Multicollinearity with lagged dependent variable

[Julia Raciniewska]

Hello,

I have a question about testing for multicollinearity. I have panel data for 11 industries over 12 years and I am using FE model. When performing a RESET test I had omitted variables and when I included the lag of my dependent variables, this was no longer a problem. However, now when I run:

xtset industry time
xtreg employment L1.employment computer_use rd_spending price_computer i.t1, fe cluster(industry1) robust 
estat vce, corr

there is high correlation between computer_use and L1.employment. Is that a problem?
i.t1 is a dummy to indicate years of crises

[Carlo Lazzaro]

Julia:
see https://www.statalist.org/forums/for...tted-variables, #6.