Can I use baseline averages of mediators for controls in a Difference-in-Differences (DiD) analysis?

External question — Cross Validated Stack Exchange Author: shepdog03 Original post: https://stats.stackexchange.com/questions/664294 License: CC BY-SA 4.0 — https://creativecommons.org/licenses/by-sa/4.0/ Adaptation: HTML converted to plain text; contact email addresses removed. I'm attempting to estimate the impact of the Belt and Road Initiative on inflation using staggered DiD. I've been able to get parallel trends to be met using controls unaffected by the initiative but still affect inflation in developing countries including corn yield, an inflation targeting dummy, and regional dummies. However, this feels like an inadequate set of controls, and my results are nearly all insignificant. The issue is how the initiative could affect inflation is multifaceted, and including usual monetary variables may introduce post-treatment bias as countries' governments are likely to react to inflationary pressure and other usual controls, including GDP growth, trade openness exchange rates, etc., are also affected by the treatment. My question is, could I use baselines of these variables (i.e., 3-year averages before treatment) in my model without blocking a causal pathway and would this be a valid approach? Some of what I have read seems to say this is OK, whilst others indicate the factors are most likely absorbed by fixed effects. Any help on this would be greatly appreciated.