The idea is not to reduce cab counts to the point where people call towncars instead; that's likely a worse outcome, because the towncar spends only ~1/3 of it's time full: the other 2/3rds are spent driving to and from the towncar's parking space.
Ideally cabs would spend near zero time driving around empty. Today this is solved by medallions, imperfectly, but tomorrow this can be solved by technology.
What can't be solved by technology is the number of cabs trolling around. You can have perfect GPS based dispatch, with your next fare always ready to get in your cab just as you're dropping off the last fare, and there would still be huge problems, because there's nothing to stop this ideal set of 100% in-use cabs from being diluted into five times as many 20% in use cabs. People will take the closest cab and drivers will drive around to make sure they're the closest cab.
You can charge a per-mile tax on empty taxis, but that just raises cab fares for citizens (or forces caps) for no good reason. Just set the optimal number of cabs by fiat.
There is no person, committee or even algorithm that can set the correct supply better than a freely trading market. This has been known for a long time.
>In some respects, this is a natural monopoly.
No, it's an artificial monopoly. What we are talking about here is exchanging rides for money. That is the market - paying someone to drive you on a point-to-point basis.
If the market were left to it's own devices, and the only regulations were around traceability, security and the usual laws against blackmail etc, then you'd have a floating market with different prices per time of day and per trip. This would actually work against congestion by pricing it in.
Instead we have a central government committee deciding supply by whatever they use, and fixing prices. The natural outcome of fixed supply and fixed prices is sub-standard service and a shortage of available product. Which is exactly what cab users everywhere experience.
I disagree: the number of cabs in circulation is a complex function of the market, but the ideal number of cabs is straightforward to calculate but won't be found by the market.
Why not? Given the choice between an optimal number of cabs (fully utilized) making a larger profit and an excessive number of cabs (underutilized) making a tinier profit, the market will always choose the latter. Markets will tend to attract competitors and thus supply until profit margins drop to near zero. In both cases prices are the same, revenue is the same, but in the latter case the pie has more slices, increasing congestion. There is less profit per taxi in the latter case, especially if prices drop with increased competition, but as long as there is at least an iota of profit, the number of cabs will not decrease to the most efficient level.
Taxing congestion itself is a good idea generally, and helps increase road efficiency, but this won't result in the optimal number of cabs, either. It will result in increased prices which will cut demand and thus supply, but given this new demand, it won't magically remove competion such that fewer cabs are servicing more passengers, relatively speaking. The number of cabs will still remain at the maximum number possible while still remaining profitable. For any given cab, the amount of time spent idling is still the same--too high.
The process that would stop 100% in-use cabs from being diluted into fives times as many 20% in-use cabs is that the 20% in-use cabs would all be losing money.
If congestion is the problem, then directly charge for that - London has for example implemented a time-based congestion charge for the city centre.
Ideally cabs would spend near zero time driving around empty. Today this is solved by medallions, imperfectly, but tomorrow this can be solved by technology.
What can't be solved by technology is the number of cabs trolling around. You can have perfect GPS based dispatch, with your next fare always ready to get in your cab just as you're dropping off the last fare, and there would still be huge problems, because there's nothing to stop this ideal set of 100% in-use cabs from being diluted into five times as many 20% in use cabs. People will take the closest cab and drivers will drive around to make sure they're the closest cab.
You can charge a per-mile tax on empty taxis, but that just raises cab fares for citizens (or forces caps) for no good reason. Just set the optimal number of cabs by fiat.
In some respects, this is a natural monopoly.