Elusive Symmetries
a response to Jill Carlson, and a meditation on money.
Last week, Jill Carlson penned an editorial at Coindesk with the deliciously provocative title, Cryptocurrency is Most Useful for Breaking Laws and Social Constructs. And provoke it did, sending the tetchier cohorts of cryptotwitter into a frenzy. Why would you give us a bad name!, It’s going to be useful for everything!, and so on and so forth, and everything in between. I’ve heard enough conference talks where people say things like, “if you decentralise the law, then justice becomes code,” to find the frank realism in Jill’s article highly refreshing. If you haven’t already, please read it before reading this.
To briskly recap, “cryptocurrency does not solve mainstream problems.” They have low throughput, slow settlement, and high overhead. These are not flaws but deliberate trade-offs to secure censorship resistance, making such systems primarily attractive to the censored, who tend to be rather outside the mainstream. She didn’t go very far out her way in the article to say that this isn’t a bad thing, per se — probably because she didn’t anticipate the flurry of cry-babying that followed exactly such a misunderstanding — but that is worth stressing here too, as the implication was obvious to grownups. I’m not sure I disagree with much of what Jill said. Rather, I take issue with what she didn’t say. There is an elusive symmetry to cryptographically secure systems of which I believe Jill treated only the plain vanilla half. I want to consider the half we don’t talk about in polite company.
Even the most straightforward public key encryption systems present a conceptual symmetry that is often forgotten. By signing a message with a recipient’s public key, we ensure only their private key can decrypt it. Alice knows the message is to Bob and only Bob. This is probably what first comes to mind when we contemplate ‘encrypting messages’. But leaving it at that is somewhat narcissistic. Alice ensures the security she cares about the most, but what about Bob? By signing it with our private key, we ensure in addition that only our public key can decrypt it. Bob knows the message is from Alice and only Alice. Of course, this is all very simplified and there may be exceptions or mitigating factors to most of this if practically implemented. But conceptually, there is a pleasant, if elusive, symmetry to the security such an encryption scheme allows.
Jill’s examples of censorship resistant and privacy preserving technologies are telling in their obviously one-sided appeal. If it were not at all common for governments to prevent users accessing illicit online material, but were rather far more pervasive a problem that governments spoofed the source of the illicit material to mislead the reader, Tor would be just as useful. But as this is not the problem that is most obviously overcome by Tor, this seems like a strange way to frame the benefits of its cryptographic security. This symmetry is elusive, but is there nonetheless. Similar comments can be made for Signal and BitTorrent, but the alternative framing comes across as even stranger.
Is the security appeal of cryptocurrencies so one-sided? It might seem so at first, but we need to be clear on what exactly we mean by ‘censorship resistance’. I would put forward that the chief benefit of Bitcoin, for example, is that transfers are accepted if and only if they are valid. One consequence of this is that, valid transfers are always accepted. This is censorship resistance. But the two are not equivalent. It also follows that, invalid transfers are never accepted. We might call this ‘integrity assurance’.
Jill draws on many clear use cases that rely on resisting censorship, and, I think rightly, concludes that these are so esoteric as to rule out a path to mainstream adoption almost by definition. Most people’s payments are not censored, and a world in which they are, and hence in which Bitcoin is the most useful method of transacting, “is a very scary place indeed.”
But she does not seem to consider integrity assurance to be a use case. How might such a use case manifest? Alluding once more to the security symmetries of Signal or Tor, does this mean a situation in which we demand that the payment is really coming from the person we think it is? Probably not, because with the transfer of value, what we ultimately care about is simply that we have the value and can transfer it once again at a future date. A coherent payment system works with necessarily fungible units. You will never be turned away from a shop because of the transaction history of your dollar bills.
This confusion arises from a false premise. Payments are not the only use case of money. They are arguably not even well-defined in isolation. We use money not only to transfer value but to store it also. If you intend to store the value received in a payment, it is very much in your interest to ensure that the payment is itself valid. You don’t want to turn around however many years later, attempt to liquidate your value, and discover it was counterfeit all along. If you want to be absolutely sure your outflows are resistant to censorship, then you had better assure the integrity of your inflows in the first instance. We store value so we can later transfer it, and we transfer value so that others can store it.
This regression points to what I think was the gravest omission in Jill’s post. By focusing on payments rather than stores of value — on censorship resistance rather than integrity assurance — she cedes the conceptual battleground to cryptocurrencies’ natural opponents: those who think economic wellbeing is best understood as the rate at which resources are transferred, rather than the rate at which we increase our ability to usefully combine them. If you care about flows alone, rather than the complex interplay of flows and stocks, then you will obviously be discouraged by Bitcoin’s clunkiness and attracted to slicker payments mechanisms. And if you are unlikely to face censorship in your payments, Jill’s comments apply perfectly. Your uses are probably mainstream, and Bitcoin’s difficulties with scale, speed, and cost will likely hinder your intended use case.
But what if you want a store of value? Bitcoin’s qualities are readily compared to ‘mainstream’ alternatives. Obviously, the comparison is highly favourable. Every fiat money in history has had its value inflated away to the point that it is simply not a viable vehicle for ‘storing value’ over the long term at all. And obviously, this cannot happen with Bitcoin. It cannot inflate beyond the transparent schedule of its security mechanism, because invalid transfers are never accepted.
If you think about what money actually is, not by reciting textbook-ish axioms about what you would like it to be, but rather how people use it in real life, you find it is simply the Schelling point for universal credit. It is an IOU that everybody is willing to redeem (primarily because everybody is willing to redeem it) and hence that at every instance of its transfer is re-ordained with the ascription of economic value to work actually performed.
The reason fiat systems fail so badly as a store of value is precisely that they lack the assurance of integrity. Monopoly issuers of fiat money extend themselves credit on the unwilling, and mostly unknowing, behalf of everybody else. No work has been done that anybody is willing to redeem. No economic value has been created, or contracted to be created, to match the token now in issuance. Bob does not know his message of value transfer is from Alice and only Alice; it is not from anybody. The issuer has executed a man-in-the-middle attack on the structure of economic exchange.
And note this isn’t an old man yells at cloud rant against money creation in general in which I insist all economic activity must be conducted with gold. The money created by credit extension can be perfectly legitimate if the risk of the maturity transformation is priced freely by interest, borne by the equity holders of the lending institution, and mitigated with collateral they understand. But it is not if the risk is priced by political expediency, borne by nobody, and collateralised by everybody.
And so we must ask ourselves whether integrity assured store of value is a use case worthy of driving mainstream adoption. It really depends. Value is subjective, and all costs are opportunity costs. In Bitcoin’s lifetime, the practice of actually holding Bitcoin has been fraught with real risk to reliable value storage, while it hasn’t been straightforwardly obvious what the opportunity cost of holding it has been. The contemporary flavour of fiat issuance has been a firehose into financial markets that has done wonders (so far) for traditional mechanisms for storing value without yet leaking into consumer marketplaces to any noticeable or worrying degree — perhaps with the exception of property yields in financial hubs, which are mysteriously left out of exactly such official calculations …
And yet, if you think market capitalisation weighted valuations can indefinitely expand faster than market capitalisation weighted returns on equity; if you think non-financial corporate debt can indefinitely grow faster than productive output; if you think negative interest rates will be tolerated indefinitely in the face of clear alternatives for value preservation; if you think economic wellbeing ought to be measured by ‘growth’ in consumption and all economic meddling ought to optimise this value to the neglect of real life price signals and sustainable capital formation; you will eventually be in for a rude awakening.
It might seem natural to be distracted at this point by demanding an investment recommendation based on this fascinating insight. I offer none. Not just because that would be terribly irresponsible, but because I genuinely have no idea how this will all play out. I don’t think anybody does, and I don’t think anybody can. When you distort the price and value of time, demanding to know how this meddling will unravel in time is frankly pretty unfair. But the beauty here is that we don’t need to. Any traditional ‘investment recommendation’ would necessarily exist within the financial system to whose grave flaws I am pointing, and no matter how directionally correct your position, the market can stay irrational longer than you can stay solvent. I am not touting how best to play this financial system. I am touting another one.
Saifedean Ammous has said on numerous occasions, I think to the irritation of some of his more hardcore fans, that Bitcoin’s biggest competitive threat would be a widespread return to genuinely sound fiat money. Like much of what is good in Jill’s piece, this realism forces us to consider that none of this is axiomatic. We can readily check. Does Saif’s threat seem likely?
Well, in the time it took me to write this essay, the Federal Reserve counterfeited more than half the outstanding value of Bitcoin just to smother warning signs that price signals in the market for time are very, very wrong. We live in a world in which the incoming president of the European Central Bank can blithely say that, “we should be happier to have a job than to have our savings protected,” and ‘mainstream’ commentators react not with perplexed horror but with a shrug. Will savers shrug when their savings are not protected? When the value of their past time is decimated to boost consumption? Will the opportunity cost of holding an asset that retains value independently of the meddling of bureaucrats be so unclear? Is this really not ‘mainstream’ enough a problem to address? Is ‘saving’ breaking a social construct, or is it itself a social construct that has been broken?
Jill claims she does not want to live in a world where cryptocurrency has found mainstream use. But we already live in that world; it just hasn’t been evenly distributed. The causes of this world are right in front of us, but the effects are spread out in time because the price of time cannot be trusted. Until it is distributed, or distributing much faster than currently, the dream of mainstream use is likely fanciful because the opportunity cost will simply not be obvious to most. Improving payments bit-by-bit will be the best we can do, and no doubt will be worthwhile work: we store value so we can later transfer it, and we transfer value so that others can store it. But Jill speaks of ‘metrics’ and ‘adoption’ as if cryptocurrency were an app or a file standard. Presumably one would treat ‘adoption’ in such cases as a function of time. Doing so here is to cross our wires and will leave us hopelessly confused.
Censorship resistance is nice to have, but integrity assurance creates the fundamental use case of fixing the price and the value of time itself, and keeping it fixed forever. We can absolutely judge based on mainstream adoption, but we can also afford to wait.
follow me on Twitter @allenf32
