Granger causality test for panel data

[Laura Ramirez]

Dear statalist,

I am taking a look at the paper of Nelling and Webb 2008. They run a granger causality test described as follows

Score_t = a_0 + a_1Score_t-1 + a_2Perf_t + a_3Perf_t-1 + e_t
Perf_t = b_0 + b_1Perf_t-1 + b_2Score_t + b_3Score_t-1 + e_t

If the coefficients a_2/b_2 or a_3/b_3 are significant they conclude there is causality.


Have you ever come across to this method for testing causality? Would you suggest other method?


Best regards,


Laura R

[Carole J. Wilson]

I have not see this method (can you give the full citation?). I have used the approach described by Hood, Kidd, and Morris (2008) Political Analysis 16:324-344. Some years ago I implemented it in Stata and found it fairly straight forward (comparing constrained and unconstrained models).

[Laura Ramirez]

Hi Carole,

Thanks for your reply. I'll take a look at that paper. The full info of the one I found is

Corporate social responsibility and financial performance: the “virtuous circle” revisited, Review of Quantitative Finance and Accounting, February 2009, Volume 32, Issue 2, pp 197-209

[Laura Ramirez]

Carole,

Since you have already run the test, could you please explain what do they authors mean is step 3 "Create a set of slope parameters by multiplying each unit-specific variable by lags of the independent variable"

Thanks!!