Life Underwriting – adopting to a digital world

Posted by

(Originally written on LinkedIn Aug 2018 – updated here with minor edits)


Why does the traditional actuarial view feel inadequate for disruptions?

Some crazy predictions on underwriting innovation

Two factors driving underwriting innovation: digitization, blockchain-driven digital co-operatives

Underwriting Innovation as seen through Life Underwriting drivers: 

  1. Data & experience
  2. Customer Behavior
  3. Sales ease
  4. Cost & margins
  5. Regulations
  6. Insurance needs and enabling framework

Why does the traditional actuarial view feel inadequate for disruptions?

We actuaries are hardwired to look into the past to estimate the future. The rapid digitization of our lives is quite exciting but has no precedence. Certainly so in life underwriting with rapid advances in understanding ourselves at our biological cell level, or even more profound at the quantum level, and advancement of wearable tech. This advancement creates a kind of FOMO as a traditional actuarial approach to comprehending these advances begins to feel inadequate. 

Some crazy predictions on underwriting innovation

Having dabbled in teaching and being reasonably tuned into tech disruptions, below are some of my takes on how life underwriting might adopt tech innovations. Of course, these are purposefully outlandish predictions to provoke debate within us:

1. In the not-distant future, you hunt for insurance cover through P2P groups; these mutual groups are enabled by white-label insurance platform enablers helping such P2P setups.

2. Us actuaries and underwriters will work on a fee-based model, getting paid by the white label carriers on a pay-as-you-go basis. We would have our irrefutable history of projects, quality assessment scores, and recommendations by our clients on the blockchain (hopefully something to boast!).

3. Insurance would cost way less than now, asymptotically approaching pure risk and some liquidity costs. Blockchain economy (BCE) will break the traditional insurance structures and recombine them into totally new constructs. BCE creates 10X volumes with 1/10th of the cost, as seen in other finance sectors. 

4. Actually, you wouldn’t even notice the providers much, as your own personal AI Bot is taking decisions on your behalf, within overall limits you control. You will have seamless banking, investing, property/business/home insurance (no motor insurance, you wouldn’t be driving anyway) ‘umbrella interface’ optimized through AI for your income and risk appetite. 

5. I would have earned a million dollars by spinning off a start-up.

Ok, the last one was a bit stretched!! But I don’t see why someone young budding actuary can’t make it!  

I was speaking in Data Science & Analytics Seminar recently at Bengaluru organized by the Institute of Actuaries of India (presentation here). I seriously ran out of time. Data on computing power steroids is creating a complex kaleidoscope of impacts; my fascination only makes it impossible to contain it into only so many slides. So here is an attempt to say in a blog-style to do some justice to these slides (having had the liberty of pruning what I wanted to say!).

The falling cost of computing creates two distinct phenomena that directly impact life UW. 

 Underwriting digitization: 

If digitization enables precision medicine, why not underwriting? A spittle = 1.5GB of genomic data, $80 watch = your heart/exercise and sleep patterns 24/7. The current underwriting is based on blood-panel and splits humanity into 20–40 data points; at this resolution, millions look very much alike. N=millions. Instead of this pixelated outline of a human, you now can get a high-resolution image of an individual. N=1. 

Humanity in Hi-Def will look qualitatively different; this would be a game-changer for UW. Tech biggies are already in this space; Google DeepMind has a blockchain around Clinical Trials, a particular use case with Electronic Health Records. In addition, global pharma giants Pfizer, Amgen, and Sanofi are adapting blockchain in recruitment for clinical trials, using AI to match profiles for the clinical trials.

Digital co-operatives

Second is more profound BCE-enabled digital co-operatives, which might ultimately dethrone current corporate structures. The decentralized private-key system provides a layer of trust & immutability at a meager cost with absolute safety. This trust significantly takes away distributor margins (distributors/enablers were needed to create confidence in the market), which drives most of our financial structures. BCE designs out there is an early indication of the possibility of breaking parts of financial frameworks to recombine them into totally new systems. Etherisc (, a BCE ICO, deserves a separate post! 

This future is blasphemy to current players, including regulators. Insurance is all about pooling risks. It’s about differentiating, not discriminating. And significant margins are required to enable this social need. All other frameworks currently being disrupted have very similar raison d’être logic for comfort, so why would insurance be different?

What BCE is changing is the scale. Massive drop in the cost of any economic utility creates more than compensating demand, creating totally different enabling frameworks possibly only at the low cost (think of the marginal cost of bank transfer for paying for your 10 INR (roughly 15 cents) tea at a roadside shop, now in India and ten years earlier). This value creation usually succeeds a good amount of destruction of the value of current players.

Underwriting Innovation as seen through Life Underwriting drivers: 

To begin to comprehend the disruptions, it might help to check each driver of the current life UW framework. What are the drivers of life underwriting? In my mind, I see something like below (kind of actuarial control cycle:

Life underwriting drivers

Data & Past Experience

Data is a critical driver. UW practices are based on census, industry portfolios experience, and data from medical research. But this data is not timely as well as not direct

Population and similar aggregate statistics are highly grouped, whereas the medical data samples are too small to cover good breadth. 

Also, the underwritten cohorts with any Insurer (say, + 200% rated diabetic lives) are rarely followed through to validate the cover cost, lacking credible follow-up validation. Hence, there is no direct measurement of underwriting effects until a generation of policies runs off. 

So, what could be changing? All of these: Ownership, access, analyses, regulations of data. You own your health data. If digital money becomes transaction data owned through private keys rather than banks, then your digital health data too becomes money-like. You share only what’s needed, and you get back their assessment of your health added to date, rightfully owned by you. With high level trust and confidentiality, changing landscape of medical studies, because of ability to match profiles, this is a disruptor on its own.

Underwriting will also need to cope with blockchain wallets with healthcare as a vertical market. The way money exchange has moved into a predominantly app-based experience from the cash-check era. With private data securely available for risk assessment, underwriters need to think differently to handle humans in ultra-Hi-Def health view. It’s early days. The data juggernaut is in the embryonic stage of data gathering, and only the crudest correlates have been uncovered so far:

Data accumulates exponentially, correlations emerge inexorably, and the traditional way of underwriting suddenly seems quaint.

Customer behavior

Target market profiles drive current underwriting practices as insurance is sold. Each would be underwritten differently; savings (fixed/variable types), term products, unit-linked (and such unbundled products), online products. But this traditional push model is being disrupted. 

Chat-bots, impersonal interactive web tools will dominate, with access to better personal information. At one extreme, complete adaptability, customization, and super quick turnaround time for all needs might create a tech-immersed sedentary group that isn’t healthy. On the contrary, an informed and healthy class of customers might become a significant customer base. In that sense, the contrasts within the groupings by the current framework will widen. Current Straight Through Process structure for rule-based and segregation (for pricing pools) resulting from pixelated version would be outdated. 

So how do we adapt our snail world to this? Further, fundamental assumptions of sharing risk over peers as a cohort and across generations might need a new moral and economic lens. For example, how will we ensure funding risk cover for needy poor, that now stand out as poor health class because unaffordability drives them unhealthy?

Possible future? Digital contracts that offer flexibility in the change of sum assured and other conditions on the fly, and the contract also adjusts the premium (Lemonade just launched their digital policy project)? Future generations, shy of long-term commitment, will drive pay as you go insurance & hence we need continuous UW? 

With more and more social data available, social trends will also influence lifestyles. Therefore, we will need new underwriting rules based on sufficiently large and detailed anonymized group data. In addition, the underwriting framework must adapt from a push-based sales perspective to large hybrid P2P group insurance.

Sales ease

Current UW implicitly assumes multiple communications. For example, insurance products with long-term individual financial needs require financial UW on top of health assessment. This complication requires experienced and informed distributors, needs a significant amount of information exchange, and takes time.

FANG (Facebook, Amazon, Netflix, Google) have changed marketing. Why not insurance? Lemonade and home-grown Acko are rewriting game rules. Insurance distribution will become impersonal and tech-driven. Life-cycle stage needs will be changing with new social norms driven by digital economy soldiers. Insurance may still need to be sold, but with less effort. As firms become transparent (Lemonade), trust in the digital ecosystem grows. Global Techs are getting in this space of delinking savings and protection (unbundling) to meet transparency and digital ease.

Underwriting rating system needs rethinking – ‘standard’ concept not so rigid. With reduced info cost and trusted instant access to medical reports, medicals may be things of the past, relying on regular check-ups, augmented by social data AI. Read lesser forms to be filled and instant underwriting. 

Underwriting will adapt to new risk pools through totally different distribution models. On-the-fly UW? Current distributor margins significantly gone, Bitcoin (or like) denominated insurance? Tools/ interfaces to continuously tweak the right insurance for you rather than current frozen contracts – AI/human mix of long-term association?

Cost & margins

UW has significant setup costs: Need to develop and continuously fine-tune underwriting philosophy and systems of rules. Accommodating a new product design with medical/ financial risk combos and training distributors costs a good fortune. Seasoned underwriters cost even more. This cost needs to be and has been covered by margins built-in the premiums. Because of this expensive framework, small-ticket insurance becomes unaffordable, leading to a large insurance gap, especially for the poor socio-economic groups.

This is where the digital economy will cash in and create disruption. PayTM (bank), Uber (transport), Oyo (hotel), Alibaba (retail): each of them disrupted core industries by creating digital structures that are operating at way below traditional margins. Why not insurance?

What’s changing? Innovative service models work at a lower cost with a better feel. Markets are being liberated from boundaries – offshoring creates access to cheap knowledge workers, increasing computing power removes inefficiencies within the traditional structures.

What does this mean for underwriting? Current structures aren’t good enough; innovation is required to handle a fast-changing lifestyle and data-driven behaviors. The underwriting philosophy needs to be agile to accommodate increasingly rapid medical advances. The systems need to evolve to interact with multiparty systems. The outcome will be a fluid and responsive underwriting philosophy and process. And all in the face of immense cost pressure and competition.

Future? AI and cloud will lead to change; many InsurTechs are chipping away at these challenges. Underwriting systems with API capabilities will become system/framework agnostic. AI will play a more prominent role – streamlining – onboarding – directing to competent UW – backward claims integration. Shared information pools become easy (IBM blockchain) with more helpful analytics. 

Far out in time, professional UW guilds might be formed providing white-labeled services – pay per UW case?


Regulations focus on fairness and equitability, desire fraud prevention, and fair arbitrage of information asymmetry between insurers and the insured. But the current framework is significantly onerous with attached costs – limits the insurance design & underwriting framework. The insurers crib about regulations on the surface, but they have enjoyed this very effective entry barrier. Regulations, post-2008, have increased focus on systemic failure. Ironically, more disclosures focusing on privacy protection and capital have increased the frictional cost and strengthened the entry barrier. More the case for decentralized structures, disruption may happen earlier in insurance than banking.

What’s changing? Innovations are boosting stagnate GDPs, and globally mobile innovators are making Regulators competitive. Moreover, strong currencies backed by democracies actively encourage InsurTech innovations and set up regulatory sandboxes. 

What does this mean to UW? Regulators may relax core activities restrictions by allowing personal data in a non-discriminatory way. Private key-public networks will ensure data safety. Lots of data also might mean systems/ratings/procedure changes, needing further investment and costs – connector APIs. Regulatory costs and speed of regulating will be adapting faster too enabled by new technologies, the trust in the system may even reduce the need for the regulations. 

Bundled need and enabling framework

Insurance is a service facilitating sharing of financial risk. This facilitation results in the reduction of risk capital collectively (without insurance, individuals will need to set aside more money for rainy days). But the delivery of this capital reduction needs an elaborate framework consisting of needs/ preferences research, pricing experts, selling, underwriting, policy/claims management, all with high regulatory costs. Complex setup is needed because understanding and managing the pooled risk is more complicated than simple money exchange. So you get why the insurance business is risky, requires a long gestation period, capital incentive, and needs deep pockets.

In the current framework, the capital required to support aggregation and the cost needed to establish and run the insurance carrier setup are inseparable. This intermixing of costs is very similar to the inefficiency of clubbing your insurance need and savings needs; both need separate approaches. 

The result is that the end-user poorly understands insurance and the framework. In addition, the complexity drives forced trust through many checks and regulations. This high cost is an effective entry barrier leading to a legacy mindset resistant to change. In this sense, the underwriting is slow to evolve and inherently costly. 

This complexity BCE is poised to attack. BCE is enabling demographic trust at a nominal cost. Anything of value can be tokenized and traded. Tokenized trusts enable breaking the financial services verticals into their components and recombining to create entirely novel structures. Like Oyo split the property ownership, asset maintenance, marketing, payment, and customer complaint handling.  

At a minimum, the insurance business can be split into expected risk value, capital costs, and transaction cost (insurer management cost). The first two are systemic, and the third one is specific to entities generating entrepreneurial revenue. These can be separately tokenized through decentralized structures with standard protocol. BCE can enable very low cost and trustworthy interfaces for handling conflict of interests, information asymmetry, and access to risk pools. 

BCE can create entirely new structures for life underwriting too. New underwriting setups need to be far more agile, scaleable, agnostic to most price differentiators. They will need to work with different systems providing data and rule engines, handshake with distributors, analytical data services, or even fee-based underwriters. The amount of computing power behind all this will need to be seriously superior to now, but we will get there soon. 

Insurance professionals like actuaries or underwriters become independent advisers, working as need-based services with their credibility established by immutable records of work quality. Independent service providers may enable new constructs like underwriter guilds. Community of users and companies, enabled by BCE, might evolve into large P2P structures supported by agnostic companies. Different BCE systems will be making interactions possible between pure risk markets, capital markets, and independent SMEs.

If these predictions seem to be exaggerated, remember those days when the internet began. Not many of us didn’t think of possibilities beyond sending e-mails back then! 

Some of the early indications below:


I currently work full-time at Swiss Re, Bengaluru. The blogs and articles on this website are the personal posts of myself, Balachandra Joshi, and only contain my personal views, thoughts, and opinions. It is not endorsed by Swiss Re (or any of my formal employers), nor does it constitute any official communication of Swiss Re.