Centralization is A Cost
Decentralization is the Default, Centralization is a Cost
As Erik Voorhees is fond of saying “it’s faster to FedEx an anvil to Japan than send a wire there.” Despite the invention of the internet, wire transfers take about as much time as they did 100 years ago. Why?
Technological improvements no longer speed up our financial system because its efficiency is not limited by technology. The financial system’s structure limits its efficiency. The reason for this inefficient structure is to mitigate risks inherent to centralization.
The default state of human interaction, financial or otherwise, is decentralized. Introducing a centralized intermediary with the power to interfere introduces risk. If you were a medieval merchant, why would you keep your money with a banking house? They could lose it, especially when they were openly using it to lend to other people. Even if they didn’t lose the funds they could steal them, or the banking house could burn down, etc.
On the other hand, if you were in London and wanted to take advantage of a deal opportunity in Paris your “decentralized” method of closing that transaction would require sending a caravan loaded with your gold and goods on a boat across a channel and then through questionably safe country roads in a foreign country. During that time, you’d just be hoping your counterparty is completing their side of the transaction. With a banking house large enough to hold respected branches in both cities, however, that transaction could be achieved with a swift messenger and a few strokes of the pen.
Sure, you would rather that the laws of physics permitted you to simultaneously teleport your goods with your counterparty according to unbreakable, pre-set, agreed rules, but that would be science fiction. Given that, it makes sense to accept the risks inherent in centralization to make that kind of speed of transaction possible.
Historically, we rightly accepted the risks of large centralized institutions with the power (and responsibility) of keeping track of who owns what within their own ledgers. Accepting that centralization made transfer of ownership and value across long distances and across time so much more efficient that it was worth the risk that came with centralized ledgers.
From Centralization Risk to Inefficiency and Cost
The modern financial system is a marvel of organically developed human cooperation. The history of that system’s development has been the conversion of centralization risk to overhead costs and inefficiencies. We prefer a cumbersome system if that reduces the risk of total loss from malevolent or incompetent actors.
This is so inherent to the history of finance that it’s even reflected in its architecture. If you were in a town in early America, the bank might have been the most expensively built, nicest structure in your town (perhaps excluding the Church). This isn’t about flaunting wealth; it’s more about the fact that you’re more likely to leave your life savings in the hands of someone who has spent 1000x those savings on a building. It would be hard for an itinerant banker in a covered wagon to compete. Banks are, by design, expensive institutions.
The existence of a trusted institution with pooled capital from everyone in town opens up the myriad wonders of capital formation. As long as you agree to let the bank hold your savings, and to take charge of keeping the ledger of value transfers to and from your savings, you get to access those wonders. You can now pull forward future income to get a mortgage for your farmstead, or a loan to start a small brick making operation. In short, banks make a much better life possible for you.
Unfortunately, the only way to do that historically has been by subjecting your entire financial life to the risks of centralized intermediaries.
There is a problem here. Not everyone is an expert in monitoring the risk of centralized financial entities. And not everyone should be. Any society that can only function properly if every person is an expert in assessing the risks of a financial institution is a failed society. This has led to systematic problems that continuously rear up throughout financial history.
To give an example, in the banking panic of 1907, a particular mismanaged bank began to collapse due to a failed attempt to short squeeze shorts that didn’t exist (for the sake of convenience, we will consider trust companies as largely the same thing as banks in this description). It hardly seems fair that anyone who deposited at that institution should lose their life savings in a bank run– could they reasonably be expected to have anticipated that?
The same lack of universal expertise around risk management also meant that other people became concerned about banks in general, causing a run on all the banks. This could even bring down perfectly solvent banks. In that particular instance J.P. Morgan famously put up a large part of his personal fortune to recapitalize banks (at extremely attractive terms), in addition to rather dramatically forcing many of the trusts into a deal to support each other’s liquidity.
So, not only did the lack of perfect knowledge and expertise in every bank depositor mean that some of them were exposed to mismanagement they didn’t anticipate, it also nearly led to the destruction of many other banks that were perfectly solvent (if not 100% liquid).
Episodes like this helped to inspire the creation of the Federal Deposit Insurance Corporation in 1933, a government entity that acts as a lender of last resort so that individual Americans don’t have to be personally responsible for judging the risk of their banks. Centralization presented a risk, a new additional expensive institution was created to mitigate that risk (replacing a risk with a cost).
But the creation of new, expensive, centralized entities isn’t the only way centralization risk has been mitigated. A large part of the mitigation comes, ironically, from a sort of “artificial regulatory decentralization.” Legally enforcing delays and multi-step interactions and checks to achieve fairly simple transactions can allow for a system where so many separate entities have to touch a transaction that we end up less exposed to a single malevolent or incompetent actor.
In our original example of the medieval merchant, the banking house could “steal” the merchant’s goods by simply changing the ledger to put it into someone else’s account. If, however, close to a dozen separate entities were required to achieve a transaction, such a clerical theft would be significantly more difficult (the same applies around the risk of loss due to a clerical error rather than crime).
This is the higher level goal of regulations: splitting up brokers, custodians, transfer agents, etc. and legally enforcing the requirement to use all of them in a transaction–as well as onerous monitoring and licensing for each of those entities. It creates a more resilient system by diversifying the centralization risk — but at the cost of inefficiency.
Above is a diagram of a simple securities transaction. The system is arbitrarily and artificially complex, expensive, and inefficient–but the risks inherent to centralization have been mitigated.
This is the kind of cumbersome, byzantine clerical and regulatory infrastructure that makes it impossible for us to speed up the financial system or make it meaningfully more efficient. And the above diagram only breaks down the non-public entities involved in such a transaction. The largest and best functioning capital markets in the world are in the United States, where there are no fewer than 18 federal agencies regulating financial markets, each enforcing their own set of required practices and employing their own bureaucratic armies (and even that number doesn’t cover the additional state-level agencies that would be involved).
The history of the development of the modern financial system is one of catastrophes caused by centralized risk, and then ever-more-byzantine new regulatory requirements to mitigate that risk. Even despite all of the protections we’ve developed over the centuries, we still had the 2008 financial crisis. And in response, we got the 2,300 page Dodd-Frank act, which has led to the creation of an entire business model of regulatory consultant firms paid millions of dollars to help financial institutions understand whether they are compliant with a law nobody understands.
Centralization risk is converted to cost and inefficiency through onerous regulation, creating a better financial system. The better financial system finds new surface area for bad outcomes from centralization risk. The cycle goes on.
A Return to Decentralization
The game-changing reality of blockchains is that they actually allow the merchant in the initial example to live in the sci-fi world of simultaneously teleporting his goods with his counterparty according to unbreakable, pre-set, agreed rules. And, they can do this without the introduction of centralization risk through an intermediary.
At a fundamental level, what is possible has changed, and the risks of centralization that our financial system has developed to mitigate can now be fully obviated.
We’ve hit a limit on how efficient we can make our financial system unless we take advantage of a technology that changes the rules and risks around which that system was built. Blockchain is the only way to actually break through to a step-change improvement. Otherwise we will only be making marginal improvements, or picking different spots on the tradeoff between inefficiencies and centralization risk.
With that said, the idea that blockchains will achieve a full break from the past, and that the future blockchain based financial system will emerge from an alien and unrelated evolutionary line, is misguided (although it is popular with crypto ideologues). This is not how new technologies are adopted.
Standardization like the Gauge Act of 1846 didn’t just help unlock the value of the new technology of rail cars by allowing for united rail systems, they also standardized them to the width of Roman Chariot. Blockchains will not onboard the global financial system until regulation begins to catch up, regulation won’t meaningfully change until the largest pools of capital in the world begin to adopt blockchains, and the largest pools of capital in the world will only lead regulation by increments.
And, yes, many of the institutions and entities required by our current system will evaporate. Many more will simply evolve. For the same reason that a society in which everyone must be an expert in assessing banking risk in order to function properly is a failed society, a society in which everyone must be able to assess smart contract risk or practice expert opsec around seed phrases is also a failed society.
We will automate away much of the investigatory and enforcement needs of regulatory entities, and also many of the arbitrary bureaucratic entities built to reduce centralization risk, but folks will still need to be told by some entity they can trust about what is and isn’t safe.
Why?
Unlike most articles arguing that blockchains are the infrastructure of the future financial system, this one is intended to be ideologically agnostic. Do I think that blockchains will lead to a more open, free, accessible, fair and transparent financial system? Yes (I in part quoted Erik Voorhees to open this so I could circle back and point the reader to his recent Keynote at Permissionless II which eloquently states the ideological case). I also think that we will not bring about that future financial system through ideological conversion.
Most people don’t care how the financial system works now, and most people won’t care how it works in the future when it all runs on blockchains. But, that new financial system will be significantly more efficient and resilient. And, the institutions and pools of capital that constitute much of the modern financial system today care quite a lot about efficiency and resilience.
