Barrier Options: Knock-ins, Knock-outs, and How They Work

Barrier options used to be exotic. Now they are one of the most heavily traded option types. Chapter 23 goes deep on them. Simple enough to understand, useful enough to be everywhere, tricky enough to mess up if you do not pay attention.

Take a vanilla option and add one condition: if the stock hits a certain level before expiry, something happens. That level is the barrier.

The Two Main Types

Knock-out (out) options pay off like a regular option at expiry, but only if the barrier was never triggered. If the stock touches the barrier at any point, the option dies. Gone. Worthless.

Knock-in (in) options are the opposite. They only come alive if the barrier is triggered. If the stock never reaches the barrier, the option expires worthless, even if the payoff would have been huge.

Then we add direction:

• Up: the barrier is above the current stock price
• Down: the barrier is below the current stock price

Combine these and you get four basic types: up-and-out, up-and-in, down-and-out, down-and-in. Each can have a call or put payoff. That gives you eight basic barrier options.

Why Would Anyone Buy These?

They are cheaper. An up-and-out call costs less than a vanilla call because you give up the payoff if the stock goes too high. The closer the barrier to the current price, the cheaper the option. More risk of losing the payoff, bigger discount.

Knock-in options appeal to people who believe a certain level will be reached. A down-and-in put is a cheap way to bet on downside if you are confident the stock will drop to the trigger level.

Double Barriers and Rebates

A double barrier has both an upper and lower level. In a double knock-out, the option dies if either is hit. Sometimes a rebate (consolation payment) is paid when the barrier triggers, either immediately or at expiry.

How to Price Them: The PDE Approach

Here is the beautiful thing about barrier options. They are path-dependent (the payoff depends on whether the barrier was hit), but the path dependence is weak. You do not need an extra variable. The option value still depends on just two things: the stock price and time.

The option value V(S, t) still satisfies the regular Black-Scholes equation:

$$\frac{\partial V}{\partial t} + \frac{1}{2}\sigma^2 S^2 \frac{\partial^2 V}{\partial S^2} + rS\frac{\partial V}{\partial S} - rV = 0$$

The barrier shows up not in the equation itself, but in the boundary conditions.

For an up-and-out option with barrier at $S_u$:

• Solve the Black-Scholes equation for $0 \le S \le S_u$
• On the barrier: $V(S_u, t) = 0$ (the option is worthless there)
• At expiry: the regular payoff (say, max(S - E, 0) for a call)

For a down-and-out with barrier at $S_d$:

• Solve for $S_d \le S < \infty$
• On the barrier: $V(S_d, t) = 0$

If there is a rebate R, the boundary condition changes to $V(S_b, t) = R$.

The In-Out Relationship

An elegant identity: $V_{\text{in}} + V_{\text{out}} = V_{\text{vanilla}}$.

Either the barrier gets triggered or it does not. Either way, you get the vanilla payoff. So if you price the knock-out, you get the knock-in for free by subtracting from the vanilla price.

Closed-Form Formulas

When volatility is constant, there are explicit formulas. For a down-and-out call with barrier $S_d$ below strike E:

$$V_{\text{do}} = V_v(S, t) - \left(\frac{S_d}{S}\right)^{2r/\sigma^2 - 1} V_v\left(\frac{S_d^2}{S}, t\right)$$

where $V_v$ is the vanilla Black-Scholes price. A neat trick: if $V_{BS}$ solves Black-Scholes, then $(X/S)^{2r/\sigma^2 - 1} V_{BS}(X^2/S, t)$ is also a solution for any X. Similar formulas exist for all barrier types, but these formulas are rarely used in practice.

The Volatility Problem

This is where barriers get really tricky. Wilmott shows an up-and-out call priced with volatilities of 15%, 20%, and 25%. Near S = 100, the curves seem to intersect. Vega appears to be zero. Looks like volatility does not matter.

Deeply misleading. The real question is not “which constant vol?” but “what if vol moves around?” The worst case: low volatility near the strike (small payoff) and high volatility near the barrier (more likely to knock out).

The gamma changes sign. Near the strike, gamma is positive. Near the barrier, negative. When gamma flips sign, simple vega is meaningless.

Practitioners handle this with various workarounds:

• Two-volatility approach: Use the implied volatility from a vanilla option with the same strike for the payoff component, and the implied volatility from a one-touch option at the barrier level for the knock-out probability
• Implied volatility surface: Price the barrier option using the full surface of implied volatilities from all traded vanillas, ensuring consistency across instruments
• Stochastic volatility models: Let volatility itself be random

There is no standard model. Different desks use different approaches. Margins can be tight on barrier options because they are so common, but the market is not liquid enough to just read off prices.

The Hedging Nightmare

At the barrier, delta is discontinuous and gamma is effectively infinite. Your stock is at 119.50, barrier at 120, you are short an up-and-out call. Stock hits 120, option dies, you need to unwind your hedge instantly. But the stock might gap to 122 before you can react. That gap is called slippage.

Wilmott shows you can account for expected slippage by shifting the effective barrier slightly. The corrected boundary condition reduces the option value by the expected slippage cost.

Static Hedging: A Better Approach

Practitioners use static hedging instead: build a portfolio of vanillas that mimics the barrier option. One technique uses put-call symmetry and the reflection principle. For a down-and-in call where barrier equals strike, hedge with a vanilla put at the same strike. Perfect hedge when rates are zero, small residual otherwise.

When barrier and strike differ, hedge with puts struck at $S_d^2/E$ (the “reflection” of the call’s strike in the barrier). The common practical approach: buy a vanilla call matching your short up-and-out call. If knocked out, you keep a long call position that partially offsets losses.

Additional Barrier Varieties

The chapter covers several more exotic versions:

Intermittent barriers appear and disappear on a schedule. The barrier might only be active on certain days.

Parisian options require the stock to stay beyond the barrier for a specified time before triggering. This makes manipulation harder and hedging easier but increases dimensionality.

Soft barriers knock out gradually. If the barrier range is 100 to 120 and the stock reaches 110, only 50% of the contract is knocked out.

Reset barriers turn into a different barrier option when hit. One barrier leads to another.

Outside barriers (rainbow barriers) trigger based on one asset but pay based on another. Clearly a multi-factor problem.

The Takeaway

Barrier options are the simplest exotic options and the most liquid. They fit neatly into the Black-Scholes framework with modified boundary conditions. Closed-form solutions exist but are unreliable in practice because barrier options are extremely sensitive to volatility assumptions, especially when gamma changes sign.

The real skill with barriers is not pricing (everyone can do that). It is hedging. Delta hedging is nearly impossible through the barrier. Static hedging with vanillas is the practical approach, and the combination of static hedging plus an implied volatility surface is what most practitioners use.

If you can handle barrier options well, you have a solid foundation for everything else in exotic derivatives.

Previous post: Exotic Derivatives: Beyond Vanilla Options

Next post: Strongly Path-Dependent Derivatives: When History Matters