I wrote the following long form essay for an industry publication. Enjoy!
Disruption in Insurance: Harder than it Looks
I think it was AirBnB that did it best. And I suppose Uber. These two companies introduced a new way of conducting very old businesses, transforming the experience where most would never think transformation was possible. Taxis in particular are horribly contaminated by regulation and special interest politicking. If they are vulnerable, isn’t insurance? Isn’t everyone?
I am a reinsurance broker and one of the things we do is link startup insurance businesses with the capital and partnerships they need to execute. After a while of doing this, you notice even dissimilar ideas fail for similar reasons, which I’ll get into below. As a salesman I am of course an optimist so think of this essay as a list of critiques you, as an entrepreneur, need to overcome to leave your dent in the insurance universe. I hope you do!
I used the word disruption up there and it (the word) is everywhere these days. However, if you were Clay Christensen, the Harvard Professor who coined the term “disruptive innovation” in a 1996 HBR article and subsequent book, you’d probably be frustrated. The term has become massively popular, of course, but widely misunderstood to mean something like ‘tech startups win!’, which isn’t what he was going for. For one, all businesses use technology and new entrants have always used it better than incumbents. Yet all new entrants aren’t ‘disruptive’. Uber has brought dramatic change but Christensen himself is on record saying he doesn’t think it’s disruptive, in the sense he meant for the word.
Real disruption, as defined by Christensen, is scary. For most companies, being relentlessly focused on the customer is a good thing: know what they want and deliver that with focus, discipline and low prices. Or maybe not! In Christensen’s telling, disruption reveals this strength to also be a deadly weakness.
He defines disruption as entering a market from an overlooked customer base, of which he says there are two kinds: low-end footholds and new-market footholds. Low-end footholds start with “less demanding customers” in that they consume much simpler, lower quality products than the mainstream market. The disruptor then progresses up to higher quality customers. In new-market footholds, the disruptor tackles non-consumption first by cultivating customers who had never used the product, then moves into an incumbent’s turf later. In both cases, the end game is offering the final group of customers an equivalent product to the incumbent’s but at a dramatically lower price.
Startups are trying to do this all the time and they are failing all the time. I picture an incumbent smugly chuckling to himself as a constant buzz of startups tap against the window over and again until they drop. What do they fail? Maybe the incumbent thinks discipline of execution and focus on the customer keep it from wasting time on these fancy pants tech ideas. The poison in that chalice is that a company can get in the habit of dismissing everything new and assuming an old ‘bad’ idea can’t ever become good. Venture capitalist Marc Andreessen likes to roll out the old dotcom bubble whipping post, Pets.com, and note that Chewy.com, an almost identical business, was recently sold for 3.3bn. Timing is more important than creativity and timing cannot be planned.
In insurance, the best example of real disruption comes from developing countries in the form of microinsurance. Check this out from the Microinsur website:
We have introduced new forms of protection for emerging customers, including micro-health, political violence, crop and mobile insurance all over the world. In each case, we didn’t start by designing a product in a board room – we visited our customers in markets and villages to understand how they cope with the variety of risks in their lives. The result of this client-centric approach is a new suite of solutions, and the opening of a new market for insurance.
This hits both of Christensen’s entry points: new and low end! In listening to Microsinsur’s founder Richard Leftly’s insnerds.com interview, it’s clear he is aware of this, talking about the power of delivering insurance solutions at phenomenally low overhead costs. A good idea, then. How about timing?
Microinsurance only becomes disruptive if it moves upmarket and the challenge there is regulatory: most insurance buying is forced. Compulsory purchasing makes customers stupid, especially at the low end of the market where the smartest bargain hunters live. Regulators set the rules for what customers can buy and don’t want change. And they have good reason! New products make it hard to tell when someone is underpricing to win, later to collapse in a mess the regulator needs to clean up.
Are regulators ready to try something different? I don’t see it but it’s common for insiders to miss signals of coming revolution. That’s the problem with timing, you need to try (and probably fail!) in order to answer the question: why now?
I think that when people say disruption they really mean disintermediation. Disintermediation is probably the number one ‘swing for the fences’ strategy for any business. Everyone deals with intermediaries and endures their transaction costs for the benefit of accessing a market. Disintermediation removes the transaction costs but also removes the market! Let’s take three examples:
- Cutting out brokers (going direct to consumer or agents)
- Cutting out insurers (Automated underwriting)
- Cutting out everyone (peer to peer insurance)
Each has been around for decades but none has taken off, really. Here are some ideas for why.
*Cutting out Brokers*
Example: direct insurers and reinsurers
Your broker protects you from getting screwed. Screwed means buying something when an identical product could be had for less. Brokers protect by forcing insurers into the ring to fight for your business. Now, insurance is also complex, requiring some expertise to ensure coverages and terms really are identical before you compare prices. Such expertise means brokers cost good money! Is it worth it?
Let’s come at this a different way by asking “do brokers cause more competition or does more competition cause brokers to appear?” I say the latter: brokers are a symptom of competition, not a cause. And competition is not constant, which means the strength of the broker also rises and falls. That, in turn, means there are two ways to legitimately cut out a broker: 1) remove the need for competition; 2) get insurers to compete without a broker.
Removing competition is not as crazy as it sounds. Consider my business, reinsurance: over the last 30 years the market has had several big negative surprises, causing incumbents to question their understanding of certain businesses, opening up room for several waves of startup reinsurers with distinct appetites. No question those were competition-enhancing episodes.
In between those episodes we’ve witness the reverse: appetites converging as the claims environment settled down. This means carriers consolidate to save on costs since they find it harder to secure an underwriting advantage against others’ identical view of risk. We call this a soft market.
Fewer differentiated options in the market makes a direct relationship more appealing. That said, a complete unification of strategy among reinsurers in volatile business will probably never happen. And even then brokers will always have deeper relationships with the marketplace than a reinsurance buyer and that means better deals. A disempowered broker is still the best place to keep from getting screwed.
In direct insurance things are a bit different, particularly in non-catastrophe-exposed personal lines where products are homogenous and margins are thin. There, new entrants are unable to replicate the data of the incumbents and so bear enormous risk of unknowingly underpricing their business.
Critical mass of is not a new problem in insurance. A hundred years ago the market solved it by banding together to create rating bureaus, centralizing the analysis and often prescribing rates themselves. It was an era much more comfortable with collusion and monopoly than today. That solution is showing its age in many lines: the data needed to put together a state of the art personal lines rating plan has exploded and is growing still. ISO, the heir of the rating bureaus, isn’t keeping up. So market participants have two options: strengthen ISO or be huge. They chose the latter.
The real technological development is that direct carriers are passing on discarded customer leads to competitors, in effect creating bilateral marketplaces. GEICO pioneered a version of this by generating a customer, stripping out the auto and partnering with other carriers (mostly large) to sell the other products. The additional cost of passing a lead to another carrier’s system is zero so why not have a world where large direct carriers are all linked up and whichever portal you enter generates an extra fee for that carrier? That would be a market without a broker but can carriers really stop themselves from manipulating the prices somehow or denying access to the market until your reject a higher priced policy first?
As always, cutting out the broker begs the question: how do you know you’ve got the best deal?
*Cutting out Insurers*
Example: catastrophe hedge funds cutting out insurers and reinsurers and going direct.
Insurance deals in risk. Risk isn’t real, really, like buildings or dishwashers or iPhones are real. Risk is a bad outcome we can’t see coming. Since we can’t see it we can’t avoid it.
That’s nice, says the insurer’s accountant, but that doesn’t help me fill out this tax return. How much money are you making and if you tell me none of it is real again you’re going to jail? Well insurers start with revenue for taking the risk (premium) but no claims payments for a while. What should the accountants do with that? Well, before insurers there were two options, a legal one and an illegal one. The legal option is to declare all the policy limits as obligations and put them on your balance sheet. But we can’t just pay today for every bad outcome that might happen tomorrow, there are too many possibilities! So enter the illegal one – to just pretend our way to an answer.
Boy the second one sounds way better. But we still have risk! Now the risk is that the insurer really is insolvent and we don’t know it yet. Who takes this risk? Nobody wants it, least of all policyholders who need their claims paid. So the state takes the risk and puts it into a box called a guarantee fund. They charge for that little trick, of course, demanding money and regulation.
Insurer disintermediation is a really about cutting out the regulator but then someone else has to take the insolvency risk. A workaround exists: using a ‘front’, being a business that takes no risk and only supplies regulatory services. All the other functions: underwriting, finance, claims processing and claims risk are all administered by a variety of agencies, consultants and reinsurers.
Fronting companies are really small, financially speaking, because the premium all goes to pay the agents and consultants and reinsurers. There remains this tiny little probability of the reinsurers, agencies and consultants going out of business or not honoring their various promises. They call it tail risk. But the front is riskless, so who takes that?
Great question! In the (non-fronting) wild, the insurer takes it and keeps a few risk absorbers against bad outcomes: premium, capital and the guaranty fund. The thin front doesn’t keep the premium base, being the first and biggest of those absorbers. As a consequence, modern fronts tend to be incredibly selective of what they accept and/or push as much of this as possible onto the other counterparties.
But as soon as you need to quantify the insurance limits exposed you’re back to the pre-insurance era of accounting, declaring all the limits and capitalizing them on your balance sheet. It really doesn’t work:
There are something like 29 million small business in the US. Each has a GL policy for, say, $1m limit and pays probably something like $1,000 for it. In theory, then, there is 29 trillion in GL limit outstanding for 29 billion in premium. That’s a 1000x disparity.
That’s why insurers aren’t asked to hold collateral against all that limit exposed. But the scope for something going wrong is huge and finding a way to absorb tail risks is the challenge in disintermediating insurers. Policyholders with valid claims aren’t going to get a haircut. Shareholders are tiny compared to the limits exposed. To disintermediate the insurance company you need to answer: who gets the tail risk?
*Cutting out Everyone*
Example: reciprocals, pools, captives, lemonade.
Here’s a cartoon example of how insurance works: we pay premium to an insurer who pays it right back to us as claims. Well, they also sit on it for a few years and keep about a third of it to fund the system: brokers, underwriters, finance departments, taxes. A third?! For giving us back our own premium? Yes.
To an economist, literally everyone in the industry, not just brokers but insurers, reinsurers, adjusters, etc, are intermediaries: they aren’t ‘doing anything’, just shuffling money around, keeping some for their trouble.
Exciting opportunity alert! Let’s cut them all out, go peer to peer and pocket the savings. Not a new idea, I’m afraid. Legend has it that the first insurance companies had exactly such a structure: merchants pooling their losses from sunken ships. So literally every other feature we see of the insurance industry was deliberately invented: brokers, underwriters, finance departments, actuaries…
Let’s defend all our jobs by first stepping back a bit. Insurers protect us from random chance (God) and moral hazard (other customers!).
Protecting us from God is about getting enough volume. House fires are too volatile to pool among 10 or 15 friends. And bigger pools are hard to coordinate so you have to pay someone to help with that. Enter the insurance company processing department and perhaps reinsurance for extreme cases.
I’d argue every other feature of our business exists to protect our premium dollars from moral hazard (ie our fellow customers). Think of the merchant in our primordial insurance pool who lies about his ship sinking and pockets the cash. An insurer is really just a pool of money that each insured has a claim to. Is it so surprising some will greedily eye that cash when they think nobody is watching? Defending the pool against wrongful claimants is what the industry does.
But what about customer service! Lemonade in particular uses the idea of insurers’ poor service as its key promotional message. Even granting that customer service is poor, root problem isn’t that insurers are evil, it’s that customers lie! Even forgetting simple fraud, how about the insured that doesn’t repair his sprinkler system because he has an insurance policy? Moral hazard is incredibly tricky (expensive) to identify. Even then, the higher expense of oversight yields massively larger benefits in reducing the cost of invalid or avoidable claims. This makes insurance cheaper! The chance that you need the customer service is pretty low. The chance that you want those hundreds of dollars of premium back in your pocket is pretty high.
So you can think of the insurance market as having found a balance between low premiums, which customers like, and suspicious and skeptical customer service, which customers don’t like. Cut all that out and you have to answer: how do I protect customers from each other?
So It Can’t Happen?
Now, of course Uber and AirBnB, in particular, looked completely stupid at first. Lots of very serious people had lots of really good reasons for why they would fail. And look at them now! The problem is that those serious people, wrong on those two counts, are normally right. Startups capture the imagination exactly because they’re so outrageous and unlikely. In a mostly market economy like ours, the world is complicated and dramatic changes are very rare. Want to make the highest probability prediction for next year? Start with “more of the same ahead”.
The answers you need to the questions above need to be better than the answers the industry already has:
- “How do you know you’ve got the best deal?” I test the market with a broker
- “Who gets the tail risk?” Insurance shareholders and guarantee funds
- “How do I protect customers from each other?” underwriting and claims management
- “if disruption, why now?”
The answer to the last is to usually point out any number of the cutting edge technologies of the day and scoff at how slow incumbents are to adopt them. I agree there is opportunity there.
But even startups launched by insiders underestimate the reason insurance is different: you don’t know your costs up front. Without a track record of success regulators, rating agencies and reinsurers will slam the door shut. Building that track record takes time. And during that time incumbents catch up and startups often assimilate, are acquired or blow up.
What’s a startup to do? I say don’t fight against the basics of insurance, use them to your advantage instead. There are times new entrants are allowed in. We call them hard markets and they can work as laboratories of insurance innovation. Incumbents lose faith in their understanding of the risk and deliberately retreat. The system is begging for someone who can be more nimble and solve the market problem.
As I write this perhaps the Caribbean is a good choice, decimated by Irma and Maria? These are challenging situations because along with whatever technology a new entrant wants to bring to the market they need to solve a pricing problem for the risks. But the entrant has a pricing problem no matter when they enter. At least they’re on even ground with everyone else in a hard market!
”This article was first published in the Journal Of Reinsurance, a publication of the IRUA – www.irua.org – and is reproduced with permission.”