concurring.
I concur in the result reached by Justices Dimond and Matthews regarding the unconstitutionality of the tax exemption statute.1
A. Nature and Extent of Infringement of Rights
The most significant dispute between the parties here is over the nature and extent of the right infringed.2 The Zobels contend that the three-year prior filing requirement is essentially a “durational residency” requirement, thus triggering strict scrutiny and requiring a compelling state justification. The state insists that the requirement does not discourage exercise of the right to migrate.
In my view, the appropriate focus is whether, and to what extent, the operation of the statute will have the effect of penalizing United States citizens for exercising their constitutional right to migrate between states.3 Although precise figures are lacking, the state has conceded that the exemption will have a differential impact *432and I reach the same conclusion.4 Proportionately, a new resident is much more likely than an old resident to be subjected to an income tax under this scheme.
The parties have devoted a significant portion of their arguments to the subject of the “intent” of the legislature in passing this enactment-/, e., whether or not the statute was “intended” to discriminate against new residents. The notion that a plaintiff must show that a statute was “intended” to discriminate against a particular group is based on a line of United States Supreme Court cases dealing with racial discrimination: Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976), and Village of Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977).5
Based on the considerations set forth in footnote five, I would not require such a showing when the effect of the legislation is shown-or is stipulated-to impact unequally on groups according to whether or not they have exercised their constitutional right of interstate migration. I would not make a showing of discriminatory intent a necessary condition of finding that a classification penalizes an individual's exercise of the constitutional right of interstate migration.
Both the nature and the extent of the infringement must be examined. Here, this *433translates into assessing exactly how the unequal impact “penalizes” the exercise of the right of interstate migration. I think that this tax can be properly characterized as an income tax levied only on an individual’s first three years of contact with Alaska; or, as its differential impact is uncontested, a tax levied primarily on new residents. As such, I think that the scheme tends to penalize the exercise of an individual’s right to migrate into Alaska in two ways. Its most obvious effect is the financial disincentive to enter: a new resident is subjected to three years of a tax burden which is considerably heavier, on an individual basis, than it would be were the same aggregate tax amount to be levied on new and old residents alike.6 Secondly, it has a symbolic value which cannot be ignored. Whether or not the people of Alaska mean to do so, it projects to other United States citizens a marked hostility to new entrants by imposing a tax on them which is not imposed on long-term residents.7
Thus, I conclude that the differential impact on old and new residents effected by the exemption scheme clearly does penalize the exercise of an individual’s right of interstate migration, in the manner noted above. As such, the state’s purposes must be weightier, and the “fair and substantial relationship” between means and ends more narrowly drawn than if the statute did not penalize the exercise of the right of interstate migration (or if it penalized this to a lesser degree); but these concerns need not reach the level of being necessary to promote a compelling state interest, as we are not applying the “strict scrutiny” test.
B. State Purposes and the Relationship Between These Purposes and the Classification
Having assessed the nature and extent of the right infringed upon, the next step under Erickson is to analyze the purposes of the statute and assess the extent to which the classification chosen has a “fair and substantial” relationship to these purposes.
The state advances six purposes for the tax exemption provisions:
(1) the plan should not relieve individuals entirely of their obligation to help defray the costs of government;
(2) the plan should not result in the total dismantling of the state’s income tax audit and collection staff;
(3) the plan should not redistribute the income tax burden so that it falls only on the very highest income earners, but should retain the existing progressive structure;
(4) the plan should not saddle individuals with an empty and burdensome annual filing requirement;
(5) the plan should not result in a revenue collection bureaucracy out of proportion in size and cost to the amount of revenue collected; and
(6) the plan should not result in a “windfall” to persons who receive the protections, services and benefits provided by the state that are closely related to the earning of income, but who have never before contributed to the costs of providing those programs.8
It seems clear that purposes (2) through (5), although obviously legitimate state interests, are not related to the specific provision here at issue, the prior filing require*434ment. The Zobels do not contest the power of the state to reduce taxes, to maintain some tax system infrastructure, to maintain a progressive tax system, to reduce or eliminate the filing requirement, or to cut back on a disproportionately large tax bureaucracy. None of these, however, are “fairly and substantially related” to the line drawn by the legislature between those who have filed three, two, one, or no tax returns in the state. These purposes may explain why the legislature chose to reduce, without eliminating, the income tax requirement, but they shed no light on why the legislature chose to gear this reduction to a prior filing requirement, which is the question that must be answered.
A stronger case can be made for finding a relationship between the prior filing requirement and purposes (1) and (6). As I interpret the argument, the state contends that every resident of Alaska should, at some point in his or her income earning career, be subject to an income tax. Long-term residents have already been so subjected, as they have been liable for past taxes; and, to achieve some basic equity between that group and those first, second, and third-time filers who have not yet borne their share of the burden, this classification is required.
The Zobels assert that this purpose-the notion of introducing equity between old and new residents based on past tax contributions-is impermissible under the rulings of the United States Supreme Court. In Shapiro v. Thompson, 394 U.S. 618, 89 S.Ct. 1322, 22 L.Ed.2d 600 (1969), the Court rejected a “past tax contributions” argument offered to justify a durational residency requirement for welfare benefits:
[The state’s] reasoning would logically permit the State to bar new residents from schools, parks, and libraries or deprive them of police and fire protection. Indeed it would permit the State to apportion all benefits and services according to the past tax contributions of its citizens. The Equal Protection Clause prohibits such an apportionment of state services.
394 U.S. at 632-33, 89 S.Ct. at 1330, 22 L.Ed.2d at 614 (footnote omitted). These same concerns were repeated in dicta in a footnote in Vlandis v. Kline, 412 U.S. 441, 450 n. 6, 93 S.Ct. 2230, 2235 n. 6, 37 L.Ed.2d 63, 70 n. 6 (1973).
The state argues that the Court did not say that past tax contributions could never legitimately form the basis for a statutory distinction; and that the equal protection issue raised by making distinctions based on past tax contributions does not turn on the legitimacy of the purpose, but rather on whether the distinction can be rationally related to the purposes of the statute. In Shapiro, such a distinction could not be rationally related to the purposes of the welfare statute, but it can be so related here.
I reject the Zobels’ argument that distinctions based upon past tax contributions are absolutely impermissible. The Supreme Court’s affirmance of Starns v. Malkerson, 326 F.Supp. 234 (D.Minn.1970), aff’d, 401 U.S. 985, 91 S.Ct. 1231, 28 L.Ed.2d 527 (1971), stands for the proposition that the Supreme Court has approved such a distinction between residents and nonresidents, based on former tax contributions, in the context of differential tuition charges as to resident and nonresident students. I think that, in some situations, a classification based on past tax contributions would be unquestionably permissible; e. g., a tax rebate made retroactive for a reasonable period could not be struck down on the equal protection ground that it distinguished past taxpayers from past nontaxpayers based solely on past tax contributions (although I express no opinion as to the other grounds on which such a statute might be attacked).
On the other hand, it cannot be ignored that this “purpose” raises potential constitutional questions in- many contexts, and is clearly impermissible in some (e. g., welfare programs). I would resolve these conflicting considerations by ruling that distinctions based on past tax. contributions are permissible, at least in the limited context of statutes designed to grant tax relief; but considering the possible constitutional in*435firmities with this purpose noted by the Supreme Court, this “purpose” must be placed among the weaker of the vast spectrum of state purposes, and normally it will not be given a great deal of weight when the Erickson balancing test is struck. Of course,' this “weight” is only significant in relation to the other elements of the Erickson test. If the statute is tightly drawn so that there are no substantial gaps in the means/ends relationship, and so that its impact, in terms of “suspect classifications” and/or constitutional rights infringed, is minimal, then this purpose may prevail.
Having found the “purpose” legitimate in this limited context, I turn to an examination of the relationship between the means and the end here. The Zobels point out, and the superior court found significance in the fact, that the only requirement is that of having filed a return, not of having incurred tax liability. Thus, the relationship is not based on past tax contributions directly, but rather on past tax return filing. In effect, it is somewhat over-inclusive. The state is correct in pointing out that over-inclusiveness does not necessarily render a statute unconstitutional. See Commercial Fisheries Entry Commission v. Apokedak, 606 P.2d 1255, 1267-68 (Alaska 1980). That is true; the over-inclusiveness is not dispositive. But it also is not irrelevant in assessing the extent to which the relationship between the classification and its purpose is a fair and substantial one.
C. Striking the Balance
The final step in the Erickson approach, striking the balance, is not an easy task here. However, I find that the above analysis of the various elements to be weighed tips the scale in favor of the Zobels. This conclusion is reinforced by my reading of United States Supreme Court case law which has dealt, under privileges and immunities analysis, with other states’ enactments of similar statutory exemption schemes.
These are the factors which I find weigh in favor of the Zobels: The statute does, as the state has admitted, have the effect of placing a disproportionate share of the tax burden on those who have recently exercised their right to migrate, and this, although not triggering the “strict scrutiny” approach, does constitute a substantial infringement of a well-estabalished constitutional right. I think it can fairly be characterized as penalizing that right. The only proffered purpose which I find to be at all related to the distinction drawn here is that of establishing some rough notion of equity between former and future Alaskan taxpayers. This purpose is permissible in this limited context (/. e., tax relief legislation), but it is not among the weightier of permissible state concerns. And considering that the relationship between means (the classification) and ends (establishing equity) is rendered somewhat less fair and substantial because the classification is not based on past tax liability, but rather on past filing of returns, I find that the balance comes out in favor of the Zobels.
In so ruling, I think it appropriate to derive guidance from the reasoning of the United States Supreme Court cases dealing with analogous problems in the privileges and immunities context. Generally, privileges and immunities analysis is applied to distinctions between residents and nonresidents, whereas distinctions between long-term and short-term residents are analyzed under the equal protection or due process clauses.9 However, this court has recognized that privileges and immunities case law and reasoning may be helpful in assessing parallel equal protection claims. Lynden Transport, Inc. v. State, 532 P.2d 700, 706 (Alaska 1975). I find merit in the Zo-*436bels’ contention that a state probably has more authority generally to draw distinctions between residents and nonresidents than between two sets of bona fide residents based on length of residency.10 Thus, cases striking down a distinction between residents and nonresidents are fairly persuasive authority for the proposition that that distinction cannot be drawn between long-term and short-term residents.
Two United States Supreme Court cases stand for the proposition that a state cannot place greater tax burdens on non-residents than on residents based solely on residency. In Travis v. Yale & Towne Manufacturing Co., 252 U.S. 60, 40 S.Ct. 228, 64 L.Ed. 460 (1920), the Court found invalid a New York taxing system which, although it taxed residents and nonresidents at the same rate, granted New York residents some exemptions which were not available to nonresidents. Nonresidents were given a credit for taxes paid to their state of residence, and New York apparently thought that the state of residence would grant substantially equivalent personal exemptions, thus equalizing the tax burden between residents and nonresidents. However, some states, notably Connecticut and New Jersey, did not not have an income tax, and so of course residents of those states received neither the tax credits nor the exemptions. In effect, residents were granted exemptions which nonresidents were not. The Supreme Court had little difficulty striking this tax mechanism down:
Whether [non-New York resident taxpayers] must pay a tax upon the first $1,000 or $2,000 of income, while their associates and competitors who reside in New York do not, makes a substantial difference. Under the circumstances as disclosed, we are unable to find adequate ground for the discrimination, and are constrained to hold that it is an unwarranted denial to the citizens of New York. This is not a case of occasional or accidental inequality due to circumstances personal to the taxpayer, but a general rule, operating to the disadvantage of all nonresidents, including those who are citizens of the neighboring states, and favoring all residents, including those who are citizens of the taxing state.
252 U.S. at 80-81, 40 S.Ct. at 232, 64 L.Ed. at 470 (citations omitted).
In a more recent case, the Court struck down the New Hampshire Commuters Income Tax in Austin v. New Hampshire, 420 U.S. 656, 95 S.Ct. 1191, 43 L.Ed.2d 530 (1975). The statute in that case imposed a four per cent tax rate both on income which non-New Hampshire residents earned within the state and on income which New Hampshire residents earned outside the state. (Income which New Hampshire residents earned inside the state was not taxed.) Again, however, it was the exemption provisions which altered the impact of the tax. The income which New Hampshire residents earned outside the state was exempted in three situations: (1) if such income was taxed by the state from which it was derived; (2)'if it was exempted from taxation by the state from which it was derived; or (3) if the state from which it was derived did not tax such income. In short, no resident of New Hampshire was taxed on his out-of-state income. Nonresidents were still taxed at four per cent or the rate at which the nonresident’s home state would tax such income had it been earned in the home state, whichever was less. Again, the Court had little difficulty striking this scheme down, noting, “The overwhelming fact, as the State concedes, is that the tax falls exclusively on the income of nonresidents; . . . . ” 420 U.S. at 665, 95 S.Ct. at 1197, 43 L.Ed.2d at 537-38.
*437I find these cases to be persuasive in reaching the conclusion that the exemption scheme in the case at bar cannot stand. If New York and New Hampshire cannot manipulate their tax exemption provisions to impact disproportionately on nonresidents, Alaska should not be able to follow this same course concerning newly arrived residents.
Finally, I agree with Justice Dimond on the severability issue presented by the cross-appeal-i. e., that the invalid portions of the statute (§§ 1 and 4-9) are severable, and §§ 2, 3, 10 and 11 may stand.
Based on these considerations, I agree that the judgment of the superior court as to the tax exemption statute should be affirmed.
.I agree with Justices Dimond and Matthews that the tax exemption statute would fall into the “rational basis” tier under the federal two-tier equal protection test. I also agree that our own state equal protection test as applied in the right-to-interstate-migration context in several prior cases [Hicklin v. Orbeck, 565 P.2d 159 (Alaska 1977), rev’d on other grounds, 437 U.S. 518, 98 S.Ct. 2482, 57 L.Ed.2d 397 (1978); Gilbert v. State, 526 P.2d 1131 (Alaska 1974); State v. Adams, 522 P.2d 1125 (Alaska 1974); State v. Wylie, 516 P.2d 142 (Alaska 1973); State v. Van Dort, 502 P.2d 453 (Alaska 1972)] has been modified by State v. Erickson, 574 P.2d 1 (Alaska 1978), which announced the test which should be applied in equal protection cases.
The test we announced in State v. Erickson, 574 P.2d at 12, is essentially one of balancing. On the one hand, the court must assess (1) the legitimacy of the state purpose purportedly furthered by the provision, and (2) the extent to which the relationship between the end (the asserted purpose) and the means (the classification chosen) is fair and substantial. On the other hand, the court is to determine the nature and the extent of the infringement of individual rights allegedly caused by the classification. Id.
. There is no need to decide the question of whether the right to interstate migration is protected under the state constitution as well as the federal constitution. See Alaska Const, art. I, § 21 (“The enumeration of rights in this constitution shall not impair or deny others retained by the people.”). I note that the federal right itself has no explicit source. Shapiro v. Thompson, 394 U.S. 618, 630, 89 S.Ct. 1322, 1329, 22 L.Ed.2d 600, 612 (1969) (“We have no occasion to ascribe the source of this right to travel interstate to a particular constitutional provision.”) (footnote omitted). Even assuming that the right to interstate migration is purely federal, I see no conceptual barrier to invoking the state equal protection clause to protect it.
. I note that there is no necessity to find that the system actually deters travel; rather, the focus is on whether or not the classification system operates to penalize those persons who have exercised their constitutional right of interstate migration. Memorial Hosp. v. Maricopa County, 415 U.S. 250, 257-58, 94 S.Ct. 1076, 1081-1082, 39 L.Ed.2d 306, 314-15 (1974).
. The state conceded below that more new residents would be paying the tax than old residents. Clearly, the appropriate comparison is not that of absolute numbers, but of ratios; but we can take judicial notice of the fact that the majority of the population in Alaska is constituted of “old residents,” and thus, proportionately, a new resident is much more likely to be paying the income tax than an old resident.
. This authority is not controlling for three reasons: first, it is based on federal equal protection law rather than state equal protection law. This court could choose to reject the Washington v. Davis approach, for state equal protection claims, even in cases of racially disproportionate impact.
In other contexts, we have held that a prima facie case of discrimination may be established by a showing of differential impact, as long as the individual can show that he or she is a victim of this differential impact. See Brown v. Wood, 575 P.2d 760, 768 (Alaska 1978), modified, 592 P.2d 1250 (Alaska 1979) (plaintiff showing that she is paid less than male colleagues for comparable work established a pri-ma facie case of discrimination and thus shifted the burden to the employer to show that the difference in pay is based on other considerations); compare Johnson v. State, 607 P.2d 944, 947-48 (Alaska 1980) (adopting same approach in the area of alleged racial discrimination in sentencing, requiring a showing that the individual’s sentence was probably higher than that which would have been imposed upon a defendant of a different race with a like criminal history who committed a similar offense). Here it is uncontested that the Zobels are prejudiced by the differential impact in this case.
Second, the authority deals with equal protection cases in the context of race, not equal protection in the context of classifications based on exercise of constitutional rights. I am not aware of any case which has required such a showing of “legislative intent” when the strict scrutiny standard is triggered, not by a racially disproportionate impact, but rather by a disproportionate impact on those who have or have not exercised certain constitutional rights; and there is some authority to the effect that Washington v. Davis and Arlington Heights are arguably “limited to equal protection claims involving alleged racial discrimination.” Socialist Workers Party v. Chicago Bd. of Election, 433 F.Supp. 11, 14 n. 8 (N.D.Ill.), aff’d and modified on other grounds, 566 F.2d 586 (7th Cir. 1977) (per curiam) (adopting the district court opinion), aff’d, 440 U.S. 173, 99 S.Ct. 983, 59 L.Ed.2d 230 (1979). Outside the equal protection context, there is authority to the effect that when an individual alleges that a particular administrative act (e. g., a discharge, a criminal prosecution, etc.) is in retaliation for that individual’s having exercised some constitutional right, he or she must show that his or her conduct was constitutionally protected, and that this conduct was a “substantial factor” or a “motivating factor” in the decision as to the administrative act; and Washington v. Davis and Arlington Heights have been cited as authority for this. See, e. g., Mt. Healthy City School Dist. Bd. of Educ. v. Doyle, 429 U.S. 274, 287, 97 S.Ct. 568, 576, 50 L.Ed.2d 471, 484 (1977). I do not find this apposite.
Third, even if this court followed Washington v. Davis, and chose to extend it to this area, the showing of intent is required to trigger the “strict scrutiny" analysis under federal law. Since I would not employ the “strict scrutiny” analysis under either federal or state law, I would not require a showing of intent as a necessary prerequisite to the Erickson approach.
.It might be argued that there is no disincentive to entry because other states impose income taxes, and thus a prospective new resident is not any worse off in Alaska than he or she was in the state of origin. However, a showing of deterrence is unnecessary to a finding of infringement of the right of interstate migration. Memorial Hosp. v. Maricopa County, 415 U.S. 250, 257-58, 94 S.Ct. 1076, 1081-1082, 39 L.Ed.2d 306, 314-15 (1974). Additionally, not all states impose an income tax. State Tax Guide (P-H) 228 (1979).
. There is also a possible argument that the statute creates a disincentive to long-term residents to move out of Alaska, thus dampening their exercise of the right to migrate. The Zobels do not complain that this aspect of the system is a factor in their own decision of whether to exercise the right of interstate migration.
. The state goes on to explain that it is necessary only to recast these negative statements into positive ones to establish the actual purposes of the tax exemption plan.
. This pattern does not always hold true. See, e. g., Baldwin v. Montana Fish & Game Comm’n, 436 U.S. 371, 98 S.Ct. 1582, 56 L.Ed.2d 354 (1978) (both privileges and immunities grounds and equal protection grounds used in attack on distinction between residents and nonresidents).
As the Zobels here are residents of Alaska, they concede they have no standing to directly invoke the protection of the privileges and immunities clause. However, they contend that the reasoning in privileges and immunities cases is persuasive here.
. A durational residency requirement, which draws a distinction between new and old residents based on the length of their residency, must be carefully distinguished from a residency requirement, which draws a distinction between residents and nonresidents. Generally, a state has much more authority to draw distinctions between residents and nonresidents than between long-term and short-term residents. See Vlandis v. Kline, 412 U.S. 441, 452-53, 93 S.Ct. 2230, 2236-2237, 37 L.Ed.2d 63, 72 (1973); Fisher v. Reiser, 610 F.2d 629, 635 (9th Cir. 1979).