There are many changes affecting insurance at the moment and actuaries are at the heart of many of them. Yet as actuaries ride this wave of transformation, are they ready for one of its most important outcomes: a new era of accountability in insurance. It could present the profession with some profound challenges.
A new era of accountability in insurance is emerging on two key fronts. Firstly and most obviously, there are the legislative changes. 2018 will see the introduction of the Senior Managers and Certification Regime and the General Data Protection Regulation. Together these move accountability in insurance to a more personal and detailed level.
Secondly, there is the changing nature of societies becoming more digitised. In the past, accountability was largely orientated around the steps that firms had taken to organise and direct their business. In today’s more digital society, that’s changing. Consumers are increasingly concerned with the outcomes that firms are generating. Gone soon will be the days when firms can just point to the right policies and processes, and say that their house is in order. Instead, they will need to account for the impacts their decisions are having.
Some Fundamental Decisions
Actuaries have a significant involvement in the design and operation of the pricing and liability models within insurance firms. Both of those models are being revolutionised by the data and analytics now available. This is resulting in actuaries making some quite fundamental decisions about how premiums are calculated and how claims are settled.
And the SMCR recognises this, designating many of the roles that senior actuaries occupy as ‘senior management functions’. This results in them being clearly identified on the ‘management responsibilities map’.
So when it comes to the accountability of a typical insurer, what type of outcomes are likely to attract the most attention from consumers and regulators? In broad terms, they are outcomes that are unfair or discriminatory. Neither will be new to actuaries, for they’ve been part of legislative and regulatory frameworks for many years now. So what’s new?
New Types of Pricing Decisions
What’s new are the decisions being made in pricing and claims, driven by the opportunities presented by the myriad forms of data and analytics now being used by insurers. Decisions such as around price optimisation, which result in price and risk no longer being aligned. And decisions around claims optimisation, where loss and settlement are also likewise seeing their traditional alignment reassessed. Both are controversial practices that have already been producing outcomes that some are judging as unfair and discriminatory.
Then there’s the overall drift towards more personalised policies and the concomitant price volatility this produces for the ordinary policyholder. It’s an outcome that some think sits uncomfortably with the benefits often emphasised for personalised insurance. And for some, that price volatility can produce what is called a micro-financial crisis, in which the premium or the excess become unaffordable during a period of low savings. There was evidence of this in parts of the UK during the run-up to the launch of Flood Re. And some civil society groups are indicating that their networks are picking up more signs of it in the last year or two.
Market developments like this are being connected with growing concerns about access to insurance. And while it is a problem that continues for vulnerable sectors of society, it is also now seen as problem that will affect many of us at some point in our financial lives. ‘Going without insurance’ is a decision more and more people are having to consider.
Market Confidence
What this adds up to are a range of outcomes that can influence consumer sentiment about insurance. And when that happens, the regulator becomes concerned about market confidence. Read this link for an example of how this is emerging in the UK.
The UK regulator is signalling to insurance firms that their pricing strategy needs to be based around a clear set of principles. And any such principles would be expected to reference the overall ‘Principles for Businesses’ upon which accountability in insurance is based. So an insurer’s pricing principles should talk about things like fairness, and be based upon compliance with equality legislation.
Yet as was outlined above, accountability relies on more than just a nicely written up set of principles. It rests upon those principles being translated into the rigorous design of systems, the testing of outputs against those principles, and the monitoring of outcomes as well. Questions marks certainly exist over how well this is happening at the moment (more here).
What Accountability in Insurance Means
It is incumbent on senior actuaries then to look at how their pricing and liability models are being managed. Are new data and analytical developments being assessed against a template of clear principles? Is fairness testing being carried out? Is there ongoing monitoring for discriminatory outcomes? Does their modelling look at the implications of pricing decisions on vulnerable groups? This is certainly what accountability in insurance encompasses.
And these questions need to be addressed in ways that can be shared with key audiences outside of the firm. For example, if the Financial Ombudsman Service wants to establish whether a particular policy’s price was fair (more here), then they will expect to see evidence of this. Ditto for the Financial Conduct Authority on a portfolio basis.
A Greater Level of Transparency
This new era of accountability is being built upon a greater level of transparency. And with increasing use of artificial intelligence in pricing and liability modelling, actuaries will find the delivery of that transparency an increasing challenge. So there’s a tension here, between the opportunities that AI gives to pricing decisions, and the explainability of those decisions to external audiences.
Responding to those increased levels of eternal scrutiny will not always be a comfortable experience for actuaries. Take fairness for example. Actuaries have a tendency to overly focus on one dimension of fairness, being the fairness of merit. This is because of the profession’s long standing concerns about adverse selection. Yet consider the wider impacts of personalisation and optimisation. If you individualise a price, and detach price from risk, then the whole selection debate changes.
Fairness of merit becomes joined with fairness of access and fairness of need. You only have to listen to experts on algorithmic fairness to realise that merit is just one of many fairness dimensions.
Some Difficult Ethical Dilemmas
If you bring accountability, transparency and explainability together, and then combine it with the pretty detailed level of ethics training that I understand some UK actuaries are undergoing, then it is quite easy to see a time in the not too distant future when some difficult ethical dilemmas will be faced. Consider claims optimisation for example. If it progresses from experimentation to widespread adoption, will some actuaries be prepared to ask simple (but awkward) questions such as “should we really be doing this?” or “I know we could do this, but is it the right thing to do?”
The actuarial profession is approaching an important ethical crossroads. From one direction is coming the incredible opportunities that data and analytics present for pricing and liability strategies. From another direction comes the new era of accountability in insurance, with transparency and explainability tucked in close behind. And from another direction is the heightened understanding their training is giving them in the ethics of insurance. No wonder the UK’s Financial Reporting Council recently identified the work of actuaries in insurance as one of its key ‘hotspot’ risks
The challenge of course lies in how individual actuaries face up to the tough decisions that they are sure to encounter at that ethical crossroads. How they respond will set the scene for how trust in insurance develops over the next five years.
