You can pay prizes after the fact for especially impressive art. Or even award grants in advance for especially impressive art-proposals. Similar to how we fund most academic research. That's likely to be a lot more effective than generalized subsidies if most art turns out to be just mediocre and forgettable.
There are too many subsidy receivers for anyone to ever visit in their life, even if they spend 40 hours a week 50 weeks a year visiting them.