Okay, we're live. Welcome back to the Bank of America U.S. Financial Services Conference. This is sort of what I would say is the insurance sleeve. We have an opportunity here to talk to Selective Insurance with CEO and Chairman John Marchioni and CFO Patrick Brennan. So we're really pleased to have them here. And if you're in the audience, you can ask a question at any time. I don't have an order to how I want to conduct things, so feel free and comfortable. That's fine. And let's see what we can learn. Thank you for being here, gentlemen.
Thank you for having us.
Let's talk about the last couple of Selective. Let me just lay it out. Selective has had an unbelievable history. The fact that there is any sort of volatility is actually a bit of surprise to what's been a much longer period of great stability and comfort from investors. How do we parse the last couple of years and with the reserving actions? Where do we stand right now in terms of footing?
Yeah, I would say, first of all, I appreciate the comments because we are proud of that history. And the long-term history will celebrate our 100th anniversary as a company this year in 2026. We're proud of that. But we're also proud of the more recent track record over the last decade to decade and a half of very strong growth and very consistent returns, and averaging a little over a 12% ROE across a decade, decade plus, is something we're very proud of and proud of being that consistent growth and profitability story. Now, to your point, the last two years have been more challenged in terms of some of the reserving actions we've taken on the very recent accident years. But I would say the overall objective is to maintain that long-term consistency and track record.
Because when it comes to casualty lines of business, and in 2024 it was predominantly our General Liability line, and then 2025 was Commercial Auto Liability, those are longer-tailed lines. And the ability to be a consistent performer over the long term requires that you react quickly when you see emerging trends. And I think the whole idea of lawsuit abuse, social inflation is pretty common in the industry. I don't think anybody is suggesting it doesn't exist. I think, generally speaking, the commentary relative to severity trends in those commercial casualty lines is high and has been increasing. And we've been reacting to that very quickly. And as a result of that, we've been taking action on more recent accident years based on relatively immature data on both a paid and a incurred basis.
We think that's the appropriate posture to adopt when you have a trend environment like we have right now that doesn't have a lot of history that you could look at to predict what it's going to look like going forward. We're trying to make sure that we stay on top of it, recognize that emergence, make sure the most recent years are booked where they should be. That ensures that our run rate profitability going forward is sound. I think when you look now a year later, 2024 is when we took this significant action on General Liability line. That was predominantly, almost entirely the post-pandemic years. More importantly, if you recall, we also meaningfully increased the 2024-year in the 2024 calendar year. That put us on very solid footing.
You saw relative stability in General Liability line throughout 2025. We think that says that that was the right approach for us to take. Then if you look at what happened in 2025, it was similar, maybe a little bit smaller magnitude Commercial Auto, where we didn't just respond to the most recent years. We also increased the current year to make sure that our run rate profitability is sound.
Is the goal to get the numbers right within a confidence interval of X or to be somewhat ahead of trend? What does a good, solid footing actually mean for the state of the reserves? What's the goal, I guess?
The goal is to have stability in reserves. And while we had a history of, call it, 15+ years of favorable emergence, that wasn't planned. That's how ultimate frequencies and severities played out. We are always planning to book the right amount. And we want to make sure that's the case. Because our process is all interconnected: reserving, fees expected, loss ratios, expected loss ratios, feed pricing indications. Pricing indications drive your pricing strategy and your risk selection strategy. And to the extent you fall behind or to the extent you're too conservative in your booking actions, it could create challenges downstream that you want to avoid. So our overall intent is to always make sure we book our best estimate every quarter, make sure that is properly fed into our planning process, and then our booked loss ratios are accurate.
So in all of this, I would say that Patrick showed up just in time for all this fun. So I guess through about 15 months or I don't know exactly, maybe we'll call it a year plus. Tell us about the past year, what you've learned, what's been changed, what you came with, what you're surprised to have at your disposal.
Yeah, it's a great question. So I think one of the things that we talk about at Selective is having a long-term view. John talked about our ROE performance over time. Even this past year, we wrote at a 14.2-point operating ROE. So certainly feel good about those results in the context of a long-term track record of producing north of 12 points of ROE. I think coming into this role, I had maybe lower expectations in terms of how much data we might have relative to where I was coming from, where it feels like I was sort of awash in data. And I got to tell you, I am really pleasantly surprised. I think we punch above our weight. We have a ton of data available to us.
I think the fun and exciting part for me is to figure out how to help the team think about unlocking even more value from that. Because I have a different perspective, and others do as well. I think there's tons of opportunity there. But even at that, we've been utilizing the data in ways that help us identify what's driving loss cost, and then what do we do about it? Where do we apply pricing when we need to apply pricing in a way that's granular, that's targeted, that's going to drive outcomes that we are seeking? I feel really good about that as well, the way that the data is being used. I think as we go forward, from an actuarial perspective, from a pricing perspective, we want to make sure that we have precision pricing. You can never be perfect at pricing.
There's always ways to get better. I think that continuous improvement mindset is something that has existed prior to my showing up. It's part of every conversation that we have on a go-forward basis. When you look at the operations efforts that we have, we want to get better information in the hands of decision makers at the time that they make decisions, leveraging tools and technology, continuing the investments that we've made in the past to really amplify their ability to execute against that. Then culturally, I've already touched on the continuous improvement mindset. John talked about it on our most recent conference call, this relentless focus on the fundamentals. Make sure we're doing all the little things right. We are always looking for new ways to measure and check those boxes to ensure that we're doing that.
And so I think in sum, there's tremendous opportunity in front of us. We have the assets. We have the tools. While we've had some challenges in the last, call it, two years, we've also really positioned ourselves to take advantage of where we're going on a go-forward basis by making investments. For example, in our technology budget, if you look at what we spent in 2023 on discretionary projects and IT, we've more than doubled that in our budget for this year to give a sense of putting our money where our mouth is and really driving towards enhancing those capabilities that we have.
So it's been a very unusual five or six years, from the pandemic to the reopening through inflation, the courts being closed, the courts being open, whatnot. And a lot of what's driven the actions over the past couple of years has been paid claims emergence in immature years. There might be some false signaling there, or maybe there's not. I mean, you're going to see the data come through, and you're going to take action. Is there something that we're learning about the data as it continues to evolve? Are these real signs of a real persistent loss trend that is rising? Or is just the way claims are emerging different from the way it's been in the past?
Yeah, I would say there's clearly a different pattern. And I think it's gone on for long enough now, post-pandemic, that I think we could call this an elevated trend driven by different claims patterns than we've seen before. And it's not just the most recent years. But when you look at where average severities are at this point, in our mind, and we're very transparent in our trend assumptions, we have all-in severity trends on the casualty side at about 9%, and excluding workers' comp, closer to 10%. And I think we've got higher conviction as time has passed over the last three accident years that those are real. We see those in our paid and our incurred dollars. I think the question is, at what point do we inflect?
At what point do we reach a more normalized run rate where the year-over-year increase doesn't continue to be the same as it has been? It's interesting to me, and we've cited this before, and this has not been the case for us. But in 2024, when we booked the significant GL reserve adjustment, that was all in the post-pandemic years. The industry, and we'll see what 2025 looks like. But I wouldn't be surprised if it's similar to what 2024 looked like when Schedule Ps all come out. In 2024, the industry in total added $10.5 billion to GL reserves. Almost half of that was from the pre-pandemic period. Our pre-pandemic years held up well.
I think what happened was, throughout the pandemic, with this dramatic drop-off in frequency, there was a false sense of security that that higher severity in those pandemic years was just anomalistic, and it was tied to that really low frequency. The reality was it was settling into a new run rate. Now, we've cited paid data in the more recent accident years as a driver. I would also suggest incurred data is telling you the same thing. This is why single data points on the reserving side, you never want to overly react to one single data point. Because when you see paid data emerge like that in very immature accident years, the leverage effect on those when you project those years to ultimate is significant when you just look at historical development patterns. You're also looking at incurred data.
But your incurred data, you have to evaluate whether or not you think case reserves, which are a big driver of incurred losses, whether case reserves are stronger or weaker or the same as historical patterns would have told you. That's important. You have to understand whether or not there's a shift in your disposal rates. Because if you ignore disposal rates, especially when disposal rates are declining, you're going to look at paid data, and you're going to look at incurred data. And it's going to give you a false sense of how these years are actually going to develop to ultimate. You have to look at litigation rates. So all of those factors alongside of each other, I think we put together, so it's not just the paid claims data in those more recent years.
It's all of those factors that we would suggest gives us confidence that the trend assumptions we are incorporating are sound and are very reflective of the most recent years. Now, an additional point, and we may mention this on the third quarter earnings call, is rather than just rely on our own historical practices and our own evaluations, and we've always had a Big Four accounting firm do their own review of our reserves twice a year. That's been historical practice, continues to be. But we brought in some other independent firms to review all of our actual reserving practices, our planning process that leads to expected loss ratio selections, and a review of our claims performance on both a closed and an open claims basis. And I think that sort of reiterated to us a couple of things.
One of which was the positive was that third-party's review of reserves, our carried reserves, were above their central estimate. So that's another important data point. But I think it also validated for us because we're talking about three external firms now that have a broad lens into the market is they validated two things. Number one, the trends we're seeing on the severity side are very similar to what they're seeing across the industry. And number two, our practice reacts much more quickly or puts more weight on the most recent experience. So within the actuarial practice, you could look at longer-term averages or shorter-term averages, and they're going to give you very different answers.
If you look in the current environment, a seven-year average, weighted average, or a seven-year progressive average is going to give you a very different answer than putting all of the weight on the most three most recent accident years. So our process and our approach has put more weight on those recent years. And we think because of the change since the pandemic, that's the right place to be.
Well, you said you don't want to focus too much on a single data point. I want to focus on a single data point. State called New Jersey. And so how much of all this information is New Jersey specific? First of all, how much is it driven what's happened to you? Two, how much is the experience in New Jersey non-reflective of the rest of the country? And how much of it is?
Yeah, so let me make a couple of comments. I'm sure Patrick will want to weigh in. Let me take this in two dimensions. First of all, for personal auto, for us, it's a smaller book. New Jersey is a bigger portion. It's 30%. All of the reserve adjustments we've taken in 2025 relate to New Jersey. Our performance outside of New Jersey on the Personal Line side is much more reflective of our target experience there. That's a small portion of the portfolio. In Commercial Auto, which is where we really focused our attention relative to New Jersey, New Jersey represents about 15% of our Commercial Auto premium and vehicles, and therefore roughly a similar amount in terms of loss dollars. There are a number of states that have higher susceptibility to social inflation because of the environment, judicial environment, legislative environment, regulatory environment.
New Jersey is one of them. So there was a RAND study done in 2024 on social inflation, evidence of social inflation. It identified seven states as those that had the highest increase in jury awards, trial values, trial awards. Half of them are in our footprint, half of them aren't. So New Jersey is one of them. New York is one of them. Pennsylvania is one of them. I think Illinois might be the fourth. But then the three that aren't are California, Texas, and Florida. So you have this group of states, some of which we're exposed to, some of which we aren't, that have had higher impact of social inflation because of their environments. But what everybody has seen and we've seen is this is widespread. It's just exacerbated in certain places like New Jersey.
I would say that we singled some of this out over the last four or so years. The New Jersey legislature has enacted a lot more pro-friendly plaintiffs' bar changes to statute that individually don't mean much, but in the aggregate have made it much more fertile ground for the trial bar. So for instance, minimum limits for personal auto increased, minimum limits for vehicles over 26,000 lb, not a big portion of our book, less than 10%, but required to carry $1.5 million of limits, pre-suit disclosure of limits, pre-suit disclosure, lowered the bad faith standard for uninsured motorist claims. So as a result of this and by the way, New Jersey always had a much higher litigation rate than the rest of the country in Commercial Auto.
So you have this confluence of statutory changes that have made it more fertile ground and I think resulted in a higher level of activity there. But I want to reiterate this point. The severity trends we're seeing are evident across the country. And in fact, the litigation rates we see across the country Commercial Auto are actually increasing. Whereas in New Jersey, albeit much higher, it's been relatively stable from a lit rate perspective.
Yeah, and on the litigation rate, our litigation rates in New Jersey tend to develop to an ultimate rate that's about twice countrywide for Commercial Auto, just to give a sense. And when you look at industry ALAE as an example, you can see that the ALA E for Commercial Auto tends to be higher in New Jersey than in other states. And in fact, in the last two or three years, has been on an uptrend for the industry from a Commercial Auto perspective, but also from another liability occurrence perspective. So there is evidence that there are environmental differences there that we're seeing and that some of those trends are manifesting across the industry and not just with us.
Markets are fickle. They react to the company that does things first is often the one that's most penalized. Do you believe that the way you've reserved for things are going to demonstrate that you were ahead of the curve and that you have reasonably, by pricing actions seen by your competitors or things that have not emerged, that you have evidence that what people are arguing is a selective issue, will be able in 24 months to be diagnosed, that you were just ahead of a much broader trend?
Yeah, I'll stop short of predicting other companies' performance. What I will say is I think our track record suggests that we get to the ultimate for an Accident Year quicker than the industry broadly. We've got slides in our investor deck that show that General Liability. I expect that when the ultimate story is written on these more recent Accident Years, that that continues to hold. Now, listen, if you stack up our General Liability and Commercial Auto Liability relative to our peer group, I think you see fairly similar results. I think what we could be criticized of fairly is we are overweight the lines that are most impacted by social inflation. We have a bigger portion of our premium Commercial Auto Liability and General Liability.
The reality is, if you look over the course of the last two years, the majority of the margins in the industry have come from Personal Lines, commercial property, workers' compensation prior year development, favorable development, and then some of the specialty lines, which went through a significant correction, professional liability, cyber, some of those. The margins on GL and Commercial Auto Liability have not been strong in the industry. We just happen to be overweight. I think it's appropriate for us to make sure that that internal subsidization that's happening for everybody in the industry with property results outweighing some of the pressure in these casualty lines, we want to make sure that we're addressing that and staying ahead of that curve.
So if we look out to guidance, I actually am not a big fan of guidance. I think it causes people to work to guidance and not work to numbers sometimes. But you have a long-term sort of 95% combined ratio. Rule of thumb, I think, is better than guidance, maybe. And you're running ahead of that right now. That means you might need more price or whatnot. What is the sort of near-term plan to get the company to the margins that you want to be achieving?
Yeah, I think if you look at our underlying or ex-cat combined ratio guidance and compare it to the underlying accident year in actual underlying accident year in 2025, you'll see loss ratio improvement of about 120 basis points. So we were at 91.8 underlying accident year combined ratio. So ex-cat, ex-development in 2025, our guidance is a 90.5-91.5 ex-cat. So let's just take the midpoint to make it easy. There's about 80 basis points of improvement to the midpoint. But we also mentioned that the expense ratio is up about a half a point, 40 to 50 basis points. So you can see the loss ratio underneath that is down about 120 to 130 basis points. When you look at overall and again, we guide across all of our, we don't do individual underwriting segment combined ratios. We guide to an overall combined ratio.
So I want to just focus on overall pricing and loss trends. The last two years, our written rate all in is about 9.5%. Our trend assumption has been between 7%-7.5%. So you got about 200 basis points of rate over trend. You apply that to the loss ratio, call it 63% of that. And it gets you to right around that in terms of improvement. And that's in our 2026 guidance. And based on our expectations, not just for pricing, but for ongoing mix improvement. And this gets to the point Patrick was making about granularity of execution to drive mix of business improvement, which gives you loss ratio benefit in addition to rate relative to loss trend. That's the forward path to get us to that 95.
I think about that mix of business, part of the mixing is Personal Lines. Part of the mixing is Excess and Surplus Lines. In five years, how does the steady state and steady state's a bad term, but is Selective a different company than it was five years ago in terms of what it provides to the market?
I think we're the same company, what we provide to the market. But I think in the profile of the company five years down the road and 10 years down the road looks different from a distribution of premium and revenue and income than it has. And I think, as I mentioned earlier, our concentration in GL and auto liability served us well when property was the big pressure point. But now it's under pressure. I think we want to make sure we continue to build more diversity in our business. And what that means varies by segment. So in commercial lines, we've invested significantly in building out our geographic footprint. So seven or eight years ago, we were in 22 states. We're now in 36 states. And quickly, we'll reach 40 states.
So there's geographic diversification, which takes some of those more challenging regulatory environments and makes them a little bit less of an impact. There's also diversification in the segments that we already operate in outside of the construction arena, which has been very good to us. And we're going to continue to be a player in various construction classes. But there's a lot of other segments we write that have a different mix of premium across the main lines of property, auto liability, and workers' comp. So there's diversification potential there inside of commercial lines. E&S has gone from nothing to 13% of our business over the last 10 years or so. And as we've talked about, we just recently opened up that product to our retail agency partners. And it's traditionally been wholesale. So there's opportunity to expand distribution significantly.
There's opportunity to expand product and appetite on the non-admitted side, but also on the admitted specialty side. That'll be a more diverse part of our book going forward. And then the pivot from mass market , Personal Lines, which we don't think we can be a strong competitor in, to the mass affluent market. And you can see our average home values of new business has been running around $1 million. That's the segment of the marketplace we think we could be a winner and a key player there. And that'll be a diversification play for us as well. And there's geographic expansion potential there as well. So five years down the road, 10 years down the road, I would expect you see more diversity in our portfolio. And we think that allows us to be that consistent player again long term.
You mentioned geographic diversity in the context of companies where the regulatory or legal environment are making it difficult. There's a lot of talk about affordability right now in insurance, particularly among Personal Lines. There's also, to my mind, that the places that are harder to price, you can charge more for because and whatnot, to the extent to which moving into easier geographies, is that really like a thing? Can you get a better as of if you pull out of hard geographies, prices go up because availability goes down? I realize that New Jersey might be a hard place to write. But ultimately, can't you charge a lot in New Jersey and get paid to do it?
Yeah, as long as yes is the short answer. And I don't know. So leveraging those environment differentials, I don't think, is a long-term strategy. I think our view is if you're a good underwriting and pricing company, you can operate in any environment. And we've long even with everything we just talked about in New Jersey Commercial Auto, New Jersey and New York, two of our longest standing Commercial Lines states, despite all of the challenges in those states, are two of our most profitable states across the entire portfolio because you've got very good property business there. It's not that these are terrible markets. And if you know how to operate in those different markets, you could be very successful.
So that's, I think, the most important point is you don't want to try to figure out because the more attractive markets where you have lower rates of litigation, you got good regulatory environments, everybody's there competing as well, which is putting downward pressure on margins. So I think that's the important point. I think the affordability question is an important one. And I think this is where, from a public policy perspective, we just have to do a better job as an industry of expressing what's actually happening here, which is all of this excessive litigation is resulting in higher costs that consumers are paying for, personal consumers and commercial consumers are paying for. But the benefits of that incremental dollar of loss cost is not going back to the claimant. It's going to our defense costs, which are higher.
It's going to the plaintiff's attorneys who are representing them. The actual portion of that additional dollar that they're paying for is not flowing back to them in terms of the benefit. I think our ability as an industry to refine that message in a way that's easily consumable will start to change the public policy landscape. You'll start to see more of what happened in Georgia, which is the situation gets so bad that that message starts to resonate with public policymakers. They understand that's negatively impacting their economy. They do something about it.
Let's pivot the topic a little bit to your distribution partners. I guess retention, as you've been raising prices, has been compromised a little bit. That's always going to happen. Is that steadying out at this point in time right now? And in terms of one thing I always want to know, you don't have to mention comparison, but I always during periods of lower retention, where is the business going? It's something that I always am interested in knowing.
Yeah, so I would say it's certainly. I'm not going to suggest agents love it when you raise prices, especially if your price increases are a little bit above where the rest of the market is. But our approach has always been to be as open in our communication and as targeted in our execution so that you're not just taking significant rate increases across the board. And you're having early conversations on a portfolio basis about the accounts that you're focused on getting a significant rate level or moving the account, have any agent move the account, and those accounts where you have room to be more flexible because your view of forward profitability is better. And as long as we maintain that strong level of communication and the granularity in our focus, agents will come along with us.
With regard to where the business goes, listen, we operate in a highly fragmented and a highly competitive marketplace. Every account that you want to achieve outsized rate on or you specifically say to an agent, we would like you to move this account because we don't think we can get to the right price or we don't think the controls are adequate, there will be somebody there to gobble that account up. It's the nature of our business. It's the nature of the fact that on the Commercial Lines side, there's no pricing power because there's so much fragmentation. I think what's happening of late is as the property market is starting to soften off of some really good results, I think companies are looking for growth and struggling for growth. As a result of that, you see some aggressive behavior on new business.
I think we're seeing that. I know that hasn't been the common commentary out there. There's not an account that we try to up-price significantly that somebody else is not going to come in and write. The agent's going to find a home for it.
You probably saw yesterday, insurance, as the segment was down, insurance distribution was down 9% yesterday. It's not really up today. So the dead cat bounce. There's a little bit, but here and there, people are really worried about AI and whatnot. And I don't exactly agree with all the statements, but I think some of the things. And one of the things I wonder, are insurance commission rates sticky? Or should we expect that commission rates may not people still use agents, but they might not be able to make as much money as a percentage of premium in the past? Is that reasonable? Or you don't see that in the cards at this point in time right now?
I don't expect that to be the primary driver of margin improvement in the near term. Agents provide significant value to their partners, balance sheet partners, underwriting partners like us because they develop deep customer relationships over time. As a result of that, they're compensated for that. I do think you've seen some shifting in certain lines of business downward on a base commission basis. You've seen a little bit more movement away from profit sharing arrangements to more guaranteed base commission. In total, that's the movement hasn't been that great. I don't think that's going to be a significant lever in the near term. It's always part of the conversation in terms of the overall value chain and managing that expense ratio. I think the reaction to distribution based on AI was a little bit of an overreaction, quite honestly.
I think the demise of that distribution model has been predicted for 30 years for different technology advances. It really hasn't come to fruition in the commercial lines space. It hasn't come to fruition. I'm not sure this.
It's not for your homeowners either, although I do question, I don't really think it's that much harder of a transaction to insure someone's home than it is to insure their car. So I think there, and if I'm a florist with a storefront and three employees in a delivery van, I'm not so sure that direct distribution can't solve for that person also.
There's a risk selection consideration. In auto, there's enough homogeneity that you could develop individual class plans and have those fairly accurate, whether it's home or it's commercial lines. I'll call it housekeeping in home. And I'll call it management of a firm in commercial. Two florists are not the same. There's a management team or an individual owner in a small business situation that manages that business a certain way that either makes the likelihood of a loss more or less based on how well controlled that account is. And this applies to small business accounts as much as it applies to larger accounts.
For instance, you're going to write a school or a daycare facility, being able to evaluate the quality of that account and therefore the appropriate pricing of that account without knowing not do they have hiring practices and background checks, but do they actually employ them on a consistent basis? And knowing that allows you to understand what's a high-end class exposure or a low-end class exposure, what is a well-controlled account or a less well-controlled account because the pricing is so varied in commercial lines. And I'm not suggesting technology can't ultimately replicate what a human does in that situation, no different than being able to understand the cleanliness of a house and the upkeep of a house by seeing the inside of it. There's technology that allows you to do that.
That makes two homes that have everything on the exterior that looks the same and the value is the same. Then they're in the same ZIP code. The susceptibility to loss is very different based on how that's maintained.
Not only that, I think there's no symbol set for homes. So in the auto insurance business, you might have a symbol for a 2024 Honda Civic Sport. You know what it costs to repair that. You can get a real sense of truly homogeneous risks in that segment. Homes have various vintages. They're built by different builders. In addition to the pieces that John talks about, about its occupants, but its own construction, there's so much variety there as well.
Arguably, the stock is, on conventional metrics, more attractive than it's been from a valuation standpoint in many years. What kind of flexibility do you have as an enterprise to do something about that?
Yeah, so we've been active in share repurchase during the course of 2025, even a little bit so far in 2026 as well. We recognize that we need a certain amount of capital to run the business. We want to invest in a growing and profitable business for the long term. We also know that it's important to return capital to shareholders as and when we can do so. We have a dividend policy or dividend program. We target 20%-25% of earnings over the long term that we return through that mechanism to the extent that we have additional capital. We like the trade, if you will, or where we are from a valuation standpoint, we will come into the market and invest in our stock. As I said, we've done that through the course of 2025.
We returned something in the neighborhood of $100 million worth of capital to shareholders through dividends and share repurchases last year. So we feel like that's appropriate. But our first and foremost, we want to invest in a growing business.
All right, there's time for one question on the floor. I always want to leave one. But this looks like a very, very anxious crowd. Nobody has questions. They're all so satisfied. I guess we'll leave it there. Thank you very much for your time.
Thank you, [Jeff].
We appreciate it. Everyone, have a good lunch. We'll come back in the afternoon.
Thank you.
Thank you.