Admiral Group plc - bigger by being better

Admiral Group plc is one of the largest car insurance providers in the UK. The company is expanding into Spain, Italy, France, the USA and Mexico. The company has the highest per-policy profit within its peer group which is a clear indication that the company has competitive advantages.

Unit economics

Of the 31 million drivers in the UK, about 4 million are insured through Admiral. According to the Association of British Insurers (ABI) the average annual premium in the UK is £520 and the average claim is £3 000. Averages can be deceiving though. Teens sometimes pay a premium well in excess of £1 500.

Selling car insurance to a large group of 65-year-olds requires effective marketing and low overhead costs. Bear in mind that the gross value of the 12-month contract for this group is roughly £300 and the cost of cliams is £180. The administrative and marketing costs are a relatively high fraction of the value of the contract and profit depends on the ability to keep these expenses well below £120 per policyholder. Of course, if you're selling a customer multiple lines of insurance, some of the expenses can be spread across all those products.

An effective strategy would be to keep costs below £100, charge a large and easily-identifiable low-risk group £350 per annum and pretty soon you’ll have a portfolio with many thousands of profitable policyholders. If you don't raise prices too much, they are likely to renew the contract which keeps marketing costs low, creating a virtuous cycle.

There are a number of online insurance companies doing exactly that. By keeping costs low, they earn a decent profit while taking market share from the incumbents. These low-cost insurers are Admiral, eSure and Hastings. Direct Line was the original low-cost car insurance company in the UK but after various mergers the company now sells its insurance through agents. These agents charge fees for their services and this drives up costs.

eSure, Hastings and Direct Line target the low-risk segment of the population. They correctly identify the low-risk 49-year old married school teacher and gain market share by undercutting the large conglomerates like Aviva and RSA. These large conglomerates have expense ratios of more than 25%. Their car insurance operations are not profitable. They make up for this by selling multiple lines of insurance.

For its part, Admiral has the highest per-customer profit by far. Hastings is a formidable competitor that is obviously targeting low-risk customers with rock bottom costs. Hastings also has lower per-policyholder revenue. Admiral's premiums are £100 above the premium charged by other low-cost insurers. The company also pays out roughly £50 more in claims.

In sum, Admiral adds value by correctly pricing risk for higher-risk customers.

Pricing risk in the low-risk segment of the market relies more on assessment of homogenous risk within certain groups. Individual case experience is less important than portfolio experience. Another way to put it is that claims for married school teachers are fairly randomly distributed within that group. Accidents are just that, accidents. Knowing that they are married and are teachers is enough to price risk for this large group. Developing a sophisticated underwriting model for subsets of this group is probably not worth the effort. Your time is better spent tracking the cost of headlights and airbags for the most popular cars driven by teachers and using those inputs to asses the expected cost of claims.

For newly-licensed teens though, case-based assessment is more rewarding. The risk is more heterogeneously distributed. There are some 400 000 teen drivers in the UK (age 17-19). That's roughly 1 percent of the 40 million UK license holders. These teens are involved in 9 percent of fatal and serious crashes. Within that group, there are about 10 000 drivers causing accidents at a rate of once a month! If just 1 000 of these idiots somehow find their way into your insurance portfolio then that’s a £30 million problem. It would be the end of your low-cost strategy. Conversely, an insurer with a predictive model that can filter out these drivers from the much larger pool of teen drivers has a significant advantage when calculating the correct rate for a teen driver.

This requires proprietary models based on enriched data and years of experiential evidence within that niche. These models allow insurance underwriters to analyze millions of different combinations of variables. In this context those would be things like age, years licensed, age of the car, web browser used, occupation, time of day, address, etc. These variables all help to correctly price risk. Once an insurer has established itself within a given high-risk niche, it can further refine its pricing model with incremental data and (artificial) intelligence.

A would-be competitor faces years of unprofitable operations before it can hope to develop its own and hopefully superior predictive models. What's more, a new entrant will initially attract those customers that are dissatisfied with the price offered by the incumbent. That is a problem in itself if the incumbent is doing a half-decent job pricing risk.

Is it sustainable?

In the short term, Admiral generates excess profits because it has superior underwriting sophistication. The company is better at pricing complex risk. In the long term management can maintain this advantage by:

1) Maintaining a strict focus on relatively complex short-tail insurance. Maintaining the dynamic pricing models requires rapid feedback. With car insurance, an error in your model becomes apparent within weeks. That is not the case with policies spanning decades such as life insurance. Profitability in the long-tail insurance business depends more on superior investment of the float. Agile development of sophisticated pricing models works best when your whole enterprise is focused on short-tail insurance within complex segments of the market.

2) Having a patient culture when it comes to developing the required underwriting sophistication. Developing superior predictive models requires many years of deliberate sub-scale operations in a given market to collect data and refine your pricing models. You scale up if and when you've found you can consistently undercut your peers while writing profitable insurance in a given market segment.

3) Maintaining a focus on the core strength (underwriting know-how) and outsourcing the rest. Admiral cedes a very large part of its portfolio to reinsurers. In a sense the company is an online retailer for large reinsurance companies and profits are promptly returned to shareholders in the form of a growing stream of dividends. This leaves Admiral's management free to focus on what it does best and that is pricing complex risk. It is worth noting that many of Admiral's senior managers, including its CEO, have a background in underwriting.

With the possible exception of Geico, which is part of a large reinsurance company, there is no insurance company that has such a strict focus on underwriting sophistication. The business model of most insurers is about avoiding losses and investing the premium at a profit. Their CEOs typically have a financial background. Their favorite game is exploiting the spread between the cost of float and the return generated from investing the premium. This has important effects on company culture. As an example, Admiral has a reputation for settling claims as quickly as possible. Dragging on litigation is expensive and settling quickly reduces costs. At other insurers, there are incentives to drag on litigation. This allows the insurer to invest the float for longer. This works well when the value of a policy is much higher (as with life insurance). The legal department of an insurer will find it difficult to do both well.

In sum, most insurers try to get better by being bigger. Admiral gets bigger by doing one thing better. That thing is correctly pricing complex risk. This focus is embeded in the company's culture and it is difficult to replicate.

The measure of Admiral's moat.

Compared to its peers, Admiral's annual per-policy profit is at least £45 higher. Admiral's loss ratio is significantly lower than the industry average and at least 5% lower than the next-best underwriter, Directline. These are the most important measures of the strength of Admiral's moat.

First published 15 feb 2019 by batbeer.

Sources and further reading:

https://boughtbymany.com/news/article/how-much-does-uk-car-insurance-cost/

https://www.abi.org.uk/products-and-issues/choosing-the-right-insurance/motor-insurance/age-and-motor-insurance/

http://www.brake.org.uk/news/15-facts-a-resources/facts/488-young-drivers-the-hard-facts