This has been a disturbing year for airline passengers. Just before New Year's Day an American Airlines 757 approaching Cali, Colombia, smashed into a mountainside, killing everyone aboard. In May a ValuJet DC-9 plowed into the Florida Everglades, leaving no survivors. Then in July came the explosion of TWA Flight 800, with another 230 deaths. While there is no common causal factor linking these disasters, by the time of the TWA crash politicians were falling all over themselves to argue that the government needs to get tougher on air safety.
Longtime opponents of airline deregulation found the ValuJet crash a convenient opportunity to slam the low-fare airlines that pose the greatest competitive threat to the formerly cartelized major airlines and their high-cost, unionized workers. Another faction consisted of the good-government types who attack the Federal Aviation Administration for a built-in conflict of interest: a statutory duty to "encourage and foster" commercial aviation while regulating its safety.
President Clinton and Transportation Secretary Federico Peña, who last year were taking credit for the billions of dollars saved by air travelers each year thanks to low-fare airlines, did a quick 180-degree turn following the ValuJet crash, ordering the FAA to ground the airline. Despite horrendous crashes, including five accidents in five years for USAir, no major airline has ever been grounded, as ValuJet was on June 17.
Before jumping to conclusions about whether the FAA has been too tough or too lax, it's important to consider the relevant facts. First of all, there's a long history of FAA policy changes driven by fatal crashes. During the 1950s the FAA's predecessor agency, the Civil Aeronautics Authority, resisted implementing radar separation ("positive control" of airspace) until midair collisions in 1956, 1958, 1960, and 1965 got its attention. Virtually every major air safety advance, including ground proximity warning systems in cockpits, was preceded not by a careful weighing of evidence but by a series of crashes, followed by political demands that something be done. The National Transportation Safety Board investigates each accident and makes numerous recommendations, but most of them are not adopted by the FAA.
It's clear that the FAA has refrained from taking actions that might have prevented certain crashes. Advanced ground proximity warning systems that would have given the pilots at Cali three times as much warning are available at $40,000 per plane, but they have not been mandated. For years the FAA has seen in-flight fires caused by hazardous cargo, including oxygen generators like those suspected of causing the ValuJet crash, but has done nothing to improve fire warning and suppression systems or to authorize its inspectors to spot-check cargo for hazards.
On the other hand, when it came to making a (post-crash) decision about ValuJet, there was a case for singling out this airline. ValuJet has amassed the worst accident record in the industry: 4.2 accidents per 100,000 flights. That's three-and-a-half times the 1.2 per 100,000 rate for all low- fare startups. Worse, it's 14 times the 0.3 accident per 100,000 flights for the major airlines. And in its intense scrutiny of ValuJet operations, the FAA found numerous safety lapses, such as inadequate supervision of maintenance contractors and 31 flights by a jet with an inoperative weather radar.
But even those statistics don't necessarily mean ValuJet should have been closed down by the FAA. As consumers, do we want an air safety regulator that takes the economic health of the industry into account, as the FAA is mandated to do, or do we want a safety regulator more like the Food and Drug Administration or the Occupational Safety and Health Administration, one that focuses solely on preventing accidents? Despite the understandable desire for zero accidents, we would not actually want to pay the costs of an all-out effort to attain that goal. There are any number of changes in aircraft design, maintenance procedures, and flight operations that could theoretically increase safety, but at large cost. We need to remember that, even though commuter and low-fare startup airlines have higher accident rates than the major airlines, flying on them is still much less risky than driving. If the FAA forced every possible safety improvement onto airlines, that would raise their costs (and hence airline fares) enough that many of the 47 million people who flew low-fare airlines last year would go back to driving--at higher risk of injury or death.
There are no easy solutions for better air safety. But depoliticizing safety policy making and enforcement would lead to more rational decisions. Congress should remove the FAA's statutory duty to "encourage and foster" commercial aviation, so it won't be reluctant to take costly but sensible safety steps before a spate of crashes. But the FAA should still be required to prove the reasonableness of its regulations in terms of their relative cost per potential life saved, taking into account the potential for more injuries and deaths due to travelers who switch from planes to cars because of higher fares.
A second step would be to remove the FAA's other conflict of interest: its operation of a major component of the aviation system, the air traffic control system. Divesting that system to an independent corporation would put air traffic control at arm's length from the FAA, just as airlines, airports, and aircraft manufacturers are. At least 16 other countries, including Australia, Britain, and Germany, have successfully spun off air traffic control.
Another modest reform would be to require the FAA to publish data on the safety records of individual airlines--both accident rates and violations of air safety regulations. Most of the travelers who currently choose low-fare (but slightly riskier) airlines would probably continue to do so, given the relatively small risk differences and large price differences. But the potential loss of even a small fraction of their customers due to the safety ratings would give those airlines an additional incentive to improve their performance.
These changes would help, but as long as airline safety remains in the hands of a bureaucracy, answerable to politicians, air safety will still be politicized. Its bureaucrats will still have incentives to cover their tails by means of paperwork, which a long history of critical reports from the General Accounting Office makes clear has been a chronic FAA problem. And politicians will still have incentives to overreact in one direction or the other, as we've seen with the ValuJet and TWA situations. Couldn't we design a system with better incentives for rational behavior?
Imagine for a moment that there were no FAA. What would prevent some airlines from cutting corners and taking risks few of us would want to accept? To find an answer, consider who would have the most to lose in such a world. Clearly, it would be insurance companies, who would bear the brunt of the risk. Since no one would operate an airline without insurance, the insurance companies could face ruinous exposure for liability and replacement costs due to an increased number of crashes. Therefore, the insurance industry would have to engage in aggressive loss prevention activities, as insurers currently do in factory safety and fire protection. (There is no federal agency that regulates your local fire department; that is done by the nonprofit Insurance Services Office.)
What would probably emerge is a nonprofit entity, funded by and answerable to the insurance industry, that would set air safety standards. And that would dramatically change air safety incentives. Consider the apparent cause of the ValuJet crash. One of its maintenance contractors improperly labeled hazardous oxygen generators and illegally loaded them aboard the doomed DC- 9. It turns out that the nine different FAA regions have nine different hazardous materials policies-- and none has the authority to open or inspect packages to see if they contain hazardous materials. An FAA memo described this problem well before the ValuJet crash, but no action was taken because nobody's money or job was on the line.
If an insurance safety organization identified such a problem, it would have a strong financial incentive to solve it, in order to prevent future losses. And the airlines would have a financial incentive to comply with the insurance entity's safety standards because doing so would lower their premiums. (Today, airlines typically fight proposed safety requirements because of their cost.) Some airlines might opt for a higher standard than others, seeking the best balance between insurance costs and safety expenditures. They might be rewarded with a published safety rating from the insurance entity, analogous to the Underwriters Laboratories symbol on electrical appliances.
On the other hand, it's unlikely that an insurance safety agency would push standards that imposed exorbitant costs on airlines for minuscule benefits; that would not serve the interest of the airlines, their passengers, or the insurers. But since it would not be in anyone's interest to have planes falling out of the sky, we could expect reasonable, science-based tradeoffs between safety improvements and cost.
An insurance-based system, which would depoliticize airline safety, should be the ultimate goal. Until we can generate support for that route, reforming the FAA by simplifying its role, requiring cost-benefit analysis, and making it publish safety data is probably the best we can do.