Airline Safety: The Last Decade

The paper presents safety data for the period 1977–86 from more than 30 U.S. domestic airlines and 80 international flag carriers. These statistics are examined in conjunction with others from an earlier MIT study about the previous two decades. The primary safety measure used is “death risk per flight,” which weights each fatal accident by the proportion of passengers killed and exploits statistical evidence that the risk arising from a nonstop flight is uncorrelated with its route length. The main conclusions of the analysis are: (1) The U.S. domestic trunklines can continue to lay claim to being the safest group of airlines in the world. In 1977–86, the death risk per flight on these carriers was roughly one in eleven million. This represents a factor-of-four improvement since the early 1970's and a factor-of-ten improvement since the early 1960's. (2) But while U.S. jet travel has become safer in recent years, the improvement would probably have been greater still in the absence of airline deregulation. The reason for this assessment relates to the new jet carriers spawned by deregulation. Most such carriers have perfect safety records so far but, collectively, the recent entrants averaged twelve times the death risk per flight in 1979–86 of the established trunklines. This statistically significant excess cut by nearly one-half the recent progress in domestic jet safety. (3) International air travel was also far safer in the last decade than in preceding ones. But the flag carriers of Communist bloc and Third-World nations averaged eight times the death risk per flight as those from industrialized First-World nations. This factor-of-eight discrepancy, which has persisted with little change throughout the last quarter century, cannot be explained by differences in route structure.