SIGI - NASDAQ

Good day, and welcome to the Selective Insurance Group Second Quarter 2025 Earnings Conference Call. [Operator Instructions] As a reminder, this call may be recorded. I would now like to turn the call over to Brad Wilson, Senior Vice President, Investor Relations and Treasurer. Please go ahead.

Good morning. Thank you for joining Selective's Second Quarter 2025 Earnings Conference Call. Yesterday, we posted our earnings press release, financial supplement and investor presentation on selective.com's Investors section. A replay of the webcast will be available there shortly after this call. John Marchioni, our Chairman of the Board, President and Cheif Executive Officer; and Patrick Brennan, Executive Vice President and Chief Financial Officer, will discuss second quarter results and take your questions. John and Patrick will reference non-GAAP measures that we and the investment community use to make it easier to evaluate our insurance business. These non-GAAP measures include operating income, operating return on common equity and adjusted book value per common share. The financial supplements on our website include GAAP reconciliations to any referenced non-GAAP financial measures. We will also make statements and projections about our future performance. These are forward-looking statements under the Private Securities Litigation Reform Act of 1995, not guarantees of future performance. These statements are subject to risks and uncertainties that we disclose in our annual, quarterly and current reports filed with the SEC. We undertake no obligation to update or revise any forward-looking statements. Now I'll turn the call over to John.

Thanks, Brad, and good morning. We delivered an operating return on equity of 10.3% this quarter with excellent investment income, which increased 18% from the prior year period. Excess and surplus and Personal lines produced strong results in the quarter, with both segments reporting quarterly and year-to-date combined ratios at or below our 95% long-term target. Overall, our insurance segments grew 5%, reflecting our disciplined underwriting and pricing strategy in an increasingly competitive market. In Standard Commercial Lines, renewal pure price increased 8.9%, and we continue to execute targeted underwriting and claims actions that I will describe later in more detail. We recorded $45 million or 3.8 points of unfavorable prior year casualty reserve development related to general liability and commercial auto. This pushed our overall combined ratio for the quarter to 100.2%, including an assumed 6 points of catastrophe losses. Given this reserving action and those in 2024, I want to provide context around our process, current risks and how we respond when new loss data emerges. In addition to comprehensive quarterly reserve reviews, we conduct semiannual independent reserve assessments and periodically engage third parties for benchmarking and methodology reviews. Ultimately, reserve development reflects the quality of our initial loss picks. Historically, those have held up quite well. Focusing on the more mature 2015 to 2019 accident [Audio Gap] 2024, we increased our other liability occurrence ultimate losses by 5.5% from our initial picks compared to an average 14% increase for the industry. In the 2021 to 2023 accident years, we increased our ultimate losses by 13% compared to the industry's 5%. While industry trends in recent years have not matched pre-pandemic levels, there have been significant reserving actions, which we believe point to industry-wide pressure on this line of business. Schedule P data confirms that we tend to respond early to loss emergence even for relatively immature accident years for longer tail lines. We added new slides in our investor presentation to highlight this point. For more mature accident years and other liability occurrence and commercial auto liability, our booked loss ratio at the third year-end evaluation for a given accident year is similar to the loss ratio at year-end 2024. For the industry, there has been a more meaningful amount of unfavorable development after the third year-end evaluation. This responsiveness informs our view when setting prospective loss trends and pricing strategy. Our casualty mix of business is higher than our peers. This has been a benefit when property lines have been challenged and our historical catastrophe losses and volatility are lower than the industry average. However, the ongoing industry-wide social inflationary environment has an outsized impact on casualty lines, particularly on claims involving bodily injury. We have been steadily increasing our loss trend estimates for casualty lines in recent years, largely in anticipation of social inflationary impacts on claim severities. Those assumptions are embedded in the current accident year loss ratios we are reporting. In our quarterly reserve review, the frequency and severity estimates included in our ultimate loss selections are key metrics. Claim frequencies are our earliest profitability indicator, and we have a robust monitoring process that includes reviewing actual and expected claim counts. Through the first half of the year, overall accident year 2025 casualty claim frequencies are consistent with or in some cases, better than our initial expectations. Workers' compensation claim counts in particular, have been notably lower than expected. However, we responded to elevated recent accident year paid emergence this quarter. In general liability, we recorded $20 million of unfavorable prior year development, primarily tied to the 2022 and 2023 accident years in the umbrella and product sublines. Higher auto liability severities have increased the frequency of claims piercing the umbrella layer. The product line has also experienced elevated litigation rates and paid severities in recent years, impacting both loss and allocated loss adjustment expenses. In commercial auto, we responded to elevated paid severity emergence in the quarter, strengthening reserves $25 million, primarily related to the 2022 through 2024 accident years. The loss trends we are seeing are broad-based with impacts that across most geographies and major industry groups. Our 2025 loss ratio includes assumptions for escalating severity trends. Even with the reserve increases in commercial auto and general liability for prior accident years this quarter, we remain comfortable with the ultimate severity trend implied by our current year loss ratio selections. We have adopted several strategies in recent years to address social inflation challenges that involve pricing, underwriting and claims. Related to pricing, we continue to seek and achieve overall renewal pure price increases above expected loss trend. I'll discuss pricing in more detail in our segment results, and we continue to execute these increases in a granular fashion. Within underwriting, we continue to take actions to maximize our well-known strategic competitive advantages. Our strong distribution partner relationships, our unique field-based operating model and our sophisticated tools. These actions include tightening underwriting guidelines for select liability exposures, including certain contractors coverage offerings, managing limits in challenging jurisdictions, reducing the number of new umbrella lines with a particular focus on reducing limits greater than $5 million, increasing minimum premiums in general liability and umbrella, trimming underperforming classes or risks with emerging exposures and prioritizing new business in better performing segments. Contractors is our largest industry segment and has a higher mix of general liability and commercial auto exposure. We have built strong expertise in this industry segment and remain comfortable in our ability to produce consistent growth and profitability. However, we continue to invest in diversifying our business mix and geographic footprint. In addition to diversification within Commercial Lines, our efforts to expand our E&S business and Personal Lines mass affluent strategy will contribute to a more balanced portfolio in the future. We remain focused on the fundamentals and claims. Our adjusters specialized by claim type, size and jurisdiction. For example, we have a limited number of adjusters assigned to Georgia bodily injury or New York labor law cases, to drive greater insight into best practices in analyzing these higher-risk claims and defending related litigation. To address social inflation, we have increased the review of cases going to trial, boosting the use of second opinions, engaging jury consultants, conducting mock trials and using roundtables to gain further insights about potential outcomes. We also have created an internal task force to evaluate our fraud beta review processes and gain insights about where to invest additional investigatory resources. And we're currently in the process of developing attorney representation claims models to replace our existing claims litigation models to more quickly identify which claimants are likely to seek representation. Our pricing strategies and underwriting refinements contributed to slower premium growth in the quarter. Overall, renewal pure pricing across our three insurance segments was 9.9%, up 80 basis points from a year ago. We will continue to maintain a balanced approach and make investments to support future growth. However, we believe emphasizing improving underwriting margins and tempering the top line in the current environment is prudent. Turning to segment performance. Standard Commercial Lines reported a 102.8 combined ratio, including 4.8 points of unfavorable prior year casualty development. Renewal pure price increased from a year ago to 8.9%, led by general liability at 11.9%. Commercial auto renewal pure price was 10.4% and Property was 7.8%. Property renewal pure price increases slowed in the quarter compared to our recent run rate, reflecting broader market conditions and improved profitability. Renewal pure price excluding workers' compensation, was 10.2%. Retention for the quarter fell 2 points to 83% due to our rate increases and underwriting actions, along with an increasingly competitive environment. Excess and surplus lines grew 9% this quarter, driven by an average renewal pure price increase of 9.3%. The segment's combined ratio was 89.8%, and we see continued growth opportunities for this segment. We have deployed deliberate E&S strategies and introduced new products over time, expanded our brokerage business and invested in operational efficiency. We are now in the early stages of giving our retail agents access to our E&S offerings. We do not expect to realize immediate significant growth in this latest effort, but believe it will facilitate additional growth capacity over time. The Personal Lines combined ratio was 91.6%, 26.5 points better than a year ago. Our rating and non-rating actions to reposition this book are continuing to gain traction. We are emphasizing growth in states with adequate rate levels. And as a result, Personal Lines net premiums written declined 5%. However, target business grew 16% in the quarter, with nearly all new business being in our target mass affluent market. Renewal pure price for the quarter was 19%. We expect rate changes will remain above loss trends, but moderate in comparison to those achieved in 2024 as the portfolio moves closer to achieving a long-term target profitability. In summary, we delivered a 12.3% operating ROE through the first half of the year and remain focused on executing our risk management strategies while driving long-term profitable growth. Loss trends remain elevated, but we are confident in our ability to quickly identify and address areas within our control and deliver consistent underwriting margins over the long term. Now I will turn the call over to Patrick, who will provide more details about our financial results.

Thanks, John, and good morning, everyone. For the quarter, fully diluted EPS was $1.36 and non-GAAP operating EPS was $1.31. Our underwriting performance was breakeven, but our return on equity was 10.7%, and operating ROE was 10.3% due to the investment portfolio's continued strong performance. The GAAP combined ratio for the quarter was 100.2%, which was elevated primarily due to the 3.8 points of unfavorable prior year casualty reserve development that John discussed. Catastrophe losses were 6.7%, which was better than anticipated and 1.7 points better than the prior year period. We continue to see the benefits from profitability improvement actions in Personal Lines as we execute our mass affluent strategy. As expected, access and surplus lines delivered another strong quarter. In Standard Commercial Lines, we're focused on executing appropriate underwriting actions and rate increases to address the current environment and position us to achieve target margins. Our overall underlying combined ratio for the quarter was 89.7%, an improvement of 170 basis points from the prior year period. Year-to-date, the underlying combined ratio was 90.8% which is up 20 basis points from the first half of 2024. Non-catastrophe property losses were 15 points year-to-date, which is 170 basis points better than a year ago and reflects continued benefits from property lines earned rate and the tightening of terms and conditions over the last few years. Year-to-date, these benefits are more than offset by a 140 basis point increase in current year casualty loss costs. The expense ratio increased by 60 points, primarily driven by higher expected employee compensation after last year's lower profit-based payouts. We remain focused on expense discipline and deploying capital to support scale, enhance decision-making and operational efficiency. Second quarter after-tax net investment income was $101 million, up 18% from a year ago. This income generated 13 points of return on equity, up 50 basis points from the second quarter of 2024. We continue to position our investment portfolio conservatively and have made no significant changes to our investment strategy. Total fixed income and short-term investments at quarter end represented 92% of the portfolio and had an average credit quality of A+ and duration of 4.2 years. We delivered strong operating cash flow in the quarter, allowing us to make over $750 million of new investments. The average new purchase yield was an attractive 5.7% pretax and the quarter end average pretax book yield was 5%. We expect this embedded book yield to provide a durable source of future investment income. Turning to capital. We ended the quarter with $3.4 billion of GAAP equity and $3.3 billion of statutory surplus. Book value per share increased 9% in the first half of the year driven by our profitability and a $1.74 per share reduction in after-tax net unrealized fixed income security losses. Debt to capital was 21.1% below our internal threshold of 25%. We continue to return capital to our shareholders by issuing quarterly dividends and opportunistic share repurchases. During the second quarter, we did not repurchase any shares of common stock and $56 million remained available under our repurchase authorization as of June 30. Effective July 1, we successfully renewed the casualty excess of loss and property per risk treaties that cover Standard Commercial Lines, Standard Personal Lines and E&S. The casualty excess of loss treaty covers all our casualty business and provides $87 million of protection above a $3 million retention. We increased the first layer to a $3 million retention, up from $2 million and continue to retain a portion of the first layer through co-participation. The remaining treaty layers were fully placed without co-participation. We also renewed our property per risk treaty, now providing $95 million of coverage in excess of a $5 million retention for losses on a per risk basis. The $30 million limit increase from the expiring treaty reflects growth and higher insured values. Pricing in key terms and conditions were within our expectations heading into the renewal. In light of results through the first half of the year, our revised 2025 guidance is as follows: we expect our 2025 combined -- GAAP combined ratio will be between 97% and 98%, up 1 point from prior guidance. Our guidance includes 6 points of catastrophe losses and the impact of prior year casualty reserve development reported through the second quarter. It assumes no additional prior year casualty reserve development and no further change in loss cost estimates. We do not make assumptions about future reserve development as we book our best estimate each quarter. After-tax net investment income of $414 million -- excuse me, $415 million, up from prior year guidance of $405 million. Our guidance includes an overall effective tax rate of approximately 21.5%. Our guidance assumes an estimated $61.5 million of fully diluted weighted average shares, including those repurchased in the first quarter, and assumes no additional repurchases under our existing share repurchase authorization. With that, I'll now turn it over to Q&A. Operator, please start our question-and-answer session.

[Operator Instructions] Our first question comes from Michael Phillips with Oppenheimer.

Appreciate the new slides, John. I guess, I want to ask a question on a slide that's been in your deck for a while, and that's where you show kind of the excellent above average and below average retention in pure price. That slide hasn't changed much over the past year. And I guess, I wonder if, if we look forward to maybe the next year, should it change specifically the below average, very low jack-up 20% rates and see retention fall to the floor. Why not see that? And I guess, maybe partly the answer could be your comments on broad-based. Is this social inflation issue? Is it across your entire book? Or is it more concentrated in kind of what you're labeling these below average risk.

Yes. Thanks, Mike. Appreciate the question. And that's a slide we have shown for a long time. I think that points to a mix of business improvements over time as you continue to have a lower retention and a higher rate on that low and very low bucket that we highlight there, and higher retentions on the excellent and above average buckets. That is a lever we continue to push. I think we always want to be mindful of the fact that unlike when you have actual rate plans in Personal Lines or in small commercial that you don't deviate on an account-by-account basis, you're going to push that a lot harder, but there's also an underwriting overlay -- subjective underwriting overlay that happens across those buckets. So on a directional basis, what we see there is, what we would expect and like to see, but that is a dial we will continue to turn, but it's not as absolute and pure as it might be in a true rate plan like you would have in small Commercial -- automated small Commercial or Personal Lines, where you don't have discretionary pricing and underwriting judgment overlay on top of your model output.

I mean, I guess, part of the question then would be, are the issues you're seeing, are they across your entire book? Or are they more concentrated in certain accounts that you kind of want to win yourself off of?

Yes. I would say -- and Mike, I appreciate your pushing on this point, and I know we've said this before, this is something we continue to evaluate. What we've been seeing and reacting to with higher paid emergence is evident across industry classifications and it's evident across geographies. And I think that's a very important point. It applies to our commentary relative to commercial auto, and General Liability as well. Now that said, our focus here is making sure we're responding appropriately to what we think is an industry-wide societal shift driving social inflation, but also recognizing there are things in our control that we have to have a continuous improvement mindset around making sure that individual risk selection decisions, individual claim decisions are getting to the right outcome on a regular basis. So there's a refinement on that front, that we think will continue and we should continue to focus on. But what we're seeing here is, in fact, broad-based. And I think, this -- based on the comments we've seen in the early write-ups on our -- after our earnings release, I think it's important to reiterate this point. There's no question that we have a higher mix of Commercial Auto and General Liability than some of our peers and for the industry as a whole. Those two lines, which are those two that are more impacted by social inflation, Commercial Auto and General Liability are about 64% of our Commercial Lines premium and just over 51% of our total premium. Now our thesis that the severity emergence that we're seeing is driven by social inflation as opposed to something idiosyncratic to our particular portfolio, is something we believe in, but it's also a thesis we continue to evaluate critically. And how do we do that? First is making sure -- with the volume of risk metrics we have on our portfolio, across our portfolio, making sure that we have relative stability in the portfolio and relative stability in the pricing of our portfolio on a risk-by-risk basis. That continues to be the case. Well, let me go a step or too deeper in terms of how we continue to validate this thesis. As I know many of you do, we do deep industry analysis on frequency and severity trends. So we do deep Schedule P analysis for these lines of business for our peer group individually and collectively and for the industry as a whole to validate that the trends we're seeing in frequency and severity match what the industry is seeing more broadly. We see that, and we see a consistent pattern. And the reason we put those additional slides into the -- the investor presentation is to show at least on a historical basis that we tend to react more quickly. And you see that for other liability occurrence, and you see it for Commercial Auto liability, looking back to those more mature years that were pre-pandemic. Now I understand that's history. And at some point, we will learn whether or not that same pattern holds when you look at these more recent accident years that we're reacting to. But I think that's an important data point when you look at how our estimates have held up and how quickly we've gotten to what our current view of ultimate is for those prior accident years. So I mean, this is an important topic for us to make sure we spend the appropriate amount of time discussing. But trust me, we continue to validate our thesis that this is, in fact, widespread and industry based.

Okay. That's perfect. I guess if I could, and I also appreciate that it's more than pricing, you talked about your underwriting actions and some claims things that you're doing. But should we be surprised at all, if I heard the numbers right, on the GO pricing was 11.9%. It's kind of in line with last quarter, did not see more pricing on the GL's piece.

Yes. I think that number -- that's a blended number. It includes the underlying GL and the umbrella GL is just under 11% in that number on a year-to-date basis and umbrella is closer to 14% on a year-to-date basis inside of that number. That number has actually moved on a pretty significant sequential basis. Now I'll also make this point. And I think despite what you might be hearing on a consistent basis across the industry around recognition of social inflation and higher severity trends and views around casualty pricing. Our pricing stands at, call it, roughly 11% on GL, is negatively impacting our conversion rate on new business, and it's put a little bit of downward pressure on our retention. You can see our retention while still strong, has ticked down to 83% for Commercial Lines. I think that's the best indication as to where pricing is relative to the market. And to the extent there's not yet full recognition on the higher severities in the more recent accident years. You might not see that fully reflected on an industry basis, and that's a competitive dynamic that we're dealing with. But as we've said, on multiple occasions, we have high conviction in our view of the more recent accident years, and that feeds our high conviction on our pricing stance. And we're willing to make that trade with regard to a little bit more top line pressure in an effort to achieve our profit objectives.

Our next question comes from Bob Jian Huang with Morgan Stanley. .

Maybe just want to unpack on the Commercial Auto reserving a little bit here. If we revisit the most recent Schedule P, you lowered your initial aspects in the most recent accident year. And I think at the time, the explanation as it was due to a pricing improvements. Just given the reserve charges that have been taken, is there a way for us to really think about the assumption changes going forward? In other words, how -- is there a way to help us gain a little bit more comfort in terms of what is the Commercial Auto assumptions? And then how we can think about the losses and the potentials going forward from here?

Yes. So our -- I think -- and I went through this last quarter, I'll kind of hit the high points again with regard to Commercial Auto, in particular. If you look at our assumed loss trend on average over the last four prior accident years, so '21 through 2024. Embedded in our expected loss ratios was an average assumed loss trend of about 8%. And I'm talking Commercial Auto Liability, Commercial Auto bodily injury in particular. And over that same 4-year time period, our average renewal pricing on Commercial Auto liability was just over 10%. And I think when you look at the loss ratio improvement that you've seen in our reported results on an accident year basis, it was really that continued gap over an extended period of time between that rate of just over 10% and the assumed loss trends at right around 8%. And as we evaluate the more recent years and the emergence that we've seen, we continue to have a similar view about loss trends with -- across our casualty portfolio, which have been increasing on an overall basis. But in Commercial Auto, in particular, we had been embedding a higher loss trend. And I think that's the best way to think about run rate performance on a Commercial Auto basis. Now again, we evaluate emergence on a quarterly basis. To the extent that changes on a go-forward basis, we'll change our pricing stance. But when you look at the Commercial Auto pricing we've been getting, we think that's a sustainable level. And that's where the industry continues to be from a pricing perspective.

Okay. So maybe just a follow-up on my point then. In this case, so is it fair to say that, despite the charges you took here, you still feel that 8% and the 10% are appropriate assumptions go forward? Or do you think that 8% might need to move up? Or have you moved up that 8%? I'm assuming no, but just kind of curious if that's the case.

We would say those are still reasonable assumptions, based on everything we continue to see.

Our next question comes from Paul Newsome with Piper Sandler.

Good morning. I hope you guys are all doing well. I was hoping you could walk me through a little bit more detail about the exit that you set for the quarter and I guess, so far this year. I was curious if there's any thoughts about raising that accident year Peg, given just the uncertainty of what we have seen with social inflation. And obviously, there's a mix here to account for. But maybe you could just give us some thoughts on that and why they actually might not rise as much or what do you think -- would you thought?

Yes. No, thanks, Paul. I appreciate that. So when you look at where we are and the trend assumptions we made, and we give you the rolled-up number with regard to casualty and casualty ex-comp, we then compare that to what are we observing in the more recent accident years in terms of severity trends, and that's sort of our test around whether or not we have a level of confidence in the current year assumptions that we need. And based on that, that's why we remain comfortable because the assumed severities that were incorporated into our expected loss ratios for GL and Commercial Auto, and across our casualty portfolio in total in terms of expected loss ratios remain in line with what we're observing in the most recent accident years, which is the most elevated point of the diagonal. And I think that is where our confidence remains. Now the other point, I think is important to make just to bring you back to last year, we also boosted and I'll focus on GL in particular, in '24 year when we took action on prior years, we boosted our GL expected loss ratio by a little over 7 points for the '24 accident year. That then got incorporated into our view of 2025. So that was an important step from our perspective, and allowed us to make our -- put our best foot forward in terms of staying ahead of this higher severity mergers we were seeing. And I think that's another important consideration and why we remain comfortable with our book loss ratios for the current year.

Yes, super helpful. I noticed the workers' comp combined ratio maybe I misread this, I apologize asking a stupid question, popped up for the quarter on the combined. Anything there? I would have thought that would be a little bit different from the kind of social inflation issues, which was much more consistent we've seen in the results in the quarter?

Yes. No, thanks, Paul. So this was something I came up last quarter as well. And if you look at the '24 year, so what you're pointing to in Q2 was also there in Q1. And so it's there on a year-to-date basis. If you look at our '24 -- 2024 accident year for workers' comp, it was in the range of roughly 97%. So when you strip out the favorable development impact, it was around 97%. And as we've consistently done based on the fact that we have been pointing to a flattening frequency trend in workers' comp, we assumed no frequency improvement from 2024 to 2025. We talk about average severities. I think you're seeing this pretty consistently across the industry, on the medical -- on the medical side over severity of about 5% and earned rate that's been running slightly negative, call it, in the 3% range. So flat frequency, severity up about 5%, rate down about 3%, you just take that 97% and roll it forward and that gets you to that both combined ratio or book loss ratio that you're pointing to. Now also highlight in my prepared comments when we talked about frequency for the '25 year we specifically called out workers' comp as having frequency favorable through the first 6 months. Now granted, it's 6 months. And I don't think we're ready to declare that a trend, but it might indicate that the flattening trend that we had been seeing or observing in the more recent accident year is not continuing through 2025. So a lot there. Hopefully, that gets to the heart of your question.

No, it does. If I could sneak in one, just one more. Is there anything unusual about your excess casualty here on grower book from a limit or terms and conditions that would make it anything different than the rest of the industry? I suspect the answer is no, but I don't think I've ever asked the question. .

I would say the answer is no. And in fact, if anything, the biggest difference might be on a peer-to-peer comparison basis is our umbrella is entirely supported. So we don't write any umbrella where we don't have the underlying Auto or GL or both. And I think that gives us better earlier insight into frequency and potentially severity of our umbrella portfolio. The profile is a lower limits profile, so 95% of our umbrella and our grow is about roughly $400 million direct premium portfolio. The 95% of the policies have a limit of $5 million or less and about half of them have a limit of $1 million. So that's the overall profile of our umbrella portfolio.

Great. I always appreciate the help and really always respect your thoughts and comments. .

Our next question comes from Mike

Just back to the kind of the thinking through the reserve additions relative to -- I know you said that the industry is likely kind of behind on adding reserves. And I think we wouldn't -- most people wouldn't disagree direction with that comment. The industry has been adding to their social inflation reserves for the years now. But I guess just kind of think structurally about selective. You've said that you've seen a flattening frequency trend in more comp. I feel like most, almost no carriers out there have cited that and frequency of just be something, it's just more black and white, I would think. And you've also said that you just have more social inflation exposure, which might be due to the contractors' book. So I guess I'm just trying to tease out, I mean, [indiscernible] is unique in that. I feel like you do have more of a special sauce with contractors. Do you feel that it's fair to paint the picture that some of this is unique to selective due to your mix of business?

Yes. Well, I think -- and again, it's 6 months worth of data. It could very well be that, that flattening frequency trend was a 1-year phenomenon. Now again, I think time needs to pass to understand whether is directionally. But if you look prior to that, we had seen a pretty consistent downward trend in frequency, albeit maybe on a slightly lower amount than it had been across the industry. So I think we should acknowledge that. The fact that our overall portfolio is more weighted towards construction would follow through to our workers' comp portfolio. And I think there's no question that through the pandemic period and post pandemic, and the influence of remote or hybrid work on frequencies and other segments of workers' comp didn't impact construction. So I think that probably would explain some of the difference with our portfolio relative to frequency change. But I think it's important to note that -- and I cited our accident year combined ratio for workers' comp in 2024 of around the 97%, I know reported combined ratios tend to get all the headlines. But if you look across the industry, across the country based on NCCI and other state bureau reports, the accident year for the industry in 2024 was right around 100. Now again, the older accident years have continued to emerge favorably for the industry and for us as well. And to the extent that continues, our current year assumptions might prove conservative, but that's how we approach a line with a tail like that.

Okay. That's helpful color, John. Maybe as a follow-up, you've been in this business for decades, John. We've seen charges in the last 5 of the last 7 quarters. Is there -- I feel like, I can't see that in Selective's history historically. So it feels like this is somewhat unprecedented. But I mean, do you -- is there -- is it -- what should we be looking at? Is it -- is this normal that just takes a lot of time, years for underwriters to really get their and actuaries to get their hands on on the loss trend? Or is it -- should we be looking at, it's just really going to depend on whether social inflation continues to move north? Are there any indicators or signs of that can give us confidence that there won't be further meaningful changes going forward?

I would say that the greatest confidence you should get from our actions is the fact that we're reacting to very recent and quite immature accident years, in terms of when you look at the actual paid and even case in and paid data from those accident years, it's relatively immature. So we're looking at recent paid emergence patterns and projecting a smaller inventory of paid claims to ultimate and we're reacting to the more recent accident years. And I think that's what probably makes us a little bit unprecedented for the industry is it's -- this activity is being driven by very immature accident years on longer-tailed lines, just to put it in perspective, and I'll give you rough numbers. For the 2023 accident year, our expected -- our ultimate expected dollars and you look at what percentage of that is paid at this point for the '23 accident year, for GL ex products, it's like 22%, for products, it's like 17%. And for Commercial Auto, it's just over 30%. So these are immature years and you're reacting to pay data and your actuarial methods on a paid basis are highly levered. When you see severity emerge like we've seen as an industry have seen, and that creates a higher level of uncertainty. And I think that is what makes us a little bit more unprecedented. Now when -- and I can't speak to what others might be doing. When you see something like this, your actuarial indications and your actuaries are going to wind up putting more weight on the more recent years from an emergency perspective as opposed to traditionally maybe weight was spread across the prior 7 years to identify paid emergence trends, but they're going to put more weight on the last 3 years, and that has a leveraged effect on a small amount of paid data. So I know I'm deep into the reserving process, but I think it's instructive here to get to the root of your question. This is not us trying to get our arms around old accident years. And I think, when you look at the industry in total, and this is from Dowling analysis, in 2024, the industry in total added $10.5 billion to other liability reserves, $10.5 billion. Almost half of that or around $5 billion was added to the pre-pandemic years. And I think that says a lot in terms of how some of these lines emerge over time. Now for us, that has not been the case. We have not seen any further emergence since we acted on the pre-pandemic years at the end of 2023 and with a $55 million reserve adjustment. So I don't know that I'm giving you anything that's going to give you high conviction for us or for the industry. But I think what makes this different is we're talking about very recent relatively immature accident years for longer tail lines of business that everybody is trying to determine when the inflection point or when severity trends will start to flatten out.

Okay. I think that helps. So I'll have to follow up and use your insights and look at more data. But so you're saying you are seeing paids increase a bit higher than expected in some of these lines. Is that correct in more recent accident years?

Yes. Yes. That's driving the entirety of our reserve adjustments over the last few quarters, it's paid emergence in the more recent accident years. This is not frequently driven. This is not older accident years. It's paid emergence. And you're seeing it in cures as well in course including case reserves, but it's more pronounced in paid emergence.

Okay. Got it. And then lastly, just pivoting to the -- you mentioned in your prepared remarks that no surprise that commercial property pricing is decelerating a bit of I think, absolute levels that are fairly healthy due to healthy profitability. As you -- is that a phenomenon that would potentially continue given it does appear the industry is earning healthy profits on commercial property? Or any viewpoints or insights there?

Yes. I would -- and I've seen some of the other industry commentary, which I would generally agree with. I think at the higher end of the market, including layered programs, share programs, we don't -- which is not really play. I think there's probably been more contraction from a pricing perspective in our part of the market. I do think you've seen some contraction. I think that will continue. I think it will remain above where property loss trends are which in the beginning of the year, as we talked about, we had property loss trends about 3.5%. So I think there's still some potential margin expansion there. We also have the overlay of potential tariff impacts that many are talking about and I think at least building into their forward view of potential loss trend increases -- and I think we still have the level of catastrophe volatility across the industry. So I think that will temper the drop. But I would expect property pricing to continue to float a little bit lower but remain favorable to loss trends as casualty goes higher.

[Operator Instructions] Our next question comes from Meyer Shields with KBW.

Great. I have two questions on the Commercial Lines. First, with Commercial Auto, we saw the same loss ratio improvement year- over-year in the second quarter as we did in the first. But if you raise the accident year '24 loss pick, and presumably, that needs some sort of catch-up for the first quarter, shouldn't that improve -- accelerate sequentially?

Yes. Now I think I remember the first point is the $25 million was spread across three accident years through '24. And also remember what I had mentioned earlier, which was what we had embedded into our '25 loss trend assumption across our casualty lines, I think that's the primary areas that I would focus you on. And remember, our practices, and we've done this before, we've done it with Commercial Auto over the course of the last 10 years, which has raised the current year loss ratio when we saw an amount of pressure that we thought was the right thing to do. We did it in General Liability last year. We've done it in Commercial Auto liability in the past. So we're certainly open to doing that. but we haven't seen evidence at this point with a 24-year to the level that leads us to think differently about where we're booking '25.

Okay. That's fair. Second question, I guess, I'm looking at the BOP business, and it's good to see that there's been no adverse reserve development there. But I'm wondering why that casualty side hasn't faced the same sort of social inflation that we're seeing in General Liability.

Well, it's much more aligned for us that BOP liability is something that we evaluate every quarter as a line of business basis. But I do think your point raises another point relative to industry comparisons, which is we report all of our General Liability in Schedule P as general liability. We don't include any of that in CMP, a number of our peers do incorporate their GL business that's written on a companion basis, companion policy basis in CMP, which makes it hard to get a full picture of GL performance, because it's co- mingled with property performance, which has been improving. But BOP liability, it's a different mix of business for us. It's a smaller line of business, but it's one that we evaluate on a quarterly basis, the liability portion like we do other lines of business.

Okay. I didn't realize that. So the BOP premium or the BOP the line of business that you report, that's just the property component?

No, no. It's a property liability are in there. All I'm saying is from a reserve review process, we evaluate that subline of BOP on a quarterly basis from a frequency severity and a carry reserve perspective.

Okay. And I guess you either -- so you're not seeing the manifestation of social inflation. I guess the full limit profile is different. .

It's -- yes, it's a different portfolio of business, and it's a much more portfolio. So even movements that could happen favorable or unfavorable, wouldn't really be noticeable on an overall basis.

Our next question comes from Michael Phillips with Oppenheimer.

I just want to make sure I got this, John. The $20 million GO 2022, '23 accident year, was that entirely in the -- it looks like about $370 million of excess umbrella. It's entirely in the umbrella business, not the primary?

No, it's both. It's -- yes, I think we highlighted products. So when we evaluate GL reserves, we look at products, non-products and umbrella. We were pointing out that the $20 million was leaning more from products and umbrella. And remember, our products portfolio is because of our construction mix tends to be predominantly completed operations type claims as opposed to your traditional kind of heavier exposure consumer products that you would think of in a more traditional sense.

Okay. So that's what I'm asking. So that $20 million was umbrella and products, but not necessarily in your primary geo on the liability occurrence.

Yes. I would say, generally speaking, we haven't seen emergence that required us to act with regard to non-products, which was the primary driver in 2024 actions that we took.

Thank you. I'm showing no further questions at this time. I'd like to turn the call back over to John for closing remarks.

Well, thank you all for joining us this morning. We always appreciate the engagement and the questions. And as always, please feel free to reach out to Brad if you have additional questions. Thank you all.