Good day, ladies and gentlemen, and welcome to the Q3 2017 Arch Capital Group earnings conference call. At this time, all participants are in a listen-only mode. Later, we will conduct a question and answer session, and instructions will follow at that time. If anyone should require assistance during the conference, please press star then zero on your touchtone telephone. As a reminder, this conference call is being recorded. Before the company gets started with its update, management wants to first remind everyone that certain statements in today's press release and discussed on this call may constitute forward-looking statements under the federal securities laws. These statements are based upon management's current assessments and assumptions and are subject to a number of risks and uncertainties. Consequently, actual results may differ materially from those expressed or implied.
For more information on the risks and other factors that may affect future performance, investors should review periodic reports that are filed by the company with the SEC from time to time. Additionally, certain statements contained in the call that are not based on historical facts are forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995. The company intends the forward-looking statements in the call to be subject to the safe harbor created thereby. Management also will make reference to some non-GAAP measures of financial performance. The reconciliation to GAAP and definition of operating income can be found in the company's current report on Form 8-K furnished to the SEC yesterday, which contains the company's earnings press release and is available on the company's website. I would now like to introduce your hosts for today's conference.
Mr. Dinos Iordanou, Mr. Marc Grandisson, and Mr. Mark Lyons. You may begin.
Thank you, Crystal. Good morning, everyone, and thank you for joining us today. This past quarter, natural catastrophe significantly impacted the industry with three major hurricanes, large earthquakes in Mexico, and other ongoing events that are likely to make 2017 one of the costliest, if not the costliest for the insurers in history. Although we were impacted by these losses, our diversified platform and good investment performance made this quarter an earnings event for us, and I'm pleased to note that our book value per share increased by a penny in the quarter to $59.61. Based on expected industry catastrophe losses in the range of $80 billion-$100 billion, Arch estimates after-tax net losses of $320 million from all events in the third quarter. We have arrived at this estimate through a combination of top-down model industry loss estimates and a bottom-up review of reported losses from our insureds.
I am sure many of you are puzzled, as I am, that by aggregating all the reported losses so far and adding an estimate for those companies and markets that have not yet reported, we cannot get anywhere close to the bottom end of the model range. While those losses are significant, this quarter again demonstrated a core principle of Arch's risk management philosophy. As expected, returns from property and property cat risks have declined over the past several years. Our underwriters were focused on risk-adjusted returns and accordingly significantly reduced their writings. Implementing this approach is not easy, as competitors in both the traditional and alternative markets have accepted business at margins we deem inadequate. However, the weakness of these margins are only exposed in a volatile catastrophe year like 2017.
As a result of these catastrophe losses in the third quarter, we are reporting a net loss of $52 million or $0.39 per share. On an operating basis, a net loss of $107 million or $0.79 per share. Arch remains in positive territory for the nine months with net income of $363 million and operating income of $260 million for the nine months ending September 30, 2017, and we expect a positive year by year-end. Turning to third quarter results. Our reported combined ratio was 110% in the quarter on a core basis. Mark Lyons usually defines that, and in a moment, he'll give you the definition, and includes 30.7 points of catastrophe losses, partially offset by five points of favorable prior year development from all three segments in the quarter.
Catastrophe losses pushed our reinsurance segment combined ratio to 127.4 combined ratio in the third quarter. Although excluding catastrophe activity and prior year development, the combined ratio improved to 96.9 sequentially in the third quarter of 2017 as large attritional loss activity in our property unit moderated. Catastrophe losses also increased our reinsurance group's combined ratio in the quarter, which rose to 138.7% combined ratio. Excluding catastrophes and prior year development, the combined ratio was 98.9 in the third quarter compared to 99.4 in the second quarter of 2017. This quarter's catastrophe events also demonstrates one benefit of a diversification into a mortgage insurance business with earnings from mortgage substantially offsetting losses in our other two segments.
While we expect a temporary increase in delinquency notices to occur from these natural catastrophes, we believe the language of a master policy generally preclude liability for the mortgage insurer when a home has suffered extensive physical damage, and historically, actual losses from catastrophic events such as Hurricane Katrina have had minor effect in the MI industry. Our mortgage segment, excluding prior year development, improved its combined ratio slightly to 42.2% from last quarter. Integration of the U.S. primary mortgage operations continued to progress very well and it remains on or slightly ahead of our targets. Marc Grandisson and Lyons will give you more flavor on that. Net investment income for the third quarter of 2017 was $94.1 million or $0.70 a share, increased marginally from the second quarter.
As you know, we manage our investment portfolio on a U.S. dollar basis, which on a U.S. dollar basis, was a positive 160 basis points for the quarter and 126 basis points on a local currency basis. Our equity and alternative portfolios were the principal drivers of the quarter's returns. Before I turn the call over to Marc Grandisson, I would like to discuss our PMLs. As we mentioned last quarter, we are also reporting to you our exposure to mortgage risk from a systemic stress event, or what we call a realistic disaster scenario, or RDS, which at the end of the quarter stood at 15.7% of tangible common equity. We have begun using tangible rather than stated equity as a result of the United Guaranty Corporation acquisition, as we believe that this is a more prudent risk management base.
Our property cat exposures are substantially the same as last quarter with our one in 250 year peak zone, which continues to be the Northeast PML, which is at 6.6% of tangible common equity. You will see additional PML numbers in our 10-K which is being filed. I will now turn over the call to Marc Grandisson to comment on our operating units and market conditions before we go to Mark Lyons for financial reporting. Then we'll take your questions. Marc?
Thank you, Dinos. Good morning to you all. This was an eventful quarter for the millions of people directly affected by catastrophes as well as the insurance industry and for those of us at Arch. Before discussing the events of the quarter, it's worth commenting on some of the recent management changes that have occurred here. One of our senior executives and an important member of our management team, David McElroy, decided to retire. Fortunately, our deep bench included our highly regarded colleague, Nicolas Papadopoulo, who agreed to take on the insurance leader's role. In turn, this allowed Maamoun Rajeh to step up to lead the reinsurance group. Both have been with us since 2001 and are terrific executives with a proven track record.
We thank Dave for his leadership and are pleased that he has agreed to continue making his contributions to Arch as a respected senior advisor to the company. I am also looking forward to Nicolas and Maamoun flourishing in their new roles. Turning now to the third quarter cat event, I would like to add my thanks to our underwriting teams who have demonstrated discipline as reflected by the decrease in our property writings over the last five years in response to declining premium rates. In cat-exposed business lines on a gross basis, Arch wrote over $800 million of property and marine premiums in 2017. At the right risk-adjusted price, we have available capacity for additional property risk.
To put again our capacity in perspective, our one in 250 single event PML is still very low at 6.6% of tangible common equity, which allows us to increase property writings should pricing improve materially. We believe that the third quarter cat event will prove difficult to assess, especially on the insurance side. The potential issues with flood and business interruption coverages as well as the assignment of benefits issue in Florida create uncertainty in the estimation process. As a result, we estimate a greater share of our aggregate loss will come from the insurance group. Turning to current property market conditions, we are still evaluating our tactics as the market remains in flux. If 2005 is any guide, and we have some reason to believe that it could be, it will take several months for the market to find an equilibrium.
For cat exposed business, we believe that substantial rate increases are required to achieve an acceptable risk-adjusted return. Focusing on the P&C insurance market conditions, they remain challenging, although we have seen some rates stabilize towards the end of the third quarter, particularly in the property sector. Most other areas had continuing rate adequacy erosion. After factoring in rate changes of a positive 190 basis points and an overall loss trend year-on-year of approximately 200 basis points, we had small margin erosion of 10 basis points for all lines in the third quarter in our U.S. P&C insurance operations. Our low volatility businesses continue to achieve rate increases while the more complex, high capacity, more commodity-driven lines of business continue to see rate decreases.
Our third quarter view, consistent with the last several quarters, was that most areas of the P&C insurance had expected returns that are below our threshold to grow our writings in these lines. Turning now to reinsurance. We continue to focus on the opportunities that have relative rate strength. We are hopeful that more favorable returns will be available in property, but as always, we will have to see how the January 1st renewals settle before we have a clearer view of the opportunity. Our reinsurance net written premium increased by 35%, largely due to a specific loss portfolio transfer transaction, as well as some reinstatement premiums from the cat events. Excluding the effect of these distortions, our growth was more modest at 8%. This growth is due to our seizing niche opportunities such as motor and some specialty reinsurance.
Our property writings decreased due to market conditions as our reinsurance group continued to focus on margins. Now switching gears to mortgage insurance or MI. As stated in prior quarters, the earnings contribution from our MI segment is again proving to be a diversifying offset to difficult conditions in our P&C operations. Our MI segment expense ratio improved to 20.6% at the end of the third quarter while our new insurance written or NIW in the U.S. was $17.7 billion for the third quarter, a slight increase of 2% over the second quarter, largely due to our targeted decrease in single premium business. Although not all MI companies have reported yet, we estimate that Arch U.S. MI's market share may have dipped slightly below 24% in the third quarter of 2017. This is consistent with our expectations as we focus on improving the risk-adjusted returns of our mortgage portfolio.
In the third quarter, 80% of our U.S. NIW came through our risk-based pricing platform, which as of last week, is fully integrated into a single RateStar module. We are writing primary U.S. MI business with an expected ROE still above our long-term target of 15%. The overall quality of the risks written remains very strong, and we continue to experience favorable developments in our U.S. MI reserves, consistent with what you may have heard from others. Arch wrote five new U.S. GSE Credit Risk Transfers, or CRTs, bringing our total risk in force from them to approximately $2.5 billion at the end of the third quarter of 2017. Average yields in the CRTs remain healthy, and ROEs are above our long-term targets. Competition in this space is heating up, and this could affect our risk appetite for new writings in coming quarters.
We executed our first mortgage capital markets transaction this week, Bellemeade 3, which is a risk management tool in helping manage our capital. We view the market's growing acceptance of these securities as confirmation that the mortgage market is originating products of very high credit quality with low risk of default. We will be preparing for opportunities that the market allows, but as always, we will be disciplined and opportunistic. Here's Mark with the more detailed financial analysis. Mark?
Great. Thank you, Mark, and good morning all. I'll make some summary comments for the third quarter, all on a core basis. As a refresher, the term core corresponds to Arch's financial results excluding Watford Re or the other segment, whereas the term consolidated includes Watford Re. I did notice that a lot of the analyst reports, the preliminary analyst reports, were pulling combined ratios that are the consolidated combined ratios, which is really not accurate. That's 180 basis points more than core. If you take the 11% rather than 100% of Watford, it only moves up 20 basis points to 110.2, and that's the proper way to look at it. It's worth noting that operating earnings per share for the quarter reflect, in accordance with GAAP, the use of basic shares rather than fully diluted shares since the company incurred an operating loss.
This translated to nearly a $0.03 increase to the operating loss per share, since approximately 4.4 million fewer shares were utilized in the operating EPS calculation. On a year-to-date nine-month basis, fully diluted shares are utilized since the company has positive operating income, also per GAAP. Claims estimates recorded in the third quarter from 2017 catastrophic events net of reinsurance recoverable and reinstated premiums, as Dinos mentioned, were 30.7 loss ratio points compared to 1.3 loss ratio points in the third quarter of last year on the same basis. Approximately $348 million of pre-tax losses and nearly $320 million of after-tax losses emanated from hurricanes Harvey, Irma, and Maria, along with the Mexican earthquakes, with the balance reflecting minor adjustments to estimates on catastrophic events that occurred in the first half of the year.
Approximately 62% of the quarter's catastrophic loss estimates stem from the insurance segment and 38% from the reinsurance segment. In total, Hurricane Harvey accounted for 37% of the quarter's cat losses, Hurricane Irma accounted for 45%, and Hurricane Maria 16%, and 2% in total for the balance. As for the California wildfires that occurred during the fourth quarter, our early read is that it will be about $25 million to $30 million, which is roughly equal to our quarterly cat load. As for prior period, pure net loss reserve favorable development, 5.4 loss ratio points was reported in the quarter, led by the reinsurance segment with approximately 60% of the total, the mortgage segment accounting for approximately 35% of that favorable development, and the insurance segment with about 5%.
Approximately 70% of the mortgage segment favorable development emanated from the U.S. primary first lien portfolio. About 22% stemmed from net favorable development, resulting mostly from subrogation recoveries by the second lien portfolio that came over as part of the UGC acquisition, and that is a runoff operation. The reinsurance segment net favorable development was across short, medium, and long tail lines and was scattered throughout the 2002 to 2014 underwriting years. The overall calendar quarter combined ratio on a core basis was 110%. When adjusting for cats and prior period development, the core accident quarter combined ratio was 84.4% compared to 93.4% in the third quarter of 2016, driven mostly by the mortgage segment's accident quarter combined ratio of 41.2%.
The reinsurance segment accident quarter combined ratio excluding cats of 96.9% compares to the third quarter of 2016's 96.4%, while the insurance segment's accident quarter combined ratio excluding cats was 98.9%, as Dinos mentioned, compared to 97.7% in the third quarter of 2016. The reinsurance segment also participated in a $45 million premium-sized retroactive reinsurance transaction, as Marc Grandisson noted, that contained sufficient risk transfer under GAAP for insurance accounting treatment. The reinsurance segment calendar combined ratio is 4.5 points lower this quarter due to the inclusion of this transaction. Without this transaction, the reinsurance segment calendar quarter loss ratio would be 40 basis points higher. The expense ratio would be 4.1 points higher. This should be considered when examining the versus combined ratio. Now, this transaction also created distortions on a current accident quarter basis.
When adjusted, the reported current accident quarter loss ratio and combined ratio of 63.5% and 96.9%, respectively, becomes a lower 58% accident quarter loss ratio and 96.4% combined ratio, thereby revealing stronger underlying fundamentals. The reported insurance group accident quarter, excluding cat, loss ratio increased approximately 30 basis points quarter-over-quarter. After controlling for large attritional losses, actually increased by approximately 80 basis points due to the lower level of such losses this quarter versus the third quarter of 2016. As a result of the ongoing competitive conditions in the P&C markets, we continue our conservative approach towards current accident year loss picks. However, the difficult conditions in the insurance and reinsurance markets were more than offset by the continued improving profitability of the mortgage segment, amplified with their net earned premium to being a larger proportion of the total.
The mortgage segment's accident quarter combined ratio, as stated earlier, improved to 41.2% from 55.8% in the third quarter of last year, and their net earned premiums represented similar to the last few quarters, about 25% of the core net earned premium, compared to only 9.1% in the third quarter of 2016. Remember that in the mortgage segment, accident quarter has a different connotation than in the P&C world, as a more similar concept to claims made businesses in the P&C space, since the notice of default defines the assignment to the appropriate quarter. Earlier, Marc Grandisson had commented on the Bellemeade 3 mortgage-linked note that was executed this week. The cost associated with this 10-year term cover will be approximately $11 million of ceded premium for the first year, first fiscal year, and will then reduce as the underlying unpaid principal balance is amortized through the securities.
Similar to last quarter, there were some non-recurrent costs in the third quarter resulting from the UGC acquisition. This quarter, such non-recurrent costs totaled $3 million even. In contrast to last quarter's $2.7 million and the first quarter of 2017's $15.6 million. The sources of costs emanated from severance, outplacement, and trailing UGC transaction costs. During the third quarter, there were 56 mortgage employees that were noticed for an October 1st termination date. In accordance with GAAP, similar to last quarter, the severance costs associated with these employees were accrued. This brings the year-to-date employee reduction total to 338, and additionally, 28 contractors have been eliminated in the quarter, bringing that year-to-date total to 87.
When combined with the actions taken in the first and second quarters of 2017, the cumulative quarterly run rate employee salary savings are $8.3 million per quarter, which would be $33.2 million on an annual basis. We will continue to comment in future quarters about any other actions that are taken and their associated financial impact. Pure severance costs for the first nine months of 2017 totaled $13.2 million, and given the nature of these expenses, and consistent with last quarter, we have excluded this $3 million from operating income as they are not part of our true operating performance. As we respect the effective tax rate with our changing portfolio and geographic mix, the third quarter of 2017 tax rate on pre-tax operating income of -7.4% requires some clarification.
The US property and casualty companies within the US tax group, although incurring material underwriting losses due to the catastrophic events discussed previously, had these losses overshadowed by the gains emanating from our US mortgage unit. This put the US tax group in a tax-paying position even though the company overall sustained operating losses. The underlying effective annual tax rate grew to 19% even for the year for the same basic reason, which is a lower level of estimated full-year operating income than was forecast as of last quarter. The increase in the estimated effective annual tax rate causes the first two quarters of 2017 to be reevaluated at this now higher rate, and this impact reduced earnings by $0.20 per share. Therefore, this tax adjustment is directly derivative from the catastrophic loss estimates.
As a reminder, our tax rate is affected by varying mixes of income, by geographical distribution, and any associated changes in local tax rates. As for after-tax operating income earnings per share accretion realized in the third quarter from the UGC acquisition, we examined our results with and without the impact of the acquisition, and the accretion remains consistent with past quarters and continues to move towards the initial 35% target. In a similar vein, I'd also like to clarify, as I did last quarter, some aspects of the profitability contributions for our three underwriting segments: insurance, reinsurance, and mortgage. As stated earlier when discussing the earnings per share accretion, the relative underwriting income is distorted by the catastrophic events this quarter.
If you adjust for long-term cat load, as well as allocating intangible asset accretion and debt servicing costs and dividends and so forth by units, on a year-to-date basis the mortgage segment accounted for 60% of pre-tax operating income, and the balance being 40% split, with the reinsurance segment seeing 23% and insurance segment 17%. Management, however, continues to evaluate performance for the operating segments primarily by underwriting income and views and manages the investment function on a total return basis across all three segments. This alternative view that we just provided may provide you additional insight into our sources of overall profitability. It should also be stated that the increased profit streams provided by the mortgage segment permitted our after-tax catastrophe losses to only represent less than two quarters of operating earnings using the previous four quarters as the reference baseline.
On a GAAP basis at September 30th versus year-end 2016, our debt to total capital 19.3%, and total debt plus preferred to total capital ratio was 26.3%, down 240 basis points from year-end 2016. This leverage reduction was due to a combination of paying down $100 million of our revolver facility debt this quarter and the growing common equity over the last nine months driven by retained earnings. During the third quarter, we partially redeemed $230 million of our Series C preferred shares. These shares carried a dividend rate of 6.75% and were replaced with a new Series F that achieved 5.45% annual dividend. This amounts to a $3 million annual savings. There was an associated charge of $6.7 million to net income, not to operating income or book value, to recognize the original costs associated with the Series C issuance as a result of the redemption.
Core operating cash flows were $440 million, up $19 million relative to the third quarter of 2016. This increase was primarily driven by growth in net premiums collected, primarily in the mortgage segment, partially offset by a higher level of net paid losses in the P&C operations. Book value per share increased that mighty $0.01 per share, and tangible book value increased 0.4% sequentially this quarter. Tangible book value growth outstripped book value per share growth due primarily to intangible asset amortization. With these introductory comments, we're now prepared to take your questions.
Thank you. Ladies and gentlemen, if you have a question at this time, please press the star and then the number 1 key on your touchtone telephone. If your question has been answered or you wish to remove yourself from the queue, please press the pound key. If you are currently using a speakerphone, we would like for you to please pick up the handset before asking your question. Again, ladies and gentlemen, that's star 1 to ask a question. Our first question comes from Elyse Greenspan from Wells Fargo. Your line is open.
Hi, good morning.
Good morning.
My first question, just going back to some of your initial commentary on the conference on the market outlook. Dinos, it's in the past, and Marc as well, you guys have mentioned you kind of told your underwriters, "Put your pens down. Let's wait for the market to turn." Are you thinking about telling them to get more excited and get ready to write more business at January one? Or is it your expectation, if rates are more, and I guess this is more a reinsurance question
More in the impacted lines. Would that be something that we wouldn't really get more firmer pricing until some of the U.S. renewals later on in the year?
It's a good question, but it's a complicated question. Let me say this. We look for moments like this as a company. By being patient in the years that pricing is not good, we get excited when pricing might get much better. We don't know yet as to where this thing is going to go. Let me share some facts with you. If you add the reported losses so far, it's $31 billion plus a number of major facilities haven't reported, meaning the retro alternative market, Liberty Mutual, Berkshire Hathaway, State Farm, et cetera. Even if you come up with some good estimates for those losses, we might get to 45, maybe we get to 50. It's a long way from 80 to 100.
I believe that either the models, they're predicting much bigger losses, which nobody seems to have that point of view, because everybody's agreeing that this is an $80 billion to $100 billion aggregate event for the industry, or maybe even more. Stay tuned in further development. If we go back and we look at historically, Wilma developed by 68%, Sandy developed by 70% for the industry, and Katrina, we got it almost right. It only developed by about 20%. I anticipate upward development to happen. I believe that we were very prudent in establishing our numbers because you have to do bottom-up and a top-down approach, and we actually weighted more the top-down approach, which is, how big is this? What is our market share? Where do we have exposure?
As Mark said, we were cautious with our insurance group, and we put what we believe is reasonable numbers for us. Depending what happens in the next quarter, and it depends what happens to pricing, I think we'd be ready to do quite a bit more if the returns, they would be acceptable to us. With that, I'm going to turn it over to Mark because he works with the units more day to day, and he will give his comments, too.
Yeah. I think more practically, what's happening right now is we're in a planning process, right? We have a portfolio, which I mentioned, a substantial portfolio of property that's going to be renewed over the next 12 months, hopefully renewed over the next 12 months. Right now, both our insurance teams and our reinsurance teams are actually going through the portfolio and assessing which one would need to get a further rate increase or a higher rate increase, and sort of planning ahead as to how we're going to react. The problem that we're seeing right now is there aren't that many renewals, as you pointed out, for the reinsurance up until the 1st of January. We have a lot of time and a lot of discussion, and as you can appreciate, there's a lot of positioning by the various players in that market.
We'll probably not know realistically how the reinsurance market plays out until very late in December. We're going to have all the things laid out in front of us, knowing exactly how to react. On the insurance side, we've seen a couple of things emerge, a couple of rate changes. The first thing we started hearing is there's no more rate decrease, which is a good place to be. The rate increases are sedate for now. They're coming up, single to double digit, but it's very sparse. There aren't that many things renewing right now. Again, we're very much on a reactive mode right now, evaluating on a weekly basis. I know Nicholas and Maamoun and their respective teams are talking constantly as to what's going to happen.
Yeah. One more comment about the available capacity in the marketplace is very hard to be estimated because some players, they have significant capacity, only because they believe that they have good support from the alternative markets and the retro market. The underpinning of their capacity on a gross basis is because that market was significant, and it was at reasonable pricing. As a matter of fact, we were buyers in that market. We have no view yet. That's what is still in our minds. We have no view as to what's going to happen to the market, and what will be the reaction of companies that they said, "Hey, I can do more on a gross basis because I got all this protection behind me, so I can be a little more aggressive in the marketplace and maintain customer relationships," et cetera.
It's all interrelated, and as Mark said, our teams are there. We're willing and able. We have capacity, but it will depend all on expected returns based on pricing.
Okay, great. I appreciate all the color. A couple of other numbers question. In terms of the tax rate, I guess, should we just expect the go back to around 14% in the fourth quarter and onward?
Well, think of it this way, Elyse, the 19% contemplates a full year view. If the fourth quarter is more average with a normal cat loads, and we talked a little about California wildfires, the 19% should hold.
Okay. In terms of the mortgage segment, can you tell us because kind of what the delta in terms of some of your additional severance costs, what expenses we could expect to come out in the fourth quarter versus the third quarter in terms of your operating expenses?
Well, I quoted salary expenses to you, that's a run rate of 8.3 per quarter. To the extent that there's any other actions contemplated there, we only talk about that once they've happened for a lot of employee morale and other aspect reasons. There are associated additional costs associated with other kinds of compensation, employee benefits, and so forth. We really don't comment on that. There is a lot of ongoing work with IT that over the next couple of years, when redundant systems start to be peeled away, all those savings and associated license costs will come to fruition.
In terms of the acquisition cost ratio itself also came down in the quarter. How should we think about that in terms of the mortgage business going forward?
Well, first off, you should think about them in total between AC and OPEX, because you may think back to what we did at the end of the year, last year, because it's not like a direct sales force. Some of that goes into acquisition. Remember, as a result of the purchase, all the DAC got really written off. It's really buried within the intangible assets. As one and subsequent, as we write more business that's on a single basis, it creates a UPR. That UPR grows, DAC grows, and that has to be accreted over time. You should be seeing an increase in the acquisition ratio.
Okay, that's great. Thank you very much.
Sure.
Welcome.
Thank you. Our next question comes from Joshua Shanker from Deutsche Bank. Your line is open.
Yeah, thank you. I was wondering if you guys could help me out a little bit and think about the move to getting less exposure to property cat over the last call it 5 years, even longer, maybe. How much of that is relying on the retro market to be affordable, and how much of that is in the gross premiums of the company?
Well, I'll give you my comments and Marc might add to it. Listen, independent of the retro market, we always look what is the pricing and what is the market accepting as an expected return on that business. When that business start going to single-digits expected returns, we're starting to lose a lot of interest. Having said that, we do have a customer base that we want to maintain and continue to service. That's when you look around and you say, maybe I can buy more protection if that protection is available, and it will allow us to maintain relationship. The reduction, clearly the reduction in our writings, it was driven by our view that the pricing of that business was not adequate for us and for that reason.
You can't go to zero, as a matter of fact, it's more difficult on the insurance side to cut as much back than on the reinsurance side because you got brokers, agents, relationships, or customers. In essence, sometimes you do hedge your bet by buying more protection, and we've done all of those things. If the market improves significantly, you might see a different approach. We might keep a lot more net, and we have capacity to do it. Also you might see us expanding our exposure base because we like the pricing. Don't forget, we'll understand. We're in the business to deploy capital and make money for shareholders, and we're not unwilling to take risk. We just were not unwilling to take risk at inadequate pricing.
Josh, the way we look at the reinsurance purchasing on the reinsurance side, they're still partners of ours, and we fully expect them to be there going forward next year and the year after the market were to present itself. I would say that they certainly helped us managing the net exposure because the market was indeed getting softer for the last 5 years as you know. I would say that between our growth, our appetite to the market and relying on our partners, probably a 50/50 split between our management of the exposure. That probably allowed us to stay a bit longer while not be overly reliant, if you will, on the retro placement. Because at the end, they were nice to have.
I think that our partners will still be there for the long haul, but we're not relying on this to ride the business going forward.
Well, thank you very much. One other unrelated question. In terms of the amount of earnings being suppressed in the UGC transaction due to your reinsurance relationship. Sort of two questions. One, I think at the investor day, you spoke about the potential that you might think about doing that for a longer period in terms of a reinsurance relationship with AIG. I just want to hear an update about that. Two, AIG provides data about how much that relationship is helping their P&L, but it seems to jump around a lot. Is that revenue steadier in your books, or can we use AIG's numbers about UGC to understand the degree to which Arch is currently under-earning its potential?
Good question. I think I'm going to turn it around a little because I think your statement actually answered your question. They don't have exact mirror accounting to us. They come up with their own views and estimates, but I think from an EP or earned premium, net earned premium, ceded earned premium to them assumed earned premium. I think that's a reasonable way to look at it. Clearly, we're in it starting the downside of the effect of the AIG quota share diminishing each quarter.
Whether you would renew it, or is there any talk about doing that?
No, there's not a renew of that. That was 2014 through 2016.
You weren't doing anything on the 2018 year with AIG or anything like that?
Well, if we did, it wouldn't be bound yet, and I wouldn't be speaking about it.
Of course.
Everything is possible, Josh.
Thank you very much.
Okay.
Thank you. Our next question comes from Amit Kumar from Buckingham Research. Your line is open.
Thanks, and good morning, and thanks for taking my questions.
Amit, speak up because we can barely hear you.
Is this better now?
Yes.
Oh, thank you. Two quick follow-up questions. Number 1 is just going back to the discussion on industry losses. I want to be clear that I understand this. Do you feel, based on your statements, that the industry loss will eventually get to the models numbers we're hearing about? Do you have a feeling that these models, and there was a lot of divergence between the numbers, overestimated the numbers? I guess I want to understand if we are overestimating it, clearly the market does not turn, and any optimism is a bit overdone.
Yeah. First and foremost, all I'm saying is that everybody seems to congregate against the $80 billion-$100 billion, both the modeling agencies, if you take their average or point estimates for each of the storms, and also a lot of our competitors, including us. I think the models probably they're projecting the right number. I don't know every company's book. I'm just making aggregate comments. Something doesn't add up. Either that number is going to come down and your hypothesis is correct. If it comes down, people they're not going to feel it as much on their P&L, so in essence, the market correction might be toned down. If you go to historical performance, models never overestimated losses in the past.
Usually, our early estimates as an industry, they were below what they ended up. I'll leave you the judge of whether you think it's going to happen. I think we're going to have an effect that is not going to be felt totally for another two or three quarters before we know where these things are going to end up.
Moving away from modeling for one second, Amit, I think if you park aside Maria, because this one has probably the most uncertainty in terms of modeling and ultimate projection, it's really hard for us internally based on the modeling and based on what we know in terms of damage, what happened, it's very hard to even consider that both Harvey and Irma are less than $25 billion each. That already takes us to the $50 billion. Before even considering Maria and the other events that happened. That's why we're sort of building from there to, it's not a far reach to get to $70 billion, $80 billion, $90 billion. I guess only time will tell. I think it's very hard in isolation to just look at these two losses and think that they're going to be like $15 billion each. It's very difficult for us.
That's the implications of the industry losses that we've seen reported so far, that these two losses will be a lot less than they seem at this point.
Got it. That's actually helpful. The only other question I had was, I guess, going back to what Elyse and maybe Josh were asking. If you look at the sub-segments in the reinsurance segment, the interplay between different segments. Even based on the market opportunities, should we be rethinking about, I guess, the total top line and the returns differently? Is that premature, or how should we think about, I guess, the capital allocation between the different segments at this juncture?
It's very difficult because I can't answer you because I don't know myself. What guide us is market pricing. At the end of the day, if you tell me what it's going to be on January 1, maybe I can give you some projection. Not knowing that, I really don't know. What we gear our people to do is to look for every opportunity available. We'll evaluate it, and if we need to be very agile and move capacity to one area versus another, we're there to do it. If we get another big event, and fourth quarter is another catastrophe, and the whole world turns upside down for January 1, I can tell you, there is a lot of actions we're going to take, including looking for additional capital and being the business for our shareholders.
That's what we get paid to do, and we're willing to do it. Marc?
Yeah, right now we're evaluating exactly this because back in 2005, it was pretty clear that the risk-reward was much more advantageous to the reinsurance team. We, at that time, allocated 80% of the cat capacity to the reinsurance team, and 20% therefore for the insurance group. This time around is different, right? Again, we don't know where the reinsurance market's going to go. A lot of these losses are retained within the company, so there might be a different outcome on the insurance side. Right now, as we speak, the two guys leading our insurance and reinsurance are going through it to see what kind of return they will be expecting next year, and therefore giving us a plan of action as to what they're going to do.
Okay.
They still have to collect information as we speak.
The big question mark here, especially on the insurance side, is how much capacity has existed through these MGA facilities. That is somebody having the pen for somebody else, and at the end of the day, their compensation is, I got to write more, I got to write more, I got to write more, because they're commission-based compensation. Will these facilities survive the event and get renewed, and that capacity is available, or they go by the wayside, or they get terms that they move the market upwards? We don't know that, because some of these facilities, as it's toning down what's happening in the market, they haven't expired, or they might require a six-month notice. They're going to be used until they can be used.
If they're in place now and you don't think it's going to be renewed, you're going to use it up to the last minute. We don't know. There is a lot of things up in the air. We have our heads to the ground and we're looking at all these opportunities. It's not a clear picture yet.
Yeah, got it. I'll stop here. Thanks for the answers and good luck for the future.
Thank you.
Thank you. Our next question comes from Brian Meredith from UBS. Your line is open.
Yeah, thanks. A couple questions here for you all. The first one, I'm just curious, and maybe I missed it. Did the RDS that you provided in the mortgage include the recent Bellemeade transaction or not? If not, what would that look like with the recent Bellemeade transaction?
It does include it. What would be the number without it? Do you have the number, Marc?
Well, did you quote what it was?
The percentage?
15.7. It's not much of a movement of it. Its benefits from that's $200.
Yeah, $200 million.
Yeah.
Yeah.
Okay, great. I'm just curious. You guys have Watford Re out there, which is more of a liability facility. Are there any thoughts of creating a facility that's more dedicated to severity or cat, if indeed the market doesn't rise enough to maybe be acceptable to you on a kind of a net basis, putting a lot more cat risk on your balance sheet? Are there opportunities potentially to do something more from a gross perspective and get the income?
We have two considerations there. One is, if the market improves significantly, we'll use a lot of our own capacity, but also we'll be very much interested in managing third-party capital because we don't want to change our risk profile. We've done that after Hurricane Katrina with Flatiron Re. If the market doesn't move, and there is people that they will be willing to get our underwriting skills and they're willing to accept maybe a little less return than we will, we're not opposed to managing money in that fashion either. That's not our preferred outcome. We like the market to get hard so we can write more on our balance sheet and maybe write for partners also.
Brian, I mean, I just appreciate you allowing me to put the plug out there in the marketplace. If anybody's looking to deploy capital in the space, we'd love to talk to them.
Yeah. Thank you for that.
Just a question, and I guess you kind of alluded, I kind of was asking maybe you can put a specific spin. How much more rate do you kind of need in property cat to kind of take more on a net basis, do you think?
Well, right now the returns on the space are in the mid to high single digits. It's not
Expected return.
Expected return. They're not extraordinary. We've actually asked our team, and that's what we're going through right now. We think that to get to a 15+ return, we would need roughly 30%-35% rate increase. We need a substantial increase in rates. What people forget is we have to be careful with looking at a rate change in isolation. I think it's been mentioned on other calls is that we're at a pretty low level compared to history for the last five or six years. This is why, in 2005, when rates went up 10%, 15%, 20%, we were in a very different market. Pricing was a lot better, where it comes from. Now it's not as good by any stretch of the imagination. We will need substantial rate increase to really fully deploy it.
Even then, Brian, as you know, we're very careful with our capital management. We will have to see it, and have a good clarity of it before we commit fully to this. I think it's going to take a gradual price increase. It will take time. It's going to be a dynamic process for us to evaluate as we go forward. If we close your eyes and you roll the tape and you say the rates have gone up 50%, 60%, then I think we would have a lot more appetite to take on a net basis. We'll sort of have to work our way towards this, as we see if the market allows us.
Brian, that would be a composite because you have to take into account the underlying ceding company rate changes as well as what might be on the cat layer.
That's right. Yeah, it's true.
Right. I guess adding on to that, what do you think the possibility of something like that happening, given it doesn't seem like there's been a change in the perception of risk with these events? The last time we had that type of rate increase, you had massive changes in the models.
Well, I won't be too quick to make that judgment yet. I think smart people are going to step back and look at the events and say, "Should we change our mind about how risky this business is and what kind of returns we should expect?" I think there is a recognition right now, Brian, as we speak, when we hear from our producers and from even the buyers of insurance or reinsurance for that matter, there's a recognition that rates need to go up. I think there is a consensus building slowly but surely. The question is how much, if it does go up. I think to answer your question, I would need to know will there be some increase in loss estimation in a few players? Will there be some change to the rating agency perception of risk? Will there be some change to the modeling?
There's a lot of things, unfortunately, Brian, that need to happen for everything to converge to one area, which was more the case in 2005 if you remember. We had two or three things converging at the same time, which really helped it. Frankly, we might be sitting in the next call talking to you guys about this and still not know fully where it's going. In 2005, it took really between May or June, March to June of 2006 to really see the market take hold and really find its footing. It takes a little while longer than we would expect, unfortunately.
Right. Yeah, it was also the release of the new models came out right around that time, right?
You got it.
Great. Thanks.
Sure.
Thank you. Our next question comes from Jay Cohen from Bank of America. Your line is open.
Yeah. Most of my questions are answered. Just, I guess one follow-up on the Bellemeade transaction. Do you see this as the first of others, given that this is a unique thing and you got it through the other one, next one might be a little easier?
Yeah, Jay. It's a good question. We do see that as an ongoing piece of our repertoire to risk management in the balance sheet. Yes.
Yeah. It's not the first one, it's the third one because United Guaranty did two before us, we've done the third one, we're going to use that as a risk management tool, as we might use also traditional reinsurance as a risk management tool.
Got it. Thanks.
You're welcome.
You're welcome, Jay.
Thank you. Our next question comes from Meyer Shields from KBW. Your line is open.
Thanks. Two question on Bellemeade. Is the ceded written premium a one quarter event, or is that going to play out?
Meyer, we can't hear you.
Hey, Meyer, can you please speak up?
Sorry.
You're not coming through. Go ahead.
Is it better?
Hello. You're much better. Much better. Yes.
Okay, great. Sorry about that. I wanted to know whether the ceded written premium for Bellemeade is going to just impact, I guess, the fourth quarter, or will it endure over the life of the contract?
No, that goes over time. You're going to see cessions associated with that every quarter.
Every quarter. The first year is the $11 million.
Yes, that's right.
Then it cascades down.
Correct. Think of it as a laddering. We do another Bellemeade, its first payment will be higher and decrement. Then if we do another one in 2018, there'll be a first payment that's higher and it will decrement.
Yeah, Meyer, the way it works is that, if you look at quarterly, the amortization of the limit, it amortizes over time. It's 5% of risk in force. If there's persistency of 80% per year, you would expect 80% of premium to be paid the next year, then 64% of the $11 million the third year, and so on and so forth. That's how the bond works. The $368 will amortize over time.
Okay, perfect. That's very helpful. Just going to the retroactive reinsurance transaction, that's the second in two quarters. Is there like a building market for that would align with broadly stating reserve redundancies across the industry?
You talking about the-
MI or? No.
No, he's talking about the reinsurance transaction market, the retroactive reinsurance.
The retro. Oh, the loss proportion.
For transfer.
Yes.
Then we're very happy, very pleased with Premia. Premia is way ahead of its initial plan, and it's testament to our team's efforts in working very diligently. I think that like we said before, there's a lot of books of business, specifically on the liability. This is a liability transaction, that's why it's booked in the casualty unit. There's a lot of books of business who have issues and warts, and people are trying to, and I can't blame them, trying to find a new home for it to just move away from it and just put it behind them. This one is really meant to bring finality to that client. I think we're going to see more of these.
I think in a sense of capital management and earnings management, you can expect more companies to look at it. There is a very healthy flow of offers on the book for our premium folks.
Okay, fantastic. Thanks so much.
Thank you.
Thank you. Our next question comes from Kai Pan from Morgan Stanley. Your line is open.
Thank you, and good afternoon now. My first question is on the casualty lines. I just wonder, do you think these increase in the property lines would spread out to the casualty lines? Do you see any sort of change in term underlying loss cost trends?
You got multiple questions. First, we hope, but we don't see it. Second, it's needed, but we don't see it. Your third question, yes, loss cost, even though we're getting slight rate increases on average, we're having difficulty in including us, which we're pretty conservative in our underwriting, maintaining the same level of profitability. As a matter of fact, we lost another 20 basis points of margin between what we believe the loss cost escalation was versus what kind of rate increases we got on average. Marc, you want to-
No, we're not seeing it right now. We think it should happen, but actually, it might actually have a perverse reaction as a result of property lines having losses. People might look at a casualty in a professional line. Now, for us, casualty, Kai, to be clear, encompasses more, especially on the reinsurance, more than just GL. It also encompasses professional lines. It could have a perverse reaction that people use it as an excuse to get price decreases for accounts that haven't had a casualty loss in a while because, well, look at the property accounts. They're giving you losses, we're not giving you accounts. I shouldn't probably be using these arguments on a call to give it to our clients, but I'm sure they'll be using them.
All right. Second question on capital management, you paused buyback for the UGC transaction. Do you think you will return to buyback in 2018 in light of potential market pricing environment to get any better? You might find other place to deploy your capital.
I think if I could just start. I think, Kai, it's completely derivative to the other discussions we are having on property cat and the opportunity. If the opportunity is strong, then there's clearly a decreased likelihood of doing that on a return. They're totally in balance with each other. We need to see what the market is before we can really answer that.
Kai, if I had to guess, I think there are going to be price movements that it will cause us to have more opportunities in the market. We don't know. Stay tuned. Ask the same question in the fourth quarter.
I will. Thank you.
Yeah. Thank you.
Thank you. Our next question comes from Ian Gutterman from Balyasny Asset Management. Your line is open.
We're not telling you what's for lunch yet.
That's okay. I actually had a question for you. Are you familiar with the game Where's Waldo, Dinos?
No.
My kids are, certainly.
Okay, that's good. We'll go with that. What I found out, I know where the missing losses are. Waldo took the profits from selling all his books, and he wrote a bunch of reinsurance. When you find Waldo, you're going to find $50 billion.
Right. Okay.
It could be in London, it could be in Bermuda, it could be in Germany. Waldo likes to hide lots of places. Keep an eye out for Waldo, and you'll find the losses.
When you find it, just call us because then it help us with our strategy going forward.
Yep.
You can look, too. He could be anywhere. He might be delivering lunch today. You never know. My first question is I'm sure you listen to some of these calls, and everyone's strident that they've learned from 2011, they've learned from 2005, and they're not going to late report this time. Where do you think the issues are? A lot of the primary companies are even saying we've closed most of our claims already. It's so obvious we know what our inventory is. There's nothing that can surprise us. You mentioned program. Outside of that, are there other reasons we should see late reporting on what the primary companies are calling simple storms, at least outside of Maria?
Well, listen, I can go into a lot of directions. Flood always has been a big problem. Look at Sandy and how long did it take, et cetera. Second, if you go to Florida, there's a lot of snowbirds that haven't even gone down yet to start looking for repairing their homes. This assignment of benefits issue in Florida, it's going to have escalation of losses. If they know they're closing, and all that is news to me. At the end of the day I'm only going by history, and history has told us that there's always been an underestimation, and we had more positive escalations than negative. People, instead of taking reserves down, they added to over time. Listen, if it's less, it's less, and then maybe my number is too high. How do I know?
Ian, I think the other thing you have to keep in mind is, if you have a portfolio of homeowners, very straightforward, very plain vanilla, and you had a lesser amount of risk, it's probably more likely that you'll be able to close that file a bit quicker. No, the area where we think there's a lot more variability is on the E&S and on unoccupied buildings and the sort. That's going to take a while for everybody to really figure out what the coverages are going to be. You have insurers who are possibly a bit more sophisticated and a bit more better equipped to fight with the insurance companies. Certainly, we're seeing it with the AOB phenomenon in Florida. That certainly doesn't help matters. I guess you have to put things in perspective, and it depends on who you talk about.
I would agree, I would echo what Dinos has said. A lot of the insured population will not probably have anything settled or finalized in terms of loss estimation and indemnity paid for another six to nine months. It takes a while.
You're going to see real losses. A restaurant who had no real damage to the restaurant, but the parking lot was flooded, and customers couldn't have access to the parking lot, and he was going to claim business interruption because of the flooding, et cetera. He was in a zone 5, which is not considered a flood zone, so he had no exclusion on the policy, and his deductible was pretty low. All these things are going to take a long time to get resolved. Like I said, it's our view, and usually not a lot of people agree with our views, but more often than not, we are right.
I agree with your view, Dinos. I think we're in the minority, but I agree with you. Related to that, the one that surprises me the most so far is that most people have Maria as their lowest loss storm, and that's the one where I would think the greatest risk is of BI because in Houston, like you said, there are some issues with getting back on your feet, but it's not that hard, right? In Puerto Rico, who knows how long, right? Why isn't every BI limit on the island a full loss?
Yes. This is very uncertain. I've asked our underwriters right in the area, this week actually. Again, it's still extremely opaque. There's still a lot of no information. A lot of areas don't have power yet, and things are not just back into an order, and it's going to be a long time before we figure out. This is where even a cat event, Ian, could be a long tail event, a long tail phenomenon.
Right.
Which is-
Definitely
Things that we forget as an industry sometimes.
Fair enough. The other comment you made earlier, Dinos, about the gross versus net line underwriting, the companies who are relying on retro so they didn't have to shrink their gross. I don't know how many calls you've listened to this week, but pretty much everyone who's been doing that has said on their calls, I'm simplifying here, "We're going to keep our net lines where we are, and we're going to look to grow our gross." It feels like people want to double down on that strategy, which, if that's the case, A, I'm not sure where they think they're getting all this extra incremental capacity. If so, doesn't that suggest it's harder to get pricing if, sort of to Brian's question, right? Where's the pain? No one wants to shrink.
Well, it's on the premise that the net to gross can work, meaning that there is a robust retro market or a quota share market that is going to reduce their net exposure. We haven't heard from that market yet. We knew it was in the $20 billion range, and some people, they might be estimating maybe 50% or even maybe 75% of it might be gone.
Yeah. Ian, I think you're exactly right. I think not only on the loss estimate from the side and what the ultimate loss is going to be on the industry, to add even more complexity, you're quite right, and we talk about it all the time, this alternative capital, it's a relatively newer phenomenon to our segment, and it brings a little bit more, quite a bit more actually, uncertainty as to what's going to happen. I would add that it might increase the volatility of what could happen, which could be good for us in a way as well. It remains to be seen.
Absolutely. Yep.
Yep.
Absolutely. All right. A couple of Arch specific things. First, the retroactive contract, should I guess that that was you took a share of the deal Premia wrote?
Yeah, 25% of it. Yeah.
Okay, got it.
25, yep.
Looking at your recoverable on the balance sheet, should I assume most of that growth was from the hurricane so that your gross was about twice your net?
Bingo. Yes.
Okay, good.
You do your homework.
Yeah, pretty good, Ian. Yeah.
I try. It was a late night last night, but I try.
Ian, I'm sorry, but A, for the record, I'll say you weren't one of the guys that went to the 111.8 combined ratio.
Exactly. The one part that surprised me on your losses was just the composition. I think I know why, but I just want to hear to make sure. I was surprised at the amount of the insurance. Just like a simple thing I looked at is your insurance loss from these events were basically equal to your 2011 plus Sandy. Right? Is that just because of the geography of things? You're in certain businesses that have more property now than you did then. I guess I was just surprised. The reinsurance isn't surprising at all, but the insurance is a little higher than I thought.
We're an E&S writer.
We believe the flood losses, they're going to have all these questions that I have raised, the business interruption, et cetera. We've been cautious of estimating a loss. Same thing in Florida. We believe that this assignment of benefits is going to have an escalation of maybe up to 30% on the cost of repairs. We factor all that in, and that's why you see more on our insurance team.
Ian, just to add further, if we go through the losses that are reported, we look on the reinsurance side. You see, even if there was some creep up or some factoring the AOB, the flood or the business interruption, it's hard to see a lot more creeping up into the reinsurance layers. It's a pro rata effect on the insurance side. We had to do a more prudent selecting the losses, reflecting those uncertainty between the insurance and the reinsurance. I think there's a lot more uncertainty on the insurance side at this point in time.
Ian, to marry what Marc's comment to some of the ones earlier on that proportional aspect, the extent that the market losses, the industry losses start to decrease because of the proportionality on the insurance side, that'll be shared.
Up or down.
Some programs on reinsurance or retro markets in some companies might be, think of it as event aggregate excesses, and you're really saving for the reinsurer more than you're saving for your net.
Yep, that makes sense. Okay, got it. Very good. Thank you. Enjoy lunch.
Thank you.
Thanks.
Take care.
Thank you. I'm showing no further questions from our phone lines. I would now like to turn the conference back over to Dinos Iordanou for closing remarks.
Well, thank you all, looking forward to talking to you next quarter. Have a wonderful day.
Ladies and gentlemen, thank you for participating in today's conference. This concludes the program. You may all disconnect, and have a wonderful day.