Thank you. Peter Hill, you may begin your conference.
Good morning, and thank you for joining our third quarter 2017 financial results conference call. Yesterday, after the market closed, we issued our quarterly release. If you didn't receive a copy, please call me at 212-521-4800 and we'll make sure to provide you with one. There will be an audio replay of the call available from about 1:00 P.M. Eastern time today through midnight on December 1st. The replay can be accessed by dialing 855-859-2056 or +1 404-537-3406. The passcode you will need for both numbers is 18690169. Today's call is also available through the investor information section of www.renre.com and will be archived on RenaissanceRe's website through midnight on January 9th. Before we begin, I'm obliged to caution that today's discussion may contain forward-looking statements and actual results may differ materially from those discussed.
Additional information regarding the factors shaping these outcomes can be found in RenaissanceRe's SEC filings to which we direct you. With us to discuss today's results are Kevin O'Donnell, President and Chief Executive Officer, and Bob Qutub, Executive Vice President and Chief Financial Officer. I'd now like to turn the call over to Kevin. Kevin?
Thanks, Peter, and thank you all for joining today's call. Last night we released third quarter earnings. As you know, it was a busy quarter resulting in a reduction in book value of 11.6% and a reduction in our tangible book value per share plus accumulated dividends of 12%. This quarter, although difficult, was not surprising and validates both our view of risk and our long-term strategy. The driver of our performance was, of course, the multiple catastrophic events occurring in the third quarter, namely Hurricane Harvey, Hurricane Irma, and Hurricane Maria, and the Mexico City earthquakes. Before moving on to a discussion of the quarter, I would like to extend my deepest sympathies to everyone affected by the Q3 large loss events.
In Mexico, the U.S., and especially in the Caribbean, life has still not returned to normal for many, and it is our hope that a significant number of claims we have already paid goes a little way towards speeding recovery. I will discuss the Q3 large loss events and their impact on RenaissanceRe in greater detail after Bob speaks. First, I would like to address our gross to net strategy, the role of retro markets, and the increased cost of risk capital. Our gross to net strategy was tested by the Q3 large loss events and performed well. We pre-announced a net negative impact of $625 million for the Q3 large loss events. As you saw in our earnings release, we now think that number will be closer to $615. Our gross position on these events, however, is about $2.2 billion.
This means that in excess of two-thirds of our gross losses have effectively been ceded to retrocessionaires, shared with third-party capital, or offset by reinstatement premiums. In addition to preserving our capital, this strategy was highly efficient, costing us only 50% of our premium, 20% of the expected profit on the associated business. Even after significant recoveries, however, we still have abundant retro capacity remaining in our program, and we continue to offer substantial capacity to clients. The retro markets, and especially the collateralized markets, have absorbed a large share of the Q3 large loss events, and a material portion of their capital will either be impaired or locked up. The ability of some of these funds to recapitalize and trade forward will be heavily dependent on rate. Investors experiencing large losses will need material price increases before they agree to reinvest.
This could cause disruption at January 1st when roughly three-quarters of the retro market renews. This is not a problem from our perspective, as our demand for retro is highly elastic. Using our integrated system, we are able to source the most efficient capital available. Sometimes that is our own, sometimes it's third party, and sometimes it's retro. If the retro market hardens sufficiently, we will happily transition from a buyer to a seller in the future. While the cost of retro is increasing, our weighted average cost of capital is decreasing. Over the last several years, we have taken advantage of record low interest rates to replace high cost debt and preferred shares at a lower cost.
This gives us the option to put more of our own capital at risk if we are paid sufficiently to do so. Another reason we aren't as exposed to the traditional retro market is that we have many long-term aligned partners. When we need to augment our own capital, I think of the potential sources as falling into 2 categories: short-term trades and long-term partners. Around 75% of our ceded premium is with our long-term partners. This is capacity we can count on being there year after year because we stand alongside it on both the profits and the losses. Having this long-term capacity has allowed us to grow over time and provides us flexibility when market opportunities arise. I have been saying for years that in many lines of business, rates have reached unsustainably low levels.
This was fueled in part by the extended drought in large catastrophic events, especially in the U.S. That drought is now over. It appears as if 2017 could be the third year having more than $100 billion of insured losses over the previous 15. To put this in perspective, the worldwide annual expected insured catastrophic loss is between $50 billion and $60 billion. Eight of the last 17 years, however, have come in under $40 billion, even on a trended basis with three under $20 billion. There is a significant amount of variance in the results, and when the low loss years cluster, this can be confused for a lack of volatility. Years like 2017 are not outliers. However, there is a far more volatility in our sector than many appreciate. We expect to have industry loss similar to 2017 at least every 10 years.
As a sector, we haven't been paid for this volatility for too long now. Making matters worse, low prices in property cat have affected almost every other line in the P&C industry, with companies writing diversifying business to help offset property cat rate decreases. The cost of risk capital needs to go up, and its impact will reach beyond loss-affected property deals. This quarter was a needed reminder that ours is a volatile business, and the vendor models cannot substitute for good underwriting. Allocators of capital are better positioned to be able to determine which underwriters were skillful and which were not, and return expectations should adjust accordingly. Looking ahead, we are hopeful that 2018 will provide greater opportunities than 2017. We're optimistic about our prospects for profitable growth and our preferential position in the market.
With both rated and collateralized balance sheets and unequaled access to efficient capital, we are ready and willing to trade forward to January 1 across all our platforms in the form our customers desire. With that, I'll turn the call over to Bob for a look at our financials, then I'll come back on and share a bit more on our business performance before we open it up for questions. Bob?
Thanks, Kevin, and good morning, everyone. As Kevin noted in his opening remarks, the Q3 large loss events caused significant damage throughout the affected regions and continue to present wide-scale humanitarian challenges. From an insured loss perspective, these events combined to produce the largest single quarter loss in RenaissanceRe's 24-year history. We recorded a net negative impact to our consolidated financial results of $615 million in the third quarter of 2017. Recall that net negative impact includes the sum of estimates of net claims and claim expenses incurred, earned reinstatement premiums assumed and ceded, lost and earned profit commissions, and redeemable non-controlling interest. Included in the net negative impact for the quarter was $534 million associated with hurricanes Harvey, Irma, and Maria, and the Mexico City earthquake.
It also included $81 million associated with aggregate loss contracts where cumulative losses under the respective contracts reached their retention points during the quarter. In an effort to provide transparent disclosures, we included aggregate losses in our net negative impact figure for the quarter as they were meaningful to our results. Our best estimate of losses from the large cat events in the third quarter would be largely responsible for triggering losses under these aggregate contracts. However, the aggregate losses in and of themselves are not necessarily attributable to a specific event in a traditional sense. There remains meaningful uncertainty with respect to our estimate of losses from the large cat events and the aggregate loss contracts given the limit features and the impact of our retro book. At this time, I'd like to highlight a few metrics that give an overview of our financial performance for the quarter.
I'll then provide some additional detail of our segment results, our investment portfolio, and capital activities before I turn it back over to Kevin. For the quarter ended September 30, 2017, we reported a net loss of $505 million or $12.75 per diluted common share and an operating loss of $547 million or $13.81 per diluted common share. On a year-to-date basis, we reported an annualized ROE of negative 7.4% and an annualized operating ROE of negative 11.7%. During the quarter, our book value per share decreased 11.6% and our tangible book value per share including accumulated dividends decreased by 12%. On a year-to-date basis, our book value decreased by 7.8% and our tangible book value per share including accumulated dividends decreased by 7.3%.
For additional details of our quarterly and year-to-date results, I would refer you to our earnings release and financial supplement which we issued last night and can be found on our website. Let me now shift to our segment results, beginning with the property segment, followed by casualty and specialty. Within our property segment, gross written premiums were up 171% for the third quarter of 2017 compared to the third quarter of 2016 and included $165 million of reinstatement premiums associated with the large events. Excluding the impact of reinstatement premiums written in 2017, our property segment gross premiums written would still have increased 34%, with our other property class of business up 64% and our catastrophe class of business up 14%. The increase in our other property class of business was mainly due to increased participation on a select number of deals and certain new transactions we found attractive.
Our catastrophe line of business typically does not see major renewals during the third quarter, but we were able to grow the book slightly, including some backup covers while exercising underwriting discipline given prevailing market terms and conditions. Our property segment incurred an underwriting loss of $750 million and a combined ratio of 323%, compared to underwriting income of $103 million and a combined ratio of 40% in the comparative quarter. The underwriting results on our property segment were dominated by the impact of the Q3 large loss events. These combined for $809 million in underwriting losses and added 252 points to the combined ratio in our property segment. Overall, our property segment performed as expected following the occurrence of the large catastrophe events in the quarter. We continue to believe we have the right people, systems, and strategy to execute through market cycles.
Moving on to our Casualty Specialty segment, where in the third quarter of 2017, gross premiums written were up 1% relative to the third quarter of 2016. We were able to selectively grow new and existing business within certain casualty lines of business. Mostly offsetting this increase was a decrease in gross premiums written in our financial lines of business, primarily the result of a large in-force multi-year mortgage reinsurance contract written in the third quarter of 2016 that did not reoccur in the current quarter. With the growth we've experienced to date in the top line, we continue to execute on our Growth to Net strategy, having ceded out 32% of our Casualty Specialty premiums, given current market conditions.
The Casualty Specialty segment incurred an underwriting loss of $43 million and a combined ratio of 120% in the third quarter of 2017, compared to underwriting income of $9 million and a combined ratio of 95% in the comparative quarter. A key driver of these results was the impact of hurricanes Harvey, Irma, and Maria, and the Mexico City earthquake, which drove the current accident year underwriting results in our Casualty Specialty segment. Positively impacting the Casualty Specialty segment combined ratio during the quarter was a 7-point decrease in the underwriting expense ratio. Net premiums earned in the Casualty Specialty segment during the quarter were up $37 million, and underwriting expenses were relatively flat as we continue to leverage our existing expense base while selectively growing this book of business.
It is important to note that following a quarter that saw the return of a number of significant loss events, we continue to evaluate our reserves for developing trends and remain comfortable with our processes and overall reserve adequacy. Turning to investments. In the third quarter, we recorded total investment result of $82 million, generating an annualized total return on our investment portfolio of 3.4%. Included in this result was net realized and unrealized gains on investments of $42 million and net investment income of $40 million. Our equity portfolio continued to perform well during the quarter, generating both realized and unrealized gains as markets delivered positive returns. Our net investment income was comprised mainly from our fixed maturity securities and benefited from higher average invested assets, modest increases in interest rates, and a tightening of credit spreads during the quarter.
Net investment income for the quarter also included some unrealized losses in our cat bond portfolio, which was impacted by the events of the third quarter. Our investment portfolio remains conservative with respect to interest rate, credit, and duration risk, with 89% allocated to fixed maturity and short-term investments with a high degree of liquidity and modest credit exposure. The duration of our investment portfolio was 2.6 years, and the yield to maturity on the fixed income and short-term investments was 2.2% at September 30th, 2017, more or less flat compared to the end of last quarter. Our Strategic Investment portfolio, managed by our Ventures unit, again produced positive returns overall for us, and we continue to be satisfied with the long-term fundamentals of the companies we own. Now turning to our capital management activities during the quarter.
Following a string of significant cat events, it is a testament to our capital management strategy that our balance sheet remains liquid and our capital position remains strong. Our access to capital also gives us the flexibility to pursue strategic investments and capital management activities as they may arise. Our ventures team continues to actively build relationships with high-quality, long-term investors, as well as looking for new strategic transactions that can enhance our underwriting franchise. Overall, our capital management actions reflect a quickly evolving market, and we believe we have developed a unique agility to deploy capital where it is needed most. Once again, our trusted long-term investment partners and our joint venture vehicles supported our efforts. They recognized the leadership we bring to the property cat exposed market and immediately stepped up with an additional capital deploy.
For example, we quickly and efficiently raised $250 million of new equity capital in DaVinci from third-party investors, and Upsilon received additional funds to support its core customers. On the share repurchase front, prior to the arrival of Hurricane Harvey, we were active in the market for our common shares, repurchasing $39 million during the quarter, which brings our total year to date purchases up to $189 million. Our approach to capital management has not changed. With the potential for improved pricing conditions in many of the markets we serve, we will first and foremost look to deploy capital into underwriting and business opportunities that may meet our risk-return hurdles. At this time, I'd like to mention that commencing with our first quarter 2008 financial supplement, we will no longer separately disclose the underwriting results of our Lloyd's platform.
We manage our business at a segment level and with the results of our Lloyd's platform getting picked up in our property and casualty specialty segments as appropriate. As such, we will continue to provide what we feel is appropriate transparency into our segment results and associated market commentary on our earnings call. From a disclosure perspective, this brings Lloyd's in line with other locations and underwriting platforms across our organization. Finally, before turning the call back to Kevin, I would like to extend our deepest sympathies to all those affected by the devastating California wildfires. They have resulted in loss of life and caused significant damage throughout major portions of the state. It is still very early days for this loss, with initial industry loss estimates ranging anywhere from $2 billion-$3 billion to $6 billion-$8 billion, and potentially higher.
As we work through our initial assessment, our early expectation is that given the complexity of these events, the industry losses will come in closer to the higher end of published industry loss estimates. There is significant uncertainty with respect to the nature and magnitude of these losses, and we will continue to monitor information from clients, industry participants, and other sources as it becomes available. With that, I'd like to turn the call back to Kevin.
Thanks, Bob. I'll divide my comments, starting with property, then casualty, and then we'll open it up for questions. We broke another hurricane drought in the third quarter, this time in U.S. landfalling major hurricanes. The last year a major hurricane made landfall in the U.S. was in 2005 with Wilma. This year we had three, Harvey, Irma, and Maria. We also had several large earthquakes, including in Mexico City. Multiple hurricane records fell in the third quarter, such as experiencing 53 named storm days. In addition, while not a record, there were five major hurricanes, including four that reached Category 4 or 5 strength. This season was driven by warm waters and low wind shear and otherwise near-perfect conditions for storm formation. As is typically the case, each of these storms had very different characteristics, hit different risks in different geographies, and will develop very differently.
Unlike more concentrated losses, such as the 2004 Florida hurricanes, the Q3 large loss events will affect a broad swath of the industry, and consequently will have wide-ranging impacts on market conditions, affecting primary, reinsurance, and retro, both in the U.S. and internationally. Starting with Harvey, which made landfall in Texas on August 25th as a Category 4 storm. This was really more of a flood than a wind event. While its wind field was relatively small, Harvey dumped up to a record 50-plus inches of rain over a broad expanse of Houston. To put this rainfall in perspective, over 25 trillion gallons of water fell on Texas and Louisiana, which is enough to fill the Chesapeake Bay. Harvey looks to be about a $30 billion industry loss, which is around a 20- to 30-year return period for the Gulf region.
Even though Harvey is predominantly a flood event, the private market is exposed on both the residential and commercial side, including a significant auto loss. While this loss primarily affects our property cat book, it will also affect our other property and casualty businesses. Next up was Hurricane Irma, which made landfall on September 10th in the Florida Keys as a Category 4 storm, then made a second landfall over Marco Island as a Category 3 storm. If Irma had tracked a handful of miles north, it would have not weakened over Cuba. In all likelihood, it then would have made landfall on the heavily populated east coast of South Florida as a Cat 5, rather than over the Everglades as a Cat 3. This would have been a true one-in-100 event, with the potential to cause more than $100 billion in loss.
Irma looks to be about a $25 billion industry loss, which is around a 20- to 30-year return period for the southeast U.S. Irma is predominantly a wind event, even though there was significant flooding. Average claim severity outside of the Florida Keys appears to be relatively low, and in many cases is coming in under applicable hurricane deductibles. Finally, at least as far as hurricanes in the third quarter go, was Hurricane Maria, which made landfall in Puerto Rico on September 20th as a strong Category 4 storm. Maria looks to be at least a $35 billion industry loss, which is a 100-plus year return period for Puerto Rico. While Puerto Rico is located in the Caribbean, as a U.S. territory, it will impact the U.S. reinsurance towers of many large U.S. insurance companies.
Due to infrastructure issues, we expect that real losses will take longer than average to fully develop. I often say that our value proposition extends beyond price, and we had another opportunity to demonstrate that again this quarter. As each of the quarter's hurricanes was developing, our underwriters, along with our team of scientists at WeatherPredict, closely monitored the storm, its potential for strengthening, and the most likely track it would take. Throughout this process, we made sure to reach out to those of our clients and brokers most likely to be affected. After the event, in addition to rapidly prepaying claims, we were able to provide core clients footprints of their portfolios run against our proprietary industry database. The speed and skill of our people and our systems post-event is testament to our decades of experience in responding to events just like these.
Third quarter also experienced several large earthquakes, including Mexico City. This loss does not appear to be as destructive as originally thought. That said, earthquake losses are very long tail in nature, and it's not uncommon to have significant development over an extended period. We will be monitoring both Texas and Florida closely for signs of assignment of benefits issues and other adjuster fraud. To date, there's been little indication that this has occurred, but there is still opportunity for fraud to begin to creep in later in the process. Insurance companies are acutely aware of this problem, however, and are taking steps to identify and minimize fraudulent claims. We also saw significant demand surge around adjuster fees. Due to the short time span between Harvey and Irma, there was intense competition for loss adjusters, driving up the fees insurance companies need to pay for their services.
While not a big driver of loss, it will result in increased loss adjustment expenses. I would like to briefly address how our independent view of risk incorporates the commercially available catastrophe models and our expectations around the frequency of Q3 large loss events. We spend considerable time and resources understanding the strengths and weaknesses of the vendor models. Consequently, all of the third quarter events were within our expectations. These were not extreme tail events. For example, in Harvey, we recognized the potential for significant flood losses and that this potential is not always sufficiently captured in the models. In Maria, we understood the vulnerability Puerto Rico faced to major hurricanes, and while more of a tail event, this loss was not surprising. This approach is consistent with our aspiration to be the best underwriter, as we believe that being so results in superior shareholder value.
In our casualty segment, gross premiums written were relatively flat quarter on quarter, but we experienced strong net premium earned growth of 21% as our mortgage book continues to earn through. We improved operating leverage in casualty again this quarter with our operating expense ratio down about one percentage point. Our casualty segment also experienced losses from the Q3 large loss events, although to a lesser extent than our property segment, and these losses primarily affected our marine and energy books. Overall, I'm pleased with the portfolio we've built in this segment. I take a long-term view on the casualty business and recognize its benefits toward maximizing shareholder value. As you know, margins on this business have been compressing, and the team is working hard to build attractive positions focused on long-term value.
Similar to our gross to net strategy and property, we cede one-third of our gross premiums on this book, which gives upfront profit to limit downside. This quarter saw the benefit of this strategy as we enjoyed significant retro recoveries, especially in marine. We have been keeping a close eye on loss trends in the casualty space. For example, we've been underweight commercial auto and overweight financial risk, which is consistent with our strategy of constructing a portfolio that is more attractive than the market average. Going forward, in addition to the business affected by the Q3 large loss events, we anticipate that some of the more challenging areas of the market will adjust and the positive trend in certain casualty lines will accelerate.
Casualty is a key aspect of our value proposition to our customers, who we believe want a reinsurer who makes a credible commitment to cover a wide range of their risks over reasonably long time periods at consistent exposure-based prices. Being able to provide a suite of products beyond property cat is essential to this value proposition. Over the long term, I believe this business is accretive to shareholder value. Our ventures unit continues to contribute both to our broader results and to our ability to execute our gross to net strategy. Once again, for example, the strategic investments managed by our ventures unit had positive returns this quarter. As Bob Qutub noted, we raised capital in our DaVinci vehicle at October 1st. DaVinci is fully funded and ready to renew existing business and grow if opportunities present.
Also, at October 1st, we were able to raise additional funds in our Upsilon joint venture to support an attractive deal with a core customer. We're also ready to trade forward in Upsilon, and to the extent there are attractive opportunities at January 1st, we will be in a position to transact. The key aspect of our consistent aligned approach with our joint venture partners appreciate is that they will have the opportunity to benefit alongside us in any market opportunities in 2018. We currently have multiple offers to bring in new capital, but we will remain aligned with our long-term partners. The third quarter was a great opportunity to demonstrate to our customers that our value proposition extends beyond price and includes many value-added services both before and after large events.
At January 1st, our customers will also realize the value of trading forward with a long-term, trusted partner with unsurpassed access to efficient capital. In 2018, our hope is that rates will adjust to the point where equilibrium is achieved, neither too low nor too high. Whatever the future, however, we have the platform, the people, and the capabilities to continue our leadership in the industry. Thank you, and with that, I'll turn it over for questions.
At this time, I would like to remind everyone, in order to ask a question, please press star then the number one on your telephone keypad. We will pause for just a moment to compile the Q&A roster. Your first question comes from the line of Elyse Greenspan from Wells Fargo. Your line is open.
Hi. Good morning. My first question, appreciate all the color around the hurricanes and kind of the outlook, Kevin, what kind of price expectations do you have heading into the one-one renewals? Just how do you kind of see based on discussions with clients the pricing environment shaping up?
Thanks, Elyse. Firstly, 2018 prices are going up. I think the comments I made about there being a broad swath of the industry affected. If you look at the insureds affected by Harvey is a different pool of insureds affected by Irma, then again, those affected by Maria are different. As you move further away from loss-affected layers, it is always a little harder to predict what will change and how much price will shift. When I think about price changes, I worry less about the market and more about our strategy, I look back as to the way we have positioned ourselves going into 2018. As I mentioned, we lowered our GAAP capital cost by refinancing our debt and our preferreds. We are coming in with a strong GAAP capital position.
Our economic capital model, which is the model that our underwriters use to deploy capital, we are representing a higher cost of capital in our economic capital model. The reason for the difference is because we have made different assumptions on the assumed, sorry, the ceded retro supporting that portfolio. In 2017, the cheapest capital we had in our economic capital model was the ceded purchasing that we did, we believe that that is unlikely to be available in the same form in 2018. That is good news for us because we have cheaper GAAP capital, we have higher expected margins. The spread between the cost of our capital and the opportunities in the market is greater, which will inure the benefit of our shareholders. While I am less concerned about the overall change in the market, I think the change in market pricing will be reasonably broad.
I think we have better access than anyone else in the market, we have unlimited and unrestricted access to efficient capital to bring to those opportunities.
Okay. As you say that there's a better margin outlook, obviously with your casualty and specialty business, the margins within that business have been in excess of 100%. How do you think about potentially putting more capital towards your property and specifically property cat business and maybe shifting away from casualty and specialty business if the catastrophe market does get a lot better?
I think let me start just talking about casualty, then I'll talk about the effect on capital. We've had a few glimmers of hope in the casualty market with some of our November 1st renewals, where we're seeing increased underwriter discipline in thinking about the risk that's being assumed. I believe that in the casualty space, we are not at a long-term sustainable margin, but I think about it over 10 years, and at over 10 years, I believe we will have sustainable margins. I think there's two areas that we'll be focused on for pricing and casualty as much of our book is proportional. One is encouraging primary companies to continue to accelerate the rate increases for their insurance books, which will inure to our benefit. The second is thinking about whether the cedes are at appropriate long-term levels.
I have reasonable confidence we'll have some rate enhancement on the primary books. I think it'll take a little longer for us to have clarity as to whether cedes will respond favorably to market pressures as well. With regard to writing more property or casualty, so if both are better, we will write more of each. The capital allocation to our casualty and specialty business remains quite low, and it will continue to be low, particularly if we find more opportunity in property cat, as property cat drives the tail of the distributions. On a marginal basis, our casualty specialty returns still look quite good. The more property we write creates more room for casualty and specialty from a capital allocation perspective. I think standalone returns look better for both property, and we're optimistic about casualty and specialty.
I think from a capital allocation perspective, we're in a very strong position.
Okay, great. Were you guys surprised that within your cat bonds in your investment portfolio only lost about 5% in the quarter?
No, we would expect in events like this cat bonds to do take an impact. The losses that we talked about were unrealized. You can see it in our supplemental where it was down about $16 million, and it was reflected in our net investment yield. Again, it's unrealized.
Okay, great. One last question. Kevin, you did give the growth versus net losses for RenRe for the quarter. How much of the cede was to third party versus traditional markets?
By third party, you mean how much of it is collateralized recovery?
Yes.
I think it's roughly about half a billion of collateralized recovery against, I think it's $1.2 billion of total recovery. Ballpark.
Yeah, around, right.
Okay, thank you very much.
Yep, thank you.
Your next question comes from the line of Kai Pan from Morgan Stanley. Your line is open.
Thank you and good morning. The first question on alternative capital market. DaVinci lost about $223 million, and you raised more than that. That should show you the readily available alternative capital out there. Would that impact the magnitude and duration of potential price increases? I just wonder when you discuss with your capital provider for the reloading, what kind of pricing expectation that you would deploy that capital?
Do you want to start, Bob?
One thing just for clarification, Kai, in the supplemental, the actual loss for DaVinci that we disclosed is $255. $223 you're referring to is the non-controlling interest on the investors, just for clarification. It's on page seven of the sup.
Kind of in layman's terms, we kept DaVinci the same size, effectively facing the market. I think about the capital raise in DaVinci kind of as business as usual, to be honest, where our normal process is at the end of the year, we dividend back to the investors the earnings. When there's an event, we put a capital call out to refund the balance sheet back to the levels it was prior to the events. I think from that perspective, we think DaVinci has good opportunities going into year-end. It's a slightly different type of vehicle, so we're not doing a traditional capital raise as one would expect with some of the more traditional collateralized funds. This is people who have been with us for a long time, and it's part of our normal process of managing capital.
Yeah. I was wondering, is that showing a broader appetite for risk still from the alternative capital markets?
We consider DaVinci effectively to be closed. We do have substantial interest in investors trying to get into DaVinci, but we had that last year as well. I think the appetite that investors have to see, to take, to share in our underwritten risk is high for 2018, but to be honest, it was very high for 2017 as well.
Okay. On your gross and net strategy, you have two third of your gross recovered from the retro market. I just wonder, given the potential rising cost of retro, would you be able to maintain your gross exposure or grow that?
That's a great question. We absolutely can maintain our gross exposure. We like the gross book that we wrote. We like the net book more, we used the retro to enhance and optimize the portfolio. As I mentioned in my comments, I think of the as a split between the retro that we purchase or the risk that we share is probably a more accurate way to think about it, where about 70% of the risk that we share is with what I consider to be long-term partners, 30% is more of a trading account. We have perfect elasticity as to whether we renew the trading account, it'll be very much price dependent. The 70% that I consider to be partner capital is capital that will participate in a better market in 2018 just alongside us like they did in 2017.
We are not subjected to the same client-facing swings from retro as others because we build our book to make sure that we have a consistent customer-facing appetite and manage our net risk through partner capital and trading capital.
Great. Last one, if I may, is on your underlying loss ratio in the property segments. It looks like it increased a lot year-over-year. I'm just wondering, is there particular sort of non-cat large losses, which is not including in your disclosure?
I think you're referring specifically to other property within our property segment. I think we provide other property as a breakout in the property segment because we have an attritional reserving component to that, which would be difficult to extract from a property cat-only representation. The other property and the property cat are highly linked, where much of the other property capacity that's put out is in conjunction with property cat lines that are written. When I look at the performance of that book, I actually look at it from a property segment perspective, and I'm less worried about the allocation of the loss ratio between other property and property cat.
Okay, great. Well, thank you so much.
Yep. Thank you, Kai.
Your next question comes from the line of Mike Mumma from Buckingham. Your line is open.
Thanks. Good morning. I like that name better. Just going back to the discussion on capital management. Obviously, you're opining that you look at the market conditions and then revisit, and you were buying back before HIM. Based on the different renewal cycles, I guess 1/1, 4/1, 6/1, et cetera, are we thinking of revisiting the capital management discussion later in 2018, or would you be able to sort of come up with a thought process after the 1/1 renewal?
There's a couple parts to that question. Let me start off with the comments in my prepared text were, as you pointed out, we're looking to deploy capital here in the fourth quarter as opposed to returning capital. We feel comfortable about our balance sheets being fully capitalized and taking advantage of the opportunity. Regarding 2018 and returning capital, I think really it's what those opportunities early on in 2018 present themselves for opportunities to deploy capital. We can look later on in 2018 and what our decisions will be then.
Is it the point
The second part of your question is really in the context of what will the parity be between the returning capital and the rate increases, and there is a relationship there.
Yeah. I think our normal process is highly integrated between our underwriting and our capital management. You had mentioned the 1/1, 4/1, and 6/1 renewals. We've already recreated our portfolios for the full expectation of price change for 2018. That gets periodically revisited, adjusting it upward or downward based on the assumptions we had prior to each major renewal. That will ultimately inform whether we're allocating more capital to the business or potentially looking to purchase shares back.
Right. The second question I had was, obviously you probably sounded a bit more optimistic on the pricing equation than what I would've thought based on the industry losses as a % of total capital. I'm curious, do you have a view on the overall industry losses sort of drifting upwards, down as time progresses? I think a lot of us here are somewhat scratching our head at trying to figure out the missing portion of the losses. I know that some of it is going into the alternative market. Even when you look at the math, it seems that based on what we know today, the numbers are still not getting there. I was curious if you had an opinion on that.
I share your confusion at kind of the highest level as to the disappearing nature of these losses. The thing I would say is if you look at our gross loss of roughly $2.2 billion, between what we've prepaid and what's been reported, we're at about 10%.
There's a lot of latitude as to how one believes they've been impacted by these events. The fact that we're at relatively low levels compared to what I believe will be close to $100 billion, isn't wildly surprising. Often in the periods shortly after loss, there is a wide gap. It tends to close over time. It does seem as if the gap is potentially a little bit bigger. I think you touched on an important component here, though, is there's usually a question as to what's the market opportunity. In going into 2018, there's a big question as to how much capital is going to be there to support it. I think from the rated capital perspective, most rated carriers are on very solid footing going into 2018.
The lack of transparency as to how much capital is impaired, how much capital is locked up from the collateralized markets, is adding uncertainty to the overall loss estimate for the combined events and also for the supply and demand dynamics going into 2018.
Fair point. Final question. I know there were some questions already on the casualty book. I know I've asked this before. In retrospect, do you think the Platinum acquisition is achieving what you would've thought you would achieve on day one or has it taken longer to get to the point what you would have outlined in your initial plans?
We put out a list of objectives when we purchased Platinum, and I feel as if we've achieved each of the goals that we outlined. I think the question is, did we set the bar too low for what we hoped to achieve? I think, when I look at where the company is, and a lot of that is because of the benefit of us purchasing Platinum, I feel like we're in a much stronger position going into 2018 because of our strong casualty platform than we would've been had we not executed on Platinum.
Fair point. I'll stop here. Thanks for the answers and good luck for the future.
No, thank you.
Your next question comes from the line of Joshua Shanker from Deutsche Bank. Your line is open.
Thank you. Forgot to get out of queue. Everyone asked my questions, but good luck in the new year.
Thanks, Josh.
Thanks, Josh.
Take care.
Your next question comes from the line of Brian Meredith from UBS. Your line is open.
Hi. Yeah. A couple of questions here for you. First one, Kevin, I'm just curious, how much rate do you think you need on the property cat reinsurance business to meaningfully increase your net exposure without the benefit of the cheap retro out there that's probably not going to be there going forward?
I think that's an iterative question. I think we will purchase retro. I look at 2017, we probably had more income statement protection than we will likely have in 2018. It's not that we won't have any trading account retro, but it'll be structured differently. I think, again, we'll look at the spread between our cost of capital and what we're being paid for risk and make sure that's adequate. If we are changing the way we're purchasing retro, I think a natural thing to ask is, well, if we're assuming more income statement volatility, that must be in exchange for something, and that will definitely be in exchange for higher ROE in 2018. One, we'll be taking more risk capital than exposing it, and two, we'll be taking more income statement volatility because a lot of the retro we purchased in 2017 was down low.
Got you. Okay. Maybe a better way of saying that is how underpriced do you think property cat reinsurance is right now relative to your kind of cost of capital?
Again, as a market or as our book, I don't feel as our book is underpriced. I look at the way it's going to be, again, how we structure the portfolio. We'll go in, we'll continue to rerun our portfolios every night, make sure we understand how we're using capital. On every marginal deal, we understand whether we're enhancing or reducing not only our existing margin, but the target margin we have for 2018.
Got you. Okay, that makes sense. I'm just curious, again, you kind of mentioned you thought this was going to have implications outside of just the loss impacted areas if you look back at the KRW and what happened in 2011 that didn't happen.
Yes.
Why different this time?
Two things I'll say. One is the cat market tends to react more to U.S. events than non-U.S. events. That's just the nature of the beast. Actually more importantly, if you go back to 2011, rates were frankly much higher, and there was less need for rate enhancement than there is in 2017, where we've had several years of rate declines. I think there's just more of a recognition that it's been a buyer's market, and at least in the conversations we're having with our clients, we're finding them receptive to the fact that we need to come to an equilibrium. We're not looking to push prices too high. We're getting to a point where margins are at a better balance between the buyer and the seller.
Right. Well, would KRW be a better kind of parallel here because rates were very cheap before then?
Yes. We did actually see quite a bit of movement in KRW.
Not outside the U.S.
Correct. As I said in my comments, I think as you move further from loss-affected layers, it is more difficult to predict the rate change. I think in looking at our international footprint on the primary reinsurance for CAT is pretty low at this point because rates have gotten to a point where there's not much business that is fitting within our attractive return profile. I think we've got upside there. I'm not sure, even with reasonable rate changes, it's going to attract us to write a bunch more international. Most of the international risk that we've taken over the last several years has come through our retro account. I do think retro rates will move pretty substantially, both internationally and U.S.
Makes a lot of sense. One other quick one here. California wildfires, is that going to be one event or multiple events for reinsurers?
That's a question that is probably going to be figured out in the near term. I think it's pretty complicated. There are wordings about how to connect discrete fires to determine whether they're a single event or multiple events. As we've seen before, for retro, it's probably one event. For insurance or reinsurance, it's going to take some time to figure it out. I think that loss is pretty complicated generally, just to give some color on it. It's easy to see the homes that have been totally disrupted, but increasingly we're hearing reports of soil contamination, and the remediation of that and how that's going to be handled could be additive to the cost. Then also the toxicity of the smoke damage to adjacent buildings is becoming more of a topic of our insureds as to how they're thinking about remediating those losses. There's uncertainty with that.
Finally, there's a role potentially to be played by the utilities as to whether they are deemed to have been a source of ignition, and that can increase the casualty component of this loss, but also through subrogation, potentially lower the property component. I think we're monitoring all these things and trying to determine how it's ultimately going to play out, but there's still much to be answered about how this will play through.
Great. Thank you.
Sure.
Your next question comes from the line of Jay Cohen from Bank of America. Your line is open.
Yeah, couple of questions. Some of the backup covers that you wrote in the third quarter, are those annual policies? Are they for a shorter duration?
The more substantial ones, the meaningful ones are short duration.
We should think about sort of next year, third quarter being, other than the reinstatement, a kind of a tough comp from a premium standpoint, just from a modeling aspect.
I think that's accurate. Unless we have a third quarter that looks like 2017.
Let's hope not. Second question, expense ratio. I assume some of the lower expense ratio is due to lower bonus accruals?
No, our expense ratio, I mean, the acquisition ratios are growing with the business, but on the operational expense, there's two factors over time that have been driving that down. One, in 2016, you can see we really effectively integrated Platinum so that doesn't exist on a comparative basis. The other element of what dropped costs from I think high 140s to low 130s was just our focus on costs and trying to leverage the platform that we talked about. You saw that most notably in the casualty and specialty. The corporate costs, last quarter, I think the question was asked, what's the burn rate there? 4-5, and that's what we printed again, 4-5 on our corporate costs for this quarter, which are down from one-time events last year.
You didn't change your bonus accruals in the third quarter because of these losses?
Look, the bonus accruals are going to be determined at the end of the year in the fourth quarter.
Okay. That's helpful. Last question, debt to capital, I guess debt plus preferred to capital. Do you have a target for that number?
Right now, we feel very comfortable that what I said earlier in my comments was that our balance sheets are fully funded, and we're ready to deploy capital into the coming year. Where we see more opportunities, we have access to capital if we need it if those opportunities prove even more significant.
That's not a constraint at this point?
No.
Okay. Thanks.
Thank you.
Your next question comes from the line of Meyer Shields from KBW. Your line is open.
Thanks. Kevin, when you were talking about your economic capital cost, you mentioned that the cost of retro is going up. That's not inconsistent with what we've heard from other participants. Was the retro market underpriced last year, or is there a new assessment of risk?
The comment I made is we ceded 20% of our expected profit on 50% of our premium. I think one can make a determination. They may have a very different capital model than we do, but that was not retro we would have written for ourselves. Looking into 2018, though, I think what we purchased in 2017 will not inform what we purchase in 2018. I don't look at us as having a series of renewals coming up for the trading account, as we'll look to construct the portfolio on a gross basis then optimize it with the ceded opportunities that we have, that we either do or do not have. I think from year to year, it's less of an important comparison than it is in certain other lines of business.
You mean from your perspective, not the market overall?
From our perspective, correct. I think there are a fair number, or others potentially rely on retro to write their gross book. We are happy to write our gross book and rely on that trading account retro to optimize the net.
Okay. That makes sense. The second, I guess, smaller point. You talked about demand surge for claims adjusters, I guess particularly following Irma.
Yeah.
Is there demand surge in actual materials costs or anything like that?
I think less so than in other events. I think when you see something like more substantial damage in broad, widespread, severe damage is going to be more demand surge. My comments are really thinking about Hurricane Harvey and Hurricane Irma for that. Puerto Rico, I think demand surge is going to be extraordinarily important element of the overall loss. The difficulty of getting materials there, the difficulty of moving materials around, the lack of infrastructure. There's a lots of things that I think put a negative skew on how bad that loss can potentially get.
Okay, that's helpful. One last question, if I can.
Yep.
No, I think I'm okay. Thank you very much.
Oh, okay. Thanks, Meyer.
Thank you.
Your next question comes from the line of Ian Gutterman from Balyasny. Your line is open.
Hi, thank you. Kevin, I guess first maybe, I went back and looked at 2005 and 2011 and even Sandy and so forth, typically your market share on a gross basis, what was low 1%, and this time on a gross basis, you're low 2%. It's definitely a much bigger share than you normally take of a loss. Is there any story to that? Is it just, "Hey, Southeast and Caribbean wind, of course, we're going to have more exposure than foreign quakes or something in the Northeast?" Or has there been sort of a change in how you position the book post Platinum where, because when we're diversified and we have more retro, we can take a bigger gross share?
Yeah, I think we've never targeted the 1%. I know it's a bogey that's used to estimate losses for us from time to time. There's been no structural shift in the way that we've composed the book. I think it is important that we do try to keep a consistent face to the market and use retro to protect the net or to optimize the portfolio so that we're not transferring the pricing and the capital uncertainty to the customer, but providing certainty there and then managing it on our balance sheet. I don't see it as a shift in anything that we've done. I think it's frankly, probably just a unique set of outcome from a set of aggregate losses in a quarter. There's nothing that I would point to say there's a structural shift in the way that we've built the portfolio.
Perfect. Just want to make sure. The aggregate covers, can you give us a sense of if you were to see more gross development from basically, if we're starting to see the gap between the disclosed numbers and the $100 billion start to close, is there more limit left under those that could go against you?
Yes, there is more limit left under some of those aggregates. I wouldn't point to those as being particularly exposed to a change in the gross loss. I think we went through our normal process of a top-down analysis to estimate the loss as well as a bottom up. Each of those aggregate contracts have been looked at individually, and we came up with our best estimate. Often our best estimate is significantly higher than the customer's reported loss at this time. I don't think we have uniquely different exposure to those than we do to the occurrence losses within each of the events.
Got it. Are those for named storms only, or could the wildfires also get you on some of those contracts?
Those contracts can respond to the wildfires as well. This is more of a detail point, but one of the reasons to not include the aggregates and separate them out is as new loss events occur, you can kind of forget time and then take all the events over the contract life and apportion the loss to them, forgetting that there's an occurrence retention.
Right.
As subsequent events can happen, it can actually go back and adjust previous events to be lower because some of the limit needs to be moved to the new event.
Yep.
That could artificially build in favorable development if they're not disclosed separately and continue to be allocated to the events.
That makes sense. On the topic, I think it was Mitch who asked, it's just a follow-up. The idea that there seemed to be a lot of missing losses, and yeah, I know that's always the case, but it seems much more dramatic on this one, right? I mean, arguably half the losses are missing.
Yeah.
Are there things you see? Obviously, you can't see how other people are picking numbers. Maybe some of your clients who are telling you what they expect, but is there a sense you have for what's different? Is it that reportage are coming in slower and maybe in the past There was a heavier aspect of reported, and that got us closer. This time you need to put up more IBNR because the reporteds are slow? I'm just trying to get a decent story. I guess I'd have to hear one for why we have such a big gap.
Let me touch on each of the events, I'll give kind of an overall perspective. If you take Irma's kind of the most traditional of the cats. It's broad coverage and reasonably modest losses. It's much more of a traditional type assessment as to what the loss will be. As I mentioned in my comments, I think, the Mexican earthquake was probably a little smaller than what was initially expected. I think our scientists believe that there's a higher chance because of where the main Mexico City earthquake occurred, that liquefaction could play a role. That always takes longer to report. Harvey is, again, much more of a flood event, as we all know. I think the thing with Harvey to watch is to see if there are large risk losses that emerge from Harvey.
That's probably been a little slower than expected in not seeing some of those major risk losses emerge. Maria is just a tricky one. I think of all the ones where there's the potential for there to be the biggest disconnect between what's ultimately recognized on balance sheets and what's reported so far, it's probably Maria for the reasons I mentioned earlier.
Right.
I think additionally, when there's this many events in a quarter, it is very complicated to come up with a net risk exposure. Even for us, across the events, our retro programs are largely shared. When first Harvey happened, you put your retro allocation against that. Once Irma happens, then you're doing an optimization between two. Mexico, Maria, you're doing optimization across four. I think as thinking about it with the net number reported and then having this many events, there's going to be a lot of movement potentially between the net numbers, but not necessarily as much on the gross, depending on how the ceded allocations are distributed.
That makes sense. If I can throw a hypothetical out, and let's say that you guys are being conservative and everyone else is being as accurate as normal, and therefore the industry losses are just a lot less than we think. Let's say it comes in that it's only $50-$60, not $80-$100. Does that tell us anything about the cat models? Is it possible the cat models just are overestimating? Just like we got used to them underestimating most events, maybe they've actually overestimated all three of these events and we need to have sort of the opposite of 2005, where the cat models went up, now the cat models need to go down. Is there anything indicating that?
I think of the cat models different than the cat modeling firms making an estimate of the industry loss.
Okay.
When we talk about return periods for these events, that's us going into our model, making an assessment as to what we think the loss is, and then putting a return period on it. We think that's reasonably accurate. We may have the industry loss wrong, but the return period to industry loss ratio will be right. I think they're assessing the loss, and even just take the wildfires. There's a big difference between what AIR is reporting and what RMS is reporting, which I think is interesting knowing that California's a pretty thoroughly modeled state.
Right.
I think one needs to separate their estimate of the insured loss to the precision one can extract from properly using the model.
Got it. Okay. Just last one real quick is, I know I'm not going to probably get you to change your disclosure on PMLs or cat loads or anything like that, but if I can maybe ask in a simpler way as a proxy. If I were to look at your last 10 years' cats and just add them up and divide by 10 as a % of EP or capital or whatever, is the last 10 years unrepresentative of what you would expect for the next 10? I know the book only shifts and you may be buying more retro or less or this or that, but is that a reasonable starting point for us? Or is there something you would say, no, that's just a terrible way to do it?
I've never done that, so I can take a look at it. I think of just very simply, much of the risk we write has a expected loss lower than 10. You would need a longer timeline in order to have a better understanding as to how the book is exposed. If I were to think about it, kind of doing what I would call a burning cost model to understand a cat book, I would tend to bias myself to a much more of a stochastic approach using a much more robust event set and longer time frames.
Of course, I was just trying to use something that I have access to, but I understand. All right. Thank you for the answers.
Okay, thanks.
Thanks.
This concludes our question and answer session. We'll turn the call back over to Kevin O'Donnell.
Thank you very much for your attention and for your participation in the call. We hope that you found it informative, and we look forward to speaking to you after the quarter close. Thank you.
This concludes today's conference call. You may now-