Restricted OLS (Help!)

[Jenner Kerr]

Hi

I'm replicating some academic work on decomposing the drivers of wage growth and am unsure about how to code this into Stata.

The general form of the regression is below:

Yt = constant + X1t-1 + Yt-1 + Yt-2 + Yt-3 + Yt-4 + X2t + X3t + X4t + error

So, it's an ARDL regression with 4 AR components and four other explanatory variables. The difficult bit is that the paper imposes restrictions on the coefficients (below) and I'm unsure how to estimate the regression model with these coefficient restrictions.

- The sum of the coefficients on the lagged components (i.e. sum of coefficients on X1t-1, Yt-1, Yt-2, Yt-3 and Yt-4) are all equal to one.

- The coefficients on X1t-1 and X4t are equal.


Any help would be greatly appreciated!

Thanks,
Jenner

[Ken Chui]

That sounds like constrained regression, try check help cnsreg. If you read the PDF help document, Example 2 seems relevant.