One way to approach that is to have a mindset of confronting the status quo. What I've observed in my nineteen years is a great deal of cultural inertia that exists at the company level and at the societal level. I'm not a credential actuary, but I’ve talked to associates about the changes to the data available to do different types of analyses and technology and a lack of willingness to react to that. Adopting change, confronting it, conceiving that there is change is a major component to it.
I would also say it comes down to kind of skill sets and capabilities; amplifying the idea that beyond technical actuarial acumen. There's a great deal of skill required as it pertains to the technology and mathematical concepts that might fall outside of the traditional actuarial realm. Also, the softer skills and understanding the role of the actuary in the broader strategic decision making at an insurer.
There are factors that are relatively in control of the actuaries like honing the well-rounded set of skills in those dimensions. That, to an extent, is controllable by the actuary and the actuarial department. There's the broader ecosystem as a second factor; the rest of the organization an actuary is working in. That's a huge factor that is less in their control; to what extent the non-actuaries give a seat to the actuary at the strategic table. How much are the actuaries lab mice kept in a corner versus are they embedded in the company’s strategic decision making process from beginning to end? I think it’s a philosophical or cultural difference by company. It also varies by the line of business. There are some factors that, to an extent, are controlled by the actuaries themselves but then there's the entity within which they're operating.
Where I work right now, at Teradata, we talk a lot about this notion of model of one, and we have clients that have millions of models. I push back internally to the non-insurance people at Teradata that, for statistical credibility, we're not going to have a million pricing models. But do I see the industry as a whole is going in the direction of more models. If you consider a typical scenario of fifty states, legal entities, and five or six coverages, you're looking at hundreds of models that should be built and perpetually maintained for pricing purposes. Right now, many carriers put many of those models on the back burner. You look at their rate filings and they’re stale for years. But I do see some niche carriers challenging the status quo of how these pricing models are built and calibrated. Now insurers are talking about building and managing thousands of models, whereas before it was ten. I see that dimension of data science and model operationalization becoming more ingrained in the actuarial function, at least from a pricing perspective.
One is to remember and keep in mind that the exams are crucial. I’m an actuarial drop out, but I credit much of my understanding to that upbringing. The numerous times I had to flip through annual statements, I use that day to day now when I go to look up and analyze a company. That’s something my peers cannot do, especially for companies that are not publicly traded. As far as advice to actuaries, look beyond the exams. No matter how many exams you take, there are other elements required to have that well rounded skill set and experience. Acknowledge and recognize what's happening around actuaries in the world of the cloud, artificial intelligence, and automated machine learning. A lot of what actuaries do, they don't necessarily call it AI or ML, would more or less be considered under those terms. In other words, don’t sell yourself short. Don't be afraid to elbow your way in there to show non actuaries what you’re capable of. Actuaries represent some of the most intellectually sharp professionals in the insurance business and they grasp concepts in a certain way that few others can. And if they want to broaden or change their horizon – for example in the digital transformation space – they can absolutely do so.
I'll use pricing as an example. Let’s look at a bread and butter generalized linear model to predict pure premium, claim frequency, or claim severity for many lines of business. When it comes to the actual mechanics, the latest algorithms to look at all the co-linearities, and the marginal value of the nineteenth variable, a data scientist might be better equipped to address the technical model build. The actuary understands the same concepts, but they are also thinking about how the rate must be sufficient to cover my claim liabilities, but not unfairly discriminatory. The actuary is thinking about how product management is going to consume this and file it, how California is going to be more difficult than Ohio, getting the rates through, and so forth. How actuaries and data scientists work together comes from decomposing what the technical, organizational and cultural aspects of solving a particular problem are, then aligning that to the respective relative strength of data scientists and actuaries.
We can also use claims analysis to illustrate. Let’s say you give a data scientist one hundred claim variables across millions of historical claims and ask for a target. The target variable being, what's the cost going to be? What is the claim liability going to be at twelve months or twenty-four months? A data scientist is going to be stronger than your typical actuary at taking those hundreds of variables and working with your data and Amazon S3 buckets, doing the modeling, and using Sagemaker to implement the analytics. An actuary will know about what it means for the end result, how it impacts the carrier solvency, or how much capacity is available to sell new business to drive geographic expansion. That’s another example of acknowledging the multidisciplinary skill sets needed to solve a particular problem and understanding where the actuary is stronger, and how data scientists can fill a relative deficiency in the actuarial world.
The short answer is innovations that are adopted and have a tangible impact. It’s one thing if you have a better genie coefficient on a model but can you say that you helped drive “x percent” improvement of loss ratio? Innovation is impactful when it has changed the outcome of a particular business process or KPI in a way that you can reasonably attribute it to that innovation. I see impactful innovation as driving tangible improvement, financially or in the human impact.
I've seen a couple of startups in the last four years founded by actuaries who grew up in different roles and feel the way things have been done for decades is broken. They have to call this person in the data department who's been on payroll since 1982, and get the answer from them on why this particular data field is this way. Why is there such a dependency on the way things have been done? It gets back to the cultural inertia to adopt change for a variety of reasons.
There are actuaries that have been bold and removed themselves from the traditional insurance institution to create innovative software to address these widely acknowledged pain points. I’ve worked in the actuarial field and there's such an entrenchment of ways to do things historically. These startups have had a great deal of impact negating that. They’ve had a big impact with clients as far as being able to churn out a larger volume and a number and quality of models in a variety of capacities.
I think that begs the question, what is the objective of an actuary? To oversimplify it, they do pricing and reserving. Both of those roles are incredibly important to the end consumer, whether it's directly the personal consumer or corporate clients in the form of commercial insurance. If I use that as a premise, the way to be more influential is to bring greater efficiencies in pricing. It’s the age-old problem of matching premium to risk. Regardless of how exactly you want to define that, the more that actuaries can objectively tie premiums to what the actuarial risk is, risk of loss, in such a way that the purchasers of insurance have greater transparency into why they are being charged a certain premium and what they can do to control that. That’s a way to be influential. You can help make the insurer more profitable, but there's also the human component to it; the broader impact that an actuary can have in society. Sure, it’s about making people's lives less painful at work by automating things, but also having a broader positive societal impact.
