For the long T panel, a possible option is the panel data techniques that account for possible cross-section dependence. See for instance: https://www.sciencedirect.com/science/article/abs/pii/S0304407620301020.

Hope this helps. Thank you.