OPINION OF THE COURT
ROSENN, Circuit Judge.
We are called upon in this appeal to consider a question of widespread importance
to both employers and employees under the Fair Labor Standards Act (FLSA or the Act), 29 U.S.C. §§ 201-216b (1976): How is the exempt status of managerial employees from the overtime pay provisions of the FLSA to be measured when those employees perform nonexempt work during a strike? The district court answered this question by adopting a “workweek standard” proposed by the Secretary of Labor (the Secretary) by which the exempt status of managerial employees would be determined through an evaluation each week of '4¿he “primary duty” of the employee.
This case arose in the context of the use by appellant, Western Union Telegraph Company (Western Union), of managerial employees to perform non-managerial duties during a strike by rank and file employees. The district court held that Western Union must determine whether the “primary duty” of the managerial employee during each week of the strike was managerial and thus exempt, or non-managerial, and therefore subject to the overtime pay provisions of the FLSA. Western Union appeals. We reverse.
I.
This case has its factual roots in a strike in 1971 by Western Union’s entire rank and file work force. In order to continue limited services, particularly those connected with domestic defense communications, Western Union used approximately 2,100 managerial employees to perform the struck-work. A substantial number of them spent more than 50 percent of their time during this period on tasks normally performed by the rank and file. Managerial employees are those working in a bona fide executive, administrative or professional capacity.
Under FLSA, the general rule is that employers must pay overtime compensation to employees working over forty hours per week. 29 U.S.C. § 207(a)(1) (1976).
Managerial employees, however, are exempt from the overtime pay provisions by virtue of section 213(a)(1) of the Act.
Western Union did not pay overtime to its managerial employees for hours worked in excess of forty per week while performing the non-managerial struck-work. The Secretary of Labor petitioned the district court for an injunction
to prohibit the non-payment of overtime by Western Union to its managerial employees and for a determination of how much retroactive overtime pay was owed. The issue before the court was whether a managerial employee could lose his exempt status by performing non-managerial struck-work. The district court held that the administrative and professional employees could indeed lose their exempt status, but that under the emergency exemption to FLSA,
executive employees retained their managerial status. The court held that administrative and professional employees who work more than forty hours per week at non-managerial duties must be paid one and one-half times their regular rate of pay for the hours of overtime worked.
On appeal, we upheld the district court’s determination that the administrative and professional personnel could lose their exempt status during the strike, and because Western Union abandoned its emergency
exemption argument for executive personnel, we held that executives could similarly lose their exempt status.
Brennan v. Western Union Telegraph Co.,
561 F.2d 477, 484 (3d Cir. 1977) (Seitz, C. J.). We remanded the case because,
inter alia,
“the district ' court [had] not yet specifically decided if, under the regulations, the determination of whether an employee is ‘employed in a bona fide executive . . . capacity’ is to be made with respect to each week separately or with respect to a broader period of time.”
'''Id.
at 483.
On remand, the Secretary proffered a workweek standard under which an employer must in each workweek in which a managerial employee works more than forty hours, determine whether the managerial employee performed primarily exempt managerial work or nonexempt, non-managerial work. If in any given week of the strike at Western Union, an executive, administrative or professional employee performed primarily nonexempt duties, the company would be required to pay overtime at the rate of one and one-half times the managerial employee’s salary for each hour worked /over the forty-hour limit. The district court adopted the workweek standard and continued the injunction. Western Union appeals from this order as it pertains to high-salaried managerial employees whose eligibility for exemption can be determined under the “short test” of the final provisos of 29 C.F.R. §§ 541.1(f), 541.2(e)(2) and 541.-3(a) (1979). In assessing whether the district court properly adopted this workweek standard, it is necessary to examine first the regulations defining managerial employee status.
II.
Whether an employee is a managerial employee may be determined in one of two ways under the FLSA regulations promulgated by the Department of Labor. Executive, administrative or professional status may be determined under a “long” or a “short” test. Under the “long test,” whether an individual is a managerial employee is determined through various job-related criteria. For example, an executive is an employee who:
(1) has management as a primary duty;
(2) customarily and regularly directs the work of two or more employees;
(3) has authority to hire or fire or who has weight in the determination of other employees’ job status;
(4) exercises discretionary powers on a customary and regular basis;
(5) does not spend beyond a certain percentage of his hours in each workweek performing non-executive duties;
(6) earns at least $155 per week.
See
29 C.F.R. § 541.1(a)-(e) (1979).
Under the “short test” for high-salaried managerial employees, managerial status is determined by reference to the employee’s salary level and his “primary duty” as an employee. For example, an employee who earns not less than $250 per week (as opposed to $150 per week under the long test) and whose “primary duty consists of the management of the enterprise . . . , and includes the customary and regular direction of two or more other employees therein,” 29 C.F.R. § 541.1(f), is deemed to have met the requirements for executive status.
Only the short test definitions of managerial employees are at issue in this case. The dispute between Western Union and the Secretary is whether the “primary duty” of a managerial employee under the short tests can be determined in such a narrow time frame as a workweek. The district court indicated that the Secretary’s workweek standard was entitled to deference unless Western Union could show that it was “plainly erroneous and unreasonable.” Although admitting that the question was a “close one,” the court deferred to the Secretary’s proffered workweek standard.
We believe that the district court erred in deferring to the Secretary’s position because the workweek standard is in (reality not an “interpretation” of the gov-erning statute or of the existing short test regulations, but rather a substantive amendment of the regulations. As such, we believe the Secretary must engage in a rulemaking procedure conforming with the notice and comment provisions of the Administrative Procedure Act, 5 U.S.C. § 553 (1976).
III.
At the heart of these proceedings is the question of how to measure the “primary duty” of high-salaried managerial employees when they deviate from their normal work course and perform nonexempt work. There is no difficulty in determining when exempt status is lost through the performance of nonexempt work by a managerial employee under the long tests. If a long test managerial employee spends more than 20 percent of his time in a given workweek performing nonexempt work, the managerial exemption is lost.
See
29 C.F.R. §§ 541.1(e); 541.2(d); 541.3(d). No such formula is found in the short test provisos for managerial employees. Indeed, the district court conceded this stating:
It is clear that the language [of the provisos] is far from conclusive on the issue of whether “primary duty” should be determined on a workweek basis or some longer period. The defendant stresses that “primary duty” cannot be read as “primary strike duty.” On the other hand, the proviso cannot be read to exclude the workweek time frame. The Secretary, while noting that the week is the time frame for measuring salary, does concede that the regulation does not say that “primary duty” is to be ascertained on a weekly basis.
As to this proviso, the following can be said: It does not mandate the adoption of either party’s position in the matter in explicit terms.
The district court thereupon deferred to the Secretary’s workweek standard as an “interpretation” of primary duty under the short tests because it was not “unreasonable or plainly erroneous.”
The fundament of the Secretary’s position is that although the short test provisos do not themselves provide the source of the workweek standard, the presence of the workweek standard in other parts of FLSA, including the long tests, makes it reasonable to apply a workweek standard in measuring primary duty under the short tests. The workweek is used to gauge the obligations of employers to pay the minimum wage and any overtime that might be earned by hourly employees. Although a workweek standard is indeed a fundamental part of FLSA, the exemption of managerial employees from the general coverage of the Act confirms our view that they were not within the scope of Congress’ concerns in enacting this legislation. The reason is plain: the underlying and declared purpose of FLSA is to eradicate from interstate commerce conditions detrimental to “the maintenance of a minimum standard of living necessary for health, efficiency, and general well-being of workers.” 29 U.S.C. § 202(a) (1976). Thus, FLSA bans from interstate commerce products produced through the imposition of lengthy hours of work at low wages.
United States v. Darby,
312 U.S. 100, 109-10, 61 S.Ct. 451, 454-55, 85 L.Ed. 609 (1941).
Specifically, the Act was manifestly designed to place a floor under wages and a ceiling over hours of employment. The limitation in hours also served the purpose during a time of extraordinary unemployment and depressed wages of spreading the work and providing extra compensation for overtime.
White v. Witwer Grocer Co.,
132 F.2d 108, 110 (8th Cir. 1942). Setting a floor under wages and a ceiling over normal hours of employment had the further purpose of eliminating from commerce the evils attendant upon low wages and long
hours of service.
McComb v. Farmers Reservoir & Irrigation Co.,
167 F.2d 911, 913 (10th Cir. 1948),
modified,
337 U.S. 755, 69 S.Ct. 1274, 93 L.Ed. 1672 (1949). Thus, the forty-hour provision of the Act is an essential element in spreading the work among the rank and file and compensating an employee in a specific manner for the strain of working longer than forty hours.
Bay Ridge Co. v. Aaron,
334 U.S. 446, 470-71, 68 S.Ct. 1186, 1199-1220, 92 L.Ed. 1502 (1947). There is nothing in the Act that demonstrates Congress had these concerns in mind for managerial employees and that it, therefore, intended that the workweek standard also be applied to them in measuring their exempt status.
Managerial employees, particularly executives, generally direct and operate a business enterprise. Because they have the power to direct, supervise, and manage the operations and are paid accordingly, granting managerial employees exempt status must have been a recognition that they are seldom the victims of substandard working conditions and low wages. Congress apparently acted on this basic principle of industrial organization, realizing that compensation and working conditions of managerial employees normally would be significantly above those of the hourly wage earner. Therefore, there was no need for Congress to be concerned about the number of hours worked per week or the hourly wage for them. But under the Secretary’s theory, a high-salaried executive of the company who is paid a weekly salary as much as $4,000 or $5,000 would be entitled to overtime pay at one and one-half times his weekly salary for time worked in excess of forty hours, if during any workweek, he spent the major portion of his time on nonexempt work. We see no basis in the underlying theory or policy of FLSA which warrants such a result. We therefore conclude, that the workweek unit, as a yardstick for evaluating FLSA obligations to nonexempt workers, does not
ipso facto
imply that a workweek standard must necessarily be applied in determining whether employees, admittedly managerial, have lost their exempt status.
The Secretary also argues that the presence of a workweek standard in the long tests for measuring exempt status means that exempt status was intended to be similarly measured under the short tests. We have found no evidence that the short tests were so intended. The only mention of the workweek in the regulations is in the last requirement of the long tests.
See
29 C.F.R. §§ 541.1(e); 541.2(d); 541.3(d). The-, careful inclusion of the workweek standard
j
in one subsection of the long tests and its
j
exclusion in the other four criteria of the/ long tests, as well as the short tests, indicates that the exclusion was intentional.' Under the usual canons of statutory construction, where Congress, or in this case an administrative agency,
has carefully em-^ ployed a term in one place and excluded it
\
in another, it should not be implied where
J
excluded.
See Federal Trade
Commission v.
Sun Oil Co.,
371 U.S. 505, 515, 83 S.Ct. 358, 364, 9 L.Ed.2d 466 (1963);
City of Burbank v. General Electric Company,
329 F.2d 825, 832 (9th Cir. 1964).
Moreover, our interpretation is consistent with the structure and substance of the regulations and effects a result compatible with economic reality. First, subsections 541.1(e); 541.2(d); and 541.3(d) in which the workweek standard is included are the only requirements, in either the long or short tests, which have specificity and potential for routine weekly calculation — the percentage of nonexempt work performed. However, no mention of the workweek standard is made in the remaining criteria.
See
page 1249
supra.
The practical realities of industrial operations support the omission. One would hardly expect to evaluate an executive employee whose primary duty was management of the entire work force, or of a department, on a weekly basis. Such employees are not normally required to keep time sheets or punch time clocks. Under such a system, reporting and identify
ing the managerial or non-managerial status of employees on a weekly basis appears incongruous and ponderous, to say the least.
Second, the Secretary’s own regulations provide that a high-salaried employee meeting the short test is deemed to meet all the requirements in paragraphs (a) through (f) of section 541.1.
See Wainscoat v. Reynolds Electrical & Engineering Co.,
471 F.2d 1157, 1163 (9th Cir. 1973). Indeed, the regulations state that “[i]f an employee qualifies under this proviso, it is not necessary to test that employee’s qualifications in detail under [the long test].” 29 C.F.R. § 541.119 (executives).
See
29 C.F.R. § 541.214 (administrators); 29 C.F.R. § 541.315 (professionals); L. Weiner, Federal Wage and Hour Law, 117, 119 (1977). The policy behind such a separation of the tests is evident from the tests themselves: short test managerial employees earn approximately $100 per week more than long test managerial employees and there is accordingly less need to make a detailed examination of their managerial duties because it is likely that at the higher salary rate, the managerial employee is indeed managerial. We read the Secretary’s regulations as an expression of intent to keep the long and short tests separate and accordingly reject any attempt to import a workweek standard from the long tests to the short test provisos.
Nor is there anything in the definition of primary duty which compels its evaluation in any particular time frame. The regulation defining primary duty provides:
A determination of whether an employee has management as his primary duty must be based on all the facts in a particular case. The amount of time spent in the performance of the managerial duties is a useful guide in determining whether management is the primary duty of an employee. In the ordinary case it may be taken as a good rule of thumb that primary duty means the major part, or over 50 percent of the employee’s time. Thus, an employee who spends over 50 percent of his time in management would have management as his primary duty.
29 C.F.R. § 541.103. We note the conspicuous absence of any time frame in which the allocation of the employee’s time between managerial and non-managerial duties is to be assessed. The regulation does not specify whether 50% of the employee’s time each
week,
each
month,
or each
year
is to be the yardstick by which primary duty is measured. Indeed, the regulation appears aimed at a holistic approach to primary duty, for it goes on to state:
Time alone, however, is not the sole test, and in situations where the employee does not spend over 50 percent of his time in managerial duties, he might nevertheless have management as his primary duty if the other pertinent factors support such a conclusion. Some of these pertinent factors are the relative importance of the managerial duties as compared with other types of duties, the frequency with which the employee exercises discretionary powers, his relative freedom from supervision, and the relationship between his salary and the wages paid other employees for the kind of nonexempt work performed by the supervisor.
Id.
This language indicates to us that the 20 percent limitation on nonexempt work in the long tests is a specific exception from the general concept of primary duty which does
not
envision the number of hours performed at a given task as necessarily the bellwether of managerial status. Primary duty itself does not specify any particular time frame for its evaluation and may depend on a number of factors exclusive of hours worked at a given task.
We are left, therefore, with the ineluctable conclusion that the regulations themselves provide no basis for measurement of primary duty in any specific time frame. The provisos themselves, the long tests and the definition of primary duty in no way compel the measurement of primary duty in any particular time frame.
IV.
Although the courts have the power to interpret the FLSA,
Rachal v. Allen,
376 F.2d 999 (5th Cir. 1967);
Walling v. LaBelle S.S. Co.,
148 F.2d 198, 202 (6th Cir. 1945), we do not have the judicial power to legis
late new provisions into the Act. In
E.C. Schroeder Co. v. Clifton,
153 F.2d 385, 390 (10th Cir.),
cert. denied,
328 U.S. 858, 66 S.Ct. 1351, 90 L.Ed. 1629 (1946), the court aptly summarized the proper role of judicial interpretation of FLSA:
The Act is to be accorded a liberal construction. But the function of judicial interpretation of a statute is to bring out and give effect to that which is already in it, latent or otherwise. It is not to add new provisions, substantive or otherwise, which the legislative tribunal in the exercise of its permitted choice omitted or withheld.
Of course, in any given case it may be exceedingly difficult to distinguish between judicial “interpretation” and judicial “legislation.” A similar problem is especially acute in administrative law where an agency may have both adjudicatory and legislative or rulemaking powers.
See
K. Davis, Administrative Law Treatise, § 7:2 (2d ed. 1979). Courts have struggled to determine whether an agency is acting in a judicial fashion or whether it is engaged in substantive rulemaking.
See, e. g., NLRB v. Bell Aerospace Co.,
416 U.S. 267, 94 S.Ct. 1757, 40 L.Ed.2d 134 (1974);
NLRB v. WymanGordon,
394 U.S. 759, 89 S.Ct. 1426, 22 L.Ed.2d 709 (1969);
Hoffmann-LaRoche, Inc. v. Kleindienst,
478 F.2d 1 (3d Cir. 1973);
see
Robinson,
The Making of Administrative Policy: Another Look at Rulemaking and Adjudication and Administrative Procedure Reform
118 U.Pa.L.Rev. 485 (1970).
In this case, we are not faced with distinguishing between agency adjudication and rulemaking, for no adjudication has occurred at the agency level. Rather, the suit brought by the Secretary originated in the federal district court. Thus, the problem is whether we may properly exercise our adjudicatory powers in resolving the conundrum now before us or whether the Secretary should have instead proceeded by rulemaking.
' We believe that the Secretary is in effect asking us to legislate a workweek standard into the concept of primary duty under the guise of “interpreting” the short test provisos. The difference between a purely judicial inquiry and a legislative proceeding was aptly summarized by Mr. Justice Holmes in
Prentis v. Atlantic Coast Line Co.,
211 U.S. 210, 226, 29 S.Ct. 67, 69, 53 L.Ed.2d 150 (1908):
A judicial inquiry investigates, declares, and enforces liabilities as they stand on present or past facts and under laws supposed already to exist. That is its purpose and end. Legislation, on the other hand, looks to the future and changes existing conditions by making a new rule, to be applied thereafter to all or some part of those subject to its power.
Although the Secretary is endeavoring to charge Western Union with overtime for its managerial employees, it cannot do so until it is first settled as to what time frame primary duty will be measured under the short tests. Thus, this suit at the present time does not attempt to resolve a particular controversy based on the law as it presently exists, but rather seeks to have this court enunciate a new rule which will then be applied for the first time to Western Union. The Secretary has never before proffered the workweek standard as a measurement for primary duty under the regulations. We are not asked at this time to apply the primary duty test of the short test provisos to Western Union employees performing struck-work. Rather, we are asked to determine
how
primary duty is to be measured. This is in essence a legislative determination which we believe only the Secretary has the authority to pursue through the designated procedures delegated by Congress.
Our conviction that the Secretary is in effect offering the workweek standard as a substantive amendment to the regulations is underscored by the Secretary’s position that if the workweek standard is approved, it will be applied outside the narrow context of this case,
i. e.,
a strike. In the Secretary’s view, recognition of the workweek standard will affect all short test managerial employees regardless of the context in which they perform nonexempt work. Thus, each employer will have to evaluate on a weekly basis whether a short test managerial employee who works more than 40 hours per week and performs non
exempt work, has management as his primary duty. If the employer determines that the employee’s primary duty was not managerial, overtime will have to be paid in accordance with FLSA.
• The Secretary’s proposed utilization of the workweek standard would have wide ranging repercussions throughout the industry and commerce of the nation and would significantly alter current practices under the Act. For example, high-salaried managerial employees do not have rigidly defined job duties nor are they required to keep records on a daily or weekly basis of their time spent performing a given task. The incorporation of the workweek standard into the short test provisos would impose additional recordkeeping burdens on both management and managerial employees, because then the amount of time spent performing managerial tasks would become important in determining primary duty.
More importantly, the entire concept of exempt status for short test managerial employees would take on new character. Theoretically, the exempt status of a high-salaried employee could shift from week to week. This might require a significant readjustment of workloads, payrolls and costs on the part of management in order to gauge how much overtime is economically feasible.
The foregoing demonstrates to us that the Secretary is attempting to utilize the instant case to promulgate a rule of widespread effect, not only as to employers confronted with work stoppages, but for all industry and commerce. Indeed, the Administrative Procedure Act defines a “rule” as “the whole or part of any agency statement of general or particular applicability and future effect designed to implement, interpret, or prescribe law or policy . . .” 5 U.S.C. § 551(4) (1976). Merely because one employer, Western Union, is the target for application of the workweek standard, does not lessen the nature of the essentially legislative action proposed by the Secretary.
See Hoffmann-LaRoche, supra,
478 F.2d at 13 (although administrative adjudicatory proceedings involved concrete factual situation, facts and inferences were used to formulate legislative-type judgment of prospective operation). When an administrative agency attempts to exercise a legislative function delegated to it by Congress, it must do so in compliance with the provision of the APA, 5 U.S.C. § 553. Section 553 requires,
inter alia,
that general notice of the proposed rule be published in the Federal Register, 5 U.S.C. § 553(b) and interested persons be given an opportunity to comment on the proposed rule, 5 U.S.C. § 553(e). The purpose of rulemaking procedure under the APA was succinctly stated by Judge Van Dusen in
Texaco, Inc. v. Federal Power Commission,
412 F.2d 740, 744 (3d Cir. 1969): “Section 553 was enacted to give the public an opportunity to participate in the rule-making process. It also enables the agency promulgating the rule to educate itself before establishing rules and procedures which have a substantial impact on those regulated. . . . ”
We believe a rulemaking procedure is particularly appropriate in this case because of the widespread ramifications that measurement of primary duty on a workweek basis under the short test provisos will have. Such ramifications implicate factual matters which are appropriate for elucidation and debate in a rulemaking procedure. For example, we believe the desirability of the workweek standard must be assessed in the context of any recordkeeping and related burdens imposed on employers for managerial employees as well as the economic impact that might result from a chameleonic exempt status that might occur in a workweek time frame. Also, whether particular managerial employees may have lost their exempt status from week to week because of extended hours of nonexempt work is a factual question which involves considerations of the nature of the employee’s work and his salary level. We believe that labor, management and other interested parties should have an opportunity to present their views on the desirability and feasibility of measuring primary duty for short test managerial employees in a workweek time frame. This will enable the Secretary to make an informed determination of what standard is appropriate for the short test provisos.
V.
In this court’s earlier decision, we determined that exempt status could be lost through the performance of nonexempt work, but left the task of determining whether primary duty should be measured on a workweek basis to the district court. Our conclusion that the adoption of any particular time frame for evaluating primary duty under the short test provisos involves the district court in rulemaking requires us to reverse the district court’s order interpreting primary duty on a workweek basis. There is, however, a continuing injunction against Western Union requiring it to pay overtime to those managerial employees who perform nonexempt work. But since it will be impossible, until a rule is promulgated, to determine
when
exempt status is lost under the short tests, we direct the district court to dissolve the injunction, without prejudice to whatever rights the parties may have upon conclusion of the rulemaking procedures.
The judgment of the district court will be reversed and the case remanded for further proceedings not inconsistent with this opinion. Each side to bear its own costs.