CNA Financial Corp

NYSE:CNA

Earnings Call Transcript 2018-Q3

[Operator]

Good day morning and welcome to CNA's discussion of its 2018 Third Quarter Financial Results. CNA's third quarter earnings release presentation and financial supplement were released this morning and are available via its Web site, 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 lines for question.

Today's call may include forwarding-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 the earnings release and is in CNA's most recent 10-K on file with the SEC.

In addition, the forward-looking statements speak only as of today, Monday, November 5, 2018. CNA expressively disclaims any obligation to update or revise any forward-looking statements made during this call. Regarding 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 Web site. With that, I will turn the call over to CNA's Chairman and CEO, Dino Robusto. Please go ahead, Sir.

[Dino Robusto, Chairman and CEO]

Good morning everyone. I am pleased to share our third quarter results with you today, which reflect our ongoing underwriting improvements. Core income for the third quarter was $317 million or $1.17 per share. Our best result in eight years and our core return on equity was 10.5%. This brings our year-to-date core earnings per share to $3.19 and year-to-date core return on equity to 9.5%.

We had another good combined ratio which was 94.2% for the quarter. Over the past seven quarters, I have described the goal of growing underwriting profits by institutionalizing and enduring expert underwriting culture here at CNA, which would enable CNA to be a top quartile performer on a sustained basis.

Key ingredients I have consistently updated you on are stronger talent and governance, building feedback loops that institutionalize our collective expertise across the value chain dramatically elevating our engagement with agents and brokers, making additional investments in technology and analytics. And every importantly, doing all of this while embedding a disciplined expense management culture throughout the company.

Our journey towards this goal is progressing well. Our expense ratio has improved almost 2 points in the last two years even as we continue to invest in talent and technology and analytics. Our underlying loss ratio was 61.1% for the third quarter and 60.8% for the first nine months. Although we view this loss ratio as stop quartile, our objective is to sustain strong performance over the long term by focusing on consistently improving areas that can generate better results.

Indeed, our total P&C underlying loss ratio of 61.1% includes an underlying loss ratio of 66.3% for International. I mentioned last quarter that we have been re-underwriting the portfolio of our Lloyd's operation and we are exiting underperforming segments to focus our efforts on writing more of our new business and our longstanding profitable target markets such as Healthcare Professional E&O.

Although our actions will decrease our writings in our Lloyd's operation, we will of course continue to work to grow our international business in areas with track records of profitability within Canada, Europe, and the U.K. However, the growth in other areas of International will likely not affect the runoff of the unprofitable business in our Lloyd's indicated in the near term.

P&C gross written premium, excluding our large warranty captive, grew 3% in the third quarter, but as a result of additional reinsurance, purchase [technical difficulty] reduced potential volatility in some property areas and in some smaller segments that we are targeting to grow. Our net written premium was down 1% in the quarter.

Notwithstanding, the lower net written premium in the third quarter for the full nine months of 2018, our net written premium growth is a healthy 5%, a function of strong execution across all of our production metrics. Specifically in the quarter, rate for P&C operations was +2%, up one point from the second quarter. It was driven by Commercial which generated 2 points of rate compared with +1 in the second quarter as property, liability, and worker's comp all improved one point in the quarter.

Our commercial rate excluding work comp for the quarter was +3%. Rate increases in Specialty and International were similar to the second quarter results of 2% and 3% respectively. Written renewal premium change in the quarter for P&C was +4% and +3% on an earned on basis which slightly exceeds our long-run loss cost trends. Retention was 82% for the third quarter, down a point from the second quarter, which was primarily a function of the re-underwriting in our Lloyd's operation or as retention for our Specialty and Commercial U.S. operations were each at 84%.

New business in the quarter was up 8% from a year ago as our disciplined effort to strengthen relationships with our best brokers and agents continue to pay benefits. Four the third quarter, our next pretax catastrophe losses were modest at $46 million, approximately, 2.5 points on a loss ratio with Hurricane Florence losses at $35 million. Before I hand it over to James, I want to comment on long-term care which we will be devoting a good part of our prepared remarks today.

You will have seen in the materials that we released this morning that we completed our annual growth premium valuation analysis as well as our long-term care claim reserve review in the third quarter this year. Given the increased investor focus on long-term care in 2018, we thought it was prudent to address our position now rather than waiting until the fourth quarter.

Going forward, you should expect these annual studies to be completed in the third quarter. James will provide more detail on the results of the GPV analysis. And I believe that the outcome as well as the additional disclosure provided, reflects the conservatism in our reserving assumptions as well as the active management of book which continues to give up confidence in our reserve position.

And finally, we are pleased to announce our regular quarterly dividend of $0.35 per share. And with that, here is James.

[James Anderson, CFO]

Good morning, everyone. Our Property & Casualty operation produced core income of $305 million, up 83% from the prior year quarter. Pretax underwriting profit of $100 million was consistent with recent quarters driven by a steady underlying combined ratio and $60 million of favorable loss reserve development. Our expense ratio improved to 33.3% and is in line with our current run rate.

Moving to each of our P&C segments, Specialty's combined ratio was 87% for the quarter including 7.7 points of favorable development driven by surety as well as management liability and financial institutions. You recall that beginning in the first quarter of this year, we began reviewing our surety reserves more frequently than once per year during the third quarter.

As a result, our prior period development in Specialty is spread throughout 2018 rather than being concentrated in the third quarter. Year-to-date, Specialty has generated a little more than 6 points of favorable reserve development versus 2017 year-to-date of 7 points. Specialty's underlying combined ratio in the third quarter was 92.3. Year-to-date, Specialty's overall combined ratio is 87.1.

Our Commercial segment's combined ratio in the third quarter was 97.4%. This result included 3 points of catastrophe losses, a good result in a quarter that historically had significant cat activity. Commercial third quarter underlying combined ratio was 94.3%. Year-to-date, Commercial's all-in combined ratio is 97%. Our International segment generated a combined ratio of 103.9% in the third quarter driven by property losses in Harvey.

International's year-to-date combined ratio is 101.8%. Our Life & Group segment produced $32 million of income this quarter. This marks 11 straight quarters of stable results since unlocking that occurred in the fourth quarter of 2015. Long-term care morbidity experience continued to be consistent with our reserve assumption. Persistency was also favorable.

The third quarter results also include the impact of our annual claim reserve review. Our review historically has been completed in the fourth quarter. The results of the claim review indicated favorable morbidity, specifically severity outcomes continuing to be better than expected. The outcome of the claim review added $24 million after tax to the results.

Now as Dino indicated, I am going to spend a few minutes providing additional detail about our long-term care business and will be referencing Slide 13 through 18 in our earnings presentation. At first a few fundamentals regarding the operations, CNA stopped writing new individual policies in 2003 and stopped excepting new group's around the same time.

As of the end of the quarter, the book had approximately 161,000 individual active lives and 164,000 group active lives. This is a mature book that has seen over 100,000 claims. A significant number compared with today's number of active lives, providing with us with what we believe is very credible claim experience. Our GAAP reserves of $11.8 billion have $182 million of embedded margin, a level down slightly from last year which I'll discuss in a moment.

While our statutory reserves are $13.7 billion and include nearly $2 billion of a margin, this statutory margin provides significant cushion to our capital position. We have very actively managed this business particularly over the last five years where we had significantly increased our actuarial resources, developed in-depth analytics, actively pursued rate increases and invested heavily in our claim management and related processes. This proactive approach across all fundamental aspects of the business provides what we believe is a firm basis for the confidence we have in our reserve position, recognizing that we must remain vigilant regarding changes in trends.

On slide 14 of our earnings presentation you can see the individual block is shrinking, down 15% since 2015, through normal mortality as well as some policyholders choosing to lapse their policies rather than pay for rate increases. The active lives in our Group block have declined 29% since the end of 2015, coinciding with when we stopped allowing groups to add new members, and with many policyholders subsequently choosing to convert to a limited benefit policy. In other words, lapsing their policy but still retaining a small benefit as a result of subsequent rate increases and other factors.

Moving to our recent claim trends at the bottom part of slide 14, you will note open claim counts in our individual block have been fairly steady in recent years. Because the average age of our individual block is approaching the average age of a new claimant, we believe open claims will reach peak levels within the next year or so, another indication of the maturity of the block. As I indicated earlier, our Group block has a lower average age, and therefore we expect group claims to continue to increase as the block matures.

It is important to remember that while these policies were sold on a group basis, the policyholders themselves are individuals, with claims that we expect to behave the way our individual block experience has. Through our expansive analytics and insights into the 100,000 individual claims, will inform claims expectations on our Group block as well.

Slide 15 shows the key characteristics of our long-term care blocks. You'll note that the average age of the individual long-term care block is 79 years old, and the average age of the new claimant is 84, again indicating that this block, which accounts for 90% of our reserves, is very mature. While the Group block is less mature with an average age of only 64, it also has a lower average level of benefits. For example, there are very few lifetime benefit policies, and only 15% have inflation protection. On slide 16 and 17, you can see the results of our gross premium valuation analysis, as well as our key GPV reserving assumptions, which should give you a better sense of the reserving position.

Beginning with morbidity improvement, we now have no future morbidity improvement embedded in our best estimate assumptions. Prior to this year, we had been assuming 1% improvement annually until 2021. While we have seen favorable morbidity in recent years, future uncertainty justifies a more prudent approach, and thus we removed the remaining improvement from our best estimate. This change alone reduced our GAAP margin by $258 million, which in addition to other changes in morbidity assumption based on experience in 2018 resulted in an overall $213 million reduction in margin for morbidity.

Within the category of persistency, we also changed our mortality assumption in a more conservative direction this year. In prior years, we had aligned morbidity and mortality improvement through 2021, after which time we assumed no improvement. This year, in addition to removing morbidity improvement, we extended mortality improvement out to 2024, which reduced GAAP margin by $86 million. We believe our revised assumptions on morbidity and persistency reflect a measured prudent approach that appropriately reflects the relative uncertainty of such assumptions.

The discount rate used in our reserves did not change much overall, staying just under 6% on a tax equivalent basis based on a nominal yield of just under 5.7%. But it improved in near-term years based on higher yield, in 2018, and a shift in our new money allocation away from tax exempt municipals toward investment-grade corporate bonds post-tax reform. In the out years, we reduced our 10-year treasury yield assumption to 425 in 2025, a level we then hold into the future. The combination of these discount rate changes added $17 million to the GAAP margin.

Finally, regarding future premium rate increases, as we have previously discussed, we only include rate increases that have either been filed and not yet approved or that we plan to file as part of the current rate increase program. In 2018, we have outperformed our initial rate increase assumptions and are continuing to file for additional rate increases, which combine to add $178 million to our GAAP margin. Future and improved rate increases now account for a total of $200 million in our best estimate assumptions. But while we limit the amount of unapproved rate increases in our reserves, make no mistake; we intend to seek rate increases over time if and when they are justified.

In addition to rate increases, we have given policyholders options to reduce their benefits in lieu of a rate increase, which in the past 40% of policyholders have chosen to do when given the option, or even to convert to a limited benefit policy which allows them to stop paying premium and keep a modest amount of benefit, an option that many Group policyholders have elected in recent years. Both of these policyholders' elections meaningfully reduce our exposure.

So, overall, after revising our morbidity and mortality assumptions and gaining benefit from achieving higher rate increases than anticipated, the GAAP margin decreased from $246 million to $182 million. To conclude on long-term care, 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 $20 million in the third quarter. Pretax net investment income was $487 million in the third quarter, compared with $509 million in the prior year quarter. The changes were driven by limited partnership portfolio which produced $23 million of pretax income, a 0.9% return, compared with $51 million last year. Pretax income from our fixed income security portfolio was $459 million this quarter, which was essentially flat with the prior year quarter. The pretax effective yield on the fixed income portfolio was 4.7% in the quarter, consistent with prior periods.

Fixed income assets that support our P&C liabilities had an effective duration of 4.5 years at quarter end, in line with portfolio targets. The effective duration of the fixed income assets that support our long-duration Life & Group liabilities was 8.2 years at quarter end. Our balance sheet continues to be extremely strong. At September 30, 2018, shareholders' equity was $11.5 billion or $42.41 per share, and shareholders' equity excluding accumulated other comprehensive income was $12.3 billion or $45.20 per share, an increase of 7% from year-end 2017 when adjusted for $2.95 of dividends per share paid so far this year. Our investment portfolio's net unrealized gain was $1.6 billion at quarter end.

In the third quarter, operating cash flow was $514 million. We continue to maintain a very conservative capital structure. All of our capital adequacy and credit metrics are well above their internal targets and current ratings.

With that, I'll turn it back to Dino.

[Dino Robusto, Chairman and CEO]

Thanks, James. Before we move to the quarter-and-answer portion of the call, let me briefly leave you with some summary thoughts on our performance. Core income for the third quarter was $317 million, our best results in eight years. We had pretax underlying underwriting income of $92 million, giving us $282 million for the first nine months of the year, up 33% over the same period last year. Our underlying P&C loss ratio was 61.1% for the quarter and 60.8% year-to-date. Total P&C written renewal premium change was plus 4% for the quarter, with the rate up one point from the second quarter at plus 2%.

Long-term care produced $32 million of core income for the third quarter driven by the favorable impact from our annual long-term care claim reserve review. We completed our annual GPV for long-tear care, and there is no unlocking even as we added conservatism to our morbidity and mortality reserving assumptions. Our 2018 third quarter core return on equity is 10.5%, and net income return on equity is 11.7%, and we announced our regular quarterly dividend of $0.35 per share.

With that, we'll be glad to take your questions.

[Operator]

Thank you. [Operator Instructions] And we'll take our first question from Josh Shanker with Deutsche Bank.

[Josh Shanker, Deutsche Bank]

Good morning everybody.

[Dino Robusto, Chairman and CEO]

Morning, Josh.

[Josh Shanker, Deutsche Bank]

Morning. My question, of course, long-term care, just trying to understand a frame of reference for understanding, maybe in layman's terms, the backward-looking morbidity improvement that was posted in 4Q '17 and the cancellation of the assumption on bidding improvement going forward in this quarter. Can you give a frame of reference about how powerful those two things were? Obviously one had about five times as much impact on your reserves, or I should say five times as much impact on changing the reserve amount than this recent charge. I'm trying to understand is there a percentage of a number of lives or a duration of claim site incident that we can understand in terms of that?

[James Anderson, CFO]

Yes, Josh, this is James. Let me try to give you a little bit of perspective there. So if you think about the morbidity improvement component that we took out this quarter, it was really three years' worth of 1% morbidity improvement that we expected to happen. So a very kind of narrow slice of improvement that was left in that assumption, versus what we did at the end of 2017 was we were reevaluating our overall morbidity assumptions coming out of the 2015 unlocking. And if you recall, last quarter, I talked a little bit about the fact that what we were seeing in 2016 and 2017 in morbidity was much better than what we had assumed at the end of 2015.

And so we reevaluated and readjusted the overall morbidity assumption as the end of last year, and that effect will persist for a long period of time versus the change we made this quarter, which was to take it out for the next three years.

[Josh Shanker, Deutsche Bank]

Well, if you've seen a steady improvement from '14 - Yes, there's just a few follow-ups. If you see an improvement from '14 to '15 to last year, what motivated you to take out the improvement going forward, it seems like you see a trend in there, why would you change that assumption going forward.

[James Anderson, CFO]

I guess I would split it out a little bit. So the morbidity improvement assumption I think of as kind of a baseline assumption that we put in place at the end of 2012, and we were assuming for the following 10 years that we were going to see this underlying improvement in morbidity, versus what we saw all-in in morbidity at the end of 2017 was a stark difference from what we had assumed in 2015. So think about morbidity improvement as a small baseline trend that we were assuming versus all of the other factors that go into morbidity that were being adjusted at the end of 2017.

[Josh Shanker, Deutsche Bank]

All right, I'll try and digest that. And on the P&C side, between Specialty and Commercial, you're approaching or added 2% renewal rate pricing comparison to the year-ago period. How does that compare with what you think the lost cost trend are? And I know people sometimes say that exposure equals rate or whatnot. Are we losing margin in writing business today?

[Dino Robusto, Chairman and CEO]

Okay, so Josh, Dino. Let me give you a couple of [technical difficulty] lost cost trends, about 2.5% for our P&C portfolio overall. Within it we have about 2.5% trend on medical inflation, largely benign. Cost of goods inflation first-party and auto, stable at about 2% to 3%. In terms of legal trends, we've talked about in the past we've seen some higher verdicts, healthcare hospital and ageing services, as well as some casualty lines. So the earned renewal premium change, which is about 3%, is slightly higher than lost trends, so that's good. But obviously you need to have that continue and persist. I think it's safe to assume it's not going to persist indefinitely, the underwriters are keenly aware of the dynamic.

So we have to continue to push for rates across our portfolio, not only to mitigate what might be margin compression if we can sustain it, or if the lost cost trends that have been relatively benign pick up. But also it's to a certain extent a catch-up, right. If you think about the 10 to 12 quarters, starting in 2015, we had negative rates in each of those quarters and lost cost trends were positive. So I think in general I'd say there's still an awareness in the marketplace across the distribution network, across insured, that rate is needed by the insurance companies, and that's what we're going to continue to do.

At this particular juncture, if you use the earned renewal premium change, as I said, slightly above, and we'll see how that plays out. We'll just keep pushing.

[Josh Shanker, Deutsche Bank]

Okay. Well, thank you for all the clarity.

[Operator]

Thank you. We'll now move on to our next question from Bob Glasspiegel with Janney Montgomery Scott.

[Bob Glasspiegel, Janney Montgomery Scott]

Good morning, CNA. A question on what sort of outside actuarial review or signoff did you have this quarter on the long-term care?

[Dino Robusto, Chairman and CEO]

Bob, we did not do a third-party review this quarter. The last one was did was at year-end of last year.

[Bob Glasspiegel, Janney Montgomery Scott]

Okay. What's the regular pace that you do the outside actuarial reviews, remind me?

[Dino Robusto, Chairman and CEO]

There's really not a regular pace. We do it when - just basically when we feel like we need a second opinion. So we have done it, as I mentioned, in end of last year, we probably did it twice in the two or three years before that. So it's not quite annually, but it's probably more than every other year.

[Bob Glasspiegel, Janney Montgomery Scott]

To what extent do you use outside actuaries to work through? I mean a lot of your competitors are doing the same process, and maybe there's more industry numbers to look at than just your own company as well to augment it.

[Dino Robusto, Chairman and CEO]

I'm sorry, your question was?

[Bob Glasspiegel, Janney Montgomery Scott]

Do you have any outside actuaries advising you at all on this, I guess is the question.

[Dino Robusto, Chairman and CEO]

We do, but I would say it's not on a regular basis. It's when we feel like we want a second opinion on the assumptions that we're using. So it's, like I said, it's more frequently than every other year, but not annually.

[Bob Glasspiegel, Janney Montgomery Scott]

Got it. My thought process was that a lot of your competitors are looking at this too, and you have a lot of data obviously to look at, and your own data is the most useful for sure. But I was curious to the extent to which you're gauging how your assumptions compare with your competitors. And as you know, it's hard to get at this improving versus baseline because you need to know what the baseline in the original assumptions are, so it's really hard to get sort of - for us from the outside to get a sense of how your book is performing relative to others. And I was just curious how you'd try to do that.

[Dino Robusto, Chairman and CEO]

I think that's fair comment, Bob. I guess what I would say with our particular book, given that we unlocked our GAAP assumptions in 2015, I think our periodic or quarterly results are going to be the easiest way for you to see how we're performing against our expectations or assumptions that were set in 2015. So we continue to talk about the fact that we've had 11 quarters of stable results. Basically what you should take from that is that means our actuals are performing at expectations.

[Bob Glasspiegel, Janney Montgomery Scott]

No, I got it. It's obviously - and it's encouraging that you're encouraged by the trends to the extent to which we can see how your trends are consistent with how others are interpreting it. It's just a big challenge to us. I want to switch gears here on a little bit of a meltdown in Q4 in some of the macro investment vehicles. And Michael, anything that you could comment on what you're seeing to date in Q4, if you survive those two shocks?

[Dino Robusto, Chairman and CEO]

Look, with regard to Michael, we don't expect that to be major catastrophe for us. We do expect that it will be larger than Florence, so that's little bit what we are seeing early on hearing from Michael.

In terms of investment performance, clearly October was not a great month for the stock market. We would expect that's going to have an impact on our LP portfolio, but there is still two more months remaining in the investment markets that we are going to have to see play out. So it's certainly too early to see how Q4 may look from an investment standpoint.

[Bob Glasspiegel, Janney Montgomery Scott]

Last question, you remind me you are not exposed to leverage loans in the partnership portfolio, were you to material extent?

[Dino Robusto, Chairman and CEO]

Not too, the majority of our LP portfolio, 70% of it is hedge funds, which is primarily equity-related. We do have - the rest is private equity, some of would be credit-related, but not a lot of exposure to leverage loan.

[Bob Glasspiegel, Janney Montgomery Scott]

Thank you.

[Operator]

Thank you. And we will move on to our next question from Jay Cohen with Bank of America Merrill Lynch.

[Jay Cohen, Bank of America Merrill Lynch]

Thank you. A couple on long-term care, the first is, you obviously described the book is having favorable characteristics, this review gave you even more confidence; how realistic is it to have some sort of risk transfer here, given the favorability of the book that you described?

[James Anderson, CFO]

Hi, Jay, that's a great question. I think that is going to continue to be an evolving story. I think as interest rates continue - hopefully they continue to increase, we may see more asset manager like folks trying to find ways to accumulate long duration assets, which clearly long-term care fall into the that category. And so, that risk transfer market, if you will, will likely continue to develop. It's certainly right now I would call it at an infancy, we've see a couple of transactions, but I would not call it a robust market at this point. So, we will continue to look at it, but most importantly, we are going to try and manage this book, so that should those opportunities come around, we are going to be you know, positioned such that we are going to be the belle of the ball.

[Jay Cohen, Bank of America Merrill Lynch]

Got it. And then secondly, on a stat side, is it fair to say you're - maybe I will just ask the question, what are the assumptions you have for things like morbidity, mortality in your stat reserves versus your GAAP reserves?

[James Anderson, CFO]

Our stat reserves have never been unlocked. So they include no improvements for morbidity or mortality.

[Jay Cohen, Bank of America Merrill Lynch]

Got it. That's the difference in the cushion there I assume?

[James Anderson, CFO]

I think that's part of the difference, but really the biggest difference is the discount rate case. So, on a statutory basis, we don't set our own discount rates. The discount rates are regulatory prescribed, and the discount rate is - it's about 200 basis points lower on the stat side than it is on the GAAP side, even though it's the same asset base and they're producing the same income.

[Jay Cohen, Bank of America Merrill Lynch]

Got it. That's helpful. Thank you.

[James Anderson, CFO]

Sure.

[Operator]

Thank you. We will take our next question from Meyer Shields with KBW.

[Meyer Shields, KBW]

Thanks. So, one quick one on long-term care, because I just want to make sure I understand it. So is it fair to assume that your current morbidity expectation for 2019 is the same as 2018, but it's lower than it was before the year-end '17 review?

[James Anderson, CFO]

I think when you look at it from a morbidity improvement standpoint, Meyer, that's exactly right.

[Meyer Shields, KBW]

Okay, thanks. I was just trying to ask two things, so two really small questions; one, the acquisition expenses ratio have been climbing a little bit in specialty and commercial, is there a noise or is there any underlying increase in acquisition costs?

[James Anderson, CFO]

I mean it's a little bit of both. That is going to move around a bit. We do have some - I would say, a little bit of noise this particular quarter, but partly we do pay some contingence, and it's based around our performance. So, as our loss ratios have improved, we paid out a little bit more to our agents and brokers.

[Meyer Shields, KBW]

Okay. Is that improvement evaluated on a quarterly basis?

[James Anderson, CFO]

It is, yes.

[Meyer Shields, KBW]

Okay. And then, Dino, can you talk a little bit about what you're actually seeing in terms of - I guess, workers' compensation, and general liability, where we are hearing some chatter about emerging differences in loss trend compared to what we'd seen in the last couple of years?

[Dino Robusto, Chairman and CEO]

Yes, fair question. So I make these observations about work comp, but we've been seeing negative sort of mid single-digit frequency trends over the past several quarters, which is less negative than a year ago. Now, while we've seen some pockets, where frequency has increased, the negative frequency trend overall is still favorable to our long run trend assumptions, because we did not lower our long run frequency assumptions despite the actual frequency consistently more negative than our assumption. Severity, which shows a little bit more fluctuation quarter-to-quarter than frequency as you would expect, as over the past 18 to 24 months been slightly below, our long run severity assumption.

So work comp continues to perform better than expected, which is you know, given us that headwind in pricing, our rates in the third quarter was minus 3%. Look, I mean at the end of the day, we feel good about where we are with comp, you know, we closely monitor this line of business, and I think we were comfortable with the level of analytics we have to sort of react to changes in severity and frequency early and accordingly, but we are conservative in how we react to those things.

The trend we've talked about that we were explicit about because we have a meaningful portfolio of healthcare business, was that we clearly - for over a year now have been seeing some larger verdicts that have been impacting hospital and ageing services, E&O. So what we have been trying to do is improve our terms and conditions. We've been pushing and we've been achieving some significant rate increases, we've been raising deductibles, and when we haven't been successful in doing that, we walked away from accounts; we stock on prior quarters and showed you some of the rate and retention results. I think today what we say, if you were to just take a look at that book, which is really where we've seen the increased trends, you know, we think our actions improved the bottom line, we have a book of solid insurance, but look, it's still not at required returns because of some of the legal trends, these larger verdicts that we are seeing, we are cognizant of that. And so, we are just going to continue to push for rate and deductibles. I think when you look at how it would impact our healthcare portfolio, it's a good news with certain expenses for us is we have a strong market position in healthcare, because of the years of experience, and we can act swiftly and appropriately, and we are seeing some progress. But look, there's clearly more to do, given some of those trends, to your question.

[Meyer Shields, KBW]

Okay, thank you.

[Dino Robusto, Chairman and CEO]

Thank you.

[Meyer Shields, KBW]

Thank you very much.

[Operator]

Thank you. And we will now take our final question from Gary Ransom with Dowling & Partners.

[Gary Ransom, Dowling & Partners]

Yes, good morning. I wanted to talk a little bit more about the expenses ratio, maybe you can remind us on where you're trying to gap, we've seen some improvement, but I think you're looking for even more, and maybe within that you can talk a little bit about technology investments that you've mentioned, what's on the table, what's left to be done, and what more are you doing on that front?

[Dino Robusto, Chairman and CEO]

Hi. Gary, it's Dino. So, a couple of observation, we didn't targeted a specific expenses ratio, and it had more to do with - you know, we have to balance on the one hand, efficiencies we are going to gain and where in particular I felt we needed to continue to make investments in things like technology analytics, but also in the talent that we've talked about, I think. As I said on the prepared remarks, it's not about two points of run rate sitting at around 33.5, but we still think there're operational efficiencies that can be achieved. As we adopt more of - you know, via technology analytics, and also just the classic sort of division of labor processes that you can use across your supply chain.

So I think it will help the numerator, or at the very least, Gary, it's going to keep the numerators stable while we continue to grow the business, which case the denominator in turn will help us out. So I think it isn't going to be a straight line down, impacts the denominator and we do have to look at those investments. We talked about the - on technology the Atos investment and the use of Atos on the infrastructure which obviously had some the heavier cost upfront, and it's going to continue through the fourth quarter, but at the end of the day it is then going to probably generate 10 plus million of savings in terms of what specifically technology analytics, it's clear to watch, and tour recent hire with Michael Costonis for technology analytics. We continue to invest in the people. We continue to invest, and this is just evolving tremendously quickly, and so I just think it - I will say today that we're going to continue to those investments, I will say, three months from now, I will say a three years from now, and for you know, however long you'll be asking, because it just continues to evolve.

The question in some of that technology like we talked about Atos is can you find new operating models, consumption models, that allows you to upgrade the capabilities and the per unit cost, if you will, managing the bit of information actually goes down, and some of our applications that were migrating to the cloud will over time confidently bring that unit cost down, but there's clearly investments we have to continue to make, and we are going to continue to do it. I think lot of focus on operational efficiencies, and we can keep that numerators stable, we have been growing the business, as I said in the prepared remarks, we are up five points net written premium at nine months, and so we anticipate it's going to continue to improve. And look, we will continue to be transparent and give you our progress reports on the expense ratio in the upcoming quarters.

[Gary Ransom, Dowling & Partners]

Do you have any sense where you stand right now if you try to compare how good your systems and technologies are stacked up against other peers? I mean the same way you're trying to get to top quartile on performance…

[Dino Robusto, Chairman and CEO]

Yes.

[Gary Ransom, Dowling & Partners]

-- where you stand on that?

[Dino Robusto, Chairman and CEO]

We hear a lot of companies and carriers talking about their investments that they're making. I'm presuming those are all improving. I would say - I mean there isn't a consultant that you can't approach the deals and works in this industry that doesn't repeat consistently the tremendous legacy laid in platforms that the insurance companies have, and we have to have legacy latent platforms, this big infrastructure change where it's obviously a big foundational piece at evolving that legacy over time. So I suspect what you would find is if you take a broad definition of technology and analytics, digital, some people are going to be more advanced in a certain area for a certain line of business. We might be a little bit better on some of this infrastructure we recently did, or maybe some of our pricing analytical models. Some people might be a little bit more advanced on claims, but I think there is clearly enough evidence based on all of the consultants working to help the insurance industry and the insurance by the insurance tech firms. But there is still a lot of legacy laid in platforms, and so, I don't know, I just - I think you know, where are we - where do we have to go, we have to continue to adjust, there is no question about it. We are highly focused on it, and we spent a lot of time talking to be insured tech firms and other people capable in this area, including several of the very large tech firms to find out how they can help us together, how can we evolve this, and we've got ways to go, so - we will keep you posted as we make more significant - or bring more significant developments to technology and analytics.

[Gary Ransom, Dowling & Partners]

Thank you very much for those answers.

[Operator]

Thank you. And that does conclude today's question-and-answer session. I would like to turn the conference back over to CNA for any additional for closing remarks.

[Dino Robusto, Chairman and CEO]

No, thank you very much, and we will talk to you in a quarter.

[Operator]

Thank you. That does conclude today's conference. Thank you all for your participation. You may now disconnect.