Is This Any Way to Run an Airway?
Aviation disasters make headlines. The Federal Aviation Administration doesn't. This indictment cites case after case to prove FAA responsibility for major crashes. With airline deregulation leading to a surge in air traffic, replacement of the FAA is imperative, and this article shows how it can be done.
On December 1,1974, TWA Flight 514 approached Dulles International Airport outside Washington. D.C. Although it was morning, stormy weather conditions required an instrument approach. At 11:01 the FAA's enroute controller cleared the flight to descend to 7,000 feet and contact Dulles approach control The captain complied, making contact with the FAA center that handles takeoff and landings. After the captain reported being level at 7,000 feet, the controller said, "TWA 514, you're cleared for VOR/DME approach to runway 12." Acknowledging the clearance, the captain let down the landing gear and told his first officer, "1,800 is the bottom." The first officer began the descent to 1,800 feet, as indicated on the approach profile chart.
Studying his charts while the 727 descended, the captain pointed out that on his plan view chart, 3,400 feet—not 1,800—was specified as the minimum altitude. But then, in response to the first officer's query, he affirmed "When he [the controller] clears you, that means you can go to your initial approach altitude," which was 1,800 feet. "Right," agreed the flight engineer.
Suddenly the altitude alert horn sounded—once, then again. A glimpse of the ground appeared through the rain. The first officer shoved the throttles forward The altimeter warning horn gave a sharp blast. "Get some power on " shouted the captain. The horn sounded again. At 11:09:22 Flight 514 plowed into 1,700-foot-high Mount Weather, 25 nautical miles northwest of Dulles. All 85 passengers and seven crew members were killed and the Boeing 727 was destroyed.
The National Transportation Safety Board's investigation of the crash raised serious questions about the adequacy of the nation's air traffic control system, operated by the Federal Aviation Administration. To begin with, the altitude restriction due to Mount Weather was shown on only one of the pilot's two charts. And the controller had cleared the pilot to descend without mentioning any altitude restriction. In defending its controller, the FAA maintained that, because the flight was not considered a "radar arrival," the controller was not required to warn the pilot of any altitude restrictions. Thus, the controller's instruction of "cleared for the approach" meant, to the controller, "Take care of yourself." Yet, as the NTSB pointed out, "It has become commonplace to clear pilots to descend below the altitudes published on the terminal area…and instrument approach charts." A pilot will therefore tend to disregard a charted minimum if he has received an approach clearance without an altitude minimum.
Further, such key terms as "cleared for approach" and "radar arrival," according to the NTSB's findings, "have no established definitions and mean different things to controllers and pilots." A number of pilots, including the chief of the AirForce instrument flight center, testified that if they had received Flight 514's clearance, they would have descended to 1,800 feet as well. (Ironically, six weeks before the TWA crash, another fight approaching the same runway had received the same clearance and had made the same response—fortunately, close enough in to have missed Mount Weather. The crew reported the incident to their airline management as a safely problem, but the FAA itself had no procedure or system for reporting such unsafe conditions or incidents.) The NTSB concluded—39 years after the federal government took over responsibility for air traffic control—that "it is essential that a lexicon of air traffic control words and phrases be developed and made available to all controllers and pilots."
Earlier in 1974, on March 3, a Turkish Airlines DC-10 operating as Flight 981 took off from Orly Airport in Paris. As it reached 13,000 feet an improperly fastened rear cargo door blew off, leading to explosive depressurization of the fuselage. The force of the rapidly exiting air caused the cabin floor to collapse, severing all control cables to the vertical and horizontal tail surfaces. Hopelessly out of control the plane dived for the ground, crashing at nearly 500 miles per hour into the pines of Ermenonville Forest, 23 miles northeast of Paris. All 334 passengers and 12 crew members lost their lives.
The McDonnell Douglas DC-10 design had been certificated as "airworthy" by the FAA on July 29, 1971. This official governmental seal of approval meant that the giant plane complied with Part 25 of the Federal Air Regulations,which imposes minimum design standards for transport category aircraft. As such, it could legally be used in commercial service—and be insured.
Yet unbeknownst to aviation insurers—or to DC-10 passengers—even before the plane was certificated, a battle over the cargo door had been raging behind the scenes. On May 29, 1970, 93 days before the DC-10's first flight, a fuselage pressure test failed when the aft cargo door blew out. The manufacturer's "fix"—a small vent door within the cargo door, designed to prevent pressurization if the door were improperly latched—was approved by the FAA.(The agency's certification board misunderstood how the vent door operated, however, erroneously believing it to be a pressure-relief mechanism, which it clearly was not.)
Once the DC-10 entered airline service, operators began experiencing trouble closing the door properly. McDonnell Douglas issued a series of Service Bulletins to all DC-10 operators recommending various modifications aimed at making sure the door would be properly closed. But the initial "fixes" were not fail-safe. The cargo door could be forced closed without being properly latched and the cabin could still be pressurized. This is what had happened to American Airlines Flight 96, a DC-10 that departed Detroit on June 12, 1972. Climbing out at 11,700 feet over Windsor, Ontario, the improperly latched cargo door blew out, collapsing the rear floor and cutting the vital control cables. Miraculously, the flight crew was able to regain control using only engine thrust and ailerons, bringing in the crippled plane into an emergency landing.
When a life threatening safety defect is discovered, the FAA has the power to issue an Airworthiness Directive (AD), an order requiring all operators to make the necessary fix. After the Windsor incident, the FAA's western regional office prepared an AD mandating design changes to make the door fail-safe. But by order of FAA Administrator John Shaffer, the AD was never issued. To spare McDonnell Douglas and the airlines bad publicity, a "gentleman's agreement" was worked out whereby the company would issue an Alert Service Bulletin recommending various fixes. Nearly two years later, the DC-10 that crashed in Paris was found to lack a key part called for by the Service Bulletin—a part that would have prevented the door from being forced closed, as it was the day of the crash.
The crashes just described—as well as the recent San Diego collision of a 727 and a private plane—illustrate serious FAA failures in carrying out its two principal functions: operating the air traffic control system and ensuring the safety of aircraft design, maintenance, and operation. How well these tasks are carried out is, literally, a matter of life and death. It is generally taken for granted that unless the government looked after these aspects of aviation, there would be chaos and carnage in the skies. Yet many aviation observers maintain that we have chaos and carnage in the skies today, not in spite of the FAA's efforts, but because of its incompetence. Once we see, however, what the FAA is responsible for and how it operates, it becomes clear that there are other ways to get the job done.
How did the federal government get the job of trying to guarantee safety and order in the skies? After World War I there was a glut of surplus biplanes, and everyone who could scrape together a few dollars and some gasoline wanted to get into aviation. The general public and the insurance industry could hardly be blamed for considering the whole field rather frivolous. Indeed, aviation's image problem was the major concern of serious enthusiasts of the 1920s and 1930s—and led ultimately to federal involvement.
Aviation's first trade group, the Aeronautical Chamber of Commerce (later to become the Aircraft Industries Association), asserted in its 1921 Aircraft Year Book the need for "proper legislation" to "assure capital that it is entering a business project instead of a romantic adventure." Further, the group noted that "rates of insurance are likely to remain unsatisfactory so long as no competent federal agency exists to determine the airworthiness of craft or the competency of pilots."
Aviation enthusiast Herbert Hoover, then secretary of Commerce, became the industry's chief advocate within the government, drafting a bill that year to set up in his department a bureau to regulate the new industry. The proposed bill made the Commerce Department responsible for designating and maintaining all public and commercial air routes, inspecting and licensing aircraft and pilots, and establishing and administering rules of the air.
But the bill languished in Congress for five years. During this time the Army Air Service set up an Airways Section that attempted to fill the vacuum. Its involvement was justified, in part, by the federal government's airmail operations—first in government planes, and from 1925 onward via private operators under contract to the Post Office. The Air Commerce Act of 1926, Hoover's long-tended bill, finally established "federal sovereignty" in the airspace of the United States, laying the basis for today's FAA.
From 1926 until 1938 the Bureau of Air Commerce established by the act laid out airways between various cities, installing lighted beacons and emergency landing fields along each one, publishing maps, and setting up radio and telegraph stations to make weather information available at airports. By the early 1930s airborne radios were being added to planes, and the first radio-based "blind landing" system was tested by the bureau in 1934-35. By 1935 several airlines had begun developing a means of traffic control at the Newark, Cleveland, and Chicago airports; the next year, this function was taken over by the bureau.
In 1938, after several years of depression-era lobbying, the federal government provided an economic bail-out for the struggling airlines: the Civil Aeronautics Act of 1938. This law converted the Bureau of Air Commerce into the Civil Aeronautics Authority, with greatly expanded powers. Besides the traffic control and safety functions, the new agency was given vast economic powers—to assign specific airlines to specific routes so as to minimize "destructive" competition, to limit entry to the industry, to regulate fares, etc.
Two years later these economic regulatory powers were split off into a separate Civil Aeronautics Board (CAB) while the air traffic and safety functions remained in what was renamed the Civil Aeronautics Administration (CAA). The CAA was kept within the Commerce Department until 1958 when, in the wake of a spectacular midair collision over the Grand Canyon, it was reorganized as the Federal Aviation Agency (FAA) and made independent. In 1967 it was incorporated into the new Department of Transportation and its name changed to the Federal Aviation Administration.
Air traffic control and air safety certification are the FAA's two main functions. Nearly half of its 60,000 employees are involved directly in air traffic control (ATC) operations. They staff some 423 airport control towers, 25 air route traffic control centers, and 347 flight service stations. Another 10,000 technicians and engineers install and maintain the various radars, instrument landing systems, and navigation aids (such as VOR radio navigation beacons).
Nearly all airline flights and some general aviation (private) flights operate under instrument flight rules (IFR) at all times. This means they must file a flight plan with the FAA and are kept track of from takeoff to landing by the ATC system—although, as the crash of TWA 514 demonstrates, this "positive control" is sometimes less positive than the FAA would have us believe. All airspace around airline airports and above 18,000 feet is under FAA control and must be flown IFR. Outside of this airspace, general aviation aircraft may fly by visual flight rules (VFR)—essentially, on a "see and be seen" basis: VFR pilots receive flight briefings and weather reports from the FAA's flight service stations but are otherwise on their own. Since many IFR flights operate below 18,000 feet, there is an obvious potential for conflict between VFR and IFR traffic—a conflict the FAA has not been able to resolve.
Besides operating the ATC system in an attempt to avoid collisions and crashes, the FAA carries out the remainder of its safety functions by means of licensing and certification. All civil aircraft operated in the United States must be certificated as airworthy by the FAA. The agency maintains several detailed books of minimum design standards—such as Part 25 for commercial transports—against which each new aircraft is measured. Theoretically, FAA inspectors are involved at all stages of aircraft design and testing; in practice, however, the major manufacturers carry out most such inspections themselves on a kind of honor system. Under the system known as delegation option authority (DOA), "the manufacturer makes a finding of compliance with the applicable regulations without full participation of the FAA." Critics charge that the DOA system builds in an unacceptable conflict of interest.
Once an aircraft is certificated, the FAA is also concerned with its safe operation. All commercial airlines must operate FAA-approved maintenance programs, in which all work must be signed off by an FAA-licensed mechanic. Everyone else involved in operating the aircraft—pilots, flight engineers, navigators, dispatchers—must also be licensed by the FAA. Various volumes of the Federal Air Regulations (FARs) set forth the requirements for different types of licenses—for example, the basic pilot categories of student, private, commercial, and airline transport. The FAA also licenses pilot and mechanic schools and instructors. Pilots must pass periodic medical exams to keep their licenses, and airline pilots must fly with an FAA check pilot once each year.
In 1970 the FAA received a new certification mandate—airports. Under the Airport and Airway Development Act of 1970, it was authorized to issue operating certificates to air carrier airports; 500 larger airports were certificated in 1973 and 400 smaller ones in 1975. Those airports receiving federal funds under the act are supposed to comply with FAA issued safety standards, but half of the 500 large airports "had to be given temporary exemptions from some of the requirements, mostly for firefighting and rescue equipment," in order that airline operations could continue past the certification deadline set by Congress. The act established a system of user taxes to build a trust fund for modernizing airports and the ATC system. The FAA now administers the trust fund, parceling out grants to airports and purchasing new air navigation equipment.
Besides these major functions, the FAA has responsibility (shared with the Environmental Protection Agency) for regulating aircraft noise and engine emissions and supervises the anti-hijacking efforts of the airlines. In addition, it operates two airports serving Washington, D.C.— Dulles International and Washington National.
Reading the FAA's brochures, one gets the impression that the ATC system is a smoothly functioning technological marvel. Talking to pilots, controllers, and others in aviation produces quite a different picture. What emerges is a description of a classic bureaucracy, unable to keep up with technological change, wasting resources, stumbling along from crisis to crisis, pressured by lobbyists, harassed by politicians, seeking protection in the bowels of the civil service system. "Even the other federal agencies look down their noses at the FAA," says one longtime air safety expert, "because it's filled with incompetents who don't know how to manage."
Apparently, this has been typical of the FAA/CAA throughout its history, and especially so since World War II. The war produced major technological advances in communications and electronics—most notably, radar. Keeping modern jet aircraft safely separated and operating in all weather conditions would be unthinkable without radar. Yet the CAA resisted the implementation of radar until the mid-1950s. Even today (as the case of Flight 514 brought out) radar is not the primary control device in the ATC system; even at jet speeds and with today's high levels of traffic, the pilot in all cases is still supposed to look out for himself.
Bureaucratic safety rhetoric aside, the agency's primary spur to action has been crashes, not a desire to take advantage of new technology as it becomes available. Radar technology in the early 1950s made it possible to keep aircraft under "positive control" in busy airspace. Yet it took the spectacular midair collision of a TWA Constellation and a United DC-7 over the Grand Canyon in 1956 to get the CM to institute positive control above 24,000 feet—and then only during restricted visibility. Several midair collisions later, in 1958, the CAA stiffened the rules to require all traffic between 17,000 and 30,000 feet, on certain routes, to file IFR flight plans, regardless of the weather. Yet gaps in the system remained. In 1960 a TWA Constellation and a United DC-8 collided over Staten Island while one was approaching La Guardia and the other Idlewild (now JFK). Inadequate communication between the airport approach controllers was blamed for the crash, leading (finally) to a system of positive radar transition between en route and approach control centers, and later to a common approach control center for all three New York area airports. It was not until a 1965 midair collision over Carmel, New York, that positive control of all airspace over 18,000 feet altitude was instituted by the FAA.
This lurching from crisis to crisis would have been quite unnecessary had the CAA/FAA followed the high-caliber technical advice made available to it by a succession of industry/government technical advisory committees. Beginning with the Radio Technical Committee for Aeronautics, in 1948, and continuing with the Huff Committee in the 1950's and the Alexander Committee in the late 1960's, these high powered bodies have laid out an evolutionary plan for making full use of the state of the art in communications and electronics to produce an ATC system adequate to the task. The CAA/FAA has paid lip service to these reports and adopted portions of their recommendations,but its overall response has been too little, too late.
In 1962 the FAA came up with a 10-year plan for automation of the ATC system. With a congressional mandate to accomplish the task within 10 years, the FAA plowed ahead, basing the system on IBM 7090 computers—which even IBM protested would be obsolete by the time they were installed and—Raytheon radar displays of a relatively unproven design. Rushed into service under severe time pressure, the system's many hair-raising failures were soon legendary. It became standard procedure to shut down the system entirely whenever the president's Air Force One was flying, relying instead on the obsolete (but more reliable) manual system.
Throughout the 1950s and 1960s air traffic volume grew by leaps and bounds.The number of airports receiving commercial service rose substantially. Yet many of them remained unequipped with such essential safety features as Instrument Landing Systems (ILS) and radar. During the decade from 1962 to 1971 there were 20 fatal accidents involving nonprecision landings (those without proper instruments) at air carrier airports; nearly 700 people died in those accidents. In some cases the airports involved did not yet meet the FAA's arbitrary criterion for the minimum amount of traffic needed to justify an ILS installation. In other cases, the criteria had been met, but FAA budget priorities meant the agency had not yet gotten around to installing an ILS. As of 1971 only 20 percent of the runways of this country's 530 airline airports had complete ILS systems, and only two-thirds of them even had control towers.
An ILS costs several hundred thousand dollars. Until the airport/airways trust fund was set up in 1971, the FAA continually cried poor-mouth as its excuse for not installing more landing aids. Yet in 1968 Rep. Fletcher Thompson of Georgia pointed out that the FAA in a three-year period had spent $3 million on renting airplanes, over and above its own 100-plane fleet. That would have bought many an ILS. In 1971 the FAA's plan to modernize and enlarge its private fleet called for spending $123.5 million—a sum that would have financed ILS installations at over half the remaining unprotected airports.
Skimping on life saving landing aids while enlarging it's private fleet is all too typical of the FAA's priorities. The agency justifies its fleet primarily on the grounds that it is needed for continually checking the performance of the of the thousands of navigation aids (primarily VOR stations) across the country. Yet aviation experts point out that most such devices either work or don't work—their performance characteristics change very little over time (unless some new terrain obstruction were to appear suddenly and and block the VOR signal pattern). Hence, their on-off status can be monitored remotely on the ground. The FAA's monitoring flights consume huge amounts of fuel ($5.6 million worth in 1974), and in many cases interfere with traffic patterns, causing increased increased risk of collisions. But what bureaucracy could resist the prestige of having it's own air force?
Today the FAA is installing a new generation of automation at its en route ATC centers. Called radar data processing (RDA), the new system utilizes narrow band radars interfaced with a sophisticated computer system. The computer generates a comprehensive set of symbols to represent each aircraft on the controller's scope—flight number, altitude, transponder code, etc. The system concept is a good one. Unfortunately, RDP has been put into operation at many centers without adequate testing and debugging—and without backup computers. Controllers' scopes go blank without warning. Symbol generators quit working.
"I wish I had a dollar for every time the radar has quit," says controller Mike Rock. "I have seen radar fail in two or three sectors at one time. This means that approximately 100 planes are in danger of colliding."During a one-month study period in 1974, 11 near-misses that resulted directly from problems with RDP were reported to the Aviation Safety Institute. "The problem for a midair accident of major magnitude will increase unless the FAA takes positive action to correct problems and decelerate implementation of of…RDP," stated ASI's John Galipault in 1975. His view remains the same four years later.
Yet the system the FAA is trying to perfect lags years behind what today's technology permits. Much of the problem of congested airways is created by the old-style navigation system still in use. It consists of thousands of VHF omnidirectional radio transmitters (VOR stations) at known locations. Instruments on board the plane permit a pilot to fly on a "radial" from one VOR to another, thereby getting from City A to City B by Zigzagging from one VOR to the next. Mathematically, knowing the distance and angle from more than one VOR makes it possible to pick any course to fly—not just a radial path from one VOR to the next. What's needed is an on-board computer to make the necessary calculations. This technique, known as area navigation (R-NAV), vastly expands the amount of airspace available to plans on a noninterfering basis. Prototypes of R-NAV equipment have existed for more than 20 years, and operational systems have been available for over a decade. Yet the FAA's ATC system—despite billions spent on automation—is still not equipped to handle large numbers of pilots setting their own courses, off the time-honoured VOR radials.
What have industry observers said of the FAA's performance as manager of the ATC system? In 1970 Robert Hotz, respected editor of Aviation Week, summed up the FAA's progress over the decade of the of the 1960s, finding that an automated ATC system "appears to be as far away as ever from its 1961 goals."Why? For one thing, there is the "technical incompetence and slothful leadership of the FAA and its predecessor agencies" and "the swarms of veteran airways employees who came into the agency in the era of of lighted beacons and low-frequency ranges and have added little technical knowledgeability since then." The FAA bureaucracy, said Hotz, "has frustrated the airlines, exasperated the Congress to the point where funding has been reluctant, virtually driven the avionics industry out of the market, and concentrated primarily on its self-serving preservation."
An investigation that same year by the House Government Activities Subcommittee noted that the "FAA as an organization has more independent empires than medieval Europe.…[it] lacks any feeling of urgency. The FAA simply does not move forward. All too often in the past, progress has been the result of tragedy." And as a political entity, the FAA can avoid accountability for its failures by blaming prior administrations. Between 1961, when the automation program began, and 1970 there were four FAA administrators, "none of whom can be held responsible for the failures of his predecessors," noted Hotz.
The FAA would have us believe that things have changed in the 1970s. Certainly the agency's PR efforts have improved. But in 1975, in the wake of the crash of TWA 514, a 10-man task force appointed by Transportation Secretary Brinegar reached many of the same conclusions Hotz did, albeit phrased in bureaucratese.The FAA's headquarters operation is "large and unwieldy and may serve as a detriment to FAA's performance of its safety mission," they concluded. The agency's rule-making process is "cumbersome and burdened with delays," and there is a "scarcity of top flight technical talent." Its advanced technology program is "relatively immediate and short-term in outlook"—that is, it fails to come to grips with fundamental long-range changes such as implementing R-NAV. And the Brinegar report pointed out that the FAA's dual role—as both operator of the ATC system and safety regulator—requires "completely different technical skills," implicitly raising the question of whether the FAA should continue with air traffic control.
Considering that the executive secretary of the task force was, the acting FAA administrator, and that its members were all from the federal government, the mild tone of its criticisms is understandable. Similar phraseology characterized a 1976 report by the General Accounting Office, which faulted the FAA for serious planning and management difficulties (for example, for not knowing whether planned improvements in the ATC system were actually worth what they would cost).
No such restraint marked the report of another group appointed by the FAA in 1975—the Special Air Safety Advisory Group (SASAG), consisting of six retired airline pilots. Its purpose was to determine what could be done to prevent further crashes like that of Flight 514, especially during the approach and landing phase. After observing 600 flights on 27 airlines and talking to controllers, pilots, and support personnel, SASAG submitted its findings in July 1975. "New hazards have developed which are a product of the very systems designed to make flying safer," they reported. "Attempts to compensate for natural deficiencies sometimes have brought on additional unrealistic rules or procedures, rather than technological improvements."
Commenting on FAA approach procedures, SASAG termed them "hopelessly complicated, impractical, and contradictory," quoting one pilot to the effect that terminal area maneuvering "is a sophisticated game of blind man's bluff, and the only real protection the crew has is common sense." Nonstandard terminology was frequently observed, as in the case of Flight 514. It was found that up-to-date information on hazardous weather was often not provided to flight crews.
But it was the ATC system itself that SASAG found most at fault. The ATC system "has created hazards, slowed traffic. restricted productive flight by all segments of the aviation industry, and used energy in frightening amounts," they found. More specifically:
• "The ATC system has loaded the cockpit with extra work to the extent that, during critical phases of flight, the air traffic function uses most of one crew member's time, removing him as a useful navigational and systems management assistant."
• "The atmosphere in an IPR room is a jumble of confusion, with demands on the human of the highest order."
• "The system is a jumble of people, radar screens, communications lines and stacks of paper strips, people communicating by voice, by radio to impersonal aircraft…while in the control room people are milling about, talking and creating distractions."
• Controllers often show disdain for pilots and evidence "lack of understanding of the pilot's job," creating a "conflict between ground and air that promotes an emotional response in an environment where it cannot be tolerated."
Summing up, SASAG concluded: "The list of Air Traffic System problems is long and detailed, but basically they reduce to these: a system too dependent on the human element; a system that has grown from old concepts with complex fixes applied to it in an attempt to accommodate its inadequacies. This, in turn, has created a monster of procedures, rules, methods, and confusing interplay between people who are separated by distance, technical knowledge, and understanding of each other's problems."
What to do? Obviously not just continue to patch up the present system. "The direction our government is taking toward future air traffic control is not commensurate with available technology. A program of top priority by an independent body is required to create a system that will reduce human error to a minimum." The question is how to go about doing that. SASAG suggested that a study be conducted "to determine whether the air traffic system would be operated more efficiently with advanced technology as an independent public company."
Could air traffic control be provided by a corporation? There is no reason why not, in principle. Air traffic control is a valuable service, as is communications or transportation. It requires the careful and efficient organization of people, equipment, and procedures to carry out the business of keeping aircraft safely separated from one another and assisting them in making safe takeoffs and landings. Removed from political control freed of the civil service mentality, adequately financed, and staffed by competent technical and managerial talent, an ATC corporation could take full advantage of today's (and tomorrow's) technology.
It isn't only pilots, such as the members of SASAG, who have suggested that ATC be transferred from the FAA to an independent corporation. The Professional Air Traffic Controllers Organization (PATCO ) has been advocating just such a change since 1969. If the controllers themselves no longer wish to work for the FAA bureaucracy, with all its civil service job security, that must tell us something!
The plan PATCO endorses was first suggested in 1968 by Glen A. Gilbert, a long-time aviation consultant. Gilbert, who was the Bureau of Air Commerce's very first controller back in 1936, proposes a COMSAT type corporation, funded 50 percent by taxes and 50 percent by user charges. It would be an "independent government corporation," operating out and responsible directly to Congress. As set forth in Gilbert's 1975 study, the U.S. Air Traffic Services Corporation would:
• operate the ATC system and perform various safety functions independent of the executive branch of the federal government;
• manage its own finances, derived from general taxes and user charges, subject only to congressional overview;
• function in a manner comparable to the best private-enterprise management and organizational concepts;
• establish its own allocations of costs to users, to meet one-half of its operating expenses;
• establish its own safety regulations and internal rules, regulations, and procedures.
Gilbert considers establishment of such a corporation essential if the ATC system is to meet the growing needs of aviation in a safe and cost-effective manner. It would also, he believes, establish and maintain a high level of morale, initiative, and productivity on the part of its personnel, be rapidly responsive to constantly changing requirements of airspace users, and provide for direct and continuing decision making participation by the users.
The idea of a COMSAT-like corporation taking over ATC functions most recently surfaced in, of all places, an official FAA document. The study, Aviation Futures to the Year 2000, was prepared for the FAA by The Futures Group, a think tank in Glastonbury, Connecticut. One of the study's five alternative scenarios features a highly automated ATC system,operated by a "COMSAT-like quasi-governmental authority" rather than by the FAA.
SASAG, PATCO, Gilbert and The Futures Group have all correctly identified the problem, and the general thrust of their proposal— put ATC on a businesslike basis cannot be faulted. But why the government involvement? It's true that Gilbert stresses the model of relatively successful COMSAT rather than the more typical quasi-government corporations—Amtrak, Conrail, and, needless to say, the US Postal Service. All three of these are responsible to Congress, and all are far from models of good management and cost-effectiveness. To involve the government in any way is to seriously risk defeating the basic purpose of the change—which is to transfer ATC from a politically controlled bureaucracy to an efficient business.
One possible argument for government involvement is that government aircraft, notably those of the military, use the the ATC system too. Another is that there is some kind of "public benefit" involved in the ATC system apart from benefits to specific users and that general taxpayer funding is therefore appropriate for at least part of the costs. These questions were addressed in depth by Department of Transportation studies during during 1972 and 1973. Economists made a detailed cost-allocation study of the ATC system and of FAA operations generally, as part of an unsuccessful administration campaign to move 100 percent user-charge funding of the existing FAA/ATC system. In brief, the studies concluded that about 50 percent of the costs of the system are occasioned by air carriers, 30 percent by general aviation, and 20 percent by military use. No general benefits for nonusers could be identified.
The DOT studies established quite clearly which users contribute which costs of the system, and a variety of methods of charging them is possible. Military and other governmental users could be charged just like any others and could participate, as customers, in the long term ATC planning process as they do now—perhaps more effectively, since they would be dealing with a far more responsive and competent organization.
The biggest obstacle to setting up a 100 percent user-financed ATC corporation would be political opposition. Whereas commercial aviation is approximately paying its share of ATC costs today (in the form of existing user charges), general aviation is paying only one-seventh of its share. Consequently, its lobby groups—the Aircraft Owners & Pilots Association, the National Business Aircraft Association, and the General Aviation Manufacturers Association—have conducted an extensive, and so far successful,campaign against efforts to increase user charges to full-cost levels. These organizations can be expected to lobby strenuously against a private ATC corporation.
Apart from such political obstacles, could it really work? Could a private corporation actually be trusted with the responsibility for thousands of planes and millions of lives? The fact is, such operations have already been demonstrated—they have worked and they are working, right now. In Berne, Switzerland, stands the headquarters of Radio-Schweiz, A/G, the nonprofit corporation that provides air traffic control services in Switzerland. Radio-Schweiz, though its start-up costs were underwritten by the Swiss government, is fully private and is 100 percent user-charge funded, for both capital and operating expenses. In Mexico in the 1940s a private, nonprofit corporation known as Radio Aeronautica de Mexico,S.A.(RAMSA) was set up jointly by the various Mexican airlines, which became its stockholders. Although the airlines were later nationalized, RAMSA continued as the ATC services company, funded entirely by user charges, until last October when it was nationalized by the Mexican government (the controllers went on strike for three weeks, in protest, but to no avail). A similar company, RACSA,was set up in Cuba after World War II and is still in operation, though Cuban airlines are also nationalized. The British firm of International Aeradio, Ltd. for many years has operated ATC services in parts of the former British empire, paid for by user charges. It currently provides most ATC services in the booming Persian Gulf area. Its I.A.R. Caribbean, Ltd. subsidiary operates in Commonwealth countries in the Caribbean.
The United States itself came close to having an ATC company and even now has a firm in a logical position to provide such services. The company is Aeronautical Radio, Inc.—ARINC. It is a not-for-profit firm owned jointly by the airlines to provide communications services between planes in flight and ground stations and between various airline ground stations. In 1977 ARINC did $85 million worth of business, up six percent over 1976. Its air-to-ground operations include some 2,100 installations, providing voice radio communications. ARINC also provides electronic switching connecting 28 computer systems. This is the system that allows a United Airlines agent, for example, to make a reservation for you on Piedmont as easily as on United. The switching system transmits some 40 million messages per month. The world's largest private-line intercity communications network, serving 135 airline users at over 1,300 locations, is also operated by ARINC. It uses 17,717 leased circuits comprising nearly 5 million voice channel miles.
ARINC actually predates the existence of air traffic control by seven years. In 1929 several airlines held experimental licenses from the Federal Radio Commission (predecessor of the FCC) to carry out air-to-ground radio communications. The licenses permitted experimental operations but granted no rights use the frequencies on a regular basis. Late in 1929 representatives of these airlines met with the FRC in Washington to ask for permanent radio frequency assignments. Unfortunately for them, the military had gotten there ahead of them. A year earlier the FRC had assigned to the military a large portion of the frequency spectrum designated for aviation purposes. As a result, the requests of the individual airlines far exceeded what was then available.
At this point the airlines' legal counsel, Louis Caldwell (himself a former secretary of the FRC) suggested a cooperative approach. Since FRC rules would not permit licensing of frequencies by an informal association, it became necessary to form a corporation. As a result, in December 1929 Aeronautical Radio, Inc. came into existence as a Delaware corporation. The original stockholders were Pacific Air Transport, Western Air Express, and Universal Airlines. The charter provided that the corporation could conduct business in all phases of aeronautical communications and navigational activities. Only airlines could become stockholders, and no one airline or affiliated group could own more than 20 percent of the stock. Users of ARINC services, including private planes, would be charged in proportion to their use at rates sufficient to cover all expenses.
In its early years ARINC was something of a skeletal organization, existing primarily as the FRC/FCC licensee and operating mostly as a coordinating body. Airline personnel installed and operated all the actual communications equipment. ARINC, however, took a leading role in developing improved radio navigation equipment during the 1930s. Operating in the interest of all the airlines, it served as the interface with companies such as Western Electric, Bendix, and RCA that pioneered in developing the hardware. It was ARINC, rather than the government, that led the way in developing such systems as airborne VHF radio, omnidirectional navigation beacons (VOR), and "blind landing" systems (ILS), often overcoming the reluctance and conservatism of the Bureau of Air Commerce and later the CAA.
From 1930 to 1934 ARINC provided the only voice communications between air and ground—en route, during approach, and at airports. When the need for controlling traffic ("fight following") became apparent in the Chicago area in 1934, it was ARINC that led the way in coordinating airline efforts. In 1935 ARINC set up interline agreements among the airlines serving Chicago and Newark to provide for a unified ATC procedure at each airport. The first Airway Traffic Control Center was set up at Newark on December 1, 1935, with costs shared among the airlines in proportion to airport use. Airline personnel trained at Newark for setting up other centers—the second one at Chicago (in April 1936) and the third at Cleveland (in June 1936). Each center controlled airline traffic within 50 miles of the airport, using the airline/ARINC radio operators at each field. The fledgling air traffic control system apparently worked quite well.
Parallel with these developments, however, pressures were building toward a greatly expanded federal role in aviation, pressures that would culminate in the Civil Aeronautics Act of 1938 with its substantial economic assistance to the depression-plagued airlines. As part of this gearing up, the Bureau of Air Commerce asserted federal responsibility to establish "a uniform and centralized system of airway traffic control…to direct and coordinate the progress of all flights, whether government, civil, or commercial, over the Federal Airways."In line with this thinking, on July 6,1936, the bureau took over the three existing traffic control centers and by the end of the year had established five more. The airlines did not protest this take-over; indeed, they welcomed it. At the height of the depression, here was one less item they had to include in their hard-pressed budgets. ARINC, as a "subsidiary" of the airlines, was in no position to object to this loss of its fledgling ATC business. Besides, it had its hands full developing new forms of communications systems.
ARINC was not out of the ATC business altogether, however. First of all, up until 1946, all CAA traffic controllers still communicated with ARINC, which transmitted their messages to the flight crews. Only in the late 1940's and early 1950s did the CAA develop its own ATC radio links, leaving the ARINC channels strictly for company messages. But while losing one ATC role, the late 1940s firmly established ARINC in another.
Until the late 1940s overseas flights still carried radio-telegraph operators, who sent and received Morse Code messages. The CAA was committed to continuation of this obsolete method and built new long-range radio-telegraph stations in San Francisco and New York. When United and Northwest received new routes from the West Coast to Hawaii in 1948, ARINC and the airlines decided the time was at hand to switch to voice communications. ARINC built four high-powered, high-frequency overseas stations—at Seattle, San Francisco, Los Angeles, and Honolulu. Needless to say, voice communications proved to be technically sound, and the change eliminated the need for a telegraph operator in the cockpit—a substantial cost savings to the airlines. Over the next few years this process was repeated in the Caribbean and then on the North Atlantic. The CAA eventually shut down its radio-telegraph stations and now contracts with ARINC for all ATC communications from the five overseas "gateways": New York, Miami, San Juan, San Francisco, and Honolulu.
ARINC also became further involved in ATC overseas during this period of time. It was ARINC that provided the advice and expertise leading to the formation of USA in Mexico and RACSA in Cuba. ARINC also worked with aviation groups in Venezuela, Colombia, Peru, and China to set up similar companies, not always with the same degree of success. Wherever an organization was successfully created (whether private or government), both communications and ATC services were provided, with funding based on user charges. Decades of interfacing with the bureaucratic CAA had taught ARINC valuable lessons about the efficiency of government versus that of the private sector.
Today, ARINC stands in marked contrast to the FAA. It is recognized as a world leader in aeronautical communications. Its specifications for standard avionics equipment are in use throughout the world. It continues to be a leader in advancing the state of the art and has set up a for-profit subsidiary, ARINC Research Corp., that does $10 million a year in applied systems engineering. ARINC, or a company modeled after it, could readily take on the task of operating and upgrading America's ATC system.
The examples of ARINC, RAMSA, Radio Schweiz, and the others demonstrate quite clearly that ATC services can be provided by a fully private company, funded entirely by user charges. The risks of creating another Amtrak or US Postal Service by insisting on government participation in or sponsorship of an ATC corporation are far too great. To ensure the highest degree of safety and cost effectiveness, a truly private-enterprise operation is clearly the only way to fly.
Given the FAA's record with the ATC system, it's not surprising to hear complaints, also, about its competence as a safety enforcer. Aside from its inherent limitations as a large government bureaucracy, what makes the agency ineffective as a safety regulator? Critics point out that the FAA is faced with a built-in conflict of interest. The Federal Aviation Act directs it to "encourage and foster the development of civil aeronautics and air commerce," in addition to promoting safety. The FAA interprets this to mean looking after the economic well-being of the aircraft and airline industries. Yet in the day-to-day businesses of designing planes and running airlines, the potential for conflict between economics and safety is, unfortunately, very real.
The FAA's safety efforts are embodied in its Federal Air Regulations (FARs). The FAA's role is to promulgate, via FARs, the minimum required safety standards, leaving it to manufacturers and operators, respectively, to decide whether and to what extent to exceed the requirements. Any such system is only as good as the adequacy of the standards themselves and the adequacy with which they are enforced. The FAA can be faulted on both counts.
The airworthiness FARs (design standards) frequently lag far behind the state of the art. For 14 years, from 1961 to 1975, the basic FARs remained unchanged, while technology proceeded apace. To patch up the gaps, the FAA issued various Special Conditions on an ad hoc basis; as manufacturers came up with new design features, the FAA thought up new Special Conditions that would accommodate them. The design of the DC-10, for example, required 30 pages of Special Conditions, covering such features as its unique three-bogie main landing gear, its emergency exits, fire-proof materials, flight controls, and propulsion system. The industry tail wags the FAA design standard dog.
As noted earlier, the FAA does not itself carry out most of the detailed work of aircraft certification. On large (transport) aircraft, about 70 percent of the inspection and compliance work is delegated to the manufacturer; for light aircraft there is up to 90 percent delegation. What this means is that certain company personnel wear one hat during regular work (for example, as test pilot) and another hat when acting as the FAA's representative (as FAA certification pilot). The potential conflict of interest is clear.
One of the first public exposures of the inadequacies of the FAA-supervised design process occurred in 1970 when two young aeronautical engineers, associates of Ralph Nader, published a report on light-aircraft crashworthiness. On the basis of their carefully researched study, James Bruce and John Draper concluded that "one-half of all general aviation fatalities…were unnecessary,…the victims of negligence in the field of crash safety engineering." Bruce and Draper noted that, although the human body can withstand loadings of up to 20 times the force of gravity (20g) vertically and 40g laterally, the FAA requires light aircraft cabin structures to be able to withstand only 3g upward, 9g forward, and 1.5g sideward. The standards, of course, are minimums, not maximums, and manufacturers are free to do better.
Indeed, some do exceed the standards. Both Beech and Helio design their planes' cabins using a tubular frame that can withstand high crash loads and configure the wings so that failure is possible only outside the fuselage. All fuel tanks are located in the wings, to minimize the danger of a cabin fire in the event of a crash. The other two major light-plane builders—Cessna and Piper—build their cabins out of sheetmetal, meeting only the minimum FAA crash-load criteria. It's not that they don't know how to build safer airplanes—both build cropdusters, in which crash survivability is an essential sales feature. But nobody mentions comparative safety levels in selling business and pleasure craft. So Cessna and Piper continue on their way, translating a lighter but less safe structure into slightly lower prices or slightly higher performance. The customer takes safety for granted because he knows all planes on the market have received a certificate from the FAA. He therefore assumes them to be equally safe. The light aircraft manufacturer's trade group, GAMA, strongly opposes higher FAA design standards.
Despite Beech's good record on crashworthy design, the company's safety consciousness leaves much to be desired. In 1961 it discovered that during certain maneuvers the fuel intake ports on its Bonanzas could become exposed, sucking in air instead of fuel and causing the engine to fail. For four years the company did nothing, meanwhile bringing to market a new luxury plane, the Baron, using the same fuel system. Finally, in April 1965 it issued a Service Letter (analogous to a Service Bulletin) describing the problem in innocuous terms and stating, "This does not create a hazard but can be disconcerting." Six month later, in October 1965, came the first of a series of Baron crashes, caused by unexpected engine failure due to unporting of the fuel intake.
The CAB investigator (the CAB investigated crashes prior to creation of the NTSB in 1967) test-flew a Baron in February 1966 under minimum-fuel conditions and identified the unporting problem. In January 1957 the CAB recommended an evaluation of the Baron's fuel system design. But the FAA was not impressed. It maintained that no FARs had been violated in the Baron's design and that the unporting problem could only occur during maneuvers that the plane was never intended to perform. Meanwhile, Barons continued to crash.
Finally, in 1968 an FAA test pilot flew a Baron, repeating the earlier CAB pilot's maneuvers. His recommendation: ground the Baron until a design change could be made. But his report never saw the light of day at the FAA. By the end of 1968, after five Baron crashes, the FAA finally issued an AD: it required, not a design change, but a warning placard in the cockpit. Although Beech finally changed the design of all new Baron fuel systems in 1970, it was not until two of the crash cases were litigated—one settled out of court for $325,000 and the second resulting in a jury verdict of $21 million—that Beech in 1972 issued a modification kit to correct the problem on pre-1970 Barons. The FAA did not require the modification, nor did Beech provide it as a routine product improvement; owners who wanted it had to buy and install it at their own expense.
A similar reluctance on the FAA's part has been evident in cases involving commercial transport aircraft. In July 1970 an Air Canada DC-8 crashed on landing at Toronto, killing 109 people. The first officer had accidentally deployed the spoilers while the plane was still 60 feet off the runway, causing it to slam into the pavement. The spoilers are designed to kill the wing's lift once the the aircraft has touched down; it is extremely dangerous to deploy them in flight. Yet The DC-8 design was not fail-safe in this regard; the spoiler handle could easily be moved into the "deploy" position instead of the "arm" position, as had occurred in the Air Canada crash.
In August the manufacturer, Douglas Aircraft, sent a cable to all DC-8 operators warning that spoilers should be used only on the ground; no mention was made of the Air Canada crash. The FAA's western regional office prepared an AD calling for a warning placard to be placed beside the spoiler handle in the cockpit reading "Deployment in Flight Prohibited ." Douglas, Eastern Airlines, and the Air Transport Association protested the AD as unwarranted, but it was nevertheless issued in December 1970. In October 1971 a European DC-8 crashed due to spoiler deployment 15 feet in the air; no lives were lost, though the plane was damaged. In November 1972 a Japanese Airlines DC-8 crashed near Moscow under conditions suggesting in-flight spoiler deployment; 61 people perished. In June 1973 an Icelandic DC-8 crash landed at JFK due to inadvertent spoiler deployment at 40 feet; 38 people were injured and the plane sustained major damage.
In August 1973 the FAA's eastern region drafted a proposed AD to make spoiler activation impossible in the air and sent it to the western region to be issued. The western region rejected it, because Douglas was about to issue a Service Bulletin on its own. The SB was issued—as a routine product improvement, not a safety item—in October 1973. Finally, in January 1974, four years after the first crash, the western region gave in and issued an AD making the Douglas SB mandatory.
The pressure on FAA regional offices to cancel ADs in favor of nonpublicized SBs can be intense, as the Turkish Airlines DC-10 and the DC-8 spoiler cases demonstrate. Captain Brian Power-Waters, a 22-year veteran of airline flying, cites additional recent cases in which proposed ADs were canceled after manufacturer protests:
• Lockheed L-1011 rudder pedals were jamming, causing serious control problems.
• Lockheed L-1011 auxiliary power unit exhaust shroud seal was a potential fire hazard.
• Lockheed L-1011 passenger oxygen masks failed to deploy when needed.
• Boeing 747 escape slides failed to operate normally.
The FAA takes the economic interests of the aircraft manufacturers very seriously. Consequently, one can wonder how objective FAA certification criteria would be even in the absence of the delegation-of-authority system.
The FAA, in enforcing its design standards, gives every appearance of following the classic pattern of a regulatory agency captured by the industry it allegedly regulates. Aircraft and component manufacturers have built up longstanding relationships with FAA personnel and offices. One former FAA official notes that "the smart manufacturers shop different FAA districts to get their hardware certified because each district has a different attitude toward standards." The FAA's 1975 task force acknowledged the existence of this region-shopping problem.
In recent years the inadequacies of the FAA's certification activities have been coming to light. In 1973 the General Accounting Office released the results of its year-long investigation of certification in the light-aircraft industry. The FAA's failure to closely monitor the design process, said the GAO, "has contributed to design safety hazards in light aircraft." Not only are planes with design defects certificated, but once the weaknesses are discovered, the FAA takes inordinately long to correct them, as we have seen. The GAO cited (without naming names) the Beech Baron case as one of several prime examples. It also pointed out that because the FAA does not itself participate in the design and test phase, it cannot influence the design until problems show up in the field, after the plane is in service.
In 1974 the House Special Investigations Subcommittee, headed by Rep. Harley Staggers, issued a massive report documenting the failure of FAA safety efforts. The report cited numerous examples of what it called administrative foot-dragging by the FAA in requiring manufacturers to make changes to improve safety. Its actions to promote air safety "have been unreasonably delayed, or omitted entirely, because of an over solicitous attitude on the part of some within the agency concerning the economic well-being of the aircraft industry or the air carriers." Other members of Congress agreed. Sen. Howard Cannon noted that the handling of the DC-10 cargo door problem "calls into question the entire regulatory relationship" between the FAA and the industry. Sen. Vance Hartke concluded simply, "The FAA is a prisoner of the industry it is supposed to regulate."
After the Turkish DC-10 crash, the GAO conducted another investigation, looking into the matter of "regulatory lag." It documented the extent of the problem, finding, for example, that it took an average of 28 months from the time a safety rule was first proposed until the final version was issued and that the rule would typically have been under study for another year prior to the rule-making process actually beginning. During its one-year study GAO identified 47 NTSB recommendations "most critical to the preservation of life" that the FAA had not acted on and asked the FAA to report their status. It took the agency three months just to figure out their status and reply to the GAO!
Regulatory lag was still a live issue in 1977. After the collision of two 747s at Tenerife, Rep. Elliot Levitas reported finding 135 instances in the previous eight years in which the FAA rejected NTSB recommendations and 281 cases in which it agreed with the board but had not implemented them fully—including 95 dating back three or more years. Economic-hardship claims by the manufacturers and airlines were frequently the major factors in causing the delay, modification, or rejection of the recommendations.
The same issues that apply to aircraft design—inadequate or outmoded FARs, delegation of authority, economics versus safety—also apply to airline operations. Various FARs define minimum qualifications for operating personnel, spell out what are considered to be safe operating conditions, and specify minimum maintenance requirements.
How do these FARs measure up? To begin with, the FAA attempts to cover itself by making rules spelling out every thing possible."We are regulated to death," states Captain Power-Waters. But new FARs don't replace old ones; they're simply added on to an ever growing jumble. As a result, "Most of the regulations in use today are as ancient and useless as the landing aids that were put into operation 30 years ago. The 747s operate under the same regulations originally drafted for the DC-3," says Power-Waters. Many are absurdly deficient on their face. FAR 121.687(b), for example, requires that every flight carry a copy of the "latest available" weather report for each stop. If the weather teletype is out of service and only last week's weather is available, it is perfectly legal to use, so long as it's the latest available! And with the proliferation of regulations, "There isn't a pilot in the sky today who doesn't break at least one regulation every time he flies. It is practically impossible to comply with one regulation without infringing on another."
One indication of the inadequacy of the FARs is provided by the FAA itself. The agency's rules for operating its own fleet are far more stringent than the minimums required for commercial airlines. FAA pilots must have at least 48 hours of pilot time within the prior 12 months to legally fly a plane; there is no such air-time rule for commercial pilots. They need only have made three takeoffs and landings within 90 days; FAA pilots must have made five. All FAA pilots who fly jets over 25,000 feet must obtain physiological training—for example, simulated decompressions in a pressure chamber and other emergency procedures. No airline crews get such training.
Other FARs set forth minimum maintenance standards. These cover both routine, scheduled maintenance and procedures for dealing with in-service failures. Maintenance is carried out by airline personnel; all such work must be at least supervised and signed off by a mechanic licensed by the FAA. The agency maintains a corps of inspectors who spot-check the airlines' work, looking at log books and talking with supervisors. It is impossible, given the FAA's limited personnel, to check out everything. Each of the 300-odd maintenance inspectors is required to check only four aircraft logbooks per month, covering only the previous two weeks. Inspectors work only a regular eight-hour day, five days a week, while airline operations continue round the clock.
The system "works" (more or less) because airlines can't afford to take serious shortcuts, since this would end up costing more than it saved; the overall record-keeping and inspection system provides for tracing failures back to the responsible parties in the event a problem is discovered; and the FAA can penalize violators by fines of up to $1,000 a day. Nevertheless, many observers think the FAA is doing as sloppy job of maintenance supervision. Some airlines are more conscientious about maintenance than others, but this information never becomes public. Mechanics who observe violations of safety regulations have no means of reporting to the FAA (unlike pilots, who—at least during certain years—have been able to report near misses and other incidents to the FAA under an "immunity" program).
The conflict between safety and economics makes itself felt most strongly when it comes to keeping flights on schedule. If certain aircraft components have become inoperative, the key question becomes, Is it safe to fly without item X, or must the flight be delayed while it is repaired or replaced? The Air Transport Association and the FAA have developed for each type of commercial aircraft a Minimum Equipment List. The MEL specifies the equipment that must be operational for a flight to proceed. Obviously, trade-offs must be made in establishing such lists—dollar losses in delayed flights and angry customers must be weighed against the potential of lost lives and destroyed planes. Critics charge that the FAA yields too readily to trade-group pressure in making up MELs that keep flights operating at the expense of safety. Crashes have occurred because flights have been dispatched with equipment that was inoperative but legal according to the MEL.
Further problems occur when the MELs themselves are not strictly adhered to. The International Association of Machinists contended in recent testimony before the House Government Activities and Transportation Subcommittee that "in a lot of instances where scheduled departures (which are money) are involved, the manual and MELs go out the window.…[The FAA inspectors] are rarely, if ever, around. The check and balance system gets out of kilter when an over-eager management-level supervisor ignores or signs off an aircraft problem in order to make a scheduled departure." Captain Power-Waters cites many examples of pilots of certain airlines being disciplined for refusing to depart on schedule due to inoperative equipment.
When violations of FARs do come to light, the FAA too often responds with a mere slap on the wrist. In 1970 Mohawk Airlines received notice from the FAA that they had violated a number of FARs, covering such items as inspection of fire extinguishers, flap actuators, rudders, and navigation system components. Some of the required inspection intervals had been exceeded by hundreds of flight hours, and ILS was two and a half years overdue for inspection. Despite the seriousness of the violations and the FAA's ability to to levy fines of up to $1,000 per day for each violation, it offered to settle for only $50,000. Even then, Mohawk was able to negotiate a settlement of $20,000.
What we have, is a system whose incentives are badly out of whack. The manufacturers and the airlines have an economic incentive to cut corners—not in major, obvious ways, to be sure, but in those borderline areas where cost trade-offs must always be made. And the record makes clear that a number of these trade-offs have led to tragedy. The FAA has no clear-cut incentive to promote higher levels of safety. To begin with, it has the dual mission, by statute, of promoting both safety and aviation's economic well-being. And then, of course, it is a government bureaucracy. That its basic motivation is, for the most part, self- preservation was candidly admitted two years ago by one FAA staffer during NTSB's hearings on the crash of TWA 514. "Safety suffers when everybody is more interested in protecting their tails than trying to make air travel more safe," he said. Why this attitude? "FAA is running from millions of dollars in liability," he continued. "If changes are made, the assumption is made that the changes are an admission of responsibility. Consequently, change is resisted."
The FAA has painted itself into a comer. As the government agency supposedly taking care of air safety, it is assumed to know what it is doing. Insurance companies and the travelling public assume that all airline and aircraft are equally safe, because they all have the FAA's continuing seal of approval. Any sort of consumer consciousness of safety has been driven out of the system. All the while the FAA's outdated, contradictory, minimum standards are exceeded by some, adhered to by others and flouted by still other. The crashes keep occurring. Lives continue to be lost. TWA 514 (92 lives); Turkish Airlines Flight 981 (346 lives); Pan Am and KLM Boeing 747s at Tenerife (581 lives)—and many, many others.The conscientious manufacturers and airlines are indistinguishable from the comer-cutters, in the minds of the public and the insurers, because the FAA has blurred the distinctions and cast its deceptive blessing over them all. We all pay for it—in higher fares thanks to higher insurance premiums, and in lives. Isn't there a better way?
Agreement that the FAA is failing in its safety responsibilities is widespread outside of the agency. The FAA's deficiencies are just too glaring to be ignored. Yet suggested solutions are few, and largely unsatisfying. Activist groups like the Aviation Consumer Action Project and the staffs of various congressional committees all favor "reforming" the FAA in some manner—finding better managers, hiring more competent engineers, removing its mandate to promote commercial success, etc.
But the bottom line of all these suggestions would still leave us with a government bureaucracy. It would still suffer from a civil service mentality. It would still be subject to political control. It would still rely largely on industry personnel. Even if it started out with a clean slate—brand new FARs, completely new, well-qualified staff, a new charter, a bigger budget—it would be only a matter of time before the classic regulatory agency pattern repeated itself and we'd be faced with the same problem all over again.
The fundamental problem is that of incentives. The incentives facing an agency like the FAA ultimately turn out to be those of self-preservation. It preserves itself best by developing smooth working relationships with the industries it regulates. The industries, in turn, are motivated by self-interest—to make profits selling planes and carrying passengers.
The FAA's only motivating tool, for either its own personnel or those of industry, is fear—fear of violating regulations. Yet fear is an ineffective motivator. Modern management has come to rely effectively on participative, rather than authoritarian, means of motivation, rewarding good performance rather than punishing mistakes. The incentive structure needs to be turned around so that both industry and the external entity are motivated by positive goal attainment rather than by fear of noncompliance.
But is this possible? We are so used to thinking about forcing compliance with safety rules that the idea of motivating compliance may seem fanciful at first. But let's suppose, for the sake of argument, that there were no FAA, that government had not been given the role of enforcing air safety. What then?
Studied from this angle, the dimensions of a solution become apparent. If there were no government safety regulation, there might be lots of crashes; that would mean lots of lawsuits against airlines and aircraft producers; and then the insurance industry would be on the hook. Since aviation underwriters would be the ones ultimately accountable for the costs, it is they who would have the greatest economic incentive to make sure flying is safe.
Is there any reason to think the insurance industry would be able and willing to do so? History provides an enlightening example. In the 1860's the fire protection industry was in something of the state aviation was in during 1920s. Fire departments were relatively haphazard outfits, frequently thrown together by well-meaning but untrained volunteers or by small, undercapitalized companies. Fire insurance rates were chaotic, varying greatly from place to place and bearing no apparent relationship to actual loss potential. To bring some order to the industry, representatives of the major fire insurance companies got together in 1866 and organized the National Board of Fire Underwriters.
At its first annual meeting, in 1867, NBFU urged its local affiliates to inspect buildings, fire departments, and water supplies as factors in determining insurance rates. In 1873 the seventh annual meeting recommended that local boards appoint standing committees to examine local fire departments and water supplies "and to endeavor to have needed improvements made therein." Already, a feedback effect was taking place. In 1890 the national board appointed an inspector of fire departments, fire patrols, and water supply. The first such inspector began work on July 1,1890, inspecting 52 cities in his first 11 months on the job.
But obtaining compliance was still a problem. NBFU resolved to work on methods of inducing cities to follow through on the inspector's recommendations, noting that "until they do take measures for improvement, we are able…to take such measures as are consistent with our own safety as underwriters, whether by means of rates, or a refusal to accept certain risks." But no formal system of adjusting rates based on safety levels was put into effect at that time; the primary means of seeking compliance was the personal influence of local insurance agents on local government officials. The inspection reports were, however, distributed to rate-making associations, as well as to local officials.
In 1902 a member of NBFU's Fire Department Committee suggested that a standard "for perfect equipment and water works service" be developed, to serve as a guide for rate making. The National Fire Protection Association was asked to "formulate standards of fire protection in cities and properly guide them so as to form a basis for rating." The resulting document, "Standards for Grading Town Public Fire Protection," was published on October 29, 1903. It established seven classes of fire departments and set down standards for equipment, water supply, and other factors. Over the next 12 years the standards formed the basis for a considerable upgrading of the NBPU's inspection work, and several engineers were added to the staff. Finally, in December 1915 a Standard Schedule for Grading Cities and Towns of the United States was issued to insurance companies and rating bureaus across the country.The results of NBPU's grading of a city were then used by the rate bureaus to assign each city (or portion of a city) to a specific insurance rate class. Thus, the better a city scored, the lower its insurance rates would be.
This basic system, upgraded periodically to reflect new knowledge, technology, and procedures, continues to operate today. Responsibility for conducting the periodic grading of cities was transferred from the NBFU to the American Insurance Association in 1966. In 1971 a nonprofit service agency, the Insurance Services Office (ISO) was created, and responsibility for fire grading became one of its primary duties. Every city and county in the United States graded approximately once every seven years by an ISO engineering team, using the Grading Schedule for Municipal Fire Protection—a direct descendant of the earlier schedules. The Grading Schedule provides for a detailed technical evaluation of a city's water supply, fire department, fire service communications, and fire safety control. Within each category, deficiency points are assigned for each instance in which a particular feature falls short of the ideal set forth in the schedule, and insurance rate bureaus then set fire insurance rates accordingly.
This system provides a direct economic incentive for the city managers to invest in safety. There is no conflict between economics and safety and no need for a regulatory bureaucracy to issue edicts and threaten fines. To be sure, trade-offs are made, as they always will be. There can be no guarantee of absolute safety, under any system, and there will always be some additional safety investments judged not worth what they would cost. But the value of the ISO system is that it provides positive incentives via lower insurance rates for taking reasonable safety precautions. Cities are not forced to follow the ISO's recommendations, but they (their citizens) must pay the price, in higher rates and loss potential, if they don't.
What works for fire insurance could also work for airline property and liability insurance as well as for aircraft product liability insurance. If there were no FAA, the insurance underwriters would have a substantial incentive to follow the fire underwriters' example and get into the safety standards business in a serious way. An Aviation Underwriters Safety Organization (AUSO), similar to ISO, would very likely come into existence, developing standards for aircraft design, maintenance and operation. Airlines and manufacturers would have an incentive to work with AUSO's engineers and inspectors, since doing so would lead to lower insurance rates.
But wouldn't the AUSO in time come to resemble the FAA—a cumbersome bureaucracy that's captured by the industry? Although such an outcome is conceivable, it must be viewed as highly unlikely. To begin with, the AUSO would be free of political control and of civil service rules, so its tendency to operate as a cover-your-tail bureaucracy would be minimized. Second, as a voluntary organization, its standards would not have the force of law and would probably not be minimums—indeed, it's quite possible that several alternative levels of safety would be specified (as with the ISO), among which airlines and manufacturers could choose, with corresponding differences in insurance rates. Because of this, the perceived value in "capturing" an AUSO would be far less than in the case of the FAA. In addition, AUSO's value to the insurance industry would be only as good as its reputation for integrity—as in the case of Underwriters Laboratory. If it began compromising its safety judgments under pressure from airlines, for example, strong corrective pressures would be exerted by the underwriters.
Why hasn't the insurance industry taken on such a role already, given the FAA's deficiencies? To begin with, any direct industry-wide system for linking insurance rates to safety practices would be illegal under present US law! In the late 1960s the major aviation insurance underwriters signed an antitrust consent decree forbidding any form of "collusion" in rate setting. Since that time, the personnel of the various underwriters have bent over backwards to avoid even the appearance of discussing business together. The consent decree does not prohibit exchanging safety information per se—and some amount of information transfer does take place. (Several major underwriters recently joined with the National Safety Council to write a handbook on ground safety at airports.) What's lacking, however, is any attempt by the industry to set rates that reflect adherence to safety standards. Any joint attempt to do so would violate the consent decree.
But what about individual company efforts? Spokesmen for the underwriters say this is simply not feasible under present conditions—which include taxpayers' subsidization of safety, at several billion dollars a year. Under these conditions, insurers' rates are much lower than they would be in a free market—where the insurers would bear the cost of safety-standard setting and enforcement. In fact, according to John Lind, safety director of the US Aviation Insurance Group, rates today are at "rock bottom" levels, making it economically infeasible for any one firm to offer discounts for customers adhering to higher safety standards than the FAA mandates.
In addition to these factors, the industry structure also diminishes the underwriters' incentives to make changes. Because the possible losses in a jumbo jet disaster are so huge, underwriters make extensive use of reinsurance. In other words, a particular insurance policy with large potential risk of loss is in effect divided up among companies that insure the insurers. A reinsurer—like Lloyds of London, Swiss Reinsurance Co., and General Reinsurance Co.—shoulders part of the direct insurer's risk in exchange for a share of the premium. When the Pan Am and KLM 747s collided at Tenerife, for example, Swiss Reinsurance was responsible for 3 percent of Pan Am's loss and 1.5 percent of KLM's. Dividing the risks into such small packages makes reinsurance feasible and helps keep direct insurance premiums low. But at the same time it also reduces the incentive of underwriters to police individual airlines, manufacturers, or governments. That incentive will not exist at all until the insurance industry's subsidies are removed—by getting rid of the FAA.
There would also be other pressures promoting air safety in the absence of the FAA. As noted earlier, the existence of FAA certification casts a kind of security blanket over the entire aviation industry, destroying any real safety consciousness on the part of consumers—light-aircraft purchasers, corporate transportation departments, and individual ticket buyers. This false sense of security has removed safety entirely as a factor in the competitive marketplace, much to the delight of manufacturers and airlines, who since the 1920s and the beginning of commercial aviation have felt the need to sidestep people's fear of flying. Yet safety ought to be a vital factor in consumer decisions—at least as long as the real differences of the kind that exist today among planes and airlines.
Several existing organizations serve as illustrations of the kind of information services that would be in demand in the absence of the FAA. The Aviation Safety Institute in Worthington, Ohio, gathers information on safety-related incidents from pilots and air traffic controllers. These people can call a toll-free hot line and report, anonymously, on near-misses, inoperative instrument landing systems, radar-scope failures, etc. The ASI feeds these data into a computer for later analysis and publishes a fortnightly newsletter warning of unsafe conditions (see box). Only a handful of people subscribe to the newsletter today, but in the absence of the FAA, such a newsletter would be marketable and highly in demand, especially by private and corporate pilots.
The Aviation Consumer Action Project is one of a variety of Ralph Nader-sponsored consumer activist organizations in Washington. It conducts studies of the ways in which the government is or is not serving the interests of aviation consumers, as opposed to those of airlines and manufacturers.Needless to say, it has a very fertile field to explore. In the absence of the FAA, an organization like ACAP could do a great deal to keep public awareness focused on safety differences among various airlines, thereby providing consumer pressure aimed at improvements. Safety would become a marketing angle, much as steak dinners and flying pubs are today.
The FAA's failure, in both air traffic control and air safety, is no mere academic issue. It is literally a matter of life and death. And it's an issue of great urgency in 1979. Congress recently passed a deregulation law substantially relaxing the CAB's economic control of the airlines. Over the next decade there will be a continuing increase in both the number of planes and the number of airlines in operation. More small airfields will get scheduled service. And one thing is patently clear. The FAA cannot cope with such expansion.
In testimony before the House Government Operations Subcommittee last year, John Galipault of the Aviation Safety Institute stated: "We believe that the FAA has progressively taken itself out of the role of leader in aviation safety.…We do not have confidence in the FAA's ability to develop responsive, meaningful, and valid requirements for the new 'deregulated carrier fleet.' Not only will it bog down in this effort, but given that the new 'fleet' is defined and can function, the FAA will not have the means for performing certification, surveillance, and quality assurance that it rightly should be doing."
Congress, in short, must bite the bullet. It has finally seen fit to move toward restoring a free market in air transportation. But this new free market must not be shackled with an inadequate, out-of-date air traffic control system or a safety enforcement mechanism that doesn't work. The FAA should be given a decent burial and the private marketplace allowed to take over its functions. An ATC corporation and an Aviation Underwriters Safety Organization can do the job—if only they're allowed to come into being. The alternative is increasing chaos and carnage in the skies.
Editor Robert Poole, Jr. holds two engineering degrees from MIT and spent seven years in the aerospace/defense industry. This article was researched under a grant from the Cato Institute, but all views expressed are solely those of the author. The author also wishes to express his appreciation to Aeronautical Radio, Inc., the American Insurance Association, and the Library of the National Air Space Museum at the Smithsonian—all of which gave him access to their archives for research purposes.
Where Not to Fly If you fly a lot, your chances of being involved in a midair collision or having some sort of malfunction occur are a lot greater on some routes and at some airports than at others. But the FAA won't tell you that. They want to preserve your naive trust in government protection.
That leaves the job to small, under-funded outfits like the nonprofit Aviation Safety Institute In Worthington, Ohio. The ASI operates a 24-hour-a-day toll-free hot line by which pilots and air traffic controllers can report near-misses, safety violations, and equipment failures—anonymously, without fear of FAA or company reprisals. The ASI stores the data in its computer and sends out notices of particular hazards in its biweekly newsletter, Monitor.
REASON asked ASI president John Galipault to prepare a few computer printouts showing comparative levels of hazard. The results, tabulated here, are quite revealing. At the nine busy airports from which ASI receives regular reports, the number of near-misses in 1977 ranged from three to eight. When these numbers are adjusted for relative traffic levels, the rate of near-miss occurrence ranges from a low of 7.89 per million operations (takeoffs and landings) at Dallas–Ft. Worth Airport (DFW) to a high of 30.3 at New Orleans (MSY). These data are not comprehensive, since they are based on voluntary self-reporting by FAA and airline employees. Most likely, actual rates of near-misses are much higher. And data from some of the country's busiest airports—LAX and JFK among them—were not available, since ASI receives no regular reports from them. But travelers would do well to avoid New Orleans and Philadelphia, in comparison with the other airports listed here.
The ASI printouts also cover 10 of the FAA's air route traffic control centers, which monitor and control flights between cities. Here, too, the variation is substantial, ranging from a low of 0.56 near-misses per million IFR (Instrument flight) operations controlled by the Cleveland center (ZOB) to a high of 9.04 at Jacksonville center (ZJX)—16 times as high! Even more shocking is the rate at which "system errors"—human mistakes or equipment failures—occur at ZJX. The Jacksonville center's system error rate was a whopping 23.83, vastly higher than second-place Chicago center (ZAU) at 3.49 or third-place Oakland (ZOA) at 3.37. The Jacksonville center controls a stretch of the very busy New York–Miami route from south Georgia to north Florida. So if your winter vacation includes this route, maybe you'd be better off taking the train. Information of this kind, allowing travelers to assess their own risks, is carefully suppressed by the FAA. If air traffic control and air safety enforcement were provided privately, there would likely be quite a market for such data. What multinational corporation would route their $800,000-a-year chief executive along the Jacksonville center's route when they could find an alternative, safer routing? The ASI incident-reporting program can be seen as prototypical of the kind of information the market would supply, once safety became a marketing factor. Relative malfunction rates of Sabreliners versus Learjets versus Jetstars, engine failure rates of 747s versus DC-10s and L-1011s, radar failure rates at various traffic control centers—all would be exposed to public scrutiny rather than hidden away In the bowels of the FAA.
This article originally appeared in print under the headline "Is This Any Way to Run an Airway?."