Can Bitcoin Markets Add and Subtract?
I find it fitting that the winner of the Nobel Prize in Economics for 2017, bitcoin’s annus mirabilis, is Richard Thaler. Thaler is known for his work on irrationality and the role of psychology in determining market movements.
In 2001, as the market approached the nadir of the dot-com bubble, Thaler co-authored a paper that asked the question: Can the Market Add and Subtract? The question certainly feels applicable to cryptocurrency markets. Nowhere is this more evident than in considering the bitcoin fork that occurred in August.
In this 2001 paper, Thaler and his coauthor Owen Lamont explore inefficiencies in the pricing of technology stocks. The basic idea that they explore is that when companies are split, the broken-up units should, in aggregate, be priced on par with the company pre-split. But that does not always happen.
The example of Palm and 3Com is the most well known and is fairly intuitive. Palm (remember Palm Pilots!?) was owned by a company called 3Com. In 2000, 3Com sold some of its stake in Palm to the public in an IPO (this is called a carve-out). 3Com also announced its intention to eventually spin off the rest of Palm to 3Com shareholders at a rate of 1.5 Palm shares for every 3Com share they owned. In other words, everyone who owned 1 share of 3Com would wake up the morning of the spin off with not only 1 share of 3Com but also 1.5 shares of Palm.
Intuitively, the following inequality should therefore have been true on the day of the carve out:
Price 3Com > Price Palm x 1.5
This is assuming that there is some fundamental value to 3Com in addition to its stake in Palm. Since equity cannot be priced negatively (and really rarely trades at zero), this is reasonable.
But this is not how it went down.
In the words of Lamont and Thaler:
“The day before the Palm IPO, 3Com closed at $104.13 per share. After the first day of trading, Palm closed at $95.06 a share, implying that the price of 3Com should have jumped to at least $145 (using the precise ratio of 1.525). Instead, 3Com fell to $81.81. The stub value of 3Com (the implied value of 3Com’s non-Palm assets and businesses) was minus $63. In other words, the stock market was saying that the value of 3Com’s non-Palm business was minus 22 billion dollars!”
The IPO was widely publicized, so it’s unlikely that there was an informational inefficiency at play. Thaler and Lamont explore the possibility that the overpriced stock was merely too expensive to short for arbitrage to make economic sense — but as they note, this does not explain why there were buyers of the overpriced stock. This mispricing lasted for weeks, despite it being publicized in the media.
What does all of this have to do with cryptocurrency and hard forks?
The closest parallel to a hard fork in existing financial markets is a stock spin off. In both cases, a pre-announced event occurs wherein the owner of a single asset ends up not only with the initial asset but also an additional asset.
Last August, the bitcoin network experienced a major fork, testing the rationality of the crypto market. The results were unsurprising (the market is irrational), but noteworthy nonetheless. This fork resulted in owners of bitcoin not only possessing the original, but also a new asset called Bitcoin Cash.
The equality that should intuitively have held true for the Bitcoin Cash fork, then, is the following:
Price BTCPreFork ≈ Price BTCPostFork + Price BCH
Just before the fork actually occurred, the price of bitcoin should have roughly reflected the price of the two resulting assets. There are some explanations for why the price might have deviated: there may be some discount in the price to reflect the risks of the impending fork or the possibility that the fork would not actually occur. But broadly, the immediate pre-fork price of bitcoin should have been about equivalent to the collective prices of post-fork bitcoin and the newly created Bitcoin Cash.
This is not about whether forks are accretive. Whether you support cash, or gold, or classic, or core — undeniably a fork results in the existence of two assets with some value. Efficient market hypothesis would say that the collective value should approximate the value of the single parent asset just prior to the event.
But this is not what happened. Bitcoin barely moved on the day of the fork. Yet Bitcoin Cash closed the day with a market cap of over $6 billion. This $6 billion in value emerged from one moment to the next… at the moment that the fork occurred. Even assuming the market viewed the event as accretive, it was foreseeable for hours, days, weeks, before it happened. Why was none of this priced in?
The only thing to do is return to Thaler:
“To explain that, one needs investors who are (in our specific case) irrational, woefully uninformed, or endowed with very strange preferences.”
Thanks to another UChicago affiliate, John Loeber, for providing feedback and to Shelley Pearson for bringing the Palm / 3Com example to my attention.