Good blockchain governance is at least a decade away.

Elaine Ou writes that a blockchain’s lack of a centralized governance mechanism is its main feature. Nobody controls Bitcoin, its core governance decisions were locked-in since v0.1, and that’s a plus. Other cryptocurrencies that experiment with more centralized decision making authorities miss the point: it’s the anarchy after the initial rules are set that’s the true feature. The post hit upon something I’ve been thinking about for a while, and I wish to push back on this view.

Contrary to the belief that blockchain’s lack of governance mechanisms is its core feature, I believe that blockchain governance needs decades of iteration, and our understanding of governance remains crude. To quote Matt Levine:

I am starting to think that the right way to think about blockchain and cryptocurrencies and tokens might be the way you’d think about stocks, if stocks had just been invented. Here in 2017, it is uncontroversial to say that the invention of limited-liability joint stock companies utterly changed how humans organized their economic activities, and had huge impacts on economic productivity and on society as a whole. But it is also fair to say that the first few hundred years of stocks were mostly fraud and irrational speculation. Similarly with cryptocurrencies and initial coin offerings, it is possible both that proponents’ grandiose claims about how they will transform finance and society will turn out to be true, and that most crypto stuff in our lifetimes will be nonsense.

I think this view is correct. We’re going to have to spend at least a few decades iterating upon the governance mechanisms for blockchains, just as we spent a few centuries iterating upon the governance mechanisms for corporations.

A brief history of corporate governance

Why did it take a century to perfect corporate governance? When you read any history of joint-stock corporations, they almost universally start with the Dutch East India Company (VOC, in Dutch initialization). The VOC was the first perpetual joint-stock corporation of a form that we would recognize as “equity” today, but it wasn’t the first equity-issued corporation. Prior to the VOC, corporations were formed in Antwerp via equity issuances for a single purpose: to finance a single ship’s voyage. When (if) the single voyage completed successfully, the corporation was liquidated and all equity holders received a share of the profits. The VOC was instead formed for a renewable 21 year duration, in response to the growing power of the British East India Company, which was a royally-chartered corporation from a geopolitical rival (and thus a powerful monopoly that threatened this entire ecosystem) with a huge capital base.

We speak of the VOC as the “first modern corporation” but it was instead the culmination of decades of corporate experimentation until the veterans of multiple corporate expeditions pooled their resources (and knowledge) into the “final” incarnation. It was preceded by a long culture of single-ship equity issuances. This view is consistent with the idea that it’s not the first movers you care about, it’s the last movers, and the VOC was the last mover, with one key governance twist that changed everything.

Why was perpetual incorporation so important? As I’m not a corporate historian, it’s hard for me to say for sure, but I’ll hazard some speculations. First, the proto-corporations raised capital frequently. I don’t know if the Antwerp exchanges charged a suspiciously fixed 7% for capital raises, but it can’t have been cheap. Second, corporations typically funded a single ship, which doesn’t bode well for carrying over institutional knowledge to the next. I believe that the key enemy to a culture of innovation is complacency. A perpetual corporation retains knowledge even if individual captains retire wealthy after a single successful journey, which didn’t permit much institutional knowledge to carry over to the next voyage.

It’s also interesting to note that corporations had interesting skirmishes with the legal system. For instance, when the VOC captured a Portuguese ship in a sudden burst of entrepreneurship 1603, nobody quite knew what to do. Sure, the Dutch were at war with Portugal. But Mennonite Dutchmen objected, and launched the first activist shareholder campaign. This objection raised the question of what rights, exactly, shareholders had. I’m not sure there was any satisfying resolution there, but we know what today’s institutions would say about that: modern shareholders have literal voting rights. Back then, the board was made up of various politically powerful stakeholders — basically just representatives of a few influential Dutch city governments. The concept of the board having a fiduciary duty to shareholders was still centuries away.

I don’t want to overly simplify things: perpetual corporate personhood wasn’t a “one weird trick” that solved everything. The regular equity turnover meant that the regular capital raises were regular referendum’s on executive officers. When the VOC dropped that, they traded away a lot of existing institutional governance for it. Creating different institutions that would bring back the good features of regular checks and balances while retaining the institutional benefits of long-lived executive ownership clearly took humanity a while longer.

Today, the modern corporation is very well mapped out. We’ve evolved norms that include the idea that equity holders hold residual financial claims; that courts should largely stay out of litigating decisions by executive officers; that boards should be elected by shareholders to hold executive officers accountable; and that debt holders should have seniority. And beyond that, we’ve established an ecosystem of institutions that ensure good behavior on all parties—from regulatory oversight by the SEC, to financial oversight by activist funds and private equity raiders. I don’t want to make it seem as if these concepts were deeply theorized before they were implemented. Most academic financial theory is retrospective, for example, the Modigliani-Miller theorem came along to explain capital structures long after we established it.

I also don’t want to paint a misleading history of straightforward linear growth in societal institutions. By no means do I think the corporation has been perfected. E. Glen Weyl and Eric Posner do raise legitimate questions about whether the rise of pooled index funds break some of our existing institutional checks. I remain fascinated by Matt Bruenig’s arguments that we’re building the institutions that will finally allow socialist ownership of the means of production. And the Chinese Communist Party might be the most innovative capitalist of them all, for example by rotating the heads of competing state-owned firms with no warning.

I want to make clear that evolving the right mix of institutions - in the case of well functioning capital markets - has taken somewhere between one and two centuries. Expecting cryptocurrencies to do the same in one decade is too ambitious. And we will have to do so with lots of random experimentation in governance structures, which will one day be explained with sound theories.

Blockchain Governance

I believe that Bitcoin (and other contemporary Blockchains) suffer from a litany of design failures that will doom each of them individually. Some of these doomed coins may still persist in value for decades. Nevertheless, the core Blockchain concept has value, and will have a lasting impact. I will focus on Bitcoin and Ethereum in my criticisms, although I am not optimistic about other coins and chains. First of all, it’s not correct to say that Bitcoin is anarchic, and has no governance mechanism. It simply rests all governance in the hands of the mining mob. I don’t mean that the miners hold all the power. But the miners are the only party that can make protocol changes. The elevation of mining interests in the governance mechanism at the expense of the other stakeholders, long-holding speculators and transacters (those who wish to use it as an efficient medium of exchange but want no exposure to the underlying asset class), is probably not optimal.

Because Bitcoin has a diverse set of stakeholders, for Bitcoin to function well at anything, it has to align their varied incentives, not randomly choose one faction as supreme. The theory that miners will do the right thing because they wish to rationally maximize the value of the asset, and that the asset price depends on increasing the transaction volume smacks of flawed Coasean reasoning[1].

But, even if we left miners as the sole authority, we need better governance than just bruteforced hash-rate fights. And as the SegWit negotations demonstrate, miners still need to find ways to build common knowledge around their common intentions to coordinate, and that requires governance, if only for ritualistic significance to build that common knowledge for coordination purposes.

Unlike Bitcoin, Ethereum has a governance mechanism that works better in the benevolent monarch of Vitalik Buterin. The Ethereum DAO hard fork was helped by the existence of a monarch (again, for coordination purposes, even if he has become a monarch by the mining-supremacy “constitution”). It is a proto-form of governance, superior to the raw Athenian mob democracy of Bitcoin, but still nowhere near what a good system of institutional checks can deliver.

For blockchains to succeed, we’ll have to see chains that build institutions that respect all the stakeholders, and have clear views as to how they are balanced. It’s not that we need “stakeholder capitalism”. It’s fine to have a very opinionated governance design that only protects one faction (a board with a fiduciary duty to maximize the value of the coin?). But having those views is important. These chains will have to have clear views as to their long term market value, if not in absolute terms, at least by clearly stating equations that they expect to hold for the long term coin value with the transaction rate: perhaps, even encoding rules based monetary policy directly into the protocol. I imagine that the required governance institutions will be very different based on what the coins actually wish to deliver. Tokens that are tradeable for goods and services will probably require very legalistic institutions that are similar to corporate governance. And tokens that seek to be replacement currencies will probably need more political institutions that can perform the common knowledge tasks required to sustain shared belief in a currency.

I don’t want to say that Ethereum requires Delaware corporate law and Bitcoin needs the Federal Reserve. The future is not simply the past on repeat. But my point is that the design space for these institutions is vast, and advancements in our understanding will take at least a couple decades. I remain optimistic.

Acknowledgements

I’d like to thank Emin Gun Sirer for explaining the Ethereum DAO fork, and the power held by various factions in the Bitcoin and Ethereum world. I’d like to also thank Dan Wang, Kevin Kwok, and Dave Moore for helpful comments on multiple drafts.

Footnotes

[1] The Modigliani-Miller theorem is frequently mis-criticized on the basis of its “laughable assumptions”. Briefly, it states that, in the absence of agency costs, there is no difference in capital structure between raising equity and raising debt. But the real point of the theorem is a proof-by-contradiction exercise which reveals that the whole song and dance about complex equity structure is in order to solve agency costs. If we didn’t have agency costs we wouldn’t need all this complication! But empirically, when we look around, we do see complex institutions that care very deeply about fine tuning capital structure. Thus the agency costs must be real. Coasean theorems also have this unintuitive structure, and these arguments are 40% of the fun of going to law school.