Eli Dourado

Bitcoin through the lens of portfolio theory

I’ve had some productive discussions with Noah Smith about Bitcoin volatility and the roles of money, to the point where we now seem to mostly agree. But it’s hard to leave it there, so at the risk of eroding some of the gains in agreement, let me press Noah on whether volatility is really so important for adoption as a medium of exchange. I’m not convinced, and to bolster my argument, I will cite something Noah knows a lot more about than I do: modern portfolio theory.

Noah says for something to come to be used as a medium of exchange, it needs to be not that volatile in terms of real goods and services. Nobody wants to go to the ATM to withdraw cash, only to discover when they arrive at the store that prices have changed, and they need to go back to get more. People like the money in their wallets to have stable value.

I think this is either not true or is a behavioral anomaly. What matters to a rational agent is the efficiency and variance of their entire portfolio. Part of your portfolio is held at your brokerage firm. But part of it is also in your wallet. From a portfolio theory perspective, it makes no difference which assets are held where. As long as investors are willing to hold Bitcoin in some form, they should be willing to hold it in their wallets instead of in their brokerage accounts.

It is true that today we use money that has low volatility. But that is probably because we happen to use media of exchange that are also units of account. Let’s take for granted that Bitcoin will not be a unit of account while it remains volatile. This discussion is solely about the medium of exchange. If we’re separating the medium of exchange from the unit of account, I don’t see why a rational agent would refuse to exchange a relatively volatile asset.

To be clear, if people are holding more volatile assets in their wallets, the rest of their portfolio will have to become marginally less volatile. My only point is that it is irrelevant where the overall volatility of the portfolio is located.

It may be the case that Bitcoin is not an asset that any rational investor would want to hold, or a rational investor might not want to hold enough of it to function as a wallet. This seems unlikely to me. I haven’t seen measurements, but I think Bitcoin probably has low or even (in the long run) negative beta. This means that it can probably increase the risk-adjusted return of a market portfolio through its hedging properties. I don’t hold a huge fraction of my net worth in Bitcoin, but I also hold more value in Bitcoin than I tend to keep in cash in my wallet. Clearly, I think this is rational, or I wouldn’t do it.

Incidentally, portfolio theory is also why arguments about Bitcoin hoarding are so lame. Some pundits have made the claim that people will not spend a deflationary currency because the value is expected to go up in the future. This is totally, totally false. If people really believe the value of the asset will go up in the future, they can simply hold more of it in the rest of their portfolios. Hold enough so that you don’t mind spending marginal Bitcoins. Spend and rebalance the overall portfolio. Problem solved.

If portfolio composition is all that matters, then why don’t people simply keep their money in a Wealthfront or Betterment portfolio and just write checks against those balances? That’s a good question. The answer is that I would love to do that, but Wealthfront and Betterment do not offer that yet. I hope they do soon. The more serious answer is that rapid asset reallocation is a new thing and as a society we are not accustomed to it yet. It may yet happen. And sadly, a part of the answer is that a portion of the population is so liquidity constrained that this doesn’t make sense for them. But overall, I think it is a good idea.

The bottom line is that once you separate out the unit of account from the medium of exchange, money becomes a funny thing. We all have an intuition that low volatility is necessary for a thing to be money, but that’s really unit-of-account bias. I think that upon reflection, volatility is not so important for the medium of exchange.