Nerdish/wonkish: Let me follow up on endogeneity of policy interventions
In this case, we can sign the bias. Taking care of endogeneity would likely INCREASE the effectiveness of the interventions
Example: Geneva imposes mini-lockdown b/c it anticipates future increase ... 1/4
In this case, we can sign the bias. Taking care of endogeneity would likely INCREASE the effectiveness of the interventions
Example: Geneva imposes mini-lockdown b/c it anticipates future increase ... 1/4
https://twitter.com/LorenzKueng/status/1330089596147949571
... in cases, but say Vaud does not and doesn't intervene. Suppose both had correct expectations about the virus BEFORE Geneva's intervention
If we then after the intervention compare cases of Geneva with Vaud., we will underestimate how successful Geneva was, b/c Vaud ... 2/4
If we then after the intervention compare cases of Geneva with Vaud., we will underestimate how successful Geneva was, b/c Vaud ... 2/4
... is the wrong counterfactual. The right - but impossible - one would be Geneva w/o the intervention in a parallel universe
What can we do?
We can search for other cantons that had similar trajectories of infections as Geneva but didn't intervene (or intervened later). 3/4
What can we do?
We can search for other cantons that had similar trajectories of infections as Geneva but didn't intervene (or intervened later). 3/4
These other cantons would be a better control/comparison group than the Vaud. Their cases would grow faster than in the Vaud
Comparing Geneva to these cantons would give us a better answer to the initial question: "How well does a mini-lockdown work?"
[End nerdish/wonkish]
4/4
Comparing Geneva to these cantons would give us a better answer to the initial question: "How well does a mini-lockdown work?"
[End nerdish/wonkish]
4/4
• • •
Missing some Tweet in this thread? You can try to
force a refresh
