March 29, 2022
How do futures/forwards term structure in TradFi instruments differ from crypto and how does it drive other factors like volatility? There may be similarities in market structure but ultimately differing fundamentals create significant differences.
Understanding futures term structure - especially in regards to the options market can provide investors with a better understanding of where to deploy risk. When trading crypto options on venues such as Deribit, its easy to think in terms of ‘where do I think spot will be then’ rather than in terms of what are the drivers of this in terms of spot/futures premium/volatility.
Let’s take a look at how the forward curves have performed in TradeFi markets like crude oil, gold, and VIX futures first as a start.
Commodities have a cyclical supply and demand component. When oil futures hit negative prices in Q2 2020, supply overwhelmed the market, and there was no room to store crude oil for delivery. Tanks, tankers, and salt mines were full… taking delivery and finding appropriate storage became so expensive that the value of raw crude itself was outstripped by storage costs hence negative oil.
In today’s market, near-dated crude oil contracts are trading at a premium to deferred contracts due to supply constraints tied to the disruption of oil production in Russia. This has placed crude oil futures into what is known as backwardation.
Backwardation is the term used to describe the curve structure of a futures product when the near-term contracts are priced higher than the longer-term contracts. Backwardation is the inverse of “Contango, " the natural state for most futures curves (some exceptions in the commodities markets), and contango exists when near-term contracts are priced lower than longer-dated expiries.
How does backwardation correlate with the volatility of the asset?
Short-term futures and long-term futures will have different realized volatility profiles; this is in “addition” to the normal ATM implied volatility term structure.
Think about it like this. Production solutions have time to “come online” with long-dated expirations. Therefore short-dated futures will have the highest RV while long-term futures will have muted reactions, causing “futures-spreads to move around”.
This means that front-end implied volatility in oil futures is often much higher than back-end volatility - especially in times of stress - since the front-end vs. back-end futures-spread can move significantly.
The gold futures term structure is often viewed as a proxy to the treasury yield curve, and gold may also serve as a currency or a proxy for the U.S. Dollar. Being easy to store, small and non-toxic gold allows traders to assume possession from delivery easily. Except for rare physical demand bottlenecks, risk-free rates often dictate the gold futures term structure.
Although the implied volatility term structure continues to exist, the realized volatility of gold futures should be nearly identical between one and the other, as is not the case with oil or VIX futures.
Therefore in gold - your volatility curve exists not really driven by futures-spread or futures-spot volatility - but by the spot volatility or the expectations of spot volatility in the future.
The VIX future (30 day implied volatility of the SPX index) is often used as the barometer for equity risk sentiment. This term structure consistently flips between contango and backwardation in given volatility regimes. Should cash VIX exceed 30, the VIX futures term structure will almost always be in backwardation. Should cash VIX be below 15, the futures term structure will consistently be in contango.
VIX futures are driven by this axiom: “Volatility is a mean-reverting asset class.”
In response, long-dated VIX futures are best approximated by long term median/mean cash VIX - they are “anchored” - while “cash VIX” and short-term futures are driven by today’s current events, free to assume nearly any value.
In this way, VIX (and equity volatility) term structure is similar in that front-end, broader market stresses drive the front-end while the back-end is more anchored by this mean reversion thought process.
Here’s where BTC futures and the associated volatility landscape significantly differ.
First, the cost of capital in crypto differs drastically from TradFi. Should traders assume a “Cash and Carry” position in the futures market, their trade will be delta-neutral and reflect stable-coin “rates” more than anything. These rates (AKA “Cash and Carry”) have traded between -20% APY to + 50% APY for a 90-day basis.
Historically, this basis rate is driven by leverage demand. If the market is bullish, traders are willing to pay substantial borrowing costs for futures leverage. If the market is bearish, traders will increase short exposure and drive the basis into negative territory. These effects are highly correlated with underlying spot prices.
Think of Crypto as a “FOMO” market…
As the market goes higher, traders pile into the long trade with more feverishness.
Long-dated futures can exhibit even HIGHER RV than spot, as basis changes affect the forwards.
This causes the BTC implied volatility term structure to retain a contango structure EVEN in higher volatility environments.
Notice the chart above. Here we display the 30-day and 180-day ATM volatility term structure.
Notice that 30-day/180-day contango structures have been present in 45%/65% IV environments and IV 80%/100%. This is quite dissimilar from VIX.
Another fundamental principle to keep in mind is the siloed liquidity and thin markets. The futures basis between Okex and Deribit don’t necessarily need to converge. A differential can be acute in the short term and persist indefinitely as the potential for corrective cross-capital flows can remain obstructed.
The perpetual, also viewed as a short-term future, becomes the hedging vehicle of choice for options traders and market-makers. The most liquid contract with the shortest funding duration, the perpetual is often used as THE delta balancing vehicle.
Again, traders are siloed into their venues. For practical purposes, Deribit risk-management system cannot “see” a trader’s BTC collateral at Bitstamp as part of a hedge for a portfolio of short-calls. Therefore, traders must hedge locally.
Why do we care?
It means that perpetuals will exhibit a higher volatility than cash/spot prices.
We are consistently witnessing an RV of differential of about +5pts between the cash and perpetual markets.
Another way to think about it is that cash price and futures risk premium (i.e. cost of leverage) are positively correlated. This means we see an increased realised volatility in the future which of course should mean that longer-tenor options should be more expensive.
Where does implied vol. stand today?
The above chart displays Deribit’s “DVol” index (a 30-day IV methodology similar to CBOE’s VIX index) compared to VIX and OVX (S&P 500 and Oil 30-day IV).
Given the macro events driving today’s market, it’s no surprise that VIX and OVX are drastically higher year-over-year. It is surprising to see a nearly identical DVol index down nearly 50% from its May 2021 peak.
Does kicking Russia out of SWIFT, El Salvador making BTC legal tender, or Central banks leaning towards CBDCs have no impact on BTC volatility?
Even if BTC was completely unaffected, US dollars themselves should surely be subject to more volatility (and therefore dollar-denominated BTC).
We don’t need to predict volatility paths to find interesting volatility trades.
Revisiting the chart above, we have a flat term structure in a relatively low volatility environment. This is a sort of “no-mans-land.” Meaning we could either reprice higher, lead by a backwardation shape along the way, or quickly steepen into contango and send 30-day IV below 50%.
Looking to long-dated volatility for opportunity, we see a historically cheap ∆25 call skew (Call IV minus Put IV).
As mentioned earlier, BTC (and crypto in general) has characterized itself as a “FOMO” market, meaning traders become highly interested in longs as BTC rallies.
Historically, this has caused long-term option skew to be priced with large positive premiums.
Today, combined with relatively low ATM implied volatility, long-term option skew is negative, and 90-day futures basis is only about 4.5% p.a.
Meaning, all these factors combined, long-dated call options are cheap. As we’ve proven in the past couple of years, anything, I mean literally anything, can happen.
Should the wild animal spirits turn bullish on BTC again, expect these three factors to come back with a vengeance: Implied Volatility, Skew, and Basis. And more importantly these factors are positively correlated. Should volatility head lower, 30-day options can lead the drop as Contango reaches a 15pts gap between our select maturities.
To us, a combination trade between these factors looks appealing…something like “Diagonal” structures (i.e. buy back-end options funded by front-end options).
Example Implementation of Diagonal
Net Offer: ~6.7% (priced as below using Paradigm’s GRFQ)
You can buy the diagonal package with single click execution on Paradigm at a “better than screen” offer.