Static Hedging: Set It and Forget It Risk Management

Delta hedging is wonderful in theory. You adjust your position continuously, and risk vanishes. In practice, it is messy. You have to trade at discrete times. Transaction costs eat your lunch. And for some contracts, like barrier options or anything with a discontinuous payoff, the required hedge ratios become absurd. You end up buying and selling enormous quantities of the underlying at exactly the wrong moments. Chapter 60 of Wilmott’s book introduces static hedging as the cure for many of these headaches.

What Is Static Hedging?

The idea is simple. Instead of constantly adjusting your position in the underlying (dynamic hedging), you buy or sell a set of more liquid contracts up front and leave them alone until expiry. You set it and forget it. The static hedge is designed to offset the risk in your exotic position, reducing the need for aggressive dynamic hedging.

This is not a replacement for delta hedging. It is a supplement. You put the static hedge in place to handle the big risks, and then delta hedge whatever residual risk remains. The residual is much smaller and better behaved, so the delta hedging becomes easier, cheaper, and less sensitive to model errors.

Wilmott describes three forms of static hedging: payoff matching, vega matching, and a more sophisticated approach using non-linear pricing equations.

Payoff Matching: Build a Replica

The first form is the most intuitive. Take your exotic option and try to replicate its value using traded vanilla options.

Wilmott uses an up-and-out barrier call as the example. This option pays like a regular call at expiry, unless the stock price hits some upper barrier first, in which case it is worthless. The trick is to build a portfolio of vanilla puts and calls that has the same value as the barrier option at expiry and zero value along the barrier.

Start by matching the payoff at expiry with one vanilla call. Then work backward in time. At each intermediate date where vanilla options exist, add another traded option to zero out the value at the barrier. You end up with a handful of vanilla options that approximately replicate the barrier option. Between the matching points there is residual risk, but it is much smaller than the original exotic’s risk. You delta hedge this small residual, and your life is much easier.

Vega Matching: Insensitivity to Volatility

The second approach comes from a different angle. The reason delta hedging is imperfect is mostly because we do not know the true volatility. If we did, delta hedging would work fine even at discrete intervals. So maybe instead of trying to replicate the exotic’s payoff, we should try to make our position insensitive to volatility.

You calibrate your volatility model to the market, price your exotic, then bump each traded vanilla by $1. If the exotic changes by $1.50 when a particular vanilla moves $1, sell 1.50 of that vanilla. Repeat for every traded vanilla. In theory, your portfolio is now insensitive to volatility changes. In practice, the deterministic volatility model itself is questionable (Wilmott calls it “scientifically dodgy”), so the hedge inherits those weaknesses.

The Real Power: Non-Linear Models

Both of the above methods are less than satisfying because they rely on models that may not be right. The rest of Chapter 60 is about static hedging when the governing pricing equation is non-linear. This is where things get really interesting.

We have seen several non-linear models in earlier chapters. Transaction costs make the Black-Scholes equation non-linear. Uncertain parameters (volatility in a range) give a non-linear equation. Crash modeling and optimal speculation strategies both produce non-linear equations. These models share a crucial property: the value of a portfolio is not the sum of the values of its parts.

In the linear Black-Scholes world, a long call plus a short call equals zero. In a non-linear world, a long call might be worth $6.89 and a short call might cost you $9.63 (because the worst-case volatility is different for each). The value depends on the sign of the position and on what else is in your portfolio. This is not a bug. It is a feature that gives us bid-offer spreads as a natural output of the model.

Reducing the Spread Through Hedging

Wilmott works through a clear example. Volatility lies between 20% and 30%. A 100-strike call, 6 months to expiry, stock at $100. Using the non-linear uncertain volatility model, the selling price is $9.63 (Black-Scholes at 30% vol) and the buying price is $6.89 (Black-Scholes at 20% vol). This spread is way too wide. No customer will accept it.

But suppose 90-strike and 110-strike calls are traded in the market at an implied volatility of 25%. Their market prices are $14.42 and $4.22 respectively. If you sell the 100 call and statically hedge by buying 0.5 of each of the 90 and 110 calls, the residual payoff is much smaller than the original call payoff. It looks like a small butterfly.

Now solve the non-linear pricing equation for this residual payoff. The short price of the 100 call becomes $8.74 and the long price becomes $7.36. The spread has shrunk from $6.89-$9.63 to $7.36-$8.74. This corresponds to an effective volatility range of 22-27% instead of 20-30%. The static hedge dramatically tightened the spread without changing our uncertainty about volatility.

The reason is simple: the residual payoff is much smaller than the original, so the non-linearity has less to work with. Less payoff means less sensitivity to whether volatility ends up at the high or low end of the range.

The Optimal Static Hedge

Here is where Wilmott asks the killer question. We used 0.5 of each hedging option because it looked symmetrical. But why not 0.4 of the 90 and 0.7 of the 110? Or some other combination? Since the model is non-linear, different hedge ratios give different values for our exotic option.

So there must be a best static hedge. The one that minimizes the selling price (if we are selling) or maximizes the buying price (if we are buying). This is the optimal static hedge.

For the selling case, the optimal hedge turns out to be buying 0.51 of each call, giving a marginal value of $8.73. Very close to the symmetrical guess. For the buying case, the optimal hedge is completely different: selling 1.07 of the 90 call and selling 0.34 of the 110 call. The marginal value is $7.47.

The optimal hedges for long and short positions are generally different. This makes sense: the non-linearity treats long and short differently, so the best hedge portfolio depends on which direction you are trading.

Automatic Calibration

One elegant consequence: you never need to “calibrate” to market prices. If you try to price a vanilla that is already traded, the optimizer simply buys one of that option as the hedge, leaving zero residual. The price automatically equals the market price. It calibrates to bid, offer, and even liquidity effects, all by construction.

Path-Dependent Options

Static hedging becomes more complex for American and barrier options because these contracts can change state during their lives.

For American options, there are two states: exercised and not exercised. The static hedge that is optimal while the option lives might not be optimal after exercise. Wilmott handles this by solving two coupled non-linear problems, one for each state, and then optimizing the static hedge over both.

For barrier options, the two states are “active” and “retired” (triggered). The key insight is that the static hedge carries over from the active state to the retired state. When the barrier is hit, the option disappears but the hedging instruments remain. Whatever hedge you chose must still be sensible after the barrier triggers. The optimization accounts for this.

For portfolios of barrier options, the number of states multiplies. Two barriers give four states. Three barriers give eight. For n barriers, you have 2^n states. But if the barriers are nested (not intermittent), there is a natural hierarchy that reduces the computational burden.

Why This Matters

Static hedging with non-linear models addresses several real problems simultaneously. It gives natural bid-offer spreads. It reduces model risk by making option values less sensitive to uncertain parameters. It automatically calibrates to market prices. And it provides a systematic framework for choosing which instruments to use as hedges and in what quantities.

The core insight is that non-linearity, which most people view as a complication, is actually an opportunity. It is what makes optimal static hedging possible and meaningful. In the linear Black-Scholes world, the hedge does not matter because the value of a portfolio is always the sum of its parts. In the non-linear world, clever hedging can genuinely create value by tightening spreads and reducing worst-case losses.

Previous post: Speculating with Options: The Non-Hedger’s Perspective

Next post: The Feedback Effect: When Hedging Moves the Market