From YCombinator, June 29:
Thoughts on Insurance
Two years ago, I printed up Chubb’s 10k and started reading.1 I’d become interested in the property and casualty insurance industry through a number of conversations with my father-in-law, who is a commercial broker. While I’d thought a bit about health insurance before that, it was mostly in the context of my own access to it, and the never-ending debate around Obamacare.
As I read Chubb’s financials, industry reports, Warren Buffet’s letters, and various blogs I came to realize that the insurance industry was both far more complex and rife with opportunity than I’d assumed. While I’ve always been attracted to fractured and regulated markets, nothing quite mimics insurance in its scope, nuance, and size. I wasn’t the only person thinking about this, as the number of recent insurance tech companies indicates.
Here are a few core problems built into the structure of insurance today.2
Insurance is fundamentally a data problem. Insurance carriers set their rates based on actuarial models designed to predict the likelihood of future events. Without access to oracular powers, these models rely on the best available data when they are built and as they are updated.
This data is necessarily incomplete. It only becomes more incomplete as time goes on and more events pile up in a very large and unordered world. For instance, a home insurer would be able to better predict the likelihood of fire if it knew the details of every overloaded power strip in every home in its portfolio. It does not.
The issues faced by carriers with data extends to the setup of carrier/broker relationships. These relationships necessarily involve lots of paper because there are few simple systems that easily integrate the two sides. This means that information about customers is often relayed poorly, misunderstood, or simply ignored.
Trying to figure out all of the different players in the industry is difficult at best.
When it comes to the distribution side, I can’t do it any better than Kyle did here: https://medium.com/@kylenakatsuji/so-your-startup-wants-to-sell-insurance-a0167581f7b1.
However, there are even more players involved behind the scenes which are important to understand if you want to uncover opportunities:
Reinsurer – There are companies that purchase insurance risk from carriers. They are critical to the system because insurers will often find that they are overexposed to a given risk (like that presented by hurricanes in the Gulf Coast) and will need to offload some of that risk. Reinsurers traditionally purchase risk from carriers and from other reinsurers. Reinsurers have also begun expanding the types of risk that they will purchase and the stage at which they’ll do it, sometimes acting nearly identically to carriers.
ILS buyers – ILS are Insurance Linked Securities. The most of famous of these are CAT (catastrophe) bonds. These are created by insurers and reinsurers who wish to syndicate risk beyond the insurance world. This is done by creating a bond which pays an interest rate and defaults in the case of particular event. The market for these is currently fairly small, and the bonds are generally purchased by hedge funds. This market will likely expand over time as capital continues to look for yield.
Fronting carriers – These are carriers that form partnerships with other entities, like MGAs (Managing General Agent, defined in Kyle’s post above), wherein the MGA writes risk using the regulatory framework of the fronting carrier, and then immediately sells the risk to a third party. This structure allows entities that could not otherwise sell insurance – whether through business choice, lack of regulatory capital, or lack of expertise necessary to form a carrier – do so as long as the fronting carrier agrees.Aye lad, but if you're any good at capital allocation, the float is where the real money is made.
Fronting carriers are not capitalized in the same way that large carriers are, as they don’t hold risk on their own books. They generally collect a fee – for the use of the regulatory framework – from the entity finding and pricing risk.
The complexity of the structure of the Insurance market creates 3 other problems, below:
Each of the players in the structure needs to get paid. Premiums are the primary source of revenue moving into the system, which means that every dollar paid in by customers has to be split between all of the value providers. After paying for broker commissions, fronting costs, reinsurance, customer service, claims processing, there’s often around 50% of the original premium dollar left to pay claims – which is the primary purpose of an insurance company....MUCH MORE
See also A16Z and the gang:
Thursday, September 22, 2016
Insurance: The FT's Izabella Kaminska Will Probably Not Be Going to This Year's Andreessen-Horowitz Christmas Party.
Last year, when Ms. Kaminska was pointing out* that Andreessen-Horowitz investee 21inc. ($116 mil from A-H, Khosla et al) seemed to be another solution-in-search-of-a-problem company, I was reasonably sure she wouldn't be invited to the 2015 party.
Now this latest pretty much rules out her attending the 2016 get-together as well.
First some background. From our January 2015 post "Andreessen Horowitz On Insurance: "Software rewrites insurance" (nudge, nudge)":
From Andreessen Horowitz:
Insurance is all about distributing risk. With dramatic advances in software and data, shouldn’t the way we buy and experience our insurance products change dramatically? Software will rewrite the entire way we buy and experience our insurance products — medical, home, auto, and life. Here’s how:By changing the way insurance companies price risk
So many more signals are available for insurance companies to better price the premiums we should pay. Drivers that drive carefully in safe neighborhoods vs. recklessly through accident-prone intersections ought to pay different amounts to insure the same car — but all that data isn’t reflected in an annual odometer reading. Water damage is one of the top sources of claims for home insurance customers: Why don’t we charge customers with water sensors less, since if they know water is leaking, they can stop it before the damage gets expensive to repair.
New data sources, better data, ongoing data reporting — all are possible now with mobile phones and inexpensive Internet of Things devices....
And today [Sept. 2016] at FT Alphaville:Also:
Breaking insurance models with big data
In the brave new world of machine learning, big data and artificial intelligence, no good deed will go unnoticed and no bad deed will go unpunished. Or so at least the dream goes.
The proposition here is simple. Soon enough, telematics companies will gather data from all our connected devices, fitbits and cars, scrutinise it intricately, then determine whether we are “good” or “bad” agents. Good behaviours will be rewarded with cheaper insurance policies, bad ones will be penalised. The relative cost of being a bad agent, meanwhile, will incentivise good behaviours, eliminating evil from our world forever. Amen.
If you thought that was a far fetched vision, however, you’d be mistaken. The burgeoning telematics sector is well on its way to partnering up with insurance companies, automakers and more. And in almost all cases the companies believe actuarial or insurance services are the best path towards the monetisation of their data intensive business models.
There’s only one problem. Personalising insurance contracts to this degree undermines the whole concept of insurance....MORE
P2P insurance firm Lemonade launches out of stealth, powered by chatbots, morals, and big bucks
The Big Questions We'd Better Figure Out, Part 2: Algorithmic Discrimination and Empathy