More precisely, it is possible to replicate the final value of a futures contract (on Bitcoin or any other asset) up front, by trading in the spot market, even though a futures contract’s payoff is uncertain until it expires. Whatever the future’s value ends up being at its expiry, you can put together in advance a portfolio of holdings that will have the same value simply by borrowing a certain amount of cash and using it to buy in the spot market a corresponding amount of the asset on which the futures contract is written; if you do this, the value of your holdings in cash (after paying interest on your borrowings) and in the asset are guaranteed exactly to equal the value of the futures contract at expiry, no matter what price the underlying asset ends up trading at then. So if there is any deviation in the value of the futures contract (at any point in its lifetime) from the aggregate value of the cash and assets in the portfolio that replicates its final value, then there is an opportunity for a pure arbitrage. Consequently trading by arbitrageurs bounds futures prices in a tight range determined by the variables that set the cost of putting together this replicating portfolio: the price of the underlying asset, the interest rate on cash and the yield on the asset (if any), the time until the maturity of the futures contract, and cost of transacting in these markets.

This arbitrage relationship is about as hard they get, so trading futures is tantamount to trading the underlying asset itself, since changes in price in the futures market will be translated directly via arbitrage into changes in price in the spot market for the underlying asset, and vice versa.

> by borrowing a certain amount of cash and using it to buy in the spot market a corresponding amount of the asset on which the futures contract is written;

So if 1 BTC@januar17 is priced 10 000 USD, you borrow 16 000 USD and buy 1 BTC. How does that help you?

> if you do this, the value of your holdings in cash (after paying interest on your borrowings) and in the asset are guaranteed exactly to equal the value of the futures contract at expiry,

Why? If the spot price at expiry is different from 10 000, let's say it sunk to 5 000, then you have to give back cash and you have 1 BTC that is now worth 5 000.

Could you illustrate how the equivalence comes out?

> So if 1 BTC@januar17 is priced 10 000 USD, you borrow 16 000 USD and buy 1 BTC. How does that help you?

You don't offer 1 BTC@january 17th for 10,000 USD if 1 BTC spot costs 16,000 USD.

> Why? If the spot price at expiry is different from 10 000, let's say it sunk to 5 000, then you have to give back cash and you have 1 BTC that is now worth 5 000.

You don't have to give back any cash. You sell the BTC, that gets you 5,000 USD, you give the 5,000 USD to the owner of the futures contract (cash settlement of the value of 1 BTC), they pay you the agreed-upon 10,000 USD (in reality, those two payments are netted, so they pay you 5,000 USD, that's it), and you pay back your 10,000 USD loan.

I do wonder how counterparty risk will be priced into the futures contracts. There've been a number of cases - Mt. Gox, Bitfinex, any Chinese/Korean/Zimbabwean exchange if you're not a resident of that country - where you can sell a Bitcoin for a nominal price on the exchange but not be able to withdraw that money into the fiat financial system. That's led to structural differences in prices between exchanges which don't get arbitraged away because potential arbitrageurs can't get their cash into or out of the exchange. What happens if that happens to institutional investors trying to settle futures contracts?

Right now 1 BTC on Gemini (the reference exchange for CBOE Bitcoin futures) is $16,600 and the futures price is $17,800 for a 1 or 2 month contract. Since the risk free rate is about 0.125% monthly and this spread is 7.2%, it looks like the market is pricing the counter-party risk at about 7% for one or two month durations.

That seems pretty high to me though, you can arbitrage on Coinbase and Gemini which are both US based regulated exchanges (they have licenses, insurance, and audits) with what I would expect to be less than 7%/month risk. Perhaps there just aren't many arbitragers with access to the futures market and knowledge of Gemini/Coinbase yet. I'd arbitrage it myself but I have no retail brokerage access to shorting the future contracts.

Further evidence the market just isn't acting efficiently yet: volume of 3.43k contracts today, which is only $61M. The 24 hour trade volume of the top 5 Bitcoin exchanges is $5 billion.

good points, the market is pricing the counter-party risk at about 7%

> Right now 1 BTC on Gemini (the reference exchange for CBOE Bitcoin futures) is $16,600 and the futures price is $17,800 for a 1 or 2 month contract.

> Since the risk free rate is about 0.125% monthly and this spread is 7.2%,

> it looks like the market is pricing the counter-party risk at about 7% for one or two month durations.

But if a lot of traders think this way, they may have high execution cost when the contract expires and they try to sell the spot. Specifically, a lot of selling pressure by this kind of trade will push down the spot price and the traders may not be able to sell the spot at the price of the futures cash settlement. So, there is a connection but only up to a point.

Of course there a probably traders doing other kinds of trades that have a different effects on the prices, but it's quite tricky to analyze, compared to a physically-settled contract.

Because it is cash settled, if multiple traders are using BTC to hedge their futures and all sell their BTC at futures expiration they'll move the price introducing risk into the transaction.

Excellent explanation - thank you.

But doesn't this sort of bypass the true scarcity of the actual bitcoins?

No.