American International Group at Lehman Brothers Property/Casualty & Life Insurance CFO Summit - Accounting, Regulation & Washington Panel Discussion - Final.

Original Source: FD (FAIR DISCLOSURE) WIRE

ERIC BERG, ANALYST, LEHMAN BROTHERS: I think we would like to do is allow everyone else to sort of wander in, and in the interest of keeping on schedule I will get started. Well, good morning. Because I don't know everyone in the room, allow me to introduce myself. My name is Eric Berg; I am the life insurance analyst here at Lehman Brothers.

On behalf not only of my own team, but also that of Jay Gelb, the property casualty analyst at the firm, who is also my partner on this event, I want to welcome you to Lehman Brothers and to this beautiful new auditorium. Greetings too, to everyone who is listening in on the Web.

What I would like to do for the next say five minutes, really no longer than that, is to give you a flavor for the history of this event so that you can understand what we hope to achieve this morning. Several months ago, Jay and I sat down to map out future events for our work together. We quickly agreed that one thing that the world probably doesn't need is another, let's call it, traditional insurance conference.

The idea of bringing together, say, Chief Executive Officers to deliver their standard presentation was not all that exciting to us, because it struck us as kind of having a me-too flavor to it. There are just a lot of events like it. We saw it as an event that would not be particularly helpful or new to investors and therefore would really not be particularly helpful to Lehman Brothers.

Jay and I agreed, in other words, that in order really to give investors something of value, something that they would remember and hopefully look forward to each year, we would have to put on a conference that was entirely different from what Wall Street is currently giving the buy side. And ideally we would want to build on the foundation that it is provided each year by others.

So that is kind of the larger view that I would like to draw out for you. What we're going to do today is to take the big picture, the strategic view, that is provided regularly by CEOs at other more general conferences, and we are going to try to complement that or build on that by homing in on the most important financial and accounting and regulatory and risk management and investment issues facing the companies.

The individuals whom we have asked to help us understand these CFO-oriented issues, they are the CFOs of some of the largest property casualty and life insurance companies in the United States. I don't think I am being -- exaggerating. I don't think I am exaggerating when I say we're truly fortunate to have an exceptional panel. In my more than 10 years as a sell side analyst, I can't remember a time when so many CEOs from so many large companies got together in one venue to discuss the most important issues of the day.

So that is exactly what we hope to do this morning. We are going to move away from strategy to focus squarely and solely on the issues that are front and center on the minds of insurance companies' Chief Financial Officer.

Without question, the conversation, at least on my side of the business, life insurance, is going to get technical at times and it could be narrow. I suspect it will be the same for Jay, but that is okay. Because what we are hoping is going to happen today is that you will leave today's event thinking that you learned more and in greater depth about the financial issues facing insurance companies than you have from any other conference.

In other words, if you leave Lehman thinking that today's conference was not only different and special, but really packed with good detailed information about the finances of insurance companies, then Jay and I will have succeeded in our goals of helping you be more effective analysts and portfolio managers.

I would like to finish up by briefly discussing the format. There is not going to be any prepared remarks. We have said that all along. Instead, Jay and I will be moderating panel discussions, the broad topics of which are in your agenda. While we certainly will be getting the conversation going and will keep it going, if need be, this morning is really for you, the investors. If we have our way, Jay and I will ask the first few questions and then turn it over to you to ask as many questions as you possibly can deliver in the allotted time.

So my team and I will be back midmorning to moderate the life discussion. Again, welcome to Lehman Brothers. It is my pleasure to have Jay lead the property casualty discussion. Welcome.

JAY GELB, ANALYST, LEHMAN BROTHERS: Thanks, Eric, and good morning. I am Jay Gelb, Lehman's senior property casualty insurance analyst. It's my pleasure to introduce the property casualty insurance panel at Lehman's inaugural CFO summit. We have with us on the panel today Steve Bensinger, CFO of AIG; Dan Hale, CFO of Allstate; and John Vollaro, the CFO of Arch Capital. As a reminder we're going to be having separate group luncheons for Allstate, Arch, and Principal Financial. That is 1 o'clock today on the 32nd floor.

The subject of our first property casualty panel is managing catastrophes, which has certainly been a hot topic for the sector. So let me kick it off. The property casualty insurers will suffer hurricane losses approaching $60 billion this year, which is the worst ever for the industry. My first panel for the question is, what lessons have you learned from the 2005 hurricane season, and what will you do differently to manage catastrophe losses going forward? So why don't we start if off with Steve Bensinger from AIG?

STEVE BENSINGER, CFO, AMERICAN INTERNATIONAL GROUP: Thank you very much. Good morning, everyone. It's nice to see some of you here bright and early. I would say in terms of lessons learned, I would categorize them as follows. One, I think we have all learned that the catastrophe models underestimated the frequency and the severity of wind activity. I would say that any model that we have seen dramatically in some cases underestimated the expected costs of the events that occurred this year, most significantly Katrina. We could talk a long time about that, but I will just -- why don't I just maybe point out what I think are the lessons? Then we can go into them in as much depth as you like.

The next one is demand surge. We saw it last year, but it certainly reared itself again this year in a very significant way. I would say that the industry at large and the models as well also underestimated the effect of demand surge on both materials, labor, and all of the ancillary issues associated with the rebuilding efforts.

Then we get to what is happening in terms of geological activity at this point in time. There are many theories out there. Some of the later ones, the more recent ones that are being espoused are theories of multi-decadal cycles. In other words, are we now entering or have we entered a period of increased activity of both frequency and severity?

If you look at what happened this year, in the space of a very short period of time we had three tropical storms that entered the Gulf; and within 48 hours all three of them became Category 5 hurricanes. That is pretty unprecedented from everything that we have seen in the past.

If you look from a historical perspective, seven of the most 10 costly wind events have now occurred within effectively the last year, between August of 2004 and October of 2005. The seven top 10 largest insured events from a wind standpoint ever.

Another lesson that I would say was infrastructure damage. Katrina, because of its characteristics and where it struck, actually caused infrastructure damage that was more closely associated and more typical with an earthquake than with a hurricane. You saw a lot of that type of damage that you would not ordinarily associate with wind but more quake. So what does that mean? How many times is that going to recur?

Wilma, as another example, had a very prolonged power loss that it caused, and that caused a significant business interruption loss for companies like AIG. So in terms of the consequences -- those are some of the significant lessons learned.

In terms of consequences I think from our perspective, we are going to be looking to continue to reduce our zonal aggregates. We're going to have to increase and we are increasing deductibles. And of course, there will be effects on pricing on both the reinsurance, primary, and excess and surplus lines level.

JAY GELB: Great, thanks for that overview. Dan, what are your thoughts?

DAN HALE, CFO, ALLSTATE CORPORATION: I love Steve's answer; I'd like to say that. I think there is also a couple of other levels we can look at the lessons learned from this past year, past two years, but particularly with respect to this year and Katrina. When you think of the industry and our customers, policyholders, I think one of the lessons is obviously that states and our federal government clearly was not prepared, is not prepared to handle a major, major catastrophe.

I am sure some of you are aware that Allstate is one of the companies that has been lobbying to make sure that other states are doing what Florida has done, in terms of the Florida Hurricane Catastrophe Fund, and that as some of the representatives from there have already done in terms of sponsoring a bill in Congress that is to make sure there is some sort of federal backstop. I am not talking about a bailout --- something that is funded by the policyholders and states, states and federal government -- but something that is paid in advance, so that we don't have this shoveling money out the back of Air Force One as a way of solving the problems from a major, major catastrophe.

So I think, that at that level, that is something that all of us will continue a debate on, what should be done to improve the preparedness and the responsiveness to major catastrophes.

JAY GELB: Could you focus on that issue a little more, Dan, the paid in advance? Do you mean setting up catastrophe reserves before they happen, or is it something else?

DAN HALE: I'm talking about state funds, as the Florida Hurricane Catastrophe Fund, which could be funded from partial payments, if you will, additional amounts from policyholders. Could be from other participants, real estate mortgage lenders, other participants in the process in these areas.

Again, there is a lot of debate as to how that could be done, in addition to the bonding capacity of states. Then the states could be part of the funding for a federal reinsurance program, if you will, the backstop kind of program. But that again is one of the broader perspectives in terms of lessons learned.

I think as Steve has indicated, from an industry perspective clearly we see that there is more we should be doing in terms of working with the modeling companies. We all want to point fingers at them, but I think the models have improved dramatically since they were first initiated before and after Andrew; and the they continue to update, upgrade, as new information, as new events occur.

But each of those items that Steve enumerated are areas that we need to continue to work to improve upon. There is clearly a real difference between what the models did, it appears to me, and depending on the information, the data that was provided by various insurers, between commercial carriers and personal lines carriers.

For instance, the actual Hurricane Katrina number that we ran just prior to landfall with our data came out fairly close, from a personal lines perspective again, to what our final estimate or at least our third-quarter best estimate was that we reported. So the quality of data, how clean it is, etc..

But there is still -- there is no doubt that the items that Steve enumerated, when you think about demand surge, there is more that needs to be done there. Mold is another issue that has historically not been covered. Autos, boats, not been covered. Loss adjustment expenses are different. The additional living expenses perhaps was not appropriately considered for some companies; and you have to do your own work with the models.

In addition when you talk about major, major catastrophes, there is definitely the issue of having to bring in temporary help, if you will, outside adjusters. That is additional expense; so all of those things need to be appropriately factored in by insurance companies.

We are working with modelers to do that, to do it more effectively than we have in the past. But also from a personal company perspective, just as Steve indicated for AIG, it is clear to us we have to do a better job of more aggressively managing our exposure. While we had a some significant progress, we clearly need to do more, and we are on the hunt to do more, using all of the levers that are available including reinsurance.

So I think, again, he covered it quite well. But it is something that all of us will have to work harder on going forward, to be sure that we're comfortable with our exposure to major catastrophes.

JAY GELB: Dan, could you drill down a little more in terms of the levers to pull? Can we look at the 2004 hurricane season in Florida? There were changes to reinsurance; there were essentially having customers go elsewhere; plenty of other steps you took. How do you plan to address that overall in the Gulf in the hurricane (multiple speakers) ?

DAN HALE: Again, in terms of rates -- and as you know we file for rate increases -- significant rate increases. We got part of what we wanted; and part in Florida, at least, is now being negotiated.

In addition to that, though, when you think about the other items that Steve mentioned, increasing wind deductibles, we did get off about 95,000 policies in Florida. We brokered those policies out, bringing in another carrier to the extent possible to make the transition as easy for our policyholders as possible.

But those kinds of actions you have to look at, depending on the regulatory environment in each of those coastal states, the ability to get rate increases, and the cost of reinsurance.

JAY GELB: Great. John, your thoughts?

JOHN VOLLARO, CFO, ARCH CAPITAL: Thank you. We come at it from a slightly different perspective in terms of lessons learned, since we came through this probably in better shape than most people. And that is, that diversification matters. Since Arch was recapitalized in 2001, we have been developing a broad mix of business, both split between reinsurance and insurance; it is pretty much a 50-50 balance. Then within those areas, a fairly balanced book across numerous lines of business, both property and casualty.

As a result of that, we came out, as I said, the quarter with a lot less than a lot of other people have felt. I think we are going to talk about the models later. We can talk about that, but we have always had a healthy respect for the amount of variability around the mean that you get when working with the models, and really believe that thick margins are one of the things that are required to deal with that inherent variability.

As we have been discussing, there are a tremendous amount of variables that have to be counted here. And clearly the models failed very miserably on the commercial side with respect to dealing with issues like business interruption, in particular, along with other things.

So we think the answer to that is, you need thicker margins, whether that is thicker margins at the primary level -- and we are not a personal lines company; we are a commercial specialty lines company as well as a reinsurer. But we think you need it at the primary level. You need -- and when I say thicker margins that doesn't necessarily have to all be price. It can be in the form of price as well as underwriting adjustments with respect to deductibles and limitations on what happens in an occurrence.

So because we have a diversified platform, our view is, one, we're never going to expose too much of the Company based on modeled results; and two, if we're going to deploy our capital to business where modeling is driving the pricing, we think you have got to have very, very healthy margins when you do that. Margins get very thick, we will deploy more capital; the margins are not as thick, we will withdraw the capital.

JAY GELB: Great, thanks for the answers. Next question. Reinsurance pricing is expected to rise significantly in the wake of this year's hurricane losses, particularly in short-tail lines. First, what is your outlook for the reinsurance market for 2006? Second, what changes do you plan to make to your reinsurance programs? And finally, what is your ability to pass on higher reinsurance costs to your customers? Why don't we start with (multiple speakers) reverse order.

JOHN VOLLARO: Sure. Since we have an interesting perspective and I am sure you guys do too, because I know at least in Steve's case that they both write and buy. In both -- in our case, in any event, we write more reinsurance than we buy.

But our view on the market is that with respect to the property lines, particularly in the energy area but even in the property cat area from a reinsurance standpoint, prices are going to go up, and in some cases very dramatically. We think they are going to go up on the primary level, and that will allow primary insurers to be able to pay for the increased cost of their reinsurance.

But we think when you divide up those price increases between insurance and reinsurance, it is going to be skewed more towards reinsurance than it will be towards the primary insurers. So what I mean by that is, on our primary insurance business, we will expect to retain more and probably pay more.

We will have higher costs. We think we can pass them on. In a more normalized market we might be able to pass on at least an equal amount; in this case we probably will not. But we think even with that margins -- for cat-exposed areas, I am talking primarily -- margins will improve.

The offshore energy area, there are a lot of things that have to be done, including significant price increases; and we think that area will also be attractive.

A few things are going to drive this, in our view. First of all, buyers -- first of all, the modeling forms terms are taking a new look at the models. We expect at least to see fairly substantial increases in what the models will now produce for events relative to what they would have produced a year ago. That will probably increase the demand for reinsurance purchasing capacity.

On the supply side, the rating agencies have now taken a much harder look too, in particular, at how they view catastrophe risk. One of them has gone so far as to state that cat losses are the single biggest threat to the industry in general. So we believe that there will be significantly higher capital requirements for every unit of cat risk that someone takes on, 2006 versus 2005. We believe that is going to constrain capacity on the reinsurance side.

It will affect some companies more than others. The rating agencies have now, I believe, are taking a different look and now believe that you do need a diversified model. For those companies whose models are more concentrated and monoline in focus on property and property cat, their capital requirements are going to go very dramatically, which will inhibit their ability to even retain the aggregate exposure they have. So we think that is going to be a driver of increased pricing.

Third, we think cedants when they buy -- and again, we are both a buyer and a seller, so we sort of look at it the way we look at -- are going to look to diversify their own purchasing among more reinsurers than they did before. I think in 2005 and earlier, there were certain insurers who were putting down very large lines, and that was being accepted by buyers. We think that is going to change; that is going to further, again, constrain capacity.

So as a result, we think the market in general in cat-exposed areas will be very, very good. We think non-cat-exposed areas, property will go up somewhat, but not dramatically, and that will depend very importantly on geography.

With respect to what happens to the nonproperty lines, even though one can -- we would look out and think that over time, as the increased capital requirements for property put a strain on everyone's balance sheet, that capacity may have to be reallocated away from nonproperty lines over to the property lines. We believe that is a rational and logical thesis. So far we have not seen any evidence to support that.

We think at a minimum it should lead to at least a leveling off in what was going on, from a rate decrease standpoint; it should stabilize the market. But we don't see any tangible evidence of significant hardening the way we are seeing right now, on the property and the marine lines.

JAY GELB: Great, thanks John. Dan, as a significant buyer of reinsurance, you probably have a different perspective. Why don't you lead us through what you are thinking about on the reinsurance front?

DAN HALE: After the experiences of this past hurricane season, as John has indicated, the cost of reinsurance will definitely increase. Now, since he would not tell me how much I can't pass that along to you but --

JOHN VOLLARO: We can talk about that later.

DAN HALE: But there is no doubt there will be an increase. It certainly depends on the location; it depends on structure of transactions, etc. The good news for us is that in the recent past we have significantly increased our reinsurance purchases; and we did so on a multiyear basis. So last year we entered into three-year contracts in all of the states that we have -- some seven states overall. Three-year contracts for all but Florida, where there is a two-year contract there. So basically margins are locked in based on exposure changes etc. is (ph) the only incremental cost you would get.

But going forward into this next season, we do expect to expand our reinsurance purchases and we are looking to do so. It will depend on, again, structure. We may even do a countrywide cover, capacity price or terms issues, as far as that type of a transaction. But there's no doubt there will be some increase. As we have done in the past we will continue to evaluate the trade-offs as we look at our overall capital requirements.

We do a -- I think most of you who have listened to our presentations in the past -- that enterprise risk management approach to developing economic capital on a stochastic basis. Obviously one of the big risks is catastrophe risk. We use the models, and we are having to modify those models to be sure that we now factor in what will be increased -- I'm sure there will be some increase in capital requirements for the PML exposure of the Company.

So as we continue to look at that, we will evaluate the reinsurance cost in various locations, different structures, and we will buy where it makes economic sense for us.

JAY GELB: Great. Steve, your thoughts?

STEVE BENSINGER: I think John and Dan have covered this subject very well, but maybe I will just add a couple of things. AIG, as you know, is a very large buyer of reinsurance. We are also, as John noted, a seller of reinsurance. Everything that John and Dan indicated in terms of pricing terms and conditions on the reinsurance side, we agree with. They will all be tightening.

You asked whether we thought that those costs would be passed along to the ultimate insurance buyers. In our case, I would say that they will. The predominance of our property catastrophe exposure is in our excess and surplus lines business. We are, by far, the largest E&S player in the world, and most of the losses that we suffered and most of our exposure in that area are in E&S. You can react much more quickly in the surplus lines area than you can in other areas of primary insurance, and the like.

So we are already seeing, because we're anticipating and are already experiencing increases on the reinsurance cost, we are showing solid 10% to 35% increases right now in our own products that we sell. It could go higher than that in some cases, depending on where it is. It is higher.

I talked earlier about the fact that we're modifying deductibles, and we are also on both flood and also wind substantially. Where they might have been 2% to 3% before for wind, now they are going up to 5%. So significant changes in the underlying terms and conditions, as well as pricing, which is another way to pass along the increased costs.

We do not have a very large homeowners book in the personal lines business, so that has not been as significant an issue for us as it has for some other companies like Dan's.

I would share the sentiment that the rating agencies -- and I think we will talk about this a little bit more -- are becoming much more vigilant in assessing the capital required for these risks. That is something that we all have to take into account.

John mentioned diversification. That helps AIG as well. Because of the very vast diversification of AIG's portfolio of businesses, even an unprecedented quarter like the third quarter of 2005, which resulted in us posting a $1.6 billion after-tax loss for the third-quarter catastrophe events, was still less than half of what we reported in net income for that quarter and represented less than 2% of our shareholders equity.

So we have the capacity to be able to absorb these types of issues; and to the extent that the reinsurance arena moves too far away from us, we will increase our aggregate and our net retentions, and we will control them on the front end, as I talked about earlier.

JAY GELB: Steve, I believe it was mentioned in the third-quarter earnings release, a comment about the reinsurance market as a whole in terms of capacity and availability of that. Can you drill down a little more in terms of what your thoughts are at this point, going into the 1/1 renewals, since AIG is a large buyer of reinsurance?

STEVE BENSINGER: Yes, absolutely, Jay. Right now, we are not seeing any issues in terms of capacity. I think the issues are going to be in terms of how much it costs, and whether or not that make sense for us from a cost-benefit standpoint. That is what we will be assessing. Capacity doesn't seem to be an issue. I will say that the reinsurance community has probably in 2005 lost more than the capital that has been replenished so far.

Even some of the new entrants that have been formed will now be, I think, exposed to the rating agency issues. If you look at some of the assumptions that the rating agencies are using, one in particular is now going to stress a company's capital for not only one event, but two events. What they are looking at is a 1-in-100 wind event and a 1-in-250 quake event all occurring within the same period; and then another series of those in the same period.

So that is going to -- if that is applied rigorously -- will result in a very low leverage on some of those new players in terms of how much they can write against their capital. It is going to be hard to achieve the types of returns that I am sure the shareholders of those institutions would like to see, without some significant changes in pricing. So it is more of a pricing issue from our perspective than a capacity issue, as we see it.

JAY GELB: Okay. Dan, to circle back to one issue on the reinsurance, to what extent do you think Allstate can pass along higher reinsurance cost to customers in the homeowners line?

DAN HALE: It depends on locations. But I know states where we have purchased reinsurance, virtually all of those, including New York, we are able to pass along the cost. Where we can't, that is where we have to be more focused on all the other levers that we have to pull. But so far, again, on those coastal areas for the most part those are states where we can pass them along. There may be a little bit of a lag in terms of timing, but it is doable.

JAY GELB: Can you do it in the Gulf states?

DAN HALE: Yes.

JAY GELB: Higher reinsurance costs? Okay, great. Let's focus on the next question, which will turn to the rating agencies. The rating agencies have been tough on the insurers and reinsurers, particularly in the wake of the hurricane season. They are taking a much tougher stance in terms of managing the exposure.

How are you adjusting to these new factors in terms of your modeling? You can be as specific as you want to be, versus the requirements put up by the rating agencies. How does that make you rethink exposure analysis? Why don't we start with Dan Hale from Allstate.

DAN HALE: Sure, Jay. At the time the hurricanes were striking, as some of you may recall, we did get put on credit watch. As we went through the entire third quarter and finalized our numbers at our rating agency meetings, all of the rating agencies reaffirmed our ratings. We had a few changes in outlook, but they all affirmed our ratings.

The situation with Allstate was much more one of a high-quality problem, if you will. Because of our underlying profitability run rate for the performance of the business, our combined ratio, if you exclude catastrophes, continues to run about 80%, 80% to 81%. We continued to generate significant cash flow; build, generate capital. The rating agencies were well aware of that.

The high-quality problem we had was we wanted to continue to dividend maximum dividends from our insurance subsidiary, Allstate Insurance Company, the personal lines company. We wanted to maximize the amount of dividends that we give to the parent, to use to repurchase shares and for other uses. In doing that, we did not want to -- we wanted to make sure, first of all, that we were not stranding excess capital in that subsidiary next year, because of the dividend limitations from a statutory point of view as a result of the hurricanes in '05.

So that high-quality issue with the rating agency discussion was clearly we want to continue the dividend maximum amounts to the parent, and to repurchase shares. We have a $4 billion share repurchase program underway. We continued in the third quarter to repurchase shares and we expect to complete that program, another roughly 1.8 billion to be completed in '06 in accordance with the scheduled date that we announced when we announced that program.

So the discussion with the rating agencies was, we want to continue to repurchase shares; we want to maximize the dividends up to the parent. To keep them comfortable and to show them our commitment to the business, we put in place a capital support agreement, which said at the parent we will keep X amount of dollars that will gradually decrease over the next year. But to be sure that they are comfortable there is enough support for the personal lines business. Again as a result of that, all of our ratings were affirmed.

I think the rating agencies from our perspective, we have an excellent relationship and have, over the past few months, taken all of the rating agencies through demonstrations of our stochastic economic capital modeling capability. As we discussed earlier, those models will have to be updated for the latest changes in the cat modeling of some of those items that need to be refined going forward. Undoubtedly that will cause the amount of capital to increase.

The real question is, to what extent does the amount of capital you need to hold decrease as we reduce our female (ph) exposure through the other actions that we have talked about. So there is a trade-off there. And the rating agencies are very comfortable with how we manage the business, how we look at risk on an overall enterprise basis, how we are incorporating catastrophe losses and potential losses into those numbers, and the ability to continue to build capital going forward and generate excess capital to continue to buy back shares as we have in the past.

JAY GELB: In terms of your discussions with the rating agencies, are they counting on you to buy more reinsurance or is that not part of the rating analysis? …