For a recent Talking Trends webinar, we invited a panel of Willis Towers Watson experts to discuss some of the hot topics that North American property & casualty (P&C) insurers will face in the next few years.
Alongside the regular challenges of running a business, dealing with catastrophes and serving customers, the insurance industry has had a number of unique challenges in 2020 and 2021. These have included working remotely, planning return(s) to work, regulatory challenges, the effects of social inflation on claims, and the unanticipated changes in loss and premium patterns.
So, what do these macro socioeconomic and industry trends mean for how insurers better engage existing and new customers, become more effective in how they make and implement decisions, and ultimately unlock new sources of revenue and profitability? That was the starting point for questions posed to our panel by Americas P&C Sales and Practice Leader Katey Walker.
The panel:
Katey Walker: Efficiency and effectiveness have been big topics for insurers in 2021. Given that our expectation is that the focus on these areas will continue, what impacts do you foresee for core functional areas?
Charlie Carwin: If you look at the past trend in capital modeling, everyone got excited with all the possibilities and what they could do with models. To take the analogy of a car, companies started building their own engines, adding their own transmission and own braking systems, and perhaps deciding if they wanted air conditioning or not.
The end result was large and expensive to maintain models that were producing value — but not necessarily in line with the costs associated with them.
So, a key trend is simplification: how insurers are approaching quantification and the modeling of risk. Companies are more and more interested in off-the-shelf models, the buy-versus-build-type situation. To go back to the car analogy, they're not interested in building their car; they're more interested in driving it and extracting value that way.
Associated with that is simplification on the IT side. “The great resignation” has been an influential factor in companies’ interest in cloud solutions or other solutions that are more scalable and that allow them to reduce the reliance on owned hardware and IT teams.
Michael McPhail: Picking up on what Charlie said about off-the-shelf solutions, that’s true across the entire insurance process.
Whether it’s a policy or claims administration system, many carriers recognize the benefits of simplifying their processes. A big issue often is that they have all these old legacy systems. Trying to move completely off the legacy system can be hard. Starting afresh is a big challenge but might be justified by the benefits in some cases; however, the use of application programming interfaces (APIs) in off-the-shelf software is making the task of connecting legacy and new systems much more straightforward.
Katey Walker: Beyond simplification, the other big word we hear in tackling efficiency is automation. How is that changing the landscape?
Scott Gibson: I’d describe it as not so much changing the landscape but enabling it. So, if we go back to Charlie’s car analogy, for example, we need some sort of road — a layer that’s going to provide the ability to use the insights of a model in multiple ways.
That’s where automation comes in, particularly in light of the proliferation of different tools and data sources and the push from regulators for transparency in understanding how information is being used to affect the customer.
Automation can play a key role because, when you think of compliance risk, for example, it's not a value-added operation. But it needs to be done, and it needs to be done accurately and to a deadline. Automation is a great opportunity to streamline and provide efficiencies.
In addition, the bots can do the repetitive tasks so that users can focus on extracting the value out of those processes. And that takes us back to simplification, where instead of having long-drawn-out, complex and highly manual processes, automation cleans up the edges so that carriers can focus on the sources of added business value.
Katey Walker: Let’s extend these concepts of simplification and automation. How do these change what actuaries and we, as consultants, spend our time doing?
Yi Jing: From a reserving standpoint, there are more uncertainties than ever. I think companies are really looking for actuaries to provide insights about how to deal with these uncertainties and, more so, how to quantify them.
So, for example, for COVID-19, how do we measure the direct COVID-19 claims? How do we assess or reflect the indirect COVID-19 impact? And in October, we saw inflation hit a 30-year high, so how do we account for future inflation and our reserves if that's not in the historical data?
There is certainly a need to go beyond some traditional methodologies and explore new methods or tools to help address some of these challenges and questions both in the shorter and longer term.
Katey Walker: While we’re talking about insight and circling back to the idea of driving value, all of you have worked for insurers, so what are your perspectives on ways for insurers to capitalize on both?
Scott Gibson: One aspect was our project list kept growing and growing, and we could never knock enough things off of it to whittle it down. That's kind of how I moved into the business process excellence space. And I think that challenge is becoming even more dramatic because the pace of change is accelerating in the industry; the management questions are getting larger and larger, including now, for example, around climate risk.
The best part of my role now is I get to ask if a company could refine a process and smooth it over; if eight days became five days, what is the opportunity to create more value for the business?
Michael McPhail: Senior insurance executives want to know the answers to “why” questions. “Tell me why trends are happening this way. Tell me why there is adverse claims development,” and so on. These historically have been hard questions to answer.
You can do some wonderful analysis, but if you can't answer that question, it really doesn't matter. And this is where I think predictive analytics can be really useful because, oftentimes, when someone can't answer that question, they start scrambling and point to some sort of anecdotal evidence that becomes a story that may or may not be true.
With data science becoming more common, and senior executives understanding analytics better and better, they're looking for more robust answers. So, I think combining predictive analytics with communicating results clearly is a significant opportunity to add value.
Katey Walker: But can predictive analytics be overdone?
Michael McPhail: Yes, but don’t get me wrong, predictive analytics are a wonderful thing and companies benefit hugely from using them, in my opinion. The danger is becoming fixated on things being perfect — which takes way too long. I had a coworker who once told me that you can get a 95% solution at 20% of the time and cost of a perfect one, and I’ve remembered that. Models aren’t meant to be perfect; they’re meant to be useful.
Katey Walker: Pricing is always a hot topic, so could we turn for a moment specifically to the opportunities you see in this area. What are some of the key aspects for potential pricing improvements in your view?
Scott Gibson: I come at this from the perspective of how to optimize what is a big workflow for insurers.
With rate indications, for example, there tend to be discrete points in time when insurers are considering the indication information for a particular state and line of business. What would be helpful is to move to a more continuous flow of information, so that the data from underwriting, claims, reserving or wherever drives when those changes are required and which pieces of information are most important to the indication. It could also automatically bring in maybe new, relevant external and internal trend information to drive a faster and more efficient pricing process.
Michael McPhail: One area is just quickening decision making. A big part of that is looking for broad strokes rather than trying to be overly precise. If I think about a rate indication, there are a lot of components that an insurer could spend a whole lot of time on, such as picking loss trends. We can use automation to make that faster while maintaining accuracy.
Another opportunity is to be able to better incorporate information from the claims and underwriting teams on a real-time basis. If underwriters are continually making risk-based adjustments, let's say on schedule rating outside the filed rating algorithm, how can that information get back into the pricing process to potentially improve the pricing models? This is far from an isolated example, as insurers have tended to do the pricing process in silos with limited communication and sharing of information between teams.
And finally, one area that I think is more directly related to improving the actual pricing process itself is considering the impacts of pricing changes on a portfolio over many periods, rather than just the filing period. Companies often make short-term decisions based on annual profitability targets but don’t necessarily factor in that a rate increase will likely have an impact on customer behavior and could lead to premium leakage. Predictive analytics can definitely help improve portfolio management by thinking about how customer behavior is impacted over many periods.
Katey Walker: Any thoughts on how simplification, automation and analytics can enhance the reserving process?
Yi Jing: When companies are doing quarterly — or even monthly — reporting, actuarial departments can be very stretched and are faced with the long project lists that Scott mentioned earlier.
So, the opportunity, as we see it, is to better use people's time by potentially automating tasks such as selection of loss development factors. Actuaries can then spend more time doing deep-dive analysis to reveal some of the drivers of the results being seen. This might include implementing diagnostic tools to pick up trends, assess those trends and find anomalies.
Another area attracting investment in our experience is better management reporting, so that senior managers have a better understanding of what the numbers mean through the use of automated executive dashboards and suchlike.
Michael McPhail: I’d just add a comment on the overlaps in the use of predictive analytics in giving indicators for action on pricing and reserving. For example, we’ve started working with insurers to use machine learning to understand the drivers of prior year development that can be verified with claims teams for explanation to senior management.
Katey Walker: Related to management insight, visualization has been a buzzword in recent years. So where does that fit into the future role of sharing results?
Michael McPhail: No question, it’s huge. My opinion is that it should be simple to understand and see what’s trying to be communicated, e.g., the drivers of a result. If it takes five minutes to explain the visualization, that’s self-defeating.
Yi Jing: I’d echo that. The clearer you can make the information jump out of a graph related to, say, adverse development in various lines of business and frequency severity trends, the better it is.
Katey Walker: A lot of what we’ve been talking about comes back to transformation. Each thing is changing what insurers do and how they do it. So where do companies start if they consider transformation as perhaps too great a challenge?
Scott Gibson: Absolutely. So in a lot of cases, the best bet is start small; look for easy wins or changes to a process. Automation is often a good solution because it doesn’t involve going back and fixing a process, not to mention it can build some momentum for further transformational activities. Later, it’s always possible to then go back and review the process itself.
Charlie Carwin: I feel there are pretty established best practices in capital modeling, but what is changing is that senior managers actually need to be able to see the results — and in a timely manner.
So when I think about visualization, for example, I think of it as just making sure that there's the marginal impact of capital from a material decision on every report. So it shouldn't be determining the decision; it shouldn't be this is what's making the decision by any means. But it should be available and should be impacted.
But for that to actually happen, you have to have capital models that are responsive. It can't take two days to run it, or you can't have a process where, for example, it takes three weeks to get an answer to a question such as: “What is the capital impact of changing the reinsurance program?”
And so, if I think about the components of transformation, one is the out-of-the-box model, as I mentioned before. Then you have automation, where some of the more manual processes are no longer necessary, where you can actually have the robots do that work and make everything much more efficient and get to the answers more quickly.
The what-if questions are always going to exist with, for example, questions about the impact of inflation and changing operational risk coming to the fore in the last couple of years. Where capital is really coming into play is that the models have to be ready to answer those questions, or at least quantify the marginal impact of capital. That's a big difference compared with even five years ago.
Katey Walker: So going back to the buy-versus-build discussion, where do you think we’re going from here?
Charlie Carwin: Historically, actuaries have liked to program; I like to program. But insurers can get a lot more value out of actually using the models than creating the models. This applies equally to larger companies that are looking for ways to reduce some of the risks, including key man risk; reduce maintenance costs; and have ready-made scalability. So, it's definitely driving buy versus build.
Michael McPhail: Away from the capital modeling side, similar arguments can be applied to open-source platforms, which have become quite popular in areas like pricing.
The problem is when someone leaves suddenly there's no one around to pick up the pieces on the coding and the governance chain is broken. That's part of why people are moving to more off-the-shelf solutions because they are starting to recognize some of the pitfalls.
Katey Walker: Does stability present the same transformational issues in reserving?
Yi Jing: Yes, I think a lot of companies are investing in a better tool to do their reserving, and I think also tightening documentation of processes for others to follow.
Let’s not forget that, particularly recently, people turnover across the industry has been high. With the pandemic, people may feel changing jobs is a little easier given that you don't have to relocate. Retaining talent is part of maintaining that stability, and insurers can engage people with more challenging, interesting and exciting work.
To that end, let the robots be robots, taking over some of the tedious and repetitive work. I think that’s important to bringing in young talent to the insurance industry.
Katey Walker: Obviously we’ve been living through the pandemic. How will it challenge companies’ efficiency and expense structures?
Yi Jing: As people have worked from home extensively, this has obviously posed efficiency challenges, but I’d also bring this back to my point about engaging people remotely with satisfying work.
Charlie Carwin: I’d point to the need for a different focus on different risks: deflations, whether the asset mix is correct and how premium volumes are changing — amongst other things.
Beyond these, I go back to the term I used earlier — “the great resignation.” I think we're seeing that continuing because of things like key man risk, joining up complex processes remotely, and the benefits of flexibility and agility that the pandemic has highlighted.
Linked to the heat companies are seeing in the job market, I think there’s a general push for trying to make sure that actuaries are doing the type of work that actuaries are fit to do, not necessarily the work that actuaries have to do. Automation and making processes more efficient and more focused on extracting the value can clearly contribute to that.
Scott Gibson: Echoing those comments, I’d add that we've actually seen that the companies that have already adopted automation were more resilient after the pandemic hit.
Another interesting thing is how the pandemic has affected the work/life balance. People have been home and, because they're home, they can go run an errand more easily than if they were in an office environment. So, it's creating a sort of an asynchronous working environment that is exacerbated by time differences. You can't necessarily knock on the cubicle next to you and expect the person to be able to tackle a task immediately. Robots can fill that void because they're going to work 24/7. They don't get tired.
But it’s not a case of replacing people. The majority of companies, when we’ve polled them, are actually looking to augment human productivity, not replace it.
So the pandemic has definitely been a potential accelerator for automation and transformation projects, helped by the fact that travel and other expenses have fallen significantly.
Katey Walker: A perpetual hot topic is M&A. Do we see a shift in how companies are approaching buying and selling?
Yi Jing: At the beginning of the pandemic, we thought M&A activity would be low given all the uncertainties. It turned out it was, and remains, a hot market. I think that says the industry still has excess capital and sees challenges as opportunities.
Features of the market include a strong interest in specialty business. Moreover, we’re seeing many companies are putting a lot more focus on underwriting strength. The question they want to get answered is: “Is the business as good as the seller is telling them?” Another facet of transactions is companies looking for adding network platforms, such as the Liberty's recent acquisition of the State Auto.
Going forward, I think having an efficient due diligence process is very important. Clients are certainly looking for us to provide better and quicker insights.
Katey Walker: So, to wrap up, I’d like to ask each of you for your take on the next big thing in the coming three to five years.
Charlie Carwin: It’s already happening, but transition to the cloud. I see that growing over the next five years along with automation of the less fun, more repetitive aspects of back-office insurance work.
Scott Gibson: I’d say it’s automation plus. Right now, many companies are automating a specific process or part of a process but have seen the benefits and want to spread their automation wings further.
Michael McPhail: Going first is an advantage on this question, so I agree that simplification and automation are where things are heading.
Yi Jing: I think about what I would like, and that’s knowing, for example, that come quarter- or month-end, my full complement of reserving data will be available the next day — or perhaps more realistically, won’t take weeks or months to produce.
What we believe the webinar discussion illustrates is that insurers’ journey to the future and being best in class does require a deep understanding of all the critical elements involved from strategic to tactical, combined with the technology to make it happen.
If you would like to discuss any of the issues or comments with us, please do contact us.
