BOUDIN, Chief Judge.

Plaintiffs, employees of Baxter Healthcare Corporation of Puerto Rico ("Baxter-PR"), appeal from the district court's decision granting summary judgment to defendants on plaintiffs' Title VII claim. Ramos v. Baxter Healthcare Corp. of Puerto Rico, Inc., 256 F.Supp.2d 127, 150 (D.P.R. 2003). That claim charged Baxter-PR, and its U.S. parent — Baxter, International, Inc. ("Baxter") — with discriminating based on race and national origin in setting the terms of plaintiffs' pension plan. The background facts are undisputed.

Baxter is a major manufacturer of health care products. Together with its numerous subsidiaries, it employs about 45,000 people worldwide. It maintains a pension plan, called the Domestic Plan, covering employees of Baxter and many but not all of its U.S. mainland subsidiaries. The Domestic Plan permits early retirement (i.e., before a worker reaches 65), but does not provide full retirement benefits to employees who retire early; until 1990, the Domestic Plan mechanically reduced benefits in set amounts depending on how early an employee retired.

In 1990, Baxter revised the Domestic Plan to employ a point system for early retirement that, in addition to age, took account of how many years the worker had been at Baxter or a covered subsidiary. The net effect was to make the reduction less, and benefits greater, for early retiring workers who had served longer. This in turn made early retirement more attractive for long-serving workers covered by the Domestic Plan. Baxter says some workers are better off with the change and others worse off.

About 6,000 Baxter employees work in Puerto Rico, mostly for Baxter-PR. Employees of Baxter-PR are covered by the so-called Puerto Rico Plan, but this plan, like the Domestic Plan, is controlled by Baxter through its so-called Administrative Committee, the members of which are appointed by the Compensation Committee of Baxter's Board of Directors. Although the plans had differed in the past, just prior to 1990 the retirement terms of the Domestic and Puerto Rico plans were similar. However, when Baxter converted its Domestic Plan to the point system, it did not make a comparable adjustment in the Puerto Rico Plan.

In 1995 Baxter began the closing of a major plant in Carolina, Puerto Rico. A number of Baxter-PR employees, most of whom worked at the plant, responded in 1998 by filing the present law suit in the federal district court in Puerto Rico. The defendants were their employer, Baxter itself, two other Baxter subsidiaries doing business in Puerto Rico, and the members of the Baxter committee that managed both plans. A principal claim, and the only one pursued on appeal, is that the different treatment of early retirees under the respective plans violates Title VII. 42 U.S.C. § 2000e-2(a) (2000).

On motion for summary judgment, the district court ruled in favor of the defendants. Ramos, 256 F.Supp.2d at 150. Although class certification had been sought, that issue was not discussed and neither side complains. In a nutshell, the district court analyzed the evidence adduced on the Title VII claim and concluded that there was no evidence of intentional discrimination sufficient to warrant a trial. After a brief analysis, the court also rejected a claim of disparate impact discrimination. This appeal followed.

Our review of the granting of a motion for summary judgment is de novo, drawing reasonable inferences in favor of the non-moving party, here the plaintiffs. Zapata-Matos v. Reckitt Colman, Inc., 277 F.3d 40, 42 (1st Cir. 2002). Title VII, so far as pertinent, says an employer may not "discriminate" in compensation or terms of employment based, inter alia, on "race" or "national origin." § 2000e-2(a)(1). Baxter assumes without argument that intentional discrimination in pension plan terms against Puerto Ricans would violate Title VII.

Baxter is less ready to concede that the Domestic Plan is more favorable to employees than is the Puerto Rico Plan. It suggests that some employees are better off under the latter than the former. But its documents also show that it would have cost Baxter-PR a substantial sum to switch over to the new point system. So we will also assume arguendo (it does not alter the outcome) that it would be "discrimination" under Title VII to withhold the point system if the motive in whole or in part was hostility to Puerto Ricans. See Rodriguez-Cuervos v. Wal-Mart Stores, Inc., 181 F.3d 15, 18-20 (1st Cir. 1999).

This brings us to the evidence of motive. Typically at this stage Title VII cases undertake a relentless survey of McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), and its main sequels, but we can be spared this exercise. Assuming the plaintiffs made out the so-called prima facie case of discrimination, see id. at 802 — a point which Baxter disputes on the ground that not all of its mainland subsidiaries enjoy access to the point system — the defendants provided a facially plausible reason for not adopting the point system in Puerto Rico. The reasons were the added cost, the financial climate, and the view that the benefits offered by Baxter-PR were already competitive in Puerto Rico. See Feliciano de la Cruz v. El Conquistador Resort Country Club, 218 F.3d 1, 6 (1st Cir. 2000); Zapata-Matos, 277 F.3d at 45.

Whether the evidence in this case created a jury issue on improper motive depends simply on analyzing the sworn testimony and documents presented on summary judgment and deciding what commonsense inferences they permit. If the plaintiffs lack direct evidence of animus, it may yet be possible to infer it circumstantially, depending on the facts; and if the defendants gave explanations that a jury could find to be false, this will count against them in the mix of inferences. Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 148-49, 120 S.Ct. 2097, 147 L.Ed.2d 105 (2000); Thomas v. Eastman Kodak Co., 183 F.3d 38, 57 (1st Cir. 1999), cert. denied, 528 U.S. 1161, 120 S.Ct. 1174, 145 L.Ed.2d 1082 (2000).

Here no direct evidence exists of discriminatory intent — for example, there are no statements by Baxter management disparaging Puerto Ricans. Instead, plaintiffs concentrate their attack on the alleged inadequacy of Baxter's reasons for its differential treatment, on supposed discrepancies in its explanations, on its failure to produce documents, and on the inference that weak explanations are given to cover wicked motives. We set forth the sequence of events within Baxter in more detail and then consider the plaintiffs' arguments.

In 1988 and 1989 a Baxter task force reporting to the Board's Compensation Committee undertook a review of Baxter's retirement program. According to the affidavit of a consultant, Gregory Hansen, who participated in the exercise, the decision to adopt the point system was made because the prior system created a strong disincentive for long serving employees to retire early. The politely unspoken inference is that Baxter wanted to encourage retirements of employees whose long service also meant high salaries.

In all events, plaintiffs do not challenge the explanation but only the failure to extend the same courtesy to employees of Baxter-PR. According to Hansen, by the time of a June 6, 1989, meeting of the task force, Baxter had decided to adopt the point system for the Domestic Plan as of January 1, 1990. When Hansen asked in the meeting whether the system would also be adopted in Puerto Rico, he was told that there might be a delay because of translation and related problems but that his firm should develop an estimate of the increased costs if the Puerto Rico Plan were to be similarly amended.

Other Baxter documents created at the same time suggest that Baxter was inclined to delay any such change in Puerto Rico until, among other things, it could complete a survey of overall benefits and compensation for its Puerto Rico operations. An affidavit from Mary Barker, a former Vice President of Benefits Management at Baxter, says that Baxter's compensation at different locations around the world differs depending upon "cost and competitive market conditions" and that the benefits package in Puerto Rico differs in various respects from those in other locations — some favoring Puerto Rico and some not.

In all events, later documents and testimony covering the 1990-1993 period show that Baxter created a new task force on compensation early in 1993 to review benefits for Puerto Rico employees, and it conducted a study during the year. The study concluded that Baxter's "base salaries in Puerto Rico [were] 93% of the comparator group" and that overall Baxter's "benefits [were] generally competitive." In addition to the study, the task force also priced various enhancements for the Puerto Rico Plan — the point system was estimated to cost $900,000 per year — but there was "no consensus" among task force members on the question whether to adopt the point system.

In her deposition, Mary Barker said that various task force proposals, including the possibility of switching to the point system, were considered by senior Baxter management in October 1993. She said that there were cost concerns about new benefits in Puerto Rico and that at the time business conditions were unfavorable: Baxter's stock price was not doing well and tax-law changes then under review were thought to have potentially adverse effects on Puerto Rico operations. Among the objectives of senior management was to be sure that Puerto Rico costs remained competitive with Baxter's other worldwide operations. Ultimately, the decision was made not to adopt the point system for the Puerto Rico Plan, although other benefits changes were adopted.

The following year Baxter created an advisory counsel made up of representatives of its business units operating in Puerto Rico. This council eventually recommended that Baxter adopt for Puerto Rico units the point system, or a new 401(k) type contribution plan, or both; but Baxter decided not to do either, partly because of cost and partly because a separate 1995 study of its competitors in Puerto Rico showed that Baxter's pension benefits were well above the median. The same study showed that Baxter's overall compensation package in Puerto Rico was below the median but this was, importantly, because Baxter did not offer the 401(k) type plan. Baxter did adopt such a plan for Puerto Rico in 1998.

Thus far there is nothing ominous in the story. About the most one could say — if this were all — is that Baxter, although sometimes following a uniform approach, appears for the most part to have varied its compensation regimes based on location and that Baxter considered adopting the point system in Puerto Rico, but concluded that it would be costly and was not needed to retain its employees. This shows that Baxter's management were profit-oriented capitalists, but not that they intentionally discriminated against Puerto Ricans within the meaning of Title VII.

In attacking this scenario and the district court's grant of summary judgment, plaintiffs' very able brief presents four main themes: that the failure to extend the point system to Puerto Rico contradicted a settled Baxter policy of uniformity as to benefits; that allegedly inconsistent statements were provided as to certain events; that there are suspicious failures by Baxter to produce documents or remember events; and that the district court relied on factual propositions that were fairly controverted and drew inferences against plaintiffs.

Plaintiffs' broadest position, and the foundation for most of their other arguments, is that Baxter had adopted prior to 1990 a general policy of uniformity of treatment in compensation for all employees of Baxter enterprises whether located on the mainland or in Puerto Rico. From this it follows, according to the plaintiffs, that purported cost concerns about extending the point system to Puerto Rico were pretextual and that a reasonable inference from this deceit is an illicit motive.

Whatever the logic, the premise is wrong. It is based mainly on statements in a Corporate Human Resource Policy and Procedure Manual adopted in 1988 described as having "corporate-wide application." The manual says that "[u]nless otherwise approved," each Baxter unit is to "participate" in employee benefit programs including retirement programs designed by Corporate Benefits. Plaintiffs take this as a commitment to uniform compensation at all locations, making the discrepant treatment of Puerto Rico a suspicious deviation.

But the manual, apart from other routine disclaimers, says that uniform treatment is provided "unless otherwise approved." The evidence is clear that Baxter did in certain respects vary benefits by location both before and after 1990; and the history of the Puerto Rico operations vis à vis mainland operations is a tangled one. The claim that uniformity was a strict policy from which this case was a unique, suspicious deviation is wrong on the facts and not one that a reasonable jury could accept.

Plaintiffs' second argument is narrower and comes closest to making out a conventional case against summary judgment. Focusing primarily on the 1990 decision itself, plaintiffs point to one statement, already mentioned, that the extension of the point system to Puerto Rico was being delayed for six months or a year because of translation and related problems; and, charging "pretext," plaintiffs contrast this with the indefinite delays that followed and the concern with cost that was given as the reason for in-action.

This is a fair argument — a false reason coupled with other facts can suggest an illicit real reason, Zapata-Matos, 277 F.3d at 45 — but context here greatly blunts its force. The focus of the 1988-89 task force was on the Domestic Plan, and Hansen in a memorandum discussing the June 6, 1989, implementation meeting says that he "brought up the subject of the Puerto Rico plan and whether comparable changes will be made." He then says that Baxter said it "might want to at least delay 6 months and maybe a year" because of translation and related issues.

Then, the formal minutes of the June 28, 1989, task force meeting record that Hansen again raised the issue and got the following reply:

Judy Coffey Hedquist explained that, from a time perspective, it would not be possible to implement and communicate changes to the Puerto Rico Plan by January 1, 1990.

Herb Walker pointed out that it would also be prudent to defer implementation so that the results of several employee surveys on compensation and benefits currently distributed to Puerto Rico employees become available.

A memorandum by Hansen the next day says: "A survey is being conducted on compensation and benefits in Puerto Rico. It was decided not to extend the new early retirement or death benefit provisions to the Puerto Rico plan currently."

In other words, from the outset, cost and competitiveness were at issue. No juror could reasonably draw from this supposed "discrepancy" an inference that Baxter lied about its reasons for not extending the point system to Puerto Rico and must have harbored animus toward Puerto Ricans. Even plaintiffs' foundational premise — Baxter's supposed policy of uniformity — is undercut by the documents.

Next, the plaintiffs say that Baxter failed to produce documents to back up its story and that the failures undercut or cast doubt on the reasons given by Baxter for its actions. In particular, plaintiffs complain about the absence of the formal minutes of the June 6, 1989, meeting of the task force; any formal study underpinning the decision to adopt the points system for the Domestic Plan; any cost or competitive study leading to the refusal to do so for Puerto Rico; and the results of any Puerto Rico surveys on compensation and benefits.

Starting with the most specific complaint — the absence of June 6, 1989, task force minutes — there is simply no indication that such minutes ever existed, or if they did, were deliberately withheld. The task force appears to have been a working group, not a formal administrative unit like the Administrative Committee that ran the pension plans (all of whose minutes were produced). Baxter did produce an agenda for the June 6, 1989, task force meeting and Hansen's summary, and, although Baxter did produce minutes for at least one other task force meeting, nothing suggests that any further documentary record of the June 6, 1989, meeting exists.

The lack of any formal study of adopting the point system as part of the Domestic Plan may or may not be surprising, but, again, there is no indication that Baxter deliberately withheld any documents in its possession from the plaintiffs. The point system was one of a number of issues considered by the task force; and anyone familiar with corporate decision-making culture would not be surprised that the remnants would be nothing more than agendas, scattered references in summaries and minutes, and a few memories secured by depositions.

Plaintiffs also point out that a stated reason for deferring the point system for Puerto Rico was to await alleged surveys of compensation and benefits being conducted in Puerto Rico. Yet the March 8, 1991, minutes of the Administrative Committee suggest that surveys and other studies were underway, and there is no dispute that by 1993 Baxter had surveyed employees in Puerto Rico and also obtained a study of benefits from comparable companies in Puerto Rico — information made available to the plaintiffs.

The last "missing" item is any document formally explaining the decision not to extend the point system to Puerto Rico. But, as with many cases of inaction, there is no reason to think that any such formal document existed. What the documents taken together show is an initial decision to defer the issue for study, a later determination that the cost of extension would be considerable, and at least two benefits studies indicating that Baxter-PR's retirement benefits matched those of other companies in Puerto Rico.

This brings us to plaintiffs' final claim, a general — but then elaborately documented — claim that the district court relied on facts not proved, ignored evidence favorable to plaintiffs, and drew inferences in favor of defendants. An introductory table provides one to two dozen examples (depending on how one counts) together with record cites and then refines or restates a number of them in text. This is certainly a helpful way to present the material.

The difficulty is that on close examination the examples given tend to dissolve. The first, and most important, group reflect plaintiffs' main premise that from the outset Baxter intended a single retirement regime embracing the mainland and Puerto Rico, deviating at the last minute for pretextual reasons. But we have already explained why the plaintiffs' premise is faulty, resting as it does on a misreading of the manual and an oversimplification of Baxter's history and practice as to uniformity. Several other criticisms rest on the absence of documents ( e.g., surveys) and the alleged contradictions in reasons given by Baxter in June 1989 — both points already discussed. And some of the examples are just argument: e.g., that Christmas bonuses were admittedly paid in Puerto Rico and not in the United States but were required in Puerto Rico by statute.

In the end plaintiffs' collection of examples fails because whatever the quibbles, there is simply no evidence that Baxter management acted out of animus to Puerto Ricans. Plaintiffs say in their table that this ignores evidence that a Baxter-PR official was concerned about the fairness of the decision and that Mary Barker was concerned about employee reaction. But neither comment provides a basis for an inference that Baxter management itself was concerned about anything beyond saving money.

Even with reasonable inferences drawn in plaintiffs' favor, the district court correctly held that there was no evidence of animus — and that means that the discriminatory treatment claim fails. What is left is that part of the evidence that shows that employees of Baxter-PR — most and perhaps almost all of whom are Puerto Rican — were treated differently than Baxter's mainland employees, who we will assume are mostly not Puerto Rican. This brings us to plaintiffs' alternative disparate impact theory of violation.

Plaintiffs claim that regardless of intent or lack of racial discrimination, Baxter's dual plan approach has the impermissible effect of disfavoring Puerto Ricans. See Int'l Bhd. of Teamsters v. United States, 431 U.S. 324, 335 n. 15, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977) (disparate impact prohibits "employment practices that are facially neutral in their treatment of different groups but that in fact fall more harshly on one group than another and cannot be justified by business necessity"); see generally Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971).

It is true that disparate impact liability can be imposed regardless of whether an employer possesses an intent to discriminate, Watson v. Fort Worth Bank Trust, 487 U.S. 977, 987, 108 S.Ct. 2777, 101 L.Ed.2d 827 (1988), but Title VII also says:

Notwithstanding any other provision of this subchapter, it shall not be an unlawful employment practice for an employer to apply different standards of compensation, or different terms, conditions, or privileges of employment pursuant to a bona fide seniority system . . . or to employees who work in different locations, provided that such differences are not the result of an intention to discriminate because of race, color, religion, sex, or national origin.

42 U.S.C. § 2000e 2(h) (2000) (emphasis added). In other words, different treatment in different locations is permissible absent an intent to discriminate.

What legislative history there is tends to confirm this view, see 110 Cong. Rec. 12,723 (1964) (statement of Sen. Humphrey); and the few appellate cases interpreting the "different locations" language of § 2000e-2(h) also appear to support this reading. See Russell v. Am. Tobacco Co., 528 F.2d 357, 362-63 (4th Cir. 1975), cert. denied, 425 U.S. 935, 96 S.Ct. 1666, 48 L.Ed.2d 176 (1976) ("If . . . Leaf and Branch are different locations, and if the differences in treatment of workers at the different locations are not due to an intention to discriminate, the company's refusal to allow Leaf employees to transfer on the basis of company-wide seniority would not violate the Act.").

In addition, there are numerous cases, including Supreme Court decisions, construing the companion "bona fide seniority system" language of § 2000e-2(h) — those cases hold that an employer who provides different levels of compensation for employees pursuant to a bona fide seniority system cannot be held liable under a disparate impact theory — plaintiffs must prove an intent to discriminate. E.g., Pullman-Standard v. Swint, 456 U.S. 273, 289, 102 S.Ct. 1781, 72 L.Ed.2d 66 (1982); Trans World Airlines, Inc. v. Hardison, 432 U.S. 63, 82, 97 S.Ct. 2264, 53 L.Ed.2d 113 (1977); Int'l Bhd. of Teamsters, 431 U.S. at 348-56, 97 S.Ct. 1843. The two provisions of subsection 2(h) can fairly be read together.

The subsection itself is not surprising. Location is often a proxy for differences in cost and other competitive circumstances; and while Congress could have made those circumstances a separate defense, the difficulties of showing that a difference in pay precisely correlated with a difference in cost would be formidable. Cf. Texaco Inc. v. Hasbrouck, 496 U.S. 543, 561 n. 18, 110 S.Ct. 2535, 110 L.Ed.2d 492 (1990) (discussing the difficulties of establishing the cost justification defense under the Robinson-Patman Act). In effect, different locations are simply a safe harbor in cases where there is no intentional discrimination.

Affirmed.