Necessary condition to introduce a lagged dependent variable in a dynamic panel data with sem
Hello community,
We obtained data from a survey that asks for past 3 years indicators, and we took 3 waves of the same survey completed in year 2008, 2011, and 2014., which we will name from now on p1, p2,and p3 respectively. We also retained sample firms that completed the survey over all three waves to create a balanced panel dataset.
The question we have is whether we can introduced a lagged dependent variable from the previous panel (i.e. a DV from 2008 as a lagged DV for the panel in 2011), though the time interval is 3 years. The DV is measured as the innovation performance of the firm. We expect dynamic panel data and conduct the analysis using sem or gsem command. And the formulation is given below.
where, for unit i at period p, y, is the dependent variable and x is the independent variable, y(p-1) is a lagged dependent variable; β and δ are the regression coefficients; and ε it is an equation disturbance term.
Thank you very much in advance.
We obtained data from a survey that asks for past 3 years indicators, and we took 3 waves of the same survey completed in year 2008, 2011, and 2014., which we will name from now on p1, p2,and p3 respectively. We also retained sample firms that completed the survey over all three waves to create a balanced panel dataset.
The question we have is whether we can introduced a lagged dependent variable from the previous panel (i.e. a DV from 2008 as a lagged DV for the panel in 2011), though the time interval is 3 years. The DV is measured as the innovation performance of the firm. We expect dynamic panel data and conduct the analysis using sem or gsem command. And the formulation is given below.
where, for unit i at period p, y, is the dependent variable and x is the independent variable, y(p-1) is a lagged dependent variable; β and δ are the regression coefficients; and ε it is an equation disturbance term.
Thank you very much in advance.
Comment