In these consolidated appeals we must decide whether a restrictive covenant in a contract of employment is enforceable under state law, and whether its attempted enforcement constituted a violation of sections 1 and 2 of the Sherman Act.1 We find that the restrictive covenant is enforceable and not a violation of section 1 of the Sherman Act; we do not address the substance of the section 2 issue because the issue has not properly been framed for appellate review.

Butler Service Group, Inc. (Butler), and Consultants and Designers, Inc. (C & D), are two of a large number of firms in the technical service industry. This industry may be viewed as a variant of the employment agency industry. Firms in this industry serve as middlemen and providers of information in the market for highly skilled, relatively mobile technically trained workers in fields such as engineering, designing, drafting, and data processing. These workers are generally referred to as "job shoppers." A technical service firm (TSF) recruits job shoppers from a national and in some cases an international market.

The job shoppers are recruited to service the needs of client firms. The client is likely to have a need for short-term, highly skilled technical workers that it cannot satisfy from its local area. Though the employment offered to job shoppers is temporary, it may last for several years. The TSF specializes in bringing together firms seeking these kinds of workers with individuals seeking this sort of work.

Job shoppers, though they work on the client's premises and under its supervision, are paid by the TSF and are in that sense employees of the TSF. The TSF earns its income by charging the client firm a price per employee per pay period which exceeds what it must in turn pay the employee. Unlike a traditional employment agency the TSF receives no finder's fee from either the job shopper or the client firm for the initial placement of the employee with the client. Because of the diminished job security and the lack of other employee benefits, client firms, in order to attract job shoppers, must be willing to pay the job shopper approximately thirty percent more than they pay their direct employees performing the same task.

TSFs generally structure their relationships with job shoppers and client firms through contract. In serving as an intermediary in the market between job shoppers and client firms, TSFs must not only compete with one another, but face competition from employment agencies and from disintermediation, i.e., firms and workers searching for and contracting with one another directly without a TSF acting as an intermediary.

In the early 1970's, the Tennessee Valley Authority (TVA) embarked on a program calling for the design and construction of several nuclear power plants. By 1979, TVA had thirteen of these plants under construction in the states of Tennessee and Alabama, and its nuclear program was the largest in the United States. To complete its program and to augment its permanent staff, TVA turned to TSFs to supply contract personnel for the design and construction of its nuclear power plants.

In 1976, Butler obtained a contract to supply technical employees to TVA. The contract between Butler and TVA provided that TVA would pay Butler an hourly wage to be received by each employee, plus a markup. These payments were "[d]eemed to cover all wages and salaries; all general, administrative, and overhead expenses; any and all other direct or indirect costs or expenses, including Workmen's Compensation, Federal and State Unemployment Insurance and FICA, and profit." The contract provided that during its term neither TVA nor Butler would hire the other's employees. The contract contained no restriction on TVA's solicitation or hiring of job shoppers after the expiration of the contract. The contract was terminable by either party on seven days' notice. The contract was extended a number of times. The final extension commenced in April 1979, and continued the agreement until May 3, 1980.

The relationship between Butler and its job shoppers was similarly constrained and conditioned by contract. Butler's contract with its job shoppers working at TVA contained the following clause:

Employees shall not accept employment directly or indirectly by client for a period of ninety (90) days following the completion of assignments to said client and/or assignment to the work of said client performed, on employer's premises out in the field without the written consent of employer.

In 1979, TVA requested a number of TSFs, including Butler and C & D, to submit proposals for contracts to supply employees to TVA upon the expiration of Butler's contract in May 1980. In March 1980, TVA decided to award the new contracts to C & D and H.L. Yoh Co., and advised Butler that its contract with TVA would end on May 3, 1980. After receiving notice that its contract would be terminated, Butler, in April 1980, notified each employee assigned to TVA that its contract with TVA was being terminated and that it would enforce its restrictive covenant and would attempt to find a new job for every employee who was interested. Butler gave C & D, Yoh, and CDI Corporation2 notice of its restrictive covenant and of its intent to enforce the covenant. Butler offered to enter into a "bridge or extended contract" with TVA during which C & D and Yoh would replace the Butler employees on an individual basis as they could. Butler also offered to the successor TSFs to negotiate a waiver of its claimed rights under the restrictive covenant.

Having learned that C & D had begun soliciting Butler's TVA employees, and that both C & D and Yoh intended to staff positions under their TVA contracts by employing Butler employees at TVA, Butler sued C & D and Yoh in the United States District Court for the Southern District of New York. Butler sought damages for tortious interference with its contracts with its employees and an injunction to prevent C & D and Yoh3 from attempting to hire away Butler's job shoppers at TVA. Butler applied for an ex parte temporary restraining order, which was issued on May 6, 1980.

After the restraining order was entered, TVA accepted Butler's offer of a bridge contract of ninety days duration terminable on seven days' notice. This contract was intended to keep Butler and its job shoppers in place for a period of time during which C & D and Yoh would gradually replace Butler's job shoppers with their own. On May 22, 1980, the district court, having been informed of the bridge contract, concluded that Butler's claim for injunctive relief was moot and therefore dissolved the temporary restraining order.

On June 9, 1980, C & D and several former Butler job shoppers (referred to collectively as C & D) brought an antitrust action in the Northern District of Alabama seeking injunctive relief and damages under the theory that Butler's attempted enforcement of its restrictive employment covenant violated sections 1 & 2 of the Sherman Act.4 On June 11, 1980, C & D amended its answer in the New York case to include a counterclaim alleging tortious interference by Butler with C & D's contract with TVA, abuse of process, malicious prosecution and unjust enrichment. C & D, joined by Yoh, moved the New York court to transfer the case to the Northern District of Alabama, and on August 29, 1980, the court granted the motion. See 28 U.S.C. Sec. 1404(a) (1976).

On February 10, 1981, the court below consolidated the New York and Alabama cases, and set the cases for trial on the limited question of the validity and enforceability of Butler's restrictive employment covenant under state law and section 1 of the Sherman Act. A bench trial was held in July and on November 3, 1981, the court handed down a memorandum opinion, in which it found that the covenant was unenforceable under state law, but that neither the covenant nor its attempted enforcement violated section 1 of the Sherman Act.

The court entered a partial final judgment, in conformance with its opinion, under Fed.R.Civ.P. 54(b). The court also denied Butler's motion for summary judgment on C & D's claim under section 2 of the Sherman Act and certified its ruling for appeal under 28 U.S.C. Sec. 1292(b) (1976). All parties appealed.

Three related issues are raised for our review in this appeal. Each of these issues revolves around the restrictive covenant in the contract between Butler and the job shoppers it hired and employed at the TVA's work sites. The first is whether this restrictive covenant is enforceable under state law. The second is whether it constitutes a "contract ... in restraint of trade" under section 1 of the Sherman Act. The third is whether, given the factual record already established and our analysis of those facts, Butler is entitled to summary judgment on C & D's claim under section 2 of the Sherman Act that Butler "monopolize[d], attempt[ed] to monopolize, or conspire[d] ... to monopolize" the relevant part of the technical service industry.

The parties agreed at the trial that the validity and enforceability of Butler's restrictive covenant was to be governed as to the Alabama employees by the stricter of Alabama and New Jersey law, and as to the Tennessee employees by the stricter of Tennessee or New Jersey law. We concur in the finding of the court below that there is no significant difference in the law of the three states. The district court correctly stated that:

Notwithstanding the vastness and variety of cases, the touchstone of this analysis has not varied since the grandfather case of Mitchell v. Reynolds cited as 1 P.Wms. 181, 24 Eng.Rep. 347 (O.B.1711).... Lord Macclesfield applied a "reasonableness" or balancing test which has survived, with many enlightening applications, to this day. We thus sail into that cherished legal haven, "Is it reasonable under the circumstances?"

In judging reasonableness, the district court looked both to the Restatement (2d) of Contracts, Sec. 188, and to the appropriate case law. The Restatement fleshes out the legal standard a bit. It states that the covenant "[i]s unreasonably in restraint of trade if (a) the restraint is greater than is needed to protect promisee's legitimate interest, or (b) the promisee's need is outweighed by the hardship to the promisor and the likely injury to the public." As the district court correctly noted, "[p]recedents are of less than usual value because the question of reasonableness must be decided on an ad hoc basis. There is no inflexible formula. Allright Auto Parks, Inc. v. Barry, 219 Tenn. 280, 409 S.W.2d 361 (1966)."

It is clear that the district court was correct in its finding that the time and area restraints are not unreasonable. The covenant is limited to a period of ninety days and to the specific client worksite at which Butler placed the individual employee. It would be difficult to conceive of a more limited post-employment covenant. Nonetheless, the covenant cannot pass muster if it fails to meet the reasonableness standard enunciated in the Restatement.

In evaluating the reasonableness of any covenant not to compete, it is essential that we determine in what business the promisee is engaged. We cannot otherwise determine whether the employer is attempting to prevent ordinary competition, something which "[t]he employer has no legitimate interest" in preventing, Whitmyer Bros. v. Doyle, 58 N.J. 25, 274 A.2d 577, 581 (1971), or whether he has some legitimate protectible interest.

While in a technical sense the relationship between Butler and its job shoppers was employer-employee, this characterization of the relationship does more to obscure the relevant issues than to enlighten them. It is more revealing to recognize Butler's role as that of the much maligned but time-honored middleman. Butler served as an intermediary in the market for highly-skilled, technically-trained workers. There were individuals seeking such employment and firms seeking to employ them. In a pristine world in which information is costless, there would be no need for the essential service that Butler performs. The client firms and the prospective employees would each know what the other has to offer and could negotiate a contract instantaneously and costlessly. That, however, is not the nature of the world in which we live. Butler and C & D, and the fifteen other firms that applied for the contract with TVA, as well as a myriad of other firms in the technical service and related industries, exist and flourish because information is a costly and valuable commodity.

Whatever ancillary services Butler supplied to TVA, it was as a provider of information that Butler performed its essential function. Butler's role was to gather, distill, and provide to TVA information on available and suitable people for TVA's positions and simultaneously to provide to prospective job shoppers information about TVA and the positions it was offering. Thus, in this sense, Butler was serving as a form of an employment agency. In the market for the information in which Butler dealt, it had to compete with: (1) all other technical service firms; (2) conventional employment agencies; and (3) disintermediation by firms and workers. Disintermediation is the actualization of the ever-present cry to eliminate the middleman, i.e., direct solicitation, negotiation and contracting between the firm and the worker. Eliminating the middleman is at first blush a facile and attractive alternative. However, middlemen exist because they provide a useful and highly-valued service. If Butler's covenant with its job shoppers is to have any justification, that justification must be to protect Butler's role as a middleman in the market for information between job shoppers and client firms.

Once (1) a job shopper has been informed about a client and vice versa; (2) the job shopper has expressed a desire to work for the client and vice versa; and (3) the employment has commenced and each is satisfied with the other, the primary mission of the technical service firm is complete. At that point, barring some form of contractual constraint, the client firm and the job shopper could sever their relationships with the TSF and consummate a relationship with one another. The job shoppers and client firms could thereby opportunistically appropriate the work product of the TSF without paying it the full value of its services.5 If the TSF is to receive a reward for the services of the job shopper proportional to the latter's contribution to the client firm, the TSF must find some contractual means to protect its future income stream from the ravages of opportunistic disintermediation.

The covenant between Butler and TVA not to hire one another's employees and the restrictive covenants in Butler's contracts with its job shoppers were clearly crafted to prevent such disintermediation. But for these contractual constraints the job shoppers and the TVA could get the benefit of Butler's services without paying the full price of those services. The application of Butler's restrictive covenant after the termination of its contract with TVA was designed to protect against TVA terminating Butler and then hiring either directly, or indirectly through another technical service firm, the Butler job shoppers. TVA had the power and the right to terminate Butler, but not to abrogate Butler's property interest in its job shoppers. Butler's rights under its employment contracts need not have been enforced. They could have been sold. Regardless of whether Butler would have struck a deal with the job shoppers, TVA, or the successor TSF, Butler was nonetheless protected by the covenant from opportunistic termination of the relationship by TVA. Therefore we conclude that Butler had a legitimate interest in protecting from opportunistic appropriation its investment in acquiring the information necessary to carry on its business, and that the covenant was reasonably well crafted to carry out that task. This satisfies the first prong of the Restatement's reasonableness test.

As to the second prong, whether either the job shoppers who are held to this covenant, or third parties, including the public at large, are sufficiently adversely affected by the covenant to render it unenforceable, the answer must be no. The TVA and all the parties to this action, with the exception of Butler, would gain if the restrictive covenant is held to be unenforceable. If the covenant is enforced, Butler has a property interest in the ninety days of employment of its ex-job shoppers at TVA. TVA, the job shoppers, and C & D could each gain by the extinguishment of this property right. In that sense, Butler's loss must be someone else's gain, and Butler's gain is someone's loss. Such a pecuniary loss to C & D, the TVA, or the job shoppers cannot serve to invalidate an otherwise valid covenant. The value of this loss has already been compensated for presumably in the consideration for the contracts negotiated between Butler and the TVA, Butler and its job shoppers, and C & D and the TVA.

The essence of the inquiry under the second prong of the Restatement should be directed at the effect of the covenant on the general public. The proper posture from which this inquiry is to be undertaken is not after the termination of Butler by TVA (ex post), but rather before the formation of the contracts between Butler, on the one hand, and its job shoppers and TVA on the other (ex ante). In precisely the same spirit that it was in TVA's and the job shoppers' interest to eliminate Butler after Butler had performed its services, and before it had received all the income from those services, so too it was in the public's interest at that time that the covenant not be enforced. The public suffers a loss as a result of the enforcement of the covenant. The loss is not, however, that the nuclear power plant at which the job shoppers were employed will suffer any delay in completion. Rather, the loss to the public, if any, will come in the form of higher electric rates. The covenant grants a property right to Butler which would doubtless have been negotiated away, resulting in a wealth transfer from one or more of the other concerned parties to Butler. The cost of purchasing this waiver might ultimately have been passed on to TVA consumers.

In applying the second prong of the Restatement test, and assessing the effect of this restrictive covenant on the public at large, we must focus on its effect at the time the contract was formed. The question before us is whether on balance the public gained or lost as a result of Butler's ability to secure itself from opportunistic behavior on the part of its client TVA and its job shoppers working there. There is some optimal investment for society in the resources required to find and place technical workers at places such as TVA. The goal of the legal system in this regard is to provide a framework and structure out of which the incentives to the individual firms making the decisions will correspond to those required to achieve the optimal investment for society. If the firms in the industry can receive the benefits of investment without paying the costs they will tend to overinvest. On the other hand if they pay the costs but cannot be assured of receiving the benefits they will tend to underinvest.

The protection that Butler believed it acquired as a result of this restrictive covenant permitted Butler to invest at a level closer to the optimum in the assets required to find personnel for TVA. By protecting the future stream of income that Butler would receive from its job shoppers, Butler was induced to invest more in searching for job shoppers than it otherwise would. Thus the public at large can be expected to gain from the enforcement of this and similar restrictive covenants to the extent that these covenants encourage optimal investment. The covenants therefore survive the second prong of the test provided by the Restatement supra.

Our conclusion is not that the enforcement of this restrictive covenant is the only or even the best way for Butler to protect its legitimate interest in the future employment of its job shoppers. Nor does our conclusion that it is reasonable under the circumstances imply that it is in Butler's ultimate self-interest to employ this covenant. Such a determination would be beyond the intellectual power of this or any other court. Ultimately it is the market which will be the final arbiter of the efficiency, or lack thereof, of this covenant. If Butler should persist in offering this covenant and its competitors do not, the market will have the opportunity to choose between them. What we are dealing with are contracts made between and among consenting adults and corporations. Presumably they will act in such a way as to maximize their individual welfare, and it would be presumptuous and harmful if we were to substitute our ex-post judgment for their ex-ante choice. Therefore we hold that the restrictive covenant in the contracts between Butler and its job shoppers at TVA is enforceable under state law. We reverse the district court's contrary holding and remand for further proceedings consistent with this opinion.6

C & D argues that Butler's enforcement of its restrictive employment covenant is in violation of section 1 of the Sherman Act either under a per se theory or as a result of rule of reason analysis. Even under the rather peculiar standards of this much abused cause of action, C & D's claim is both bizarre and frivolous.

Ever since the decision of the Supreme Court in United States v. Addyston Pipe & Steel Co., 85 F. 271 (6th Cir.1898), modified, 175 U.S. 211, 20 S. Ct. 96, 44 L. Ed. 136 (1899), there has been an unbroken line of cases holding that the validity of covenants not to compete under the Sherman Act must be analyzed under the rule of reason. In United States v. Empire Gas Corp., 537 F.2d 296, 307-08 (8th Cir.1976), cert. denied, 429 U.S. 1122, 97 S. Ct. 1158, 51 L. Ed. 2d 572 (1977), the court held that restrictive covenants in employment agreements were not a per se violation of section 1 of the Sherman Act. See Alders v. AFA Corp., 353 F. Supp. 654, 656 (S.D.Fla.1973). C & D, however, makes three arguments for treating Butler's attempt to enforce its covenants as per se illegal.

First, C & D would have this court treat the application of restrictive covenants by Butler as a form of a group boycott and thereby per se illegal under Klor's Inc. v. Broadway-Hale Stores, Inc., 359 U.S. 207, 209, 79 S. Ct. 705, 707, 3 L. Ed. 2d 741 (1959). This is a creative but misleading characterization of Butler's action, and we remain thoroughly unconvinced. In a post-employment covenant the employee is required to boycott some person, activity, place, or combination of those factors. It is possible to characterize and treat any post-employment restrictive covenant applied to more than one employee as a group boycott, but such characterization supplies a label and little else. As the court said in Cesnik v. Chrysler Corp., 490 F. Supp. 859, 866 (M.D.Tenn.1980), "[t]he attachment of the group boycott label does not necessarily require as a consequence an application of the per se approach...." The per se doctrine has developed to deal with those economic practices that are clearly in violation of the letter and spirit of the Sherman Act and totally devoid of redeeming value and should not be extended to restraints of trade that are of ambiguous effect.

The labeling of a restraint as a group boycott does not eliminate the necessity of determining whether it is a "naked restraint of trade with no purpose except stifling competition," White Motor Co. v. United States, 372 U.S. 253, 263, 83 S. Ct. 696, 702, 9 L. Ed. 2d 738 (1963). In this case, as our analysis in section A above reveals, we are dealing with a restraint that has a legitimate and valid business purpose; therefore, appending the group boycott label does not make this restraint an appropriate candidate for per se treatment.

C & D's second argument for per se treatment is that "[g]iven the finding of the district court that the sole purpose of Butler's application of the covenant was to avoid or eliminate competition," Butler's action violated the standard set out in United States v. Realty Multi-List, Inc., 629 F.2d 1351, 1364, 1366-67 (5th Cir.1980), in which this court examined whether a real estate multiple listing service which limited membership in the service on the basis of a variety of criteria was a naked restraint of trade not ancillary to any legitimate business purpose.7 C & D's argument is a distortion of the district court's finding. The district court found that Butler threatened "enforcement of its covenant for the purpose of causing TVA to reconsider and award its 1980 contract to Butler."

While we are not completely satisfied with the district court's finding, it is by no means equivalent to the practice this court condemned in Realty Multi-List. In Realty Multi-List, we also stated that "when a practice tends to reduce competition of this type, but nevertheless operates to make the market more efficient--thereby aiding in the reduction of prices and the better allocation of resources, for example--then it may still be found, under the rule of reason, to further the [Sherman] Act's goals in aiding competition." 629 F.2d at 1364. We are persuaded that the alleged exclusionary practice and intent in this case is not so obviously egregious as to obviate the need for a rule of reason analysis under Realty Multi-List.

C & D's final argument is that Butler's application of the covenants was invalid per se under Addyston Pipe & Steel Co., since Butler had no legitimate business purpose in enforcing these covenants. We cannot concur: first because our analysis in section A above reveals a legitimate purpose, and second because whether or not Butler had a legitimate business purpose in enforcing these covenants is precisely the sort of issue which is suitable to rule of reason determination.

The purpose of the per se rule is to dispose of cases quickly without the more detailed and costly inquiry required under the rule of reason. It is to be applied, however, only when history and analysis have shown that in sufficiently similar circumstances the rule of reason unequivocally results in a finding of liability under the Sherman Act. The per se doctrine has developed to deal with those economic practices that are clearly in violation of the letter and spirit of the Sherman Act and totally devoid of redeeming value, and should not be extended to restraints of trade that are of ambiguous effect. Northern Pacific Railway Co. v. United States, 356 U.S. 1, 4-5, 78 S. Ct. 514, 518, 2 L. Ed. 2d 545 (1958). To use the per se rule as a means of avoiding rule of reason analysis when it is unclear what the result would be under the rule of reason would subvert the intention and purpose of the per se rule.

Under the rule of reason it is necessary to establish that the covenant had an adverse impact on competition. As this court said in H & B Equipment Co. v. International Harvester Co., 577 F.2d 239, 246 (5th Cir.1978):

The first step in establishing an unreasonable restraint of trade is to show anticompetitive effect, either in intrabrand or interbrand markets.... Absent anticompetitive effect, an unlawful intent will not establish a rule of reason violation, Chicago Board of Trade v. United States, 246 U.S. 231, 238, 38 S. Ct. 242 , 62 L. Ed. 683 (1918), nor will the use of unfair methods of competition.

Similarly, in Kestenbaum v. Falstaff Brewing Corp., 575 F.2d 564, 571 (5th Cir.1978), cert. denied, 440 U.S. 909, 99 S. Ct. 1218, 59 L. Ed. 2d 457 (1979), we held:

[T]he rule of reason standard hinges the ultimate legality of a restraint on whether the plaintiff has demonstrated an anticompetitive effect which is not offset by a need to achieve a procompetitive benefit or justification.

The questions of whether Butler's enforcement of its covenant has any anticompetitive effect and whether this effect is outweighed by any procompetitive benefits are not really separate from one another. Our analysis will consequently treat these issues as a single problem to be analyzed and resolved as a unit. An anticompetitive or procompetitive effect must be judged in relation to a market, for it is in that conceptual space that competition takes place. What is the relevant market of inquiry?

C & D is challenging the enforcement of the restrictive covenant contained in the employment contracts of Butler's job shoppers at TVA. Butler is enforcing a contractually reserved right by enforcing the covenant. It is perhaps too obvious to warrant mention that reserving to oneself the use of resources secured by contract must be harmful to competitors, whether or not it is harmful to competition. In the same way that a firm that excludes its competitors from using its factory and machinery is harming its competitors, so too Butler is harming its competitors at TVA by excluding them from access to its job shoppers. The relevant inquiry is not whether its present attempt to exclude adversely impacts competition but rather whether its acquisition of the power to exclude competitors had a sufficiently adverse impact on competition to constitute a violation of section 1 of the Sherman Act. It is the formation of these contractual rights which either does or does not have an adverse impact upon competition.

Butler's restrictive covenant has an impact on two sets of parties, the job shopper and the client firm. The record is clear that Butler serves a relatively small portion of the technical service market. In this market it must compete for job shoppers on the one hand, and client firms on the other. As an example we note that in competing for the 1980 TVA contract Butler had to compete against sixteen other firms. Thus, in competing for job shoppers the offer of a contract containing such a restrictive covenant seems, at first blush, to place Butler at a competitive disadvantage rather than an advantage vis-a-vis its competitors. Similarly when bidding for contracts with client firms the presence of such restrictive covenant in its contracts with job shoppers should, at first blush, place Butler at a competitive disadvantage.8 It is therefore difficult to see how these covenants impermissibly hurt either Butler's competitors or competition for the TSF contract with TVA. When Butler offered its employment contracts to its job shoppers they had as an alternative all the other means of obtaining employment in their chosen fields. Similarly, when Butler offered its services to TVA it too had as an alternative all the other means of obtaining workers for its nuclear projects. It would be ludicrous to suggest that at the time these contracts were formed Butler had any sort of monopoly power either over TVA or the individuals who later became Butler's job shoppers at TVA. Therefore to the extent that the job shoppers and TVA disabled themselves by signing contracts with Butler that reserved to Butler for ninety days employment rights over Butler's job shoppers at TVA, we must presume that it was in exchange for valuable consideration. The proof of the procompetitive benefits of the restrictive covenant is in the pudding. Having to compete against vigorous competition both for job shoppers and client firms, Butler succeeding in winning TVA contracts and in supplying a workforce. Thus, though Butler offered both the workers and the client firm facially less attractive contracts, it must have thus been enabled to provide its services at lower cost, because both sets of parties accepted the contract.

It confuses and obfuscates the issue to ask whether Butler had an unfair competitive advantage at the stage of bidding for the 1980 contract. Nonetheless, assuming, arguendo, that the covenant is enforceable, Butler had no competitive advantage of the type disfavored by the Sherman Act at the stage of bidding for the 1980 contract. The covenant gives Butler a contractual right to prevent, for ninety days, the employment of its job shoppers at TVA. There is presumably some market price at which Butler will relinquish that right. Once it is recognized that this right can be sold, its value will be factored into every bid, including Butler's, for the TVA contract, and all parties will be competing on an equal footing.9

A competitive advantage can only exist for a party if the cost to that party is less than the cost to other parties. The existence, or lack, of a property right in the continued employment of its 140 job shoppers at TVA cannot confer such a cost saving on Butler or anyone else. This property right has a market value. That market value is a cost to whomever wins the contract. It is obviously a cost to any party other than Butler, since it must pay Butler the value of a waiver of this property right. It is less obviously, but just as certainly, a cost to Butler. If Butler wins the contract, then it loses the power to sell its waiver of its rights under the covenant to someone else.

The only way in which Butler could have had an unfair "advantage" in the bidding would be if the restrictive covenant was enforceable but not transferable. In that event Butler would have had an incentive to bid lower than it otherwise would in order to protect its fragile property interest in the job shoppers it had placed at TVA.10 Even if this were the case, it would hardly constitute an unfair advantage in the sense proscribed by the antitrust laws. The only unfairness would be to Butler, since it would be forced to lower its price below the level it would ordinarily charge. Therefore, we conclude that the restrictive covenant in Butler's contracts with its job shoppers survives a rule of reason analysis, as it does not have a sufficiently adverse impact on competition.

The district court's denial of Butler's motion for summary judgment against C & D's section 2 Sherman Act claim was certified for our consideration on appeal pursuant to 28 U.S.C. Sec. 1292(b) (1976). An appeal was taken and the matter is now before us.

Section 1292(b) requires that the district court recite that the order it is issuing "involves a controlling question of law as to which there is substantial ground for difference of opinion." This circuit does not require that the district court, in certifying an issue for appeal, append a statement explaining specifically the reasons for certification. See Rothenberg v. Security Management Co., 617 F.2d 1149, 1150 (5th Cir.), rehearing denied, 623 F.2d 711, cert. denied, 449 U.S. 954, 101 S. Ct. 359, 66 L. Ed. 2d 218 (1980). "However, when the case is of such a nature that the reasons for the ... certification are unclear, it may be necessary for adequate appellate review that the district court's reasons be stated." Id.

The court below has provided us with some instruction as to the rationale of its certification. The district court noted that the issues raised by the motion for summary judgment were deeply entwined with the issues already addressed in this case. Further, the court stated that it was bound by Poller v. Columbia Broadcasting System, Inc., 368 U.S. 464, 467, 82 S. Ct. 486, 488, 7 L. Ed. 2d 458 (1962). Poller involved the reversal of a grant of summary judgment in a Sherman Act case on the grounds that "[s]ummary judgment should be entered only when the pleadings, depositions, affidavits, and admissions filed in the case 'show that ... there is no genuine issue of material fact and that the moving party is entitled to a judgment as a matter of law.' Rule 56(c) Fed.Rules Civ.Proc." id.; and there were genuine material issues still in dispute in that case.

The problem in this case is that the district court has not told us specifically why it denied summary judgment. The court did not articulate a legal ruling which might serve as the focus of this appeal. It is well nigh impossible, therefore, with this open-ended record, for us to divine what was in the court's mind.

We reluctantly conclude that permission to appeal on this issue was improvidently granted. The order granting the right to appeal is therefore vacated.

In conclusion: we affirm the district court on the section 1 Sherman Act claim; reverse on the enforceability of Butler's restrictive post-employment covenant; vacate the grant of a right to appeal the denial of summary judgment on the section 2 Sherman Act claim; and remand for further proceedings consistent with this opinion.

AFFIRMED in part; REVERSED in part; VACATED in part; and REMANDED.

I concur in the result.