Can You Weather the Bet?

Can prediction markets double as insurance against climate catastrophe? In this episode, Andrew Wilkinson and Patrick Brown of Interactive Brokers unpack how forecast contracts are reshaping risk management- where saying “yes” or “no” to a hurricane could mean millions on the line.

Summary – IBKR Podcasts Ep. 276

The following is a summary of a live audio recording and may contain errors in spelling or grammar. Although IBKR has edited for clarity no material changes have been made.

Andrew Wilkinson 

Welcome to today’s episode in which we’re going to discuss forecast contracts in the context of global climate and the application of insurance. Joining me to discuss is Patrick Brown, who joined the IBKR team earlier this year. Welcome Patrick. 

Patrick Brown 

Thanks a lot for having me. 

Andrew Wilkinson 

You’re very welcome. Let’s get straight into it.How can an organization use a forecast contract in a way similar to a disaster insurance contract? 

Patrick Brown 

Yeah, so just backing up for a second—prediction markets have long been recognized for their ability to provide information by aggregating and representing the collective wisdom of the crowd, where the crowd is typically represented by speculators seeking to profit on their own unique insight. But there are a bunch of applications of prediction markets that tend to be less discussed, that are more directly applicable to those actually participating in the markets. 

And so one of the clearest examples of that is the hedging of financial risk due to weather or natural disasters. So akin to disaster insurance contracts, at Forecast X they’re offering various prediction markets for natural disasters at local scales, such as Will a major hurricane make landfall in Miami-Dade County in 2025? And as you can imagine, there are numerous organizations that would face broad exposure to such an outcome that they might want to insure against. 

So those would include things like utility companies, commercial and residential real estate owners, businesses with a bunch of franchises in the area—say, theme parks—municipalities, and insurers and reinsurers themselves. For these organizations, buying a large number of yes contracts on a disaster strike like that event question would mean a large payout immediately after the event occurs. And so this is very similar to purchasing what’s called parametric insurance, which is triggered by weather parameters like recorded wind speed—as opposed to the type of insurance we’re all more familiar with, which is called indemnity insurance, that covers the direct demonstrable financial losses that an organization would incur. 

Do you mind if I give you just an example of how this might work? 

Andrew Wilkinson 

Go ahead. 

Patrick Brown 

So let’s imagine that the true probability of a major hurricane making landfall in a given county this year is 10%. That would mean that for a forecast contract, the price of an initial yes contract would be listed at about 10 cents. And so let’s suppose you’re an organization with exposure to that event and you want—just to throw out a number—$10 million in coverage should that occur. Then buying a yes at 10 cents would pay out a dollar for each contract, for a net of 90 cents per contract. So if you wanted a $10 million net payout if the event occurred, the organization would have to buy a little over 11 million yeses for a little over $1.1 million. 

Then, if the event occurs, you net $10 million. And if it does not, the yeses are worth nothing—so you lose $1.1 million. This is roughly analogous to an insurance arrangement where the organization paid $1.1 million in premium for $10 million in coverage, though the cost of the insurance is all paid upfront. I should note that for simplicity I’m making these numbers round, and this example ignores the additional attributes that would change the numbers slightly—one to the disadvantage of the participant and one to the advantage. 

To the disadvantage Forecast X charges a fee of 1 cent per each pairing of yes and no contracts.To the advantage Forecast X also invests the collateral and passes all of that back to the members each month, which can be used as an incentive coupon. For Interactive Brokers, the incentive coupon is currently at a 3.83% annualized rate, which accrues daily on the dynamic market price of the contract and then is paid out monthly. 

But that’s the general idea of how that would work from the perspective of someone wanting to hedge the risk. 

Andrew Wilkinson 

Gotcha. Okay. Since forecast contracts resolve based on a simple yes or no event trigger, they seem to share a lot of characteristics with traditional parametric insurance. Is that correct? 

Patrick Brown 

So there are shared advantages and shared disadvantages. The most commonly discussed advantage of parametric insurance is just the fast, automatic payout—they occur almost immediately after the event takes place. 

That’s in contrast to indemnity insurance, which is going to involve claims, paperwork, and loss adjustment that can be negotiated for very long periods of time and cause delays—even months or years after the event occurs. This is especially true after major natural disasters; we saw this during Hurricane Harvey, where claims adjusters were essentially overwhelmed. 

The other main advantage of parametric insurance is that it entails very broad coverage with no strings attached. If the event occurs and you get a dollar per contract, the money can be used for anything the organization deems appropriate. 

So you can cover less tangible risks that are not nearly as quantifiable as just physical damage. It could cover general business disruption or supply chain disruption, lost tourism, lost property values, lost tax base for public entities. These speed and flexibility aspects would be especially appealing to public entities like municipalities—especially if they’re interested in allocating funds for immediate disaster relief efforts in any way they deem necessary. 

Another kind of advantage is that it would eliminate concerns about relying on political systems. You don’t have to hope and wait for the federal government to declare a state of emergency in order to mobilize resources—you just get those funds no matter what. Disadvantages, though.. the main disadvantage of parametric insurance is referred to as basis risk, which is essentially the mismatch between the payout you would receive based on the weather parameter and the actual financial losses that are incurred by the organization. 

If a storm strikes just outside the defined region or falls slightly below the strength threshold, the contract would fail to trigger—even if the buyer suffers losses. But it’s important to note that basis risk cuts both ways An organization could emerge relatively unscathed from a disaster that did trigger yes and still receive the full yes value, of course. 

And I think it’s important to note that basis risk is a lesser issue for entities with broad exposure across a region. Local governments or insurers or reinsurers are more likely to find that trigger to be representative of their losses. 

Nevertheless, I think entities would generally consider any type of parametric insurance as a complement to other forms of indemnity-based insurance—not a replacement. And so forecast contracts would probably be viewed the same way. 

Andrew Wilkinson 

Patrick, why might somebody buy the no side of the contract? 

Patrick Brown 

So first, I think it’s important to point out that noes could be viewed as revenue downside insurance for a lot of entities—so noes themselves could be insurance. These would be things like retailers like Home Depot or Lowe’s—anything that’s going to provide services during a disaster roofing, siding, and window contractors, restoration and remediation contractors that deal with water damage, mold cleanup, debris removal, that type of thing. 

So that’s one thing—noes could be considered insurance themselves. But more generally, the buyers of noes would include basically anyone interested in receiving the risk premium offered by those purchasing the yeses. 

If we go back to our previous example where the true hurricane landfall probability is 10%, let’s say the yeses get bid up to, instead of 10 cents, 20 cents. That’s because organizations are willing to pay a premium to transfer risk off their books—which of course they are, or insurance would not exist. 

Essentially, those looking to hedge the risk are bidding up the yes prices, which is inherently pushing the noes down. So noes become undervalued, or they go on sale. In our example, someone could purchase a no at 80 cents when it’s really worth 90 cents. That means, on average, that person would receive 90 over 80—that’s a 12.5% return. And that could be in a very short time. If we’re talking about hurricanes, that could be purchased in, say, July, and it expires in December—so a 12.5% return in six months. 

The idea is this strategy could be scaled as well. You’d purchase a suite of undervalued noes that are uncorrelated with each other. That provides a way to achieve relatively high returns that are likewise uncorrelated with the broader business cycle, with mainstream equities and bonds. 

You could almost consider these noes as an asset class in and of themselves that diminishes volatility and downside risk of a broader portfolio. We’ve seen hedge funds, pension funds, sovereign wealth funds, and even university endowments interested in catastrophe bonds for these same attributes. 

Another entity that might be interested would be any type of impact-driven investors—like ESG-oriented funds or climate resilience funds—that could argue investing in noes provides the capital that facilitates climate resilience while still delivering a relatively good return. 

And the really interesting part about this is that those are all institutional investors, but Interactive Brokers would allow retail investors as well to purchase these noes. So it’s basically anyone interested in receiving the risk premiums for taking on the fractional exposure of some disaster risk can participate. 

Andrew Wilkinson 

Patrick, is there enough liquidity? Are people using forecast contracts in this way? 

Patrick Brown 

Yeah, so these markets do require broad participation to function effectively for this purpose. That’s why Interactive Brokers is working to raise awareness about the utility of these contracts for this purpose. Right now, there’s not a lot of participation, but we hope that getting the word out will increase liquidity rapidly. 

In the meantime, though, there is a first-mover advantage to these markets because newly emerging markets with low participation are certainly more likely to be mispriced, which represents a greater opportunity to capitalize on these mispricings and generate profits early on. 

Andrew Wilkinson 

Mr. Patrick Brown, Head of Climate Analytics at Interactive Brokers, thank you for joining me today. Thanks, Patrick. 

Patrick Brown 

Sure. Thanks for having me. 

Andrew Wilkinson 

And to the audience—don’t forget to subscribe wherever you download your podcasts from. Bye for now.