Citizens for John W. Moore Party v. Board of Election Commissioners of the City of Chicago, Defendants
794 F.2d 1254
7th Cir.
1986
Check Treatment
EASTERBROOK, Circuit Judge.

John W. Moore and some of his friends circulated petitions that put him on the ballot of a primary election in March 1982; Moore sought nomination as the Democratic Party's candidate for state senator from the 16th senatorial district. Although Moore qualified for the ballot, he withdrew before the election. Then Moore and essentially the same group of friends formed the “Citizens for John W. Moore Party” and collected signatures to place Moore on the ballot for the general election in November 1982 as a candidate for the Illinois House of Representatives in the 31st house district. Moore collected 3,829 signatures; he needed 1,500 valid signatures to qualify.

The Board of Election Commissioners threw out 1,493 signatures for reasons Moore does not challenge and another 975 because Moore himself had collected them, leaving Moore 139 signatures short. (Moore told us at oral argument that almost all of the Moore Party’s signatures are invalid under § 10-4, although only Moore’s were challenged.) The Board set aside the signatures Moore had collected because the Moore Party was the second party for which Moore had solicited signatures in the 1982 election season — the Democratic Party being the first. The Board concluded that these signatures did not count, citing § 10-4 of the Illinois Election Code, Ill.Rev.Stat. ch. 46 § 10-4, which provides: “[N]o person shall circulate or certify petitions for candidates of more than one political party ... to be voted upon at *1256 the next primary or general election.” Moore and his allies simply were unaware of § 10-4, and therefore they did not try to collect signatures with the aid of people who had not circulated petitions for another party.

Moore contends that § 10-4 is unconstitutional. The district court denied Moore’s request for an injunction that would have put him on the ballot in 1982. The case is not moot, however. See Storer v. Brown, 415 U.S. 724, 737 n. 8, 94 S.Ct. 1274, 1282 n. 8, 39 L.Ed.2d 714 (1974); Moore v. Ogilvie, 394 U.S. 814, 816, 89 S.Ct. 1493, 1494, 23 L.Ed.2d 1 (1969). The district court held a trial on Moore’s request for a declaratory judgment. The court heard from expert witnesses who both explained the genesis of the statute and gave differing versions of its effects. The court concluded that § 10-4 reduces the confusion that may occur if circulators switch sides in the same campaign. It also helps keep intra-party squabbles within the party by discouraging people who have participated in a primary from running as candidates of new parties. 599 F.Supp. 662 (N.D.Ill.1984).

The court conceded that the statute affects rights secured by the first and fourteenth amendments but found the effect small. Moore came within 139 votes of success even though ignorant of the rule, and the “evidence does not prove that there was a shortage of circulators or that the burden imposed outweighs the statutory objectives and benefits.” 599 F.Supp. at 670. Although the court thought “strict scrutiny” unnecessary, it held that because the “statute did not prevent Moore’s candidacy” (ibid.) and the state’s objectives are substantial the statute would survive such scrutiny. It concluded: “The statute requires that circulators make a choice of political parties and remain with that party or independent candidate throughout an election. This is not an unconstitutional burden.” Ibid.

After we heard oral argument, we certified to the Supreme Court of Illinois the question whether “person,” within the meaning of § 10-4, includes a candidate soliciting signatures on his own behalf. 781 F.2d 581 (7th Cir.1986). On January 23, 1986, the Supreme Court of Illinois entered an order declining to answer this question of law. We therefore proceed to decision on the assumption that “any person” includes Moore — an assumption the parties, the Board, and the district court share. This treatment does not imply, however, that the constitutional analysis of candidate-circulators and other circulators is identical. We discuss this difference in Part IV.

I

The principal effect of § 10-4 is to make it harder for someone to run on two party lines during the same election season or to defect from a party and run as an independent. A person who remains in the same party may use one group of solicitors to obtain signatures for as many offices as he seeks, in primary and general elections. A person who changes parties or switches from a party to an independent candidacy will need to attract another group of people to circulate petitions.

Although § 10-4 creates this relative disadvantage, it does not affect the total number of signatures the candidate must solicit. Moore needed 600 signatures to get on the primary ballot for the senate, and he needed 1,500 to get on the general election ballot for the house whether or not he switched parties. Section 10-4 does not affect the number of signatures needed and therefore does not affect the number of circulator-hours needed to obtain the signatures. It affects only the identity of those who circulate petitions. A candidate who switches parties in an election campaign must demonstrate a somewhat broader base of support by being able to attract a new group of circulators.

Moore pins his principal hopes on persuading us to apply an elevated standard of scrutiny to § 10-4. He argues that the statute should be subjected to strict scrutiny and held unconstitutional because not sufficiently well tailored to its ends and because the state did not introduce at trial *1257 evidence proving that it achieves its objectives. Failing that, Moore would settle for some intermediate level of scrutiny.

Moore does not want for cases, some in the Supreme Court and many in the lower courts, that have applied strict scrutiny to laws restricting access to the ballot. These cases reason that laws controlling access affect rights of both speech and association, hence the need for extraordinary scrutiny. Paradoxically, some of the cases announcing the most rigorous scrutiny also sustain the most onerous statutes. See Storer v. Brown, 415 U.S. 724, 94 S.Ct. 1274, 39 L.Ed.2d 714 (1974); and American Party of Texas v. White, 415 U.S. 767, 94 S.Ct. 1296, 39 L.Ed.2d 744 (1974), two cases to which we return.

The defendants do not want for cases applying quite relaxed scrutiny to statutes affecting candidates’ right to run. For example, Clements v. Fashing, 457 U.S. 957, 969-70, 102 S.Ct. 2836, 2846-47, 73 L.Ed.2d 508 (1982) (plurality opinion), applies the step-at-a-time doctrine to a statute that imposes on would-be candidates the price of resignation from their current jobs. The step-at-a-time doctrine is a lax form of scrutiny; a court using it dismisses an argument that the law is irrational with the response: “Sure, the law is not very sensible in its current form, but the legislature is entitled to address one problem at a time and doubtless will get round to fixing things up later.”

The most recent decision of the Supreme Court suggests that courts should avoid putting decisions in terms of a “standard of scrutiny.” Anderson v. Celebrezze, 460 U.S. 780, 789, 103 S.Ct. 1564, 1570, 75 L.Ed.2d 547 (1983), instructs lower courts to “consider the character and magnitude of the asserted injury to the rights protected” by the constitution, to “identify and evaluate the precise interests put forth by the State”, and then to decide whether the interests justify the restriction. The decision is functional, not turning on the form of words, and we have therefore held it important to follow what the Supreme Court does rather than attempt to articulate any formal methodology. See Hall v. Simcox, 766 F.2d 1171,1174 (7th Cir.), cert. denied, — U.S. —, 106 S.Ct. 528, 88 L.Ed.2d 459 (1985). See also Dart v. Brown, 717 F.2d 1491, 1498-1504 (5th Cir.1983), ce rt. denied, — U.S. —, 105 S.Ct. 105, 83 L.Ed.2d 49 (1984). Here, as in Hall, we start by identifying the important patterns in what the Court does.

1. The Court does not demand that the state prove at trial an empirical basis for its judgment that the regulation in question is useful. It will accept a logical justification. When scrutiny is elevated, as in Craig v. Boren, 429 U.S. 190, 199-204, 97 S.Ct. 451, 457-60, 50 L.Ed.2d 397 (1976), or strict, as in New York Times Co. v. United States, 403 U.S. 713, 91 S.Ct. 2140, 29 L.Ed.2d 822 (1971), the Court often demands that the state substantiate on the record the basis of its decision. When the basis is open to fair debate — will the publication of the Pentagon Papers lead to the death of soldiers?, will the addition of names to the ballot lead to confusion and the election of inferior public officials?— the demand for proof concludes the case in favor of the person relying on the constitutional right.

Yet all of the cases about the regulation of elections from Williams v. Rhodes, 393 U.S. 23, 89 S.Ct. 5, 21 L.Ed.2d 24 (1968), through Anderson ask only whether the state’s justification is logically sound and consonant with common experience. The Court accepts justifications such as “maintaining the integrity of the various routes to the ballot” (Storer, 415 U.S. at 733, 94 S.Ct. at 1280) or “regulating the number of candidates on the ballot to avoid undue voter confusion” (American Party, 415 U.S. at 782 n. 14, 94 S.Ct. at 1307 n. 14) as “compelling” (ibid.) without proof that a given number of candidates is “too confusing” or that a particular statute does (or does not) have an identified effect on the number of candidates. See Hall v. Sim-cox, supra.

The sorts of interests states assert to justify regulation of elections often cannot be reduced to proof. What would be the *1258 standard by which to tell how much confusion is too much? How. would we know whether a given change in the rules would help the voters select better candidates?— if we could even begin to know what a better candidate is? A requirement of proof, like a requirement that the state’s interests quantitatively outweigh associational interests (how much does an interest weigh?), would either condemn almost all regulation or give courts far too much discretion to substitute their opinions for those of the political branches. The Court has allowed states to justify regulation of elections with logical rather than empirical support in order to preserve the fundamentally political character of these choices.

The acceptability of logical rather than empirical justification does not imply that anything goes. The “justifications” offered for economic regulation often must be taken with a wink or not at all. A law may be designed to protect a politically powerful group, yet the Court will swallow a feeble explanation that it somehow serves more noble purposes. See Minnesota v. Clover Leaf Creamery Co., 449 U.S. 456, 101 S.Ct. 715, 66 L.Ed.2d 659 (1981); Kotch v. Board of River Port Pilot Commissioners, 330 U.S. 552, 67 S.Ct. 910, 91 L.Ed. 1093 (1947); cf. Vance v. Bradley, 440 U.S. 93, 110-12, 99 S.Ct. 939, 949-50, 59 L.Ed.2d 171 (1979). Compare Cass R. Sunstein, Naked Preferences and the Constitution, 84 Colum.L.Rev. 1689, 1713-17, 1727-30 (1984), with Frank Michelman, Politics and Values or What’s Really Wrong with Rationality Review, 13 Creighton L.Rev. 487 (1979). Political choices respond to influence, not solely to deductive logic. A court that demands consistent, principled lawmaking demands the impossible of a multi-member political body. An effort to deny to pressure groups the fruits of their influence may lead them to secure legislation that appears more neutral but that leaves others worse off still. Cursory scrutiny leaves the daily conflict of political forces to the political arena in which it belongs. But when the result of this process trenches on specific constitutional rights, such as the rights of speech and political association, the Court takes a closer look. The Court does not accept imagined justifications in voting cases. So, here, we do not wink. We require of Illinois a logical justification for § 10-4, the real justification and not a front for more sinister objectives, even though we do not insist on evidence establishing how the statute achieves its objectives. (We discuss the justification for § 10-4 in Parts II and IV. We are concerned here only with the kind of justification that will be necessary.)

2. Several cases say that the state should select the “least restrictive alternative” when regulating elections. Now “least restrictive alternative” could be a formula for consistent invalidation, because “[a] judge would be unimaginative indeed if he could not come up with something a little less ‘drastic’ or a little less ‘restrictive’ in almost any situation, and thereby enable himself to vote to strike legislation down.” Illinois State Board of Elections v. Socialist Workers Party, 440 U.S. 173, 188-89, 99 S.Ct. 983, 992, 59 L.Ed.2d 230 (1979) (Blackmun, J., concurring). It is always possible to do less and achieve less. If the Court were unconcerned about requiring the state to achieve less, it would strike down statute after statute. Yet it has not done this. The Court has not used a “least restrictive means” test to undercut the more tolerant approach we have outlined above. See Hall, 766 F.2d at 1174.

The point of looking at alternative methods of regulation is that the state must justify each incremental restriction in light of its incremental benefits. See Anderson, 460 U.S. at 789, 103 S.Ct. at 1570. The court must ask what function the rules serve and how this would be affected by an alternative. If 95% of the benefit of a rule is achieved by the first 5% of restriction, this implies that the remaining 95% of the restriction — the real bite of the rule— achieves almost nothing. It is a familiar problem in regulation. A reduction in the amount of benzene in the atmosphere from 10 parts per million (ppm) to 1 ppm may save 1,000 lives per year at the cost of $1 *1259 billion per year. Cf. Industrial Union Department v. American Petroleum Institute, 448 U.S. 607, 100 S.Ct. 2844, 65 L.Ed.2d 1010 (1980). Yet it may turn out that a reduction from 10 ppm to 5 ppm would save 950 lives at a cost of $100 million. The marginal gain from the “strict” 1 ppm rule is 50 lives at a cost of $900 million, which puts the justification for the strictness in a different light. In economic affairs a court does not require scrutiny of this sort in the absence of statutory command; when the “strictness” of a regulation comes at substantial cost in constitutionally protected entitlements, however, a comparison of marginal gains against marginal losses is necessary. See Stevenson v. State Board of Elections, 794 F.2d 1176, 1180-1181 (7th Cir.1986) (concurring opinion).

3. A court may study the workings of a statute either in the large (how it influences the conduct of people in general) or in the small (how it applies to the plaintiff). One common line of argument in cases under the first amendment is that the state must justify not only the usual application of the rule but also the nature of the effects on (and the possibility of an exception for) the plaintiff. Ballot-access cases like this one are first amendment cases (see Anderson, 460 U.S. at 786-88 & n. 7, 103 S.Ct. at 1568-69 & n. 7) so that in order to decide whether to examine the systematic effects of the statute or the effects on Moore in particular, we must consider the evolution of first amendment doctrine.

For many years the Supreme Court generally examined laws that affect speech and association (as ballot access laws do) in the small. The Court was suspicious of rules, and if the rule overshot or undershot in any respect — as all rules do — the Court was apt to find a defect. No more. The question today is whether the rule is adequately justified, not whether the plaintiff’s case is a paradigm for application of the rule or whether the law would fail if the plaintiff were treated differently. E.g., Regan v. Time, Inc., 468 U.S. 641, 104 S.Ct. 3262, 3272 n. 12, 82 L.Ed.2d 487 (1984); Clark v. Community for Creative Non-Violence, 468 U.S. 288, 104 S.Ct. 3065, 3070-71, 82 L.Ed.2d 221 (1984); Heffron v. International Society for Krishna Consciousness, Inc., 452 U.S. 640, 652, 101 S.Ct. 2559, 2566,69 L.Ed.2d 298 (1981). Cf. United States Postal Service v. Council of Greenburgh Civic Associations, 453 U.S. 114, 135, 101 S.Ct. 2676, 2688, 69 L.Ed.2d 517 (1981) (Brennan, J., concurring); Goldman v. Weinberger, — U.S. —, 106 S.Ct. 1310, 1316 & n. 6, 89 L.Ed.2d 478 (1986) (Stevens, J., concurring). In Anderson the Court analyzed the arguments for and against the rule (a requirement that independent candidates file for office long before parties had selected their candidates) by asking whether they made sense over the run of cases. The dissent asked whether the rule made sense when applied to Anderson himself. See 460 U.S. at 808-12, 822-23, 103 S.Ct. at 1587 (Rehnquist, J., dissenting). Justice Rehnquist took the majority to task for looking past Anderson’s case to the general operation of the statute. We must follow the majority’s path and inquire how § 10-4 usually operates, not how it affected Moore.

4. One variant of the argument that § 10-4 is both overbroad (as applied to Moore) and underinclusive (because it does not pursue to the hilt the interests the state asserts on its behalf) is a contention that the statute is not part of a comprehensive plan. Indeed it is not. It is an occasional piece of legislation. Moore invites us to judge it harshly as a result. But Clements, which applied the step-at-a-time doctrine to legislation affecting elections, is sufficient reply. The constitution does not require a state to adopt comprehensive plans or none at all. It is enough if the law the state adopts serves permissible purposes. See also Members of the City Council of the City of Los Angeles v. Taxpayers for Vincent, 466 U.S. 789, 104 S.Ct. 2118, 80 L.Ed.2d 772 (1984) (rejecting an argument that a city may regulate political handbills only as part of a comprehensive plan).

*1260 II

So we come at last to § 10-4, and we begin as Anderson directs with an assessment of the burden the law imposes on protected interests. The district court found the burden slight, and we concur. The statute applies only to circulators who switch parties during an election season. With the exception of its effect on the candidate himself — to which we return in Part IV — it does not diminish the pool of people available to circulate petitions, a pool the district court found to be ample. It does not increase the number of signatures anyone needs to get on the ballot or the number of circulator-hours necessary to appear once, twice, or more. The statute requires candidates such as Moore who switch parties during an election season to tap a larger number of circulators (one set of circulators per party), but it does not demand a greater total effort from circulators. The statute never bars a candidate from the ballot. Anyone may make one run for a single post; Moore himself secured a position on the ballot for a primary election without any hindrance from § 10-4. And although § 10-4 makes it harder to set up a new party after the primary season, because some circulators are locked-in to the old party, it does not impede the formation of a new party during or before the primaries. 1 The statute does not affect the ability of “used” circulators to speak. They may campaign for anyone to their hearts’ content; Moore’s friends could have done everything for him in his second race except collect signatures. Section 10-4 does undercut the value of the circulators’ speech to Moore and his friends, and what undercuts the value of speech implicates the first amendment. Buckley v. Valeo, 424 U.S. 1,15, 96 S.Ct. 612, 632, 46 L.Ed.2d 659 (1976). But this was true, as well, in cases such as Heffron and Community for Creative Non-Violence, in which the Court allowed regulation by neutral rules.

Now consider why the state might think § 10-4 useful. In evaluating the justifications we proceed, for the reasons developed in Part I, by testing the logical justification for the statute as applied to candidacies in general, not by inquiring whether the state proved empirically a need to regulate the Moore Party’s particular acts.

The statute is logically related to a number of permissible purposes, some of which the Supreme Court has called “compelling.” The rule promotes the cohesion of political parties. Circulators are the cadre of any political movement. A party needs a cadre to exist. Once a circulator acts on a party’s behalf to solicit signatures, he remains the party’s agent (if he is active at all) for the electoral season. This helps political parties to act as entities selecting and offering candidates to the voters, interests the Court has found of the highest order. See American Party and Storer. It induces candidates to cast their lot with (or against) a party for one political season at a time, also an interest the Court has held important. American Party, 415 U.S. at 781, 94 S.Ct. at 1306 (the state “may insist that intraparty competition be settled before the general election”); Storer, 415 U.S. at 733-36, 94 S.Ct. at 1280-82; cf. Rosario v. Rockefeller, 410 U.S. 752, 93 S.Ct. 1245, 36 L.Ed.2d 1 (1973) (state may require a voter to select a single political party and make the selection binding for one electoral season); Jenness v. Fortson, 403 U.S. 431, 442, 91 S.Ct. 1970, 1976, 29 L.Ed.2d 554 (1971).

Circulators engage in personal, often high-pressure, solicitation. There is always some potential for deceit; there is also a potential for confusion if a circulator identified as the agent of one party suddenly solicits signatures for another party or an *1261 independent candidate. The Court has recognized a difference between personal solicitation and speech that is more abstract. Compare Zauderer v. Office of Disciplinary Counsel, — U.S. —, 105 S.Ct. 2265, 85 L.Ed.2d 652 (1985), with Ohralik v. Ohio State Bar Association, 436 U.S. 447, 98 S.Ct. 1912, 56 L.Ed.2d 444 (1978). The constitution permits greater regulation of personal solicitation in light of its greater potential to confuse. This, too, supports § 10-4.

The state also asserts that § 10-4 helps prevent maneuvers that could affect the quality of the candidates who will be on the ballot. Suppose, for example, the Whig Party decided that the strongest candidate of the Free Silver Party would be Smith, then running against Jones in a primary. The Whigs would be delighted if Green could enter the primary, draw votes away from Smith, and allow Jones to prevail and represent the Free Silver Party in the general election. The Whig Party might be willing to dispatch its cadre to carry Green’s petitions, putting him on the primary ballot. But if this meant locking valuable supporters into the Free Silver Party for a year, the Whigs would be much less likely to try this maneuver. The district court found that § 10-4 discourages tactics such as this. Perhaps, as Moore insists, such tactics do not happen very often, but dirty tricks are not unknown in American politics, see Grimes v. Smith, 585 F.Supp. 1084 (N.D.Ind.1984), aff'd, 776 F.2d 1359 (7th Cir.1985), and there is a logical relation between § 10-4 and the prevention of one kind of dirty trick.

Moore’s argument about dirty tricks carries across the board. He maintains that § 10-4 is not very effective in achieving any of the asserted ends. The statute does not preserve the integrity of parties, he contends, because it still permits candidates to switch sides after the primaries; only a sore-loser statute — a rule that a candidate who loses in a primary election is disqualified from running for another party or as an independent candidate in the general election — would achieve the state’s objective. The statute does not prevent confusion because it does not regulate the sort of information canvassers provide or otherwise examine their sales pitches; it is not necessary to prevent confusion because sometimes the solicitation is focused so cleanly on the candidate personally (as Moore asserts it was in his case) that the possibility of confusion never arises. So, Moore concludes, the statute as it stands is both too weak and too strong.

This is a form of the “least restrictive alternative” argument, and we concluded in Part I that an election law need not be “least restrictive.” It is enough if the statute achieves permissible ends without pointless restrictions on speech and association. Section 10-4 is not pointless or plainly excessive in relation to its ends. Moreover, the form of “least restrictive alternative” argument Moore has advanced is misleading. Every time a state does less than the maximum it achieves less too, and so it opens itself to the sort of argument Moore makes — that the statute achieves too little to be valid. As we have rejected the argument that the state must do less (the traditional “least restrictive means” approach), so we reject the argument that the state does itself in by attempting to make do with less.

It may be easier to see that § 10-4 is a less restrictive alternative. It is less restrictive than two kinds of rules the Supreme Court upheld in Storer: disaffiliation statutes and sore loser statutes. A disaffiliation statute provides that a candidate must resign (disaffiliate) from one party a substantial time before running on the ticket of another or as an independent. The statute in Storer required disaffiliation a year in advance, and the Court found this a legitimate way to secure party unity and focus the party’s energies on intra-party competition designed to produce the best candidate. Cf. Rosario v. Rockefeller, supra (sustaining a statute requiring a voter to resign from a party 11 months in advance of voting in the primary of another party). The sore loser statute upheld in Storer barred the loser in a party’s primary election from running in the general *1262 election on any ticket or as an independent. The Court invoked the same set of interests; we suppose it also would have sustained a sore loser statute that treated someone who withdrew from a primary in the face of impending defeat (as Moore did) the same way it treats candidates who endure through election day and lose then.

Either a disaffiliation statute or a broad sore loser statute could have kept Moore off the ballot as a candidate of the Moore Party in 1982. Illinois does not have such a law. 2 Illinois has chosen instead to discourage rather than prohibit successive candidacies in one electoral season. This choice prevents Illinois from achieving its objectives as fully as it could; it also means that the state constrains the exercise of speech and associational rights less than it could. A state may choose the less restrictive — and less effective — path without transgressing constitutional limits.

III

Moore’s remaining arguments against the statute as a whole need not long detain us. He says that § 10-4 is too vague. But the language (“no person shall circulate or certify petitions for candidates of more than one political party ... to be voted upon at the next primary or general election”) is sufficiently precise in operation. The State Board of Elections issues a guide to candidates for each electoral season. The guide for 1982 contained a step-by-step outline for circulating petitions, and part Il.b of the guide states that a circulator “[m]ay not circulate for more than one party ... [and] may not circulate for an independent candidate or candidates in addition to a political party candidate in the same election”. Doubtless both statute and guide leave marginal questions unanswered (for example, does “same election” mean the very same post or does it include other posts in the same electoral season?), but not even the Code of Federal Regulations, for all its prolixity, answers all questions. Administrative agencies may overcome vagueness through the accumulation of precedents, just as courts do. See Civil Service Commission v. Letter Carriers, 413 U.S. 548, 93 S.Ct. 2880, 37 L.Ed.2d 796 (1973). Illinois is pursuing that permissible path.

Moore argues, and the district court found, that the Chicago Board of Elections once refused to interpret § 10-4 in the way it applied that statute to Moore, citing constitutional problems. On another occasion the Cook County Board held that § 10-4 does not prohibit a person who circulated for one party in a primary from changing sides in the general election. See Marcus v. Nimrod, No. 82 CO EB 3, quoted at 599 F.Supp. 666. The Chicago Board has changed its tune, and the brief filed in this court on behalf of both the State Board of Elections, which has general supervisory authority over local boards, and the Attorney General of Illinois, assures us that the uniform current practice is to interpret § 10-4 as prohibiting a switch within an electoral season. The district court found that this interpretation is correct. The constitution does not require unvarying application of any statute, especially so when the statute is applied by a multitude of local bodies under the guidance of a statewide agency. If Moore had made a showing of detrimental reliance on the cases he cites, we might have a difficult problem. But he did not; he was blissfully ignorant of § 10-4. The problems lurking within § 10-4 did not injure Moore, and in light of his experience in 1982 cannot injure him in the future. He knows, however painful the *1263 acquisition of knowledge may have been, that the State Board construes the statute broadly.

Moore’s contention that § 10-4 favors the major parties fares no better. He apparently believes that a circulator may carry petitions for the two major parties in the same season but may not do so for a major party and a minor party. The language of § 10-4 does not support this belief, and the candidates’ guide prepared by the State Board of Elections likewise treats all parties alike. A provision in § 10-2 states that “political party,” as used elsewhere in the election code, “shall mean any political group which shall hereafter undertake to form an established political party in the manner provided” by law — in other words, all political parties are treated alike under § 10-4.

IV

Although we have rejected Moore’s challenges to the structure of § 10-4 and its application to new or small parties, we must also consider whether the statute is constitutional as applied to candidates who circulate their own petitions. Moore contends that candidate-circulators are a special case because the first amendment rights of the speaker-circulator are greatest when the circulator is the candidate. The candidate may wish to convey the message “I’m a grass roots candidate” or “I’m willing to do my own work”, messages that are harder to send if § 10-4 forbids candidates from doing their own work. A candidate’s personal approach to the voters for support may be inseparable from the message of his campaign and essential to its success.

On the other side of the balance, the state’s interest in regulating the candidate’s circulations is significantly less. The candidate is identified with himself and his message, not with a party. As a result, candidates as circulators do not threaten confusion and are less likely to play dirty tricks. To pull a political stunt a party usually needs a cadre, not a single candidate. We think it an open question whether the justifications the state has offered for § 10-4 are even applicable to candidate-circulators, let alone sufficient to support an abridgment of the constitutional interests at stake. The question is difficult. Although the burden on candidates may be great, some of the greatest impediments — disaffiliation statutes, sore loser statutes, resign-to-run statutes — apply exclusively to candidates. Cases such as Buckley v. Valeo, supra, treat candidates and their organizations as one entity for most purposes; to restrain the organization is to restrain the candidate, and the reverse.

We do not say how a focused challenge to the application of § 10-4 to candidates as circulators should be resolved. Our difficulty is that the findings of the district court concern the application of § 10-4 to supporter-circulators exclusively. The district judge did not separately assess the state’s justifications for restraining the speech-related conduct of the candidate. The parties’ briefs have devoted only slightly more attention to the differences between candidates and other circulators. Moore has done enough to preserve the constitutional issue for decision, but not enough to make it appropriate for us to decide the question without the benefit of the district court’s views.

On remand the district judge may determine that candidates and non-candidate cir-culators have materially identical first amendment interests and that, despite the different roles they play in the electoral process, both are subject to the state’s proffered justifications. Or he may hold that Moore the candidate failed to carry his burden in this regard. On the other hand, he could decide that the justifications offered for § 10-4, as applied to candidate-circulators, are inadequate or nonsensical. The record might support any of these conclusions. The district court should make them first, and there will be an occasion for appellate review later.

To the extent the district court held § 10-4 constitutional as a general rule, the judgment is affirmed. To the extent the *1264 district court held § 10-4 constitutional as applied to candidates circulating petitions for their own candidacies, the judgment is vacated, and the case is remanded for further proceedings consistent with this opinion. Appellants will recover half of their costs.

Notes

2

. In 1983 Illinois passed a sore loser statute now codified in § 10-2. It provides: "A candidate for whom a nomination paper has been filed as a partisan candidate at a primary election, and who is defeated for his or her nomination at the primary election, is ineligible for nomination as a candidate of a new political party for election in that general election.” This statute apparently would not have interdicted Moore’s candidacy under the Moore Party’s banner in 1982 because Moore withdrew from the primary before he could be defeated. At all events, this statute does not assist Moore to obtain a declaratory judgment that § 10-4 is unconstitutional. Quite the contrary, by disqualifying candidates who could run under § 10-4, the new sore loser statute diminishes the adverse incremental effects of § 10-4.