ill, we used a system of Tobit regressions that involved censoring through a conditional mixed process estimator procedure developed by Roodman(2007).
The procedure estimates multi-equations where the dependent variable of each equation may have a different format. The model is recursive with three structural equations and the estimation is a full information maximum likelihood (FIML).
The methodological approach is justified by our attempt to sort out the issues and concepts related to the econometric analysis of data such as health expenditure that contains a large number of zero expenditure observations, which can lead to a number of estimations bias when using inappropriate estimators.
A Seemingly Unrelated Equations (SUR) system of two Tobit models for household out-of-pocket health expenditure and the catastrophic health expenditure gap combined with a linear equation explaining household health status (measured as an ill-health score by counting the reported diseases within the household) is implemented. This accounts for correlations among unobservable factors affecting our endogenous dependent variables in addition to dealing with censoring issues.
The empirical model can be written as follows:
Where H∗i represents a censored variable measuring the distance from the catastrophic expenditure threshold for the householdi , observed only for a household whose health expenditure is above the threshold. The endogenous I∗i
corresponds to the total out-of-pocket household health expenditures. S reflects the (number of diseases in the household) measure of health status, inversely related to the aggregate household health stock. Xi, Mi and Ei are sets of exogenous household socioeconomic, environmental and control variables, such as location and household size, that enter into consideration in the specification of the endogenous variables. εHi εIi and εSi are random disturbances of the equations. The empirical model is jointly estimated.
Thanking YOU
with warm regards
Comment