Reference no: EM132438174

Assignment -

1. Load the data file 'health_inclass.csv', conduct simple regression without correcting for endogeneity, and try to answer the question of whether having health insurance leads to higher or lower medical expenses. In this exercise, add more variables from the data, you can create dummy variables, add meaningful interaction variables. Try at least three models, and find the best one among the three, interpret the model results. Present all the three model results, and answer the following questions:

(1) Based on what metrics did you choose the "best" model?

(2) Do you think the endogeneity of the $HealthIns$ variable still exists? Why or why not?

2. Suppose the $HealthIns$ is still endogenous, even with your "best" model, use 'SSIRatio' variable as your instrument, and conduct the following exercises

(1) Use 'ivreg()' conduct the 2SLS estimates for your "best" model, while correcting for endogeneity of the $HealthIns$ variable.

(2) Compare the results from this model with those from the simple OLS approach, in terms of model fit, parameter interpretations, and your answers to the question "whether having health insurance leads to higher or lower medical expenses."

Attachment:- Assignment Files.rar