Bank failure & Staggered Difference in Difference

[Helen Chang]

I aim to study a particular firm’s activity (dependent variable) following bank failures (Bank_fail). That is, using the bank failure as an exogenous shock. There can be multiple bank closures within a year at the state level.
The data that includes firm characteristics is a firm-year panel data. Bank failure data (bank-state-year level) will be matched to each firm based on the state of the firm’s headquarters in the corresponding year.

1. Matching bank failures to firms: Is this an appropriate method to link bank failures to firms?
2. Staggered difference-in-differences analysis: Given this setup, is it feasible to implement a staggered difference-in-differences (DiD) regression? If so, could you help with the model specification?

Any guidance or suggestions would be greatly appreciated. Thank you in advance!

[George Ford]

I wouldn't think bank failure would be an exogenous shock. It's seems more likely to be a slow process. Pre-trends could be an issue.

But, it seems like a staggered did approach would provide estimates (of something).