A technological transformation

08-09-2021

A technological transformation

The captive insurance industry is facing a revolution, shaped by disruptive technologies such as blockchain and artificial intelligence. Marcus Schmalbach of Ryskex reports.

“A smart contract can be set up to automatically make a payout if a certain event occurs.” Marcus Schmalbach, Ryskex

The insurance industry and the captive sector are in the early stages of a technological transformation. Since the formation of insurtechs such as Lemonade and Hippo, the industry has been shaken up, with traditional insurers now looking at the digital transformation of their processes and business models.

What is really going on? Is it all just hype, or is it the evolution and revolution of the captive industry? Let’s take a look at some of the most important technologies.

Factor 1: artificial intelligence + big data

The first important technological development to examine is the use of artificial intelligence (AI) algorithms to mine large datasets, known as “big data”. The scientific method was developed in the 17th century, powered the startling advances of the enlightenment in the 18th and still underpins most academic reasoning today.

The method has three steps: abstraction, modelling and reapplication. When a scientist is faced with a messy, complex real-world problem the first step is to simplify or generalise it through abstraction.

By making certain assumptions, omitting minor variables and ignoring feedback effects, the problem can be modelled in a simplified form and captured in a theoretical construct with mathematical notation. The specifics of the situation are replaced by x’s and y’s to make a general model.

Once the model is constructed it can be used to make predictions. Putting in different values for x and y will illustrate the range of possible outcomes. So, reapplying this to the real-world problem, some useful conclusions can be drawn, policies implemented, and inferences made.

To summarise, the traditional scientific method involves a trip from the specific to the general, followed by some theoretical number-crunching and then a reapplication from the general to the specific.

There are problems with this approach. An oft-heard complaint is “that’s ok in theory but it does not work in practice”. The problem normally lies in the first step: the abstraction.

Once you move from the specific to the general you intentionally leave the context behind. But the context is sometimes the most interesting part. Things taken out of context will often be misleading That’s often a reason that models fail—the contextual detail left behind in the abstraction phase was very important.

There is an alternative approach that is now commonly used in AI algorithms. Instead of ignoring the context, the algorithmic approach is focused on it. It effectively reduces the three steps of the scientific method to a single step: from problem direct to solution.

To launch any new insurance product, you need claims data in order to model risk. Without data you cannot offer insurance, but with no insurance offered you cannot get data. The AI + big data approach gets around this conundrum by using data scientists to create an index from non-traditional sources, thus obviating the need for historic claims datasets.

It is wonderful to see that various students in the risk and insurance industry have already recognised this trend for themselves. One example is the Q-squared Project, a student-led organisation that participates in quantitative and qualitative research surrounding quantitative finance, risk analysis, business analytics and linking this with AI technology.

Factor 2: the platform business model

The second technological development is the platform business model. The titans of the internet are all platform businesses; Apple, Facebook, Google, eBay, Uber and Airbnb have all been remarkably successful in exploiting this type of business model.

It took new technologies such as mobile phones, social media and the cloud to properly unlock the full potential of this approach, but the underlying concept of a platform business is actually quite ancient.

Let’s start earlier and go back to London in 1867 and the Hop Exchange. It has a vast open atrium with three tiers of balconies overlooking it, designed to allow “open outcry”—traders on the floor and merchants on the balconies shouting their orders to each other (the Lloyd’s building in London has a similar atrium and balcony design).

Victorian developers built it in a burst of progressive optimism hoping to capture and consolidate the hop trade inside its walls. But the hop factors and merchants already had their own various premises and saw no reason why they should move.

The Hop Exchange was an attempt to cast economic activity in an architectural form. Sadly, it did not work. This building, designed to house speculators, was itself a spectacular failure. Not a single hop was ever traded there.

The Hop Exchange was a Victorian version of a platform business. The aim is to create an environment in which buyers and sellers can meet and conduct business.

Normal businesses design and build a product and then try to sell it. With platform businesses, the products are created by the users and only the rules and the tools of exchange are controlled by the platform owner.

Competitors for normal businesses are those companies who have similar products, and so product differentiation is a key goal. For platforms, competitors are those who have the same pool of users as you.

So, the competitive race is to grab as many users as you can, as fast as possible, to achieve critical mass. The dominant platform then gets major economies of scale due to network effects, which is why Google has a 90 percent market share of the internet search market.

Platforms have oversight but they don’t have foresight—they don’t know what content will be put on the platform, but they do define how it is put there. Traditional insurance policies try to use foresight to anticipate possible future events and then exclude them or draw causal chains in anticipation of particular types of damage.

Parametric insurance, on the other hand, does not need to establish chains of causality. The only thing that is important is that the event happened, not what caused it. There is no foresight needed. That makes a platform-style business ideal for parametric insurance.

How to get to critical mass? The Hop Exchange failed because there were no clear benefits for the merchants over using their own premises. In order to attract business to a new platform there must be clear advantages over the traditional way of doing business.

A platform that allows risk holders to trade insurance-linked securities (ILS) has two of these: it is cheaper due to the cost savings in the claims process, and it is easier because of the third technological revolution.

Factor 3: blockchain technology

The third important development in technology is blockchain-style distributed ledgers, which is the enabling technology behind smart contracts.

Traditionally, financial transactions require a trusted third party to facilitate the exchange of payments and securities. This financial clearing house acts as central counterparty between the buyers and sellers, acting as a guarantor for the transaction. The counterparty risk is transferred from each of the participants to a trusted and highly regulated central authority.

Often the actual assets in question, such as stocks and shares, are held in a central securities depository. Trading shares then just becomes a book entry change rather than a physical transfer of certificates. Identity and ownership are verified by a centralised database owned and controlled by this trusted middleman.

A smart contract does away with the central database and uses a distributed database on a decentralised peer-to-peer network instead. This transfers the trust element from a central counterparty to the cryptography behind the technology.

Bitcoin was the first example of this distributed ledger technology, known as blockchain, but there have been many other blockchain variants since then such as Ethereum, Cardano and Litecoin.

While Bitcoin’s main appeal was in the anonymity of its transactions, Ethereum is designed for commercial use as an enabling technology to support smart contracts which can be settled in Ether currency. The public and private identifying keys that are required for Ether transactions are stored in a cryptocurrency wallet. The audit trail of these transactions is embedded in the blocks in the chain as a permanent, tamperproof record.

The type of information that is stored in this blockchain ledger can be items, actions and permissions. As an example, the items can be the clauses of a contract, the actions can then specify how a transfer of funds will take place and the permissions will work as the trigger for that transfer to occur.

A smart contract can be set up to automatically make a payout if a certain event occurs—no muss, no fuss. There is no need for verification by a human third party. This makes it the perfect underlying technology for parametric insurance.

Since the payout terms and criteria are baked into the code you have a fast, secure and transparent system which should help rebuild trust and also be much cheaper, as processing costs are dramatically reduced.

Shaping the industry

The captive insurance industry, and the wider re/insurance industry, are under scrutiny. The technologies presented above will shape the future of the industry.

The industry is set to take advantage of the technologies, and these technologies will enable the move of insurance into the era of risk financing.

Ryskex, Marcus, Schmalbach, blockchain, insurance, business, platform, technology

Captive International