The plaintiffs, Boston-area businessmen John Roche and Mark DiRico, invested in a fish farm in Prince Edward Island, Canada. The venture failed, and they sued the court-appointed receiver which sold them the farm and the bank which had originally financed the project. A jury found against plaintiffs on their claims of common law fraud and misrepresentation for failure to meet their burden of showing proximate cause. The district court, initially finding plaintiffs' allegations actionable under Mass. Gen. Laws ch. 93A ("chapter 93A"), found after trial that defendants had committed unfair and deceptive trade practices, but that their offending acts did not occur "primarily and substantially" in Massachusetts, as required by chapter 93A. Plaintiffs recovered nothing. Defendants' motion for attorneys' fees was also denied. Both parties appeal on the chapter 93A issue, and defendants appeal from the denial of attorneys' fees. We affirm.

We recite the facts as the jury and district court could have found them. Cambridge Plating Co. v. Napco, Inc., 85 F.3d 752, 756 (1st Cir.1996). "Where specific findings are lacking, we view the record in the light most favorable to the ruling, making all reasonably supported inferences." United States v. McCarthy, 77 F.3d 522, 525 (1st Cir.1996).

The story starts in February 1987, when Aquacare A.S., a Norwegian firm, conducted a study for its subsidiary Seasprings Farms Ltd., on the viability of a land-based fish farm in Prince Edward Island, Canada ("PEI"). Aquacare prepared a prospectus (the Aquacare I report), which presented general information about the biological and technical aspects of the proposed operation. The prospectus also contained financial projections and asserted that the production capacity of the contemplated facility would be 131.8 tons of fish in the first year of operation and 360 tons annually thereafter.

The Royal Bank of Canada financed the construction of the farm, which was operated by a firm called Marine Harvesting, Ltd. ("Marine"). Fish were first introduced into the facility in December 1987, but there were construction delays and the plant was not completed until June 1988. Things went badly from the start. The fish did not grow as fast as anticipated and had unexpectedly high mortality rates. By October 1988, the fish had not yet reached market size (4 lbs), as expected.

In October, Cleve Myers, Marine's President, retained Aquatech Systems, A.S., another Norwegian aquaculture firm, to perform an independent study of the farm's problems. Myers wanted to know, among other things, how many tons of fish the farm was capable of producing. Dr. Michael Smith, a biologist from Aquatech, came to the farm on October 30 and spent three days making observations, conducting tests, and speaking with employees. Dr. Smith was not shown the Aquacare I report, but the Aquacare operating manuals were made available to him. The result of Dr. Smith's analysis was a forty-four page report (the Aquatech report), dated November 18, 1988.

The Aquatech report criticized the design of the farm. The report stated that "until experience over a fully operational annual cycle has been gained," the farm could produce, at best, only 200 tons of fish per year. Dr. Smith's report also made several recommendations for improving production. For instance, it recommended that a higher proportion of fresh water be pumped into the tanks, for purposes of both temperature control and salinity control. This suggestion deviated from the specifications in the Aquacare operating manuals. The 200-ton projection was premised on the assumption that the recommended changes would be implemented.

Meanwhile, Osler, Inc., one of Marine's two fifty-percent shareholders, was in receivership, and A.S. Bergens Skillingsbanken, the other fifty-percent shareholder, was refusing to inject additional capital into the venture.

On November 30, Myers turned the keys to the farm over to Lou McGinn, the Loan Officer in charge of the project at the Royal Bank, and requested that the farm be placed in voluntary receivership. Myers stated that without operating capital he was unable to care for the inventory or pay the staff. He also stated that he had "received a consultant's report regarding the production capabilities of the plant which indicate [sic] that the facility is not viable as it presently is financed." McGinn stated that Myers had told him that he received "a consultant's report" which indicated that "the plant was capable of producing only approximately 200 tons of product per year as opposed to the 360 tons of product per year upon which its financing and viability had been initially contemplated and established." In a letter of January 4, 1989, to McGinn, Myers wrote that:

after the [Aquatech] report indicated an expected tonnage, which was not viable, I received confirmation from the directors and major shareholders to advise the bank and request a receiver. Both they and I considered it wrong to continue operations with this information in our possession.

McGinn understood from Myers that the Aquatech report was one of the factors that led Marine to decide not to inject any more money into the project and to opt for receivership. McGinn informed his supervisor of the development, and the supervisor noted in a memo that "apparently a recent study ... places some question on the viability of the project."1

At the Royal Bank's request, the Supreme Court of PEI appointed Deloitte & Touche2 as receiver of Marine's assets. Karen Cramm, a partner in Deloitte's Halifax, Nova Scotia office, was in charge of the receivership. Cramm took instructions on the handling of the assets from the Royal Bank, specifically from McGinn, as well as from the PEI court. The Royal Bank was the sole secured creditor of Marine, with a $ 2.8 million note outstanding. The expectation was that none of Marine's other creditors would get anything out of the sale of the assets; there would be a shortfall.

Cramm immediately took possession of Marine's assets and reviewed the company's financial records. Farm employees were terminated as employees of Marine and rehired as employees of Deloitte, as receiver for Marine. Deloitte decided to sell the farm by means of a public tender. Cramm immediately began preparing an information package which could be sent to potential investors. Around this time Myers showed Cramm both the Aquacare I report and the Aquatech report. Cramm read both reports. She included neither of the reports in the information package, though later she would offer the Aquacare I report alone to prospective purchasers.

Deloitte placed advertisements in various newspapers, including the Boston Globe, describing the farm and inviting interested readers to contact Deloitte for further information. The ads stated that the "business is intended to be sold on a going-concern basis."

In early December, after learning that Marine had become insolvent, Aquacare contacted Cramm and told her that it wanted access to the plant to conduct a review of plant operations. Aquacare had heard about the Aquatech report, which criticized Aquacare's design and operational specifications. Cramm agreed to Aquacare's request. Aquacare then issued two unsolicited reports, the Aquacare II report on December 15 and the Aquacare III follow-up report on December 20, attempting to rebut the Aquatech report. Aquacare II reviewed plant operations and praised the design and operational specifications of the plant. Aquacare II reasserted that 360-ton production "can be achieved without problem," provided there was proper stocking of fingerlings (young fish), and recommended a fingerling acquisition plan. Aquacare II also blamed Marine's management for the plant's failures. Aquacare III explicitly refuted the Aquatech report. Aquacare did not bill Deloitte for these reports.

Meanwhile, John Roche, a Massachusetts real estate developer, had been seeking a business opportunity in aquaculture. After reading the Globe ad, he called Cramm from his home in Massachusetts. The next day she sent him the information package. Roche discussed the farm with two of his business associates, Mark DiRico (the chief engineer of a packaging supply business) and George Call, who both expressed an interest. Roche and Call decided to visit the facility in Canada. On January 3 or 4, 1989, they toured the facility with Cramm. During this visit, Roche asked Cramm why the previous investors had failed. She responded that they had run out of money. Roche and Call also met with Cliff Yorston, the Plant Manager, who told them that some of the fish were ready for sale and that all the plant needed was an infusion of new capital.

Cramm offered the visitors a copy of the Aquacare I report. Before receiving it, they were required to sign a disclaimer, which stated, among other things, that they were aware that Deloitte "had not, in any way, attempted to verify the information contained herein." Cramm stated that she offered the Aquacare I report to prospective purchasers because "it was an informative package that included a layout of the plant facility" and "it talked about, in general, the aquaculture operation." She explained that the disclaimer was intended to protect Deloitte against claims involving any inaccuracies in the information contained in the report, such as the financial projections, growth charts, and profitability analysis. She had not reviewed that information and had no opinion as to its accuracy.

Cramm admits she did not offer Roche and Call a copy of the Aquatech report, but claims it was among the financial statements that Roche and Call were invited to view.

Roche and Call next met with McGinn at the Royal Bank of Canada to discuss financing. McGinn, like Cramm before him, told the two Americans that the farm went into receivership because the previous investors had run out of money.

Roche and Call returned to Massachusetts, where Roche showed the Aquacare I report to DiRico. During the next week, Roche had several telephone conversations with McGinn about potential financing. McGinn recommended that Roche retain a local Canadian lawyer, Scott MacKenzie, and a local Canadian accountant, Stan MacPherson, to assist him in dealing with various government agencies. Roche did so.

On January 11, Roche and Call, this time accompanied by DiRico, met with McGinn and Cramm at the Royal Bank in PEI. DiRico asked McGinn for "sales figures"; DiRico claims that McGinn gave him a copy of the Aquacare I report, telling him that "all the figures are in here." DiRico says he was not asked to sign a disclaimer. Cramm denies that DiRico was given a copy of the report at this meeting. DiRico asked McGinn about the plant's troubles; McGinn responded that mismanagement was to blame.

Roche and his associates also went to see the farm. Cliff Yorston, the Plant Manager, pulled a large fish from one of the tanks and told the two visitors that he had a whole tankful of similar "beauties" which could be sold for $ 3.75 a pound in Boston. Roche and his associates planned to generate operating capital for the plant through the immediate sale of inventory.

Roche and his partners made an offer of $ 2.9 million the next day, subject to Royal Bank financing. But Cramm, after securing the Royal Bank's support, rejected the offer, instead choosing to accept a lower all-cash offer ($ 1.4 million) from a group led by Hirsch Spiegleman. The Spiegleman group's offer, unlike the Roche group's offer, did not require Royal Bank financing.

On or around January 17, Cramm told Roche that his offer had been rejected. McGinn also asked Roche if Roche's group would still be interested in the event that the Spiegleman deal fell through.

After his bid was accepted, Spiegleman requested a copy of the Aquatech report (which he referred to as the "200 ton report"). The record does not reveal how Spiegleman, unlike Roche, knew about the Aquatech report. Cramm informed Aquacare of Spiegleman's request and asked for permission to give him a copy of the Aquacare II and Aquacare III reports at the same time.

In mid-February 1989, the Spiegleman deal collapsed. The Roche group hoped to take over the Spiegleman offer rather than going through the process of submitting a new bid. Seeking financing, Roche and his partners met with Amy Hunter, a loan officer at Boston Private Bank & Trust Company (Boston Private). Roche gave Hunter a copy of the Aquacare I report. Hunter understood that Roche was presenting the report as his business plan. Roche told Hunter that the fish farm went into receivership because the previous owners had run out of money. Roche also told Hunter that he planned to raise operating capital by immediately selling fish.

Hunter next called Cramm and McGinn to confirm what Roche had told her. Cramm told Hunter that the farm's equipment was state-of-the-art and all that was needed was the right management. McGinn echoed these thoughts, adding that Roche would get half his investment back in the first three months without even having to invest in any new fish. Hunter asked McGinn if he knew of any downside risks to the farm or whether he had any projections for the farm. McGinn responded that he did not know any downside risks other than bad management and that Roche's ideas sounded good to him. He did not offer any other projections. On February 28, Boston Private approved a loan of $ 1.2 million to Roche, secured by property Roche owned in the Boston area.

Cramm, however, decided not to allow Roche to take over Spiegleman's offer. Instead, she re-tendered the farm to all potentially interested buyers, including Roche. Consequently, she edited the original information packet and sent it out on March 1 to everyone who had expressed an interest in the farm.3

On March 2, Cramm received from Yorston, the farm's Plant Manager, a copy of a recent memorandum written by Aquacare (the Aquacare IV report). This memorandum had been prepared by Aquacare for its subsidiary, Seasprings. It accidentally fell into Yorston's hands because Aquacare, intending to send it by fax to Seasprings, sent it to the fish farm's fax number. The Aquacare IV report contained scathing criticism of Deloitte's management of the farm. Aquacare asserted that Deloitte, by attempting to maintain the status quo pending sale, was actually reducing the value of the farm's assets. For instance, several tons of fish had matured past marketability. Aquacare believed that Deloitte was wasting money by feeding these unmarketable fish. Additionally, Deloitte had failed in a timely fashion to purchase fingerlings to restock the farm and future harvests would be lower. Aquacare believed that under Deloitte's management the value of the inventory as well as the farm's 1989 and 1990 earning potential were decreasing daily. Aquacare implicitly disavowed its own earlier production capacity projections, made in the Aquacare I report and reasserted in the Aquacare II report, at least as to 1989 and 1990.4

Roche and DiRico visited the farm again on March 8. Roche told Yorston that if his bid was accepted he would need to start selling fish immediately to raise working capital. Yorston replied that he had $ 150,000 worth of fish ready for shipment to Boston. Roche also believed that, once the farm was purchased, McGinn was prepared to provide working capital in the form of a Royal Bank operating line. On March 10, Roche and his partners offered $ 1.4 million. This offer was accepted on March 13.

Roche called Yorston on March 13 to inquire again about the prospects of immediate sales. Yorston faxed a response to Roche. The cover sheet said, "Here are some projected cash flows for 1989." The attached projections anticipated sales in March of $ 154,000 and total sales for the year of almost $ 2.5 million.5 These projections had not been prepared by Deloitte or by Marine, but by the Spiegleman group. However, someone had obliterated the identifying heading at the top of the page so that, when the projections were sent by fax, the source of the projections was not identified to Roche. Yorston had earlier told Roche that he was not authorized to release any information about the plant without permission from Cramm.

On March 17, Scott MacKenzie, the Canadian attorney hired by Roche to help buy the farm, faxed two newspaper articles to Roche, along with a brief note saying that Roche might find the attached articles "interesting." One of the articles discussed Marine's collapse, quoting Myers as saying "[t]he Aquatech report was devastating and the only option open to Marine Harvesting Ltd. was to request that the bank place a receiver on site." Roche filed the articles away without reading them.

On March 20, the Canadian court approved the sale of the farm. Roche formed a corporation in Canada, International Marine Fisheries, Ltd. (IMFL), to make the purchase. Roche instructed Yorston to fill orders for shipment to Boston. Yorston was able to fill the first few orders of 2 to 4 pound fish, but the buyers in Boston complained that the fish were too small. Yorston was unable to fill the first order for 4 to 6 pound fish. He then resigned. Scott Taylor, formerly the farm's Chief Biologist, took over Yorston's position, but he was also unable to fill these orders. In April, Taylor told Roche that the larger fish were all unmarketable because they were too old and that the other fish were too small for sale. Taylor predicted that the smaller fish would be four pounds by June and could be sold then. By June, the smaller fish still had not grown to four pounds. Taylor was fired. The new Plant Manager, Richard Gallant, recommended, in order to cut expenses and see if the fish would grow faster, that the fish be placed in cages in the bay. This was done in July.

Roche and his partners injected cash of their own into the enterprise to save it and also took out an operating line of credit from the Bank of Nova Scotia in the amount of $ 500,000. Roche personally guaranteed the loan. During the summer, Roche and DiRico bought Call's share.

Roche came to think the fish farm was a failure. He began negotiating with a local veterinary college which was interested in buying the farm for research purposes. He hired Deloitte to help him set up this deal. Stan MacPherson and Cleve Myers, both recently hired by Deloitte, worked on the project.

In September 1989, Roche was at the fish farm. Tipped off by a former secretary, he found the Aquatech report in a desk drawer. Attached to the report was a 1988 note from Cleve Myers, the former president of Marine, to a prospective fish dealer. The note said that the farm:

was placed in receivership Friday afternoon. The fish have not grown fast enough, as you know, plus I have received a specialist's report from Norway stating that we can only grow 200 tons annually (just enough to lose one million dollars each year).

After reading the report and the attached note Roche concluded that he had been defrauded.

In November 1989, the Bank of Nova Scotia called the loan and demanded that Roche, as personal guarantor, make payment to the bank. He did not do so. The fish farm closed in December, and went into receivership in February. The Bank of Nova Scotia sued Roche and DiRico on the defaulted loan.

Meanwhile, IMFL, the plaintiffs' Canadian corporation, was dissolved in May 1990. Roche and DiRico brought this action against Deloitte and the Royal Bank in November 1990.

Roche and DiRico sued Deloitte and the Royal Bank on two theories: common law fraud and misrepresentation, and "unfair or deceptive acts or practices in the conduct of any trade or commerce," under Mass. Gen. Laws ch. 93A, § 2.

The common law counts were tried before a jury. Since the defendants did not consent to a jury determination of the chapter 93A claim, that claim was decided by the district judge. The jury returned a verdict for the Royal Bank on all the common law counts. As for the common law claims against Deloitte, the jury found that Roche and DiRico had proved there were negligent misrepresentations, but that the plaintiffs were ninety-nine percent negligent compared to the one percent negligence of Deloitte; therefore the claim foundered on proximate cause grounds and there was no recovery. The district court ruled that defendants had engaged in misrepresentations contrary to the standard of chapter 93A,6 but that the actionable conduct did not take place "primarily and substantially within the commonwealth" of Massachusetts, as chapter 93A requires. Mass. Gen. Laws ch. 93A, § 11. The court thus ruled in favor of defendants on the issue of chapter 93A liability.

Plaintiffs, while defending the district court's predicate rulings, appeal the court's ultimate decision on chapter 93A liability. They assert that the deception occurred primarily and substantially in Massachusetts.7 Defendants, by contrast, defend the "primarily and substantially" ruling, but challenge a number of the district court's predicate rulings. Specifically, defendants argue that Roche and DiRico are not the real parties in interest because the deception, as alleged, was perpetrated on the Canadian corporation established by Roche and DiRico to purchase and operate the fish farm, not on the two men as individuals. They also argue that, under forum state (Massachusetts) conflicts principles, Canadian law, rather than Massachusetts law, should have been applied to the suit. Finally, they argue that, even under Massachusetts law, plaintiffs had not successfully shown actionable conduct under 93A. Because we affirm the ruling below on the ground that the conduct was not primarily and substantially in Massachusetts, there is no reason to address defendants' other arguments in any detail.

The question of whether the deception occurred primarily and substantially in Massachusetts is one of law, which we review de novo. Compagnie de Reassurance d'Ile de France v. New England Reinsurance Corp., 57 F.3d 56, 90 (1st Cir.1995); Clinton Hosp. Ass'n v. Corson Group, Inc., 907 F.2d 1260, 1264 (1st Cir.1990). This review, however, cannot be conducted in a vacuum. We first determine exactly what constituted the unfair or deceptive acts.

The district court found that six acts or omissions constituted actionable deception on the plaintiffs. Whether a certain act or omission (or cluster of acts or omissions) is actionable under chapter 93A is a question of fact. Brennan v. Carvel Corp., 929 F.2d 801, 813 (1st Cir.1991); USM Corp. v. Arthur D. Little Sys., Inc., 28 Mass.App.Ct. 108, 546 N.E.2d 888, 897 (1989). Review is for clear error. After defining the deception, the district court then employed the pragmatic, functional analysis spelled out by this court in Clinton Hospital, 907 F.2d at 1266, in order to ascertain whether the deception occurred primarily and substantially in Massachusetts.8 Neither party challenges the district court's methodology in determining the "primarily and substantially" question by examining only the specific acts (or omissions) of misconduct, instead of, say, by examining the larger context of the parties' dealings. We therefore use the same methodology. See also Gilleran, The Law of Chapter 93A § 3:15, at 32 & n.59.1 (1995 Supp.) ("The relevant wrongful conduct to be considered for purposes of personal jurisdiction under 93A is that conduct which violated 93A.") (citing cases).

The district court found that the following six acts/omissions constituted the deception:

i. Cramm's response on January 4, 1989, to a question from plaintiffs about the problems of the fish farm, indicating that the previous operators had run out of money, but omitting that the investors withdrew after reading the Aquatech report;

ii. McGinn's statement to the same effect;

iii. Cramm's failure to offer plaintiffs a copy of the Aquatech report when she offered them a copy of the Aquacare I report;

iv. Cramm's failure, after March 1, 1989, to disclose to plaintiffs the Aquacare IV report, which showed that Aquacare no longer stood by the projections it had made in the earlier Aquacare I report;

v. McGinn's statement that buyers would be able to sell fish immediately to generate operating capital; and

vi. McGinn's failure to tell Hunter about the Aquatech report, after offering his opinion as to the viability of the farm, when she asked if he had any more information.9

Both parties challenge the district court's findings on what constituted the deception. Plaintiffs urge us to give a broader construction, while defendants complain that the district court saw deception where there was none. We have reviewed the record and are mindful that there is conflicting testimony on many points. The judge saw the witnesses on the stand and had the opportunity to weigh their credibility. See Fed.R.Civ.P. 52(a).

Defendants argue that the district court erred in finding that defendants deceptively failed to disclose the Aquatech report. Defendants assert that the non-disclosure claim is an impossibility here because plaintiffs had knowledge prior to buying the farm about the very document the defendants are accused of failing to disclose, the Aquatech report. This knowledge, the defendants essentially argue, is based on either of two alternate theories. First, defendants assert that plaintiffs had imputed knowledge of the report: their agent MacKenzie knew about the Aquatech report through a newspaper article and that knowledge could be imputed to the plaintiffs.10 Second, defendants assert that the plaintiffs had constructive knowledge: MacKenzie had faxed a copy of the article to Roche (a week after plaintiffs' offer was accepted), and Roche's knowledge could be assumed, without relying on an agency theory (although Roche did not read the article until long after).11 Even assuming these theories are valid, defendants have still failed to demonstrate reversible error. Prior to the closing, plaintiffs at most knew of the existence of the Aquatech report. Defendants' deception here was in not providing plaintiffs with a copy of the report given that plaintiffs had been provided with a copy of the rival Aquacare I report. Under these facts the district court did not err in finding that defendants deceptively failed to disclose the Aquatech report.

Defendants also argue it was clear error for the district court to conclude that the farm's original investors pulled out because of the Aquatech report, a conclusion upon which the non-disclosure claim rests. Defendants are correct that the record does not reveal that the investors pulled out solely because of the unfavorable Aquatech report. However, the report clearly played a major role in the investors' decision to place the plant in receivership. At the very least, Cramm and McGinn knew that the previous investors pulled out just after receiving the Aquatech report, and that the report played a large role in this decision. Myers wrote McGinn on January 4, 1989, saying that both the directors and the major shareholders thought it was wrong to continue operations after receiving the information in the Aquatech report as to the expected tonnage. There is no clear error.

Plaintiffs, in turn, challenge certain findings of non-deception as clearly erroneous. They argue that the advertisement placed by Cramm was deceptive because it described the farm as a "going concern." A "going concern" is "[a]n enterprise which is being carried on as a whole and with some particular object in view" or "an existing solvent business, which is being conducted in the usual and ordinary way for which it was organized." Black's Law Dictionary 691 (6th ed.1990). The district court determined that the use of the term "going concern" here was not deceptive. Plaintiffs rely entirely on Dr. Smith's findings in the Aquatech report in asserting on appeal that the farm was not a "going concern." In Dr. Smith's opinion, say plaintiffs, the farm's pumping system was incapable of operating properly at optimum levels and the farm was, as a general matter, not state-of-the-art.

Plaintiffs' charge of error fails on two grounds. First, Dr. Smith is one aquaculture expert among several whose conflicting views about the farm became known in this case. Indeed, this is the very reason his report should have been available to the plaintiffs: the plaintiffs had been given a different report by a competing aquaculture firm which expressed a far rosier view of the farm's operations and production capacity. While the plaintiffs' non-disclosure claim is rooted in the variety of opinions available about the viability of the farm, their "going concern" argument presents Dr. Smith's opinion as the single truth. But most damaging to plaintiffs, Dr. Smith's report never suggested, let alone stated, that the farm was not a "going concern." His report indicates only his belief that the farm was more poorly designed, and consequently a riskier investment, than its designers thought.12

Finally, that the farm ultimately failed (and thus, presumably could not properly have been described as a "going concern" at some point during its decline) is not proof that it was doomed to failure at the time the ad was placed, at the start of the receivership. Indeed, implicit in the plaintiffs' overall case is a claim that their venture might not have failed if the receiver had managed the farm better during the term of the receivership and if the selling price better reflected the actual value of the assets. Precisely because the value and production capacity of the farm were so hotly disputed at trial, the judge did not commit clear error in finding that the farm was a "going concern" at the time the Globe ad was placed. The farm's ultimate failure does not change this.13

These factual findings lead to the question whether the deception occurred primarily and substantially in Massachusetts. Contrary to the standard burden of proof on jurisdictional questions, here the burden is on defendants to show that their misconduct occurred primarily and substantially outside Massachusetts. Mass. Gen. Laws ch. 93A, § 11; Compagnie de Reassurance d'Ile de France, 57 F.3d at 90; see Gilleran, supra, § 3:16, at 65. This is because " § 11 provides an exemption from 93A liability, available as a defense, rather than a jurisdictional prerequisite to suit." Kansallis Fin. Ltd. v. Fern, 40 F.3d 476, 481 (1st Cir.1994). The district court's conclusion that defendants satisfied this statutory burden is one of law, reviewed de novo. Compagnie de Reassurance d'Ile de France, 57 F.3d at 90; Clinton Hospital, 907 F.2d at 1264.

In Clinton Hospital, this court read the sparse body of Massachusetts precedent as support for a pragmatic, functional approach for determining whether actionable misconduct occurs primarily and substantially in Massachusetts. The Clinton Hospital court distilled the approach down to three basic factors: (1) where defendant committed the deception; (2) where plaintiff was deceived and acted upon the deception; and (3) the situs of plaintiff's losses due to the deception. Id. at 1265-66. The district court, applying the Clinton Hospital factors to this case, reasoned that the deception here occurred primarily and substantially in Canada.

This court has previously recognized that the first factor is the least weighty of the three factors. Compagnie de Reassurance d'Ile de France, 57 F.3d at 90; Clinton Hospital, 907 F.2d at 1265-66. The district court was aware of this and ruled that the first factor weighed against the plaintiffs because most of the defendants' misrepresentations and deceptive acts were committed in Canada. This is plainly correct. Indeed, the district court understated the point: all the misrepresentations occurred in Canada. Nevertheless, plaintiffs argue on appeal that the "clear preponderance" of defendants' deceptive conduct occurred in Massachusetts. This argument, however, is premised on a certain set of acts which the district court did not consider part of the deception. Specifically, plaintiffs assert that the Boston Globe ad and the two information packets were deceptions which defendants committed in Massachusetts. But the district court did not err in finding that the ad and the packets were not misleading.

As to the non-disclosure of the Aquatech report and the Aquacare IV report, there is a metaphysical dilemma in ascertaining where this occurred. The very problem is that disclosure did not occur. The district court read Clinton Hospital as support for the proposition that non-disclosure occurs where the defendant who fails to disclose is located. This was obviously Canada. But Clinton Hospital does not answer this question. Where the defendants' obligation to disclose certain documents arose out of their voluntary disclosure of other documents, and where the parties' dealings were fairly evenly split between two places, it is most reasonable to say that the non-disclosure occurred equally in both places. However, since all the actual misrepresentations, as detailed above, were made by the defendants in Canada, the first Clinton Hospital factor still weighs in favor of defendants.14 The district court concluded, and we agree, that the bulk of the defendants' unfair and deceptive acts and omissions were in Canada.

The second Clinton Hospital factor requires an analysis of where the plaintiffs received and acted upon the deceptive or unfair acts or practices. As the district court properly recognized, this factor, though framed by the Clinton Hospital court as a single factor, really requires two distinct inquiries.

As to where the plaintiffs received the deception, the district court concluded that it was primarily Canada. In coming to this decision, the court declined to consider the misleading statements made by McGinn to Hunter during their telephone conversation because the district court believed Hunter "was not acting as an agent of the plaintiffs when she received the misrepresentation, and there is no evidence she relayed any of the misrepresentations to the plaintiffs." Plaintiffs complain the court erred in not including this telephone conversation as a deception. This is an arguable issue, but we resolve it against plaintiffs. In terms of strict agency law, Hunter was acting for Boston Private, not for plaintiffs, when she made the call. There is no evidence that Hunter told McGinn or Cramm that she was calling on Roche's behalf or that she would transmit the information to Roche. Thus, while there was a deception of Hunter, it was not of plaintiffs. As to the district court's second rationale--that no misrepresentations were conveyed by Hunter to plaintiffs--it is inapplicable on these facts. McGinn concealed from Hunter the very information which, Hunter testified, she would have relayed to Roche if McGinn had been more forthcoming.

The only other misrepresentation received in Massachusetts was the Yorston fax, which, under the district court's own evidentiary rulings, may be considered in this analysis. On the other hand, there were three misleading statements made by the defendants and received by the plaintiffs in Canada during the plaintiffs' initial visits there. And, as discussed above, the defendants' failure to disclose was, for these purposes, received by the plaintiffs equally in Massachusetts and Canada, as the parties' overall dealings were evenly divided between those two places. Therefore, while the plaintiffs received the deception in both places, we agree with the district court that slightly more of it was received in Canada.15

The second Clinton Hospital factor also requires us to examine where plaintiffs acted upon the deception. Indeed, in Clinton Hospital, this court recognized that "the critical factor is the locus of the recipient of the deception at the time of reliance." Clinton Hospital, 907 F.2d at 1265-66; see also Compagnie de Reassurance d'Ile de France, 57 F.3d at 90. Under varying circumstances this may or may not be the same place as the place where the plaintiff received the deception. Here, the district court concluded that the plaintiffs' actions in reliance were evenly divided between Massachusetts, where they negotiated a loan secured by Massachusetts properties, and Canada, where they hired a lawyer and accountant and formed a Canadian corporation to purchase the farm, and that this sub-factor thus favored neither party.

The district court thus concluded that the two-faceted second factor, like the first factor, weighed in favor of Canada, and consequently in favor of the defendants. We agree.

The third Clinton Hospital factor is the location of the loss. The parties advance radically different conceptions of the loss here. Plaintiffs assert that the loss was the loss of the Massachusetts property that secured the Boston Private loan and that was ultimately foreclosed on when the plaintiffs defaulted on the loan. Defendants conceive of the loss as the loss of the money which was invested in the farm in Canada.

The district court adopted the plaintiffs' definition. The court's reasoning was heavily influenced, perhaps even mandated, by its earlier ruling on the scope of the damages which plaintiffs would be allowed to seek in this action. In that earlier ruling, the court rejected defendants' argument that the plaintiffs as individuals were not the real parties in interest in this case by defining the alleged losses narrowly, allowing plaintiffs to seek recovery only of losses they incurred personally in capitalizing the corporation and not those losses suffered by the corporation in Canada. The district court felt obligated, for the sake of consistency, to use the same conception of the loss in this context. The court thus concluded that this third factor favored the plaintiffs. We agree.

The district court balanced the three factors and found that defendants--in whose favor the first two factors weighed--had met their burden of proving that the deception had occurred primarily and substantially outside Massachusetts. Although the pragmatic, functional analysis is not necessarily limited to the three factors, it is significant that two of the three identified factors weigh in favor of the defendants here. Cf. Central Mass. Television, Inc. v. Amplicon, Inc., 930 F. Supp. 16, 27 (D.Mass.1996) ("[W]hen 'place of injury' is the only factor weighing in favor of a claimant, the admonition of Massachusetts courts that liability under chapter 93A is not to be imposed lightly is particularly relevant.").

Defendants moved, after trial, for attorneys' fees in the sum of $ 865,000. The district court, in a written memorandum, denied this motion. Defendants appeal.

The district court first noted the dispute between the parties over whether United States or Canadian law should apply to the fee-shifting claim. Then, assuming arguendo that Canadian law applied, as defendants urged, the court interpreted Canadian law to grant judges "absolute and unfettered" discretion in attorneys' fees claims. Exercising this discretion, the district court determined that an award of fees would be "inequitable" because of defendants' intentional misconduct, and it denied the motion.

Defendants base their argument--one wholly based on Canadian substantive law--on the underlying premise that Canadian substantive law applies to the entire dispute between the parties. They make no argument that they are entitled to fees under Massachusetts law. The district court, however, rejected the defendants' choice of law position, applying Massachusetts law to the plaintiffs' chapter 93A claim. The court consulted Canadian law in considering, and disposing of, the claim for the sake of argument only. It is unnecessary to resolve the choice of law issue, as the denial of attorneys' fees was, under Canadian law, well within the trial court's discretion given the circumstances of this case.

Defendants unsuccessfully argue that the district court overestimated the degree of discretion enjoyed by Canadian courts in the fee-shifting context. Under Canadian law, "[a] successful litigant has by law no right to costs." Orkin, The Law of Costs § 202.1, at 2-3 (2d ed.1993).16 "Although [the successful litigant] may have a reasonable expectation of receiving [costs], this is subject to what some cases have termed the court's absolute and unfettered discretion to award or withhold costs." Id. This discretion is "to be exercised according to the circumstances of each particular case." Id.

Defendants also argue that the court abused its discretion by basing its denial of fees on its finding that defendants' conduct was violative of chapter 93A. They claim that Massachusetts law is irrelevant to Canadian attorney fee determinations and thus that the district judge erred in considering conduct that was violative of Massachusetts law. Defendants misapprehend the district court's reasoning. The point is that their conduct was wrongful, not that it violated a particular statute. In denying the motion, the court noted that "defendants acted intentionally to withhold material information from plaintiffs." Chapter 93A was merely the lens through which the court examined defendants' conduct. Certainly, under the broad discretion Canadian law accords judges in attorney fee determinations, the district judge was entitled to consider this misconduct.

Defendants prevailed on plaintiffs' chapter 93A claim because they successfully asserted that the misconduct occurred primarily and substantially outside Massachusetts. Their motion for attorneys' fees was denied because the district court, exercising its discretion, believed it would be inequitable to award fees to parties whose wrongful conduct had been established at trial. It was entirely proper, in this context, for the district judge to consider defendants' intentional misconduct in deceiving the plaintiffs as a circumstance militating against an award of fees.

For the reasons expressed above, the judgment of the district court in favor of defendants and the denial of defendants' post-trial motion for attorneys' fees are both affirmed. Parties to bear their own costs.