This article by Nick Lamparelli originally published on InsNerds.com
Recently, I wrote an article at InsNerds.com where I outlined a simple modeling framework I use when I try to predict how a new insurance product or new insurtech startup is likely to perform. In this article, I will walk through an example to give you a play-by-play on how to put this simple mental model into exercise.
Can Peer-To-Peer Insurance Succeed?
Peer-to-peer business models really came into their own in the financial arena, where companies such as Prosper and Lending Club were able to create platforms that allowed individuals to loan funds directly to one another. As a Prosper investor, I still recall how neat it was that I could loan a family $25 and be part of a pool of like-minded people who were looking to help others and make a little bit more money than a bank account. (disclaimer: you can lose your money too. I have had several borrowers default, and you will need to make up for it on those accounts that don’t default).
Investors, always a group looking for the next big unicorn, have applied principles of P2P to others businesses as well, such as car sharing and file sharing. Even digital currencies such as Bitcoin are P2P based. Not surprisingly, investors and entrepreneurs are looking into whether P2P would work well in the seemingly tattered insurance industry. Companies such as Lemonade, Guevara, and Friendsurance are already selling policies and making a name for themselves and also getting a lot of investor attention. InsNerds.com was very lucky to have Dylan Bourguignon of So-sure insurance, a complete P2P insurer write an article for us on the topic (be sure to read this article if you want a breakdown from the point of view of an insurer).
Let’s walk P2P insurance through the model framework and see what all stakeholders need to see.
The Exposure component is the one that deals with claims; past, present, and future. The P2P model looks to reduce the frequency and severity of losses by reducing the desire to policyholders to make bogus claims. Because policyholders in a P2P model have some affiliation with each other, the hypothesis is that this connection will prevent policyholders from harming their peers. This seems intuitively possible. If it is true, what would that mean to the insurance coverage?
Fraud is estimated to add 10% to losses in Property & Casualty insurance. That is a significant sum, which would equal approximately $ 34 billion per year! Fraud is most typically investigated in workers compensation, auto and health insurance (not necessarily in this order). Traditional insurers spend a lot of money investigating and rooting out fraud. Big data vendors such as Verisk and Valen have commercial models available for both workers compensation and auto-even homeowners insurance isn’t immune. Reports of widespread false claims after Hurricane Katrina were documented.
The difference between what traditional carriers do and what P2P offers is that P2P subtly promises to remove the fraud BEFORE it happens, while today’s fraud is only caught during the fraud or afterwards. If P2P can fulfill this promise, there is a tremendous amount of value it can provide to the market.
If I were an investor, I would look for companies that can show that their P2P network has very tight ties. As the network gets larger, it seems unlikely that the strong ties can be maintained, and so, you begin to lose the ability to have shame or other social pressures keep fraud under control. Any technology that can strengthen the ties to large portfolio scale could be immensely valuable.
I’ve written about Lemonade in the past, and while they no longer consider themselves P2P, their initial “technology” was to group like-minded homeowners or renters together, and give an excess year end profits to a charity of their choice. If you are following along with where I am going with this, you may see some of the flaws in the model. First, homeowners insurance isn’t in the big 3 for fraud, so the potential benefits are not nearly as big as if they tried auto or workers compensation. Second, I didn’t really see any proprietary technology that could give them a leg up on any competitors. From all of the press releases, their P2P networks seemed easy to copy, as is their charity angle. That they dropped their P2P marketing seems to have confirmed that that part of their business model probably would not have produced worthwhile value. As an investor, I’d like to see a direct line to fraud reduction and truly big potential to drop the investments now being committed to detecting fraud post-event.
Before leaving EXPOSURE, my opinion of P2P is that it needs to bring some new type of configuration of insurance that meets needs not currently being met. The insurers mentioned above are tackling industries with heavyweight competition. I see an opportunity for P2P to unite common insureds in a way that provides coverage or risk reduction in areas where coverage is difficult to obtain or just doesn’t exist. In California, earthquake deductibles are very large. It seems reasonable that property owners could unite to buy coverage to protect each other against losses arising from the combined deductibles. There’s a similar case to be made for flood as well. I imagine these P2P insurers almost acting as public captives covering very niche risks for similar insureds.
The distribution component of the framework deals with how customers are marketed and sold. In the P2P model, there is a heavy emphasis on the social element to distribute. Like-minded insureds telling other like-minded insureds to join. Most P2P insurers are direct to consumer. Thus, P2P insurers must depend heavily on their insured network to do much of the heavy lifting for them, whether that’s through word of mouth or via social media.
If I were an investor looking into this area, I’d want to see some proof of concept that value can be created here to some scale. Brokers get paid well for a reason: it is expensive to find and maintain insurance customers. Advertising on Facebook is more expensive than you think, and if you are using Adwords, you are competing against GEICO, State Farm and other large insurance companies. Good luck with that.
Ultimately, I think distribution will directly depend on the product development and what was discussed in EXPOSURE above. P2P insurers must be able to differentiate themselves. Take Lemonade for instance. As a home and rental insurer, is Lemonade different than a traditional home insurer? Yes. Is it 10x better? I don’t think so. The product is nearly identical, only the customer experience is truly different. It is exceptional, but will that alone be enough to drive customers to buy policies? I think it will, but not by enough of a margin for Lemonade to deliver “uber” like returns. That’s not happening.
Insurance is a capital intensive business. To start a plain vanilla insurance company in most states requires $5-10 million in surplus capital. This is capital that is above and beyond capital that is used to pay for claims. That capital must be invested into the highest quality securities (generally government bonds and AA corporates). Any startup that is more complicated than “vanilla” needs more capital. And any expansion into other states will require still more capital. All of this capital is needed even if you only have one policy on your books and even if you are ceding all of your business to reinsurers. Startup insurers are behind the eight ball right from the get go and are at massive disadvantage when compared to the big guns. State Farm has surplus in the tens of billions of dollars. Those are funds State Farm can invest and through which generate investment income which can be used to offset other costs in their premium. Startup insurers can’t do that and are very vulnerable to any large loss and thus require heavy reinsurance partnerships. And that isn’t cheap! For startups, cost of capital is very high, and those costs must be reflected in the premium.
This is why Lemonade’s expansion across the US is head scratching. Though Lemonade is not a P2P, as a startup insurer, much of their newly acquired capital for this expansion, is sitting in bonds. Unless there is some other news that we are not privy to, using B-round capital as a bond portfolio does not seem to be a great use of funds. This is a lesson for other P2P’s. An entire P2P strategy can collapse if their capital structure is not maximized. If I were an investor looking at this field, I’d want the P2P insurer to be partnering with a capital source that already has scale, so that the P2P can focus on product differentiation and distribution.
P2P insurers have a terrific advantage in this area. Being born in the digital age means that these insurers can skip over legacy systems and go directly to an entirely modern platform. I would want to see seamless integration and movement of data between marketing, binding, policy issuance, accounting, and claims management. I would want the see the ability to easily capture data at the front end, augment data during the lifecycle, and the ability to put that data to work in integrated plug and play models.
For P2P insurers, Lemonade is providing the blueprint for how this should be done. (And by the way, big time kudos to Lemonade for being so transparent and allowing curmudgeons like me to nitpick their business model). Lemonade’s integration of Chatbots to eliminate human intervention in the purchasing of insurance and the filing of claims seems to be an operations winner thus far. In this model, we should expect to see overhead expenses drop. Expenses associated with losses should also drop as well (this part really belongs in EXPOSURE, but it seemed more fitting here in OPERATIONS). This is one component where if the P2P insurer was not able to show significant decreases in expense, then I would know something is terribly wrong.
I love the concept of P2P insurance. But I don’t think it will ultimately become a great way to invest venture funds. I just don’t think the returns will justify the risks. P2P insurers should be able to provide significant value in operations. If they can differentiate product development, they should be able to attract customers who would be interested in their products. BUT…I think P2P insurers are not going to find very large markets for their niche products. Because of this, distribution costs will be higher than they expect, and they will suffer from capital costs unless they form the right partnerships. Those really inexpensive Lemonade rates likely won’t last. P2P prices may not end up cheap as capital and distribution costs overwhelm advantages obtained in potential decreases in fraud costs and operational efficiencies.
P2P insurance is full of potential, and as a model, will behave more like traditional MGAs. The potential for supersized returns are not likely.
Nick puts P2P insurance, including Lemonade, through is model. No way to summarize this one, you gotta read it. It's good!