Markets, enterprises, and data: blockchains and the role of disintermediation

Data and information technology have already reshaped the global economy more than we realize, fundamentally changing the way markets and enterprises interact. What effects have we seen so far, what can we expect, and how can we separate the good from the bad? An unrolled tweetstorm of mid technicality, originally posted at @ecoinomia.

The most important economic contribution of blockchain might be that it shifts actual marketplaces closer to the economic abstraction of a market. Obversely, this might also become blockchain’s most important contribution to economics.

The Buttonwood Agreement established the New York Stock Exchange as a P2P network between brokers by cutting out the auctioneer middlemen. It also set a 0.25% fee. Today the NYSE is a subsidiary an Atlanta-based IT company which runs exchanges.

Let me use this example as an anchor to discuss the contested boundary of markets vs enterprises (hierarchies), and the role of processes vs mechanisms in that contest.

Keep your markets free and tidy. Photo by Arturo Rivera on Unsplash.
Standard transaction cost/contracting theory considers enterprises as solutions for opportunism (TCE) or hold-up (CT). Short version: things can go wrong, contracting parties might take advantage, so it makes sense to agree in advance who gets to decide. Voila: hierarchy.

Discussions on the “thick middle” tends to focus on hybrid contractual arrangements like joint ventures, consortia or limited partnerships. I’m more interested in the idea of a marketplace as an enterprise.


I’m sure you’ve seen some kind of variant of what I call “Goodwin’s meme.” In 2015, Tom Goodwin (not to be confused with Mike Godwin of the eponymous law) somewhat breathlessly realized that “Uber, the world’s largest taxi company, owns no vehicles”, etc.

Entertainingly, he talked about Uber, AirBnB, Facebook, etc. in the context of disintermediation. Today we call their business models, somewhat incorrectly, platform strategy: the platform as a shared virtual space humans can dock onto to get together and exchange things.

This platform of course has been around forever. In ancient Greece it was called an Agora: a public space for gathering, philosophizing, politicizing, and to buy and sell stuff. A marketplace for ideas and actual, commercial wares. Tom Goodwin inadvertently rediscovered the market.

Before Zuckerberg and Kalanick recognized the potential of creating algorithm-based marketplaces, economists have discovered the same. And before economists, or maybe around the same time (1995), Craig Newman of craigslist did.

Craigslist is an interesting example because as an accidental creation at the dawn of the world-wide web, it is still run as a sort-of idealistic marketplace, revenue-generating but not profit-maximizing, mostly community-driven but privately owned.


Facebook, Uber, Google, Alibaba, Intercontinental Exchanges (the parent of NYSE and a dozen-ish other exchanges) are of course purely rent-driven operations. Economic rents = profits reaped from exclusive access. Uber is laughably bad at it, but the others make boatloads of money.

NYSE in the meantime operated as a consortial non-profit entity (a gentlemen’s club) until the mid-2000s, when it started to look into reorganizing itself as a for-profit. Since then it has become a ping-pong ball of corporate interests until ending up as a subsidiary of ICE.

[Side note: Intercontinental Exchange, or ICE, created an operating income of $2.1B out of $4.5B in revenues in 2016. Nice money if you have it.]


When Goodwin looked at Facebook etc. as disintermediators he probably thought of consumers vertically integrating into production (aka prosumers) and of direct, algorithm-facilitated interaction. But you also directly buy your veggies from the farmer at the farmers’ market.

So how does this square with our perception that Facebook, Google, etc are indeed rent-seeking intermediaries, members of the internet-age middleman club, or, as I like to call them: the Sicilian Water Mafia?


Sicily is a mountainous, hot, agrarian place. Water means life. Mafia, opportunistic as they are, realized that and took control of the water flow. If you are a farmer and invested your livelihood into bringing your harvest to 90% of fruition you’d pay everything to reach the last 10%.

Farmers are locked in in three ways: time, place, and product. You cannot simply wait, move, or replace water with something else. This offers a perfect opportunity for rent extraction, and Mafia are world champion rent-extractors.


But back to the quandary how the Facebooks and Ubers can be disintermediating middlemen, free services and “mafia” at the same time.

If you go to a farmer’s market, you don’t expect the market operator to hand you the produce. You expect three things: stewardship, custodianship, and, if they’re well organized, a “Big Board” that shows you the ongoing or past deals. That’s all you need to venture out on your own.

Stewardship means picking the right Agora, resolving squabbles, and keeping out the bad apples — both figuratively and literally. Custodianship is about cleaning up before and after. In a nutshell it’s about keeping the market not only free but also tidy.

The big players of the internet age all do these jobs to various degrees. For the most part the big board has been replaced by individualized little boards, like what Bloomberg did to the NYSE. App stores, media feeds, chats, timelines, search results for taxicabs or anything.

The interaction part has not been disintermediated as much as it has been automated. Matching seekers with offerers is done at scale, based on algorithms that are tweaked all the time with very limited human intervention, simply because humans are slow (and expensive).

The intervention usually comes when algorithms have to be tweaked to improve business or engagement targets. Custodianship is a labor-intensive exercise and the focus of public criticism because the cleanup efforts are either too lenient or too strict.


Which finally gets me to the point I promised to tackle at the outset: how mechanisms vs processes have influenced the balance btw enterprises and markets, how enterprises have taken over markets and sucked up the rents that were previously dispersed to the participants and…

What “blockchain” can do to help!! Thanks to all who are still awake.


From Ronald Coase we know that markets and enterprises are competing modes of allocating scare resources to hopefully beneficial outcomes. My point so far was that there is a significant overlap in these modes, and that enterprises tend to swallow markets, like paper wraps rock.

So let’s talk a bit about what sets them apart. For one, markets have no till, there’s no cash which can attract grabby hands. All cash flows immediately go to market participants. For two, enterprises as legal and accounting entities have outer boundaries. Markets do not.

Also, even both are intertemporal things, we associate markets with “spot” and enterprises with “continuous”. Markets have mechanisms, enterprises have processes. This is partly due to the economic framing of a transaction as a simultaneous, instantaneous exchange of things.


It might be worth defining the difference between process and mechanism, seeing that they are frequently used synonymously. A process is a step-by-step instruction set designed for human intervention. A mechanism is designed to operate autonomously, even to keep humans at bay.

Processes are too “rigid”, “inflexible” or “bureaucratic” if they don’t allow a certain amount of human autonomy and leeway. Mechanisms are “unpredictable” or “rigged” if they can be gamed.

This is admittedly somewhat of a distinction in degree, and we are continuously moving in a direction where processes become more and more like mechanisms until they reach full autonomy.

The autonomous driving folks I worked with hated the word “autonomous”, as every time they came up with a nifty improvement someone cried that “this isn’t truly autonomous” because the car still had a steering wheel and, more importantly, a brake pedal.

It is not an accident that the NYSE moved from a “market” to an “enterprise” mode at a time when it was beleaguered by more technologically advanced exchanges and had high capital requirements to catch up and move from human-centric open outcry to electronic trading.


So if we want to apply a Coase/Williamson/Hart transactional/contracting framework to the P2P age (and I believe it is still by miles the best we have), we have to contend with some recent developments which run counter to the predictions based on opportunism and hold-up.

1. Buttonwood-type markets are subject to being swallowed by enterprises if 
- competence at operational efficiency beats competence at stewardship and custodianship
- capital requirements favor a for-profit setup
- individual transactions are too trivial for payments

2. Even in the presence of holdup, asset investments and complex contracting, enterprises are divesting themselves of their supply chains, viz. automotive and smart phones. I discussed this here.

So the challenge for blockchain aficionados of the institutional sort and blockchain governance scholars: How will any of this change if we acquire an accounting system, a process platform, and a unit of account that spans across markets?

The original tweetstorm appeared on @ecoinomia. The text has been slightly fixed and it might still evolve into a real essay. Oliver Beige is an industrial engineer turned economist (PhD Berkeley, MBA Illinois, MSIE Karlsruhe) who is focusing on how technologies like machine learning or blockchain might change the value creation between markets and enterprises.