Helen Yates: Thank you for coming along this morning. Today we’re going to be talking about reinsurance buying strategies. To begin with, I would like to ask what you consider to be the key elements of a successful reinsurance buying strategy.

Alan Fowler: The first key element is to know your portfolio. We try to differentiate ourselves from our competitors by providing the best quality data, analysing it, knowing what our exposures are, whether from a catastrophe or risk point of view, then assessing what our appetite and reinsurance needs are.

It is then a question of translating that into something I can put in front of our executive and propose what we need to go out and buy.

This is translated into instructions to the brokers that are appointed to place the cover and conversing with the reinsurers over quite a long period. People seem to think all this happens in the space of a month. Most of our treaties are renewable at 1 January and our data collection is well into its cycle now because it takes about six months to get everything correct and in the market.

The final piece is making sure everything is complete by the renewal date to afford contract certainty, all the wordings and the signed lines, so that our reinsurers know the detailed outcome. It is a long process. And we increasingly have the brokers and the capital markets coming to speak to us to see if there is anything of a more non-traditional aspect that might be appropriate.

Helen Yates: Thank you. Does anyone have any additional elements they feel should be part of a reinsurance buying strategy?

Andrew Brooks: To pick up on Alan’s point, one of the things we are very keen on is the type of aggregate and the type of data we are getting at the level that the original business has been written at, whether it is on a primary or an excess basis. Since Katrina, I think it has been far easier to get that quality of data. Obviously, that flows through into the whole purchasing strategy of what we think our tolerance levels are for different types of hurricanes or series of quakes there are throughout the year.

Helen Yates: Has contract certainty had any effect on the timing in terms of when you start to sit down and write out these policies?

Alan Fowler: I am not sure it has changed the timing in any significant way. It probably means you start a bit earlier. The issue is the market and its circumstances at placement. If you take 2005 for example, it was difficult because of Katrina, Rita and Wilma and the whole market was late. No reinsurer really wanted to show its hand too early then. Going back some years, we were all probably slightly more relaxed, not about having the cover fully closed, but about wordings and signings. Those now all have to be completed by renewal, and we have been doing that for the last three years. Making sure everything is in place is the newer challenge.

Nigel Ralph: We sell reinsurance and we have seen an enormous acceleration of the process. The data comes through much earlier, and that is not before time. Companies such as ourselves operate in various markets. We are in Bermuda, Stamford in the US and in London. We can see different speeds in terms of the way each market works. I hate to say it but in certain elements, London has been quite slow and I think elements of the market have taken advantage of how slow London is. Perhaps slow is not the right word, but the brokers do delay putting the information into the market, knowing they can get an answer in a very short space of time. I am pleased to see that we are now getting the data in and the contracts signed before the effective date, and that is not before time. It is a big improvement on both sides of the equation.

Eric Paire: There is also an enormous impact on the contract in terms of all the work done on data which is done far sooner than before and we are more involved with our clients. It is effectively preparing something six or seven months in advance. At the same time, the pressure at the other end of contract certainty means that the last bit of the work also has to be done under more pressure because we now have a clear deadline by which everything needs to be signed.

Andrew Brooks: It also depends on whether you are buying in the subscription market or not. Much of the business we purchase in the reinsurance market has 100% placements, so contract certainty has not made a huge difference. For those in the subscription market it has made a big difference. We would question whether that has had a diverse effect on the security that is allocated at the end of the placement because the broker is so keen to sort out the signing we sometimes do not get the feedback we used to get on the security we would require. You get signings given and you find out that sometimes better security was signed down because they responded slower than other securities.

Simon Kilgour: I think we have said consistently that contract certainty is not contract quality. As a lawyer who has spent 15 years dealing with the problems of people not buying enough reinsurance or being in difficulty for arbitraging, the market has been very good at managing its numbers or counting numbers but has paid scant attention to the words in the contract. That is why there are quite often many reinsurance disputes, and most often after large losses. That is the history of the industry as a lawyer sees it. Contract certainty is not going to remove the obligation for whoever is responsible for the programme to check its terms and make sure it fits together.

Certainly over the last five years, we have spectacularly seen that many of the entities in London and Bermuda are becoming much bigger, which obviously affects their buying strategy, but it makes the task of coordinating the buying of the reinsurance more complicated. Some organisations may have a lot of reinsurance programmes. Part of the challenge is making sure that they mesh together. Whether it is the buyer or the reinsurer, everyone has to understand how the programme is supposed to operate. That is not just about having a contract but understanding it.

Helen Yates: I suppose as these programmes become more complicated and companies become ever larger and more global, having this organisation-wide approach to your buying strategy must be quite a challenge. Does enterprise risk management (ERM) help with that?

Eric Paire: It definitely pushes it in the right direction. Many people tend to think that ERM is only about models but actually it is really a holistic view of the risk. It is making sure that you deal with the basic questions of how much capital you need, what it costs and how you will use it. ERM essentially helps you answer those three questions. In that respect, what it does for the global reinsurance programme is that it pushes the companies that use ERM as a business tool to look at their reinsurance buying globally, because the things they are doing with one specific subsidiary will not really help them at the corporate level. In that sense, it pushes it in the right direction.

Andrew Kail: I would also say that if you go back a few years, it was the regulator that was pushing more for ERM to be implemented into insurance companies. Essentially there was a regulatory driver to certain aspects of it. You now find companies saying more frequently that having been through that, they are in a better place. The rating agencies are now more interested in ERM and what they do, and it is helping them make better quality decisions, not only in terms of underwriting. It also allows them to select a more relevant reinsurance buying strategy. They have a better picture of what is happening right across the organisation and probably some data and metrics that allow them to make decisions based on more than just the intuitive feel they had a few years ago.

Alan Fowler: Picking up on what Simon said before, the World Trade Center 2001 incident was the big wakeup call on the wording side. Since then, we have focused far more on getting wordings correct in the run-up to renewal, rather than in a hurry on 31 December.

The bit I missed out earlier on the strategy was the process. From a sign-off point of view we have very clear licensing sign-off so that any cover that we purchase has to be signed off by me and where appropriate by our chief executive and the board. Going back to the ERM point, every purchase we make is looked at from a capital value point of view. Is there a capital benefit or not from a group perspective? Things have changed a fair amount in recent years.

Peter Grant: From a rating agency standpoint that is something we have noticed as ERM has developed over the past couple of years. Five years ago when we had discussions with companies about their reinsurance purchasing plans it was really all about minimising the downside and how much of a capital hit they would be willing to take for an event. Now the discussion is more along the lines of the risk/return trade-off for each incremental pound of reinsurance spend, if you like, and very much more around trying to optimise the portfolio.

Whether reinsurance forms part of that will obviously depend on how cost efficient the reinsurance is at a given point in time. They are also looking at reducing their gross aggregates as an option or looking at capital markets-based solutions or various other vehicles. Whereas the process five years ago may have been that the only option was a traditional reinsurance retrocession type player, I think there now appears to be a more sophisticated dialogue taking place within many entities.

Helen Yates: Has the view of the value of traditional reinsurance changed at all in the industry since Hurricane Katrina? Last year, there was quite a lot of risk retention, and retrocession was very expensive, and then obviously these new capital market vehicles became very popular. Has the perception of reinsurance changed at all?

Gary Wells: I think the market pushed risk retentions up following Katrina. People bought the reinsurance but they did not have to buy it at a higher level. That was just fundamental market forces. The markets are softening now so you will probably see retentions coming down a little. There is still a big role for traditional reinsurance. Obviously, there is quite a lot of interest in cat bonds because they offer an alternative avenue to gain protection; they offer diversification in the investment community and there is also a significant amount of money out there looking to invest. The reinsurers are quite keen to offer cat bonds because there is a lot of appetite for them. They are getting pretty good prices on that relative to a traditional reinsurance contract.

Nigel Ralph: I think that is true, but it is largely true in relation to the US market and the energy market as opposed to the European or UK markets for instance. With Katrina, Rita and Wilma, the main effects were in terms of pricing. There was less impact in markets outside the US.

Gary Wells: You are right. The reason the US has gone forward more on cat bonds is they have indices there to allow transparent pricing. The problem with the UK and Europe is they do not really have any indices that the whole market can look at and say whether an index is valid and given a particular result it would pay up accordingly. There are moves afoot to put a European index in place.

Alan Fowler: I agree, but another reason cat bonds have done well in the US is because the traditional product price went up so much they actually looked competitive. We look at cat bonds every year but we have yet to model one that is competitive against a traditional product. They are getting closer though.

Helen Yates: At the moment I know of one cat bond that is covering UK floods.

Alan Fowler: I am not saying there are not any, but then you look at the time it takes to put one in place, the transactional costs, the indices issue and the question of what is the trigger. If I am buying reinsurance I want an indemnity trigger and not a parametric trigger, unless I am certain that the two run in tandem. I would question whether a company with a UK cat bond would have triggered the cover via the Yorkshire floods, for instance.

Eric Paire: On the point of the parametric versus the indemnity trigger, that is a very key point. As brokers, we spend increasingly more time helping our clients who are considering going the parametric route to assess the basis risk, and not just within a given model framework, but also taking into account the model risk itself and the data. That is clearly a key element. I would agree with your comment on the US. It is not surprising that the market has developed a lot over there because when you have rates on lines at the level they are, it is effectively easier to fit into that.

My last comment is that there should be convergence between the two in terms of price because as we were saying about ERM before, it is going to make the reinsurer, if they have an economic rationality, price their cover in a way that should not be that different from the way the capital markets price their capital. The question we are effectively left with is the indices. We need indices in Europe if we effectively want to develop a bigger capital market there. I would definitely welcome the creation of an index so we can do that.

Helen Yates: Would that be an index similar to the Property Claim Services (PCS)?

Eric Paire: That would be the idea, because really if you are talking about a parametric index anybody can do it; you do not need a market-wide effort because a parametric index is basically about trying to understand the way the losses occur. Anyone can design their own parametric index to fit their own portfolio. From our point of view, that does not need a market-wide effort. A PCS-type index would need a market-wide effort because of the data collection.

Gary Wells: However, it has to be seen to be independent. There needs to be a mechanism so that the market will buy into this index.

Simon Kilgour: Does that create an opportunity for the traditional reinsurance underwriter who is not so slavishly tied to market indices? Is this not the problem with cat bonds that ultimately, unless they can be modelled, cat bond providers will be reluctant to take risk? For instance, as I understand it, terrorism is not something where they are willing to take that risk, but obviously it is a big PML [probable maximum loss] exposure.

Eric Paire: It has a lot to do with understanding the risk and the data. The data you can transmit to the capital market has to be relatively standardised, and you usually do not expect all these investors to assemble all the detail given a specific situation. Terrorism in particular is an area where you can model it only so much, but up to a point it is really up to the underwriting skill of the people who are spending time and energy understanding the exposure.

Andrew Brooks: The quality of data in the US now is far superior to that we get from anywhere else in the world. Also, when you model outside the US, if you look at the PMLs that you get from a model in relation to the 100% aggregates, it still throws up some quirky numbers where you do question whether that is the right number. Europe, for example, covers a large area, and because it has not had the same frequency of windstorms, the models are relatively untested.

Andrew Kail: If you can only use a cat bond in the short term in areas such as Florida windstorm, then with regard to whether there is still a traditional role for reinsurance you would have the rest of the market. Because you would be talking about whether you can extend it into windstorm risk in Europe. That is just another piece of the market. You have not even talked about casualty risk yet or non-European windstorm risk. Therefore, it strikes me that there will still be a marketplace for the traditional side even if the cat bond market grows and becomes exponentially bigger than it is today.

Helen Yates: So do you think non-traditional reinsurance could ever replace traditional reinsurance?

Andrew Kail: Not in its entirety. “Never” is a big word, but it is very difficult to see the capital markets having the appetite to extend their offering.

Jonathan Barnes: There are a few things I would like to chime in on there. First of all, on the terrorism point, I think the modelling for the severity point of view is there. You can get your head around severity risk. I think frequency is the problem. There has been a cat bond transaction that included terrorism and that was the Golden Goal transaction for the football World Cup. That was not excluded from that and the modelling agent specifically looked at and modelled that risk.

Going back to the earlier discussion about the application in the US, I think the reason why these covers have been purchased mostly for US carriers is simply that the exposure in the US is so much greater than elsewhere. As I understand it, the 1% probability loss for US wind is approximately $100bn and for the European wind I think it is probably roughly $25bn or $30bn, depending on which model you use. There is a much greater demand for protection. Obviously, when you are looking at losses of $100bn you want to have collateralised indemnities rather than counterparty credit risk. I think that is the driver for the issuance of bonds in the US particularly. I also think the credit issue is the single most important driver for the broking sector as a whole. Looking at these extreme events, the buyers of protection really do value certainty.

Andrew Brooks: I think Jonathan is absolutely right. After Katrina, we estimated you could probably buy up to $10bn-$40bn in the US on the conventional basis. After that it is a question of your tolerance as to how much capital you want to protect above that, and definitely the only way to do it was either cat bonds or industry loss warranties. I think industry loss warranties have been far more readily available because of the speed that you can transact them. Our reluctance to go the cat bond route is that in two and half months the market can change substantially. We tried to buy one last year and the model kept being recalibrated as we were buying it. Two and half months is too long a time.

Jonathan Barnes: The other point I wanted to make was about the basis risk. Obviously, basis risk is a problem and the buyers need to feel comfortable that they have squeezed out that basis risk. One of the lessons I am taking away from these recent floods in the UK is that there is a basis risk inherent in the traditional product as well. Buyers of protection have spent a lot of money buying protection for catastrophe events in the UK and Europe and the main focus of that has been for European windstorm risk. Here we are seeing very significant losses arising from these floods and the majority of that damage is not going to be relieved by reinsurance at all. I would say that is basis risk in the traditional product. Indeed, rather than purchasing reinsurance in the traditional way on an each and every event basis, maybe we should be looking at annual aggregate covers. Indeed, from the buyer’s perspective frankly he does not know whether he is going to be hit by a frequency of smaller losses, a single major event, or a combination of the two.

Alan Fowler: I agree. We have an aggregate weather cover and it is engaged.

Jonathan Barnes: Well, you will be the hero then!