Nearly 58,000 Applied. Only 2% Made It: Inside the Controller Shortage the FAA Can't Fix

By Jim Kerr, AviatorDB

AviatorDB analyzed 94,000 NTSB accident records, 149,000 NASA safety reports, and FAA certification data to investigate why the air traffic control system is losing more controllers than it can replace — and what the federal safety data shows about the consequences.

---

On any given day, 45,000 aircraft move through the national airspace of the United States. In fiscal year 2024, the FAA managed 16.8 million flights — a number that has grown 10% in the past decade alone. Every one of those flights depends on a voice on the other end of a radio frequency: an air traffic controller, sequencing departures, separating arrivals, and keeping metal apart in a system that operates with zero margin for error.

The FAA currently employs 13,164 controllers. The agency's own collaborative staffing target puts the deficit at 3,544 certified controllers below what is needed. More than 40% of the FAA's 290 terminal facilities are understaffed. Nineteen of the 30 largest facilities — the ones that handle 27% of all commercial operations — are staffed below 85% of their targets. Those 19 facilities account for 40% of all flight delays in the national airspace system.

The shortage is not new. In 1981, before the PATCO strike, approximately 17,500 controllers managed 5 to 6 million flights per year. President Reagan fired 11,345 striking controllers on August 5, 1981. The FAA claimed it would take two years to rebuild the workforce. It took eleven. The system has never fully recovered. Today, 25% fewer controllers handle three times the traffic. Each controller in 2025 manages roughly 1,178 flights per year — compared to 315 in 1981. That is 3.7 times more traffic per person, with fewer people, in a system that was already strained before the first picket sign went up.

The Pipeline Few Survive

In fiscal year 2022, 57,956 people applied to become air traffic controllers. Cumulatively, roughly 200,000 have applied since 2020. Of those, fewer than 10% met the initial qualification requirements. Only 4% of applicants with no prior ATC experience who applied between 2017 and 2022 completed the hiring process and entered the FAA Academy in Oklahoma City.

The Academy can train approximately 1,800 students per year — a physical bottleneck of simulators, instructors, and classroom space. Of those who enter, 30% wash out. In fiscal year 2017, the failure rate peaked at 41%. The survivors are then assigned to one of 313 FAA facilities for on-the-job training that lasts two to four years. At that stage, only 61% certify as full controllers. At the most demanding En Route Centers — the facilities that manage high-altitude traffic between airports — the certification rate drops to 46%. Fewer than half make it.

The Government Accountability Office confirmed the math in a December 2025 report: only about 2% of all applicants ultimately become Certified Professional Controllers. The arithmetic is straightforward — 4% reach the Academy, 70% of those graduate, 70% of those certify at their facility. The total time from application to certification averages 5 to 5.5 years.

In 2025, a 43-day government shutdown forced the Academy to suspend training. Approximately 500 trainees were lost — not failed, not washed out, but simply removed from a pipeline that was already hemorrhaging.

At New York TRACON, designated N90 and one of the busiest air traffic facilities on Earth, the staffing picture is a case study in collapse. The facility is operating at 54 to 58% of its authorized staffing level, with 113 certified controllers against a target of 195. Two-thirds of trainees assigned to N90 wash out — a 67% training failure rate. Only 8 operational supervisors remain against 30 authorized positions. In July 2024, the FAA transferred responsibility for Newark's airspace from N90 to Philadelphia, not because Philadelphia was better staffed, but because New York could no longer handle its own.

A System That Penalizes the People Trying to Help

In 2014, the FAA replaced its validated Air Traffic Selection and Training aptitude test with a Biographical Questionnaire — a change intended to widen the hiring pool for a profession that was already struggling to attract candidates. The new test carried a 90% failure rate. Weeks after the change, a supervisor in the National Black Coalition of Federal Aviation Employees was recorded on audio sharing questionnaire answers with prospective applicants. NBCFAE officials conducted workshops walking candidates through the correct responses. From outside the system, it looked like cheating. From inside a pipeline that was rejecting 98% of applicants for a profession facing a critical shortage, it looked like someone trying to get more controllers into a system that desperately needed them.

The change eliminated the Collegiate Training Initiative pathway — a network of FAA-accredited college programs that had trained candidates specifically for ATC careers. Graduates who had invested years and significant tuition preparing for the profession were forced to compete on the same biographical questionnaire as applicants with no aviation background. A class-action lawsuit, Rojas v. FAA, reached the Ninth Circuit in 2019.

The Department of Transportation Inspector General investigated and cleared the supervisor in 2016, concluding the findings did not warrant federal prosecution. In March 2025, Transportation Secretary Sean Duffy reopened the investigation. No one has been fired. No certifications have been revoked. The pipeline remains broken, and the people who tried to work around it face fewer consequences than the system that broke it.

Who Is Guarding the Sky

In 1973, a landmark psychological study examined the effects of chronic occupational stress on a single profession: air traffic controllers. That study gave the English language a word that has since entered everyday use — burnout. More than fifty years later, the profession that defined burnout is still being consumed by it.

A December 2024 study by Southern Illinois University Carbondale surveyed active controllers and found that approximately 20% suffer from moderate to severe anxiety — four times the rate in the general population. More than 10% showed signs of major depression. The suicide rate among controllers is roughly 30 per 100,000, three times the national average. Between 2014 and 2023, 22 FAA controllers died by suicide. Seventy percent were veterans with more than 20 years of service. FAA data from 2022 identified chronic fatigue from 10-hour shifts as the primary factor in 75% of controller suicides.

Controllers who recognize they need help face a system designed to punish them for asking. The FAA's own Mental Health Aviation Rulemaking Committee, in its April 2024 final report, identified the core problem: the process disincentivizes honesty due to the often-immediate grounding and removal from duty until recertification. A controller who discloses depression or anxiety is pulled from the position, moved to administrative duties at lower pay, and enters a recertification process with no guaranteed timeline. The Mental Health in Aviation Act of 2025 passed the House in September, directing the FAA to modernize its approach. The Senate has not acted.

In 2011, NASA completed the most comprehensive fatigue study ever conducted on air traffic controllers — surveying 3,268 controllers and monitoring more than 200 at 30 facilities across the country. The findings were alarming. Controllers averaged 5.8 hours of sleep per day during work weeks. Seventy-six percent of controller schedules produced chronic tiredness that created, in NASA's language, pressure to fall asleep. Approximately 20% of controllers reported committing a significant error — a loss of separation, a missed handoff, a wrong clearance — in the previous year. More than half attributed those errors directly to fatigue. Among controllers working six-day schedules, the error rate exceeded 30%.

The FAA suppressed the study for nearly four years, releasing it only in August 2015. The National Transportation Safety Board had called the FAA's response to controller fatigue unacceptable in 1989 — twenty-two years before NASA proved the Board right.

Today, 41% of FAA facilities require controllers to work six-day weeks at least once per month. Several require it every week. Controllers average 400 to 500 hours of overtime annually. The standard rotation — known as the Rattler — compresses two day shifts, two evening shifts, and a midnight shift into four calendar days, with as little as eight hours between shifts including commute time. Controllers accept it for the three-day weekend. The science says it is dangerous.

Morale across the ATC workforce is at what NATCA, the controllers' union, describes as historic lows.

It Is Not for Lack of Trying

The FAA has met or exceeded its controller hiring targets every year since 2023 — 1,512 in fiscal year 2023, 1,811 in 2024, and 2,026 in 2025. The Academy set a record in July 2025 with 550 students training simultaneously. The hiring process has been trimmed from eight steps to five, saving four months. Trainee salaries were increased 30%, to $47,026 per year. A new retention program offers retirement-eligible controllers a 20% lump-sum bonus for each year they continue working. An enhanced Collegiate Training Initiative, launched in October 2024, allows graduates from nine accredited universities to skip the Academy entirely and proceed directly to facility training.

None of it is enough.

In fiscal year 2023, despite meeting the hiring target, the FAA achieved a net gain of 15 certified controllers. In 2024, with a record 1,811 hires, the net gain was 108. The deficit has held stubbornly between 3,500 and 3,800 certified controllers every year — unmoved by the hiring surges. The average annual attrition over the past decade has been approximately 1,330 controllers — through retirement, training failure, resignation, and removal — while average annual hiring over the same period was 1,210. The FAA hired only two-thirds of the controllers it projected needing between 2013 and 2023.

The National Academies of Sciences, Engineering, and Medicine projects that under current plans, the system will gain fewer than 200 net additional controllers by 2032. Former FAA Administrator Mike Whitaker stated publicly that air traffic control towers will never reach full staffing under the current system.

Although compensation for controllers working overtime at the busiest facilities can reach $200,000 to $250,000 per year, the profession competes for talent against airline cockpits — where pilots earn an average of $280,570 per year with federally mandated duty-time limits that cap flight time at 1,000 hours annually and require minimum rest between assignments. Both pilots and controllers hold the lives of passengers in their hands. One profession has legal protections against overwork. The other averages 400 to 500 hours of mandatory overtime per year on top of a full schedule, with no equivalent safeguard.

The crisis is not confined to the United States. The International Federation of Air Traffic Controllers' Associations projects a global shortfall of 71,000 controllers by 2034. Nav Canada has been forced to periodically close towers at Kelowna and Winnipeg, delaying more than 300,000 passengers since April 2024. European ATC delays have cost airlines and passengers an estimated 16.1 billion euros over the past decade. No country has solved the controller staffing problem. Every country is getting worse.

A Compounding Problem

This is not a new crisis. It is a compounding one — and AviatorDB's analysis of 94,000 NTSB accident records shows a pattern of controller-linked incidents stretching back decades, growing more dangerous as staffing thins.

AviatorDB identified 5,108 accidents in which air traffic control was cited in the narrative or probable cause. Of those, 159 were fatal. The data reaches back to the earliest digital NTSB records and forward to the present, and the message is consistent: when controllers are overloaded, undertrained, or simply not there, people die.

On February 1, 1991, USAir Flight 1493 landed on top of SkyWest Flight 5569 at Los Angeles International Airport, killing 34 people (DCA91MA018). The NTSB's probable cause cited the failure of the Los Angeles air traffic facility management to implement procedures that provided redundancy. The controller was working an overloaded position. That accident prompted staffing reforms at LAX. It did not prompt systemic change.

The same pattern repeated. On August 27, 2006, Comair Flight 5191 attempted to take off from the wrong runway at Lexington, Kentucky, killing 49 people (DCA06MA064). A single controller was on duty — the other was on break. On August 8, 2009, a sightseeing helicopter and a Piper PA-32 collided over the Hudson River near Hoboken, New Jersey, killing nine (ERA09MA447). The Teterboro Airport controller had been on a personal phone call and failed to provide a traffic advisory. On July 17, 2022, two aircraft collided while maneuvering to land at North Las Vegas Airport, killing four (ERA22FA318). The NTSB cited the controller's failure to provide timely and adequate traffic information to either aircraft.

Then came the near-misses — a cascade of them. In January 2023, an American Airlines Boeing 777 crossed an active runway at JFK while a Delta Boeing 737 was taking off. In February 2023 at Austin, a FedEx Boeing 767 and a Southwest Boeing 737 came within 150 feet of each other on the same runway. In April 2024, also at JFK, a SWISS Airbus was cleared for takeoff while a ground controller simultaneously cleared four other jets to cross the same runway. In fiscal year 2023, the FAA recorded 503 significant air traffic control operational incidents — a 65% increase over the prior year. Runway incursions reached 1,760, with 23 classified in the most serious category, up from 16 the year before.

On January 30, 2025, PSA Airlines Flight 5342 collided with a U.S. Army Black Hawk helicopter on approach to Reagan National Airport in Washington, D.C. (DCA25MA108). Sixty-seven people were killed. The controller on duty was working a combined position — simultaneously managing helicopter routes and fixed-wing aircraft during peak traffic. It was a staffing configuration that exists at Reagan National and facilities across the country because there are not enough controllers to staff each position independently.

Reagan National had already logged two additional NTSB-investigated incidents in the months that followed — a runway incursion involving an American Airlines Airbus in May 2024, and a loss of separation between a USAF formation and a Delta Airbus in March 2025. The pattern at a single airport mirrored the pattern across the system.

The confidential reports that controllers and pilots file with NASA's Aviation Safety Reporting System tell the story from inside the towers. All of the following were filed in 2025.

A controller at the Los Angeles Center, working a closing shift with all five sectors combined to two people and 25 aircraft inbound to Las Vegas, wrote: I turned to my coworker and told him I was at a loss, I have no idea what to do with all these aircraft, there is literally nowhere to put them. What I experienced last night was incredibly dangerous, and I do not say that lightly.

A tower controller in Colorado, forced to work all positions simultaneously, warned: Eventually a controller that is over saturated from working too many positions is going to make another mistake the CIC won't catch and someone will die just so we could have a standalone CIC.

A TRACON controller reported in April 2025: The level of volume and complexity was too great for this configuration, but staffing forced us to work it combined. We are being pressured to work unsafe levels of traffic.

During a government shutdown in November 2025, a controller at Miami Center — already short-staffed before the furloughs — nearly caused a midair collision and filed a report that began: I am honestly shook up about the possibility of having a mid air collision.

A trainee at Boston Center, in July 2025, summed up the state of the system: The NAS is slowly starting to crack.

Please Do Something

After the January 2025 crash at Reagan National that killed 67 people, a pilot at LaGuardia Airport filed a confidential safety report with NASA in August 2025. The report described a pattern of controllers pushing operational limits at one of the busiest airports in the country. It ended with a warning that reads less like a safety filing and more like a plea:

The pace of operations is building in LGA. The controllers are pushing the line. On thunderstorm days, LGA is starting to feel like DCA did before the accident there. Please do something.

In 1981, the United States had 17,500 air traffic controllers managing 5 million flights. In 2025, it has 13,164 controllers managing 16.8 million. The pipeline that is supposed to replace them rejects 98% of applicants, takes half a decade to certify the rest, and produces a net gain that barely registers against the tide of retirements, washouts, and resignations. The controllers who remain are working six-day weeks on six hours of sleep, developing anxiety at four times the national rate, and dying by suicide at three times the national rate. The FAA has tried hiring surges, pay raises, streamlined processes, retention bonuses, and training reforms. The math does not work. It has not worked since 1981, and every year, 16.8 million flights depend on the people left standing.

The National Airspace System is not slowly starting to crack. It has been cracking for 45 years. The question is no longer whether it will break. It is how many more people will be on board when it does.