There is a tendency to describe blockchain technology either as an optional add-on to existing core processes of an industry or a complete industry-wide disruptor.
Neither of these descriptions are not going to lead to the timely and optimal application of blockchain in insurance. In the former description the add-on could arguably be achieved more effectively with existing technology; in the latter the industry-disruptor narrative leads to a collective wait-and-see/fast follower approach.
The way out of this conundrum lies in defining the insurance industry inherent, core process in extremely reduced and functional terms. This definition can then be compared to a plausible future state of blockchain technology to make more actionable inferences on blockchain and insurance.
This definition is: Insurance hedges against reality. For the purposes of this article the focus will be on reality defined as the state of things as they actually exist, as opposed to a notional idea of them. Reality, as it pertains to human health, sickness, life and death is too complex to navigate without using some form of abstraction. Abstraction is essential for decision making but abstraction also takes what it gives. A map, as an example of an abstraction, can give too much information; too little information or the wrong information altogether. Either way, each abstraction has its inherent limitations of which decision makers should stay cognizant of.
Awareness of the insurance industry’s current map and its limitations will give actionable insight into the role blockchain technology will play within the insurance industry.
The current map of the insurance industry and its limitations
The map that an insurer uses to make the messy reality of human activity related to health, sickness, life and death into something that can inform decisions are highly advanced actuarial models. These maps are indispensable but from an overly simplified point of view they have three inherent limitations:
- One-way – the actuarial models provide accurate decision support on trends within a population. Nevertheless it is unable to dynamically affect the behaviour of that population due to a one directional data flow from population to model. This makes the creation of positive or negative feedback mechanisms that are native to the actuarial model impossible
- General – the actuarial models can make inferences about a person based on general demographic indicators but can’t draw inferences from their individual behaviour. This makes it impossible to provide products that are customized for an individual’s actions within a particular context
- Proprietary – the actuarial model is also an insurance company’s core intellectual property and can therefore not be shared with other companies. This limits the opportunity to collaborate and add more value through a powerful business model template known as double-sided markets
Just like a map that is not at the correct scale, actuarial models to some extent, and divorce insurance companies from reality. My thesis is that as blockchain technology matures it will enable a movement towards reality-based insurance when used to augment existing actuarial models.
Blockchain as a step towards reality-based insurance
In layman’s terms a blockchain is a novel type of ledger that can be used to maintain its own tokens and automatically execute transactions in these tokens based on real-world events via smart contracts. These unique features in combination with the rapid advances in privacy-centric cryptography[1], such as zero knowledge proofs and anonymized cryptocurrencies, will eventually make insurance (1) two-way, (2) granular and (3) shared:
- Two-way – the most powerful – albeit overlooked feature – of blockchain technology is the ability to maintain a feedback mechanism via the conditional transfer of a native token
This incentive mechanism explains the rapid growth of cryptocurrencies such as Bitcoin or Ethereum. Participation in these networks leads to automated payouts in the native currency or tokens and therefore incentivizes participant behaviour that leads to more of these payouts. In Bitcoin and Ethereum the most common example is mining (providing processing power) but it could also be as simple as accepting payment in these tokens for services rendered.
A very basic insurance-related example is KPMG Australia’s FitChain initiative. When users do a particular amount of steps that is monitored by their FitBit their account is credited with tokens on a blockchain that can then be redeemed for a coffee. This can be extended to a host of healthy behaviours (which is monitored in different ways) and lifestyle-specific rewards for lifestyle-based incentives. Instead of a one-way directional flow the actuarial model can be enhanced with this token mechanism to influence behaviour through token payouts. The results of the payouts can be monitored on a population level – without endangering user privacy – to enhance the actuarial model. An automated feedback loop of behavioural information and incentive is therefore created. Incentivizing healthy individual behaviour holds long-term benefits at a societal level such as less dependency on healthcare budgets. For the insurer it allows a mechanism to rapidly adjust policies to support consumer lifestyle goals.
- Granular – The startup Trov offers on-demand insurance on a per device basis. It is the shape of things to come but it demands a lot of manual intervention from the customer to customize it for their particular and temporary context. Customers are increasingly generating more and more contextual data from their phones, wearables and payments. By means of using blockchain technology and the anticipated privacy-centric features this contextual data can be delivered to dynamically adjust their insurance based on the risk associated with their current individual behaviour. The benefit to the customer being that the customer only pays for the individual behavioural risk they face instead of paying the average for their demographic. This logic can of course be extended to the provision of highly customized product combinations and premiums towards individuals instead of a demography
- Shared – Within the insurance industry, there is an inherent disincentive to cooperate. This is due to the possibility of the actuarial model becoming public knowledge. A shared blockchain infrastructure enables cooperation with other organizations while divorcing the result of individual actuarial models from the model itself. By using a shared blockchain infrastructure supported by a shared database for larger documents insurers would then be able to share information to collectively provide one integrated customer experience that includes the entire lifestyle of the consumer. Sharing of relevant information from different types of insurance would also provide more accurate individual risk assessment and policy pricing. This integration would also make it possible for InsurTech companies to provide their services to a broader customer base without being dependent of only one insurer. The incumbent insurer benefits from being able to keep their focus on their core competencies
A governed and collaborative approach, between insurers, towards building this shared infrastructure would require the creation of a consortium. The consortium’s activities would include but not be limited to creation of similar scoping, design, creation of unified data standards, joint public relations and coordination with regulators.
Conclusion
An anticipated critique of the abovementioned points would be: “We can already do this with centralized databases and don’t need a blockchain for this.” Fact of the matter is centralized databases don’t support cooperation between organizations nor between silos within organizations. Also, their ability to elegantly handle any level of transactional complexity is naught. Most importantly – machine learning and advanced data analytics hold untold potential for the insurance industry. When centralized databases and their localized versions are out of sync (as they often are) the whole system grinds to a halt and that potential is left unrealized.
For reality-based insurance to work there needs to be a singular, shared and constantly updated version of the state of the truth. With its distributed design with one constantly updated and coordinated version of the truth blockchain technology will form the bedrock for reality-based insurance and provide the privacy-centric environment in which the untapped potential of machine learning and data analytics can be unlocked.
No singular insurer can make this future vision a reality seeing that blockchain technology – by its very architecture – is inter-organizational. A governed approach towards collaboration in the form of a consortium is the most efficient way of making the future happen.
Author
Gys Hough is senior consultant in digital ledger services at KPMG in the Netherlands. He has worked at a pan-European level on the future of blockchain in the transactional industries and authentication. His current focus is on adding value by creating industry-specific use cases that leverage the distributed, immutable and tokenized nature of blockchain technology.
[1] George Samman’s The Trend Towards Blockchain Privacy: Zero Knowledge Proofs is an easy to understand primer on this subject.