Usa Group Loan Services, Incorporated, Usa Group Guaranteeservices, Incorporated, and Usa Group Enterprises,incorporated, Doing Business As Usa Group Enrollmentfinancial Aid Services, Plaintiffs-appellants, v. Richard W. Riley, in His Official Capacity As Secretary Ofthe United States Department of Education, and Theunited States Department of Education,defendants-appellees, 82 F.3d 708 (7th Cir. 1996)

The federal government has an enormous program, administered by the Department of Education, of subsidizing student loans. The loans are made by banks but are guaranteed by state and private agencies that have reinsurance contracts with the Department, making it the indirect guarantor of the loans and thus inducing banks to make what would otherwise be risky loans. The proceeds of the loans are used to pay tuition and other expenses; so the colleges and other schools whose students are receiving these loans are also involved in the federal program. Like so many government programs, the student loan program places heavy administrative burdens on the entities involved in it--the lenders, the guarantors, and the institutions. A whole industry of "servicers" has arisen to relieve these entities of some of the administrative burdens. As agents of the educational institutions, the servicers maintain records of the institution's student loans. As agents of the banks, they collect the loans from the students as the loans come due and dun the students when they are slow in paying. As agents of the guarantors, the servicers keep track of defaults and make sure that the banks comply with the various conditions for triggering the guarantees. In any of these roles a servicer who makes a mistake can end up costing the federal government money. If the servicer remits loan moneys to a school for the tuition of a student not eligible for a loan, or fails to pursue a defaulting student, or honors an invalid claim by a bank for reimbursement from a guarantor, federal money is disbursed in violation of the regulations governing the student loan program.

Mistakes and outright fraud by servicers, some resulting in large losses of federal money, led Congress in 1992 to amend Title IV of the Higher Education Act to authorize the Secretary of Education to "prescribe ... regulations applicable to third party servicers (including regulations concerning financial responsibility standards for, and the assessment of liabilities for program violations against, such servicers) to establish minimum standards with respect to sound management and accountability." 20 U.S.C. § 1082(a) (1). See S.Rep. No. 58, 102d Cong., 1st Sess. 2225 (1991). The Secretary has done this, see 34 C.F.R. Parts 668, 682 (1994); Dept. of Education, Student Assistance General Provisions, 59 Fed.Reg. 22348 (Apr. 29, 1994), esp. pp. 22405, 22408-10, and the servicers have brought this suit to invalidate portions of the regulations on substantive and procedural grounds. The district court rejected the challenge, and the servicers appeal.

The challenged provisions make servicers jointly and severally liable with their customers (lenders, guarantors, and institutions) for violations of the statutes, regulations, or contracts governing the student loan program. To be liable, the servicer must itself have violated a statute, regulation, or contract. But it is not a defense that the violation was inadvertent or even that it could not have been avoided at reasonable cost, though given the complexity of the rules and regulations governing the program, and the volume of transactions, mistakes are inevitable even if all due care is used. Although liability is thus strict, as the servicers complain, the regulations use the term "joint and several liability" in a special sense. The usual meaning is that if two or more tortfeasors produce a single injury, the victim can sue any of the tortfeasors for the full amount of his damages and collect that amount from the tortfeasor he has sued. That tortfeasor may or may not have a right to obtain contribution or indemnification--a right to a sharing or shifting of the cost of liability--from the other tortfeasors. But under the challenged regulation the Department may go against a servicer only if unable to collect the overpayment from the servicer's customer. The servicer's liability is thus a back-up liability.

The servicers have led with their weakest argument, indeed an argument so weak as to border on the frivolous. It is that the word "minimum" in the statute means "minimal," so that the regulations are invalid if they impose on the servicers more than the bare minimum of duties that is consistent with the statutory purpose. For this argument they offer no authority, because there is none. A minimum is a floor, not a ceiling. The statute requires the Secretary of Education to establish standards of accountability below which the servicers may not fall but nowhere says or hints that he must set those standards at the lowest possible level. What is true is that if he set them unreasonably high the servicers would have a compelling argument that the standards were invalid, but this argument, which the servicers also make, would owe nothing to the inclusion of the word "minimum" in the statute. This is not to say that the word does no work, only that it does not do the work that the servicers think it does. Had the word been deleted, it would be arguable that the standards set by the Secretary are exclusive, so that no other agency, state or federal, could impose additional standards. Cf. Campbell v. Hussey, 368 U.S. 297, 82 S. Ct. 327, 7 L. Ed. 2d 299 (1961). The standards would be both floor and ceiling. "Minimum" shows they are just a floor. Florida Lime & Avocado Growers, Inc. v. Paul, 373 U.S. 132, 147-48, 83 S. Ct. 1210, 1219-20, 10 L. Ed. 2d 248 (1963).

Is it unreasonable to impose on the servicers the extent of liability that the Secretary's standards impose? A question not mentioned by either party though clearly germane is how much less the servicers' liability would be in the absence of the standards. Perhaps not much less; perhaps no less. It is true that even a negligent mistake in servicing a transaction is unlikely to give rise to common law liability for consequential damages. Evra Corp. v. Swiss Bank Corp., 673 F.2d 951 (7th Cir. 1982); Rardin v. T & D Machine Handling, Inc., 890 F.2d 24 (7th Cir. 1989). A student loan goes sour. The bank that made it asks the guarantor to reimburse it for the loss. The servicer hired by the guarantor to process claims fails, through negligence, to notice that the bank has not complied with all the conditions of the guaranty. As a result of the oversight the money is paid to the bank and the guarantor is reimbursed by the government. The government later discovers the mistaken payment, is unable to obtain repayment from either the bank or the guarantor, and sues the servicer in tort. The suit would be to recover the amount of money that the servicer had mistakenly authorized to be paid to the bank, rather than the much greater amount of damages, classically consequential, resulting from an unforeseeable disruption of the student loan program--the sort of "for want of a nail the kingdom was lost" liability that cases like Evra and Rardin cut off.

The common law liability in the example that we have just given would not be strict in theory, because the government would have to prove that the servicer had been negligent. But it would be likely to be strict in fact, by reason of the doctrine of respondeat superior. A clerk's careless mistake in overlooking the condition that the bank in our hypothetical example had failed to satisfy would be attributed to his employer, the servicer, even if the servicer had done everything it possibly could have done, through care in the hiring, training, and supervision of its employees, to prevent its clerks from being careless. The servicer's liability would be joint and several with the guarantor's and the bank's because joint and several liability is the rule in tort law. The servicer's liability might even be primary, if the bank or guarantor were thought merely to have broken their contract rather than to have committed a tort. As regards shared or primary liability the regulatory scheme that the servicers are challenging does not go as far as the common law, because the scheme makes their liability secondary--so maybe it is here that "minimum" has its bite. The scheme goes further than the common law with regard to the standard of liability, however, since there could be cases in which no one employed by the servicer had been careless--the mistake in authorizing payment might be due entirely to unverifiably false data furnished by the customer--yet the regulation would make the servicers liable, as the common law would not. How many such cases there are likely to be, however, can only be conjectured.

This excursus into tort law has not been a digression. It bears directly on the servicers' argument that the regulatory scheme will actually harm rather than benefit the student loan program in the long run. It will, they argue, force them to raise the price of their services in order to offset the expected cost of the additional liability created by the regulation, and the higher price will eventually be shifted forward in one way or another to the students and the schools. Maybe yes, maybe no. Higher costs of servicers need not in fact translate into significantly higher interest rates for student loans or into other cost increases to students or schools, and this apart from the fact that fees chargeable to students are regulated and interest rates set by statute. 20 U.S.C. § 1077a; 34 C.F.R. § 682.202. Liability will raise the costs of the less careful servicers relative to the more careful, and competition will force the former to swallow much or all of their higher costs or else lose their business to the more careful servicers, whose own costs may be little affected by the increased liability.

And to the extent that the regulation merely codifies, rather than adds to, the servicers' common law liability, there will be no additional costs to the servicers, costs they might be able to shift to their customers or might have to swallow. Probably there will be some additional costs--if there were none, it would be difficult to understand why the servicers are fighting the regulation so tenaciously. The regulatory scheme creates an administrative procedure that is less cumbersome than ordinary civil litigation for enforcing the servicers' liability, and adds fines to the usual tort remedy of compensatory damages. 34 C.F.R. §§ 668.84(b), 668.88-.92, 668.95, 682.413. Again, though, these costs may be passed forward to the servicers' customers. Or may not; and anyway a regulation can be challenged as arbitrary or capricious even if the regulated firms can deflect the costs of compliance onto their customers. Still, the lower the cost of the added liability either to the servicers or to their customers, the less vulnerable the regulation is to the servicers' challenge. It was their burden to show what increment of liability the regulation imposes over and above their ordinary common law liabilities for the careless mistakes of their employees and that the increment is so great as to make the regulation arbitrary and capricious. They have not even tried to carry this burden.

The other half of the servicers' challenge to the reasonableness of the regulation attempts to raise a question about the regulation's effect on care. The argument here is that strict liability and even the bobtailed form of joint and several liability that the regulation imposes are unnecessary to minimize mistakes in the administration of student loans. The principal difference between liability for negligence and strict liability is that the latter imposes liability for those mistakes that could not be avoided by the exercise of due care--mistakes, in other words, the costs of which fall short of the costs of preventing them. The mistakes will therefore be no fewer. Strict liability will merely shift the cost of the mistakes from the government to the servicers, who may be able to shift them forward to their customers, from thence to the students and schools, and perhaps eventually back to the government, making the regulation futile. All this is possible but ignores the fact that the servicers' preferred position is not negligence liability rather than strict liability but either no liability or liability capped at the fees charged to their customers. This position if accepted would impair their incentive to avoid making mistakes--would give them less incentive to do so than common law liability might do. Had they proposed to the Secretary, if only as a back-up to their preferred position, the alternative of negligence liability, their argument against strict liability would have greater force. Strict liability does not necessarily create any greater incentive to take care than negligence liability. But it does create a greater incentive to take care than no liability, or liability with a ceiling lower than the cost of the harm inflicted by a liable defendant.

The servicers also complain about the scanty articulation by the Secretary of the grounds for the challenged parts of the regulation. They point out that he did not conduct or commission studies on the likely impact of the regulation on the servicers. And that is true. But since this was notice and comment rulemaking, the servicers had every opportunity to conduct their own studies and put the results before the Secretary for his consideration. They failed to do so, and we have warned elsewhere than when an industry opposes a regulation on a ground that requires data for the ground to be convincing, they had better obtain and submit the data. Morales v. Yeutter, 952 F.2d 954, 960 (7th Cir. 1991). And the Secretary can hardly be faulted for having failed to expound on the differences between strict liability and negligence liability, since negligence liability was not the alternative urged by the industry.

The remaining arguments are procedural and the main one is that the Secretary adopted the challenged regulation in violation of the conditions of "negotiated rulemaking," a novelty in the administrative process. The 1992 amendment to the Higher Education Act, under which the regulation was promulgated, required that the Secretary submit any draft regulation to a process of negotiated rulemaking, to be conducted in accordance with recommendations made by the Administrative Conference of the United States and codified in 1 C.F.R. §§ 305.82-4 and 305.85-5 and with "any successor recommendation, regulation, or law." 20 U.S.C. § 1098a(b). A "successor law" to the Administrative Conference's recommendations had in fact been enacted in 1990. It is the Negotiated Rulemaking Act, 5 U.S.C. §§ 561 et seq. It is to expire later this year but was applicable to the servicer rulemaking. The Act and the Administrative Conference's recommendations authorize the agency, in advance of the notice and comment rulemaking proceeding, to submit draft regulations to the industry or other groups that are likely to be significantly affected by the regulations and to negotiate with them over the form and substance of the regulations. The hope is that these negotiations will produce a better draft as the basis for the notice and comment proceeding. The 1992 amendment to the Higher Education Act made negotiated rulemaking mandatory in proceedings implementing the amendment, as we have seen.

The servicers argue that the Department negotiated in bad faith with them. Neither the 1992 amendment nor the Negotiated Rulemaking Act specifies a remedy for such a case, and the latter act strongly implies there is none. 5 U.S.C. § 570. But even if a regulation could be invalidated because the agency had failed to negotiate in good faith, this would not carry the day for the servicers.

During the negotiations, an official of the Department of Education promised the servicers that the Department would abide by any consensus reached by them unless there were compelling reasons to depart. The propriety of such a promise may be questioned. It sounds like an abdication of regulatory authority to the regulated, the full burgeoning of the interest-group state, and the final confirmation of the "capture" theory of administrative regulation. At all events, although the servicers reached a firm consensus that they should not be liable for their mistakes the Department refused to abide by its official's promise. What is more, the draft regulations that the Department submitted to the negotiating process capped the servicers' liability at the amount of the fees they received from their customers, yet when it came time to propose a regulation as the basis for the notice and comment rulemaking the Department abandoned the cap. The breach of the promise to abide by consensus in the absence of compelling reasons not here suggested, and the unexplained withdrawal of the Department's proposal to cap the servicers' liability, form the basis for the claim that the Department negotiated in bad faith.

We have doubts about the propriety of the official's promise to abide by a consensus of the regulated industry, but we have no doubt that the Negotiated Rulemaking Act did not make the promise enforceable. Natural Resources Defense Council, Inc. v. EPA, 859 F.2d 156, 194 (D.C. Cir. 1988) (per curiam). The practical effect of enforcing it would be to make the Act extinguish notice and comment rulemaking in all cases in which it was preceded by negotiated rulemaking; the comments would be irrelevant if the agency were already bound by promises that it had made to the industry. There is no textual or other clue that the Act meant to do this. Unlike collective bargaining negotiations, to which the servicers compare negotiated rulemaking, the Act does not envisage that the negotiations will end in a binding contract. The Act simply creates a consultative process in advance of the more formal arms' length procedure of notice and comment rulemaking. See 5 U.S.C. § 566(f).

The complaint about the Secretary's refusal to adhere to the proposal to cap the servicers' liability misconceives the nature of negotiation. The Secretary proposed the cap in an effort to be accommodating and deflect the industry's wrath. The industry, in retrospect improvidently, rejected the proposal, holding out for no liability. So, naturally, the Secretary withdrew the proposal. A rule that a rejected offer places a ceiling on the offeror's demands would destroy negotiation. Neither party would dare make an offer, as the other party would be certain to reject it in order to limit the future demands that his opponent could make. This concern lies behind the principle that settlement offers are not admissible in litigation if the settlement effort breaks down. Fed.R.Evid. 408. By the same token, the negotiating position of the parties in negotiated rulemaking ought not be admissible in a challenge to the rule eventually promulgated when the negotiation failed.

The servicers argue that they should be allowed to conduct discovery to uncover the full perfidy of the Department's conduct in the negotiations. Discovery is rarely proper in the judicial review of administrative action. The court is supposed to make its decision on the basis of the administrative record, not create its own record. There are exceptions, summarized in Animal Defense Council v. Hodel, 840 F.2d 1432, 1436 (9th Cir. 1988), amended, 867 F.2d 1244 (1989), and the main one has some potential applicability here: discovery is proper when it is necessary to create a record without which the challenge to the agency's action cannot be evaluated. E.g., Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 420, 91 S. Ct. 814, 825-26, 28 L. Ed. 2d 136 (1971); Edgewater Nursing Center, Inc. v. Miller, 678 F.2d 716, 718-19 (7th Cir. 1982). Negotiated rulemaking does not usually produce a comprehensive administrative record, such as notice and comment rulemaking, or a cease and desist order proceeding, or a licensing proceeding would do, any more than a settlement conference will usually produce a full record. Some discovery was conducted in the district court in order to present a picture of what went on at the negotiations between the servicers and the Department. The servicers argue that if only they could get access to the notes of certain participants in the negotiating sessions they could demonstrate additional instances of bad faith on the part of the Department.

Their conception of "bad faith" reflects, as we have noted, a misconception of the negotiation process. It is not bad faith to withdraw an offer after the other side has rejected it. If as we doubt the Negotiated Rulemaking Act creates a remedy as well as a right, we suppose that a refusal to negotiate that really was in bad faith, because the agency was determined to stonewall, might invalidate the rule eventually adopted by the agency. But we do not think that the Act was intended to open the door wide to discovery in judicial proceedings challenging regulations issued after the notice and comment proceeding that followed the negotiations. If as in this case the public record discloses no evidence of bad faith on the part of the agency, that should be the end of the inquiry. Cf. Citizens to Preserve Overton Park, Inc. v. Volpe, supra, 401 U.S. at 420, 91 S. Ct. at 825-26. A contrary conclusion would stretch out such judicial proceedings unconscionably. The Act's purpose--to reduce judicial challenges to regulations by encouraging the parties to narrow their differences in advance of the formal rulemaking proceeding--would be poorly served if the negotiations became a source and focus of litigation.

AFFIRMED.