Great paper by Ahonen & Palanne
showing with exhaustive data on Finnish car mileage that horizontal distributive effects matter a lot under carbon pricing.
Only 50% of the variance is explained by "observables".
(Weird wording though since they can observe everything.)