POSNER, Circuit Judge.
This antitrust suit under section 1 of the Sherman Act, 15 U.S.C. § 1, pits a building-materials dealer, Walters, against a manufacturer of prefabricated farm buildings, Morton. The complaint, filed in 1978, alleges that Morton tied building components to its trademark, assigned its franchised dealers (one of whom was Walters) exclusive sales territories, limited the prices the dealers could charge consumers, and eventually took over the dealership function, terminating Walters in the process.. There also is a pendent claim for breach of a duty of fair dealing under Wisconsin law. In 1981, shortly before trial was to begin, Morton moved for summary judgment. The district judge referred the motion to a special master, a member of the Milwaukee bar. After considering for 17 months the voluminous materials that the parties had submitted in connection with the motion, the special master submitted his report, which recommended granting the motion except with respect to the claim regarding exclusive territories. In a brief order the district judge adopted the master’s recommendations in their entirety and therefore dismissed all but the exclusive-territories claim, which he retained for trial. He then directed entry of final judgment for Morton on the dismissed claims upon an express determination that there was no just reason for delay, and Walters has appealed under Rule 54(b) of the Federal Rules of Civil Procedure.
We consider first, on our own initiative as we must, whether the dismissed and retained claims are separate claims within the meaning of Rule 54(b); if they are not, the judgment of dismissal cannot be appealed as a final, judgment and Walters must wait to take its appeal until the exclusive-territories claim is tried. Our recent decisions emphasize that trial judges do not have carte blanche to certify partial dispositions for immediate appeal under Rule 54(b). See A/S Apothekernes Laboratorium v. I.M.C. Chem. Group, Inc., 725 F.2d 1140 (7th Cir.1984); Minority Police Officers Ass’n v. City of South Bend, 721 F.2d 197 (7th Cir.1983). The rule itself makes clear that a district judge may enter an appealable judgment only if it disposes of a “claim for relief” that is “separate” from the claims not disposed of. But unfortunately the meaning that the draftsmen intended “claim for relief” to bear is not clear, see Local P-171, Amalgamated Meat Cutters v. Thompson Farms Co., 642 F.2d 1065, 1069-71 (7th Cir.1981); 10 [702]*702Wright, Miller & Kane, Federal Practice and Procedure § 2657 (2d ed. 1983), and the Supreme Court has not yet attempted a compendious definition. See Sears, Roebuck & Co. v. Mackey, 351 U.S. 427, 76 S.Ct. 895, 100 L.Ed. 1297 (1956); Cold Metal Process Co. v. United Engineering & Foundry Co., 351 U.S. 445, 76 S.Ct. 904, 100 L.Ed. 1311 (1956); Liberty Mutual Ins. Co. v. Wetzel, 424 U.S. 737, 743 n. 4, 96 S.Ct. 1202, 1206 n. 4, 47 L.Ed.2d 435 (1976). The recent decisions of this court have experimented with a practical approach based on what we conceive to be the most important purpose behind Rule 54(b)’s limitations — to spare the court of appeals from having to keep relearning the facts of a case on successive appeals. The approach is: if the facts underlying different claims are different, the claims are separate for Rule 54(b) purposes. For if there are different facts (and of course different issues) consideration of the appeals piecemeal rather than all at once will not involve a duplication in the efforts required of the judges to prepare for argument in, and to decide, each appeal. The gains from forcing the consolidation of appeals will therefore be small and will be outweighed by the benefits of an immediate appeal in resolving the parties’ rights with respect to a part of the controversy between them. By the same token, if there is a great deal of factual overlap between the decided and the retained claims, they are not separate, and appeal must be deferred till the latter are resolved.
Applying this approach to the present case, we find potential but not actual factual overlap between the charges that the district judge dismissed and the single charge, involving exclusive territories, that he retained; therefore we have jurisdiction of the appeal. A charge that limiting the territories in which one’s dealers can sell violates section 1 of the Sherman Act is now evaluated under the Rule of Reason. See Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36, 97 S.Ct. 2549, 53 L.Ed.2d 568 (1977). This means among other things that the plaintiff must show that the defendant has market power, see Valley Liquors, Inc. v. Renfield Importers, Ltd., 678 F.2d 742, 745 (7th Cir. 1982), as this is a prerequisite to being able to restrain trade unreasonably. The Supreme Court’s most recent tying decision requires the plaintiff to make a similar, maybe identical, showing in order to prove that a tying arrangement is unlawful, even though tying still is referred to as a per se offense. See Jefferson Parish Hospital Disk No. 2 v. Hyde, — U.S. -, 104 S.Ct. 1551, 1559-60, 80 L.Ed.2d 2 (1984). Moreover, where as in the present case the plaintiff tries to show that the whole is greater than the sum of its parts — that an “aggregation of trade restraints” is unlawful even if the constituent restraints taken one by one are not — and one of the parts is territorial exclusivity, the issue of unlawful “aggregation” logically requires consideration of the facts concerning territorial exclusivity.
All these are reasons why in the usual case a challenge to a scheme of restricted distribution states only one claim for Rule 54(b) purposes. But this case is unusual in that the particular grounds on which the special master and the district judge disposed of the charges that are involved in the appeal do not overlap, factually or legally, the charge of territorial exclusivity that he retained for trial. A clue to this is that the parties’ briefs in this court tell us nothing about the exclusive territories — how large they were, how exclusivity was enforced, how long exclusive territories were in effect, and for what reasons and with what consequences. Although the voluminous record submitted with the appeal contains much on these questions, that is only because the appeal record is the same record that was before the master and the district judge. The parties have made no reference to the parts of the record that relate to exclusive territories. The district judge dismissed the tying charge not because he found that Morton lacked market power, an issue that as we have said would also be highly relevant to the lawfulness of territorial restrictions, but because he found that the tying and [703]*703the tied products were not separate products and therefore that there could be no tie-in. He dismissed the resale price maintenance and vertical integration allegations (the latter involving Morton’s takeover of the retail distribution of its buildings) on evidentiary grounds unrelated to the lawfulness of exclusive territories. And with all of the specific charges other than the territorial charge dismissed, the “aggregation of trade restraints” charge fell of its own weight, like a conspiracy charge when all but one of the alleged participants have been exonerated.
Since, for these reasons, this appeal is unencumbered by any facts bearing on the retained claim, we are entitled to regard that as a separate claim. Its retention by the district court therefore does not deprive us of jurisdiction over the appeal from the dismissal of the rest of the plaintiff’s ease. This brings us to the merits. We begin with the tie-in charge.
Almost every product can be viewed as a package of component products: a pair of shoes, for example, as a package consisting of a left shoe and a right shoe; a man’s three-piece suit as a package consisting of a jacket, vest, and pants; a belt as a package consisting of a buckle and a strap. As shown by the last of these examples, it is possible to describe a product as a package of components even if the components are physically integrated at the point of sale to the consumer. Indeed, much of what is called “manufacturing” is the assembly, i.e., physical integration, of components, often components manufactured by other companies. The manufacturer of an airplane, for example, does not build the whole thing from scratch. Usually he will build just the airframe, consisting of the fuselage, wings, and tail, and he will buy the engine, the wheels, the radar and other electronic equipment and much else besides from other manufacturers and will then assemble the various components to create the finished aircraft. Very few products today are produced like a statue that is made by pouring molten bronze into a cast — the ultimate single product. But to hold therefore that every composite product is a tie-in, subject to the hostile scrutiny to which antitrust law still subjects tie-ins, would place industry under a vast antitrust cloud, and has been rejected. In Johnson v. Nationwide Industries, Inc., 715 F.2d 1233 (7th Cir.1983), for example, we held that a condominium and a contract for the provision of management services to the condominium owner were a single product, so that the owner could not complain, at least under antitrust law, that the developer who had sold him the condominium had forced him as a condition of being allowed to purchase it to agree to buy the developer’s management services as well. See also Foster v. Maryland State Savings & Loan Ass’n, 590 F.2d 928, 932-33 (D.C.Cir. 1978).
The problem is that there is no obvious way of deciding whether a product is a single product or an assemblage of components. The practice has been to classify a product as a single product if there are rather obvious economies of joint provision, as in the left-shoe-right-shoe example. See Principe v. McDonald’s Corp., 631 F.2d 303, 309-11 (4th Cir.1980); Sullivan, Handbook of the Law of Antitrust 449 (1977); Comment, A New Approach to the Legality of Franchising Tie-Ins, 129 U.Pa.L.Rev. 1267, 1293 (1981). Although this approach seems to take what would otherwise be a matter of defense and make its absence a threshold requirement of the offense, it does serve to screen out many silly cases. A different but related approach was taken recently in the Jefferson Parish case (decided after argument in our case), where the Supreme Court held that products are separate for tie-in purposes if there are separate markets for each product. See 104 S.Ct. at 1562-63. The link to the older view is that if there are not separate markets, this is evidence that the economies of joint provision are overwhelming.
The test of Jefferson Parish disposes handily of the shoe example, but could be thought to place decisions such as Johnson under a cloud, since there are separate markets for condominiums and for real estate management services. Only time will [704]*704tell how far the “separate markets” approach of Jefferson Parish will be pushed. There are separate markets for sugar and for sugarless breakfast cereals, but it would be surprising to find that a sugary-cereal was a tie-in (sugar tied to cereal), assuming the seller refused to sell a sugar-free version. The belt example also becomes problematic under the separate-markets approach. Belts are rarely sold without buckles; but surgical operations are even more rarely sold without anesthesia (held in Jefferson Parish to have been tied to the hospital’s operating rooms). There may be overwhelming economies of joint provision for most customers and yet enough customers with idiosyncratic demands to encourage small markets tailored to their needs to emerge, as has happened with ornamental belt buckles. The separate market for ornamental buckles resembles the separate market for anesthesia, which exists because a patient can contract separately with the surgeon and with the anesthesiologist. We doubt that, even after Jefferson Parish, belts are tie-ins of buckles to straps; yet we cannot be sure where the separate-markets test will lead.
Still, we are reasonably confident that under that test as under the older, more intuitive test a building is a single product and not a tie-in of the walls, floors, roof, windows, and so forth; for most of these components are not sold in separate markets, though some are. And as Walters’ counsel acknowledged helpfully at argument, it does not matter whether, as with hobby kits of various sorts (and even motorcycles, see Flaminio v. Honda Motor Co., 733 F.2d 463 at 465 (7th Cir.1984)), the manufacturer, instead of assembling the components himself at the factory, ships the unassembled components, and the actual assembly is done by the customer or by a dealer (as was the case here, before Morton took over the dealer function). It is enough if the end product is a single product; and a single building is a single product.
All this is more or less common ground among the parties and we have gone over it at such length only to dispel any impression that a tie-in case about prefabricated buildings must be based on a claim that the prefabricated parts were tied to each other. That is not Walters’ claim. Its claim is that the Morton name, the Morton trademark, is the tying product and that the building, the kit of prefabricated building components, is the tied product. The difficulty with this argument is that a product and its name are inseparable. It is one .thing to say that a manufacturer of copying machines who requires his customers to buy from him the copying paper that is used in the machines is conditioning the sale of the machines on the customer’s purchase of a distinct product; it is quite another to say that General Motors lets you use the name Buick on condition that you buy the car to which the name is attached. That is a fantastical description of the transaction, and the cases reject the proposition that a tie-in claim can be based on it. See, e.g., California Glazed Products, Inc. v. Burns & Russell Co., 708 F.2d 1423, 1430 (9th Cir.1983); Dart Industries, Inc. v. Plunkett Co., 704 F.2d 496, 500 (10th Cir.1983); Hamro v. Shell Oil Co., 674 F.2d 784, 788 (9th Cir.1982); Krehl v. Baskin-Robbins Ice Cream Co., 664 F.2d 1348, 1352-54 (9th Cir.1982); Redd v. Shell Oil Co., 524 F.2d 1054, 1056-57 (10th Cir. 1975); Kugler v. AAMCO Automatic Transmissions, Inc., 460 F.2d 1214 (8th Cir.1972). The only appellate case we can find in which such a claim was accepted is Warriner Hermetics, Inc. v. Copeland Refrigeration Corp., 463 F.2d 1002, 1015-16 (5th Cir.1972). To accept it would be to impose in the name of antitrust a regime of compulsory licensing of trademarks — an absurd project. Moreover, since a trademark that denotes a product is rarely licensed apart from the product — and was not in this case — the separate-markets test of Jefferson Parish is not satisfied.
However, some cases, including several in this circuit, do treat as tying products trademarks that name not a product manufactured by the trademark’s owner but a service which he provides to consumers [705]*705through a system of franchised retail outlets; the tie-in consists of requiring franchisees, as a condition of being allowed to use the trademark, to buy distinct products supplied by the franchisor. See, e.g., Photovest Corp. v. Fotomat Corp., 606 F.2d 704, 722 (7th Cir.1979); Ohio-Sealy Mattress Mfg. Co. v. Sealy, Inc., 585 F.2d 821, 834 (7th Cir.1978); Milsen Co. v. Southland Corp., 454 F.2d 363, 365 (7th Cir.1971); Siegel v. Chicken Delight, Inc., 448 F.2d 43 (9th Cir.1971); Susser v. Carvel Corp., 332 F.2d 505 (2d Cir.1964), cert. dismissed, 381 U.S. 125, 85 S.Ct. 1364, 14 L.Ed.2d 284 (1965). The trademark is analytically separable from the tied product, so there is not the same absurdity in treating the trademark as a tying product as there is when the trademark is simply the name of the alleged tied product, as in our example of the Buick name and the Buick automobile. “Chicken Delight,” in contrast, was the name of a fast-food franchise operation, not the name of a product; and the franchisor, the owner of the “Chicken Delight” mark, manufactured none of the things (which included mixes, fryers, and packaging supplies) that it required its franchisees to buy from it as a condition of their being allowed to use the name. Siegel v. Chicken Delight, Inc., supra, 448 F.2d at 48 n. 4. It was a “case of a franchise system set up not to distribute the trade-marked goods of the franchisor, but ... to conduct a certain business under a common trade-mark or trade name.” Id. at 48 (footnote omitted). The name and the tied products thus were separable in a way that the name Buick and the car sold under that name are not. Or so at least it can be argued; for the Chicken Delight line of cases is not universally delectated. See, e.g., Principe v. McDonald’s Corp., supra, 631 F.2d at 309; Baker, The Supreme Court and the Per Se Tying Rule: Cutting the Gordian Knot, 66 Va.L.Rev. 1235, 1315-17 (1980). And the emphasis in the Supreme Court’s recent decision in Jefferson Parish on the importance of proving that the owner of the tying product has real market power may doom the franchise trademark cases, as they mostly involve highly competitive retail industries, such as the fast-food business. But we shall not have to get into the broader questions here. All we shall have to decide is whether Morton Building, the name the defendant has given to its prefabricated farm buildings, is more like Chicken Delight or more like Buick — bearing in mind that since this question was decided on summary judgment we cannot answer it in Morton’s favor if there is any material fact in dispute.
Morton manufactures some building parts, such as sheet-metal walls and wooden corner posts, buys others from other manufacturers, and ships the unassembled parts to retail outlets for assembly into finished buildings on the farmer’s property. The parts when assembled make a rather austere building, even for a farm utility building, and often the dealer will customize the building by adding things that are not part of Morton’s kit. A Morton Building is therefore quite likely to contain some parts not supplied by Morton, just as an automobile might have some accessories that had not been supplied by the manufacturer, such as a dealer-supplied car alarm. But whether few or many farmers embellish their Morton Buildings, the inference that Morton really is engaged in manufacturing buildings is inescapable, and we think this fact defeats any effort to separate its name from the product that it is alleged to be tying to it. Maybe Walters is capable, as it alleges, of building from parts it gets elsewhere a farm building identical to the Morton Building; but it does not follow that it can put Morton’s name on it, when Morton wants to manufacture the Morton Building itself. If Morton were just selling blueprints for a building and it told the people it sold them to, “You must buy your bathroom faucets from us if you want to use the blueprints,” then it might be possible to separate design and product. But if you happen to be in the business of selling faucets, a dealer cannot force you to let him sell, under your name, faucets he gets elsewhere.
To summarize, we hold that there was no tie-in because there were no separate prod[706]*706ucts (or services) tied together. But we emphasize again that it is not alleged that Morton tied accessories to the building kits that it sold, so we need not decide at what point prefabricated building parts cease to be a single product.
The next question is whether Morton violated the per se rule against resale price maintenance. Several times each year Morton would advertise special deals for its buildings. The advertising was directed to the consuming public and mentioned special prices at which consumers could buy the buildings from dealers such as Walters, listing them by name. Morton gave dealers a discount from the wholesale price to make it easier for the dealers to offer consumers the special retail price that Morton had advertised. It was important to Morton that dealers not charge a higher price than the advertised price, because if they did Morton’s advertising would lack credibility — indeed, would be deceptive and create ill will. So, according to Walters’ affidavits, Morton took various steps to see that its dealers sold its buildings at no more than the advertised price. These steps included threatening the dealers with termination if they went above that price, offering if they did to sell directly to the public at the advertised price, and checking up on the dealers to see whether they were charging more. Walters also charges that Morton made it hard for the dealers to comply with its desire that they sell at the advertised price by giving them a smaller discount from the wholesale price than the difference between the regular retail price and the advertised price.
The usual form of resale price maintenance, the form that invited condemnation from the Supreme Court in Dr. Miles Medical Co. v. John D. Park & Sons Co., 220 U.S. 373, 31 S.Ct. 376, 55 L.Ed. 502 (1911), involves setting a minimum price below which dealers may not sell rather than a maximum price above which they may not sell. It is minimum price fixing that creates the analogy to a dealers’ cartel upon which the per se rule against resale price maintenance rests. Although Al-brecht v. Herald Co., 390 U.S. 145, 88 S.Ct. 869, 19 L.Ed.2d 998 (1968), extended the rule of Dr. Miles to a scheme of maximum resale price maintenance, the continued vitality of Albrecht is in doubt after Continental T.V., Inc. v. GTE Sylvania Inc., supra, held that assigning dealers exclusive territories is not unlawful per se. The newspaper dealers in Albrecht had exclusive territories, and putting a ceiling on their prices was, as Justice Stewart pointed out in his dissenting opinion, a way for the Herald Company to prevent the dealers from exploiting the monopoly power that territorial exclusivity gave them. See 390 U.S. at 169, 88 S.Ct. at 881. But at a time when exclusive territories, at least if imposed through a sale contract (as in Albrecht) rather than an agency contract, were illegal per se, as had been held a year before Albrecht in United States v. Arnold, Schwinn & Co., 388 U.S. 365, 87 S.Ct. 1856, 18 L.Ed.2d 1249 (1967) (the decision overruled by Sylvania), it was not a good defense of maximum price fixing that it was ancillary to a system of exclusive territories. Now that assigning exclusive territories to dealers is lawful if reasonable, a manufacturer-imposed price ceiling intended to limit the power that exclusive territories give dealers to raise prices regardless of what other dealers in the manufacturer’s product are charging may also be lawful in some cases — a proposition suggested by Professor Pitofsky in The Sylvania Case: Antitrust Analysis of Non-Price Vertical Restrictions, 78 Colum.L. Rev. 1, 16 n. 59 (1978), and supported by the First Circuit’s decision in Eastern Scientific Co. v. Wild Heerbrugg Instruments, Inc., 572 F.2d 883, 885-86 (1st Cir. 1978). The defendant in that case had a lawful system of exclusive territories but allowed its dealers to sell outside their territories provided they did not sell below a certain price. This provision was held lawful because it allowed a little more competition than if the territories had been airtight. The justification for maximum prices within exclusive dealer territories is even more straightforward: it forces price within the territories closer to the competí-[707]*707tive level. See Easterbrook, Maximum Price Fixing, 48 U.Chi.L.Rev. 886, 890 n. 20 (1981). It is true that, after Sylvania was decided, Albrecht was cited approvingly in Arizona v. Maricopa County Medical Soc’y, 457 U.S. 332, 347-48, 102 S.Ct. 2466, 2474-76, 73 L.Ed.2d 48 (1982) (4-3 decision), but only as supporting the proposition that horizontal price fixing is not saved from condemnation merely because the prices fixed are maximum rather than minimum prices, see id. at 348 and n. 18, 102 S.Ct. at 2475 and n. 18, a matter of no significance here. Indeed, the footnote can be read as supporting a distinction between horizontal and vertical maximum price fixing.
But we need not explore this maze further. Since the lawfulness of Morton’s territorial exclusivity has not been adjudicated, it is premature to conclude that maximum price fixing would be a lawful incident to it. We mention the point to avoid creating the impression that we consider maximum price fixing always and everywhere unlawful (unless horizontal, in which event it is indeed illegal per se, as held in Maricopa, 457 U.S. at 348, 102 S.Ct. at 2475). That seems to us an open question after Sylvania. But even if we assume that maximum price fixing would be unlawful in this case, as it might well be if Morton’s exclusive territories were unlawful (an issue not before us on this appeal), we must still determine whether the conduct attested to in Walters’ affidavits is price fixing. The special master thought that Morton’s efforts to compel its dealers to comply with the advertised price, and the dealers’ responses, did not create an agreement to adhere to that price; and without agreement there can be no violation of section 1 of the Sherman Act. The picture painted by Walters’ affidavits is one not of agreement but of Morton’s unsuccessful efforts to force Walters to agree. However, Supreme Court pronouncements which appear still to be authoritative despite the considerable changes in the Court’s interpretation of the antitrust laws in recent years indicate that when a seller goes beyond announcing a policy of refusing to deal with price cutters (or in this case price gougers) and of cutting off those dealers who do not comply with the announced policy, it has, for section 1 purposes, conspired with any dealer who complies with the policy. See, e.g., United States v. Parke, Davis & Co., 362 U.S. 29, 80 S.Ct. 503, 4 L.Ed.2d 505 (1960). Here some dealers did comply; and Walters’ affidavits if true indicate that Morton went beyond simply announcing its policy and terminating noncompliers, and used persuasion and policing to obtain that compliance, as in Yentsch v. Texaco, Inc., 630 F.2d 46, 49-55 (2d Cir.1980). See also Monsanto Co. v. Spray-Rite Service Corp., — U.S, -, 104 S.Ct. 1464, 1471-73, 79 L.Ed.2d 775 (1984).
But there is a more fundamental ground, alluded to by the special master, which requires that we uphold the dismissal of this part of the complaint. Walters concedes as it must that it is perfectly lawful for a manufacturer to advertise his product to the ultimate consumer, whether or not he sells directly to the consumer, and to mention in that advertising the retail price of the product — the only price the consumer is interested in. The law of products liability has come to recognize what economists and businessmen have long known — that manufacturers are in a direct economic relationship with the ultimate consumers of their products even if they are not in privity with them. See, e.g., Henningsen v. Bloomfield Motors, Inc., 32 N.J. 358, 412-16, 161 A.2d 69, 99-101 (1960). This insight cannot be limited to tort law. The demand for a manufacturer’s product depends significantly on the retail price; the lower that price is (other things being equal), the more his product will be demanded by consumers and therefore by the dealers who are his immediate purchasers. Hence the manufacturer has a vital interest in the retail price. Maybe that interest is not great enough to allow him to place a cap on the retail price but Walters concedes that it is great enough to allow the manufacturer to advertise a retail price.
[708]*708This concession is very damaging to Walters. If it is lawful to advertise a retail price, it should be lawful to take at least the minimum steps necessary to make that advertising beneficial. It would be pretty embarrassing for a manufacturer who had advertised a special retail price to be bombarded by complaints from consumers that dealers were refusing to sell to them at that price. Such refusals would make the advertising misleading and might even expose the manufacturer to legal sanctions under the Federal Trade Commission Act or counterpart state regulations. So if retail price advertising by manufacturers is to be feasible the manufacturer must be allowed to take reasonable measures to make sure the advertised price is not exceeded. These measures include trying to persuade dealers to adhere to the advertised price and checking around to make sure they are adhering. These are the respects in which Morton is alleged to have gone beyond the simple announcement of policy and refusal to deal with noncompli-ers that would be permissible even if it were trying to get its dealers to adhere to its suggested retail prices across the board. See Monsanto Co. v. Spray-Rite Service Corp., supra, 104 S.Ct. at 1469. They are the minimum steps that Morton had to take if its advertised price was to have any value at all, and they are therefore lawful. Cf. Bruce Drug, Inc. v. Hollister, Inc., 688 F.2d 853, 859-60 (1st Cir.1982).
It is true that dicta in Butera v. Sun Oil Co., 496 F.2d 434, 437 (1st Cir.1974), based on Pearl Brewing Co. v. Anheuser-Busch, Inc., 339 F.Supp. 945 (S.D.Tex.1972), suggest that for a supplier to condition a cut in the wholesale price to his dealers on their agreeing to pass it on to the consumer by lowering their prices, which is a little like what is alleged here, is a form of price fixing. But the more recent decisions hold to the contrary. See Lewis Service Center, Inc. v. Mack Trucks, Inc., 714 F.2d 842, 846 (8th Cir.1983); AAA Liquors, Inc. v. Joseph E. Seagram & Sons, Inc., 705 F.2d 1203, 1206 (10th Cir.1982). And in none of these cases did the manufacturer have the additional argument that Morton has here: Morton had to pressure its dealers to lower price in order to maintain the credibility of its price advertising — a form, by the way, of constitutionally protected speech. See Virginia State Bd. of Pharmacy v. Virginia Citizens Consumer Council, 425 U.S. 748, 96 S.Ct. 1817, 48 L.Ed.2d 346 (1976) .
All this is not to say that simply by advertising retail prices a manufacturer gets the right to fix retail prices. If Morton had tried to fix its dealers’ prices at times when it was not advertising special deals at specific prices, its efforts would not be privileged just because at other times it did advertise retail prices. Nor could Morton require its dealers to comply with minimum prices. And therefore it could not forbid them to sell at less than the advertised prices. See Lehrman v. Gulf Oil Corp., 464 F.2d 26, 39-40 (5th Cir.1972). But it did not do any of these things.
Even if what Morton did was price fixing, Walters could not challenge it. A private plaintiff can complain only of an antitrust injury. Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 489, 97 S.Ct. 690, 697, 50 L.Ed.2d 701 (1977) . It is not enough that the plaintiff would not have been injured but for the defendant’s alleged violation of the antitrust laws; the injury must be the sort of thing that the antitrust laws are intended to forbid. The plaintiff in Brunswick was complaining that the defendant’s acquisition of a rival would enable the rival to become a more efficient company, and therefore a more effective competitor of the plaintiff. The Supreme Court held that the plaintiff could not complain about the injury. It would have been different if the complaint was that the acquisition would lead to the rival’s engaging in predatory pricing or some other method of competition with the plaintiff that the antitrust laws are intended to prevent.
There is nothing esoteric about the Brunswick rule. It is the application to antitrust law of venerable principles of tort [709]*709causation illustrated by Gorris v. Scott, 9 L.R. Ex.-125 (1874). The plaintiffs animals, which were being transported on the deck of the defendant’s ship, were washed overboard in a storm. They would have been saved if the deck had been penned, as required by statute. But since the purpose of the statute was to prevent contagion, not drowning, the defendant was not liable. Similarly, although the plaintiff in Brunswick would not have been hurt if there had been no acquisition, the acquisition did not hurt it in the way that the framers of the antitrust laws had in mind. In the present case, even if Morton did violate the prohibition against fixing its dealers’ prices, the only harm to Walters came from the fact that competing dealers (or Morton itself) would lower their prices to consumers if Walters did not. There is no suggestion that the lower prices would have been below cost; they would have been lawful prices. This case thus is stronger for the defendant than Gorris. The loss of the animals was a pure social cost, while the loss to Walters from lawful price competition was a gain to consumers. Walters will not be heard to complain about having to meet lawful price competition, which antitrust law seeks to encourage, merely because the competition may have been enabled by an antitrust violation. See MCI Communications Corp. v. American Tel. & Tel. Co., 708 F.2d 1081, 1161 (7th Cir.1983); Page, Antitrust Damages and Economic Efficiency: An Approach to Antitrust Injury, 47 U.Chi.L.Rev. 467, 491 (1980).
The remaining antitrust claim that the district court dismissed is that the tie-in, resale price maintenance, and exclusive-territories allegations, taken together, make out an independent violation of section 1 of the Sherman Act, because they establish a scheme to destroy Morton’s independent dealers in order that Morton could take over the retail sale of its buildings, as eventually it did. Our conclusion that there was no tie-in and no resale price maintenance may seem to eliminate all possibility that there was an unlawful scheme of greater compass than the territorial exclusivity, which is not before us on this appeal. But it is possible to recast Walters’ argument as follows. Even if there was no tie-in in a technical sense and no resale price maintenance in the sense in which it is a per se violation of section 1, still when one combines what Morton is alleged to have done (such as forbidding its dealers to obtain, from other manufacturers, any Morton Building components that Morton supplies, and squeezing the dealers by forcing them to sell buildings at the low prices that Morton advertised and to swallow part of the discount) with its alleged motive in doing so of driving the dealers out of business, an unreasonable restraint of trade can be inferred. And it meets the threshold requirement of section 1 that there be a contract, combination, or conspiracy, since Morton managed to draw some of its (doomed) dealers into its scheme to eliminate recalcitrants such as Walters.
It does not add much, though, to call the scheme an “aggregation of trade restraints.” The term goes back to Timken Roller Bearing Co. v. United States, 341 U.S. 593, 598, 71 S.Ct. 971, 974, 95 L.Ed. 1199 (1951), and as used there, and later in United States v. Sealy, Inc., 388 U.S. 350, 354, 357, 87 S.Ct. 1847, 1851, 1852, 18 L.Ed.2d 1238 (1967), just meant that whether a restriction was illegal per se might depend on what other restrictions the defendant had imposed. A later decision in the interminable Sealy litigation, Ohio-Sealy Mattress Mfg. Co. v. Sealy, Inc., supra, 585 F.2d at 827-28, used a similar approach to conclude that a number of separate provisions in the Sealy licensing agreement, each innocuous in itself, had been combined in such a way as to perpetuate the division of markets that the Supreme Court had held was a per se violation of section 1. All -that Walters is asking us to do here is to decide whether its challenge to Morton’s system of marketing raises a triable issue under the Rule of Reason.
It is helpful to recall that at common law a conspiracy was unlawful if the conspirators used unlawful means to a law[710]*710ful end or lawful means to an unlawful end. United States v. Addyston Pipe & Steel Co., 85 Fed. 271, 293 (6th Cir.1898), modified and aff’d, 175 U.S. 211, 20 S.Ct. 96, 44 L.Ed. 136 (1899). This is a pretty encompassing definition, and if Walters cannot bring itself within either branch it is unlikely to have a cause of action under section 1. The means, we have seen, were not unlawful, unless the exclusive territories are unlawful; but if they are, Walters can get all the relief it seeks in the trial of the retained claim in the district court. Predatory pricing, alleged in Coleman Motor Co. v. Chrysler Corp., 525 F.2d 1338, 1345 (3d Cir.1975), on which Walter relies as a case where the goal of the alleged conspiracy was vertical integration into retail distribution, is not alleged here.
The end that Walters alleges is that Morton wanted to take over the retail function; in the terminology of industrial organization, it wanted to integrate forward. But vertical integration is not an unlawful or even a suspect category under the antitrust laws: “Firms constantly face ‘make-or-buy’ decisions — that is, decisions whether to purchase a good or service in the market or to produce it internally — and ordinarily the decision, whichever way it goes, raises no antitrust question.” University Life Ins. Co. of America v. Unimarc Ltd., 699 F.2d 846, 852 (7th Cir.1983). Morton manufactures some of the building components that it includes in the kits that it sells. It is therefore vertically integrated, because it could buy those components on the market instead. There is nothing abnormal about this. If Morton employs its own typists it is vertically integrated, because it could farm out its typing work to independent contractors. When a corporation that has been using a law firm to handle a particular type of litigation hires a lawyer to do the litigation in-house, it is vertically integrating into litigation services. When a law firm that has been buying a billing service from a computer time-sharing firm buys its own computer to perform the service, it is vertically integrating into computer services. Vertical integration is a universal feature of economic life and it would be absurd to make it a suspect category under the antitrust laws just because it may hurt suppliers of the service that has been brought within the firm.
A common type of vertical integration is for a manufacturer to take over the distribution of his product. Some manufacturers use wholesalers; others, for example the U.S. automobile manufacturers, do their own wholesaling and thus are vertically integrated into wholesale distribution. Some firms, IBM for example, sell their products directly to the public, rather than through independent retail outlets, and thus are vertically integrated into retail distribution. It was not improper for Morton to want to deal directly with the consuming public rather than indirectly through dealers such as Walters. Of course if it set about to do this in an unlawful fashion it would not be immune from liability just because it had an unexceptionable end in view. But that is only because, as we have noted, a conspiracy to attain a lawful end by unlawful means is actionable.
We just said that vertical integration is not an improper objective. But this puts the matter too tepidly; vertical integration usually is procompetitive. If there are cost savings from bringing into the firm a function formerly performed outside it, the firm will be made a more effective competitor. Moreover, the option of vertical integration places competitive pressure on the firm’s suppliers and buyers, who know that if they charge too much for their services the firm may decide to perform them itself. It thus increases competition in the markets for those services. Morton claims that the spiraling costs of building materials forced it to cast about for methods of economizing on its operations, and one method it hit on was to do its own retailing. This is a common motivation for vertical integration. True, some economists believe that monopolistic firms might integrate vertically in order to deny supplies or outlets to competitors, or to make it more costly for new firms to enter the [711]*711market (because they would have to enter at more than one level of production or distribution), or to facilitate price fixing with their competitors. See 3 Areeda & Turner, Antitrust Law ¶¶ 725-26 (1978). But nothing of this kind is suggested here. Walters does allege that Morton has a big name in the prefabricated farm buildings market, but there is no indication that this is a meaningful economic market that might be worth monopolizing, or that Morton’s purpose in integrating into retail distribution was to make life harder for its competitors. Its object was to make more money by reducing the cost of retail distribution, not by coercing or excluding (or for that matter colluding with) its own competitors, whoever they may be, or discouraging potential competitors. Indeed Walters’ tie-in claim is premised on the ready availability, from other manufacturers, of the building parts that Morton sells in kits from which Morton Buildings are put together. This shows that Morton has no monopoly.
The final substantive question is whether the district court was correct to dismiss Walters’ claim that Morton breached its duty under Wisconsin law to deal fairly with a party with which it had a contractual relationship. The dismissal was on the merits; and if it were not for the fact that Walters and Morton are of diverse citizenship and that the amount in controversy between them exceeds $10,000, there would be a question whether retention of this state-law claim after all the federal claims except the challenge to Morton’s use of exclusive territories were dismissed before trial was a proper exercise of pendent jurisdiction; for the relationship between the pendent claim and the only retained federal claim is somewhat distant.
Walters does not charge Morton with a breach of contract, in part because Walters has been unable to produce a copy of any written contract though it claims there was one. The charge sounds in tort. But we share the special master’s perplexity as to what tort is being charged. In Wisconsin as in many other states it is tortious for an insurance company to refuse to pay an insured his claim if the company has no colorable basis for its refusal. See, e.g., Kanzush v. Badger State Mutual Casualty Co., 103 Wis.2d 56, 61-62, 307 N.W.2d 256, 259-60 (1981); Anderson v. Continental Ins. Co., 85 Wis.2d 675, 687, 271 N.W.2d 368, 374 (1978). But the tort requires a contractual claim so incontestable that the refusal to honor it must have been in bad faith; here, if Walters has any contractual claim at all, it is so tenuous that Walters is not asserting it. Also the tort has not been extended, in Wisconsin anyway, beyond ■ the insurance setting, where the disparity in resources between insurer and insured has seemed to make the bad-faith breaking of contracts especially reprehensible. Indeed, even in that setting the tort has been defined narrowly, as we pointed out recently in A.W. Huss Co. v. Continental Casualty Co., 735 F.2d 246, 249-54 (7th Cir.1984).
Fiduciaries have a duty of fair dealing, of course; and in Estate of Chayka, 47 Wis.2d 102, 176 N.W.2d 561 (1970), this duty was extended to spouses executing a joint will — a special relationship resembling a fiduciary one. But it is a big step from this to reading into every commercial relationship a duty of fair dealing not cabined by contractual provisions, express or implied. Statutes in many states, including Wisconsin, give dealers and franchisees special protections against termination. See Wis.Stat. § 135.03. But Walters does not rely on any statute; it relies on a common law tort concept that is not yet the law in Wisconsin. The common law of Wisconsin does not make the franchisor a fiduciary of his franchisees. Murphy v. White Hen Pantry Co., 691 F.2d 350, 355 (7th Cir.1982).
This completes our discussion of the matters presented in the appeal, but before ending we want to make an observation about the procedure employed by the district court in this case. Although the use of special masters to assist federal district judges in complex litigation is not only authorized but indispensable to the effective operation of the federal courts under [712]*712current workload conditions, its scope is limited by Fed.R.Civ.P. 53(b), as follows: “A reference to a master shall be the exception and not the rule. In actions to be tried by a jury, a reference shall be made only when the issues are complicated; in actions to be tried without a jury, save in matters of account and of difficult computation of damages, a reference shall be made only upon a showing that some exceptional condition requires it.” The reference in this case was proper only if required by “some exceptional condition,” since it was not made to assist a jury and did not call for an accounting or a damage calculation; and we doubt that this demanding condition was met. Cf. Benner-son v. Joseph, 583 F.2d 633, 642 (3d Cir. 1978). In this more than in any other circuit we must be alert to the danger of overusing special masters, for it was overuse by one of our district judges that led the Supreme Court to issue a writ of mandamus in La Buy v. Howes Leather Co., 352 U.S. 249, 77 S.Ct. 309, 1 L.Ed.2d 290 (1957). Judge La Buy “stated he did not know when he could try the case ‘if it is going to take this long.’ He asked if the parties could agree ‘to have a Master hear’ it. The parties ignored this query and the next day [Judge La Buy] entered the orders of reference sua sponte. The orders declared that the court was ‘confronted with an extremely congested calendar’ and that ‘exception [sic] conditions exist for this reason’ requiring the references. The cases were referred to the master ‘to take evidence and to report the same to this Court, together with his findings of fact and conclusions of law’____ [A]ll parties objected to the references ____” Id. at 253-54, 77 S.Ct. at 312 (footnote omitted).
In this case the district judge, also on his own initiative and also because he did not have time for a long trial, and hoping that the special master’s report would (as it did) enable the number of issues to be reduced for trial, referred summary judgment proceedings to a special master who prepared a report that the judge then adopted without independent analysis. Since the judge did not refer the case to the special master for trial, and only one of the parties (Walters) objected, the delegation of judicial power was less extreme than in La Buy. But that does not mean it was in conformity with Rule 53(b). Although the materials submitted in connection with the motion for summary judgment ran to several thousand pages, this unfortunately is not unusual in antitrust and other complex litigation; yet it would be contrary to Rule 53(b), as it has been understood and interpreted against a background of concern with the cost and delay created by using masters, to refer summary judgment motions in antitrust cases routinely to masters. See 9 Wright & Miller, Federal Practice and Procedure § 2605 (1971); cf. MCI Communications Corp. v. American Tel. & Tel. Co., supra, 708 F.2d at 1169 n. 125.
But unless the delegation of judicial power to the master in this case exceeded the bounds of Article III of the Constitution, it is not reversible error, because Walters has not asked us to reverse the district court’s judgment on the ground that the reference to the master violated Rule 53(b). See, e.g., Hayes v. Foodmaker, Inc., 634 F.2d 802, 803 (5th Cir.1981) (per curiam); 9 Wright & Miller, supra, § 2606. The objection was made in the district court but has been abandoned on appeal, so we would not revive it unless we thought the reference was beyond the district court’s subject-matter jurisdiction. The district judge delegated a reporting and recommending power rather than the power of decision; it was he, not the master, who entered judgment. True, the master’s function was to evaluate a long documentary record and determine whether it presented genuine issues of material fact; the opinion he wrote was indistinguishable in form from a judicial opinion; and the district judge approved that opinion rather than write his own. Walters therefore has a valid gripe about being hustled off to an adjudicative process conducted — and at rather too leisurely a pace — by a private practitioner. If Walters had preserved its objection on appeal we might well reverse, [713]*713and we would have to do so whether it had or not if we thought the procedure was so improper as to be beyond the powers granted federal judges by the Constitution, statutes, and rules — which means, practically speaking, so egregious as to warrant reversal even though the issue was not preserved on appeal. We do not think the procedure was quite that egregious, and we naturally are reluctant to prolong the length and increase the expense of this proceeding (already six years old) by vacating the reference on our own initiative. But we trust that in the future the use of special masters in antitrust and other complex litigation will be conducted with greater sensitivity to the problems of excessive delegation of judicial power that such use can create and to the precise language of Rule 53(b), which as we have pointed out requires in a case such as this a showing of “some exceptional condition” justifying the reference to a master.
Affirmed.