The Bitcoin ETF will be here tomorrow.  Well, not really. While it is true that ProShares will be listing the first US-listed, bitcoin-linked ETF on the NYSE tomorrow, with others expected to follow, BITO and its brethren will not truly be bitcoin ETFs. The ETFs will not be directly holding cryptocurrencies – they will instead be holding listed bitcoin futures.  This structure is likely to create additional hidden costs and underperformance for those who hold these ETFs.This is not meant to single out the proposed new ETFs – those detriments are common to all ETFs that hold futures. It is important to understand why.

It is quite clear that traders expect that the long-awaited availability of US-traded, bitcoin-linked ETFs will spur demand. This undoubtedly had to play a key role in bitcoin’s nearly 50% bump in just a month, as a 1-month chart of bitcoin indicates:

Bitcoin vs US Dollar, 1-Month Chart

(Click on image to enlarge)

Source: Bloomberg

It is also quite clear that smaller traders have also gravitated to bitcoin futures, as evidenced by the near doubling of open interest in the more newly-listed CME Micro bitcoin contracts:

Open Interest in CME Micro Bitcoin Futures, Weekly Data

(Click on image to enlarge)

Source: Bloomberg, CFTC

We can debate whether the market is properly anticipating the incremental demand that will arise from the newfound availability of a US-listed, bitcoin-linked ETF. That is a trading judgment, however. We would prefer to focus upon the structural elements of futures that will specifically weigh upon the new ETFs.

The key to the likely underperformance is linked to the shape of the futures curve. Bitcoin futures have been in contango, which describes the condition when futures with longer expirations trade at a premium to those with shorter expirations. This is the normal condition for most exchange-traded futures. When someone buys a future, they assume the obligation to pay the full value upon the expiration of the contract. Because futures are typically purchased on margin, the buyer can invest the excess funds. That rate of return is priced into the longer-term futures contracts, boosting their value.

The futures curve for bitcoin, displayed below, is very much in contango, and has been for months:

CME Bitcoin Futures Curves – Today (green), Prior Trading Day (orange), 1 Week Ago (blue), 1 Month Ago (red), June 22^nd (white)

(Click on image to enlarge)

While we can see the curves rise dramatically as the price of bitcoin rose, the curves remain generally parallel.   That means they remained in contango. Even as bitcoin was correcting sharply in June, the curve remained somewhat positively sloping. This is fine for futures traders and hedgers, but detrimental for potential ETF investors. Here’s why:

ETFs that buy futures eventually need to roll those positions. This is in contrast to a typical ETF that buys shares in the proper proportion to represent the index that it is attempting to replicate. Those shares can be bought and held, with transactions only necessitated by rebalancing, creation, and redemption. Funds typically require large creation and redemption increments, minimizing their occurrence. They also charge fees for creation and redemption, which should cover the bid/ask spreads and transaction costs for the securities. In contrast, a futures ETF is investing in derivatives that expire. Since the ETF is not allowed to take delivery of the underlying commodity, they must buy contracts that expire at a later date and sell the expiring contracts. The act of replacing nearer-term contracts with longer-term contracts is called “rolling”.

Unfortunately, it can become quite expensive to roll long futures positions in markets that are in contango. The roll requires selling lower-priced near-term contracts and replacing them with more expensive longer-term contracts. This means that the fund will be forced to replace its expiring contracts with a smaller number of long-term contracts unless it happens to have enough excess cash on hand to pay the additional price.   This will cause significant slippage. Using current prices, one November 2021 bitcoin future has a value of $314,575, while a December contract has a value of $316,475.If the current level of contango holds, which means that it will either cost the fund about $2,000 to roll the contract or the fund will only be able to purchase about 994 contracts for every 1,000 that it holds. Either way, that will hurt results. And if you think that sounds insignificant, bear in mind that the $2,000 cost annualizes to about 7.5%. That is a very significant drag! 

The only way this would not be a drag on performance would be if bitcoin futures slipped into “backwardation”, which is when near-term contracts trade at a premium. This can occur when the market perceives a short-term scarcity of a commodity. Yet this has not been the case for bitcoin. It may seem counterintuitive because one of the features of bitcoin is that the supply will ultimately be constrained. But the supply of futures is unconstrained. Futures contracts are opened and closed at the clearinghouse when two traders consummate a trade, and while traders have position limits, the clearinghouse does not.

There are always costs to rolling contracts. At a minimum, there is a bid/ask spread and commissions. Similar to when index ETFs rebalance, these expenses become a drag on performance. If this were the only cost, this would be considered part of the normal slippage. Yet we have just shown that the roll is likely to cause about 7.5% annual underperformance. That is a significant price to pay for the privilege of buying a fund to hold bitcoin futures on your behalf.

Does this invalidate the bitcoin futures ETFs completely? I believe that it can be a useful trading vehicle since it should do a reasonable job of replicating the daily moves in bitcoin. I would, however, be very wary of it as an investment vehicle because of the potential for systemic underperformance. And this of course sidesteps the question of whether the act of listing these ETFs has become a “buy the rumor, sell the news” situation.