FARAH ISMAIL: Hi, everybody. I'm Farah. I'm based out of New York. I lead the commercial lines team at WTW. I've been in the industry for about 16 or 17 years. I started in actuarial, moved over to the data science side, and now and very much focused on the consulting and technology side. So I'm very excited to talk to you all about bureau rates today.
RAFAEL COSTA: Happy to have you here with us today, Farah. And Ben.
BEN WILLIAMS: Hi, everyone. My name is Ben Williams. As Rafael said, I'm a Director with WTW Insurance Consulting and Technology. I'm based in Chicago. I've got about 20 years experience in the insurance industry, and I'm a specialist in predictive analytics and applications.
RAFAEL COSTA: Great to have you here, Ben. So Ben, we'll start with you. Can you explain the primary role of rating bureaus in the commercial insurance industry?
BEN WILLIAMS: Sure. So rating bureaus and their close cousins advisory organizations, they are organizations that collect information from insurers and provide materials such as advisory rates or loss costs. And more broadly, rules, forms, product language, and also data products such as cubes showing losses by different segments, and so on, reports on industry trends and additional services.
And they're particularly widely used in commercial lines. Less so in personal lines because risks and coverages are more homogeneous than in commercial lines. So think, for example, of the breadth of possible commercial property risks versus a private home.
And some examples are ISO, AIS, NCCI, PCRB, DCRB. Throughout this podcast, there are situations in which I might say bureau rates. I might really mean loss costs as well. So just keep that in mind.
RAFAEL COSTA: And, Farah, what are the key benefits to startups or even large insurers of using bureau rates for commercial insurance?
FARAH ISMAIL: Sure. Great question. So bureau rates play a really important part in commercial insurance. Their benefits can vary depending on the size and maturity of the carrier. For startups, carriers often rely on bureau rates because they typically don't have enough proprietary homogeneous data to build robust rates on their own.
But bureau rates give them a really strong foundation for pricing since the rates themselves are based on large homogeneous data sets that have been collected and cleansed, aggregated, and analyzed by rating bureaus. So it's usually a really great starting point without having a need for any of external or extensive internal data.
For larger carriers though who do have their own data and much more sophisticated pricing models, they often layer proprietary models on top of bureau rates to fine-tune pricing and better align with their risk appetite and strategic goals. And so this approach tends to combine the robustness of the bureau data with the carrier's unique insights.
But even with the largest of carriers, they still often will use bureau rates as a benchmark even if they do have enough data to develop their own proprietary rating manuals and rating algorithms from scratch, just because so many carriers already rely on them. And so it helps them to ensure that they stay competitive in the market and provide a really great benchmark for market alignment and things like that.
RAFAEL COSTA: That makes sense. Well, with any benefit, there are usually also drawbacks. So Ben, what are the drawbacks of relying on bureau rates for commercial insurance pricing?
BEN WILLIAMS: Before I go into the drawbacks, a few other advantages I would add to what Farah said. We find that they are generally accepted. So underwriters will think of them as being right. So that helps to begin with.
They're often filed with regulators. So that makes it easier to get your own rates filed. Often there will be accelerators and other materials that exist for major policy admin systems that allow you to stand up pricing more quickly.
There are often updates in terms of form and coverage options for emerging risks like black mold and PFAS and so on. You've got to think of the alternatives as well. So if we're talking about a small insurer that doesn't have its own data, then the alternative is a me too. And by which I mean imitating a competitor's rates.
And using bureau rates is probably less work. And me too may not be that easy because of the transparency of the file rates and so on. Also, it reduces the need for analytical skills within an insurer. You can still stand up rates without actually having to have a strong analytics team.
But as you said in the question, there are advantages. And Farah already talked about some of these. So the bureau rates are based on a proxy for industry data. And that's not going to be a strong proxy for the industry everywhere.
So that means that those bureau rates may not reflect an individual insurer's business accurately because they'd be based on potentially low or biased market share and certain products or segments. There are situations in which the bureau rates may not be as up to date as they could be. So there could have been a change in loss experience in certain segments that isn't reflected in those rates.
Something else that needs to be kept in mind is that bureau rates, again, because they aim at universality, they need to be simplified to make them broadly applicable. And those simplifications may mean that they're missing something very important for certain insurers.
Another thing is, and I think Farah might touch on this again later, they're based on contributed data. And some insurers struggle to provide quality data. So if the data they're based on is questionable, then they're going to be questionable as well.
Other potential issues are there can be a real lag. So think of property deductibles. And some bureau rates I know of, they-- the deductibles weren't updated for a very long time. So there was a lag there. Also, if an insurer is using bureau rates and a lot of the other insurers in the same space are using the same rates, then there's not going to be much room for pricing differentiation. So you lose that potential for competitiveness.
And finally, bureau rates can be very expensive. And that varies from bureau and advisory organization from one to another. But there can be a reasonable cost associated with their use.
RAFAEL COSTA: Farah, do you have any additional thoughts on the drawbacks for insurers?
FARAH ISMAIL: I concur with everything that Ben had said already. I just wanted to reiterate that while bureau rates provide a really strong foundation, there are some potential drawbacks that insurers should consider. And so there's five key reasons why they should consider some of these things.
The first is limited customization. So bureau rates are often based on industry-wide data which might not fully reflect the individual carrier's unique risk appetite or underwriting philosophy or portfolio mix. And so there could be some misalignment with their individual strategic goals.
There does tend to be a bit of a lag in reflecting emerging risks because bureau rates, like Ben said, really rely on historical data in formal filing cycles. And so they don't necessarily capture new or evolving risks as quickly. And so that could just be like a consideration to keep in mind.
Oftentimes, carriers rely on bureau rates really, really heavily without any adjustments. And so it makes it harder for insurers to differentiate themselves in the market if they're all relying on the same baseline data set, especially for smaller carriers. And so that's just something to consider.
And then finally, it's regulatory complexity. So zero filings are often subject to regulatory approvals which can limit flexibility for carriers that want to move quickly or innovate in certain lines of business. And so it might be slightly outdated. And so we just generally recommend that folks definitely start with bureau rates, but try to supplement it with any proprietary data or models to stay competitive or responsive to market changes.
RAFAEL COSTA: Now, on that note, Farah, how can large insurers develop their own models to fine-tune pricing? So they would start from the bureau rates. There would be very little differentiation, as you were explaining. So the larger insurers might have their own data, might be able to fine-tune that to create differentiation. Do you have any examples on how they can develop that fine-tuning?
FARAH ISMAIL: Yeah. So large carriers often have access to extensive proprietary data which often allows them to build pricing models on top of those bureau rates. And so there's a couple of different ways that they can do that.
One is loss experience analysis. So they can analyze their own historical claims data and identify patterns in frequency and severity. And you could use that to adjust the bureau rates to reflect your own unique loss experience, which might be a little bit different than an industry average.
The other way that you could do that is segmentation or risk differentiation. And so large carriers are often able to develop these granular segmentation models that look at factors like industry class or geography or risk characteristics. And so you're able to basically apply predictive analytics to distinguish between two businesses that might be in the same class but with slightly different safety practices and things like that.
The third thing that I'll talk about today is just using more advanced techniques, like machine learning, to predict future losses based on a much wider range of variables beyond what bureau rates typically consider. And so you can use third-party external data sources, like economic indicators or weather patterns, to refine your pricing and provide more dynamic pricing strategies and things like that.
RAFAEL COSTA: Now, Ben, do you have any thoughts on this differentiation and examples from large insurers?
BEN WILLIAMS: Sure. Well, I was actually going to say there are a couple of approaches to using predictive analytics to develop pricing models. And one would be the ground up approach. Farah talked about that a little.
So modeling either frequency and severity or modeling loss costs directly. And that has the advantage that the models are going to reflect the individual insurers' data. In doing this, you're setting aside any bureau rates and just building your own from scratch.
And so this is likely, again, so it's going to reflect your own data, your own risks. It's likely going to be more transparent. They can be transparent, but others can be rather opaque. They will likely look very different from competitor rates. So you could have some interesting competitor dynamics in terms of how you stack up versus competitors in terms of price and how that varies by segment.
And in order to understand that, so if the insurer has access to bureau rates, and they can use those to understand how their competitive position is going to play out. Now, I mentioned that an advantage of modeling ground up is that the model will reflect your own data. That's also a downside. Because no insurer has infinite data, and there will be segments where they're going to be more or less thin, where they can't get good information from the model.
The other approach is effectively loss ratio modeling. So using-- and Farah mentioned this. So using current bureau rates or loss costs as the starting point and modeling deviations from those. So this has some good advantages. So it's conceptually appealing because you're using bureau rates except where your data says that you shouldn't. Or where the bureau rates are high or low, and where there's enough data to tell you that that difference is significant.
I already mentioned that underwriters tend to like bureau rates because they're based on industry data. So it's going to-- something you can explain to an underwriter. Say, well, our data says that in this segment, we should be charging more or less. And you can explain why.
It can be easier to manage dislocation if you've been using bureau rates in the past because that's your starting point, and you'll be modifying just a handful of factors potentially. There are some downsides as well. Those bureau rates and loss costs are going to vary between states. So if you're modeling loss ratios, it can be hard to understand, to separate signal from noise.
Also, understanding the results can be harder. So if you're modeling frequency or severity, you can often relate the patterns you're seeing in the models back to the underlying business. But if you're modeling a loss ratio, that can be harder to interpret.
A couple of mechanical issues. In order to take this approach, insurers need to be able to rewrite all their business on their most recent bureau rates. And some carriers struggle with this. They just don't have the machinery in place that allows them to perform those regular calculations. And of course, they need to have expertise with data and model if this is something new to them, and they might not have that experience.
Bureau rates continue to be widely used, and the larger insurers do tend to be the ones attempting to model deviations. And there are still a minority of insurers who are modeling ground up.
RAFAEL COSTA: Well, you both were explaining that there is a strong reliance on bureau rates broadly by the insurance industry. Now, how do the rating bureaus themselves ensure that the rates, forms, and the data products remain relevant and remain up to date? Farah, what are your thoughts on this?
FARAH ISMAIL: Yeah. So bureau rates have a very structured process to make sure that they keep their rates and forms as current and relevant as possible. And so they are continuously gathering lost cost data from carriers on an ongoing basis. And they are trying to cleanse and aggregate and analyze that data to identify trends in claims severity and frequency and exposure and things like that on a continuous basis.
They also tend to be very methodical in updating their rates and forms through formal filings with regulators that are constantly updated to reflect changes in industry experience and inflationary trends and evolving risk factors and things like that.
And they're continuously working with carriers and industry experts, like the actual insurance carriers, the actuaries and other stakeholders, to make sure that their products are aligning with real world underwriting practices and market conditions and things like that, as well as tracking regulatory or legislative changes and court decisions and economic shifts and things like that that could impact risk profiles.
And taking all of that information and incorporating it into their filings and forms and things like that on an ongoing basis. And so it is a very robust data analysis and continuous feedback loop to make sure that they are collaborating with the industry, with regulation, and with legislation to make sure that they're as accurate and useful as possible.
RAFAEL COSTA: Now, to wrap us up, Ben, how should carriers be using bureau rates to add value?
BEN WILLIAMS: The answer really varies with the insurer. So if we're talking about a startup or a smaller insurer that has no or very little data, potentially doesn't have an analytics team, then not a lot of option beyond using bureau rates, and they will do a decent job.
I'd say the larger and more established the insurer gets in terms of business volumes, having the relevant teams to be able to perform analytics and do the right filings and implement deviations from ISO, the more they should be exploring those deviations and at least understanding where their own experience differs from rate bureaus.
And where their experience is differing, they should be exploring ways to reflect that in the rates that they are deploying. But we do see bureau rates rules, forms, all the other products continue to be widely used. And that suggests that the advantages they offer, the benefits outweigh any of the disadvantages that we've mentioned.
But we consider it to be an ongoing evolution. We do see more and more exploration of deviations, especially among the larger insurers and some insurers filing their own ground updates as well.
RAFAEL COSTA: Now, to wrap us up, Farah, what are your thoughts on this?
FARAH ISMAIL: Bureau rates are most valuable when carriers treat them as a foundation and not necessarily the final answer. And so as a general rule, my recommendation tends to be to use bureau rates to provide that reliable starting point, because they do have that large homogeneous data set that's been cleansed and analyzed by rating bureaus that are keeping up with industry practices and industry trends and things like that.
And using bureau rates as that baseline really helps carriers avoid gaps when they don't have enough proprietary data in-house. But in general, I would generally recommend that carriers apply their own experience, risk appetite, and strategic pricing models on top of bureau rates when they can to better reflect their own specific loss experience and segmentation and expense loads and competitive positioning in the market to make sure that everything is as close to their proprietary underwriting guidelines as possible.
But even the largest carriers that are able to build their own proprietary rates do tend to use bureau rates at minimum as a benchmark to make sure that they remain competitive and aligned with industry norms and things like that.
The one thing I did want to call out as an aside is that there are tools and technologies that really allow carriers to automatically ingest bureau rates into their systems that does make it a lot easier to layer in those proprietary adjustments and make sure that you're still consistent across your different products and portfolios and are being as efficient and accurate and compliant as possible.
RAFAEL COSTA: Well, based on what you both said, I think we can safely say that the bureau rates do add value. That was the question that we posed in the beginning of this podcast. But of course, you have to use it cautiously. And you have to use it-- have to use it correctly.
And that brings us to the end of today's discussion on bureau rates. Thank you, Ben.
BEN WILLIAMS: Thank you, Rafael.
RAFAEL COSTA: And thank you, Farah.
FARAH ISMAIL: Thank you. I really appreciate the time.
RAFAEL COSTA: And thanks, everyone, for listening. If you enjoyed this episode, don't forget to subscribe and join us next time on (Re)thinking Insurance.
SPEAKER: Thank you for joining us for this WTW podcast featuring the latest perspectives on the intersection of people, capital, and risk. For more information, visit the Insights section of wtwco.com. This podcast is for general discussion and/or information only. It's not intended to be relied upon. And action based on or in connection with anything contained herein should not be taken without first obtaining specific advice from a suitably qualified professional.
