COFFIN, Circuit Judge.
This case concerns the constitutionality of section 3 of the Massachusetts TakeOver Bid Regulation Act, Mass.Gen.L. ch. HOC (1986), and the efforts of plaintiffs-ap-pellees Hyde Park Partners, L.P. and Hyde Park Holdings, Inc. (collectively “Hyde Park”) to gain control of defendant-appellant High Voltage Engineering Corporation (“High Voltage”).
Section 3 of the statute provides that prospective takeover bid “offerors” who fail to disclose their intent to gain control of the target company before acquiring five percent of its stock may not make a takeover bid for that target until one year after the failure to disclose.1 High Park failed to comply with the disclosure provision, and sought an injunction in federal district court enjoining the Secretary of the Commonwealth and High Voltage from enforcing section 32 on the grounds that it violates the Supremacy and Commerce clauses of the United States Constitution.3 Relying solely on the Commerce Clause, the district court, 676 F.Supp. 391, granted a preliminary injunction, prompting High Voltage and the state to bring this appeal.
Hyde Park’s acquisition of High Voltage’s outstanding stock surpassed 5% on December 9, 1987. It is not disputed that Hyde Park did intend to gain control of High Voltage at that time.4 Hyde Park did not disclose its control intent until December 21, 1987, by which time it purportedly owned 6.46% of the stock. Enforcement of section 3 would therefore impose a moratorium on Hyde Park’s takeover bid until December 9, 1988. The preliminary relief granted by the district court enjoins enforcement of that section so that Hyde Park may promptly commence purchases of stock pursuant to its tender offer. After this appeal was taken, Hyde Park commenced its tender offer for any and all outstanding High Voltage stock on January 6, 1988. The deadline for expiration of the offer has recently been extended until February 12, 1988.
I. Removal and the Anti-Injunction Act
According to the decision below, the procedural history of this case is as follows:
At 11:05 A.M. on December 21, 1987, Hyde Park filed in the federal district court its verified complaint seeking declaratory and injunctive relief. Hyde Park requested an ex parte hearing on a motion for a temporary restraining order, but Judge McNaught decided not to handle the matter on an ex parte basis. He set a hearing for the next day at 11:00 A.M. At approximately 4:45 P.M. on the 21st, High Voltage presented a complaint in Middlesex Superi- or Court, seeking to enjoin Hyde Park from purchasing any shares of High Voltage common stock. Such an action is precisely the sort of state statutory enforcement [841]*841that Hyde Park sought to enjoin in its federal action. The state court declined to grant High Voltage’s request for a temporary restraining order that afternoon because of the insufficiency of facts in the complaint.
The next morning, two hours before the federal hearing was to commence, High Voltage presented an amended complaint to the state court clerk. At 9:20 A.M. on December 22, a certified petition for removal of the state court action was hand-delivered by counsel for Hyde Park to the Mid-dlesex County court clerk. Copies of the removal petition were distributed to counsel for High Voltage at 9:45 A.M., and a certificate of removal was filed in the federal court. Notwithstanding the removal petition, the state court proceeded to its hearing on High Voltage’s motion for a temporary restraining order. At some point on the 22nd, High Voltage, in hopes of defeating complete diversity, amended its state court complaint to add other plaintiffs, including a New York corporation, in support of a motion to remand the removed case. Counsel for Hyde Park filed a “Notice of Nullity” in the state court, alleging that this amended complaint was not verified, and thus improper. On December 23rd, Superior Court Justice William H. Welch granted High Voltage’s motion for a temporary restraining order, and scheduled a state court hearing on a preliminary injunction for December 29th. In a memorandum issued the next day, he wrote that, despite the removal petition, “the court believe[d] it ha[d] a right to complete the hearing on the T.R.O. initially requested December 21, 1987.”
Judge McNaught conducted a hearing on Hyde Park’s federal application on December 22nd and 23rd, at the end of which all counsel agreed that the motion be converted into one for preliminary injunction. Judge McNaught was informed on December 24th of Justice Welch’s “temporary restraining order.”
On December 29th, Judge McNaught denied High Voltage’s motion for remand in the state case, and granted the preliminary injunction requested by Hyde Park in the federal case. That injunction, the merits of which are before us now on appeal, in essence prohibits the very state court enforcement action which had already been commenced by High Voltage. Thus the question arises whether we are faced with a substantial conflict implicating the Anti-Injunction Act, 28 U.S.C. § 2283.5
But concerns about the proper balance between federal and state courts, and the deference due the latter by the former, all but evaporate under inspection of the peculiar removal procedures that transpired in this chaotic flurry of litigation. Under 28 U.S.C. § 1446(e) removal is “effected]” as soon as the petition is filed, bond is posted, notice is given to all adverse parties, and a copy of the petition is filed with the clerk of the state court. At that point, “the State court shall proceed no further unless and until the case is remanded.”
In this case, therefore, removal was effected at 9:45 A.M. on December 22nd, when the last of the removal requirements was met. At that point, the jurisdiction of the state court “absolutely ceased, and that of the [federal court] immediately attached.” Steamship Co. v. Tugman, 106 U.S. 118, 122, 1 S.Ct. 58, 60, 27 L.Ed. 87 [842]*842(1882). The subsequent hearing and temporary restraining order by the Superior Court were void ab initio. Polyplastics, Inc. v. Transconex, Inc., 713 F.2d 875, 880 (1st Cir.1983). In fact, the state court had a “duty ... to proceed no further in the cause. Every order thereafter made in that court was coram non judice, unless its jurisdiction was actually restored.” Tugman, 106 U.S. at 122, 1 S.Ct. at 60. This is so even if it eventually is determined that removal was improper. Id. Thus, even if the district court later were to decide to remand because of the addition of a non-diverse state plaintiff, the temporary restraining order previously issued is of no legal effect. Judge McNaught’s preliminary injunction does not deprive High Voltage of the protection of that earlier restraining order, because the latter “was a nullity anyway, with or without the order against further [state] proceedings.” Polyplastics, 713 F.2d at 880.
Even if the temporary restraining order had properly been entered prior to effectuation of removal, that order would still have had no binding effect on Judge McNaught. 28 U.S.C. § 1450 establishes the authority of the federal district court to dissolve or modify all injunctions and orders issued prior to removal. See Granny Goose Foods, Inc. v. Brotherhood of Teamsters, Local 70, 415 U.S. 423, 437, 94 S.Ct. 1113, 1123, 39 L.Ed.2d 435 (1974); Ex parte Fisk, 113 U.S. 713, 725, 5 S.Ct. 724, 729, 28 L.Ed. 1117 (1885). Judge McNaught’s preliminary injunction of December 29th effectively terminated the temporary restraining order earlier granted. This does not violate the Anti-Injunction Act, because the restraining order was itself transferred to federal court jurisdiction once the case was removed. 28 U.S.C. § 1450. Removal was proper even if it was done in order to escape the strictures of the Anti-Injunction Act. See Bondurant v. Watson, 103 U.S. (13 OTTO) 281, 287-88, 26 L.Ed. 447 (1880).6
We now proceed to consider the merits of the appeal.
II. Standard of Review
We review the grant of a preliminary injunction for an abuse of discretion. Planned Parenthood League of Mass. v. Bellotti, 641 F.2d 1006, 1009 (1st Cir.1981) (quoting Charles v. Carey, 627 F.2d 772, 776 (7th Cir.1980)). Such a review may encompass a consideration of whether the district court applied an improper legal standard or misapplied the law to particular facts. Id.
The district court granted Hyde Park's motion for a preliminary injunction based on the following conclusions:
(1) Hyde Park is likely to succeed on the merits of its claim that section 3 of Mass. Gen.L. ch. HOC conflicts with the Commerce Clause;
(2) Hyde Park will suffer immediate irreparable harm if the injunction is not granted because the state has declared that it intends to enforce section 3, which would at least delay, and likely prevent, a takeover bid, causing Hyde Park to lose a valuable right to take over a target company in accordance with applicable federal law, and exposing Hyde Park to the risk of incurring civil and criminal liability;
(3) the injury to Hyde Park outweighs any harm that would be inflicted by granting injunctive relief, since the effect on a target company of a change of ownership is speculative and an injunction may prove beneficial to its shareholders — without a takeover bid, they may lose the opportunity to sell their shares at a more favorable price; and
(4) the public interest, the safeguarding of which is the sole interest of the state in this case, would not be adversely affected [843]*843by the granting of the injunction against the enforcement of the Act, which, it appears likely, is unconstitutional. See Memorandum and Order on Motion for Injunc-tive Relief, December 29,1987 at 6-9. Our primary consideration on appeal is whether there was an abuse of discretion by the district court concerning Hyde Park’s likelihood of success on the merits.
III. Likelihood of Success on the Merits
High Voltage and the state argue that section 3 is valid under the Commerce Clause and, further, that it is not preempted by the Williams Act. We will address each argument in turn.
A. Commerce Clause Analysis
Article I, section 8 of the Constitution gives Congress “Power ... [t]o regulate Commerce with foreign Nations, and among the several States, and with the Indian Tribes.” If a state regulates interstate commerce in a manner inconsistent with that prescribed by Congress, the state regulation is preempted by the federal law, and is therefore constitutionally impermissible. Where Congress has not acted directly, nothing in the Commerce Clause explicitly prohibits the states from regulating interstate commerce. Nevertheless, the Supreme Court has established a doctrine, sometimes denominated the “dormant Commerce Clause,” under which the states are barred from regulating interstate commerce in a manner which significantly interferes with the national economy. See Gibbons v. Ogden, 22 U.S. (9 Wheat.) 1, 6 L.Ed. 23 (1824); Cooley v. Board of Wardens of the Port of Philadelphia, 53 U.S. (12 How.) 299, 13 L.Ed. 996 (1851); Baldwin v. G.A.F. Seelig, Inc., 294 U.S. 511, 55 S.Ct. 497, 79 L.Ed. 1032 (1935); Dean Milk Co. v. City of Madison, 340 U.S. 349, 71 S.Ct. 295, 95 L.Ed. 329 (1951). Perhaps the most eloquent and comprehensive analysis of dormant Commerce Clause doctrine was provided by Justice Jackson in H.P. Hood & Sons v. DuMond, 336 U.S. 525, 69 S.Ct. 657, 93 L.Ed. 865 (1949). In that opinion, he wrote for a majority of the Court:
While the Constitution vests in Congress the power to regulate commerce among the states, it does not say what the states may or may not do in the absence of congressional action, nor how to draw the line between what is and what is not commerce among the states. Perhaps even more than by interpretation of its written word, this Court has advanced the solidarity and prosperity of this Nation by the meaning it has given to these great silences in the Constitution.
Id. at 534-35, 69 S.Ct. at 663-64. As Justice Cardozo noted, state action must not “neutralize the economic consequences of free trade among the states,” Baldwin, 294 U.S. at 526, 55 S.Ct. at 501, because the Constitution “was framed upon the theory that the peoples of the several states must sink or swim together, and that in the long run prosperity and salvation are in union and not division.” Id. at 523, 55 S.Ct. at 500. In light of these principles, the Court “consistently has rebuffed attempts of states to advance their own commercial interests by curtailing the movement of articles of commerce, ... while generally supporting their right to impose even burdensome regulations in the interest of local health and safety.” H.P. Hood & Sons, 336 U.S. at 535, 69 S.Ct. at 663.
Just last year, in CTS Corp. v. Dynamics Corp. of America, — U.S. —, 107 S.Ct. 1637, 95 L.Ed.2d 67 (1987), the Court articulated a basic analysis for testing a statute’s constitutionality under the dormant Commerce Clause, applicable to review of takeover statutes. In CTS, an Indiana statute provided that the owner of control shares in certain defined corporations acquired voting rights only to the extent granted by collective resolution of the other shareholders. Id. 107 S.Ct. at 1641. Control shares were defined as shares that, but for the operation of the statute, would bring the acquiror’s voting power in the corporation to or above any of three thresholds: 20%, 33V8%, or 50%. Id. The practical effect of the statute was “to condition acquisition of control of a corporation on approval of a majority of the pre-existing disinterested shareholders.” Id. (footnote omitted). Six justices conclud[844]*844ed that the statute did not violate the Commerce Clause. The Court adopted a two-pronged test for its Commerce Clause analysis: (1) Does the statute discriminate against interstate commerce? and (2) Does the statute adversely affect interstate commerce by subjecting activities to inconsistent regulations? 107 S.Ct. at 1648-49.
In the case at bar, section 3 of the Massachusetts Act is non-discriminatory, because it applies equally to both intrastate and interstate offerors.
We are also persuaded that section 3, as applied in this case, poses no risk of inconsistent regulation by different states upon particular takeover activities. Section 1 of the statute defines a “target company” as a corporation organized under the laws of, or having its principal place of business in, the Commonwealth of Massachusetts. As to an out-of-state corporation, the statute would not likely survive Commerce Clause analysis under CTS because the corporation might be subject to regulations in both Massachusetts and its state of incorporation as to the same transaction. See 107 S.Ct. at 1649. But the particular target in this case, High Voltage, has its principal place of business in Massachusetts and is incorporated there. No other state could have a legitimate interest in regulating takeover bids for a corporation neither incorporated in that state nor having its principal place of business there. Therefore there is no risk of conflict or inconsistent regulation on these facts.7
Traditionally, the next step in the Commerce Clause analysis would be to balance the local benefits of the statute against any incidental burdens the statute might impose on interstate commerce. Pike v. Bruce Church, Inc., 397 U.S. 137, 142, 90 S.Ct. 844, 847, 25 L.Ed.2d 174 (1970) (“Where the statute regulates evenhandedly to effectuate a legitimate local public interest, and its effects on interstate commerce are only incidental, it will be upheld unless the burden imposed on such commerce is clearly excessive in relation to the putative local benefits.”). See also Edgar v. MITE Corp., 457 U.S. 624, 643-46, 102 S.Ct. 2629, 2641-43, 73 L.Ed.2d 269 (1982).
There is, however, some support for the view that any Commerce Clause analysis now is, or at least ought to be, complete with the two above-described inquiries, and that the Pike balancing test is obsolete. See, e.g., CTS, 107 S.Ct. at 1652-53 (Scalia, J., concurring); Langevoort, Comment— The Supreme Court and the Politics of Corporate Takeovers: A Comment on CTS Corp. v. Dynamics Corp. of America, 101 Harv.L.Rev. 96, 103-04 (1987); Regan, Siamese Essays: (I) CTS Corp. v. Dynamics Corp. of America and Dormant Commerce Clause Doctrine; (II) Extraterritorial State Legislation, 85 Mich.L.Rev. 1865, 1866-68 (1987). The balancing test does seem to have been abandoned by the CTS Court in cases where the state law merely regulates intrastate “corporate governance.” However, the CTS Court did not indicate that the Pike balancing test was to be abandoned in cases where the legislature’s purpose was not simply to regulate intrastate corporate governance.8 The Court specifically distinguished the Indiana statute in CTS from the type of statute invalidated in MITE, which prohibits entities “from offering to purchase, or from purchasing, shares in [state] corporations, or from attempting thereby to gain control.” 107 S.Ct. at 1652.
We can glean, then, some basic tenets of dormant Commerce Clause doctrine. A state may impose certain nondiscriminatory regulations on sales of goods within its borders, even if those goods originated out of state. See, e.g., Minnesota v. [845]*845Clover Leaf Creamery Co., 449 U.S. 456, 101 S.Ct. 715, 66 L.Ed.2d 659 (1981) (upholding statute that forbade the sale of milk in Minnesota in plastic nonreturnable containers). The state may also regulate internal corporate governance, even if such regulation has a substantial effect on interstate commerce, as long as the provisions are not discriminatory on their face. See CTS. But the direct regulation of interstate transactions of goods triggers the Pike balancing analysis, even if the goods originate in the legislating state. See, e.g., Pike (applying balancing test to, and invalidating, Arizona statute that required that all cantaloupes grown in Arizona be packed in standardized containers, even if cantaloupes were to be sold out of state); MITE, 457 U.S. at 643-46, 102 S.Ct. at 2641-43 (using Pike balancing analysis, Court invalidated Illinois statute that proscribed tender offers in certain situations).9
Section 3 of the Massachusetts statute falls in this latter category. The state creates corporations and empowers them to send shares of stock into interstate commerce. Now it seeks to attach various strings, consisting of conditions and prohibitions, to the sale of those shares throughout the United States. The penalty provision of section 3 of the Massachusetts Act places a one-year moratorium on takeover attempts as a sanction for noncompliance with section 3’s disclosure provisions. It is therefore analagous to the statute invalidated by a majority of the Court in MITE, since it directly prohibits the interstate sale of shares. The disclosure provision itself, while not falling in the category of prohibiting sales, also is distinguishable from the sort of internal corporate “governance” statute found in CTS, because it deals with disclosure of the offerors’ control intent, rather than with regulation of stockholders’ power “to define the rights that are acquired by purchasing their shares.” Id. 107 S.Ct. at 1650. Thus, it conditions the interstate transfer of goods, and falls somewhere in between the regulations in MITE and CTS.
Accordingly, we shall apply the Pike balancing test to both provisions. In striking this balance, we are sensitive to the CTS Court’s emphasis on any possible intent the legislature may have had to benefit stockholders. We also, in a more traditional vein, focus on the actual local and interstate effects of section 3.
We begin with the disclosure provision. High Voltage and the state assert that section 3 benefits local investors by ensuring that they have the opportunity to obtain the premium, usually paid in connection with acquiring control shares, during the “toehold” period when an existing control intent would otherwise have remained secret.
This purported interest is undercut, however, by a number of factors. First, the notion of local shareholder protection was questioned by a majority of the Court in MITE, which recognized that such protection applies to all shareholders, not simply those within the state. “While protecting local investors is plainly a legitimate state [846]*846objective, the State has no legitimate interest in protecting nonresident shareholders.” 457 U.S. at 644, 102 S.Ct. at 2641.
Second, Hyde Park contends that the disclosure provision will actually hurt investors by decreasing the overall number of takeover bids made for Massachusetts companies. The disclosure before 5% allegedly eviscerates the “toehold” position of the potential bidder, a position that normally would enhance the credibility of the acqui-ror with the market and the target management and increase the offeror’s ability to wage a proxy fight. Hyde Park also argues that the profits máde on the “control premium,” profits which would disappear under section 3, are often a necessary incentive to offset the enormous costs of a takeover bid.
It is not clear whether a decrease in takeover bids would help or hurt local shareholders. Hyde Park argues that if takeovers decline, the asserted “control premium” for the stockholders would disappear. Hyde Park also contends that fewer tender offers would reduce shareholders’ opportunities to sell their shares at the takeover premium rate (or in a market where prices are elevated by the buying activity). See MITE, 457 U.S. at 643, 102 S.Ct. at 2641 (interstate commerce would itself be unduly burdened by a decrease in takeover bids because management will lose the incentive to perform well and because the “reallocation of economic resources to their highest valued use, a process which can improve efficiency and competition, [would be] hindered”). The CTS Court, however, while acknowledging the force of this argument, noted that the increasingly “divergent views in the literature ... reflect the reality that the type and utility of tender offers vary widely.” 107 S.Ct. at 1651 n. 13. Thus, the Court reflected, “there is no reason to assume that the type of conglomerate corporation that may result from repetitive takeovers necessarily will result i[n] more effective management or otherwise be beneficial to shareholders.” Id. (emphasis in original). This dictum reflects a burgeoning “sense of intellectual ambivalence” toward the value of takeovers, see Langevoort, The Supreme Court and the Politics of Corporate Takeovers, 101 Harv.L.Rev. at 102-03 & nn. 40-43. It is therefore difficult to decide at this stage of the proceedings in this case whether a decrease in the number of tender offers would or would not benefit Massachusetts shareholders and Massachusetts corporate governance.
High Voltage further contends that the disclosure provision will serve another local interest by encouraging investment in Massachusetts corporations. Such an assertion suggests that a further purpose of the Massachusetts statute is to protect local businesses from outside takeover by discouraging tender offers.10 This would seem to be weak support for appellants’ case, because a unanimous Supreme Court has held that mere protection of local economic interests is generally an impermissible purpose under the Commerce Clause. See Lewis v. BT Investment Managers, Inc., 447 U.S. 27, 43-44, 100 S.Ct. 2009, 2019-20, 64 L.Ed.2d 702 (1980). But in MITE, two justices joined the Court’s judgment on the understanding that a state may have a substantial interest in protecting against large-scale upheavals in the labor market, disturbances that could seriously disrupt not only the labor market, but the local cultural, charitable, and educational life as well. 457 U.S. at 646-47 & n. *, 102 S.Ct. at 2642-43 & n. *. (Powell, J., concurring); id. at 655, 102 S.Ct. at 2647 (Stevens, J., concurring in judgment). See also Regan, Siamese Essays, 85 Mich.L. Rev. at 1871 & n. 33.11
[847]*847Assuming that the protection of local business stability is in some instances constitutionally permissible, Hyde Park counters by asserting that the fostering of investments in Massachusetts-based businesses allegedly produced by section 3 protections is illusory, since there would be less incentive to invest in those companies that might be less efficiently managed due to the dearth of free market mechanisms. Also, the securities marketplace might “discipline states with overly protectionist takeover laws by devaluing the shares of corporations located there.” Langevoort, The Supreme Court and the Politics of Corporate Takeovers, 101 Harv.L.Rev. at 109. See also the sources cited in id. at 109 n. 63.
In summary, we believe that the Court’s opinion in CTS, which the district court appears to have discounted, has relevance to the case before us and indicates that at least some of the reasoning upon which the district court based its determination of the likelihood of the invalidity of section 3 on Commerce Clause grounds perhaps should be reconsidered, or at least qualified. It is not entirely clear that enforcement of the disclosure provision of section 3 would substantially chill tender offers in Massachusetts. But even if it did, it is hard to judge whether the resultant benefits to state shareholders and to corporate stability would be outweighed by the disadvantages that would be suffered by those same stockholders and the burdens that would be placed on the interstate market.
In view of the substantial arguments we believe each side exhibits on the merits, we cannot say with great confidence that Hyde Park has shown a likelihood of success on the merits of its claim that the disclosure provision of section 3 works a violation of the dormant Commerce Clause.
The penalty provision of section 3 is a different story, however. Rather than merely conditioning sales, and thus deterring them, it prohibits certain sales outright. Section 3 bars an offeror from making a takeover bid for one year after failure to disclose control intent before the 5% threshold. Such a ban is like the direct restraint imposed on bids in the Illinois statute in MITE, and is thus distinguishable from the CTS analysis, which explicitly dealt with an act that did not prohibit any entity from purchasing any shares or attempting to gain control. 107 S.Ct. at 1652. The effect on interstate commerce, then, is direct and substantial.
High Voltage and the state respond that one who fails to comply with what they contend is a minimally burdensome disclosure requirement has little basis on which to complain of the penalty for noncompliance. Furthermore, they argue, the penalty assists stockholders by deterring violations of the disclosure provision.
We see no reason why we should adopt the state’s argument that, if the disclosure provision itself is constitutional, the attendant deterrent should be set aside only if it is preempted or not rationally related to deterring violations of the disclosure provision. The Supreme Court has held that “the extent of the burden [on interstate commerce] that will be tolerated will of course depend on ... whether it could be promoted as well with a lesser impact on interstate activities.” Pike, 397 U.S. at 142, 90 S.Ct. at 847. This principle would seem to apply even to so-called “penalty” provisions.
The particular penalty prescribed here seems to be precisely that which is most disruptive of the commerce interests identified in MITE and CTS. Penalties analogous to the recission and damage remedies provided in section 9 of the Massachusetts statute presumably would be adequate to enforce the disclosure requirement. A heavy fine for nondisclosure surely would provide an equivalent deterrent, but it likely would not so clearly and directly offend the interests to be protected under the Commerce Clause. While the state might [848]*848not be required to provide the least restrictive alternative, it certainly is required to fashion its penalty in a manner that is solicitous of the interests of the investors who are the putative beneficiaries of the statute. Although the deterrent value of section 3 might inure to the benefit of shareholders, the chosen penalty itself will only serve to harm the investment options of those shareholders by precluding them for up to one year from selling to the non-disclosing offeror at a tender offer premium rate.
Notwithstanding that we previously upheld the section 3 penalty against a preemption challenge, Agency Rent-A-Car, Inc. v. Connolly, 686 F.2d 1029 (1st Cir.1982), we left open the possibility that this provision might implicate substantial Commerce Clause concerns. Id. at 1040. We continue to believe that the provision presents “very serious and substantial questions, leaving its validity very much in doubt.” Id. CTS does not allay our doubts. The statute in that case, we reiterate, did not, according to the Court, “prohibit any entity — resident or nonresident— from offering to purchase, or from purchasing, shares in Indiana corporations, or from attempting thereby to gain control.” 107 S.Ct. at 1652. The penalty moratorium in section 3 of the Massachusetts statute, by contrast, does just that. We think that the burdens on interstate commerce occasioned by this particular deterrent are substantial. On the other side of the equation, the deterrent is the only benefit to the local investors. The means used to realize that deterrent themselves hurt the shareholders rather than help them. On balance, the burden on interstate commerce seems, at least at this juncture, likely to outweigh whatever negligible local benefits that might result.
In summary, we must review whether the district court abused its discretion or erred as a matter of law in determining that Hyde Park would likely succeed on the merits of its Commerce Clause argument. We have taken note of the arguments advanced concerning the purported benefits relative to the alleged burdens of section 3. We find that although it cannot be said with certainty that the disclosure provision of section 3 itself violates the dormant Commerce Clause, the penalty prescribed for violations of that provision is not likely to pass constitutional muster.
B. Preemption Analysis
We believe in addition that section 3 is probably preempted by the federal Williams Act, 82 Stat. 454, as amended, 15 U.S.C. §§ 78m(dHe) and 78n(d)-(f). The district court did not face this issue. For this reason, undoubtedly, the major focus of this appeal has been the Commerce Clause. Yet counsel have minimally addressed preemption in their briefs, and we believe it necessary to address it herein. We are not precluded from doing so by Agency Rent-A-Car, Inc. v. Connolly, 686 F.2d 1029 (1st Cir.1982). In that case, we held that the penalty provision of section 3 was not preempted by the Williams Act on the presumption, stipulated by the parties, that the disclosure provision was consistent with the Williams Act. In this case, Hyde Park challenges that presumption. Agency Rent-A-Car therefore is no bar to our preemption analysis.12
The Supreme Court has recently summarized the variants of preemption doctrine.
The Supremacy Clause of Art. VI of the Constitution provides Congress with the power to pre-empt state law. Pre-emption occurs when Congress, in enacting a [849]*849federal statute, expresses a clear intent to pre-empt state law, Jones v. Rath Packing Co., 430 U.S. 519 [97 S.Ct. 1305, 51 L.Ed.2d 604] (1977), when there is outright or actual conflict between federal and state law, e.g., Free v. Bland, 369 U.S. 663 [82 S.Ct. 1089, 8 L.Ed.2d 180] (1962), where compliance with both federal and state law is in effect physically impossible, Florida Lime & Avocado Growers, Inc. v. Paul, 373 U.S. 132 [83 S.Ct. 1210, 10 L.Ed.2d 248] (1963), where there is implicit in federal law a barier to state regulation, Shaw v. Delta Air Lines, Inc., 463 U.S. 85 [103 S.Ct. 2890, 77 L.Ed.2d 490] (1983), where Congress has legislated comprehensively, thus occupying an entire field of regulation and leaving no room for the States to supplement federal law, Rice v. Santa Fe Elevator Corp., 331 U.S. 218 [67 S.Ct. 1146, 91 L.Ed. 1447] (1947), or where the state law stands as an obstacle to the accomplishment and execution of the full objectives of Congress. Hines v. Davidowitz, 312 U.S. 52 [61 S.Ct. 399, 85 L.Ed. 581] (1941).
The critical question in any pre-emption analysis is always whether Congress intended that federal regulation supersede state law.
Louisiana Public Service Comm’n v. Federal Communications Comm’n, 476 U.S. 355, 106 S.Ct. 1890, 1898-99, 90 L.Ed.2d 369 (1986). See also Palmer v. Liggett Group, Inc., 825 F.2d 620, 624-25 (1st Cir.1987); Massachusetts Medical Soc. v. Dukakis, 815 F.2d 790 (1st Cir.1987).
There is no explicit indication that Congress intended that the Williams Act would completely preempt state law. In fact, 15 U.S.C. § 78bb(a) provides in pertinent part:
Nothing in this chapter [chapter 2B — Securities Exchanges] shall affect the jurisdiction of the securities commission (or any agency or officer performing like functions) of any State over any security or any person insofar as it does not conflict with the provisions of this chapter or the rules and regulations thereunder.
It also appears undisputed that most of the other preemption problems are not present here; for instance, it is possible for entities to comply with both the Williams Act and section 3 of the Massachusetts statute. Therefore, we review section 3 only to determine whether “there is outright or actual conflict between federal and state law,” and, most importantly, whether section 3 “stands as an obstacle to the accomplishment and execution of the full objectives of Congress.” Louisiana Public Service Comm’n, 106 S.Ct. at 1898. Accord CTS, 107 S.Ct. at 1644; MITE, 457 U.S. at 631-32, 102 S.Ct. at 2634-35 (plurality opinion).
According to the Supreme Court, “the sole purpose of the Williams Act was the protection of investors who are confronted with a tender offer.” Piper v. Chris-Craft Industries, 430 U.S. 1, 35, 97 S.Ct. 926, 946, 51 L.Ed.2d 124 (1977). Before passage of the Williams Act in 1968, hostile tender offers were not regulated by the disclosure requirements of the federal securities laws. The Act was designed to give the investor fuller information regarding the tender offer, so that the investor can make knowledgeable decisions about whether to sell. The three-judge plurality in MITE acknowledged that the basic concern of the Williams Act was the protection of investors. 457 U.S. at 633, 102 S.Ct. at 2636. The plurality went on to say that “a major aspect of the effort to protect the investor was to avoid favoring either management or the takeover bidder.” Id. Hyde Park seizes on this “neutrality” ideal for the proposition that section 3 is preempted by the Williams Act if it in any way tips the balance between offeror and management.
But the Court further explained in CTS that neutrality between management and bidder is the means to the end of investor protection, rather than the objective itself. Justice Powell, writing for a majority of the Court, shed some light on the Williams Act principles to be honored in preemption analysis:
As the plurality opinion in MITE did not represent the views of a majority of the Court, we are not bound by its reasoning. We need not question that reasoning, [850]*850however, because we believe the Indiana Act passes muster even under the broad interpretation of the Williams Act articulated by Justice WHITE in MITE. As is apparent from our summary of its reasoning, the overriding concern of the MITE plurality was that the Illinois statute considered in that case operated to favor management against offerors, to the detriment of shareholders. By contrast, the statute now before the Court protects the independent shareholder against both of the contending parties. Thus, the Act furthers a basic purpose of the Williams Act, “‘plac[ing] investors on an equal footing with the takeover bidder,’” Piper v. Chris-Craft Industries, 430 U.S. at 30, 97 S.Ct. at 943 (quoting the Senate Report accompanying the Williams Act, S.Rep. No. 550, 90th Cong., 1st Sess., 4 (1967)).
107 S.Ct. at 1645-46 (footnotes omitted). See also Piper, 430 U.S. at 29, 97 S.Ct. at 943 (“Congress was indeed committed to a policy of neutrality in contests for control, but its policy of evenhandedness does not go ... to the purpose of the legislation_ Neutrality is, rather, but one characteristic of legislation directed toward a different purpose — the protection of investors.”); MITE, 457 U.S. at 646-47, 102 S.Ct. at 2642-43 (Powell, J., concurring) (incidental protection of local management does not contradict Williams Act’s neutrality policy); id. at 655, 102 S.Ct. at 2647 (Stevens, J., concurring in judgment) (same).
The Court acknowledged that the Indiana statute at issue in CTS might, by deterring tender offers, tip the balance between management and offerors to the benefit of the former. Nonetheless, the Court held that protection of management that is incidental to protection of investors does not per se conflict with the purpose or purposes of the Williams Act.
Of course, by regulating tender offers, the Act makes them more expensive and thus deters them somewhat, but this type of reasonable regulation does not alter the balance between management and offeror in any significant way. The principal result of the Act is to grant shareholders the power to deliberate collectively about the merits of tender offers. This result is fully in accord with the purposes of the Williams Act.
107 S.Ct. at 1646 n. 7 (emphasis supplied).
This is the clearest indication of the test to be applied to anti-takeover statutes. Our task is to identify the principal result of section 3, and to distinguish between that effect and any consequences that are merely secondary or incidental. One probative test is to determine if the regulation alters the balance between management and offeror in any significant way, but this is not in itself dispositive. Our focus should remain on determining whether the disclosure provisions are beneficial to the investors caught between management and offerors.
For reasons expressed below, we conclude that the disclosure provisions of section 3 are likely to act primarily to the detriment of shareholders, because the deterrent to tender offers caused by the disclosure outweighs any additional information that investors might attain because of the earlier dissemination of control intent. Because it is so difficult accurately to determine the costs and benefits to investors from such regulation, we would usually be reluctant to decide that shareholders are hurt by such legislation. See supra at 846. We are aided, however, by the fact that Congress explicitly and carefully established a preferred balance in the Williams Act at a point where the benefit gained by shareholders from early disclosure is not outweighed by the detriment to those same investors if there is too great a deterrent to tender offers. Section 3 of the Massachusetts Act in effect second-guesses the balance struck by Congress, by altering the point in time at which control intent should be made public. There is, in other words, an “actual conflict between federal and state law” concerning a matter that directly implicates the central purpose of the Congressional Act. We need not undertake too detailed an economic analysis of the effect on investors of the disclosure law, because Congress itself has already resolved definitively what scheme of disclo[851]*851sure regulation will best vindicate its goal of benefitting shareholders.
We begin our analysis by looking to the asserted purposes of the Massachusetts statute. The preamble to the enactment of chapter HOC states that the ostensible purpose of that chapter is, in part, to protect shareholders.
Whereas, The deferred operation of this act would tend to defeat its purpose, which is, in part, to protect shareholders of corporations organized under the laws of the commonwealth or having principal places of business within the commonwealth, therefor it is hereby declared to be an emergency law, necessary for the immediate preservation of the public convenience.
1976 Mass. Acts 121.
High Voltage and the state contend that the principal effect of section 3 is to protect shareholders by putting them on an equal footing with a takeover bidder, thus letting investors know of a bidder’s existing, but undisclosed, intent to acquire control of the corporation. This will, it is argued, give the shareholder a more accurate view of the true value of its holdings, and thus allow the investor to reap the value of the “control premium” on the shares sold prior to the tender offer.
Pursuant to the Williams Act, on the other hand, one who acquires more than a 5% stake in a corporation must publicly disclose the existence of such a stake, including whether the purpose of the purchases is to acquire control, within ten days after the acquisition of such a stake. 15 U.S.C. § 78m(d)(l).
The Williams Act thus treats differently the open-market transactions preceding the tender offer and the sales made pursuant to the tender offer itself. As to the latter, the Act requires that disclosure of holdings be made by the time of commencement of the offer. 15 U.S.C. § 78n(d)(l). But as long as the tender offer has not yet been made, disclosure of 5% holdings, and of control intent, need not be made until 10 days after such a threshold is reached. See Ludlow Corp. v. Tyco Laboratories, Inc., 529 F.Supp. 62 (D.Mass.1981) (explaining differences between tender offer acquisitions and pre-tender offer open-market acquisitions under Congress’s “two pronged, approach”).
In introducing the bill that was to become the Williams Act, Senator Williams of New Jersey described the different treatment of open-market acquisitions:
In addition, this bill provides that anyone acquiring the beneficial ownership of more than 10 percent of a class of equity security registered under the Securities Exchange Act file with the SEC substantially the same information as that Required with respect to a cash tender offer; however, such a filing would not be required until 7 days after the acquisition. This statément would have to be sent tó each exchange on which the security is listed and to the principal executive offices of the issuer of the security. Substantial open market or privately negotiated purchases or shares may precede or accompany a tender offer or may otherwise relate to shifts in control of which investors should be aware. While some people might say that this information should be filed before the securities are acquired, disclosure after the transaction avoids upsetting the free and open auction market where buyer and seller normally do not disclose the extent of their interest and avoids prematurely disclosing the terms of privately negotiated transactions.
113 Cong.Rec. 854, 856 (1967). The bill, as enacted, enlarged the delay in disclosure from seven to ten days. Two years later, Congress lowered to 5% the threshold of equity securities that triggers disclosure. Pub.L. Ño. 91 — 5é7, § l(á), 84 Stat. 1497 (1970).
This legislative history thus reflects congressional consideration of the optimal balance between investor awareness of control intent prior to acquisition of a certain percentage of stock and the operation of a free and open auction market. It indicates a congressional judgment, in balancing the relative merits, that open-market transactions Should be unregulated prior to the acquisition of a 5% stake and during the [852]*852ten-day period after such acquisition. Any earlier disclosure would be, in Senator Williams’s term, “premature.” Most importantly, this balance was struck for the benefit of investors. The Williams Act was intended to provide shareholders with the best of both worlds — disclosure substantial enough and early enough to ensure fully informed choices, but not so early that those choices will be unduly restricted by the chilling effect on takeover bids.
It cannot be assumed that this point of equipoise was randomly or uncarefully chosen by Congress, such that a slight variation (for instance, that in the Massachusetts statute) would be consistent with the congressional intent. Faced in Senator Williams’s original bill with a seven-day grace period for disclosure of 10% holdings, Congress specifically changed the Act to provide a ten-day period, and then amended the Act two years later to change the threshold to 5% of the target’s stock. “[W]here there is in existence a pervasive and comprehensive scheme of federal regulation ... pre-emption follows in order to fulfill the federal statutory purposes.” Merrill Lynch, Pierce, Fenner & Smith v. Ware, 414 U.S. 117, 139, 94 S.Ct. 383, 396, 38 L.Ed.2d 348 (1973). The Williams Act admittedly is not an exclusive scheme for regulating takeover bids. It does, however, establish a “pervasive and comprehensive scheme” for regulating disclosure of open-market acquisitions prior to a tender offer.
Section 3 of the Massachusetts law intrudes upon this carefully regulated scheme. The effect of section 3 is to remove from a purchaser who intends to acquire control the right to increase its stake beyond 5% before it must disclose the existence of such a stake pursuant to the Williams Act at the end of ten days. This likely will discourage takeover attempts to a much greater extent than that envisioned by Congress. We can deduce, then, that, according to the calculations of Congress, the investor will be made worse off. The essential purpose of the Williams Act is therefore likely to be defeated, and we must conclude that section 3 has a high probability of being preempted.13
This conclusion is supported by further evidence that section 3 is designed to assist management, rather than investors.14 Even assuming that investor protection was one of the legislature’s purposes in enacting section 3, and that such protection does in some way result, we believe that those benefits are, at best, merely incidental to the principal effect of altering the balance between management and offerors in a manner that ultimately, by restricting too many takeover attempts, works to the detriment of investors.
Obviously, incumbent management also receives advance notice of an acquiror’s control interest via the disclosure. This gives management more time to prepare defensive maneuvers designed to defeat the takeover bid. More significant is the fact that section 1 of the statute excludes negotiated, “friendly” takeovers from the definition of takeover bids. In effect, management can opt in or out of section 3 coverage, a powerful weapon in fending off a hostile takeover attempt. It seems clear that this provision was designed at least in part to favor management. Not providing shareholders with disclosure information in the friendly takeover situation also conflicts with Congress’s determination to protect investors by respecting their autonomy. By giving management the option of [853]*853effectively waiving the disclosure and penalty provisions of section 3, the Massachusetts Act lets management decide for investors instead of letting investors decide for themselves. Not only does this indicate that the statute does not assist investor autonomy in the friendly takeover situation, it also suggests that the disclosure rules in hostile takeover situations are themselves principally designed for the benefit of management, with “control premium” protection for shareholders at best an incidental or secondary concern.15
We conclude, therefore, that the most likely principal result of the disclosure provisions of section 3 would be to alter the balance carefully constructed by Congress between providing helpful information to investors and encouraging tender offers that also may be beneficial to investors. The imbalance caused by section 3 works against the best interest of the shareholders, as that interest has been identified by Congress in respect to pre-tender offer disclosure of “open market” securities acquisitions.
Because the disclosure provision itself likely is preempted, it is but a short logical step to concluding that the penalty provision is probably preempted as well. As we explained in the Commerce Clause discussion, supra, the only possible benefit to investors from the one-year moratorium is the deterrent against violations of the disclosure provision of section 3. Once it is established that those provisions are themselves not primarily beneficial to investors, there is no longer any doubt that the penalty also cannot be in the interest of shareholders, since it only serves to decrease their ability to take advantage of tender offers.
This result is not foreclosed by Agency Rent-A-Car. In that case, we held that the penalty provision was not preempted by the Williams Act. This holding, however, was explicitly premised on the parties’ stipulation in that case that the disclosure provision itself was beneficial to investors. See id., 686 F.2d at 1036 (“Agency does not challenge the disclosure ... requirements of Chapter HOC”; those requirements are “concededly valid”); id. at 1038 (“The validity of section 3’s regulation of creeping tender offers is not in itself attacked.”); id. at 1038 n. 10 (“the disclosure provisions in the context of this case are assumed by the parties to be valid”). That presumption is not in effect here; the disclosure provision has itself been challenged on preemption grounds. Thus, the factual predicate of our decision in Agency is missing, and our decision there is not binding. See also supra at 848-849 & n. 12.
IV. Irreparable Injuries and the Public Interest
Having concluded that Hyde Park has shown a likelihood of success on the merits,16 we turn to the other elements of the preliminary injunction standard. If the injunction were to be overturned, appellants would undoubtedly move to enforce the penalty provisions of section 3, which would delay appellees’ tender offer for at least one year. The harm resulting from this delay would be substantial and irreparable. It is obvious that timing is everything with tender offers, and that such a delay would effectively kill the takeover bid in its present form. At the very least, it would allow management to institute substantial defensive measures on which it otherwise could not depend.
If we are wrong about the constitutionality of the statute, on the other hand, there will certainly be harm to the management of High Voltage, because of the costs of [854]*854the impending takeover. Whether the company itself or its shareholders would be injured is not so certain. First, the disclosure violation has already occurred; it cannot be undone. Second, the deterrent to future violations would be intact if the constitutionality of the law is ultimately upheld. Finally, it is impossible to say whether this particular takeover bid would or would not be to the benefit of the stockholders. The irreparable harm to High Voltage, therefore, is speculative. See Public Service of New Hampshire v. West Newbury, 835 F.2d 380, 383 (1st Cir.1987).
“[Sjhould the statute be valid, the public interest would obviously be adversely affected by a preliminary injunction enjoining its enforcement.” Agency Rent-A-Car, 686 F.2d at 1035. Just as obviously, should the statute be unconstitutional, the public interest would be adversely affected by denial of such an injunction.
In this case, we are not persuaded that the district court abused its discretion in finding that High Voltage’s arguments concerning its irreparable harm and the harm to the public interest did not outweigh the harm Hyde Park would suffer were its takeover attempt delayed by a denial of the preliminary injunction.
Accordingly, the order of the district court granting a preliminary injunction is affirmed.