Stata provides a variety of diagnostic tests for regression models. These tests can be used to check the assumptions of OLS regression and to identify potential problems with the model. Some of the diagnostic tests available in Stata include:

• Heteroskedasticity tests
• Normality tests
• Autocorrelation tests
• Multicollinearity tests

Hello Chuck A Arize João Bento I use "parmby "xi:ardl variable1 variable2, maxlag( 2 2) nocons ec1 regstore(res)”, label norestore" command. After I forecast the ARDL model I try to examine the autocorrelation using the Breusch-Godfrey LM test (estat bgodfrey) and I get "time variable not set, use tsset varname" code (r111). I'm facing the same problem when I conduct the other diagnostic tests.

In stata 14, you can run the ardl/ecm without setting the time variable. I am using stata 17, and I can do same without setting the time variable. However, the documentation says you should set the time variable before running the ardl model.

So it's possible that's the cause of the error as indicated in the error message you got.

If your time variable is "quarters" just go with "tsset quarters" before you run the model.

I hope this helps.

Stata diagnostic tests are used to check how well your data meet the assumptions of OLS regression. These tests include a test for heteroskedasticity, a test for cross-sectional dependence, and diagnostic plots