Let’s Talk About Arbitrage – Bitcoin Futures Edition

Let’s get one thing straight right off the bat: this is not a post where I’m going to debate the merits of Bitcoin or discuss it in the comments.  Take your zealotry elsewhere.  I want to talk about the effect that the launch of Bitcoin futures trading may have on Bitcoin.

In case you missed it, the CME announced the imminent launch of Bitcoin futures. The cliff notes are:

  1. Each contract will represent 5 Bitcoin
  2. The contracts will be cash settled to the BRR (Bitcoin Reference Rate)
  3.  The CME will also publish the BRTI (Bitcoin Real Time Index)

Now, I’ve written a few posts about arbitrage before, in the case of the S&P 500 and also in a closed end silver fund, $PSLV – I would suggest checking out those posts if you’re looking for some arbitrage background.  I am not an expert in Bitcoin, but I am pretty well versed in the concept of arbitrage, which is a constant across asset classes.

First of all, the fact that the BTC Future is cash settled does not mean that it doesn’t affect the price of the underlying, in this case: Bitcoin.  In order for supply and demand imbalances in BTC-F to impact BTC, all we need is a mechanism to arbitrage the two products.   The easier that mechanism is, the easier the arbitrage is to perform, and the less we should expect the futures to trade out of whack with the spot (BTC).  In the S&P 500, for example, the futures settle to the opening print for all the index components on the 3rd Friday of the month.  It’s easy for any legit trading desk to “achieve” this settlement price, so it’s easy for said desks to trade dislocations between the value of the ES futures and the underlying S&P 500 index.  Again, for more on this, see the whole post I wrote years ago.

So how easy will it be to trade the Bitcoin settlement price?  the CME link above tells us about the BRR – the Bitcoin Reference Rate – which determines the futures settlement value:


CME’s partner, Crypto Facilities, gives us some more detail about the input exchanges:


To summarize, the BRR will be calculated by taking all of the trades on 4 different exchanges (Bitstamp, GDAX, itBit and Kraken), calculating the VWAP (volume weighted average price) for 12 separate 5 minute periods, and then taking the arithmetic average of those 12 VWAPs.  Now, that doesn’t sound at all trivial to me to replicate for an arbitrageur, but note that CryptoFacilities explicitly claims that the calculation is geared toward “real-time replicability in underlying spot markets,” so they clearly think that sophisticated traders will be able to replicate this benchmark.

Now how does trading of BTC-F (futures) impact BTC?  Let’s assume that the BRTI (their Bitcoin Index) accurately represents the price of Bitcoin.  When the price of the future (BTC-F) exceeds (ie: trades rich) its fair value (for simplicity, let’s just say the fair value is the BRTI, although we’d need to make some adjustments for the time value of money based on the margin required, and if we want to get real philosophical about it, we might have to adjust for any possible forks during the duration of the contract – but let’s not get ahead of ourselves), arbitrageurs will want so sell the BTC-F and buy the underlying BTC.  Excess demand for BTC-F translates into demand for BTC.   The arbitrageur can do this easily, I presume – she doesn’t have to trade the Bitcoin at the underlying BRR: that’s not something the arbitrageur worries about until settlement, and who knows – she might have traded out of her position by then due to arbitrage in the opposite direction.

Are you with me so far?  So excess demand for BTC-F has a mechanism to clearly flow through to underlying BTC.  How about excess supply of BTC-F?  If everyone wants to sell BTC-F, and the futures trade cheap to their fair value, the arbitrageur will need to buy the futures and sell BTC.  This might not be quite so easy: if the arbitrageur is already long BTC, she can sell the BTC that she is long, but is there a robust market for shorting BTC with the kind of size and security an arbitrageur would need to execute?   I am told that there are ways to short BTC on some exchanges, but I don’t think this shorting methodology is robust yet.   If the arbitrageur cannot easily sell or short BTC to hedge her futures position, she will be less willing to readily step in and “correct” the situation where BTC-F is trading cheap – where there is an excess of BTC-F sellers.   Thus, it seems likely that the arbitrage is more readily executable in one direction: to correct “rich” BTC-F futures – which means that excess demand for BTC-F will translate into upward BTC pressure more readily than excess supply of BTC-F would translate into downward BTC pressure.

Now, what happens if the arbitrageur doesn’t have the opportunity to trade out of her position?  In order for the “arbitrage profit” to be realized while minimizing execution risk, the arbitrageur will need to trade out of her position at the BRR (Bitcoin Reference Rate – see above) on the day of expiration:  trading on 4 exchanges, over a period of time, at what seems to me like a difficult-to-replicate average, but what the index providers say is an average designed to be replicated.   I guess that remains to be seen.  The harder the index is to replicate with real trading, the more “slippage” risk the arbitrageur has, and the more dislocation from fair value the arbitrageur will demand to initiate the trade in the first place.

As soon as it was announced that the CME would begin trading Bitcoin futures, you saw the usual suspects lament that this would give The Government / The Cartel / The Illuminati the ability to suppress Bitcoin prices.   Reality, I am guessing, will prove to be the opposite.   If an evil cartel wished to sell BTC-F relentlessly with no regard for profit, either the futures would trade cheap to fair value, or at some point arbitrageurs would step in and try to right the mispricing (putting pressure on the underlying BTC): I am guessing that many arbs are already long BTC currently, and if the price is right they could buy the futures and sell their BTC.  So could this crush the price of Bitcoin?  Maybe, temporarily, until expiration – then what happens?   Either 1) the evil cartel buys back their short BTC futures, in which case the arbitrageur will now be in position to sell BTC-F and buy BTC, reversing the initial downward price impact, or 2) the evil cartel lets his manipulative short futures expire: now the arbitrageur is left with a short BTC position which she would try to buy back at the BRR where her futures settle  – and the (initial downward) price impact is still reversed.

My guess would be that, just as $GLD enabled the masses to easily get gold exposure, which flowed through to the underlying price of gold and was a massive positive force for the gold market, synthetic BTC derivatives like BTC-F will, at least initially, result in speculators seeking long exposure from the product, which will result eventually in demand flowing through to the underlying BTC, via the mechanism I explained above.

So You Want To Understand Index Arbitrage

Misinterpreting PSLV’s Premium

Bitcoin Futures FAQ



As of November 1, 2017 the estimate of initial margin that CME Clearing would require is 25-30%, though this is subject to change