Streetwise Professor

Imagine my great surprise at reading a Bloomberg piece titled: “Blockchain, Once Seen as a Corporate Cure-All, Suffers Slowdown.”

That was sarcasm, by the way.  I’ve long and publicly expressed my skepticism that blockchain will have revolutionary effects, at least in the near to medium term.  In my public speaking on the topic, I’ve explored the implications of three basic observations.  First, that blockchain is basically a way of sharing/communicating information, which can in turn be put to various uses.  Second, there alternative ways of sharing/communicating information, with different costs and benefits.  And third, it is necessary to distinguish between sharing information within an organization and between organizations.

Much of the hype about blockchain relates to the potential benefits of more efficient sharing and validation of information.  But this does not address the issue of whether blockchain does this more efficiently than alternative means of sharing/communicating/validating.  As in all institutional/technology issues, a comparison of alternatives is necessary.  This comparison has been sadly lacking in public discussions of the potential for blockchain, beyond incantations about blockchain eliminating the need for trusted third parties which is (a) often untrue (in part because trusted parties may be required to enter information into a blockchain, and (b) is not necessarily a feature, because trusted third parties may be able to operate more efficiently than consensus based systems employed on a blockchain.

The most developed implementation of blockchain (Bitcoin) involves very large cost to solve a particular problem that (a) is unique to cryptocurrency, and (b) is not necessarily important in other contexts–namely, the double spend problem in crypto.  Maybe blockchain is the best way to solve that particular problem (which itself begs the question of whether cryptocurrency`is an efficient solution to any economic problem), but that doesn’t mean that it will be a more efficient way of solving the myriad types of opportunism, fraud, and deceit that plague other kinds of transactions.  Double spend is not the alpha and omega of transactional challenges.  Indeed, it might be one of the most trivial.

Thinking in Williamsonian transaction cost terms, where the transaction is the unit of analysis, transactions are highly diverse.  Different kinds of transactions are vulnerable to different kinds of information and opportunism problems, meaning that customized blockchain approaches are likely necessary.  One likely cause for the waning enthusiasm mentioned in the Bloomberg article is that people are coming to the recognition that customization is not easy, and it may not be worth the candle, compared to other ways of addressing the same issues.  Relatedly, customization makes it harder to exploit scale economies, and recognition of this is likely to be making initially enthusiastic commercial users less keen on the idea: that is, it may be possible to use blockchain in many settings, but it may not be cost-effective to do so.

The siloed vs. cooperative divide is also likely to be extremely important, and the Bloomberg article mentions that issue a couple of times.  The blockchain initiatives that do seem to have been implemented, at least to some degree, as with Maersk in container shipping or Cargill with turkeys, are intra-firm endeavors that do not require coordination and cooperation across firms, and can exploit the governance structure that a firm has in place.  Many of the other proposed uses–for instance, in trade finance, or in commodity trading, both of which require myriad parties in a single transaction to communicate information among one another–are inherently multilateral.

This creates all sorts of challenges.  How can commercial rivals cooperate?  How are the gains from cooperation divided?–this is a problem even when participants supply complementary services, such as a trading firm, banks providing trade finance, and the buyer and seller of a commodity.  As oil unitization has shown, battles over dividing the gains from cooperation can dissipate much of those gains.  Who gets to see what information?  Who makes the rules?  How?  How are they enforced? What is the governance structure?  How is free riding prevented?  Who pays?

Ironically, where the gains from cooperation are seemingly biggest–where there are large numbers of potential participants–is exactly where the problems of coordination, negotiation, and agreement are likely to be most daunting.

I’ve drawn the analogy between these cooperative blockchain endeavors and commodity exchanges, which (as I showed in a 1995 JLS paper) were formed primarily as ways to reduce transactions costs via cooperative rule making and enforcement.  The old paper shows that exchanges faced serious obstacles in achieving the gains from cooperation, and often failed to do so.  Don’t expect blockchain to be any different, especially given the greater complexity of the transactional problems that it is being proposed as a fix.

Thus, I am not surprised to read things like this:

“The expectation was we’d quickly find use cases,” Magnus Haglind, Nasdaq’s senior vice president and head of product management for market technology, said in an interview. “But introducing new technologies requires broad collaboration with industry participants, and it all takes time.”

or this:

Most blockchains also can’t yet handle a large volume of transactions — a must-have for major corporations. And they only shine in certain types of use cases, typically where companies collaborate on projects. But because different businesses have to share the same blockchain, it can be a challenge to agree on technology and how to adopt it.

One of my favorite illustrations of the hype outstripping the reality is the endeavor launched with much fanfare in the cotton market, where IBM and The Seam announced an endeavor to use the blockchain to revolutionize the cotton supply chain.   It’s been almost two years, and after the initial press releases, it’s devilish hard to find any mention of the project, let alone any indication that it will go into operation anytime soon.

Read the Bloomberg article and you’ll have a better understanding of R3’s announcement of an IPO–and that they might have missed their opportunity.

In 2017 and a little before, Blockchain was a brand new shiny hammer.  People have been looking everywhere for nails to pound with it, and spending a lot of money in the effort.  But they’re finding that many transactional problems aren’t nails, that there are other hammers that might do the job better, and there are other problems that require many parties to agree on just how the hammer is to be used and by whom.  Given this, it is not surprising that the euphoria is fading fast.  The main question that remains is in what shrunken domain will blockchain actually be employed, and when.  My guess is that the domain will be relatively small, and the time until employment will be pretty long.

Ethereum wunderkind Vitalik Buterin delivered a rant against centralized crypto exchanges:

“I definitely personally hope centralized exchanges burn in hell as much as possible,” Buterin said speaking to TechCrunch.

When bitcoin, the original cryptocurrency, was founded in 2008 by the anonymous Satoshi Nakomoto, the point was to create a decentralized financial future that renders middlemen useless. Nearly 10 years later, the centralized exchanges — those folks sitting in the middle of buyers and sellers — are among the most powerful players in the market for digital currencies such as ethereum and bitcoin.

Bloomberg News estimates they brought in $3 million a day last year. And exchanges such as Gemini and Coinbase are expanding at a clip, bringing on talent from Wall Street.

“It’s hard to ignore the irony that an asset created to allow decentralization is currently almost completely traded on centralized exchanges,” Peter Johnson, a vice president at Jump Ventures, said in an interview. Buterin, however, wants the crypto community to focus more on decentralization so that cryptos can more frequently trade peer-to-peer. Buterin’s remarks come as so-called decentralized exchange gain more attention.

Like many of his arrogant ilk, Buterin ignores the lesson of Chesterton’s fence: why does this thing you do not like and do not understand exist?

Yes, blockchain and cryptocurrencies allow peer-to-peer transactions.  They were largely designed to facilitate such transactions.  For some, the motivation is ideological: an anarchic belief in radical decentralization, and a deep distrust of centralized institutions.

But just because blockchain and related technologies reduce the costs of peer-to-peer transactions, doesn’t mean that such transactions are cheaper than centralized trading on exchanges.  Transacting requires finding a counterparty.  It requires negotiating a price (for a standardized thing, like a Bitcoin–negotiations of other terms for more complex things).  Negotiating a price is costly when information about value is diffuse, so in a decentralized setting not only is it necessary to search for counterparties, it is advantageous to search for information about prices to (a) find the best price, and (b) to be able to negotiate with better information about value .

Centralization reduces the cost of finding a counterparty.  It enhances competition, which tends to reduce bargaining costs.  It leads to better and more symmetric information about prices, which also tends to reduce bargaining costs.  Further, centralized markets can support specialized intermediaries–market makers–who specialize in smoothing out idiosyncratic temporal imbalances in buy and sell order flow, which further reduces trading costs.

Because of these features, centralized trading is frequently an emergent outcome of individual decisions, and one that economizes on transactions costs.  This is clearly what is happening in crypto world.  Indeed, the main puzzle at present is why there are so many exchanges.  The centripetal forces of liquidity will likely result in a huge consolidation in this space.

Buterin and others are attempting to find ways of mitigating some of the disadvantages of bilateral trading (bilateral just being another, more conventional, way of saying “peer-to-peer”).  Reducing the ways of finding people who want to take the other side of a transaction, for example.  But I am highly skeptical that these measures will overcome the inherent advantages of centralizing trading of homogeneous things that large numbers of people want to buy and sell pretty much 24/7, to the point that peer-to-peer will supplant centralized trading.  Buterin can rant all he wants, but centralization is here to stay, and if anything, this segment of the market will become more centralized.

Buterin’s error is seemingly the opposite of those who bewail the lack of centralization in some markets, e.g., those who want to make swaps trading more centralized and who rail against bilateral OTC transactions, but it is really the same mistake. Those who see too much centralization in some markets, and those who see too little in others, fail to recognize that trading mechanisms are emergent orders that develop diverse niches to accommodate the fact that transactors and transactions are heterogeneous.  Centralization is efficient for some transactors and transactions: bilateral/OTC for others.  That’s why we see both.

(This is a point I made at a Platts blockchain conference in November, BTW.  The theme of my talk was where decentralization can work, and where it is likely inefficient.  Trading of standardized instruments was one of the main cases I discussed.)

Alas, the ignorance of techno-geniuses is not limited to trading mechanisms.  One of the supposed benefits of blockchain that is that it allows the ownership of anything–a painting, a house, you name it–to be divided into shares, with the fractional interests recorded in an immutable register, and traded peer-to-peer.  That is, block chain facilitates equitization of assets.  A breakthrough!

Uhm, not really. The benefits of equitizing assets and risks has been long, long understood by economists.  In particular, it has long been understood that equitization facilitates more efficient risk sharing.

But long ago, economists also recognized that despite these apparent benefits, in fact very few assets and risks are equitized.  A vast literature has come up with explanations why.  Information and incentive problems–moral hazard and adverse selection–are notable among these.  A prosaic example: If I sell off shares in my car, what incentive do I have to maintain it properly and to economize on wear and tear and to reduce the probability of theft?  Who pays for maintenance? Who decides on what maintenance is needed?  When I sell the shares, I am likely to have better information about the value and condition of the vehicle, which would subject the buyers to an adverse selection problem, meaning that I am likely to get a low price for the shares–so why bother selling them?  There are other transactions cost problems associated with measurement (who verifies exactly what the asset is?) and opportunism and governance and control.   Related to the centralized trading point, if an asset is highly idiosyncratic/unstandardized, the desire to trade fractional shares will be small.

A potentially slightly cheaper way of recording and transferring fractional ownership does not address these far, far, far more fundamental impediments to equitizing (or should I say, “tokenizing”?) assets and risks.  But the coder geniuses miss the forest for the trees.  They see the issue that their technology can address, and think that it will be revolutionary, only because they do not understand the broader economic issues in play, and therefore think everyone born before them or who does not code must be an idiot.

No, not really.  They are looking at the capillaries, and missing the heart, veins, and arteries.

It reminds me of the Mark Twain quote: “When I was a boy of fourteen, my father was so ignorant I could hardly stand to have the old man around. But when I got to be twenty-one, I was astonished at how much he had learned in seven years.”  Except seven years haven’t passed for the Buterins of the world, and frankly, I seriously doubt that they will.  Instead, they inhabit a techno-Groundhog Day.

All of this is symptomatic of blockchain hype and froth.  There is an indication that we have reached peak hype.  R3, a bank-led blockchain consortium, is contemplating an IPO.  To me this is a signal that those on the inside of blockchain development, especially in the area where its benefits have been particularly hyped (finance/payments/settlement/fintech) understand that the reality will never match the hyperbole, so it’s best to sell out while hyperbole reigns supreme.  (Yes, they claim that they are being approached by those looking to buy the whole thing, but take that with a big grain of salt–I view it merely as part of the sales pitch.  “This is a hot little property right now.  Better get in before someone snatches it away.”)

In brief: don’t be the greater fool.

I think that blockchain and DLT will have some viable commercial applications.  But I am highly confident that they will not be nearly as revolutionary as the True Believers claim.  This is in large part due to the fact that it is clear that the True Believers have an extremely narrow, blinkered understanding of the broader economic issues associated with transacting, ownership, risk transfer, incentives, and governance.  Blockchain may address some issues, but many–if not most–of these issues are secondary or tertiary, not fundamental.  Some things are done more efficiently in a centralized fashion–the trading of standardized instruments being one.  Some things are not equitized/tokenized not because it is technically infeasible/prohibitively costly to issue and record fractional interests, but because fractional ownership entails substantial incentive and information problems.

So don’t believe the hype.  And take a pass on those R3 shares, if they do come to market.

Addendum: the dominance of crypto exchanges is even more remarkable, given how they, well, pretty much suck.  They are hardly comparable to modern futures or equities trading exchanges.  Yet people still strongly prefer to trade on rather clunky platforms with major potential security issues where you can’t easily convert digital into fiat currency and which are likely rife with manipulation than peer-to-peer.  That tells you something.

This article isn’t explicitly about blockchain, but in a way it is. It describes the halting progress of the LME’s LMEShield digital warehouse receipt system, and ascribes its problems to the Alphonse-Gaston problem:

“It’s all about liquidity and a tipping point,” said an executive with a warehouse company.

“If Party A is using the system, they can’t trade if Party B is not using the system. It’s a redundant system until market masses are using it.”

After you, Alphonse. No! After you, Gaston! I insist! (Warning! Francophobia and Mexaphobia at the link!)

In essence, LMEshield is attempting to perform one of the same functions as blockchain–providing a secure, digital record of ownership. LMEshield’s technological implementation is not blockchain or distributed ledger, per se. It is a permissioned network operated by a trusted party, and does not employ a distributed ledger, rather than being an unpermissioned network not involving a trusted party, run on a DL.

But it is not that technological difference that is causing the problem: it is the challenge of coordinating the adoption of the use of the system. This coordination problem exists here even though there is an entity–the LME–that has an incentive to promote adoption and build a critical mass so that tipping takes over. Despite a promoter, the virtuous cycle has not taken hold. The adoption problem is even more challenging without a promoter.

This problem will be present in all attempts to create a secure digital record of ownership, regardless of the technology used to achieve this goal.* There is more than one technology to skin this cat, and it is not the technology that will in the end determine whether the cat gets skinned: it is the ability to overcome the coordination problem.**

And as I’ve noted before, if tipping does occur, that just creates another conundrum: tipping effectively creates a natural monopoly (or at best a small numbers natural oligopoly), which raises questions of market power, rent seeking, and governance.

Bitcoin world is providing an illustration of the challenges of governance (as well as raising questions about the scalability of blockchain). Block size has become a big constraint on the capacity to process transactions, leading to a spike in transactions charges and long lags in processing transactions. There are basically two proposals to address this issue: expand the size of blocks, or allow some processing to occur off the blockchain. This has divided Bitcoin world into camps, and raises the possibility that if the dispute is not resolved Bitcoin could experience a hard fork (i.e., split into two or more incompatible networks). Miners (mainly Chinese) want to expand block size: Core (the developers who maintain the software) want to externalize some processing. Both sides are talking their book–go figure–which illustrates that distributive considerations and politicking, rather than efficiency, will have an outsized effect on the outcome.

Keep both LMEshield and the Bitcoin block size debate in mind when somebody offers you pie-in-the-sky, techno-evangalist predictions of how blockchain/DLT Is Going to Change the World. It’s not the technology alone that matters. Indeed, that may be one of the least challenging issues.

Also keep in mind that there is nothing new under the sun. The functions that blockchain/DLT are intended to–or dreamed to–perform are inherent in all transactional settings, including in particular financial and commodity transactions. Blockchain/DLT is a different way of skinning the cat, but the cat has been skinned one way or ‘tother since the dawn of these markets. Maybe in some cases, blockchain/DLT will do it more efficiently. Maybe elements of blockchain/DLT will be blended into more traditional ways of performing these functions. Maybe some applications will prove resistant to blockchain/DLT.

But the crucial thing to keep in mind is that blockchain/DLT will not banish the fundamental challenges of coordinated adoption and governance of a system that scales. And note that if it doesn’t scale, it won’t replace existing systems (which do), and if it does scale, it will pose the same organization, governance, and market power issues that legacy systems do.

*There is no guarantee that DLT is even a technological advance on existing technology. As I discuss further on, some implementations (notably Bitcoin) have exhibited severe problems in scaling. If it don’t scale, it will fail.

**I own a cat. I like my cat. I like cats generally. If colloquialisms offend, this is not the place for you. Particularly colorful if somewhat archaic American colloquialisms that I learned at my grandfather’s knee. Alphonse and Gaston is also something I picked up from my grandfather. Today it would cause an outbreak of fainting, shrieking, and pearl clutching (though maybe not–after all, the characters are Europeans), but if you can’t separate the basic comedic idea from what was acceptable in 1903 (the year of my grandfather’s birth, as well as the date of that clip) but isn’t acceptable today, you’re the one with the issues.