> New “efficiency features” regularly get introduced in ILS and written into contracts. One of the most transformative has been the use of parametrics. Unlike traditional insurance, which calculates payouts based on actual losses (what’s called indemnity), parametric insurance uses preset triggers to determine whether money gets released. During an interview, a London-based parametric expert gave me this example of a parametric scenario: If, during a hurricane, wind speeds off the Florida coast hit a predetermined trigger speed — say 175 mph — at a trigger distance of two miles offshore within a preset longitude and latitude grid, the payout is, in theory, immediate. No actual damage need occur; the trigger measures just need to be met.
Wow, that is absolutely begging for exploitation.
Whoever controls the authority reporting these figures now controls whether these bonds pay out. That in turn means that whoever holds those bonds has a huge financial incentive to manipulate what that authority says.
Put another way, if you're holding a bond that will cost you $100 million if a hurricane windspeed hits 175 MPH, then you have $99 million bucks that are worth spending trying to get the NOAA to say anything but that.
In Turkey the mandatory earthquake insurance for homeowners is owned by the government. It triggers parametrically (> 7 magnitude). In one case at least [1] the government office responsible for announcing the magnitude, AFAD, declared it lower than this threshold although other countries and Turkish research institues measured it as 7.0 . At the end the insurance payout was so much more limited even for people who kist their house and loved ones.
Sure. Getting reliable data all parties can agree on is one of the biggest challenges to parametric insurance. The data sources I hear about most often are US governmental agencies – and this is a problem, since the US political system is not stable enough to finance its governmental agencies reliably. (Most recently I recall concerns around budget and staffing cuts for NOAA and USGS.)
That said, sane practice for parametric insurance is to have redundancy in data sources, and an agreed procedure for settling differences in conclusions resulting from relying on either of them alone.
The same incentive exists for economic figures (inflation linked bonds) and market prices (cash settled derivatives), and it's seemingly not an issue, and those are far easier to game than physical measurements like wind speed or whatever.
> and those are far easier to game than physical measurements like wind speed or whatever.
I’m not so sure about that. I bet that we could tamper with an anemometer somewhere out in a field. Easiest is to put brushless motor with a propeller next to it and blow propwash on it. More technically difficult is to tamper with the signal between the sensor and the station, or MitM the station.
If you are careful and only modifying the measurments when the weather is already crummy they might not even suspect.
Sounds like you do see an issue after all. Is it obfuscated? Hidden in complexity? Is it artificial to appear as one thing while indirectly, as a side effect, leading to another?
I thought you meant it could be gamed in the other direction...
Technically, if all the bond-holders get together, they could spend up to $99M building an extremely powerful, extremely tiny fan pointed directly at the sensor.
Taking your point of view now makes sense on why to defund NOAA, fire everyone that's not going to toe the line, and then have those that will parrot the necessary info to keep from paying out. Make Weather Great Again, just doesn't lend itself to a hat though
Towards the end Mark Baum who has bought insurance on house index is frustrated why the index isn’t falling even as the house prices across America are plummeting and foreclosures are rising.
The obvious counter to that is you’re going to define parameters based on multiple observers. NOAA is not the only authority on Earth who can state what a hurricane’s wind speed is.
CTO at an ILS fund. Cat bonds are essentially securitised versions of fully collateralised reinsurance contracts where the premium is the coupon plus the return on collateral. A benefit being that you can trade them. They're not usually used for speculation as stated in the article - investors are typically pension funds looking for investments that are uncorrelated to traditional financial market risk. e.g. on a US hurricane exposed cat bond you may only lose money if a huge hurricane blows through Florida, no matter what credit and equities are doing.
It's true that a lot of the deal sourcing is relationship-driven, but there is a good amount of data-driven tech involved in overlaying the insured's past claims and underwriting data on top of simulated catastrophe model output, applying your own view of climate, vendor model adjustments, hurricane activity etc.
This reminds me of the predator hierarchy (for example, see Colinvaux's "Why Big Fierce Animals are Rare"): the reinsurers spread the risk from various insurers and for various catastrophes around among a pool of meta-insurers. But this pool is necessarily smaller than that of primary insurers, and their risks more likely to be correlated (catastrophes can cause other catastrophes, and multiply primary insurers can be affected by the same catastrophe).
For that matter, I'm also reminded of credit default swaps, and Lehrer's "We Will All Go Together When We Go."
One aspect worth pointing out is that ILS are transferring insurance risk outside of the insurance industry. Appetite has gone up and down but e.g. hedge funds would normally not be available to assume insurance risk otherwise.
It's correct that the number of reinsurers is smaller than that of primary insurers. But the risk born by reinsurers is less correlated, not more. Any given primary insurer has risk clusters (domestic market, line of business, etc.). If a large catastrophe happens in their domestic market they might go bust but what are the chances that it happens simultaneously to all markets globally?
Say you're a primary home insurer in the US. If a hurricane hits you might not have enough capital to rebuild all the homes. A reinsurer which is also covering Europe, Asia, LatAm, etc. is less likely to go bankrupt. The reinsurer can cross-subsidize and use the insurance premiums from other regions to pay out the claims from the US market. All that matters is that on average the loss probabilities and severities are estimated correctly.
And this is just using one line of business as example, reinsurers are covering property, casualty, life and health which add extra layers of diversification.
Reinsurance does not only spread risk by pooling multiple insurers, but also smears out the impact of catastrophes geographically and temporally: big events in one year, in one part of the world result in more expensive reinsurance all over the globe for a few years forward, as reinsurers collectively stock up on capital again.
So while locally catastrophes can cause other catastrophes, for the most part earthquakes in Thailand does not trigger wildfires in Texas. Nor does a hurricane in Florida one year cause more hurricanes in Florida the next year.
I like to think I’m somewhat intelligent, but there’s something I don’t understand here. The article cites an example of pandemic bond holders receiving a return of 40% over 3 years and these bonds being a useful way for the issuer to secure needed funds in the event of a pandemic. Unless a pandemic happens every ~8 years, isn’t this a ridiculous and unsustainable risk premium to pay?
The class B bonds paid roughly 11% over LIBOR, so about 40% over three years, against the risk of a viral outbreak for five different families, defined as At least two countries experiencing at least 250 fatal cases increasing over twelve weeks, so the trigger did not have to be as globally-significant as COVID-19 turned out to be. That’s a pretty aggressive coupon, but the chance of a regional outbreak was also pretty high.
Based on that description it would have been triggered by COVID-19, swine flu in 2009, and I think just missed out (depending on the fine print) on SARS in 2002. That's two or three in 18 years, so losing your money once every eight years is not far off the recent performance of this kind of bond.
> She is the author of “Investable! When Pandemic Risk Meets Speculative Finance – A Cautionary Tale,” from which this article is adapted.
So I think structurally, the conclusion here is that 'cat bonds are an example of how insurers can work with abstract risks, and so any risk (such as global pandemic) could be worked with this way', and the rest of the book then examines how people are trying to actually do so with pandemic risk.
So, it's betting not on the outcome, but specifics of the event. The analogy could be sports betting where you bet on a player making a particular individual score, not just the win/loss.
At the end of the day, if it prevents bankrupting companies, countries and the insurers when an event occurs then it is a good thing.
I know of a Australian-based insurer that only operates in the far North primarily for residents who can't get reasonable premiums to cover for floods and cyclones from the major insurers. They reinsure with a global company so paying out an event in Australia is offset with no events elsewhere.
Interestingly the frequency of cyclones is decreasing, but the strength is projected to increase with "global warming": Higher water temperature = more energy. So betting on the strength makes sense.
Reading this as somebody who tries to solve betting markets with data and python, yeah, you’re right, this is just prop bets but with less data due to lower frequency events.
I honestly think there’s EV in this - you can expect some sports guys to be booking tickets to Baden-Baden next year if they find this article.
As if pandemics weren't already political enough. Let's get large corporations, investment funds and billionaires involved and give them direct stake in declaring what is or isn't a pandemic, how many deaths have happened in a certain area, what was the cause of death etc. That should end well.
> New “efficiency features” regularly get introduced in ILS and written into contracts. One of the most transformative has been the use of parametrics. Unlike traditional insurance, which calculates payouts based on actual losses (what’s called indemnity), parametric insurance uses preset triggers to determine whether money gets released. During an interview, a London-based parametric expert gave me this example of a parametric scenario: If, during a hurricane, wind speeds off the Florida coast hit a predetermined trigger speed — say 175 mph — at a trigger distance of two miles offshore within a preset longitude and latitude grid, the payout is, in theory, immediate. No actual damage need occur; the trigger measures just need to be met.
Wow, that is absolutely begging for exploitation.
Whoever controls the authority reporting these figures now controls whether these bonds pay out. That in turn means that whoever holds those bonds has a huge financial incentive to manipulate what that authority says.
Put another way, if you're holding a bond that will cost you $100 million if a hurricane windspeed hits 175 MPH, then you have $99 million bucks that are worth spending trying to get the NOAA to say anything but that.
In Turkey the mandatory earthquake insurance for homeowners is owned by the government. It triggers parametrically (> 7 magnitude). In one case at least [1] the government office responsible for announcing the magnitude, AFAD, declared it lower than this threshold although other countries and Turkish research institues measured it as 7.0 . At the end the insurance payout was so much more limited even for people who kist their house and loved ones.
As wiki page mentions in notes section AFAD declared this a 6.6 magnitude earthquake although it was 7.0 . [1] https://en.m.wikipedia.org/wiki/2020_Aegean_Sea_earthquake
Sure. Getting reliable data all parties can agree on is one of the biggest challenges to parametric insurance. The data sources I hear about most often are US governmental agencies – and this is a problem, since the US political system is not stable enough to finance its governmental agencies reliably. (Most recently I recall concerns around budget and staffing cuts for NOAA and USGS.)
That said, sane practice for parametric insurance is to have redundancy in data sources, and an agreed procedure for settling differences in conclusions resulting from relying on either of them alone.
People have already done this with NWS weather equipment for federal farm drought insurance: https://coloradosun.com/2024/09/08/patrich-esch-ed-dean-jage...
The same incentive exists for economic figures (inflation linked bonds) and market prices (cash settled derivatives), and it's seemingly not an issue, and those are far easier to game than physical measurements like wind speed or whatever.
> and those are far easier to game than physical measurements like wind speed or whatever.
I’m not so sure about that. I bet that we could tamper with an anemometer somewhere out in a field. Easiest is to put brushless motor with a propeller next to it and blow propwash on it. More technically difficult is to tamper with the signal between the sensor and the station, or MitM the station.
If you are careful and only modifying the measurments when the weather is already crummy they might not even suspect.
Replace the anemometer cups with slightly larger ones.
Manipulation of reported economic numbers has been an issue in the past, see LIBOR.
There have some governments outside of the developed world accused of manipulating inflation numbers.
There have been allegations of the same in well-developed economies as well.
Those well-developed economies' data is mostly corroborated by proxy measurements (like light levels, copper/electricity consumption etc).
> and it's seemingly not an issue
Sounds like you do see an issue after all. Is it obfuscated? Hidden in complexity? Is it artificial to appear as one thing while indirectly, as a side effect, leading to another?
The Libor scandal is not an issue?
I thought you meant it could be gamed in the other direction...
Technically, if all the bond-holders get together, they could spend up to $99M building an extremely powerful, extremely tiny fan pointed directly at the sensor.
... or purposely get sick and start coughing in crowded places.
Taking your point of view now makes sense on why to defund NOAA, fire everyone that's not going to toe the line, and then have those that will parrot the necessary info to keep from paying out. Make Weather Great Again, just doesn't lend itself to a hat though
Goodhart's law continues to ring true.
This reminds me of a scene from The Big Short.
Towards the end Mark Baum who has bought insurance on house index is frustrated why the index isn’t falling even as the house prices across America are plummeting and foreclosures are rising.
Maybe something fishy was going on.
The obvious counter to that is you’re going to define parameters based on multiple observers. NOAA is not the only authority on Earth who can state what a hurricane’s wind speed is.
So if it is 174 mph insurance pays nothing? After a hurricane destroys a town.
It depends on the details of the product. In theory you could get a product that doesn't have a cliff and it will be priced accordingly.
These are basically equivalent to options in finance.
If you can prove the data was altered or tampered with you'd have a solid fraud case though I imagine proving it is the difficult part.
CTO at an ILS fund. Cat bonds are essentially securitised versions of fully collateralised reinsurance contracts where the premium is the coupon plus the return on collateral. A benefit being that you can trade them. They're not usually used for speculation as stated in the article - investors are typically pension funds looking for investments that are uncorrelated to traditional financial market risk. e.g. on a US hurricane exposed cat bond you may only lose money if a huge hurricane blows through Florida, no matter what credit and equities are doing. It's true that a lot of the deal sourcing is relationship-driven, but there is a good amount of data-driven tech involved in overlaying the insured's past claims and underwriting data on top of simulated catastrophe model output, applying your own view of climate, vendor model adjustments, hurricane activity etc.
And I thought I was the only ILS guy reading HackerNews..
Actuary here. I don't work in ILS but I'm familiar with what you're talking about.
Me too. Wouldn’t say my comments make that obvious. But in/around the space too - portfolio manager.
TwelveSecuris, London!
This reminds me of the predator hierarchy (for example, see Colinvaux's "Why Big Fierce Animals are Rare"): the reinsurers spread the risk from various insurers and for various catastrophes around among a pool of meta-insurers. But this pool is necessarily smaller than that of primary insurers, and their risks more likely to be correlated (catastrophes can cause other catastrophes, and multiply primary insurers can be affected by the same catastrophe).
For that matter, I'm also reminded of credit default swaps, and Lehrer's "We Will All Go Together When We Go."
One aspect worth pointing out is that ILS are transferring insurance risk outside of the insurance industry. Appetite has gone up and down but e.g. hedge funds would normally not be available to assume insurance risk otherwise.
It's correct that the number of reinsurers is smaller than that of primary insurers. But the risk born by reinsurers is less correlated, not more. Any given primary insurer has risk clusters (domestic market, line of business, etc.). If a large catastrophe happens in their domestic market they might go bust but what are the chances that it happens simultaneously to all markets globally?
Say you're a primary home insurer in the US. If a hurricane hits you might not have enough capital to rebuild all the homes. A reinsurer which is also covering Europe, Asia, LatAm, etc. is less likely to go bankrupt. The reinsurer can cross-subsidize and use the insurance premiums from other regions to pay out the claims from the US market. All that matters is that on average the loss probabilities and severities are estimated correctly.
And this is just using one line of business as example, reinsurers are covering property, casualty, life and health which add extra layers of diversification.
Reinsurance does not only spread risk by pooling multiple insurers, but also smears out the impact of catastrophes geographically and temporally: big events in one year, in one part of the world result in more expensive reinsurance all over the globe for a few years forward, as reinsurers collectively stock up on capital again.
So while locally catastrophes can cause other catastrophes, for the most part earthquakes in Thailand does not trigger wildfires in Texas. Nor does a hurricane in Florida one year cause more hurricanes in Florida the next year.
I like to think I’m somewhat intelligent, but there’s something I don’t understand here. The article cites an example of pandemic bond holders receiving a return of 40% over 3 years and these bonds being a useful way for the issuer to secure needed funds in the event of a pandemic. Unless a pandemic happens every ~8 years, isn’t this a ridiculous and unsustainable risk premium to pay?
The class B bonds paid roughly 11% over LIBOR, so about 40% over three years, against the risk of a viral outbreak for five different families, defined as At least two countries experiencing at least 250 fatal cases increasing over twelve weeks, so the trigger did not have to be as globally-significant as COVID-19 turned out to be. That’s a pretty aggressive coupon, but the chance of a regional outbreak was also pretty high.
Based on that description it would have been triggered by COVID-19, swine flu in 2009, and I think just missed out (depending on the fine print) on SARS in 2002. That's two or three in 18 years, so losing your money once every eight years is not far off the recent performance of this kind of bond.
Makes sense now, thank you! I feel like the author should have mentioned this.
Not sure what the controversy here is. Catastrophe risk is the bread and butter of property insurance.
Felt like the article ended before a thesis statement.
> She is the author of “Investable! When Pandemic Risk Meets Speculative Finance – A Cautionary Tale,” from which this article is adapted.
So I think structurally, the conclusion here is that 'cat bonds are an example of how insurers can work with abstract risks, and so any risk (such as global pandemic) could be worked with this way', and the rest of the book then examines how people are trying to actually do so with pandemic risk.
But as traditional insurance, not cat bonds.
I mean eventually once you proceed down the financial tree you get reinsurance linked securities. It's been this way for decades. It nothing new.
So, it's betting not on the outcome, but specifics of the event. The analogy could be sports betting where you bet on a player making a particular individual score, not just the win/loss.
At the end of the day, if it prevents bankrupting companies, countries and the insurers when an event occurs then it is a good thing.
I know of a Australian-based insurer that only operates in the far North primarily for residents who can't get reasonable premiums to cover for floods and cyclones from the major insurers. They reinsure with a global company so paying out an event in Australia is offset with no events elsewhere.
Interestingly the frequency of cyclones is decreasing, but the strength is projected to increase with "global warming": Higher water temperature = more energy. So betting on the strength makes sense.
Reading this as somebody who tries to solve betting markets with data and python, yeah, you’re right, this is just prop bets but with less data due to lower frequency events.
I honestly think there’s EV in this - you can expect some sports guys to be booking tickets to Baden-Baden next year if they find this article.
As if pandemics weren't already political enough. Let's get large corporations, investment funds and billionaires involved and give them direct stake in declaring what is or isn't a pandemic, how many deaths have happened in a certain area, what was the cause of death etc. That should end well.
This is the dumbest idea I've read about in a long time.
Why is it dumb?