Endogeneity?

Instrumental var: lottery number gaps between individuals' poorer vs richer neighbors 🎲 🎲

(Random lottery nums, lower = higher draft chance)

Higher nums for richer neighbors (in B's circle) ➡️ fewer richer drafted ➡️ bigger poorer-richer draft rate gap

[6/8]

For example:

▶️ Doctor connections ➡️ 11% higher health exemptions
▶️ The richer had more doctor connections

If lower drafting of the rich was observed, esp. among neighbors, it was perceived as unfair

[4/8]

1️⃣ These fairness concerns in Gallup polls were well-founded:

▶️ Highest income group's draft rate = *half* that of lowest income group
▶️ Rich more likely to get exemptions

Their exemptions? More questionable...

[3/8]

Key context:

▶️ 12 million draftees and 6 million volunteers during WW2
▶️ 20% of Gallup respondents perceived the draft as unfair
▶️ Main complaints: ‘favoritism’, ‘money talks’, and ‘some use pull to get out of draft’

Based on newly digitized military records, I show 🧵

[2/8]