Hedging With Prediction Markets: What a Binary Contract Can and Cannot Protect

A prediction-market contract can hedge a real-world exposure only when the contract's payoff moves with the thing you are trying to protect. Because the payoff is binary — $1 or $0 — it is a step function, not a smooth curve. You are matching a jump, not a slope. That single fact explains both what these contracts do well and where they quietly fail.

They are good at insuring discrete event risk: a strike, a ruling, a default, an election outcome. They are poor at insuring continuous risk: a currency drifting, a margin being squeezed, a slow rise in input costs.

Two payoff shapes

A futures or options hedge has a payoff that changes with the size of the move. If you are long oil and short a futures contract, a $2 decline in crude produces a predictable offset. The hedge ratio is a slope, and you can adjust it as the exposure changes.

A binary contract pays nothing until a threshold is crossed, then pays $1. Its delta is near zero most of the time and effectively infinite at the threshold. You cannot delta-hedge it in the usual sense, and you cannot fine-tune the protection by trading the same contract. You can only choose how many digital jumps to own.

That is not a defect if your risk is itself a jump. If the question is "will the port close," a contract that pays on closure is a closer match than a continuous instrument that pays on the size of the disruption. The mistake is using a jump instrument for a smooth exposure and calling the residual "hedged."

Basis risk: the four mismatches

Basis risk is the gap between what the hedge pays and what you actually lose. In prediction markets it comes from four places.

Threshold mismatch. The contract pays if the event happens; your loss depends on how bad it is. A port closure that lasts three days and one that lasts three weeks trigger the same $1 payout but very different losses.

Definition mismatch. The market's rule may require an official announcement, a specific source, or a particular legal step. Your exposure may be triggered by the economic reality before the official process catches up.

Timing mismatch. The contract has an end date. Your exposure may persist after it, or materialize before the market is liquid enough to trade.

Correlation mismatch. The event can happen without hurting you, or hurt you without happening. That is the residual you cannot hedge away, and it is usually larger than traders assume.

A worked example

A business expects $50,000 of revenue from a port-dependent shipment before December 31. It estimates that a port strike before that date would cost $20,000 of that revenue, after mitigation. A Polymarket contract pays $1 if a strike occurs before December 31, and it trades at 0.18.

To offset the conditional loss, the business buys 20,000 Yes shares at 0.18, costing $3,600. The taker fee at the general rate of 0.05 is 20,000 × 0.05 × 0.18 × 0.82, or $147.60. Total outlay: $3,747.60.

• If a strike occurs: the contract pays $20,000. Net of the premium, the hedge gains $16,252.40. The exposure loses $20,000. Combined result: a loss of $3,747.60.
• If no strike occurs: the contract expires worthless. The business loses $3,747.60 and keeps the revenue.

The hedge converts an uncertain loss of either $20,000 or $0 into a certain loss of about $3,748. That is what insurance does. The premium is 7.5% of the exposure and 18.7% of the loss being covered.

Now the basis risk. If a strike happens but the shipment was already rerouted, the contract pays while the business has no loss — a windfall, not a hedge. If no strike happens but congestion raises freight costs by $6,000, the business is unhedged and has also paid the premium. The hedge matched the named event, not the full economic exposure.

The cost of protection

The premium is the price of certainty, and it should be compared with the loss it prevents, not with the probability alone.

In the example, 18.7% of the covered loss is expensive for a three-month policy. Annualized, it is well above most conventional insurance, which is why prediction-market hedges are usually reserved for risks that are otherwise uninsurable or for losses that would be ruinous.

Cheap protection exists only where the event is genuinely unlikely. A contract at 0.02 costs two cents per dollar of coverage, but it pays only in a narrow state of the world. If the event is likely, the market price already reflects it, and the "insurance" is expensive precisely because the risk is real.

There is also a cost that does not show up in the premium: liquidity. Hedging a large exposure means crossing the spread in a book that may be thin. The order book has no fixed size limit, but large orders move the price, so the last shares you buy cost more than the first. Size the hedge in tranches and check depth before you commit.

Correlation and the hedge ratio

For a binary hedge, the starting ratio is simple: number of contracts equals the loss you want to offset, conditional on the event. If the loss is $20,000, you buy 20,000 contracts. That covers the full conditional loss rather than the expected loss, and it is the right choice if the goal is solvency rather than expected value.

If the goal is to smooth expected costs, multiply by your probability estimate: 0.18 × 20,000, or 3,600 contracts. That costs less but leaves most of the loss unhedged in the event. Neither ratio is "correct" without stating the objective.

Correlation enters through the loss estimate. If the event and your loss are not one-to-one — because some strikes are short, or because your contract has force majeure protection — the conditional loss is lower than the headline exposure. Using the full exposure as the hedge size turns a hedge into a speculative position on the event.

Multi-outcome events and negative risk

Polymarket's negative-risk mechanism makes some hedges cheaper. In an event where only one outcome can win, one No share can be converted into one Yes share in every other outcome. That means betting against a candidate is economically the same as betting on the rest of the field, and you can hold the position in whichever form is cheaper.

The catch is that the conversion links all the legs. Your hedge is only as good as the event definition, and the "Other" outcome can absorb cases the named outcomes do not cover. Read the augmented negative-risk rules before using the conversion as a hedge.

When not to hedge

Do not hedge when the contract's rule does not match your exposure. A hedge with basis risk can be worse than no hedge, because it costs money and creates a false sense of protection. Do not hedge when the premium exceeds the loss you are preventing, unless the loss would be ruinous. Do not hedge when the book is too thin to establish and exit the position. And do not call a correlated bet a hedge: if you are buying the contract because you think it will go up, that is a position, not insurance.

A rule set

1. State the loss you are hedging, conditional on the event, in dollars.
2. Match the resolution rule, not the theme. Write the exact criterion next to the exposure.
3. Choose the objective: offset the loss, or offset expected loss. The contract count differs.
4. Compute the premium as a percentage of both the exposure and the covered loss.
5. Stress-test the basis: event happens and you are not hurt; event does not happen and you are hurt anyway.
6. Check resolution risk. A hedge that can resolve 50/50 is only half a hedge.
7. Enter in tranches and confirm exit liquidity before you need it.

The honest summary is that prediction markets are a useful supplement to conventional hedging, not a replacement. They cover named events with defined rules, at a price that reflects the crowd's probability plus a liquidity premium. Used for the right risk, they are precise. Used for a smooth exposure, they are an expensive way to feel safe.