Air Traffic Control: FAA Needs to Better Prepare for Impending	 
Wave of Controller Attrition (14-JUN-02, GAO-02-591).		 
                                                                 
Thousands of the Federal Aviation Administration (FAA)		 
controllers will soon be eligible to retire because of extensive 
hiring in the 1980's to replace striking air traffic controllers.
Although the exact number and timing of the controllers'	 
departures has not been determined, attrition scenarios developed
by both FAA and GAO indicate that the total attrition will grow  
substantially in both the short and long term. As a result, FAA  
will likely need to hire thousands of air traffic controllers in 
the next decade to met increasing traffic demands and to address 
the anticipated attrition of experienced controllers,		 
predominately because of retirement. FAA has yet to developed a  
comprehensive human capital workforce strategy to address its	 
impending controller needs. Rather, FAA's strategy for replacing 
controllers is generally to hire new controllers only when	 
current, experienced controllers leave. This does not take into  
account the potential increases in future hiring and the time	 
necessary to train replacements. In addition, there is		 
uncertainty about the ability of FAA's new aptitude test to	 
identify the best controller candidates. Further, FAA has not	 
addressed the resources that may be needed at its training	 
academy. Finally, exemptions to the age-56 separation rules raise
safety and equity issues.					 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-02-591 					        
    ACCNO:   A03625						        
  TITLE:     Air Traffic Control: FAA Needs to Better Prepare for     
Impending Wave of Controller Attrition				 
     DATE:   06/14/2002 
  SUBJECT:   Air traffic controllers				 
	     Attrition rates					 
	     Education or training				 
	     Employee retirement plans				 
	     Hiring policies					 
	     Human resources utilization			 
	     Labor force					 
	     Strategic planning 				 
	     Civil Service Retirement System			 
	     Federal Employee Retirement System 		 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-02-591
     
A

Report to the Chairman and Ranking Democratic Member of the Subcommittee on
Aviation, House Committee on Transportation and Infrastructure

June 2002 AIR TRAFFIC CONTROL FAA Needs to Better Prepare for Impending Wave
of Controller Attrition

GAO- 02- 591

Letter 1 Executive Summary 2

Purpose 2 Background 2 Results in Brief 4 Principal Findings 5
Recommendations for Executive Action 7 Agency Comments and GAO?s Evaluation
7

Chapter 1 9

Introduction Air Traffic Controllers? Responsibilities Vary by Facility and

Position 9 Staffing Levels Negotiated between FAA and Controllers? Union 12
Special Requirements Affect the Hiring and Retirement of Air Traffic

Controllers 14 FAA Relies on a Variety of Sources for Air Traffic Controller

Candidates 16 Objectives, Scope, and Methodology 18

Chapter 2 21

FAA Is Facing FAA Estimates It Will Need to Increase Controller Staffing
Levels

and Will Increasingly Lose Many Controller Specialists 21 Increased
Controller

GAO?s Analysis Indicates that Sizable Controller Attrition Is Hiring because
of

Likely 23 Higher Staffing Levels and Growing Attrition

Chapter 3 33

FAA Needs a More FAA?s Hiring Process Does Not Adequately Ensure that
Qualified

Controllers Will Be Available When Necessary 34 Comprehensive

FAA Developed Screening Test to Help Identify Potential Candidates Workforce
Plan for Air

Most Likely to Succeed 36 Traffic Controllers

Challenges Exist in Addressing Academy and On- the- Job Training Resources
and Equipment Needs 39 Exemptions to the Age 56 Separation Provision Raise
Safety and Equity Concerns 40

Conclusions 43

Recommendations for Executive Action 44 Agency Comments and GAO?s Evaluation
45

Appendixes

Appendix I: Potential Impacts of Proposed Changes to Increase Air Traffic
Control Annuity Calculations 47 Proposed Bill Would Increase Annuities 47
Proposed Bill Would Create Financial Impacts 48

Appendix II: Air Traffic Controller Schools 50

Appendix III: Retirement Eligibility Methodology and Analysis 52

Appendix IV: Methodology for Computer Simulation 53 Analysis of Separation
Trends 53 Estimation Methodology of Future FAA Controller Separations 53
Limitations 54

Appendix V: Methodology for GAO?s Survey of Air Traffic Controllers?
Retirement and Attrition Plans 56 Study Population 56 Sample Design 56
Survey Development 57 Survey Administration 58 Estimates 58 Sampling Error
58 Nonsampling Error 59

Appendix VI: GAO Survey of Air Traffic Controllers 60

Appendix VII: GAO Contacts and Staff Acknowledgments 64 GAO Contacts 64
Staff Acknowledgments 64

Tables Table 1: Numbers and Types of Air Traffic Controllers 12 Table 2:
Retirement Eligibility Requirements for Controllers 15

Table 3: FAA?s Controller Specialist Staffing Needs through 2010, by Fiscal
Year 22 Table 4: FAA?s 10- year Estimate of Controller Specialist Losses, by

Fiscal Year 23 Table 5: Sources of New Controllers, Fiscal Years 1997- 2001
36 Table 6: Current and Proposed Annuity Calculations 47 Table 7: Retirement
Annuity Calculations under Current CSRS and

S. 871 48

Table 8: Current Capacities of Air Traffic Controller Schools, as of
November 2001 51 Table 9: Survey Sample Size and Disposition 57

Figures Figure 1: Air Traffic Control System 11 Figure 2: Regional
Controller Specialist Allocations, Fiscal Year

2001 13 Figure 3: Air Traffic Controllers Becoming Eligible for Retirement

in Each Fiscal Year 25 Figure 4: Past and Simulated Air Traffic Controller
Attrition, by

Fiscal Year 26 Figure 5: Survey Estimates: Past and Estimated Air Traffic

Controller Attrition 27 Figure 6: Past and Projected Retirement Eligibility
for Supervisory

Air Traffic Controllers 28 Figure 7: Survey Estimates: Past and Estimated
Air Traffic

Controller Attrition for Supervisory Air Traffic Controllers 29 Figure 8:
Controllers Becoming Eligible for Retirement by Fiscal

Year for En Route Centers, Ten Busiest Towers, and Ten Busiest TRACONs 31
Figure 9: Locations of Air Traffic Controller Schools 50

Abbreviations

ATCS Air Traffic Control Specialists CPMIS Consolidated Personnel Management
Information System CSRS Civil Service Retirement System DOD Department of
Defense DOT Department of Transportation FAA Federal Aviation Administration
FERS Federal Employee Retirement System GAO General Accounting Office NATCA
National Air Traffic Controllers Association TRACON Terminal Radar Approach
Control OPM Office of Personnel Management

Executive Summary Purpose The Federal Aviation Administration (FAA) is
responsible for managing the

nation?s air transportation system so that the 200, 000 aircraft taking off
and landing each day can safely and efficiently carry more than 700 million
passengers per year. Because of the significant hiring in the early 1980s to
replace strikers who had been fired, many thousands of FAA?s controllers
will soon become eligible to retire, potentially leaving FAA with too few
fully trained controllers. Because of these concerns, the chairman and
ranking democratic member

of the Subcommittee on Aviation, House Committee on Transportation and
Infrastructure, asked GAO to (1) identify likely future attrition scenarios
for FAA?s controller workforce and (2) examine FAA?s strategy for responding
to its short- and long- term staffing needs, including how it plans to
address the challenges it may face.

To identify likely future attrition scenarios, we (1) reviewed FAA?s 10-
year hiring plan and associated attrition forecasts for approximately 15,000
controller specialists who actively control and separate traffic in the air
and on the ground; (2) analyzed FAA?s workforce database to determine when
the current controllers (those at FAA as of June 30, 2001) would become
eligible to retire; (3) developed a computer model to predict future
attrition based on historic levels; and (4) developed and administered a
survey to a statistically representative sample of controllers so as to
obtain information on when they might leave FAA. 1 GAO?s analysis covers
over 20, 000 controllers- the 15, 000 controller specialists whom FAA
analyzed, plus about 5,000 controllers who supervise and manage the air
traffic

control system. GAO included the additional personnel because attrition from
these positions is generally filled from the controller specialist ranks
and, thus, omitting them would understate potential attrition among all
controllers. In addition, among other things, we contacted all FAA regional
offices, the 14 colleges or universities that have controller training

programs, and the branches of the military so as to identify and discuss
various aspects of workforce planning for air traffic controllers.

Background In 1981, thousands of air traffic controllers who participated in
a nationwide strike were fired and barred by a presidential directive from

1 For this report, ?attrition? refers to controllers who leave FAA for a
variety of reasons, including retirement, removal for cause, death, or
disability.

reemployment with FAA as air traffic controllers. As a result of the strike,
FAA was forced to hire, over a 3 to 4 year period, thousands of new air
traffic controllers and to rebuild its controller workforce. FAA currently
employs over 20, 000 employees who manage the air traffic control system.
Most of these (about 15,000) are air traffic control specialists who are
responsible for controlling the take- off, landing, and ground movement of
planes. In addition, there are traffic management coordinators (about 670),
who control the flow of air traffic; front- line supervisors (about 1,900),
who work in various facilities around the country; and managers or staff
(about 2,370), who oversee and administer the air traffic control program.
Under a 1998 collective bargaining agreement with the National Air Traffic
Controllers Association (NATCA),

the union that represents the air traffic control specialists, controller
specialist staffing levels were set at 15, 000 for fiscal years 1999 through
2001 and are authorized to grow to 15,606 controllers by the end of fiscal

year 2003. FAA hires new air traffic controller candidates from several
different sources. Most candidates with no prior experience come from one of
14 post- secondary educational institutions that train new controllers for
FAA.

Once hired by FAA, most of these candidates attend a 12- week training
program at FAA?s academy in Oklahoma City and receive an average of 2 to 4
years of on- the- job training at field facilities to become certified
professional controllers. Most candidates with prior experience come from
either the Department of Defense or the pool of fired controllers who were

allowed to return to FAA beginning in 1993. Air traffic controllers are
covered under either the Civil Service Retirement System (CSRS) or the
Federal Employee Retirement System (FERS), depending upon when they were
hired by FAA. Under either retirement program, controllers are subject to
special requirements that allow them to

retire at an earlier age and with fewer years of service than most federal
employees. In addition, Congress directed that air traffic controllers are
subject to mandatory separation from controlling live air traffic at age 56
because of safety concerns, but there are exemptions to this requirement. 2

2 Although most controllers are required to stop controlling live traffic at
age 56, they can continue working at FAA in other positions.

Results in Brief Although the exact number and timing of the controllers?
departures is impossible to determine, attrition scenarios developed by both
FAA and

GAO indicate that the total attrition will grow substantially in the short
and long terms. As a result, FAA will likely need to hire thousands of air
traffic controllers in the next decade to meet increasing traffic demands
and to address the anticipated attrition of experienced controllers,
predominately because of retirement. For example, the results of GAO?s
survey of controllers indicate that approximately 5, 000 controllers may
leave in the next 5 years, a figure that is more than two times higher than
that for the

past 5 years. GAO also found that the potential for retirement among
frontline supervisors and controllers at some of FAA?s busiest facilities is
high.

FAA has not developed a comprehensive human capital workforce strategy to
address its impending controller needs. Rather, FAA?s strategy for replacing
controllers is generally to hire new controllers only when current,
experienced controllers leave. GAO?s review identified challenges that FAA
will face in trying to ensure that well- qualified new controllers are
available when needed. For example, FAA?s hiring process does not adequately
take into account the potential increases in future hiring and the time
necessary to fully train replacements. In addition, there is uncertainty
about the ability of FAA?s new aptitude test to identify the best controller

candidates. Further, FAA has not addressed the resources that may be needed
at its training academy and for providing on- the- job training at its
control facilities in order to handle the large influx of new controllers
and to ensure that FAA?s controller workforce will continue to have the
knowledge, skills, and abilities necessary to perform its critical mission.
Finally, exemptions to the age- 56 separation rules raise safety and equity
issues that FAA has not assessed.

GAO recently published a model of human capital management that highlights
the critical success factors FAA can use to manage its human capital more
strategically to accomplish its mission. 3 Along these lines,

GAO makes recommendations intended to help FAA meet its impending need to
hire and train thousands of air traffic controllers. In commenting on a
draft of this report, senior FAA officials indicated that the report was
generally accurate. The officials also commented that they would look at

3 A Model of Strategic Human Capital Management, GAO- 02- 373SP (Washington,
D. C.: Mar. 15, 2002).

GAO?s human capital management model to determine its applicability to air
traffic controller specialist staffing, and that FAA would consider GAO?s
recommendations in its planning.

Principal Findings FAA Will Likely Be Faced

FAA will likely need to hire increasing numbers of controllers over the next
with Hiring Increasing

decade to meet increasing traffic demands and to address the anticipated
Numbers of Controllers

attrition of experienced controllers. FAA estimates that by 2010, it will
need about 2,000 more controllers than are presently employed to handle
future increases in air traffic. In addition, many air traffic controllers
currently employed by FAA will likely leave their positions within the next
decade.

FAA estimates that by 2010, about 7,000 controller specialists, nearly 50
percent of those currently employed, will leave. The largest part of this
exodus will come from retirements, with FAA estimating that it will
experience retirements of controller specialists at a level three times
higher than that experienced over the 5- year period from 1996 through 2000.
GAO

analyzed aspects of FAA?s controller workforce in addition to controller
specialists and found that even more controllers might soon leave their
current positions than FAA estimates. For example, GAO?s analysis of
personnel data for over 20,000 of FAA?s controllers shows a scenario where
about 2,500 of FAA?s current controllers were eligible to retire as of

September 30, 2001, and nearly 14,000 controllers (or 70 percent of the
current controllers) will become eligible to retire by the end of fiscal
year 2011. In addition, GAO?s model for potential controller attrition
indicates that, on average, about 600 to 800 controllers will leave FAA
employment, primarily though retirement, in each of the next 10 years.
Further, results from GAO?s survey of air traffic controllers indicated that
many controllers are currently planning on leaving (predominately because of
retirement) in the near future. Of the approximately 20,000 controllers now
working at FAA, GAO estimates on the basis of its survey that approximately
5,000 controllers plan to leave by the end of fiscal year 2006. 4 This
includes an estimated 51 percent of the controllers who plan to retire when
they first

4 All estimates based on GAO?s survey of air traffic controllers are subject
to sampling error. Unless otherwise noted, 95 percent confidence intervals
for percentage estimates are +/- 5 percentage points or less, and numerical
estimates other than percentages have confidence intervals of +/- 10 percent
or less the value of the estimate.

become eligible to do so. Many potential retirees currently hold key
positions as supervisors, work in some of FAA?s busiest facilities, or both.
GAO found that about 93 percent of current supervisors will reach retirement
eligibility by the end of fiscal year 2011. In addition, by the end of that
year, FAA?s busiest centers will potentially face a significant turnover in
their current controller workforce, as about 65 percent of current
controllers in en route centers become eligible to retire by the end of
fiscal year 2011.

A Comprehensive An effective human capital process anticipates expected
attrition and

Workforce Strategy Could includes the development of a comprehensive
workforce plan that (1)

Better Prepare FAA for establishes an effective approach for hiring
individuals with the requisite

Upcoming Controller skills and abilities in time to accomplish agency
missions; (2) provides new

employees with the best training opportunities possible to maximize their
Attrition

potential; and (3) uses opportunities to retain qualified staff. FAA has not
developed such a comprehensive workforce strategy to address all of the
challenges it faces in responding to its impending need for thousands of new
air traffic controllers, thus increasing the risk that FAA will not have
enough qualified controllers when necessary to meet air traffic demands. GAO
identified several challenges in FAA?s approach to hiring, recruiting, and
training new candidates and to retaining existing ones, and these are

not fully addressed in FAA?s plans. For example, FAA?s process of generally
hiring replacements only after a current controller leaves does not
adequately take into consideration the time it takes to train a replacement
to become a fully certified controller- up to 5 years, which might result in
gaps of coverage or increased overtime. In addition, FAA?s proposal to rely
more heavily on candidates who have no previous experience (so- called off-
the- street hires) may result in additional challenges. Because it takes a

certain type of person to become an effective controller and FAA has
experienced failure rates at its training academy of as high as 50 percent,
FAA developed a screening test to help select better potential candidates.
However, recent changes have been made to the screening test to allow
additional candidates to pass it. As a result, the effectiveness of the
screening test in identifying successful candidates has yet to be

determined. FAA plans to implement the revised exam in June 2002 and, if
funding is available, plans to evaluate the success of the exam in
identifying successful candidates. A further challenge exists at FAA?s
training academy, where staff have identified equipment and personnel
requirements that will

need to be addressed to effectively handle the expected influx of new
candidates. Finally, with the rehiring of controllers who went on strike in
1981, FAA is faced with having increasing numbers of employees

controlling air traffic who are past age 56- most of the 733 rehired
controllers are exempt from this mandatory separation age. FAA has not
assessed the potential safety and equity issues associated with exempting
these or other controllers from the mandatory separation age.

Recommendations for To better respond to the challenges presented by the
need to hire

Executive Action thousands of new controller candidates, GAO recommends that
the

secretary of transportation direct the administrator of the Federal Aviation
Administration to develop a comprehensive workforce plan that includes
strategies for (1) identifying the timing of hiring necessary to ensure that
facilities have appropriate numbers of certified controllers available to
provide adequate coverage; (2) evaluating the newly developed screening test
to determine whether it is identifying the most successful candidates; (3)
addressing the capacity challenges associated with the training academy and
on- the- job training programs; and (4) assessing the potential safety and
equity issues associated with exempting potentially large

numbers of controllers from the mandatory separation age requirement. Agency
Comments and We provided FAA with a draft of this report for its review and
comment. GAO?s Evaluation

Senior FAA officials found that the report was generally accurate and
indicated that they would consider GAO?s recommendations in its workforce
planning.

Overall, FAA?s comments stressed that FAA has a working human capital
workforce strategy model that has enabled the agency to meet its staffing
goals over the past few years. FAA officials agreed that the potential for
sizable future attrition, in the range of 600- 800 controllers per year, is
likely over the next decade. The officials said, however, that although they
have

plans that extend to 2010, the uncertainty surrounding the future, as well
as labor contracts and budget constraints, limit their specific workforce
planning for air traffic controllers to fiscal years 2002 through 2004. With
general agreement between FAA and GAO that attrition will grow substantially
over the next decade, GAO believes that the workforce challenges FAA faces
exist well beyond fiscal year 2004. As such, GAO believes that sound
workforce planning demands that FAA develop a strategic vision that includes
a workable, long- term plan to meet staffing needs.

Regarding GAO?s concern about FAA?s preparedness for the future, the
officials remarked that FAA?s ability to meet its past goals is an
indication of its ability to meet future needs, and that there is nothing to
indicate that its successful performance will not continue in the future.
GAO recognizes that FAA has been able to meet its recent staffing goals.
However, the recent workforce climate for FAA could be significantly
different from that which it will face over the next decade. The report
highlights the workforce challenges, particularly the sizable anticipated
increases in controller attrition, that are likely over the next decade, and
identifies challenges in FAA?s planning that will make it difficult for FAA
to maintain its past performance. In particular, the report points out the
potential skills gap

that FAA could face in the future because its current hiring process does
not ensure that fully qualified controllers are available to replace
experienced controllers when they leave.

The officials also commented that FAA has long planned for an operational
evaluation of the new screening exam. The officials indicated that they are
currently considering two options for evaluating the effectiveness of the

exam, and that they need to decide on the appropriate option and develop an
implementation and funding plan. However, the officials noted that continued
funding for the ongoing research could not be assured. In response to this
comment, GAO revised the text of the report to recognize FAA?s efforts and
plans regarding evaluation of the new screening exam, and modified its
recommendation to clarify that the evaluation is needed as part of a
comprehensive workforce plan.

In addition, the FAA officials provided technical comments that GAO
incorporated, as appropriate.

Chapt er 1

Introduction The Federal Aviation Administration is responsible for managing
the national airspace system and ensuring the safe and efficient movement of
air traffic. In doing so, FAA controls the take- off and landing of nearly
200,000 planes per day, which carry over 700 million passengers per year. To
accomplish this mission, FAA must have a sufficient number of adequately
trained air traffic controllers working at its air traffic control
facilities.

In 1981 over 11,000 air traffic controllers went on strike and were
subsequently fired by President Ronald Reagan. Between 1982 and 1990, FAA
hired thousands of individuals to permanently replace the fired controllers.
Most of this hiring took place between 1982 and 1986. Many of these
controllers, as well as those controllers who did not participate in the
strike, are now eligible or will soon be eligible to retire from FAA.

Air Traffic Controllers? Air traffic controllers play a critical role in the
nation?s air transportation

Responsibilities Vary system by helping ensure the safe, orderly, and
expeditious flow of air

traffic in the air and on the ground. Controllers help ensure that aircraft
by Facility and Position maintain a safe distance between one another and
that each aircraft is on proper course to its destination.

Specific controller responsibilities for managing air traffic vary according
to the type of air traffic control facility. For instance, controllers who
work at airport control towers are responsible for ensuring the safe
separation of aircraft on the ground and in flight in the vicinity of
airports, generally within a 5- mile radius. These controllers manage the
flow of aircraft during take- off and landing and coordinate the transfer of
aircraft with adjacent control facilities as aircraft enter or leave an
airport?s airspace. Controllers working at terminal radar approach control
(TRACON) facilities use radar screens to track planes and manage the arrival
and departure of aircraft within a 5- to 50- nautical mile radius of
airports. At these TRACON facilities, a key function of an approach
controller is to line up and sequence airplanes as they descend into an
airport?s 5- mile radius. Controllers working at air route traffic control
centers (commonly called

en route centers) manage aircraft beyond a 50- nautical mile radius. These
controllers assign aircraft to specific routes and altitudes while they fly
along federal airways. These controllers also coordinate the transfer of

aircraft control with adjacent en route or terminal facilities. 5 The
typical en route center is responsible for more than 100,000 square miles of
airspace, which generally extends over several states. Figure 1 shows how
controllers working at the different air traffic control facilities track
aircraft during ground movements, take- off, in- flight, and landing
operations. Currently, FAA operates 339 air traffic control facilities,

consisting of 24 en route centers and 315 terminal facilities. 5 There are
24 en route centers, which include 3 center en route radar approach
facilities- facilities that combine center and TRACON operations. Terminal
facilities can include both a TRACON and a tower, which FAA categorizes as
one facility.

Figure 1: Air Traffic Control System Airport tower

TRACON En route center

En route center

En route center

TRACON Airport

Local and ground control

tower

Departure and approach control Cruising altitude

Source: GAO presentation of FAA information.

In total, about 20, 000 employees categorized as air traffic controllers
directly control and manage the air traffic system, comprising several
positions with differing responsibilities. 6 (See table 1.) This total
includes positions that actively control, or supervise the control of,
traffic (air traffic control specialists, traffic management coordinators,
and operational supervisors); and ?off- line? positions that do not control
traffic (former air traffic control specialists in management, training, or
staff positions). Table 1: Numbers and Types of Air Traffic Controllers

Number of Air traffic control position Duties and responsibilities employees

Air traffic controller specialists (ATCS) Controls and manages the
separation of air traffic in designated 15, 120 airspace or on the ground at
airports. Traffic management coordinators Controls the flow of air traffic
by determining how many planes

670 should be in designated airspace at once. Can order that planes be held
on the ground and special routings. Operations supervisors (first- line)
Provides general supervision of the controllers on duty, including 1,862

monitoring and managing the flow of traffic and distributing work among
controllers. Other (management, staff specialists, and so

Provides support services such as training, management, and 2,369

forth) administration in an air traffic control facility.

Total 20, 021

Source: GAO?s analysis of FAA?s personnel database, as of June 30, 2001.

Staffing Levels As the table above indicates, the majority of air traffic
controllers are

Negotiated between classified as specialists. These controllers are
represented by the National

Air Traffic Controllers Association, which negotiated staffing levels with
FAA and Controllers?

FAA in 1998. Under the terms of the agreement, nationwide staffing (in
fulltime Union

equivalents) for these specialists was set at 15, 000 for fiscal years 1999
through 2001. The agreement also called for 2 percent staff increases for
fiscal years 2002 and 2003, arriving at a controller specialist staffing
level of

6 The Office of Personnel Management classifies civilian air traffic
controllers in the FAA as occupational series 2152- civilian air traffic
controllers. In addition to these 20,021 employees, there are about 2,800
flight service station controllers who do not directly control or manage air
traffic but provide pilot briefing, weather reports, and emergency services
to pilots before and during flights.

15,606 by the end of fiscal year 2003. FAA has requested funding to meet the
staffing levels called for in the agreement. Under the 1998 agreement, FAA
headquarters officials and NATCA national representatives negotiate
allocation of staffing levels for the air traffic control specialists among
FAA?s nine regions. Figure 2 below shows the location of each FAA region and
the number of controller specialists allocated to each region for fiscal
year 2001.

Figure 2: Regional Controller Specialist Allocations, Fiscal Year 2001
Northwest Mountain

New England 1,350

573 Great Lakes

2,937 Central

750 Eastern

2,118 Western Pacific

2,025 Southwest

2,028 Southern

3,000 Alaskan

219

Source: GAO presentation of FAA data.

Once the regions receive their staff allocations, FAA regional managers and
NATCA regional representatives negotiate staff allocations among the various
field facilities in each region. The additional 606 controllers called for
under the 1998 agreement are to be distributed to regions and field

facilities in the same way, with FAA and NATCA officials negotiating
allocations to each region and specific facility.

Special Requirements In 1972, Congress passed Public Law 92- 297, which
authorized the

Affect the Hiring and secretary of transportation to set a maximum entry age
for initial

appointments to air traffic controller positions at the FAA. Pursuant to
this Retirement of Air

authority, FAA requires that a potential controller candidate be hired
before Traffic Controllers

reaching his or her 31 st birthday. This provision was established in
recognition of the fact that younger trainees are more successful in
completing the controller training programs, and that younger individuals
may be better able to deal with the stress of controlling air traffic. One
exception to this rule is the Employment of Retired Military Air Traffic
Controllers Program, commonly known as the Phoenix Controller- 20 program,
under which FAA commits to hiring retired military controllers who are past
the age of 30. This exception allows military controllers to stay with the
military longer before moving to FAA to continue their controller
activities.

Controller retirement is also affected by special requirements. Controllers
working at FAA?s air traffic control facilities and staff offices are
eligible to retire under two sets of retirement provisions: the general
retirement requirements for federal employees and special requirements for
controllers. Depending on when a controller was hired, he or she is covered
by either the Civil Service Retirement System or the Federal Employee
Retirement System. As federal employees, controllers under these systems can
retire if they meet certain age and years- of- service requirements. For
example, under general CSRS, a controller who is 55 years old can retire
after 30 years of federal service, or at 60 years old with 20 years of
service, or at 62 with 5 years of service.

Under the special controller retirement requirements, a controller may
retire earlier than under the general CSRS and FERS requirements if he or
she has enough service time as an active controller specialist, traffic
management coordinator, or immediate supervisor. Time in these ?covered?
positions is generally known as ?good time? because it counts toward the
special retirement requirements. Controllers can retire at age 50 if they
have spent at least 20 years in a covered position, or at any age if they
have at least 25 years in a covered position. Under these provisions,
controllers covered by CSRS are guaranteed a retirement annuity

amounting to the greater of two figures: either 50 percent of their high
average 3- year salary or the basic federal retirement annuity. 7
Controllers covered by FERS receive an annuity amounting to 1.7 percent of
their high average 3- year salary for the first 20 years of service plus 1
percent of their high average 3- year salary for each additional year of
service.

Table 2 summarizes the CSRS, FERS, and special retirement provisions.

Table 2: Retirement Eligibility Requirements for Controllers Type of
retirement Age Years of service

CSRS (applicable for most federal employees 62 5 hired before 1984) 60 20

55 30 FERS (applicable for most federal employees hired 62 5 in or after
1984) 60 20

55 to 57 a 30 Special controller retirement under either CSRS or 50 20 FERS
(service time must be in a ?good time? position) Any age 25

a Basic retirement eligibility under FERS is subject to a minimum retirement
age that varies depending on the birth date of the employee. Source: GAO
presentation of Office of Personnel Management information.

In addition to these basic retirement eligibility requirements, air traffic
controllers covered by CSRS are also subject, pursuant to Public Law 92-
297, to a rule requiring mandatory separation at age 56. Controllers

7 There is currently a Senate bill, S. 871, that proposes to increase CSRS
annuity levels. See appendix I for a discussion of the impacts of this
proposal.

covered by FERS are subject to a similar rule, pursuant to Public Law 99-
335. Under this requirement, with some exceptions, controllers actively
working in covered positions must separate by the last day of the month in
which they turn 56. 8

FAA Relies on a Variety FAA relies on a number of sources to fill its
controller positions. These

of Sources for Air sources are (1) individuals with no prior controller
training or work

experience in the air traffic control environment, (2) individuals with some
Traffic Controller

controller training but generally no actual controller work experience, and
Candidates

(3) individuals with prior controller work experience. The first group
includes individuals who respond to an Office of Personnel Management
vacancy announcement. Referred to as off- the- street hires, these
candidates must pass an OPM exam to qualify for employment with FAA and must
pass a 15- week initial training program at FAA?s Academy in Oklahoma City,
Oklahoma, before being assigned to a facility. 9 There have been no OPM job
announcements for entry- level air traffic control

specialist positions since 1992, because FAA has chosen to rely on other
sources for new candidates. FAA estimates that approximately 150 people who
responded to the last announcement and passed the OPM exam are still
eligible for employment as controllers. 10

The second group includes graduates of FAA- accredited collegiate programs
who receive initial air traffic control training prior to being hired by
FAA. This type of training introduces students to the terminology, airspace
configurations, and technical skills necessary to manage air traffic

and operate equipment. Students can receive general air traffic control
training at one of 13 schools under FAA?s collegiate training initiative
program, or specialized en route training at the Minneapolis Community

8 Most controllers in FAA are subject to this rule. However, controllers
appointed by the Department of Transportation prior to May 16, 1972, and
controllers appointed to the Department of Defense prior to September 12,
1980, are exempted. In addition, controllers covered under FERS can work
until they accrue 20 years of good time regardless of age, and CSRS
controllers can work past the age of 56 if they have not qualified for a
retirement annuity. 9 The FAA Academy in Oklahoma City provides management
and technical training to

controllers, inspectors, and other FAA personnel. 10 Over the past 5 years,
FAA has hired on average fewer than 10 candidates from the 1992 OPM list
each year.

and Technical College, formerly known as the Mid- American Aviation Resource
Consortium school (see app. II for more detailed information on the
schools). Collegiate training initiative schools offer either two- or
fouryear aviation related degrees. Unlike these schools, the Minneapolis
Community and Technical College program is not part of a broader academic
program, and the federal government subsidizes the cost of training its
students. Collegiate training initiative graduates must pass an initial 12-
week controller training program at the FAA academy to begin work at their
assigned facility, while Minneapolis Community and Technical College
graduates can immediately begin working at their assigned facilities. During
fiscal years 1997 through 2001, FAA has hired 465 from the collegiate
training programs and 291 from the Minneapolis Community and Technical
College. 11

The third group of candidates consists of controllers with previous air
traffic control experience, including both former Department of Defense
(DOD) controllers and controllers fired in the 1981 strike. DOD employs both
active- duty military controllers and civilian controllers. In general,
military controllers can leave DOD for FAA at the end of their enlistments,

as long as they do so before turning 31 years of age. 12 To help DOD
minimize military controller losses, FAA and DOD designed a program in 1999
called the Phoenix Controller- 20 program to give controllers an incentive
to stay in the military past age 30. Under this program, military
controllers can join FAA after they retire from military service. FAA may
also hire controllers who previously held air traffic controller positions
with the agency; most of them are among those fired in the 1981 controller
strike. President Reagan banned the federal government from hiring any of
these controllers, but President Bill Clinton lifted this ban in 1993, at
which

time FAA issued a job announcement for fired controllers interested in
returning to work. Candidates in this group are not required to attend
initial controller training at the academy but may be required to take
refresher training there. During fiscal years 1997 through 2001, FAA hired
793 former DOD controllers and rehired 562 controllers who had been fired in
1981. 11 FAA regional officials supplied us with hiring information for each
fiscal year since 1997.

Fiscal year 2001 reflects partial year data. 12 FAA policy pursuant to
Public Law 92- 297 prohibits hiring controllers after they have turned 31
years of age. Exceptions to this rule include retired military controllers,
civilian controllers whom DOD had hired prior to their 31st birthdays, and
re- hired former controllers, such as those fired in the 1981 strike.

Once assigned to an air traffic control facility, candidates are classified
as ?developmental controllers? until they complete all requirements to be
certified for all of the air traffic control positions within a defined area
of a given facility. It generally takes new controllers who have had only
initial controller training between 2 and 4 years- depending on the facility
and the availability of facility staff or contractors to provide on- the-
job training- to complete all the certification requirements to become
certified professional controllers. 13 It normally takes individuals who
have prior controller experience less time to become fully certified.

Objectives, Scope, and In October 2000, the chairman and ranking democratic
member of the

Methodology Subcommittee on Aviation, House Committee on Transportation and

Infrastructure, asked us to examine FAA?s efforts to address existing and
future controller staffing needs. We were asked to (1) identify likely
future attrition scenarios for FAA?s controller workforce and (2) examine
FAA?s strategy for responding to its short- and long- term staffing needs,
including

how it plans to address the challenges it may face. To identify future
attrition scenarios for FAA?s controller workforce, we (1) obtained and
analyzed FAA estimates of future retirement and attrition; (2) analyzed
FAA?s employee database to determine when controllers would reach retirement
eligibility; (3) developed a computer model to simulate

future attrition based on historic FAA air traffic controller rates; and (4)
developed and mailed a survey to a sample of current air traffic controllers
to determine their retirement plans.

FAA?s estimates: To obtain FAA?s estimates of future retirements and
attrition, we interviewed officials in FAA?s Office of Air Traffic Resource
Management who are responsible for managing the controller workforce. These
officials provided information on the data used to support FAA?s estimates
of future controller attrition. They provided estimates only for the 15, 000
controller specialists; similar estimates were not available for other
categories of air traffic controllers.

Analysis of FAA?s workforce: We used personnel data supplied by FAA to
calculate the age and service characteristics of 20, 021 air traffic
controllers who were employed as of June 30, 2001, the most recent data
available at

13 In some of FAA?s busiest and most complex air traffic control facilities
it can take up to 5 years to become a certified professional controller.

that time. These included 15, 120 controller specialists, 670 traffic
management coordinators, 1,862 operational supervisors, and 2,369 managers
and staff specialists. We used this information to determine the number of
controllers reaching retirement eligibility over the next decade. Additional
information on how we made these projections is contained in appendix III.

Simulation model of attrition: We developed a computer simulation that
projected the level of potential controller attrition through 2011. This
model used age and years of service information for the controller
workforce, in addition to past attrition rates and some assumptions about
future attrition rates, to estimate the number of future losses FAA will
face in its controller workforce. Additional information on the methodology
of the computer simulation, including the assumptions we used, is given in
appendix IV.

Survey of controllers: We mailed a survey to controllers to obtain
independent estimates of future controller attrition. After developing and
pre- testing the survey, we sent it to a statistically representative sample
of

2,100 current controllers. The survey asked the controllers about when they
planned to retire or leave the agency and about factors that could affect
their decision. We received responses from over 75 percent of our sample.
Additional information on the survey methodology can be found in appendix V.
To address the second objective of examining FAA?s strategy for responding

to its short- and long- term staffing needs, including how it plans to
address the challenges it may face, we obtained information on the
availability of potential controller candidates, FAA?s process for hiring
new controller candidates, and FAA?s training activities associated with new
candidates. To obtain information on the availability of candidates, we
interviewed

officials at FAA headquarters, the 9 FAA regional offices, the 14 college or
university air traffic control programs, and the Department of Defense to
determine the number of controllers who are potentially available to FAA.

We visited schools in California, Minnesota, New Hampshire, and Florida to
better understand their activities. We did not verify the information
provided by these sources.

To understand FAA?s process for hiring new controller candidates, we
interviewed officials at FAA?s headquarters and regional offices. At FAA?s
headquarters we focused on the activities of the Air Traffic Resource

Management office, which is responsible for monitoring air traffic
controller hiring levels. In addition, we met with officials at FAA?s Civil
Aeronautical Medical Institute to discuss their activities to develop a new
screening test for potential controller candidates- referred to as Air
Traffic Selection and Training exam (AT- SAT). In addition, we obtained
information on how FAA uses staffing standards to determine staffing levels
at its various facilities and interviewed officials with the National
Academy of Sciences about their review of FAA?s staffing standards.

To obtain information on FAA?s training activities, we visited FAA?s
training academy in Oklahoma City, Oklahoma, and discussed on- the- job
training with each of FAA?s nine regional offices.

We also interviewed officials with the Air Transport Association, NATCA, and
representatives of all nine FAA regional offices to ensure that we obtained
a nationwide perspective on controller staffing issues. Finally, we obtained
and reviewed information from the Office of Personnel Management and our
previous reports on good human capital practices in government agencies to
evaluate FAA?s workforce plan regarding air traffic controller staffing. We
conducted our review from January 2001 through April 2002, in

accordance with generally accepted government auditing standards. We
obtained oral comments on a draft of this report from senior FAA officials,
which are discussed at the end of chapter 3.

FAA Is Facing Increased Controller Hiring because of Higher Staffing Levels
and

Chapt er 2

Growing Attrition Although the exact number and timing of the controllers?
departure is impossible to determine, attrition scenarios developed by both
FAA and GAO indicate that the total attrition will grow substantially in the
short and long terms. As a result, FAA will likely need to hire thousands of
air traffic controllers in the next decade to meet increasing traffic
demands and to address the anticipated attrition of experienced controllers,
predominately created by retirements. Depending on the scenario, total
attrition could range from 7, 200 to nearly 11,000 controllers over the next
decade. GAO also found that the potential for retirement among frontline
supervisors and controllers at some of FAA?s busiest facilities may be high.

To identify likely future attrition scenarios, we (1) reviewed FAA?s 10-
year hiring plan and associated attrition forecasts for approximately 15,000
controller specialists who actively control and separate traffic in the air
and on the ground; (2) analyzed FAA?s workforce database to determine when
the current controllers (those at FAA as of June 30, 2001) would become
eligible to retire; (3) developed a computer model to predict future
attrition based on historic levels; and (4) developed and administered a
survey to a statistically representative sample of controllers so as to
obtain information on when they might leave FAA. 14 GAO?s analysis covers
more than 20,000 controllers- the 15,000 controller specialists whom FAA

analyzed and about 5,000 controllers who supervise and manage the air
traffic control system. GAO included the additional personnel because
attrition from these positions is generally filled from the controller
specialist ranks and, thus, omitting them would understate potential
attrition among all controllers.

FAA Estimates It Will In May 2001 FAA prepared a 10- year estimate of its
hiring needs that

Need to Increase included a projection of the number of controller
specialists who may be

needed in the future and estimates of expected controller losses. The
Controller Staffing

estimate shows that the number of controller specialists needed to help
Levels and Will

manage the air traffic system could grow from about 15,000 in fiscal year
Increasingly Lose Many

2001 to over 17,000 by the end of fiscal year 2010, and that losses of
controllers could increase from 428 in fiscal year 2001 to over 1,000 in
2010. Controller Specialists

14 In this report, ?attrition? refers to controllers who leave FAA for a
variety of reasons, including retirement, removal for cause, death, or
disability.

FAA Estimates It Will Need FAA estimates that future air traffic increases
will require it to hire more

about 2,000 More Controller than 2, 000 additional air traffic controllers
over the next decade. FAA bases

Specialists its future projections on a mathematical model, referred to as
the staffing

standard, which factors expected traffic levels and the amount of tasks a
typical controller can perform in a given time frame in order to estimate
the future number of controllers that FAA will need. As shown in table 3,
FAA

anticipates a growing requirement for controller specialists.

Table 3: FAA?s Controller Specialist Staffing Needs through 2010, by Fiscal
Year 2002 2003 2004 2005 2006 2007 2008 2009 2010

Total controllers 15, 300 a 15,606 a 15,906 16,139 16,363 16, 599 16, 836
17, 072 17, 309 required a Staffing levels for fiscal years 2002 and 2003
were negotiated between FAA and NATCA.

Source: GAO?s presentation of FAA data.

FAA?s controller staffing levels in fiscal years 2002 and 2003 were
established under the terms of FAA?s 1998 contract with NATCA, which
represents the controller specialists. To estimate staffing needs for fiscal
year 2004 and beyond, FAA used its air traffic control staffing standards.
The standards estimate that FAA will need, on average, about 245 additional
controllers each year from the end of fiscal year 2003 though fiscal year
2010, mainly because of increases in air traffic. 15 The standards further
estimate that FAA will need 17,309 controllers by fiscal year 2010- over 2,
000 more controllers than are currently employed.

The National Academy of Sciences examined FAA?s staffing standards in 1997.
16 It found that the standards did a reasonable job of estimating future
needs on a national or regional level, but that the standards were not as
useful in determining facility level needs. It recommended that FAA modify
its staffing process to produce more reliable facility staffing estimates.
To date, however, FAA has not fully implemented this recommendation 15 FAA
estimates that the events of September 11, 2001, will cause decreases in air
traffic through 2003. However, FAA predicts that air traffic will then
rebound and steadily increase,

creating a need for additional air traffic controllers. Future controller
staffing levels will be negotiated between FAA and the union.

16 Air Traffic Control Facilities: Improving Methods to Determine Staffing
Requirements,

National Academy of Sciences, 1997.

because of funding limitations, according to the branch manager, Resource
Management. FAA Estimates that Future

FAA?s projections show growing losses of controller specialists. FAA
Controller Losses Will Grow

included estimates of three types of losses: retirements, nonretirements
(for example, resignations, firings, and deaths), and non- attrition
(controllers who leave to take other positions within FAA, such as
supervisory and staff positions). According to the branch manager, Resource
Management, the forecast is based on historic attrition levels. Table 4
displays FAA?s 10- year projections.

Table 4: FAA?s 10- year Estimate of Controller Specialist Losses, by Fiscal
Year 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Total

ATCS retirement 153 202 246 294 335 423 569 620 666 719 4, 227 Non-
retirement 104 106 108 110 111 113 115 116 118 119 1, 120 Non- attrition 171
174 178 181 184 187 189 192 195 197 1, 848

Total estimated 428 482 532 585 630 722 873 928 978 1, 036 7, 195 losses
Source: GAO?s presentation of FAA data.

As the table shows, FAA is estimating sizable increases in controller
specialist retirements over the next decade, with retirements increasing
each year and exceeding 700 by the end of fiscal year 2010. The average
annual retirement level over the length of the forecast period is 423, which

is three times higher than the average annual retirement level of 141 that
FAA experienced over the 5- year period of 1996 through 2000. Combined with
other losses, this estimate anticipates a nearly 50- percent turnover in the
next decade from its current controller specialist contingent of
approximately 15,000.

GAO?s Analysis The scenarios shown by our analysis of retirement eligibility
trends, the

Indicates that Sizable results of our simulation model, and estimates from
our controller survey all indicate that FAA may face a sizable increase in
future attrition,

Controller Attrition Is primarily because of retirements. In addition, we
examined attrition

Likely patterns for supervisors and for controllers at the busiest
facilities because

of their importance to the national air traffic control operations, and we

found that attrition levels for these groups could be sizable over the next
decade.

Number of Employees Because many controllers were hired in the early 1980s,
FAA is facing an

Eligible to Retire Increases aging controller workforce. As of June 30,
2001, the average age of an FAA

Rapidly controller was 43, and approximately 7,400 controllers were 45 or
older. In

addition, because of the special controller retirement provisions, many
controllers may soon accrue enough years of service to meet the retirement
eligibility requirements. Because FAA?s employee database does not identify
the amount of time controllers have worked controlling traffic (good time),
we examined the eligibility of FAA?s entire controller workforce (about
20,000 employees), using both the special controller retirement provisions
and the CSRS/ FERS retirement provisions. 17 Although most of the employees
would be expected to first reach eligibility

under the special provisions (20 years of good time and age 50, or 25 years
of experience at any age), some of those employees who were older when hired
or were working at positions other than actually controlling traffic

(like training) might first become eligible under CSRS or FERS provisions
(age 55 with 30 years federal employment, age 60 with 20 years federal
experience, or age 62 with 5 years experience). Our review of the
eligibility data shows that about 2,500, or 12 percent of the current
controller workforce, was eligible to retire at the end of fiscal year 2001.
As figure 3 shows, an increasing percentage of current

controllers will become eligible to retire between fiscal year 2002 and
2011, with nearly 11,200 of the current controllers becoming eligible for
retirement over the next 10 years. In addition, those already eligible,
coupled with the nearly 11,200 additional controllers becoming eligible

over the next 10 years, will increase the number of current controllers
eligible to retire to more than 13,600, or 68 percent of FAA?s total current
controller workforce, by the end of fiscal year 2011. 18

17 Because FAA?s database does not contain information on the amount of an
employee?s good time service, to calculate retirement eligibility under the
special provisions we assumed that all of the controllers? FAA experience
was good time. The results indicate the maximum number of current
controllers who become eligible each year.

18 Although about 68 percent of current controllers may become eligible to
retire by 2011, FAA?s workforce at that time will not have this level of
eligibility because some current controllers will retire before then and FAA
will hire new employees.

Figure 3: Air Traffic Controllers Becoming Eligible for Retirement in Each
Fiscal Year

2, 000 Number of staff becoming eligible

9.8% 1, 800 1, 600 1, 400

6.5% 6.4% 5.7%

6.1% 1, 200

5.6% 1, 000

4.1% 4.2% 800

3.6% 3.8% 600 400

1.2% 1.5% 1.4% 1.6% 1.1% 200

0 1997

1998 1999

2000 2001

2002 2003

2004 2005

2006 2007

2008 2009

2010 2011

Fiscal year About 2,500 (12 percent) of current controllers are already
eligible to retire, and by

2011 another 11,200 (56 percent) of current controllers will become eligible
to retire.

Past eligibility rate each fiscal year, as a number and as a percentage of
fiscal- year total

Projected eligibility rate each fiscal year, as a number and as a percentage
of fiscal- year total

Source: GAO?s analysis of FAA?s data.

GAO Model Predicts High Our controller attrition simulation model projects
that high numbers of

Attrition Levels over the controllers will leave the workforce between
fiscal years 2002 and 2011.

Next Decade Probabilities for separation were based on controller attrition
patterns

between 1997 and 2000 and were applied to the 20, 021 controllers at FAA as
of June 30, 2001. Projections are therefore based on the June 2001
population, and there is no adjustment for new appointments. As shown in
figure 4, the simulation model predicts that about 600 to 800 controllers
will leave each year between fiscal years 2002 and 2011, which is one and
one- half to two times higher than average attrition was over the past 5
years. It also indicates that nearly 7,500 controllers (about 37 percent of
the current controller workforce) are projected to leave FAA by the end of
fiscal year 2011.

Figure 4: Past and Simulated Air Traffic Controller Attrition, by Fiscal
Year

2, 000 Number of staff

1, 800 1, 600 1, 400 1, 200 1, 000

4.4% 4.3% 4.1% 4.0% 3.9% 3.7% 3.5% 800

3.3% 3.1% 2.9% 600

2.2% 2.5% 2.2% 2.0% 2.0% 400

200 0

1997 1998

1999 2000

2001 2002

2003 2004

2005 2006

2007 2008

2009 2010

2011

Fiscal year Of the 20,021 controllers working at FAA as of June 30, 2001, we
estimate that about

7,500 controllers will leave by the end of fiscal year 2011.

Past attrition rate each fiscal year, as a number and as a percentage of
fiscal- year total

Simulated attrition rate each fiscal year, as a number and as a percentage
of fiscal- year total

Note: -- Denotes the minimum and maximum values from the simulation model.
Source: GAO simulation using FAA database.

Many Controllers Based on the results of our survey of air traffic
controllers, we estimate

Responding to GAO Survey that many controllers plan to leave FAA soon. Of
the 20,021 controllers

Plan to Leave FAA Soon working at FAA as of June 30, 2001, we estimate that
approximately 5,000 controllers plan to leave (predominately because of
retirement) between

fiscal years 2002 and 2006, and about 10,900 by the end of fiscal year 2011.
19 As shown in figure 5, we estimate that between fiscal years 2002 and
2011, approximately 1,100 controllers on average plan to leave each year,
and about 1,300 20 controllers plan to leave in fiscal year 2007 alone- also
the peak year for controllers reaching retirement eligibility. These
estimates are more than double the recent attrition levels that FAA has

19 These estimates also include some minimal amount of nonretirement
attrition- those few controllers who indicated that they would leave FAA
before they retired. 20 The 95 percent confidence interval for this estimate
extends from 1,007 to 1,656 controllers.

experienced- on average, about 436 controllers separated each year for the
past 5 years.

Figure 5: Survey Estimates: Past and Estimated Air Traffic Controller
Attrition

2, 000 Number of staff

1, 800 6.6%

6.6% 6.2%

6.1% 1, 600 1, 400

5.6% 5.7%

5.3% 5.4% 1, 200

4.2% 1, 000

3.3% 800

600 2.2% 2.5% 2.0%

2.0% 2.2% 400 200

0 1997

1998 1999

2000 2001

2002 2003

2004 2005

2006 2007

2008 2009

2010 2011

Fiscal year Of the 20,021 controllers working at FAA as of June 30, 2001, we
estimate that about

10,900 controllers plan to leave by the end of fiscal year 2011.

Past attrition per fiscal years, as a number and as a percentage of fiscal-
year total.

Estimated attrition per fiscal year, as a number and as a percentage of
fiscal- year total.

Note: -- Confidence interval: displays the upper and lower bounds of the 95%
confidence interval for each estimate.

Source: FAA?s historical data and GAO?s estimates based on survey responses.

We also estimate, based on the survey responses, that there are two time
frames for when controllers said they might leave or retire. An estimated 40
percent of the controllers said they planned to leave or retire at age 50 or
earlier, and another 26 percent said they planned to leave or retire around
the maximum 56- separation age. In addition, we also found that
approximately 51 percent of controllers said they planned to retire when
they first become eligible.

Supervisor Attrition Is Because supervisors are important to air traffic
control operations and

Likely to Increase Rapidly because they tend to be older than others
controlling traffic, we examined

retirement eligibility and survey results of supervisors at FAA as of June

2001. We found that supervisors will become eligible and said they planned
to leave FAA in very high numbers over the next decade.

We found that 1,205, or 65 percent, of current supervisors will become
eligible to retire between 2002 and 2011. (See fig. 6.) Given that 28
percent of current supervisors are already eligible to retire and that by
2011 another 65 percent will have reached eligibility, about 93 percent of
current supervisors will be eligible to retire by the end of fiscal year
2011. As a result, FAA may face substantial turnover in its supervisory
ranks over the next decade. Figure 6: Past and Projected Retirement
Eligibility for Supervisory Air Traffic Controllers

400 Number of supervisors

350 300

14.2% 250

200 150

6.8% 7.5% 6.7%

6.4% 6.1% 5.6% 100

4.6% 3.9% 2.4% 3.0% 3.0%

3.0% 50 2.1%

2.7% 0 1997

1998 1999

2000 2001

2002 2003

2004 2005

2006 2007

2008 2009

2010 2011

Fiscal year 28 percent of current supervisors are already eligible to
retire, and by 2011

another 65 percent of current supervisors will become eligible to retire.

Past attrition per fiscal year, as a number and as a percentage of fiscal-
year total

Projected eligibility per fiscal year, as a number and as a percentage of
fiscal- year total

Source: GAO?s analysis of FAA database.

In addition, estimates from our survey show sizable attrition through fiscal
year 2011. As shown in figure 7, we estimate that 770 21 supervisors (about
39 percent of current supervisors) said they plan to leave between fiscal

21 The 95 percent confidence interval for this estimate extends from 592 to
984 supervisors.

years 2002 and 2006, and 1, 503 22 supervisors (about 76 percent of current
supervisors) plan to leave FAA, primarily through retirement, through fiscal
year 2011, an average of about 150 per year. The peak year in planned
attrition is fiscal 2007, when we estimate that 221 23 supervisors plan to

leave. This level of potential attrition for supervisors is higher than in
the past 5 years, during which an average of 71 supervisors left each year.

Figure 7: Survey Estimates: Past and Estimated Air Traffic Controller
Attrition for Supervisory Air Traffic Controllers

400

Number of supervisors

11.2% 350

6.9% 9.4%

300 8.1%

8.4.% 7.4% 6.9%

6.7% 250 5.5% 5.9%

200 150 100

3.9% 3.5% 3.7% 2.9% 3.1% 50

0 1997

1998 1999

2000 2001

2002 2003

2004 2005

2006 2007

2008 2009

2010 2011

Fiscal year We estimate that about 1,500 (76 percent) of current supervisors
plan to leave FAA by

the end of fiscal year 2011.

Past attrition per fiscal year, as a number and as a percentage of fiscal-
year total

Estimated attrition per fiscal year, as a number and as a percentage of
fiscal- year total

Note: -- Confidence interval: displays the upper and lower bounds of the 95%
confidence interval for each estimate.

Source: FAA?s historical data and GAO?s estimates based on survey responses.

High levels of supervisor attrition could also affect the controller
specialist workforce. To the extent that FAA replaces supervisors who leave,
increases in supervisory retirements could further reduce the number of
experienced controller specialists available to control traffic and increase
controller specialist hiring needs in order to replace the controllers
moving

22 The 95 percent confidence interval for this estimate extends from 1,279
to 1,728 supervisors. 23 The 95 percent confidence interval for this
estimate extends from 115 to 382 supervisors.

to supervisory positions. The overall impact of supervisor attrition is
unclear at this time. Until recently, FAA was in the process of reducing the
controller- to- supervisor ratio from 7- to- 1 to 10- to- 1, through
attrition, as agreed to in the 1998 NATCA collective bargaining agreement.
This strategy would help mitigate the flow of NATCA bargaining unit
controllers into the

supervisory ranks. The outcome of this strategy is uncertain because the
Conference Report for the fiscal year 2002 Department of Transportation
Appropriations (H. Rpt. 107- 308) stated that the conferees were concerned
about the impact of the reduction and directed FAA not to reduce supervisory
staffing further. 24 FAA intends to abide by this language for this fiscal
year, and its future decisions on supervisory reductions are subject to
congressional direction. FAA?s Busiest Facilities May

Because of the crucial role played by en route centers and the busiest Face
High Attrition Levels

terminal facilities in the national air space system, we analyzed the impact
of retirement eligibility on the 21 major en route centers, the 10 busiest
airport towers, and the 10 busiest TRACON facilities. Based on our analysis
of FAA?s employee database, we found that the en route centers and the
busiest terminal facilities will experience a sizable increase in the number
of controllers reaching retirement eligibility. As figure 8 shows,
retirement eligibility in these facilities grows over the next decade.

24 For additional discussion of controller supervisors see Air Traffic
Control: FAA Enhanced the Controller- In- Charge Program, but More
Comprehensive Evaluation Is Needed, GAO02- 55 (Washington, D. C.: Oct.
2001).

Figure 8: Controllers Becoming Eligible for Retirement by Fiscal Year for En
Route Centers, Ten Busiest Towers, and Ten Busiest TRACONs

20

Percentage of staff becoming elligible to retire each fiscal year

18 16 14 12 10

8 6 4 2 0

1997 1998

1999 2000

2001 2002

2003 2004

2005 2006

2007 2008

2009 2010

2011

Fiscal year

En route centers Busiest towers Busiest TRACONs

Source: GAO?s analysis of FAA?s data.

In analyzing retirement eligibility data for the en route centers, we found
that 903, or about 11 percent, of the controllers currently at FAA?s 24 en
route centers are already eligible to retire. Additionally, the cumulative
percentage of current controllers becoming eligible to retire increases to
about 28 percent by the end of fiscal year 2006 and reaches about 65 percent
by the end of fiscal year 2011. In terms of the 21 major en route

centers, the Jacksonville center had the highest proportion of
retirementeligible controllers at the end of fiscal year 2001, with 79 of
its 376 controllers being eligible for retirement (21 percent). By the end
of fiscal

year 2006, at least 29 percent of current controllers will be eligible for
retirement at 10 centers- Albuquerque, Atlanta, Boston, Fort Worth, Houston,
Jacksonville, Los Angeles, Memphis, Seattle, and Washington,

D. C. At the 10 busiest airport towers, 76, or about 10 percent, of current
controllers are eligible to retire. The cumulative percentage rises to about
34 percent by the end of fiscal year 2006 and reaches 74 percent by the end
of fiscal year 2011. Based on our analysis for these towers, we found that

the Denver tower had the highest proportion of retirement- eligible
controllers as of September 30, 2001, with 14 of its 51 (27 percent)
controllers being eligible to retire. By the end of fiscal year 2006, 45
percent of Denver?s current controllers will be eligible to retire, and by
the end of fiscal year 2011 it reaches 90 percent, as 46 of its 51 current
controllers will reach retirement eligibility. At the 10 busiest TRACON
facilities, about 199, or about 12 percent, of

current controllers are eligible to retire. The cumulative percentage
increases to about 36 percent by the end of fiscal year 2006 and reaches
about 73 percent by the end of fiscal year 2011. Based on our analysis for
these facilities, the Dallas/ Fort Worth TRACON had the highest level of
current controllers eligible to retire at the end of fiscal year 2001, with
36 of its 147 (24 percent) controllers being eligible. By the end of fiscal
year 2006, the cumulative percentage grows to 46 percent, and by the end of
fiscal year 2011 it reaches 87 percent, as 128 of the 147 controllers
currently at the facility will have reached retirement eligibility.

FAA Needs a More Comprehensive Workforce

Chapt er 3

Plan for Air Traffic Controllers Attrition of air traffic controllers will
increase substantially over the next decade, primarily because many
controllers will retire. This condition is widespread across the various air
traffic control facilities at the FAA, and the potential for massive
turnover exists even at FAA?s most complex and busiest facilities. To
effectively deal with expected attrition, government agencies must identify
human capital needs, assess how current staff and expected future staff will
meet those needs, and create strategies to address any shortfalls or
imbalances. As we have reported, a highperforming organization typically
addresses its current and future workforce needs by estimating the
following: the number of employees it will need; the knowledge, skills, and
abilities those employees will have in order for the organization to
accomplish its goals; and the areas where employees should be deployed
across the organization. 25 We have developed a model that identifies
strategic workforce planning as a critical success factor in effectively
managing a human capital program, because

such planning can help agencies ensure that they have adequate staff to
accomplish their missions. 26

Although FAA will be faced with unprecedented numbers of retirements of its
air traffic controllers, it has not yet developed a comprehensive workforce
plan to address this issue and therefore risks having a shortage of
qualified controllers. Good workforce planning includes developing

strategies for integrating hiring, recruiting, training, and other human
capital activities in a manner that meets the agency?s long- term
objectives. FAA generally hires new controllers only when current,
experienced controllers leave, and it does not adequately take into account
the time

necessary to fully train these replacements. Furthermore, although FAA
intends to increasingly hire individuals with no prior controller
experience, its new aptitude test for potential candidates may not be as
effective in screening them as initially planned. In addition, FAA has not
provided its

training academy with the resources necessary to handle the expected large
increase in controller candidates. Finally, exemptions to the mandatory age
56- separation provision raise equity and safety issues. FAA therefore might
face a shortage of experienced controllers, leading to an increase in
overtime logged by its remaining controllers. Increased flight delays might
also result from this situation, as fewer controllers might not

25 Human Capital: Key Principles from Nine Private Sector Organizations,
GAO/ GGD- 00- 28 (Washington, D. C.: Jan. 31, 2000). 26 A Model of Strategic
Human Capital Management, GAO- 02- 373SP (Washington, D. C.: Mar. 15, 2002).

be able to safely guide the same number of flights that would be possible
with a fully staffed controller workforce.

FAA?s Hiring Process A key component of workforce planning is ensuring that
appropriately

Does Not Adequately skilled employees are available when and where they are
needed to meet

an agency?s mission. This means, in part, that an agency continually needs
Ensure that Qualified

trained employees becoming available to fill newly opened positions. FAA?s
Controllers Will Be

current hiring process does not adequately ensure that qualified Available
When

replacements are available to expeditiously assume the responsibilities of
those who retire. The main objective of FAA?s branch manager for resource

Necessary management is to ensure that controller- staffing levels meet the
levels

called for in FAA?s contract with the controller?s union (NATCA). To do
this, he estimates how many controller specialists will leave during the
year and allocates this number among regions as a target- hiring figure. On
at least a quarterly basis, he informs the officials in the regions how many
controller

candidates they are allowed to hire for that period. If attrition is lower
than expected during that period, he may tell them to delay hiring until a
later quarter. For example, in fiscal year 2001, the plan called for hiring
425 controller candidates but, because of lower- than- expected attrition
levels, FAA hired only 358 new controllers. According to this official, FAA
does not have budgetary resources to maintain and develop an employee
pipeline to ensure that fully certified replacements are available, so it
has no plans to change these hiring practices.

FAA?s approach of hiring new employees only when current employees leave
does not adequately account for the time needed to train controllers to
fully perform their functions, or for the increased retirements that are
projected in the short and long terms. The amount of time it takes new
controllers to gain certification depends on the facilities at which they
will work, but it generally takes from 2 to 4 years and can take up to 5
years at some of the busiest and most complex facilities.

The branch manager?s May 2001 hiring plan identifies a ?hiring lead time
adjustment? starting in fiscal year 2004 that provides for additional hiring
in recognition of the time necessary to train employees. The numbers
included, however, do not appear adequate to account for the large potential
increases in controller attrition. For example, in fiscal year 2004, the
adjustment is for hiring 70 extra candidates, which would respond to a

potential attrition of about 700 to 1,100 controllers in 2006, when these
new hires might be ready for certification at some facilities. In addition,
the branch manager told us that budget constraints play a key role in

determining the timing of hiring new candidates. For example, he said that
budget requests are tied to the NATCA contract amounts and that FAA had no
plans to request the additional funding necessary to go above those levels.
FAA officials also stressed that staffing management is now a partnership
between FAA and NATCA, and that this also creates constraints on FAA?s
ability to hire and place new controllers at specific facilities.

FAA regional officials, who are responsible for ensuring that FAA?s air
traffic facilities are adequately staffed, are particularly concerned about
FAA?s replacement- hiring policy. Eight of nine regional officials with whom
we spoke stated that they would like for FAA to allow them to hire new

controller staff above their authorized levels so that experienced, fully
qualified controllers will be ready when current controllers retire. The
officials were particularly concerned that significant increases in
retirement rates among veteran controllers would leave the facilities short
of qualified controllers while new trainees are hired and trained. Several
regions stated that they had made formal and informal requests to FAA
headquarters to obtain additional controllers who could be hired and trained
in advance of future retirements. In May 2001, for example, officials

from FAA?s Southwest Region formally requested 48 additional staff members
to mitigate the impact of future retirements. The region asked for new hires
at one of its en route centers to ?ensure that quality customer service is
maintained, budgetary concerns are addressed, and controller attrition is
dealt with.? In April 2002, FAA headquarters informed the region that their
request was denied because of operational constraints imposed by the 1998
agreement with the controllers? union and because of the current fiscal
year?s budgetary constraints. Furthermore, numerous FAA regional officials
told us that they were frustrated by their agency?s insistence on staffing
as close to the numbers called for in the NATCA

contract as possible. A lack of experienced controllers could have many
adverse consequences, according to several FAA regional officials. Several
regional officials stated that if a facility becomes seriously short of
experienced controllers, the

remaining controllers might have to slow down the flow of air traffic
through their airspace. If the situation became dire, FAA could require
airlines to reduce their schedules, but this would be an unlikely, worst-
case scenario, according to some FAA regional officials. Also, because there
would be fewer experienced controllers available to work, some FAA

facility officials stated that those controllers could see increased
workloads and additional, potentially mandatory, overtime. Some facility

managers told us that they expected this increased burden to result in
additional work- related stress for the remaining controllers, which would
increase sick leave usage. It could also cause experienced controllers to
retire sooner than they otherwise might. For example, based on our survey
results, we estimate that 33 percent of controllers would accelerate their
decision to retire if forced to work additional mandatory overtime.

FAA Developed Identifying sources of future potential employees with the
requisite skills

Screening Test to Help and aptitude is another key piece of a comprehensive
workforce plan. As

discussed in chapter 1, FAA historically has hired its new controllers from
a Identify Potential

variety of sources, including graduates of institutions in FAA?s collegiate
Candidates Most Likely

training initiative program, the Minneapolis Community and Technical to
Succeed

College, candidates already on a list maintained by OPM, controllers
formerly employed by FAA who were fired by President Reagan in 1981, and
former DOD controllers. Table 5 shows the sources and number of new
controllers that FAA hired between fiscal years 1997 and 2001.

Table 5: Sources of New Controllers, Fiscal Years 1997- 2001 Source 1997
1998 1999 2000 2001 a Tot al

Collegiate Training 161 50 60 119 75 465

Initiative Minneapolis Community 32 48 52 77 82 291

and Technical College Office of Personnel 4 14 149748 Management Reinstated
employees 26 12 4 7 16 65

Former controllers fired 188 289 30 41 14 562

in 1981 Department of Defense 89 355 96 136 117 793 Total 500 768 256 389
311 2, 224 a Partial year data.

Source: GAO?s analysis of FAA regional data.

DOD officials were concerned that increasing retirements of FAA?s
controllers over the next 5 years will cause greater operational problems,
and possibly affect defense readiness, if potentially thousands of DOD
controllers were to fill openings at FAA. DOD has lost many controllers to
FAA- about 35 percent of FAA?s hires in the past 5 years. FAA?s regional

officials told us that they like to hire former military controllers because
of their experience, maturity, and work ethic. DOD officials with whom we
spoke explained that these losses had resulted in cutbacks for fighter
training missions by at least one of the armed services and in the
implementation of significant retention bonuses to military controllers.
Although DOD employs both civilian and uniformed military controllers,

there remains a pay disparity between DOD and FAA. These officials believe
that the higher pay offered by FAA explains why DOD military and civilian
controllers apply for FAA controller jobs. For example, in fiscal year 2001,
the maximum base salary levels for DOD controllers were $48,730 for a DOD
military controller and $74,553 for a DOD civilian controller, while FAA
controllers could earn up to $128,386. 27 DOD officials stated that both
agencies (FAA and DOD) must meet their recruiting and retention goals to
support national security and defense requirements. To that end, DOD
officials said that the focus needs to be on the requirement

for air traffic controllers as a whole and not on two competing systems.
Along these lines, FAA headquarters officials said that because FAA hopes to
achieve a more diverse workforce, it expects to concentrate increasingly on
hiring off- the- street candidates. The success of the off- the- street
hiring

depends in large part on identifying potential candidates who have an
appropriate aptitude for controllers? work. Traditionally, FAA used the
academy?s initial entry- training program to screen out candidates who could
not become successful controllers. According to FAA officials, as many as 50
percent of off- the- street applicants have dropped out before finishing the
required training program. These officials estimated that about $10 million
per year was spent on training candidates who later failed the program. FAA
therefore developed a new screening exam, known as AT- SAT, to better ensure
that the new hires have the skills and abilities to succeed on the job. FAA
plans to require that candidates without prior experience pass the 8- hour
AT- SAT exam before they begin training at its academy. According to academy
officials, the academy is planning to rely on AT- SAT as a way to screen out
candidates unlikely to pass the academy?s

training, and it has therefore revised its training program to emphasize
teaching skill- sets rather than serving as a screening program.

Uncertainty exists regarding the exam?s ability to screen out unsuccessful
candidates and help ensure that new candidates have the aptitude to

27 Base salary figures do not include other military compensation, such as
subsistence and housing allowances; these allowances fluctuate, depending on
numerous factors.

become successful controllers. For example, FAA has recently changed the
exam to allow more candidates to pass, which creates some uncertainty about
its ability to identify successful candidates. During initial validation of
AT- SAT, FAA found that the test should predict, with a high level of
validity, that those who passed it would become successful controllers.
However, FAA found that only about 28 percent of non- FAA test subjects and
about 62 percent of active controllers could pass the test. In addition,
they found that passing rates for some applicant groups, including
particularly African- Americans and females, might be significantly lower
than the overall passing rates. Therefore, FAA concluded that the passing
score on the test was set higher than the typical controller?s job
expectations. As a result, the developers of the exam changed the weight
given to different portions of the exam and adjusted the passing score to
tie the test more accurately to the actual job performance of controllers.
According to FAA, this will result in more applicants passing the exam (68
percent are now expected to pass).

FAA plans to begin using the test in June 2002. Although FAA has not
revalidated the effectiveness of the revised exam, FAA officials stated that
they have long planned to perform an operational evaluation of the exam to
assess how well the exam works in practice, and that they are currently

considering two options for performing this evaluation. First, FAA could
correlate candidates? scores on the exam with how well they perform on a
computer simulation of actual air traffic. In order to implement this
option, FAA would have to develop a new computer- based performance measure
for the terminal environment. Officials indicated that this would cost
several hundred thousand dollars. The second option would be to validate the
exam against initial training at the academy, field training, and job
performance. This would require FAA to develop criteria for measuring
success in each of these three areas. In any case, to evaluate the exam, the
officials need to decide on an option, develop a detailed implementation
plan, and identify funding for this purpose. Officials could not provide an

estimate as to when they will decide on a specific option. Until the results
are evaluated, the operational effectiveness of the exam will be unknown.

Challenges Exist in Workforce planning should consider the approach and
resources necessary

Addressing Academy for providing new employees with the means to acquire the
knowledge,

skills, and abilities to accomplish the agency?s mission. However, FAA has
and On- the- Job

not adequately addressed the challenges associated with providing the
Training Resources and

training resources- specifically training staff, equipment, and Equipment
Needs

opportunities for on- the- job training- needed for large increases in new
hires. Most controller candidates undergo both 15 weeks of classroom
exercises at FAA?s academy in Oklahoma City and on- the- job training at the
facility where employees are assigned. As of March 2002, the academy was
staffed with 91 FAA employees and 60 contractors. This number of employees
and contractors has been used to train, on average, about 200 new hires for
each of the past 5 years. The academy?s training plan anticipates that
between 547 and 980 controller candidates might need training each year
through fiscal year 2005. To meet the projected levels, these officials
believe they will need up to 50 additional staff to provide training.

The training academy may have difficulty recruiting current controllers to
conduct portions of their training program. Academy officials told us that
their recent attempts to persuade experienced controllers to volunteer to
train new recruits have not been very successful. Academy officials
explained that the 1998 pay raise, which in some cases increased salaries
for controllers by more than 30 percent but applied only to periods when

the controllers were actually guiding air traffic, has affected the
controllers? willingness to participate. Whereas a controller was once paid
the same amount for providing training as for controlling traffic, under the
new system a controller would lose pay by becoming a trainer at the academy.
Academy officials said they recently put an announcement out asking for
volunteers to conduct training and received 31 applications. They noted that
before the pay raise they were receiving hundreds of applications for these
positions, which provided them a greater opportunity to select from a
broader pool.

Equipment deficiencies also hamper the academy. For example, the academy is
training en- route controllers on equipment that is not used at actual en-
route centers, so controllers must retrain on different equipment once they
reach their facilities. To efficiently train en- route and terminal
controllers, academy officials told us that they need a specialized en-
route

simulator lab known as a Display System Replacement lab, which costs between
$7 million to $45 million, depending on the sophistication of the model
purchased. Academy officials have been trying to obtain this

equipment for several years, and the academy has recently made another
proposal regarding this equipment. FAA headquarters is expected to decide
whether to purchase this lab in the near future. In addition, the academy
uses tower simulators to give trainees experience with controlling traffic
in a computer- simulated environment. However, academy officials said their
current simulators are often broken, outdated, and lacking in the necessary
capacity to train large numbers of new hires. The cost of the new equipment
is estimated at $2 million. If FAA does not make these investments, academy
officials said, controller candidates will need more training time when they
reach their facilities.

New controllers might also have difficulty obtaining on- the- job training,
FAA regional officials stated. New controllers are to receive their facility
training from fully certified controllers already working in that facility.
Under FAA?s current hiring system and estimated attrition rates, however,
there will be fewer experienced controllers to provide training and more

new hires in need of training. More time will thus likely be needed to train
new controllers. This situation could be particularly acute at FAA?s en-
route centers and busy terminal facilities, because it takes longer to train
replacement controllers at these facilities. Retirements at these facilities
are expected to increase the burden on the remaining experienced controller
staff.

Exemptions to the Age Ensuring that a workforce retains employees with the
requisite skills and

56 Separation abilities is another important piece of workforce planning. As
described in

chapter 1, legislation passed in 1972 stipulates that air traffic
controllers Provision Raise Safety

must separate at age 56. 28 Some controllers are exempt from the retirement
and Equity Concerns

rule, however, and continue to work beyond age 56. This practice raises two
concerns: (1) whether the skills and abilities of the older controllers have
diminished, thus potentially compromising safety; and (2) whether the
exemptions result in unequal treatment for some controllers.

28 This legislation covers those controllers under the CSRS retirement
system. Another mandatory separation provision was passed in 1986 to cover
those controllers who are under the FERS retirement system Public Law 99-
335 (codified at 5 U. S. C. 8425a).

In 1972, Congress directed that ?an air traffic controller shall be
separated from the service on the last day of the month in which he becomes
56 years of age.? 29 The House Report associated with this law justifies the
provision by stating that ?air traffic control is a young man?s business?
and that because of the natural forces of aging, magnified by the stresses
of control functions, the productive and proficient life of the controller
is

substantially less than that which prevails in most other occupations.? 30
In addition, the report states, ?the controllers themselves are convinced
that the demands of their job are so great that only young, healthy adults
can consistently do a safe, competent job of controlling the steadily
growing volume of air traffic.? The House Report further states that ?as the
controller approaches age 50 his mental faculties of alertness, rapid
decision making, and instantaneous reaction? begin a definite decline.? In

addition, the associated Senate Report 31 states, ?like skilled athletes,
most controllers lose proficiency to some degree after age 40, and in the
interest of the public?s safety, should not be retained as controllers in
busy facilities beyond the time they can perform satisfactorily.?

The law?s provision requires mandatory separation at age 56 for controllers
who separate and control air traffic; provide preflight, in- flight, or
airport advisory service to aircraft operators; or serve as the immediate
supervisors of any employee who performs these duties. These positions
include controller specialists and their first- line supervisors as well as
traffic management coordinators and their first- line supervisors. Some
controllers who separate and control traffic are exempted from this

provision, however, including controllers appointed by the Department of
Transportation (DOT) before May 16, 1972, and controllers appointed by DOD
before September 12, 1980. In addition, those controllers covered by the
FERS retirement system can continue working past age 56 until they have
reached 20 years of service in a covered position (so called good time under
the special air traffic controller retirement provisions). Similarly, on
November 12, 2001, the president signed a law allowing controllers covered
by the CSRS retirement system to work in covered positions past age 56 until
they first become eligible for retirement annuities under any

29 Public Law 92- 297 (codified at 5 U. S. C. 8335). 30 House Report 92-
516. 31 Senate Report 92- 774.

retirement scenario. 32 Our analyses of FAA?s employee database shows that
approximately 700 of those controllers currently engaged in separation and
control of traffic are exempt from the requirement and have already turned
age 56, and another 1,200 will reach 56 by December 31, 2006, if they do not

leave FAA before then. According to FAA, 287 controllers were appointed by
DOT before May 16, 1972, and are exempt from the requirement. Most of the
remaining exempted controllers were either appointed by DOD before September
12, 1980, or are covered by FERS provisions.

FAA also has the statutory authority to waive the age provision on a caseby-
case basis. The applicable law states that ?the Secretary of Transportation,
under such regulations as he may prescribe, may exempt a controller having
exceptional skill and experience as a controller from the automatic
separation provisions of this subsection until that controller becomes 61
years of age.? 33 However, according to an FAA Headquarters official, FAA
has never granted an age waiver to the mandatory separation provision.
Further, since 1995, it has been FAA?s policy not to grant any age waivers.
This official also stated that most controllers are aware of the difficulty
in obtaining an age waiver and do not even apply for one- only

seven controllers have applied for a waiver since 1995. Despite this view,
many controllers said they would like the opportunity to work past the age
of 56. Our survey indicates that many controllers would continue to work if
they were permitted to do so; approximately 31 percent of respondents cited
the opportunity to work past age 56 as a factor that could lead them to
delay their retirement plans. In addition, regional FAA officials said they
would like to have the flexibility to retain some of these experienced
controllers.

As mentioned above, safety concerns formed the basis of the age- 56
separation provision. Only limited actions have been taken, however, to
assess whether those controllers who are exempted from the provision have
adequately retained the skills and abilities necessary to perform their
duties. FAA requires all controllers to pass annual physical examinations
that test sight, hearing, and overall health conditions. No additional
tests- such as for mental acuity or changes in reaction time- are given to
controllers who surpass age 56.

32 Treasury, Postal Service, and General Government Appropriations Act,
2002, Public Law 107- 67, Sec. 640 (a). 33 Public Law 92- 297 for CSRS,
Public Law 99- 335 for FERS.

The equity issues associated with the exemptions to the age- 56 separation
rule could become more prominent in the future if FAA continues to rehire
controllers fired during the 1981 strike. In 1993 President Clinton, through
presidential directive, lifted the ban on hiring former striking employees.
In the past 5 years, FAA has rehired about 850 controllers who were fired in
1981. The average age of the 733 still working as of June 30, 2001, was 54,
and about 35 percent were aged 56 or older. The oldest was 69 as of June 30,
2001. In addition, FAA officials said that most of the rehires are exempt

from the mandatory separation provisions because they were originally hired
before May 16, 1972. Further, recently a group of controllers fired during
the 1981 strike filed a class action lawsuit alleging that FAA discriminates
against such controllers because of their age. Depending on the outcome of
this lawsuit, about 2,000 former controllers- many aged 50 and above- could
be given hiring priority.

Conclusions Although the attrition scenarios projected by FAA and us reflect
estimates of the future, and any particular estimate in any given year is
subject to

varying degrees of uncertainty, the overall results suggest that FAA will
face significant personnel challenges. If controllers leave at a quicker
pace than estimated, the situation may become even more difficult for FAA,
as it would have to swiftly replace its seasoned controllers with new
controllers possessing lesser experience. To the extent that controllers
leave at a slower pace than estimated, FAA will have a larger window of
opportunity to replenish its workforce. Ultimately, FAA?s ability to
successfully plan for and manage this situation will dictate its overall
impact on the nation?s air traffic control system and the safety and
efficiency of air travel in the United States.

The employees whom FAA will need to replace possess unique skills and are
critical to the safety and efficiency of the nation?s air transportation
system. FAA, as the agency responsible for managing this workforce, does

not have a comprehensive workforce plan to help manage the expected
turnover. An effective human capital process anticipates expected attrition
and includes the development of a comprehensive workforce plan that (1)
establishes an effective approach for hiring individuals with the requisite
skills and abilities in time to accomplish agency missions, (2) provides new
employees with the best training opportunities possible to maximize their
potential, and (3) uses opportunities to retain qualified staff. FAA?s
approach to workforce planning does not adequately address these

strategies, raising the risk that the safety and efficiency of the nation?s
air

transportation system will be adversely affected. In addition, if FAA does
not take steps to develop and implement a more comprehensive workforce
strategy, increased traffic delays and overtime costs could result. FAA?s
practice of hiring replacements for controllers only after a position is
vacated leaves the agency vulnerable to skills imbalances, with

inexperienced and uncertified controllers replacing seasoned veterans. This
situation may be exacerbated at individual air traffic control facilities
because the age and experience of controllers varies across the system,
which could cause some locations to experience additional staffing

challenges. Also of concern are the effects of the recent scoring changes
that were made to the test used to screen potential candidates. Until the
screening test results are examined, the ability of the exam to identify
candidates who will make successful controllers will not be known. Further,
the quality of the training that controllers receive could be compromised
because FAA has not addressed the human resources and

equipment needs of its training academy, despite the growing projected
student population. Finally, safety and equity issues associated with the
age- 56 separation exemptions could affect the morale of the controller
workforce and the safety of air traffic.

Recommendations for To help meet the challenges presented by hiring
thousands of new

Executive Action controller candidates, we recommend that the secretary of
transportation

direct the administrator of the Federal Aviation Administration to develop a
comprehensive workforce plan that includes strategies for:

 Identifying the number and timing of hiring necessary to ensure that
facilities have an adequate number of certified controllers available to
perform needed duties. As part of this effort, FAA should determine and plan
for the expected attrition levels and timing at each facility;

 Evaluating the newly developed screening test to determine whether it is
identifying the most successful candidates;

 Addressing the resource and equipment needs at the training academy to
help ensure that FAA is in a position to successfully train a growing number
of controller candidates; and  Assessing the safety and equity issues
associated with exempting

potentially large numbers of controllers from the mandatory age- 56
separation requirement.

Agency Comments and In commenting on a draft of this report, senior FAA
officials found that the

GAO?s Evaluation report was generally accurate and indicated that they would
consider our

recommendations in FAA?s workforce planning. Overall, FAA stressed that it
has a working human capital workforce strategy model that has enabled the
agency to meet its staffing goals over the past few years. FAA officials
agreed that the potential for sizable future

attrition, in the range of 600- 800 controllers per year, is likely over the
next decade. The officials said, however, that although they have plans that
extend to 2010, the uncertainty surrounding the future, along with labor
contracts and budget constraints, limit their specific workforce planning
for air traffic controllers to fiscal years 2002 through 2004. With general
agreement between FAA and GAO that attrition will grow substantially over
the next decade, we believe that the workforce challenges FAA faces extend
well beyond fiscal year 2004. As such, we believe that sound workforce
planning demands that FAA develop a strategic vision that includes a
workable, long- term plan to meet staffing needs.

Regarding our concern about FAA?s preparedness for the future, the FAA
officials remarked that FAA?s ability to meet its past goals is an
indication of its ability to meet future needs, and that there is nothing to
indicate that its successful performance will not continue in the future. We
recognize that FAA has been able to meet its recent staffing goals. However,
the recent

workforce climate for FAA could be significantly different from that which
it will face over the next decade. Chapter two of the report highlights the
workforce challenges, particularly the sizable anticipated increases in
controller attrition, that are likely over the next decade, and this chapter
identifies challenges in FAA?s planning that will make it difficult for FAA
to maintain its past performance. In particular, the report points out the
potential skills gap that FAA could face in the future because its current
hiring process does not ensure that fully qualified controllers are
available to replace experienced controllers when they leave.

The officials also commented that FAA has long planned for an operational
evaluation of the new screening exam, and that research associated with this
evaluation has been underway for some time. The officials indicated that
they are considering two options for evaluating the effectiveness of the
exam. The officials commented, however, that limited work has been done

on the evaluation process since 1998, and that they must determine which
option to pursue, develop a detailed implementation plan, and identify
funding for the evaluation. The officials further noted that continued

funding for the ongoing research could not be assured. In response to this
comment, we revised the text of the report to recognize FAA?s efforts and
plans regarding evaluation of the new screening exam. As such, we are
encouraged that FAA plans to conduct an operational evaluation of the exam,
once it has been implemented. However, we remain concerned that FAA has not
decided how it will conduct the evaluation or how it will fund it and has
already highlighted potential funding issues that could serve as a
constraint to performing the planned evaluation. Further, we believe that an
evaluation of the revised exam is an integral part of a comprehensive
workforce plan and have modified the recommendation to emphasize this
belief.

Finally, the FAA officials provided technical comments that we incorporated,
as appropriate. For example, we added information in this chapter to
highlight the constraints that FAA?s labor contract and budget impose on the
timing of hiring controllers.

Appendi xes Potential Impacts of Proposed Changes to Increase Air Traffic
Control Annuity

Appendi x I

Calculations In May 2001, a bill was introduced in the U. S. Senate that
would amend annuity computations for air traffic controllers retiring under
the CSRS system. 34 It would provide controllers with an additional 2
percent to their annuity computation for each year of service past 20 years.
Under CSRS, air traffic controllers are guaranteed the greater of either (a)
50 percent of their average highest 3 years salary (high- 3) or (b) the
basic CSRS annuity, which is 1. 5 percent of the high- 3 for the first 5
years, 1. 75 percent of the high- 3 for the next 5 years, and 2 percent of
the high- 3 for the remaining

years. In general, annuity calculations under the special provisions are
greater than under the basic CSRS formula until a controller has attained 27
years of service. The controllers? union supports this proposed legislation

and believes it serves as an incentive to keep controllers on the job beyond
their date of retirement eligibility. The bill, however, would create
substantial financial impacts to the federal treasury.

Proposed Bill Would The controllers? union notes that while CSRS controllers
may receive an

Increase Annuities annuity of 50 percent of their high- 3 salaries after 20
years of service at age

50 or after 25 years and any age, there is little incentive to continue
working longer because the amount of the annuity does not grow until a
controller accrues 27 years of service. For example, as shown in table 6, a
controller with the same high- 3 salary has the same level of annuity
whether he or she has worked 20 or 25 years.

Table 6: Current and Proposed Annuity Calculations Higher of current special

Proposed annuity Current years

rule annuity (50 percent (2 percent for each of service High- 3 salary

of high- 3 or basic CSRS) year over 20)

20 $100,000 $50,000 $50, 000 25 100,000 50,000 60, 000 27 100,000 50,250 64,
000 Source: GAO analysis.

Under the proposed bill, controllers would receive an extra 2 percent to
their annuities for every year of service past 20 years. As the above table
shows, this would increase the annuity calculation. In addition, the union

34 S. 871, the Federal Air Traffic Controllers Annuity Computation Act of
2001.

notes that federal firefighters and law enforcement officials receive
increased annuities after 20 years of service. The union also points out
that such an incentive could be useful in keeping controllers working
longer, which would help address the expected upcoming increase in
retirements. Proposed Bill Would

This proposal does not take into account the existing incentives that Create
Financial

encourage controllers to work past the point when they first become eligible
for retirement. First, a controller receives a full salary for each year
Impacts

he or she continues working as a controller. This level of income is
significantly higher than what controllers would receive from their
retirement annuities. For example, many controllers who are eligible for
retirement are making in excess of $100,000 per year. Assuming they would
retire under the controllers? current special rules, the differential would
be at least $50,000 in the first year. Second, controllers receive annual
pay increases like other federal employees, which increase the amount of
salary they receive and also increase the annuity levels. Table 7 shows
examples of retirement annuity calculations for an individual controller,
including projected salary increases, with 20 years of service at 50 years
of

age, 25 years of service at 55 years of age, and 26 years of service at 56
years of age. Also shown is the effect of current federal retirement rules
for other federal employees, which contain annuity penalties for federal
employees who retire before age 55. Table 7: Retirement Annuity Calculations
under Current CSRS and S. 871

Years of Current special rule

Proposed annuity (with Current federal CSRS Age service High- 3 a annuity
(50%) 2% increase) basic annuity

50 20 $100,000 $50, 000 $50,000 $32, 676 d 55 25 112,551 56, 275 67,530 b
52, 055 56 26 115,928 57, 964 71,875 c 55, 935

a Assumes 3% annual growth. b $56, 275 plus 2 percent of the high- 3 for 5
years- each year of service after 20. c $57,964 plus 2 percent of the high-
3 for 6 years. d The CSRS calculation includes an age penalty of 2 percent
per year for each year prior to age 55 that a federal employee retires. It
also assumes that employees with these ages and years of service would be
allowed to retire with immediate annuity. Otherwise, these individuals would
need to wait until turning 60 to begin receiving these amounts.

Source: GAO analysis.

The table above also shows the potential financial impacts that the bill
would create for individual controllers. For example, a controller with 26
years of service and a high- 3 of approximately $116,000 would receive an
annual increase of roughly $14,000 during each year of his or her annuity.
To determine the overall financial impact of the proposed bill, OPM

prepared an analysis of the bill?s long- term costs. OPM estimated that the
cost of the bill to the treasury had a present value of $1.7 billion. In
addition, OPM found that the higher benefit levels under the proposal could
potentially encourage somewhat earlier retirements, because employees with
24 years of service would receive the same benefit they now get with 31
years of service (that is, 58 percent of the high- 3).

Appendi x II

Air Traffic Controller Schools Figure 9: Locations of Air Traffic Controller
Schools

University of North Dakota Minneapolis Community

and Technical College Daniel Webster College

Oakdale College College of Aeronautics Community College of

Beaver County Purdue University Hampton University Middle Tennessee State

University Embry- Riddle Aeronautical

University Miami Dade

Community College Inter- American University University of

Alaska- Anchorage Mount San Antonio College

Source: GAO presentation of FAA data.

Table 8: Current Capacities of Air Traffic Controller Schools, as of
November 2001 Current

Current Type of degree

number of program

School program students

capacity

Daniel Webster College 4- year 9 20 University of North Dakota 4- year 68
119 Embry- Riddle Aeronautical 4- year 24 45 University Miami- Dade
Community College 2- year 30 80- 100

Purdue University 4- year 52 800 Community College of Beaver

2- year 139 139 County Dowling College 4- year 9 Unknown

Inter American University of 5- year 12 15 Puerto Rico Middle Tennessee
State

4- year 18 18 University College of Aeronautics 2- or 4- year 17 100

Mt. San Antonio College 2- year 40 Near capacity Hampton University 4- year
1 20 University of Alaska- Anchorage 2- or 4- year 62 90 Source: GAO
interviews with school representatives.

Retirement Eligibility Methodology and

Appendi x III

Analysis A key piece of information needed to assess FAA?s controller
staffing is knowing when controllers will become eligible to retire. To
determine when FAA?s series 2152 controllers will become eligible to retire,
we obtained selected demographic information from FAA?s personnel database

for all 22, 865 controllers (categorized as series 2152) working at FAA as
of June 30, 2001. Because we wanted to focus on those controllers who are or
could be involved in managing air traffic, we eliminated the 2,844 flight
service station controllers from our analyses, leaving a total of 20, 021
controllers.

For the 20,021 controllers on board as of June 30, 2001, we determined the
earliest date when they would become eligible to retire based on age at hire
at FAA, retirement plan, years of service at FAA, and years of pre- FAA,
retirement- creditable service. All FAA service was assumed to be in good

time and creditable toward air traffic controller special retirement
provisions (25 years of service at any age, and 20 years of service at age
50). For example, to compute the eligibility dates of controllers hired by
FAA at an age younger than 26, we added 25 years to their FAA service entry
dates. The eligibility date for controllers entering FAA from the ages of 26
up to 30 was calculated to be the date they turned 50 years of age.
Eligibility for

those from ages 30 through 35 was calculated by adding 20 years to their FAA
service entry date. For controllers entering FAA after the age of 35,
eligibility dates were based on the provisions of their retirement plans

(CSRS or FERS) and the amount of retirement- creditable service they had
before entering FAA. Whether the individual first became eligible under air
traffic controller special rules or regular CSRS/ FERS rules was dependent
on the number of years of creditable prior service he or she had upon entry
at FAA.

To provide a context for projected eligibility trends, we calculated the
number of controllers becoming eligible to retire each year between 1997 and
2001, using the same method described above, with FAA personnel data from
fiscal years 1997 through 2001. We also stratified our analyses of
controllers by facility type, supervisory status, selected location, and
position. The positions we examined included supervisors, certified
professional controllers, traffic management coordinators, controllers not
controlling traffic, developmental controllers, and trainees.

Appendi x IV

Methodology for Computer Simulation To assess the impact of estimated
separations for the career controllers between fiscal years 2002 and 2011,
we analyzed past separation trends and used these factors in estimating
future controller separations. Analysis of Separation

We analyzed FAA?s fiscal year- 1997 through fiscal year- 2000 data for
Trends

separations for the career air traffic controllers. In doing so, we
categorized past separations into voluntary retirements, other retirements,
and all other separations. For separations, we calculated years of service
by finding the difference between the service computation date and the date
of actual separation. Similarly, we calculated age at separation by finding
the difference between the date of birth and the date of separation. We
calculated the age and years of service for the air traffic controllers on
board at the end of the fiscal year similarly, but used September 30 th as
the end date.

For each 2- year interval of years of service and age of those who separated
from FAA during the period 1997 through 2000, we calculated the probability
of leaving by dividing the number who separated by the number of controllers
with a similar combination of years of service and age who were on board at
the end of the fiscal year preceding the fiscal year of separations. We also
bound all controllers into those 64 years and above, as well as all
controllers with 32 or more years of service. We then modeled the rate of
separation as a function of years of service, age, and CSRS/ FERS

retirement status. We developed an equation that estimates the rate of
separation for any age, years of service, and CSRS/ FERS retirement status.

Estimation We applied a simulation technique to each of the 20,021
controllers on Methodology of Future

board in 2001. Each controller?s age and years of service was used as input
into the model described above. Based on the model, an individual FAA
Controller

controller was considered to have separated if his or her predicted rate of
Separations

separation was less than a generated random number. If the predicted value
was greater than a generated random number, then the individual controller
was deemed not to have separated. This process was repeated for each of the
20,021 controllers. The process was then continued for those controllers who
were not estimated as having separated in 2001, but with each controller now
being 1 year older and having 1 more year of service. As before, the
controller?s new age and new years of service were used as input into the
model, and a predicted rate of separation was contrasted with a newly
generated random number to determine whether

the controller was considered as separated in 2002. A separation decision
was made for each of the remaining controllers, and either each controller
was counted as having separated in 2002 or else 1 year was again added to
both age and years of service. This process was repeated 11 times to
represent an 11- year horizon.

Because we are dealing with a process that is of a probabilistic nature
(that is, a controller may or may not have separated in any one year), we
repeated the process 100 times. The results of the 100 iterations were then
averaged to estimate the number of controllers separating in 2001 through
2011.

Limitations We developed a mathematical model in order to calculate any
individual controller?s rate of separation, which was based on three
criteria: (1)

FERS/ CSRS retirement status, (2) age at any point during the 11- year
horizon, and (3) years of service at any point during the 11- year horizon.
This mathematical model was based on the retirement rates for the same

three conditions: FERS/ CSRS retirement status, age, and years of service
for the 4 previous years. The optimization in developing the mathematical
model, known as regression analysis, is to minimize the squared differences
between the actual rates of separation and the predicted rates

of separation. An index, which is known as the squared correlation
coefficient and is bounded between zero and one, is one useful numerical
quantity to assess the strength or predictive power of the mathematical
model. A perfect fit in a model would yield a squared correlation
coefficient of 1. 00. In our mathematical model, we achieved a squared
correlation coefficient of .79. Thus, we were able to capture and predict

about four- fifths of the variability in the rates of separation for the 4
years? worth of separation data. One limitation, therefore, is that our
model does not predict with 100 percent accuracy the actual rates of
separation, although it is uncommon in real world applications to find such
a high squared correlation when dealing with behavioral data such as
separating from controller service. It is also worth noting that associated
factors such as an individual?s health, race, sex, or even children?s ages
and college

status may affect his or her decision to separate. These other factors were
either not available or not included in the mathematical model.

Another limitation in simulating the separation from service, which is based
on a mathematical model, includes the concept of using the previous patterns
of separating from service to generate the mathematical model. If the rate
of separation for those individual controllers starting in 2001 is

different from the previous 4- year patterns, then we introduce a source of
error into the simulation. As mentioned earlier, many factors are possible
in deciding to separate from service, which may include something unique or

something that for many controllers does not manifest itself until 2001 or
beyond. It is possible that the controllers who came aboard in 1982 and
beyond will separate at either higher or lower rates of separation than
those of their counterparts who began their service at an earlier time. This
variability cannot be assessed until actual rates of separation occur and
should be very closely monitored by the FAA.

Methodology for GAO?s Survey of Air Traffic

Appendi x V

Controllers? Retirement and Attrition Plans A primary objective in this
study was to determine the number or proportion of current air traffic
controllers who plan to retire each year, over the upcoming 10- year period.
To meet this objective, among other things, we surveyed a statistically
representative sample of air traffic controller personnel. We developed and
administered a survey designed to obtain the views of selected air traffic
controller personnel regarding issues associated with attrition, with
emphasis on retirement. The survey was

mailed in August 2001 to a stratified sample of 2,100 controllers. As of
February 12, 2002, we had received 1,591 completed, usable surveys. Our work
was conducted in accordance with generally accepted government auditing
standards.

Study Population FAA provided data from the Consolidated Personnel
Management Information System (CPMIS) as of June 30, 2001, for all FAA
controllers

(Series 2152). Since our primary interest was to estimate for the controller
population most likely to be directly involved in monitoring the movement of
planes, we removed from our study population 2,844 flight service station
controllers who give out weather information and pilot briefings. This left
us with a study population consisting of 20,021 air traffic

controllers. Sample Design The sample design for this study is a single-
stage stratified sample of FAA

employees in the study population. The first four strata consisted of
employees who were likely to be eligible to retire before the end of 2006.
35 The fifth stratum consisted only of rehired former employees, and a
final,

?residual? stratum was defined to ensure complete coverage of our study
population. A total sample of 2,100 employees was selected from the 20,021
employees in our study population, and we received a total of 1,591 valid
responses, for an overall response rate of 76 percent. 36 The following
table summarizes the population size, sample size, number of respondents,
and response rate for each of the sampling strata.

35 We assigned employees to one of the four ?eligible to retire by 2006?
strata based on age and years of service, as reflected in the CPMIS data
files provided by FAA. 36 Several surveys were received but not valid and
are not included among the 1,591 respondents. This includes surveys that
were returned blank, surveys that were duplicates from the same individual,
and surveys that were returned by respondents who had left FAA

before the fielding period ended.

Table 9: Survey Sample Size and Disposition Population

Sample Response Stratum size size Respondents rate

(1) Non- supervisors at en 2,079 331 247 0. 75 route centers- Eligible to
retire by 2006 (2) Supervisors at en route 547 177 152 0. 86

centers- Eligible to retire by 2006 (3) Non- supervisors at other

3,937 614 473 0. 77 facilities- Eligible to retire by 2006

(4) Supervisors at other 754 242 187 0. 77 facilities- Eligible to retire by
2006

(5) All rehired former 733 336 254 0. 76 employees (6) Controllers not
eligible to

11, 971 400 278 0. 70 retire by 2006 Total 20,021 2, 100 1,591 0. 76

Source: GAO analysis.

Survey Development In designing the questionnaire, we interviewed FAA
officials in Human Resources at headquarters in Washington, D. C., and at
the Oklahoma City

center to identify issues of interest and past work on retirement. We met
with NATCA officials and reviewed their 1999 survey about the retirement
eligibility and intentions of NATCA members in the terminal and en- route
air traffic controller bargaining unit. To further guide the development of
appropriate questions, we reviewed current literature on retirement issues
and studies. We also asked officials at FAA, NATCA, and the Federal Managers
Association to review a draft version of the survey.

To verify the clarity, length of time of administration, and suitability of
the questions, we also pre- tested the questionnaire with selected
controllers at two towers, one en- route center, and the Systems Command
Center in Herndon, Va. A copy of the Survey of Air Traffic Controllers
Retirement and Attrition Plans can be found in Appendix VI.

Survey Administration We conducted a survey between August 2001 and February
2002, using a self- administered mail- out survey. We sent a second
questionnaire on

October 2, 2001, to all initial nonrespondents in order to encourage a
higher response rate. Following this mailing, we experienced an extended
delay in returns until January 9, 2002, because mail delivery was halted on
account of the anthrax contaminations in Washington, D. C. Hence, we
extended the expected cut- off date until February 12, 2002, after a stream
of returns had tapered off. By February 12, 2002, we had 1,591 completed,
usable surveys for an

overall response rate of 76 percent. Some surveys were eliminated because
they (1) had been returned blank, (2) were duplicates from the same
individual, or (3) came from respondents who had left FAA before the
fielding period ended. We used a contractor to create a database of survey
responses. All data were double keyed during the data entry process, and GAO
staff verified a sample of the resulting data to ensure accuracy.

Estimates Estimates produced in this report are for a target population
defined as air traffic controllers in our study population. A very small
proportion (fewer

than 1 percent) of the survey respondents indicated that they were not
classified as Series 2152 air traffic controllers at the time of the survey.
Those respondents are not included in any estimates derived from survey data
in this report; therefore, the final target population for estimation is 19,
880 controllers.

Estimates were formed by weighting the survey responses to account for
effective sampling rates in each stratum. These weights reflect both the
initial sampling rate and the response rate for each stratum. As with most
surveys, our estimation method assumes that nonrespondents would have
answered like the survey respondents.

Sampling Error Because we surveyed a sample of air traffic controllers, our
results are estimates of air traffic controller characteristics and thus are
subject to

sampling errors that are associated with samples of this size and type. Our
confidence in the precision of the results from this sample is expressed in
95- percent confidence intervals. The 95- percent confidence intervals are

expected to include the actual results for 95 percent of the samples of this
type. We calculated confidence intervals for our study results using methods
that are appropriate for a stratified probability sample. For the

percentages presented in this report, we are 95- percent confident that the
results we would have obtained had we studied the entire study population
are within +/- 5 or fewer percentage points of our results, unless otherwise
noted. For example, our survey estimates that 33 percent of the controllers
would retire earlier if there were increased mandatory overtime. The 95
percent confidence interval for this estimate would be no wider than +/- 5
percent, or from 28 percent to 38 percent. For estimates other than

percentages, 95- percent confidence intervals are +/- 10 percent or less of
the value of the estimate, unless otherwise noted.

Nonsampling Error In addition to these sampling errors, the practical
difficulties in conducting surveys of this type may introduce other types of
errors, commonly

referred to as nonsampling errors. For example, questions may be
misinterpreted, the respondents? answers may differ from those of people who
did not respond, or errors could be made in keying completed questionnaires
or in the preparation of data files for analysis. We took

several steps in an attempt to reduce such errors. In addition to the steps
taken during the development of the survey and its administration, we
performed computer analyses to identify inconsistencies and other indicators
of errors, and a second independent analyst reviewed all computer programs.

GAO Survey of Air Traffic Controllers Appendi x VI

INTRODUCTION traffic controller staffing needs. responses. the return
address is

Mr. William R. Chatlos Washington, DC 20548

United States General Accounting Office Survey of Air Traffic Controllers
Retirement and Attrition Plans NOTE: Numbers to the left of each response
are the estimated percentage of all controllers that would have provided
that response based on a sample of 1,591.

The U. S. General Accounting Office ( GAO) , an independent agency of
Congress, has been asked by the Chairman, Subcommittee on Aviation,
Committee on Transportation and Infrastructure, U. S. House of
Representatives, to study the retirement and attrition of air traffic
controllers ( ATC) historically, currently, and planned. A major aspect of
this study is to assess the Federal Aviation Administration s ( FAA) air

As part of our study, we are sending this questionnaire to a randomly
selected sample of current air traffic controllers. Your name was part of
this sample. In this survey you are asked about your plans concerning
continued employment with, or retirement from, the FAA. Your responses will
help to determine what steps need to be taken in order to ensure a smooth
continuation of air traffic control service.

We will present the results of this survey in summary form, taking steps to
safeguard the privacy of your

Individual responses will not be reported in any way that would allow an
individual respondent to be identified. The number on your survey is to help
us track responses and analyze results. This survey can be answered by
checking boxes or filling in blanks; it should take about 10 minutes to
complete.

Please return your completed survey, within the next 10 days, in the
enclosed pre- addressed, postage- paid envelope. In the event that the
envelope is misplaced,

U. S. General Accounting Office 441 G Street, NW, Room 2440F

If you have any questions, please call either David Lichtenfeld at ( 312)
220- 7663 or William Chatlos at ( 202) 512- 7607. Thank you very much for
your time.

1

PROFESSIONAL HISTORY

1. 2. 3. Series 2152? ( Check one. ) ( Check one. )

43.6% 1.

Retirement Eligibility Rules Controllers working at FAA s facilities are
eligible to retire under two sets of retirement rules - the general rules
for federal employees and special provisions for air traffic controllers.

Type of retirement Special Provisions for air traffic controllers 50 20* *
Years actively controlling planes as a Controller or Operations Supervisor,
Under CSRS or FERS Any 25* - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

General Provisions for all federal employees 55 30 Under CSRS 60 20

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - - - - - - Under FERS 55- 57* * 3 * *
For those with 30 years of service or more, retirement eligibility under

6. RETIREMENT: Federal service 87.9% 1.

16. 13. On the planned retirement date you listed in Question 11, how many
years and months of

retirement applicable federal service will you have of the following service
types?

Military service _ _ _ _ _ Yrs & _ _ _ _ Mos 2152 good time _ _ _ _ _ _ Yrs
& _ _ _ _ Mos ( only FAA and

DOD civilian ATC) 2152 Not good time _ _ _ _ _ _ Yrs & _ _ _ Mos Other
Federal Service _ _ _ _ _ Yrs & _ _ _ _ Mos

TOTAL Retirement Applicable Service _ _ _ _ _ Yrs & _ _ _ _ Mos

( Please go to Question 14. )

How might each of the following actions affect the retirement date you
specified in your response to Question 11? ( Check one column for each
action. )

Action 1

a. Increased pay 3.4 % 50.1% 43.8% 2.6% b. Increased mandatory overtime
32.8% 57.0% 2.8% 7.3% c. More opportunities to work staff positions ( not
good time) 2.4% 67.3% 21.6% 8.7% d. Flexibility in shift scheduling (
trading shifts) 0.4% 71.0% 22.2% 6.4% e. Ability to work part time at FAA
3.7% 44.8% 47.4% 4.1% f. Increased work responsibilities 23.5% 71.6% 1.7%
3.1% g. Increased training responsibilities for new staff 18.0% 75.7% 1.8%
4.5% h. Increased air traffic operations 18.4% 74.5% 3.2% 3.8% i. Improved
labor- management relations 1.0% 68.8% 27.0% 3.2% j. Technological changes
in the workplace 1.6% 76.1% 19.6% 2.7% k. Approval of waivers to work as ATC
past the age of 56 0.4% 60.4% 31.1% 8.1%

( Please continue on the back on page 4. )

3 14. Under which of the following retirement

15. systems will you retire? 18.2% 2.

17. To what extent, if at all, did each of the following factors contribute
to your choice of the planned retirement date you specified in response to
Question 11? ( Check one in each row. )

Factors 1 DEMOGRAPHICS

18. What is your specific facility ID code, for example, ZAU, ORD, or C90?

Facility ID Code: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

COMMENTS Little or

No Extent

a. Difficulty getting time off or weekends off 45.4% 14.1% 11.7% 8.9% 12.5%
7.4% b. Will reach the age of 56 and must leave scope 36.2% 5.1% 2.3% 4.1%
17.4% 35.0% c. Near first date eligible to retire 30.6% 10.6% 10.4% 13.9%
24.0% 10.5% d. Dissatisfaction with work 53.7% 15.4% 11.2% 6.2% 6.3% 7.2% e.
Time to move on; date was my choice 41.5% 14.7% 11.4% 11.0% 13.7% 7.7% f.
Completed major responsibilities for children 39.2% 13.3% 11.4% 11.7% 10.6%
13.8% g. Chosen to coincide with spouse s plans/ needs 53.5% 13.4% 7.0% 6.4%
3.3% 16.3% h. Will have enough retirement savings by then 30.7% 18.6% 20.0%
13.3% 11.3% 6.0% i. Other ( Specify. ) 21.3% 0.2% 1.0% 8.8% 41.1% 27.5%

20. In the space provided, please write any comments you would like to make
about retirement and attrition. Thank you for responding to this survey.

Please place your completed survey in the enclosed envelope and mail.

4

3 Moderate

Extent 2 Some Extent

19. What is your sex? 87.7% 1.

Appendi x VII

GAO Contacts and Staff Acknowledgments GAO Contacts Gerald L. Dillingham,
Ph. D. (202) 512- 3650 Glen Trochelman (312) 220- 7729 Staff

In addition to the above, Ruthann Balciunas, William Chatlos, William
Acknowledgments

Doherty, Colin Fallon, David Hooper, Mitch Karpman, David Lehrer, David
Lichtenfeld, Mark Ramage, Raymond Sendejas, Rebecca Shea, and Amy Stewart
made key contributions to this report.

(390005)

GAO?s Mission The General Accounting Office, the investigative arm of
Congress, exists to support Congress in meeting its constitutional
responsibilities and to help improve

the performance and accountability of the federal government for the
American people. GAO examines the use of public funds; evaluates federal
programs and policies; and provides analyses, recommendations, and other
assistance to help Congress make informed oversight, policy, and funding
decisions. GAO?s commitment to good government is reflected in its core
values of accountability, integrity, and reliability.

Obtaining Copies of The fastest and easiest way to obtain copies of GAO
documents at no cost is

through the Internet. GAO?s Web site (www. gao. gov) contains abstracts and
fulltext GAO Reports and

files of current reports and testimony and an expanding archive of older
Testimony

products. The Web site features a search engine to help you locate documents
using key words and phrases. You can print these documents in their
entirety, including charts and other graphics.

Each day, GAO issues a list of newly released reports, testimony, and
correspondence. GAO posts this list, known as ?Today?s Reports,? on its Web
site daily. The list contains links to the full- text document files. To
have GAO e- mail this

list to you every afternoon, go to www. gao. gov and select ?Subscribe to
daily E- mail alert for newly released products? under the GAO Reports
heading.

Order by Mail or Phone The first copy of each printed report is free.
Additional copies are $2 each. A check or money order should be made out to
the Superintendent of Documents. GAO

also accepts VISA and Mastercard. Orders for 100 or more copies mailed to a
single address are discounted 25 percent. Orders should be sent to:

U. S. General Accounting Office 441 G Street NW, Room LM Washington, D. C.
20548

To order by Phone: Voice: (202) 512- 6000 TDD: (202) 512- 2537 Fax: (202)
512- 6061

To Report Fraud, Contact:

Waste, and Abuse in Web site: www. gao. gov/ fraudnet/ fraudnet. htm

E- mail: fraudnet@ gao. gov Federal Programs

Automated answering system: (800) 424- 5454 or (202) 512- 7470 Public
Affairs Jeff Nelligan, managing director, NelliganJ@ gao. gov (202) 512-
4800

U. S. General Accounting Office, 441 G Street NW, Room 7149 Washington, D.
C. 20548

a

GAO United States General Accounting Office

Page i GAO- 02- 591 Air Traffic Control

Contents

Contents

Page ii GAO- 02- 591 Air Traffic Control

Contents

Page iii GAO- 02- 591 Air Traffic Control

United States General Accounting Office Washington, D. C. 20548

Page 1 GAO- 02- 591 Air Traffic Control

A

June 14, 2002 Letter

The Honorable John L. Mica Chairman, Subcommittee on Aviation Committee on
Transportation and Infrastructure House of Representatives

The Honorable William O. Lipinski Ranking Democratic Member, Subcommittee on
Aviation Committee on Transportation and Infrastructure House of
Representatives

In response to your request, this report identifies potential scenarios for
future air traffic controller attrition and FAA?s plans for dealing with
such attrition. This report contains recommendations to the Secretary of
Transportation.

Unless you publicly announce its contents earlier, we plan no further
distribution of this report until 30 days from the date of this letter. At
that time, we will send copies to interested congressional committees; the
Secretary of Transportation; the Administrator, Federal Aviation
Administration; the Secretary of the Air Force; the Secretary of the Army;
the Secretary of the Navy; the Director, Office of Management and Budget;
and the Director, Office of Personnel Management. We will also make copies
available to others upon request.

Please call me at (202) 512- 3650 if you or your staff have any questions
concerning this report. Major contributors to this report are listed in
appendix VII.

Gerald L. Dillingham, Ph. D. Director, Physical Infrastructure Issues

Page 2 GAO- 02- 591 Air Traffic Control

Executive Summary Page 3 GAO- 02- 591 Air Traffic Control

Executive Summary Page 4 GAO- 02- 591 Air Traffic Control

Executive Summary Page 5 GAO- 02- 591 Air Traffic Control

Executive Summary Page 6 GAO- 02- 591 Air Traffic Control

Executive Summary Page 7 GAO- 02- 591 Air Traffic Control

Executive Summary Page 8 GAO- 02- 591 Air Traffic Control

Page 9 GAO- 02- 591 Air Traffic Control

Chapter 1

Chapter 1 Introduction

Page 10 GAO- 02- 591 Air Traffic Control

Chapter 1 Introduction

Page 11 GAO- 02- 591 Air Traffic Control

Chapter 1 Introduction

Page 12 GAO- 02- 591 Air Traffic Control

Chapter 1 Introduction

Page 13 GAO- 02- 591 Air Traffic Control

Chapter 1 Introduction

Page 14 GAO- 02- 591 Air Traffic Control

Chapter 1 Introduction

Page 15 GAO- 02- 591 Air Traffic Control

Chapter 1 Introduction

Page 16 GAO- 02- 591 Air Traffic Control

Chapter 1 Introduction

Page 17 GAO- 02- 591 Air Traffic Control

Chapter 1 Introduction

Page 18 GAO- 02- 591 Air Traffic Control

Chapter 1 Introduction

Page 19 GAO- 02- 591 Air Traffic Control

Chapter 1 Introduction

Page 20 GAO- 02- 591 Air Traffic Control

Page 21 GAO- 02- 591 Air Traffic Control

Chapter 2

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 22 GAO- 02- 591 Air Traffic Control

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 23 GAO- 02- 591 Air Traffic Control

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 24 GAO- 02- 591 Air Traffic Control

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 25 GAO- 02- 591 Air Traffic Control

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 26 GAO- 02- 591 Air Traffic Control

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 27 GAO- 02- 591 Air Traffic Control

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 28 GAO- 02- 591 Air Traffic Control

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 29 GAO- 02- 591 Air Traffic Control

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 30 GAO- 02- 591 Air Traffic Control

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 31 GAO- 02- 591 Air Traffic Control

Chapter 2 FAA Is Facing Increased Controller Hiring because of Higher
Staffing Levels and

Growing Attrition Page 32 GAO- 02- 591 Air Traffic Control

Page 33 GAO- 02- 591 Air Traffic Control

Chapter 3

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 34 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 35 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 36 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 37 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 38 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 39 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 40 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 41 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 42 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 43 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 44 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 45 GAO- 02- 591 Air Traffic Control

Chapter 3 FAA Needs a More Comprehensive Workforce Plan for Air Traffic
Controllers

Page 46 GAO- 02- 591 Air Traffic Control

Page 47 GAO- 02- 591 Air Traffic Control

Appendix I

Appendix I Potential Impacts of Proposed Changes to Increase Air Traffic
Control Annuity

Calculations Page 48 GAO- 02- 591 Air Traffic Control

Appendix I Potential Impacts of Proposed Changes to Increase Air Traffic
Control Annuity

Calculations Page 49 GAO- 02- 591 Air Traffic Control

Page 50 GAO- 02- 591 Air Traffic Control

Appendix II

Appendix II Air Traffic Controller Schools

Page 51 GAO- 02- 591 Air Traffic Control

Page 52 GAO- 02- 591 Air Traffic Control

Appendix III

Page 53 GAO- 02- 591 Air Traffic Control

Appendix IV

Appendix IV Methodology for Computer Simulation

Page 54 GAO- 02- 591 Air Traffic Control

Appendix IV Methodology for Computer Simulation

Page 55 GAO- 02- 591 Air Traffic Control

Page 56 GAO- 02- 591 Air Traffic Control

Appendix V

Appendix V Methodology for GAO?s Survey of Air Traffic Controllers?
Retirement and Attrition Plans

Page 57 GAO- 02- 591 Air Traffic Control

Appendix V Methodology for GAO?s Survey of Air Traffic Controllers?
Retirement and Attrition Plans

Page 58 GAO- 02- 591 Air Traffic Control

Appendix V Methodology for GAO?s Survey of Air Traffic Controllers?
Retirement and Attrition Plans

Page 59 GAO- 02- 591 Air Traffic Control

Page 60 GAO- 02- 591 Air Traffic Control

Appendix VI

Appendix VI GAO Survey of Air Traffic Controllers

Page 61 GAO- 02- 591 Air Traffic Control

Appendix VI GAO Survey of Air Traffic Controllers

Page 62 GAO- 02- 591 Air Traffic Control

Appendix VI GAO Survey of Air Traffic Controllers

Page 63 GAO- 02- 591 Air Traffic Control

Page 64 GAO- 02- 591 Air Traffic Control

Appendix VII

United States General Accounting Office Washington, D. C. 20548- 0001

Official Business Penalty for Private Use $300

Address Service Requested Presorted Standard

Postage & Fees Paid GAO Permit No. GI00
*** End of document. ***