Authoritarian regimes may be different, but my read of the evidence — clearly implied by Ed Glaeser’s survey — is that rioting in democracies is overwhelmingly a micro-phenomenon where better response by authorities on the ground right at the beginning could (and usually does) prevent it without any meaningful change in broad social conditions. I think the key evidence for this is the sporadic occurrence of sports-related riots. Riots often seem to originate out of protest marches, which makes the rioting seem political. But the fact that riots also often originate in sports fans’ response to game outcomes should make us think that the real issue here is the existence of a crowd of people in the streets.
Combine the crowd with some bad apples and an inadequate initial response, and you get a riot.