2. Margin, liquidity and the day you cannot get out

Derivatives 9: Derivatives in a real portfolio

The previous module measured the size of a position on the assumption that you can get in and out whenever it suits you. This one removes that assumption. An options position does not consist only of its payoff profile: it also consists of the conditions under which you will be able to unwind it, and those conditions get worse precisely on the day you need them.

Margin is a living figure

Level 6 introduced margin from the disaster side; here what matters is the planning side, because the usual error is not being unaware that it exists, but budgeting for it badly.

Clearing houses do not apply a percentage to the notional. They value the portfolio under a range of scenarios (the underlying rises and falls by various percentages, the implied rises and falls) and demand the worst result in that range. Two consequences follow:

• The requirement depends on the shape of the position, not only on its size: a spread with a well-defined long leg requires little, and the same apparent risk built as the ratio of level 8 requires several times more.
• The requirement is recalculated every day using two variables you do not control: the distance to your strikes and the level of implied volatility.

A short straddle requiring some 4,000 euros of margin per contract in calm conditions can require three times that after a session in which the underlying moves sharply and the implied jumps, without your having traded. And there is a third source: the intermediary or the clearing house can raise the percentages pre-emptively for everybody, ahead of a vote, a central bank meeting or a tense expiry.

The budgeting rule is uncomfortable and simple: the margin you must be able to pay is not today's, but the one you would be asked for with the underlying against you and volatility multiplied.

An option's spread is measured against the premium

In a large share the spread is a tiny fraction of the price. In options it is measured against the premium, a small number, so the percentage is of a different order.

An option with a bid of 1.10 and an offer of 1.30 has a mid of 1.20 and a spread of 0.20: 17% of the option's value. Buying by paying the offer and selling by hitting the bid costs that whole spread. On a contract of 100 shares that is 20 euros; on ten, 200 euros, before commissions.

That spread is not stable either. Whoever quotes you is a market maker who will immediately hedge whatever he buys from you or sells you, and his price embeds that cost. When volatility jumps, that hedge becomes more expensive and his inventory risk rises, so the spread widens. The same option can go from 1.10 against 1.30 to 2.80 against 3.90 in one morning: the premium has risen, but the exit toll has risen much faster.

A chain's volume is deceptive

Looking at the aggregate volume of a chain and concluding that the underlying has liquid options is a mistake. Two things that get confused daily are worth separating:

• Volume is the contracts traded in the session; open interest is the contracts alive. A strike can carry open interest of several thousand and not cross a single trade all week.
• That volume is not spread evenly: it concentrates in the nearest expiry, in the round strikes and in those close to the money. Outside that core there are cells with a published quote and nobody behind it willing to trade in size.

The quote on an out-of-the-way strike is published by an algorithm, and publishing it commits it to no more than the minimum: there may be two contracts at that price and the third may trade far worse. A position that was opened comfortably because the market maker wanted that inventory does not close the same way when the one who wants out is you and he does too.

Hence the rule: liquidity is checked at the specific strike and expiry you are going to use, not across the whole chain.

Legging out, the risk you add all by yourself

The combined positions of level 8 have a problem of their own. If the intermediary does not allow them to be closed as a single order, or if the combined price is bad, the temptation is to close first the leg that does have a counterparty. During that interval, which can last hours, you hold a different position from the one you thought: for example, a sold call without its bought call, that is, a naked one.

The bad day is a single day

What matters is not that they exist separately, but that they arrive together and from the same cause:

• Volatility jumps, so the margin demanded rises.
• Volatility jumps, so market makers widen the spread.
• Everybody wants the same thing at once, so the counterparty disappears at the strikes that now matter.

And there is usually a fourth piece: the intermediary restricts opening or raises margin that same day. So the margin call of level 6 and the impossibility of getting out at a good price are not independent risks that add up: they are the same event.

Real case: January 2021

During the January 2021 episode around a handful of heavily strained US stocks, several retail intermediaries restricted the opening of positions in those underlyings and raised their margin, in some cases as far as demanding the full notional, citing the greater requirements they were themselves receiving from the clearing house.

The spreads on those options widened to magnitudes where crossing them cost a very large part of the premium, and numerous strikes were left with no workable counterparty. Anyone with a combined position open ran into all three things at once: they could not add, they could not close at a reasonable price and they were being asked for more margin.

It is worth putting into numbers with an illustrative position, not taken from any real account. Ten spreads whose exit in calm conditions would have cost a spread of 0.20 per contract, that is, 200 euros, cost 1,500 euros with the spread widened to 1.50. If the net premium the month was meant to earn was 1,000 euros, the toll at the door has eaten the entire target and part of the capital. And that sum assumes there is someone on the other side, which in those days there was not always.