Add example for dynamic differences in differences

[drbenvincent]

• [ ] In the current bank run example notebook... Clarify in the current example with multiple pre/post observations that this only makes sense if we assume the treatment effect to have a step-function type impact.
• [ ] But where we expect the treatment effects to vary over time, this will not be appropriate, so we have to use dynamic differences in differences. Add an example using dynamic differences in differences to the banking example notebook
• [ ] Optionally: Add a new example notebook just focussing on dynamic difference in differences, using a different (possibly simulated) dataset.

Resources:

• https://stats.stackexchange.com/questions/504144/difference-in-differences-dynamic-treatment-group-timing
• https://theeffectbook.net/ch-DifferenceinDifference.html
• https://lost-stats.github.io/Model_Estimation/Research_Design/event_study.html

[drbenvincent]

Tagging you @NathanielF just so you are aware. I know you are working on related stuff.