Right now space on NYC streets is free, leading to a 'tragedy of the commons': Businesses are each incentivized to overuse the resource because the costs, congestion and greenhouse gas emissions, are externalities for them - Uber and Lyft don't pay for them. With no cost to supply, it also leads to an oversupply of drivers, which depresses their incomes (another reason I don't like the 'contractor' model - the drivers take on all the risk and cost).

The obvious, market-based approach is that businesses bid for street capacity (and greenhouse gas emission capacity, but that's a different political issue), to shift the costs from an externality born by the public to the people consuming the resource, the businesses. If I understand correctly, that is how medallions work (it's funny how ideas suddenly make sense when you come to them on your own). But maybe there are more modern solutions: Remember Bloomberg proposed congestion pricing; I oppose that because it distributes a public good, the public streets and transportation, to citizens unequally based on ability to pay. However, if we think of congestion pricing as dynamically priced medallions, that seems like a more efficient solution for businesses who are consuming public goods. It also would seem to make the market more open; rather than long term reservations of capacity (medallions), anyone can buy in at any time; on the other hand, that might discourage long-term investment.