Thread by @duganist

Derivative relationships 101:

Swap vs. Future

You have N USD, you buy Y BTC at, you sell N USD worth of contracts at the trade price with notional = -Y BTC. Synthetic cash. With a future there's a bond-like fixed maturity, it's a longer term rate, with swaps there's payments.

There are a few positions one can take:

Swap Cash Carry

Future Cash Carry

Long Curve (Short Swap/Long Future)

Short Curve (Long Swap/Short Future)

Cash carry means collateral is all hedged into USD, like cash, and you carry it, for the swap payments for the premium decay.

Being short a swap gives one low basis risk (the swap is unlikely to rise much more than 35 basis points over spot, and mean reverts to 0 often) while exposing one to interest rate risk (the returns are variable, the rate can go below 0 and cost the short position funding).

Being short a future gives one high basis risk (the future can fluctuate between -250 and +900 basis points in its wild lifetime) while endowing a fixed-term yield assuming one can hold the position (which in cash carry, 1x collateral, it's simple, patience is it).

So when I was managing treasury and a cash carry fund, I would alternate between swap and futures to time the markets in a soft way. I wouldn't go all-in on either side until things were really turning. This way, I was able to outperform the passive return on the swap.

However sometimes I churned and churned multiple legs, so much volume, so much fees, to rebalance and ultimately was in the same yo-yo that a daytrader going long or short might be in. It pays to use limit orders for 1/2 of the legging, and it pays rebalance infrequently.

When the futures comes down, it yields immediately a lot of basis points. This happens sometimes in options, when IV has spiked and mean-reverts to be 40% lower over a week, and the effect generally occurs in tandem with futures premium melting.

Even if swaps are still positive, though more modest, it's unlikely that swaps can out-perform the decline in basis on futures after the peak-momentum moment of any given move in the markets that has a trendline and lasts a week ore more.

All of this is more relevant to those who are levered, collateral is unhedged because they're long or short the curve, even 1x, all the mistakes are paid for twofold.

Also the rewards: swaps are positive when markets are bullish, so long collateral, get swap, earn future curve.

So one can see how even a modest 1x curve position vs. being in a dollarized yield-management position, is already much more daring. If one were 2x short swaps, to hedge collateral, and 1x long futures, it's still more basis and interest rate risk.

Rolling into a futures hedge from a swaps hedge before the market dives, but maybe after swap yields have come down, means you earn more basis points, and you don't have to panic into cash as you unwind lossy swaps.

Instead you have a safe-ish basis *position.*

If you were going to buy swaps back at -20 bps and sell quarterly futures at -50 bps, as occurred in 2018, that's a tricky situation. If you rolled swaps at +10 bps and sold futures at +80 bps, holding futures as a hedge is a safer basis risk adjusted for that profit.

In position trading, one hopes to have a position they can afford to hold, either form cashflow, or ideally, a good V shaped bounce right after your buy, great timing btw.

Basis trading has similar entrenchments to it that give a portfolio confidence.

A futures position that has collapsed to 0% still gives one the option to wait and see what else is going on. A monthly futures contract that is counting down days to expiration has less and less probably basis range to deviate from 0, so rolling to swaps is a lucrative option.

If the market jumps back to life, and swaps are levitating, futures premium might go back to 30 bps with days left, but maybe swaps are also at 30 bps, baking in those sweet payments. This is the futures-expiration roll option in practice, in my experience.

Multiple futures makes things more interesting, like a gumbo, Calendar spreads <-> two futures often have less basis risk than swap to long-dated (where the basis and trading liquidity is more loose).

Calendar spreads between say Sept./Dec. can hedge the curve collapsing.

This is like curve trading 301 though.

Main thing in curve 101 is how much yield, wait out mistakes, and curve trading 201 is time the curve decently, and going multivariable is definitely 301. 401 is spreads of spreads, 501 involves arbitraging curve/box vega against options.