In value over replacement, I explained a handy feature of option theory or really derivative pricing broadly is it models important aspects of decision-making explicitly. Especially opportunity costs:
In options, the opportunity cost can be thought of as the risk-free rate. But the risk-free rate is an instance of a category we call benchmark.
Professional investors separate alpha from beta by benchmarking to an index. We can get fancier into benchmarking by using factors. Private investments can be subject to hurdles. All of these ideas are focused on the same question:
What is the marginal contribution of an action or intervention?
This is important because that’s what we compare the marginal cost to.
I discuss later in the post that structured products appear to have a compelling pitch by a sleight of hand. They prey on our “VOR blindness” when they announce that they can’t lose money. It’s a sales tactic that ignore opportunity cost. If I return 1% in a world with a 5% risk-free-rate or even 3% inflation I’m just falling for a real vs nominal illusion. My capital has lost ground despite the 1% gain.
[If you would buy these structured notes but be unwilling to spend your bond coupons on index calls you either don’t understand that you are doing the same thing in principle or you are saying I’d rather pay someone to do this. Either is ok to admit, just have your eyes open.]
In that example, making opportunity costs explicit neuters an otherwise compelling pitch. But this can also work in reverse. We can make an uncompelling pitch favorable.
I’ll give 2 examples from the trading world.
✔️Zero or negative edge trading strategies
You’re running a large delta-neutral vol book. It spits off tons of deltas as stocks move around and gamma varies. You need to continuously hedge. Assume your explicit and implicit (ie slippage) hedging costs are 10 bps.
Imagine you came up with a mean reversion stat arb strategy that had zero pre-transaction cost expectancy. Hell, pretend the strategy has -5 bps of expectancy.
Instead of facing the “street” on all these delta hedges you could internalize them by allocating them to the stat arb book. The book is nominally delta-neutral but might lose less in expectancy over some assumed holding period than constantly turning your deltas over on the exchanges.
In other words, a strategy that would be a non-starter from an alpha POV is worth doing because it loses less than the alternative. The bogey is not “we need to make positive edge” but the explicit cost of otherwise paying 10 bps.
By properly benchmarking our decisions, we have turned an uncompelling pitch to a favorable one.
[Real-life observation: Index Trader A is short QQQ gamma and Trader B is long AAPL gamma going into earnings and the stock has a big move down. Trader B needs to buy AAPL and Index Trader B needs to sell it. AAPL is 9% of QQQ so if the index book is 11x bigger than the single stock book and they have matched greeks, then the buy and sell orders would happen to pair off. But even if they didn’t the deltas between the 2 books would be virtually paired off and the firm would hedge the residual in the open market. This saves transaction costs and slippage on gross position sizes.]
✔️Option “dissection”
In the clip below (excerpted from the large screencast), I explain how market makers use synthetic and arbitrage structures like condors and butterflies to “chunk” risk by themes. They can then remove such well-defined strategies from their main risk view so they don’t have to hedge the greeks they spit off.
It’s not alchemy as far as edge. It’s simply splitting your risk into those that can be safely cordoned off vs ones that need more management (open ended exposures to vol or higher moments of the distribution).
In a large options book you have all this open interest because you got paid to buy or sell it at one point. But now it’s just stale inventory. You have no view on it. It’s effectively random risk. But it’s expensive to flatten it all.
[Actually it’s incoherent to do so. The whole reason you have a business is because someone needs to warehouse the risk that arises due to a mismatch in timing and desire — hedge fund A wants to buy puts on Tuesday and mutual fund B wants to sell calls on Thursday. Your role is to trade with both of them and manage the vertical spread they’ve effectively foisted on you. Sequentially to boot. You were forced to be short vol for 2 days in the interim.]
Your risk management logic looks like:
a) I need to hedge
b) Hedging is expensive
c) Can I reduce hedging costs proportionally more than the risk of being less hedged adds?
In other words, your risk management criterion is against an inevitable cost. The hedge is not positive expectancy, it just needs to reduce risk at reasonable cost or reduce costs without adding risk.
Dissection reduces costs without adding risk (although it changes the shape of risk. If you are indifferent to the new shape it gives you choices and those choices need not have anything to do with being profitable — they just need to lose less.)