A nice letter in the FT recently described Uber as a form of cyber serfdom. It's a good analogy not just for Uber but for similar firms in the "sharing economy" such as Deliveroo. Whereas farmers once paid socage to their feudal lords, drivers now pay it to the owners of an app.

This poses the question: why do such feudal arrangements exist? Uber's drivers are supplying capital - the cars - and are bearing a lot of risk. Common sense says that people with more skin in the game should own the company, because they have the sharpest incentives to make it work. As Oliver Hart put it, "a party with an important investment or important human capital should have ownership rights."

So, why have we got Uber and Deliveroo rather than cooperatives of drivers or cyclists who simply hire an IT guy to write an app? I suspect there are three reasons.

One is a lack of entrepreneurship. If you're a cabbie or new migrant, you're struggling just to get by. You don't have the cognitive bandwidth to see a new profit opportunity or start a company.

Secondly, there's a standard collective action problem. Getting dozens of cabbies together to agree upon something is tricky. It's easier for the app developer to find cabbies and co-ordinate them than it is for cabbies to co-ordinate themselves. To use Oliver Williamson's analogy, a wheel-type network in which the central hub communicates with others can be easier than an all-channel network in which everybody tries to coordinate with everybody else*:

To the extent that the requisite information-processing and decision-making talents are not widely distributed, efficiency will be served by reserving the central information collection and decision-making position to the one or few individuals who have superior information processing capacities. (Markets and Hierarchies, p 52)

Thirdly, there are credit constraints. Uber began with $200,000 of seed funding. That's a lot for a team of cabbies to raise. And it's easier to raise the money when you can show backers a natty app than when you're merely hoping to buy the app - and, of course, much easier when you have a track record of a successful business behind you, as Uber's founder Garrett Camp had.

But here's the thing. Although these points explain why a coop didn't come up with Uber in the first place, they are consistent with the possibility that coops will displace it eventually. Uber has demonstrated proof of concept, and energized its contractors to try to cut out the middle man. As a result, it is facing some nascent competition. In the US People's Ride has begun as a challenger to Uber. And the Guardian reports that Yorkshire cabbies are starting their own coop. Such initiatives could spread.

In a classic paper, William Nordhaus pointed out that the profits from innovative activity usually got competed away. The question for Uber, therefore, is: does it have what Warren Buffett called an economic moat - protection against competition**? Is first mover advantage really sufficient? I'm not sure. And I'm certainly not $62.5bn sure.

Coops have started slowly in the race against cyber-feudalism, but they might win in the end.

* The collective action problem might be especially severe for larger groups, such as home-owners who use Airbnb or - eventually - owners of self-driving cars wanting to hire them out. Cyber-feudalism might have a bigger edge here.

** Apple's great designs and brand loyalty give it a big moat - which is why it has been an exception to Nordhaus's finding.