Staggered DID approach - is this the best approach

[James Shepherd]

I have proposed to assess the impact of Belt and Road Initiative on 9 outcomes variables (social, environmental, economic) across ASEAN countries (11) from 2002-2022. I was building this idea of the back of seeing a study using staggered DID approach accounting for country heterogeneity for 1 outcome variable on a greater number of countries, which I now realise was a poor idea for extending. I'm concerned now that this is not the correct methodology to apply and is unlikely to be doable. Any advice would be greatly appreciated; this project is important for me. I have seen mention of Seemingly Unrelated Regressions (SUR) and Latent Growth Curve Modelling (LGCM) may be applicable, yet I'm not entirely sure. Thanks for any ideas on how to solve this issue.

[George Ford]

For one outcome, a variety of methods may be considered. The issue with multiple outcomes is that the errors across them may be correlated, making the SE of the one-by-one outcome approach tricky. If you use a command that suest can handle, then you can use suest to address that issue. SUR would do the same. You may run into issues where the dependent variables differ in type (continuous, ordered, etc.).

Maybe someone else has done this and has some ideas.