An Interview with Nicos Scordis: A Leading Academic's Insights on Insurance Asset Management

Stewart: Welcome to another edition of the InsuranceAUM.com podcast. We have got a good one for you today. We are joined by a very well-regarded academic who has written a great deal on insurance asset management. Today, our guest is Nicos Scordis, Professor and the John R. Cox Ace Limited Chair of Risk and Insurance in the St. John's University Greenberg School of Risk, Insurance, and Actuarial Science. That is a phenomenal title, and you're also a friend and we're thrilled to have you on, Nicos. Welcome.

Nicos: Stewart, it's a pleasure. It's a pleasure to be here because you are doing a wonderful job for the industry and I enjoy getting out of the university and talking to real people like you and your audience. So thank you very much.

Stewart: Well listen, thanks for the kind words, we certainly appreciate it and we're thrilled to have you on, you have no idea. You've got real-deal academic chops, St. John's University is one of the leading risk management schools in the world, and you've written extensively on insurance asset management and we're thrilled to have you.

Nicos: I have, I have.

Stewart: But before we go too far down that path, I want to run you through the same gauntlet as all of our other guests. What's your hometown? Where'd you grow up? Your first job and a fun fact.

Nicos: So hometown, I grew up on the island of Cyprus, as you can tell from the accent, never leaves you. But right now, I live in Princeton, New Jersey and I have a small place in New York and running the numbers being in finance, I realized that I'm questioning the wisdom of doing that because the rent translates to a lot of hotel nights in New York and you don't have to make the bed either. Yeah.

Stewart: That's right.

Nicos: My first job was working at a restaurant. I was trained as a chef. So actually, I had my union card next to my PhD until I got the first publication. And then I said, "Okay, I can do this academic stuff, so I can let the union card go," hedging my bets.

Stewart: Absolutely. And that could qualify for your fun fact. How many people in your position are trained chefs? I mean, that's really cool, but lots to talk about with you. And I want to start with something that's really fundamental and that is how do insurance companies create value?

Nicos: Yeah. Well that's a great question because I used to be on that side of the fence and over the years, I'm on the other side of the fence. So first, I was thinking that insurance companies are like levered investment trusts. So they generate float, they issue the premium at a risk-free rate, then they create that float, they invest the float above the risk-free rate, boom, they make their money, their value just like a levered investment trust.

Stewart: And so let me just step in here just a second.

Nicos: Yes.

Stewart: Just to make sure that everybody knows what you're saying, so when you use the term, "float," that is the time between when the premium dollar hits the bank account of the insurance company and the time that that dollar goes out the door to pay a claim?

Nicos: Yes.

Stewart: Okay.

Nicos: Yes, yes.

Stewart: Okay. Good deal.

Nicos: Yes. And you mentioned my research and for the past 20 years or so, 25 years or so, I've been looking at how much capital do insurance companies have? How much risk do they take with that capital investments and operations? How do they allocate their assets? What is their optimum product portfolio? How much value do they generate for the risk they're taking? And starting looking at the data, it looked like it was difficult for the average insurer to generate economic value from float. I'm not saying that some insurance companies can not do that, but in the middle of the road, the bulk, the average insurer. So research suggests that what really drives value is the underwriting of the insurance company, good solid underwriting. That's where insurance companies have a competitive advantage, identifying underwriting package, reselling risk as opposed to investing. So I shifted in my view over the years to looking at investments as a companion piece rather than the value drive.

Stewart: That's interesting.

Nicos: Right.

Stewart: That is really an interesting position. I've got to get my geek hat on again. So the way that I've thought about this is that the combined ratio is effectively the cost of funds, right? And that insurance companies, and I worked for a company that was owned by Buffet, and Buffet used this term, "float," for a long time, right?

Nicos: Yes, yes.

Stewart: Companies are trying to get low cost of funds, right? So if I write at 100 combined, that means I'm making a zero underwriting profit, which the way that I have always thought about it is that my cost of funds is now zero. And so we're going to talk about leverage in just a minute, but really a bank can't get to a zero cost of funds. And if I'm writing at a 99% or a 98%, now my cost of funds is -1% or -2%. But the way that I understand this is you view the float as the risk-free rate, which is not zero. So can you unpack that a little bit and set me straight?

Nicos: Yeah, we'll unpack that, right? So in fact, there is a stream of research that makes exactly the argument you articulated, Stewart, the combined ratio is the cost of capital. And that's what we thought for the longest time and that's what I was advocating for many, many years. And then I run into a thin stream of research that's viewing the insurance contract as an option, right? And if you view the insurance as an option, things change because it's not priced as an option, but your audience are familiar with the insurance contract, right? You get nothing before the deductible. Then once you have a claim above the deductible, there is one-for-one reimbursement. And then once you hit the limit, it doesn't matter what your claim is, then it levels out. So if you think of that pattern of payment, that is the same as a bull spread, option bull spread. So you buy a call at the exercise at the deductible and you sell a call at the same time, exercise at the leave. So you could replicate that.

So if what you're selling as an insurance company is an option, what drives the value of the option is the volatility of the claim as opposed to the expected value. And this stream of research then takes this idea of the options. And a key author is Doherty, who was at the University of Pennsylvania retired, and Slazenger at Alabama who died untimely, not that there is such a thing as a timely death. So then what drives then the combined ratio in the option framework is whether the claims covariances are positive or negative in relation to the capital that the insurance company has. So as a policyholder, if buying insurance provides a hedge to my non-diversifiable risk, to my background risk, I'll be willing to pay a higher premium than what we think is actually fair. So the insurance company can generate one minus a combined ratio is positive, right? They make an underwriting profit.

Stewart: Underwriting profit, right.

Nicos: Exactly.

Stewart: Yes. Yeah.

Nicos: Right. And if however I, the policyholder, don't see the value of using insurance to hedge, it doesn't matter how low the premium is, I am not going to be interested in that insurance. So in that framework combined ratio, it just tells you how much money came in. But it's not really the cost of capital because what drives the cost of capital is how much liquidity the company has to support its claim that is going to honor those options. We don't have an exchange that does the margining call, right? So you need that surplus, you need the liquidity as a tangible indication. You bind the promise that you as an insurance company is going to pay the claim.

Stewart: That's interesting.

Nicos: Right? So it's a little bit different perspective, right?

Stewart: That's really interesting. I knew this is a good idea to have you on. All right, so you ready for the next one?

Nicos: Yeah.

Stewart: Investment policies, the statement is that they cannot be independent of the insurer's underwriting policy.

Nicos: Yes, yes.

Stewart: Can you talk a little bit about that?

Nicos: Yes. Well I don't claim to have a huge practical experience with how insurance companies actually run, right? But when I talk to friends and industry practitioners, is there a communication between those people who do the investments and those people who do the underwriting? Some people say, "Oh, yes, great communications, we couldn't ask for more." But some people say, "I wish those people actually talk to us," and goes to the other side too, "I wish those investment people actually let us know what they're doing." So my view is that no, investment policy cannot be independent of the underwriting policy. And here's what I'm thinking. At the end of the period, let's say the claim is $100, right? So if we take a risk-free rate of 5% and we discount that, then what the premium should be today would be $97.20, right?

Because otherwise, the policyholder would say, "Wait a minute, if my claims, if the cost of the insurance company is going to be $100 at the end of the period, why would I give my money for free to the insurance company? I need some discount." But then what is reserves, right? Reserves, according to the standard textbook definition, is the present value of expected insurance liabilities. So there is a lot of similarity between what the premium is at the beginning of the year and what the discounted reserve should be at the beginning of the year. So if the discount of the reserve is greater than the investment rate of the reserves, so you discount the $100 claim at 8%, but you invest that money, you, the insurance company, at 5%. In essence, you have underreserved because what money you invested would grow. You discount at 8%, so it's about $92.5. Then you take that $92 and you invest it at 5%, you being the insurance company, so it grows to about $97, not $100 which is the average claim. So you have underreserved. Now, if you're the insurance company and you underreserve, in essence, you take value from the shareholders, surplus and so on or dividends you don't pay, and you transfer them to the policyholders. And your investors are not going to like that. Now, if we take the other approach, discount rate of reserves is lower than what you invest in the reserves, then you'll have reserve. At the end of the year, you have more dollars in your reserve account than the $100 you need to pay out. So now, you are having value that loads away from the policyholders to the shareholders, right?

So you're taking the money of the policyholders and at the end of the year it's like, "Oh, excess reserves, let's put them into surplus." So the regulators are not going to like that. Or if you're a non-commercial insurer and your commercial customers are going to catch up on that pretty fast, right? And it's a competitive market. So to keep all key stakeholders happy, it seems not the discount rate must reflect somehow the systematic risk of the claims. And there is some research, I haven't seen anything recent about that, but the earliest was one done by Darcy. Darcy was a professor at University of Illinois, and he was also a fellow in the actuarial society. Smart guy. He retired now, he did his research in the late 90’s, early 2000s in the general risk and insurance that he does show that discount rate reflects in the systematic risk of claims. And that's why I don't think that the investment policy can be independent of the underwriting policy. There are other arguments too.

Stewart: Then you view that as an investment constraint.

Nicos: Yes.

Stewart: Right? In your framework, that is an investment constraint that-

Nicos: Absolutely, you're right.

Stewart: The discount rate must reflect the systematic risk of claims.

Nicos: Yes.

Stewart: Because if it doesn't, you've got an imbalance. You're either taking money from the shareholder to the policyholder or from the policyholder to the shareholder.

Nicos: Yes. Yes.

Stewart: Okay.

Nicos: Right? And actually, I like that word that you use, "constraint." See, that's why I like you because you make me look smart. You just gave me the right words there, Stewart. So you see that word, "constraint." So if you have a bunch of money, you being the insurance company, trying to invest, that's a first constraint as opposed to somebody who just has the money and is trying to generate value, they don't have that constraint. And actually talking about constraints, sorry if I jump ahead, but another big constraint is that claims, right, the time of the claims happens at different times. And the size of the claim is different than we could ever anticipate at the time the asset location is made for a particular reserve fund.

And in fact different lines, as all your audience would agree, they have a different payout duration, like homeowners, one-year duration. So 50% of the value of a claim is paid out within one year after the claim is made as compared to medical malpractice, that's 5.7 years. So about 50% of the value of that claim is paid just under six years. And I got this information, interestingly, from a benchmark survey that Guy Carpenter and Marsh published in 2017, and that was the last year they made it public, because now they're offering that benchmark service for a fee. It's a very valuable service to have. So now, we have another constraint to use your words, right?

Stewart: And you're talking the way that the industry, the colloquial term for those differences is short tail and long tail, right?

Nicos: Yes.

Stewart: So you've got short tail, which would be in... I'm going to get myself in trouble quick here, but auto is short tail.

Nicos: Yeah.

Stewart: Worker's comp is long tail.

Nicos: Yes.

Stewart: Right? So in aggregate, we've got information around how the average claim is paid out and we can graph the payout pattern by line, right?

Nicos: Yes. Yes.

Stewart: Okay.

Nicos: Exactly. Exactly. Those long and short tails. And I use duration because dealing with investments, duration comes in if an insurance company is going to use their investments to manage their interest rate risk, the asset liability management area that both regulators and actuaries care about. And its liabilities are difficult to change, right? It's your product mix, whatever an insurance company sells is their product mix. So what then needs to happen in order to match the duration of surplus is how does the insurance company allocate its assets? So really the liability duration is driving the asset duration if there is a desire to insulate asset and liabilities from small changes in interest rates. So that's another constraint, right? So now we have at least three constraints and there are more we can come up with.

Stewart: And so it occurs to me, I've had folks on the podcast say before that they view the duration of the liability as their baseline starting point and that they make a decision whether to be long or short that... I think it depends on the line and depends on a lot of things, but that certainly squares with other conversations I've had. So what about surplus? How does that play into things?

Nicos: Well let me ask you, Stewart, what is surplus? What would you think of surplus? Is it just assets minus liabilities, which are counting we call surplus? Does the surplus need to have some quality of its own to be useful for an insurance company?

Stewart: The way that I've thought about it over time is in terms of economics, less than accounting. In a perfect world or in a theoretical concept, the surplus would be if I sold all my assets today and I paid all my claims today, the amount that I have left is what's an insurance word is surplus, or what we would call capital, right? It's their capital that's left over. And I think that sometimes folks, when they're talking about asset allocation, view, "Well I've got one strategy for my reserves and I've got a different strategy for my surplus." And I think that's somewhat of a common theme in the insurance industry. I got a feeling that you have a different view of this.

Nicos: No, no, no. No, you're raising... I respect your knowledge a lot, Stewart, you know a lot about the investments and rubbing shoulders with investment professionals because I have to admit that there is very little data. There is hardly no data publicly available that we could use to run the numbers and answer those questions. So even the issue of duration, if I go through the Form 10-K of publicly traded insurance companies, there is only one that I have seen that reports both the duration of its assets and the duration of its liabilities. There is a handful that report the duration of their assets and then they have some information and you can infer the duration of liabilities. But there is no idea if this is option adjusted duration of simple duration, right? And some of those reports include their equity and some they don't. And then the issue is how do you actually measure the duration of equity? And there is some research that tries to calculate that. So it's doable.

So what you're saying is spot on. So there is a tendency to treat surplus differently than reserves. But then capital, it's fungible. You have $100, it can either be used to pay claims or it can be used to pay off debt. I cannot imagine an insurance company that's going to say, "Well we need to pay our financial debt, but this money is in the reserve so we're going to default on our financial debt," Or vice versa. "This is in the surplus that's going to pay the financial debt and we're underreserved and we're not going to pay that claims liability because it's in a different account." But what you are saying has been validated by research that there is a tendency to have mental accounts. So this pot of money goes into that account and it's for that purpose, and that's it. Close the happy hole and leave it.

And then we're like squirrels, right? Little nuts here and there is a surplus for me, to go back to your question, right? I think we touched on this earlier, is what makes credible the promise of the insurer to pay claims. So if you're going to have an insurance company, you need surplus before you go out to sell policies. And therefore, that surplus needs to be liquid. It has to be unencumbered by liabilities. And I like this asset minus liabilities, because liabilities could include the debt insurance, does include the debt insurance company has. But instead of taking the resulting number and say that surplus, I challenge to go a step further and say, "What asset from this surplus, once I sell it in a fire sale, it would halt its value because otherwise I have to discount the surplus that the company holds because it's not liquid enough to provide that guarantee of payment of claims."

And in fact, research, one name that comes in, Epermanis and Harrington, forgot when they did this paper, they were able to document that when the perceived quality of the insurance product declines, they required premium goes up. I know it's a long answer, I'll shut it up. If surplus, Stewart, is supposed to be liquid to fulfill its economic function, then there is only a limited number of liquid assets that an insurance company can get. And everybody knows that liquid assets command a lower risk premium than less liquid assets. So here is another investment cost, right?

Stewart: We can't forget about regulations, right?

Nicos: Yes. Yes.

Stewart: So regulators prefer less risky assets. I say that because they have a lower capital charge for things like governments than they do for things like private credit, right?

Nicos: Yes. Yes.

Stewart: And so it appears based on that framework that regulators prefer things that they view as less risky. How do you work that into the equation?

Nicos: Yeah, yeah. Well you're right. We almost forgot the regulators. I had a student several years ago, he was doing his master of science and did a thesis on whether the regulations, the investment regulations of the state of New York for the surplus of an insurance company allow the insurance company to invest on the efficient frontier. And your listeners are well familiar with that, the efficient frontier is the combination of all assets that would allow us to achieve the highest return at a given level of risk, right? So if I have a combination A, by changing the combination, I can go to a combination B. It might be the risk remains the same and by risk the original efficient frontier used risk as volatility, which then all pins a whole other can of worms because standard deviation volatility as a measure of risk is a poor measure of risk. And if you want, we can talk a little bit about why.

So what my student did with his thesis assumed the hypothetical insurance company and programmed all the constraints of the New York State regulations, run the numbers, and what he found, he found that the insurance company under the reserve requirements can achieve the efficient frontier, right? So there is no penalty, it's not a suboptimal investment, but low on the efficient frontier which is consistent with intuition if we look that their regulations favor bonds, fixed income. And talking about bonds, it's very difficult to create economic value, achieve alpha with bond investments. I mean they're well priced, formation around them is plentiful, the markets for them are efficient. So we could actually say that's another investment constraint in how money is invested in insurance companies. You do have that regulation that says you need to hold bonds for the account you deem reserves. I'm not sure if there are any surplus regulations, right?

Stewart: So if I understand that, the work from your student, the insurance company can touch the efficient frontier, lower on the curve. But as you move up the risk curve, then the insurer's opportunity set is going to pull away from the efficient frontier.

Nicos: Yes. Yes. Yes.

Stewart: Okay. So an insurance company, based on the student's research, shows that an insurance company can achieve portfolio efficiency for a lower risk but cannot do so on in a higher risk profile.

Nicos: Yes, because you have that constraint that you need to buy bonds.

Stewart: Right. Yeah, makes sense.

Nicos: Yes. So if you try to maintain that percentage and then open up the risk on the assets that you are allowed to, it doesn't work. But that was several years ago. That was several years ago.

Stewart: That's terrific.

Nicos: Yeah.

Stewart: Okay. So how about this? This is a long podcast and a good one. I'm massively geeking out here.

Nicos: Yeah.

Stewart: So insurance companies, there's always this battle between “Am I investing for yield or investment income? Or am I investing for total rate of return?”

Nicos: Yeah.

Stewart: What work has been done there?

Nicos: Not much really. And the reason is that it's very difficult to calculate investment return or investment yield from the data. The only information we have publicly available are companies like Conning or S&P or AM Best that spend a lot of time, a lot of resources, they have access to companies and they do some average calculations. Now, many of my colleagues use the NAIC database, the National Association of Insurance Commissioners, and it's very granular. It allows to make impressive models, very accurate work. But the problem, the challenge rather, with that data, Stewart, is that it only reflects the U.S. business of an insurance company. And ultimately, one has to answer the question, "Do I believe that the data is subsidiary?" Let's say Farmers Insurance Company is a wholly-owned subsidiary of Zurich Insurance Company. "So do I accept that the data that Farmers reports on their NAIC data is truly an independently created, arrived at data by decisions the managers of Farmers make? Or is it driven by some capital allocation or reinsurance allocation or a directive that comes from Zurich?"

And some data, the answer is yes, some data, it's a stretch to say yes, it's an independent decision. That's why I said the data are not there to answer. Does a global company like CHOP or AIG or Barclay’s, I just picked three companies that came to mind, I'm not getting any conversation to plug any particular company, their NAIC data, how accurately does it reflect their non-US operations, right? So there isn't much to that, there isn't much done. All I can say is that the solvency system that the NAIC has, the IRIS ratios, the solvency surveillance system, one of the ratios they have is investment yield, not total return. And then another ratio they have is the operating ratio, which is combined ratio minus investment yield, investment income divided by net premium error. So there is a regulatory incentive to use investment yield.

And then from what we know from finance, it's easier to predict the investment yield of an asset thing of dividend yield as opposed to the total return, which probably the analysts like that certainty of projection. So I come empty on that one, should they invest for yield or for return? But if you are managing the business, if you're creating value from underwriting, then it makes sense, I’m thinking aloud, it makes sense to use investment yield, right? Because it's the cash flow you generate. But then if you're the insurance company and you're trying to generate economic value to generate alpha from your investments, I don't know, can you generate alpha by just focusing on investment yield? Or you probably need to be looking at total return, right?

Stewart: Interesting.

Nicos: Yeah.

Stewart: Interesting.

Nicos: So I don't really know. I don't really know the-

Stewart: That helps.

Nicos: The argument. Yeah.

Stewart: Yeah.

Nicos: Or I cannot make a stronger argument. I have more questions than answers on that particular yield versus return.

Stewart: So when I look at our framework for today's podcast, there's some information about NAIC data. As we wrap up here, do you want to talk about what conclusions you've got out of there? And then I guess where it ends up is: all things considered, can a typical insurance company generate a fair return and can a typical insurance company achieve durable alpha?

Nicos: Yeah. No, we've been talking for almost an hour. People are busy, we can always do this. I can only different topics, but quickly. All things considered, I think it is likely that insurance companies do generate a return that is appropriate for the risk, they trade risk for return. And there is evidence or that, there are at least three good studies that would suggest that. But when it comes to generating economic value, that alpha, I'm not saying that no investment advisor can do that because it's difficult enough to generate a sufficient return for the risk you take. I'm not downplaying the expertise of investment advisors to just be able to achieve that risk and return fairness, right? But then when it comes to alpha, because of all those investment constraints that we went through, I don't see how it's possible to do the investments that you would give you that alpha. But I don't know, I don't have the data. We're just reading the tea leaves here, right? Yeah.

Stewart: I love it. I've had a great time having you on and I want to have you back.

Nicos: Love to.

Stewart: And so I mean I just get really, really geeked out in the conversation too. Thanks for being on, man, really.

Nicos: Most welcome. Thank you, Stewart, I appreciate it. Thank you. You do a great job for your audience. Thank you.

Stewart: Thanks. Thanks so much, I appreciate it. We've been joined today by Nicos Scordis, Professor and the John R. Cox Ace Limited Chair of Risk and Insurance at the St. John's University Greenberg School of Risk, Insurance and Actuarial Science. Please follow us on all the major podcast platforms. Rate us, please review us, we appreciate that very much. Thanks for listening. My name's Stewart Foley and this is the InsuranceAUM.com podcast.

The InsuranceAUM.com podcast is available on