Big data, used in insurance product design, presents actuarial opportunities but also potential threats, both to the profession and to public interest, says Brian Gedalla
'Big data', after 'Solvency II', is probably the hottest topic at actuarial seminars in recent years. Every self-respecting consultancy has an opinion to give or some cute software to sell.
But what does it mean in the insurance world and should actuaries be worried about it?
If you were to ask the pricing managers from personal lines insurers what big data meant to their firms and how they were taking advantage of new data sources, my guess is they would say their companies had been big data users for decades, utilising ever more sources in increasing volumes and sophistication. We are constantly told that new sources, such as social media and data mining techniques, are key to identifying potential customers and developing pricing models, but this is where conduct risk and business ethics come in.
One of the principal aspects of conduct-risk awareness we encourage insurance businesses to adopt is the need to always keep customers at the heart of every consideration. That involves thinking about customers throughout the lifecycle of a product, from the earliest concept meetings to the design phase and pricing process and on into sales, post-sales and claims. Most importantly, it also includes identifying the target market for a product right at the start of the process.
How comfortably does that sit with new methodologies, largely based on data mining techniques that set out to 'find' new marketing opportunities otherwise hidden in the data?
In truth, none of this is new as insurers have been on that path for decades. I started out in the personal-lines pricing department of a major insurer when firms still considered themselves to be 'tariff' or 'non-tariff', even though the tariff system had actually been abolished some years earlier. The concept of providing insurance 'from cradle to grave' still held sway. The arrival of new entrants selling direct to specific market segments was met with much disapproval by the old hands.
These newcomers realised the old paternalistic approach meant that young customers were charged less than office premium and overcharged in later years to balance the account. By aiming only for 35- to 50-year-olds with good records and with no young drivers on the policy, they could carve that specific market away from the big players and rapidly establish themselves.
Today's price comparison websites, price optimisation techniques and other methods are only the logical extension of the same process, which is possibly gathering pace as well as being applied to other classes of insurance beyond the traditional motor.
Could market segmentation as an underwriting and pricing tool become over-segmentation to the detriment of whole groups of customers? Nobody is disputing the cleverness of modern actuarial modelling, but if teenage drivers cannot afford to buy insurance, then how is the industry that created this situation serving the public good? If house-purchasers, encouraged to buy properties in new estates, find their homes are uninsurable owing to flood or other risks only identified by sophisticated models after purchase, is the industry serving them? And what of private medical insurances that become essentially unaffordable at precisely the point in life when their consumers most want them?
Over-segmentation
In short, is micro-segmentation damaging, or even destroying, the very principle of insurance as a pooling of risk? Is the marketing ploy of 'why pay for his mistakes' nothing more than the latest manifestation of 'I'm all right, Jack'?
This leads to the question of actuarial responsibility. Going back to my early role in personal lines, there wasn't an actuary in sight. Most pricing was done by underwriters with a sprinkling of statisticians to help out. But then the pricing models started to get more complex. Actuaries have been very successful at creating roles as pricing experts, and pricing actuaries have held the reins in most major insurers' pricing functions for many years. Good thing too, I hear you say!
There are good reasons to agree with you. Actuaries are professionally trained, members of a body that regulates their behaviour, subject to a code of conduct and held to account under their disciplinary scheme. The IFoA is now a participant in the Joint Forum on Actuarial Regulation (JFAR), under the aegis of the Financial Reporting Council and alongside the three main industry regulators (the Prudential Regulation Authority, the Financial Conduct Authority and the Pensions Regulator). JFAR has set out to identify the risks actuaries contribute to and how actuaries can help to mitigate them. It provides a useful forum in which all the regulators can discuss these issues and agree how to address them.
Perhaps more importantly, actuaries are now trained to see themselves as key players in the management of the companies they serve. It is right that today's young and aspiring actuaries no longer want to be back-room technical specialists but dream of a seat on the board and an office in the C-suite. But this comes with an increased sense of responsibility for all of the business' activities, way beyond the models and the numbers, to ensure that the business meets all of its obligations, particularly to its customers. Failure to do so can and should bring due censure from both regulators and professional bodies. That increased understanding of professional duty and responsibility could well be one of the best protections against customer abuse.
We can reasonably expect that an actuary will start from the principle that the product and pricing structure must meet customer, as well as business, needs and that customers will not be unfairly disadvantaged by the design. Conduct risk means putting customers' interest first and that is what is now demanded of businesses. It certainly doesn't mean "the risk of being fined by the Regulator", as has been known to appear in some firms' business plans and risk assessments!
Could the role of the actuary be under threat? Actuaries are expensive to recruit as graduates, while the cost of training, providing them with support through their exams and meeting their salary expectations makes them much more costly than almost any other group of employees. Once they have qualified, their expectations again make them very expensive.
Profession of choice
Are actuaries necessarily the people of choice for businesses trying to take advantage of the latest theoretical developments in data exploration and model design? How many firms are turning to PhD students and post doctorates to do this work? In many cases, these people will be markedly cheaper than actuaries, and certainly less onerous in terms of development commitment than actuarial students. They may also have skill sets no less applicable, and perhaps more so, to the task at hand.
Will we see specialist data scientists take over roles that we have come to think of as the 'traditional' preserve of the pricing actuary? Will firms start to feel that they no longer need their head of pricing to be an actuary at all?
If that happens, who will monitor their behaviour? Such specialists need have no professional membership or qualifications. They might choose to be members of a learned society such as the Royal Statistical Society. They might even choose to take that society's 'Chartered Statistician' status, but they don't have to. Without a formal professional qualification, they have no obligation to take part in continuing professional development (CPD) - even an affiliate member of the IFoA has no CPD obligation. Likewise, they have no requirement to attend professional skills training, no professional standards to uphold and no professional body to hold them to account if they fall short of the standards actuaries are expected to maintain.
While firms are free to decide to employ actuaries, data scientists, statisticians or whoever they please, it is for actuaries to make the case that their professional education and standards make them much better equipped to take a business forward than others whose skills are perhaps purely technical.
There is a bright future out there for actuaries, but you need to make it happen. If other professions take it from you, it is unlikely to lead to the best outcome for customers.
Brian Gedalla is employed by the Financial Conduct Authority. He is a chartered statistician and affiliate member of the IFoA
The views expressed in this article are the author's alone and should not be taken as those of the FCA