Analysis covid impact

[Ayi Ayayi]

Hi
I have date pre. -covid from 2007 to 2009 and data post covid from 2020 to 2022. I want to analyze the impact of the covid on the performance (ROE) energy sector. The companies pre-Covid and post covid are the same. Performance is the dependent variable.. As for the independent variables I have among other, Liabilities, size and so on..
What kind of impact analysis can I conduct. Simply put will be the appropriate econometric model to use

[Jared Greathouse]

You can't do meaningful causal analysis here. If I have data from 1900 to 1910 and data from 1960-1999, these are two totally different time periods across which no meaningful temporal comparisons can be made, so, my advice to you is that you should find another topic.

[Clyde Schechter]

While I definitely agree, taking the content of #1 literally, with Jared Greathouse in #2, I wonder if perhaps O.P. meant to say he has data from 2017-2019, as well as 2020-2022. While, even then, it would be dicey to attribute differences between those eras just to covid, as there is really no control group possible for the covid pandemic, nevertheless one might be able to say something with a straight face.

[Ayi Ayayi]

Hi
my bad I meant 2017 to 2019.
Thanks to both of you

[Jared Greathouse]

Okay I stand corrected then!!!!!! Ummmmmm........ I guess it's possible, a little, but you couldn't use Stata. You'd need R/Python's CausalImpact which implements Bayesian Structural Time Series which I an definitively not an expert on, as these don't demand a control group.

Alternatively, you could use comparison industries that were conceivably less impacted by the pandemic, I guess, but this is unlikely.