oversight

Airline Scheduling: Airlines' On-Time Performance

Published by the Government Accountability Office on 1990-06-15.

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

                   United   States   General   Accounting   Office
                                                                     .
                   Report to Congressional Requesters
dA0

June   1990
                   AIRLINE
                   SCHEDULING
                   Airlines’ On-Time
                   Performance




 GAO/RCED-90-154
                    United States
GAO                 General Accounting Office
                    Washington, D.C. 20548

                    Resources, Community, and
                    Economic Development Division

                    B-1971 19

                    June 151990

                    The Honorable Glenn M. Anderson
                    Chairman, Subcommittee on Investigations
                      and Oversight
                    Committee on Public Works and
                      Transportation
                    House of Representatives

                    The Honorable William F. Clinger, Jr.
                    Ranking Minority Member, Subcommittee
                      on Investigations and Oversight
                    Committee on Public Works and
                      Transportation
                    House of Representatives

                    As requested in your letter of March 2,1989, and as agreed in subse-
                    quent discussions with your offices, we assessed whether (1) airlines
                    had increased scheduled flight time to keep reported delays at a min-
                    imum in response to the Department of Transportation’s (nor) on-time
                    performance reporting requirement and (2) nor verifies that flights
                    omitted from the on-time data because of mechanical problems have in
                    fact experienced mechanical problems. To address these issues, we
                    interviewed nor, Federal Aviation Administration (FAA), and airline
                    industry officials, and analyzed statistical data maintained by nor on
                    airlines’ on-time performance.


                     Airlines have adjusted the amount of time they schedule for flights to
Results in Brief     develop schedules that are as short as possible yet still allow flights to
                     be on t&e. However, the airlines’ on-time performance, on average, has
                     not improved since the reporting requirement began in September 1987.
                     In addition, while airlines are permitted to exclude flights with mechan-
                     ical problems from the on-time data, nor does not verify that these
                     flights experienced mechanical problems. Specifically, we found the
                     folIowing:

                   . Airlines adjusted scheduled flight time and made other changes to
                     streamline flight operations and improve on-time performance. A flight
                     is considered “on time” if it departs from or arrives at the gate within
                     15 minutes of its schedule. However, on the basis of the l&minute crite-
                     rion, there is no clear trend toward improved on-time performance for



                     Page 1                              GAO/RCEXM@lM   Airlines’ On-Time Performance
               Is197119




               the reporting airlines. Since the reporting requirement began in Sep-
               tember 1987, the airlines’ average monthly on-time performance has
               fluctuated, ranging from a low of about 66 percent to a high of about 86
               percent. Generally, the percentage of on-time flights declines during the
               winter months when poor weather interferes with scheduled departures
               and arrivals. Airline industry and government officials agree that poor
               weather is the principal reason for late flights. In 1988, about 80 percent
               of flights arrived on time. In 1989, slightly more than 76 percent of the
               flights arrived on time.
             9 DOTmonitors the number of flights excluded from the on-time perform-
               ance data for mechanical problems, but does not verify that these flights
               had mechanical problems. Flights delayed or canceled because of
               mechanical problems are not required to be reported to nur in the on-
               time data, but the mechanical problems must be reported to FAA. How-
               ever, FAA does not record the mechanical problem data in a way that
               allows nor to crosscheck the information. According to nor’s on-time
               statistics, about 5 percent of all scheduled flights, or an average of over
               23,000 flights per month, are excluded from the on-time data because of
               mechanical problems.


               nor’s investigation of airline scheduling practices during 1986 and 1987
Background     resulted in the on-time reporting requirement. At that time, bcrr found
               that airlines often scheduled unrealistic flying times because the com-
               puterized reservation systems used by travel agents gave priority list-
               ings to flights with the shortest elapsed time. Travel agents usually book
               passengers on one of the flights that appear on the first few display
               screens of a reservation system. Flights with longer scheduled elapsed
               times would appear on subsequent screens and agents would be less
               likely to book these flights. nor officials concluded that unrealistic
               scheduling was an unfair and deceptive trade practice. In August 1987,
               they obtained a commitment from the computerized reservation system
               vendors to stop listing flights in the order of scheduled elapsed times.
               This eliminated an incentive for airlines to underestimate flight times.

               In September 1987, nor began recording the on-time performance of U.S.
               airlines to provide consumers with information on airline flight timeli-
               ness. DOTrequires the largest U.S. airlines-those having more than 1
               percent of total domestic scheduled-service passenger revenues-to
               report departure and arrival data for each non-stop flight on a monthly




               P8ge 2                              GAO/%CEB9@154 Alrllnea’ On-Time Perfo-ce
                        B-197119




                        basis.’ Each flight is identified by airline, flight number, and route (line
                        of travel between the flight’s origin and destination). For example,
                        Trans World Airlines’ flight 0012 between Boston and Los Angeles, via
                        St. Louis, is counted as 2 separate flight operations, one from Boston to
                        St. Louis and another from St. Louis to Los Angeles (TW/0012/BOSSTL
                        and TW/OO12/STLLAX). DOT’SOffices of Economics and Aviation Infor-
                        mation Management jointly receive, compile, and monitor the airlines’
                        flight delay data. D&S Office of Consumer Affairs issues a monthly
                        report, the “Air Travel Consumer Report,” showing each airline’s
                        performance.

                        In reporting to D(JT,airlines are allowed to exclude flights experiencing
                        mechanical problems from the on-time data to ensure that airlines do
                        not operate unsafe planes for the sake of maintaining a good on-time
                        performance record. Flights canceled for reasons other than mechanical
                        problems, such as poor weather conditions or lack of flight crew, are
                        included and counted as late in calculating on-time statistics. Even
                        though DOTranks the airlines according to their on-time performance. it
                        has not established a required level of performance.


                        To improve on-time performance, airlines adjusted scheduled time and
Airlines Adjusted       streamlined flight operations, according to MJTand airline officials.
Flight Schedulesand     However, the data do not show a trend toward better on-time perform-
                        ante, based on the 15-minute criterion. Airline officials claim that
Operations to Improve   weather and air traffic control system problems are the main reasons
On-Time Performance     why flights are not always on time. Each airline’s monthly on-time per-
                        formance, based on the 15minute criterion, is provided in appendix I.


On-Time Performance     The on-time statistics show that airlines’ performance has fluctuated
                        from month to month. On-time performance was lowest in December
Fluctuates              1987, when 66 percent of flights were on time. The highest performance
                        was achieved in September 1988, when about 86 percent of the flights
                        arrived on time. As shown in figure 1, flights have arrived on time over
                        70 percent of the time since February 1988, but on-time performance
                        generally was lower in 1989 than in 1988.


                        ‘These airlimes, currently 12, account for about 90 percent of domestic operating revenues .&
                        required by 14 C.F.R. part 234, the airlines must report non-stop flights operating to and from I hv
                        largest U.S. airports-those with at least 1 percent of total domestic scheduled-serwce pawwwr
                        enplanements (currently 29). However, all 12 airlines have voluntarily provided data for thrlr mtm
                        domestic systems.



                        Page 3                                           GAO/RCEB9@154 Airlines On-Time Prrfornwwe
                                            B197119




                                                                                                                 --
Figure 1: Airlines’   On-Time Performance   -                .     _,,,.                 .-
                                            100     Pofeml




                                                     -           1987
                                                     111.        198s
                                                     -           1989

                                            Source. DOT’s Air Travel Consumer Reports.


                                            On-time performance tends to decline during the winter months when
                                            poor weather interferes with scheduled departures and arrivals. For
                                            example, between November 1987 and March 1988,66 to 79 percent of
                                            the flights arrived on time, while from April 1988 through October 1988,
                                            82 to 86 percent arrived on time. On-time performance then generally
                                            declined during the winter months, November 1988 through March
                                            1989, when 72 to 78 percent of the flights arrived on time. In the subse-
                                            quent months, through October 1989, on-time performance rates were
                                            generally higher but did not improve to the levels of the preceding April-
                                            October period.

                                            Part of the explanation for the poorer performance can be attributed to
                                            factors beyond the airlines’ control. For example, in June 1989, unusu-
                                            ally heavy rains on the East Coast caused delays at airports in the East
                                            that affected traffic throughout the nation. Flights arrived on time only
                                            72 percent of the time during June 1989, compared with 84 percent in
                                            June 1988. Also, pilots for one of the nation’s largest airlines, Cmted
                                            Airlines, decided to slow down flights during the summer of 1989 as



                                                Page 4                                        GAO/ltCEDSElM   Airlines’ On-Tlmr Prrhmance
                         8197119




                         part of a labor dispute with the airline’s management. United’s on-time
                         performance for the 1989 summer months (June-August) averaged just
                         over 60 percent, compared with about 81 percent for the summer of
                         1988. According to DCJT,  the United pilots’ action caused the flights of
                         other airlines to be delayed as well.

                         Industry and government officials agree that poor weather is the prin-
                         cipal reason for late flights. Airline officials also claim that airport con-
                         gestion is a major cause of flight delays, but WT and FAA officials, while
                         acknowledging that congestion contributes to some delays, characterized
                         it as a secondary factor.


Airlines Have Adjusted   When on-time performance reporting began, some airlines simply added
                         time to their flight schedules to improve on-time performance. However,
Scheduled Flight Times   increasing the scheduled flight times also increased the airlines’ oper-
                         ating costs, since flight crews are paid according to scheduled, not
                         actual, flight time. Also, by increasing the scheduled tihe for each flight,
                         fewer flights could be scheduled for an aircraft. Therefore, airlines
                         adjusted schedules to be as short as possible yet still allow the flights to
                         arrive on time, according to airline officials. Rather than simply adding
                         more time, some airlines tried to improve on-time performance by
                         shifting frequently late flights to less-congested time slots, streamlining
                         baggage handling, and reducing the time required to fuel and prepare
                         their aircraft.

                         The airline industry has not set an on-time performance standard. but
                         airline officials told us that public awareness of on-time performance
                         provides incentive to do as well as possible. Airlines with relatively good
                         on-time performance records often publicize their standing in their
                         advertisements. In deciding how much time to allow for a flight. the air-
                         lines use past experience and historical trends and schedule different
                         amounts of time for flights covering the same route depending on the
                         time of year and on expected airport congestion. However, m and F.u
                         officials noted that airlines schedule their flights assuming the most
                         favorable flying conditions.

                         DUFSon-time statistics indicate whether flights are completed within the
                         15-minute criterion, but do not show how late the flight was. Consumers
                         cannot distinguish between flights that are moderately late and those
                         that are severely late. Also, m does not assess the data to determme
                         whether the amount of scheduled time has increased over time in an
                         attempt to improve on-time performance. We found that for a sampltl of


                          Page 5                              GAO/RCEDS@l54 Airlines’ On-Time Performance
                        5197119




                        over 200 non-stop flights during October 1987, 1988, and 1989 airlines
                        added and reduced scheduled time for the same flights. However, we
                        found no relationship between changes in the amount of time allowed to
                        complete a flight and on-time performance, and longer flight times did
                        not always result in better on-time performance.

                        nor’s Assistant Secretary for Policy and International Affairs recently
                        reviewed the on-time performance data and recommended that more use
                        be made of the performance data. For example, the information could be
                        used to identify unrealistic peak-period flight schedules, congestion pat-
                        terns at hub airports, and flights that are repeatedly late or canceled. On
                        February 28, 1990, the Secretary of Transportation adopted this recom-
                        mendation and directed DOT’SOffices of Economics and Aviation Infor-
                        mation Management to develop computer programs and techniques to
                        further analyze the on-time data. As part of this analysis, DOTplans to
                        assess how late flights are and whether scheduled flight times have
                        increased since the on-time reporting began.


                               Offices of Economics and Aviation Information Management mon-
DOI’ Monitors On-Time   itor the on-time performance data the airlines submit each month, but
Data but Does Not       do not routinely verify that flights excluded from the on-time data
Verify Mechanical       because of mechanical problems were reported to FAA or that these
                        problems actually occurred.g D&S Office of Consumer Affairs and Office
Problems                of Inspector General have reviewed some flight records to verify that
                        mechanical problems occurred and were documented as required.
                        Results from these surveys show that some flights were not accurately
                        reported and documentation was not always sufficient to support the
                        claim that mechanical problems caused the delay or cancellation. In
                        March 1990, nur and Fu began to develop procedures that should allow
                        DOTto verify that flights excluded from the on-time data for mechanical
                        reasons were reported to FM.


DOT Monitors On-Time    ncrr checks each airline’s monthly an-time report for such things as com-
                        pleteness (scheduled flights are reported as required) and rejects a sub-
Data for Errors and     mission if the error rate is greater than 0.01 percent. If its data are
Omissions               rejected, the airline must correct the data and re-submit its report. After
                        accepting the data, DOTlooks for anomalies, such as sudden changes in

                        hairlines are required to report mechanical problems to FAA by filing either a Mecharwal Kellabllity
                        Report or a Mechanical Interruption Summary as outlined by 14 C.F.R. 121.703 and 14 C F H
                        121.705.



                        Page 6                                          GAO/RC3ED~lM         AJrlhed On-Ttmr Prrfo-ce
5197119




the distribution of actual arrival times. For example, if an airline sud-
denly began reporting a large number of flights arriving 13 to 1-I min-
utes after the scheduled arrival times (on time according to DOT’S1.5
minute criterion), DCTIwould investigate to determine whether the airline
adjusted actual arrival and departure times to show better performance
than was actually achieved. According to MJTofficials, to date, the on-
time data have not indicated sudden changes in near on-time perform-
ance that would require such an investigation.

DOTestimates the number of flights excluded for mechanical problems
by comparing the flights completed (as reported in the on-time data) to
the flights scheduled to operate (as listed in the Official Airline Guide).
Flights scheduled to operate but not reported in the on-time data are
assumed to have experienced mechanical problems, according to the nor
officials. On the basis of this assumption, flights excluded from the on-
time data because of mechanical problems have averaged over 23,000
flights per month, or about 5 percent of the flights scheduled to operate
each month. :

While DOTdoes not routinely verify flights affected by mechanical
problems, it has investigated this issue. Beginning in 1988, DCT’SOffice
of Consumer Affairs surveyed flights omitted from the on-time data. DCT
officials selected samples of 5 or 6 flights from each of the 12 reporting
airlines and traced them back to the airlines’ records. They found that
about half of the flights had been either incorrectly excluded from the
on-time data or were not sufficiently documented to support the claim
that a mechanical problem caused the delay or cancellation.

In 1989, the DCTT Inspector General (IG) also audited the on-time perform-
ance data. The audit objectives were to determine (1) whether data sub-
mitted by the airlines are accurate, (2) whether flights not reported due
to mechanical problems are appropriately documented, and (3 ) whet her
nur’s on-time reporting system provides accurate information. To verify
the data, the IG traced a sample of flights for each of the 12 reporting
airlines back to the pilots’ logs.’ To assess the accuracy of nur’s
reporting system, the IG reviewed DOT’Scomputer programs and edit pro-
cedures and processed sample data to test the system.

 ‘Overall, the percentage of flights omitted each month from the on-time data has flUmdtt*i   -bung
from about 4 percent to over 7 percent. Individual airline rates of mechanical exclusions -h w *lrnllar
month-to-month variations.
‘The IG reviewed March 1989 flight data for all airlines except Eastern Air Lines. Auguu?rt1~9 &ta
were used for Eastern, since its flights were reduced significantly in March and the ~mrnr*l~rr~~
                                                                                                WOW-
quent months because of the strike by the Eastern machinists and pilots.



Page 7                                           GAO/WED-9@164 Airlines’ 011-b            Prrforrmnce
                       El97119




                       The IG’s draft report states that no-r’s reporting system provides accu-
                       rate on-time performance information for the flight data the airlines
                       report to DOT.Regarding the accuracy of the airlines’ data submissions,
                       the IG reported that the airlines accurately reported arrival and depar-
                       ture times. However, the IG found that of 3,903 flights excluded by the
                       airlines from the on-time data because of mechanical problems, .564
                       should have been reported and included in the on-time data.’ The IGesti-
                       mated that if these flights had been included, 11 airlines’ on-time per-
                       formance would have been slightly less-by 0.1 to 1.9 percent-than
                       reported.


DOT Has Not Verified   Flights delayed or canceled for mechanical problems average about
                       23,000 each month, or about 5 percent of all flights scheduled to
Mechanical Problems    operate. Airlines are not required to report these flights in the on-time
                       data submitted to m, but they must report the mechanical problems to
                       FAA. However, MJTcannot readily verify that the flights were reported to
                       FAA because FAA does not record the data on mechanical problems in a
                       way that allows D(JTto cross-check the information. Specifically, F-AA
                       does not require airlines to include flight numbers when they report
                       mechanical problems. DUI’needs these numbers to verify that the flights
                       were scheduled to operate but were not reported in the on-time data.
                       Furthermore, FAA does not enter the reports on mechanical problems
                       into a consolidated computer data base that would streamline the verifi-
                       cation process. Because of the volume of flights affected by mechanical
                       problems, manual cross-checking would require significant amounts of
                       time and resources.

                       In January 1990, D&S Assistant Secretary for Policy and International
                       Affairs recommended to the Secretary of DCJT that FAA’s reporting system
                       be revised so that flights excluded from D&S on-time data becamseof
                       mechanical problems could be verified. According to DOT,FM began to
                       computerize the maintenance records on mechanical problems more than
                       2 years ago but stopped because of budget constraints. The Secretary of
                       Transportation recently adopted this recommendation, and FAA and DOT
                       have started to develop procedures for recording and consolidatmy the
                       data on mechanical problems in a way that would allow DOTto \.et-lfy the
                       exclusions. According to DCX,it may require a year to develop and imple-
                       ment the revised reporting procedures.

                       ‘The IG’s audit methodology used simple random sampling       to &ect a sample of each ANIUW- Whhts
                       and 95-percent confidence levels for statistical projections. See appendix III for additlonni ,I+-.I.A on
                       the methodology.



                       Page 8                                            GAO/WED-90-154 Airlines’ On-Tlmr Pwfwmance
                            5197119




On-Time Data Include Non-   Flights canceled for non-mechanical reasons are reported to DOTas late
                            flights in the on-time data. These non-mechanical cancellations average
Mechanical Flight           about 4,700 flights, or just over 1 percent of flights each month. Reasons
Cancellations               for non-mechanical cancellations are not identified but include adverse
                            weather, lack of crew, and other factors, such as economic cancellations
                            or unexpected runway closures.‘, Information on non-mechanical cancel-
                            lations for each month is provided in appendix II.

                            According to DOTand airline officials, most of the non-mechanical can-
                            cellations appear to be due to weather conditions since they follow a
                            seasonal pattern. Also, flight cancellations tend to occur in clusters at
                            specific airports on specific dates, which generally correspond to the
                            storm patterns. However, some non-mechanical cancellations may be
                            due to an airline’s attempt to use aircraft more efficiently when mechan-
                            ical problems occur.

                            As a result of inquiries from consumers, during 1987, officials from
                            bur’s Office of Consumer Affairs conducted several on-site investiga-
                            tions to assess whether non-mechanical flight cancellations occurred for
                            economic reasons. The DUI’investigation found that, in some cases, air-
                            lines reassigned planes when a mechanical problem occurred so that the
                            fewest number of passengers were inconvenienced. While economic and
                            efficiency factors are involved in such a decision, uur does not consider
                            these to be economic cancellations. In these cases, flights with relatively
                            few passengers might be canceled so that the aircraft can be used for
                            flights with relatively more passengers whose planes might have experi-
                            enced a mechanical problem. Many factors are involved, but “inconve-
                            niencing the fewest passengers possible” was reported to be the airlines’
                            primary consideration in making the decision. uur concluded that there
                            was insufficient evidence to demonstrate that airlines had been engaged
                            in a deceptive practice of canceling flights for economic reasons.


                            The purpose of the on-time reporting requirement is to provide con-
Conclusions                 sumers with information on airlines’ timeliness, but the on-time per-
                            formance statistics do not indicate how late flights are or provide
                            information on why some flights are canceled. While LXX’Sreporting
                            system measures whether flights depart and arrive within 15 minutes of
                            their scheduled times, it does not indicate whether actual flight times


                            “m defmes economic cancellations as the practice of canceling a flight to save fuel and I WH I I MS
                            when there are only a few passengen booked on the flight.



                             Page 9                                         GAO/RCELb9@164 Airlines’ On-Time Prrformnnce
                    El97119




                    are closer to or further from scheduled times since the on-time perform-
                    ance reporting began. Furthermore, the on-time statistics do not provide
                    information on changes that airlines may make to scheduled flight times
                    to facilitate on-time performance. In addition, while non-mechanical
                    flight cancellations are included in the on-time data, airlines do not
                    report the reasons for or extent of such cancellations. In its effort to
                    increase the use of the on-time data, bcrr plans to assess how late flights
                    are and whether scheduled flight times have increased since the on-time
                    performance reporting requirement began. Including the reasons for
                    non-mechanical flight cancellations could also provide additional con-
                    sumer information on airlines’ performance.

                    uur has not been able to readily verify that flights excluded for mechan-
                    ical problems actually experienced these problems or were reported to
                    FAA. Because of the volume of mechanical problem data and the fact that
                    the data are not recorded in a way that facilitates verification, ucrr has
                    not routinely verified this information. The Secretary has directed the
                    Administrator of FAA to implement procedures that will record and con-
                    solidate the mechanical problem data in a way that will allow uor to
                    verify that flights excluded from the on-time data were reported to FAA.
                    These changes should provide the information DOTneeds to monitor the
                    cause of delays.


                    We recommend that the Secretary of Transportation, as part of the
Recommendation to   effort to increase the usefulness of the on-time data, assess the feasi-
the Secretary of    bility of requiring airlines to report the reasons for non-mechanical can-
Transportation      cellations. Airlines currently report flights canceled for non-mechanical
                    reasons, but would need to include the reasons for such cancellations in
                    their monthly on-time data so that DUFcould assess the type and fre-
                    quency of each reason.



                    officials agreed with our overall conclusions and reiterated that plans
                    are underway to improve and further use the on-time data. They also
                    acknowledged that including the reasons for non-mechanical flight can-
                    cellations would provide additional airline performance information.
                    However, they noted that it would be difficult for uur to verify these
                    reasons, given the time and resources required to do so. To minimize the
                    time and resources needed to verify this information, we suggest that
                    DOTconduct periodic spot checks to verify that the reasons provided by
                    the airlines support the non-mechanical cancellations. If such spot


                    Page10                              GAO/BCEDgO-1M Airlines’ On-Time Performance
El97119




checks indicate repeated reporting errors, a more extensive verification
process would be warranted.


In conducting our review, we interviewed ~xrr, FAA, and airline officials
and obtained the most recent on-time performance data. We performed
our work in accordance witl&enerally accepted government auditing
standards under the direction of Kenneth M. Mead, Director of Trans-
portation Issues, who can be reached at (202)276-1000. Appendix III
discusses in more detail the scope and methodology we used in compiling
this report, and appendix IV lists other major contributors to the report.

We are sending copies of this report to the Secretary of Transportation,
the Administrator of the Federal Aviation Administration, and other
interested parties.




J. Dexter Peach
Assistant Comptroller    General




 Page 11                            GAO/ECED9&164   Ahlined On-Time Perfotmance
Contents


Letter                                                                                            1
Appendix I                                                                                       14
Percentageof Flights
That Met the On-Time
Performance Criterion
Appendix II                                                                                      16
Flights Canceled for
Non-Mechanical
Reasons
Appendix III
Objectives, Scope,and
Methodology
Appendix IV
Major Contributors to
This Report
Figure                  Figure 1: Airlines’ On-Time Performance                                   4




                        Abbreviations

                        DOT       Department of Transportation
                        FAA       Federal Aviation Administration
                        GAO       General Accounting Office
                        IG        Office of Inspector General
                        RCED      Resources, Community, and Economic Development DiveIon


                        Page 12                           GAO/‘ECEDWlM   AMned   On-Time Perfwmance
Page 13   GAO/RCEDfMJlW Airlines’ On-Time Prrformnnce
Appendix I

Percentageof Flights That Met the On-Time                                    -
Performance Criterion

                            .,              -
                                 Alaska     America
               Month/year             Air      West    American    Continental
               Sept 1987            79.0        73 4        84 5            81 1
               Ott 1987             75 2        74 9        86 1            84 4
               Nov.1987             75.4        77 1        83 2            74 5
               Dec. 1987            59 2        76.4        73 1            60 5
               Jan 1988             76.4        03 9        75 7             648
               Feb 1988             76 9        88.7        80 8       ~__ 67 7
               Mar 1988             756         89 9        83 8              71 6
               Apr.1988             77 7        90.8        850               81 5
               May 1988             81.3        90.9        87 6
                                                             ~---.___.        80 1
               June 1988            82.9        92.8        882               869
               July 1988            86.0        93.9        859               830
               Aug. 1988            03.7        91 5        859               868
               Sept.1988            86.9        92.9        078            ___86 3
               odt. 1988            77 9        89.1        83 8              81 9
               Nov 1988             74.4        79.2        78 2              77 1
               Dec.1988             68.2        76.4        75 6              75 7
               Jan.1989             76.0        81.5        77 4              786
               Feb.1989             78.2        77.5        65 4              75 2
               Mar 1989             77.4        84.9        70 4              78 i
               Apr 1989             87 5        92.8        85 9              85 3
               May 1989             87 0        91.7        81 7 - -.-__ 80.4
               June 1989            80.4        90.8        76 2          ~__ 72 5
               Julv 1989            82.2        89.4        83 8              80 0
               Aug. 1989            75.8        87.0        83 6              79 3
               Sept. 1989           83.7        86.7        846         ~~~ 83 6
                                                                      -- .~~
               Oct.1989             79.0        77.1        858               83.4
               Nov.1989             80.1        76.9        85 5              79 9
               Dec.1989             69.7        72.9        74 1              71 6




               Page 14           GAO/RCED-W154 Alrllnes’ On-Time Prrfomce
                                             Appendix 1
                                             Percentage of Plight.9 That Met the On-Time
                                             Perfommnce Criterion




                                         Pacific
Delta   Eastern         Northwest    Southwest             Pan Am       Piedmont          Southwest          TWA          United       USAir       Average
 72.3       80.4              69.0          705                 74.3           80.3               82.4        784            792         67 4           766
 77 5       83.0              76.5          60.3                79.2           83.2               85.2        794           80.7         77 3-----      803
 70 1       766               73.0          73.3                74.6           73.2               82.7        77 5           79.8        732            75 1
 61 8       69.5              63.3          57.6                77 3           67 2               742         635           62.6         71 9
                                                                                                                                        -____           664
 65.6       61 5              61.6          81.6                72.6           62.4               85.0        65.5           69.8        73.2      -~ 692
 73.6       70.6              61.7          90.6                80.1           75.0               88.5        694            73.2        746            747
 77.3       75.6              75.2          92.3                78.8           78.7               86.3        744            78.4        785            788
 85.6       75.5              842           91.1                76.5           81.0               90.3        81 5           81.8        779            826
 87.4       76.2              83.9                 a            68.8           77.5               92.5        81.6           80.5        70.7           81 9
 87.6       82.8              84.4                 a            72.2           83.4               90.8        79.0           81.1        766            84 3
 86.1       77 1              79.2                 a            68.8           74.2               91.5        76.6           809         744            81 5
 86.9       80.8              78.4                 a            74.3           81.9               90.5        78.6           81.0        750            830
 88.0       SO.5              75.8                 a            83.2           85.3               89.3        81 9           840         83.7           856
 87.8       SO.3              80.7                 a            81 .l          84.3               82.8        81 5           79.7        79.7           83.4
 79.8       85.3              73.8                 a            77.8           77.5               77.0        79.0           73.8        76.6           77 7
 79.3       88.1              77.4                 a            75.1           71.6               79.6        73.9           69.0        766            76 1
 77.1       82.6              78.8                 a            76.3           74.3               77.0        72.1           69.2        792            76 7
 73.4       80.5              80.4                 a            73.7           68.8               71.0        71.7           65.6        76.0           72.4
 76.7              I>         82.1                 a            69.1           56.1               75.3        66.8           66.4        730            72.3
 84.2       92.9              89.7                 a            70.9           70.7               84.2        80.8           80.2        85.6           034
 81.7       87.2              83.6                 a            69.6           71.0               77.9        79.0           73.9        79.4           795
 72.0       80.1              75.1                 a            64.4           62.8               75.8        78.6           63.0        69.4           72.2
 79.1       86.4              77.6                 a            73.9           67.6               85.4        79.8           62.3        722            769
 79.0       86.9              74.6                 a            75.9           65.7               84.6        78.3           60.1        62.6           74 6
 82.7       80.6              74.6                 a            77.7                  c           86.9        82.2           76.1        71 7           79 6
 83.8       83.9              77.3                 a            76.8                  c           82.6        82.4           76.9        70.3           795
 80.1       81.3              76.7                 a            76.9                  c           81.4        81.8           78.1        684            780
 71.4       69.1              76.7                 a            73.5                  c           81.6        67.8           70.7        594            70.2

                                             Note: A flight IS considered on time if it departs from or arrives at the gate wrthrn 15 mtnutes of IIS
                                             scheduled trme. Arrknes are requrred to report on-trme performance data for non-stop flights from the
                                             largest U.S. arrports (currently 29). In addition, the arrlines have chosen to report thus data for their entire
                                             domestic systems and that information is rncluded in this table.
                                             %ot avalable; Pacrfic Southwest merged wrth USAir.

                                             “Eastern drd not report usable data because its strike Interrupted flight operations. Subsequent months
                                             reflect reduced operatrons
                                             CNot available; Ptedmont merged with USArr




                                             Page 15                                              GAO/RCEDBO-154 AlrUnes’ On-Time Pedormance
Appendix II

Flights Canceledfor Non-Mechanical Reasons ‘-



                                                                    Total flights           Flights canceled
               Month/year                                            performed            Number          Percent
               Sept 1987                                                 427 570            3.864
               Ott 1987                                                  448.620            3 oo, -.--~~~~~~-0709

               Nov 1987                                                   422.803             5 191 -             17
               Dec.1987                                                   440,403            11.493               26
               Jan 1988                                                   436,950            15,755               36
               Feb 1988                                                   412,579             7,323               18
               March 1988                                                 445.080             3,123 ---~~-          07
               April1988                                                  427,325             2,414                 06
               May 1988                                                   435,916            2,627                  06
               June 1988                                                  431.299             1.203                 Oi
               July 1988                                                  441.118            2.750    -             06
               Aug 1988                                                   446,769            2.478                  06
               Sept.1988                                                  424,075            2,038                  05
               Ott 1988                                                   441,670            2,306                  05
               Nov 1988                                                   420,861            3.517        ~~-....__ 08
               Dee 1988                                                   438,454            4.629                  11
               Jan.1989                                                   440.022            7.040                  16
               Feb.~1989                                                  395,176            8,106                  21
               March1989                                                  414,833            7,035                  17
               April1989                                                  405,604            2.086                  05
               Mav 1989                                                   416,160            2,523                  06
               June 1989                                                  406,293            3,722                  09
               July 1989                                                  417,166            2.953          --
                                                                                                            -~      07
               Aug.1989                                                   426,085            2,692          -~ __ 06
               Seot.1989                                                  415,068            5,195                  13
               Oct. 1989                                                  437,134            4,211                  10
               Nov.1989                                                   417,821             3,497     -_-         08
                                                                                                             -. __
               Dec. 1989                                                  429,490            9,282                  22
               Total                                                  11,962,344           132,054       ~___
               Monthly averaqe                                            427,227             4,716               11
               Note: Flights canceled for non-mechanical reasons are reported as part of the on-time dala counted as
               late, but also Identified as canceled flights.




               Page 16                                         GAO/RCED9@164 Airli~~es’ On-Time Prrfomxance
Ppe
%i%ves,   Scope,and Methodology


              On March 2, 1989, the Chairman of the Subcommittee on Investigations
              and Oversight, House Committee on Public Works and Transportation,
              and the Ranking Minority Member of the Subcommittee requested that
              we review the Department of Transportation’s (Dar> Office of Consumer
              Affairs. Specifically, the requesters asked us to focus on the data DOT
              maintains on airlines’ on-time performance. To respond to the reques-
              ters’ concerns, we obtained information on airlines’ scheduling practices,
              on-time performance statistics, and flights affected by mechanical and
              non-mechanical problems.

              To respond to these concerns, we interviewed officials in bur’s Office of
              Consumer Affairs, Office of General Counsel, Office of Economics,
              Office of Aviation Information Management, Federal Aviation Adminis-
              tration, and Office of Inspector General. We also obtained the most
              recent statistics maintained by nor on airlines’ on-time performance and
              consumer complaints regarding airline service.

              We discussed flight schedule changes and adjustments airlines made to
              reduce delays and improve their on-time performance with officials
              from American, America West, Delta, Eastern, United, and USAir air-
              lines. We also evaluated a sample of 216 nonstop flights to assess
              whether changes had been made to the time scheduled for each flight.
              We selected eight flights from each of the 27 largest airports for which
              on-time performance data were reported during October of 1987, 1988,
              and 1989. For each airport, we identified two routes with a large
              number and frequency of flights. Then, we selected four flights for each
              of the two routes, which were operated by at least two airlines and
              scheduled at various times throughout the day. For these flights, we
              compared the scheduled flight time and on-time performance data
              reported for October 1987,1988, and 1989.

              We interviewed officials from u&s Office of Inspector General (lo) to
              obtain information on their audit of the data on airlines’ flight delays.
              The overall objective of the IG audit was to determine the accuracy of
              flight delay data reported to DOTand the effectiveness of DCYPS   system in
              controlling the reported data and producing the on-time performance
              information. The IG’s audit assessed the accuracy of the on-time flight
              data submitted to DOTand whether the flights excluded from the on-time
              data because of mechanical problems were appropriately documented.
              The audit included statistical sampling (simple random sampling) to
              select a sample of the reported flight data and a separate sample of the
              excluded flights. The IG'S sampling methodology provided audit results
              used for statistical projections at the g&percent confidence level.


               Page 17                            GAO/WED-BtS164 Alrllnee’ On-‘Woe hdomunce
Appendix III
Objectives, Scope, and Methodology




Our review was conducted between July and December 1989 in accor-
dance with generally accepted government auditing standards.




 Page 18                             GAO/RCED9&154   AlrUnes’ On-Tinw Pwf~rmance
Appendix IV

Major Contributors to This Report


                        James Noel, Assistant Director
Resources,              Francis P. Mulvey, Assistant Director
Community, and          Nancy E. Oquist, Evaluator-in-Charge
                        John C. Johnson, Evaluator
Economic
Development Division,
Washington, D.C.




                                                                ..




                         P8ge 19
    c




.




        Requests for copies of GAO reports     should be sent to:

        U.S. General Accounting    Office
        Post Office Box 6015
        Gaithersburg, Maryland     20877

        Telephone   202-275-6241       ”

        The first five copies of each report   are free. Additional   copies are
        $2.00 each.

        There is a 25% discount    on orders for 100 or more copies mailed to a
        single address.

        Orders must be prepaid by cash or by check or money order made
        out to the Superintendent of Documents.