The Not Unreasonable Podcast

Steve Mildenhall on The Macro History of The Insurance Market

March 11, 2021 David Wright Season 1 Episode 51
The Not Unreasonable Podcast
Steve Mildenhall on The Macro History of The Insurance Market
Show Notes Transcript

This episode marks the return of Steve Mildenhall, principal at Convex Risk, former Assistant Professor of Actuarial Science at St John’s University and former CEO of Analytics at Aon. This time, Steve is bringing an amazing dataset that he has developed showing the longest sweep of history in insurance I have ever seen. You can see the deck we go through here. This show is probably best consumed as a video, which you can see here

You can see show notes at notunreasonable.com. 


Twitter: @davecwright
Surprise, It's Insurance mailing list
Linkedin
Social Science of Insurance Essays

David Wright:

My guest today is Steve Mildenhall principal at convex risk, a former assistant professor of Actuarial Science at St. John's University, and former CEO of analytics at ayon. Today, we are talking about the history of the macro environment of the insurance market as part of a course Steve is designing on pricing insurance risk. Steve is also writing a book about pricing insurance with risk was john major due out this fall, Steve, welcome back to the show.

Steve Mildenhall:

Thanks for having me.

David Wright:

So we're talking about the history of, of insurance, the macro environment of insurance. And I think that we're gonna go so far back in history, I think, which I'm really excited about. But it's worth asking the question, why bother? What can we learn?

Steve Mildenhall:

So that's a that's a great question. And let's just see, how far back Are we gonna go here, I got this chart that shows the ratio of property casualty premium to GDP all the way back to 1923. And partly, I did this simply because I could I, when I was at St. John's, they've got a wonderful insurance library, because the Davis library, and it has all of the am best books, all the aggregates and averages books, and you just it was simply there. I had had this child previously backed, I think the 1968 is reasonably easy to come by. And I was just like, Hmm, I wonder what happened, you know, before that. And so I was just motivated just to it just because it was there. And I love data, I wanted to see what it looked like. But in terms of what, yeah, what's the point? What can we get out of this? insurance is data driven, right? It's all about the losses. And statistics like this shed a lot of light on the structure, the management, the evolution of the insurance industry. And as you know, as we, as we look at it, this, this type of thing is sort of electronically and easily available back to 1996. And you see a lot of studies that, you know, therefore they they use that data, and I felt like, Well, you know, I personally, I started in the industry in 92. And for those around them, that was Hurricane Andrew, which was a huge, huge deal. And then you had 94 was Northridge and I, I gotta get back to 92. And then, you know, keep going back, we want insurance companies to run and make good on their promises for, you know, 2550 100 years, you need a long term time history to sort of get the requisite set of events in there that you can, you can truly sort of test again. So I have a lot of data driven reasons why this sort of thing

David Wright:

was good. And no, I can't help but say, as I look at the graph, that Northridge Hurricane Andrew, even if you skip ahead to Hurricane Katrina, they just don't matter. I mean, talk about irrelevant How disappointing, right for the amount of like the amount of ink that has spilled over these very, you know, prominent catastrophe events. totally irrelevant, I would say in the macro history.

Steve Mildenhall:

Yeah. Great, great point. And yeah, I just overlaid a few, a few events here to sort of see what what seems to matter and what what doesn't matter. You can see, there was a little tiny blip after Andrew but it was part of a Down down down after the LM x spiral in the, in the 1980s. You know, oil crisis was something Harvey actually at the end there, you know, 2017, Harvey, Irma, and Maria had a sort of fairly significant uptick. Obviously, the Second World War had an uptick. But yeah, a lot of these events that are sort of so huge as you lift through them, maybe didn't move the dial necessarily on this.

David Wright:

Yeah. And you also get the idea of a sense of scale of hard market, right. So the, the vert, the slope of the vertical line, and the length of that vertical line tells you real how awful it was. And I was, as I was looking at this, too, I mean, I could probably talk about this very graph all day. But I remember to a podcast interview I did with cash, what was his name? Paul, in gray, Pauline gray, and Pauline gray told me the story about like 1980 to 93. And he talked about how he was having to double the premium on a particular policy and raise its deductible, he was an access policy or something like that, by like, 5 million bucks, right? And then and then sort of making up prices. I mean, right? You know, an actuary, right, you're thinking like, okay, so if I double the premium, and I increase the deductible, like, you know, how rate adequate EMI I mean, I don't know, you go through the book, right. And then then that policy is still lost money. And so you have this like period of what we describe as a very hard market, but down the down down the slope, we continue to go. I mean, think of the agony, and that's also a period to where, where GDP, nominal GDP is rising pretty fast because of inflation. So you kind of have, you know, the numbers were all the place. And, I mean, you see the consequences at the end at that bottom where he says element spiral, which I would think is probably a result not a cause of the hard market, right? That was kind of like the end the beginning or The end of the beginning. That was the

Steve Mildenhall:

Yeah, the trigger. I mean, this period here is just remarkable. If you look at this, you're you're moving over a percentage point of the entire economy move into the TNC sector in two years. It's just astonishing, right? Yeah. And you had, so all sorts of things happened in 86, you know, absolute pollution exclusion claims made form that was tax reform, whatever sort of changed a lot of the dynamics, but there was a huge hole here that people needed to dig out of and, and you know, your story about Paul, doubling the premium. The thing that you don't see from this is an at, you know, I experienced this in the sort of post post World Trade Center, much smaller, hard market that happened after Oh, one, people were doubling the premium, raising the deductible and lowering their limit, right. So this doesn't really capture necessarily the sort of unit cost of ceding risk into the insurance market.

David Wright:

Maybe it's clearly Yeah. And you think like the scale there again, so let's look at the early 2000s. You had one of those in the mid 70s, the hard market of a similar scale to that one.

Steve Mildenhall:

Right that that was the soul you had a huge stock market decline that happened after the oil crisis. 7374 we'll see later on was actually a low point for surplus, adequacy for the for the industry. And then the word liability issues. I think it was a med mal crisis then. So in fact, that one looked a little more maybe like the WPC kind of dissolve was sort of reserve driven. And then you had inflation sort of picking up during that time, I'm sure people discovered that they were somewhat under reserved, you know,

David Wright:

I'm also struck as I, you know, I mean, Steve, when he first showed me this graph, I was just like, big smile on my face, because I've seen the ones. And I have actually, in various presentations over the years tried to recreate the one you were talking about the best one that goes back to the 60s. And I was always like, Man, that's so cool. You can go back. I wonder if we go back farther in here. you done it? And then you look, what do you learn? The 1920s had the biggest heart market of all, right? I mean, there you go more than, you know, percentage tune one and a half percentage points of GDP from 1922. Late 1930.

Steve Mildenhall:

What? So maybe, right. I think the other interpretation there is that that was the beginnings of the insurance industry, you know, automobiles was sort of coming on line. And really, I think that that that piece was probably more reflective of more things be interesting should then had. So workers comp laws, I think I looked up I think, like Liberty Mutual was founded, I think, in the sort of,

David Wright:

yes, State Farm as well. Yes. And

Steve Mildenhall:

it's so and they were in direct response to workers compensation laws coming into into play where count your employees that had to had to provide this coverage for their, for their workers. So probably a lot of that 20s. Boom, was actually more insurance, right? And then you got to see the same thing, I think, yep. So then you have the Great Depression, which obviously, changes everything and your Second World War, probably people aren't running around suing one another, as so much of the economy is militarized at that point that your normal rules of insurance probably don't apply. But I think the period sort of from about 1950, up through, and you can, in fact, if I draw the line on the next page, you know, you've got this long period from sort of just after the Second World War, let's call it into the sort of mid 1990s, where there was a sort of in the industry was on the finding coverage, front foot, if you will. And so I remember reading that, like Hank Greenberg's book about AIG. He describes being in the late 60s, and your new DNO lawsuits were coming up sort of new mechanisms for sort of pain finding people liable. And he talked about, you know, wanting to find coverage to sort of stand behind those, those potential tours. And so there was potentially, you know, there was an expansion of coverage, there was a period of enormous economic growth. And it was a sort of fairly industrialized growth, right, that you had, you know, there was a lot of big pieces of machinery moving around, and that tends to cause damage and sort of require insurance. So I think that period, yeah, the watershed probably was at six, there's another version, I got to this, I don't have it here, where I look at losses that are sort of seeded into the industry, and they're almost V shaped, they go up until 86. And then they kind of come down on the other side. And part of that is the way the tax code used to be written prior to 86. You got you basically got a deduction for your underwriting loss. And if you made investment income, that was all good. And then in 86, they changed it so that you got taxed basically on a sort of operating income basis where you have to discount for tax purposes. And so that made it less advantageous to just kind of push stuff through the insurance industry which I think Paul You know, a lot of business out and then in addition, there was a there was you know, as best as an environmental wouldn't really be kept people begin really becoming aware Have those at the Lloyds crisis going on? And people may be moved away from this, hey, how do we find coverage to now it was I want to restrict coverage, right? Absolute pollution exclusion, you could argue claims made form, which was kind of an attempt to do that. And the industry mate moved on to a sort of more defensive posture in terms of trying to, you know, expand. I'll

David Wright:

say it was a lot there.

Steve Mildenhall:

What

David Wright:

can we just talk for a second about the detail of your point there about the change in the tax law? So you're I think that the sentiment you're expressing is that companies were less loss averse, they're more willing to make underwriting losses, because there's a tax advantage to that they could deduct that from their tax bill, is that, right?

Steve Mildenhall:

Yeah. So I started working at a company that had a very complicated net present value calculation that incorporated all the tax. And I'm not sure exactly where it stood, you know, Visa V, sort of tax reform. But certainly at one point, there was a sort of, you know, it was almost like, the more money you lost, the more money you made, because you made it up from the tax benefit. So there was you could write stuff with a big underwriting loss. And obviously, let's not forget here, early 80s, right, you could buy treasuries, and they were yielding 10. I mean, investment income was a serious thing back then. And so yeah, there was there was a lot of kind of cash flow underwriting as it's called, because you could make so much money on the, on the investment piece, and then the tax code was

David Wright:

another thing that was, you know, sort of fascinated by is knowing during your, during the period of your orange line, you have I think that what you're kind of alluded to, there is a lot of let's call it legal entrepreneurship, right? Where we are doing the the court system is redefining, pretty deep kind of like, I mean, moral code of a society, right, where you're saying, who's responsible for what, and you'd like us to mention strict liability, where now the duty of care that certain, you know, let's thinking like medical malpractice, right, where now you can sue a doctor because a procedure went wrong. Whereas before it was, you know, it was like best effort required or something. United States, of course, being a bit of an outlier, and how strict the liability is here. But that'll happen then, too. So we're sort of discovering new uses for insurance, which is something that, you know, a lot of insurance executives these days, would love to do, right? And invading an insurance, we all want to find new products, but I imagine it's pretty scary. You know, that period. And look at that. I mean, look at that slope. And you

Steve Mildenhall:

know, also during this period you had people went from I don't know what car ownership was immediately after the Second World War, but I'm imagining you're probably less than 31 car per household, on average to more than one car per household. But you know, by the end of the period, you're the average size of a home increased significantly over this period. So there's, there's just a massive amount of more stuff that needed protection by insurance as well, I don't think it was, it wasn't just the sort of legal piece, but that obviously was very important for certainly,

David Wright:

yeah, and I think that maybe this preview, some will move to perhaps next, but you there are clearly different eras. Right? So we're looking at, like, the macro environment has been, you know, very different in different phases of history.

Steve Mildenhall:

Yeah, I mean, this one, I one more thing, just to comment on this is that, you know, this sort of idea of the underwriting cycle that I think everybody young, yeah, it all sort of buy into this concept of writing cycle. And if you if you think take the cycle to be the sort of deviation from the orange line here, it's pretty clear that you know, for a very long period of time, there was there was a cycle, right? I mean, it just oscillates nicely up and down, sort of either side of that line getting getting more More, more and more extreme as, as time goes on. And, you know, it's, it's interesting to look at that cycle, and sort of decompose it a little bit. Now here, we fall back to now I'm going anything I have, where I have line of business detail, it begins, unfortunately, 1992 I've got, I need to type it in and it takes it seems to take about three hours per year to sort of take these numbers in. So I spent a long time typing in 90 234, and five to get myself back to 92. And this is as far back I go. But here is the cycle, you know that so this this green on the right here is is kind of the last it's the you know, it's the down and up from here, right? So that's that's what we're looking at. And what we see a few things here. One is we can split this into personal and commercial, I have the line of business so personal was basically personal auto and homeowners farm and this is in there, and then commercials essentially everything else. And so you see from this, that the cycle tends to be much more of a commercial lines phenomenon than a personalized there is a little bit of a cycle. And it's sort of interesting how it's it's correlated, right, given that they are sort of largely separate markets with separate players and sort of separate issues as you think about that. But you know, the top to bottom two peak to trough here is, is maybe two point, you know, two tenths of a point, the peak to trough here, you know, six tenths of a point, right. So it's sort of a 3x Delta. And I think, you know, the late what you had in the late 1990s was a sort of overall period of magical thinking, if you will, in financial services, you know, you've got Enron and what have you, and everyone just sort of, did forget that, you know, expected losses just add up unless you take unless you do something extraordinary to take something out of the system, you can't just bundle it all together and magically make losses, you know, disappear. And then there was a lot of reliance on investment income here, you know, the roaring stock market in the late 90s, it had been going on for quite a while, sort of depressed premiums. So we see a little bit of a swing on the personalized side, but what we see we see, you know, substantially more of a swing on the on the commercial side. And then the other thing that we we see that that's now sort of interesting here is that it's been incredibly stable. it you know, going back sort of, you know, if we look at the broader picture here, sort of back eight, nine, you know, 10 plus years, that certainly this period down here from what is this, like nine to about 16, that was a very long period is there isn't a sort of comparable period, when you look back in history, that was as stable as that. And then, you know, this, the data I've got here for 2020 is a little bit of an estimate, the GDP number has come out. But the insurance number one come out for a few more months, yeah. But it's pretty easy to guess where insurance is going to come out from the third quarter number, which is what this is based on. And this, this, this uptick is actually a little bit of an illusion, obviously, because GDP was lower in 2020, because of COVID. And if you take if you if you're allowed GDP to grow at the same rate, you know, for 2020, over 19, as it did 19, over 18, this point for 2020 would actually be flat. So it's almost entirely been a feature of lower GDP rather than increased rates. So we're still in a very, very sort of stable period, which leads you to wonder, you know, are we kind of out of the out of the cycle business now? And, you know, there's arguments on

David Wright:

it. One thing that comes to mind as I look at this, which is that's a fascinating point, and try and come back to that in a second, but through this piece, but one thing that is the case of personal over commercial, is that rates are much more restricted, right? So you have more filing more regulation, less flexibility to increase and one would imagine decreased rates, I'm not totally sure about that definitely increase in rates, it's harder. So but I think if you go back to that last graph, the prior graph with the historical sweep, one thing you also see as the amplitude of the cycle increases, you see rate liberalization, right. So you see, like, super strict rate before the SE an underwriter case, they went to the Supreme Court that basically destroyed this this wall for a moment, or nearly destroyed the state regulation of insurance when federal, right where you have these effectively cartels, pricing cartels, all designed to I mean, the irony, and I just never tired of saying this, the irony of insurance regulation is it's here to protect the insurers, not the insurance, because the solvency regulation against themselves, of course, we're giving too good of a deal to their customers. And you can you can predict, then, you know, there's two sides of the debate, one would say, rate liberalisation where on the one hand, they'll say liberalize rates, because you're going to improve the deal for customers. And then the other side says, Don't liberalize rates because they are going to screw customers, right, the exploitation. And turns out both are right. And they can be right, because you have a cycle, where you're intermittently giving customers too good of a deal, and then a really bad deal. And that mark, the amplitude of the cycle.

Steve Mildenhall:

Right, and the industry is sort of like a giant retro waiting plan. Right. I mean, rates have to be prospective. But obviously, your estimate of future rates is based on your recent experience. And so yeah, it's it's a broadly sort of self normalizing process.

David Wright:

And one thing that I've toyed with, I don't think anybody agrees with me, but I'm gonna keep talking about it is that, you know, if you look at that, that kind of, you know, the the great stagnation that's called or the the great de risking of the of the 2000s, the kind of latter half of the 2000s, that flat period you're talking about there? You know, I tend to pose this point and put it at this point as a question, which is to say, do you think that technology has zero impact on an insurance company's ability to manage risk? And some people will say, Well, no, of course it has some impact on the ability of insurance companies to manage risk. And then I would say that well, then wouldn't a dramatic improvement in information technology result in better risk management and insurance companies. And the answer probably would be Yes, a little you have to disagree with only the Steve. And then then you would, what would you see if you had a giant improvement in risk management technology, a flat curve? What do you think about that argument? Does,

Steve Mildenhall:

there's a lot of a lot of things there. So, risk will, we'll get we're gonna get on to discussing risk, because that's my kind of favorite topic here. But broadly, I think, so we've got asset risk, let's kind of take that off the off to the side, because that's sort of a background noise almost for our processes here. As we think about underwriting risk, there's sort of three levels to it. There's, I know what the price should be. But the actual results I get are way different from what I expect, okay. And that's basically catastrophe risk, right? If I'm writing Florida homeowners book of business, I might have exactly the right rate for the 100 year average. But next year, holy cow, it's going to be way off what it what it should be. So we've got that sort of event risk level, then there's a risk, which is, I'm pretty sure I know what the price should be. But it can't necessarily get in the market. And I think a lot of what we see here with the cycle is reflective of that second type of pricing risk. And in fact, they think that, you know, the industry is sort of reasonably competitive. And it what that competition does is it kind of forces companies, right to the edge of what they can credibly price, right, there's always an incentive to take any, if you've got some sort of classification rating plan to take that cell and divide it up, right, and people kind of keep dividing until at some level, that kind of reading the tears, right, and that sort of introduces risk for you. Because I know you have to worry, you know, in the old days where everyone was on, so you don't have to worry about your plan being selected against by someone else's plan today, you do need to worry about that. And that's a different, different dynamic, maybe than we've had in the past. And then, you know, the third level of risk, obviously, is I don't actually know what the premium should be. So this is the problem we have, like with a new coverage, like cyber, for example, where people are really sort of grappling with, well, you know, what, what should? What should the price for that to coverage be? And we tend to be very conservative about that isn't a new industry for the obvious reasons that you could stack up a pretty big bet, bear and you know, end up discovering that, in fact that you know, you weren't judging right.

David Wright:

So then would you say that the our ability to manage our own pricing risk, so what is your explanation for the flat curve? You know, I say, I say, Well, I mean, I wouldn't say assign 100% probability of this being true, let's call it I don't know, I think a 60% believe this. But that the, the, the, the flat, the flat curve, and the 2000s, is a result of improved measurement technology. So companies are able to check themselves more effectively from going over the cliff and dropping the rates too far.

Steve Mildenhall:

So I think Yeah, I would say there's probably two major contributors to that. One is rate monitoring came in in a really big way after 2001. Right. And I think if you go back and you read conference calls from like, the late 1990s, I don't think anyone was talking about rate monitoring. I mean, they were where they should have been. And in fact, and when it was amusing, right, if you were like pricing at that point, ever, you could do an index of how soft the market was, was directly proportional to how stupid people thought ISO and ncci was right. Oh, these rates are ridiculous, I can't possibly get this reason alone. And it didn't seem to occur that everyone was like, hey, they've got all the data. And it's not that difficult to set the overall rate level, maybe those rates are actually about right, you know, and that turned out to be the case, if you really look look at work come, particularly in your late 90s, early 2000s, it basically went from everyone was COVID crediting rates, everyone was damaging the rates, the actual law schools really didn't move, you know very much at all. So people knew what the rates were, but they sort of chose not to monitor and that became entirely unacceptable post sort of 911, you got your very transparent, you know, Bermuda, new coal started up reinsurance got stripped out of multi line companies. And there was just as an overall increase in transparency of the business, and there was an expectation from investors, but hey, I want you to start reporting on break. And that became a really big deal. People started started doing that. So I think the people were, that was one part of it. I think the other part of it is the improvement in the technology for getting capital into the industry, right. Part of the explanation for why we had cycles in the past was, it was always difficult to deploy capital into the industry, and that the transformation in that there's happened over the last sort of 25 years, it's just been enormous. I don't know the first sort of cat bonds, you know, the USA rez rebonds that they did in bottle a 9596. They probably took them two years or something to get the first one done. And then it took a long time to do those issues. I haven't heard what the latest is, but I imagine now you could probably get the whole thing done in a month, right? And you can probably set a new Co Op basically in a month. If you've got the management team. Getting the money is easy. There's a lot of people who are sort of sitting around the edge of the insurance Industry looking for that opportunity to invest? And I think that sort of availability of capital kind of puts the cap on the rate. And that's interesting, because what we've seen y'all just recently, why was h i m, and brushfires and California wildfires and all that? Why has that been a big deal and to attract capital in the collateralize, markets and whatnot? I think in part, it's maybe the first time we've had people pausing, and feeling like the cat models, and this sort of idea that, you know, how do we really know and understand this risk has been challenged? And maybe we need to think about what we do, particularly with the wildfires Do you know do how well can we can we model that how well can we keep going the draw, is global warming having some sort of macro effects here, there's actually going to be observable on the sort of insurance timeframes is made. Because people just to sort of become a little more leery about deploying capital into the

David Wright:

hospital, I would like I feel the urge to tell a story, or tell a version of that story that doesn't include property cat, right? Because I feel like I'm convinced myself, we're looking at your data, that property cat doesn't matter for cycle management. And maybe you're telling me you weren't telling me? Well, that's a consequence of the innovation. And maybe right i think Hurricane Andrew there being such a minor non event, to me is really amazing. And so but I think it might still be true that your point about innovation and in deploying capital into the insurance industry, because I think that here's a question for you, since you've been at it longer me, has reinsurance generally become more important, because from the stories I hear, anyway, the reinsurance business before say the 90s was a little bit more of a cottage industry, dominated by Lloyds and a few kind of domestic markets, and the diversity of reinsurers. And I mean, as a result of the enormous profits reinsurers made at the turn of the millennium after 911 perhaps allowed even the liability lines to to gain their access to more capital, and perhaps reinsurers ability to raise capital and deployed into even casualty because there's no greater capital vehicle for insurance than the quota share the mighty quarter share, because now you can increase your book enormously. And and the reinsurance organizations perhaps because they're just smarter or because of some other innovation, or maybe the monitoring was so good. Now they're able to be better at monitoring insurance companies that even in liability lines, we've experienced more liquidity of capital, what do you think?

Steve Mildenhall:

So I think some of what you're talking about, I obviously haven't had my finger on the pulse since 2016, when I started teaching. But I would say I think the story from 2000 to see 20, you know, the mid teens was slightly different than that. I think that coming out of 911, you actually had a catastrophic failure of communication between the primary companies and the reinsurance companies. And it essentially killed a massive amount of casual, brainy insurance. Because what happened was, there was this disconnect, where insurance companies looked at their book and like I'm running, you know, or 120 30, whatever it is, on this liability book, I really need to bring that in, you know, raise deductibles lowered limit jack rates, truly fixed that book of business. And somehow they were not able to credibly explain that to the reinsurance. And as a result, you can, you can look at the sort of mix of sessions from sort of, Oh, 102, through, you know, Oh, 506. And that market shifted the property in a quite a remarkable, quite a remarkable way. Because reinsurance and particularly alternative capital, truly is a way better way of burying that outside cat risk. And if you've got to think of it as sort of like a spiky risk on your balance sheet, and cash is the ultimate spiky risk, as opposed to a sort of a yellow plus or minus 10 point sort of volatility kind of a swing is very inefficient to bear that with equity capital, because equity capital is really expensive for a whole host of reasons that investors don't want to give insurance companies sort of permanent capital. And so if you can find something like you've got with alternative capital, where you're accessing a different set of investors, but you're, you're accessing them with a vehicle that addresses a lot of the concerns that people have around equity capital, right, so why don't people like equity capital, so long term commitment into a regulated industry, I've got double taxation, I've got a trust management is a very opaque product. I got no independent view of the pricing. Look at what a cat bond does a cat bond strips out all of those problems, every single one, right? You write it in a low regulation, no tax jurisdiction, you've got cat bonds gives you an independent view of the pricing. You're only giving them capital for two or three years and then you get it back because it's a it's a bond. You don't have to trust management because your your cap on hasn't got management, there's essentially no moral hazard. I mean, it's just, it's brilliant. So this extreme lines, why alternative capital can come in with a much, much lower cost of capital than equity capital and be really the perfect vehicle for siphoning off that risk that is only economic for you to bear on on your equity based insurance balance sheet. And whether that gets done, you know, directly to a cat bond or it gets kind of transformed through a reinsurer, or reinsurance have sort of slightly different mix of investors, and they sort of address some of those concerns, or they can address the taxation and they can address the regulation pieces, and they can be sort of more transparent like the reputed companies were. So I think it has a had a huge and transformative impact on the property side. Fortunately, at the same time, that casualty just killed you just shot yourself in the foot because the reinsurers were just reeling. Oh, yeah, no, I think your book sucks like it did in the late 1990s. And I'm not going to believe any of these things that you tell me you've done like, well, the primary companies were like, Well, I know I've done them. And if you can't give me better deals on this on this quarter show, I'm taking it now. And that's what happened, whose enormous shift of business away from Casualty. Yeah,

David Wright:

and I think I'm gonna keep trying to push the story I which i think i think we're gonna agree on this, I think, where I think the rate monitoring technology gives a, let's call analogous, if not quite as effective a benefit to reinsurers or other monitoring entities. Because I remember being in my career, early 2000s, when we were talking about rate monitoring reports, the beginning the dawn of rate monitoring reports, and they went back two years. Why? Well, because that's when they got their system. And so there's like a, there's a distinct technological moment, or like their second or third system, but they turn them over a few years because it was such an immature technology. But now you can actually like record exposure information, and you can do some kind of rudimentary exposure adjustment of your premium. And so derive at a at a at a somewhat robust rechange figure. And that was new. And, you know, that is like a very important tool for a casualty reinsurer right now, to go in and out of a deal or, you know, on level loss history and project the future. Where that I mean, it has to have had some effect. I got to think in dealing with a pricing risk. Yeah.

Steve Mildenhall:

Absolutely. Absolutely. And it gave you know, I think the presence of that was part of what gives the primary companies the confidence that they knew they didn't have to buy the reinsurance. It's like, I know what I've done in pricing yesterday.

David Wright:

And I think the one other point that you make to me is that that effect differs by line, right? So some lines are more sensitive to price and risk. And some lines are less so.

Steve Mildenhall:

Yeah, I mean, I don't want to get let's do it. So the story. Yeah, let's let's line. So y'all, we it looks like we've got this a level playing field when because when we look in total, I should also point out if you're wondering, why have I got this uptick here, and I'm flat here, on the byline view is direct. And the other view is net. So there's a little bit of a difference there between the director, nephews, but the broad as it did. When we look at bottom line, though, we see this you know, so down here, we got the total and this is all sort of very stable, virtually no other line is is stable, right. So these are showing direct premium to GDP split out by the major lines of business, on the sort of closest we get to someone who has sort of behaved, you know, you know, at least a monotonic way as homeowners, right. And homeowners obviously has had, it's been deemed by Andrew Northridge, and then your other hurricane cat. And then ironically, after 2004 2005, we had all these hurricanes, then there was a very long period with no cat three or greater making landfall. But we just dinged a death with your severe convective storm and her tornado hurricane. So that drove kind of massive rate needs on the on the homeowners side. But not none of these other lines really looks kind of like you would expect, like they're they're just like, all over the place. And this broad family, you know, connection, so could you know, your basic liability here, you know, but you've got an auto cycle. So personal auto, and then commercial auto sort of correlated. But in a you know, commercial auto in particular has been a very difficult line very stressed for a lot of people. But but possibly No wonder, right? You gave up here, what's that from about eight tenths of a eight tenths, nine tenths of a point of premium there out of commercial auto, financial guarantees, obviously, on a sort of whole different scale with the events that happened there, med Mal, you know, after the sort of giant crisis has had favorable, you know, tort reform and whatnot, favorable development, to sort of been down and down. So, so each line,

David Wright:

these are premium to GDP. So this so the thing I'm getting out of this is seeing how I mean the sign, like of the slope of the different lines of business, It's less than magnitude is just so different. I mean, it's it's incredible how diverse they are. Yes. And homeowners as a share of GDP, that's kind of fascinating. And it's flat for the last little period of time. But it I've heard stories about how, you know, maybe that's it, maybe that's an asset price thing, maybe that's just like the homes are worth more because people put their money into their homes is that, you know,

Steve Mildenhall:

well, homes have gotten bigger, you know, homes have gotten bigger people have moved to coastal states with a higher exposure, construction, you know, issues, and then a lot of sort of just over and just catch up on the, on the on the rate when modeling came in. I remember doing a rate level indication for Florida, when I first started sort of after Andrew with, with cap modeling Incorporated. And if I recall, even before you got to the hurricane load, it was just woefully inadequate. And then the hurricane that was like, 100% on top of that, because I've been thinking about what the appropriate rate and

David Wright:

the one that maybe, you know, just because the because some of these are small scale, like if you look at I mean, I don't know what like a med Mal, it's kind of small, but you know, workers comp. Wow. Like, you know, what happened?

Steve Mildenhall:

Yeah, so workers comp. It's interesting to look at these lines, in terms of the growth. So how have they grown, relative to GDP. So this is indexing everything back to 1992. So let's just go back and look at the total picture, right? So 92, is about here. Okay, we were industry, we were about three and a half percent of GDP. And now we're at about three, right? So we know, overall, if we look at all these lines, on average, we've sort of fallen somewhat behind GDP, right. And that's what we're showing here. So this is the total. And you can see, sure enough, we've fallen somewhat behind the GDP growth that we've had, which is what you'd expect, but look how different they are by line again. So these are sorted by their average growth rates, if I recall. And so work comp is by far has had by far the slowest growth overall, through this period. Now, some of this could be people taking column out of the traditional insurance sector and sort of, you know, self insuring and whatnot, large loads, deductibles, what have you. But it is interesting that so the, the, the solid black line here is the total premium growth, just so you've got that for context, the dotted line is GDP growth. And then the colored line is the line of business. And then on this one in particular, I just added levels of employment, because it looks like comp has actually tracked somewhat closer to employment growth, rather than sort of nominal monetary growth, which when you think about all the claim inflation and what have you is actually kind of stunning. You would have thought it would have outstripped that pretty substantially. But I think the other dynamic you've got going on here is that, you know, we all day average job has just Yeah, right. safer for working in an office than it is working on a construction. Yeah. conceptions?

David Wright:

Yeah. Financial guarantee, obviously one that sticks out as being weird. 2008 is kind of like maybe the only event in the history of that business. That's probably wrong. But certainly the biggest one, but

Steve Mildenhall:

yeah, yeah. Financial guarantees. Well, we'll see a couple of other exam financial guarantee includes a include mortgage guarantee in here as well. So yeah, you can see sort of giant spike there making up for the losses. How long has this been, you know, on the inexorable March, but other than financial guarantee that's had the fastest overall growth driven just by the sort of climate realities that we're, we're going through inland marine is a lot of property coverages in there as well. liability, you got the big, you know, big, big cycle that you see over time. Commercial auto, again, falling behind, way behind, and you would think, you know, I think part of this catch up is sort of us Moving now to a sort of delivery economy, which again, 20 with with COVID, and one I think, is sort of spiked up even more and personal auto, you can see a sort of tracks with commercial auto. But first of all, they're much more rational, personal, you know, tracking that pretty much with, with with the old total is such a big part of the total, obviously, in tracking reasonably well with

David Wright:

me recently in the auto lines, I think you have these sort of like the, I don't know, yin and yang developments of smartphones, on the one hand, distracted driving and then the other hand smartphone components, which are adding intelligence to the cars themselves, right, which are having a good effect.

Steve Mildenhall:

Yeah, I mean, my personal view is that the long term prospect for personal auto is that we will get to, you know, no accidents, somewhat extreme. If I want to run, get a rise out of people, I say no accidents. I just look at what's happening on the aviation industry right now. We can transport that Billions of people with no fatalities in the airline's industry, right, it happened. It's happened. It's been several years where there have been zero passenger fatalities in the US. And in your mostly Europe, North America. So Airlines has gotten incredibly safe. And I think it's people. I think there's a little bit of not thinking outside the box enough, as we imagine what's going to happen with driverless car technology. It's not a bunch of independent cars driving around anymore. It'll be a network of communicate cars, they'll be able to anticipate what they're each going to do, no one will be doing anything irrational, we'll be able to learn from mistakes. I'm sure they'll be the odd, you know, systematic, colossal screw up happen. But overall, I think you know, is that technology comes online, we're going to see very positive impacts on commercial auto, and then on trucking as well as the ability to have sort of convoys of with drivers that aren't going to fall asleep on these very long horn as a very grueling as a job to do right. Very challenging. And so the ideal for a computer just sit there and you know, pay

David Wright:

full again, drunk driving and you know, as things like autopilot, or whatever that comes, you know, just taking kind of some of that human judgment error out of it. I agree. And I can plug another note, a reasonable podcast with Haas hybrid loss data Institute where I'm gonna be releasing a video actually have a time I did with him. So stay tuned audience. We got Next, we want to Can we see this? Can we see the arrows yet? I really want to see the eras graph. I can't wait.

Steve Mildenhall:

So this is a precursor to the arrows. This is the other thing that you might look at, right? Why do we Why do we worry about all this volatility? And why do we need surplus so that we've got the cushion to pay the claims, this is how surplus has varied to GDP. And I think I mentioned that 74 was was kind of the low point they're on on the sort of level of policyholder surplus to GDP. And that's been just sort of rising pretty much as a straight line. Since then, as we've gotten more reserve intensive casualty lines, we've got more assets on the insurance company balance sheet. And you can put these two things together. This is actually john, no major tipped me off to this, I've never thought of doing this. So if we look at on the x axis here, surplus to GDP, and on the y axis, the premium to GDP and just sort of plot out the point, they naturally kind of cluster into three areas, you've got a period from inception in the 30s. And I only have the surplus data back to 1930, and have the premium data. So from from the 30s, through 68, you're in this sort of period of your lower level of penetration into the economy, right, so the premium to GDP ratio somewhat lower, and the sort of surplus to GDP levels sort of medium, right, so so fairly well capitalized period, between 68 and 69, you jump over into what I call phase two, and this is this sort of period where we were looking to find coverage, expand insurance grow, you can see that was done on relatively low levels of capitalization, right, the surplus to GDP ratio here, so in the one and a half range, historically are much lower than it was through much of the prior period. But decent levels of premium come up in growing levels of premium to GDP in into the economy. And then between 85 and 86, we had that big jump that we saw on the first chart. And that takes us over into what I called phase three, which has been this sort of period of consolidation, maybe greater concentration on solvency and sort of, you know, responsibility, if you will. And that that pushes us, you know, we we've been in that period, since 86. You just highlight that period on the right hand side here. So this is just a sort of zoom in. It's slightly different, though, because I was wondering, Well, what can we predict where is premium to GDP going to go right? Is that something you'd be interested in? Well, and this is contemporaneous, this is this year's surplus to GDP against this year's premium to GDP. But on the right hand side here, I'm going to show prior year end surplus to GDP, because that could be something that would then influence premium in the coming year. And you're going to actually see that the your face three dots, they tighten up. And they fall really quite nicely along a regression line here linking the two variables. In fact, the R squared for that regression is about 86%. So there's a pretty good regression. And you can see now that there's no extreme outliers that the dots are sort of fairly well distributed across the two sides. It's a pretty good look in regression. It's got very nice statistics. And going back to our question that, you know, we began with about, well, what's happened to the cycle of we're going to, we're going to continue to see the cycle. This very clearly says, If you want premium to jump up, you're going to need surplus to fold out right. There's it's clearly stating that and so we go back to our explanation of why we haven't seen big spikes in premium is because there's that capital kind of sitting on the edge waiting to be deployed, that essentially stops this surplus from falling too much because new capital comes in?

David Wright:

Well, it's, in some ways, it's like, I think, intellectually sacrilegious? Maybe, Steve? Because, right, because you're kind of like, throw out the manual on exposure rating or anything, right in terms of rate levels, because all you're saying that, you know, whether you like it or not, all it really matters is whether you made money last year, or kind of the last level or the last five years? Well, it is like an old mentor of mine say, Well, what are the last five years? And then now, I can tell you what the sentiments gonna be like next year. And again, it has nothing to do with anything else, you can't feel pain, right? You can't you can't you can't you can talk a market, you can't talk in any direction.

Steve Mildenhall:

Well, yeah. But I mean, you're you're an expert, you're not throwing the manual out the manuals based on the luxurious findings as well. And so it

David Wright:

was like, I mean, exposure rating. So exposure rating has this has this sort of, like higher kind of calling, right, you made the point earlier about the property cap rates, you know, where like, there's something there's some there's something called a unit of risk, which exists. And we can measure that Don mango has, you know, he says lots of interesting things about, you know, units of risk with internet, Internet of Things devices, and like, where if you can measure risk, call it you know, first define it, then measure it directly. Now, you can tell actually, how much you should charge for something. So as independent of results, there's there's like this platonic ideal of what the right rate is, and which is independent of your results, and you just go towards that right rate? Do you agree with that as possible? Given your fairness?

Steve Mildenhall:

I think there is for, you know, personal auto and dingy several lines. But I don't think there's such a thing in catastrophe reinsurance, because I think it would be like saying, hey, there's a right value for a stock. And we know that you know, the value can range. And honestly, because you look at cat reinsurance, the premium is somewhere between two and five times the expected loss, right, you see loss ratios in the sort of 20 to 40% range. So the risk premium in that is a very, very substantial piece of the whole. So I think that platonic ideal works in some lines in some areas, and it doesn't work so well in other areas. I think another way is you just need to bring in sort of market forces to tell you what you're

David Wright:

at. You know, another interesting point about this graph is that it doesn't include the alternative capital, right? So surplus wouldn't be measured there, I wouldn't think because that winds up getting reinsured. And that's outside the system. So how does that play into into the story? You think?

Steve Mildenhall:

Well, it Yeah. So I that's something I've, so this, these slides Do you mentioned at the beginning, that is is pricing insurance risk costs. I've been spending a lot of time with john on the book for the thinking about that. And thinking about how to how to price is pricing for risk is what we're talking about. We're not talking about predictive modeling and getting to the lowest cost. We're, we're sloughing that off on someone else. And assuming that heavy lifting has been done, we just want to like gross it up to get to a premium. And reinsurance is really interesting, because it's sort of a low expense. Yeah, I spent 13 years selling reinsurance and I was all over the reinsurance capital. I fully believe that I mean, it is it is a form of capital. But I had never really thought about it. It's sort of in what, what, what is capital? When we talk about capital, kind of what do we mean? What we generally mean is, we've got a unit of risk here. And I need to my my regulator tells me, my regulator, my reading agency, they tell me how much assets I need to support that risk, right? So to credibly issue, this insurance policy, I need a certain amount of assets, which is the left hand side of your balance sheet, and you need to finance it, right? So you need to how do you finance you've got two sources of financing. You've got premium you're going to collect from your insurance. And then you've got in equity that is paid into your entity by your investors. And why did the investors pay in equity because they are buying the residual value of the firm? Right? And so that sort of video the highlight if you've got an all equity balance sheet now if you bring reinsurers in what a reinsurance do reinsurers by the residual in a specific sort of subset of your book. And it could be a whole account quarter, it could be on everything, but it could be just a sort of cap layer or what have you. And one of the reasons that reinsurance is a little tricky to get your arms around is that it doesn't get booked on the balance sheet in an explicit way. So you can Alo yo you do gap accounts, you do gross reserves and you do you know you could do gross premium and then you net out the session. So it's sort of quasi gross, but you never gross your balance sheet up to reflect the limit but your purchased our reinsurance right What would actually be and you know, and you can see that sort of more in some areas than others, you get a Florida homeowners where you get these very reinsurance dependent little companies, it actually would be much more informative, to be able to look at a balance sheet gross that up. And for some types of reinsurance, you can do it, it's pretty easy cat in particular, you're about $100 million, a captain is basically$100 million a capital and I just plug it into my capital structure. But as you said, You know, I do a casualty quarter share, that's maybe sort of unlimited in multi year and the dollar is not obvious, exactly kind of how I would do that. And so I think for good reason, you know, the accounts don't do that. But it is it would be, it would make a lot of things clearer. If we could actually do that. And yeah, it would change the dynamics here. Because obviously, there's, you know, there's about $100 billion of alternative capital out there supporting, you know, a pretty good proportion of the of the cap,

David Wright:

there are a couple of ideas there that I really like talking about. So the metaphor I like to use for reinsurance is that it shrinks the insurance company for a given capital base, right. So they're just sort of less of everything in IT assets and liabilities. And I think that, so that's kind of one thing I like to say another thing that I like to point out is, you know, your comment there about not treating the limits of reinsurance, you know, explicitly, to me kind of shoots an arrow right at the heart of the kind of the miracle of insurance, right. And the miracle of insurance is that you don't count your limits, you asked the actuary what you should put as your liability, and that's what goes in there, right. And so you can destroy risk in kind of a very real accounting sense, and have a non insurance company writes an insurance policy, they call it a loan, and they put a liability on their balance sheet for the equal to that limited loan, or whatever we're gonna call it, right, the insurance company can sort of wave that away. And so they ignore limits, which is kind of the heart of the whole thing.

Steve Mildenhall:

Well, that gets to why insurance works, right? I mean, it's a people love to say there's no free lunch. If the I think actually there is a free lunch. It's diversification, diversification, because you can pull risk, essentially costlessly. And you absolutely all benefit from that. And I think one of the big and important distinctions in insurance is lines that truly diversify well, and lacks that dumb, and we know what they are right. And so if you're a Katla, and you do not diversify well, and it just doesn't work brilliantly within your typical equity based insurance company balance sheet. And that's why people want to offload that type of risk to sort of different investors, different appetites, different structures, and what have you. Whereas personal auto, you know, whichever way you look at it, that is, how many cars out there a 150 million vehicles or something in this country that are being held, they are largely independent, right, and they are not going to have an accident at the same time. And in fact, you know, the weather is really bad, it's great people stop driving. But you know, the worst of the worst is that you have some snap storm that comes through after morning rush out, everybody's got to drive home. But mostly you can see the weather coming, it's going to be terrible, you know, on truly bad weather base, the results get better because everyone stays at home. So yeah, it the limit goes away, because you you're an equity insurance company true is a brilliant structure for bearing that type of diversifiable risk. But it is not a good structure for bearing risk that doesn't diversify. And you know, that's why we have such a vibrant insurance market is making different investors that can bear that,

David Wright:

I want to kind of try and tie this back to this graph. Because, you know, what would the impact of this alternative capital B's is, you know, capital for the most part is taking out these and diversifiable risks, which I think is the story that you might endorse, then you're going to wind up with less volatile surplus, I suppose. Although we have this puzzle about the cat risk, the non diversifiable risk and kind of not seeming to really matter. So like, if you put this way, if you pull the alternative capital out of the system, would you regression change? Or if you

Steve Mildenhall:

think the line would shift outwards to the north, the North East, right? Because you would have more premium? Because this is net premium, right? So if you didn't have so much reinsurance, you'd be you'd be keeping that premium, and you would have to have more surplus, I'm not sure what would happen to the right slope,

David Wright:

because that would still be affected by your perspective, profit, I guess or whether or not you have money or whether you got it right. You know, you can think of like, the slope as being a kind of index of surprise, maybe or something. Right. So it's like, How surprising Am I finding results?

Steve Mildenhall:

is partly an indicator of sort of cap was capital adequacy, right? I mean, so you've sort of got premiums of service, basically, premium surplus ratio, this is the slide

David Wright:

so I think we only got a few minutes left here. So like, let's talk about what What we take from all this for pricing insurance risk? What do we what do we? How do we use all this fascinating history.

Steve Mildenhall:

So I think we need to learn that we've got, we talked about the insurance market, and it's not a single market, it's quite complicated. You've got these the personal commercial distinction, I think more accurately, you've got a sort of mass market specialty market distinction. In mass market, you're tending to deal with customers who they don't have a, they don't expect to have a claim. And so it's very much a transactional purchase, it's impersonal, there's very few interactions with the insurance company during the year, it can move around. And your rating of proxy business brand is super important. Because it's it's a, it's a, you know, let's face it, no one wants to buy insurance, right, there's not the top of anyone's agenda to go out and shop for insurance, you want to get it down as easily as possible. So brand becomes important, automation becomes important. And we've got that sort of side of the market. Even though we saw with PERS model, right, we've had swings, but not not so much. Then we got the commercial line side, which sort of splits into a piece of that is like your personal lines in the you know, it's the sort of small commercial side with the same type of reading technology, and then you've got the larger risks, where gradually, you know, risk low becomes a more important factor as you're providing a high limit property and what have you, but they tend to have claims year Oh, you know, year over year, like the carrying a body of claims over maybe makes them sort of stickier. So I think, you know, understanding that the different lines have different drivers. And going back to sort of where it started, as we think about pricing, your the risk, you've got event risk and cap risk, which is very real and can't be run away from that property driven. You've then got the risk that you're you know, we know all the pre we know what the premium should be. We know Don mangles platonic ideal, but can we get it in the market, which kind of comes down to people are have a tendency to push the granularity of pricing right up to a breaking point. And then you've got lines where you don't know the premier, I guess, probably, you know, the financial things in the mid 2000s. Was it was a great example of that we thought we did but in fact, you know, maybe we didn't have such a good, good idea. And there's a series of slides, you know, I guess we're about halfway through what I thought we might talk about today, which is fine, this has been a great conversation. But I think if we do part two of this, we can dig into some of those drivers by line of business and see some very interesting things in terms of different behaviors on the last side. And the premium side that really sort of informed how people underwrite in those lines, and how they think about selling, and how your loss ratio was we think about the loss ratio volatility is, is driven by loss and premium, right, and you need to separate those two things out. And look at those.

David Wright:

I think that's a great idea, Steve, inviting yourself back into my show. But I'll take it. And so we'll hold on that. But I do think that there is potentially some macro conclusions we can draw from this, because you just have to kind of I mean, you got a regression line there, Steve. So if we extrapolate this line, right, what we will see is we will see that premium to GDP will continue to decline. Right? That's one thing we'll see right? over time. And are there What does that kind of mean? Does that mean that we, with the surplus GDP declining, though somehow that's going to continue increasing? So that means like probably a narrowing of return on capital continue dropping? Right, if that keeps going? What else can we infer? Well, I

Steve Mildenhall:

think we're we're in furring is that if we're expecting that to be a big change in premium, it needs to be driven by a capital event. That's the big takeaway here, right. So we're thinking we're currently at a one a 3.15, I think the 2020 delta is between the three and 3.2, if you want to get back to where we were in 92, which was you can see is like three point, you know, 5% of GDP, you this would be suggesting that you'd need to see, you need, you need to take surplus to GDP down from 4% of two and a half percent, which is what 15 out of 40, which is a big old, you know, a big proportion. And you can just imagine what the rating agencies are thinking about that if we, you know, our surplus levels go down that much. So, you know, is that likely to happen? Or as we think about growth in the industry, should we be thinking, Hey, you know, we're, we're gonna grow with GDP. And so that, you know, that has implications because it means I'm probably if I do organic growth, I'm stealing someone else's lunch, when and that's going to be tough. There's going to become, you know, so how am I going to develop that competitive advantage that I'm going to be able to win business? Well, we can see a couple of ways that people do that in the auto space. So we've got the inexorable March of, of Geico and progressive and you No low expenses right? They're coming into the fight over the premium dollar, five points is already in their pocket like more than five points in many cases through expenses. So some people have to you know, decided to do that. Other players are going with a cost of capital advantage. You can look at the on your Mutual's come in. And overall, they kind of run out about the same but the Delta comes out with policyholder dividends, it gives them you know, essentially an expense advantage over the stocks because they don't have to pay maybe so much for their their cost of capital. So I think there's an interesting, competitive,

David Wright:

I think it's a great lead in to to, to our next conversation, because to me, one thing of extraordinary competition, extreme competition, which is I think, what increases as you know, we have benign and we're having a kind of a weird thing happening right now. But I think overall, a benign period means removed GDP will continue dropping, that means you probably Nisha five, many more pieces of business and so you have to look in a really granular manner, at everything. But with that, we will close my guest, Steve bell to hell. Thank you very much, Steve. Look forward to having you back.

Steve Mildenhall:

Thank you very much.