Thursday, April 5, 2018

Crypto assets

A large portion of what one might read about Bitcoin and company online falls into a few broad classes:
  • Bitcoin is totally a bubble!
    • It's unfolding just like the dotcom bubble around the turn of the century, but a lot faster
    • Bitcoin isn't actually worth anything
  • Bitcoin is totally not a bubble!
    • The last time it crashed it was way, way up a year later.  Do you really want to miss out?
    • Cryptocurrencies will take over the world.  The world's money supply is worth tens of trillions of dollars, so BTC is conservatively worth at least $1 million (or something along those lines).
  • You're all missing the point.  It's not Bitcoin, it's the blockchain.
    • Blockchains are useful because they provide a secure, shared, decentralized ledger
    • Even if a particular currency fizzles out, there's still value in the technology
While researching a separate post, which I may or may not end up actually posting [I did, more or less -- see next post], I ran across this much more nuanced take by Adam Ludwin on the blog for  If nothing else, I love the disclaimers at the top, a bullet list ending with several "Very few people <in some group> understand what's going on" and finally "It’s very possible I don’t understand what’s going on".

In that spirit, a disclaimer or two of my own: I'd never heard of Chain or Adam Ludwin before this and I have no financial interest in it or in any crypto currency.  I do note that Chain's tagline is "We build cryptographic ledgers that underpin breakthrough financial products," so it's pretty clear what Ludwin/Chain's basic stance is (and, of course, that's a good thing).

The piece itself is to some extent in the third bucket above, except with a lot more thought and detail and without the usual air of certainty.  The thesis, as I understand it, is that Bitcoin and things like it are properly assets, not currency, and their value lies in supporting a certain class of distributed application where various people contribute resources to the application as a whole, and/or consume such resources.

To make this work there has to be some sort of ledger that everyone trusts.  One way to do that is to put someone -- say, a bank -- in charge of the ledger.  Crypto assets provide an alternative, namely a secure ledger with no single, centralized authority, which can be trusted without having to trust all the other participants, or even any particular participant.

Ludwin is careful to point out that this is not always, or even often, worth the trouble.  A decentralized, secure ledger incurs significant overhead, since one way or another you have to be sure that the various parties can agree on what's in the ledger.  More importantly, at least in my view, a decentralized system is by definition not governed by any single authority, meaning governance has to be provided directly by the participants.  There should be cases where this is worth the trouble, but it's not a given.

Bitcoin itself is probably not the best example of what Ludwin is driving at.  Better examples (which Ludwin cites) would be Filecoin (a token for tracking who is providing and using storage in a decentralized cloud) and Ethereum (a generalized platform for crypto assets).  Ludwin attributes special significance to Bitcoin anyway on the basis that it currently represents the biggest chunk of actual money associated with crypto assets.  I'm personally skeptical of this argument, and "first mover advantage" arguments in general, but it's not totally unreasonable.

What follows are my own thoughts, which should all start with "It’s very possible I don’t understand what’s going on", but in bigger, bolder letters.

One of my fundamental concerns with "cryptocurrencies" in general and Bitcoin in particular is that it seems hard to draw a direct line from the value of secure, decentralized ledgers as a service to the value of the associated asset.  The ledger technology may be valuable in general, but that doesn't mean that any particular crypto asset is valuable.  Which tokens you're using is essentially a matter of what ledger you're writing things on, but what makes a ledger valuable?

Before computers, actual paper ledgers carried records of billions of dollars in assets, but the price of paper didn't rise as a result.  It came down over time as paper-making and printing became more efficient.  Very few people cared who printed the ledger, so long as the layout was usable.

One of the key ideas behind Bitcoin and some but not all other tokens is that they will become more valuable over time because there is a strictly limited supply.  But if the price of BTC in dollars keeps going up because everyone is hodling on to them, why use it?  Why not use something else, or spin up your own tokens on Ethereum?

At the end of the day people hosting the ledger will need to be paid for the computing resources they provide, but this should be a modest cost relative to the amounts recorded on the ledger.  One way or another there will be a flow of money from users of the ledger to providers of the service.  In some cases these will be the same people.  If I buy some capacity from a cloud provider and use it to participate in a ledger that I also use to record transactions of whatever nature, then I'm paying the cost of the computing resources to the cloud provider, and that may be about it.

This is assuming that participating in a ledger is relatively cheap.  This is decidedly not the case for proof-of-work based systems like Bitcoin.  Adding a new block to the Bitcoin blockchain currently gets a reward of 12.5 BTC plus whatever transaction fees are included (currently nominal).  At current prices that's somewhere around $80K per block, or around $60 per transaction.  This reward is effectively covered by inflating BTC rather than by direct payment, but miners are most certainly getting paid and are most certainly spending reserve currency on equipment and electricity.

However, I see no structural reason for a ledger to be expensive, unless proof-of-work really is the only way to solve the double spending problem. Without proof-of-work you're computing cryptographic signatures and copying bytes over a network.  This isn't free, but it's not that expensive either.

Which leaves us with this.  If secure, distributed ledgers are expensive, that's one more reason not to use them.  If they're cheap, then it's hard to see how people will make huge piles of money off them.  The optimum is probably in the middle somewhere, with people making modest amounts of money for providing a useful but fairly mundane service.  It's still possible for individuals to make serious money in such an environment, but more in "old-fashioned" ways such as providing better service or a marginally lower price to a lot of customers.

1 comment:

David Hull said...

I'm not going to start a whole new post for this, but it wasn't quite right to say "unless proof-of-work really is the only way to solve the double spending problem". The solution to the double spending problem is a secure ledger -- decentralized or not. That's exactly what's going on whey you pay with an ordinary debit card: the bank records that you've spent the money, and can't spend it again, and people trust the bank.

Cryptocurrencies use the same solution -- a secure ledger -- but implement it in a decentralized way. This is closely related to the "byzantine agreement" problem in computing, but with different enough parameters that it's probably a distinct problem. Bitcoin and others use proof-of-work to implement a distributed, secure ledger, but there are other possible options out there, for example proof-of-stake.

I'm explicitly not taking a stand on which kind of secure distributed ledger is best, or when or whether a distributed ledger is better than a centralized one. The only point here is that "solving the double-spending" problem is a bit of a red herring.