Why derivatives markets need oracles

I was in a meeting recently where the topic of oracles for derivatives markets arose, and someone proposed that an alternative to needing an external oracle for derivatives settlement might be to use the internal valuation of the product at expiry. They specifically suggested an auction, similar to the way the LSE runs intraday SETS auctions and closing auctions.

This absolutely won’t work for the final / expiry settlement of a derivatives market!

Since I have recently also had a colleague ask me to explain why this was the case, I thought I’d share here the analogy I gave him.

Imagine you and a mate make bet with each other on Trump winning the next election (imagine you’re betting against, your mate is for). So, just like other derivatives markets, this “contract” plays out over time. Let’s imagine the odds go your direction and it’s not looking likely Trump wins. Then, come election day the result is in. The bet should now expire and the bit of information critical to who won or lost is an external piece of information, pre-defined and agreed at the start of the bet. For something like this defining exactly which news source isn’t important, but for real world derivatives, the source of that settlement data is always clearly specified.

Now, let’s imagine that we didn’t use the external information to close the bet. Let’s say, we set it up so you and your mate could agree between each other who’d won by some kind of market mechanism. A simple market mechanism could be whoever puts up the most money on the for/against side of the bet in the final minute before the results are announced (or indeed immediately after). Now, the “winner” is decided by who has the most money, not by the actual result. Indeed, let’s say Trump lost (meaning you should rightfully win the bet), but your mate had a Grandma who’d just won the lotto and was willing to temporarily lend him a stack of cash; she could place that money on the “Trump will win” side of the bet and now it looks to the settlement engine that the market has voted that as being the outcome.

It’s important to remember the critical difference between spot and derivatives With derivatives, the value of a derivative is derived from the value of something else. And that something else’s value is critical to settling the market. For example, it could be the value of a weather station reading or the closing price of a stock on the LSE or the volatility of another market’s prices. Regardless, when deciding whether to enter a derivatives trade, the definition of the very thing you are buying and selling (e.g. a futures, an option) needs to contain precise definition of what the underlying is and what external data is used to ascertain the value of that underlying. Without that external data, the definition of your derivative is meaningless. If you were settling based off a final auction, then the derivative you’re trading isn’t in fact “whether Trump won”, it is rather “who has the most money to influence the final outcome”. So, definition of your underlying speaks to the very thing you are trading. This is in stark difference to spot markets, where you are trading and exchanging a fungible “thing” at the point of trade and the value of that thing is standalone.

Now, I’ve talked about settling at expiry, but what about interim settlements, which markets instigate to avoid credit risk problems? The price used for these interim settlements absolutely can use intra-market forces (like final traded price at a certain time). Auctions are a good option for this, as is the VWAP of trades over a period or a random close time, so as to make the ability to use speed to set the final price more difficult. The price used for these interim settlements is important because it sets the margins required for a trader to hold their open positions - and of course, traders would rather hold onto their money that put it up as margin. Note that the interim settlement processes can only use internal market pricing mechanisms, because the traders know that the “day of reckoning” is coming and so that any of the contracts they buy or sell during that price determination period will one day be liable for the expiry settlement… i.e. the contracts’ value is ultimately pinned to that.

I’ve only described simple derivatives here. Some derivatives have as part of their product definition, a series of readings of a series of underlying values. Either way, if the product definition of the derivative defines that, then when traders are trading they have “multiple days of reckoning”… and that’s OK, it’s upfront and understood. All of these will require external data at all those points.

To conclude I’ll say that the cool thing about derivatives is that there is so much scope in defining them - so they offer a broad creative expression of risk. As blockchains emerge as infrastructure on which to run global finance, it’s fun to think of all the things that could be underlyings and consider if hedging against that value is useful… for example, a Vega market could be made on the total tx per second that Vega is achieving over a defined period. This may help node operators or market makers hedge some revenue risk. But more on that kind of fun thinking at another time…

Would love any questions / comments / challenges to what I’ve written above… and if anyone. has better ways of explaining what I’ve attempted to, I’d love to hear that too!

1 Like

Is this true for, say, physically settled futures? Intuitively it would seem that the “oracle” in this case is unnecessary. I have to find the asset to deliver if I hold the contract at expiry, but as long as I can do so it seems like no external price source would be needed?

Therefore, along those lines, my thinking is that the oracle is, in the case of a cash settled future, really providing the rate needed to convert someone’s liability in, say, bitcoin to the settlement currency (perhaps US dollars) at a fair price.

Is that true or am I missing something? If so, is it generalisable for other products which are physically settled?

1 Like

Great point!

So, physically settled futures are defined as a pre-agreement to swap, at expiry, an amount of one asset in exchange for receiving a (pre-agreed) amount of a different asset. In this case, you don’t need an oracle to define the amount being exchanged at expiry, and in this way they are more like traditional “forward contracts”. And so, I agree they don’t need an oracle at expiry. In saying that, however, this also means there’s no need for any kind of settlement auction or other final price setting mechanism. The key thing to “grounding” this contract in real value is a robust physical settlement process.

Speaking of physical settlement processes, I just took a look at the CME physically settled FX futures - check out the length of time and steps required for settlement!!

So, I’ve traded physically settled derivatives in the power and gas markets, but it means something a bit different when you’re literally sending gas to a power plant because someone’s exercised a call option (usually as part of a large, structured product!). I’ve never traded exchange based vanilla contracts that are physically settled, and it seems to me options could also be physically settled :thinking: . Am looking forward to us getting physical settlement into Vega. It seems to make even more sense given physical settlement is quick, easy, free (unlike the onerous schedule CME outlines). And since the settlement is an exchange of one token for another… and one day those tokens may represent a claim on a physical asset… this also gets us another step closer to commodities type physical settlement powered on blockchains :smiley: