Adrian Leonard investigates the current developments in the satellite sector.

Statistics show why capital has withdrawn from the satellite insurance sector at such an alarming rate. In the five years from 1998 to 2002, total market premium for launch and in-orbit risks was $3,662m, while total paid and outstanding claims were $6,451m, according to broker Willis Inspace. In the eleven years to 1997, the market had built up a cumulative underwriting profit of £1,290m, but by the end of 2001 the balance since 1987 was a loss of £1,474m. In just four years, the market burned through its credit balance with paid and incurred losses of $2.8bn.

Key eventsUnderwriters fled. Theoretical capacity for launch and in-orbit risks peaked at more than $1.3bn in 1999, but today brokers would have to work hard to identify $750m (perhaps $800m if Berkshire Hathaway, which occasionally writes satellite risks, was to show interest). According to broker Willis Inspace, total 'working' global capacity (excluding Berkshire) is about $365m for launches, and $310m - and falling - for in-orbit risks. "We predict that capacity will shrink to the level where it may be difficult to reach the sum insured, particularly for in-orbit risks," said Inspace Managing Director Philip Smaje. "Since late September, a small number of players have been telling us that they are considering writing only launch risks. Insurers generally renew their quota share treaties at year end, and I expect they will feel some pressure."Sabine Holl, Senior Space Underwriter at Swiss Re, concurred. "Capacity has fallen dramatically," she explained "and at best I expect it to remain at current levels, but I would not exclude further shrinkage. Some markets are still dependent on reinsurance products which are no longer available, or they are available only at a price." Simon Clapham, Active Underwriter of BRIT Syndicate 2987 and lead underwriter of the BRIT Space Consortium (formerly the Marham Consortium at Lloyd's) also predicted that further restraint by reinsurers could squeeze supply. "In the past, a lot of the professional reinsurance market wrote more than one treaty. They are consolidating their lines going forward," he reported.Terms and conditions have significantly improved, so it is not unreasonable to expect new entrants, but none have put their money on the table as yet. "With markets pulling out and rates peaking, underwriters that have not suffered the losses of the last five years may consider coming in," said Mr Smaje. Some companies have been exploring the possibilities, he reported, but so far none have reached a conclusion. "We might see one or two modestly sized capacity lines come in, but will see even more disappearing. The net effect for next year's capacity will be a reduction," he predicted. That is unhappy news for satellite operators, who are already facing new, strict terms, high prices, and potentially insufficient supply. And although total capacity may be sufficient to cover the best risks, it is not available for everything that goes up. "We estimate that the world market capacity available at the moment for a good launch risk is around $300m to $350m, but that only applies to good risks," said Ms Holl. "One problem we see now is a new generation of launchers," she explained, citing changes such as a new upper stage for the Ariane 5. Manufacturer the European Space Agency says Ariane's new upper stage "will carry 25 tonnes of liquid oxygen and liquid hydrogen to feed a new powerful Vinci engine which will be able to carry out multiple restarts in flight and which will have enough power to put a dual payload of 12 tonnes into geo-stationary transfer orbit." For insurers, that means more untested launch vehicle technology. Underwriters do not quickly forget that the maiden launch of the Ariane 5, in July 1996, resulted in a catastrophic failure and fiery explosion. Ariane 4 had completed more than 50 successive, successful launches.Ariane isn't the only altered rocket. Lockheed Martin's Atlas has graduated from the stalwart model 2, which had an unblemished record, to Atlas 5, with numerous variants. Boeing's Delta 3 failed, and was followed by Delta 4, which has not taken off as Boeing would have liked. The Proton has got larger, and returned to commercial success - but it remains unfamiliar. "Launch is the riskiest phase, so launch vehicle selection has the biggest impact on the rate," Ms Holl said. "Only in very rare cases is this the other way around, when a highly sophisticated satellite is put on a very reliable launcher."

Modern risksNew technology requires more technical analysis and is inherently more risky, so it attracts higher rates. The higher the sum insured the client needs, the harder it is to find the capacity. If, on top, the risk is complex from a technical point of view, the rate goes up. This is good news for the market, which has adopted a much more technical basis for its pricing. "We see a very differentiated approach to underwriting, which takes into consideration the technical specifications of the risk, and also the full range of premium components we need to write this business in a profitable way," Ms Holl reported. New launch vehicles are attracting much higher rates. "Maiden and early launches are showing a significant price difference," Mr Smaje said. "The difference between the best and the worst launcher is probably a 100% rate loading for the launch." Satellite insurance prices are returning to their peak. "We are going towards rates of 16%, 18% and 20% [of insured value], and the latest losses will help to maintain the increase in the rates," said Mario Montelatici, Head of Aviation Underwriting at reinsurer Converium. Is it enough? Mr Smaje has two answers. "Launch rates are as high as they have been in several years, and conditions are more restrictive. The view of the insurers, rightly or wrongly, is that launch rates are sufficient to make money, but that in-orbit rates are marginal, despite the increases. We are seeing pressure on the rating of in-orbit cover, and moves to more restrictive coverage." In-orbit losses have not been kind to insurers. Once accounting for about 40% of losses, they now comprise 60% or more, in part because satellite manufacturers have cut corners in pre-launch testing, bringing construction time down from about three years to about two ("magically," in the words of one underwriter). Regardless of build quality, in-orbit risks are more complicated. "The underwriting process is more time consuming, with more questions on the health of the satellite," Mr Smaje said. "When insurers don't like the risk, they may impose policy restrictions... Simply increasing rates is not really possible. Some operators are saying cover is not viable at higher rates, and are looking at coinsurance or increased retentions." Policy conditions have changed dramatically. Coverage periods have been universally limited to one year. End-of-life power and fuel margins have been tightened up, while the constructive total loss point for functionality has been increased to 75%, from 50% in the soft market, and some policies placed are carrying a 90% constructive total loss (CTL) limit. Some markets would like to abandon CTL points altogether, since doing so would take salvage value out of the equation. Some underwriters believe that a specified loss of functionality should simply lead to a specified partial indemnity, rather than triggering a payment and negotiation of salvage values. Others want much more of the salvage revenue to go the underwriters. Terrorism has been subjected to an explicit exclusion, and a "generic defects" clause has been discussed. The latter must be in part a result of the catastrophe of Boeing's 601 and 702 series satellites, and the malfunctioning power arrays they carry. "They have affected the discussion in a significant way," Mr Smaje said. "The 702 is a bit of a black cloud hanging over the sector, and it will take some time to unravel." Six 702 satellites are affected, with a total insured value of $1.621bn. So far, only Thuraya Telecoms has resolved its 702 claim. It sought $371m for the CTL of its Thuraya-1 satellite, the first 702 launched, but sources say it settled $241.15m, or 65% of the insured value. It is likely that the remaining five losses will be settled, underwriters and brokers say, although it is thought that insurers have reserved for the full amount. XM Satellite Radio, which has two faulty 702s in orbit, recently stated: "XM has insurance claims in process relating to the power degradation trends experienced by the satellites. A group of XM's insurers recently denied these claims, asserting that the satellites are still performing above the insured levels and the power trend lines are not definitive; these insurers also allege XM failed to comply with certain policy provisions regarding material change and other matters. The company... will proceed to settlement discussions, arbitration or litigation (as needed) to recover the insured losses." Meanwhile, XM has made plans to launch two new satellites, and PanAmSat, which owns two 702s, has also filed notifications of claim. One possible outcome is subrogation against Boeing. However, the space sector has previously agreed not to make such claims, and continues to operate under the principles of no-blame which were established in 1963 when President Kennedy pledged to put men on the moon. Anyone who set foot on Cape Canaveral had to sign an inter-party waiver of liability, and the tradition continues, so unless insurers can prove gross negligence or wilful misconduct, there is little recourse to subrogate. If subrogation is unsuccessful, the small satellite insurance market will be left to absorb the 702 loss, which, based on the precedent of the Thuraya settlement, is likely to top $1bn. In the four years beginning 1999, insurers have incurred losses of nearly $1.5bn more than their paid claims, according to statistics from Willis. Presuming some of the resulting reserves have been posted at 100%, insurers could benefit from reserve releases if the settlement pattern is favourable. Terms and conditions, as well as rates, have been cleaned up dramatically, but the market may have further legacy problems to contend with. A number of multi-year policies are still in force (covering almost all the 601s and 702s, for example). Some coverage periods will extend another three years, leaving plenty of time for additional soft-market underwriting to impact future years. "It is possible that there could be losses in the next two to three years under the old multiyear policies," Ms Holl said. However, insurers are optimistic that the build quality of satellites is improving. "We have indications that manufacturers have spent tremendous sums to fix the defects they have on their spacecraft, but that will only be of benefit to future launches," she revealed.

More state researchIn addition, some believe that the successful application of satellite technology in recent military action will result in an increase in government-sponsored research and development for military purposes, which has obvious civilian benefits. However, it isn't always easy for underwriters to get to grips with new technology. The US Defense Trade Regulations (formerly the International Traffic In Arms Regulation) prevent the transfer of many technologies from the US to people in other countries, and as a result sometimes limit the information supplied to underwriters about new satellite technology, and slowing the risk assessment process. One market insider said the rules have made buying US-made satellites so unattractive that Europe and Japan have become the dominant satellite manufacturers. Meanwhile, the remaining space insurers struggle on. The underwriting year 2002 looked good until the last few months, when four losses hit back-to-back, including the launch failure of an Ariane 5 rocket that destroyed Eutelsat's Hot Bird 6. The loss ratio may remain below 100%, however. So far, 2003 has seen some positive settlement news on the 702s, and underwriters will have heaved a sigh of relief when SES Americom's AMC-5 satellite regained its earth lock. Insurers did open their wallets for SES in 2003, when they paid ¤291.5m for the ASTRA 1K launch failure and ¤45.0m for a battery failure in ASTRA 1G, as well as for other 2002 losses. However, this year has seen only two significant losses so far. The failure of a propulsion system on PanAmSat's Galaxy IVR (a Boeing 601) led to a $160m claim (underwriters wisely excluded losses related to the propulsion system on another PanAmSat 601, which also failed). An electrical failure on 19 September of Loral Skynet's eight-year-old Telstar 4, manufactured by Lockheed Martin, was the second loss. Munich Re, the world's largest satellite insurer, led the latter risk, which cost $141m. Space insurers made a few hundred million dollars in 2002, and stand to make even more this year, barring further catastrophic losses. However, it isn't enough yet to achieve payback of the astronomical soft-market losses that wiped out a decade of profits unmatched by insurers in other lines.By Adrian LeonardAdrian Leonard is a freelance insurance journalist and a regular contributor to Global Reinsurance.