‘Clearing’ price technology
With new technology many areas of life are changing, some think for the better and some think for the worse. This is equally true in the reinsurance market where technology is coming in to try to make things more efficient and ‘better’, often in the hope of reducing the price for coverage.
One of these new technologies is called Tremor. It is a new platform which puts buyers of cover together with sellers of cover, but more than that, it allows buyers of cover to ‘clear’ at the lowest price. This is how the capital markets works when they issue notes: With a note issue investors are asked how much they would purchase of an issue at a price of x, at a price of x-y, at a price of x-2y, etc. The placing agent can then select the lowest possible price where he has sufficient purchasing interest to sell the entire issue, referred to as the ‘clearing price’. This can go 1 step further; if the issuer thinks that the clearing price is really good the issuer may decide to increase the size of the issue. This would push up the clearing price, meaning that the issuer can continue to increase the size until either (i) it doesn’t want any more or (ii) the increased clearing price is at a ceiling it has set itself. Tremor has very cleverly created a system which allows for placement of reinsurance in this manner. The cedent puts up how much cover it would purchase at different prices in the form of a price/demand curve, and the reinsurers would equally all put up how much cover they would offer at different prices, again in the form of a price/demand curve. The Tremor software then puts these curves together to find the amount of cover purchased at a clearing price accepted by the cedent where the entire amount of cover is placed. This would at first instance appear to be highly efficient, and most certainly getting the lowest price for the cedent. Interestingly enough this is not the first time that the concept of getting the lowest ‘clearing price’ has been raised. A couple of years ago Marsh tested a platform where reinsurers could ‘bid’ (as in an auction) to offer capacity at a certain price; this was only a one-way clearing, not allowing the cedent to increase the amount of limit purchased, but not entirely dissimilar to Tremor’s offering. The author is not sure whether the Marsh platform was ever used for live transactions, however he has never seen any transactions on it.
Is this good for the industry?
In order to answer this question, one must first understand the dangers inherent in reinsurance pricing and then how the industry has priced in the past to avoid such dangers. One then needs to understand how reinsurance is an unwritten multi-year undertaking between the cedent and the reinsurer.
Pricing for reinsurance is fundamentally different to pricing for other products. If you are making a widget you know exactly your ‘cost of sales’ i.e. the cost to produce the widget. If it costs 10 to produce the widget you know that any price above 10 provides a profit and any price below 10 results in a loss for each widget sold. A key issue with reinsurance is that you do not know your ‘cost of sales’ ahead of time. Essentially your ‘cost of sales’ is a function of the expected loss and the cost of capital. However the expected loss is not a definite figure. Reinsurers can model the risk, however even the best models have a high degree of uncertainty about them. Taking some recent examples: (i) World Trade Centre attack – not modelled; (ii) A key driver of the Katrina loss was the breaking of the levees and subsequent flooding – this was not modelled; (iii) The Ike loss was heavily inflated because it retained its hurricane strength winds into the Mid-West – this was not modelled; (iv) The severity of the Christchurch quakes were inadequately modelled as the fault line running under Christchurch was unknown; (v) The Tohoku earthquake was a category 9.1 quake when it was believed that the maximum quake on that fault line was only 8.7-8.8 (the numbers are close but on a logarithmic scale this makes a big difference). Also the damage done was, to a large extent, caused by the ensuing tsunami and, at the time, none of the vendor models modelled tsunami; (vi) The Thai Floods loss was not modelled, and when the claims from that event starting coming from Japan (under JIA clauses covering Japanese companies’ plants in Thailand – a reason why Solidum struggled to offer policies in Japan), no Japan portfolios had considered exposure outside of Japan. In short the modelled figures are not correct – they provide a very good insight into the loss potential, but still need to be loaded by experienced and knowledgable underwriters who truly understand the policies they are underwriting and the risks that they bring. Ultimately the modelled expected loss is NOT the ‘cost of sales’ as some would have you believe. The key here is the ‘experienced and knowledgable underwriters’ – writing purely to the model will ultimately end in tears.
The above danger was, to a great extent, mitigated in the past as risks which were syndicated were priced by lead underwriters. A lead underwriter is an experienced and knowledgable underwriter well known as such in the market. The lead underwriter would do a lot of work to price the risk and would drive the claims management process should claims arise. Once the broker had succeeded getting one or more good lead names onto the placement, he would approach so-called following markets. Whereas they would have underwriting expertise, they would not have the knowledge and experience of the lead. They would typically be smaller entities with smaller line sizes. With a well respected lead the broker would usually quite easily fill out the rest of the placement with following markets. Such a system provides a control on pricing – it helps to avoid scenarios where entities without the required knowledge, potentially relying on models alone, push down the price to uneconomic levels where the ‘experienced and knowledgable underwriters’ are unwilling and unable to write business.
The danger was further mitigated by the ongoing relationship between reinsurers and cedents and the concept of ‘pay-back’. There has always been an unwritten understanding between reinsurers and cedents that, over time, the reinsurer will make more money than he loses. The fact that the reinsurance market was limited in the number of players further ensured that this held true. These concepts resulted in long-term relationships between cedents and reinsurers. If there were losses in a year, the cedent would always offer the reinsurer to continue, typically at a higher price, so that he could recoup his losses. Following a loss the cedent would not merely go somewhere else for a lower price because the reinsurer who paid the loss would have a long memory and would not accept them back at a later date, maybe when the cedent desperately needed that reinsurer to reinsure them. Conversely, at renewal time with no losses in recent years, the cedent would say he has build up a so-called ‘bank’ with the reinsurer and argue that the price should go down. This concept of relationship was critical in keeping the market in balance over quite literally hundreds of years.
Coming back to ‘market clearing’ services such as Tremor; the concept of ‘bids’ from the reinsurers where the only thing that matters is the price per dollar of capacity absolutely kills the concept of the relationship between cedents and reinsurers. It totally commoditises reinsurance capacity which will ultimately end in greater litigation – without a relationship the only ‘understanding’ between the parties is the contract and it is impossible for reinsurance contracts to cover all eventualities as by its very definition reinsurance is about assuming an unknown risk. Also without a relationship and the concept of payback, reinsurers won’t care about their clients, they will no longer be trading with friends that they have known for decades (the longest relationship the author has ever come across was a 50-year contract between Allianz and Munich Re, with an option to extend for a further 50 years! That really is a long-term relationship). There will no longer be the possibility to find a solution as friends and trading partners.
The ‘market clearing’ services also remove the concepts of ‘lead’ and ‘following markets’, and the pricing discipline that came with it. On platforms such as Tremor all reinsurers vie just on price. This means that less experienced and knowledgable underwriters are putting out prices on business. If they don’t price correctly, possibly pricing to the modelled expected loss without sufficient load, they will price the business too cheaply. The lead markets therefore will not be able to write business at the correct rates until those who are under-pricing business lose too much, which may take years as the events covered remain relatively infrequent. By which time there may be new ‘innocent capacity’ willing to step into the breach and assume the business at under-priced rates. Without pricing discipline it will only be a matter of time before the market finds itself in great difficulty, with no capacity available anywhere at any price, as investors pull out believing that the market can never be profitable.
As an anecdote, in the early days of Allainz Risk Transfer it was company policy to ask whether many others were looking at the same transaction, for if they were it was immediately declined. The reason for this was that when lots of reinsurers are looking at and bidding for a transaction and you put out a price for it there are 2 possible outcomes, both of which are unattractive: (i) you miss something in your analysis and price lower than everyone and get the deal; or (ii) you do your analysis, price correctly, you are not the lowest so you spend a lot of time and don’t get the deal. This simple rule possibly encapsulates why one should not participate in such ‘market clearing’ services. Reinsurance has always been about taking unknown risks for a price where over time the reinsurer knows that they will make money through a symbiotic relationship with the cedent. This simple fact protected the reinsurer’s equity holders, or, in the case of an ILS fund, the investors in that fund, thereby ensuring that the investors were sure to make a positive return over time.