Samir Shah on Innovating Capital

WEBVTT
00:00:00.450 --> 00:00:11.820
My guest today is Samir Shah, co founder and CEO of ledger investing former chief reinsurance officer and head of capital market today. AIG. Samir started his career as a pension actuary and management consultant. Samir, welcome to the show.
00:00:12.210 --> 00:00:13.589
Thanks. Thanks very much, David.
00:00:13.810 --> 00:00:16.329
First question.
00:00:13.810 --> 00:00:40.600
So, insurance is a promise to pay and solvency regulation. And here I'm lumping together the regulators themselves and also rating agencies, which I, you know, I think jointly administered the solvency regulation and it kind of informal sense, they exist to ensure these promises to pay are met. Now to increase liquidity, and we can get into that if you'd like, unrated insurance and reinsurance sits outside of that system. So I'm wondering, should insurers be worried about the promise to pay?
00:00:41.740 --> 00:00:47.020
Yeah. So, you know, there's it's a good question.
00:00:47.320 --> 00:01:04.209
There's a short, technical answer to this. But I think there's a longer answer, that is more fundamental. So that the short technical answer is this.
00:00:58.479 --> 00:01:04.209
The Yes, insurance is a promise.
00:01:06.700 --> 00:01:24.700
However, it is, you know, in most insurers, reinsure, and so they get promises for other people. Yeah. And it's promises that are built on promises, and the reinsurers do retro. So it's like promises built on promises, built on promises.
00:01:24.730 --> 00:01:25.120
Yes.
00:01:25.359 --> 00:01:28.150
And, and hopefully regulators are managing that.
00:01:28.150 --> 00:02:10.689
But it's really a network of promises. And the real risk to a consumer, or to you know, commercial buyer is whether this network fails in some systemic way. And we've seen stresses of that, you know, once in a while, but but generally that, notwithstanding the fact that it's promises built on promises, it's performed pretty well, for a long, long time. And I recall reading, I think it was a report that said the generally it acts like a double A credit, because even when insurers fail, which they fail more frequently than double A credit, somebody comes in and buys them up and sort of honors their obligations. Okay.
00:02:11.020 --> 00:02:14.229
That's what they have now.
00:02:11.020 --> 00:02:28.360
Promises built on promises, but it's worked pretty well. When you go outside of that into the ILS market. It's not really for liquidity, it's, it's to tap into a deeper pool of capital.
00:02:23.949 --> 00:03:17.680
And so the short technical answer to your question is that when it goes there, it's not really outside the system. But because it's not a rated balance sheet, they fully collateralized their obligations. So what's better than a promise is to say, Here's the money, the maximum amount of money that could ever owe you. And here it is sitting in a trust. And you can draw on that money anytime you want, without my permission to pay your claims, that's actually much better than a than a promise. And so in that sense, in that very narrow sense, the, for a consumer for the insured, if I know that I'm getting a promise from an insurance company that's backed by real collateral sitting someplace, rather than another network of promises, that's safer for me.
00:03:19.000 --> 00:03:24.219
And so, you know, let me pause there, that's the The short answer.
00:03:25.229 --> 00:03:41.340
So I want I want to, like, you know, push on one part there, where you, you push back on my, my, my characterization of liquidy being the problem. So I want to see, I'm gonna keep going on that for a second. And we can and round back, we will spend an entire hour talking about unpacking your answer. They're similar, I suspect, which would be wonderful.
00:03:41.370 --> 00:03:58.710
Actually, I you know, David, I really think this is a very fundamental, I'm glad you start from the consumer in short perspective. Because you're right, I think we could spend hours on from that perspective, what's happening in the market, and whether it's a net good or neutral, bad.
00:03:59.469 --> 00:04:02.199
So away we go.
00:03:59.469 --> 00:05:01.029
Now, now, I think liquidity is important, I think liquidity is critical. Because if I think about, you know, the capital providers here, these kinds of large pools of capital, their problem is they're saying, if I want to write insurance risk, then the regulator keeps it, and it's very hard to get it back out again. Right. And so if I'm going to capitalize a liability policy, or even a property policy for that matter, you know, there's a world in which the consumer doesn't want to say it doesn't want to commute their policy, right? So they don't want to say I'm done. You know, if anything happens in the future, you know, you can you don't have to pay any more to consumers want you know, they buy a promise that promises good forever, right? If asbestos happens, you know, 50 years from now, they can still claim and so the investors though don't like that because they're saying Well, I have I have customers who want their money back at some point. And so I need to be able to provide that kind of liquidity to them. So to me like the real call it the financial innovation, of of collateralized reinsurance is to provide liquidity in a certain sense for Insurance liabilities.
00:05:02.439 --> 00:05:34.959
Yeah, I see your point. But I think you're also pointing out that they want their money back. So they're not offering you know, they're there. It's providing liquidity to the investor at the expense of the insured is what you're pointing out, right? I think that however, what's happening is that that initial promise is still the one that the insured has to rely on. And it's the network of promises behind them.
00:05:34.959 --> 00:05:59.170
And so what happens is, when the insured needs to offer liquidity to somebody, they're essentially saying, Okay, I'm taking back that risk. And now it's my price. So it's always been my promise, it's always going to be my promise. Now, how much to what extent Am I going to rely on other promises, that's really where the liquidity comes in.
00:05:55.360 --> 00:06:02.079
When you go to a reinsurer, that promise can be good forever, technically, it's good forever.
00:06:02.079 --> 00:06:51.490
But practically, it's as good as long as long as the company is, is solid. And on the other side, it says, Good, as long as the investor doesn't want their money back, if the investors want their money back, that's great. So I take that promise back. And so this isn't the issue of this isn't an issue that should affect the insurance, because what's happening is that some portion of the forever part is being capitalized by money sitting in a trust, whereas currently, none of it is being capitalized by money sitting in a trust, right, all of it is just built on promises. In your, there's no more promise. So even having that for a short period of time is better than not having it at all.
00:06:52.350 --> 00:08:04.829
So let me like, let me see if I can, you know, summarize this. So we have two models, one where you have an insurance balance sheet, and that insurance balance sheet does not have $1, for every dollar of limit, right? Because that would be too big. You know, they write the limits that the ratio of premium to limits is, you know, something very large, limits premium. And so you can't really do that for most insurance policies. But the regulator says, okay, you're, you're we're gonna break the rules, I called the miracle of insurance, right? The miracle of insurance is the accounting rules, say if you have an obligation, you have to put money in your balance sheet to fund that obligation to finance that obligation, right. And an insurance regulator says you don't, we're gonna ask an actuary what they think is going to happen. And that's how much you put in the bank, right. And so there's no money there. And it's not in a trust so much. But it let's call it, you know, equivalently restricted, to a trust. So the whole point of the trust is that it mirrors you know, the restrictions and capital flows of an insurance company. And then we just watch that to make sure that they're not screwing up their estimates, or we have the actuary sign stuff, and they get in trouble if they mess it up. We have this independent body kind of quasi independent auditing actuaries, and regulators. So they're all kind of watching each other, as you say, the network of trust.
00:08:02.670 --> 00:08:34.590
And then we have this sort of, outside of the system, we have this collateralization of the limits, which as you point out, is probably better, because there's no need to have an actuary tell you whether it's solvent or not. It's like, it's all there. The money's in the bank for property. But the problem though, is that like, as I'm pointing out, is it It ends, like they shut it down? Like we closed the no hurricane happened? Like, we pulled the money out of the trust, and now there's nothing left. And and there's, I mean, I prepared a lot of these collateralized reinsurance negotiations, where it's like, when do you, you know, when do you actually call it quits? And then, you know,
00:08:34.000 --> 00:09:07.000
yeah, I think we're okay, I see your point. But I think we're mixing a couple of different dynamics here. So let me try to generalize to get all of these dynamics in, let's say that in that insurance company, that they have $100 of capital that they have to maintain, and you characterize it as sort of a miracle, but it's really based on the law of large numbers rarely, right? And so, if not, for the law of large numbers, nobody would be insured. Right?
00:09:02.950 --> 00:10:09.460
So it's, it's a net good thing for insurance insurance, that there is such a dynamic as a law of large number that allows them to say only hold $100 even though all my limits add up to you know, 50 times. So okay, now they have $100 of capital. Now, the question is, what is the best place to get that $100 of capital? So one choice is in the in the choice that the industry has used for a long time, is it's going to be permanent capital, because after all, it's a promise. So it's nice to say that I have permanent capital, okay. But, you know, as we as we just agreed that you know, that$100 a capital, sometimes what they do is say, I'm not gonna keep $100 myself, I will take$40 of that and get it from On promises from other insurance companies, who are also going through these law of large numbers exercise and holding certain amount of capital, and quite likely less than 40.
00:10:11.019 --> 00:10:55.690
Right? So now I have $60 of capital, I'm relying on $40 of capital from somebody else, I promise, okay. Now I could, instead of working on a promise, I could go to the ILS market and have $40 a capital sit in a trust. And, and you're right, there's an end to that. But essentially, when that ends at any moment in time, when that ends, unless I'm renewing it, I have to have a hundred dollars of capital again. Okay. So and that's really the way the industry works is they're always supposed to have enough capability to have the equivalent of $100 of capital.
00:10:50.740 --> 00:11:39.580
So, so the question isn't so much. You know, when does that end, it's like, if you, if you created a Venn diagram of the amount of capital, let's say on the y axis, and the x axis is like time, so you need a certain amount of capital for each moment in time. For the amount of business you write, it decreases over time. And you say, well, for a portion of that, I'm going to get an ILS market. And it means that I've planned for not having that at some point. And, and, and I'm taking care of the rest of the capital. So you know, the, the insured is not worse off, when they can go to the ILS market, they're better off in two ways.
00:11:39.580 --> 00:11:43.539
One is that, first of all, this this promise is more reliable.
00:11:43.750 --> 00:11:49.509
The second thing is, the whole reason to do this is because there's more capital, and it can be potentially cheaper capital.
00:11:50.950 --> 00:12:57.159
And so let me introduce that dimension of it as a... So the overall perspective of what insurance should care about what's happening right now in the industry, or if it happened for a long time is that, you know, there's, I know, somewhere between one and a half and $2 trillion of capital in the property, global property casualty industry, depending on how you count, you know, real capital that's fungible over life and property Casualty. And it's split over 1000s of balance sheets around the world. And notwithstanding the fact that there's one and a half to $2 trillion capital, there is a lot of risks that is still uninsured. And you can see, you know, California, there's only 10 to 12% of the market actually buys earthquake insurance, even though there's you know, people say there's a lot of capital, there are many institutions ciceri, or you and others who who monitor the economic loss versus the insured loss. And there's always a huge gap there.
00:12:57.639 --> 00:13:00.669
And that gap is increasing what they call the protection gap.
00:13:01.659 --> 00:13:18.159
And the reason for those gaps, which all hurt the all of us as consumers of insurance, is that we're not able to insure a lot of things at an effective price.
00:13:12.730 --> 00:14:40.690
And so the issue is the cost of that capital. And when you go beyond these several 1000 balance sheets that are highly concentrated, you know, and when you have limited, you know, when you have these small balance sheets, as the risk goes up, as the more you write the the cost increases exponentially. So if you can tap into the capital markets, which is a much, much deeper market, and can better absorb these concentration, as you write more risk, the cost doesn't go up exponentially goes up linearly. And so what that should do is open up the possibility for offering insurance at more economic rates. So the net plus to an insured is, first they have a promise, but they also have a price for that promise. So one thing is they want to make sure the promises kept. The second thing is they want to be able to insure more at a good price. And I think the capital markets does both of those things. Because for the portion of the poor promise that they are the capital markets are taking on, which is not only amount, but also time, like I'm going to give you $100 but I'm going to give it to you for a year, and then I want it back. So for that portion, it's more reliable, and it's cheaper. As a long term basis, it could be cheaper.
00:14:40.740 --> 00:15:38.970
Let me let me come back to an earlier point and we're gonna rake back over all this again. Okay, we're gonna keep doing this, because there's just so much here, I think and and I've always had a few things that have bothered me about this. And so, let me let me tell you one thing. So let me let me try Let me try and summarize something you said a little earlier and come back to it and let's see if see where we go. So you mentioned earlier that when you read a promise as an insurance company, you you don't necessarily capitalize that promise 100% of internal capital, you're not raising equity to fund at all, you're buying reinsurance to offset some of that capital. So reinsurance is lowering the amount of capital you need, and therefore, reinsurance is a form of capital. And then we were making the observation or an observation that the reinsurers are doing the same thing. And so they actually don't have, you know, they don't have as much capital to back that obligation to you, as you would have had yourself because it here's, you know, the sort, we didn't use the word yet. But the source of this, what I'm calling the miracle of insurance is diversification. So diversification is a source of capital. So you're saying the premium from this side and the premium from that side over there, they offset each other, only one of them can go wrong.
00:15:39.240 --> 00:16:05.850
Therefore, we can think of like the premium balance that we have as a as kind of a source of capital itself, right? Because that's what's primarily backstopping the claims is our premium portfolio. And so if you can use diversification to as a source of, of de risking capital, therefore, right, then you need less of it. Right, and on it goes. And so what's what's amazing to me about this, is that the capital markets, I'm going to come to this cost of capital question in a second.
00:16:06.090 --> 00:16:23.399
But it's amazing to me that they are capitalizing the insurance liabilities way more than an insurance company's like there's like, I mean, by an order of magnitude, right, more capital is coming into the system, although the you know, the trade is that it's coming out, it's not in there permanently, right.
00:16:23.399 --> 00:16:59.879
So it's in there for a temporary period of time. And, you know, we're sort of like implicitly talking about your business here, here, all over the place, I know that your big idea is right, because you one of the things that you guys are innovating on, is how and in what when do you actually need that much capital. And you know, because because there's there's 10x, more capital here, let's all sounds great, but you don't need that for the entire length of the public policies, liabilities, because it tapers off and get more certainty. And there's, we should express that certainty in the form of a reduced capital need. So to me, like the magnitude of capital has just gone so far up. And again, my feeling is the trade for that is that you're introducing some liquidity and they should be they want to be able to pull it out in exchange for like capital.
00:17:00.620 --> 00:18:05.809
Yeah. Okay. So I think my sense is that most of these comments are a function of the current ILS market. This is the cap on market. And so it's really about cat risk and how it works, which I don't think is the, the model for the potential for capital markets. And so let me explain the distinction between these two things. The current market is been was primarily motivated by risk management objectives, whereas the the huge opportunity that I'm talking about for insurance is from a capital management perspective. And the difference is this that, from a risk management perspective, what happens is an insurance company takes on risk and says to itself, well, I don't like the earnings volatility profile of my company, I have a lot of concentration here. And so I would like to decrease that earnings volatility, and I will then reinsure it to somebody else. And so they usually buy.
00:18:06.859 --> 00:19:07.039
So excess of loss type covers in order to decrease that volatility. And, and when they do that, they are actually end up paying more then holding it on their balance sheet from a cost of capital perspective. But they do, they're willing to do that, because the net effect of that is lower return and lower risk. And that's the trade off that all the insurance companies make. And so that's a risk management perspective. And so when you buy cap on, or when you when you sponsor a cap on your, you know, you're doing a top layer, you end up quite frankly, paying more than what you would accept because there's friction, there's overhead of reinsurers, etc. Now, capital management perspective says, I am now an insurance company, I'm comfortable with my risk profile, I've done whatever I need to for risk management purposes, and I need to hold the$100 of capital, okay. $100 against limits that are, you know, 10,000 I know some multiple of that $100. Right.
00:19:07.490 --> 00:19:53.839
And, and I only want to source capital that is $100. And not more than that. Okay. And so what happens is, if you take the bottom layer of the capital tower, and say, not... now you're not talking about risk, you're just saying I hold the$100 of capital, and which ends up being you know, some small fraction of the reserves are a small fraction of the premium, but unwilling to go out and say I want, you know, you know, 30 or $40 of that I'm going to go into the capital markets, so they raise 30 or $40 of capital.
00:19:48.289 --> 00:20:01.880
That system doesn't mean that there's more capital overall.
00:19:53.930 --> 00:20:14.960
What what happens is when you're when you're going into the ILS market for top layers, you end up over collateralizing. Because if that top layer was kept inside of an insurance company, the amount of capital they would contribute to it would be a fraction of the full limit.
00:20:15.559 --> 00:20:57.950
Whereas when you go into the cat bond market, you collateralize the full limit. But when you do a bottom layers, that's less the case. And so, you know, I've gone through an exercise where I've tried to optimize, like, where in the capital tower should you source capital on a collateralized basis, and it's always the bottom layers, because the bottom layers are, are less fungible than the top layers. Right? I mean, you get the maximum diversification benefit on the top layers on a balance sheet, you know, if something has a 1% chance of happening, and you have 10 product lines, and you have 1% chance of something happening bad on 10, product lines, they're not all going to happen.
00:20:55.369 --> 00:21:01.009
And so you hold enough capital for maybe one or two product lines to go bad. Not all 10
00:21:01.039 --> 00:21:01.519
interesting.
00:21:01.519 --> 00:21:05.839
When you get to the bottom, you actually have to hold something for meaningful.
00:21:06.109 --> 00:21:18.259
Yeah. So can I kind of make kind of make Can I just make a point here, just let me put this in my words, because I think this is pretty important idea. And it comes at something that I like thinking about sometimes, which is to say that capital markets are actually not very good at tail risk, right?
00:21:18.259 --> 00:21:18.710
Because
00:21:18.710 --> 00:21:19.069
correct,
00:21:19.069 --> 00:21:20.599
because they, you know, you look at options,
00:21:20.599 --> 00:21:21.619
they're not efficient at that,
00:21:21.690 --> 00:21:37.109
yeah, they don't want or they don't want, because they only have to put capital up against this one shot tail. And it's like, you know, what you want to be able to do is you want me to put all the tails into a pot, right, you can fund any one happening. And now the diversification versus that it's most powerful, right?
00:21:37.750 --> 00:21:51.130
Yeah,that is the most efficient. And so from a capital perspective, you're right. collateralization is, is less efficient than a balance sheet, always, by definition.
00:21:51.430 --> 00:22:24.849
But the difference here, the offsetting factors are that it can be on a long term basis, a lower cost of capital, and it can be more of it. And both of those things are actually the what's been, from my perspective, big issues in the insurance industry. One is the protection gap. You know, we were saying like, I used to work for AIG, and even AIG, a big company, you know, couldn't offer limits above a certainty.
00:22:24.849 --> 00:22:58.029
You know, there were insureds who couldn't get as much insurance as they wanted, notwithstanding the fact that you're a big company. And so, going into the capital markets helps with that. And on a long term basis, it should be cheaper, because in there are there are a bunch of reasons why it can be cheaper. One is that, first of all, it can be right now, the risk is not traded.
00:22:50.740 --> 00:23:04.089
Equities traded, but really you don't have a very transparent liquid market for insurance with insurance risk are very opaque.
00:23:04.299 --> 00:23:04.630
Yes.
00:23:05.230 --> 00:23:15.069
And this is the reason why this is the main reason why insurance capital is not a commodity. It capital is a commodity in most industries.
00:23:11.380 --> 00:23:26.259
Insurance is one of the largest, the last large industry, where capital is not a commodity. And the main reason is because risks are not transparent. And there's a lot of middlemen like you say, you have to go to an actuary.
00:23:26.529 --> 00:23:33.039
And the fact is, even if you can go to actuary if you go to three different actuaries they give you three different answers, right? I mean, even though it's a professional
00:23:33.039 --> 00:23:33.640
or six different answers
00:23:34.349 --> 00:26:00.509
I mean, all of those things hurt when you're trying to commoditize capital. And so transparency of risk and liquidity in which would then create true liquidity, being able to trade it, you know, the reason why people say I want my capital back, because they can't really there isn't really a secondary market. For these things like there is for most other robust sort of asset classes, like you and I. And I compare this to the credit markets, the credit markets are our.... create the precedent here for ILS. Remember, banks used to be just like insurance companies. They used to warehouse risk, they were very opaque, very relationship based, and securitization, open that open those markets up and what happened as a result of securitization, consumers were able to access credit in much more real like right now, you and I can just go on our website, enter some information and get a loan. We don't even have to go to a bank. And who's the other person who's going to give us a loan somebody else like I could borrow something and you could be the guy who is lending me the money into access to credit has benefited society and through through securitization through commoditization of credit risk capital. And, and one of the key factors that allowed for that transformation is standardization and transparency and credit risk, you know, using FIFO scores, for example, you know, created a great standard and made things transparent. And at the portfolio level, using, you know, when JP Morgan's risk metrics team started using stochastic var, as a way to manage portfolio credit risk, that, that allowed people to take portfolios of credit risk, understand the risk, read it and then traded, this is the same opportunity in on in the insurance side is if we can make this risk transparent. We can go into a much deeper capital market. And not only can we bridge the protection gap, but I think that it will make this entire thing much more efficient it can you can have cheaper insurance, and you can have more of it.
00:26:00.900 --> 00:26:37.380
Yeah, so let me let some, you know, a few things in there, again, that I am very happy. I mean, you're almost predicting the list of things I wanted to talk to you about. So one of them is this liquidity of insurance liability. So I think that's this huge idea, you know, the kind of ground that in some transactional, reinsurance people listen to this. The hit rate for loss portfolio transfers is very, very low. I mean, how many of these things as a reinsurance broker in those days in my career, did I ever go out, and, you know, they just never turn out. I mean, they do sometimes, of course, but very rare, because of the information asymmetry is always my, you know, my explanation as well.
00:26:34.049 --> 00:27:18.299
But a very much is a lack of a market or a lack of a, you know, especially compared to other asset markets, as you're pointing out there, there isn't any liquidity insurance liabilities. But to me, like, the lesson I take from that is that is to wonder, you know, why we should presume that we can create liquidity in liabilities anywhere, you know, that I mean, to me, like, we got to take that seriously, right. And I see property risk, I see a way that we've gotten through that through the, you know, the most extraordinarily successful insure tech boom of the last century has been cat modeling, right? Because that's created some transparency into risk. And as the simplest form of cat models, you stuff on a map, right? And just, you know, pretend a hurricane and how much of them do you lose? Right, so we can actually see where it is.
00:27:15.809 --> 00:27:54.990
And you can actually get some insight into into the risk characterised with the portfolio, maybe not so much the underwriting characteristics, definitely risk characteristics, certainly in the tail, and that that allowed some liquidity to occur, right, I think there's unlock certain amount of liquidity because we have transparency in risk, such a thing doesn't exist, kind of a kind of anywhere else. And so if I come to your, you know, to the to ledger, which focuses on liability lines, and you know, I know it's a bit broader than that, and probably overly narrowing it, but you can tell me how you want to characterize it. But you know, presuming to be able to create some kind of liquidity in the capital structure, to me is incredibly ambitious.
00:27:55.470 --> 00:28:02.400
Yeah. It may be ambitious. But but that's what startups do. They have bold mission.
00:28:02.420 --> 00:28:04.339
Yeah. I'm a supporter, by the way.
00:28:04.509 --> 00:28:08.980
Yeah. Can you do it?
00:28:04.509 --> 00:28:31.000
And? And so let me we're doing it actually. And so let me explain how you get transparency first is, is getting information on a real time basis. Okay, so when our transactions you know, in to date, all of our transactions have been with MGAs going through fronted carriers.
00:28:31.000 --> 00:30:05.559
And what we do is connect into their policy and claim admin system directly. And so every day in a completely automated fashion, we have the same information, meaning that our investors have the same information that the underwriters and claims adjusters have. So same information at the same time, okay. This This, by the way, there's also a precedent for supply chain this information connected connectivity in other industries, the supply chain industry, in manufacturing supply chains became more efficient, when they were able to link up when somebody pulls a product off a store shelf, if the the retailer has the information and then the you know, the wholesaler has information or if the warehouse has information, then the manufacturer has information and the supplier to the manufacturer has information that entire if everybody has information at the same time, it reduces the risk to the last person in the supply chain. We have something like that year in the insurance industry, somebody writes the risk, and then several later steps later, somebody actually provides the capital and there is a huge gap there in terms of time, even on cat you know, the fact is that you know, I'll tell you a quick story about this thing and then I'll come back to the other point. So I used to work at at Validus and Validus.
00:30:01.059 --> 00:31:00.940
You know, the founding team, in fact, had done a fantastic job getting all their cat moldings run every day. Okay, so on a real time basis, they had this information of their PMLs from the night before. This. I mean, this is a while ago, they did this, I think probably the first time the first now, and they would in those days, they would go around with their Blackberry, and they would show rating agencies, you know, look, look, we can, you know, we're really on top of risk. Now, what what people didn't realize, of course, was the data that they were getting was from companies like AIG, which had data that was nine months stale. So the fact that they were getting information on a real time basis, though, so what what we have an opportunity now to do is that on, on non cat lines, cat, it's less important because it's, nothing really happens.
00:30:57.789 --> 00:31:54.940
But when you're doing auto or workers comp, everyday, the loss ratio moves and I used to on the first transaction, I used to have an app on my phone, which would tell me exactly what happened the night before, what was the loss ratio? What was the, you know, the claim count, all that information happens real time. So that's one part of it is information on real time basis is needed for commoditization. The second part is analytics. The analytics in the industry is, you know, has been based on actual science and actual science has done a fantastic job of creating a fantastic, good, common good for society over more than 100 years. However, because it's a profession, by its very nature, it is subjective. It's, you know, two different actuaries, as you know, two different actuaries will get two different results, even though it's it, they use a scientific process.
00:31:55.539 --> 00:32:06.819
And that's not great for commoditization of capital, you know, it's a little like having two economists, you know, sort of have different views.
00:32:03.400 --> 00:33:24.039
However, over the last 10 years, and actually, when it comes to predictive modeling more longer than that, we have sort of transitioning from actual science to much more modern statistical tools. You know, people like the, you know, throw around the term AI too much. But let's just say it's much more modern statistical methods that can, in fact, be back tested. So the methods that we use, you know, we have data for the entire industry, for the last 30 years. So if we have a model that forecasts for example, commercial auto liability, we can back test that model on 1000s of insurance over the last 30 years and see how well it works. We can't do that with an actuary. So if you go to an actuary today and say, you know, tell me what your forecast is, I'm commercial law auto for something, I can't back test that actuary. Right. So that is not that was not an opportunity 1020 years ago, and that is an opportunity opportunity now. So when you mix real time information, with more objective, and, and transparent in the modeling itself, notwithstanding the cap models have, you know, the market has gotten used to count models, but the current models themselves are, quite frankly, very opaque.
00:33:20.440 --> 00:35:09.309
They're prepared proprietary black boxes, yeah. Right. And in, and they really can't be back tested, I got, you know, there are times when you can sort of simulate, you know, prior events, but you really can't, you know, truly back test, you know, it's a back test in sort of air quotes. So compared to that, the modeling that we're doing, and our modeling is not proprietary, we give it away. In fact, we we also to the extent that investors have the capability, we not only give them the input data, but we give them the our Python code or our code and say, Look, if you want to rerun our math, and and we produce a risk report, that level of transparency, and replicability has not existed in the insurance industry, ever. And in the end, there were no incentives to create that transparency, because that transparency benefits the ultimate capital provider who can now provide capital more effectively, and therefore also the insured. But it doesn't really benefit all the middlemen in between, right, you and I, and we've all been middlemen in between. and middlemen make more money, the more opaque something is, and the more bespoke it is the you know, the more it's based on relationship, the greater the value the middlemen have. And so this issue about liquid you know, creating, you know, liquidity or real time information is is ambitious, but it's absolutely doable. We're doing it now. I mean, we you can see on the portfolio's that we've done every day, what's happening and it will create the conditions for secondary trading. Whereas for cat, even though there's some secondary trading on cat, it's really mainly for liquidity purposes.
00:35:06.730 --> 00:35:14.559
It's not because people have different points of view on what's happening because there's nothing's actually happening.
00:35:11.469 --> 00:35:49.210
You know, you don't getting real time changes in exposure, or real time views in on the annoying cat risk. But if you're holding auto portfolio, workers comp portfolio during the pandemic, you could get real time information, you could see that the frequency was going down, and you could capture some of those gains and trade it with somebody else. That is not only possible, we're doing it and so that makes insurance look more like credit and follow that precedent in terms of commoditization of capital,
00:35:49.000 --> 00:36:41.800
so if I'm an underwriter, I'm here I'll tell you what I don't like about that story. Right. And by underwriter I mean, like, you know, reinsurance underwriter or even an insurance underwriter for that matter. I'm gonna say, you know, listen to me, you're, you know, that sounds good. I agree with the earlier quote, version, earlier comments you made, which said that the revolution in, in the capital markets for let's say, credit, secondary market for credit, was precipitated by a behavioral score, which is the credit score, the credit score is this thing, which is designed to measure my propensity for repaying loans, and it's like, alleges to or arrogated, you know, to to like, tell a disinterested party, how trustworthy I am, right? It's a clear human institution and credit score, like it's, it's a profoundly like, I don't know, it's almost like a dystopian controversy. I mean, people don't talk about it like this, but that's how I think of it.
00:36:41.800 --> 00:37:31.480
It's like, you know, this is a score, which represents how good of a person I am. From a bank's perspective. I mean, there's like a morality embedded in the credit, right. And we have no such thing in insurance, because insurance is like a different kind of moral behavior. Because like, there's, it's not just a default on a loan, like the purpose of insurance is the default on the, you know, the equivalent of default, which is to take claim under certain circumstances, right. And so there's way bigger gray area for what, you know, good behavior and bad behavior, right. So it's like, way harder to do. And we don't have one, we don't have a proper insurance score, I mean, people will hack away at their credit score and give you the insurance score. But this is, the basis is too far off, that there's no such thing as a true insurance score to like, you know, like, I like to joke with my father in law, who probably never submitted insurance claim, under any circumstances, just like on a person, right? You know, house burns down here, we build it himself, you know, I don't want to you know, and some people like that other people, you know, are not like that.
00:37:32.500 --> 00:37:48.699
There's whole segments of the marketplace, where you know, you're doing is fighting with your insurance all the time, highly adversarial relationship, we have no way of capturing that. And data goes my underwriters response to your, your, you know, your point there, and so absent that data, we will still need underwriters.
00:37:45.969 --> 00:37:52.030
And so, you know, if you're not measuring the right thing, then you're not measuring anything yet.
00:37:52.860 --> 00:38:58.949
Yeah. Okay. So, okay, a couple things. One is, when I brought up the FICO score, I wasn't using it to suggest that we should have an insurance score that, that demonizes people into numbers or anything like what's been in none of that is needed in order to make my argument about the value or the ability to to rate risk. I'm what I'm doing is illustrating the role that transparent information real time information and transparent out analytics has, in getting investors comfortable with risk and rating of the risk rating of the risk is important. Now, the in most of what I'm also talking about is portfolio modeling, because that's where investors provide capital, whereas in credit, they actually because it's an asset, you can actually do match up one loan with one, you know, one borrow with one lender year, you can't, right.
00:38:59.489 --> 00:39:04.619
So, so we're talking about at the portfolio level, always never at the individual.
00:39:04.650 --> 00:39:19.110
Yeah, I'm sorry, can I break in for one sec, just say, a really important feature of insurance is you can't it's the trading the promise itself is really hard. Right? So the the promise, yeah, so like the Novation of a policy and transferring to another carrier like that is freakin hard to do.
00:39:19.289 --> 00:39:22.260
Whereas it seems like in capital markets, you'd have to have the service or
00:39:23.320 --> 00:40:33.489
it's hard to do because the information is opaque and the analytics or methodology is not reliable. I mean, those are two issues. But those are not unsurmountable issues. There's no way that that you you need to keep it opaque in order to provide a good and be able to rate risk. There's no that's not an inherent requirement in assessing insurance rates. That'd be kept opaque. Or that'd be made subjective. You know, you know, the fact is that, you know, the gathering of data and improving the reliability. I mean, there's millions of examples of this stuff. You know, notwithstanding the fact that you know, Vikas, though to may D humanize people just mean simply because how successful it was it was ended up being used too much for the wrong things. And but that's not a problem with the fact that there was a FIFO score.
00:40:33.519 --> 00:40:35.500
Yeah. And it's still being used effectively.
00:40:35.530 --> 00:40:56.650
Yeah, I don't want to critical if I could, by the way, because I do I do kind of believe in the optimistic version of the financial revolution, you know, that the picture you painted earlier, was one of my co success, which I I'm on board with, you know, there are some implications about that philosophical implications, maybe we're a little uncomfortable with but to me, that's part of the trade. You know, like, I mean, we got we got some outcomes, um, you know, which one to clear, like, I'm supportive of the whole? I
00:40:56.650 --> 00:41:15.519
don't know, I don't know, I think it's good to have these debates, because you're right. This is, again, I, I love the fact that you started from a consumer perspective, a lot of times these dialogue start, you know, that I have are about, well, what does this do to reinsures? And what does this do to, you know,
00:41:15.550 --> 00:41:18.280
nobody cares about reinsurers, Samir. They don't even know they exist!
00:41:19.050 --> 00:42:21.449
Exactly, exactly, we should start with the most important people, which, in fact, there's two customers, I think, the people that people have the rest of the people who have the capital, sure, most of the conversation is about the middlemen, rather than rather than those two endpoints. And we have not been servicing those two endpoints. But you're, I think that, you know, predictive modeling, as an example, has, you know, has grown, right? I mean, we can't deny the fact that predictive modeling has grown, there may be some people who want to debate whether it's, it's ethical to use predictive modeling, and you know, what factors go into that and all that, but, but the fact that you can do predictive modeling, and the growth of that, I mean, is is going to benefit the insurance industry, the same kind of thing happens at the portfolio level, which is where we're focused on, we're doing applying the same concepts, but at the, at the portfolio level.
00:42:21.690 --> 00:43:23.039
And now, we tend to the fact that, you know, not everything can be modeled, but I think right now, to think that you absolutely need an underwriter, a person can better process risk, with their own biases, rather than a mathematical model, sort of, you know, is, is sort of dismissive of all the problems with, you know, people, you know, like, you know, you know, I've worked with underwriters a lot, and there's all kinds of biases and underwriting and, and you can eliminate those biases and mathema. Math has its own biases as well. So I don't want to dismiss it. But my point is that the difference between the two is that math can be made, can you can back test it, you can make it transparent, you can lay out the logic very clearly. And let the investor say, gauge the strengths and weaknesses of the that math and price the wrist, which is what they're doing right now in the in ILS market.
00:43:23.400 --> 00:43:43.889
They know AIR and RMS and EQ, er, are not perfect models, they have sort of understood the strengths and weaknesses, and they price through that good investors can price through good investors don't need a perfect model, they need a very transparent model that, that where they can gauge the strengths and weaknesses. And they will, they will price it.
00:43:43.889 --> 00:44:55.800
And I think the same thing applies in non cat risk, in fact, more so in non cat risk, because in non cat risk, things move every day. And you can, you know, one other point to make here about non cat risk is the structures that we that we use are all based on loss ratios, action, your loss ratios, remember, because and we do quota shares because we're doing the bottom layers of the capital tower, not the top. And so what the investor is betting on is loss ratio. If you look at loss ratios in any product line for any insurance companies over the last 30 years, and we have we've done that we've looked at all you know, you see that they generally trend from year to year, they move up and down in trend and they mean revert okay, right. I mean, we all know that loss ratios can't get too high before they come back down. They can't get too low they have to come back up. So that's the fundamental dynamic of this industry. The fact that loss ratios trend and mean revert, as opposed to cat risk, which you know, you know, can blow up on on any day and, and so
00:44:57.900 --> 00:45:09.989
can I can actually prospectives my purse there. It was Just to say that I agree with you in the aggregate, you know, at a, I don't know, some high level of aggregation, they mean revert, but oh, I mean company level they do that.
00:45:10.440 --> 00:45:12.480
Okay, we're still pretty high level log aggregation, right?
00:45:12.480 --> 00:45:35.010
Because, you know, you have a portfolio when companies kill divisions, now they kill divisions like they will, they will, they will eliminate a portfolio because I don't believe this can make money anymore. And alpha goes, great companies go bust occasionally as well. So like, I'm thinking like, imagine the s&p 500 stock index, right, so the index goes up and up and up. But some companies come in and out, right? So the composition can change the loss ratio over time.
00:45:35.010 --> 00:45:38.250
And if you pick the wrong horse inside of that competition, you can still lose.
00:45:38.518 --> 00:45:49.108
Yeah, if you're picking horses, but first of all, the fact that divisions can be cut, supports the idea that loss ratios can't go up forever.
00:45:45.688 --> 00:46:54.298
If your loss ratios are going up, and you're not able to figure it out. Either you bring it down, or you're dead. Right, right. And so you're out of business. So there's no you know, so it goes to somebody else who can was more skillful in that. And so the point still remains is that loss ratios trend up and down, and they mean revert. And as an investor, you know, I think the opportunity now is not to seek alpha by picking specific underwriters and trying to gauge who's better or worse, the opportunity is to get broad access to this risk, which creates alpha in the sense that this is diversifying risk, which is, you know, the only free lunch you can have in the capital markets, right? It's, this is the law of large numbers. And the idea is that if you get a good cross section of industry, of risk, that mean reverts, and you're in it strategically for a long time, and the information is transparent, and the analytics are objective, and reliable and transparent, then you have all the conditions for commodities market,
00:46:54.719 --> 00:47:26.369
you wind up with an insurance company in another kind of characterization of that. And I want to come back to there's a few things I want to make sure we touch on before we run out of time, hopefully, we'll make them. So one of them is you mentioned earlier about cost of capital advantage, right for capital markets. And here's something that bothers me about that. Which is to say that the pension funds, let's just call it pension funds, ultimate capital providers are cheaper and writing cat bonds, let's say then they are, but but they're the same people who are buying the equities. So why are they giving a better deal on this side of the house, then?
00:47:27.079 --> 00:48:26.869
Yeah, it's a good point. So it gets to the question of what risk are you taking when you buy equity versus when you buy ILS? So when you're, when you buy insurance company stock, the fact is that you think you're buying something that is insurance, so it should be diversifying with the fact that stock prices move up and down with the market, they usually have a beta close to one. So insurance company stocks are not the diversifiers that you know, that you would think they are. But if you buy a cap on that would have a beta of zero compared to the border asset classes. Do you know, the interest rates or stock markets?
00:48:22.340 --> 00:48:26.869
And so are we doing?
00:48:26.900 --> 00:48:31.699
Are we sure about that? So like, here's, let me let me kind of pick an idea.
00:48:28.550 --> 00:48:43.849
Because like, you know, I wonder how much of that is just sample error? Because like, if you look at the the financial crisis, regular liquidity crunch and cat bonds, right, you might say, well, that's not because of the underlying risk, but I would say, Well, hey, too bad, you know, the prices still dropped.
00:48:40.730 --> 00:48:56.329
You know, and because, you know, showing it there is asset market correlation, because if I think about insurance company, you know, the proper correlation to me, would be between insurance company results and cat bond results as opposed to earnings, right, as opposed to asset prices, because, you know, I'm getting that right, right.
00:48:56.400 --> 00:49:04.920
Yeah. Look, I agree.
00:48:56.400 --> 00:49:14.099
So, okay. Okay, let me So, in fast moving markets, in the end, all prices are correlated, right. That's the point you're making. I agree with that. Okay.
00:49:14.849 --> 00:50:05.670
However, that doesn't change the fact that, you know, so, you know, we talked about correlation as if it's sort of like one number that applies all the time. It's correlation in different markets, right. It's, it's more like a copula. It's like correlation in fast moving tail risk scenarios, as opposed to correlation in normal day to day markets. Man, so the thing is that the, the price of these insurance linked securities, the underlying risk, which, which obviously affects the performance as well, you know, there are two things that affect the performance, the underlying losses, and then the price of that risk. So the price of that risk can be correlated But the underlying losses are not.
00:50:00.179 --> 00:50:21.239
Right. Right. And so in normal markets, it provides significant diversification benefits. But in fast moving, sort of tail risk scenarios, ultimately all prices will will, we'll move together.
00:50:21.269 --> 00:50:38.429
So but if you if you wind up, your vision is realized, right, and you're saying, now we have a index of the market at a certain layer of performance through where you're most capital efficient, why would your results not be the same as an insurance company?
00:50:33.059 --> 00:50:39.480
earnings? And so not command any difference in capital costs?
00:50:40.050 --> 00:50:52.800
Yeah. Oh, yeah. So I think this is why I think that insurance. One of the reasons why insurance company cost of capital is high is like I said, it's there, it's very opaque.
00:50:53.610 --> 00:51:41.820
You know, I, I was a chief risk officer for the gold, you know, for a couple of different companies. And, and it was hard enough for me to understand all the risk and, and then be able to explain to my management team and then to explain to the board, the investors don't really, with all due respect to investors don't really have a clue. It's because they don't have information. And they don't you know, that, you know, they trust middlemen. And I think just that for you know, that, if anything happens, you're even if you keep everything inside of insurance company, if you can give information on a real time basis, and use a methodology that is more transparent and reliable and more objective, that will decrease the cost of capital, even inside of insurance company as well. And so, I'm not suggesting that the construct of an insurance company is itself inefficient.
00:51:38.940 --> 00:51:44.250
I'm saying the way insurance companies are run now is inefficient.
00:51:44.309 --> 00:52:07.230
That's super interesting, because there's another thing that I wanted to make sure we touched on, which is to say, was observation that I make. And, you know, one of the things I like to say is when when any company hatches initiative, particularly one that says let's go to the capital markets, because it's cheaper capital, right? And and then I say, but here's one thing we need to keep in mind, what are we going to do when the traditional market beats the capital markets on price?
00:52:04.380 --> 00:52:25.530
Because that's something you see all the time, right, they'll say, we'll just be cheaper. And to me, like, why we should take that seriously where they you know, you're telling me a story now for why the traditional markets could do so because they're actually their cost of capital for such a risk is the same as the capital markets, because they're the capital is ultimately the same. And if they have transparency into the pricing, they should be just
00:52:25.679 --> 00:52:30.989
one part of it. But I'm not done with that. And so that's one part of the theory.
00:52:27.780 --> 00:52:30.989
Oh, okay. So that's one reason.
00:52:31.199 --> 00:53:15.210
However, there is inefficiency in the value chain, because we all know that there's and we I think we all acknowledge now, because I hear everybody kind of been saying it for the last five years. There's too many hops. In the insurance industry, there's too much friction to get to its capital. And so we have a lot of overhead. So what happens is, let's say that you have a big insurance company that uses an mga the mgh is doing the underwriting and they're doing the claims, but the insurance company has their chief underwriting officer, they have a chief risk officer. And then they reinsure some of it in the reinsurers as the chief underwriting officer and chief risk officer, and everybody is remodeling that same risk in order to manage their capital.
00:53:10.860 --> 00:53:30.869
And but if everybody had the same information chain, and everybody was plugged into the same information chain, and the modeling was done once, then you can go, you can have many different layers of capital wouldn't create that overhead.
00:53:28.019 --> 00:53:51.059
Right. And so that's another problem with the with the current construct of the industry is that it's, there's just way too much overhead, and again, to the benefit of all the middlemen, but not necessarily to the benefit of the insured or the capital provider. And, and, and that's a lot, actually, I mean, you know,
00:53:51.960 --> 00:53:56.730
but let me, let me let me try and tell a sympathetic story on that one.
00:53:53.460 --> 00:55:01.920
So, you know, much earlier on, you mentioned, at the very beginning of our conversation, we talked about how there's this, like, I forget the word to use ecosystem of promises, right, that supports the initial promise of an insurance company, and I'm with you on that. Now, to me, the promises are trust, but verify, right. And so, you know, you're talking about duplication of effort, I'm saying, You sure. But that's how, you know, right? You measure the same thing a whole bunch of different times. And so each of these, you know, I think the ecosystem of the insurance industry, the capital structure, right, so, you know, I had a conversation with Terry Vaughn, you know, ages ago on this podcast, we put a link to that, where, you know, I made the observation that there's regulators, there's rating agencies, there's reinsurers right? And there's even like agencies and stuff, who are always assessing how people are doing and how are you doing your job properly? And they're all kind of doing the same thing, which is to say, Are you properly pricing your risk and and you know, one of them can fail and often one of them does fail, right or even two, but you have so many of them that somebody is going to blow the whistle and then you know, or even investors are looking at this stuff, right. So, you know, duplication of effort, sure, but diversification of oversight is
00:55:01.750 --> 00:55:31.960
No, I don't think there is that. So here's why I think that because I think it would be right, if everybody was using the same information and had access to the same information and redoing the same things. But they're not. You know, the fact is that, you know, I recall that, you know, when I was in AIG, I used to, there were times when I would go out with a field engineer, who would go to a building that we were underwriting, and we would go to the roof, and we would walk around the roof, and we would check all these things.
00:55:32.320 --> 00:55:44.199
There's a lot of information that was being gathered there to underwrite that building, when it was reinsured. In an XOL program or a cat program, they didn't have that information.
00:55:44.860 --> 00:56:35.230
There's no way they had that information, they had aggregated level of information. Okay. And so what's happening is that at each layer there, they're getting aggregated information, and they're doing more portfolio analysis, it would be better, wouldn't it be better if everybody got the same information, and you still have the same number of people, so you have rating agencies get the same information, the regulators get the same information, and the capital providers get the same information, the insurance, get the same information, and you in and you have one set of analytics, and everybody knows how, what the strengths and weakness of analytics that would be consistent with what you're describing, having many eyes, re reviewing, sort of the sort of the wisdom of the crowd sort of approach to, to looking at the same information and coming to some conclusion, that would be valuable. But right now, you don't have that?
00:56:35.469 --> 00:56:37.840
Well, I would say, so I'm with you on that comment.
00:56:37.900 --> 00:56:48.489
But I would broaden the concept information here. Because reinsurance underwriters, they have a job, many of them are very good at their job. But they don't necessarily even if you gave them all information, they don't know what to do with it.
00:56:48.730 --> 00:57:08.289
Right? So what they do instead is they evaluate the people. And same with regulators, or rating agencies, they sit down with your presentation of the of the CEO or whoever it is, and they you know, they'll pepper you with some questions and see how you do on those questions. That is a different source of information. It's an aggregated obscure obfuscated information about the portfolio, but it is a distinct source of information, right? Do I trust the degree?
00:57:08.519 --> 00:57:49.860
Yes, they may come in, they may be coming at a different angle. But are they also evaluating their compensation plan and their bonus pool and to what extent they get promoted? If they say no to a risk? If it doesn't meet those requirements? They're not doing any of that. Right. So when you're evaluating an underwriter, you're not involved evaluating the entire system under which they are being underwritten? No, and and those are very different from company to company and from underwriter to underwriter. And you first of all, that will take a lot of time. And there's a lot of subjectivity in this analysis.
00:57:46.739 --> 00:57:58.230
If instead again, if you had something that was objective, if every if you had the same information that the underwriter had, you don't have to evaluate the underwriter you can evaluate. So
00:57:58.500 --> 00:58:02.010
their decision Yeah, if you can, yeah, you can Monday morning quarterback?
00:58:02.699 --> 00:58:12.179
Well, it's on Monday. Yeah, so yeah, you, you you're, it's got to be better.
00:58:05.909 --> 00:58:52.679
If you have all the same information that the underwriter has. And you can invest, you can not only analyze information, but then you can also analyze the underwriters reaction to that information. So imagine, like, you know, in our deals, we have an mga that's connected up to a capital provider is the, you know, so that they, you know, the mga has underwriters, right? And so, now the capital provider is able to see that information is able to see how the underwriter reacted to that information, or the claims person reacted to information and what their net result is, that is much more effective.
00:58:44.820 --> 00:58:57.719
Sort of overview, then, simply saying, I don't know what it is, but let me look at your CV, and you know, you have a great reputation. Well,
00:58:57.719 --> 00:59:06.840
I mean, so another way of putting that so again, you know, I'm inspired by the breadth of the ambition, depth of the ambition here, right?
00:59:03.719 --> 00:59:15.929
We'll be able to like atomize, an insurance company's portfolio, you know, analyze the decision making process and a granular fashion, and then we aggregate it holy cow, but let me let me kind of tell you, some
00:59:15.000 --> 00:59:20.789
MGAs are doing that now. Right? I mean, they're getting information on a real time basis. Yes,
00:59:20.449 --> 01:00:03.980
yes. But I'm talking about evaluating a company reinsurance company. But anyway, yeah, I might disagree with you. The thing that you gain to me with this is, are the things the way around that if you're a regulator or rating agency or reinsurer is look at it. You know, you mentioned CV, another way of saying look at the track record, which is why it takes so long, right sometimes, where you know, it's why you have to trot out somebody who's done it before in order to get that a rating out of the gates if you want to start up a new company, because now you're kind of pinning your track record to this. And and if you can sustain a track record over time, your point there about analyzing compensation structures and like the innumerable features and variables that go into successful insurance. You That all is too hard to do for human.
01:00:00.380 --> 01:00:10.010
So instead we just say, Okay, well, I'll just pull up my arms, and I'm going to wait five years and see how you do. And then you can come back. And then I'm like, give it what you want.
01:00:08.179 --> 01:00:10.010
Right? Yeah. Which is?
01:00:10.280 --> 01:00:31.400
Okay, so let's look at track record. Now, would you rather have you're going to invest in a portfolio? Would you like to see the track record by actually seeing the losses in the premium of that portfolio for the last 10 years? Would you like to see the the the profitability of the company that was insuring that portfolio?
01:00:32.760 --> 01:00:36.510
Well, of course, you'd like to look at the risk, right? Yeah, if you can find a way to aggregate the risks,
01:00:36.539 --> 01:00:50.519
you are getting a track record here. Yeah. But you're getting the track record of the thing that you're investing against, rather than a proxy. So all these discussions about, you know, evaluating the underwriter and all that, which is the best we can do now.
01:00:47.940 --> 01:02:07.710
Right? It's been done, because that's the best we can do not because that is the best way to do it. In the reason that's the best we can do is because they don't have access to information for information if I was a reinsurance. Look, I don't want to talk to the underwriter show me the loss ratios, and show me the detail loss run, show me the premium and show to me over the last 10 years. That tells me something. And show that to me, not only for your company, but for every other company. Now, that tells me not only what you're doing but how you're doing compared to in a competitive market of other you know, other companies. All that is available now. It's available. It's it's, it's not being? Certainly it's not looking even in the sample example you're providing. Who's doing that evaluation? Is it the person who's providing the capital? No. It's the the capital, the person who's providing capital is trusting some management team. And that management team has a hiring underwriter who's trusting another management team and their underwriter. And they may be trusting an mga. Okay. So it can't be more efficient, then the capital providing getting the information that the the, that's at the front of the line?
01:02:07.739 --> 01:03:04.050
Well, certainly not. And in fact, you know, so here, we're kind of re deriving a whole bunch of cultural features of insurance, right, where we're saying that, you know, if the original sin is is information, opacity, then the only the only way to evaluate strategy is to wait and see how the strategy goes, right. And, and, and now, like, if you're an underwriter that has a good track record, that's a precious thing. Because it took it's very expensive thing to build in terms of time. And so now you're not really willing to change too much about it, because or take risks, right? career risks, strategic risks, because if something fails, now that's pinned to your forehead, and you've got to live with that, that for a long time. And so it creates a sign of wood in this kind of like conception to what we think, but create a kind of an inertia that fights against change or doing things differently, because, you know, this too expensive to recreate a track record with something new if it just goes wrong, even by chance?
01:03:05.250 --> 01:03:15.300
Yeah, well, I agree that there's a lot of cultural inertia in the industry. So, but I'm not thinking that that's necessarily a good thing.
01:03:16.469 --> 01:03:19.889
Yeah, I hope I wasn't coming out in support of it there. I was just saying that it existed.
01:03:21.989 --> 01:03:45.570
And so inertia exists. And so and that will dictate the pace at which new things are absorbed. But we also know that another dynamic of the insurance industry is, quite frankly, insurance industry has not been known to be leaders in innovation. But they are known to be good followers.
01:03:45.900 --> 01:03:47.789
Yeah, conservative, right.
01:03:47.820 --> 01:04:21.420
Yeah, they're conservative, right. And so the, you know, the ones who will lead this innovation are the ones who are going to make all the money or make a lot of money, and then and then the rest of the industry will follow. And that's fine. When you're trying to drive change, whatever system it is, you don't try to achieve change by going up the entire system, you only need 510 percent of the system to embrace the change. And that becomes a catalyst for everything else.
01:04:21.570 --> 01:04:57.300
Well, there, you know, I should say that there is one piece of the insurance the conservatism which I, I like or I'm sympathetic to, which is there is a certain amount of intellectual humility, based it baked into this, right? where, you know, it's, you know, risk is a risk is infinite and all kinds of crazy things can happen. And so, you know, that's a great idea. One way I like to put it is, innovation is indistinguishable from underpricing in the short term, right. It's like you got a great idea. Tremendous, but you might be blowing up tomorrow and I just can't tell because I don't I don't know enough about what's really going on. And so sounds plausible. Let's wait and see.
01:04:53.639 --> 01:05:02.400
Right and and so like, I think that that's a cultural feature.
01:04:57.300 --> 01:05:08.219
That is It is a as a residual of like six settler survivorship bias in there. I don't think that people want to be that way.
01:05:05.130 --> 01:05:17.130
I just think those have survived are that way. And a lot of the others, you know, not everyone, but many people who have great ideas are underpricing for reasons they didn't understand and explode. You know? Yeah.
01:05:17.480 --> 01:05:38.389
Yeah. It's. So again, you know, what's related to what you just described is, I used to wonder, and we used to have debate with lots of colleagues about how much you know, when you're rating an underwriter? How much of that is luck? And how much of that is skill?
01:05:38.780 --> 01:05:39.110
Yep.
01:05:39.619 --> 01:06:52.610
Right, given the inherent randomness of the underlying process? And, you know, the thing is, there are statistical methods to try and assess that not like in a yes, no fashion. But in degree, using sort of, you know, Bayesian methods, you can as new at you have a priori estimates, and you have new evidence comes in, and each time evidence comes in it, you know, sort of, it's like credibility theory, there is a way to kind of, to, to gauge what is luck, you know, at least get some sense of, statistically, to what extent something is random versus something is meaningful that that's really the root of all the, you know, the statistical inference methods is to is to separate signal from noise in these things, I think, you know, what we're seeing right now, at least with the investors that we're talking to, they value the fact that this is direct connection into risk. And then they value that they're getting information that's much more transparent and objective than other things that they've done.
01:06:53.510 --> 01:07:30.019
And also, in the current environment, because there is so much overhead in the industry, it's hard, it's, it's, it's possible to get rich returns that are higher than what they should be for comparable D rated risk, which is out of the cap on market was, you know, I you know, you know, back in like, you know, 2012 1314 15 during that time, yeah, I was fortunate to be in the market at that time, the spreads were much higher than what they should be.
01:07:25.670 --> 01:07:49.250
Eventually it converged. And right now, we're in a similar environment, that the spreads are higher than what they should be. And there's opportunities for investors to take advantage of that. And I don't expect that all they will all do it or, but I think only 510 percent need to try it. And then we'll see.
01:07:50.010 --> 01:08:04.530
Well, I have to say, I have not covered everything. I wanted to cover it, but we are at time. So Samir, anything else about ledger? How any asks of the audience, what can we do to help you be looking I don't know,
01:08:04.530 --> 01:08:35.159
I hadn't really thought about a specific asked, I think that I'm hoping that people will appreciate and believe that it is doable, this idea that you can get information that is much more transparent, much more reliable, and you can tighten up the value chain. This is not. You know, this is not speculative in that sense that all of these things are doable, and people are willing to sort of at least look into it.
01:08:35.569 --> 01:08:38.810
Great. My guest today is Samir Shah. Samir, thank you very much for your time,
01:08:39.109 --> 01:08:39.829
David. My pleasure.