Table of Contents
Perpetual futures, often called “perps,” are very popular in the crypto markets. Many people extoll the virtues of perps. There is even a recent Bloomberg article arguing this innovation should come in to traditional finance. Here we are going to argue that perpetual futures are more like a beginner-level product.
Rather than being innovative, perpetual futures are simply more convenient because they hide some of the realities of trading. In crypto markets, this hides important costs from traders in ways that do not exist in traditional finance.
There are idiosyncratic areas in traditional finance that could usefully employ perps. But they are the exception rather than the rule.
Convenience is great. And finding the right way to package up a risk for convenience is a valuable skill in finance. But hiding what is really going on isn’t a step forward. It’s difficult to understand how a perp product could be made consistent with the general principles of disclosure and transparency required in retail markets. Let’s take a look.
All Positions Have Carry
All positions have a “cost of carry” to hold. If you own Turkish Lira paying 10% and are short USD paying 4% you have what is known as a “positive carry” position earning 10-4=6%. These rates are annualized, so if the spread was instead 3.65% you’d be earning 1/100th of a percent every day.
The same applies to stocks, bonds and commodities. If you own a bond earning 8% or a stock with a dividend yield of 8% in a currency with a deposit rate of 3% you have a 5% positive carry. For financial products like stocks, bonds and currencies we can work out the carry rate from some other market. Maybe this is the stock’s dividend yield, or the currency’s deposit rate or the bond’s yield. But it’s something.
This gives us a thing called a “forward price.” If your position is marked at 100.0 today and has a positive carry of 12% the forward price in 1 month is 99. Why 99? It is expected to earn 12% for one month on 100: 1%. Why isn’t it 101? Surely the owner would love to sell at 101 in one months time. But consider the buyers position. They can buy at 100 today. Then they’ll earn that 1 unit over the next month. So if they are only going to get possession in a month they will pay just 99 today.
What is essential here is that both parties are indifferent between trading at 100 today and 99 in a month. That tells us we have the correct forward price. For a commodity like oil or gold we can’t look up a coupon or deposit rate. What we do instead is look at forward prices and work out what the carry rate is. Why? Because oil takes months on a ship to get to it’s next destination. It makes a lot more sense to trade oil for delivery today vs two months from now than it does to quote some kind of oil “yield curve” which pretends oil can arrive or depart on any given day.
Of course mathematically these things are equivalent but we are not going to get into that level of detail. What matters is that we can convert funding markets into a forward prices and vice versa.
Serial Futures Make This Obvious
Traditional futures expire along a strip of dates. Maybe at the end each month, or quarterly, or on some other fixed schedule. But they correspond to a strip of future dates for delivery.
And because there is a fixed expiration date, anyone that wants to hold a position longer than that expiry must “roll” their position. If you are long, this means selling the near-expiry contract and then buying the next one out. If the underlying position is positive carry, as we worked through above, we expect the next one out to be trading at a lower price. Thus the “positive” carry. If the carry is negative you’ll be buying back at a higher price.
This is, admittedly, more than zero work. And many people prefer perps because they can skip this step. But — and this is essential — this process is within a trader’s control and transparent. As we will now see perp traders still experience carry — it is just less transparent and harder to hedge.
Perp Funding Hides It
The “funding” process on perps works by settling up the difference between the perp and the underlying periodically. Here are explanations from BitMEX and Binance.
Let’s just assume there is a 1 hour period between these funding rounds. Further let’s assume we are long a contract on XYZ/USD where XYZ yields 10% and USD yields 4%. We expect there to be some kind of 6% positive carry here but where is it?
Say the XYZ/USD spot and perp are sitting at 100.0 right before the funding process kicks off. What are our choices?
- First we can just buy the XYZ/USD perp at 100.
- Or we can buy XYZ vs USD at 100, earning 10% for an hour on our XYZ and paying 4% on our USD.
If both cost 100 it is pretty clear the second strategy is better. If we do that we earn an extra 4%/(24*365)=0.046%. Yeah that is not much but this is per hour.
What should happen is the perp should trade that little sliver below 100. Then we are indifferent between buying or selling the futures and the spot.
But in practice the exchange we are trading on snaps the spot and futures prices at the same time and assess some kind of charge. This presents two problems:
- Per-period funding is likely to be tiny and difficult to track.
- There is little, if anything, traders can do to hedge this risk.
Let’s take them in turn.
Obfuscation
Your funding is hidden inside a string of hard-to-monitor tiny adjustments. And note that it is possible with noisy spot and perp markets for the real carry hour-to-hour to deviate a lot from the level you expect. Almost certainly over a longer period of time it will average out – but there is no reason to expect the snap at any one moment to match. If you hold a position for 2 hours this “average” is taken over two numbers. If we were talking about government bonds then maybe things are stable enough such an average damps the volatility. Here? Not so much.
Further, in traditional markets nobody tracks spot oil price ticks against futures price ticks and interprets moment-to-moment shifts in the spread as changes in carry. In crypto we see massive swings in funding rates depending on market sentiment. Holding a perp for a short period of time exposes you, as discussed above, to the random sampling problems inherent in the observation mechanism. Worse still, holding such a position for longer exposes you to persistent large swings in funding rates.
The moment-to-moment changes are noise. Market participants will quote the spread saying things like “I will sell you 3 month forward at 2 over the spot” and negotiate on the spread. And some high-frequency trading firm will get quotes on that spread and try to wriggle money out of the random deviations in the spot vs futures prices. But those are deviations to be ironed out by arbitrage, and not to be taken seriously as changes in carry or funding.
End users never see this stuff. Not because it is hidden from them or some secret insider-only source of profit. End users do not want to deal with this mess. You know you need to roll a futures position so you watch the roll price and trade it. Your roll is locked in when you trade. This is entirely within your control, transparent, and simple to hedge. You can even lock in carry in the future! Say you hold a position expiring in March and want to lock in the May->June carry today because it looks good. You can keep holding your March position and then sell May + buy June. Then, whenever you end up holding a May position, it is pre-rolled.
Difficult to Hedge
When we have a strip of futures with different fixed expiry times each trader can decide to roll when they want. Yes, if you wait until the last moment they may be forced to trade. But these things are generally listed months, if not years, ahead. You may not like the price at which you are forced to roll but you can tell exactly what it is at any time and generally control your own positions.
Not so with perps.
By shoving all the complexity of this process in the funding calculations, we lose control. There is no way to “roll” a perp because it has no expiry date. We can mitigate our risk by selling our perp position before the funding and holding the underlying instead.
Or can we? Most people use perps for the leverage. They cannot hold the same sized cash positions because they do not have that much money. The “hedging solution” is out of reach for precisely those traders that make up the target market.
Convenience yield
There is no question perps are easier to use than traditional fixed-expiry futures. And sometimes easier is better. But here you are giving up not just the ability to carefully manage your own risk. You are also injecting noise into your carry and funding calculations.
Truthfully adding perps to the TradFi product suite is not difficult. Many derivatives in traditional markets resemble more-than-closely a perpetual futures position. Many banks’ systems have been able to handle such payoffs for decades. Those are not the problems. Rather the difficulties are:
- Why would any professional want this product?
- For a non-professional how can we satisfy disclosure requirements given the weird, somewhat backwards, way that perps get their carry?
The answer to 1 is easy: essentially no professional wants to give up their ability to manage risk. Issue 2 is more complex. Products embedding this sort of convoluted funding mechanism do exist in traditional finance. But they generally involve elaborate and visibly-struggling-to-be-transparent procedures precisely because the target audience ends up being non-professional and cannot reasonably be expected to unwind all of this mess themselves.
Many people will say this is condescending and merely gating access to sophisticated products. They will further argue finance should be democratized and allow people to make their own decisions. Fine. Generally we are big fans of allowing people to voluntarily waive rights and protections if they so wish.
But if that same person does not understand the shortcomings of perps they are asking to be taken advantage of. And that needs to be considered when drafting regulations.
TradFi perps are less of a problem
When there is a liquid funding market and perp exchanges are highly regulated these are smaller problems. We can rely on arbitrage trading to keep the cash vs perp spread in line with the underlying product’s funding rates. There is also likely to be less random noise in large liquid markets.
Schiller’s original paper on perpetual futures explicitly deals with cases where spot prices are hard to find and forward prices may not exist at all. Think real estate or GDP. If there are no forward prices then we are not really talking about a generic replacement for derivative finance and funding is not a major issue. There are cases where these are useful.
But for cases like money-markets, currencies and commodities we already have – and will always have –developed forward markets. Overnight indexed swaps already provide exposure to nicely averaged and compounded funding rates where a single leg’s underlying reference rates average hundreds of trillions or more in money-market transactions.
Perps can offer a convenient way for non-professional end users to access such markets. But to be fair these products will require a lot of disclosure work and controls around the funding mechanisms. For professionals that need to manage their funding risks perps offer nothing new beyond simply another end-client revenue opportunity.
But crypto has no funding markets!
Indeed. And this is the real problem. Because there is no government-bond equivalent here funding markets will never develop like they have in traditional finance for decentralized products. We could see proper funding markets for stablecoins and other tokenized real-world-assets where the connection is via a centralized and trusted party. But then we are just doing traditional finance on a new database.
We are not going to see funding markets with the size and liquidity of fiat currencies for decentralized tokens like BTC or ETH because nobody can offer a backstop like a central bank. Perps hide a lot the complexity in crypto that exists precisely because certain structures cannot be built. Why are funding levels so linked to sentiment in DeFi but not in TradFi? At least it part because there is nobody to stabilize funding rates. The high level of noise in perp funding–and the attendant increase in need to hedge such exposures – is a native feature of DeFi.
You can argue that is a good thing or a bad thing. And both sides of the argument have merit. What you cannot reasonably argue is that hiding this complexity from market participants is a good idea. More people should make an effort to graduate from perps to proper fixed-expiry futures.
