Nos. 1022, 97-7783United States Court of Appeals, Second Circuit.Argued: January 28, 1998
Decided: April 30, 1998
Appeal from a judgment entered June 5, 1997, in the United States District Court for the Southern District of New York, Peter K. Leisure, Judge, granting partial summary judgment in favor of appellees.
The judgment is vacated and this case is remanded.
[EDITORS’ NOTE: THIS PAGE CONTAINS HEADNOTES. HEADNOTES ARE NOT AN OFFICIAL PRODUCT OF THE COURT, THEREFORE THEY ARE NOT DISPLAYED.]
Ellen B. Fishman, Office of Corporation Counsel of the City of New York, New York, New York, (Paul A. Crotty, Corporation Counsel of the City of New York, Leonard Koerner, and Alan M. Schlesinger, of counsel) for Appellants.
Joan Stern Kiok, New York, New York, (Robert N. Felix, New York, New York, of counsel) for Appellees.
Before: CALABRESI, CABRANES, and HEANEY, Circuit Judges.
HEANEY, Senior Circuit Judge:
 The City of New York (City) and the City Department of Juvenile Justice (DJJ) appeal a judgment of the United States District Court for the Southern District of New York (Peter K. Leisure, Judge) granting summary judgment in favor of DJJ employees on the issue of liability under the Fair Labor Standards Act (FLSA) of 1938, as amended, 29 U.S.C. §§ 201-219. Appellants contend that the application of the FLSA violates the Tenth Amendment, the employees are exempt from the FLSA, and that the appellants complied with the FLSA. Because the record is not sufficiently clear regarding certain practices under the employment contract, we vacate the district court’s judgment and remand this case for further proceedings.
 Appellees, current or former employees of the DJJ, worked as Houseparents or Senior Houseparents in non-secure detention facilities for children under the age of sixteen. Prior to April 1, 1986, two types of Houseparents were employed by the City. The first was employed by the City Department of Social Services (DSS) and regularly worked a 120-hour workweek, comprised of five 24-hour shifts. The second was employed by DJJ and regularly worked a sixty-hour workweek, comprised of five 12-hour shifts . Prior to April 1, 1986, Houseparents were paid the same annual salary regardless of whether they worked the DSS 120-hour workweek or the DJJ sixty-hour workweek. Obviously, the effective hourly rate differed greatly between DSS and DJJ Houseparents.
 The Houseparents were represented by Local 371, District Council 37, American Federation of State, County and Municipal
Employees, AFL-CIO (Union). On April 15, 1986, the FLSA was made applicable to public employers. In order to address the disparity between the workweeks of DSS and DJJ Houseparents and to conform to the newly applicable FLSA, the Union and the City entered into arbitration to reform their employment contract.
 The Union took the position that both DSS and DJJ Houseparents should work forty-hour workweeks at then current salary levels. Citing lower prevailing wages in the private sector, the City suggested that Houseparents work 120-hour workweeks at then current salary levels. On February 24, 1987, the arbitrator issued an award that set sixty-hour workweeks for all Houseparents and maintained pre-FLSA annual salaries. In arriving at his decision, the arbitrator was concerned that the 120-hour DSS workweek, relieved only by a forty-eight-hour respite, threatened the health of the Houseparents and the children in their charge. (Arbitrator Op. and Award at 13-14.) On the other hand, the arbitrator pointed out that the Union had recently received a substantial wage increase and Houseparents would receive a significant windfall if he were to adopt the Union’s position in its entirety.
 The sixty-hour workweek and pre-FLSA annual salaries were continued in subsequent collective bargaining agreements that also provided for percentage wage increases. Under the new arrangement, the Houseparents’ hourly rate was calculated as follows: first, the annual salary for each Houseparent was divided by 52.2 to get the weekly salary. Next, the weekly salary was divided by seventy to get the hourly rate. The figure seventy was used to show that the first forty hours of work were compensated at straight time and the last twenty hours of work were compensated at time and one-half. The twenty hours of overtime were, therefore, expressible as thirty hours at straight time.
 On November 22, 1993, a group of DJJ Houseparents brought suit against the City and DJJ for money damages and declaratory judgment, claiming that the hourly wage computation established by the arbitrator and subsequently included in collective bargaining agreements violated the FLSA. In June 1995, the Houseparents moved for partial summary judgment as to whether they received the correct overtime rate under the FLSA. The City and DJJ cross-moved to dismiss the claim. The district court granted the Houseparents’ motion and found that the regular rate calculated under the employment contract failed to comply with the FLSA.
 In characterizing the arbitrator’s award, the district court said that “the arbitrator derived the regular hourly rate using the postulate that the annual salary must not change, and any subsequently derived hourly rate, based on this arbitrator’s regularly hourly rate, is founded on the postulate that the annual salary must not change.” (Op. and Order at 4.) The district court disagreed with the arbitrator’s decision and held that:
 The only reasonable interpretation of the pay scheme of DJJ Houseparents and Senior Houseparents prior to FLSA applicability is that they were paid a fixed annual salary based on a 60 hour workweek, and that this salary should be divided by the number of hours actually worked to derive the regular hourly rate. To derive the regular hourly rate using the arbitrator’s postulate assumes that prior to FLSA applicability the DJJ Houseparents and Senior Houseparents were paid a premium of time and one-half for overtime work. This assumption is unreasonable in light of the
Congressional purpose to change the status quo ante of freedom of contract. Thus, the arbitrator’s postulate circumvents the statutory scheme, as do later contracts based on the arbitrator’s postulate. Accordingly, the Court finds that the computation of the regular hourly rate in those contracts is erroneous and violates the FLSA.
 (Id. at 7.) The City appeals.
 We review a grant of summary judgment de novo. Kulak v. City of New York, 88 F.3d 63, 71 (2d Cir. 1996). Summary judgment is appropriate when there is no genuine issue of material fact and the moving party is entitled to judgment as a matter of law. Id. at 70. The party moving for summary judgment has the burden of showing the absence of a genuine issue of material fact. Adickes v. S. H. Kress Co., 398 U.S. 144, 157 (1970). In examining this question, the court resolves all ambiguities and draws all factual inferences in favor of the nonmoving party. See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255 (1986).
 Section 207(a)(1) of the FLSA provides, in relevant part: Except as otherwise provided in this section, no employer shall employ any of his employees . . . for a workweek longer than forty hours unless such employee receives compensation for his employment in excess of the hours above specified at a rate not less than one and one-half times the regular rate at which he is employed.
 29 U.S.C. § 207(a)(1). This case requires us to answer the following questions: (1) does the application of the FLSA to the appellants violate the Tenth Amendment; (2) are the Houseparents exempt from the FLSA; (3) could the parties maintain pre-FLSA wages once the FLSA became applicable; and (4) does what in fact occurred under the employment contract comply with the FLSA?
 With regard to the Tenth Amendment, the appellants argue that because New York has a highly regulated system of collective bargaining and impasse procedures, it is unconstitutional to apply the FLSA in this case. In support of their position, appellants contend that the rule in Garcia v. San Antonio Metropolitan Transit Authority, 469 U.S. 528 (1985), would no longer command a majority of the United States Supreme Court. We are bound by this precedent and decline to engage in the speculation suggested by the appellants. Moreover, the Tenth Amendment is not implicated because appellants have failed to show that New York was incapable of participating “`in the national political process or that it was singled out in a way that left it politically isolated and powerless.'” See Reich v. State of New York, 3 F.3d 581, 589 (2d Cir. 1993) (quoting South Carolina v. Baker, 485 U.S. 505, 513 (1988)). Additionally, there is no reason to believe that applying the FLSA to this case “`would upset the usual constitutional balance of federal and state powers.'” Reich, 3 F.3d at 589 (quoting Gregory v. Ashcroft, 501 U.S. 452, 460 (1991)). Consequently, we reject appellants’ Tenth Amendment challenge to the FLSA’s applicability in this case.
 Appellants next contend that Houseparents are subject to either of two FLSA exemptions. We note from the outset that the appellants bear the burden of proving that Houseparents fall within an FLSA exemption and we do not give such exemptions generous application. Reich, 3 F.3d at 586-87 (citations omitted). First, with limited exceptions not applicable in this case, “any employee employed in a bona fide executive, administrative, or professional capacity” is exempt from the overtime provisions of the FLSA. 29 U.S.C. § 213(a)(1). The appellants claim that “Senior Houseparents” fall under this exemption because they perform evaluations for subordinate Houseparents and other supervisory tasks. The district court rejected this argument and, citing 29 C.F.R. § 541.115, found that Senior Houseparents were more like “working” supervisors because they spent most of their time performing the same tasks as Houseparents. Title 29 C.F.R. § 541.115(b) provides:
 Such employees, sometimes known as strawbosses, or gang or group leaders perform the same kind of work as that performed by their subordinates, and also carry on supervisory functions. Clearly, the work of the same nature as that performed by the employees’ subordinates must be counted as nonexempt work and if the amount of such work performed is substantial the exemption does not apply.
 Id. We see no reason to upset the district court’s ruling that Senior Houseparents performed essentially the same work as Houseparents and therefore affirm its determination that Senior Houseparents do not fall under the “executive, administrative, or professional capacity” exemption.
 Second, the appellants contend that the Houseparents fall into the “domestic service” exemption. The district court rejected that contention. Appellants argue that since the FLSA became applicable to public employers, neither Congress nor the Department of Labor has created or expanded exemptions to more appropriately fit the public employer context. They further contend that the district court placed undue weight on a Department of Labor Regulation and Letter Ruling. Title 29 C.F.R. § 552.3 defines “domestic service employment” as “services of a household nature performed by an employee in or about a private home.” Id. The letter ruling, WH-638, dated November 25, 1975, expresses a similar view that the domestic service must take place in a private home or residence to warrant an exemption. (See Jt. App. at A238.) Appellants suggest that the district court’s interpretation effectively excludes all public employees from the ambit of this exemption because they do not work in private homes or residences.
 It is clear that Houseparents did not work in a private home or residence. We note that it has been more than ten years since the FLSA became applicable to public employers, and we assume that Congress and the Department of Labor have affirmatively chosen not to adopt the extension of the domestic service exemption advanced by appellants. Consequently, we reject the appellants’ argument and affirm the district court’s ruling that the Houseparents did not fall within the FLSA’s domestic service exemption.
 As to the question of whether the parties could agree to maintain pre-FLSA wages, its resolution turns on the meaning of “regular rate” and its proper context under the FLSA. The FLSA broadly defines the regular rate as “all remuneration for employment,” 29 U.S.C. § 207(e), and lists certain exceptions that are not relevant here. The Supreme Court has said that the regular rate is “the hourly rate actually paid the employee for the normal, non-overtime workweek for which he is employed.” Walling v. Youngerman-Reynolds Hardwood Co., 325 U.S. 419, 424 (1945) (citation omitted). The Department of Labor has adopted this definition. See 29 C.F.R. § 778.108.
 In order to evaluate an employer’s FLSA compliance, wages must be expressed in terms of a regular hourly rate, which is accomplished by “dividing [the] total remuneration [except statutory exclusions] . . . in any workweek by the total number of hours actually worked.” 29 C.F.R. § 778.109. While a mechanical application of this rule to the present case might support the district court’s reading of the FLSA, the regulations specifically address the situation where an employee’s wage is expressed in non-hourly terms: “Where the salary covers a period longer than a workweek, . . . it must be reduced to its workweek equivalent.” Id. § 778.113(b). Once this is done, the regular rate “is computed by dividing the salary by the number of hours which the salary is intended to compensate.” Id. § 778.113(a) (emphasis added). This regulation does not, however, address the situation of an employee who works more than forty hours a week
and whose base salary may already include an overtime premium.
 Where, as here, an employee regularly works the same amount of overtime, “it is, of course, proper to pay him, in addition to his compensation for nonovertime hours, a fixed sum in any such week for his overtime work, determined by multiplying his overtime rate by the number of overtime hours regularly worked.” Id. § 778.309. “As in all cases of employees hired on a salary basis, the regular rate depends in part on the agreement of the parties as to what the salary is intended to compensate.” Id. § 778.323 (emphasis added). While an agreement between the parties is instructive, the Supreme Court has held that it is not necessarily determinative where there is a sham wage agreement. 149 Madison Ave. Corp. v. Asselta, 331 U.S. 199, 204, modified on other grounds, 331 U.S. 795 (1947). Similarly, 29 C.F.R. § 778.108 provides that an employee’s regular rate of pay “cannot be left to a declaration by the parties as to what is to be treated as the regular rate for an employee [but] must be drawn from what happens under the employment contract. The Supreme Court has described it as the hourly rate actually paid the employee for the normal, nonovertime workweek for which he is employed.” Id. (citations omitted). These authorities do not resolve whether the salary of an employee who regularly works more than forty hours per week should or should not be presumed to include an overtime premium.
 In regard to overtime compensation, 29 C.F.R. § 778.202(a) provides that:
 Many employment contracts provide for the payment of overtime compensation for hours worked in excess of 8 per day or 40 per week. . . . If the payment of such contract overtime compensation is in fact contingent upon the employee’s having worked in excess of 8 hours in a day or in excess of the [statutory maximum per week], the extra premium compensation paid for the excess hours is excludable from the regular rate under section 7(e)(5) and may be credited toward statutory overtime payments pursuant to section 7(h) of the Act.
 Id.; see also 29 U.S.C. § 207(e)(5) (providing for exclusion of overtime premiums in calculating the regular rate); Bay Ridge Operating Co. v. Aaron, 334 U.S. 446, 465-71 (1948) (same). Consequently, if the pay rate set under the contract already includes an overtime premium for excess hours, that premium is not counted in determining the base rate. In other words, employees are not entitled to “double” overtime. Therefore, if the annual salary was properly intended by the parties to account for both a regular rate and an overtime rate, the contemplated arrangement is in compliance with the FLSA.
 As a general rule, nothing in the FLSA prevents an employer from contracting with its employees to pay them the same total wages received prior to FLSA applicability, so long as the regular rate equals or exceeds the minimum wage. Walling v. Belo Corp., 316 U.S. 624, 630 (1942). In fact, as long as the minimum wage is respected and the agreed-upon rate is not a sham, “the employer and employee are free to establish this regular rate at any point and in any manner they see fit.” Youngerman-Reynolds, 325 U.S. at 424. Therefore,
[a] wage plan is not rendered invalid simply because, instead of stating directly an hourly rate of pay in an amount consistent with the statutory requirements, the parties have seen fit to stipulate a weekly wage inclusive of regular and overtime compensation for a workweek in excess of 40 hours and have provided a formula whereby the appropriate hourly rate may be derived therefrom. The crucial questions . . . are whether the hourly rate derived from the formula . . . was, in fact, the “regular rate” of pay within the statutory meaning and whether the wage agreement under consideration, in fact, made adequate provision for overtime compensation.
 Asselta, 331 U.S. at 204.
 While it is true that the FLSA contemplates setting the regular rate through wage negotiations, a regular rate will not be recognized when it is set
in a wholly unrealistic and artificial manner so as to negate the statutory purposes. Even when wages exceed the minimum prescribed by Congress, the parties to the contract must respect the statutory policy of requiring the employer to pay one and one-half times the regular hourly rate for all hours actually worked in excess of 40.
 Walling v. Helmerich Payne, Inc., 323 U.S. 37, 42 (1944). After all, the regular rate is not
an arbitrary label chosen by the parties; it is an actual fact. Once the parties have decided upon the amount of wages and the mode of payment the determination of the regular rate becomes a matter of mathematical computation, the result of which is unaffected by any designation of a contrary “regular rate” in the wage contracts.
 Youngerman-Reynolds, 325 U.S. at 424-25.
 There can be no doubt that an employer and employee may set out to maintain pre-FLSA wages in order to comply with the FLSA. Based on the above discussion, if (1) the minimum wage is maintained; (2) the hourly rate is in fact the “regular rate” of pay within the statutory meaning and is not calculated in a wholly unrealistic and artificial manner that does not reflect actual practice; and (3) the employment contract in fact made adequate provision for overtime compensation, appellants will have been in compliance with the FLSA. We read the district court’s opinion as granting summary judgment on the grounds that the parties were prohibited from maintaining pre-FLSA wages in an employment contract in order to comply with the FLSA. Because the law supports such employment contracts, we reverse the district court.
 In regard to the final question of whether what actually occurred under the wage contract — as opposed to the mere terms of the contract — reflected an FLSA-sanctioned arrangement, we attempted to resolve the matter at oral argument by asking the parties to submit letter briefs. In those briefs, we specifically asked the parties to address what happened in the event a Houseparent worked less than a full workweek and had no compensatory time remaining in his or her account.
 Appellees’ letter brief was unilluminating as to what the actual practice would be in such circumstances, and appellants likewise asserted that there was nothing in the record to answer our question. However, counsel for appellants stated that based on “conversations with [DJJ], the Office of Labor Relations, and the Office of Payroll Administration,” she had ascertained the following practices. In the event that Houseparent A worked only eleven hours of her regularly scheduled twelve-hour shift, had not accrued any paid leave, and did not make up that hour during the workweek, according to the City, she would have been docked 1/60th of her usual weekly salary — the pro rata share of her weekly salary — not her overtime rate. If, in addition to his regular twelve-hour shift, Houseparent B elected to work an extra hour to cover for Houseparent A, according to the City, he would have been compensated for one hour at the overtime rate. Therefore, when Houseparent A only worked fifty-nine hours, she would have been docked at the pro rata share of her weekly rate. When Houseparent B worked sixty-one hours in a workweek, he would have been credited with one hour at the overtime rate. With this pay adjustment discrepancy in mind, our only remaining inquiry is whether what occurred under this arrangement complies with the FLSA as discussed above. Specifically, we address whether the situation violates the dictates of Asselta, 331 U.S. at 204, and its prohibition of sham wage contracts.
 The facts of Asselta are similar to those here. The initial agreement in Asselta (Sloan Agreement) generally paid employees a weekly rate for a workweek in excess of forty hours and did not distinguish between a regular and overtime rate. Id. at 201. As that agreement approached its expiration, the parties were unable to reach a new agreement and the case was certified to the War Labor Board. Id. at
201-02. The parties entered into a new agreement (Board Agreement) based on the Board’s disposition of the case. Id. at 202. The Board Agreement provided for a workweek of fifty-four hours for watchmen and forty-six hours for regular employees. Id. In Asselta, the employer and employees used essentially the same formula as in the present case to determine the hourly and overtime rates: “The hourly rates for those regularly employed more than forty (40) hours per week shall be determined by dividing their weekly earnings by the number of hours employed plus one-half the number of hours actually employed in excess of forty (40) hours.” Id. (quotation marks omitted). While this may have been an acceptable arrangement had the parties adhered to it, the Supreme Court held that the practice of the parties in Asselta was inconsistent with the agreement. Id.
 In Asselta, employee absences were attributed either to an excusable or inexcusable cause. Id. at 203. If it was an excusable cause, the employee was paid at the formula rate with the provision that six of the hours worked were compensated at the overtime rate, regardless of whether the total hours worked during that week were more or less than forty. Id. If the employee’s absence was inexcusable, he or she was apparently paid according to the formula rate. Id.
 The Board Agreement further provided that all regular employees except watchmen who worked in excess of forty-six hours were to be paid at one and three-quarters the formula rate. Id. In other words, regular employees who worked in excess of forty-six hours received one overtime rate for hours worked between forty and forty-six, and a higher rate for all hours worked in excess of forty-six. Watchmen had a similar dual overtime rate. Id. For all hours worked over fifty-four, watchmen were paid twice the formula rate. Id. Rather than receiving the formula rate, part-time workers’ hourly wages were calculated by dividing the weekly wage for full-time workers by the regular workweek. Id. at 205. Therefore, part-time employees were paid a pro rata share of the regular workweek rather than according to the formula rate. Id.
 The Supreme Court agreed with the employees that the formula rate was not the regular rate. Rather, the regular rate was substantially obtained by dividing the weekly wage by the hours scheduled during that week. This view was bolstered by the fact that part-time workers received a pro rata share of the weekly wage rather than a straightforward calculation of hours worked multiplied by the formula hourly rate. Additionally, the Court pointed to the fact that the hourly rate was not consistently applied. For example, an employee whose absence was excusable could receive overtime compensation when that employee worked less than forty hours in that week. The Supreme Court stated that: “The payment of `overtime’ compensation for non-overtime work raises strong doubt as to the integrity of the hourly rate upon the basis of which the `overtime’ compensation is calculated.” Id. at 205. Also, there was evidence in the Board Agreement that the wages were calculated in such a way as to effectively tie time-and-a-half pay to a forty-six-hour workweek, rather than, as required by the FLSA, a forty-hour workweek. “Overtime over forty-six (46) hours is paid at a rate of time and three-quarters in an effort approximately to equal the overtime to which an employee would ordinarily be entitled, if it were computed on the basis of time and a half after a forty-six (46) hour week.” Id. at 207 (alteration in original) (internal quotation marks and citation omitted). The Court also noted that the payments made to
employees that were meant to represent the retroactive application of the Board Agreement to the period the agreement was being negotiated were “in no sense determined by application of hourly rates derived from the formula.” Id. at 209.
 Considering the Asselta facts and the hypothetical discussed above, the appellants similarly might not have adhered strictly to the formula rate. While Houseparents who worked in excess of sixty hours would have been properly compensated at the overtime rate, those who worked fewer than sixty hours would have been docked at a lower rate — namely, a pro rata share of their weekly salary — than they should have been. For example, if we assume a formula rate of $10.00 per hour, the weekly salary for a Houseparent is [($10.00 x 40 regular hours) + ($15.00 x 20 overtime hours)] or $700.00. To take the above hypothetical, the workweek in which Houseparent A worked only 59 hours, she would have been docked [$700.00/60] or $11.67 rather than the proper amount of $15.00.
 While it is possible that the pay-docking practices described above could reflect that, as in Asselta, the hourly rates established in this case were not genuine regular rates under the FLSA, we lack a sufficient record to decide whether the contracts at issue were sham contracts. The description of docking practices offered by counsel for the City in her letter brief was based on off-the-record conversations, whose source within the relevant offices has not even been identified. Moreover, we have no way of knowing whether the practices described were the product of mere speculation on the part of those spoken to by counsel, or whether the question of pay docking is one that has actually come up and is regularly faced. There is reason to believe that the question of pay docking may be a purely speculative, previously unconsidered one because adjustments to a Houseparent’s accumulated “compensatory time” were apparently the primary mechanism for adjusting a Houseparent’s wages when more or fewer than sixty hours were worked.
 Under these circumstances, we vacate the judgment of the district court and remand for further development of the record on the question of pay docking, which may shed light on whether the terms agreed to by the parties were a sham; at the court’s discretion, of course, it may also allow the introduction of any evidence it would find helpful in determining whether the contract terms genuinely reflected actual practice under the contract. While the district court must decide this matter in the first instance, we make the following observations.
 If the district court determines that no Houseparents were docked their pro rata weekly salary and that compensatory time was the exclusive and proper means by which their salary was adjusted for less than a 60-hour workweek, then the agreement was not a sham and the employment contract complied with the FLSA. Even if the district court determines that Houseparents were docked their pro rata weekly share for less than a 60-hour workweek, the employment relationship — which otherwise, as far as we are aware, does not exhibit any irregularities of the sort found in Asselta — might still be in compliance with the FLSA if in fact pay docking was an inconsequential issue or if the court can fairly say that “the hourly rate derived from the formula here presented was, in fact, the `regular rate’ of pay within the statutory meaning.” Asselta, 331 U.S. at 204 (emphasis added).
 For the aforementioned reasons, we vacate the district court’s judgment and remand the case for proceedings consistent with this judgment.