April 20, 2014

avatar

The low-transaction-fee argument for Bitcoin is silly

A common argument advanced by Bitcoin proponents is that unlike banks and credit cards, Bitcoin has low (or even zero) transaction fees. The claim is a complete red herring, and in this post I’ll explain why.

Let’s assume for the purposes of argument that Bitcoin transaction fees are, in fact, zero. There are small mining-related transaction fees, but it seems plausible that these fees will always be far smaller than those associated with traditional banking.

Why do banks and credit cards charge those annoying fees? A major reason is fraud. Banks eat the cost of fraudulent transactions, but pass on the cost to the customer by taking a cut of each legitimate transaction. Fraud is not an artifact of a particular system that we can design away — it is inherent to every form of money handled by humans. To compare Bitcoin meaningfully with traditional banking, then, we must ask how big fraud-related losses are for Bitcoin users.

Framed this way, the comparison is not a happy one for Bitcoin. From thefts of wallets to hacks of Bitcoin exchanges, fraud in the Bitcoin ecosystem is rampant. It only gets worse when we add sources of risk other than fraud. A recent study found that 45% of Bitcoin exchanges shut down. Several of the rest have suffered attacks and losses.

Granted, one can’t read too much into the numbers at this point, as Bitcoin is still in its pimply-faced adolescence and might very well grow up to be a responsible adult. But there are two major ways in which fraud and other losses in Bitcoin compare quite unfavorably with traditional banking, in a qualitative sense.

First, in the libertarian tradition of cypherpunk technologies, Bitcoin shifts a lot of responsibility from institutions to users (of course, that’s part of what’s attractive about it). But users are, and will always be, dramatically worse at security than institutions. There’s a reason why most of us don’t keep our money in cash under our pillows, or worse, carry it around in a briefcase.

Second, also in the libertarian tradition of cypherpunk technologies, ownership of bitcoins is defined by crypto and enforced by code. As such, transactions are final, and there is no possibility of recourse to legal or social mechanisms to reverse theft. Put another way, reducing the problem of fraud prevention to that of computer security cannot possibly be an improvement.

Because of these problems, as Bitcoin becomes more mainstream, it looks like users will interact through services like Coinbase (which is like Paypal or an online bank, except Bitcoin-based) instead of owning bitcoins directly. This hybrid model will alleviate the problem somewhat, but not entirely. If and when these services mature enough to offer any kind of fraud protection, they will have to start charging higher fees.

In other words, Bitcoin by itself is not a payment mechanism, but can serve as the foundation for one. Thus, the claim that Bitcoin payments have low transaction fees is a category error due to looking at the wrong layer of the system. Considering fees together with risk, my prediction is the Bitcoin will remain more expensive than traditional payment services.

To summarize, Bitcoin-based payments seem inherently and categorically more fraud-prone and loss-prone than the money transfer systems it competes with. Hyping the absence of transaction fees as a benefit is like touting a new car that is faster and lighter, but neglecting to mention that it’s because the body is made out of plastic.

As a final thought, cross-border bank transfers have particularly high transaction fees, and it is possible that Bitcoin could take a chunk out of this market. But I wonder if this might be better understood as an end-run around regulation, likely to get shut down once governments wise up to it.

Thanks to Josh Kroll for comments on a draft.

Comments

  1. Gordon Mohr says:

    The fees of traditional banks and credit cards aren’t just fraud insurance, they’re oligopolistic rents.

    BofA wants $25 to send a *domestic* wire transfer, and that can still take a couple days to arrive. (They also charge $12 to *receive* a *domestic* wire transfer.) The fees for international transfers are even higher.

    With Bitcoin, any amount can be transferred anywhere in the world for less than $0.05. The receiver can see it’s been sent in seconds, and gets strong confidence it’s irreversible in tens of minutes.

    For that much speed and savings, I can self-insure against a lot of risks. Or, as with the CoinBase or other hosted-service model, I can now choose a different mix of convenience/cost/risks/insurance, outside the restricted and expensive options of legacy banks and long-unexamined conventions.

    Also, many online vendors deal with a lot of blatant chargeback fraud. Rather than being controlled by the supposed superiority of institutions in security, this kind of fraud is enabled by their policies, and they’ve shifted much of the fraud costs to vendors. It’s the incumbent power of banks and payment-networks, also somewhat cemented in place by regulation, that lets them pull off this shift — not pure economic rationality, which would let other contractual arrangements compete. Sometimes, both sides of a transaction will prefer the immediacy and irreversibility of an “I hand you cash” transaction – it solves some trust issues better than the “everything’s reversible” model.

    Bitcoin (and other emerging competitive payment options) give us a chance to renegotiate these tradeoffs. Fraud is a real problem, yes, but it’s also been used as a boogeyman by incumbent financial institutions to limit challenges to all the ways they’re squeezing maximum rents from their customers.

    • Arvind Narayanan says:

      Good points, thank you.

    • paul says:

      On the other hand, even if you’re right that the current cost of Bitcoin transactions (including fraud protection) is low, that says nothing about the future. Pretty much every newcomer to an established market has included low price as one of their points of competition, and pretty much every newcomer that has grown enough to become an established player has at some point decided that substantial price advantages are no longer necessary to their business model.

      I would expect in particular that, in addition to intermediaries, you’ll find security and backup vendors proliferating as (if) Bitcon becomes widespread enough to support the marketing effort. And then pretty soon you’ll find a regulatory (private or otherwise) infrastructure growing to hold down the more obvious frauds by intermediaries…

  2. pfogg says:

    Doesn’t the transaction fee vs. risk profile just approximate cash transactions? If two trustworthy parties engage in a cash transaction, the risk is very small (e.g. accidental passing of counterfeit money), although risk scales with the size of the transaction (carrying around lots of cash is inherently risky, as noted). Stores perform cash transactions with arbitrary people off the street, and the risk-related costs of that are priced into the products. Cash can be placed in an institution, and for varying transactional costs can be applied in various ways (electronic transfer, debit card, automatic bill pay), which offer different types of protection.

    The institutionalization of Bitcoin transactions operates as an added feature for added cost, but permits transactions analogous to direct exchanges of cash, and in cases where the participants can reasonably determine that the risk of a direct transaction is sufficiently low, this is a clear win.

    Users will have to learn how to safely handle Bitcoins, but a great many people have learned how to handle cash safely (and have learned when to use institutional mechanisms instead). I think it’s a little early to assess how well people can learn to handle electronic transactions when there is an option of direct, low-transaction-cost exchanges with no safety net.

    • Arvind Narayanan says:

      I like your analogy, but it also points to a difference. It takes a huge amount of effort to learn to use cash safely and navigate financial institutions. Parents, schools and peers all play a role. We don’t necessarily realize how much effort it is because it’s part of growing up.

      Will society ever accept Bitcoin in the same way? I doubt it. The evidence so far shows that typical users will not invest much time into getting educated about software security, even though the benefits are immense.

      I’m sure Bitcoin-based payments will be a good option for technically sophisticated users. But I’m skeptical that it will make the jump to the mainstream.

  3. Eugene says:

    You say “banks eat the cost of fraudulent transactions.” This is incomplete and misleading. You are comparing apples to oranges.

    Most nations allow banks to create their money in the form of loans and, with fractional reserve banking, most of those banks keep little to nothing in reserve (for example, Canada, has no reserve requirement whatsoever and banks can make as much money as they want). Furthermore, banks collect vast swaths of interest for that privilege of making up money, in addition to collecting operating fees and investing our deposits. Fraud does not affect their bottom line.

    The risk of fraud, even if it factors into a bank’s balance sheet, is still ultimately born by the holders of the currency. For example, counterfeiting still causes inflation. Theft from your bank account does not get your money back unless that law is on your side, and only then if the responsible party is capable of paying damages. Fraud in Bitcoin presents the same risk to it’s holders, without many of the disadvantages of the popular fractional reserve banking and insured-credit models.

    As for credit cards, they simply build an insurance premium (and other costs) into every transaction. Noone is prevented from operating a similar service for Bitcoin – but it is a matter of taste and would be like comparing apples to oranges. Bitcoin is comparable to cash or gold, not credit. You have compared it to credit.

  4. Dude says:

    Hello Sir,

    Considering banks and Government work with cash would You be so kind to explain the difference between money in institution safe and single person pocket? And are You arguing that Visa takes 2.75% because someone in Uganda mugged some foreign tourist? Hahahaha

    I think we can very well assume that thiefs and stealing is much wider spread than bitcoin wallets mugging, so Your point is simply useless.

    Bitcoin has much lower operating costs than any other fin institution due to its efficiency, all the gray stuff related issues are more rampant in RL than on network..

    Kind regards

  5. Francisco Ribeiro says:

    People hardly trust in money to store their wealth these days. While hipsters try to explain the ‘efficiency’ of bitcoins over Internet, common people who can, are buying gold because in real life they miss the solidness of the economy from other times. While this happens some geeks unawarely brag about the instability of value of bitcoins with its ups and downs. They also fail to understand that a fundamental element of any trade unit to be adopted at a world-wide scale is its value being clearly understandable and tangible to most people and not just some skilled mathematicians.

    Bit coins on current economy are like a rollercoaster build on top of a vulcano with a special deal for smugglers.

  6. Jeff Thompson says:

    Yeah, and I’m subsidizing thieves when I send money to my friends and mom – even if there is no risk of fraud in those transactions. I’m afraid you missed the point – there are plenty of transactions in which neither party is worried about fraud – yet in the current system we pay the ‘fraud tax.’

  7. Nicholas Weaver says:

    The other problem is irreversibility. The modern financial system goes out of its way to ensure that, “if its electronic and big its reversible” whenever possible, because this allows fraud detection and remediation. With irreversible transactions, you can only do fraud prevention, which is a much harder problem.

    This is why its actually so hard to steal money from someone’s bank account online: You need to recruit money mules to take the transfers out in cash immediately and then western-union the money to crooks.

    So it will ALWAYS be high friction to turn dollars into bitcoins: you must go through some irreversible transaction step before someone will sell you bitcoins, because otherwise the bitcoin seller will be destroyed by fraud, since they will be left holding the bag.

    As a result, if you want to buy BitCoins with bank transfers, you can’t on any appreciable volume or speed (e.g. one startup allows it… $100/day max, and it takes 4 days). And the best way to buy bitcoins in the US involves cash drops at CVS!

  8. Richard says:

    The problem is… banks and credit card companies don’t eat the frauds. They pass them onto the consumer or the merchant. For example, if I sell an item online, in good faith, but the purchaser has used a stolen credit card, then I, the merchant, will suffer the chargeback. Personally, I’d prefer the same guarantees as cash: you take responsibility to look after it properly; if it’s in your posession, it’s yours; if you want to transfer it, you can do so without overhead, and without tracking.