Belleville Shoe Manufacturing Company redeemed 550 shares of its stock from Marian Short in 1977, paying $273 per share. Twelve years later Short filed this suit under Sec. 10(b) of the Securities Exchange Act of 1934, 15 U.S.C. Sec. 78j(b), and the SEC's Rule 10b-5, 17 C.F.R. Sec. 240.10b-5, contending that Belleville (a closely-held, family corporation) and its controlling shareholders had defrauded her about the value of the shares. They painted a gloomy picture of the firm's prospects in 1977; it has prospered since. Short depicts her brother Homer Weidmann, an officer of Belleville, as her principal business adviser and contends that she delayed filing suit because Weidmann repeatedly assured her that she had received a fair price in 1977.

The district court dismissed the case under Fed.R.Civ.P. 12(b)(6), concluding that Short could not escape the statute of limitations. Short relies on state and federal tolling rules. Defendants respond that we should not inquire into tolling because Sec. 13 of the '33 Act, 15 U.S.C. Sec. 77m, which they ask us to apply, sets three years as an unyielding limit. Defendants ask, in other words, that we cease looking to state law as a source of periods of limitations in securities cases, on the authority of Agency Holding Corp. v. Malley-Duff & Associates, Inc., 483 U.S. 143, 107 S.Ct. 2759, 97 L.Ed.2d 121 (1987), which applied a federal statute of limitations (from the antitrust laws) to cases under the Racketeer Influenced and Corrupt Organizations Act (RICO), 18 U.S.C. Secs. 1961-68.

* A

For many years we have applied to cases under Rule 10b-5 statutes of limitations borrowed from state blue sky statutes, adding an overlay of tolling principles from state and federal law. Davenport v. A.C. Davenport & Son Co., 903 F.2d 1139 (7th Cir.1990), is the most recent in a line going back at least to Parrent v. Midwest Rug Mills, Inc., 455 F.2d 123, 125-27 (7th Cir.1972). See also, e.g., Norris v. Wirtz, 818 F.2d 1329 (7th Cir.1987); Teamsters Local 282 Pension Trust Fund v. Angelos, 815 F.2d 452 (7th Cir.1987); Suslick v. Rothschild Securities Corp., 741 F.2d 1000 (7th Cir.1984); Goldstandt v. Bear, Stearns & Co., 522 F.2d 1265 (7th Cir.1975); Tomera v. Galt, 511 F.2d 504 (7th Cir.1975). Like the other courts of appeals, we looked to state law because Congress has not enacted a statute of limitations for Rule 10b-5, and could hardly have been expected to--for the right of action under Rule 10b-5 was created by the courts rather than Congress. Section 10(b) itself grants rulemaking authority to the SEC; because it does not create a right of action, it is not accompanied by a statute of limitations.

Because Congress was silent, we applied the principle, derived from the Rules of Decision Act, 28 U.S.C. Sec. 1652, that when federal law is deficient, state law fills the gap. One of the venerable applications is to obtain periods of limitations from state law. E.g., Auto Workers v. Hoosier Cardinal Corp., 383 U.S. 696, 704, 86 S.Ct. 1107, 1112, 16 L.Ed.2d 192 (1966) (collecting cases). This is so predictable that Congress may anticipate it, so that when it does not enact a federal statute of limitations it means to leave in place the background rule that state law applies. See Reed v. United Transportation Union, 488 U.S. 319, 109 S.Ct. 621, 625, 102 L.Ed.2d 665 (1989); Agency Holding, 483 U.S. at 147, 107 S.Ct. at 2762; DelCostello v. Teamsters, 462 U.S. 151, 158, 103 S.Ct. 2281, 2287, 76 L.Ed.2d 476 (1983). Federal courts are so accustomed to turning to state periods of limitations that we (and our colleagues in other circuits) did this on auto-pilot, without discussing whether something differentiated securities laws from other statutes. See also Ernst & Ernst v. Hochfelder, 425 U.S. 185, 210 n. 29, 96 S.Ct. 1375, 1389 n. 29, 47 L.Ed.2d 668 (1976); Herman & MacLean v. Huddleston, 459 U.S. 375, 384 n. 18, 103 S.Ct. 683, 688 n. 18, 74 L.Ed.2d 548 (1983), both recognizing the prevailing practice.

Yet there are differences. One is that Congress could hardly have anticipated that courts would turn to state law for a period of limitations, when Congress did not create the right of action in the first place. We are not dealing here with a problem of deciding whether Congress meant to depart from a norm; we have a problem of the courts' creation. See United Parcel Service, Inc. v. Mitchell, 451 U.S. 56, 68 n. 4, 101 S.Ct. 1559, 1567, 67 L.Ed.2d 732 (1981) (Stewart, J., concurring); DelCostello, 462 U.S. at 158-59 n. 12, 103 S.Ct. at 2287 n. 12. Federal courts have an obligation to create stable periods of limitations for their handiwork. Smith v. Chicago, 769 F.2d 408 (7th Cir.1985).

A second difference is that Congress has not been silent. The securities acts are jammed with statutes of limitations. For every statutory right of action there is a corresponding statute of limitations. The Securities Act of 1933 gave the purchaser two years from the date of discovering the fraud, but in no event more than ten years from the date of sale. 48 Stat. 84 (1933). When Congress enacted the '34 Act, it amended this statute (now Sec. 13 of the '33 Act) and added four more, one for each of the new express rights of action. See Secs. 9(e), 16(b), 18(c), and 29(b), 15 U.S.C. Secs. 78i(e), 78p(b), 78r(c), and 78cc(b). In 1988 it added another for inside trading cases, 102 Stat. 4680-81, to be codified as Sec. 20A, 15 U.S.C. Sec. 78t-1. There can be no presumption that Congress meant courts to look to state law, when every time Congress has authorized a federal remedy it has also created a federal period of limitations.

Third, turning to state law for periods of limitations creates special problems under the securities acts because the acts do not apply in the first place unless the transactions occurred in interstate commerce. At least two state statutes therefore could be applied to any case. Courts must use conflict-of-laws principles to pare the number down to one. Congress enacted a national rule against fraud because it believed that national law ought to govern multi-state transactions. Returning to state law to fetch a period of limitations is inconsistent in spirit with this decision.

All of this would be by the by if the Rules of Decision Act requires federal courts to use state law. But two cases decided after the courts of appeals began to apply state law under Rule 10b-5 hold that it does not. DelCostello concluded that a six-month period drawn from the National Labor Relations Act governs "hybrid" suits alleging an employer's breach of contract coupled with a union's breach of the duty of fair representation. Agency Holding decided to apply the four-year period of limitations in the antitrust laws to suits under RICO. These cases call for fresh examination of the question whether to turn to state law in securities cases. The third circuit conducted such an examination and concluded that Sec. 13 of the '33 Act is the appropriate statute. In re Data Access Systems Securities Litigation, 843 F.2d 1537 (3d Cir.1988) (in banc). The Securities and Exchange Commission likewise believes that federal law supplies the period of limitations, although it thinks that the new Sec. 20A rather than Sec. 13 is the closer match. Brief for the United States as Amicus Curiae in Lebman v. Aktiebolaget Electrolux, No. 88-1144 (October Term, 1988), cert. denied, --- U.S. ----, 109 S.Ct. 3214, 106 L.Ed.2d 564 (1989). The time is ripe for reconsideration in this circuit too.

"[W]hen a rule from elsewhere in federal law clearly provides a closer analogy than available state statutes and when the federal policies at stake and the practicalities of litigation make that rule a significantly more appropriate vehicle for interstitial lawmaking", DelCostello, 462 U.S. at 171-72, 103 S.Ct. at 2294, quoted in Agency Holding, 483 U.S. at 148, 107 S.Ct. at 2763, it is appropriate to select a federal statute of limitations. Securities law satisfies both branches. The statutes of limitations in the securities acts are addressed to the conduct at stake in this case--securities fraud. Section 13, in particular, covers the express right of action in Sec. 12, which concerns incorrect statements and material omissions in the course of selling securities. Congress has spoken directly to the appropriate statute of limitations for litigation of the sort Short presses. When the statute contains not only a statute of limitations for every express right of action, but also an express right of action quite similar to the implied right of action, then the federal statute "clearly provides a closer analogy than available state statutes".

Difficulties in selecting an appropriate state statute also influence the "practicalities of litigation". Even when turning to state law, courts should endeavor to find a single period of limitations "that can be applied with ease and predictability in all 50 States." Owens v. Okure, 488 U.S. 235, 109 S.Ct. 573, 578, 102 L.Ed.2d 594 (1989). That task has proved daunting in securities cases. Different courts have looked to different bodies of state law for inspiration. Some borrow from the statute generally applicable to fraud; some borrow from state blue sky laws; some courts use a little of each, depending on the particulars of the claim under Rule 10b-5. The ABA's Committee on Federal Regulation of Securities, Report of the Task Force on Statute of Limitations for Implied Actions, 41 Bus.Law. 645, 659-66 (1986), collects the decisions from around the circuits. Some states have more than one fraud or blue sky statute, complicating matters. States also revamp their laws more frequently than Congress amends the Securities Exchange Act. Illinois, for example, amended the statute of limitations in its blue sky act in 1985, and one of the questions with which the district court wrestled is the extent to which the alterations apply to a claim that accrued in 1977. (The district judge, disagreeing with four other district judges in the circuit, concluded that Short had at most two years after January 1, 1986, the effective date of the amendment. See also Davenport, 903 F.2d at 1141 n. 5; Norris, 818 F.2d at 1333 n.*.) Translating state periods of limitations to federal court also has led to difficult questions concerning whose tolling and estoppel doctrines apply: the state's, the federal court's, or both? Davenport, Norris, and many other cases grapple at length with the interaction of state and federal tolling rules.

Although this circuit regularly uses state blue sky laws, other circuits have been equally committed to different courses. From the perspective of practitioners litigating cases originating in many states (especially class actions), the situation is a nightmare. Lawyers and courts alike devote untold hours to identifying proper state analogies and applying multiple (conflicting or cumulative) tolling doctrines. "This uncertainty and lack of uniformity promote forum shopping by plaintiffs and result in wholly unjustified disparities in the rights of different parties litigating identical claims in different states. Neither plaintiffs nor defendants can determine their rights with any certainty. Vast amounts of judicial time and attorneys' fees are wasted." ABA Committee, 41 Bus.Law. at 647. Loud calls for reform issue from scholars and the bar. With a unanimity unmatched in any other corner of securities law, everyone wants a simpler way--and to everyone that means a uniform federal statute of limitations. In addition to the ABA Committee's plaint, see, e.g., Thomas Lee Hazen, 2 The Law of Securities Regulation Sec. 13.8 (2d ed. 1990); Bateman & Keith, Statutes of Limitations Applicable to Private Actions Under SEC Rule 10b-5: Complexity in Need of Reform, 39 Mo.L.Rev. 165, 183 (1974); Ruder & Cross, Limitations on Civil Liability Under Rule 10b-5, 1972 Duke L.J. 1125, 1142; Schulman, Statutes of Limitation in 10b-5 Actions: Complication Added to Confusion, 13 Wayne L.Rev. 637 (1967); Note, Statutes of Limitation for Rule 10b-5, 39 Wash. & Lee L.Rev. 1021 (1982). As Professor Loss puts it, Fundamentals of Securities Regulation 1168-69 (1983): "This reference to state law makes for a great amount of utterly wasteful litigation.... Would it not be eminently more consistent with the overall statutory scheme to look to what Congress itself did when it was thinking specifically of private actions in securities cases than to a grab-bag of more or less analogous state statutes?"

Norris examined this subject and concluded that the statute of limitations in Sec. 13 fits actions under Rule 10b-5 far more closely than the collection of statutes absorbed from state law. 818 F.2d at 1332-33. We did not jettison our existing approach, however, deeming it "too late for an inferior federal court to turn back the clock", 818 F.2d at 1333. When we decided Norris, the courts of appeals were unanimous in thinking that state law supplies the statutes of limitations under Rule 10b-5. A month later the Supreme Court decided Agency Holding. Six months after that the third circuit broke ranks in Data Access, relying on Agency Holding and the analysis in Norris. Now that there is an established conflict among the circuits and the SEC has endorsed the use of federal law, we need not be so meek. Although stare decisis has its claims, Agency Holding took a bold step that invites the inferior federal courts to reconsider their handling of adjacent fields. We hold that federal and not state law supplies the statute of limitations in suits under Sec. 10(b) and Rule 10b-5.*

We leave for the future all questions concerning retroactive application of this decision. Retroactivity has not been briefed, and the retroactive application of decisions affecting periods of limitations is a question of some subtlety depending on the nature of reliance interests. See Chevron Oil Co. v. Huson, 404 U.S. 97, 92 S.Ct. 349, 30 L.Ed.2d 296 (1971); St. Francis College v. Al-Khazraji, 481 U.S. 604, 107 S.Ct. 2022, 95 L.Ed.2d 582 (1987); Goodman v. Lukens Steel Corp., 482 U.S. 656, 662-64, 107 S.Ct. 2617, 2621-22, 96 L.Ed.2d 572 (1987); Hill v. Equitable Trust Co., 851 F.2d 691 (3d Cir.1988) (applying Data Access retroactively); cf. American Trucking Ass'ns, Inc. v. Smith, --- U.S. ----, 110 S.Ct. 2323, 2343-45 (Scalia, J., concurring), 2347-52 (Stevens, J., dissenting), 110 L.Ed.2d 148 (1990). It is enough for the moment to observe that new rules of law apply in the cases that establish them and that Short cannot have relied on Illinois law, because she claims to have been unaware of the basis for litigation until a short time before filing suit. Effects of reliance on our former holdings may be determined when the plaintiff demonstrates reliance.

Federal securities laws contain two candidate statutes of limitations. The first is Sec. 13 of the Securities Act of 1933, as amended by the Securities Exchange Act of 1934, which provides:

No action shall be maintained to enforce any liability under section 11 or section 12(2) unless brought within one year after the discovery of the untrue statement or the omission, or after such discovery should have been made by the exercise of reasonable diligence, or, if the action is to enforce a liability created under section 12(1), unless brought within one year after the violation upon which it is based. In no event shall any such action be brought to enforce a liability created under section 11 or section 12(1) more than three years after the security was bona fide offered to the public, or under section 12(2) more than three years after the sale.

Section 11 deals with errors and omissions in registration statements, including the portion of the registration statement that circulates as the prospectus. Section 12(1) deals with sales "in violation of section 5"--that is, unregistered sales of securities required to be registered. Section 12(2) addresses all other forms of materially incorrect or misleading selling literature and oral communications in the sale of a security. Data Access selected Sec. 13 as the appropriate federal law because misstatements and omissions that violate Sec. 12(2) are the closest match to the subjects of Rule 10b-5. (Short could not have sued under Sec. 12 itself; she complains of fraud by the buyer, while Sec. 12(2) is addressed to fraud by the seller. Still, Sec. 12(2) applies to the kind of conduct involved here.)

The second candidate is Sec. 20A(b)(4) of the Securities Exchange Act of 1934, added by the Insider Trading and Securities Fraud Enforcement Act of 1988. This provides:

No action may be brought under this section more than 5 years after the date of the last transaction that is the subject of the violation.

Section 20A(a) defines "the subject" as violating the '34 Act or its rules "by purchasing or selling a security while in possession of material, nonpublic information". The Solicitor General, representing the SEC in Lebman, asked the Supreme Court to use this statute for all actions under Rule 10b-5, on the rationale that inside trading violates Rule 10b-5, so that employing Sec. 20A(b)(4) just extends it from some to all Rule 10b-5 actions.

Choosing between these statutes is difficult. Section 13 is attractive because it covers Sec. 12(2), which describes the offense to which Rule 10b-5 is applied: fraudulent misstatements or omissions in connection with the purchase or sale of securities. Rule 10b-5 came into being when Sec. 13 was the only game in town. Any court choosing a federal statute of limitations between 1946 (the date of the first decision recognizing a private right of action under Rule 10b-5) and 1988 necessarily would have selected Sec. 13. It is also the statute Data Access picked. To pick Sec. 20A would be to create a fresh conflict among the circuits. Because a principal reason for choosing a federal law is to replace confusion with certainty, we are reluctant to depart from the statute that will be employed in the third circuit.

On behalf of Sec. 20A one may say that five years represents Congress' latest thoughts and that because trading on material inside information violates Rule 10b-5, the only way to apply a single statute of limitations to all actions under that Rule is to use Sec. 20A(b)(4). Clever lawyers could describe even a garden variety fraud as "purchasing or selling a security while in possession of material, nonpublic information"--that is, the truth--blurring the line between the offenses and making the choice of a period of limitations difficult. Moreover, Sec. 20A offers plaintiffs more time than does Sec. 13. Section 10(b) and Rule 10b-5 require the plaintiff to prove scienter, which a plaintiff under Sec. 12(2) need not; perhaps the additional years are appropriate to the extra requirements of litigation under Rule 10b-5.

Interests in uniformity tug both ways: if we choose Sec. 20A we shall have one statute of limitations for all Rule 10b-5 cases in the seventh circuit, but different rules in the seventh and third circuits. If we choose Sec. 13 we shall be in line with the third circuit, but different kinds of claims under Rule 10b-5 will be subject to different periods of limitations. The latter diversity is likely to be tolerable. To obtain the five-year period of Sec. 20A, the plaintiff would have to sue under Sec. 20A, establish trading by an insider (Belleville Shoe's insiders did not trade with Short), and accept the limitations of that statute, including the restriction of damages to the defendants' profit (as opposed to the plaintiff's loss). In the great bulk of cases, there will be no ambiguity.

One interest cuts strongly in favor of Sec. 13: it is a statute of repose, while Sec. 20A appears to be a statute of limitations. Section 13 gives one year from discovery, but "[i]n no event shall any such action be brought to enforce a liability created ... under section 12(2) more than three years after the sale." (Emphasis added.) Conduct by the defendant that postpones the date of discovery therefore does not extend the period of suit by more than two years. Data Access concluded that equitable tolling does not postpone the accrual of the claim, or equitable estoppel does not bar the invocation of Sec. 13. See also Admiralty Fund v. Hugh Johnson & Co., 677 F.2d 1301, 1308 (9th Cir.1982); Summer v. Land & Leisure, Inc., 664 F.2d 965 (5th Cir.1981); Bloomenthal, Statutes of Limitations and the Securities Acts--Part I, 7 Sec. & Fed.Corp.L.Rep. 18, 21 (1985) (collecting legislative history); ABA Committee, 41 Bus.Law. at 655. Cf. Darms v. McCulloch Oil Corp., 720 F.2d 490 (8th Cir.1983); Aldrich v. McCulloch Properties, Inc., 627 F.2d 1036 (10th Cir.1980) (both interpreting language in other statutes identical to Sec. 13 as precluding any extension beyond three years).

Courts say that equitable tolling does not apply under Sec. 13, see Hazen, 1 Law of Securities Regulation Sec. 7.5.4 at 340-42 (collecting cases), but this is not strictly accurate. It is better to say that equitable tolling and related doctrines do not extend the period of limitations by more than the two-year grace period Sec. 13 allows. Norris, 818 F.2d at 1332. Congress did not obliterate these valuable doctrines so much as it set bounds on the length of delay. It may be that equitable estoppel could preclude resort to the one-year period after discovery, even though no tolling of the three-year outer limit is possible; we need not decide. Unless the "in no event more than three" language cuts off claims of tolling and estoppel at three years, however, it serves no purpose at all--what other function could be served by such language in a statute that starts the time on discovery?

Section 20A, by contrast, appears to be an ordinary statute of limitations. The SEC believes that it is a statute of repose as well as a statute of limitations, see Lebman amicus brief at 14 n. 17, but it is not so clear to us. The language is in the traditional form of a statute of limitations, and the Solicitor General did not cite any legislative history indicating that it is more potent than the usual variety. (No case has been decided under the period of limitations established in Sec. 20A, and none is likely to be until after 1993.)

The difference between statutes of limitations and statutes of repose is substantial in securities litigation. Indeed, the SEC's submission that Sec. 20A should be used across the board appeared to depend on its conclusion that Sec. 20A, like Sec. 13, is a statute of repose. All of the express provisions in the '33 and '34 Acts (other than Sec. 20A), are drawn as statutes of repose, and deliberately so. Prices of securities are volatile. If suit may be postponed indefinitely on equitable grounds, then investors may gamble with other people's money. An investor in Short's position may sell her shares for a price certain. If the firm does poorly, she keeps the money; if it does well, she sues and asks for the increase in value. Congress chose one year after discovery, and a cap of two additional years on tolling principles, in order to curtail the extent to which the securities laws permit recoveries based on the wisdom given by hindsight. This was why in 1934 Congress repealed the period in the original version of the '33 Act (two years from discovery, and in no event more than ten) and replaced it with the current Sec. 13 (one year from discovery, and in no event more than three). See H.R.Conf.Rep.No. 1838, 73d Cong., 2d Sess. 32, 36, 42 (1934); 78 Cong.Rec. 8198-8203 (May 7, 1934) (Senate amendments carried to the conference); J.S. Ellenberger & Ellen P. Mahar, 6 Legislative History of the Securities Act of 1933 and Securities Exchange Act of 1934 6565-66, 6718, 6993 (1973), and 7 id. at 7743-44. Investors then have a more powerful incentive to investigate rather than accept another person's word without question, as Short says she did. Prudent investors almost always can sniff out fraud (or enough smoke to justify litigation) within three years. Section 13 cuts off only the claims of the most trusting or somnolent--or the most wily, those who wanted to wait as long as possible.

Setting an outer limit for claiming fraud or material omissions in the sale of securities is an important aspect of Sec. 13, one of the principal reasons Congress designed the statute that way. Section 20A, by contrast, uses for inside trading cases a measure of damages that the passage of time does not affect. Section 20A(b)(1) provides that the "total amount of damages ... shall not exceed the [insider's] profit gained or loss avoided in the transaction or transactions that are the subject of the violation." Profit gained or loss avoided is a contemporaneous measure: the difference between the price the insider realizes and the market price of the securities after the news is released. Wisconsin Real Estate Investment Trust v. Weinstein, 781 F.2d 589, 596-97 (7th Cir.1986); SEC v. MacDonald, 699 F.2d 47, 52-55 (1st Cir.1983) (in banc); Gerstle v. Gamble-Skogmo, Inc., 478 F.2d 1281, 1306 n. 27 (2d Cir.1973) (Friendly, J.). This sum is fixed within days after the trading, sometimes within hours. No one has a free put or call of securities; no one can use delay in bringing suit to speculate on the firm's future prosperity. Section 20A may use a five-year period, and be drafted as an ordinary statute of limitations, precisely because the reasons for setting the shorter one-and-three period for fraud are not present. To select Sec. 20A(b)(4) as the rule for all Rule 10b-5 actions, just because a five-year period has been deemed appropriate for one atypical kind of violation, would let the tail wag the dog. Congress did not say that Sec. 20A applies to all cases under Rule 10b-5; it did not amend Sec. 13; it limited the scope of Sec. 20A's statute of limitations to a particular and atypical kind of violation. Section 20A(b)(4) has been prescribed for inside trading cases; run-of-the-mine frauds under Rule 10b-5 should be governed by the statute in force when Rule 10b-5 came into being, the statute designed for the kind of conduct about which Short and most other 10b-5 plaintiffs complain: Sec. 13 of the Securities Act of 1933.

Application of Sec. 13 to this case is straightforward. Short sold her shares to Belleville Shoe on May 5, 1977. Although she contends that she did not discover the fraud until late 1988, the second sentence of Sec. 13 sets three years after the sale as the outer limit for litigation. That time expired in 1980. Suit was not filed until 1989. It is untimely.

Short's second federal claim arises under RICO. She contends that Belleville and the other defendants carried out a pattern of racketeering acts (that is, of securities fraud) by acquiring the stock of four minority investors in 1977. Short, Louise Lampert Larsen, and Jean Lampert Sundheim sold their stock in spring 1977, and Rosalind Wagner that fall. These sales, according to Short, establish a "pattern" of fraud characterized by the necessary "continuity plus relationship", H.J. Inc. v. Northwestern Bell Telephone Co., --- U.S. ----, 109 S.Ct. 2893, 2900, 106 L.Ed.2d 195 (1989). The district court disagreed, holding that four transactions in close proximity 12 years ago do not establish the "continuity" element in the absence of a threat of enduring fraud.

We approach the question with diffidence, because "continuity plus relationship" is such an ambulatory phrase. Still, it is impossible to stretch the events of 1977 into a "pattern" of fraud. Belleville Shoe's board of directors authorized the purchases at a price of $273 per share in a single resolution, dated May 1, 1977. The four investors accepted Belleville Shoe's offer at different times--Wagner accepting in the fall because she was out of the country when the other three said yes. One offer followed by four acceptances hardly threatens "continuity" of any kind. The Court said in H.J. that "[p]redicate acts extending over a few weeks or months and threatening no future criminal conduct do not satisfy this [continuity] requirement." 109 S.Ct. at 2902. This language supplied the basis of the holding in Sutherland v. O'Malley, 882 F.2d 1196 (7th Cir.1989), that a course of conduct stretching over five months did not amount to "continuity". See also Management Computer Services, Inc. v. Hawkins, Ash, Baptie & Co., 883 F.2d 48 (7th Cir.1989).

Short tries to rescue her case by characterizing the threat of further misconduct against minority shareholders as the source of continuity. She does not identify any such investors, however, or explain how they were threatened. The complaint does not mention any further purchases of stock under circumstances suggesting fraud. RICO is not designed to transfer to federal court all claims of oppression of minority investors in violation of fiduciary duties established by corporate law. Short may pursue such contentions in state court; all of her claims under state law were properly dismissed, for lack of an independent jurisdictional basis, following the failure of her two federal claims. Short insists that Illinois' tolling rules would allow her to recover from Belleville Shoe and her relatives 13 years after the events. She may test that belief in the appropriate forum.

AFFIRMED.

POSNER, Circuit Judge, concurring.

I have no disagreement with the court's opinion or judgment, and join both, writing separately only to express reservations about the principle of "borrowing" a period of limitations from one statute for use with another (here, a rule of the SEC, but the principle is the same) that has none. The Supreme Court's unwavering attachment to the principle leaves us no choice but to apply it, but that attachment is beginning to be questioned, Agency Holding Corp. v. Malley-Duff & Associates, Inc., 483 U.S. 143, 157-70, 107 S.Ct. 2759, 2767-74, 97 L.Ed.2d 121 (1987) (concurring opinion), and I would like to add my small voice to the chorus.

Borrowing a period of limitations from one statute to use with another that doesn't have its own limitations provision is a matter of which round peg to stuff in a square hole. It artificially truncates the court's choice of an appropriate period, one well suited to the particular statute under consideration. It also runs the risk of applying one unprincipled legislative deal to a problem entirely outside the scope of the deal. Suppose Statute A specifies no period of limitations. Statute B regulates analogous conduct, and has a six-month period. But maybe B has such a short deadline for suit only because the interest group that opposed the enactment of B had enough muscle to block a longer deadline that would have made more sense from a neutral standpoint. In that event, to borrow B's limitations period for use with A will project the interest-group pressures that deformed B into a completely new area of conduct.

Whether or not courts are aware of this danger, they do attempt to correct for it by considering, as part of the borrowing procedure, the suitability to the substantive rule under consideration of the limitations periods in the various candidate statutes of limitations. But this places a lot of balls in the air. The considerations bearing on the suitability of one limitations period compared to another include the difficulty of investigating potential violations, the possibility that the consequences of wrongdoing will be delayed, the opportunities for wrongdoers to conceal the wrong, the rate at which evidence of wrongdoing and also evidence pertinent to the alleged wrongdoer's defenses is likely to decay, the sophistication of the relevant tribunals in handling stale evidence, the desirability of freeing court time for fresh claims, the interest of potential defendants in repose--that is, in knowing after a definite period has passed that they no longer have to worry about being sued--and the effect on the deterrence of statutory violators of reducing the time for bringing suit. When these considerations are mingled with the underlying question of the similarity between the conduct in the case at issue and the conduct to which the various candidate statutes of limitations apply, we have the essential condition for standardless, discretionary judgment: a multifactor test with no weights on the factors. Since courts cannot be expected to converge on a uniform outcome when they are operating under such a standard, predicting what statute of limitations will be borrowed is impossible and as a result extensive litigation often is necessary before a definitive conclusion on the limitations period emerges. It may not come until a Supreme Court decision is rendered resolving an intercircuit conflict that was years in the brewing.

As we noted in Hemmings v. Barian, 822 F.2d 688, 689 (7th Cir.1987), judges feel that they have to borrow an existing statute of limitations rather than lay down a period of limitations as a matter of federal common law because it would be arbitrary to pick a term of years. They feel in other words that enactment of a limitations period, because of its inescapable arbitrariness, is a legislative rather than a judicial task. Legislators can be as arbitrary as they please within the broad limits set by the Constitution but judges are supposed to reason to their conclusions, and how can you reason to 3 years over 2, 300 days over 240, 20 years over 15? You cannot; but neither can you reason to the right statute of limitations to borrow. The imponderables are so numerous that in the end the borrowing judgment, too, is inescapably arbitrary, as the present case illustrates.

What is to be done? There are several possibilities. One would be for courts to drop the mask and create statutes of limitations for claims that lack them. There is precedent for this in the judicial doctrine of laches, which on at least one occasion was firmed up into a period of definite length. Smith v. City of Chicago, 769 F.2d 408 (7th Cir.1985). Yet even then we borrowed the limitations period in an existing statute. The difficulty with creating a statute of limitations ex nihilo lies in the fact that judge-made rules come in the first instance from the bottom of the judicial hierarchy, rather than being imposed from on high. In the early days of a new substantive statute that specified no limitations period, each district judge would be selecting a period of limitations for the individual case, each court of appeals would be selecting the period for the cases in the circuit, and only when the Supreme Court got a case would the definitive period finally be declared. The improvements over the borrowing procedure would be modest; they would be chiefly in the area of judicial candor.

An institutional solution is necessary, and two possibilities come to mind. One, which wouldn't work in this case because the cause of action postdates the enactment of the statute under which it arises, would be for Congress to adopt a rule that every statute shall contain a statute of limitations. This course would be superior to Congress's adopting (as sometimes proposed) a general catch-all statute of limitations, because there is no single period of limitations that would be suitable for the entire range of causes of action in statutes that do not specify a period. Such a rule might, it is true, be ineffectual without a nagging agency within Congress to enforce it, some counterpart to the Congressional Budget Office, as proposed in the Report of the Federal Courts Study Committee 21-22, 89-93 (April 2, 1990)--unless the Supreme Court adopted Justice Scalia's view that if a statute contains no period of limitations, there just is no deadline on suing. Agency Holding Corp. v. Malley-Duff & Associates, Inc., supra, 483 U.S. at 164, 107 S.Ct. at 2771. This would cause considerable havoc but could be mitigated by a rule (well within the power of judges to create, I believe) that, in the absence of a statutory limitations period, the courts will apply the equitable doctrine of laches even if the cause of action is legal rather than equitable. Then the defendant could defend by showing that the plaintiff had unreasonably delayed in bringing the suit and that the defendant had been hurt by the delay. There is precedent for applying laches in cases at law. Piper Aircraft Corp. v. Wag-Aero, Inc., 741 F.2d 925, 939 (7th Cir.1984) (concurring opinion).

The second institutional possibility would be for Congress to delegate to the Judicial Conference of the United States, or to some new agency modeled on the Sentencing Commission--perhaps my hypothetical "nagging agency"--the power to adopt by regulation a period of limitations for any statute that does not have one. This would lift the burden from Congress of having to specify a limitations period in every enactment and would shift it to an expert body that could avoid the delay of litigation. It would be a great improvement over borrowing.