Bitcoin continues to surge in buzz and price (194,993 coin transfer, US Senate hearings and astronomical price and total capitalization). This gets me to thinking: what in finance terms is Bitcoin? It claims aspire to be a currency, but what is it actually behaving like?
To put my cards on the table: I am not really a fan of Bitcoin and do not own any. To say it stronger I do not desire any Bitcoin. Sure: I would love to have purchased it a year ago, but that is only because I would (like many) like to have paid around $20 a year ago for something that is pricing around $700 now. Or I would love to loan $20 to receive $700 back a year later. But that in no way convinces me that the same coin will price at $24,500 a year from now. If I had a way of knowing such a thing I would buy Bitcoin. If I had a way of predicting if or when Bitcoin might collapse I might research how to short it. But I don’t know any such thing either way.
However, (even though it has been true for the last five or so years) I do not think now is always a good time to buy Bitcoin. Ponzi schemes all run on the variant of the aphorism “The best time to plant a tree is 20 years ago, and the second best time is now.” So unless you intend to assert Bitcoin is a Ponzi scheme you don’t want past performance to be your sole justification of purchase (you need to find and reason about underlying mechanisms). Don’t confuse holding means of production (like a tree) with holding assets (like money, Gold or Bitcoin). See: Buffet’s story of holding all the world’s gold for 100 years (the punchline being 100 years later all you have is still all of the world’s gold, but if for a similar price had you seized all of America’s crop-land and much more you would expect to hold a lot more at the end of 100 years).
I actually really do like the timing analysis Locklin tried in trying to fit a log-periodic power law. Usually pure technical signals are not to be trusted- but in this case the analysis explicitly assumes the market is purely driven by a feedback loop (which, if true, would relate distribution of future price to pure technical indicators). If you find the assumption plausible the analysis is interesting (and if you find the assumption implausible the analysis is not relevant). Locklin correctly identifies that an overly good log-periodic power law fit may be a signal of a possible market phase change, and does a much better job of model fitting than my optimization library did. The missing issue: you need to calibrate such a signal: how many times in the past did you see it that strong without a crash (something you can estimate from earlier data, but it is a lot of work) and what is the unknown base probability of crashing (a Bayesian issue, without this you can’t tell if a region of time that is a thousand times more likely than average to crash is actually likely to crash or not, and this is something hard to estimate). But overall- it was a good analysis, which I think has something useful to say about the relative probabilities of a commodity price crashing over different time-periods. It was merely pre-mature to move it from a forecast of relative propensity to an actual single time prediction of outcome. But I will grant: a technical analysis essentially based on the assumption the Bitcoin market is a Ponzi scheme is always going to predict a crash (it is just interesting to see if it can make a specific forecast of when the crash will occur or not).
That got me thinking: are there analogies other than Tulips panics and Ponzi schemes that may apply and may have different conclusions? Can we assume something else? And if we assume something else what different possible conclusions could we entertain?
The answer is yes: Bitcoin can be thought of as a variation of a semi-closed-end index fund (or semi-open-end index fund) where the underlying assets are electricity and computer hardware (on a forced changing price schedule). If you believe this analogy then by analogy Bitcoin could behave like an index fund that is forced to appreciate due to the designed rapidly increasing entry expense (in terms of computer hardware and electricity) in creating Bitcoins (which perhaps gives us some of the behavior of an open-end fund with a rapidly increasing entry fee).
Notice in this “analysis by analogy” I am avoiding math, and avoiding a lot of details about Bitcoin (in particular its celebrated distributed and cryptographic underpinnings). I am trying to just get some qualitative consequences assuming some financial properties. If we knew for certain the principles apply we could take our conclusions as reasonable predictions. But we have not established our index-fund analogy actually holds, so our conclusions are only conditional or subjunctive (not to be relied upon until you can establish they are talking about the real world, and not talking about consequences of axioms that don’t actually hold). Leaving that aside: what to we mean by closed/open funds?
An open-end fund such as an open-ended index fund can be explained and reasoned about. Take a simplified S&P 500 open-end index fund for example. For every $1 invested directly in the fund the fund manager buys a $1 basket of the target assets (in this case the stocks representing the index in the proper proportions). If an investor wants $1 back out of the fund then the fund sells the appropriate basket and gives back the money. Both the actual buying and selling of baskets are relatively painful propositions so more often investors enter and exit the fund by buying and selling shares of the fund itself on a market. So shares in the fund are priced as an asset driven by supply and demand. The fund’s price can float on its own (commanding a premium or representing a discount with respect to the matching underlying basket). But the fund price will tend to track the value of the underlying basket due to the threat of arbitrage. If the fund is cheap relative to the basket then it is tempting to buy the fund while selling a matching basket of stocks (pocketing the difference in value) and if the fund is expensive relative to the basket then it is tempting to sell the fund while buying a matching basket of stocks (again, pocketing the difference in value). Remember for an ideal (not real) open-end fund there are two ways invest: give money to the fund and have them create a new share for you (by having them buy more of the underlying assets) or buy somebody else’s share of the fund on the market. And there are (in theory) two ways to cash-out: demand your share be redeemed from the fund or sell your share of the fund to somebody else.
An open-ended fund’s price can (and will) differ from the price of the underlying basket due to market friction. The fund may not always be accepting capital (so the buy may not always be possible) and may return capital even less often (or not at all). There are also opportunity costs (setting up the ability to be ready to buy/sell at any time), and actual trade costs. But the principle is arbitrage or even the potential for arbitrage sets the fund price (at least on the long run, Benjamin Graham: “In the short run, the market is a voting machine but in the long run it is a weighing machine.”).
A closed-end fund is a stranger beast (and in fact pre-dates the open-end funds in the United States). Essentially a closed-end fund is like an open-end fund except: after the initial offering (and purchase of underlying assets) no new shares are created (no new money is allowed into the fund) and no shares are redeemed (nobody gets their money back out directly). The only way to get into such a fund after its creation is to buy a share of the fund from a third party who wants to sell theirs. The only way to get your money out is to sell your share to somebody else.
At first blush a closed-end fund sounds like an open-end fund with just far fewer opportunities to enter and exit. But these seemingly small changes in features mean there is no arbitrage setting the price (except the opening of the fund and the thread of a final liquidation of the fund). Unless you trigger the fund liquidation (essentially a hostile takeover and cutting up the fund company for the underlying assets) the underlying assets may as well not exist. For all you know the fund managers could have burnt or embezzled them all when the fund was created. You can’t tell the difference because they never given them back (because they probably will poison pill, burn on embezzle away all assets if you do manage to trigger fund liquidation). So a closed-end fund behaves a lot like an open-end fund, yet doesn’t have low-friction arbitrage and is indistinguishable from having destroyed its underlying assets (instead of holding them).
In a crude sense this is similar to Bitcoin. With Bitcoin any electricity used in Bitcoin mining is in fact wasted or destroyed. Any hardware (CPUs, GPUs, FPGAs, ASICs, fully custom chips) depreciates rapidly (especially to the degree it is customized to Bitcoin mining). So you can (with some effort) think of BitCoin as a semi-open end fund with rapidly increasing entrance fee. You can’t arbitrage and get the electricity back or un-depreciate obsolete specialized hardware. So the pricing, if it has any stability, would be stable for the same (unknown) reasons closed-end funds have historically tracked their underlying basket. Bitcoin in fact has one advantage over true close-end funds in being semi-open (allowing the creation of more shares/coins at a rapidly increasing level of expense): it signals demand. A ration actor who is considering investing $700 in hardware, electricity and management to produce a total of 1 Bitcoin could be convinced to instead buy a Bitcoin for a similar price. So Bitcoin has an enhanced version of what the markets call “price discovery” in that any public knowledge of proposed hardware or electricity spend can be used to signal demand and influence future price.
So I would say a Ponzi scheme is not the only plausible analogy to Bitcoin. Bitcoin could also be a semi-closed-end index fund with a forced price schedule (due to the designed rapid increase in mining cost per Bitcoin). I don’t know which analogy really holds, so I don’t know which of the alternate conclusions to draw (will Bitcoin crash, or will it follow a forced price increase schedule for a longer time).
Data Scientist and trainer at Win Vector LLC. One of the authors of Practical Data Science with R.