Array Hi,

I am writing a thesis on the effect of higher stringency on Covid-19 mortality in US counties. I am comparing all counties that border each other across state borders, and aiming to understand if there was an effect of stringency. After I have run these regressions, I want to try to understand the causality, as a rise in Array

may cause stringency to increase. To do this I am using political party as my instrument, however I am a bit confused on how to use ivreg2. Below is the code line I wrote, and the output it gave. Just wondering if this is correct, and if it is how do I interpret these results?

ivreg2 WeightedMortality StringencyIndex (Party=Pairs)

This is the regression. The thing I am confused about is which way around Party and Pairs go. If they are this way round then these are the results... Where it says Party has been instrumented.

However if I switch around Pairs and Party (Pairs refers to the pair of counties that border each other over state borders, so Florida Escambia and Alabama Baldwin are pair 1, then the next counties that border are pair 2 and so on).

ivreg2 WeightedMortality StringencyIndex (Pairs=Party)

The results are different and it says Party is an excluded instrument. Which way around is correct? How do I interpret these results?

Many thanks to anyone who is able to help!!

Best wishes

Charles