'Actuaries': What Do They Do?
It’s a common question which every student actuary grapples with time and time again – whether from the interview panel for a City job or from that inquisitive aunt around the Christmas dinner table.
It’s a question which doesn’t have a hard and fast answer, and whose answer itself continues to change as years go by, particularly as the world around us develops and becomes more digital. So, in the current time, how would we answer this question?
Going back a few decades, the role of an actuary was to analyse books of data for, say, an insurance firm, to try and find insight from them and deduct trends using mathematical concepts which could be applied into the future for premium pricing or reserving. You could say actuaries were “the original data scientists”, as coined by ex-IFoA president John Taylor, who dealt with datasets and tried to convert them into logical models. I’d take a guess that the actuaries of those days had no idea of the sheer extent of data which would be available in the future and the countless ways, many of which yet to be discovered, that this data could be utilised and exploited for the benefit of consumers and, in particular, insurers and financial institutions.
Back then it was simple policyholder data – age, sex, health status. Nowadays the parameters of data which insurers can possibly employ spans everything from an individual’s heart rate variability to the number of times they brake on a stretch of A-road; from the number of times a person enters or leaves their home to the litres of water drank by an individual in a day. The possibilities are endless, and with more and more connected devices being used by people every day, the quantities of data will only continue to pile up.
What happens to this data? We all hear about the big tech companies of today, from Google and Facebook to Apple, cashing in on this data, particularly to learn the interests and habits of an individual to tailor advertisements and suggestions for them. It’s data scientists who work with these large swathes of data, using programming techniques to sift through the garbage to find the gold, in this case, useable information. Why, then, can’t the insurance industry do the same and tailor products and premiums to their consumers based off an individual’s data?
It’s no surprise that insurance companies are racing to incorporate data analytics into their businesses, as the market shifts from old traditional methods to the new digital-driven future, and the growth potential is massive. However, there’s a fear that they aren’t doing it fast enough, with new InsurTech start-ups popping up, capitalising on this new, expanding market, leaving the old established firms moving at a snail’s pace. PwC state that whilst 74% of insurance companies see FinTech innovations as a challenge for their industry, only 28% of them are exploring partnerships with FinTech companies. The growth has been huge, with the global InsurTech market being valued at USD 2.72 billion in 2020, which is expected to continue growing annually by 48.8% from 2021 to 2028.
The IFoA itself recognises the use of telematics as an essential step forward as firms look to better manage their risk. Car insurers that collect individual data from their policyholders not only guarantees individual risk profiles are evaluated more accurately, but also enhances customer engagement and increases their retention rates as insurers can provide personalised offers to their customers. One of the UK’s current top InsurTech firms is Bought By Many, a pet insurer that uses search and social media data to sell insurance and disrupt insurance distribution, and offers online form-free claims, a key feature that consumers are shouting out for. It’s clear that the insurance world will never be the same as it was a decade ago, and actuaries can’t afford to remain stagnant and must follow the trend seen industry-wide and the clearest way that they can do this is to involve themselves in data science!
Actuaries vs Data Scientists
So, what’s stopping a student actuary from simply taking a role as a data scientist? Or why can’t insurance firms start employing data scientists to actuarial roles? The two roles can be quite similar; both actuaries and data scientists deal with large data sets. However, actuaries use the data to help understand and assess financial impacts of risk, applying their in-depth knowledge of mathematics and statistics, whilst data scientists use programming techniques to understand and simplify large sets of data.
While there is substantial overlap between the two jobs, in the context of business requirements, the insurance expertise and technical knowledge acquired by actuaries from the professional examinations is still highly desired and necessary; their knowledge of the intricacies of the insurance world is unrivalled. But in a world where big data will become the norm, actuaries need to be able to understand data science and be able to work in tandem with data scientists in order to deliver the core aim of insurers – develop efficient products to give value to both the insurer and the customer.
It’s not a case of actuaries versus data scientists, but rather actuaries with data scientists. A team where a torrent of data is broken down and understood by data scientists and where this data is then manipulated and analysed by actuaries will lead to a very powerful team in the future, especially when there is common knowledge of programming.
A 2018 study by PwC in the United States was carried out to investigate how the two roles would marry within a property and casualty (P&C) (re)insurer. The study looked at the predicted makeup of traditional workstreams within a (re)insurer by 2030, showing just how much the role of data scientist is predicted to infiltrate the traditionally actuary-focused teams; this is visualised in Figure 2. It finds that the long-established actuarial presence within insurance pricing and reserving will soon be shared with data scientist colleagues, as their more sophisticated automated and data driven processes will take centre stage.
A particular expansion for data scientists will be seen in the marketing and distribution of insurance products, as algorithmic science and AI develop to further modernise operations and increase efficiency for insurers. What remains clear from the report, however, is that no matter how much data scientists penetrate the sector, actuaries will still be required to give professional opinion, sign off on important predictions and ensure models are appropriate considering all variables. You simply cannot code your way past actuarial expertise!
What’s more, in the same publication by PwC, a case study was performed, seeing just how well the integration of data science and actuarial functions worked. While initially the new group, consisting of just data scientists and business analysts, was positive, the lack of insurance proficiency among the data scientists led to misunderstandings and incorrect predictions. In response, the analytics group was merged with a subset of the actuarial team, consisting of 40% actuarial and 60% data science personnel, headed by a data scientist.
The new group were able to build a claims severity prediction model, analyse client retention data to identify what led to customers leaving and develop a fraud scoring model to support fraud prevention. The insights delivered by this new team enhanced value to the overall business, with the actuarial-specific domain knowledge complementing the data scientists’ greater prediction and programming expertise, allowing the company to effectively compete with InsurTech companies by developing new innovative products.
Why Actuaries Need to Know Data Science
All of this points to one thing – actuaries need to know data science! Whilst no one expects an actuary, who has gone through many years of gruelling study and examinations, to then pursue another degree in data science, all actuaries should introduce themselves to it in some way. It’s without doubt that actuarial teams of the future will need data scientists, and it’ll be the actuary that knows and understands programming languages, data analytics and predictive analysis that will provide great value to the firm by collaborating effortlessly with their data scientist colleagues.
We shouldn’t see the dominance of data science simply as a requirement for the profession to catch up with the times, but rather see it as an opportunity to further expand the reaches of the actuarial profession. Data science knowledge provides the ability to re-engineer what actuaries do in traditional fields, enabling further bespoke products for the individual customer based on their data. If customers know that their insurance products truly reflect their habits and lifestyles, they will be happy in the knowledge that they are getting the best value for money which may lead them to insure other aspects of their lives which they may not have considered before.
It’s not just in the traditional fields; opening up to data science gives actuaries a platform to expand into other fields. The speed at which big data has exploded onto the scene has also caught other industries by surprise, and now more than ever, having personnel who are accustomed to dealing with data, and who can help them translate it into procuring value and boosting firms’ performance, is a must. And no other individual is as good for the job as a data scientist actuary. New fields for actuaries include the technology sector, telecoms and climate change. With the development of autonomous vehicles, actuaries are the perfect choice for ride-sharing apps such as Uber to analyse and help manage the risks they encounter whilst developing the new project.
Let’s remind ourselves of what data science involves; according to Simplilearn, data science is the domain of study that deals with vast volumes of data using modern tools and techniques to find unseen patterns, derive meaningful information, and make business decisions. These tools and techniques include modelling, statistics, and more importantly, programming and machine learning. The average data scientist must be competent with a number of programming languages – names such as R, C++ and MATLAB come to mind – and various machine learning algorithms – these include regression, clustering and decision trees. In contrast, the extent of the programming ability of an actuary currently involves Excel, and maybe SQL or R. Just like the digital world around us, the world of programming languages continues to develop and expand, and we must keep up to date with new languages and programmes if we want to avoid lagging behind other industries in the new digital age.
One prominent newcomer is Python, which has been revered as the best programming language for machine learning due to its ease of use, versatility, efficiency and speed. It’s no surprise that it’s the fastest growing programming language in the world. In a Society of Actuaries article by Andrew Webster, founder of an actuarial technology company, he emphasises the need for any potential actuary to learn programming languages, especially Python, emphasising that if actuaries want to explore non-traditional opportunities, having Python skills provides a competitive advantage over other candidates.
Luckily the Institute hasn’t been ignorant to the changes happening in the industry and the digital world around us more generally, as they continue to acknowledge, support and research the growing importance of data science to the actuarial profession. It’s fantastic that the IFoA have introduced a Certificate in Data Science alongside the Web Science Institute at the University of Southampton. This training course teaches students the concepts, tools and techniques used in data science, and how they can be applied to the actuarial profession. It’s hoped that more companies will invest in the future of their workforce and see the benefits of putting new actuaries through this course; a data science-equipped actuary can bring untapped potential to any insurer. With any luck, the Institute will take one step further, and realise the benefits of incorporating data science into the existing framework of actuarial examinations and accreditations.
Whilst the years of study and exams affords an actuary with unrivalled expertise, they cannot afford to be complacent. It’s clear that in this era of unprecedented change, actuaries, like many other professions, face a new, shifting environment. The era of big data and machine learning shouldn’t be a daunting proposition for the profession, but rather an exciting, innovative chapter which beckons new opportunities. And unlike some careers facing such change, we know the key to unlock this new chapter, with which we can remain on the cutting edge of this transformation – data science!
After all, if the “original data scientists” don’t adapt and evolve, they’ll end up being the “last actuaries”.
Enter your text here...
Enter your text here...