concurring.
I agree with the majority’s disposition of this case. I write separately only because I believe that the bench, bar, and public would benefit from a more detailed explanation of the decision, particularly in light of the somewhat confusing state of current Article I, section 20, jurisprudence.
This case is the latest skirmish in a conflict that began more than 25 years ago over unequal funding of kindergarten to twelfth grade (K-12) school systems across the state. Earlier cases1 considered (and rejected) challenges to the funding of school districts. The challenge here is to the funding of Education Service Districts (ESDs), which, like school districts, are geographical entities that receive revenue for K-12 education but, unlike school districts, provide the funds for such supplemental services as audits, staff development, and special programs. ORS 334.125; ORS 334.175. Plaintiffs are students in schools served by the Region 10 ESD, encompassing Crook and Deschutes counties. They sought a declaratory judgment that the state, by providing fewer funds to the Region 10 ESD than to others, violated their right to equal treatment guaranteed by Article I, section 20, of the Oregon Constitution and the Equal Protection Clause of the Fourteenth Amendment. The state responded that the funding system reflects a permissible legislative policy choice to achieve funding equity gradually by phasing in decreases to “richer” ESDs and increases to “poorer” ones, instead of achieving equity immediately. After a hearing on cross-motions for summary judgment, the trial court granted defendant’s motion, denied plaintiffs’, and entered judgment for defendant.
The facts are all matters of public record and are not in dispute. The Region 10 ESD receives less money per student per year than every other ESD in the state. When this case was tried in late 2000, for example, plaintiffs’ ESD received $125 per student, the best-funded ESD received *6$1,991, and the average for all ESDs was $245. At that time, the operative ESD funding formula created less inequity than earlier formulas, but it did not contain any mechanism for achieving complete equity. However, after this case was tried, the 2001 Legislative Assembly created such a “transition” mechanism. Or Laws 2001, ch 695, §§ 2-8. That statute, if its terms are met, will bring ESD funding to a level of equity at the beginning of the 2005-06 school year that matches the level achieved by school districts and approved by this court in Withers v. State of Oregon, 133 Or App 377, 891 P2d 675, rev den, 321 Or 284 (1995) (Withers I) and Withers v. State of Oregon, 163 Or App 298, 987 P2d 1247 (1999), rev den, 331 Or 284 (2000) (Withers II). It does so by establishing roughly equal “revenue targets” for each ESD, assigning each ESD a “revenue gap” that is the difference between its “base” level (the current funding level as of 1999) and its target, and progressively reducing each ESD’s gap by increasing the funding for ESDs whose gaps result from underfunding and decreasing the funding for ESDs whose gaps result from overfunding. Or Laws 2001, ch 695, §§ 2-8. That system, according to plaintiffs, violates state and federal equality guarantees.
The Supreme Court has not explicitly and self-consciously construed Article I, section 20, under the principles set out in Priest v. Pearce, 314 Or 411, 415-16, 840 P2d 65 (1992): “There are three levels on which [the] constitutional provision [at issue] must be addressed: Its specific wording, the case law surrounding it, and the historical circumstances that led to its creation.” However, the existing case law surrounding section 20 has already considered that provision’s specific wording and the circumstances of its adoption. See, e.g., State v. Clark, 291 Or 231, 236-37, 630 P2d 810, cert den, 454 US 1084 (1981). The second item in the Priest list of levels therefore incorporates the other two.
Further, I presume that neither this court nor the Supreme Court would say that whatever Article I, section 20, “meant in 1857, it means precisely the same thing today” — as the Supreme Court said regarding Article I, section 17. Lakin v. Senco Products, Inc., 329 Or 62, 72, 987 P2d 463 (1999). That is because the framers of the Oregon Constitution, whatever else their virtues, had a conception of equality that *7contemporary legal (and moral) principles has emphatically repudiated. If this court or the Supreme Court were to interpret Article I, section 20, as the framers intended, the court would have to conclude that section 20 permits official invidious governmental discrimination based on race, ethnicity, and gender, which, in turn, would require overruling a significant number of cases and interpreting Oregon’s equality guarantee to provide many fewer protections than the minimum required by the Equal Protection Clause of the United States Constitution.
The framers revealed their understanding (or misunderstanding) of equality clearly and often. For example, they submitted to the people a referendum that would have allowed them to ratify the constitution, including section 20, at the same time that they decided to enter the union as a slave state. Or Const, Art XVIII, § 2. As it happened, the people rejected that option but chose instead — again, without any apparent qualms induced by the guarantee of equal privileges and immunities — to reject slavery but prohibit black people then living in the state from holding property or making contracts and to prohibit any “negro, or mulatto not residing in this State at the time of the adoption of this Constitution” from entering or residing here. Or Const, Art I, § 34 (banning slavery); Or Const, Art I, § 35 (repealed Nov 2, 1926) (imposing disability on current “negro” or “mulatto” residents; banning further emigration). The framers proclaimed that “[a]ll elections shall be free and equal” and, in the next sentence, restricted the franchise to white men. Or Const, Art II, §§ 1-2 (1859). They prohibited Chinese and other nonwhite “foreigners” from owning property. Or Const, Art I, § 31 (repealed May 26, 1970).
The record of their debates contains more evidence, if more evidence than the constitutional text itself were necessary, that the framers’ conception of equality, perhaps unremarkable at the time, is not ours. For example, they spent much of the afternoon of September 15, 1857, discussing whether “Chinamen” were better or worse than “negroes.” Charles H. Carey, ed., The Oregon Constitution and Proceedings and Debates of the Constitutional Convention of 1857 359-62 (1926). One delegate argued that “the *8negroes far surpassed, morally and physically, the China-men; if there were any class of thieves who understood their profession thoroughly it was the Chinamen.” Id. at 362. Another maintained that the two races were equally repugnant; he wanted to “consecrate Oregon to the use of the white man, and exclude the negro, Chinaman, and every race of that character.” Id. Another promised to “vote to exclude negroes, Chinamen, Kanakas, and even Indians. The association of those races with the white was the demoralization of the latter.” Id. The only delegate to speak in favor of Chinese did so on the ground that his constituents would “like to have a lot of them” because “[t]hey made good washers, good cooks, and good servants.” Id. Whatever “equal privileges and immunities” meant to people with such views, that meaning can have no relevance in contemporary discourse on the subject and no bearing on the interpretation of anything that could seriously be called an equality guarantee.
The correct analysis of Article I, section 20, then, is one derived from case law. As we have noted, that analysis is a “work in progress.” Tanner v. OHSU, 157 Or App 502, 520, 971 P2d 435 (1998). For example, it appears at first glance that there are two analyses, one developed by the Supreme Court and one by this court. Supreme Court cases appear to apply a simple, two-step inquiry to determine whether government action unlawfully discriminates against a person based on his or her membership in some class.2 The first inquiry is whether the class allegedly subject to the discriminatory treatment is a “true” class, variously defined as one whose members share “characteristics which they have apart from the law in question,” for example their race, ethnicity, religion, status as veterans, or geographical residence, Clark, 291 Or at 240, or as one whose members have in common *9some “antecedent personal or social characteristics or societal status,” Hale v. Port of Portland, 308 Or 508, 525, 783 P2d 506 (1989). If no “true” class is involved, the challenge under Article I, section 20, fails. The section simply has no relevance to such classification schemes. Sealey v. Hicks, 309 Or 387, 397, 788 P2d 435, cert den, 498 US 819 (1990). If a “true” class is involved, then the second and final inquiry is whether the classification scheme has a “rational basis.” See Jensen v. Whitlow, 334 Or 412, 424, 51 P3d 599 (2002); Crocker and Crocker, 332 Or 42, 55, 22 P3d 759 (2001); Seto v. Tri-County Metro. Transportation Dist., 311 Or 456, 467, 814 P2d 1060 (1991); Sealey, 309 Or 397-98.
The Court of Appeals, on the other hand, now uses an analysis that appears to include an additional step and a different level of deference to governmental policy choices.3 We recently held that government action violates section 20 if it (1) discriminates for or against a “true” class;(2) the classification is suspicious, that is, based on immutable traits or traits on the basis of which class members are subjected to adverse social or political stereotyping or prejudice;4 and (3) the discrimination survives a “demanding’' level of scrutiny. Tanner, 157 Or at 520-23. It would therefore appear that our treatment has deviated from Supreme Court precedent in that the Supreme Court does not distinguish between “suspect” and other true classes where we do and the Supreme Court applies “rationality review” to all “true” class distinctions where we require suspicious true classifications to embody more than mere rationality.
*10In fact, however, those discrepancies are illusory. The Supreme Court cases cited above that provide the occasion for applying Article I, section 20, adjudicate classification schemes that do not involve what could possibly be called suspicious classifications. Most do not even deal with true classes. They deal with alleged discrimination against “victims of governmentally inflicted torts” and “tort victims suffering damages in excess of $100,000,” Hale, 308 Or at 524; “persons injured by products more than eight years after their initial sale,” “potential tortfeasors who injure their victims with eight-year-old manufactured products,” and “[p]ersons injured by products,” Sealey, 309 Or at 397; residents of a transportation district, Seto, 311 Or at 466-67; divorced parents, Crocker, 332 Or at 55; and public employees, Jensen, 334 Or 423. Thus, when the Supreme Court in these cases does not discuss whether the classifications discriminate against a group that is defined by an immutable trait like race or a group that historically has been subjected to prejudice or stereotyping like, for example, Catholics, that omission does not imply that the distinction is never relevant or that it does not exist; it implies only that, in the particular case, it is not relevant (because the class is not even “true,” so proceeding to a next step in the analysis is unnecessary) or not worth discussing (because if the class is true, nobody could possibly believe that it is suspicious).
When the distinction is worth discussing, the court discusses it, or at least mentions it in passing. The distinction between suspicious and nonsuspicious classifications under section 20, in fact, originates in a Supreme Court case, Hewitt v. SAIF, 294 Or 33, 45-46, 653 P2d 970 (1982):
“Like other state and federal courts, we agree that a classification is ‘suspect’ when it focuses on ‘immutable’ personal characteristics. It can be suspected of reflecting ‘invidious’ social or political premises, that is to say, prejudice or stereotyped prejudgments. Historically, the most obvious such classification, and the one recognized to be such within the special concerns that gave rise to the fourteenth amendment, was, of course, racial discrimination. * * * [C]lassifications [based on gender, like those based on race, alienage and nationality are] inherently suspect. The suspicion may be overcome if the reason for the classification reflects specific biological differences between men and *11women. It is not overcome when other personal characteristics or social roles are assigned to men or women because of their gender and for no other reason. That is exactly the kind of stereotyping which renders the classification suspect in the first place.”
(Footnotes omitted.) Further, even as it appears to use a two-step process, the court has frequently announced its continued allegiance to Hewitt and the concept of suspicious classifications. In Crocker, for example, the court held:
“When distinctions are based on personal characteristics that are not immutable, this court reviews the classification for whether the legislature had a rational basis for making the distinction. See Seto * * * (geographic classification reviewed for rational basis); Hewitt * ** (classifications based on immutable traits are suspect). Marital status is not an immutable trait. Thus, to withstand constitutional scrutiny under Article I, section 20, it is necessary only that the statute in question be based on rational criteria.”
332 Or at 55. Using almost identical language, the court in Jensen held:
“ORS 30.265 (1)[, the Oregon Tort Claims Act,] distinguishes on the basis of public employment, not any immutable characteristic such as race, religion, or alienage. See Hewitt * * * (classifications based on immutable characteristics are suspect). Because ORS 30.265(1) does not distinguish based on any immutable characteristic, it satisfies Article I, section 20, if the legislature had a rational basis for distinguishing between the classes involved.”
334 Or at 423-24.
Those quotations compel two conclusions. First, the Supreme Court distinguishes not only between true and non-true classes but also, within the universe of true classes, between suspicious and nonsuspicious classifications. Thus, the Crocker court notes that some true classes, like those based on geography, are tested for “rational basis,” while others, i.e., those based on immutable traits, are not. Second, the court indicates that true classes that are also suspicious receive a form of review that is less deferential than “rational basis” review. The negative implication of the last sentence *12from Jensen quoted above is that, if ORS 30.265 had distinguished based on an immutable characteristic, it would not satisfy Article I, section 20, merely because it had a “rational basis.” I therefore conclude that the analysis we applied in Tanner is simply a more elaborated version of the analysis implicit in Supreme Court decisions.
Under that analysis, to repeat, we examine the following four questions: (1) Does the government action provide a privilege or immunity to a person based on that person’s membership in a class? (2) Is the class a true class? (3) Is it also suspicious? (4) Is the suspicion confirmed; that is, is the governmental action based on stereotype or prejudice instead of a genuine correlation between the trait on which the classification is based and the benefit or burden conferred? If the answers to all four questions are affirmative, the government action violates section 20. Otherwise, the action does not violate section 20 if it rationally furthers some lawful governmental objective — even one that the lawmakers did not intend.5
As the present case demonstrates, however, knowing what questions to ask is one thing; what those questions mean, and what the answers are, present different and independent problems. Here, for example, defendant maintains that plaintiffs have not established that students in other districts are receiving the privilege of a better education; at most, the record establishes that their district’s system is receiving more money, and, even if we could presume that money translates directly into educational quality, the total amount of funding provided to plaintiffs’ school system from all sources differs from the total amount provided to other districts by only three percent. That argument focuses on the definition of “privilege,” a threshold issue.
*13Defendant also argues (with a good deal of cogency) that plaintiffs are not members of a true class. Defendant concedes that, under several earlier cases (including Clark, Seto, and Withers II), classes based on geographical residence are deemed to be true but maintains that statements to that effect in those cases are too broad: although residents of, say, a particular city, county, or defined neighborhood — that is, people who identify themselves as residents of a well-defined area with a preexisting self-consciousness as an entity, such as Pendleton, Lane County, or South Hills — are a true class, residents of other geographical areas (for example, the area falling within 15 miles of the intersection of Interstate 5 and Priceboro Road) are not. See Withers II, 163 Or App at 310-11 (Haselton, J., concurring) (questioning whether school district residents are a true class). If, as the Supreme Court has stated, a class created by statute is not “true,” while one based on “antecedent personal or social characteristics or societal status,” Hale, 308 Or at 525, is true, then defendant is correct: before the enactment of ORS chapter 334, creating Education Service Districts, and in particular of ORS 334.020(1)(j) establishing the Region 10 ESD, plaintiffs could not have self-consciously regarded themselves as “Region 10 residents” because Region 10 did not exist. Nor would they have regarded themselves as “residents of Crook or Deschutes Counties,” because, except for purposes of defining an ESD, residents of those two counties are not generally and widely regarded as forming a cohesive, self-conscious group.
Finally, defendant argues that, even if residents of plaintiffs’ ESD are a true class, the classification scheme is not “suspicious” because it is not based on immutable traits— plaintiffs could move to a wealthier district if they wanted. That argument reminds us that some cases such as Crocker and Jensen suggest that a suspicious classification is one based on immutable traits, while other cases such as Tanner hold that other criteria matter:
“Although the court in Hewitt referred to ‘immutable’ characteristics as being sufficient for defining a suspect class under Article I, section 20, subsequent cases make clear that immutability — in the sense of inability to alter or change — is not necessary. The court has since explained that, in addition to gender, such classes as alienage and *14religious affiliation also are suspect classes. Both alienage and religious affiliation may be changed almost at will. For that matter, given modem medical technology, so also may gender. We therefore understand from the cases that the focus of suspect class definition is not necessarily the immutability of the common, class-defining characteristics, but instead the fact that such characteristics are historically regarded as defining distinct, socially recognized groups that have been the subject of adverse social or political stereotyping or prejudice.”
157 Or App at 522-23 (footnote and citations omitted).
In fact, the present case does not require resolution of these difficult issues. The parties agree that defendant prevails if the current system of funding ESDs is rational, and they are correct. Whether or not plaintiffs are members of a true class, the class is not suspicious — regardless of whether suspicion is a function of immutability or of historical victimization. Residence in the Region 10 ESD is not an immutable trait such as race or ethnicity, nor have Region 10 residents been subjected to adverse social or political stereotyping or prejudice in the same way that religious, gender, or other historically oppressed groups have. Because the classification does not raise the suspicion that illicit legislative motives are afoot, rationality review applies.
No Oregon cases offer a definition of that standard. Because it derives from federal law, Hale, 308 Or at 524, definitions from United States Supreme Court cases are instructive. Over the years, that Court has used a variety of verbal formulations to describe the test. What they reduce to is nearly total deference to legislative classifications. Since the New Deal, the number of times the Court has struck down statutes because they fail rationality review can probably be counted on the fingers of one hand, and in most of those cases the classification, while not one of the very few labeled “suspicious,” discriminated against some easily identified, historically disadvantaged group. See, e.g., City of Cleburne v. Cleburne Living Center, 473 US 432, 105 S Ct 3249, 87 L Ed 2d 313 (1985) (striking down discrimination against mentally disabled using rational basis test); Plyler v. Doe, 457 US 202, *15102 S Ct 2382, 72 L Ed 2d 786 (1982) (striking down discrimination against children of undocumented aliens using rational basis test).
The language of deference takes many forms. One of the most frequently cited formulations announces that a law will pass rationality review if “it is evident from all the considerations presented to [the legislature], and those of which we may take judicial notice, that the question is at least debatable.” United States v. Carolene Products Co., 304 US 144, 154, 58 S Ct 778, 82 L Ed 1234 (1938); see also Minnesota v. Clover Leaf Creamery Co., 449 US 456, 101 S Ct 715, 66 L Ed 2d 659 (1981). The “debatable” proposition need not have been, in fact, debated in the legislature. As the Court has specifically held on several occasions, “[i]t is, of course, ‘constitutionally irrelevant whether this reasoning in fact underlay the legislative decision,’ * * * because this Court has never insisted that a legislative body articulate its reasons for enacting a statute.” U.S. Railroad Retirement Board v. Fritz, 449 US 166, 179, 101 S Ct 453, 77 L Ed 2d 368 (1980) (quoting Flemming v. Nestor, 363 US 603, 612, 80 S Ct 1367, 4 L Ed 2d 1435 (1960)).
Further, a statute that only partially attacks an identifiable evil (such as underfunding of ESDs) is not an irrational method of accomplishing a legislative objective. In Railway Express Agency v. New York, 336 US 106, 69 S Ct 463, 93 L Ed 533 (1949), the Court sustained an ordinance prohibiting advertising on some trucks, but not others, on the theory that visual distraction could lead to traffic accidents. To the argument that this law irrationally singled out a particular group even though distraction was equally caused by other similar advertisements, the Court responded: “It is no requirement of equal protection that all evils of the same genus be eradicated or none at all.” Id. at 110. On the same theme, the Court sustained an Oklahoma statute that created an allegedly irrational distinction between opticians who dispensed ready-to-wear glasses and other opticians. The Court held:
“Evils in the same field may be of different dimensions and proportions, requiring different remedies. Or so the legislature may think. Or the reform may take one step at a *16time, addressing itself to the phase of the problem which seems most acute to the legislative mind. The legislature may select one phase of one field and apply a remedy there, neglecting others.”
Williamson v. Lee Optical, 348 US 483, 489, 75 S Ct 461, 99 L Ed 563 (1955) (citations omitted); accord F.C.C. v. Beach Communications, Inc., 508 US 307, 316, 113 S Ct 2096, 124 L Ed 2d 211 (1993).
Using this deferential standard of review, I (like the majority) find no distinction between the transitional ESD funding mechanism at issue here and the school district transitional funding mechanisms at issue in Withers I and Withers II. The formulas are identical. Both serve to bring about funding equity in a short period of time. In Withers I, we held that “it is rational for the legislature to have concluded that immediate and drastic cuts in school district funding would have a direct and deleterious effect on the school children in those districts.” 133 Or App at 387. We reiterated that conclusion in Withers II. 163 Or App at 309. According to a “Research Report” prepared by the Legislative Revenue Office in October 2001, “The intent [of the phase-in formula] is to allow time for planning additional services if revenue increases and to reduce disruption to ESD funded services if revenue decreases.” Legislative Revenue Office, Research Report #3-01 10 (Oct 2001). That statement establishes the requisite rationality. The legislature may move gradually toward equitable funding of ESDs. Defendant’s ESD funding system does not violate plaintiffs’ rights under Article I, section 20.
Coalition for Equit. School Fund. v. State of Oregon, 311 Or 300, 811 P2d 116 (1991); Olsen v. State ex rel Johnson, 276 Or 9, 554 P2d 139 (1976); Withers v. State of Oregon, 163 Or App 298, 987 P2d 1247 (1999), rev den, 331 Or 284 (2000) (Withers II); and Withers v. State of Oregon, 133 Or App 377, 891 P2d 675 (1995), rev den, 321 Or 284 (Withers I).
Article I, section 20, also prohibits discrimination against an individual as an individual, unrelated to any class to which he or she may belong. Clark, 291 Or at 239. The state violates the prohibition against individual-based discrimination when it distributes benefits or burdens in a haphazard, standardless manner. City of Salem v. Bruner, 299 Or 262, 702 P2d 70 (1985); State v. Freeland, 295 Or 367, 667 P2d 509 (1983). That form of discrimination is not implicated in this case. Here, plaintiffs’ allegation is not that, as individual students, they receive unequal funding arbitrarily or according to no system, but that they receive unequal funding by virtue of their membership in a class: those who attend a school within the Region 10 ESD. They argue, in other words, that the state has a system, but that it is an unjustifiable one.
Earlier cases from this court applied an analysis like the Supreme Court’s current one. For example, in Jungen v. State of Oregon, 94 Or App 101, 105, 764 P2d 938, rev den, 307 Or 658, cert den, 493 US 933 (1989), we concluded that a plaintiff challenging government action had to show:
“(1) that another group has been granted a ‘privilege’ or ‘immunity’ which their group has not been granted, (2) that [the challenged law] discriminates against a ‘true class’ on the basis of characteristics which they have apart from that [law], and (3) that the distinction between the classes * * * has no rational foundation in light of the state’s purpose.”
I avoid the expression “suspect class” not only because it unnecessarily invokes federal equality jurisprudence but also because it conveys the impression that something about the class members themselves requires us to suspect them. The phrase “suspicious classification,” on the other hand, says what it means: that some systems of classifying people should cause us to suspect the classifiers. See 191 Or App at 13-14.
This default “rational relationship” test for ordinary discrimination crept into Article I, section 20, case law in Seto, 311 Or 456, despite the fact that two years earlier, in Hale, the court said that the “rational basis” test came from federal law and “has been superseded by our more recent [Article I, section 20] decisions.” 308 Or at 524. Seto cites Sealey and Clark, but neither of those cases adopts anything like a “rational basis” test. Seto, 311 at 467. In fact, nothing in Oregon case law before Seto required ordinary classifications to be any more “rational” than any other kind of statute, which is to say, not rational at all. Regardless, for better or for worse, the test now appears to be a well-settled part of Oregon equality law.