When to include time fixed effects?

[Tina Samuel]

Hello statlist,
I am running a panel regression analysis of the impact of environmental aid on exports. I am using exporter fixed effects in my model and when I check the literature I find that none has used time fixed effects. I am not sure if I should add time fixed effects or not, especially that my references have not used it, but I would like academic ground probably statistical justification as to why not to add them. Any suggestion of any tests or academic references on this?

[JJ Kovach]

Time fixed effects are often used to account for anything that is time-specific that could influence the relationship between the dv and iv's. For instance, some periods in time may have been recessionary or in a pandemic, etc. that could influence larger economic decisions and outcomes. Some recent research has raised caution on the causal interpretation of both time and entity fixed effects in the same model, with the potential exception of a balanced TWFE difference in differences model. Here are some references I found on another forum.

http://web.mit.edu/insong/www/pdf/FEmatch-twoway.pdf

https://journals.plos.org/plosone/ar...l.pone.0231349

https://www.researchgate.net/publica...for_Panel_Data