Productivity Decomposition in Stata

[Bridget Kauma]

Dear Statalist Members,

I am trying to do a decomposition analysis proposed by Foster-Haltiwanger-Krizan (2001)

I am struggling with the Between and Covariance parts of the decomposition analysis

***Foster, Haltiwanger, and Krizan (2001)**********



gen wvar=employment

gen pvar=lp /// labour productivity


*****firm level
egen s_weight=sum(wvar), by(year)
gen weight=wvar/s_weight

gen product=pvar*weight

egen Phi_t=sum(product), by(year)



*******Industry level*****

egen sj_weight=sum(wvar), by(year isic2)
gen weight_j=wvar/sj_weight

gen product_j=pvar*weight_j

egen Phj_t=sum(product_j), by(year isic2)


gen i_stayers=0
replace i_stayers=1 if i_within!=.
egen stayers_i= total(i_stayers), by(year isic2)


******Between

gen i_between_w=(weight-l1.weight)*(l1.pvar-l1.Phj_t) 

gen i_between_a=(weight_j-l1.weight_j)*(l1.Phj_t-l1.Phi_t)


egen bw= total(i_between_w), by(year) 

egen ba= total(i_between_a), by(year)

gen between= bw + ba


****Covariance

gen i_covariance_w=(weight-l1.weight)*[(l1.pvar-Phj_t)-(l1.pvar-l1.Phj_t)] 
 
gen i_covariance_a=(weight_j-l1.weight_j)*(Phj_t-l1.Phj_t)*stayers_i

egen cw= total(i_covariance_w), by(year)

egen ca= total(i_covariance_a), by(year)

egen covariance= cw + ca

I have attached the definition of the FHK (2001), for reference. Please help in how i can code it in Stata. My code doesn't seem to add up.

Best,
Bridget

[iris lee]

I also want to know how to use Stata to calculate the proposed decomposition analysis by Foster Haliwanger Krizan (2001)