Good morning, and welcome to CNA's discussion of its 2019 third quarter financial results. CNA's third quarter earnings release presentation and financial supplement were released this morning and are available via its website, www.cna.com. Speaking today will be Dino Robusto, CNA's Chairman and Chief Executive Officer, and James Anderson, CNA's Chief Financial Officer. Following their prepared remarks, we will open the line for questions. Today's call may include forward-looking statements and references to non-GAAP financial measures. Any forward-looking statements involve risks and uncertainties that may cause actual results to differ materially from the statements made during the call. Information concerning those risks is contained in its earnings release and in CNA's most recent 10-K on file with the SEC. In addition, the forward-looking statements speak only as of today, Monday, October 28th, 2019. CNA expressly disclaims any obligation to update or revise any forward-looking statements made during this call.
Regarding the non-GAAP measures, reconciliations to the most comparable GAAP measures and other information have been provided in the financial supplement. This call is being recorded and webcast. During the next week, the call may be accessed on CNA's website. With that, I will turn the call over to CNA's Chairman and CEO, Dino Robusto. Please go ahead, sir.
Thank you, Cody. Good morning, everyone. I'm pleased to share our third quarter results with you today, which reflect continued good underwriting performance, accelerated price increases, and strong growth across our U.S. operations. Core income for the third quarter was $102 million or $0.37 per share, inclusive of $170 million, or $0.63 per share after-tax charge related to the unlocking of our long-term care active life reserves, driven by our decision to reset our assumptions on the discount rate. James will provide further detail on the unlocking, as well as on the favorable $44 million after-tax outcome of our annual long-term care claims reserve review, the fourth year in a row of a favorable claim outcome. For the quarter, the P&C underlying combined ratio was 94.6%, a slight improvement to last year's third quarter results.
Strong underlying performance in commercial and specialty offset an almost three-point deterioration in international. In international, we remain confident in our belief that we are doing all the right things and have already seen some improvement in our results, including a strong underlying loss ratio of 61.4% through the first three quarters of this year, which is in line with our overall company result of 61.1% through three quarters. In addition, our re-underwriting efforts have significantly reduced our international catastrophe exposure, which allowed us to avoid catastrophe losses in Asia due to the recent events there. The higher underlying loss ratio this quarter is driven by lines of business we began non-renewing late last year, validating our previous re-underwriting decisions. Obviously, during this process, the improvement won't be a quarter-over-quarter straight line.
The P&C all-in combined ratio of 97.6% included 1.8 points of catastrophes and 1.2 points of unfavorable prior period development, principally related to a block of commercial legacy mass tort accounts we no longer write. As usual, James will provide more detail on our prior period development. Our expense ratio in the third quarter was 32.5%, nearly a full point lower than the second quarter. We are pleased with our net written premium growth of 8%, which was fueled by robust growth in the U.S. segments. Gross written premium, excluding third-party captives, was up 11% in the U.S. segments, and net written premium was up 9%. In international, gross written premium was down 2% as increases in Canada continue to be offset by re-underwriting actions in our Lloyd's syndicate.
We continue to effectively manage the rate retention dynamic, achieving higher rate increases in each business unit and doing so with steady retentions, except for specific areas such as aging services and large property, which had lower retentions. We are satisfied with the trade-off, as those areas have the highest rate needs, and we continue to walk away when we can't secure adequate terms and conditions. In the third quarter, rate for P&C overall was +6%, up two points from the second quarter. Commercial rate was +4%, up one point from the second quarter. Specialty was +6%, up two points from last quarter, and international rate was +10%, up three points. Let me drill down on these rate increases to provide more insight into our execution in the marketplace.
In our specialty business unit, rate was actually up 13% outside of our professional E&O program business, which we refer to as affinity. As we have highlighted before, our affinity business mainly represents longstanding multiyear programs and therefore are less affected by recent price movements. In our healthcare business, which has experienced higher loss cost trends for more than two years, as we have referenced on several prior calls, rate increased further to 18 points compared with 14 points in the second quarter. In public company D&O, rate was up 42 points in the quarter, a level significantly higher than the 15 points we achieved in the second quarter.
While the magnitude of this rate increase was influenced by several large accounts, we had broad increases across the book, as reflected by the median of the rate achievement distribution for the third quarter, which was double the median of the second quarter. In commercial, excluding workers' compensation, rate was 6%, a point higher than the second quarter, and included umbrella up 10%, property up 8%, and auto up 7%. International rate was 10%, compared with 7% in the second quarter, with broadly consistent rate movement in Canada, Europe, and in our Lloyd's syndicate. Before moving on to our new business results, I want to make a few additional comments on rate and loss cost trends and the resulting margin impact, given the understandably high level of interest in this interdependent dynamic.
With the six points of written rate in the third quarter, we have achieved slightly more than three points of earned rate against our long-run loss cost trends, which are just above 2.5%. Obviously needs to be sustained before we recognize any meaningful margin expansion. Particularly when you consider that we experienced almost four years of rate changes being lower than long-run loss cost trends starting early 2015. All else equal, we would have to sustain the current rate levels through mid-2021 to make up, if you will, the lost ground in pricing. In terms of the impact to margin, the correlation is oversimplified, as it would have to assume that long-run loss cost trends are the same as actual loss cost trends during that period, and that our book of business didn't change.
In reality, losses have deviated from the long-term average quarter to quarter. Moreover, we have elevated our underwriting focus in the last few years, pulling levers beyond pricing that positively impact the loss profile of the book. These two dynamics, along with the fact that some part of exposure increases during that period also acted like rate increases, explain to a large extent why the lost ground in pricing during those years didn't equate to a dollar for dollar compression in margin. Similarly, going forward, even if we sustain the current rate movement earning through mid-2021, gaining back the lost ground in pricing, it won't necessarily equate to a dollar for dollar expansion in margin. Rather, we need to incorporate all the vectors of influence on our accident year loss ratio picks, such as actual claim frequency and severity trends.
As we previously highlighted, a couple of lines have experienced higher loss trends, and based on the consistency of the pattern of deviation, led the actuaries to raise their respective long-run loss cost trends. We must also account for portfolio changes, terms and conditions changes beyond price, changes in reinsurance coverage, as well as legal, judicial, and regulatory dynamics, all of which we do in a disciplined fashion during the quarterly reserve reviews. Now, let me share the detail on actions we took on certain long-run loss cost trends, accident year picks, and reserve increases based on the overall puts and takes from the influencing factors. As we have disclosed over the past several years and discussed on prior calls, one area that we have consistently seen a more aggressive plaintiff bar has been within our healthcare portfolio, especially aging services, where they have been targeting medical malpractice claims.
We began seeing this in 2016 as claims from older accident years accelerated in both number and cost. We raised our accident year loss ratio in 2016 and began taking underwriting action at the time to mitigate the higher accident year loss ratio. Importantly, we also raised our long-run loss cost trends in 2017 because we saw the elevated frequency and severity due to the deteriorating legal climate persist, as evidenced by further adverse prior period development we took and previously disclosed. We continued to review the results in subsequent quarters and determined even more aggressive underwriting action was needed. Last year, we began to substantially increase rate while continuing to re-underwrite the portfolio. Although we have seen some improvement in the frequency trend for aging services this year, our view of long-run loss cost trends is 11%.
We will continue to push on pricing and other terms and conditions, which, as a market leader in the healthcare space, we do have the ability to execute effectively upon. We have also seen, to a lesser extent, the impact of a more aggressive plaintiff's bar in our umbrella book, specifically the auto exposure in our umbrella book, which we have also commented on during past calls. Beginning in the second quarter of 2018, we began to see consistently higher severity on our auto claims in the excess layers for accident years 2014 through 2016. As you know, umbrella claims take time to develop and recognize adverse prior period development for those accident years. We also increased the 2018 accident year loss ratio by eight points in the fourth quarter, and that essentially remains our current accident year loss pick.
In addition, we quickly began to take a number of underwriting actions, such as raising our attachment points on more auto-exposed accounts and reducing the number of umbrella accounts with larger underlying auto exposures. Even with these actions, the actuaries felt it was appropriate to raise the long-run loss cost trends for umbrella by 100 basis points in this year's third quarter to reflect that higher severity trend. The good news is we now have higher rate achievement in umbrella, and it is continuing to accelerate. Here we are today, and based on the momentum that currently exists and the overall tone of the market, I now believe it is more likely that rate increases running above our loss cost trends will persist throughout 2020. It is quite rational in light of the lost ground and the pressure on loss cost trends I just highlighted.
The sustainability of price increases is further validated when you place it against the backdrop of an exceedingly protracted low interest rate environment. Getting back to our production results, another important element of price increases we are experiencing is that they extend to new business pricing, and that has helped fuel our new business growth, which was up 10% over the same period last year. This traditional market causes a heightened level of business to be marketed as agents and brokers grapple with significant changes in terms and conditions that they fear may impact even their best-performing accounts. Our focus on re-energizing relationships with our distribution partners over the last 24 months, along with our talent investments that I discussed on past calls, has allowed us to capitalize on this dynamic, which has contributed to our new business growth. With that, I'll turn it over to James.
Thanks, Dino. Good morning, everyone. Our property and casualty operations produced core income of $241 million in the third quarter. Pre-tax underwriting profit was $42 million, and underlying underwriting profit was $95 million. Moving to each of our P&C business units, Specialty's third quarter underlying combined ratio was 92.1%, and its underlying loss ratio was 60.1% in the quarter, consistent with both the third quarter of 2018 as well as the first half of this year. Specialty's overall combined ratio was 89.8%, included 2.8 points of favorable prior period development. This favorable development was primarily in accident years 2017 and prior, driven by surety and management liability, and partially offset by healthcare being adverse. With the significant rate we're now achieving in healthcare, along with considerable underwriting efforts, we are confident that we're getting this book back under control.
Nonetheless, as Dino just highlighted, we will be cautious regarding the recognition of margin improvement until we see the benefits of rate manifest themselves in our actuarial analysis. Specialty's gross written premium, ex third-party captives, grew a healthy 9% in the quarter. Our Commercial segment's underlying combined ratio was 93.8% in the quarter, and its underlying loss ratio was 61.5%, which is a point higher than the third quarter of last year, but a slight improvement compared with the first half of this year. The third quarter overall combined ratio in Commercial was 101.6%, including three points of catastrophe losses and 4.8 points of adverse prior period development.
As Dino mentioned, the majority of the prior period reserve charge in Commercial, in fact, $35 of the $40 million, came from a block of legacy accounts from accident years 2009 and prior, and are unrelated to the New York revival statute legislation. This adverse change was primarily driven by a reevaluation of expected reinsurance recoveries on those reserves, in addition to an increased case reserves on a handful of accounts. I highlight the revival statute because there's been a lot of discussion about it. However, it is early in the process for us because we are primarily in an excess position in the areas where we may have potential exposure. We will continue to evaluate this as we get more information. Aside from the legacy development, Commercial had $5 million of adverse prior period development driven by umbrella.
Commercial's gross written premium, ex third-party captives, grew 13% in the quarter. The underlying combined ratio for our International segment was 105.3% in the third quarter. As Dino pointed out, the first three quarters of 2019 show an improving picture with an underlying combined ratio of 98.9%. In the third quarter, the underlying loss ratio was 67.3%, 1 point higher than the third quarter last year and several points higher than the first half of 2019, driven by large property losses in our Lloyd's syndicate and in Europe. As we've noted in previous calls, the improvement in international will not manifest itself in the results overnight. The expense ratio deteriorated 1.7 points year-over-year due to the reduction of earned premium from our re-underwriting efforts. The all-in combined ratio was 107.4%, including 1.7 points of catastrophe losses and minimal prior period development.
Our P&C expense ratio of 32.5% was slightly below our current run rate, as the quarter's results included some favorable items and acquisition expense. Our life and group segment produced a core loss of $122 million in the quarter. This result includes an after-tax charge of $170 million related to the unlocking of our long-term care active life reserves, partially offset by a $44 million after-tax gain resulting from our annual long-term care claim reserve review. The claim reserve review, which is the review of our current claim population, was favorable, driven by lower than expected claim severity. As Dino mentioned, this is the fourth year in a row the result of the claim review was favorable. On slide 13 of our earnings presentation, you can see the results of our gross premium valuation analysis. The most significant change in the analysis was the discount rates.
Given current investment yields, we reduced our near term expectation for new money yields, in addition to lowering our expectation of normalized new money yields for 2025 and beyond. Last year's analysis assumed the 10-year Treasury yield would get to a normalized level of 4.25%. This year, we reduced that expectation to 3.75%. This 50 basis points drop in new money yield expectation reduced the all-in discount rate to 5.50% on a nominal basis and 5.76% on a tax equivalent basis. This was the driver of the $280 million reduction in GAAP margin due to discount rates. Consistent with the results of the claim reserve review, morbidity provided a $32 million favorable change to GAAP margin. Moving to persistency, there are two dynamics to note.
First, margin improvement was reduced by $166 million over the course of the year, driven by policyholder response to rate actions, which has resulted in our earning a portion of last year's margin. In other words, benefit reductions and lapses over the past year at a current earnings contribution and reduced future expected earnings, which is what margin is. Of course, the reduction of active policies also reduces the risk over the long term. In addition to this dynamic, mortality rates for policyholders not on claim have been slightly lower than expected, which we reacted to in this analysis. Lowering the expected mortality going forward was the primary driver in the remaining $68 million reduction to GAAP margin from persistency.
Finally, regarding future premium rate increases, as we've previously discussed, we only include rate increases that have been filed and not yet approved, or that we plan to file as part of a current rate increase program. Over the past year, we've outperformed our previous rate increase assumptions and are continuing to file for additional rate increases, which combine to add $58 million to our GAAP margin. Our best estimate assumptions currently reflect $230 million of future unapproved rate increases. While we limit the amount of unapproved rate increases anticipated in our reserves, we will continue to seek rate increases over time if and when they are justified. Summarizing our gross premium valuation analysis. Every year, there is movement in each of these variables driving the margin.
In this year's analysis, the low interest rate environment drove the GAAP margin below zero, causing an unlocking of the active life reserves and a charge to earnings. Before moving on, I'd like to give you a few important statistics that I think lend credibility to our assumption setting process. It's been nearly four years since our assumptions were last unlocked. Over the course of that time, our active policy count is 5% lower now than we expected it would be. Our open claim count is slightly lower than expected, and the total dollar amount of paid claims is 2% lower than expected. Three broad and important metrics, how many policies remain, the number of open claims, and the total amount paid over the last four years, were all better than the assumptions set in 2015.
On slide 15 of our earnings presentation, we've updated the data that we first provided during last year's third quarter earnings call. The active lives in both the individual and group blocks continue to decline, down 21% and 33% respectively since 2015. On the bottom left of slide 15, you'll note open claim counts in our individual block have been fairly steady in recent years. We believe open claims on the individual block have essentially plateaued, another indication of the maturity of this block. To conclude on long-term care, slide 16 shows the key characteristics of our long-term care blocks. You'll note that the average age of our individual block is 79 years old, and the average age of a new claimant is 84. Again, indicating that this block, which accounts for 85% of our reserves, is very mature.
While the group block is less mature, with an average attained age of 65. It has a lower level of benefits. For example, there are very few lifetime benefit policies, and only 15% have inflation protection. Overall, our block is mature, well-managed, and we continue to have confidence in our long-term care reserves. Our corporate segment produced a core loss of $17 million in the third quarter. Pre-tax net investment income was $487 million, the same amount as the prior year quarter. Our limited partnership and common equity portfolios produced pre-tax income of $18 million, a 0.9% return, and a result that is roughly half of our quarterly average. Pre-tax income from our fixed income portfolio was $462 million, slightly higher than the prior year quarter. The pre-tax effective yield on the fixed income portfolio was 4.8%, in line with prior periods.
Fixed income assets that support our P&C liabilities had an effective duration of 4.1 years at quarter end, in line with portfolio targets. The effective duration of the fixed income assets that support our life in group liabilities was nine years at quarter end. Our balance sheet continues to be extremely strong. At quarter end, shareholders' equity was $12.1 billion or $44.66 per share, and our unrealized gain position increased to $4.2 billion due to the decline in interest rates. Shareholders' equity, excluding accumulated other comprehensive income, was $12.0 billion or $44.14 per share, an increase of 6% from year-end 2018 when adjusted for the $3.05 of dividends per share paid during the first three quarters of this year. In the third quarter, operating cash flow was $466 million.
We continue to maintain a very conservative capital structure, and all of our capital adequacy metrics, as well as credit metrics, are well above their internal targets and current ratings. Finally, we're pleased to announce our quarterly dividend of $0.35 per share. With that, I'll turn it back to Dino.
Thanks, James. Before we move on to the question and answer portion of the call, let me leave you with some overarching thoughts on the quarter. Our actions on long-term care reflect our continued prudent management of this portfolio. Our underlying P&C loss ratio was 61.7% for the quarter and 61.1% year to date, while the expense ratio improved to 32.5%. U.S. net written premium grew 9%. We achieved 6 points of rate in the quarter, 2 points higher than the second quarter. Based on the current momentum, we believe rate increases will persist above our long-run loss cost trends throughout 2020. With that, we'd be glad to take your questions.
Thank you. If you'd like to ask a question, please signal by pressing star one on your telephone keypad. If you're using a speakerphone, please make sure your mute function is turned off to allow your signal to reach our equipment. Once again, that is star one if you'd like to ask a question. I'll pause for just a moment. We'll take our first question from Jay Cohen with Bank of America.
Yes, thank you. A couple questions. I guess, first on the international side, when would you expect the earned premium to kind of fully reflect the actions you took to get out of certain lines of business? When will that stuff be kind of off your books?
Well, we started the non-renewal, Jay, Dino, sort of late 2018. You non-renew them as their renewal dates come out. That plays out through the course of the year. As we indicated in some of the prepared remarks, it's a dynamic process, and there's some additional business that we're non-renewing, which is normal for the process when we see something and we don't think we're going to get the right terms and conditions, we're going to non-renew it. We think the non-renewal of accounts is going to take through sort of the end of the second quarter of next year, and that obviously has to earn itself out, which is probably sometime through 2021. Just keep in mind, we mentioned it before, there are certain lines we exited, like political risk and large project construction engineering risks that have a multi-year tail.
Some of those, Jay, effectively can stay on the portfolio. I think you're looking at 2021.
Yeah. No, that makes sense. Okay. The second question, I think I know the answer to this, but just to double-check. In the accident year loss ratio, was there any material current year catch-up? Did you reassess the first half result in any segment that might have influenced the current quarter-reported accident year loss ratio?
No. Jay, when you look at the accident year commercial underlying loss ratio, and you're comparing it Q3 to Q3, keep in mind that in Q4 of last year, we increased the umbrella, and we also increased the property. That's obviously played itself out because it was in the Q4. You're seeing it elevated against Q3, which was the quarter before we made those changes.
We made a slight tweak in specialty, Jay, just around aging services as Dino was reflecting earlier.
Got it. No, that makes sense. That's great. Thanks for pointing that out.
Thank you. We'll now take our next question from Joshua Shanker with Deutsche Bank.
Good morning, everybody.
Morning.
Morning.
I wanted to go back to some of Dino's prepared remarks. First of all, on the 2.5% long-term loss cost for an estimation, if I look over the past decade, obviously, been a decade of lower than long-term loss cost trend, how does that 2.5% stack up against history, I guess?
Pretty consistently. Keep in mind, I think we went through it, or James did on the last call, keeping in mind because overall, it can seem low at 2.5%, but we have a large portfolio of the professional E&O affinity portfolio that has run probably over the full decade, a long-run loss cost trend that was slightly under 1%. It might be a little slightly higher than history because of more comp, which has been particularly lower in the recent years. You put those two things together and you got some pretty consistent long-run loss cost trends.
Is the last decade a good decade to use as the base of assumption?
I guess, Josh, I would say the last decade is the best decade that we have to use for our assumptions. Going back further than that, the market was quite a bit different than it is today.
Was our book.
Yeah.
That's more the part of the issue.
I think, we're going to consistently look at this every quarter to see if there's anything that's changing, and as Dino reflected in his remarks, we're going to tweak specific line items and specific lines of business as we see them change. Certainly, the trend has been holding for quite some time.
That was the other part of my question. You talk about the fact that the negative pricing over the past few years had not resulted in deterioration because of a combination of the exposure acting like rate, and I guess the re-manicuring or I don't know what you want to call it, the changing of the mixing your portfolio. You said going forward, those things will still be an issue, so you won't have a dollar-for-dollar impact. Are you saying that you expect the gap between rate and loss cost trend will be reflected by even lower loss ratios because then you'll layer on top of that business mix and exposure? I was just trying to understand the dollar for dollar. You're going to get more than your bang for the buck.
Yeah
with these rate increases?
The comment on the go forward was a comment on symmetry that it hadn't necessarily been, and loss cost trends are going to be what they're going to be actual. You might not end up with exactly the dollar for dollar. The point being that often on these calls, there's a lot of conversation of what I consider to be the oversimplified correlation of rate and long run loss cost trends, and I was just simply suggesting that there's so many vectors that influence it, including judicial, regulatory, and you play all of that out. We're going to be cautious in how we move the margin. It is in that vein that I was referring to it.
Look, long-term, you're always improving the portfolio, and you're always hopefully getting exposure increases that act like rate.
Sure.
Over the long run, should you always expect to have better margin than the rate over loss cost trend would indicate?
I'm not sure if I'm following that exactly, Josh.
Josh, I don't think that we would expect that per se. I think, and just going back to the first part of the question, I think part of what we were trying to get across in that messaging on the dollar-for-dollar margin is that with all the moving pieces, we're actually going to be cautious.
Yeah.
Just as Dino said. We're not going to take margin improvement as soon as we see sustained earned rate above our loss cost trend. It's actually likely to play out slower in terms of margin improvement rather than faster.
Okay. That answers my question. Thank you very much.
Sure.
Thank you. We'll now hear from Gary Ransom with Dowling & Partners.
Yes, good morning. I wanted to dig in on some of the loss cost trends also. You mentioned in healthcare the number is something like 11%. We've heard a lot of anecdotes about how aggressive the plaintiff's bar has been. I wonder if you have any thoughts specifically about how some of the external effects like litigation funding and medical financing companies may have been at least a partial driver of those trends. That's all part of a question of you think it's 11 today, but maybe it's 15. That's sort of what I'm kind of driving at. What gives us comfort that we're kind of in the right place on these loss trends?
Gary, Dino, I'll start, and James may want to jump in also. We're not really seeing in the portfolio the effects of legal funding. You do see its impact from a few different areas. First of all, the plaintiffs' bar's really targeted this industry. You can see it in the ad campaigns and the marketing. Inviting, if you will, more claimants to come forward, and that is happening. Keep in mind, Gary, we've been at it for over two decades, so we can see the difference.
Also, there have been some larger jury awards, and although a lot of these cases never make it to court, what it does is embolden the plaintiffs' bar, based on what they've seen in some of these jury awards, not to settle upfront for what was potentially a lower amount for a similar type case in the past. Intriguingly, it also affects the adjusters and the defense attorneys who are also going to incorporate, if you will, the higher verdicts into their settlement calculus. As I indicated, we're seeing a little bit of less frequency, but we're going to wait. Some of that is a function of obviously all the re-underwriting that we have done. Then, we'll watch to see how the longer and loss cost trend and whether that sustains itself, the slight improvement we're seeing on frequency.
Keep in mind, you've got some very significant rate increases, which eventually will be also sort of factored in. We feel good about the actions we are taking, how we're leading the market. It's a combination of underwriting actions, deductibles, wording changes, tightening wording, and then also getting a lot of rate. You play this forward and quarter for quarter, we take a look at this thing. As I tried to suggest, we've acted consistently, and we will act up and down as this moves forward. I'm not really sure what else I could add to give you even more clarity.
No, that's helpful. I realize it's mostly anecdotal. Another question on the legacy reserve charge. I'm not sure I understood, James, what you were saying about the reinsurance recoverable that you had somehow taken down. Did I get that right?
That's right, Gary. As part of the reserve review, not only were we looking at our open claim inventory, but we were also looking at the ceded recoverables that we had on older claims. It turned out we had just overestimated those ceded recoverables on some of the older claims. That's what came through.
Was there anything consistent about these claims? Were they from some class of business that was similar, or were they just scattered random claims?
No. The review was really old product liability cases with multi-claimants involved, things like public nuisance and food additives, which is really what's in that bucket of claims. The reserve review focused around, as I said, both the claim reserves as well as the recoverables.
Okay. Is there any-- I can't remember if you do an A&E charge as well at this point that might affect the accounting with-
No. We will do our annual asbestos and environmental review next quarter.
Next quarter. Okay. Thank you.
From the first quarter.
All right. Thank you very much.
Thanks.
Thank you. Once again, as a reminder, that is star one if you'd like to ask a question. We'll hear now from Meyer Shields with KBW.
Great. Thank you very much. Good morning, and thank you for all of the detail you've provided on loss trends. That's been very helpful. Can you walk us through what you're seeing now in terms of rates and loss trends for workers' compensation?
I would say it's been pretty steady. The rate in workers' comp continues to be kind of mid-single digits negative. It's slightly better in the third quarter, but not materially. When we look at trends, severity trends continue to be benign and stable, as we mentioned last quarter. Frequency has flattened compared to what it was doing the last several years, which again, is the same as it was last quarter. No real change quarter-over-quarter.
Okay, perfect. I was hoping you could sort of outline the exposure that the affinity book has to these worsening trends. It sounds like you're not seeing anything deteriorate, but is there exposure if the trial bar settles on sort of this E&O pocket?
Our affinity book really is not I mean, the kinds of risks that are inside there are much smaller. They tend to be pretty homogeneous accounts, but many small accounts, and don't fall into the types of coverages where we've seen the plaintiffs' bar attack.
Also, our risk control, I mean, it's five, six decades, as we've indicated before, and all of that makes a difference. Also, it doesn't have, because of the professional E&O, the medical cost exposures that you would see on comp, some of the auto-related umbrella. That makes a difference given the medical undertones in some of the others, Meyer.
Okay, perfect. Thank you so much.
Thank you. That does conclude today's question and answer session. I'd like to turn the conference back over to Mr. Robusto for any additional closing remarks.
Great. Thank you very much for attending, and thanks for your questions.
Thank you. That does conclude today's conference. Thank you all for your participation. You may now disconnect.