A capital idea: ERC executive proposes a new method for pricing reinsurance.

Donald Mango, director of research and development at GE's Employers Reinsurance Corp., has developed a new way to deploy reinsurance capital and to price reinsurance. In his prize-winning paper, Capital Consumption: An Alternative Methodology for Pricing Reinsurance, Mango says he is challenging "many fundamental conceptual underpinnings of reinsurance pricing."

Although his paper has won three awards, including the 2004 Charles A. Hachemeister Prize from the Casualty Actuarial Society, his concept is not part of widespread industry practice.

For decades, reinsurers have been taking a risk-based, capital allocation approach in pricing their product. But Mango's method eliminates the need for capital "allocation and release. It evaluates the actual contract cash flows in various scenarios, evaluates the operating deficits in each scenario as contingent capital calls on the capital pool, and reflects the expected cost of these calls as an expense load.

"Allocation treats segments as if stand-alone, with less capital," the paper explains. "This means being in a portfolio is like being on your own, but you have to support less capital. Consumption, on the other hand, treats being an a portfolio like being stand-alone, with access to potentially all the capital, but with the added wrinkle that all the other segments have similar access fights."

Those who advocate the allocation method wonder how else reinsurers could make performance evaluations or investment decisions, as well as decide where to grow or shrink their books, Mango said. But his approach is a valid alternative, he stressed. "Both approaches are based upon the premise that riskier segments must pay for their risk. Both approaches are also dependent upon a sound portfolio risk model, the true foundation of stochastic reinsurance pricing," the paper says.

Best's Review spoke recently with Mango about his new method and what its use could mean to reinsurers.

Q. What is the basic premise in Capital Consumption?

That you need to treat the capital in an insurance company less like capital that's being invested in a project--which is the current approach--but more like how we manage shared assets around the world, a shared asset being, for example, a country club, golf course, a national park, a large hotel, a reservoir. When seen in this context, many things become clearer.

When you look at it that way, you see that it's not about investing capital in these products, which then gets returned, Instead, you're establishing a franchise and you're letting people use it. And the intended use of that franchise, meaning the capital, is not to damage it.

It's similar to staying at a lovely hotel on the beach--they're not expecting that you're going to destroy the room or many rooms or the hotel entirely. The expectation is you'll use the room, then you'll leave, and you'll pay them for that use. The payment will be a function of how many rooms you've used and bow long you've occupied them. These are all concepts that are very understandable in most other businesses, but have been difficult to put into practice in insurance because we've been trying to apply that project-based allocation approach. If we think of it more as a shared asset, then it becomes very clear.

Q. What are the key points of this new pricing approach?

Number one, you don't have to allocate capital to price. That's probably been the thing that has created the most controversy, at least in actuarial circles or risk analysis circles. This is because capital allocation was seen as a mandatory requirement in order to price, and so all the debate was about how do you do it, and how do you calculate it.

Number two, it really represents a pricing methodology that utilizes all the information that's produced as a part of what I would call best practice reinsurance pricing techniques used today, and previous approaches that weren't taken full advantage of.

Third, the title does represent another dimension of the impact on a company's capital of writing business. The previous approaches were all focused on what I would call allocating required capital, rating agency capital or something like that. But the second way a company's capital can be impacted by writing business is when the results deteriorate and the capital gets used up and moved from the capital account to the reserve account--this is known as reserve strengthening. It happens all the time, everybody knows about it, but all the pricing approaches had ignored it. I see it as a second dimension to the picture, so I really look at this as trying to fill out the whole picture.

Q. Why is your suggested method considered so revolutionary?

Because of that second dimension. It's a paradigm expansion--I wouldn't say it's a shift. It's like moving from a single dimensional picture to two dimensional.

Q. How did you arrive at this method?

It took years of struggle. Through a lot of reinsurance pricing work, a lot of reinsurance pricing model development, I saw how reinsurance was priced at several leading reinsurers. I saw what they were doing and I just had a nagging sense of dissatisfaction. Everyone was using variations on the same thing.

The signposts are all found, especially in Question 3 of the paper," What does being in a portfolio mean?" That to me was the biggest problem--that this was just not addressed sufficiently in the current approaches that I'd seen. What seemed to be happening in the allocation approach was you'd take the total capital, divide it up and give every piece of the portfolio a little bit. So being in the portfolio treated segments as if they were stand-alone, with less capital. That's OK, there's nothing wrong with that, but here's the problem: If the results go bad enough, every portfolio in the company, every line of business could draw all the company's capital down. That aspect is never reflected anywhere.

When I do the presentation on this, the best analogy I've found is that it's like the difference between having a small swimming pool in your back yard and joining a swim club. You and I and 50 other people are part of a swim club, so we've all got access to this big lovely pool. The problem is if we all go at once, it's very crowded and nobody likes it. Or you call have your own little pool and it's all yours, but it is not the same.

Q. And the probability of all of us going to the pool at the same time is low?

Exactly. The view changes from taking the total, cutting it up, and giving everybody a piece. My approach says, 'No, no, the whole stays intact and you all have simultaneous overlapping rights.' And the reason it works for insurance is, we're all holding onto it, we're not trying to pull it away. But when the results go bad, that's when all of a sudden a certain product will drain it.

Q. What's a layman friendly example of how your pricing method--consumption--would be applied in reinsurance?

The easiest one to understand is a property catastrophe contract--a high-layer excess contract--meaning it very rarely will be hit. The insured pays a small premium to get a big limit that they probably will never use, or hopefully never want to use. Under this approach, the first thing you do is preserve the outcome scenario, make it simple by saying it's a high-layer deal, so 98% of the time it's going to have no loss, 2% of the time it's going to have a loss, and when it has a loss, it's going to have a full-limit loss. It's very typical of these high-layer deals. It's either going to get blown away, completely obliterated, or it won't get hit.

The capital allocation approach to pricing this would be to calculate the required capital, by any number of methods, and this will be an amount that is probably a fraction of that limit, something like 10% or 20% of that limit. Then we calculate the expected return, the expected net present value, and divide it by that capital amount. That's the return on equity or ROE. And that only uses one scenario, an expected scenario. It collapses all the scenario information, meaning ff there are many scenarios, the approach forces them down into one scenario like your best case or expected case. Because that's all that the capital allocation framework can handle, a single scenario showing you invest this much and you get it back over time plus some return, and that's the ROE calculation.

Under the consumption approach, you actually see what each outcome would be. If there's no loss, I keep the premium and everybody's happy. Under the loss scenario, it" I collected $1 million and I have to pay out $10 million, then I'm short $9 million. And that $9 million is coming out of capital. So I treat it as a call, just like you're calling your parents for money from college. It's almost like a line of credit. We in our pricing approach would have to come up with a charge for those calls. You calculate the expected capital call costs, and you deduct them like an expense.

So instead of taking a capital number and dividing the return by the capital--that's how the allocation approach works--you're ending up with a risk-adjusted net present value. This is a lot closer to what would be called a value-added framework, meaning it tells me what the economic value of this deal is once I've factored in all the costs. This capital call cost is like a hidden cost, if you will. It's not an expense item that shows up in the annual statement, but it's a way of penalizing or expense loading contracts. This is another way to evaluate reinsurance contracts and maybe it captures things that the previous approach did not capture, and maybe it's valuable to at least try it.

Q. What are the benefits to the reinsurer in applying your method?

It adds another dimension to their decision making and maybe this would lead to better decision making. My contention overall is that there's no single decision framework that works the best every time. I've met a lot of resistance to this where people weren't even willing to entertain the possibility that this is a valid alternative. So I ended up having to soften my presentation to ask them if they would at least consider it as a valid alternative. I tell them they can run it as a parallel test.

American Re has this in place. The paper I wrote was a result of American Re doing this. The biggest benefit is that American Re has become very explicit about risk preferences. All this function is saying is be explicit about your risk preferences.

Q. You mentioned that some companies have been very resistant to this. Why such strong resistance?

With scientific expansions in general, this is the reaction. And that's fine. My father is a chemist and I got to see first hand that this is the way science is. When there's dissatisfaction with an approach, people try things at the fringes of that approach until somebody makes a leap forward. And then there's resistance to the idea. But actually some of the strongest opponents early on to this pricing method are slowly accepting it as a parallel approach. It's a big leap forward.

Capital Consumption: An Alternative Methodology for Pricing Reinsurance is available online at www.ercgroup.com. Click on GE ERC Company Spotlight.

Key Points

* Donald Mango, director of research and development at GE's Employers Reinsurance Corp., has written a prize-winning paper that challenges fundamental underpinnings of reinsurance pricing.

* Mango's most controversial idea is to treat an insurance company's capital like a managed shared asset. Currently, capital allocation is seen as a mandatory requirement in order to price.

* His method uses information produced from best practice reinsurance pricing techniques as well as previous approaches that he feels were not fully explored.