The cleanest way I know to think about carry is to ask a boring question about a position. If nothing moves, if the price sits exactly where it is right now and never budges, what do I earn or pay just for holding it? That number, positive or negative, is the carry. Everything else is speculation about where the price goes. Carry is the part of the return you collect for showing up, and once you start seeing it that way, you notice it hiding in almost every market you touch.
What makes carry interesting is that it looks different in each asset class even though it is the same idea underneath. The mechanics change, the plumbing changes, but the question stays identical. You are always being paid, or charged, to keep a position open across time. So let me walk through where it lives.
The same trade in four costumes
In FX, carry is the interest rate differential. If you borrow in a low-rate currency and hold a high-rate one, you earn the gap, and as long as the exchange rate does not move against you, you keep that gap. This is the oldest carry trade there is, and it has funded and blown up more hedge funds than most people realize. The return if nothing moves is the rate spread. The risk is that the currencies move, and they tend to move hardest against you at exactly the wrong time.
In bonds, carry shows up as rolldown plus the yield you clip. If the yield curve is upward sloping, a bond you buy today will, all else equal, roll down to a lower yield as it ages, and a lower yield means a higher price. So you earn the coupon, and you earn the price appreciation from the bond aging into a cheaper part of the curve. Hold a five year note for a year on a normal curve and you are now holding a four year note, which typically yields less, so it is worth more. That rolldown is carry.
In commodities, carry is the roll on the futures curve. When the curve is in backwardation, the front contract trades above the later ones, so a long position rolls into cheaper contracts over time and earns positive carry. When it is in contango, you pay to roll, and that is negative carry that quietly bleeds a long position even when spot goes nowhere. Anyone who held a long oil futures position through a steep contango learned this the expensive way.
In crypto, carry is perp funding. Perpetual futures never expire, so instead of a settlement date pulling the price toward spot, there is a funding rate that longs and shorts pay each other, usually every eight hours. When the perp trades above spot, longs pay shorts, and the classic cash and carry trade is to go long spot and short the perp, collecting funding while carrying no directional exposure. The return if nothing moves is the funding rate. It is the crypto version of the exact same thing FX traders have done for decades.
Why they all break at the same time
Here is the part that took me too long to internalize. These four trades feel diversified because they sit in different asset classes, but they are not really diversified in the way that matters. They share a risk profile, and it is an ugly one.
Every carry trade earns a small, steady, positive return most of the time and then hands back a large chunk of it in a sudden, violent move. The payoff looks like picking up coins in front of a slow train that occasionally speeds up. FX carry unwinds when risk appetite collapses and everyone rushes back into funding currencies at once. Bond carry gets hit when the curve steepens sharply or rates gap higher. Commodity carry flips when a supply shock reprices the curve. Perp funding, which can run pleasantly positive for weeks, snaps negative and stays there when leverage flushes out of the system.
The problem is these unwinds often correlate. A broad deleveraging event hits FX carry, credit, commodity longs, and crypto funding basically together, because they are all, at root, short volatility and long the assumption that things stay calm. When calm breaks, they break as a group. So a book that looks diversified across four asset classes can turn out to be four expressions of one bet, and that bet is that nothing bad happens all at once.
Sizing it so one unwind does not end you
None of this means carry is a bad trade. It has paid well over long stretches, and the coins are real. It means you have to size it like the thing it actually is, which is a strategy that will, on some schedule you do not get to choose, give back a year of gains in a week. A few rules I try to hold myself to.
- Size to the crash, not the average. Ask what happens to the whole sleeve in a coordinated unwind, then set position sizes so that scenario is survivable and boring, not fatal. If a bad month can take the book down by half, you are too big.
- Do not count the four legs as four independent bets when you budget risk. Treat cross-asset carry as closer to one position wearing different clothes, and haircut your diversification benefit hard.
- Watch the level of carry itself as a warning. When funding, curve slope, or rate differentials get extreme, that is usually the market paying you more because the risk of an unwind is rising, not because you found free money. Rich carry is a sign to trim, not to press.
- Have an exit that does not depend on your judgment in the moment. Carry unwinds are fast, and the instinct to hold and collect one more funding payment is exactly what turns a drawdown into a wipeout. Pre-commit to trimming when volatility spikes or when your carry buffer stops covering realized moves.
- Keep the funding side liquid. In crypto especially, the cash and carry trade is only as good as your ability to unwind the perp leg when funding flips, and thin books at the wrong moment can eat the whole spread.
The practical workflow I like is to measure carry the same way in every market, as a plain annualized number, so you can compare a bond rolldown against a perp funding rate on one screen and see which leg is actually paying you and which is just risk you are being underpaid to hold. On the crypto side I lean on Blockcircle for the funding and basis data across venues, because perp funding is where carry moves fastest and where a stale number costs you the most. Then I size the whole thing against the bad day, not the good month, and I try very hard to be bored by it.
Carry rewards patience and punishes size. If you respect the crash risk and keep the sleeve small enough that a coordinated unwind is an annoyance rather than an obituary, it is one of the more durable things you can do with capital. The moment it feels like easy money is usually the moment to check whether you are just standing closer to the train.