Why don’t fractured-rebel regimes try to share power across armed groups? Finding #4: They TRY, but FAIL! 77% of fractured regimes try to share power outside of their own group; 96% of these attempts fall apart /8

Finding #1: FRACTURED-rebel regimes are *three times* more likely to break down in a given year than UNIFIED-rebel regimes. Statistically significant difference holds across regions & time periods, not driven by social revolutions, leftist rebellions, etc. /6

📊NEW DATA SET: Every rebel regime in the world between 1900-2020. We identify 84 rebel regimes in 61 countries, almost all authoritarian. INCLUDES weak, short-lived, and failed-state cases /5

But!☝️Mass violence also causes chaos and instability. Why should we necessarily expect the resulting regimes to rebuild durable order? /3

Common story: Rebel/revolutionary autocracies are exceptionally durable /2

🚨 New in @apsrjournal.bsky.social! 🚨How do the origins of authoritarian regimes shape their subsequent trajectories? @kbclarke.bsky.social @jackpaine.bsky.social and I are excited to share "Rebel Victory and Authoritarian (Dis)Order" /1

Summary table here /7

In light of Trump making the 3rd term comments on Sunday, I thought it would be interesting to revisit the broader data on term limit evasion attempts. In 2020, Mila Versteeg (UVA Law) and I worked with some law students to collect data on strategies incumbents use to circumvent term limits /1

Graph below shows that democracies always have high concession rates. Autocracies have historically had high rates of NO concessions, but recently have seen a *big increase* in strong concessions. The recent global increase in strong concessions is partly driven by trends in autocracies /15

Also, the probability of a concession is HIGHEST when the margin of victory is between 5-15%. Losers are more likely to challenge results when the election is very close /11

We also coded how long it takes (number of days) the loser to concede. Most candidates who concede do so within 2 days. /9

And this graph shows trends over time. Dark blue line is the rate of strong concessions, red line is the rate of no concessions. Strong concessions have increased in recent years /8

This graph shows that’s there’s a lot of within-country variance in concession rates! Dark blue = strong concessions, light blue= weak concession /7

🚨New paper alert!🚨 @anthlittle.bsky.social and I are excited to share “Presidential Election Concessions.” We present a new global dataset on concessions in presidential elections in all countries worldwide from 1980 to 2024 /1