Railroad Safety: New Approach Needed for Effective FRA Safety Inspection Program

Published by the Government Accountability Office on 1990-07-31.

Below is a raw (and likely hideous) rendition of the original report. (PDF)


         I ..--l.“- I_.-_._-.- .--.. -----.---   I__-.
.llllV      I!)!)0

                                                         RAILROAD SAFETY
                                                         New Approach Needed
                                                         for Effective FRA   ~
                                                         Safety Inspection

Resources, Community,   and
Economic Development    Division


July 31, 1990

The Honorable John D. Dingell
Chairman, Committee on Energy
  and Commerce
House of Representatives

The Honorable Carl Levin
Chairman, Subcommittee on
  Oversight of Government Management
Committee on Governmental Affairs
United States Senate

In response to your request, this report addresses the adequacy of the Federal Railroad
Administration’s (FM) railroad safety inspection program. Specifically, the report discusses
FRA’Sinspection coverage standards, how FRAuses data to target inspection activities, follow-
up actions on inspections, and uniformity in the application of safety regulations by
inspectors. The report makes recommendations for improvement in F’RA’Sinspection program.

As agreed with your offices, unless you publicly announce its contents earlier, we plan no
further distribution of this report until 30 days from the date of this letter. At that time, we
will send copies to the Secretary of Transportation; the Administrator of the Federal
Railroad Administration; and other interested parties. We will make copies available to
others upon request.

Our work was performed under the direction of Kenneth M. Mead, Director, Transportation
Issues (202) 275-1000. Other major contributors are listed in appendix IV.

Assistant Comptroller General
Executive Summ~

                   In 1989, 249 Federal Railroad Administration (FRA) and 110 state rail-
Purpose            road inspectors were responsible for inspecting the nation’s 580 rail-
                   roads, which owned approximately 20,000 locomotives, 1.2 million
                   freight cars, and 258,000 miles of track. These inspectors conducted
                   about 63,300 inspections and found over 378,000 safety problems. Con-
                   cerned about railroad safety, the Chairman, House Committee on Energy
                   and Commerce, and the Chairman, Subcommittee on Oversight of Gov-
                   ernment Management, Senate Committee on Governmental Affairs,
                   asked GAO to assess the effectiveness of Fm'S safety inspection program.
                   This report, the third in a series, focuses on FRA’Sinspection coverage
                   standards, how FRA uses data to target railroads for inspection, follow-
                   up actions on inspection results, and uniformity in the application of
                   safety regulations.

                   As an agency of the Department of Transportation (OCR), FRA is respon-
Background         sible for establishing and enforcing safety regulations for the railroad
                   industry. To carry out this responsibility, FRA has issued safety regula-
                   tions that railroads must follow covering track, signals, equipment, and
                   operating practices. The railroads are primarily responsible for
                   inspecting their operations to ensure that they conform to FRA'S regula-
                   tions. FRA and state inspectors, under a cooperative agreement with FRA,
                   monitor railroad compliance with these regulations by routinely con-
                   ducting inspections at railroads. If inspectors find deviations from FM'S
                   safety regulations (defects), their inspection reports identify the defect
                   thereby notifying the railroads that unsafe conditions have been found
                   that must be corrected.

                   FItA'9 safety inspection program does not provide assurance that the
Results in Brief   nation’s railroads are operating safely. GAO found that FRA did not have
                   minimum inspection coverage standards defining the frequency of rail-
                   road inspections or the size of the territory an inspector could be
                   expected to cover. Without such standards, some railroads go unin-
                   spected, and FRA does not know whether its staff is adequate. Also, FRA
                   does not systematically target inspections by integrating available acci-
                   dent, injury, and inspection data. Rather,‘Fm relies on inspectors’ judg-
                   ment and knowledge to plan inspections, which could result in high-risk
                   areas receiving decreased inspection activity. In addition, FM neither
                   requires railroads to report actions taken to correct identified safety
                   problems, nor does it have a systematic follow-up inspection program to
                   determine if railroads correct safety problems. Therefore, FRA has no

                   Page 2                     GAO/RCED-90-194 FRA Inspections: New Approach Needed
                            Executive Summary

                            record to show whether the thousands of safety problems it has identi-
                            fied have been corrected, Finally, GAO found that inspectors were not
                            uniformly applying FRA’S safety regulations, which resulted in FRA
                            regions filing different numbers of violations for the same defective

Principal Findings

Inspection Coverage         FRA has not established inspection coverage standards for the various
Standards Not Established   types of inspections. As a result, inspectors do not have guidance on
                            how often railroads’ equipment, track, signals, or operating practices
                            need to be inspected. FRA also has no standards relating to the size of an
                            inspector’s territory or how much territory an inspector could be
                            expected to cover. The lack of coverage standards has resulted in some
                            railroads not being inspected. GAO found that in 1989 some railroads did
                            not receive any type of inspection. Typically, these were the smaller,
                            short-line railroads, which have higher accident and injury rates than
                            overall industry averages. Specifically, in 1989, of the nation’s 500 rail-
                            roads, 32 received no inspection of any type, 168 did not receive an
                            inspection of their operating practices, 151 did not have their equipment
                            inspected, and 75 that owned track did not receive a track inspection.

                            Because there are no coverage standards, FRA does not know whether its
                            staff of 249 inspectors and 58 supervisory personnel, augmented by 110
                            state inspectors who conduct inspections for FRA, is adequate. FRA offi-
                            cials in the regions visited told GAO that resources were not sufficient to
                            adequately cover their territories. In addition, many state directors told
                            GAO that they were reducing or considering reducing their inspection
                            efforts for FRA due to the elimination of all federal funding for state
                            inspection programs in 1989. Such reductions will further affect FRA’S
                            ability to adequately cover the nation’s railroads.

Data Not Analyzed to        GAO’S 1987 management review of DOTfound that safety indicators need
Target Inspections          to be developed in order to target inspection resourcesi One safety indi-
                            cator is the number of times that railroads are found to be in compliance

                            IDepartmentof Transportation:EnhancingPolicy and ProgramEffectivenessThrough Improved
                            Management(GAO/RCED-87-3,Apr. 13, 1987).

                            Page 3                           GAO/RCED-99-194 FRA Inspectiona: New Approach Needed
                        Executive Summruy

                        or non-compliance with FRA'S regulations. Those in frequent non-compli-
                        ance should be targeted more often for inspection. FRA, however,‘does
                        not analyze existing inspection and accident data to target railroads for
                        inspection. Rather, the selection of railroads for inspection is based on
                        individual inspector judgment and knowledge, not on an analysis of
                        existing safety data to determine where safety problems are occurring
                        and to direct inspections to those problems.

                        GAO  found little relationship between changing accident trends, another
                        safety indicator, and FRAinspection activity. For example, while the
                        number of Union Pacific accidents in Idaho more than doubled between
                        1986 and 1988, the total number of Union Pacific inspections in Idaho
                        decreased nearly 38 percent.

                        Also, in 1980, the Congress required FRAto submit a system safety plan.
                        As part of the plan, FRAwas directed to develop an inspection method-
                        ology that was to consider, among other things, carrier safety records,
                        population centers, and volume of track and equipment used in pas-
                        senger and hazardous materials routes. As a result, FRAdeveloped the
                        National Inspection Plan, but the plan has not fully met the congres-
                        sional mandate. The number of inspections in the plan are based on the
                        number of inspections previously conducted rather than on a method-
                        ology that uses information on carrier safety records or the volume of
                        passenger and hazardous materials traffic in specific locations. FRA has
                        not developed the methodology the Congress envisioned because it lacks
                        information on the volume of track and equipment used in passenger
                        and hazardous materials routes, which it is now in the process of

Ineffective Follow-Up   FRAdoes not know whether thousands of the safety defects it identified
                        have been corrected because it does not require railroads to report
                        actions taken to correct safety defects. Railroads voluntarily report
                        actions taken to correct defects identified during some FRAinspections,
                        but this voluntary reporting is incomplete. GAO found that between 1986
                        and 1988, track inspectors found 360,683 track defects and signal
                        inspectors found 34,813 signal defects. FRA had no record of actions
                        taken by the railroads to correct nearly 40,000 (about 11 percent) of the
                        track defects and 6,039 (about 14 percent) of the signal defects. Some
                        railroads report actions taken to correct equipment and operating prac-
                        tices defects, but FRA does not maintain records of these reports. FRA,
                        therefore, does not know whether the approximately 6,000 operating
                        practices and 200,000 equipment defects cited each year were corrected.

                        Page 4                      GAO/RCED-99-194 FRA Inspections: New Approach Needed

                          FWA,as a matter of practice, does not routinely perform follow-up rein-
                          spections to determine whether safety defects are corrected. In those
                          cases where an FRA reinspection was performed, GAO found examples of
                          previously cited track and signal defects not corrected by the railroad.
                          Upon reinspection, these defects were cited as serious violations.

Application of Safety     The Federal Rail Safety Act of 1970, as amended, requires that FRA
Regulations Not Uniform   inspectors uniformly apply safety regulations throughout the railroad
                          industry. GAO'S analysis of inspection activities for 1986 through 1988
                          showed that uniform application of safety regulations was not being
                          achieved, which results in FRAregions filing different numbers of viola-
                          tions for the same defective safety condition.

                          GAO found numerous examples of one FRAregion filing many more viola-
                          tions than another for the same defective condition between 1986 and
                          1983 In 1988, one FM region cited one certain track defect 312 times
                          but filed no violation reports, while another region found the same
                          problem 433 times and cited it as a violation in 166 cases. GAO also found
                          that a number of inspectors had a pattern of not citing railroads for vio-
                          lations Uniformity problems exist because guidance has been unclear
                          for issuing violations and training has been inadequate.

                          GAO is making a number of recommendations to the Secretary of Trans-
Recommendations           portation that will provide FRA with a better measure of whether the
                          nation’s railroads are operating safely. These recommendations include
                          defining inspection coverage, refining its approach to targeting inspec-
                          tions, and establishing a follow-up inspection program. GAO is also rec-
                          ommending actions to ensure that safety standards are uniformly

                          GAO discussed its findings with the Administrator, FM, who generally
Agency Comments           agreed with GAO'S findings, especially the need to make the inspection
                          approach less random and more scientific by using available data. He
                          also said that FRA was hiring a Director of Training and Communication
                          who would (1) design training programs for both new and current
                          inspectors to ensure that they have analytical capabilities and (2) keep
                          inspection manuals up-to-date to better ensure uniformity in applying
                          safety standards. As requested, however, GAO did not obtain written
                          agency comments on a draft of this report.

                          Page6                       GAO/RCED-90-194 FRA Inspections: New Approach Needed

Executive Summary                                                                                        2

Chapter 1                                                                                                8
Introduction                 FRA’s Approach to Railroad Safety
                             Prior Railroad Safety Reports
                             Objectives, Scope, and Methodology                                         10

Chapter 2                                                                                               13
FRA’s Inspection             FR.A Has No Inspection Coverage Standards
                             Safety Data Not Used to Target Inspections
Program Not Effective        Ineffective Follow-Up and Reinspection                                     23
                             Conclusions                                                                25
                             Recommendations                                                            26
                             Views of Agency Officials                                                  26

Chapter 3                                                                                               28
FM   Inspectors   Do   Not   Regional Differences Exist in Enforcement                                  28
                             Causes of Lack of Uniformity                                               29
Uniformly Apply              Conclusions                                                                31
Safety Standards             Recommendations                                                            31
                             Views of Agency Officials                                                  31

Appendixes                   Appendix I: Reliability Assessment of FRA Inspection                       32
                             Appendix II: States Participating in FRA Inspection                        35
                             Appendix III: Major Contributors to This Report                            36

Tables                       Table 2.1: FRA and State Track and Equipment
                                 Inspections, 1988
                             Table 2.2: FRA Resources Used to Complete System                           22
                                 Assessment Field Work
                             Table 2.3: Time Between Railroad Written Response and                      23
                                 FRA Follow-Up Review
                             Table 2.4: Total Track and Signal Defects and Violation                    24
                                 Reports as Compared to Defect and Violation Reports
                                 With No Follow-Up, 1986-88
                             Table 3.1: Enforcement Actions by Inspector Discipline,                    30

                             Page 0                     GAO/RCED-W-194 FRA Inspections: New Approach Needed

Table I. 1: Number of Record Header Fields and Detail                        33
    Fields Reviewed
Table 1.2: Estimated Error Bounds by RIRS Data File                          34


AAR         Association of American Railroads
Dm          Department of Transportation
FRA         Federal Railroad Administration
GAO         General Accounting Office
RIRS        Railroad Inspection Reporting System

Page 7                      GAO/RCED-90-194 FRA Inspections: New Approach Needed
Chapter 1


                    The Federal Railroad Safety Act of 1970, as amended, directed the Sec-
                    retary of Transportation to prescribe regulations for all areas of rail-
                    road safety. The Secretary has delegated these responsibilities to the
                    Federal Railroad Administration (FRA). F’RA’Ssafety mission includes (1)
                    establishing federal rail safety rules and standards; (2) inspecting rail
                    carrier track, signals, equipment, and operating practices; and (3)
                    enforcing federal safety rules and standards. The nation’s railroads are
                    primarily responsible for conducting safety inspections of their equip-
                    ment and facilities. FRA’S responsibility is to monitor the inspection
                    activity of the railroads. The railroads employed approximately 35,000
                    track and signal personnel, and 75,000 motive power and equipment
                    maintenance people in 1987.

                    To carry out its safety mission, FRA has established five inspection disci-
FRA’s Approach to   plines and eight regional offices under the direction of an Associate
Railroad Safety     Administrator for Safety in Headquarters. Inspectors specialize in one
                    discipline and are generally not cross-trained. Each discipline and a brief
                    description of the inspectors’ responsibilities follow:

                    track: enforcement of Federal Track Safety Standards, including com-
                    muter track;
                    locomotive power and equipment: all aspects of the design and operation
                    of rolling railroad equipment, including locomotives;
                    operating practices: enforcement of federal operating regulations, car-
                    rier rules and practices, and train operations;
                    signal and train control: all aspects of signal switching systems, train
                    control, and locomotive cab signal devices; and
                    hazardous materials: all aspects of the rail transportation of hazardous
                    materials, including the enforcement of safety regulations at shippers
                    and railroads.

                    The primary responsibility of the inspectors is to conduct routine
                    inspections of railroads. FRAalso conducts system assessments where
                    inspectors from all disciplines are brought together to conduct an assess-
                    ment of all aspects of a railroad’s operation. FRA inspectors also investi-
                    gate accidents and complaints.

                    The regional offices are staffed with a director, safety specialists in
                    each discipline, and inspectors. The safety specialists are responsible for
                    assisting inspectors in planning and conducting inspection activities in
                    their disciplines. When inspections or complaint investigations reveal
                    noncompliance with the laws, inspectors are to list each noncomplying

                    Page 8                      GAO/RCED-90-194 FRA Inspections: New Approach Needed
                        Chapter 1

                        condition (defect) in an inspection report, When the inspector identifies
                        defects that pose an immediate safety hazard, he is to prepare a viola-
                        tion report that is submitted to FRA'S Office of Chief Counsel. The Office
                        of Chief Counsel is to process the violation and assess the railroad a civil

                        To inspect the nation’s rail system, FRA had, as of December 1989, a total
                        of 249 inspectors: 66 track, 86 equipment, 41 operating practices, 33
                        signal, and 33 hazardous materials. Its 249 inspectors and 68 supervi-
                        sory personnel must inspect a rail system consisting of approximately
                        600 railroads, 20,000 locomotives, 1.2 million freight cars, and 258,000
                        miles of track. In addition, FRA has also relied on 110 state inspectors in
                        33 states to perform inspections for FRA under a cooperative agreement.

                        In 1989, FRA and state inspectors conducted 63,278 inspections and cited
                        over 378,000 defects. FRA also conducted system assessments of the
                        Montana Rail Link Railroad and the Oklahoma, Kansas and Texas Rail-
                        road in that year. In 1988, FRA and state inspectors conducted 65,309
                        inspections and cited over 360,000 safety defects. FRA also conducted
                        two system assessments in 1988-one of the Guilford Transportation
                        Industries and another on the New Jersey Transit Rail Operations.

                        Three prior GAO reports noted problems with FRA’S safety program. Our
Prior Railroad Safety   April 1989 report analyzed the accuracy of accident and injury data
Reports                 reported by railroads to FRA. We discovered problems with the accuracy
                        of the data submitted by railroads and made recommendations that will
                        improve the accuracy of reporting by railroads. FRA agreed with our rec-
                        ommendations and is taking actions to improve reporting that include
                        conducting more records inspections and requiring railroads to set up
                        internal controls for reporting.’

                        In our November 1989 report on hazardous materials transportation, we
                        noted that FRA had an inadequate number of hazardous materials inspec-
                        tors to ensure that railroads and shippers were complying with safety
                        regulations.” Based on our report, the Secretary of Transportation
                        reported an inadequate number of hazardous materials inspectors as a
                        material weakness under the Federal Managers’ Financial Integrity Act.

                        ‘Railroad Safety:FRA Needsto CorrectDeficienciesin ReportingInjuries and Accidents(GAO/
                              _ 9- 109,Apr. 5, 1989).
                        “Railroad Safety:DOTShouldBetter ManageIts HazardousMaterialsInspectionProgram(GAO/
                        m A- -43) Nov.17,1989).

                        Page 9                            GAO/RCED-90-194 FRA Inspections: New Approach Needed
                        Chapter 1

                        The Secretary reported that the insufficient number of inspectors pre-
                        vents FRA from providing adequate coverage of hazardous materials
                        shippers and railroads. Also in response to our recommendations, FRA
                        has stated that it will hire more hazardous materials inspectors and
                        rewrite its Hazardous Materials Enforcement Manual to revise its
                        approach to shipper and carrier inspections.

                        In April 1990, we reported on FRA’S inspection activities in one of its
                        regional offices”. In that report, we concluded that the hazardous mater-
                        ials inspection program was hampered by the lack of written inspection
                        goals and possibly by inadequate inspector resources.

                        The Chairman, House Committee on Energy and Commerce, and the
Objectives, Scope,and   Chairman, Subcommittee on Oversight of Government Management,
Methodology             Senate Committee on Governmental Affairs stated that they were con-
                        cerned about safety on the nation’s freight and passenger railroads and
                        requested that we assess the effectiveness of FRA’S safety inspection
                        program. Our overall objective was to assess the adequacy and effi-
                        ciency of FFLA’Sapproach to safety inspection activities. More specifi-
                        cally, we determined whether FRA had coverage standards indicating
                        how often railroads and equipment should be inspected, how FRA uses
                        inspection data to target railroads for inspection, actions taken by rail-
                        roads to correct identified safety defects, and whether inspectors were
                        uniformly applying FRA’S safety regulations when conducting inspec-
                        tions Because of our November 1989 report on hazardous materials
                        inspections, we did not include hazardous materials inspections in this

                        We conducted our review at FRA headquarters offices, and at four of its
                        eight regional offices (Atlanta, Georgia; Chicago, Illinois; Fort Worth,
                        Texas; and Portland, Oregon). We selected these four regions to provide
                        a balance of regions with stable or rising accident rates and provide
                        nationwide coverage. We also discussed FRAinspection activities with
                        officials from the National Transportation Safety Board, the Association
                        of American Railroads, the Soo Line Railroad, and state railroad officials
                        for all 33 states participating in the FRA state inspection program.
                        Appendix III contains a listing of the states that conduct FRAinspections.

                        “Railroad Safety: MoreFRAOversightNeededto EnsureRail Safety in Region2 (GAO/
                              _ 0_140,Apr. 27, 1990).

                        Page 10                          GAO/RCED-90-194 FRA Inspections: New Approach Needed
Chapter 1

We reviewed federal laws and regulations on railroad safety, and FRA’S
General Manual and inspection manuals to determine FRA’Sinspection
requirements, policies, and standards. We analyzed data on FRA inspec-
tion activities contained in the Railroad Inspection Reporting System
(HIRS) database. Because of the importance of this automated system to
our findings, we conducted a reliability assessment of the database at
FRA’Sdata entry contractor, I-NET in Bethesda, Maryland. We found the
accuracy of the data to be high. The methodology and results of this
assessment are contained in appendix I.

To assess the coverage and uniformity of FRA inspections, we inter-
viewed regional safety directors, deputy regional directors, district
chiefs, specialists, and inspectors. We discussed the same matters with
the directors of the Office of Safety Analysis and the Office of Safety
Enforcement and their staffs in headquarters. We analyzed RIRS data on
individual inspectors and data on railroads provided by the Office of
Safety Analysis.

To assess how FIU uses inspection data to target its inspection activities,
we reviewed FRA’S National Inspection Plans, and interviewed regional
safety directors and specialists. We also interviewed and obtained data
from headquarter’s Office of Safety Analysis officials responsible for
the plans.

To assess the efficiency of FRA’S railroad system assessments, we
reviewed the three assessments of Class I railroads conducted in 1987
and 1988. We compared the results of system assessment findings to
routine inspection data in the RIRS. We also obtained data on the cost of
system assessments from the Office of Safety Enforcement and inter-
viewed FRAspecialists and inspectors who took part in system

We conducted a telephone survey of the program directors for the 33
states who under a cooperative agreement with FRA conduct inspections
for FRA to discuss their perceptions about the program and determine
their future level of participation. We also discussed FM-state coordina-
tion with FRA regional safety directors and safety specialists in each of
the four regions visited.

We conducted our work from April 1989 to June 1990 in accordance
with generally accepted government auditing standards. We discussed
the factual information in this report with the FRA Administrator and
other top level officials responsible for railroad safety. A discussion of

Page 11                     GAO/RCED-99-194 FRA Inspections: New Approach Needed
    Chapter 1

    their comments are contained in the body of the report. As requested,
    however, we did not obtain official agency written comments on the
    report draft.


    Page 12                    GAO/RCED-99-194 FRA Inspectione: New Approach Needed
Chapter 2

FRA’s Inspection Program Not Effective

                      FFU’Sinspection program is not effective in ensuring that the nation’s
                      railroads are operating safely. FRA'S inspection program does not (1)
                      include inspection coverage standards prescribing how often railroads
                      should be inspected; (2) target railroads for inspection on the basis of
                      problems identified through analysis of existing accident, injury, and
                      inspection data; and (3) require railroads to report actions taken to cor-
                      rect safety problems identified during FRA inspections. Because FRA'S
                      inspection program is ineffective, many railroads go uninspected. Also,
                      FRAdoes not know if (1) its inspector workforce is sufficient to carry out
                      its safety mission, (2) high-risk railroads are not targeted for inspection,
                      and (3) safety problems identified during inspections are being

                      FRAdoes not have minimum inspection coverage standards that pre-
FRA Has No            scribe (1) how often railroads should be inspected, (2) the size of an
Inspection Coverage   inspector’s territory, (3) the number of inspection locations an inspector
Standards             can reasonably cover, and (4) how often inspection locations should be
                      covered. Neither do FF&i policies and manuals set minimum standards
                      describing adequate levels of inspection coverage. For example, there is
                      no clear standard for how often a passenger or hazardous materials
                      route should be inspected, FRAis in the process of collecting data on
                      regional inspection points in an effort to determine the size of each
                      inspector’s territory and how often inspections are conducted in that
                      territory. Because no coverage standards exist, individual inspectors
                      decide what constitutes adequate coverage. Without coverage standards
                      railroads go uninspected, and FRAcannot accurately determine how
                      many inspectors it needs. Furthermore, it is uncertain whether FRA can
                      continue to rely on assistance from state inspectors and how this might
                      affect the number of inspectors FXAneeds for proper coverage.

Many Railroads Not    Our review of FRA inspection data showed that during 1989 some rail-
Inspected             roads received no inspection of any type. Because of the size and diver-
                      sity of the railroad industry, it may not be possible for FRA to cover each
                      railroad in all five inspection disciplines each year. Typically those rail-
                      roads not inspected are the smaller, short-line railroads, but on occasion
                      railroads with significant activity are also not inspected. The presence
                      of short-line railroads, which have had consistently higher accident and
                      injury rates than the industry as a whole is growing. The number of
                      train miles of smaller railroads has increased 38 percent between 1986
                      and 1988. These railroads accounted for about 11 percent of the rail

                      Page 13                     GAO/RCEMW-194 FRA Inspections: New Approach Needed
    Chnptm 2
    F?U’s Inspection Program Not Effective

    activity in the United States. Many other railroads did not receive an
    inspection in one of the disciplines. For example, in 1989

l for railroads that operated at least 1,000 train miles, 32 did not receive
  an inspection of any type;
. operating practices inspections were not performed at 168 of the
  nation’s 680 railroads;
l of 415 railroads owning track, 75 with a total of 1,164 miles of track did
  not receive a track inspection;
. of 484 railroads filing accident or incident reports, 265 were not
  inspected to verify the accuracy of their reporting; and
. the Alaska Railroad, a regional-sized railroad with over 316,000 passen-
  gers, 600 miles of track, and 1,300 cars and locomotives, did not receive
  a rail safety inspection of any type. Approximately 49 percent of this
  railroad’s freight-ton miles’ involve hazardous materials shipments.

    FRA  Office of Safety officials said that FRA inspectors could not ade-
    quately inspect the nation’s entire rail system in a given year. Officials
    in each FRA region we visited told us that they could not adequately
    cover their assigned territories. Different FRA regions have different
    interpretations of what constitutes adequate inspection coverage. For
    example, one regional signal specialist defined adequate coverage as
    inspecting every signal at least once every 1 to l-1/2 years, while a
    signal specialist in another region defined adequate coverage as
    inspecting every signal once every 2-l/2 years, Additionally, one
    regional equipment specialist defined adequate coverage as inspecting
    50 percent of the cars operating in the region each year, while another
    regional equipment specialist defined adequate coverage as inspecting
    20 percent of the cars operating in the region each year. Without cov-
    erage standards, inspectors and regions’ judgments in deciding what
    constitutes adequate coverage can vary widely.

    Vacancies in existing inspector positions have contributed to coverage
    problems. In some cases, inspector positions remained vacant for more
    than a year resulting in coverage reductions or no coverage for some
    locations. For example, one track inspector stated that, because of
    vacancies, he could only inspect mainline track once every 10 months,
    instead of twice a year, as he considered necessary. Because of vacan-
    cies, total FRA inspections, which numbered about 71,000 in 1986,
    dropped each successive year to about 63,300 in 1989. In an effort to
    solve this problem, FRA hired 35 new inspectors in 1989. However, until

    ‘A unit of measurewhich measuresoneton of freight moving onemile.

    Page 14                           GAO/RCED-90-194 FRA Inspections: New Approach Needed
                                     Chapter 2
                                     FRA’s Inspection Program Not Effective

                                     FHAdefines the level of coverage it wants in each discipline, it will have
                                     no basis for determining whether its inspection resources are adequate
                                     to provide that level of coverage.

Continued State                      The Federal Railroad Safety Act of 1970, as amended, augmented the
Participati .on Uncertain            federal inspection force by allowing state inspectors meeting FRA certifi-
                                     cation standards to perform track, equipment, signal, and operating
                                     practices inspections, provided the state agreed to participate in a coop-
                                     erative program with FRA by following FRA inspection procedures, filling
                                     out FM inspection forms, and citing railroads for defects and violations.
                                     As of September 1989,33 states employed 110 inspectors: 62 track, 35
                                     equipment, 11 operating practices, and 2 signal. Data on inspections con-
                                     ducted by state inspectors is recorded in FM'S inspection data base.

                                     FRAuses state inspectors to provide significant amounts of track and
                                     equipment inspection coverage. As shown in table 2.1, states inspected
                                     45 percent of the approximately 400,000 miles of track inspected by FRA
                                     and state inspectors in 1988. States also inspected 33 percent of the
                                     approximately 1,146,OOOcars inspected by FFtA and state equipment
                                     inspectors. Because of the large number of inspections performed by
                                     state inspectors, 26 of the 33 state program directors stated that FRA
                                     could not adequately cover their states if the state programs did not

Table 2.1: FRA and State Track and
Equipment Inspections, 1988                                                 Miles                                 Cars
                                     FRA    region
                                      ..______-.._                  FRA                 State               FRA                State
                                     - _._-- --                    17,188              12,277              56,488             22,837
                                     2--..-..- ------         _-
                                                                   29,050              29,091             165,867             172,449
                                     3                             32.613              28.328             124.150              81.973
                                     4 --.------
                                     ._..._              . . --- 29,579                32,918             121,572                     0
                                     5                             43,886              18,687              93,377
                                                                                                           ___-.   ~----       31,278
                                     6                             38,825              30,482             128,816              21,662
                                     7                              8,206              15,826              37,766 -_           26,630
                                     8                             21,125              12,263              40,831              19,869
                                     Total                       220,472             179,872             788,887             378,898
                                     Source,   FRA   Railroad Inspection Reporting System

                                     Until fiscal year 1989, FRA paid a portion of each state inspector’s
                                     salary. Since that time, FM has not funded the state inspector program.
                                     As a result, several states altered their inspection programs or men-
                                     tioned possible future changes. Ten state directors said that they

                                     Page 16                                  GAO/RCEDSO-194 FRA Inspections: New Approach Needed
                        Chapter 2
                        FRA’e Inspection Program Not Effective

                        responded to the elimination of funding#by increasing state-related
                        inspection activities and decreasing m-related inspections. The poten-
                        tial exists for FRA’S inspection program, particularly in track and equip-
                        ment, to be seriously limited if states completely abandon their railroad
                        inspection programs.

                        Many state directors we contacted complained about communication and
                        coordination problems between FRAand their states. State program
                        directors also said that coordination of inspection activities and commu-
                        nication between state and federal inspectors is poor. Fourteen of the 33
                        states we contacted characterized FRA’Scoordination of inspection activ-
                        ities with their states as poor or inadequate. Good communication is
                        important because state and FFWinspectors may not have distinct, sepa-
                        rate inspection territories. Poor communication causes inspectors to
                        overlap inspection activities making the program less efficient. State
                        directors, FRA inspectors, and railroad officials we talked to cited
                        instances of state and FRA inspectors inspecting the same railroad loca-
                        tion only days apart.

                        State directors do not expect that poor communication between federal
                        and state inspectors will improve in the future. In fact, four state direc-
                        tors said that the relationship between federal and state inspector
                        offices will deteriorate further given the elimination of federal funding.

                        In our 1987 management review of DOT,we noted that FRAneeds to
Safety Data Not Used    develop safety indicators to effectively deploy its inspector resources.2
to Target Inspections   One safety indicator is whether or not railroads are inspected. As noted
                        earlier in this report, FRA does not have minimum coverage standards
                        that has resulted in some railroads going uninspected, especially in cer-
                        tain disciplines. Another safety indicator is railroads with poor safety
                        histories. FRA does not use available data to target routine inspections
                        toward high-risk locations and railroads with poor safety histories
                        making inspections less effective. FFL4targets routine inspections poorly
                        because FRA’S Office of Safety sends regional offices accident data that
                        safety specialists often do not analyze. Instead, the specialists rely on
                        individual inspectors to independently schedule their own inspection

                                                    : EnhancingPolicy and ProgramEffectivenessThrough Improved
                        Management(GAO/             3S,Apr. 13, 1987).

                        Page 16                           GAO/RCED-99-194 FRA Inspections: New Approach Needed
                          Chapter 2
                          FRA’s Inspection Program Not Effective

                          Another safety indicator could be the volume of usage of track and
                          equipment. FR.A created the National Inspection Plan (NIP) in response to
                          the Congress’ directive in 1980 to submit a safety plan, As part of this
                          plan, FRAwas required to develop a methodology to determine the fre-
                          quency and schedules of safety inspections, giving appropriate priority
                          to track and equipment involved in passenger trains and hazardous
                          cargos, The methodology was to also analyze and take into account the
                          safety records of the railroads, location of track and equipment, and
                          volume of usage of track and equipment. FRA has yet to develop such a
                          methodology because it does not gather information on volume of traffic
                          on individual passenger and hazardous materials routes.

                          Whether a railroad is experiencing a particular safety problem or
                          problems could be another safety indicator and can be determined from
                          an analysis of routine inspection data. FRA’S approach to system assess-
                          ments is inefficient because it does not use safety data from its routine
                          inspections to target assessment activity to problem areas. Instead, FFL4
                          conducts hundreds-sometimes       thousands-of    additional routine
                          inspections during system assessments.

Routine Inspections Not   Instead of using the safety data it collects to target its inspection efforts,
                          FR.A relies on its inspectors’ personal judgment and knowledge to deter-
Targeted to High-Risk     mine where inspections are needed. We found little relationship between
Railroads                 changing accident trends and FRA inspection activity. As a result, rail-
                          roads with increasing numbers of accidents did not receive additional
                          inspection coverage.

                          Because of the dependence on individual inspectors, we found many
                          instances in which FRA did not respond to an increase in accidents on a
                          railroad by increasing inspections. In fact, in many cases inspections
                          actually decreased, indicating a misallocation of inspection resources
                          because the worsening level of safety on a railroad should require more,
                          not less, inspection resources to determine the cause for the rise in acci-
                          dents. For example, while the number of CSX accidents due to human
                          error increased in Tennessee by nearly 67 percent between 1986 and
                          1988, FRA decreased operating practices inspections on the railroad by
                          about 45 percent. At the same time in Tennessee, FRA increased oper-
                          ating practices inspections on the Norfolk Southern Railroad by 41 per-
                          cent, even though Norfolk Southern reported only four accidents due to
                          human error in 1986 through 1988. We found comparable examples at
                          other railroads for other inspection disciplines.

                          Page 17                           GAO/RCED-90-194 FRA Inspections: New Approach Needed
Chapter 2
FRA’s Inspection Program Not Effective

FFU’S  current approach to routine inspections is primarily based on the
inspector’s knowledge of his territory. Regional specialists and district
chiefs we interviewed emphasized that inspectors are in the best posi-
tion to determine which railroads need inspecting because they are more
familiar with increases and decreases in traffic on carriers throughout
the year and know where the problems are in their territory. Thus, FRA
relies on the inspectors’ judgment to schedule their inspection activities.
These schedules are rarely revised by regional or headquarter’s

FRA   recently has experienced a significant turnover in inspectors, which
affects the inspectors’ overall level of experience and knowledge of their
territories. In 1989, FRA hired 35 of its 249 inspectors. According to the
Director of the Office of Safety Analysis, by the end of 1990 approxi-
mately half of FR4’S inspectors will have been hired within the last 2
years. This condition raises serious questions about FM’S current non-
systematic approach to inspection activity because much of the inspec-
tion workforce will not have FRA experience or specific knowledge of
their territories.

FRA’S  Office of Safety Analysis does little analysis of its safety data to
highlight problems or trends for the regional offices. Instead, regional
offices receive a monthly, unanalyzed listing of all accidents which
occurred in the region. Each discipline has a specialist who is respon-
sible for analyzing this data and adjusting inspection efforts accord-
ingly. Regional specialists we interviewed, however, did not
systematically analyze this data and rarely changed inspection activity
because of it.

FRAOffice of Safety Analysis officials stated that FRA needs to improve
its targeting of inspection resources based on an analysis of all available
safety data. These officials explained that FRA currently does not have
an effective methodology to consolidate safety data contained in several
reporting systems because they were designed independently to address
various safety and policy reporting requirements within FILL They also
stated that FRA needs to develop additional data on carriers’ volume of
traffic in specific inspector territories. With this data and the integra-
tion of existing data, these officials said that they could better target
inspection resources.

Page 19                           GAO/RCED-90-194 FRA Inspections: New Approach Needed
                          Chapter 2
                          FRA’s Inspection Program Not Effective

NIP Does Not Prioritize   In response to the 1980 congressional directive to submit a safety plan,
High-Risk Routes          FRAdeveloped the NIP. FRA also developed the Statistical Analysis
                          Report (SAR) as one element of the NIP. The SAR shows the percentage
                          of accidents each railroad in a region experienced during the previous
                          13 quarters. Each region receives a copy of the SAR to use in developing
                          their portion of the NIP. The FRA official who created the SAR method-
                          ology in the early 1980s said that it was to be a starting point in com-
                          plying with the congressional directive. We found that FRA has not
                          improved the SAR methodology by collecting data it needs on passenger
                          and hazardous materials movements in specific locations. As a result,
                          FR4 has yet to fully develop a methodology for scheduling inspections
                          that incorporates data on track and equipment involved with passenger
                          trains and hazardous cargos, volume of usage of track and equipment,
                          and location of track and equipment in population centers as directed by
                          the Congress.

                          FRAregional officials at the four regions we visited did not document or
                          analyze inspection results on passenger and hazardous materials routes
                          in preparing the NIP. Although officials in each region stated that such
                          routes should be inspected twice a year, none of the four regions could
                          provide us with the total amount of inspection coverage of these routes
                          in 1988 nor the findings of those inspections. According to the regional
                          safety specialists we interviewed, such information was not tracked or
                          used in determining the number and types of inspections. Instead,
                          inspectors were expected to know where such major routes were in their
                          territories and provide them adequate coverage.

                          FRAdoes not target high-risk locations through the NIP. For example,
                          one FRA region planned to decrease its track inspection coverage of the
                          Union Pacific’s track from 4,100 miles in 1988 to 2,880 in 1989 even
                          though (1) the number of Union Pacific track-caused accidents in the
                          region had nearly doubled between 1987 and 1988, (2) Union Pacific
                          carries heavy hazardous materials traffic over more than 2,800 of its
                          6,158 miles of track in the region, and (3) Amtrak trains operate over
                          740 miles of Union Pacific track in the region.

                          In addition, the NIP currently does an inadequate job of targeting high-
                          risk areas because FRA does not collect data on specific locations that
                          carry a high volume of passenger and hazardous materials traffic.
                          Rather, FRA depends on inspector knowledge of their territories to target
                          high-risk locations. FRAOffice of Safety officials stated that they plan to
                          collect data within the next year that would enable them to target high-
                          risk areas in the 1991 NIP.

                          Page 19                           GAO/RCED-90-194 FRA Inspections: New Approach Needed
                       Chapter 2
                       FRA’s Inspection Program Not Effective

NIP Not Used as an     FRA does not use the NIP as a management tool to effectively plan, mon-
Effective Management   itor, execute, and evaluate FRA'S inspection program. We noted this in
                       our 1987 management review of DOT, and the same situation currently
Tool                   exists.

                       Regional plans receive a minimum amount of review by the Office of
                       Safety and are not integrated to provide a national inspection approach.
                       We found instances where regional submissions were identical to the
                       previous year’s submission and not changed by FRA headquarters. For
                       example, one FFtA region planned to inspect, for every railroad in the
                       region, exactly the same number of cars, locomotives, and equipment
                       records in 1989 as it planned to inspect in 1988. FRA Office of Safety
                       officials responsible for developing the NIP estimated that regional sub-
                       missions were changed approximately 10 percent of the time. They also
                       acknowledged that the NIP needs greater national integration and that
                       they are working toward that goal.

                       Planned system assessments are often not included in the NIP.
                       According to FRA'S Acting Director of the Office of Safety Enforcement,
                       this occurs because FRA must respond to immediate safety problems and
                       congressional concerns, and needs to reserve judgment on the selection
                       of railroads for system assessments based on these factors. However,
                       regional officials stated that if system assessments are not included in
                       the NIP, the number of planned inspections is not representative of
                       what may actually occur if a system assessment is conducted in that
                       region. For example, in our analysis of FRA inspection data, we found
                       that in 1987 one FRA region exceeded its planned car inspections of the
                       Soo Line Railroad by nearly 70 percent because it conducted a major
                       part of the Soo Line system assessment. In contrast, the same region
                       inspected approximately 29 percent of the Union Pacific cars it had
                       planned to inspect in that year, in part because of the number of inspec-
                       tions conducted during the Soo Line system assessment.

                       We also found that FRA regions and the Office of Safety did not monitor
                       their performance against the plan for specific railroads. Instead,
                       regional specialists tracked inspection totals for each discipline and
                       relied on inspectors to ensure coverage of specific railroads. The Office
                       of Safety tracked performance quarterly on the basis of total number of
                       inspections and the average time taken to complete each inspection.
                       Neither approach ensures that the inspections are adequately covering
                       individual railroads because FRA regions do not know how often a spe-
                       cific railroad was covered in a given year relative to planned activity.

                       Page 20                           GAO/RCED-90-194 FRA Inspections: New Approach Needed
                               PI&i’s Inepectlon Program Not Effective

                               Finally, FRA does not assess each region’s performance at the end of the
                               year to determine if regional goals were met. FRAhas not developed an
                               evaluation mechanism to judge how actual inspection performance com-
                               pares to planned performance. Our analysis of FRA inspection data
                               showed that FRA regions often greatly exceeded or did not meet their
                               planned inspection activity for specific railroads. For example, one FRA
                               region exceeded the number of inspections of Burlington Northern’s
                               track records that it planned to inspect in 1988 by 413 percent. At the
                               same time, it inspected only 12 percent of the records for Florida East
                               Coast Railroad’s track that it planned to inspect, Another FRA region
                               planned to conduct 182 inspections of operating rules on the Burlington
                               Northern in 1988 and 69 such inspections on the Southern Pacific Rail-
                               road. However, the region actually conducted twice as many inspections
                               as planned for the Southern Pacific but only half as many as planned for
                               the Burlington Northern.

FM’s Approach to System        During system assessments, FFtA often finds problems with a railroad’s
AssessmentsNot Efficient       operation that it could have identified if it had analyzed existing data
                               from its routine inspections. Our analysis shows that, had FRA analyzed
                               existing inspection data for each of the railroads involved in the most
                               recent system assessments, it would have realized that it had already
                               identified safety problems and could have eliminated the expenditure of
                               a tremendous amount of resources. By analyzing existing inspection
                               data, FRA could have focused its attention during the system assessment
                               on finding causes of the safety problems that were already identified.
                               We reviewed the three most recent system assessments conducted on
                               Class I railroads that were issued by the time of our review: the Soo Line
                               (1987), Conrail (1987), and Guilford Transportation Industries (1988).
                               Examples of what we found in the three inspection disciplines of oper-
                               ating practices, track, and signal follow:

                           l During the Conrail system assessment, about 14 percent of all operating
                             practices defects involved failure to report accidents. Inspection data
                             prior to the assessment, however, shows that FRA had already discov-
                             ered the problem and should have been looking for the cause of the
                             problem. In both 1986 and prior to the assessment in 1987, FM cited
                             Conrail for failure to report accidents at a rate over double the national
                             average for 1986 and 1987.
                           . During the Soo Line assessment, about 12 percent of all track defects
                             involved defective rails. Inspection data prior to the assessment reveals
                             that in 1986, about 11 percent of all Soo Line track defects involved
                             defective rails-higher   than the national average of about 7 percent.

                               Page 21                            GAO/RCED-90-194 FRA Inspections: New Approach Needed
                                     Chapter 2
                                     F&~‘S Inspection   Program   Not Effective

                                     Prior to the assessment in 1987, nearly 10 percent of all Soo Line track
                                     defects concerned defective rails, while the national average was
                                     approximately 7 percent.
                                   . During the Guilford assessment, about 24 percent of all signal defects
                                     concerned signal wires on pole lines and cables. Inspection data prior to
                                     the assessment reveals that in 1987, about 36 percent of all Guilford
                                     signal defects involved wires on pole lines and cables - more than three
                                     times the 1987 national average of about 12 percent. In 1986,16 percent
                                     of all Guilford signal defects involved pole lines and cables, which
                                     exceeded the national average of 12 percent in that year. If FRA had ana-
                                     lyzed prior inspection data for Guilford, it would have discovered the
                                     problem earlier and could have brought it to Guilford’s attention.

                                     FRA’S  inspection force consists of about 300 inspectors and specialists,
                                     and system assessments require a large number of these resources. Rec-
                                     ognizing this in 1986, FRArequested five new deputy regional director
                                     positions to handle the increased work load caused by system assess-
                                     ments. Table 2.2 shows the level of resources used for the three assess-
                                     ments of Class I railroads we reviewed.

Table 2.2: FRA Resources Used to
Complete System Assessment Field     Dollars in thousands
Work                                                                                                  Inspections
                                     Railroad                         Inspections involved              performed         cost
                                     Conrail                                            60                  4,152        $1,277
                                     Guilford                                           40                    771           214
                                     So0 Line                                           62                    652           194
                                     Source: FRA Office of Safety Enforcement

                                     Several FRA regional officials and inspectors we interviewed were crit-
                                     ical of how system assessments are currently conducted and of the
                                     resources involved. One specialist who participated in the Conrail
                                     system assessment commented that the assessments direct valuable
                                     inspector resources away from routine inspections and are not highly
                                     productive. He added that it would be more beneficial for FRAto analyze
                                     current inspection data to identify problem areas on particular railroads
                                     rather than to exert such a large effort reviewing the entire operations
                                     of a railroad. One acting regional director believed that system assess-
                                     ments were too large, involved inspectors from too many regions, and
                                     covered disciplines that may not be problem areas.

                                     Because current system assessments are labor intensive and FRA does
                                     not use inspection data in planning and conducting these assessments,

                                     Page 22                               GAO/RCED-99-194 FRA Inspections: New Approach Needed
                                           Chapter 2
                                           FRA’s Inspection Program Not Effective

                                           we believe FRA is not efficiently using its inspection resources. By sys-
                                           tematically planning and conducting system assessments using the data
                                           it has, FRA could conduct more system assessments and better investi-
                                           gate problem areas to bring to the attention of top railroad management.

                                           Follow-up on system assessments has not been timely. The Department
                                           of Transportation (DCW)Inspector General (IG) in 1984 recommended
                                           that FRAestablish formal guidelines for system assessment follow-up.
                                           The IG noted that follow-up reviews were the key to the effectiveness of
                                           system assessments. In response, FFtA issued assessment guidelines,
                                           which state that follow-up activity must begin 8 to 10 weeks after FFtA
                                           receives the railroad’s ‘response.

                                           We analyzed the system assessments conducted since the 1985 guidance
                                           was issued to determine whether follow-up reviews were conducted in
                                           accordance with the guidance. We found that only two of the eight
                                           follow-up reviews met FRA’S criteria. The six follow-up reviews that did
                                           not meet FRA’S criteria experienced delays of at least 6 months.

Table 2.3: Time Between Railroad Written
Response and FRA Follow-Up Review                                                                                  Met FRA follow-
                                           Railroad                                            Elapsed time        up criteria?
                                           1. Burlinaton Northern                                   4 weeks        Yes
                                           2. SEPTA                                                10 weeks        Yes
                                           3. Metro North                                          40 weeks        No
                                           - Norfolk & Western                                     60 weeks        No
                                           5. Conrail                                              76 weeks        No
                                           6. Guilford                                             36 weeks        No
                                           7. New Jersey Transit                                   48 weeks        No
                                           8. So0 Line                                             52 weeks        No
                                           Source: FRA Office of Safety Enforcement

                                           FRA has no assurance that railroads are correcting problems identified in
   -- _ Follow-Up                          its routine inspections because there are no requirements that the rail-
and Reinspection                           roads respond in writing to indicate that an identified defect has been
                                           repaired. Even in the absence of requirements to report corrective
                                           actions, railroads voluntarily respond in writing to most track and signal
                                           defects indicating that corrective actions have been taken. Although
                                           some railroads also report corrective actions for equipment and oper-
                                           ating practice defects, FRA maintains no record of these written
                                           responses. In addition, FRA does not perform many reinspections, an
                                           additional tool for verifying whether a safety defect has been corrected.

                                           Page 23                               GAO/RCED-99-194 FRA Inspections: New Approach Needed

                                            Chapter 2
                                            FRA’a Inspection Program Not Effective

                                            As a result, FRA cannot be certain that the railroads have corrected the
                                            safety defects it has identified.

                                            For track and signal defects, railroads generally respond in writing to
                                            FRAindicating that the identified safety problem has been corrected.
                                            However, given the large number of track and signal defects found
                                            during inspections, thousands of them had no written response to FM. In
                                            1986 through 1988, FRA identified about 361,000 track defects for which
                                            nearly 40,000, or 11 percent, had no recorded railroad response as
                                            shown in table 2.4. Furthermore, of the 88,000 track defects identified
                                            as serious. enough to warrant a $5,000 civil penalty, 11 percent had no
                                            recorded response. Additionally, 46 percent of the signal defects with
                                            associated civil penalties of at least $6,000 did not have a recorded

Table 2.4: Total Track and Slgnal Defects
and Vlolation Reports as Compared to                                                     Track 1986-88
Defect and Vlolation Reports With No                                                               No railroad          Percent non-
Follow-Up, 1986-88                                                                    Total          response              rerpon8e
                                            Total Defects                           360,683             39,964                  11.1
                                            Defects with $5,000 Penalty              88,401               9,534                  10.8
                                            Violation Reports                            .769                128                 16.6
                                                                                          Signal 1986-88
                                                                                                     No rallroad        Percent non-
                                                                                       Total           response            response
                                            Total Defects                             34,813               5,039                 14.5
                                            Defects with $5,000 Penalty                  219                  101                46.1
                                            Violation Reports                            428                  76                 17.8
                                            Source: FRA Railroad Inspection Reporting System

                                            Reinspections are a tool for determining whether a safety problem has
                                            been corrected. In a limited number of cases where FRA inspectors per-
                                            formed a reinspection, we found many instances in which the reinspec-
                                            tion revealed that previously cited defects had not been corrected by the
                                            railroad. Not correcting identified safety defects is one of the factors FRA
                                            inspectors use in deciding whether a defect should be cited as a violation
                                            and a civil penalty assessed. Because of limited reinspections and the
                                            absence of a requirement that railroads respond in writing indicating
                                            defects have been corrected, FM has little assurance that railroads are
                                            actually correcting defects.

                                            In 1988, FRAconducted 57,435 track, equipment, signal, and operating
                                            practices inspections of which 1,527, or 2.7 percent were reinspections.

                                            Page 24                              GAO/RCED-SO-194 FRA Inspections: New Approach Needed
              Chapter 2
              FICA’SInspection Program Not Effective

              Only one of FRA’S five disciplines, signal, has an established standard
              indicating when an inspector should conduct a reinspection. In that dis-
              cipline, however, only 126 of the 6,699 inspections conducted in 1988,
              about 2 percent, were reinspections.

              We found several instances in which an FRA reinspection revealed that
              previously cited serious safety defects had not been corrected by the
              railroad. For example, one FRAtrack inspector found numerous defects
              during an initial inspection of a high density hazardous materials route
              near a populated area. During a subsequent reinspection, the inspector
              found that the railroad had failed to correct several of the defects.
              According to the inspector, any one of these conditions could have led to
              a catastrophic accident. As a result of the reinspection, the inspector
              cited the railroad for violations in each case.

              We found major problems with the way FRA was implementing its rail
Conclusions   safety inspection program. FRA has not established minimum inspection
              coverage standards or standards for the size of an inspector’s territory
              or how much an inspector could be expected to inspect. These standards
              are needed to determine the size of the inspector workforce. FRA also
              relies on state inspectors to provide inspection coverage, but poor com-
              munication and coordination have caused problems. In addition, the con-
              tinuation of the state program is in question due to the elimination of
              federal funding.

              FRA  does not take a systematic approach to both its routine inspections
              and system assessment activities by using available safety inspection
              data. FRA has not fully integrated available inspection, accident, and
              injury data to target high-risk locations for routine inspections in order
              to achieve the maximum effectiveness from its limited inspection
              resources, FRA’S reliance on inspector judgment and knowledge to exe-
              cute its inspection program has resulted in high-risk areas receiving
              decreased inspection activity. In addition, FRA has not used the NIP as a
              management tool to assess the effectiveness of its program, and the NIP
              does not meet a congressional directive that FRA develop a methodology
              to prioritize its inspection activity to high volume passenger and haz-
              ardous materials routes. FRAdoes not use its safety data to efficiently
              use inspection resources during system assessments. FRA has not fol-
              lowed up on its system assessment findings in a timely manner, which
              further reduces their effectiveness.

              Paye 25                           GAO/RCED-99-194 FRA Inspections: New Approach Needed

                      Chapter 2
                      FRA’s Inepectlon Program Not Effective

                      FRAhas no requirements that railroads respond to FRA by sending notifi-
                      cation of actions taken to correct safety defects that were identified
                      during routine safety inspections. Thus, FRA has no assurance that the
                      railroads are correcting unsafe conditions. In addition, a possible tool to
                      determine if railroads are correcting safety defects-reinspections-is
                      seldom used.

                      In order to make FRA’S railroad safety inspection program more effective
Recommendations       in ensuring that the nation’s railroads are operating safely, we recom-
                      mend that the Secretary of Transportation direct the Administrator, FRA
                      to take the following steps:

                  l Establish a minimum inspection coverage standard for each of its
                    inspection disciplines and determine the number of inspectors necessary
                    to achieve this standard of coverage. In determining the number of
                    inspectors needed, FRA needs to determine the projected size of the state
                    inspector workforce and include this in its coverage analysis.
                  l Resolve communication and coordination problems with the states by
                    routinely exchanging inspection plans and periodically meeting with
                    state directors and inspectors to discuss their inspection activities.
                  . Improve its National Inspection Plan by developing a methodology that
                    incorporates past inspection results and prioritizes inspections on the
                    volume of traffic on passenger and hazardous materials routes.
                  l Use its safety data to target high-risk railroads and locations for routine
                  l Redefine the approach to system assessments by using existing inspec-
                    tion data to detect known areas of weakness and assign inspector
                    resources to determine the underlying causes of these weaknesses.
                  . Complete system assessment follow-up reviews within the timeframes
                    established by FR.4 criteria.
                  . Establish an effective follow-up program that would include (1)
                    requiring railroads to report actions taken on FRA inspection findings, (2)
                    determining what reinspection levels are needed to ensure railroads are
                    responding to inspection findings, and (3) attaching civil penalties for
                    failure to report corrective actions.

                      We discussed the facts in this report with the Administrator, FRA, and
Views of hgency       top level safety officials who generally agreed with our findings, espe-
Officials             cially the need to change the inspection program. Specific comments on
                      the topics covered in this chapter follow.

                      Page 26                           GAO/RCED-99-194 FRA Inspections: New Approach Needed

                        Chapter 2
                        FRA’s Inspection Program Not Effective

Need for Inspection     The Administrator stated that, from his perspective, the current inspec-
CoverageStandards       tion approach is too random and that inspection coverage standards are
                        needed. He added that FRA is trying to hire new inspectors to reach their
                        authorized ceiling of 361 inspectors. The Administrator also said that he
                        is in the process of hiring a new Director of Training and Communica-
                        tion to train entry-level employees in a comprehensive 2-week orienta-
                        tion program. The new director will also be responsible for designing a
                        new training program for regional directors and inspectors that would
                        include training on how to analyze and interpret data needed to estab-
                        lish inspections coverage standards.

Targeting Inspections   The Administrator agreed with our finding by saying that FRAneeds to
                        take the randomness out of its inspection strategy. He stated that what
                        FRA needs is a more scientific and strategic approach to inspections
                        based on the information currently available. He added that inspectors
                        need to be computer literate and schooled in analytical techniques so
                        they can recognize trends in safety data and react accordingly.

Inspection Follow-Up    The Administrator stated that we should not be advocating a loo-per-
                        cent reinspection effort because it is not feasible given current staff. FRA
                        safety officials stated that requiring railroads to report their corrective
                        actions on equipment defects and reinspecting them is not feasible
                        because equipment is moved on a day-to-day basis. These officials added
                        that requiring railroads to report their corrective actions may pose an
                        undue paperwork burden on the railroads. FRA'S Office of Chief Counsel
                        officials stated that reinspections are not needed because inspectors
                        monitor defect ratios to determine compliance and that the enforcement
                        program also ensures compliance.

                        We are not advocating loo-percent reinspection, but a process through
                        which railroads submit reports to FRA indicating that defects have been
                        corrected. As it is now, FFU does not know whether railroads are cor-
                        recting the defects found by FFIA inspectors. Furthermore, a sampling
                        procedure could be designed for follow-up to determine whether correc-
                        tive actions have been taken.

                        Page 27                           GAO/RCED-90-194 FRA Inspections: New Approach Needed
Chapter 3

FRA Inspectors Do Not Uniformly Apply
Safety Standards

                           The Federal Railroad Safety Act of 1970, as amended, states that fed-
                           eral rail safety laws and regulations “shall be nationally uniform to the
                           extent practicable.” FRA’Sstated policy is that uniform application of the
                           act is essential for effective program management. We found that FM
                           and state inspectors did not apply safety rules and standards uniformly
                           throughout the railroad industry. Our analysis of inspection activities
                           between 1986 and 1988 showed that uniform application was not
                           achieved because of outdated or minimal written guidance, limited
                           training, and minimal coordination among safety specialists, This has
                           resulted in FRA regions filing different numbers of violations against rail-
                           roads for the same defective safety conditions.

                           Uniformity in enforcement of rail safety regulations does not exist
Regional Differences       between FRA’S eight regional offices. Our review of FRA inspection data
Exist in Enforcement       between 1986 and 1988 revealed numerous examples of one FRAregion
                           filing many more violations than another-often     concerning the same
                           defective condition. For example:

                       . In 1988, one FRA region cited railroads for inadequate track inspection
                         records 312 times but filed no violation reports. Another FRAregion
                         found the same problem 433 times in that year and cited it as a violation
                         in 166 of those cases.
                       . In 1988, one FRA region cited improperly secured center plates-a
                         serious equipment defect with a $6,000 civil penalty-72 times but filed
                         only one violation. Another FRA region found this serious defect 46 times
                         in 1988 but cited it as a violation in 16 cases.
                       . In 1987, one FRA region cited signal defects for “shunting sensitivity”-a
                         serious signal defect with a $6,000 civil penalty-13 times with no vio-
                         lations. Another FRAregion, however, found this same serious defect
                         eight times and cited it as a violation all eight times.

                           Several FRA headquarters officials, regional specialists, and inspectors
                           expressed concern about the level of uniformity between FRAregions. In
                           one case, an equipment specialist we interviewed complained about the
                           number of defective cars and locomotives entering his region from
                           another region. He said that this occurs because his inspectors were
                           much more aggressive in enforcing regulations than the other region’s
                           inspectors, noting that his region wrote many more equipment

                           Page 28                     GAO/RCED-99-194 FRA Inspections: New Approach Needed
                       Chapter 8
                       FRA Inspectors Do Not Unif-ormly Apply
                       Safety Standards

                       FRAdoes not achieve uniform application of rail safety regulations
Causesof Lack of       because agency enforcement manuals are outdated or provide minimal
Uniformity             guidance as to when violations should be written, training is limited, and
                       there is minimal coordination among safety specialists to ensure a
                       common approach among inspectors.

                       Written guidance provided to inspectors is generally outdated and pro-
                       vides vague criteria as to when a violation should be written. Only one
                       enforcement manual has been updated and issued within the last 7
                       years. For example, FRA’S equipment manual dates back to 1975 and the
                       operating practices and hazardous materials manuals date to 1983. The
                       track enforcement manual was updated in 1987 and FRA’S general policy
                       manual has recently been updated but has not yet been issued to inspec-
                       tors. FRA officials agreed that the manuals were outdated and stated that
                       they are in the process of hiring a new Director of Training and Commu-
                       nication who will be responsible for, among other things, keeping the
                       manuals current.

                       In addition, these manuals only outline general factors, such as the seri-
                       ousness of the defect and the compliance history of that railroad, which
                       the inspector should consider in making his decision to issue a violation.
                       Each inspector is expected to use his own judgment in deciding when a
                       violation should be cited using FFU’Ssafety manuals and his experience
                       as guides. As a result, one inspector may cite a serious defective condi-
                       tion but not file a violation report while another may find the same
                       serious condition and file a violation report.

                       We found in each inspection discipline that FRA has inspectors who did
                       not file a single violation against any railroad over a 3-year period. In
                       addition, we found that numerous inspectors filed one to three violation
                       reports-one per year or less-during the same 3-year period. We also
                       found a number of inspectors who filed many violation reports while
                       conducting approximately the same number of inspections as those
                       inspectors who filed almost no violation reports. For example, of the 105
                       track inspectors who performed at least 300 inspections during 1986-88,

                   l   16 filed no violation reports,
                   l   39 filed one to three violation reports, and
                   l   9 filed more than 20 violation reports.

                       We found similar disparities with equipment inspectors. For example,
                       one FR4 equipment inspector filed violation reports in 37 percent of his
                       413 inspections during 1986-1988, while another FRA equipment

                       Page 29                          GAO/RCED-90.194 FRA Inspections: New Approach Needed
                                    Chapter 3
                                    PRA Inspectors Do Not Uniformly Apply
                                    Safety Standar&

                                    inspector filed no violations as a result of his 465 inspections, Table 3.1
                                    provides a breakdown of violation reports written by FRA and state
                                    inspectors by discipline for inspectors who conducted at least 300
                                    inspections during 1986-88.

Table 3.1: Enforcement Actions by
Mpector Discipline, 1986438                                                                Number of Inspectors
                                    Violation reports                       Track      Equipment Signal           practices    Total
                                    0                                           16              4     1                    5     26
                                    1-3                                        39               7     4                   5      55
                                    4-10                                        29                7      11              IO      57
                                    11-20                                       12               13       8               6      39
                                    21-30                                        6               17       4              IO      37
                                    31-40                                         1              14       0               6      21
                                    41-50                                        2                9       1               2      14
                                    Over50                                       0               46       0               2      40
                                    Total                                      105             117       29              46     297

                                    Source:   FRA   Railroad Inspection Reporting System

                                    This lack of uniformity also exists between FRA inspectors in the same
                                    region and discipline. For example, between 1984 and October 1988, an
                                    FRA equipment inspector in one region made 759 inspections involving
                                    1,738 locomotives and found 109 violations. Another equipment
                                    inspector in the same region made 778 inspections involving 1,750 loco-
                                    motives and filed one violation.

                                    FRA depends on inspectors to make decisions as to when a violation
                                    should be cited but provides them with limited training in the enforce-
                                    ment of federal safety regulations. For example, we found that of the
                                    approximately 300 FFtA regional personnel, 147 had received one or no
                                    FRA training classes between 1985 and September 1989. In fact, 50 track
                                    inspectors hired by FRA prior to 1986 had received no FRA training
                                    between 1986 and September 1989. According to specialists and inspec-
                                    tors we interviewed, the level of their training was inadequate to ensure
                                    uniform application of the rail safety regulations. FRA headquarters offi-
                                    cials concurred that such training had been inadequate. In addition,
                                    because of recent inspector turnover, FRA has a large number of new
                                    inspectors who have little experience in enforcing federal safety regula-
                                    tions, which could exacerbate the problems we identified.

                                    Page 30                                 GAO/RCED-90-194 FRA Inspections: New Approach Needed
                      Chapter 3
                      FRA Inspectors Do Not Uniformly Apply
                      Safety Standards

                      Coordination between FRAsafety specialists to ensure the uniform appli-
                      cation of safety standards is also lacking. According to regional direc-
                      tors and specialists we interviewed, FRA should periodically hold
                      meetings with all the safety specialists in each discipline to ensure a uni-
                      form enforcement approach. One track specialist stated, for example,
                      that all eight track specialists had only met once in the last 6 years.
                      Instead of such specialist meetings, FR4 currently holds safety confer-
                      ences involving field staff from one or two regions. According to FRA
                      regional officials and inspectors, these conferences sometimes provide
                      verbal guidance on FRAenforcement. However, FRAhas no assurance
                      that such verbal guidance guarantees national uniformity since it is pro-
                      vided to only a few regions at a time and is not in writing.

                      FRA inspectors have not uniformly applied safety regulations throughout
Conclusions           the railroad industry. Inspectors in some FRA regions often cited serious
                      safety problems as violations while inspectors in other regions rarely
                      cited violations for the exact same safety problem. We believe that FRA
                      has not provided adequate guidance or training to its inspectors to
                      ensure the uniform application of the regulations as required by law.

                      In order for FRA to assure better uniformity of inspections, we recom-
Recommendations       mend that the Secretary of Transportation direct the Administrator,
                      FRA, to increase training, especially for new inspectors, and to issue
                      formal guidance to inspectors reemphasizing the need for uniformity in
                      citing violations.

                      The Administrator, FRA, agreed that inspectors are not uniformly
Views of Agency       applying the safety regulations and that there is a wide variance. He
Officials             recognized that inspectors will vary because each may interpret the reg-
                      ulations differently, but his goal is to lessen the range of variance. To do
                      this he is considering the following actions:

                  . Common training for all regional directors and deputies in Washington
                    on how to apply the regulations.
                  . Common training for all specialists in each discipline.
                  l More regional conferences to disseminate information.
                  l Updated manuals by the new Director of Training and Communication.
                  . Frequent meetings between regional directors and the railroads in their
                    region to discuss safety problems and enforcement actions.

                      Page 31                          GAO/RCED-90-194 FRA Inspections: New Approach Needed
Appendix I

Reliability Assessmentof F’RAInspection Data *

                     Many of the analyses contained in this report rely on data contained in
Overall Assessment   FFU’Sautomated Railroad Inspection Reporting System (RIRS). Because
                     we relied on this data, we conducted a limited reliability assessment of
                     the RIRS as required by Government Auditing Standards. We found the
                     accuracy of the data to be sufficiently high to be usable in performing
                     the analyses contained in this report- analyses not normally performed
                     by FRA in managing its inspection program.

                     Although the data are not independently verified, we observed a variety
                     of controls for detecting and correcting errors throughout the data entry
                     process. The data entry contractor uses edit, logic check, quality control,
                     review, and correction controls to ensure correct data. However, because
                     100 percent of the data are not independently verified by different data
                     entry clerks, some portion of typing errors will not be detected and cor-
                     rected by the controls in place.

                     To test the reliability of the specific data used in this report, we selected
                     a sample of data elements, comparing data in the automated system
                     with original paper inspection reports. Based on this sample, and the
                     statistical analyses performed, we concluded that the overall accuracy
                     of the data elements in the fields we reviewed is quite high.

                     FRAemploys a contractor I-NET-located      in Rockville, Maryland, for
System Overview      entering data into the Railroad Inspection Reporting System. I-NET per-
                     forms data entry on a multi-user system utilizing microcomputers.

                     I-NET inputs data to disk from batched      inspection forms. On a weekly
                     basis, data which have passed through      I-NET’s internal controls are
                     written to tape and sent to the National     Institutes of Health’s computer
                     system for inclusion in FRA’S inspection   database.

                     I-NET staff count and sort inspection report forms by form type, region,
Data Receipt and     payroll ID, and report number, and record them in a daily receipts log.
Input                They visually scan documents for completeness, illegible information,
                     and duplicate report numbers. If the form is incomplete or in error, I-
                     NET staff either return the form to the inspector or contact the
                     inspector by telephone for correction. After batching the forms by type
                     of inspection, the batches are entered into the computer by data entry

                     Page 32                      GAO/RCED-99-194 FRA Inspections: New Approach Needed
                                     Appendix I
                                     Reliabtltty Assessment of FRA
                                     Inspection Data

                                     I-NET uses various types of automated edit routines to flag incorrect
Quality Control Edits                data. These edits fall into two basic types: real-time edits during data
                                     entry, and batch edits done some time after the data are entered. Real-
                                     time edits include field edits that allow only valid codes or values to be
                                     entered into a field. Batch edits include special programs that are run to
                                     test data that have been entered into the computer. We observed 23
                                     weekly batch programs for detecting discrepancies, missing data, or
                                     duplicative data.

                                     Because FRA does not retain original copies of inspection reports for
Results of Reliability               more than one year, our review was limited to inspection reports for the
Test                                 January 1989 to August 1989 period. We selected a cluster sample of 75
                                     inspections from each of four files in FRA'S inspection database: signal,
                                     track, equipment, and operating practices. For each inspection selected,
                                     we traced the automated database information to the corresponding
                                     hardcopy source report to determine whether the inspection information
                                     had been accurately recorded. Our review included data from the header
                                     information for each record, as well as the detailed information on
                                     safety defects listed on the report, A single header record could be asso-
                                     ciated with multiple detail records identifying the defects cited on the
                                     inspection report. When an inspection had more than one detail record,
                                     we verified the information for all the detail records. The number of
                                     fields on the header and detail records we reviewed are shown in table

Table 1.1: Number of Record Header
Fields and Detail Flelds Reviewed                                              Number of Fields Reviewed on
                                     File                             Header Record                     Detail Records
                                     Signal                                       13                                    3
                                     _____---                                     13                                    3
                                     Eaubment                                      8                                    4
                                     Operating practices                          23                                    3

                                     Based on the number of errors found on the header and detail records,
                                     we determined an upper bound, lower bound, and best estimate of the
                                     accuracy of the data in the Railroad Inspection Reporting System.
                                     Results of this analysis, shown in table 1.2, indicated a high level of data

                                     Page 33                         GAO/RCED-90.194 FRA Inspections: New Approach Needed
                                       Appendix I
                                       Reliability Assessment of FRA
                                       Inspection Data

Table 1.2: Estimated Error Bounds by
RIRS Data File                         Figures in percent
                                                                                   Estimated Data elements with wrong information
                                                                                   Elements     Errors        Best Lower Upper
                                                                                    reviewed    found    estimate bound bound
                                       Signal                                             1,518            0            0.0       0.0        0.2
                                       Track                                              3,654            2            0.1     <O.l         0.2
                                       Equipment                                          3,840            5            0.1     <O.l         0.3
                                       Operating practices                                2,028            2            0.1     <O.l         0.4

                                       Note: Estimated lower and upper bounds are at the 95.percent confidence interval. That is, if we had
                                       checked all the data elements in the selected fields and calculated the percent of incorrect data ele-
                                       ments, 95 times out of 100 this percent would lie between the upper and lower bounds given. We used
                                       a cluster sampling design in which we randomly selected an inspection report, then verified all the
                                       selected data elements in the header and detail records for that report. However, results from formulas
                                       normally used with cluster sampling gave unrealistic results (lower bounds less than zero, for example).
                                       Therefore we based our estimates on the hypergeometric distribution, which may somewhat misstate
                                       the upper bound. Estimates based on the hypergeometric distribution assume that the data elements
                                       selected for review represent a simple random sample.

                                       Page 34                                GAO/RCED-90-194 FRA Inspections: New Approach Needed
States Participating in l?RA Inspection Program

                Alabama                                Nevada
                Arizona                                New Hampshire
                Arkansas                               New Jersey
                California                             New York
                Connecticut                            North Carolina
                Florida                                Ohio
                Illinois                               Oklahoma
                Iowa                                   Oreaon
                Kansas                                 Pennsylvania
                Louisiana                              South Carolina
                Maine                                  Tennessee
                Maryland                               Texas
                Michigan                               Utah
                Minnesota                              Virainia
                Missouri                               Washinnton
                Montana                                West Virginia


                Page 36       GAO/RCED-90-194 FRA Inspections: New Approach Needed


Appendix III

Major Contributors to This Report

                        Roy J. Kirk, Assistant Director
Resources,              John S. Kalmar, Jr., Assignment Manager
Community, and          Sara Moessbauer, Operations Research Analyst
Development Division,
Washington, D.C.

                        Gregory G. Booth, Regional Management Representative
Chicago Regional        Robert M. Ciszewski, Evaluator-in-charge
Office                  Timothy F. Hannegan, Site Senior
                        Rosa Maria Torres-Lerma, Evaluator
                        Patricia M. Barry, Evaluator

(a4sela)                Page 36                  GAO/RCED-WI-194 FRA Inspections: New Approach Needed
‘I’htw is a ZYo tliscouut~ ou orders for 100 or more copiw     mailed t.0 it
sirigle addrrss.

Ortlt~rs must. tw pwpaid by cash      or by check or mout~y order   made

out. Lo t.lw Su~~t~rintentlelIt of Documents.
-.__ ..-.._ l,-.II-------