Death Toll from Air Crashes at All-time Low in 1999

Air Safety Week, Jan 10, 2000

Cautionary note: One good year does not make a trend

The low number of passenger deaths marked 1999 as one of the safest years in the half-century history of large-scale commercial air transport.

On all Western-built aircraft with 15 seats or more, jet or turboprop-powered, some 435 passengers were killed in 58 known accidents in which the aircraft were written off as total losses. The figure increases by 77 additional passenger deaths when losses involving Eastern-built aircraft are factored in, bringing the 1999 total to 512 deaths on revenue passenger flights. Even so, the death toll is lower than the 585 passengers killed in 1946 - the base year marking the advent of air travel for the masses. In this respect, the lower death toll for 1999, despite the exponential increase in air travel in the 53 years since, represents a significant achievement.

"After a couple of bad years, it's very nice to see a good year," said Paul Hayes, director of air safety for Airclaims, Ltd., the UK-based company that tracks such trends. Its annual special bulletin traditionally provides an accident overview of the year just ended. Airclaims' Dec. 31 bulletin was no exception, summarizing 1999 as follows: "Apart from the relatively low number of passenger fatalities other, well-established, trends generally continued during the year. The number of major losses in 1999 was much as expected, while the cost to insurers... continued its relentless climb. The current estimate of incurred losses in 1999 at $1.8 billion is virtually right on the trend line."

These are preliminary figures, but they should hold generally. Hayes said the less-serious accidents tend to be under-reported. "A couple more (accidents) come out of the woodwork every year," he explained.

Even so, the overall situation is pretty well established. The 299 passengers killed on Western-built jets is the lowest since 1984 (15 years ago), when only two passengers were killed on revenue jet flights, according to Airclaims. Despite this record, Hayes cautions against making too much of the record in one year. "It's the trends that you have to watch, and work on," he declared. For example, the data show a very gradual decrease in the number of spectacular, fatal crashes, from about 7-10 per year in the 1970's to about 8 annually in the 1990's. The Airclaims counts of destroyed aircraft, 23 jets and 35 turboprops in 1999, include "constructive hull losses," or those aircraft damaged to the point where they were written off.

The gradual decrease in the annual toll of smashed machinery is not matched by the human toll. "The number of passengers killed does not show the same decrease," Hayes observed. Aircraft are generally bigger today; sheer demand and yield management systems have pushed up average load factors.

Unquestionably, the cost of accidents is increasing in financial terms. "Passenger settlements are increasing way ahead of the rate of inflation," Hayes said. When more expensive airplanes crash, those losses mount, too. The estimated $985 million cost of major hull losses for Western-built jets made 1999 the worst year ever, with a noticeable increase over the $900 million cost for 1998. The same trend held for turboprop-commuter aircraft: hull losses in 1999 cost $136 million - a significant bump over the $98 million cost in 1998 (but not as bad as the $143 million cost after a rash of accidents in 1994).

From the Airclaims tabulations, a few points bear mention:

* The death toll was dominated by a few accidents. Of the 23 jet losses in 1998, only five involved a total loss of all passengers and crew aboard and, of these, three crashes involving cargo or ferry flights were not carrying any passengers. Nearly 70 percent of the 299 passengers killed in 1999 were aboard EgyptAir Flight 990, which crashed October 31st off Nantucket.

* Most of the passengers survived. Four of these 23 jet losses in 1999 might be categorized as survivable accidents, as not all passengers were killed. In these less-than-wipeout cases, 83 of the 833 passengers aboard were killed. In other words, nine out of ten passengers survived. That survival rate is encouraging. A study by the National Transportation Safety Board (NTSB) of 17 survivable accidents over the period 1983-1996 showed that about 8 out of 10 passengers survived accidents involving fire and serious injury and substantial damage or loss of the aircraft (see ASW, Feb. 1, 1998).

* Larger aircraft may be more survivable. While everybody was killed in 5 of 23 big jet accidents, all aboard were killed in 12 of 35 turboprop/commuter crashes in 1999. The difference suggests that 80 percent of the big jet cases were survivable, but less than 70 percent of the crashes in the regional airline category were survivable. Admittedly, this impression is based on a one-year snapshot, but the crashworthiness of smaller aircraft may merit further examination.

Of note, 1999's laudatory accident record must be viewed in context. New technology is helping to avoid crashes. For example, the enhanced ground proximity warning system (EGPWS) on an Air France B777 helped the crew avoid a potential impact during a February, 1999 approach to landing at Rio de Janeiro, Brazil (see ASW, May 24, 1999). Other close calls occurred. A number of near-crashes could have dramatically altered the picture for 1999. For example, two 747's nearly collided April 1, 1999 at Chicago's O'Hare International Airport (see ASW, April 26, 1999). There also were a number of reports last year of aircraft on final approach nearly landing right on top of aircraft waiting at the end of the runway for takeoff. Were it not for alert aircrews, the outcomes could have been fireballs rather than near-misses (see ASW, July 26, 1999). In the inimitable words of Terry Kelly, manager of safety policy for NAV CANADA, "The absence of accidents is not proof of the absence of hazards."