Insurers Tripping Over the Blockchain
Insurance underwriters like to think they’re hard-headed and precise number crunchers, when in fact the aspects of their profession they seem most fond of are those where they get to practice voodoo, poking through the entrails, reading the tea leaves and deciding the insured’s fate. IT tends to tear the bones from their beards and expose uncomfortable statistical realities. I’ve talked about their fear of Artificial Intelligence. The blockchain is another technology whose time appears to be at hand and it’s going to make voodoo underwriting harder to practice.
(Fortunately none of the underwriters I work with read me; but that has nothing to do with why I’m poking fun at the profession.)
An insurance product offers a transfer of risk in certain circumstances. The precise commitments, obligations and coverages of a policy are just agreements between parties – intangible, data – and all that an insurance company’s business rests on. That makes smart contracts (agreements that execute automatically when the terms are met) in secure and irrefutable ledgers a natural fit for the insurance industry. Truly, I’d been planning to write this article for a while, conceiving a wide survey of the insurance industry’s adoption of emerging blockchain technology. I was expecting to write pages and pages on its enormous, disruptive potential for this industry. But he who hesitates is scooped, and it looks very much like the blockchain has already emerged in this domain. There is now a wealth of analysis and prognostication out there, including white papers, blogs and industry consortia. In fact, TechCrunch stole my thunder six months ago with an article that covers the most innovative areas of blockchain insurance applications, and they even took the “Future of Insurance” title that I’ve used to serialize my insurance-related posts in the past. If you’re still a little vague about how this exciting technology will shake up our stodgy industry, follow the links in this paragraph, including Deloitte’s readable piece, this summary from McKinsey and these thoughts from PwC.
All the same, insurers will move slowly into the blockchain for some good reasons. Insurance is a conservative, risk averse industry, yes; but the hesitation isn’t simply resistance to change. It requires enormous confidence in the security, privacy and availability of the architecture in order to put corporate data and customer personal information into a blockchain which must by its nature be public and widely distributed. Hard questions are appropriate – but they’re also answerable by reference to intrinsic characteristics of the technology. And the enthusiasm we’re seeing suggests that assurances have been satisfactorily delivered. But these are the basic questions a business would have about any technology; as usual financial services have concerns that go a little deeper. These are some of the risks associated with this technology which are not so easily set aside:
Performance and scalability. To companies slinging billions in premium, Bitcoin still looks like a toy. Bitcoin’s blockchain currently handles three or four transactions per second and can probably comfortably handle seven. Compared to the 2000 tps (and peaking at ten times that) which the major credit card companies process it doesn’t look like something for the blue chips. Ethereum can handle more transactions and is theoretically much more scalable; but it has even less of a track record with which to demonstrate its performance and reliability. With blockchains maintained open source, resolving these issues is often messy and uncertain. Nonetheless, the blockchain community is keenly aware of the necessity of getting over these limitations and it will only be a matter of time.
Error correction. Some companies are nervous about writing to a ledger that has no way of fixing a typo, or worse. Last summer hackers stole many millions of dollars worth of Ethereum’s cryptocurrency, ether. The only way to restore it was to revert the Ethereum blockchain back to a point prior to the theft, a very controversial move, and not something you would likely contemplate if serious companies were handling core business on the platform. To avoid such incidents in the future Accenture has proposed an editable blockchain that allows an administrator to make corrections. Many feel this undermines the technology since its strength depends on the elimination of central authorities and trusted third parties. Without that, it’s just another database and you have to trust your insurance company to maintain it with integrity. If it’s not editable, the integrity is built in. But given real world exigencies, it may be a necessary compromise.
Human Judgment. Notwithstanding my quips about witch doctors, underwriters are generally ethics-driven professionals who have had extensive training and take their responsibilities to their insureds seriously. The complexity of an insurance contract doesn’t always respond appropriately to those fiduciary duties or acknowledge circumstances that argue for different rates or endorsements. It can be hard enough with present systems for a human being to adjust a policy to fit the real world; blockchain-based policies push in the other direction, modeling an idealized customer using inflexible rules. Blockchain applications can commodify niche products, enabling highly specific rating but reducing or eliminating any human touchpoints. Smart contracts and parametric insurance programs eliminate, by design, the opportunity for an underwriter’s supervision of individual transactions. And peer-to-peer models’ strength come from a design that relies less on an underwriter’s professional judgment to evaluate a potential customer. So how is the underwriter to discharge their duty to a client they’ve never even had an email from? System design including checks and balances is important, and I foresee some lively discussions with my underwriting counterparts as they seek to temper the technology’s potential for hands-off policy management with a professional underwriter’s oversight.
“Works in theory” is never going to be good enough for companies whose customers rely on them when the very worst happens. The track record established by Bitcoin and Ethereum doesn’t inspire a fiduciary’s confidence, not yet. Expect to see cautious adoption in supporting systems like forms libraries and rate filing systems before blockchains move into the core functions like policy administration and claims management. But move into these they will because insurance technology demands systems with the built-in trust, confidentiality and security that comes out-of-the-box with blockchains, and their more advanced features enable truly innovative products. It’s inevitable that blockchains will soon be found in the tool box of any insurance industry CIO.
« Weakness in the Media Distribution Business Model Ransomware: It Certainly Does Make Me WannaCry »
