That's not the same as what he's suggesting though is it? He's suggesting a market based approach where you present possible policies to the market and arrange it so that the best one for your metric (e.g. reducing inequality) is also the best one to invest in. Then you enact the policy with the highest price.

Essentially you enlist the modelling skills of all those extremely highly paid quants to figure out which policy is the best.

I imagine it would be quite hard to set up - especially in a way that keeps the incentives you want intact - but I think it's a very interesting idea and deserves to be tried.