U.S. Federal Labor Relations Authority

Search form


United States of America




In the Matter of








Case No. 98 FSIP 19




    Local ZKC, National Air Traffic Controllers Association (Union) filed a request for assistance with the Federal Service Impasses Panel (Panel) to consider a negotiation impasse between it and the Department of Transportation, Federal Aviation Administration (FAA), Kansas City, Missouri (ZKC or Employer) resulting from an agency determination to terminate a 5-4/9 compressed work schedule (CWS) under the Federal Employees Flexible and Compressed Work Schedules Act of 1982 (Act), 5 U.S.C. §§ 6120-6133.(1)

      Following investigation of the request for assistance, the Panel determined that the impasse should be resolved on the basis of an informal conference to be followed by a single written submission from each party if the dispute were not resolved during the informal conference. Accordingly, Supervisory General Attorney Ellen J. Kolansky held an informal conference with the parties on January 7 and 8, 1998. Although the parties explored settlement possibilities during the informal conference, the dispute was not resolved. The parties then submitted written statements, and the Panel has now considered the entire record.


    The Employer regulates air commerce and transportation on the ground and in the air. In Air Traffic Control, the Employer is engaged in managing air traffic in a safe, orderly, and expeditious manner. The Kansas City Air Route Traffic Control Center (Kansas City Center) is one of about 20 such centers nationwide. The Union represents 360 bargaining-unit employees at the Kansas City Center who are part of a nationwide, consolidated unit of 14,000. The affected employees work as air traffic controllers (controllers or ATCs) who monitor radar screens chiefly to maintain separation between aircraft. They are covered by a master collective bargaining agreement (MCBA) which expired in August 1997, but under a provision in the MCBA, its terms remain in effect until a successor agreement is implemented. At this time, successor agreement negotiations are underway.(2)

    At the Kansas City Center, the Employer is responsible for monitoring 195,000 square miles of airspace. Controllers are organized into teams that are assigned to one of six geographic areas. Between 49 and 52 controllers are assigned to each team. Within each area, there are a total of 40 different sectors which are further subdivided into 3 altitude levels. Controllers must be certified in a particular sector to be permitted work in that sector. To work at the full performance level (FPL), controllers must be certified in four sectors within an area. As a result, these employees are not fungible, and thus they cannot substitute for other ATCs in sectors where they lack certification. The work, maintaining airplane separation, cannot be deferred to a later time; should a sector become too crowded, airplanes may be rerouted to different sectors.

    The Employer provides service to aircraft on a 24-hour, 7-day-per-week basis. To meet mission requirements, employees are assigned to shifts established under a basic watch schedule.(3) The length of the shift depends on whether the employee has selected a compressed, flexible, or straight-8 schedule. At ZKC, in addition to the 40 percent of ATCs permitted to work a 5-4/9 CWS, six employees work a straight-8 schedule. The balance, nearly 60 percent, work an 8-hour day with flexible arrival and departure times known as a flexible work schedule (FWS), and may earn credit hours.(4) By agreement, some 21 time slots have been set aside to permit ATCs to work credit hours without prior approval.(5) Employees also may swap particular shifts so long as they are certified to work in each other’s sector. If an employee is turned down for annual leave, that employee is permitted to get a qualified substitute who can earn credit hours for that time period.


    In accordance with section 6131(c)(3)(C) of the Act, the sole issue in dispute is whether the finding on which the Employer has based its determination to terminate employees’ 5-4/9 CWS is supported by evidence that the schedule has caused an adverse agency impact.(6)


1. The Employer’s Position

    The Employer contends that the 5-4/9 CWS has had an adverse impact under each of the three criteria specified in the Act: Regarding service to the public, the Employer’s chief concern is safety; "adequate staffing levels is . . . a necessary ingredient to maintaining an acceptable margin of ‘safety’ for air travelers." Although "the Agency has an ample number of Air Traffic Control Specialists working at the Kansas City Center, it just can’t get them in the right place at the right time . . . because of compressed work schedules." In particular, there are too few controllers during peak traffic periods of 7 a.m. to 10 a.m., and 7 p.m. to 11 p.m. The situation is further exacerbated by pressures from having to cover absences on these shifts caused by a 51 percent increase in leave use(7) and employees’ use of preapproved credit hours. The latter creates a "pseudo" 5-4/9 schedule which is of little benefit to the Employer, since supervisors who are unable to predict whether employees will actually decide to work credit hours must, by contract, adequately staff shifts 28 days in advance.(8) Current staffing shortages cause managers to fear they may be unable to provide an adequate response to exigencies that arise unpredictably. A recent example is a power outage that occurred at ZKC on December 18, 1997. In one area with only nine ATCs on duty that day and no radar, radio frequencies, or telephones working, a supervisor reported being more afraid than at any time in his career. Such incidents increase employees’ stress and absences.

    As to costs, approximately $310,191 per year is wasted because controllers on a 5-4/9 schedule work only 9 shifts per pay period instead of 10, and have a RDO once every 2 weeks. Additional waste occurs daily between 1 and 4 p.m., when two to three extra controllers are on duty.(9) The cost of holiday and Sunday pay for employees on CWS is up by $19,000. In addition, a survey of managers and supervisors shows that they use some 4,702 hours annually to administer the schedule; 40 percent of these hours is attributable to CWS.

    With respect to productivity, while important refresher training, briefings, and staff studies used to occur while employees were working overtime, supervisors’ affidavits state that 75 percent of the overtime budget, or $68,339, is now being used to cover staffing shortages on the evening shifts. This is consistent with the staffing-to-traffic data that it has provided to the Panel indicating that there is a greater demand for ATCs between 7 p.m. and 11 p.m. Because the current CWS results in too many ATCs at the workplace at midday, they spend 5 percent less time on position (TOP)(10) than they would under a straight-8 schedule, and more time sitting in the break room. Its studies show that TOP is decreasing annually for employees on all schedules. Finally, the Employer is concerned about operation errors.(11) An October 18, 1995, FAA survey of "Air Traffic Systems Effectiveness" found that even the ATCs at ZKC attribute such errors to staffing, among other things.

2. The Union’s Position

    The Union asserts the Employer has failed to meet the burden of proof required under the Act, and has presented a case consisting "primarily of broad generalizations." In this regard, it challenges the Employer’s data, in part, because a large portion was collected after May 5, 1997, the date the Employer notified the Union that the CWS was having an adverse agency impact.(12) Furthermore, since this case concerns a decision to terminate a schedule, the Panel has indicated in previous cases that employers should rely on actual documentation, not the mere speculation relied upon here of a predicted inability to meet operational needs. Notwithstanding the "doom and gloom picture" the Employer paints, with respect to getting the job done, the Center’s employees are often commended for providing "outstanding" service to the public, including during the December 18, 1997, power outage at ZKC.(13) The Employer’s concern about the level of staffing is also misguided, as 8 or 9 ATCs per area would signify an adverse impact, not the 12 to 15 controllers now available during peak hours. If the Employer believes that levels are too low, it has the prerogative to determine appropriate staffing levels on all shifts. In actuality, "the only way to rectify the supposed evening shift staffing shortage would be for the parties to renegotiate the basic watch schedule . . . not to terminate the compressed work schedule." Despite the Employer’s predictions of a potential inability to handle air traffic, based on actual experience under the schedule, the Employer has identified no disasters, complaints, loss of service, or incidents related to the 5-4/9 CWS. Its staffing-to-traffic data also are suspect because the monthly average of traffic density it is based upon does not take into account day of the week, time of year, and sector variations. The Employer’s claims of evening shift shortages are, therefore, completely unsubstantiated.

    On the matter of costs, the Union asserts that the 5-4/9 CWS actually resulted in $35,000 worth of savings in salary expenditures, offsetting $19,000 in extra holiday and premium pay expenses the Employer attributes to CWS. That the Employer’s overtime budget has been capped at $100,000 annually undercuts any allegation that rising overtime use provides evidence that the Employer has met its burden under the Act. Regarding the time supervisors contend is used to administer the schedule, the Employer does not indicate what portion of the 4,200 hours is attributable to FWS.

    Since the number of controllers who handle air traffic has remained constant during the period under the schedule while air traffic volume has increased by at least 10 percent, productivity has actually improved. In this regard, TOP data is an unreliable measure of productivity because, among other reasons, supervisors can influence how long a controller is on position just by granting or denying breaks, and it is based on flawed sign-in and sign-out data. The latter is substantiated by an "external study" that discredits the validity of such data. Moreover, employees working CWS actually remain on radar positions longer than 8-hour employees. As to the number of operation and deviation errors such as "less than adequate separation," these figures rise and fall year-by-year without any apparent pattern to suggest causation and, therefore, such data are inconclusive. Furthermore, the error data provided by the Employer do not track closely the time period during which CWS has been in effect, nor is the data separated by attribution to work schedule type. There also are no comparison data provided from the period prior to the April 1992 introduction of CWS which would definitively allow one to conclude that such schedules produced an increase in error rates. Historically, as recited in the report the Employer provides from Gateway (one of the six geographic areas at ZKC), errors are mainly associated with inattentiveness during periods of light traffic. Finally, it is simply bogus to suggest that the Employer is losing an entire shift each pay period for the 144 ATCs on CWS when the length of their workweek remains 40 hours.


    Under section 6131(c)(3) of the Act, the Panel is required to take final action in favor of the agency determination if the finding on which the determination is based is supported by evidence that a CWS has caused adverse agency impact. Having considered the record before us, we conclude that the Employer has not met its statutory burden. In arriving at this conclusion, we are acutely aware that this case stands out among those the Panel has previously decided on this subject because of the critical nature of the job controllers perform to ensure the safety of airplane crews and passengers. We also recognize, as discussed in the December, 1997 report provided by the Employer, Avoiding Aviation Gridlock and Reducing the Accident Rate, A Consensus for Change (see footnote 11), that as air traffic grows, the FAA faces increased challenges to not only maintain, but to improve travel safety.

    Without losing sight of the paramount importance of these safety-related concerns, and the unacceptable consequences of a mistaken decision in this matter, the Panel’s task nevertheless is to evaluate whether the evidence upon which the Employer based its May 1997 determination to terminate the schedule establishes that the 5-4/9 CWS caused an adverse agency impact. After careful scrutiny, we find that the data the Employer presents do not establish that the affected employees’ compressed schedules have resulted in diminished service to the public, reduced productivity, or increased costs to the agency. The relevant period in question is April 1996 through May 5, 1997, the period prior to the adverse agency impact finding during which 40 percent of controllers were working the 5-4/9 CWS. In our view, it is reasonable to expect that data from this period would be presented along with data originating before the CWS was implemented so that a direct comparison could be made. In the alternative, data from the relevant period could be compared to corresponding data regarding the productivity, level of service, and costs associated with the efforts of employees working other types of schedules during that same period. Instead, the Employer mainly presents information by fiscal year which, at best, overlaps the target period. For example, it offers a report on operation errors and deviations which is dated FY 1991 through FY 1995.(14) To cite another example, the TOP data the Employer provides to demonstrate that productivity has been reduced is also listed by fiscal year, or was collected after the finding was made (May 23 through 29, 1997) and, therefore, does not track the appropriate time frame.(15) This imprecision does not appear to be inherent in the nature of the subjects in question, but rather arises because of the methods the Employer used to collect and present its evidence. These defects render much of the Employer’s data, at best, inconclusive, and not in accordance with the basic framework established under the Act.

    Turning next to the Employer’s evidence that overtime has increasingly been used to cover staffing shortages during evening shifts rather than training needs, the graphs and charts provided are essentially bare statements that do not document the change in allocation of this resource or reference what specific training opportunities, if any, were lost. The affidavit the Employer cites from an Operations Supervisor, dated January 6, 1998 (8 months after the Employer made its finding), who attests to the change, references his review of overtime data with no corroborating records.(16) Without more, the Panel cannot reasonably conclude that adverse agency impact has been established, particularly in light of the fact that ATCs on CWS work the same number of evening shifts in a pay period as those working straight 8-hour shifts.

    By contrast, the published articles and corroboration by managers present at the informal conference support the view that ATCs, regardless of the type of schedule they work, are meeting the public’s safety needs by maintaining separation between aircraft. Such performance continues despite a 10-percent (or greater) increase in air traffic in recent years. Controllers also reportedly respond to fast-breaking, difficult conditions such as weather changes and traffic shifts. In addition, we note that at least 14 other centers, manned by ATCs performing identical work, permit their employees to work CWS; at least one-third allow participation rates above the 40-percent rate at ZKC. Other cost and productivity data presented--holiday and premium pay increases and hours spent administering the schedule, to name two--are also unpersuasive in demonstrating adverse agency impact. In this regard, while administering employees’ work schedules is undoubtedly burdensome to some supervisors, it appears that this is more a matter of the sheer multiplicity of options currently available (and other prevailing constraints), and not solely attributable to CWS. In any event, such burdens do not rise to the level of establishing adverse agency impact.

    Notwithstanding our conclusion that the Employer has failed to meet its statutory burden, we recognize that serious pressures and problems exist at ZKC related to scheduling and staffing levels which may have safety-related consequences. As previously indicated, CWS is but one factor within the parties’ complex menu of work schedule options and other mandates. In addition to the CWS, the menu includes FWS, credit hours with the preapproval feature (and substitutes being permitted to earn credit hours when they take another employee’s shift), rising leave use by a maturing workforce, and constraints such as the 2-hour limit for TOP, and the 28-day assignment notice requirement, among others. Although based on the evidence presented, the elimination of the CWS is not a permissible solution to the Employer’s perceived problems, we are persuaded that the key to addressing its legitimate concerns resides in the annual negotiations just getting underway between the parties over the basic watch schedule. In this regard, the MCBA clearly states that "the basic watch schedule must satisfy coverage requirements." The Panel is confident that the parties’ mutual interest in ensuring the maximum level of safety will enable them to craft satisfactory solutions. If this is not the case, however, and the Employer gathers new evidence which demonstrates that the CWS has had an adverse agency impact, following the negotiations specified in § 6131 (c)(3)(B) of the Act, the parties are not prevented from again seeking the Panel’s assistance. Nor are they prevented from bringing a case under the Statute should they reach an impasse during their current negotiations over the basic watch schedule.


    Pursuant to the authority vested in it by section 6131 (c) of the Federal Employees Flexible and Compressed Work Schedules Act, the Federal Service Impasses Panel, under section 2472.11(b) of its regulations, hereby orders the Employer to continue air traffic controllers’ 5-4/9 compressed work schedules.


By direction of the Panel.

H. Joseph Schimansky

Executive Director

January 30, 1998

Washington, D.C.


1.Under a 5-4/9 schedule, employees work eight 9-hour days, one 8-hour day, and have one regular day off (RDO) each pay period.

2.The Union reports that Article 34, Hours of Work, which includes references to alternative work schedules (AWS) is one of a number of articles being put on an expedited track at these negotiations.

3.The basic watch schedule is a generic schedule which establishes staffing levels for the year. In Article 32, Section 1, of the MCBA, it is defined “as the days of the week, hours of the day, rotation of shifts, and change in regular days off.” The same section of the contract states that “the basic watch schedule must satisfy coverage requirements.” Once the schedule is set, employees bid for available shifts based on seniority as determined by the Union. Employees work a 2:2:1 shift rotation pattern: two evening shifts (e.g., 1 p.m. to 9 p.m.), two day shifts (e.g., 5 a.m. to 2 p.m.), and one midnight shift (e.g., 11 p.m. to 7 a.m.). In January 1997, the parties agreed to roll over the previous year’s watch schedule with some minor adjustments. That agreement permitted employees to bid for their yearly schedules.

4.Article 34, Working Hours, Section 7, of the MCBA, defines credit hours as “any hours worked under a flexible AWS which are in excess of any employee’s basic work requirement and which are worked at the election of the employee after approval by the Employer.” Employees are permitted to earn 8 credit hours per pay period and to carry over 24, the maximum allowable under the law. Section 8 reads: “Procedures for using credit hours shall be the same as those for approving annual leave requests.”

5.Under the Act, when credit hours are permitted, normally the employee requests permission in advance to earn credit hours based on the availability of work.

6.5 U.S.C. § 6131(b) defines adverse agency impact as:


(1) a reduction of the productivity of the agency;

(2) a diminished level of the services furnished to the public by the agency; or

(3) an increase in the cost of agency operations (other than a reasonable administrative cost relating to the process of establishing a flexible or compressed work schedule).

Under section 6131(c)(3) of the Act and section 2472.11(a)(2) of its regulations, the Panel is required to take final action in favor of the agency head’s (or his delegatee’s) determination to terminate a compressed work schedule if the finding on which the determination is based is supported by evidence that the schedule has caused an adverse agency impact. If, however, the finding on which the determination is based is not supported by evidence that the schedule has caused adverse impact, the Panel, in accordance with section 2472.11(b) of its regulations, shall take "whatever final action is appropriate" to resolve the impasse. The Act’s legislative history clearly indicates that the Employer bears the burden of proving adverse agency impact. S. REP. NO. 97-365, 97th Cong., 2d Sess. at 16.

7.Regarding leave use, since the average age of controllers is 42, up from 31 a decade ago, as a group, they tend to earn the maximum amount of annual leave, 8 hours per pay period.

8.At the informal conference, the Employer suggested that, with respect to the reduced number of shifts an employee works under the CWS, its predicament is similar to that of the Palo Alto Veterans Affairs Hospital, which convinced the Panel to implement at once the 8-hour, 5-day-per-week shifts which “allowed for more effective around-the-clock coverage” by security guards. Department of Veterans Affairs, Palo Alto Health Care System, Palo Alto, California and Local 2110, American Federation of Government Employees, AFL-CIO, Case No. 96 FSIP 1 (March 20, 1996), Panel Release No. 385.

9.The Employer states: “There are 16 hours of work because of the traffic between the hours of 7 a.m. and 11 p.m. Two 8- hour shifts cover that 16-hour period. With two sets of 9- hour shifts, one beginning at 7 a.m. and the other ending at 11 p.m., it is impossible to eliminate that 2-hour overlap.”

10.Time on position means the actual time during a particular shift when the controller is logged on to the radar screen or performs certain other mission-related activities. By contract, controllers must be given breaks at least every 2 hours except in emergency circumstances.

11.The Employer provided a December 24, 1997, article from the Tribune Chronicle, “Air System Flies Toward Crash” by Norman Mineta, chairman of the National Civil Aviation Review Commission and a senior vice president of Lockheed Martin, IMS. A section noted by the Employer states: “But it will not be good enough to simply retain the current accident rate while air traffic mounts. If that happens, there will be about six to seven catastrophic accidents a year by 2010, about double the three or four of today. That is unacceptable.” A Congressionally mandated report provided by the Union, Avoiding Aviation Gridlock and Reducing the Accident Rate, A Consensus for Change, December 1997, prepared by the same Commission, suggests that “[t]he FAA must establish performance measures to focus resources and hold the agency’s safety management accountable to make improvements.”

12.The Union cites Department of the Air Force, Dover Air Force Base, Dover AFB, Delaware and Local 1709, American Federation of Government Employees, AFL-CIO, Case No. 97 FSIP 107 (September 16, 1997), Panel Release No. 401 (Dover AFB), a case arising under the Act, in which the Panel found that the employer’s evidence did not support its finding that the 4-10 CWS for jet engine mechanics who overhaul engines for the C-5 cargo plane had caused an adverse agency impact. The Panel explained that evidence of adverse agency impact collected after a CWS is terminated is inconsistent with the basic framework of section 6131 of the Act; the relevant evidence should be that on which the employer bases its finding of adverse agency impact, and not evidence arising after the finding is made.

13.During the informal conference, members of the Employer’s team concurred in this view. In addition, in the October 1997 edition of the ZKC newsletter, the Air Traffic Manager noted “exceptional operational performance” by controllers. He also wrote: “We had three consecutive months this summer where we broke records for traffic handled. We also broke the two million operations mark and set numerous peak day traffic records. And to boot, we did this and broke our five-year average for operational errors with 17 errors. . . As if this wasn’t enough, we also responded to significant convective weather and several equipment and environmental issues of various impact to our personnel.”

14.Other problems with the report are that it surveys the entire region rather than just ZKC, and does not separate error totals by work schedule type. The survey of managers and employees on their beliefs about what causes controllers to make errors also is not definitive. Furthermore, as cited by the Union, a report from the Gateway area notes that controller errors are associated with inattentiveness during periods of light traffic, and not scheduling problems caused by the CWS.

15.TOP data also appear to be an unreliable measure of productivity since, among other things, a report cited by the Union states that ATCs fail to log on the radar screen at least 25 percent of the time.

16.It is interesting to note that following his discussion of problems with both overtime and credit hours, he states: “Management must accept some responsibility for allowing these programs to have the impact they do.”