All other things being equal they actually perform about the same for our use case.
Then let’s talk replication… MySql GTID/Binlog replication is stupid easy to setup. And _it just works _ and recovers really easy. You can make hundreds of near-real time replicas.
Finally the tooling. I find MySql’s tooling far easier to use than Postgres. Nearly everything is SQL. You have to memorize configuration in both cases so take this last point is an opinion.
Both are a good choice. I think would save if you need absolute single-server performance, probably Postgres is the better option, but understand you're only getting marginal gains. If you want to make hundreds of read-only replicas or use their semi-sync HA plugin, MySQL is probably way to go, but understand that multi-master writes is still not really a thing.