dissenting:
The majority extends its magic wand over the parties and waives appellant’s obligation to exhaust its administrative remedies. I believe that the court lacks the unbridled discretion it asserts and, in any event, that appellant is not entitled to such favored treatment.
Title VII of the Civil Rights Act, passed 35 years ago, was amended in 1972 to give the EEOC the exact investigative powers previously conferred 52 years ago on the NLRB. Accordingly, 29 U.S.C. § 161(1) applies to subpoenas issued as part of the investigative process by both the NLRB and the EEOC. Although it is set forth in the majority opinion, I quote it again:
Within five days after the service of a subpena [sic] on any person requiring the production of any evidence in his possession or under his control, such person may petition the Board to revoke, and the Board shall revoke, such subpena [sic] if in its opinion the evidence whose production is required does not relate to any matter under investigation, or any matter in question in such proceedings, or if in its opinion such subpena [sic] does not describe with sufficient particularity the evidence whose production is required.
The NLRB issued a regulation interpreting or implementing that statutory provision in 1947. See 12 Fed.Reg. 5657, 5660 (1947). The current version of the regulation provides that “[a]ny person served with a subpoena ..., if he or she does not intend to comply with the subpoena, shall, within 5 days after the date of service of the subpoena, petition [the regional director, or if during the hearing, the administrative law judge] in writing to revoke the subpoena.” 29 C.F.R. § 102.31(b) (1999).
Not surprisingly in 1972, the same year Title VII was amended to give the EEOC the investigatory powers of the NLRB, the EEOC issued a virtually identical regulation, see 37 Fed.Reg. 9218 (1972), the current version of which provides that “[a]ny person served with a subpoena who intends not to comply shall petition the issuing Director or petition the General Counsel, if the subpoena is issued by a Commissioner, to seek its revocation or modification. Petitions must be mailed to the Director or General Counsel, as appropriate, within five days ... after service of the subpoena.” 29 C.F.R. § 1601.16(b)(1) (1999).
The appellant makes much of the distinction between the statute’s “may” and the regulation’s “shall,” but I think that is a tempest in a teapot. The statute could not have used “shall” because it does not include the qualifying phrase “person ... who intends not to comply”; it speaks to everyone served with a subpoena. Obviously, it would make no sense for Congress to tell someone who is served and has no objection to producing information that he or she shall petition to revoke. The most reasonable interpretation of the statute, accordingly, is that it is equivalent to the regulation — that the process is obligatory if one wishes to object to a subpoena. There can be no other reason why Congress imposed a five-day limitation.1 It certainly would have been senseless for Congress to provide that a party may petition within five days, but that nothing is to stop the party from petitioning after the five days. The permissive “may” was used, not to create a five-day period that any party can disregard at will, but merely to make clear — quite sensibly — that objee-*972tions are optional. And if a party who objects to a subpoena must petition the agency to revoke within five days, it follows that the exhaustion requirement is mandatory. The majority would do well to bear in mind that the provision in question is part of the NLRB’s (and the EEOC’s) investigative powers, and it is certainly understandable that Congress should wish such agencies to be armed against efforts to frustrate and delay their investigation. If an employer against.whom such a subpoena is issued could simply ignore the agency or say, “I’ll see you in court,” the investigative process would be significantly impaired.
If the statute is to be honored, the district court should not entertain, in a subpoena enforcement proceeding, a respondent’s objection that should have been raised first before the agency.2 That is simply another way of saying the exhaustion requirement is mandatory. See McCarthy v. Madigan, 503 U.S. 140, 144, 112 S.Ct. 1081, 117 L.Ed.2d 291 (1992) (“Where Congress specifically mandates, exhaustion is required.”) (emphasis added). To be sure, the EEOC did not argue that the statute created a jurisdictional bar to the district court’s consideration of appellant’s claim, unquestionably because § 161(1) is not phrased in terms of jurisdiction. Cf. Weinberger v. Salfi, 422 U.S. 749, 756, 95 S.Ct. 2457, 45 L.Ed.2d 522 (construing a statutory exhaustion requirement that operates by divesting district courts of jurisdiction except in cases of final agency decisions). And we have said that exhaustion is a jurisdictional prerequisite “[o]nly when Congress states in clear, unequivocal terms that the judiciary is barred from hearing an action until the administrative agency has come to a decision,” I.A.M. National Pension Fund Benefit Plan C v. Stockton Tri Indus., 727 F.2d 1204, 1208 (D.C.Cir.1984) — a requirement that § 161(1) may well not meet (despite its obvious implication). But although it does not speak to our jurisdiction, the statute seems to me to impose a mandatory obligation on targets of subpoenas.
The majority not only conflates a statutory provision limiting our jurisdiction with a statute or regulation governing parties’ behavior, it also ignores the distinction between a mandatory exhaustion requirement created by statute or regulation, and the judicially-created common law doctrine of exhaustion. See Weinberger, 422 U.S. at 765-66, 95 S.Ct. 2457 (distinguishing between “statutorily specified jurisdictional prerequisite[s]” and “the judicially developed doctrine of exhaustion”). The latter is a judge-made rule that admits of various prudential exceptions; the former is not. See I.A.M. National Pension Fund, 727 F.2d at 1208; Glisson v. United States Forest Service, 55 F.3d 1325, 1327 (7th Cir.1995) (“But to the extent that [exhaustion] is a doctrine of federal common law rather than the inflexible command of a statute, it is to be applied with due regard for its underlying purpose and for considerations that may in particular cases counsel for a waiver.”). It is relatively open to us to relax the rigors of exhaustion doctrines we ourselves have created, but a statutory or authorized regulatory command is entitled to more respect — even if not phrased in judicial jurisdictional terms. It may well be (there are really no cases in point) that a non-jurisdictional but mandatory exhaustion requirement allows a court to consider certain traditional limited exceptions such as futility or agency bias. But to call the exhaustion requirement non-jurisdictional — the majority’s extensive argument to that effect is really a red herring — is not to allow a court to treat it as if it had no legal bite. The majority seems to be under the impression that it has no obligation to respect agency rules unless we are told by Congress that we lack jurisdiction to *973do otherwise. In my view, this case does not implicate the court’s basic authority, still less its jurisdiction, but rather illustrates the more familiar, but nevertheless distressing, misuse of judicial power to override rather than defer to a reasonable agency rule. We are “prohibited,” as the majority puts it, from hearing appellant’s argument not because we lack power to do so but because we are a court of law.
Even if the statute itself were thought ambiguous as applied to this case, the NLRB-EEOC regulations — interpretations of § 161(1) that deserve deference, see Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984) — unquestionably make exhaustion obligatory and clearly cover all potential defenses such as attorney-client privilege. The Supreme Court has recognized that agencies may promulgate regulations mandating exhaustion which are to be enforced by courts, see Darby v. Cisneros, 509 U.S. 137, 154, 113 S.Ct. 2539, 125 L.Ed.2d 113 (1993), and surely a reasonable regulatory interpretation or elaboration of a statutory exhaustion requirement qualifies as such. I do not see how the majority justifies ignoring this regulation’s obligatory nature. That is equivalent to holding the mandatory exhaustion regulation unenforceable and replacing it with a judge-made, equity-inspired, permissive doctrine. As this court asked once before: “If an agency rule requires, without exception, that a party must take an administrative appeal before petitioning for judicial review, on what basis may a court excuse non-compliance?” Marine Mammal Conservancy, Inc. v. Department of Agric., 134 F.3d 409, 411 (D.C.Cir.1998). My answer is: certainly not the basis made up to fit this case.
The majority sideslips Darby’s recognition that agencies may impose regulatory exhaustion requirements by limiting Darby to situations where the agency’s regulatory exhaustion requirements are authorized by statutes that speak in jurisdictional terms. But Darby expressed no such limitation; the Darby Court was not even concerned with limiting the power of agencies to create exhaustion requirements; it was concerned with limiting the power of courts to do so. Nor does Darby confuse (as does the majority) the concept of mandatory exhaustion requirements imposed on parties with statutory provisions that strip courts of jurisdiction. Thus the Court concluded that § 10(c) of the APA “has limited the availability of the doctrine of exhaustion of administrative remedies to that which the statute or rule clearly mandates.” 509 U.S. at 146, 113 S.Ct. 2539 (emphasis added).
Rather inconsistently the majority insists that it shows no disrespect for the agency’s rule. Instead, it is simply correcting the agency’s “error.” See Maj. Op. at 964, quoting United States v. L.A. Tucker Truck Lines, Inc., 344 U.S. 33, 37, 73 S.Ct. 67, 97 L.Ed. 54 (1952) (“[Cjourts should not topple over administrative decisions unless the administrative body ... has erred.”). But the majority never identifies the error it is targeting, and it should be obvious that it simply does not like the rule.
Even assuming the exhaustion requirement were not mandatory, in which case we could freely “balance the interest of the individual in retaining prompt access to a federal judicial forum against countervailing institutional interests favoring exhaustion,” McCarthy, 503 U.S. at 146, 112 S.Ct. 1081, I do not think it would be appropriate to excuse Lutheran’s default. Since Lutheran has not claimed one of the recognized exceptions to the exhaustion requirement, see id. at 146-49, 112 S.Ct. 1081— the majority takes a different tack. It reasons instead that, since the issue of the existence or scope of attorney-client and work product privilege is not one that is committed to agency “expertise,” there is no basis for enforcing the exhaustion requirement here. That conclusion rests on the erroneous premise that the only real purpose of an exhaustion requirement is *974“to allow an administrative agency to perform functions within its special competence.” Parisi v. Davidson, 405 U.S. 34, 37, 92 S.Ct. 815, 31 L.Ed.2d 17 (1972). But the leading exhaustion case cited by Parisi, quoted by the majority, is McKart v. United States, 395 U.S. 185, 89 S.Ct. 1657, 23 L.Ed.2d 194 (1969), and McKart spoke of exhaustion as having various purposes, including enabling the agency to create a factual record, to apply its expertise, and to exercise its discretion. See id. at 194, 89 S.Ct. 1657 (observing that the exhaustion doctrine’s furtherance of executive and administrative autonomy is “particularly pertinent where the function of the agency and the particular decision sought to be reviewed involve exercise of discretionary powers” or application of special expertise). More recently in McCarthy, the Court again said that “[exhaustion concerns apply with particular force when the action under review involves exercise of the agency’s discretionary power or when the agency proceedings in question allow the agency to apply its special expertise.” McCarthy, 503 U.S. at 145, 112 S.Ct. 1081 (emphasis added).
When a target of an NLRB or EEOC subpoena formally asserts an attorney-client or work product privilege pursuant to the regulation, we should expect, and EEOC’s counsel confirmed at oral argument, that the matter is escalated above the litigation attorney to a more senior official of the agency — perhaps the general counsel himself. And, as we should also expect, that might well lead to a modification (or even abandonment) of the agency’s subpoena. It is hard tu imagine any agency decisions more laced with discretion than such investigative and prosecutorial determinations. Cf. Heckler v. Chaney, 470 U.S. 821, 830-35, 105 S.Ct. 1649, 84 L.Ed.2d 714 (1985) (discussing discretionary nature of agency decisions not to undertake enforcement action). That we would not defer to such a decision, once made, does not detract at all from the importance of permitting the agency to resolve the question in the first instance. I cannot understand the majority’s disposition to dismiss this sort of agency discretion as beneath our notice, let alone respect.
Understandably reluctant to be specific about the precise nature of the sweeping discretionary power it arrogates to itself, the majority does not make reference to doctrines of equity. Yet the judicial posture the majority assumes goes far beyond even the balancing of institutional and individual interests that traditionally accompanies non-mandatory common law exhaustion. Its judgment seems ultimately to rest on little more than its determination that the “combined circumstances of this case” make it “unfair” to hold that Lutheran has forfeited its privilege claim, which I suppose translates into a new doctrine of judicial will, i.e., it pleases us not to support the EEOC’s, exhaustion rule.
But since the provision of Title VII and the EEOC regulation at issue in this case simply followed the NLRA and Labor Board regulation, § 1601.16(b)(1) perforce must be interpreted and applied just as we would treat the NLRB’s counterpart regulation (and vice versa). And the Board’s exhaustion requirement has long been enforced by the federal courts, without any reference to the majority’s broad notions of “equity.” See, e.g., Maurice v. NLRB, 691 F.2d 182, 183 (4th Cir.1982); Hedison Mfg. Co. v. NLRB, 643 F.2d 32, 34 (1st Cir.1981); NLRB v. Frederick Cowan & Co., 522 F.2d 26, 28 (2d Cir.1975). In the only circuit case on point, the counterpart EEOC exhaustion requirement was similarly enforced. See EEOC v. Cuzzens of Georgia, Inc., 608 F.2d 1062, 1063-64 (5th Cir.1979) (per curiam); see also EEOC v. County of Hennepin, 623 F.Supp. 29, 31-32 (D.Minn.1985); EEOC v. Roadway Express, Inc., 569 F.Supp. 1526, 1528-29 (N.D.Ind.1983). The majority would distinguish these cases, pointing out that Cowan and Hedison, for example, involved a level of party misconduct not found here. Yet in all of the cases, courts *975refused to entertain objections to subpoenas that had not been presented to the issuing agency; in none of them was there a suggestion that open-ended equitable balancing was appropriate or even permissible. This case then is the first serious challenge to important Labor Board and EEOC regulations that go back several decades. The majority’s superficially narrow holding — forgiving Lutheran’s failure to exhaust given the “combined circumstances of this case,” Maj. Op. at 966 should not obscure the far-reaching consequences of its decision. Every case, after all, has its circumstances, and every chancellor’s foot a different length. See generally Antonin Scalia, The Rule of Law as a Law of Rules, 56 U. Chi. L. Rev. 1175 (1989).3
Finally, even assuming we had the extraordinarily broad equitable discretion the majority claims, I believe there is no reason for invoking it in this case. Traditional equitable doctrines do not help Lutheran here. The EEOC has not waived its exhaustion argument, nor would tolling be of much use since Lutheran never made its privilege objection in writing at any time4 (as opposed to objecting after the five days passed), see Jones v. Runyon, 91 F.3d 1398, 1400 n. 1 (10th Cir.1996) (distinguishing the requirement of a timely EEOC filing from the requirement of an EEOC filing). And although the majority suggests that agency officials misled Lutheran into thinking that it was not obligated to comply with the regulation— which I take to be an estoppel notion though the majority does not say so — the majority bases this suggestion on Lutheran’s lawyer’s assertion that an EEOC investigator told him that Lutheran would be notified before the agency took any action. It does not seem, to matter that this conversation took place well after the five-day period had elapsed, or that the EEOC investigator does not recall making the statement.
A careful parsing of the majority’s opinion reveals that the only real ground upon which my colleagues rely to tilt the equities in favor of appellant is the one they “begin” and end with. See Maj. Op. at 964-65.5 That is, the subpoena stated, that “it is issued pursuant [to] 42 U.S.C. § 2000e-9,” but did not explicitly alert appellant’s Human Resources Director, to whom the subpoena was sent, that under EEOC’s regulation (and for that matter the statutory provision that § 2000e-9 incorporates), appellant had five days to object formally.
But it is well settled that inaccurate or ineffective notice from a government agency is an excuse for non-compliance with an EEOC time limit only when the agency is “required to provide notice of the limitations period.” Bowden v. United States, 106 F.3d 433, 438 (D.C.Cir.1997). Like the *976regulation at issue in Bowden, 29 C.F.R. § 1601.16(b)(1) does not require that the agency provide notice of the limitations period for objections. As we held in Bow-den, even if we thought it would be “sensible and simple” for the EEOC to list the regulation on the face of the subpoena, “the agency had no duty to do so” and thus Lutheran’s failure to comply with the regulations cannot be excused on that ground. Id.
As for the supposed unfairness to Lutheran in not receiving notice of a federal regulation, I do not know whether to laugh or cry. Any lawyer experienced in the employment law field, of course, would be familiar with the regulation, but I think we should assume that any competent lawyer receiving such a subpoena would spend the ten minutes necessary to determine the applicable law. (The majority is willing to assume that Lutheran’s Human Resources Director could locate § 161(1) of the Labor Act from the cross-reference in Title VII, but thinks it requires some great feat of legal ingenuity to locate the applicable regulation.) Even if the Human Resources Director had not sent the subpoena to Lutheran’s counsel in time (as the majority observes, the record is silent on this issue), that would be no excuse. In today’s world, any person in such a job who did not consult counsel immediately would be guilty of gross negligence. Perhaps the import of the majority’s opinion is that the NLRB and the EEOC, if they wish their exhaustion regulations honored, will have to give the recipients the administrative law equivalent of a Miranda warning, including a list of counsel who have shown competence in employment law.
The majority’s last point is that we should excuse Lutheran’s failure to exhaust because voluntary compliance with Title VII is an important value, and because that value depends on safeguarding the attorney-client and work product privilege. That conflates the merits of Lutheran’s privilege claim with the need for an exhaustion requirement. The Congress and the EEOC, which administers this investigatory regime, believes that a mandatory exhaustion requirement is a necessary and useful component of it. Though voluntary compliance is an important value, and overzealous agency investigation without regard to claims of privilege could undercut that value, it is for the agency, not for this court, to strike the balance. Why the majority cannot see (thinking ex ante rather than ex post) that its holding will actually undermine voluntary compliance — by encouraging subpoena recipients to flout agency regulations and to gamble on garnering equitable relief from a court — is beyond me. See generally Frank H. East-erbrook, Foreword: The Court and the Economic System, 98 HaRV. L. Rev. 4, 10-12 (1984). In any event, I can accept the majority’s policy preference no more than I can subscribe to its assertion of discretionary power. I dissent.
. It is rather misleading to ask as the majority does whether the appellant has "waived” its attorney-client privilege; the question is better phrased as whether the appellant forfeited its claim by not raising it in a timely fashion before the agency.
. Judge Tatel, dissenting in In Re Sealed Case, 124 F.3d 230, 239-40 (D.C.Cir.1997), powerfully argued that a loosey-goosey balancing test as applied to attorney-client privilege issues was inappropriate. The Supreme Court agreed. See Swidler & Berlin v. United States, 524 U.S. 399, 118 S.Ct. 2081, 141 L.Ed.2d 379 (1998). In that case, the clear rule favored private lawyers, whereas here the loo-sey-goosey approach adopted by the majority apparently bails out some private lawyers.
. As the majority indicates, Lutheran did inform the EEOC by letter that it thought the subpoena "improper” and that it did not intend to comply. But as I read the regulation, that letter, in addition to being untimely, was fatally deficient not only because it was not sent to the Director who issued the subpoena, but more importantly because the regulation requires the complaining party to state specifically the grounds of non-compliance, see 29 C.F.R. § 1601.16(b)(2).
.The majority thinks it also significant that Lutheran orally informed an EEOC investigator (not the District Director) of its objections and that the District Director was (the majority infers) aware of Lutheran's objections. Apparently my colleagues are willing to excuse a party's non-compliance with agency regulations when the party comes "close enough”— in this case, informal conversations with the wrong person that took place before the subpoena ever issued. That is, to say the least, a rule of administrative law of which I am not aware.