Lemonade has a huge cast of characters in our international family—from customer experience champions to product managers, underwriters, claims experts, finance pros, writers, and creative directors. Everyone has a unique role to play as we do our best to prove a simple fact: that insurance is fucking awesome.
Some of these roles can seem shrouded in mystery… and math. For instance, we’ve all heard of actuaries, but how well does the average person know what that job really involves?
We asked DJ Falkson, Lemonade’s Director of Actuarial, to break down the field for us. Don’t worry—we kept it light, breezy, and layman-friendly, as if we were chatting at a bar or dinner party (remember when those things were possible? Such fond memories)…
So, let’s say I know next to nothing about what being an actuary entails, and I’m slightly scared of math. Both are true! What’s the 10-second elevator pitch of how you’d explain what an actuary does?
DJ FALKSON: Actuaries predict the future using a combination of reading tea leaves, animal entrails, and tarot cards to—
Sorry, my lawyer is telling me to start over.
Actuaries use past insurance data combined with market knowledge to figure out how to keep insurance companies in business, by setting premium prices that adequately reflect underlying risk and setting aside enough money to pay for claims that have already happened (reserving).
Damn, I was kind of hoping there was some actual tea-leaves-and-entrails magic involved… What sort of data do you rely on? And where does it come from—does Lemonade have a secret vault of information, or is stuff that’s already out there in the public domain?
The most important data we have is past Lemonade losses and premiums. We are trying to predict the ultimate level of insurance losses in the near future, so the losses from the recent past and how they have trended over time can help inform this.
But Lemonade is a young company that doesn’t have a ton of insurance losses on the books yet. So we are also seeking out new sources of data—things we can collect internally on our customers through our A.I. bots, things from 3rd-party sources, and even things that are in the public domain.
Sounds pretty high-tech! The world is so unpredictable though—seemingly more so every day…doesn’t it get difficult trying to predict what might be around the corner?
The world is crazy and unpredictable and 2020 was a great example of this. But at Lemonade we are only focused on a subset of the risks that are out there, namely the ones we cover in our promise to policyholders. Fires, windstorms, theft, burst pipes, dog bites—these are some of the named perils that we are trying to predict.
Still, it is very difficult to predict how many claims we are going to have (aka claim frequency), and how large they are going to be on average (claim severity). And that is just at the aggregate level—we also need to predict frequency and severity for segmented groups of policyholders based on their underlying risk, or else we would just be charging the same premiums to everyone.
Another nuance is that large-scale events that affect many people are fundamentally different than the day-to-day losses any of us might incur: consider how the losses from the recent wildfires in California differ from a kitchen fire that could happen anywhere. So there are many things to consider.
A commonly cited quote for actuaries and stats people comes from famed statistician George Box: “All models are wrong, but some are useful.” Reality is very complex and impossible to predict with certainty. But we can make strides to approximate it with increasing accuracy.
At Lemonade we are trying to find a balance between classical actuarial models that have worked well for a while, while using new data sources and modeling methodologies available to us to push the envelope.
That makes a lot of sense. How does all of this play out in your day-to-day? Like, over the course of a ‘typical’ Wednesday, what would you find yourself spending your time on?
I’m only a few months in at Lemonade and I’m not sure I’ve quite settled into ‘typical’ yet, if such a thing exists here. Just a few of the things I’ve worked on so far, or will be working on soon, are building out our rating plan for the France launch; making improvements to our Netherlands rating plan; consulting with our data scientists on pricing strategy; devising pricing for an add-on to our Pet policy; and giving actuarial support to our National Expansion team to get rate filings approved.
Perhaps the most typically ‘actuarial’ thing I’ve been working on lately is refreshing our homeowners rate indications. A rate indication is a number that tells us how much we need to increase or decrease rates in order to be ‘rate adequate,’ meaning we are charging just enough to cover our expected claims costs and internal expenses, while providing a reasonable return which will allow us to support our continued growth. Having this number for all the states we write in helps our National Expansion team know where to target their resources for the coming year.
Sounds like a delicate balancing act—setting prices low enough so that we can be competitive, but high enough so that we can…you know…actually make money and thrive.
Indeed, and this is further complicated by the fact that Lemonade is a young company that has very high aspirations for growth.
Something I’m focused on is making sure we can meet those aspirations in a responsible way that will ensure we are around to pay claims 50 years from now.
However, pricing is only one aspect of the equation. We can also keep our claims costs down by being judicious about who we write policies for and under what conditions (that’s the job of underwriting), keeping fraudulent or inflated claims payments to a minimum, and managing our exposure so that we don’t get overly concentrated in any particular geographic area.
Smart decisions aided by Lemonade’s technological resources in these areas eventually percolates into pricing as we expect fewer claims in the future and can provide coverage at more affordable rates.
Not to get too in the weeds, but what’s the basic difference between an actuary and an underwriter, in the insurance industry?
Underwriters make the yes/no decision on individual policies, determining whether we choose to add them to our book of business or not. At a higher level, they steer the composition of our book—do we feel like we are overexposed in a certain area due to hurricane risk? Underwriters would be the ones to pump the brakes on adding more policies in that area.
Pricing actuaries help decide what the premium for each policy should be, given that it is accepted into the book by the Underwriting team. Policies that are more likely to have a claim should pay a higher premium to account for the increased risk. So both Underwriting and Actuarial involve the assessment of risk.
I visualize there being a line between the two functions: At some point, a policy is so risky that we don’t want to allow it on the books, even at a very high premium.
So, did you always think you were going to become an actuary?
I definitely did not always have this planned as a career path. I was not even aware what an actuary was until part of the way through college, and didn’t decide to pursue it until I had almost graduated.
I majored in Economics and had a strong interest in public policy research which culminated in me interning at a policy think tank one summer. While becoming an economist might have been an interesting and fulfilling path, I ultimately decided I didn’t want to commit the next several years to pursuing a PhD unless my heart was really in it.
So I revisited the actuarial path and quickly realized that I needed to start hitting the books: in order to become credentialed, student actuaries must take a long series of exams covering many topics such as probability and statistics, risk management, insurance pricing, reserving, and finance.
I actually found the challenge appealing… I would get to keep learning through books and papers, but also learn on the job. So I graduated and got studying. I was waiting tables and working as a bouncer at night, and studying for my first couple of exams during the day. I was able to start getting interviews with a couple of exams done, and the rest is history.
I had heard that the exams process for actuaries is famously grueling! Do most people who pass those exams go on to actually work as actuaries, or are there other places where they might commonly land?
The exams are a long process for sure. This winter I’m helping grade actuarial exams for the first time. You get assigned one question to grade… over and over and over. It does help keep certain topics sharp though, which is useful, and it’s been nice getting to network with other actuaries around the country. In normal times, they send us all to Las Vegas for a grading summit, but that obviously isn’t happening with COVID.
Most actuarial students I know of who go through the exams end up working as actuaries, at least for a time. Many may branch out and move into other parts of the insurance world, such as operations, underwriting, or research. More recently, many actuaries have branched into the data science realm.
And hey, life isn’t just about work! Tell me about some interests and passions you have beyond the job—whether or not they relate at all to actuarial science.
I read a lot, mostly fiction. I’ve been working on whittling down the unread books I have remaining on my shelf. I had a few weeks between jobs before I started at Lemonade and pretty much every day I would get an espresso and go read in Central Park. Those were nice days but I was very happy to start at Lemonade and get reacquainted with numbers.
I love trivia and watch Jeopardy almost nightly. It has been pretty surreal watching Alex Trebek’s final recorded episodes, knowing he passed already. I look forward to getting back into bar trivia in some pandemic-free future.
I’m pretty active and managed to run and bike a lot during the pandemic, which has helped keep me sane.
There’s too much TV right now and I have trouble getting into shows, so I don’t watch much. But recent series that I have seen are Nathan For You, Silicon Valley, The Marvelous Mrs. Maisel—and PEN15, which was just incredible.
Oh, and I have a greyhound named Yulia who I try to get to the park a few times a week. During off-leash hours we let her sprint at full speed, which is an amazing sight.