Barrier Options

You can already read a vanilla payoff, you’ve met delta, gamma and vega, and you have the Black–Scholes intuition that an option’s value is just the discounted, risk-neutral average of where it might land. A barrier option keeps all of that — and bolts on an on/off switch. The switch flips based on whether the underlying ever touches a chosen level, the barrier, at any point during the option’s life. That one extra condition changes everything: the payoff no longer depends only on where the stock ends up, but on the road it took to get there. Barrier options are the first path-dependent exotic most people meet, they’re cheaper than the vanilla they’re built from (you give something up for that discount), and they are the workhorse engine humming inside almost every structured note on the market. Let’s wire up the switch.

The four flavours (knock-in vs knock-out, up vs down)

Analogy. Two everyday switches, opposite wiring. A knock-in is the loyalty coupon that only activates once you cross a spending threshold — useless in your wallet until you’ve spent enough, then suddenly worth real money. A knock-out is the engine warranty that’s valid right up until the day you redline the tachometer — fully alive, protecting you, until one reckless moment voids it for good. One needs an event to wake up; the other dies the moment an event happens.

The definitions.

• A knock-out option starts alive and behaves like a vanilla — unless the underlying touches the barrier, at which point it is extinguished for good, even if the stock later wanders back into the money. It can only die once; it never comes back.
• A knock-in option starts dormant — it does nothing, pays nothing — until the underlying touches the barrier, which activates it into a live vanilla for the rest of its life. If the barrier is never touched, it simply never wakes up and expires worthless.

Now cross each of those with the direction of the barrier relative to today’s price. An up-barrier sits above the spot; a down-barrier sits below it. Two switch types × two directions × {call, put} gives the full family of eight named combos. The four switch-and-direction labels:

Stick a call or a put on each cell and you get the eight you’ll see quoted: up-and-out call, up-and-in call, down-and-out put, down-and-in put, and the rest. The name reads in order — direction, switch, right — so a “down-and-in put” is a put that wakes up if the stock falls to the barrier, and an “up-and-out call” is a call that dies if the stock rallies to the barrier.

One more piece of vocabulary you’ll hear on a desk: a rebate. To soften the sting of a knock-out, the contract sometimes promises a small fixed payment — the rebate — if and when the option gets knocked out. It’s a consolation prize: “sorry your option just died, here’s a little cash.” Many barriers have no rebate at all; when present it’s usually modest. For the rest of this lesson assume rebate = 0 unless we say otherwise.

In–out parity (the elegant identity)

Here’s the cleanest identity in the whole barrier world, and it falls straight out of the definitions. Take a knock-in and a knock-out with the same strike, same barrier, same expiry on the same underlying. On every conceivable path, ask: did the stock touch the barrier?

• If it did, the knock-out died but the knock-in woke up — so the knock-in is alive and the knock-out is worthless.
• If it didn’t, the knock-out stayed alive and the knock-in never activated — so the knock-out is alive and the knock-in is worthless.

Either way, exactly one of the pair is live, and the live one behaves like a plain vanilla. So owning both at once means you always hold exactly one live vanilla, no matter what the path does. That’s the identity:

Knock-in + Knock-out = Vanilla (same strike, same barrier, same expiry).

Rearrange it and you get the pricing shortcut desks actually use: Knock-in = Vanilla − Knock-out. Price any two of the three and the third is free.

Worked example. Say the plain vanilla call is worth $8, and the corresponding up-and-out call is quoted at $3. Then the up-and-in call must be:

No model required — just arithmetic and the no-arbitrage logic that the two halves must add back to the whole. If a market-maker quoted you the up-and-in at $6 while the vanilla and the up-and-out sat at $8 and $3, you’d buy the $8 vanilla, sell the $3 knock-out and the $6 knock-in, and bank a riskless $1 — which is exactly why they don’t.

Now make the logic physical. The island below is an up-and-out / up-and-in call, strike $100, up-barrier $120, with two toggles: knock-out vs knock-in, and a path that touches the barrier vs one that doesn’t. Walk through all four combinations deliberately — each click is a teaching moment:

1. Knock-out + “stays below barrier”. The stock drifts up to 115, never tags 120, finishes in the money — the knock-out is alive and pays its intrinsic value. Good, that’s just a vanilla so far.
2. Knock-out + “touches barrier”. The stock spikes through 120 early, then settles back to 110, still in the money — yet the readout says knocked out, worthless. That’s the gut-punch: it was right about direction and collected nothing, killed by a touch it later “recovered” from.
3. Knock-in + “stays below barrier”. Same gentle drift to 115, ends in the money — but the knock-in never woke up, so it pays nothing. Being in the money is irrelevant if the switch never flipped.
4. Knock-in + “touches barrier”. The stock tags 120, activating the option, then settles at 110 — now it’s alive and pays intrinsic.

Then notice the parity with your own eyes: on the breaching path the knock-in is the live one and the knock-out is dead; on the non-breaching path it’s the reverse. Exactly one of the pair pays on each path — add them and you’ve reconstructed the vanilla.

Why barriers are cheaper (and the down-and-out workhorse)

Parity doesn’t just give you arithmetic — it tells you why barriers are discounted. Since knock-in + knock-out = vanilla and neither price can be negative, each barrier option must be worth strictly less than the vanilla (assuming there’s any chance the barrier matters at all). The discounts on the two halves are mirror images: whatever value the knock-out gives up, the knock-in picks up, and together they refund the full vanilla price.

The intuition for each discount stands on its own:

• A knock-out is cheaper because it has an extra way to disappoint you: it can die. You’re holding a vanilla that might get extinguished mid-life, so you pay less for the chance it never survives to pay out.
• A knock-in is cheaper because it has an extra hurdle before it can ever help you: it might never wake up. You’re holding a vanilla that only exists on the paths that touch the barrier, so you pay less for the paths where it stays dormant.

The down-and-out call — the classic workhorse. Picture the most common trade in the family. You’re bullish, you want call upside, but you don’t want to pay full vanilla price. So you buy a down-and-out call: a normal call that gets knocked out if the stock first collapses to a low barrier. Your reasoning: “If the stock crashes down to the barrier, my bullish thesis is probably broken anyway — so I’m happy to forfeit the option in exactly that scenario, in exchange for a cheaper ticket today.” You’re selling off the part of the option’s value that lives in the “it crashed first, then recovered” paths — paths you don’t believe in — and pocketing the discount. That’s the whole pitch: cheap upside, surrendered only if the stock first falls apart.

Read the bottom row: price(knock-out) + price(knock-in) = price(vanilla) ties the two discounts together. The $2 you saved on the down-and-out is precisely the $2 the down-and-in is worth — value doesn’t vanish, it just gets allocated to whichever half owns the relevant paths.

The discontinuous delta & the reverse knock-out

Everything so far was about value. Now meet the reason traders lose sleep over barriers: their Greeks misbehave violently near the barrier.

Think about a knock-out call as the stock creeps toward the barrier. One side of that level, the option is a live, valuable thing worth real money. A hair on the other side, it’s dead — worth zero (or just the small rebate). So its value has to fall off a cliff over a vanishingly small price move. And delta is the slope of value against price; gamma is the curvature. When value plunges from “real money” to “zero” across a tiny interval, the slope and curvature don’t politely shrink — they blow up. Near the barrier a knock-out’s delta and gamma explode in magnitude, and the delta can even flip sign: the option that was getting longer as the stock rose suddenly wants you short as it approaches the trigger, because every step closer to the barrier now destroys value rather than adding it.

The nastiest specimen: the reverse knock-out. A “regular” knock-out has its barrier out of the money (e.g. an up-and-out call with the barrier below the strike — it tends to be cheap and gentle). A reverse knock-out puts the barrier in the money — for instance an up-and-out call with the barrier above the strike. Now picture the value profile near expiry: as the stock climbs above the strike the call gathers intrinsic value, getting more valuable the higher it goes… right up until the barrier, where it instantly snaps to zero. The option is most valuable just below the barrier and worthless just above it — a sheer cliff exactly where it’s worth the most. That’s the signature reverse-knock-out shape, and it is a genuine hedging nightmare.

When to use them

So when do you actually reach for a barrier instead of a vanilla? When you have a view on the path, not just the endpoint — and you want to pay less by surrendering the paths you don’t believe in. A vanilla makes you pay for every possible route to expiry; a barrier lets you carve out and sell back the routes you think are irrelevant, in exchange for a cheaper ticket. Two big use cases:

• Cheapen a hedge or a directional bet. The down-and-out call from earlier is the archetype: bullish upside at a discount, forfeited only if the stock first collapses (a path that, by hypothesis, kills your thesis anyway). A protective down-and-in put is another — cheap crash insurance that only activates if the market first falls to the barrier, so you only pay for protection in the regime where you’d actually need it.
• Embed them inside structured notes. This is where barriers really earn their keep. The knock-in put buried inside an autocallable note — which we’ll dissect in detail later — is what lets an issuer offer a juicy coupon: the investor is implicitly short a down-and-in put, so they pocket premium in calm markets and only eat losses if the underlying breaches a downside barrier. Almost every “capital-protected-but-not-really” note has a barrier doing the heavy lifting somewhere in its guts.

But respect the trade-offs, because they’re exactly the things that make barriers dangerous:

Putting it together

A barrier option is a vanilla with an on/off switch wired to whether the underlying ever touches a barrier during its life — which makes it path-dependent and cheaper than the vanilla it’s built from. A knock-out starts alive and dies on a touch; a knock-in starts dormant and wakes on a touch; crossing each with an up or down barrier and a call or put gives the eight named flavours. The keystone is in–out parity: a matched knock-in and knock-out sum to the vanilla, because exactly one of them is alive on every path — giving you a pricing shortcut (knock-in = vanilla − knock-out) and the reason each half is discounted. The down-and-out call is the everyday workhorse (cheap upside, surrendered if the stock first collapses). The danger lives in the Greeks: near the barrier a knock-out’s value is discontinuous, so its delta and gamma explode and can flip sign, and the reverse knock-out — barrier in the money — is a true cliff that desks can’t delta-hedge with stock, so they hedge with other options and charge for it. Use barriers when your view is on the path and you’ll embed them everywhere in structured notes — just respect path risk, the dead-on-a-technicality outcome, and the monitoring convention on the term sheet.

Next — digital (binary) options revisited and lookbacks, where we push the path-dependence idea further: payoffs that remember the best price the stock ever printed, not just where it landed.