Rail Safety: The Federal Railroad Administration Is Taking Steps
to Better Target Its Oversight, but Assessment of Results Is
Needed to Determine Impact (26-JAN-07, GAO-07-149).
Since 1980, the train accident rate has improved significantly,
but progress has leveled off over the past 10 years. Recent
serious accidents--such as one in Graniteville, South Carolina,
that led to 9 deaths and 292 injuries--elevated concerns. The
Federal Railroad Administration (FRA) develops safety standards
and inspects and enforces railroads' compliance with these
standards. This report addresses how FRA (1) focuses its efforts
on the highest priority risks related to train accidents in
planning its oversight, (2) identifies safety problems on
railroad systems in carrying out its oversight, and (3) assesses
the impact of its oversight efforts on safety. To complete this
work, GAO reviewed FRA regulations, planning and policy
documents, and safety data. GAO also contacted FRA officials in
headquarters and three regional offices and others.
-------------------------Indexing Terms-------------------------
REPORTNUM: GAO-07-149
ACCNO: A65187
TITLE: Rail Safety: The Federal Railroad Administration Is
Taking Steps to Better Target Its Oversight, but Assessment of
Results Is Needed to Determine Impact
DATE: 01/26/2007
SUBJECT: Accident prevention
Accidents
Agency evaluation
Inspection
Performance measures
Policy evaluation
Railroad accidents
Railroad industry
Railroad regulation
Railroad safety
Risk management
Safety regulation
Safety standards
Strategic planning
Transportation
Government agency oversight
Policies and procedures
Program goals or objectives
FRA National Rail Safety Action Plan
******************************************************************
** 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-07-149
* [1]Results in Brief
* [2]Background
* [3]FRA Has Made Progress in Targeting Its Oversight Efforts on
* [4]National Rail Safety Action Plan Provides Strategy for Addre
* [5]FRA Is Making a Number of Efforts to Reduce Accidents Caused
* [6]FRA Is Pursuing Several Initiatives to Improve the Detection
* [7]FRA Has Made Progress in Targeting Its Inspections on the Ba
* [8]FRA Relies Primarily on Direct Inspections to Identify Safet
* [9]FRA's Oversight Identifies a Range of Site-Specific and Broa
* [10]Several Other Organizations Have Implemented Comprehensive A
* [11]FRA Measures Its Progress in Achieving a Variety of Safety G
* [12]FRA Has Established a Range of Safety Goals and Uses Coopera
* [13]Performance Measures Support FRA's Oversight, but Informatio
* [14]FRA Has Made Changes in Response to Evaluations but Has Not
* [15]Conclusions
* [16]Recommendations for Executive Action
* [17]Agency Comments
* [18]Appendix I: Scope and Methodology
* [19]Other Aspects of Our Work
* [20]Appendix II: FRA Addresses Safety Problems through Cooperati
* [21]Appendix III: Oversight of Risk Management in the U.S. Commu
* [22]Appendix IV: GAO Contact and Staff Acknowledgments
* [23]GAO Contact
* [24]Staff Acknowledgments
* [25]Order by Mail or Phone
Report to Congressional Committees
United States Government Accountability Office
GAO
January 2007
RAIL SAFETY
The Federal Railroad Administration Is Taking Steps to Better Target Its
Oversight, but Assessment of Results Is Needed to Determine Impact
GAO-07-149
Contents
Letter 1
Results in Brief 3
Background 6
FRA Has Made Progress in Targeting Its Oversight Efforts on the Basis of
Risk 15
FRA Relies Primarily on Direct Inspections to Identify Safety Problems and
Does Not Oversee Railroads' Management of Safety Risks 27
FRA Measures Its Progress in Achieving a Variety of Safety Goals, but Has
Limited Information on the Direct Results of Its Oversight 39
Conclusions 50
Recommendations for Executive Action 50
Agency Comments 51
Appendix I Scope and Methodology 52
Other Aspects of Our Work 53
Appendix II FRA Addresses Safety Problems through Cooperation and
Enforcement 56
Appendix III Oversight of Risk Management in the U.S. Commuter Railroad,
U.S. Pipeline, and Canadian Railroad Industries 62
Appendix IV GAO Contact and Staff Acknowledgments 65
Tables
Table 1: FRA's Inspection Disciplines 13
Table 2: Key FRA Initiatives Aimed at Addressing Main Causes of Train
Accidents 17
Table 3: FRA's Safety Performance Measures 43
Table 4: Description of FRA Enforcement Actions and Frequency of Use,
January 2005 through July 2006 57
Table 5: Risk Management Framework 62
Figures
Figure 1: Train Accident Rates, 1980 through 2005 2
Figure 2: Train Miles Traveled, 1996 through 2005 7
Figure 3: Train Injuries and Deaths, 1980 through 2005 8
Figure 4: Primary Causes of Train Accidents, 2005 10
Figure 5: Human Factor and Track Accident Rates, 1980 through 2005 11
Figure 6: FRA Inspector Measuring Track Gauge 29
Figure 7: FRA Inspector Inspecting Train Cars 30
Figure 8: Inspections Conducted and Defects and Violations Cited, by
Inspection Discipline, in 2005 32
Figure 9: How FRA's Inspection and Enforcement Programs Contribute to Rail
Safety 41
Figure 10: FRA's Civil Penalties, 1996 through 2005, in 2005 Dollars 59
Abbreviations
APTA American Public Transportation Association
FRA Federal Railroad Administration
NTSB National Transportation Safety Board
PHMSA Pipeline and Hazardous Materials Safety Administration
This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.
United States Government Accountability Office
Washington, DC 20548
January 26, 2007
The Honorable Patty Murray
Chairman
The Honorable Christopher Bond
Ranking Minority Member
Subcommittee on Transportation, Housing and Urban Development,
and Related Agencies
Committee on Appropriations
United States Senate
The Honorable John Olver
Chairman
The Honorable Joe Knollenberg
Ranking Minority Member
Subcommittee on Transportation, Housing and Urban
Development, and Related Agencies
Committee on Appropriations
House of Representatives
Since 1980, the overall safety record in the railroad industry, as
measured by the rate of train accidents, has improved markedly. (See fig.
1.) However, during the past decade, the rate of improvement has leveled
off, and a number of serious accidents in 2004 and 2005 elevated concerns
about railroad safety. For example, in January 2005, a train carrying
hazardous materials collided with a standing train in Graniteville, South
Carolina, resulting in 9 deaths and 292 injuries and requiring the
evacuation of 5,400 people. In response to the leveling off of the train
accident rate, this accident, and other serious train accidents, in May
2005, the federal railroad safety regulator, the Federal Railroad
Administration (FRA), announced a National Rail Safety Action Plan to
improve its safety oversight. The plan outlines strategies for FRA to
target the most frequent, highest-risk causes of accidents, focus federal
oversight and inspection resources, and research the use of technologies
in order to improve rail safety. Since 1980, the overall safety record in
the railroad industry, as measured by the rate of train accidents, has
improved markedly. (See fig. 1.) However, during the past decade, the rate
of improvement has leveled off, and a number of serious accidents in 2004
and 2005 elevated concerns about railroad safety. For example, in January
2005, a train carrying hazardous materials collided with a standing train
in Graniteville, South Carolina, resulting in 9 deaths and 292 injuries
and requiring the evacuation of 5,400 people. In response to the leveling
off of the train accident rate, this accident, and other serious train
accidents, in May 2005, the federal railroad safety regulator, the Federal
Railroad Administration (FRA), announced a National Rail Safety Action
Plan to improve its safety oversight. The plan outlines strategies for FRA
to target the most frequent, highest-risk causes of accidents, focus
federal oversight and inspection resources, and research the use of
technologies in order to improve rail safety.
Figure 1: Train Accident Rates, 1980 through 2005
Citing concerns over several serious accidents that occurred in 2005, the
Senate Appropriations Committee directed that we assess FRA's oversight
approach. Accordingly, this report concentrates on how FRA (1) focuses its
efforts on the highest priority risks related to train accidents in
planning its safety oversight, (2) identifies safety problems on railroad
systems in carrying out its oversight, and (3) assesses the impact of its
oversight efforts on safety.
To determine how FRA focuses its efforts on the highest priority risks
related to train accidents in planning its safety oversight, we reviewed
the agency's National Rail Safety Action Plan, plans for developing new
regulations, inspection planning documents, and other key planning
documents related to targeting its oversight at these risks. We also
discussed oversight planning efforts with FRA officials. To assess how FRA
identifies safety problems on railroad systems in carrying out its
oversight, we reviewed FRA's statutory authority; regulations; and
policies, procedures, and guidance for conducting inspections and
identifying safety issues. We reviewed data from FRA on its inspection
activities for the period from 1996 through 2005. We also examined risk
management principles and safety oversight approaches used by other modal
administrations within the Department of Transportation and elsewhere that
have similar safety missions in order to determine their possible
application to FRA. To determine how FRA assesses the impact of its
oversight efforts on safety, we examined FRA safety performance measures
and evaluations of its oversight activities, determined how FRA uses this
information in making decisions about its oversight, and reviewed our
products on performance measurement and evaluation.
As part of our work in each of these areas, we interviewed program and
enforcement officials at FRA headquarters, as well as at three regional
offices covering states with the highest train accident rates. We also
discussed FRA's approach to safety oversight with representatives of
railroads, unions, and state railroad safety organizations. Our work
focused on FRA oversight efforts to reduce the rate of train accidents
rather than those to reduce highway-rail crossing and trespassing
accidents because (1) the Department of Transportation's Inspector General
has recently assessed efforts to reduce highway-rail crossing accidents^1
and (2) trespassing accidents primarily involve issues not related to
railroad safety performance. We also focused on FRA's oversight of
railroads rather than its oversight of nonrailroad companies (such as
shippers of hazardous materials by rail) because the agency's oversight
efforts focus primarily on railroads. In addition, according to FRA, most
recent serious train accidents involving the release of hazardous
materials have resulted from problems with railroad operations. As part of
our review, we assessed internal controls and the reliability of FRA's
data on its inspection activities and enforcement actions that were
pertinent to these efforts. We determined that the data elements were
sufficiently reliable for our purposes. We conducted our work from
November 2005 through January 2007 in accordance with generally accepted
government auditing standards. (See app. I for additional information on
our scope and methodology.)
Results in Brief
In planning its safety oversight, FRA is focusing its efforts on the
highest priority risks related to train accidents through various
initiatives aimed at addressing the main causes of these accidents as well
as through improvements in its inspection planning approach. The agency's
overall strategy for targeting its oversight at the greatest risks is the
National Rail Safety Action Plan. This plan provides a reasonable
framework for guiding the agency's efforts to improve its oversight. It
includes initiatives to address the two main causes of train
accidents--human factors and defective track--and FRA has pursued some
additional initiatives to address these causes since issuing the plan.^2
These initiatives--which include new regulations, research on new
technologies and approaches for improving safety, and new vehicles for
inspecting track--have the potential to reduce the risks associated with
these causes and thereby reduce the rate of train accidents. For example,
FRA has recently issued proposed regulations that, if finalized, will
enable the agency to take enforcement actions when railroad employees do
not follow key railroad operating procedures for ensuring safety and,
therefore, may help to reduce some common types of accidents caused by
human factors. However, most of these initiatives have not yet been fully
implemented, and, while some may start showing results in the next year or
two, their overall impact on safety will probably not be apparent for a
number of years. Furthermore, the ability of many of these efforts to
improve safety will depend on voluntary actions by railroads, such as the
adoption of a model FRA has developed to improve train crew scheduling
practices in order to prevent worker fatigue. In addition, the National
Rail Safety Action Plan announced a new approach for planning inspections
that relies on greater use of accident and inspection data and other data.
Under this approach, which FRA has been using for over a year, inspectors
focus their efforts on locations that, according to data-driven models,
are likely to have safety problems. This new planning approach allows FRA
to use its inspectors more effectively and better target the greatest
safety risks.
^1See U.S. Department of Transportation, Office of the Inspector General,
Audit of Oversight of Highway-Rail Grade Crossing Accident Reporting,
Investigations and Safety Regulations, MH-2006-016 (Washington, D.C., Nov.
28, 2005); and Report on the Audit of the Highway-Rail Grade Crossing
Safety Program, MH-2004-065 (Washington, D.C., June 16, 2004).
In carrying out its safety oversight, FRA identifies safety problems on
railroad systems mainly through routine inspections that determine whether
operating practices, track, and equipment, such as signals and
locomotives, are in compliance with safety standards. Through this
approach, FRA inspectors identify a range of safety problems at various
sites on railroads' systems. FRA also identifies some broad-scale problems
that affect multiple sites, primarily through analyses of accident and
inspection data, internal discussions, and some nonroutine inspections.
However, the number of FRA and state inspectors is small relative to the
size of the industry and FRA inspections are able to cover only about 0.2
percent of railroads' operations each year.^3 Also, these inspections
focus on compliance with minimum standards and are not designed to
determine how well railroads are managing safety risks throughout their
systems that could lead to accidents. The American Public Transportation
Association (APTA), the Pipeline and Hazardous Materials Safety
Administration (PHMSA), and Transport Canada have implemented approaches
to oversee the management of safety risks by U.S. commuter railroads, U.S.
pipelines, and Canadian railroads, respectively. Risk management can be
described as a continuous process of managing--through the systematic
identification, analysis, and control of risks associated with safety
hazards (such as train collisions)--the likelihood of hazards' occurrence
and their negative impact.^4 These oversight approaches complement, rather
than replace, traditional compliance inspections and, therefore, provide
additional assurance of safety. However, because we believe that FRA's
initiatives to reduce train accident rates need time to mature and
demonstrate their effects on safety, we are not recommending that FRA
adopt an approach for overseeing railroads' management of safety risks.
^2Generally, human factors are behaviors that affect job performance, such
as incorrectly setting switches. According to FRA, the term human factors
refers broadly to the role of human participation in any system and to the
ways in which human beings positively or negatively contribute towards
system performance.
FRA uses a broad range of goals and measures to assess the impact of its
oversight efforts on safety. For example, it has developed new goals to
target its inspection and enforcement efforts at reducing various types of
railroad accidents and related measures to track its progress. These
measures include the level of train accidents caused by human factors,
track defects, and equipment defects, both nationwide and within each of
its eight regions. These safety performance measures provide FRA with much
information that it uses to make decisions about its oversight approach.
However, its ability to make informed decisions is limited because it
lacks measures of the direct results of its inspection and enforcement
programs, such as measures of the extent to which these programs have
resulted in the correction of identified safety problems or improvements
in compliance. We have found that it is a useful practice for agencies to
establish measures of programs' direct results to help show their
contributions to the ultimate results the agencies seek to achieve.
Furthermore, while FRA has made some changes in its oversight approach in
response to external and internal evaluations, the agency has not
evaluated the effectiveness of its enforcement program in achieving
desired results. Evaluations can provide a broader range of information on
program performance and how to improve it than performance measures alone.
Both performance measures and evaluations can provide valuable information
on program results that helps hold agencies accountable for the
performance of their programs.
^3This figure is an FRA estimate, based on an estimation of the amount of
activity, such as train movements, on each railroad in the United States.
^4Risk is the combination of the likelihood and the consequence of a
specified hazard being realized.
To improve FRA's ability to determine the extent to which its inspection
and enforcement programs are contributing to rail safety and whether
changes in these programs are needed, we are recommending that FRA develop
and implement measures of the direct results of its inspection and
enforcement programs and evaluate its enforcement program. We provided a
draft of this report to the Department of Transportation for its review
and comment. The department did not offer overall comments on the draft
report or its recommendations. It did offer several technical comments,
which we incorporated where appropriate.
Background
Railroads are an important component of the transportation system,
transporting about 42 percent of the nation's freight (as measured by
weight). For passenger movement, railroads play a much smaller role than
do highway and air travel; however, communities are looking to different
forms of public transit for relief, particularly as highways become
increasingly congested. Demand continues to grow for both freight and
passenger transportation on railroads. In 2005, railroads traversed 790
million train miles,^5 an increase of 18 percent since 1996. (See fig. 2.)
Moreover, the Department of Transportation estimates that between 1998 and
2020, the amount of freight transported by rail (as measured by weight)
will increase by about 50 percent. Commuter and intercity passenger
railroads have also grown--providing over 522 million passenger trips in
2005, compared with 385 million in 1995. According to the Federal Transit
Administration, as of 2006, seven more commuter rail systems throughout
the country were being planned or designed.
^5A train mile is the movement of a train a distance of 1 mile.
Figure 2: Train Miles Traveled, 1996 through 2005
The railroad industry is primarily composed of 7 large freight railroads
(called Class I railroads); about 560 smaller freight railroads (called
Class II and III railroads); and 118 passenger, commuter, tourist,
excursion, and other railroads.^6 Within the industry, Class I freight
railroads predominate, representing about 93 percent of railroad freight
revenue and 69 percent of the total U.S. rail mileage. Class II and Class
III railroads include a number of regional and short line railroads that
provide freight transportation. Class II regional railroads typically
operate 400 to 650 miles of track in a region spanning two to four states,
whereas Class III short lines typically perform point-to-point service
over short distances.
On average, about 446 people have been injured and 14 people have been
killed each year over the past decade, from 1996 through 2005, exclusive
of highway-railroad grade crossing and trespassing accidents. Despite
overall improvements since 1980, gains have tapered off. Since 1992, the
accident rate has remained at about or sometimes more than four accidents
per million train miles. (See fig. 1.) In recent years, a number of
serious accidents raised concerns about the level of safety in the
railroad industry. Train accidents resulted in 1,884 injuries and 15
deaths in 2002, and 733 injuries and 33 deaths in 2005. (See fig. 3.) FRA
officials attributed the large rise in number of injuries and deaths in
2002, and subsequently in 2005, to one or two major accidents, as opposed
to a series of accidents. For example, in 2002, a derailment in Minot,
North Dakota, led to the release of approximately 200,000 gallons of
anhydrous ammonia, 1,442 injuries and 1 death. In 2005, a train collision
in Graniteville, South Carolina, resulted in the evacuation of 5,400
people, 292 injuries and 9 deaths.
^6For 2005, the Surface Transportation Board has defined Class I railroads
as railroads earning adjusted annual operating revenues of $319.3 million
or more. Class II railroads are those earning between $25.5 million and
$319.3 million, and Class III railroads are those earning less than $25.5
million.
Figure 3: Train Injuries and Deaths, 1980 through 2005
Note: Injuries were far more common than deaths during the period of 1980
through 2005; deaths ranged from 4 to 67 per year. Figures do not include
highway-railroad crossing and trespassing incidents.
Although Class I railroads have a lower rate of accidents than Class II
and III railroads, because of their size, they account for most train
accidents; in 2005, Class I railroads were involved in 76 percent of train
accidents.^7 Moreover, Class I railroads were involved in 53 percent of
injuries and 58 percent of deaths during that year.
According to FRA data, the majority of train accidents are attributable to
either human factors or track defects. (See fig. 4.) For 2005, FRA data
show that human factors and track defects were the primary causes of 72
percent of all train accidents. Those accidents caused by human factors
often result from actions such as improperly positioning switches, shoving
cars without properly checking for safe conditions, leaving cars in a
position that obstructs track, and failing to secure a sufficient number
of handbrakes. Those accidents caused by defective track often result from
defective or ineffective crossties; broken or worn switch points; and
broken, fissured or fractured rail components.
^7FRA defines a train accident as any collision, derailment, fire,
explosion, act of God, or other event involving operation of railroad
on-track equipment (standing or moving) that results in reportable damages
greater than the current reporting threshold to railroad on-track
equipment, signals, track, track structures, and roadbed. The threshold
for 2006 was $7,700.
Figure 4: Primary Causes of Train Accidents, 2005
^aInteraction of lateral and vertical forces refers to a specific type of
accident that occurs when a lateral, or sideways, force exceeds a vertical
force, or gravity, by 50 percent or more. Lateral force may be caused by
travel through curves and alignment imperfections in the track.
^bThe miscellaneous category of accident causes contains a number of
subcategories, none of which exceeds 1.5 percent.
As have overall accident rates, improvements in human factor and track
accident rates have leveled off over the past decade, achieving their
lowest rates in 1996 and 1995, respectively. (See fig. 5.) According to
FRA, the increase in the human factor accident rate in 2004 can be
attributed to increases in accidents caused by employees not following
railroad operating rules. (These rules specify various operating
procedures, such as the proper positioning of track switches, to ensure
safe operations.) However, according to FRA, these types of human factor
accidents decreased by 21 percent from the first half of 2005 to the first
half of 2006. Officials attribute this decrease largely to actions the
agency has taken to focus railroads' attention on this problem,^8
including issuing a safety
and equipment, such as wheel safety requirements and formulas that
determine maximum allowable speeds on curved track. However, some
regulations are not prescriptive and allow railroads greater flexibility
in determining how to comply. For example, if a railroad chooses to
implement a processor-based train control system or product,^10 it must
develop and receive FRA's approval for a railroad safety program plan. The
plan should include a description of risk assessment procedures and the
safety assessment process, and railroads have flexibility to adjust their
programs to accommodate the specific system or product change. In
addition, some regulations require railroads to develop and implement
safety programs, such as accident and incident reporting programs and
roadway worker protection programs.
^8In addition, according to FRA, train accident rates for 2006 will
probably appear slightly more favorable than those for 2005 because of a
single, large increase in the dollar level reporting threshold for the
year 2006. Specifically, starting in 2006, railroads are required to
report accidents resulting in railroad property damage exceeding $7,700,
up from the previous threshold of $6,700.
In 1996, FRA adopted a more participatory approach to rulemaking through
the creation of the Railroad Safety Advisory Committee. This committee is
designed to include all segments of the rail community in developing
solutions to safety regulatory issues. Currently, the committee consists
of 39 member organizations, including representatives from railroads,
railroad associations, labor, states, and agencies with railroad
regulatory responsibility in Canada and Mexico.
FRA conducts inspections to determine railroads' compliance with its
regulations. Typically, inspectors conduct inspections at specific sites
of railroad operations. For example, inspectors examine track, equipment,
devices, employee actions, or procedures and may review records maintained
by the railroad in order to determine the railroad's compliance with FRA
regulations. Inspectors generally specialize in one of five areas, called
inspection disciplines: (1) operating practices, (2) track, (3) hazardous
materials, (4) signal and train control, and (5) motive power and
equipment. (See table 1.) FRA's policy is to cite defects for most
instances of noncompliance and to encourage the railroad to comply
voluntarily.
^10A processor-based train control system or product is one that is
dependent for its proper functioning on a digital processor, such as an
onboard signal or switch control.
Table 1: FRA's Inspection Disciplines
Examples of what inspections cover (not all
Inspection Discipline inclusive)
Operating practices Railroad operations related to human factors,
including employee compliance with railroad
operating rules, railroads' monitoring of this
compliance, drug and alcohol testing of
employees, employees' hours of service, radio
communications, locomotive engineer
qualification, and accident and incident
reporting.
Track and structures^a Condition of track and structures, including
track components and geometry, railroad track
inspections, and programs to maintain continuous
welded rail track and protect roadway workers.
Hazardous materials Rail transportation of hazardous materials,
including the integrity, markings, maintenance,
and placement of tank cars; the training of train
crews; security; and emergency preparedness.
Signal and train control Signal switching systems, locomotive signal
devices, locks and brake application, including
related recordkeeping, testing, modifications,
and repairs.
Motive power and Design and operation of railroad rolling
equipment equipment, including railroad freight and
passenger car safety, locomotive safety and
maintenance, safety devices, brake system safety,
and emergency preparedness procedures.
Source: GAO analysis of FRA information.
Notes: FRA has recently established a new inspection discipline,
industrial hygiene, which covers the protection of railroad employees on
the job, including enforcement of FRA standards for occupational safety
and health. The discipline is also responsible for the operation of FRA's
internal occupational safety and health program for ensuring FRA employee
safety and health. The resources devoted to this inspection discipline are
relatively small--FRA plans to have a total of 5 industrial hygienist
inspectors. In comparison, the other inspection disciplines each have
between 50 and 90 inspectors, approximately.
aIn addition to its manual inspections, FRA has an automated track
inspection program that uses data produced by vehicles that precisely
measure track geometry.
When railroads do not comply voluntarily or identified defects are
serious, FRA may cite violations and assess civil penalties*either against
railroads or individuals*or take other enforcement actions to promote
compliance with its regulations. FRA developed the concept of focused
enforcement in the mid-1990s to ensure that inspectors know which
violations pose the greatest hazards and make enforcement decisions
accordingly. The purpose of this approach is to concentrate FRA's
enforcement efforts on the areas with the greatest potential safety
benefits. FRA's enforcement policy, as stated in its regulations,
specifies that before recommending penalties, inspectors should consider
the seriousness of the condition or act, the potential safety hazards, and
the current level of compliance of the individual or railroad, among other
things. FRA is authorized to negotiate civil penalties with railroads and
exercises this authority. For example, it settles claims annually with
each Class I railroad. FRA uses civil penalties as its primary enforcement
tool. However, it also has other available enforcement tools. These
include compliance agreements and compliance orders, special notices for
repair, emergency orders, criminal penalties, disqualification orders and
injunctions. (See app. II for further discussion.)
FRA conducts additional oversight of Class I railroads through the
Railroad System Oversight program, established in October 2005, which
addresses safety issues not subject to regulation, such as aspects of
worker fatigue, as well as safety compliance issues. Under this program,
the agency assigns an FRA manager for each Class I railroad to cooperate
with it on identifying and resolving safety issues. These managers act as
liaisons with the railroads and labor officials, analyze accident and
inspection data for their assigned railroad, and support FRA's inspection
and enforcement efforts. Finally, under this program, FRA has begun annual
meetings with the leadership of each Class I railroad to discuss its
safety performance.
The Railroad System Oversight program replaced FRA's Safety Assurance and
Compliance Program, which had emphasized using teams--consisting of FRA
inspectors and other FRA officials, railroad officials, and union
representatives--to identify and resolve safety issues, as a complement to
FRA's regular inspection activity. The agency envisioned this program as a
comprehensive approach to safety through which these representatives would
work together to identify and correct the root causes of problems across
an entire railroad. According to FRA, the program had a number of
accomplishments, such as improving collaboration among management, labor
and FRA, and encouraging railroads' voluntary cooperation in taking
corrective action on safety issues not covered under FRA's safety
regulations. However, FRA ended this program in response to concerns that
it had lost its effectiveness, inhibited enforcement actions, and shifted
regional resources away from conducting site-specific inspections, the
mainstay of FRA's safety program.
In addition to these activities, FRA conducts other types of safety
oversight aimed at reducing train accidents, such as monitoring of
railroad safety data, accident investigations and reviews and
investigations of complaints, and education efforts aimed at small
railroads.^11 Furthermore, FRA funds research and development that
supports its safety oversight, by, for example, assisting in the
development of new regulations and the revision of existing regulations.
FRA is a small agency, especially in relation to the industry it
regulates. As of July 2006, FRA had 657 full-time and part-time safety
staff, including about 400 inspectors in the field (in its regional,
district, and local offices). In addition, 30 state oversight agencies
participate in a partnership program with FRA to conduct safety oversight
activities at railroads' operating sites. Currently, about 160 state
inspectors work with FRA to conduct inspections and other investigative
and surveillance activities as needed. In contrast, the railroad industry
has about 235,000 employees,^12 219,000 miles of track in operation,
158,000 signals and switches, and over 1.6 million locomotives and cars.
^11For example, FRA conducts twice yearly training seminars for short line
railroads on how to comply with its regulations and operate safely.
FRA Has Made Progress in Targeting Its Oversight Efforts on the Basis of Risk
In planning its safety oversight, FRA focuses its efforts on the highest
priority risks related to train accidents through a number of initiatives
aimed at addressing the main causes of these accidents as well as through
improvements in its inspection planning approach. The agency's overall
strategy for targeting its oversight at the greatest risks--the National
Rail Safety Action Plan--provides a reasonable framework for guiding the
agency's efforts. FRA's various initiatives to address the two main causes
of train accidents--human factors and defective track--are promising.
However, most of these initiatives are not yet fully implemented and their
overall impact on safety will probably not be apparent for a number of
years. FRA has also recently implemented new approaches for planning its
inspection activity--based on analyses of accident, inspection, and other
data--that allow it to better target the greatest safety risks and more
effectively use its inspectors.
National Rail Safety Action Plan Provides Strategy for Addressing Highest
Priority Risks, but Impact of Safety Initiatives Is Not Yet Clear
FRA's May 2005 National Rail Safety Action Plan provides an overarching
framework for the agency's efforts to target its oversight at the highest
priority risks. The agency developed the plan in response to a leveling
off of the train accident rate in recent years and the occurrence of
serious train accidents in 2004 and early 2005. The plan outlines a number
of initiatives aimed at reducing the main types of train accidents, those
caused by human factors and track defects.^13 These efforts include some
innovative approaches for the railroad industry, such as a pilot project
for reporting close calls, that are designed to prevent accidents by
addressing safety problems that may cause them.^14 Other efforts include
new regulations, several research endeavors, and new track inspection
vehicles. We have not reviewed these individual initiatives in depth, but
believe that the plan provides a reasonable strategy for guiding and
prioritizing FRA's efforts to reduce the rates of accidents attributable
to human factors and track defects.
^12This number does not include contractor employees hired by the
railroads.
^13The National Rail Safety Action Plan also includes initiatives to
improve hazardous materials safety and emergency response capability. Most
of the serious accidents involving the release of hazardous material that
have occurred in the last several years have been caused by human factors
or track defects.
Some of the efforts outlined in the action plan are underway, and some are
planned for the near future. FRA issued a progress report to the Secretary
of Transportation on the action plan in June 2006 and intends to continue
to report on the plan's implementation. Since issuing the plan, the agency
has pursued additional initiatives to target risks posed by human factors
and track defects, including issuing new track regulations in response to
a congressional mandate and encouraging a new braking technology. (See
table 2.) The combined initiatives hold promise for reducing the risks
associated with human factors and track defects. However, these efforts
are in varying stages of development or implementation and their overall
impact on safety will probably not be apparent for a number of years. Some
individual initiatives, such as the close call reporting project, may
start showing results in the next year or two.
^14According to FRA, a close call is an opportunity to improve safety
practices in a situation or incident that has a potential for more serious
consequences. It represents a situation in which an ongoing sequence of
events was stopped from developing further, preventing the occurrence of
potentially serious safety-related consequences.
Table 2: Key FRA Initiatives Aimed at Addressing Main Causes of Train
Accidents
Cause of train
accidents Initiative FRA action and time frame
Human factors Rail Safety Action Plan
initiatives
Regulations on employee Issued proposed regulation in
compliance with October 2006. Plans to issue final
railroad operating regulation by the end of 2007.
rules
Pilot project to Plans to initiate this 5-year
establish voluntary project in January 2007 at one
reporting system to Class I railroad site. FRA is
learn from close call seeking the participation of other
incidents railroads.
Research on worker Discussed results of the research
fatigue to develop a with key stakeholders and released
model that could be a report on the results in November
used to improve crew 2006.
scheduling
Other initiatives
Pilot project to Is considering establishing a pilot
establish voluntary project in fiscal year 2008.
risk management
programs at railroad
worksites
Encouraging new braking Released a study on new brake
technology that can system technology in August 2006
prevent or reduce human and plans to issue proposed
error regulations in May 2007 to
facilitate its use.
Track defects Rail Safety Action Plan
initiatives
New track inspection Plans to deploy two new vehicles in
vehicles early 2007.
Research and Demonstrated a prototype system in
development on new October 2005. Enhanced the system
inspection technologies in 2006 and is planning further
enhancements in 2007.
Other initiatives
Additional regulations Issued a final rule on rail joint
on continuous welded inspection in October 2006. Plans
rail track to develop additional regulations
for improving management of this
type of track, but has not yet
developed time frames for this
effort.
Source: GAO analysis of FRA data.
The National Rail Safety Action Plan also outlines the agency's
development of a new approach for planning its inspections, based on
greater use of its accident and inspection data. Since issuing the plan,
FRA has made other efforts to improve its approach for planning its
inspections.^15 Starting in fiscal year 2006, FRA's new inspection
planning approach has allowed the agency to better target its inspections
on the basis of risk and to better coordinate inspection planning among
its headquarters and regional offices.
^15Other key agency plans that focus efforts on the highest priority risks
related to train accidents include the department's rulemaking agenda,
strategic plan and annual performance plan, and FRA's performance budget.
The rail safety goals and measures contained in the performance plan and
performance budget are discussed later in this report.
FRA Is Making a Number of Efforts to Reduce Accidents Caused by Human Factors
Human factor accidents result from unsafe acts of individuals, such as
employee errors, and can occur for a number of reasons, such as employee
fatigue or inadequate supervision, training, or staffing.^16 FRA has
recently issued proposed regulations aimed at reducing the most common
causes of such human factor train accidents: improper positioning of track
switches or derails,^17 shoving rail cars without properly monitoring for
safe conditions or controlling the movement, and leaving rail cars in a
position that obstructs adjacent track. Procedures for employees to follow
to avoid these types of accidents are contained in railroads' operating
rules.^18 Currently, FRA regulations contain general requirements that
railroads train employees on their operating rules and periodically test
their compliance with these rules, but do not specifically require that
employees follow the rules that can prevent these types of accidents.^19
As a result, according to FRA officials, the agency has had a limited
ability to cite noncompliance and take enforcement actions in this area.
The proposed regulations mirror established railroad operating rules that
require employees to follow procedures, such as procedures related to the
positioning of track switches, that if followed, could prevent these types
of accidents. In addition, they include further requirements for railroads
to train employees on these rules and monitor their compliance with these
rules. According to FRA, these new requirements and its ability to enforce
them will make railroad employees more accountable for following operating
rules and railroad management more accountable for ensuring that employees
do so.
^16Management decisions at the organizational level, such as decisions
regarding the allocation of resources or crew scheduling, can have
consequences in the workplace that can contribute to human factor
accidents.
^17Derails are devices used to prevent the obstruction of track by
unauthorized movements of trains or unattended rolling stock.
^18Most Class I railroads use one of two sets of standard rules: the
Northeast Operating Rules Advisory Committee rulebook and the General Code
of Operating Rules. Railroads must file their operating rules with FRA.
^19In a few cases, FRA's regulations do require some practices, such as
securing a sufficient number of handbrakes, that are in railroad operating
rules. Also, the Switching Operations Fatality Analysis working
group--made up of representatives of FRA, unions, railroads, and the
Department of Transportation's Volpe National Transportation Systems
Center--studies fatalities that occur to workers engaged in switching
operations and recommends ways that such events can be prevented.
FRA is also sponsoring a 5-year Confidential Close Call Reporting System
pilot project, through which employees of participating railroads can
provide confidential information on close calls. A neutral party, the
Bureau of Transportation Statistics, will maintain the close call data and
a team of representatives from the participating railroad, labor
organizations, FRA, and the bureau will review these data to identify
safety problems. Railroads will be expected to correct identified problems
in order to prevent accidents. The purpose of this project is to determine
the effectiveness of such a voluntary reporting system for the railroad
industry. FRA has developed plans to monitor and evaluate the performance
of the project over time; these plans include short-term and long-term
performance measures. The agency anticipates that it will have early
indications of how the program is affecting safety in the next year or
two. To date, one Class I railroad has committed to participate in this
project, at one yard on its system, and, according to FRA, two others have
expressed strong interest. Such systems have contributed to significant
reductions in accidents in some other industries, such as aviation.
FRA and the National Transportation Safety Board (NTSB) have identified
employee fatigue as a significant factor in many train accidents. Railroad
employees often work long hours and have unpredictable and fluctuating
work schedules. Under current law, these employees could potentially work
a maximum of 11 hours and 59 minutes, followed by 8 hours off duty, and
then another 11 hours and 59 minutes on duty, continually.^20 In addition,
time spent waiting for transportation at the end of a tour of duty and
being transported to a release point, called limbo time, does not count as
either duty or off-duty time and can be significant. FRA has sponsored a
study to develop a fatigue model that could be used to improve train crew
scheduling practices, has discussed the draft results with railroads and
labor organizations, and released the final report on the study in
November 2006. The agency is also taking some other actions to encourage
railroads to improve their management of employee fatigue, such as
providing funding for a new program, in use at a Class I railroad, that
tracks and analyzes crew scheduling to remedy practices that could
contribute to fatigue.
^20The law also specifies that train employees are required to have 10
consecutive hours off duty following 12 continuous hours on duty.
Through a recent investigation of a 2004 train accident in which three
people died, NTSB found that the engineer and conductor were likely asleep
at the controls and recommended that FRA require railroads to use
scientifically based principles when assigning work schedules for train
crew members and to limit crew member limbo time.^21 In recent testimony,
the FRA Administrator noted that, several times in the 1990s, the
Department of Transportation proposed legislation to repeal or reform the
hours-of-service law or to require railroads to develop fatigue management
plans, but that these bills encountered opposition and were not passed.
However, since that time, FRA has not submitted such legislation. The
agency has not yet responded to NTSB's recommendations, but has told us
that it intends to tell the board, as it has in the past, that the agency
lacks jurisdiction to issue regulations addressing hours of service.^22
While we were conducting our work, FRA was considering establishing a
pilot project that would use risk management to help reduce human factor
accidents at selected railroad worksites. Risk management can be described
as a systematic approach for identifying, analyzing, and controlling
risks.^23 The agency envisioned that, under such a project, each worksite
would collect and analyze data on precursors to human factor
accidents--such as close call incidents, employee errors, or
organizational characteristics--to better identify and correct individual
and organizational factors that contribute to such accidents and therefore
reduce the risks of such accidents occurring. The agency proposed that
funding for this project be included in its fiscal year 2008 budget
request and this proposal was approved by the department. In January 2007,
as we were finalizing our report, FRA told us that it had decided to
expand the scope of this project to include efforts to use risk management
to reduce all types of accidents, not only human factor accidents.
^21NTSB, Collision of Union Pacific Railroad Train MHOTU-23 With BNSF
Railway Company Train MEAP-TUL-126-D With Subsequent Derailment and
Hazardous Materials Release, Macdona, Texas, June 28, 2004, Railroad
Accident Report NTSB/RAR-06/03 (Washington, D.C., 2006).
^22According to FRA, it is the only safety regulatory agency in the
Department of Transportation that lacks regulatory authority over worker
duty hours.
^23Risk is the combination of the likelihood and the consequence of a
specified hazard (or threat) being realized. We have developed a framework
for risk management based on industry best practices. See app. III for a
discussion of this framework as well as comprehensive risk management
approaches in use by several other transportation agencies for overseeing
the U.S. commuter railroad, U.S. pipeline, and Canadian railroad
industries.
FRA has examined possible approaches to use for this project. According to
agency officials, one possible approach is represented by a new program
that a Class I railroad has implemented at two locations on its system. In
this program--which has received funding from FRA--a committee of
employees documents employee behaviors that could lead to unsafe
conditions, without recording names, and provides feedback to the
responsible employees to help them eliminate these behaviors. FRA has also
considered the approach used by the Occupational Safety and Health
Administration's Voluntary Protection Program. This program recognizes
individual worksites with exemplary safety records and practices,
including the identification, analysis, prevention, and control of
workplace hazards that could lead to employee injuries and illnesses.^24
As envisioned by FRA officials, this project will focus on establishing
risk management programs at three separate railroad worksites and will
include close monitoring and evaluation of these programs to determine
their impact in reducing accidents over a 5-year period. If the pilot is
successful, FRA anticipates establishing a voluntary risk management
program for the railroad industry, which would encourage railroads to
implement this type of approach on a systemwide basis in order to reduce
human factor accidents, as well as other types of accidents.
Finally, FRA has recently issued a study on a new braking technology,
electronically controlled pneumatic brakes, which improves train-handling
and decreases stopping distances by 40 to 60 percent. These brakes use an
electronic line to uniformly command brake applications and releases
throughout the train. FRA has decided that it will develop new regulations
to facilitate the use of this technology over the next decade. According
to FRA officials, improving railroads' braking systems can have a
significant safety benefit by improving the ability of locomotive
engineers to control their trains and, therefore, avoid or reduce the
severity of some types of human factor-caused accidents.^25
24In reviewing the Voluntary Protection Program of the Occupational Safety
and Health Administration, along with several other voluntary compliance
programs, we found that benefits reported by participating worksites
included reduced injury and illness rates, an improved safety culture, and
improved employee-management relations. See GAO, Workplace Safety and
Health: OSHA's Voluntary Compliance Strategies Show Promising Results, but
Should Be Fully Evaluated before They Are Expanded, [26]GAO-04-378
(Washington, D.C.: Mar. 19, 2004).
The above initiatives use a variety of approaches, some quite innovative,
for addressing the causes of human factor accidents. These initiatives,
which are in varying stages of development or implementation, have the
potential to eventually reduce these types of accidents. However, while
some may start showing results in the next year or two, their overall
impact will likely not be apparent for a number of years. Furthermore, all
of these initiatives, except for the proposed regulations on operating
rules, depend on voluntary actions by railroads, and, in some cases, labor
as well, for their success. For example, the impact of FRA's effort to
develop a model to address the problem of worker fatigue depends on the
extent to which railroads eventually use this model to improve train crew
scheduling practices. FRA has worked with railroads and labor on some of
these initiatives, but it is too early to predict their outcomes.
FRA Is Pursuing Several Initiatives to Improve the Detection and Management of
Track Defects
Railroads operate trains on about 219,000 miles of track across the United
States. This track consists of traditional jointed rail as well as newer
rails that are smooth bands of welded steel, called continuous welded
rail. Derailments can occur when rails are uneven or too wide apart or
when rails or joint bars are cracked or broken. FRA inspects track
conditions through manual inspections conducted on-foot or in on-track
equipment, and with automated track inspection vehicles that precisely
measure track and can identify problems that are difficult to detect
through other types of inspections. The agency operates one automated
track inspection vehicle that it uses in inspecting track and plans to add
two more for this purpose in early 2007.^26 According to FRA, these
additional vehicles will allow the agency to triple the miles of track
that it is able to inspect per year, to nearly 100,000 miles.^27
25FRA has also recently issued standards for processor-based positive
train control systems. These systems are an advanced train control
technology that can prevent train collisions through automatic brake
applications. They also can provide enhanced protection for
maintenance-of-way workers.
^26FRA also has two additional automated track inspection vehicles that
are primarily used for research activities but occasionally are used for
inspections.
^27However, FRA may inspect some sections of track more than once a year.
FRA is also developing an automated inspection system for improving the
detection of cracks in joint bars. Such cracks can lead to a
derailment-causing break but can be difficult to detect through simple
visual inspections. Specifically, FRA has designed and is refining a
high-resolution video inspection system that can be used in on-track
inspection equipment and will improve detection of these defects.^28
According to FRA, the technology is ready for use, was demonstrated and
refined in the field in 2006, and will undergo further enhancements in
2007. The agency expects that railroads will make use of the technology
and is reviewing how to use it in its own inspections. According to FRA,
one Class I railroad is starting to make use of this technology and others
have shown strong interest in it.
Finally, in response to a congressional mandate and NTSB recommendations,
FRA has recently finalized regulations that require track owners to
conduct detailed and periodic inspections of rail joints in continuous
welded rail track. Although FRA issued regulations in 1998 requiring
railroads to develop and implement procedures for the inspection and
maintenance of continuous welded rail track, a number of train accidents
occurred since that time in which the failure of a rail joint on this type
of track was a factor.^29 FRA officials told us that the railroads'
overall management of the condition of continuous welded rail track is a
major concern for the agency because about 20 accidents involving problems
with this type of track occur per year and these accidents are usually
serious. FRA has estimated that continuous welded rail track represents
between 99,000 and 120,000 miles of the 219,000 miles of track in
operation in the United States. The agency is working with the Railroad
Safety Advisory Committee to develop additional regulations to improve
railroads' management of this type of track.
These initiatives have the potential to reduce accidents caused by track
defects. FRA's deployment of two new track inspection vehicles in early
2007 will enable the agency to significantly expand its ability to monitor
the condition of the nation's track. The agency's development of an
automated inspection system has the potential to help prevent derailments,
provided that railroads make use of this technology. Finally, FRA's
issuance of new regulations related to continuous welded rail track and
development of further regulations in this area should eventually improve
railroads' management of this type of track, although time frames for the
development of new regulations are not yet clear.
^28FRA also has a number of other ongoing efforts to improve track
inspection capabilities. For example, since 2002 it has sponsored a Rail
Integrity Task Force--composed of experts in the railroad industry, FRA,
and the Department of Transportation's Volpe National Transportation
Systems Center--to identify best practices for rail inspection,
maintenance, and replacement.
^29In continuous welded rail track, rails are welded together to form one
continuous rail that may be several miles long. There may be joints in
this rail for several reasons, including the need for insulated joints
that electrically separate track segments for signaling purposes and the
need to replace a section of defective rail.
FRA Has Made Progress in Targeting Its Inspections on the Basis of Risk
Like other modal safety administrations within the Department of
Transportation, FRA has few resources for overseeing railroads compared
with the scope of its responsibility. According to agency officials, it
inspects 0.2 percent of railroad operations per year. FRA has developed a
new approach--the National Inspection Plan--for using available data to
target its inspections at the greatest safety risks. The agency began
using the new approach for three of its inspection disciplines (operating
practices, track, and motive power and equipment) in October 2005 and
expanded it to the remaining two disciplines (hazardous materials and
signal and train control) by March 2006. The purpose of the plan is to
optimize FRA's ability to reduce the rates of various types of train
accidents as well as releases of hazardous materials. The plan provides
guidance to each regional office on how its inspectors, who each
specialize in one of the five inspection disciplines, should divide up
their work by railroad.
Developing the plan involves two steps. In the first step, FRA
headquarters produces an initial plan for each of the agency's eight
regions. This plan specifies, by inspection discipline, numeric goals for
the level of inspection activity to allocate to each railroad, by state.
These numeric goals are derived from models--based on trend analyses of
accident, inspection, and other data--that predict, by inspection
discipline, locations where train accidents and incidents are likely to
occur within each region and provide the optimal allocation of inspection
resources to prevent accidents.^30 FRA has developed separate models for
each inspection discipline based on how well individual data elements,
such as historical information on inspection results, tend to predict
accidents. According to FRA officials, they expect to refine this new
planning process to reflect lessons learned during the first year of its
implementation.
^30We did not evaluate these models or the data on which they are based.
FRA defines train incidents as events involving the movement of railroad
equipment that results in a casualty but does not cause damage above the
reporting threshold established for train accidents, which was $7,700 in
2006.
In the second step, the regional administrators are allowed to adjust the
goals for their region on the basis of local knowledge and emerging
issues, such as recent accidents. However, according to FRA officials,
there were only a few such adjustments for fiscal year 2006. Throughout
the year, FRA monitors how the regions are meeting their goals. Starting
in fiscal year 2007, regional administrators will have a second
opportunity to adjust their inspection plans at midyear to respond to
safety issues that emerged during the first 6 months of the year.
Previously, FRA had a less structured, less consistent, and less
data-driven approach for planning inspections. According to agency
officials, each region prepared its own inspection plan, based on
judgments about appropriate priorities and analysis of available data.
However, the use of data was not consistent from region to region.
Inspectors had greater discretion about where to inspect and based
decisions about priorities on their knowledge of their inspection
territories.
The National Inspection Plan covers federal inspectors, but not state
inspectors. Other than funding training and computer equipment, FRA does
not provide funding for state inspection activities.^31 Therefore, each
state makes its own decisions about how to use its inspectors. FRA
officials told us that the agency has not included states in the National
Inspection Plan because it does not have authority to tell the states what
inspections to conduct. The 30 states that participate in FRA's state
program have varying numbers of inspectors and most conduct inspections in
some, but not all, of FRA's five inspection disciplines. According to FRA,
its regional offices coordinate with the states in their region to avoid
duplication of effort. The regional administrators may make adjustments to
their National Inspection Plan goals based on the work of state inspectors
within their region.
The National Inspection Plan also does not establish priorities across
regions and inspection disciplines, but rather, for given staffing levels
for each discipline within each region, assigns inspection levels to
railroads and states. However, FRA eventually plans to use its results to
help decide how to optimally allocate additional inspectors, as vacancies
occur or new positions are funded. According to headquarters officials,
the National Inspection Plan model played a role in a decision to allocate
additional inspection staff to the operating practices discipline in the
regions.^32 However, officials told us that they will need more time to
determine how well the plan is working before using it to reallocate
resources among the regional offices.
^31FRA provides training for state inspectors and certifies them as
qualified to perform inspections and cite violations.
The fiscal year 2006 plan resulted in various reallocations of inspection
activity within FRA's regional offices. These reallocations have allowed
FRA to better target its inspections on the basis of risk. For example, in
the track area, in some cases regions are focusing more attention on
certain railroads that have higher accident rates and worse track
conditions than others. Conversely, in the area of operating practices,
some regional offices have decreased their focus on certain railroads that
have shown good or improving performance in this area compared with other
railroads.
In fiscal year 2006, in addition to implementing the National Inspection
Plan, the agency implemented a new coordinated approach for planning
nonroutine inspection activity, by inspection discipline. Examples of
these types of inspections include in-depth inspections by a regional
office of a railroad's compliance with track standards; interregional
inspections of compliance with certain regulations, such as those related
to bridge safety, of a Class I railroad that operates in multiple regions;
and headquarters-led inspections of Class I railroads' drug and alcohol
testing programs. Some of these planned inspections are based on analyses
of data on accidents by railroad, accident causes, and inspection results
in order to define, beyond the National Inspection Plan goals, what
railroad locations and specific regulatory requirements warrant increased
attention by inspectors. Others, such as inspections of some required
railroad safety programs, are performed periodically. According to FRA
officials, the regional offices and headquarters previously planned such
inspections separately and made less use of data in their planning. Under
this new planning approach, FRA headquarters and regional offices
coordinate in developing plans for inspections they will conduct, by
inspection discipline. These plans are compiled into an agencywide plan,
and then FRA tracks the completion of these inspections.
^32Overall, the motive power and equipment discipline currently has the
highest number of inspectors (86), followed by the operating practices
discipline (76), the track and structures discipline (73), the signal and
train control discipline (61), and the hazardous materials discipline
(55). However, to help reduce human factor accidents, which account for
the highest percentage of train accidents, inspectors in the motive power
and equipment discipline conduct some inspections of operating practices.
FRA's new approaches for planning its inspection activity allow it to
better target the greatest safety risks and coordinate inspection planning
among its eight regional offices and headquarters offices. Therefore, they
allow FRA to make more effective use of its inspectors. However, it is not
yet clear whether these new planning approaches will lead to a
prioritization of inspection levels across regions and inspection
disciplines or improved safety.
FRA Relies Primarily on Direct Inspections to Identify Safety Problems and Does
Not Oversee Railroads' Management of Safety Risks
In carrying out its safety oversight, FRA identifies safety problems on
railroad systems mainly through routine inspections to determine whether
operations, track, and equipment, such as signals and locomotives, are in
compliance with safety standards. Through this approach, FRA inspectors
identify a range of safety problems at various sites on railroads'
systems, through citing defects and violations. FRA also identifies some
broad-scale compliance problems that affect multiple sites, mainly through
analyses of accident and inspection data, internal discussions, and some
nonroutine inspections.
FRA's inspections focus on compliance with minimum standards within five
separate inspection disciplines and do not attempt to determine how well
railroads are managing safety risks on their systems. APTA, PHMSA, and
Transport Canada have implemented approaches to oversee the management of
safety risks by U.S. commuter railroads, U.S. pipelines, and Canadian
railroads, respectively. These oversight approaches complement, rather
than replace, traditional compliance inspections and therefore provide
additional assurance of safety. However, we are not recommending that FRA
adopt such an oversight approach, since the agency is currently pursuing
various initiatives to reduce train accident rates. In our view, these
initiatives need time to mature to demonstrate their effects and,
subsequently, an informed assessment would need to be made about whether
additional actions are warranted.
FRA's Oversight Identifies a Range of Site-Specific and Broad-Scale Problems on
Railroad Systems
Overseeing the safety of the railroad industry is a huge task. FRA's 400
inspectors, along with about 160 state inspectors, oversee 686 railroads,
with about 235,000 employees, 219,000 miles of track, 24,000 locomotives,
1.6 million cars, 158,000 signals and switches, and 240,000 highway-rail
grade crossings.^33 As noted previously, according to FRA officials, the
agency's inspectors are able to directly observe only about 0.2 percent of
the railroad industry's operations per year. FRA carries out this
oversight responsibility primarily through inspections of railroads'
compliance with its safety standards at various locations on railroads'
systems and through cooperation and enforcement aimed at resolving
identified problems. During inspections, which are generally conducted
separately within the five inspection disciplines, inspectors examine
railroads' compliance with a broad range of federal standards. Inspectors
discuss identified compliance problems (called defects) with railroads to
achieve voluntary compliance, and cite violations--recommending that the
agency take enforcement action against a railroad--when they determine
that the problems are serious or when a railroad does not voluntarily
comply. (See app. II for a description of FRA's use of cooperation and
enforcement to resolve safety problems and improve safety.) In addition,
FRA's Railroad System Oversight managers work with Class I railroads and
labor to identify and resolve some safety problems that are not directly
related to compliance with the agency's regulations. For example, one
manager worked with a railroad and labor organization to improve the
railroad's program for communicating with roadway workers to ensure that
they are aware of and implement key safety procedures.
FRA primarily monitors railroads' compliance through routine inspections
by individual inspectors at specific sites on railroads' systems.^34 As
discussed previously, FRA inspects locations likely to have safety
problems, which it identifies using accident and previous inspection data
as well as other information. Inspectors typically cover a range of
standards within their discipline during these inspections. This
inspection approach focuses on direct observations of specific components
of the train, related equipment, and railroad property--including the
track and signal systems--as well as operating practices to determine
whether they meet FRA's standards. (See figs. 6 and 7.) Inspectors also
examine railroads' inspection and maintenance records. The railroads have
their own inspectors who are responsible for ensuring that railroad
equipment, track, and operations meet federal rail safety standards. For
example, FRA requires that railroads inspect brake systems, signal
systems, passenger equipment, track conditions, and train crews' adherence
to operating rules, among other things. According to FRA officials, the
agency's inspectors often review the railroads' records of inspection to
determine whether the records accurately represent the types of problems
FRA inspectors are finding during their own inspection activities.
^33These figures do not include railroad contractor employees or track
that is not in use.
^34Inspectors also sometimes travel along part of a railroad's system in
conducting inspections. According to FRA, from 2002 to 2005, inspectors
traveled between two locations in 17 percent of the routine inspections of
railroads.
Figure 6: FRA Inspector Measuring Track Gauge
Figure 7: FRA Inspector Inspecting Train Cars
FRA also conducts more in-depth inspection efforts that generally focus on
railroads' compliance in a particular area, such as their inspections of
employees' adherence to operating rules. These efforts often involve a
team conducting separate inspections at multiple sites, generally within
one of FRA's eight regions.^35 FRA focuses these in-depth inspection
efforts on emerging issues, identified through previous routine
inspections or analyses of accident data. FRA also periodically conducts
in-depth inspections of some systemwide programs that the railroad is
required to implement, such as employee drug and alcohol testing programs
and accident and incident reporting programs. In some cases, FRA may
conduct a systemwide in-depth set of inspections on a railroad to
determine its overall compliance within a single inspection discipline or
in several disciplines. For example, in early 2006, in response to a Class
I railroad's high accident rates, FRA conducted an in-depth set of
inspections of the railroad's compliance with operating practices, track,
signal and train control, motive power and equipment, and hazardous
materials regulations across its system. However, according to an FRA
headquarters official, the agency does not frequently perform systemwide
or multidisciplinary inspections.
^35However, in some cases, FRA conducts nationwide inspections of
railroads' compliance with specific requirements. For example, in
mid-2006, it began a set of nationwide inspections of various railroads'
compliance with requirements for notifying train crews of the types of
hazardous materials being transported on their trains, after identifying
noncompliance in this area as a problem.
In 2005, federal and state inspectors conducted a total of about 63,000
inspections.^36 According to FRA, routine inspections constituted about 75
percent of the inspections of railroads and in-depth inspections accounted
for about 11 percent. The remainder of these inspections (14 percent)
consisted of other types of activities, such as investigations of
accidents and complaints. Inspectors in the track discipline performed the
most inspections, followed by those in the motive power and equipment,
operating practices, hazardous materials, and signal and train control
disciplines.^37 This approach to oversight enables FRA inspectors and
managers to identify a wide range of safety problems. Inspectors identify
specific compliance problems--conditions that do not meet FRA's
standards--at sites they visit by citing defects. Inspectors cite
violations for those defects that they believe warrant enforcement action.
They consider a number of factors in making this decision, including the
railroad's history of compliance at that location and the seriousness of
the noncompliance (such as whether it is likely to cause accidents,
injuries, or releases of hazardous materials). Inspectors in some
disciplines cite more defects and violations than others. (See fig. 8.)
Overall, FRA and state inspectors cited about 293,000 defects and about
9,500 violations during the 63,000 inspections conducted in 2005.
^36This number includes inspections of railroads as well as of
nonrailroads (companies that ship hazardous materials by rail, tank car
manufacturers, and tank car repairers). In 2005, inspections of
nonrailroads represented 7 percent of all inspections.
^37To help reduce accidents caused by human factors, which are the leading
cause of train accidents, FRA's motive power and equipment inspectors
conduct some inspections to look for operating practices problems that can
lead to these types of accidents.
Figure 8: Inspections Conducted and Defects and Violations Cited, by
Inspection Discipline, in 2005
Note: These figures include inspections carried out by both federal and
state inspectors. Inspectors are instructed to cite defects for most
instances of noncompliance found, but have discretion in determining which
instances to cite as violations warranting enforcement action.
The motive power and equipment discipline cites almost half of all defects
and over a third of all violations. FRA officials told us that the
standards in this inspection discipline are the most prescriptive, making
defects and violations easier to find. However, these types of defects
cause a much smaller proportion of accidents than human factors and track
defects.^38 (See fig. 4.) The most frequently cited violations include
those for noncompliance with standards regarding locomotives and freight
cars, track conditions, recordkeeping on the inspection and repair of
equipment and track, and the condition of hazardous materials tank cars.
^38FRA officials have explained that operating practices inspectors have
had a limited ability to cite defects and violations because of the way
regulations in this area are written. For example, as noted previously,
the regulations contain general requirements about railroads' programs for
inspecting employees' adherence to operating rules and do not specifically
require that employees follow these rules. The agency expects that its
proposed regulations on operating rules will improve its ability to
enforce in this area, because the requirements will be more stringent than
existing regulations.
While individual defects and violations are generally for compliance
problems identified at specific locations on railroads' systems, FRA also
identifies broad-scale compliance problems, by inspection discipline, that
affect multiple locations on a railroads' system. It does so mainly
through analyses of accident data and data on defects and violations found
during inspections, communications among managers at headquarters and in
its eight regional offices on the results of analyses and inspections, and
further inspections to obtain more information about identified problems.
Agency officials told us that they hold frequent internal discussions
about emerging issues to determine where problems are occurring and plan
actions that the agency should take to address them. The agency's Railroad
System Oversight managers support this effort by analyzing systemwide
accident and inspection data, by inspection discipline, for the Class I
railroads to identify trends and emerging issues. FRA may plan and conduct
in-depth inspections to determine the scope of such issues. For example,
if the rate of human factor accidents has increased at various locations
on a railroad's system, FRA may conduct inspections of operating practices
at these locations.
Examples of broad-scale problems FRA has identified at railroads include
weak implementation within a particular state of a program for monitoring
employees' adherence to operating rules, poor inspections by a railroad of
its track in a particular region, systemic problems in reporting accidents
and incidents, and defective equipment across a railroad's system. In some
cases, FRA inspectors identify some higher-level management issues, such
as a lack of supervision or inadequate training of railroad personnel,
which could have led to the compliance problems. According to FRA
officials, they discuss broad-scale compliance problems with railroad
officials to try to get these problems resolved. For example, after an
in-depth inspection, inspectors meet with railroad managers to discuss
overall problems found and, according to headquarters officials, usually
provide a written summary of those problems. Efforts to cooperate with the
railroad to resolve broad-scale problems may be combined with enforcement
actions, usually civil penalties, for specific violations identified at
individual sites. According to FRA officials, the agency always conducts
follow-up on serious problems it has identified to ensure that they are
resolved.^39
While FRA does track and maintain data on various types of train accidents
and incidents as well as defects and violations cited by inspectors and
enforcement actions taken, the agency does not centrally track the
broad-scale compliance problems it has identified.^40 These problems are
described in some agency documents, such as inspectors' summary reports on
findings of in-depth inspection efforts; various reports prepared by the
Railroad System Oversight managers on the Class I railroads;^41 and the
agency's overall plan, by region and inspection discipline, for its
nonroutine inspection activity. As explained in the next section, this
lack of centralized tracking can impede the ability of the agency to
measure the effectiveness of its efforts to resolve identified broad-scale
problems.
FRA's five Railroad System Oversight managers also identify some broad
safety problems at Class I railroads that are not related to compliance.
They identify these problems mainly through contacts with labor and
railroad officials and FRA regional officials and try to address them
through cooperation with the railroad. For example, these managers have
worked with railroads in addressing labor's concerns about practices for
transporting train crews at the end of their shifts that may worsen
fatigue and programs for training employees on the railroads' operating
rules. In some cases, these problems were first identified under the
Safety Assurance and Compliance Program.
^39App. II contains a description of FRA's efforts to resolve such
problems through discussions with railroad officials as well as
enforcement actions. According to FRA officials, the agency uses
compliance agreements, which require railroads to take significant actions
to improve their ability to comply, when broad-scale compliance problems
are egregious and have not been resolved through other methods. FRA has
issued eight such agreements since 2000. These agreements have mainly
focused on compliance problems in the operating practices and track
disciplines.
^40According to FRA officials, its ability to track broad-scale compliance
problems is limited because its existing databases related to safety are
not integrated. The agency has an effort underway to better integrate its
existing data. The next section contains a description of this effort.
^41The Railroad System Oversight managers track the status of
nonregulatory problems, as well as some regulatory problems, that they are
working on and, in response to our request, created papers for us
describing systemwide and regional issues for each of the Class I
railroads. In addition, these managers produce quarterly and annual
reports showing safety trends--based mainly on accident data--for each of
the Class I railroads.
Several Other Organizations Have Implemented Comprehensive Approaches for
Overseeing the Management of Safety Risks in Transportation Industries
FRA officials have noted that their approach of directly inspecting safety
conditions and targeting locations that are most likely to have compliance
problems provides a safety net and holds railroad management accountable.
However, because the number of FRA and state inspectors is small relative
to the size of railroad operations, FRA inspections can cover only a very
small proportion of railroad operations (0.2 percent). Also, FRA targets
inspections at locations on railroads' systems where accidents have
occurred, among other factors, rather than overseeing whether railroads
systematically identify and address safety risks that could lead to
accidents.
Rail transportation poses a variety of potential safety hazards, including
collision or derailment; injury to workers, passengers, or nearby
residents; and damage to property or the environment. Risk management is a
systematic approach for dealing with the risks posed by such safety
hazards and has been used in the private and public sectors for decades.
It can be described as a continuous process of managing--through the
systematic identification, analysis, and control of risks associated with
hazards (or threats)--the likelihood of their occurrence and their
negative impact.^42 A framework for risk management based on industry best
practices and other criteria that we have developed divides risk
management into five major phases: (1) setting strategic goals and
objectives, and determining constraints; (2) assessing risks; (3)
evaluating alternatives for addressing these risks; (4) selecting the
appropriate alternatives; and (5) implementing the alternatives and
monitoring the progress made and results achieved.^43 Risk management can
help to improve systemwide safety by systematically identifying and
assessing risks associated with various safety hazards and prioritizing
them so that resources may be allocated to address the highest risks
first. It also can help in ensuring that the most appropriate alternatives
to prevent or mitigate the effects of hazards are designed and
implemented.
^42Risk is the combination of the likelihood and the consequence of a
specified hazard being realized. In risk management, the term "threat" is
sometimes used in place of hazard.
^43GAO, Risk Management: Further Refinements Needed to Assess Risks and
Prioritize Protective Measures at Ports and Other Critical Infrastructure,
[27]GAO-06-91 (Washington, D.C.: Dec. 15, 2005); Homeland Security:
Summary of Challenges Faced in Targeting Oceangoing Cargo Containers for
Inspection, [28]GAO-04-557T (Washington, D.C.: Mar. 31, 2004); and Rail
Security: Some Actions Taken to Enhance Passenger and Freight Rail
Security, but Significant Challenges Remain, [29]GAO-04-598T (Washington,
D.C.: Mar. 23, 2004).
Other transportation oversight organizations have developed and
implemented approaches for overseeing industries' overall management of
safety risks. In particular, during the last 10 years, APTA, PHMSA, and
Transport Canada have developed and implemented such oversight approaches
for U.S. commuter railroads, U.S. pipelines, and Canadian railroads,
respectively. These approaches complement, rather than replace,
traditional compliance inspections. APTA established a U.S. commuter
railroad oversight program in 1996, in partnership with FRA and the
commuter rail industry, that supplements FRA's inspections of these
railroads.^44 Under this program, APTA provides guidelines to these
railroads on managing the safety of their systems--including safety
risks--and audits their plans for and implementation of this management
approach. Beginning in 2000, PHMSA issued a series of requirements for
pipeline operators to develop "integrity management" programs to manage
risk in areas--such as those that are densely populated--where leaks or
ruptures could have the greatest impact on public safety.^45 The agency's
integrity management regulations supplement its minimum safety
regulations, and it inspects operators' compliance with both types of
standards. In Canada, the department responsible for overseeing railroad
safety, Transport Canada, in 2001 began requiring that railroads establish
safety management systems that include risk management.^46 Transport
Canada assesses these systems as well as railroads' compliance with its
traditional safety standards.^47 (For further information on GAO's risk
management framework and these oversight approaches, see app. III.)
^44APTA is a nonprofit organization representing the transit industry,
including U.S. commuter rail systems. APTA offered to develop this program
after FRA directed passenger railroads to develop system safety plans for
addressing hazards associated with passengers occupying the lead units of
a train. The APTA program is more extensive and is intended to cover all
aspects of system safety. FRA issued this directive in 1996, in an
emergency order, after two passenger train accidents caused 14 deaths.
^45PHMSA administers the national regulatory program to ensure the safe
transportation of hazardous liquids and natural gas by pipeline. PHMSA and
FRA are similar in several respects. For example, both oversee large
industries with relatively few inspectors. Both also oversee industries
that have relatively few deaths, injuries, and accidents (as compared to
transportation as a whole), making additional safety gains more difficult.
^46Transport Canada oversees the safety and security of Canada's rail,
marine, highway, and aviation operations.
These oversight approaches are intended to provide additional assurance of
safety beyond that provided by inspections of compliance with minimum
safety standards. They supplement uniform, minimum standards by
encouraging or requiring companies to identify and address their unique
safety risks. APTA, PHMSA, and Transport Canada have emphasized that risk
management provides a higher standard of performance than traditional
safety regulation based on compliance alone. According to APTA officials,
their approach helps companies to prioritize their actions for addressing
risk and therefore optimize safety within constraints of their resources.
According to PHMSA officials, pipeline companies' compliance with the
agency's traditional regulations ensures minimum safety performance, but
its integrity management approach has improved the ability of these
companies to systematically address the full range of safety threats to
their pipelines. According to Transport Canada officials, by encouraging
systemwide improvements in companies' safety performance in order to
address systemic causes of specific problems, its new approach helps the
agency to leverage its resources. Transport Canada has emphasized that
risk management ensures that risks are being adequately addressed and
should point railroads to areas where they could undertake initiatives
beyond their current practices that could improve their overall safety
performance. Transport Canada officials also told us that a primary
objective of their new approach is for railroads to assume more
responsibility for the safety of their operations. Similarly, APTA
officials told us that their oversight approach is proactive because it
encourages companies to identify and address potential hazards before
accidents occur.
We have reviewed PHMSA's gas transmission pipeline integrity management
oversight approach and have recently concluded that it enhances public
safety.^48 We also found that representatives from the pipeline industry,
safety advocacy groups, and state agencies generally agree that this
approach improves public safety. Operators told us that the primary
benefit of the program is the comprehensive knowledge they acquire about
the condition of their pipelines. APTA and Transport Canada officials have
told us that their oversight approaches have not been formally evaluated
to determine their effectiveness. However, according to FRA officials,
APTA's system safety oversight approach has strengthened safety program
management in the commuter rail sector. Finally, Transport Canada is
expanding its safety management system approach to its oversight of civil
aviation.
^47In addition, the European Commission has funded an effort to develop
proposed guidelines for a safety management system for rail companies of
its member nations. These proposed guidelines recommend that European
railways incorporate some basic risk management elements--including safety
performance targets, risk assessment and control, and an internal audit
process--in their safety management systems. See E.M. El Koursi, L. Tordai
and J. Rodriguez. European Commission Fifth Framework Programme, SAMNET
Thematic Network, SAMNET Synthesis Report, Safety Management and
Interoperability (SAMNET, February 2006).
While FRA does not oversee railroads' overall approach for managing safety
risks on their systems, it has taken some steps in a limited number of
areas to oversee and encourage risk management in the railroad industry.
For example, the agency has several regulations in place that require
railroads to use a risk-based approach for managing safety in specific
areas: the operation of high-speed passenger trains, the fire safety of
new passenger cars and locomotives, and the adoption of new
processor-based signal and train control technologies. In addition, PHMSA,
in consultation with FRA and the Transportation Security Administration,
has recently proposed a regulation that could lead to greater FRA
oversight of railroads' management of hazardous materials risks.^49 FRA
has also issued guidance for passenger railroads on assessing collision
hazards and risks and developing strategies for addressing them.^50 In
addition, FRA is currently working with APTA and some commuter railroads
to improve these railroads' abilities to conduct collision hazard
analyses. Finally, as discussed earlier, FRA is considering establishing a
pilot project to examine how a risk management approach could be used in
the railroad industry, on a voluntary basis, to reduce human factor
accidents as well as other types of accidents.
^48GAO, Natural Gas Pipeline Safety: Integrity Management Benefits Public
Safety, but Consistency of Performance Measures Should Be Improved,
[30]GAO-06-946 (Washington, D.C.: Sept. 8, 2006).
^49As noted previously, FRA enforces PHMSA's hazardous material
regulations as they relate to the transportation of such materials by
rail. Specifically, under this proposed regulation, which was issued in
December 2006, railroads would be required to compile annual data on
certain shipments of hazardous materials that are particularly hazardous,
use the data to analyze safety and security risks along the rail
transportation routes where those materials are transported, assess
alternative routing options, and base routing decisions on those
assessments.
^50FRA has also worked with Amtrak, which does not participate in APTA's
program, to assist it in developing a system safety plan that is
consistent with APTA's guidelines.
Although FRA is taking some steps to encourage increased use of risk
management in the railroad industry, oversight of railroads' overall
approach for managing safety risks on their systems, in addition to FRA's
existing discipline-specific compliance-based oversight, has the potential
to provide additional assurance of safety. Such an approach could help to
ensure that railroads systematically identify and address the full range
of risks on their systems and could also encourage railroads to take on
more responsibility for safety. According to agency officials, FRA is
concerned that railroads too often wait for inspectors to show up before
addressing problem areas, while FRA would prefer that they find and fix
problems on their own. However, developing and implementing such a new
oversight approach would be a major undertaking for the agency and would
also require the support and participation of the railroad industry.
While we believe that adopting a comprehensive approach for overseeing
railroads' management of safety risks, similar to the approaches discussed
in this section, can lead to improved safety, we are not making a
recommendation aimed at encouraging FRA to adopt such an oversight
approach. As discussed in the previous section, FRA is pursuing several
initiatives aimed at reducing train accident rates. In our view, these
initiatives need time to mature to demonstrate their effects and, at the
appropriate time, the department may wish to conduct an informed
assessment to determine whether additional actions are warranted.
FRA Measures Its Progress in Achieving a Variety of Safety Goals, but Has
Limited Information on the Direct Results of Its Oversight
FRA has a broad range of goals and measures that it uses to provide
direction to and track the performance of its safety oversight activities.
However, its ability to make informed decisions about its inspection and
enforcement programs is limited because it lacks measures of the
intermediate outcomes, or direct results, of these programs that would
show how they are contributing toward the end outcomes, or ultimate safety
improvements, that the agency seeks to achieve. Furthermore, while FRA has
made some changes in its oversight approach in response to external and
internal evaluations, it has not evaluated the effectiveness of its
enforcement approach. Evaluations can provide a broader range of
information on program performance and how to improve it than performance
measures alone. Both performance measures and evaluations can provide
valuable information on program results that helps hold agencies
accountable for their programs' performance.
FRA Has Established a Range of Safety Goals and Uses Cooperation and Enforcement
to Achieve These Goals
To its credit, FRA has adopted a range of useful safety performance
goals.^51 These safety goals are useful because they help the agency
target its oversight efforts to help achieve the department's goals of
reducing (1) the rate of rail-related accidents and incidents and (2) the
number of serious hazardous materials releases. For fiscal year 2007, FRA
established six new agencywide safety goals that are aligned with its five
inspection disciplines and its grade crossing efforts. These goals are to
reduce the rates of (1) accidents caused by human factors; (2) accidents
caused by track defects; (3) accidents caused by equipment failure; (4)
accidents attributable to other causes, including signal defects; (5)
hazardous materials releases; and (6) grade-crossing incidents. These
departmental and agency goals represent the key end outcomes, or ultimate
results, FRA seeks to achieve through its oversight efforts. The agency
has also recently established regional office goals that are generally
aligned with the new agencywide goals. These regional office goals help
FRA to link the oversight activities of its eight regional offices with
its overall goals.
FRA officials told us that their inspection and enforcement programs
contribute to meeting these safety goals, or end outcomes, by resulting in
the correction of safety problems and compliance.^52 These desired direct
results can be called the intermediate outcomes of the inspection and
enforcement programs, although FRA has not identified them as such. FRA
officials told us that they use a combination of cooperation with
railroads and enforcement actions to achieve the correction of safety
problems and compliance. The linkages between such program outputs and
desired intermediate and end outcomes can be demonstrated in a "logic
model" that helps to show how program activities contribute to the
ultimate results the agency seeks to achieve.^53 (See fig. 9.)
^51We have reported that agencies should create a set of performance goals
that address important and various aspects of program performance. See,
for example, GAO, Results-Oriented Government: GPRA Has Established a
Solid Foundation for Achieving Greater Results, [31]GAO-04-38 (Washington,
D.C.: Mar. 10, 2004); Managing for Results: Strengthening Regulatory
Agencies' Performance Management Practices, [32]GAO/GGD-00-10 (Washington,
D.C.: Oct. 28, 1999); and Agency Performance Plans: Examples of Practices
That Can Improve Usefulness to Decisionmakers, [33]GAO/GGD/AIMD-99-69
(Washington, D.C.: Feb. 26, 1999).
^52A number of other agency efforts--including the Railroad Safety
Oversight program, the development of new safety standards, rail-related
research and development, and initiatives to improve highway-rail grade
crossing safety--also contribute toward these end outcomes.
Figure 9: How FRA's Inspection and Enforcement Programs Contribute to Rail
Safety
Note: The program outputs and intermediate outcomes identified in this
figure are examples of the outputs and intended direct results of FRA's
inspection and enforcement and resulted from discussions with FRA
officials. FRA has not identified these as outputs or intermediate
outcomes. In addition to the agency's inspection and enforcement efforts,
its Railroad System Oversight managers work cooperatively with the Class I
railroads to achieve safety improvements not related to compliance.
FRA uses cooperation with railroads and enforcement actions in various
ways to resolve safety problems and achieve compliance. As explained
previously, problems identified in inspections can be site-specific
compliance problems or broader problems affecting multiple sites.
Inspectors try to resolve site-specific compliance problems found during
routine inspections by discussing defects with railroad officials in order
to achieve voluntary compliance. The agency's policy of focused
enforcement requires that inspectors cite violations and recommend
enforcement actions, most frequently civil penalties, for those compliance
problems that pose the greatest safety hazards. Enforcement actions can
require railroads to correct identified compliance problems as well as
deter future noncompliance. After in-depth inspection efforts at multiple
sites, inspectors meet with railroad managers to discuss overall findings
of safety problems that need to be corrected to achieve compliance. FRA
sometimes encourages or requires railroads to make broad-scale
improvements, such as in the training of railroads' track inspectors,
which could help a railroad to comply with the agency's standards. (See
app. II for more information on how FRA uses cooperation and enforcement
to improve safety.)
^53We have suggested that regulatory programs develop logic models to
develop a better understanding of how their programs deliver results, in
order to select appropriate performance goals and measures. See
[34]GAO-04-38 and [35]GAO/GGD-00-10 .
Performance Measures Support FRA's Oversight, but Information on Direct Results
Is Limited
Consistent with the Government Performance and Results Act of 1993, which
calls on federal agencies to develop performance measures to help
determine the extent to which intended outcomes are achieved, FRA has
developed a range of performance measures that it uses to track the
progress of--and provide direction to--its safety oversight programs.^54
(See table 3.) A number of these measures provide useful information about
the extent to which various desired end outcomes, or ultimate results, are
being achieved. In particular, FRA has developed a number of measures,
based on agency and regional goals, that capture important components of
its progress in achieving the department's overall safety goals. Since
these new measures are linked to inspection and enforcement activities of
its inspection disciplines and regional offices, they can provide some
useful information on progress in achieving the desired end outcomes. For
example, the agency expects that inspection and enforcement efforts in the
operating practices discipline will reduce accidents caused by human
factors, and it tracks the extent to which these accidents are reduced,
both at the national and the regional level.
^54This act is the centerpiece of a statutory framework that Congress put
in place during the 1990s to help resolve the long-standing management
problems that have undermined the federal government's efficiency and
effectiveness and to provide greater accountability for results. See
[36]GAO-04-38 .
Table 3: FRA's Safety Performance Measures
Type of measure Description
End outcomes
Departmental safety goals Rate of rail-related accidents and incidents^a
and number of serious hazardous materials
releases
Agency safety goals Rates of train accidents caused by human
factors, track defects, equipment defects, or
other (signal and miscellaneous) causes; rate
of nonaccident rail-related hazardous materials
releases; and rate of highway-rail grade
crossing incidents
Regional office safety By region, numbers of train accidents caused by
goals human factors, track, equipment, and other
causes and numbers of highway-rail grade
crossing incidents
Class I railroad safety Trends in rates of accidents of various types
performance trends for each Class I railroad
Program outputs and
management
Agency efficiency goal Ratio of safety budget expenditures on
safety-related activities, such as inspections,
versus administrative activities
Management of inspection o Various measures used in targeting
and enforcement activities inspections and determining enforcement
actions, including trends in various types
of accidents and employee injuries and cited
defects and violations, by inspection
discipline, region, state, and railroad
o Inspections conducted and costs by region
and inspection discipline
o Extent to which regions are meeting
planned targets for inspections
o Timeliness of reports received from
railroads on actions taken to remedy
violations^b
o Measures related to enforcement, including
civil penalties assessed and collected and
processing times
Source: GAO analysis of FRA information.
aIncludes train accidents as well as grade-crossing, trespassing, and
other accidents.
bFor those violations requiring corrective action, railroads are required,
within 30 days after the end of the month in which the violation occurred,
to notify FRA of the actions they have taken.
The agency has adjusted its oversight approach in response to trends in
end outcome measures. For example, it developed the National Rail Safety
Action Plan in response to a flat trend in the overall train accident rate
and an increasing rate of accidents caused by human factors. In addition,
FRA has developed measures to track the performance of Class I railroads,
and it reviews these measures quarterly to assist in making decisions
about oversight of these railroads.
FRA also uses various measures of program outputs, such as numbers of
inspections and enforcement actions, as well as some other types of
measures to manage its oversight efforts. While the agency does not track
its cooperative efforts to achieve compliance, it does track inspection
and enforcement activities. In fiscal year 2007, the agency will start
using an efficiency measure to track its progress in using its resources
on safety-related activities, such as inspections, rather than on
administrative activities. FRA headquarters and inspection staff use data
on defects and violations cited in inspections, together with data on
accidents and incidents, in planning inspection activities and making
enforcement decisions. Finally, FRA tracks whether railroads report on
actions taken to correct violations within the required time.
While FRA has developed a range of measures of end outcomes and program
outputs, it lacks measures of the desired intermediate outcomes, or direct
results, of its inspection and enforcement efforts--that is, the
correction of identified safety problems and compliance. We have found
that it is a useful practice for agencies to establish measures of
intermediate outcomes to help show programs' contributions to desired end
outcomes.^55 According to FRA officials, inspectors review reports on
corrective actions provided by railroads and may ask a railroad to
resubmit a report if they believe that it does not adequately address the
violation.^56 FRA officials also told us that inspectors always follow up
on serious problems identified--both site-specific and broader scale
problems--to ensure that they are corrected, and may cite additional
violations if they find continuing problems.^57 However, the agency does
not measure the extent to which identified safety problems have been
corrected. Without such a measure, FRA cannot determine the extent to
which its inspection and enforcement efforts are achieving the desired
direct results.
Measuring whether safety problems have been corrected is particularly
important when serious compliance problems are broad-scale, affecting
multiple sites, whether the problems are local, regional, or systemwide.
These problems are sometimes identified in reports of in-depth
inspections. For example, one set of track inspections of a Class I
railroad at various sites within a region in early 2006 led to overall
findings that these locations had deteriorating track conditions, that no
repair work was scheduled, and that the railroad's track inspections were
not adequate.^58 According to FRA officials, headquarters managers and
managers of the agency's eight regional offices frequently discuss serious
safety problems, indicated by the results of field inspections and data
analyses, to determine the scope of the problems and decide on actions the
agency should take to ensure that railroads resolve them. However, while
FRA tracks a variety of safety related data, it does not centrally track
these broad-scale compliance problems or their status and therefore lacks
overall information on the effectiveness of its efforts to ensure that
they are resolved.^59
55See [37]GAO/GGD/AIMD-99-69 and [38]GAO/GGD-00-10 .
^56For violations determined by the inspector to require corrective
action, railroads are required to provide FRA with a report within 30 days
of the end of that month on the corrective actions they have taken. These
reports also include the type of action taken, such as repair or
replacement of equipment and disciplinary action or training of employees.
FRA tracks whether these reports are received on time as well as the types
of actions taken.
^57FRA's inspection reports contain a field for recording the receipt of
railroads' reports on corrective actions but do not contain a field for
recording whether inspectors have determined, in following up on
violations, that the corrective actions are adequate. Furthermore, some
regional staff told us that inspectors do not consistently record
follow-up inspections as such.
FRA also lacks overall measures of railroads' compliance. FRA officials
told us that, while defect rates (the ratio of defects found per units
inspected) measure noncompliance found by inspectors, they cannot be used
to produce statistically valid measures of railroads' overall level of
compliance because inspections are focused on problem areas and FRA is not
able to conduct enough inspections of railroads to ensure that it is
getting a good measure of compliance.^60 Officials have emphasized that
the agency relies on inspectors' day-to-day oversight of and interaction
with railroads to track compliance. Also, as noted previously, FRA
officials, both at the headquarters and regional levels, analyze defect
data in each inspection discipline to identify emerging issues and plan
inspection activity. Finally, officials noted that the agency is planning
to use its automated track inspection vehicles to survey most of the
national track system and to monitor improvements in the condition of
track over time.
^58In contrast, defects and violations tend to be much more specific. In
this case, inspectors also found numerous defects concerning specific
instances of track not meeting federal standards and two violations
concerning defects that had gone for more than 30 days without corrective
action.
^59In fiscal year 2006, FRA compiled information for the first time on all
in-depth inspection activity by its headquarters and regional offices into
one report and began producing quarterly updates on the status of these
inspections. These updates contain some information on findings of
broad-scale problems, in narrative form, but do not provide overall
assessments of progress being made by each railroad as a whole in
addressing these problems.
^60However, FRA does include in its quarterly review of Class I railroads'
performance trends certain operating practices and track defect rates that
it has found to be related to accident rates.
We recognize that developing measures of intermediate outcomes would be
difficult and that it is challenging for regulatory agencies to develop
such measures.^61 Nevertheless, some other regulatory agencies in the
Department of Transportation have developed such measures. For example,
the Federal Motor Carrier Safety Administration measures the percentage of
truck companies that improve their performance in a follow-up inspection
and PHMSA measures the extent of improvement in pipeline operators'
integrity management programs, as indicated by successive inspections of
operators' programs. FRA officials have told us that the fact that the
agency has not integrated its existing safety-related databases has
impeded its ability to develop measures of intermediate outcomes. The
agency has an initiative underway to better integrate these databases,
including its database on accidents and incidents and its inspection and
enforcement databases, in order to better manage its information
resources.^62
Performance measures should provide agency managers with information that
helps them make decisions that improve program performance, including
decisions to adjust policies and priorities. As noted, FRA has used its
existing performance measures to make decisions about its oversight
approach in a variety of ways. However, not having measures of the
intermediate outcomes of its inspection and enforcement approaches limits
FRA's ability to make informed decisions about these approaches and adjust
them to improve performance. Intermediate outcome measures can provide
more timely information on program performance than end outcome measures,
because it may take longer for program efforts to affect end outcomes.
Measures of program results can also help hold agencies accountable for
the performance of their programs. Congress needs information on program
results to support its oversight of agencies and their budgets. FRA's new
discipline-specific and region-specific outcome measures do help ensure
accountability for results. However, without measures of intermediate
outcomes, the extent to which FRA's inspection and enforcement programs
are achieving direct results and contributing to desired end outcomes is
not clear. FRA officials have noted that they cannot attribute any drops
in accident rates solely to FRA's efforts because other factors, such as
railroads' investments in their systems, also play an important role.
^61We have reported on such challenges and how agencies have overcome
them. See, in particular, [39]GAO-04-38 ; [40]GAO/GGD-00-10 ; Managing for
Results: Measuring Program Results That Are Under Limited Federal Control,
[41]GAO/GGD-99-16 (Washington, D.C.: Dec. 11, 1998); and Managing for
Results: Regulatory Agencies Identified Significant Barriers to Focusing
on Results, [42]GAO/GGD-97-83 (Washington, D.C.: June 24, 1997).
^62FRA has tasked a contractor with developing a plan for a data
warehousing strategy for the agency that will integrate its data from
various sources. According to agency officials, this initiative will allow
FRA to better monitor its performance, through, for example, providing
scorecards and graphical tools to depict performance.
FRA can also use measures of intermediate outcomes to increase railroads'
accountability for correcting safety problems. In spring 2006, the agency
instituted annual meetings with the heads of the Class I railroads to
discuss their overall safety performance, using trends in various accident
rates. These meetings are a good step forward for the agency because they
represent an opportunity for FRA to put pressure on top railroad
executives to adequately address major problems the agency has identified.
However, without measures of the extent to which individual railroads have
addressed such problems, FRA cannot take full advantage of this
opportunity. In addition, without central tracking of these problems and
their status, FRA's ability to identify continuing or recurring problems
as well as interrelated problems and make appropriate enforcement
decisions may be impaired, especially since much of this work is performed
in eight separate FRA regional offices and in five separate inspection
disciplines.^63 According to FRA, annual meetings with the major railroads
to negotiate civil penalties focus on individual violations but also
address systemic issues that have been identified through analysis of
individual violations. However, regional officials told us that these
meetings do not generally deal with systemic issues. Readily available
information on the status of broader problems would help put a railroad's
individual violations into perspective and could help ensure that FRA
negotiates appropriate final amounts with railroads.
FRA Has Made Changes in Response to Evaluations but Has Not Evaluated Its
Enforcement Approach
Besides requiring performance measurement, the Government Performance and
Results Act of 1993 calls for agencies to evaluate the effectiveness of
their programs in achieving intended outcomes. We have found that, since
it can be challenging for regulatory agencies to measure the direct
results of their programs, program evaluations are particularly helpful in
determining these results.^64 Program evaluations are objective,
systematic studies that answer questions about program performance and
results. By examining a broader range of information than is feasible to
monitor on an ongoing basis through performance measures, evaluation
studies can explore the benefits of a program as well as ways to improve
program performance. They can also be used to develop or improve agencies'
measures of program performance and help ensure agencies' accountability
for program results.^65
63A major concern with the previous Safety Assurance and Compliance
Program, which ended in 2005, was that systemwide safety problems handled
under this program took too long to resolve.
FRA's safety oversight activities have recently undergone several external
evaluations by the department's Inspector General, as well as an internal
review, and FRA has made some changes as a result. For example, on the
basis of several reviews since 1998, the department's Inspector General in
2004 recommended that FRA develop a plan to make meaningful use of
available data to focus its inspection and enforcement activity.^66 FRA
developed its National Inspection Plan in response. Also, in 2004, FRA
established a committee to conduct an internal review of its Safety
Assurance and Compliance Program. This committee solicited the views of
various FRA managers on the strengths and weaknesses of the program. Based
on its findings, FRA terminated this program and replaced it with its
Railroad System Oversight program.
In addition, the Office of Management and Budget assessed FRA's overall
safety program in 2003.^67 Although the office found this overall program
to be moderately effective, it also found that FRA had not arranged for
independent evaluations of its design and effectiveness. In response, FRA
arranged for a review of its Railroad Safety Board process and has stated
that it will continue to have regular independent reviews of various
aspects of its safety program.^68 The agency is planning to have an
independent evaluation of its Railroad System Oversight program conducted
in the third quarter of fiscal year 2007. FRA also made some other
improvements in response to recommendations of the Office of Management
and Budget assessment, including developing its new efficiency measure and
procuring new vehicles for conducting track inspections.
^64Such evaluations can also help determine the extent to which a program
is having an impact on these outcomes versus other variables that affect
outcomes. See [43]GAO-04-38 .
^65See GAO, Program Evaluation: An Evaluation Culture and Collaborative
Partnerships Help Build Agency Capacity, [44]GAO-03-454 (Washington, D.C.:
May 2, 2003); and Program Evaluation: Studies Helped Agencies Measure or
Explain Program Performance, [45]GAO/GGD-00-204 (Washington, D.C.: Sept.
29, 2000).
^66U.S. Department of Transportation-Office of the Inspector General,
Review of Allegations that FRA Deputy Administrator Attempted to Relax
Safety Enforcement Against Union Pacific (Washington, D.C., Dec. 10,
2004); and Audit of Oversight of Highway-Rail Grade Crossing Accident
Reporting, Investigations, and Safety Regulations, MH-2006-016
(Washington, D.C., Nov. 28, 2005).
^67The Office of Management and Budget performed this assessment using its
Program Assessment Rating Tool. This tool examines factors that affect and
reflect program performance, including program purpose and design,
performance measurement and evaluations, and aspects of program
management.
Although FRA has modified various aspects of its safety oversight in
response to evaluations, it has not evaluated the extent to which its
enforcement is achieving desired results. In addition to providing
information on program performance and how it could be improved, an
evaluation of FRA's enforcement approach could help to identify data
needed to develop useful performance measures. For example, the Federal
Motor Carrier Safety Administration examined the rate of violations by
trucking companies before and after implementing a policy of assessing
maximum penalties for such violations and, based on initial analyses, has
improved its data system to be better able to analyze this trend.
Under FRA's focused enforcement policy, developed in the mid-1990s,
inspectors cite a small percentage of identified defects (about 3 percent
in 2005) as violations that they recommend for enforcement action,
generally civil penalties. While this policy relies, to a great extent, on
cooperation with railroads to achieve compliance and is intended to focus
FRA's enforcement efforts on those instances of noncompliance that pose
the greatest safety hazards, it is not clear whether the number of civil
penalties issued, or their amounts, are having the desired effect of
improving compliance. FRA officials have told us that they have not
evaluated the effectiveness of civil penalties in ensuring compliance,
noting that this would be difficult because penalty payments usually occur
after the agency's yearly settlement process. However, without an
evaluation of its enforcement program, FRA is missing an opportunity to
obtain valuable information on the performance of this program and on any
need for adjustments to improve this performance.
^68The Railroad Safety Board approves or denies requests for waivers or
special approval submitted by railroads and other parties subject to FRA
regulations.
Conclusions
The various initiatives that FRA has begun in the past year and a half to
better target its oversight--by addressing the main causes of train
accidents and better focusing inspections on problem areas--hold promise
for bringing down the train accident rate, reducing injuries, and saving
lives. Some initiatives, such as reporting of close call incidents,
encourage the railroad industry to address safety problems before they
result in accidents. However, the success of many of these initiatives
will depend on voluntary actions by the railroads and their overall safety
impact will likely not be apparent for a number of years. While FRA is
pursuing these initiatives, it has not changed its approach for conducting
inspections, which relies primarily on direct observations of operations,
equipment, and track. An additional approach that has provided additional
assurance of safety in the U.S. commuter railroad, U.S. pipeline, and
Canadian railroad industries is oversight of companies' overall management
of safety risks. Although we believe that a similar approach could help
improve rail safety, we are not recommending that FRA adopt such an
approach because its current initiatives to bring down the train accident
rate need time to demonstrate their effects.
Without measures of the direct results of its inspection and enforcement
programs, FRA cannot demonstrate how these programs are contributing to
rail safety and lacks key information that could help it improve
performance. This information could also help FRA hold railroads
accountable for addressing safety problems it identifies. While these
measures are not always easy to develop, at least one other modal
administration within the department has done so at our recommendation.
Coupled with better measures of FRA's direct results is the need to assess
the effectiveness of its enforcement program, especially its use of civil
penalties, to understand the degree to which they contribute to improved
safety outcomes and to determine whether it should adjust its approach to
improve performance.
Recommendations for Executive Action
To enhance FRA's ability to determine the extent to which its inspection
and enforcement programs are contributing to rail safety and whether
changes in these programs are needed, we recommend that the Secretary of
Transportation direct the Administrator of FRA to take the following two
actions:
o develop and implement measures of the direct results of its
inspection and enforcement programs; and
o evaluate the agency's enforcement program to provide further
information on its results, the need for additional data to
measure and assess these results, and the need for any changes in
this program to improve performance.
Agency Comments
We provided a draft of this report to the Department of
Transportation for its review and comment. The department did not
offer overall comments on the draft report or its recommendations.
It did offer several technical comments, which we incorporated
where appropriate.
We are sending copies of this report to congressional committees
and subcommittees with responsibility for transportation safety
issues; the Secretary of Transportation; the Administrator,
Federal Railroad Administration; and the Director, Office of
Management and Budget. We will also make copies available to
others upon request. This report will be available at no charge on
the GAO Web site at http://www.gao.gov .
If you have any questions about this report, please contact me at
(202) 512-2834 or [email protected] . Contact points for our
Offices of Congressional Relations and Public Affairs may be found
on the last page of this report. Staff who made key contributions
to this report are listed in appendix IV.
Katherine A. Siggerud
Director, Physical Infrastructure Issues
Appendix I: Scope and Methodology
To determine how the Federal Railroad Administration (FRA) focuses
its efforts on the highest priority risks related to train
accidents in planning its safety oversight, we reviewed FRA's
National Rail Safety Action Plan, plans for developing new
regulations, documents related to its initiatives for reducing
human factor and track-caused accidents, and inspection planning
documents. We also discussed the agency's inspection planning
approach and key initiatives with headquarters and regional
officials responsible for these areas. In addition, we obtained
information on FRA's initiatives through attending congressional
hearings on FRA's oversight and meetings of the Railroad Safety
Advisory Committee.
To determine how FRA identifies safety problems on railroad
systems in carrying out its oversight, we determined how FRA uses
its inspections, the Railroad System Oversight program, and other
oversight activities to oversee safety and identify problems. In
order to do this, we reviewed FRA's regulations, policies,
procedures, and guidance for conducting inspections and
identifying safety problems as well as reports on inspection
results. We also reviewed the agency's statutory authority under
the Federal Railroad Safety Act and reviewed Railroad Safety
Oversight program documents. We discussed FRA's oversight
activities with FRA headquarters managers and inspection
discipline specialists and Railroad Safety Oversight managers to
gain a greater understanding of how FRA identifies safety problems
on railroad systems. We also contacted three (Chicago, Forth
Worth, and Atlanta) of the eight FRA regional offices reporting
the highest numbers of accidents in their jurisdictions during
2005. We discussed this topic with administrators and track and
human factor discipline specialists from each of these regional
offices. We conducted these interviews on-site at the offices
located in Fort Worth, Texas, and Chicago, Illinois. We also
discussed this topic with officials from the three state
regulatory associations employing the greatest number of railroad
safety inspectors in order to discuss how state inspectors
coordinate their activities with FRA inspectors. (See the end of
this appendix for a list of organizations that we contacted.) We
also examined our published work on risk management and safety
oversight approaches used by other modal administrations within
the Department of Transportation as well as some other
organizations responsible for overseeing rail safety.
We reviewed FRA data on its inspection activities for the period
from 1996 through 2005. To assess the reliability of the
inspection data, we (1) performed electronic testing of required
data elements, (2) reviewed existing information about the data
and the system that produced them, and (3) interviewed agency
officials knowledgeable about the data. We discussed our
preliminary results with FRA to ensure that we were capturing the
correct information and consulted with these officials to resolve
questions on the data collection process. We determined that the
data were sufficiently reliable for the purposes of this report.
To determine how FRA assesses the impact of its oversight efforts
in improving safety, we examined the methods FRA uses to assess
the results of its oversight programs and FRA's use of this
information to make decisions about its oversight strategy. As
part of this effort, we reviewed FRA documents on its safety
performance measures and evaluations of its oversight and
enforcement activities. We also discussed this issue with managers
at headquarters and in the regional offices responsible for
developing and using performance information. In addition, we
reviewed our recommendations on performance measurement, the use
of performance information, and program evaluation and determined
the extent to which FRA's practices are consistent with these
recommendations.
We focused our work on FRA's oversight activities aimed at
reducing train accidents-such as train collisions and
derailments-rather than those aimed at reducing highway-rail
crossing and trespassing accidents. FRA's oversight of
highway-rail grade crossing safety has been the subject of two
recent Department of Transportation Inspector General audits. In
addition, both these types of accidents and trespassing accidents
involve issues not related to railroad safety performance, such as
driver awareness of grade crossing safety and individuals'
willingness to abide by railroads' warning signs on their
property. We also focused on FRA's oversight of railroads rather
than its oversight of non-railroad companies (such as shippers of
hazardous materials by rail) because the agency `s oversight
efforts focus primarily on railroads. In addition, according to
FRA, most recent serious train accidents involving the release of
hazardous materials have resulted from problems with railroad
operations.
Other Aspects of Our Work
We also examined how FRA uses enforcement and other methods for
ensuring that safety problems on railroad systems are resolved. In
order to do this, we reviewed FRA's policies, procedures, and
guidance for conducting enforcement; major enforcement actions
such as compliance agreements; Railroad Safety Oversight program
documents; and other agency documents. We also discussed FRA's
approaches to resolving safety problems, through enforcement and
other means, with officials from FRA headquarters, Railroad System
Oversight managers, and administrators and inspection discipline
specialists in three regional offices. Regarding FRA's enforcement
approach, we reviewed how FRA employs enforcement to improve
safety on railroads' systems. However, we could not determine the
extent to which railroads address systemic safety problems as a
result of enforcement because FRA does not track the broad-scale
safety problems it identifies or the extent to which these
problems are resolved. We focused on FRA's use of enforcement as
part of its overall oversight strategy and did not examine the
agency's individual enforcement actions.
We also obtained FRA data on civil penalties from FRA's
enforcement database, which includes data on assessed and
collected amounts from 1996 through 2005. We assessed the
reliability of the enforcement data by the same means that we
employed to assess the reliability of FRA's inspection data. We
determined that the enforcement data used in this report were
sufficiently reliable for the purposes of this report.
Finally, we met with or contacted the following organizations or
persons in order to obtain a fuller understanding of railroad
safety issues and obtain their perspectives on FRA's oversight
approach.
Other federal agencies:
National Transportation Safety Board
Freight railroads (Class I):
BNSF Railway Company CSX Transportation, Inc. Norfolk Southern
Railway Company Union Pacific Railroad Company
Passenger railroads:
National Railroad Passenger Corporation, or Amtrak Northeast
Illinois Regional Commuter Rail Corporation, or Metra New Jersey
Transit Rail Operations
Regional and short line railroad management companies:
Cedar American Rail Holdings, Inc. RailAmerica
Industry associations:
American Chemistry Council Association of American Railroads
American Public Transportation Association American Short Line and
Regional Railroad Association
Labor unions:
AFL-CIO Brotherhood of Maintenance of Way Employees Brotherhood of
Locomotive Engineers and Trainmen United Transportation Union
State oversight organizations:
Association of State Rail Safety Managers California Public
Utilities Commission Ohio Public Utilities Commission Texas
Department of Transportation
Canadian oversight organization:
Transport Canada
Academic expert:
Dr. Ian Savage, Department of Economics, Northwestern University
FRA uses cooperation with railroads and enforcement actions in
various ways to resolve identified safety problems and achieve
compliance. When inspectors find problems during inspections,
FRA's policy is to cite defects for most instances of
noncompliance and to encourage the railroad to comply voluntarily.
For example, issues may be addressed on site with railroad
officials during inspections. According to FRA officials,
railroads often correct identified problems immediately and if so,
these problems would not require additional action. When railroads
do not comply voluntarily or the identified defects are serious,
FRA may cite violations and recommend civil penalties or take
other enforcement actions--either against railroads or
individuals--to promote compliance with safety regulations.^1
According to FRA officials, inspectors follow up on violations or
high numbers of defects within 60 days to ensure that they are
corrected.^2
FRA most commonly uses civil penalties against companies as its
enforcement tool for site-specific violations.^3 From January 2005
through July 2006, FRA assessed about 8,600 violations for civil
penalties. FRA has other enforcement tools. These include
compliance agreements and compliance orders, civil penalties
against individuals, special notices for repair, emergency orders,
criminal penalties, disqualification orders, and injunctions. (See
table 4.) FRA uses these tools much less frequently than it does
civil penalties. For example, FRA issued 288 special notices for
repair, 118 warning letters and fewer than a dozen of all other
enforcement actions during this period.^4
^1While we reviewed how FRA employs enforcement to improve safety on
railroads' systems, we could not determine the extent to which railroads
address systemic safety problems as a result of enforcement. This is
because FRA does not track the broad-scale safety problems it identifies
or the extent to which these problems are resolved. We focused on FRA's
use of enforcement as part of its overall oversight strategy and did not
examine the agency's individual enforcement actions.
^2Often, an inspector will conduct follow-up re-inspections during other
routine inspections. However, an inspector may make a special visit to
conduct follow-up if warranted.
^3FRA's Statement of Enforcement Policy specifies that before citing
violations and recommending penalties, inspectors consider the seriousness
of the condition or act, the potential safety hazard posed by the
condition or act, and the current level of compliance of the offending
person (e.g., a railroad or individual), among other things.
^4FRA could not supply data covering a longer period on the frequency of
use of all of its enforcement actions without substantial effort.
Table 4: Description of FRA Enforcement Actions and Frequency of Use,
January 2005 through July 2006
Number of
Enforcement action times used Description
Civil penalty against a 8,606 Imposes a monetary penalty on a
railroad or other entity railroad or other entity, such as a
except for an individual company that ships hazardous
materials by rail, that violates a
rail safety or hazardous materials
statute, regulation, or order. Each
day that the violation continues is
a separate violation.
Special notice for repair 288 Orders a locomotive, freight car, or
passenger car out of service and may
require a reduction of the maximum
operating speed over defective track
segments.
Warning letter to an 118 Informs an individual that FRA
individual believes that the individual has
committed a rail safety or hazardous
materials violation and that any
future violation by the individual
will result in enforcement action.
Civil penalty against an 6 Imposes a monetary penalty on any
individual individual who willfully violates or
willfully causes the violation of a
rail safety statute, regulation, or
order.
Compliance agreement 2 States a railroad's agreement to
take a specified action to promote
compliance. The railroad agrees that
if, in FRA's judgment, the required
action is not performed, the
railroad will not contest FRA's
sanction-typically a compliance
order. See below.
Criminal penalty 1 Imposes (1) either a monetary
penalty or imprisonment for up to 2
years, or both, on an individual for
knowingly and willfully violating
certain reporting, recordkeeping, or
other regulations or (2) a monetary
penalty or imprisonment for up to 5
years, or both, for knowingly
violating the hazardous materials
statute or an implementing
regulation or order.
Disqualification order 1 Prohibits an individual from
performing safety-sensitive
functions in the rail industry for a
specified period if the individual's
violation of a rail safety statute
or regulation demonstrates unfitness
to perform such functions.
Emergency order 1 Orders corrective action where an
unsafe condition or practice causes
an emergency situation involving a
hazard of death or personal injury.
Compliance order 0 Directs compliance following
repeated failure to comply with rail
safety or hazardous material statute
or regulation.
Injunction 0 Restrains a violation of, or
enforces, a rail safety or hazardous
materials regulation or order.
Source: FRA.
FRA inspectors cite many defects, but cite comparatively few of these
defects as violations warranting enforcement action. Since 1996, FRA
inspectors have cited an average of about 4 violations for every 100
defects cited annually. According to FRA officials, inspectors cite
relatively few defects as violations warranting enforcement action because
FRA's focused enforcement policy guides inspectors to cite violations only
for problems that pose safety risks. In addition, inspectors have
discretion in citing a defect or a violation for a given instance of
noncompliance--FRA directs inspectors to first seek and obtain the
railroads' voluntary compliance with the rail safety regulations.
According to FRA officials, inspectors usually choose to provide the
railroad with information about defects they found during their
inspection, discuss these instances of noncompliance, and attempt to
obtain the railroad's commitment to improve compliance. If the railroad's
response is inadequate or the inspector finds that the problem warranting
the defect is serious, the inspector may exercise the agency's enforcement
discretion by citing a violation, recommending that FRA take enforcement
action, generally a civil penalty. The agency makes an initial penalty
assessment against the railroad based on the type of violation that
occurred.^5 FRA meets with the major railroads in an annual settlement
meeting to negotiate a final amount for all civil penalties cited in the
past year. This amount is based on the railroad's compliance history and
efforts to correct the problem, among other factors.^6 The total value of
civil penalties assessed and collected each year was higher from 2001
through 2005 than from 1996 to 2000. (See fig. 10.) According to FRA
officials, the higher civil penalty assessments after 2000 reflect the
agency's efforts to focus its inspection activity on areas that present
the greatest risk. FRA officials also told us that when the agency
initiated the Safety Assurance and Compliance Program--which emphasized a
partnership approach with railroads to improve safety--in 1995, inspectors
initially cited fewer violations. They noted that in 2001 FRA's management
pushed for inspectors to be more aggressive in citing violations. Since
2001, the amounts collected have been about 63 percent of the amounts
initially assessed. Consistent with the purpose of the federal railroad
safety laws and with federal court decisions interpreting the purpose of
such laws, the agency's goals are to promote safety and to gain compliance
rather than to maximize amounts collected. For example, FRA may agree to a
reduced penalty amount if the railroad immediately remedies the safety
problem or implements a new program to prevent a problem from recurring.^7
This is consistent with federal law. In other cases, FRA will agree to a
reduced penalty amount if its documentation on the violation is not
substantial enough.
^5In December 2006, FRA published in the Federal Register proposed
amendments to its schedules of civil penalties for each type of violation
to ensure that penalty amounts more fully reflect the risk associated with
a railroad's violation of the rail safety regulations.
^6Given the volume of civil penalty cases, FRA usually negotiates civil
penalties with smaller railroads and shippers through the mail and
telephone conferences. FRA is authorized to negotiate civil penalties with
railroads and exercises this authority by annually settling civil penalty
amounts with each railroad. The criteria for compromising with railroads
on civil penalty amounts are set in statute. In determining the amount of
a compromise, the Secretary shall consider (1) the nature, circumstances,
extent, and gravity of the violation; (2) with respect to the violator,
the degree of culpability, any history of violations, the ability to pay,
and any effect on the ability to continue to do business; and (3) other
matters that justice requires.
Figure 10: FRA's Civil Penalties, 1996 through 2005, in 2005 Dollars
Notes: Individual penalties resulting from violations are consolidated
into one or more case(s) for each railroad and are negotiated annually
with FRA and the railroad during settlement conferences.
Each year's amounts are for cases initiated in that year. Some penalties
may be initially assessed in one year, with the final assessment in
another year. FRA, in commenting on a draft of our report, provided 2006
civil penalty figures. However, we did not use this information because
there was not enough time to assess its reliability before the report was
issued.
Penalties against individuals are not included. From 1996 through 2005,
FRA issued an average of 1 civil penalty per year against individuals.
^7The Federal Motor Carrier Safety Administration follows a similar
policy. See GAO, Large Truck Safety: Federal Enforcement Efforts Have Been
Stronger Since 2000, but Oversight of State Grants Needs Improvement,
[48]GAO-06-156 (Washington, D.C.: Dec. 15, 2005).
FRA seeks to resolve broad-scale compliance problems it has identified by
first discussing them with railroad officials and elevating them, as
necessary, to obtain an appropriate response. After in-depth inspection
efforts at multiple sites, inspectors meet with railroad managers
responsible for those sites to discuss overall findings or problems. For
very serious or systemic problems, FRA may ask railroads to submit
corrective action plans and may review these plans for adequacy. FRA
sometimes discusses broad problems with railroad headquarters officials to
try to obtain corrective actions. For example, one regional administrator
told us that his region arranged a multiregion meeting with the senior
management of a Class I railroad to discuss systemic problems with the
railroad's equipment. FRA asked the railroad to submit a corrective action
plan to address the equipment problems. According to FRA officials, the
railroad has been meeting the milestones in the corrective action plan and
has been making progress in addressing its equipment problems. Also, FRA's
Railroad System Oversight managers, who act as liaisons with the Class I
railroads, help to maintain frequent communication with these railroads
about major problems that have been identified and associated corrective
actions. These efforts to cooperate with railroads to resolve problems may
be combined with civil penalties for violations found at specific
locations. According to FRA officials, when in-depth inspections result in
findings of serious problems, inspectors always follow up to determine
whether these problems are adequately resolved. However, FRA has no
central repository for data on the status of these broader problems its
inspectors have identified at railroads.
FRA officials told us that they hold frequent internal discussions among
headquarters and regional management about these broad-scale compliance
problems that have been identified--whether local, regional, or
systemwide--and actions that the agency should take to attempt to resolve
them. According to FRA officials, the agency sometimes uses compliance
agreements, which require railroads to take significant actions beyond
those specifically required by regulations to improve their ability to
comply, when broad-scale compliance problems are egregious and have not
been resolved through other methods. Compliance agreements allow FRA to
apply more leverage in trying to obtain compliance because FRA uses the
agreements as an alternative to the railroad involved undergoing an FRA
proceeding for a compliance order. For example, under recent compliance
agreements, railroads have agreed to reinstruct managers on how to test
employees' adherence to operating rules and to develop and implement track
maintenance plans to eliminate systemic track defects. FRA monitors
railroads' performance under these agreements. In some instances, FRA has
also used compliance orders to address significant railroad compliance
problems. However, these enforcement tools are seldom used. FRA has
entered into 13 compliance agreements altogether and one compliance order
since 1996, or an average of about 1 per year. In comparison, since 1996,
FRA has assessed in the neighborhood of almost 40,000 violations for civil
penalties.
This approach to resolving broad-scale safety problems has had some
success. For example, the three railroads that have entered into
compliance agreements in the past 2 years have generally made progress in
improving compliance, according to the results of FRA's follow-up.^8 In
addition, FRA officials have cited a number of other successes, including
improved compliance by several railroads with signal inspection and
testing requirements,^9 and another railroad's implementation of an
electronic system to record locomotive defects and repairs following FRA's
finding that inadequate recordkeeping had resulted in the operation of
locomotives with multiple defects. FRA's Railroad System Oversight
managers have also worked with railroads to make some systemic safety
improvements not related to compliance. For example, these managers have
encouraged railroads to make improvements related to roadway worker
communication, signal maintenance, and mentoring of new employees.
^8FRA's compliance agreements with three different divisions of one Class
I railroad have resulted in improved compliance. FRA had to extend the
compliance agreement with the third division because noncompliance
continued but has since terminated the compliance agreement because of
improved safety performance. While FRA does not have overall measures of
compliance, it sometimes, in summary reports on follow-up inspections at a
railroad, records trends in certain types of defects found.
^9FRA has noted instances of significant noncompliance with signal
inspection and testing requirements on the part of two major commuter
railroads and at least one Class I railroad.
Appendix III: Oversight of Risk Management in the U.S. Commuter Railroad,
U.S. Pipeline, and Canadian Railroad Industries
Risk management is a systematic process for assessing risks and taking
appropriate steps to deal with them. It is founded on several inspection
disciplines, including financial economics, decision science,
organizational theory, and strategic management. The National Academy of
Sciences, a presidential commission, private organizations, and others
have addressed the subject and have recognized its applicability in both
the private and the public sectors. We have developed a framework for risk
management based on industry best practices.^1 This framework divides risk
management into five major phases. (See table 5.)
Table 5: Risk Management Framework
Phase Description
Strategic goals, Establish an organization's strategic goals
objectives, and constraints and the steps needed to attain those results,
including performance measures to assess
progress. Constraints that affect outcomes can
also be identified during this phase.
Risk assessment Assess the threats to and vulnerabilities of
assets so that countermeasures may be
instituted to prevent or mitigate risks. Risks
can be assessed by various methods, depending
on the specific application and knowledge
available.
Alternatives evaluation Evaluate risk reduction methods by considering
the countermeasures and the costs and benefits
associated with each. Countermeasures can be
considered and prioritized according to a
number of factors, such as the degree of risk
reduction they afford and the cost and
difficulty to implement them.
Management selection Choose among alternative actions. Management's
active participation is important at this
phase because risk assessment tools contain
various assumptions about preferences that may
require value judgments and review at the
management level.
Implementation and Move from planning to implementing the
monitoring selected countermeasures. Following
implementation, monitoring is essential to
help ensure that the entire risk management
process remains current and relevant.
Source: GAO.
During the last 10 years, the American Public Transportation Association
(APTA), Pipeline and Hazardous Materials Safety Administration (PHMSA),
and Transport Canada have developed and implemented new approaches for
overseeing safety in the U.S. commuter railroad, U.S. pipeline, and
Canadian railroad industries, respectively. These approaches promote the
use of risk management and incorporate various elements of risk management
noted above. They also complement traditional compliance inspections.
^1To develop the framework, we reviewed risk management literature, our
reports and testimonies on this topic, and other government guidance. In
addition, we consulted with experts on risk management, risk modeling, and
terrorism. We reviewed numerous frameworks from industry, government and
academic sources. We field-tested the framework, and it was reviewed by
academic experts in risk management. See [49]GAO-06-91 .
In 1996, following two serious commuter rail accidents, APTA, with FRA
support, started a program to help its 16-member commuter railroad
properties develop and implement plans for managing the safety of their
systems. APTA had previously developed a similar program for the rapid
transit industry. Risk management is a key aspect of these plans, which
must include, among other things, goals and objectives, the identification
of hazards, an assessment of their associated risks, the analysis and
implementation of actions to reduce these risks, and internal assessments
of the effectiveness of safety management. Commuter railroads are given
the latitude to develop individual plans that conform to APTA's guidelines
but that are based on their unique circumstances. APTA audits its member
railroads' system safety management plans on a 3-year cycle. During the
first year, APTA conducts an initial assessment of the plan; during the
second year, APTA evaluates how well the railroad implements its plan; and
during the third year, APTA follows up with the railroad to see what
actions it has taken in response to APTA's audit findings.
In 2000, to better focus on safety risks that are unique to individual
pipelines, PHMSA, then operating as the Office of Pipeline Safety, issued
the first in a series of integrity management regulations that now apply
to more than 1,000 hazardous liquid and gas transmission pipelines
operating in densely populated and environmentally sensitive areas. Under
this approach, operators are required to develop programs to
systematically assess and mitigate safety threats, such as corrosion, to
pipelines located in such high-risk areas. For example, these programs
must integrate all available information about pipeline integrity and the
consequences of a leak or rupture, repair identified defects within
defined time limits based on their severity, and evaluate the need for
additional preventive and mitigating actions. In addition, operators are
required to report program performance measures to PHMSA semi-annually.
PHMSA inspects these written programs as well as their implementation in
periodic comprehensive inspections.
During the 1990s, a series of derailments raised concerns over the level
of safety in the Canadian rail industry. In 2001, following a review of
its oversight approach, Transport Canada determined that a more
comprehensive safety management assessment was necessary to minimize
safety risks, and adopted new safety management regulations under Canada's
Railway Safety Act. The new regulations require that Canadian railway
companies develop their own safety management systems, subject to
Transport Canada's review. These systems must include, among other things,
annual safety performance targets; the identification of safety issues
through a variety of methods (such as analyses of safety data and input
from employees); an assessment of the risks associated with these issues
to determine their significance; the development, approval, and
implementation of strategies for controlling these risks; and systems for
monitoring these strategies and other management-approved corrective
actions. Transport Canada audits the railroads' documentation and
implementation of their safety management systems. The level of such
monitoring varies depending on the safety performance of the railroad.
Appendix IV: GAO Contact and Staff Acknowledgments
GAO Contact
Katherine Siggerud (202) 512-2834 or [email protected]
Staff Acknowledgments
In addition to the contact named above, James Ratzenberger, Assistant
Director; Ashley Alley; Marianne Anderson; Timothy Bober; Elizabeth
Eisenstadt; Judy Guilliams-Tapia; Brandon Haller; Bonnie Pignatiello Leer;
Masha P. Pastuhov-Purdie; and Minette Richardson made key contributions to
this report.
(542077)
GAO's Mission
The Government Accountability Office, the audit, evaluation and
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 GAO Reports and Testimony
The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site ( www.gao.gov ). Each weekday, GAO posts
newly released reports, testimony, and correspondence on its Web site. To
have GAO e-mail you a list of newly posted products every afternoon, go to
www.gao.gov and select "Subscribe to Updates."
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. Government Accountability 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, Waste, and Abuse in Federal Programs
Contact:
Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail:
[email protected] Automated answering system: (800) 424-5454 or (202)
512-7470
Congressional Relations
Gloria Jarmon, Managing Director, [email protected] (202) 512-4400 U.S.
Government Accountability Office, 441 G Street NW, Room 7125 Washington,
D.C. 20548
Public Affairs
Paul Anderson, Managing Director, [email protected] (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548
www.gao.gov/cgi-bin/getrpt?GAO-07-149 .
To view the full product, including the scope
and methodology, click on the link above.
For more information, contact Katherine Siggerud at (202) 512-2834 or
[email protected].
Highlights of [57]GAO-07-149 , a report to congressional committees
January 2007
RAIL SAFETY
The Federal Railroad Administration Is Taking Steps to Better Target Its
Oversight, but Assessment of Results Is Needed to Determine Impact
Since 1980, the train accident rate has improved significantly, but
progress has leveled off over the past 10 years. Recent serious
accidents--such as one in Graniteville, South Carolina, that led to 9
deaths and 292 injuries--elevated concerns. The Federal Railroad
Administration (FRA) develops safety standards and inspects and enforces
railroads' compliance with these standards.
This report addresses how FRA (1) focuses its efforts on the highest
priority risks related to train accidents in planning its oversight, (2)
identifies safety problems on railroad systems in carrying out its
oversight, and (3) assesses the impact of its oversight efforts on safety.
To complete this work, GAO reviewed FRA regulations, planning and policy
documents, and safety data. GAO also contacted FRA officials in
headquarters and three regional offices and others.
[58]What GAO Recommends
GAO is recommending that FRA develop measures of the direct results of its
oversight and evaluate the effectiveness of its enforcement program. The
Department of Transportation did not provide overall comments on these
recommendations but did provide several technical comments, which were
incorporated as appropriate.
In planning its safetyoversight, FRA is focusing its efforts on the
highest priority risks related to train accidents through initiatives
aimed at addressing their main causes--human behaviors and defective
track--as well as through improvements in its inspection planning
approach. In its May 2005 National Rail Safety Action Plan, the overall
strategy for targeting its oversight at the greatest risks, FRA provides a
reasonable framework for guiding these efforts. Its initiatives to address
the most common causes of accidents are promising, although the success of
many of them will depend on voluntary actions by the railroads. In
addition, the action plan outlined the agency's development of a new
inspection planning approach. Under this approach, inspectors focus their
efforts on locations that data-driven models indicate are most likely to
have safety problems.
In carrying out its safety oversight, FRA identifies a range of specific
and broad-scale safety problems on railroad systems mainly by determining
whether operating practices, track, and equipment are in compliance with
minimum safety standards. However, FRA is able to inspect only about 0.2
percent of railroads' operations each year and its inspections do not
examine how railroads are managing safety risks throughout their systems
that could lead to accidents. Such an approach, as a supplement to
traditional compliance inspections, is used in the oversight of U.S.
commuter railroads and pipelines and of Canadian railroads. While this
type of approach can provide additional assurance of safety, GAO is not
recommending that FRA adopt it because its various initiatives to reduce
the train accident rate have not yet had time to demonstrate their effects
on safety.
FRA uses a broad range of goals and measures to assess the impact of its
oversight. For example, it has developed (1) new goals to target its
inspection and enforcement programs at reducing various types of railroad
accidents and (2) related measures to monitor its progress. These measures
include the rate of train accidents caused by human behaviors, track
defects, and equipment defects. However, FRA's ability to make informed
decisions about these programs is limited because it lacks measures of
their direct results, such as the correction of identified safety
problems. Furthermore, FRA has not evaluated the effectiveness of its
enforcement program.
Track Inspector Examining Rail to Assess Its Safety Compliance
References
Visible links
26. http://www.gao.gov/cgi-bin/getrpt?GAO-04-378
27. http://www.gao.gov/cgi-bin/getrpt?GAO-06-91
28. http://www.gao.gov/cgi-bin/getrpt?GAO-04-557T
29. http://www.gao.gov/cgi-bin/getrpt?GAO-04-598T
30. http://www.gao.gov/cgi-bin/getrpt?GAO-06-946
31. http://www.gao.gov/cgi-bin/getrpt?GAO-04-38
32. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-00-10
33. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD/AIMD-99-69
34. http://www.gao.gov/cgi-bin/getrpt?GAO-04-38
35. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-00-10
36. http://www.gao.gov/cgi-bin/getrpt?GAO-04-38
37. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD/AIMD-99-69
38. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-00-10
39. http://www.gao.gov/cgi-bin/getrpt?GAO-04-38
40. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-00-10
41. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-99-16
42. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-97-83
43. http://www.gao.gov/cgi-bin/getrpt?GAO-04-38
44. http://www.gao.gov/cgi-bin/getrpt?GAO-03-454
45. http://www.gao.gov/cgi-bin/getrpt?GAO/GGD-00-204
48. http://www.gao.gov/cgi-bin/getrpt?GAO-06-156
49. http://www.gao.gov/cgi-bin/getrpt?GAO-06-91
57. http://www.gao.gov/cgi-bin/getrpt?GAO-07-149
*** End of document. ***