The Dumb Money

The definitive explanation of why Bitcoin is stupid.

Pile of Bitcoin and Litecoin keychains, November 30, 2013. BTC Keychain / Flickr

Bitcoin’s pseudonymous creator, “Satoshi Nakamoto,” announced his new invention under two guises. The first was as a sober engineer, touting the benefits of his new invention for internet merchants. His original 2008 technical paper introduced Bitcoin as the solution to a practical problem. Why accept this weird new currency? Because while the existing electronic payments system “works well enough for most transactions,” it “still suffers from the inherent weaknesses of the trust based model.”

The problem, as Nakamoto described it, seems curiously modest: financial institutions can reverse charges if a customer disputes them. “The cost of mediation increases transaction costs, limiting the minimum practical transaction size and cutting off the possibility for small casual transactions, and there is a broader cost in the loss of ability to make non-reversible payments for non-reversible services.” For in-person transactions, cash solves that problem — once it’s passed to you, you have possession. But Bitcoin does it for electronic transactions. It replaces mere “trust” with “cryptographic proof.”

But then there was Nakamoto’s second guise: that of a techno-libertarian, web-forum dweller, tapping into old tropes of right-wing money crankery, presenting Bitcoin as a weapon against “fiat currency” and “fractional reserve banking.” A web-forum post by Nakamoto linked to the technical paper, but reworked the message for a different crowd. The problem again was “trust,” but here he focused on Bitcoin as a solution to public rather than private problems: “The central bank must be trusted not to debase the currency, but the history of fiat currencies is full of breaches of that trust… Banks must be trusted to hold our money and transfer it electronically, but they lend it out in waves of credit bubbles with barely a fraction in reserve.” Again, Bitcoin avoids the problem of trust in fallible or malign human institutions through the wonders of “cryptographic proof.”

Money is a social device. The things that make it work are not to be found inside the coins and notes, but in a network of connections between the people who use it. Bitcoin’s fundamental flaw is the myth that such workability can be “coded.” This is no more possible than it was to physically stamp society’s monetary relations at the mint. At best, Bitcoin will develop such a social infrastructure — in which case, it will work a lot like regular money, and so what’s the point? At worst, the properties coded into Bitcoin will make it unsuitable for a viable social infrastructure, and it will implode and remain niche — a disruptive technology that mainly disrupts its own users.

Experience so far suggests it is more likely to go the second way.

Bitcoin and the Monetary Problems

For each of the monetary problems Nakamoto writes about — real or imagined, private or public — “cryptographic proof” is his answer. That is what Bitcoin is offering, so if it is to be the future of money, that answer better be pretty good. But are Nakamoto’s problems really the fundamental problems of money?

The technical problem that Bitcoin solves is to simulate a physical coin electronically, without relying on centralized administration. When a real coin passes from hand to hand, something physical that was in one person’s possession is now in another person’s possession. When someone spends a coin, they don’t have it anymore and so can’t spend it again.

Simulating this electronically is nothing new. Most of what we use as money is not physical cash — coins or notes — but electronic money in the form of bank deposits. And in reality, this takes the form of “ledger” money, which has existed for centuries through paper accounts. Ledger money simulates coin in the sense that you can’t spend the same deposit twice, even though, unlike with hard cash, there is nothing physically stopping that from happening. The banking system could add to the payee’s account without deducting from the payer’s, with just a few keystrokes. (“Bank error in your favor!”) The reason banks don’t do so is that their deposits are liabilities — promises to pay — which they’re required to stand ready to convert into higher-level currency (such as government-issued paper money). “Copying” a deposit would merely increase a bank’s liabilities without increasing its assets. Banks do create new deposits “out of thin air,” but only when they lend, and they do so in exchange for a corresponding asset — the loan contract.

The regular electronic-payments system is a complex network, and hierarchical in nature: we pay each other mainly in bank deposits; banks settle among themselves mainly with central-bank money — the famous “fiat currency.” International payment networks, involving conversion between currencies, are still a bit janky, though they seem pretty seamless for anyone with a credit card. The whole system is administered: much of it may now be automated, but it still relies on institutions maintaining databases in the form of ledgers.

The blockchain technology Bitcoin introduced also simulates coin transfer, but in a different way — without centralized, hierarchical administration. It does this by creating public records of transactions, with copies of the blockchain existing on many computers, thus creating a hard-to-falsify record of who has title to each coin. The details aren’t important for our purposes, but you can get an overview from the original paper itself.

The blockchain does solve a problem: how to simulate a coin so that it cannot be double-spent or copied, without relying on financial institutions maintaining centralized ledgers, thus providing the alleged benefits of privacy while dispensing with middlemen and any need for human trust. The way it does this may seem pretty clumsy to a cryptographic layman like myself, and it is notoriously wasteful of electricity, but experts have called it “an incredibly clever piece of cryptographic engineering,” so I will take their word for it for the sake of argument.

But this merely solves the most basic monetary problem. True, any means of payment, at a minimum, needs to have solved this problem: when you have it, you must have it, and other people you might want to pay it to will recognize it as it. But that is just a minimum. It’s not enough to have collective recognition that, yes, that is a ten-dollar bill, two-pound coin, or Bitcoin. That can be coded into Bitcoin just as traditional currencies are stamped at the mint. What cannot be coded is:

  • Acceptability: that other people will take it in payment at all;
  • Value: what it is worth in terms of other currencies, goods and services, etc.; or
  • Liquidity: the ease with which that value can be realized by exchanging it for those other things.

Acceptability in payment generally depends on the other two things: for people to be willing to accept something in payment, they must be confident that others will see it as valuable, and that they will be able to realize that value without too much trouble.

So let’s look at how Bitcoin fares.

The Value Problem

Bitbugs often claim, like goldbugs before them, that “fiat currency” is a bad store of value. Unlike goldbugs, they are not recommending we adopt something (like gold) with nonmonetary uses or a long tradition as a safe haven, which might put a floor under the value of their alternative. Their claim rests on the idea that central banks are prone to manage currency badly, due to either ineptitude or depravity. Bitcoin, on the other hand, has no manager at all. It is coded so that mining will release new coins at a predictable and declining rate, with peak supply of 21 million Bitcoins. So, the argument goes, value has been coded into it.

So far, though, Bitcoin has fared worse in terms of value stability than almost every fiat currency in history.

But Nakamoto never promised value stability. Instead, he predicted that Bitcoin would steadily rise in value: “Instead of the supply changing to keep the value the same, the supply is predetermined and the value changes. As the number of users grows, the value per coin increases. It has the potential for a positive feedback loop; as users increase, the value goes up, which could attract more users to take advantage of the increasing value.”

That worked as a third pitch for Bitcoin — to appeal, beyond the nerds and techno-libertarians, to shoeshine-boy speculators. And this argument would ultimately prove a very important one.

Until the last couple of months, Bitcoin’s trajectory was mainly upward, and spectacularly so in 2017. Its value in terms of national currencies, and in terms of things priced in those currencies, kept rising. Nakamoto’s prediction seemed to be coming true — not because people were flocking to actually use Bitcoin as money, but because they were buying it to hold (a strategy known as “HODL” in Bitcoin’s online lingo), to speculate on further increases in value.

Wasn’t this a sign of Bitcoin’s strength? Even the bitbugs could hardly deny that there were classic speculative dynamics at work that couldn’t go on forever. But to the extent that people could give any rationale for the sustainability of Bitcoin’s value, it was the idea that eventually Bitcoin will take some substantial share of the global demand for money. Since total Bitcoin supply is capped, even a fairly modest share would support a much higher Bitcoin price than even the stratospheric valuations of late 2017. (A back-of-the-envelope calculation on this basis published at Investopedia came up with a valuation of $514,000 for one Bitcoin, assuming Bitcoin meets 15 percent of demand for money worldwide — whether that seems modest or outrageously absurd, it is good enough for cryptocurrency subreddits.) Speculators were getting in early, and over the long run would be able to offload their holdings to new users as Bitcoin moved towards mass adoption.

Unfortunately, Bitcoin’s very success among speculators made it less useful as money. A rapid, self-perpetuating rise in value can undermine a currency as easily as rapid inflation: who wants to spend something that might multiply in value in a matter of weeks? And of course, Bitcoin’s rise did eventually peak and go into reverse. Now its trajectory is uncertain from day to day, even from hour to hour.

The bitbug’s response to this is that it doesn’t matter, because Bitcoin is not used as a unit of account. Sellers do not usually price things in Bitcoin independently. They set their price in terms of some national currency, like the dollar, and let the Bitcoin price automatically adjust according to the current exchange rate — just as they do with foreign currencies when they sell to buyers around the world. Bitcoin is only a vehicle, a means of payment not intended to be held any longer than it takes to convert it into some other currency.

This is quite a climb down from any claim of Bitcoin rivaling “fiat currency.”

It is true that something does not need to be a good long-term store of value to succeed as a means of payment. But instability of a unit of account, on the other hand, can undermine a monetary system, because contracts are drawn up in terms of units of account. For contracts to do what they are meant to — lock-in aspects of the future — the parties must have some confidence in the future meaning of the numbers written into them. Imagine what would have happened to anyone contracting for a wage set in terms of Bitcoin, fixed last February for a year in advance. Maybe Bitcoin’s appreciation would mean the workers would now be rich — but more likely their employer would be out of business. Or, imagine an annual wage struck in terms of Bitcoin in December: already, the workers’ living standards would have fallen by more than half. Clearly a currency subject to so much volatility will not be adopted as a unit of account.

But, alright: draw up the contracts and price the commodities in terms of dollars, but specify that the actual payments will be made in Bitcoin according to the going exchange rate: this is Bitcoin as means of payment, but not as unit of account. Bitcoin is, in that case, not an alternative to state-issued currencies, but parasitic on them. It leaves the stability of units of account to the dastardly central banks.

But even this is doubtful. A central bank’s ability to manage the stability of its unit of account (for example, the dollar) depends on the fact that it issues a means of payment denominated in that unit of account (for example, Federal Reserve notes denominated in dollars), and so has some control over conditions of supply. It also depends on its acting as supervisor for private banks that issue deposits at par to that unit, who need to pay among themselves in the “fiat currency” base money. To the extent that Bitcoin — miraculously — were to become a popular alternative means of payment, it would undermine the unit-of-account stability it depends on.

But, as parasites go, Bitcoin is not likely to be terribly successful at reproducing, and so will not kill its host. That’s because it also makes a terrible means of payment.

The Liquidity Problem

To be a good means of payment, something needs to be widely accepted in payment, and usable with a minimum of transaction costs. That raises a chicken-and-egg problem: people usually accept something in payment because they are confident other people will accept it — so how does a new money get off the ground in the first place?

There is an out: even if you can’t directly spend the thing very widely, as long as you can easily convert it into something that is widely acceptable, it is almost as good. Convertibility has helped support the acceptance of most of what we now use as money: national currencies and bank deposits alike. The text appearing on American bills used to promise “to pay the bearer on demand” X dollars “in gold or lawful money.” Now those bills are X dollars, acceptable to the point that convertibility is superfluous. The acceptability of bank deposits still depends on convertibility at par, not least because we want to be able to pay the customers of other banks. Bitcoin has no one guaranteeing convertibility at any particular rate, so at best it fits into the category of liquid assets, like bonds and shares. But unlike them, it is intended as a means of payment. Can such a thing achieve wide acceptance?

Bitbugs used to claim low transaction cost as an advantage of Bitcoin, because no overhead needs to be paid to a managing institution like a bank. That now looks foolish. There are two parts to the cost of using Bitcoin. One is the transaction cost internal to the Bitcoin system itself, incurred to execute a Bitcoin transfer from one person to another (or one “wallet” to another — in fact, you even incur a cost to transfer your own coin between your own wallets, since that needs to be registered in the blockchain as much as any other transaction). Then there is also the cost of shifting out of Bitcoin. Most retailers that do accept Bitcoin cannot themselves make most of their payments in Bitcoin, so the ease of its convertibility into national currencies is vital.

The internal Bitcoin transaction fee is an odd thing. The would-be spender must give the blockchain processors (“miners”) an incentive to include the transaction in the next block they generate. The spender chooses how much to pay, but of course the miners are interested in including the transactions with the highest payoff for the space they take up on the block. (It gets complicated: each block has a finite amount of room for transactions. The data size (or “weight”) of your transaction depends on the number of different sources from which you received the Bitcoin you are now spending, not the amount of Bitcoin you are spending. So someone spending a Bitcoin accumulated from thousands of micro-transactions has a much “weightier” transaction — one that takes up more room in a block — than someone spending the same amount, but received in a single transaction. The weightier the transaction, the higher the fee.) Instead of paying a set fee to some dinosaur of a bank, Bitcoin users make bids on a free market to get their payments processed. You must monitor the going rate, which depends on how many other transactions are out there competing for blocks, and then trade off how much you are willing to pay against how long you are willing to wait for your transaction to clear. If you’re unwilling to get the calculator out, there are sites that will estimate how much you currently need to offer to be confident of getting your transaction included in the next x blocks.

It can seem from that last paragraph — which you may have skimmed — that Bitcoin is destined for a niche made up of the overlap in the Venn diagram between “people who enjoy operating their computer from the command line” and “people who love the idea of bidding on a market just to make a payment.” In reality, the casual Bitcoiner does the Bitcoin equivalent of using a graphical user interface, turning to a wallet service like Coinbase, which will make an executive decision on the trade-off and simply present it to the user as a transaction fee. For a long time, this was basically invisible to people, because the going fee was tiny — typically a few US cents, for processing within an hour. That changed in 2017 for the simple reason that a lot more Bitcoin transactions started happening as the bubble took off. In June, the median fee rose to around $2.50 — upsetting anyone in the habit of buying their coffee with Bitcoin. It fell back before another spike in August took the fee briefly above $5. And then, December: the median fee hit above $30 a few days before Christmas, and stood above $14 for more than a month. This, remember, was the cost of executing a single Bitcoin payment.

The speculative mania had overwhelmed the processing capacity of the mining system, making it useless as an everyday means of payment — even though the only even slightly plausible rationale for Bitcoin’s absurd value was its glorious future as a means of payment. (Sample Reddit comment from last September: “Bitcoin will soon be worth hundreds of thousands of dollars each and the idea you should waste it buying coffee makes no sense… Eventually the world will no longer accept US dollars as payment, because the world will have discovered there is a superior currency that has more demand, but until that happens, you should take advantage of people’s willingness to trade things for paper fiat.”) Since the January crash, the transaction fee has returned to more reasonable levels as the flow of transactions to process has diminished.

But these fees internal to the Bitcoin system are only part of the transaction cost. Because Bitcoin is not widely accepted, retailers who do accept it need to turn it into their regular domestic currency. Because its value is so volatile, they generally do not want to take the risk of holding it any longer than they have to. No institution has taken on the responsibility of converting Bitcoin into currency at a fixed rate — and it would be insane to do so. That alone makes Bitcoin, or any cryptocurrency, far less liquid than the competition — that is, electronic bank money denominated in a national currency. Bitcoin’s liquidity, such as it is, depends on the state of its exchanges.

At best, this would be not so far from the situation internet retailers face in accepting foreign currency. In any international transaction someone — whether the buyer, the retailer, or a financial institution in the middle — has to change currencies in the foreign-exchange market, and that adds a small cost to the purchase (around 3 percent), usually borne by the purchaser. International purchases are routine — so what’s the problem?

For one thing, Bitcoin adds such a cost even to domestic purchases, when buyer and seller share a currency and the transaction could have happened without any need for money-changing. In fact, the cost is doubled: until people start earning wages in Bitcoin, both parties have to go to extra effort: one going out of their way to buy Bitcoin, the other forced to sell it, each paying the internal Bitcoin fees as well as any fees their financial institutions charge.

But it gets worse, because Bitcoin is not just another currency. It is much more volatile: its price tends to shift around relative to national currencies much more than national currencies shift relative to each other. A retailer accepting Bitcoin directly into their own wallet is going to be exposed to changes in Bitcoin’s value between the time they accept payment and the time at which they can exchange it for local currency. The time it takes for transactions to be recorded in the blockchain is uncertain but often around an hour, and much longer in high-transaction periods. Given Bitcoin’s volatility, that means a gamble as to how much domestic currency they ultimately receive from a Bitcoin payment.

Meanwhile, the second transaction — from Bitcoin to the seller’s local currency — also depends on the state of the Bitcoin exchanges. Bitcoin’s bid-ask spreads — the gap between the price at which Bitcoin can be sold and the higher price at which it can be bought — are not terribly high on the most liquid exchanges for US dollar trades, suggesting that it is generally not hard to find someone to transact with. But it can be substantial for some currencies. This opens yet another wedge between what the buyer pays and the seller receives, on top of the transaction fee and any institutional fees. And there is no cryptographic protection from a liquidity seizure on the markets, should people suddenly become wary of Bitcoin for some reason or another.

In fact, most retailers accepting Bitcoin do not do it directly. They use institutions to handle the backend, making it (relatively) painless and (mostly) risk-free. Companies like Bitpay and Coinbase take care of the Bitcoin transaction, and simply deposit the equivalent in local currency into the retailer’s bank account. Of course, they charge for the service, one way or another: someone still has to pay the transaction fees and take the risk of holding the actual Bitcoins. To some extent, Bitcoin middlemen can reduce transaction costs, because if two users of the same service pay one another, that can happen on the middleman’s books without Bitcoin moving from wallet to wallet. That is, they cut costs by avoiding actually using the Bitcoin infrastructure.

In other words, we’re back to the ledger and the need to trust institutions. Can fractional reserve banking be far behind?

The lesson here is that institutions spring up in response to problems because money cannot manage itself, even with all the cryptographic innovation in the world. Perhaps this emergence of an infrastructure of cryptocurrency middlemen shows a certain maturation of the currency, just as banks inevitably grew up around currencies. But it squarely undermines Bitcoin’s original claim to enable payments without the need for trust or institutional guidance. (Similarly, much “buying coffee with Bitcoin” is actually buying coffee with gift cards pre-purchased with Bitcoin, adding a whole extra layer of transactions costs and illiquidity.)

Most casual users of Bitcoin engage through an online service, and so get the worst of both worlds. They pay, one way or another, for the transaction costs and risks of bitcoin, and also get the problems of trust and security that come from dealing with institutions — the very trust problems which Nakamoto had presented Bitcoin as the solution to. Anyone using an online service to manage or trade their Bitcoins must trust institutions with untested reputations, at best. The record has been rather patchy. Some choice epitaphs from the tombstones in the Bitcoin failure graveyard:

  • …filed for bankruptcy protection while 850,000 Bitcoins evaporated from customers…
  • …regulatory problems, the loss of US$100,000, and a dispute with a payment processor…
  • …struggled for several years before shutting down due to a lack of interest…
  • …significant financial loss… when a web host had an internal security breach that gave the attacker access to the wallet in which Bitcoinica stored funds…
  • …after increasingly degraded service … the operator shared that the customers’ funds had been used for speculation…
  • …reported 17,000 client Bitcoins were missing, after it lost access to its wallet.dat file…
  • …started experiencing problems with payments in November 2013, prompting Reddit users to call it “a scam”…
  • …the site went missing, claiming a “maintenance period”…
  • filing for bankruptcy after it was hacked for the second time in a year…
  • brought down by law enforcement … with twenty-one counts of money laundering and other related financial crimes…

Even the largest and most reputable platforms like Coinbase and Bitfinex “have the annoying habit of limiting withdrawals to small sums and/or suffering processing delays whenever prices move downwards.”’ The blockchain itself might offer anonymous payments, but not the institutions: at the end of February Coinbase told 13,000 customers that it would be complying with an IRS request for their transaction records. As for raw Bitcoin, forensic financial analysts recently showed that an impressive amount of payment detail can be inferred from the data written into the blockchain, making it “an effective tool for law enforcement.”

There are signs that the middlemen have run into difficulties with transaction costs. Just before Christmas, BitPay announced that it would no longer process payments to retailers smaller than $100, because smaller transactions “are quickly becoming impractical for users to send and for BitPay to process.” They soon reversed the decision in the face of customer outrage, but noted that with payments then costing upwards of $30, it did not make much sense for people to use Bitcoin for small payments anyway. Since then, the average Bitcoin transaction fee (the one internal to a Bitcoin payment, not including other costs of moving into and out of Bitcoin) has fallen back down to around ten to twenty US cents, as the mania has subsided and a blockchain code tweak has spread. But if the pace of transactions picks up again, at some point Bitcoin will reach the bottlenecks.

Not that many people are using Bitcoin for any kind of retail payment. Take a look at a current list of retailers that accept it: Overstock.com, Microsoft (kind of, maybe), Expedia, Dish Network … and then luminaries like eGifter, PizzaForCoins, Newegg, Shopify, and Roadway Moving Company. For the small companies that bother with Bitcoin, the extra hassle can be considered a marketing expense: a London sandwich-shop proprietor told Bloomberg the incentive was “a large amount of publicity” for “a few small payments.” If too many people started paying with Bitcoin, they would have to reconsider, the shopkeeper added — “but I can’t imagine that happening any time soon.”

When a Morgan Stanley report last year revealed that the number of the top 500 internet retailers accepting Bitcoin had fallen to just three, Business Insider put a positive spin on it: “Bitcoin will continue to go mainstream, but as an asset rather than a transaction method.”

But as an asset, the only slightly plausible rationale for its value was its supposedly glowing future as a transaction method.

Self-Disruptive Technology

In spite of all this, Bitcoin has non-negligible value — half as much as a few months ago, but much more than a year ago. This is almost entirely held up by speculative demand — people HODLing on because they hope to ride the price back up, or avoid realizing their losses.

There is a genuine use-case for Bitcoin and other cryptocurrency, and that is their ability to facilitate private online payments: black-market transactions, legal purchases people would prefer to keep off their credit-card statements, and the evasion of financial controls. (Adoption in China was critical to the early success of Bitcoin; for a while in 2013, yuan-denominated Bitcoin trades outpaced dollar-denominated Bitcoin trades, and it has continued to be a major venue for mining and trading in spite of official crackdowns.) Any of these reasons can make someone willing to endure the hassle and risk and transaction costs of using the clumsy currency.

But this does not justify any particular value for bitcoin, not matter how limited the supply. It competes with other cryptocurrencies and its own hard forks (new versions of the currency, like Bitcoin Cash, created when the network splits over code changes). In the unlikely event it is widely adopted, there is nothing to prevent the development of the dreaded fractional reserve banking in bitcoin. So it is likely to have a niche in the monetary system, but a small one.

For the Left, there is no reason to regret Bitcoin’s failure. Its critique of the monetary system is not our critique. There is a view that since banks and central banks are no friends of ours, anything that threatens them must be at least a little welcome. But Bitcoin does not threaten the establishment because it is a pathetic competitor with the establishment’s money, which still offers security, liquidity, and relatively low transaction costs. Bitcoin has little to offer anyone except its enthusiasts, and speculators will only hang around so long as it is plausible that its value might keep rising.

Conservatives and libertarians are obsessed with managed money because they need a scapegoat. Capitalism has problems? It’s because the state has interfered with the natural order, or repressed our freedom to transact. Socialists understand that the state’s monetary architecture is designed to maintain the stability of a system marked by inequality and alienation, but is not itself the cause of those things.

Were Bitcoin to somehow replace fiat currency and the banking system — forgetting for a moment all the reasons that is not going to happen — it would leave capitalist social relations intact, but blinder and harsher. The gold standard ultimately fell because it ran up against a rising labor movement; our present arrangements are a tenuous compromise that emerged in its wake.

Our solution leads in the opposite direction to Bitcoin: not less management but more — socialized and under democratic control. That cannot be coded. Ultimately, Bitcoin is one of those disruptive technologies that mostly disrupts its own users. We should leave them to it.