The insurance industry took its first steps into aviation just a few years after the airplane was invented. Arrangements for insuring early airline operations sometimes were in place even before client airlines existed because of insurers’ experience with the risks of other transportation modes. More than 90 years later, however, aspects of aviation insurance that are familiar to an airline’s financial risk manager may not be as familiar to its operations risk manager, although both face challenging demands to quantify the economic value of making specific investments in safety.
The main coverages in 21st century aviation insurance policies — excluding those related to war, hijacking and other perils, including terrorism — are for partial, major partial or total hull loss, meaning damage to the aircraft; liability for injury or death of passengers; and third-party liability, meaning liability for bodily injury, death and property damage external to the aircraft. Hull losses typically are paid within weeks, based on an agreed value of the airplane. For airline accidents as a whole, insurers’ third-party loss amounts — for example, payment to the owner of a building damaged by an aircraft — have been almost negligible, but with a potential for catastrophic losses in some scenarios.
Exact individual and aggregate passenger liability after an accident, however, is difficult to determine quickly. “Depending on the size of the aircraft, geographical area of operation and the relative legal requirements, [liability] limits can range anywhere from US$250 million to $2 billion,” according to Swiss Reinsurance Co. (Swiss Re). “Insurers provide these liability limits to the airline for each aircraft, each takeoff and hence each occurrence, and there is no limit to the number of occurrences covered in a given [one-year] policy period.”1
Two major airline losses underscore the concern about potential third-party liability. Swiss Re said, regarding the loss of Pan Am Flight 103, in which 259 occupants and 11 people on the ground were killed after a bomb detonated in a Boeing 747 over Lockerbie, Scotland, in December 1988, “third-party losses were caused by terrorism, the theme, which, unfortunately, many believe has grown into the pre-eminent concern in air travel today.” The largest recent loss in aviation insurance terms — about $225 million — occurred when American Airlines Flight 187 crashed at Belle Harbor, New York, U.S., on Nov. 12, 2001, following the in-flight separation of the vertical stabilizer on an Airbus A300, according to Michael Mahoney of GE Insurance Solutions.2
Recovered from economic shocks of Sept. 11, 2001, aviation insurers by late 2004 operated in an environment in which the hull value of an airliner could be valued at $1 million to $250 million, and potential liability for the passenger awards in one fatal accident could be $1.5 billion.
In a 2006 presentation to the U.S. Casualty Actuarial Society (CAS), Mahoney estimated that average airline insurance liability awards per passenger currently vary from $150,000 as a global low to $400,000 in the Asia Pacific region; $500,000 in Latin America and the Middle East; $1 million in Australia and South Africa; $1.5 million in Europe and Canada; $2.5 million in Japan, Hong Kong and Singapore; and $3.9 million in the United States.
Scenario in 1919
According to 1920 proceedings of the annual CAS meeting, England in 1913 became the first country in which an insurance underwriter issued a policy on an airplane; an insurance pool was planned. “Soon after the [World War I] armistice, the leading insurance companies [in England] combined to form a pool to take care of aviation risks,” said H.E. Feer, representing the Scandinavian Pool for Aircraft Insurance and its statistical institute. The pool and institute were set up in 1919 by about 90 companies, even before Scandinavian airline service began.3
In an early presentation about airplanes to the CAS meeting, Walter Cowles in 1919 said, “The fact that we, here in the United States, are far behind England and all other countries in the development of this most helpful competitive means [the airplane] should not deter us, as representatives of insurance interests, from laying a sound foundation and establishing a useful practice for aircraft insurance, notwithstanding present discouragements, notwithstanding a limited field and notwithstanding the lack of substantial hope for the immediate future. … We must have the aircraft. It must be developed and improved. It must be cheapened in cost and upkeep. It must be dependable. It must be practical.”
A. McDougald, commenting on Cowles’ paper the following year, urged timely accident investigation and dissemination of related data. “Only by [accident investigation] can weak points in administration, personnel and material be eliminated and the safety of the public proportionately increased. … Aircraft risks as the subject of insurance are new, and it must necessarily be some time before any dependable data can be collected on which to base equitable premium rates. In the meantime, the arbitrary rates will be governed by considerations of analogy and argument, and influenced possibly to some extent by competition.”
McDougald cited information from the British Air Ministry covering the last eight months of 1919, showing 403 airplanes in use, 35,330 flights and 8,368 hours flown carrying 64,416 passengers and 67,143 pounds of cargo. Eighteen accidents during the same period comprised four fatal accidents, eight accidents involving injury to one or more aircraft occupants, one accident involving death of a third party, and five accidents with no injuries or fatalities. There was one accident per 1,960 flights and 465 flight hours. “The most common cause of accident was engine failure, of which there were six cases,” he said.
Risk-management priorities in England then were dissemination of weather reports by radio, “the development of aerial navigation by means of directional wireless,” introduction of strip maps containing flying directions with “information as to landing grounds, aerodromes, wireless and meteorological data,” experimental lighthouses, and aviation-specific meteorological research.
Fast Forward
A 2006 survey of 51 of the world’s top 200 airlines by revenue — conducted by Airline Business magazine and Aon, found that an average of 2.1 percent of participating airlines’ total revenue was spent on risk management, with about 70.1 percent of that amount representing costs of aviation insurance premiums. Researchers estimated that the top 200 airlines would spend $5.86 billion on aviation insurance premiums. Aon said, “The survey highlighted that risk has a rising profile in airline organizations and is now managed at the very highest levels. Over three-quarters of airlines employ an individual or team in a dedicated risk-management role, nearly half of which report directly to the CEO’s office.”
Aviation insurers may offer insights to operations risk managers on their airlines’ overall scope and scale of exposure. While aviation safety professionals typically work to reduce risk in all aircraft operations that they can influence, aviation insurers see in the world a very broad range of risks, including natural catastrophes such as hurricanes, tornadoes, earthquakes, floods, hail storms, bird and other wildlife strikes; plus man-made exposures, such as those involving war and terrorism. Natural risks have generated major losses for aviation insurers. “For example, one [hail storm] struck at Kingsford Smith Airport in Sydney, Australia, in mid-April 1999, damaging 45 aircraft and resulting in a loss of $61 million,” Swiss Re said.
Aviation insurers know that passengers on a typical airline flight represent several hundred million dollars of liability exposure, with the exact amount dependent on the passenger profile. “Yet when determining the size of a loss after an accident, the types and nationalities of the passengers on board are more important than their actual number,” Swiss Re said. “The ‘type’ of the passenger refers to the status of the traveler [e.g., each person’s earning power and dependents, and the country in which court action can be brought]. … [The jurisdiction] factor is central to insurers’ exposure calculation, as compensatory damages can vary greatly from jurisdiction to jurisdiction.”
Where’s My Discount?
A question that arises among operations risk managers is whether a specific safety-related change will reduce an insurance premium the following year, which seems like a good incentive for senior management. One problem, however, is that methods of pricing the premium vary widely. “In simple terms, premium is the cost of a promise on the part of the insurer to pay the insured in the event of claims covered under a policy,” Swiss Re said. “Numerous factors — which include the frequency and severity of the potential claims — will determine the price of that commitment.”
Morton Lane, a U.S. broker-dealer, said in 2003, “There is no agreed-upon theoretical method for pricing [aviation] insurance risk. Several approaches have been designed but none can claim ascendancy over another.”4
Nick Brown, chief underwriting officer–airline insurance, Global Aerospace, said that the imperative of spreading the risk transferred from an aircraft operator to a large number of disparate insurers and reinsurers adds complexity to understanding premium pricing and the underlying economic factors. “As a consequence of the very large limits of indemnity, all airline insurance policies are syndicated among a panel of co-insurers,” Brown said. “It is important to understand that each individual insurer will have its own underwriting criteria and its own methodologies for calculating the premium for a given account.”
Technical variables familiar to airline operations risk managers are only part of the equation. “Our pricing models take into account a wide range of risk factors in addition to the basic exposure metrics (fleet values, passenger numbers, departures, etc.),” Brown said. “This includes loadings [adjustments that increase premium] and discounts which are specific to quantifiable technological factors — such as the percentage of the fleet equipped with [a terrain awareness and warning system (TAWS)] and traffic-alert and collision avoidance system [TCAS] — and also more subjective evaluations of the quality of the safety management system [SMS] or safety culture of the airline in question.
“However, the overall premium payable by the airline in question will be an amalgam of the offers of individual insurers, who will all quantify such factors in differing degrees according to their own objective or subjective pricing criteria. Additionally, simple ‘market forces’ will have a significant influence on the actual premium paid. This makes it difficult or impossible to quantify the economic value — in insurance-premium terms — of making investments in safety, at least on a prospective basis. On a retrospective basis, there is a very clear benefit in insurance terms, because the loss record of an airline will have a significant bearing on how its premium is rated. Over time, therefore, airlines [that] have poor safety management pay much higher premiums due to their claims experience and, conversely, airlines that improve their safety management and consequently improve their claims record will benefit from lower premiums.”
Running With Data
Paul Hayes, director of Ascend, said that aviation insurers influence airline management to accept the reality of risks that psychologically may seem incongruous with the safe operations they observe day after day. “Most airlines in the world are small airlines that have never had a catastrophe; in any five-year period, 90 percent have not suffered a loss,” Hayes said. “Obviously, there is great variation between them, but as a group, those hundreds of small airlines have a worse accident rate than large airlines. In a big airline, it may be that the current management does remember the last disaster the airline had. Managements change so quickly in small airlines, so there might be a view that ‘yes, there’s a chance it will happen to me but probably accidents will happen to someone else.’ Accidents are so far removed from their experience, from operations management — but at some small airlines, if they have an accident the airline is gone. Take, for example, Birgenair, a Turkish airline that had a Boeing 757 crash into the sea off the Dominican Republic, killing German tourists — the airline didn’t survive.”5
In many areas of aviation insurance practice, from exposure modeling to insurance premium pricing, external proprietary databases often are used by brokers, insurers and reinsurers, according to Hayes. “Insurers use these data strategically — their actuarial departments will look at flights being made, number of flight hours performed, average number of aircraft in the fleet, accidents they suffered, total losses, fatal accidents, number of people killed, passengers carried, revenue passenger kilometers, etc. so that they can work out the frequency of losses of certain sizes. Our data do not allow them to see that airline XYZ does all these good things [for example, TCAS, TAWS, SMS or flight operational quality assurance (FOQA)] but airline ABC doesn’t. That has to be part of the information underwriters discover or assume when they’re writing the insurance coverage. The actuarial department also will use these data tactically to look at risks and simple things, like, for this airline fleet, ‘these are the market values of their aircraft, how do they compare with the agreed values?’”
Another current application of these databases to aviation insurers’ models has been to test hypotheses of why another large aviation insurance loss has yet to occur. “A major discussion in the insurance market is how frequently aviation catastrophes are occurring now,” Hayes said. “Prior to 9/11, there was an assumption that somewhere in the world, they would get a catastrophic loss every year and a half to two years, or something like that. Six years have gone by, which is an unprecedented period. The American Airlines Airbus A300 in Queens, New York, U.S., in November 2001 was the last catastrophic loss in insurance terms of looking at the dollar cost. So actuaries are doing … statistical analyses to try and decide whether this is the extreme part of a normal distribution, or whether the trend has changed. It looks like maybe it has. Actuaries are looking at a very high level of safety. One argument put forward is that the recession in the airline industry resulted in so many older-generation aircraft being parked in the desert that we’ve got a marked change in the fleet makeup … a far higher percentage that are higher technology types, plus TCAS and TAWS are in most of the world’s fleet today.”
Influential Research
In the late 1990s, Skandia Insurance Co. — now Inter Hannover Scandinavian Branch — published an analysis of 20 years (1976–1996) of hull-loss accident data. Its research compared turbine-aircraft operators that had implemented routine flight data monitoring with those that had not, in the United States and worldwide. Skandia’s data showed that by the mid-1980s, hull losses among operators with less than seven years of flight data monitoring — expressed as a percentage of total turbine fleet — were lower than the percentage for operators worldwide. Moreover, operators with seven to 14 years of flight data monitoring had better results than all other U.S. turbine aircraft operators, and operators with more than 14 years of flight data monitoring by 1996 had the fewest hull-loss accidents, about 0.05 percent. These findings frequently have been cited by Flight Safety Foundation (FSF) — and by others in U.S. Federal Aviation Administration presentations as recently as April 2005 — in promoting the voluntary adoption of FOQA programs.
Participants in the FSF Ground Accident Prevention (GAP) project during the past five years have adopted methods similar to aviation insurers to take a closer look at the financial consequences of ground accidents and incidents to airlines, along with the risk factors discovered in loss reports. They initially found that the overall cost of global jet airliner damage was difficult to quantify because individual losses amounted to less than the deductible of their aviation insurance policies. At an early stage, the participants analyzed proprietary airline data for 274 ground accidents in 2001–2004 and found that although total industry losses could be estimated to be $30 billion, the losses from all but one of the sample accidents were less than the applicable insurance deductible, and the average loss was $250,000.
Advising Corporate Operators
Aviation insurers also may influence operations risk management within corporate aircraft operators, helping them to prioritize how they address exposures and keep them in perspective, according to Bob Conyers, vice president and manager of general aviation safety for Global Aerospace. “We offer safety services free to insured operators, for example,” Conyers said. “The most popular service is a flight operations survey, which entails a full review of management policies, training standards, operational procedures and maintenance practices. The idea is to assess a flight department’s operation compared to similar operators and to pass along ‘best practices’ — typically well beyond regulatory minimum requirements — that have been observed.”
Safety-problem recognition by an aviation insurer can generate safety recommendations to its insured aircraft operators. “Following the [fatal Gulfstream III] accident in Aspen, Colorado, U.S. [in March 2001], we have encouraged operators — and generally have been successful — to adopt higher-than-published minimums at mountain destinations,” Conyers said. “Strict adherence to higher minimums is generally supported by inclusion in the company’s flight operations manual.”
Aviation insurers also may influence insured aircraft operators to be alert to possible operational risks attributable to increases in air traffic. “Insurers in general have been successful in convincing operators of single-pilot jets and turboprops to fly with two-pilot crews,” Conyers said. “This is not attributable solely to increases in traffic, but success in this area has come primarily in the eastern third of the United States and on the West Coast, where congestion is greatest.”
Aviation insurers’ advocacy of simulator training for turbojet pilots was a classic example of positive influence, according to Ed Williams, CEO of the Metropolitan Aviation Group and chairman of the FSF Corporate Advisory Committee. “[In] the early 1960s, accident rates of both air carriers and the then–brand new corporate jets were much higher than today,” Williams said. “Training accidents, using the aircraft itself, were a particularly deadly endeavor. No one doubted the need for proper ground and flight training of flight crews, but at that time there was a definite reluctance — if not downright refusal — from a significant number of corporate flight department leaders to embrace flight simulator–based training. In fact, from the World War II and Korean War eras, there were some chief pilots who believed that they didn’t need training because of their high number of total pilot flying hours. The attitude was, ‘I’m already a highly experienced pilot with no accidents, and I don’t need the training.’”
A combination of training accidents and other accidents during the transition from propeller-driven airplanes to corporate jets took a toll on aviation insurers, who, as a group, decided that something had to be done, he said. “About the same time, FlightSafety International began developing the first flight training simulators for the newly introduced corporate jet aircraft,” Williams said. “About that same time, the air carriers had begun to utilize their simulators more and more, and their collective training accident rate — compared with using the actual aircraft — was showing a definite decrease.
The aviation insurers were inspired to begin using their collective clout toward ‘forcing’ corporate insureds to begin using the new simulators. The fact was that [training] could now come much closer to simulating very serious on-board mechanical problems, abnormal conditions, outright emergencies, low-ceiling instrument approaches and other situations than could ever be accomplished safely in an aircraft.”
So aviation insurers required their insured corporate flight departments either to begin utilizing the available simulator-based training programs or face very high premiums or refusal of coverage. “This influence evolved over 40 years into the situation today in which corporate flight departments are effectively uninsurable if they operate a turbine-powered aircraft but professional, ground-based and flight simulator–based training programs aren’t an integral part of their operations,” Williams said.
Worldwide Implications
Among advances in the airline industry capturing the attention of aviation insurers is the International Air Transport Association’s Operational Safety Audit (IOSA) program. “It is yet unclear in the current market environment if airlines gain any significant credit from IOSA accreditation in terms of short-term insurance premiums, but it seems likely that if they help build the long-term safety of the aviation industry, the price of insurance should ultimately fall,” Aon said in a fall 2006 airline insurance market review.
Access to affordable, bona fide aviation insurance coverage remains a critical issue for some aircraft operators in the developing countries with substandard physical and regulatory oversight infrastructure. “It is true that the greatest variation in operating standards is seen in developing parts of the world, and it is in these areas that insurers are most likely to make a positive intervention in order to try and improve the safety of a particular operator,” said Brown of Global Aerospace. “Typically, this involves the lead insurer commissioning a third-party expert to conduct a review of the airline’s operations and to make recommendations. The lead insurer will then require the airline to address those recommendations and, in the event of non-compliance, may issue notice to cancel coverage.”
The opposite concern, however, would be the possibility that this free market can allow substandard aircraft operators to obtain aviation insurance coverage, with a possible implication to passengers that safety standards have been met. “There are certainly airlines to whom Global Aerospace would not offer coverage due to safety concerns,” Brown said. “Insurance is a free market, however, and many of these operators will find coverage from other aviation insurers, possibly at very high insurance rates. Others will not be able to buy any coverage in the ‘mainstream’ aviation insurance market and will either operate without insurance or buy low limits of coverage from local or non-conventional insurers. This will inevitably limit the scope of such airlines [because they] will be unable to meet the insurance requirements necessary to fly into North America or Europe.”
Notes
- Chrystal, Philip; LeBlanc, Suzanne. Flight to Quality — Financial Security in the Aviation Insurance Market. Swiss Re Technical Publishing, Aviation. Swiss Reinsurance Co. September 2002. Chrystal, Philip; Fok, Marcel; Martino, Ferdinando; Peter, Andreas; Shirai, Shinji. The True Value of Aviation Insurance. Swiss Re Technical Publishing, Aviation. October 2004.
- Mahoney, Michael W. Aviation Pricing and Modelling. Casualty Actuarial Society 2006 Seminar on Reinsurance. Mahoney is senior underwriter, GE Insurance Solutions.
- Cowles, Walter G. “Aircraft Insurance.” Proceedings of the Casualty Actuarial Society, 1919, 31–51. MacDougald, A.; Feer, H.E. “Aircraft Insurance: Discussion.” Proceedings of the Casualty Actuarial Society, 1920, 328–346.
- Lane, Morton N. Pricing Issues in Aviation Insurance and Reinsurance. A paper presented to the 2003 Thomas P. Bowles Jr. Symposium of the Casualty Actuarial Society, March 2003. Lane is president of Lane Financial.
- According to a May 1996 safety recommendation by the U.S. National Transportation Safety Board, the Boeing 757 — operated as a charter flight to Frankfurt, Germany, by Birgenair — crashed into the Atlantic Ocean after takeoff from the international airport at Puerto Plata, Dominican Republic, on Feb. 6, 1996. All 189 people aboard were killed and the aircraft was destroyed.