GMM and Panel Data Models 2. Use the JTRAIN.dta database and consider the following equation: log(scrap)...

 

Transcribed Image Text:

GMM and Panel Data Models 2. Use the "JTRAIN.dta" database and consider the following equation: log(scrap) b_0+ b_1 union + d88+ d89 + b2hrsemp+ cX + e where 'scrap' represents the rate of broken products for company I in year t, union is a dummy variable indicating whether the company is affiliated with a union, d88 and d89 are dummies for those years, and hrsemp represents the training hours per employee. a) Choose additional control variables for X and justify your model choice. b) Under what assumptions is 'grant' (an indicator whether the company received a grant) a good instrument for the variable 'hrsemp'? c) Retrieve the weight matrix after GMM estimation and define any other matrices necessary to calculate the estimated value of the criterion function. d) Use the value of the criterion function to perform the Sargan-Hansen test. e) Replicate the estimation in (a) using two linear regressions. What are the differences between the estimation and the results here and in (a)? f) Conduct a test for the instrument's relevance and re-estimate the equation using first differences with the instrumental variable. Discuss your results and compare them with the previous ones. GMM and Panel Data Models 2. Use the "JTRAIN.dta" database and consider the following equation: log(scrap) b_0+ b_1 union + d88+ d89 + b2hrsemp+ cX + e where 'scrap' represents the rate of broken products for company I in year t, union is a dummy variable indicating whether the company is affiliated with a union, d88 and d89 are dummies for those years, and hrsemp represents the training hours per employee. a) Choose additional control variables for X and justify your model choice. b) Under what assumptions is 'grant' (an indicator whether the company received a grant) a good instrument for the variable 'hrsemp'? c) Retrieve the weight matrix after GMM estimation and define any other matrices necessary to calculate the estimated value of the criterion function. d) Use the value of the criterion function to perform the Sargan-Hansen test. e) Replicate the estimation in (a) using two linear regressions. What are the differences between the estimation and the results here and in (a)? f) Conduct a test for the instrument's relevance and re-estimate the equation using first differences with the instrumental variable. Discuss your results and compare them with the previous ones.