Insurance To The Power N
I went to a talk by Connor Love of Lightspeed the other day at the GSB. Connor did a great job communicating his excitement about insurance and insurtech - I’ve found myself in several rabbit holes over the last week, and the central theme of these has been insurance. I’m not a big fintech guy, not a big blockchain guy, not a big finance guy. But there’s something about insurance that’s really gripping.
Let’s get back to the basics. Insurance is simple - it’s a contract signed between two parties (a policyholder and an insurance company, both terrible descriptors) wherein the insurance company agrees to pay for covered damages the policyholder might experience in some span of time. Most of us have interacted with insurance in some form. Motor insurance, if you’ve ever driven a car. Health insurance, if you’ve visited a doctor in the United States. You pay your insurance company a premium so in theory you don’t have to worry about some set of adverse events.
It’s not always crystal clear exactly what is and is not covered (there’s often mountains of convoluted language - try reading your health insurance policy’s coverage document). And that’s for a whole host of reasons. But that’s a post for another day.
Coming back to insurance, though, on a higher level - why would one start an insurance company? Lots of reasons. Insurance is often legally mandated - so it’s not a question of whether there are customers, its a question of price. You’re helping people get back on their feet after adverse events. Mihir Desai, in his excellent book The Wisdom of Finance, talks about how it led to the development of Bayesian methods. Which is really cool.
But a large part of the answer is float. Float is just the sum total of all the money collected by an insurance company from premiums paid by policyholders.
Why is it nice to have a bunch of money just sitting around? So you can run a hedge fund with it - with essentially zero cost of capital. You’re not answerable to investors, in some sense. It’s kind of a dream scenario. And for a long time, it was. Insurance companies went famously bankrupt in the 80s because of their audacious bets, and state governments had to step in and put in some semblance of sanity. They demanded a ~10% cash on hand requirement - this is called Risk-Based Capital.
Okay all this sounds great and I kinda want to start an insurance company now. How would I do this? It depends on how much of the risk you are willing to take on. You can be an agent (take on very little risk, take a cut from premiums), an MGA (you can underwrite and manage contracts on behalf of an insurance company), or just be a full insurance company (super risky and capital-intensive).
But how are you going to be differentiated from GEICO? Better analytics? Maybe. Better distribution? Perhaps. The graveyard of failed insurers is rather large, so tread lightly. It’s also not, by any means, easy. Here’s a thread I found on someone trying to start an insurance co-operative.
But this post is entitled reinsurance. Insurance for insurance companies. I read Making a Market for Acts of God by Paula Jarzabkowski to try and uncover some of the enigma surrounding this industry. There’s frighteningly little known about it. Turns out it’s for good reason. There’s 3 centers for reinsurance globally - London, Bermuda and “Continental Europe”. To a lesser extent, Singapore. And reinsurance is a tiny industry, all things considered - everyone knows everyone else, and personal connections are paramount.
How things generally work is as follows. There’s cedents - insurance companies that want to cede their risk away. As an example, say you are MetroMile, and you think some x% of your policies are kinda sus. And you get cold feet vis-a-vis being on the hook for all those potential claims. You call your friendly neighborhood sleazy reinsurance broker, who then goes on to offer this deal to a bunch of reinsurers, the reinsurers go noodle on it for a while, and then all return with different quotes.
MetroMile looks at all the quotes, and then creates a panel, with a consensus price (not necessarily the lowest quote, but a price MetroMile thinks will have many takers). Twenty to thirty reinsurers get in on the deal, and all at the same price. Rationale here is that they are all just as invested and have just as much to lose, and you get to distribute risk. So if there’s an Act of God it isn’t necessarily going to clean out the capital of any one reinsurance firm.
These deals renew annually - and you have presumably slightly different compositions of panels year on year. There’s soft and hard cycles in the reinsurance world. An event like 9/11 cleans out a lot of people’s capital - and makes it harder for them to raise more. So the cost of capital goes up, and these costs are reflected in their quotes. A hard cycle. Versus when everyone is holding hands singing Kumbaya we’d likely be in a soft cycle where capital is cheap.
There’s all kinds of other subtleties. Often it makes sense to stick with the same reinsurers through hard and soft cycles, and there’s a you-scratch-my-back-I’ll-scratch-yours phenomena at play as well. You “reward” reinsurers for sticking with you through bad times. Jarzabkowski et al go into the details, if you’re interested.
I still have more questions than answers. Why is all this so mysterious? Why isn’t there some type of exchange where cedents post their consensus prices? Why can’t I be a part of a panel? Why isn’t this risk traded in public markets - can it not be packaged in a way to do so? What are these fancy companies in Bermuda like? Is there insurance for insurance companies for insurance? How do I meet more people in the industry? More to come!
