OPINION

1. INTRODUCTION

Kenneth Manning was denied a Tier II subsistence permit for the 2000-2001 Nelchi-na caribou herd hunt. Manning brought suit, contending that certain aspects of the regulation governing the Tier II subsistence hunting permit point system discriminate against urban hunters in violation of state constitutional provisions requiring equal access to natural resources. Manning challenged three criteria in the regulation used to identify subsistence hunters eligible for the Nelchina caribou hunt: (1) access to alternative sources of game; (2) the cost of groceries; and (3) the cost of gasoline. All three scores are based initially on the applicant’s self-reported data, but are capped based on composite data drawn from the community where the applicant resides. The superior court upheld the food and gas criteria but held that the alternative sources of game criterion was unconstitutional. We affirm.

II. FACTS AND PROCEEDINGS

Alaska Statute 16.05.258 governs the allocation of game in subsistence areas. Under this statute, the Board of Game identifies the portion of each game population that can be harvested consistent with sustained yield.1 If the harvestable portion of the population is not large enough to satisfy the needs of all subsistence applicants, the Board restricts the number of subsistence permits that it issues.2 The more restrictive group of sub*1217sistence permits are called Tier II permits.3 The Board identifies those eligible for Tier II permits through limitations based on (1) “the customary and direct dependence on the game population by the subsistence user for human consumption as a mainstay of livelihood,” and (2) “the ability of the subsistence user to obtain food if subsistence use is restricted or eliminated.” 4

5 AAC 92.070 sets out the point system for ranking Tier II subsistence hunting permit applicants. The more points an applicant receives, the greater the likelihood the applicant will receive a permit. Subsection (a) describes how to measure the applicant’s “customary and direct dependence on the game population by the subsistence user for human consumption as a mainstay of livelihood.”5 This subsection awards up to fifty points based on the number of years the applicant has “hunted on or eaten from the [relevant] game population,” or would have had he been permitted to hunt.6 An additional ten points are available based on the number of years a member of the applicant’s household has hunted or eaten or would have hunted or eaten from the game population in question.7 Manning, a veteran hunter of Nelchina caribou, was awarded the full sixty points under subsection (a) of the regulation. He does not challenge the constitutionality of this subsection.

Subsection (b) of the regulation describes the process for measuring “the ability of a subsistence user to obtain food if subsistence use is restricted or eliminated.”8 Under subsection (b), points are awarded based on the three criteria challenged in this appeal: access to alternative sources of game, or the “game ratio” (twenty points), the cost of groceries (ten points), and the cost of gasoline (ten points). Subsection (b) provides:

The “ability of a subsistence user to obtain food if subsistence use is restricted or eliminated” may provide up to 40 points. It is measured by the following indicators and points:

(1) the relative availability of alternative sources of game to the applicant’s household, which may provide up to 20 points, as measured by the formula Score = 20(I/J), in which “I/J” is the percent of the applicant’s household’s wild game that came from the Tier II population over the past five years, in which “I” stands for the amount of game harvested by hunters from the applicant’s household from the Tier II population and “J” stands for the amount of game harvested by hunters from the applicant’s household from within the state; “1/3” may be a percent up to but not exceeding G/H, in which “G” stands for the amount of game harvested by hunters from the applicant’s location from the Tier II population and “H” stands for the amount of game harvested by hunters from the applicant’s location from within the hunt area and from all reasonably accessible game hunts within 150 miles, as calculated by the department; before January 1, 2012, the provisions of this paragraph do not apply to Units 22 and 23 musk oxen hunts;

(2) the availability of food for purchase in the community where most of the applicant’s household’s store-bought food was purchased during the past year, which may provide up to 10 points, as calculated by the department’s current Tier II cost-of-food index; the number of points received by an applicant may not exceed the points calculated by the department using the cost-of-food index for the community nearest the applicant’s residence; and

*1218(3) the cost of gasoline in the community where most of the applicant’s household’s gasoline was purchased during the past year, which may provide up to 10 points; the number of points received by an applicant may not exceed the points calculated by the department using the cost of gasoline for the community nearest the applicant’s residence.[9]

Evidence before the superior court provides more detail about how this regulation is administered. The Board calculates the food and gas caps from grocery and gas surveys compiled approximately every other year by the University of Alaska. The G/H game ratio cap is based on the Board’s records of Tier II and other “big game” kills. The units used in the ratio are pounds of big game meat, with standard weights ascribed to each kill of a particular type of game. The Board determines how many permits it can issue for a particular Tier II population in a particular year consistent with sustainable yield. It then scores all the applications and calculates a minimum cutoff score that will produce the correct number of permits.10

Under the current regulatory scheme, Manning received a Tier II subsistence permit to hunt Nelchina caribou in the five seasons preceding the 2000-2001 season. Manning applied for, but did not receive, a permit for the 2000-2001 season. Questions 14 through 19 of the application for the 2000-2001 hunt solicited Manning’s Tier II scoring data. Question 14 asked how many year’s the applicant had hunted or eaten meat from this particular game population, or would have hunted this population if permitted to do so. Manning answered “36” years. Question 15a asked for the maximum number of years any one living member of the applicant’s household had hunted or eaten from this game population. Manning again answered “36” years. Question 16 asked what percent of the total amount of big game meat that all household members had harvested over the past five years within this hunt area and within 150 miles of the applicant’s residence came from the Tier II population. Manning checked the box marked “91-100%.” Questions 17 and 18 asked in what community the applicant bought most of his food and gasoline. Manning wrote down “Anchorage” in response to both questions. In response to question 19, Manning stated that Girdwood was his community of principal residence.

Based on his answers to questions 14 and 15, which showed thirty-six years of hunting Nelchina caribou, the Board gave Manning the full sixty points for “customary and direct dependence on the ... game population.”

The next component of Manning’s score was based on the game ratio. Had this part of his score been based solely on the statement in his application that 91-100% of the game harvested by his household over the past five years came from the Nelchina caribou population, Manning would have been awarded something close to the full twenty possible points.11 But Manning’s score was capped at 5.77, the maximum game ratio score available to residents of Girdwood. This cap was based on the “G/H” ratio, or the percent of big game meat that hunters from Girdwood harvested in the hunt area and within 150 miles of Girdwood that came from the Nelchina caribou population.12

The Board awarded Manning zero out of the twenty total points available based on the cost of food and gas. Manning received this score because the score for those who purchase most of their food and gas in Anchorage is zero. He would have received the *1219same score if he had stated on his application that he purchased most of his food and gas in Girdwood because the cap for Girdwood residents is also zero.

Manning’s total score was 65.77, the highest possible score a Girdwood resident could get under the regulation. The cutoff for receiving a permit in the 2000-2001 season was approximately 68.22 points. Girdwood hunters therefore had no chance of receiving the 68.22 minimum score for 2000-2001. The residents of several other communities were also ineligible for permits in 2000-2001 because the caps made it impossible for these individuals to receive a score above the cutoff.

In July 2000 Manning filed suit alleging that the use of the applicant’s community of residence in awarding points for Tier II permits violates various provisions of the state and federal constitutions. Both parties moved for summary judgment.

On May 11, 2003, Superior Court Judge Sen K. Tan issued a summary judgment decision declaring that the game ratio portion of the regulation violates sections 3 and 17 of article VIII of the Alaska Constitution, but upholding the food and gas criteria in the regulation. Both parties filed motions for reconsideration. On June 16, 2003, the superior court issued an order clarifying .that its summary judgment decision applied to the current regulation, which had been amended while the summary judgment motions were pending. In this order, the superior court also denied the State’s motion for reconsideration in all other respects and denied Manning’s motion for reconsideration. On July 24, 2003, the superior court entered final judgment declaring the game ratio portion of the regulation unconstitutional. The State appealed and Manning filed a cross-appeal. Manning contends that the food and gas criteria should also be invalidated; the State asserts that the game ratio criterion should be reinstated.

III. DISCUSSION

A. Standard of Review

We review a grant of summary judgment de novo.13 We review constitutional questions using our independent judgment.14

B. Legal Framework for Interpreting the Equal Access Clauses of Article VIII

Four provisions from the natural resources article of the Alaska Constitution are relevant to this appeal.15 The equal access clauses are article VIII, sections 3,16 15,17 and *122017;18 the sustained yield clause is contained in article VIII, section 4.19

McDowell v. State20 contains what is perhaps our most thorough discussion of the equal access clauses as they relate to statutes and regulations that limit access to subsistence fishing and hunting activities. In McDowell, we held that the portion of the 1986 subsistence statute limiting subsistence fishing and hunting activities to rural residents violated the equal access clauses of the Alaska Constitution. Our disposition in McDowell was based on two alternative grounds. In Part A of McDowell, we established that the subsistence statute’s overt residency requirement was per se impermissible. We explained that “[a]lthough [the prohibition on exclusive rights or special privileges in] section 15 pertains only to fisheries, the prevention of grants of exclusive or special privileges with respect to fish and game is also one purpose of the common use and the uniform application clauses.”21 Consequently, we concluded, “grant[s] of special privileges with respect to game based on one’s residence [are] also prohibited.”22

In Part B of McDowell, we established that the subsistence statute’s residency requirement was also impermissible under an equal protection type of analysis. But while we held that an equal protection analysis was proper and that the subsistence statute’s residency requirement could not withstand such an analysis, we did not reach a majority consensus as to the proper level of scrutiny to apply. The plurality opinion began by noting that the constitution permits the use of a system to exclude some, but not all, applicants from subsistence activities where exclusion is required for species protection reasons.23 In such cases, the opinion went on, “assuming that the exclusionary criterion is not per se impermissible, ... demanding scrutiny is appropriate.”24 Therefore, “[i]n reviewing legislation which burdens the equal access clauses of article VIII, the purpose of the burden must be at least important. The means used to accomplish the purpose must be designed for the least possible infringement on article VIII’s open access values.”25 Employing this equal protection analysis, the plurality opinion determined that the interest in ensuring “that those Alaskans who need to engage in subsistence hunting and fishing in order to provide for their basic necessities are able to do so” is an important one.26 The plurality opinion nevertheless struck down the statute because the rural-urban criterion was an “extremely crude” means of furthering this interest given that many urban Alaskans have legitimate claims as subsistence users and many rural Alaskans do not.27 However, only two justices adopted the de*1221manding scrutiny test in its entirety. Justice Rabinowitz dissented from both Parts A and B of McDowell; Justice Compton declined to comment on Part B; and Justice Moore wrote separately to articulate a different and relatively weaker “close scrutiny” test.

Like the plurality opinion, Justice Moore’s concurring opinion determined that “access to wildlife for subsistence purposes ... is [an] ... important right.”28 Unlike the plurality opinion, however, Justice Moore’s concurring opinion stopped short of declaring that this important right merited demanding scrutiny and instead maintained that the “[t]he challenged enactment ... should receive close scrutiny.”29 As explained in Justice Moore’s concurring opinion, close scrutiny requires that the statute or regulation in question “be closely related to an important state interest.”30 Because there was “only a modest correlation between the set of people who reside[d] in areas designated as rural under the Act and the set of people who [were] dependent upon subsistence hunting and fishing,”31 Justice Moore’s concurring opinion concluded that the Act was not closely related to the State’s interest in “ensuring that those who must engage in subsistence hunting and fishing are able to do so.”32 As such, the subsistence statute s residency requirement could not stand.

As we have previously noted, “[w]hen a fragmented court decides a case and no single rationale explaining the result enjoys the assent of [the majority], the holding of the court may be viewed as that position taken by those Members who concurred in the judgment on the narrowest grounds.”33 In McDowell, it was Justice Moore who concurred on the narrowest grounds and his position would therefore ordinarily be considered the court’s holding. However, Justice Moore’s concurrence expressly refused to rule out the possibility that a more stringent test was merited34 and our subsequent case law has, in fact, repeatedly articulated — although never actually applied — the plurality opinion’s stringent demanding scrutiny test.35 Ultimately then, it is not entirely clear which equal protection test carries precedential weight.

Because we hold that (1) the game ratio is impermissible under the more lenient close scrutiny test, and (2) the food and gas ratios are permissible under the more restrictive demanding scrutiny test, it is unnecessary for us to choose today between the two McDowell tests.36

*1222C. The Rule Against Residency-Based Criteria

We first must determine whether the challenged criteria are' per se impermissible because they violate the rule against residency-based criteria in Part A of McDowell. Manning argues that the caps based on community of residence as applied to an applicant’s game ratio, food, and gas scores constitute impermissible residency-based restrictions.- We disagree.

First, we note that as a factual matter Manning is incorrect in his allegation that the regulation discriminates against urban hunters and amounts to a reinstatement of the rural preference stuck down in McDowell and Kenaitze. Of the 2000 Tier II hunting permits issued for the 2000-2001 Nelchina caribou hunt, 1,493 were issued to Anchorage, Palmer, and Wasilla area residents.

Second, as a matter of law, the criteria in the regulation are not impermissible residency-based criteria. The criteria struck down in McDowell and Kenaitze created an arbitrary preference based explicitly on where one lived.37 By contrast, the criteria in 5 AAC 92.070(b) seek to measure the applicant’s relative access to alternative sources of game, store-bought food, and gasoline. The regulation calls for an individual determination as to each applicant, and an applicant who gets a higher percentage of his game from outside the Tier II population than his neighbors or who purchases food and gas in a cheaper location will receive a different score than others who live in the same community. It is true that an applicant’s game ratio, food, and gas scores are capped according to community of residence, but these caps simply reflect the economic and geographic realities in Alaska: that location affects access to game populations and the expense of food and gas. The only way to eliminate any reference to residence from a regulation that seeks to measure access to alternative game and relative food and gas prices would be to rely solely on individual, self-reported data that is not verifiable and would in any case still likely be strongly correlated to an applicant’s place of residence. Virtually any method of measuring “the ability of the subsistence user to obtain food if subsistence use is restricted or eliminated” under AS 16.05.258(b)(4)(B)(iii), including criteria (such as income), that are facially completely unconnected to residence, will likely produce results that are closely correlated with residence. . In light of this reality, Part A of McDowell must be read to prohibit only criteria that are explicitly and exclusively based on residence because such classifications are based on the arbitrary assumption that Alaskans in certain locations engage in subsistence activities while those in other communities do not.

We conclude that the residency caps in 5 AAC 92.070(b) do not violate the rule against residency-based criteria in Part A of McDowell.

*1223D. Equal Protection Analysis

McDowell’s demanding scrutiny and close scrutiny tests both require that restrictions on access to natural resources serve an important purpose. One purpose of Alaska’s subsistence statute and its implementing regulations is to ensure that Alaskans who need to engage in subsistence hunting and fishing in order to provide for their basic necessities are able to do so.38 In McDowell, we held that this interest is an “important” one for the purposes of an equal protection analysis.39 We therefore must determine the fit between this important interest and the criteria challenged by Manning — the game ratio, the cost of food score, and the cost of gasoline score.

1. The game ratio

The “game ratio” in 5 AAC 92.070(b)(1) seeks to measure the relative availability of alternative sources of game. Access to other game is clearly a reasonable criterion to use in determining an applicant’s dependence on subsistence hunting. But Manning argues that the regulation’s game ratio scoring system is not an accurate method of measuring an applicant’s access to other game. We agree.

As described in more detail above, 5 AAC 92.070(b)(1) provides for an applicant to be assigned up to twenty points based on the percent of the applicant’s household’s game that came from the Tier II population over the past five years. An individual applicant’s score cannot exceed the game ratio for his community of residence, which measures what percent of the community’s total big game catch in the surrounding area and in Nelchina comes from the Nelehina caribou herd.

As Judge Tan pointed out in his summary judgment decision, the game ratio cap does not accurately measure a community’s access to alternative game. The formula assumes that hunters’ actual use patterns reflect the relative availability of game for a community. This is not necessarily the case. Some local hunters may seek out particular Tier II game populations for non-subsistence reasons such as aesthetics, sport, or taste. A community may draw much of its game from a particular Tier II population and thus have a high game ratio score even if other sources of game are readily available. Relying on use patterns may be particularly problematic in smaller communities where hunting patterns of a small number of individuals may significantly affect the community’s game ratio score.40

The second problem with the game ratio formula is what the superior court called the “ratcheting effect”: every year that a community is excluded from the Tier II hunt, the game ratio will go down; after five years of exclusion the game ratio will be zero. Once a community is excluded because of the game ratio cap, no one in that community will be able to secure a permit unless the State decides to issue more permits. The State argues that the downward ratchet is fair because the only way a community drops out of contention for the permits is if there is an initial determination that it has access to alternative sources of game. But this argument is based on two questionable assumptions: that the community’s game ratio was an accurate measure of alternative resources to begin with, and that the availability of alternative game will remain stable once the community has been excluded from the Tier II hunt.

*1224The game ratio is a structurally infirm and ultimately inaccurate method of measuring applicants’ access to alternative food sources. We conclude, therefore, that 5 AAC 92.070(b)(1) is not closely related to the State’s interest in ensuring that Alaskans who need to engage in subsistence hunting are able to do so. As such, the regulation violates sections 3, 15, and 17 of article VIII of the Alaska Constitution.

Our holding does not imply that it is impossible for the State to craft a constitutionally acceptable regulation measuring an applicant’s alternative sources of game. To name just one possibility, a community cap based directly on the availability of other big game hunts reasonably accessible to a given community would seem to be a more accurate measure than the current regulation.41 Nor does our holding imply that such a regulation must produce uniformly perfect results. Our decision today merely holds that the current game ratio regulation is so faulty as to not satisfy the requirements of the Alaska Constitution.

2. The food and gas criteria

Manning also challenges the food and gas criteria in 5 AAC 92.070(b)(2) and (3). Under the regulation, up to ten points each are awarded based on the cost of food and gasoline in the community where the applicant’s household purchases most of its food and gas.42 Both scores are capped based on the cost of food and gas in the applicant’s community of residence.

The cost of groceries is a reasonable way to determine the applicant’s access to store-bought food if subsistence use is restricted, and the cost of gasoline is a reasonable way to measure the applicant’s ability to access alternative game hunts through gasoline-powered vehicles such as cars and snowma-chines. The residency caps are also reasonable. These caps are designed to measure the relative availability of alternative food; the fact that an individual applicant may purchase most of his food in a location that is more expensive than where he lives simply indicates that that applicant is not making the most cost-effective use of the alternative sources of food available to him. A community-wide cap makes sense because the cost of food and gas in that location is the same for all applicants.

Manning argues that these criteria are not narrowly tailored because criteria based on income would be a more accurate measure of dependence on subsistence hunting. The Board rejected using income as a criterion after serious consideration.43 The primary difficulty with an income classification is that there is evidence in the record that in rural areas members of wealthier households are often the primary subsistence hunters. They share their harvest with poorer members of the community who are unable to hunt for themselves but who depend on the game to meet their basic necessities of life.44 Therefore, income criteria would not necessarily correspond to the neediness of the end-consumer of the meat, and income-based classifications would not necessarily be more narrowly tailored than the food and gas criteria for achieving the State’s purposes. Moreover, income-based criteria would involve costly and time-consuming evaluation of individual financial data for thousands of Alas*1225kans.45 To most precisely measure an applicant’s ability to pay for store-bought food, the Board would have to take into account not just the applicant’s taxable income, but other measures of wealth and net worth in a complicated calculation similar to college financial aid determinations or individualized child support awards under Alaska Civil Rule 90.3.

We conclude that the food and gas criteria in 5 AAC 92.070(b)(2) and (3) are narrowly tailored and “designed for the least possible infringement on article VIII’s open access values.”46 They therefore survive constitutional review even if subjected to demanding scrutiny.

E. Manning’s Rule 11 Motion

During discovery, Manning made requests for admissions. In a response signed by Board official Steven Schwartz, the State replied as follows:

2. Please admit that plaintiff Kenneth H. Manning was automatically denied approximately 34.23 points ... based on his community of residence of Girdwood and because he bought the majority of his groceries and gasoline over the previous year in Girdwood and/or Anchorage.

Answer: Denied. Mr. Manning was not automatically denied any points. His scoring was based on his ability to obtain food if subsistence hunting is restricted or eliminated.

Another request asked whether the Board “predetermined a fixed amount of exclusionary points” for Girdwood residents. This request was also denied. Manning sent a letter to Assistant Attorney General Kevin Saxby accusing Schwartz of “fraud and perjury.” Saxby replied by letter, stating that Manning’s accusation was false and that “in my opinion, Mr. Schwartz has a cause of action against you for defamation, should he ever decide to pursue it.” Manning then moved for sanctions against Saxby for violations of Alaska Civil Rule 11. The superior court denied the motion during oral argument. Manning appeals the denial of that motion.

Decisions on Rule 11 motions are reviewed for abuse of discretion.47 Rule 11 prohibits attorneys from knowingly filing pleadings and court papers that are not “well grounded in fact.”48 Alaska’s Rule 11 no longer provides for mandatory sanctions,49 but sanctions may be imposed for a violation of Rule 11 under Civil Rule 95.50 It was not an abuse of discretion to deny sanctions here. The use of the words “automatically denied” and “predetermined a fixed amount of exclusionary points” in Manning’s requests for admissions were provocative, and the State’s denials to these questions as worded did not amount to false statements. Moreover, the State’s responses could not have misled Manning about the process for reviewing permit applications: in the same set of responses, the State admitted that residents of Gird-wood could not have achieved the minimum successful score. The superior court did not err in denying Manning’s Rule 11 motion.51

*1226IV. CONCLUSION

We hereby AFFIRM the order of the superior court.

dissenting in part.

I disagree with the court’s conclusion that Alaska’s game-ratio provision violates the Alaska Constitution’s equal protection and uniform application clauses.1 As the court correctly recognizes, the constitutional test we must apply is the one described in Justice Moore’s McDowell concurrence.2 But the court misapplies that test here. In my view, because the challenged regulation uses individualized criteria that are facially neutral and reasonably respond to an actual resource allocation need, the regulation would fail to pass constitutional muster only if Manning proved that the game ratio resulted in an actual injustice as applied to his ease. He has not made this showing.

McDowell considered a subsistence law that expressly discriminated on the basis of residency, categorically denying subsistence hunting rights to all urban residents. Moreover, the law at issue in McDowell applied at the Tier I level; it restricted subsistence rights in the absence of a determination that the restriction was necessary to protect subsistence resources.3 Faced with these circumstances, the McDowell plurality’s demanding scrutiny test articulated a constitutional- standard to be used when a law not specifically driven by resource needs favors one group of subsistence users over another on the sole basis of residency.

The justices who joined in Part A of McDowell took pains to emphasize that they did “not imply that the constitution bars all methods of exclusion where exclusion is required for species protection reasons.”4 Rather, they held “only that the residency criterion used in the 1986 act which conclusively excludes all urban residents from subsistence hunting and fishing regardless of their individual characteristics is unconstitutional.”5 McDowell thus specifically recognized that more latitude to regulate access would be allowed under a Tier II determination, when restrictions served actual resource-protection needs.

In his concurring opinion, even though acknowledging that the case presented an overtly discriminatory “geographical classification scheme,”6 Justice Moore articulated a more moderate understanding of Alaska’s constitutional standard. Because he believed that subsistence rights were similar to the “important right to engage in economic endeavor” addressed in State v. Enserch Alaska Construction, Inc.,7 Justice Moore used the same mid-level scrutiny test we adopted in Enserch.8 This test asks whether the challenged law is “closely related to an important state interest.”9 In light of the challenged law’s overt and exclusive reliance on the distinction between urban and rural residency, Justice Moore found that its means were not closely related to its goals. *1227Yet at the same time, Justice Moore stressed that “[a] law providing for individual determinations of eligibility would in my view be sufficiently tailored to the state’s interest to withstand a constitutional challenge.”10 As already noted, Justice Moore’s view on this point governs here.11

Unlike the law disputed in McDowell, the subsistence regulation now at issue, 5 AAC 92.070, was adopted in response to a Tier II determination and thus serves a specific resource-protection need. Furthermore, the regulation uses an individualized, facially neutral point system in determining eligibility — not an overtly discriminatory “geographical classification scheme.”12 Nothing in Justice Moore’s McDowell concurrence suggests that a rational, individualized, Tier II regulation of this kind could be deemed too loosely tailored to pass facial constitutional muster merely because, in theory, one of its measurements might not always perfectly fit its goal.

Nor does any post-AfcDoweM ease suggest the need for such a demanding means-to-end fit. In fact, our post -McDowell ease law suggests the opposite. In Gilbert v. State,13 which the opinion cites as re-articulating the McDowell plurality’s “least possible infringement” test,14 this court actually refused to apply any form of the McDowell test, even Justice Moore’s moderate “close relationship” requirement.

Like McDowell, Gilbert involved a constitutional challenge to a regulation that, on its face, adopted an overtly discriminatory classification — a commercial fishing regulation that favored the salmon fishery in Chignik over a competing, mixed-stock interceptor fishery in Stepovak in which most salmon were headed for their natal streams around Chignik.15

While McDowell had described this sort of facially disparate treatment as a “geographical classification scheme,”16 Gilbert referred to it as a “non-uniform classification!] ].”17 Citing McDowell and other Alaska cases, Gilbert acknowledged that “an analysis under [the Uniform Application Clause of article VIII] may invoke ‘more stringent review.’ ”18 Gilbert went on to observe that the more stringent test applies when the court reviews “fish and game regulations creating non-uniform classifications.”19 But the court in Gilbert nonetheless declined to apply this stringent test to the non-uniform classification before it because the competing fisheries were not similarly situated. The court instead emphasized that “[t]he regulation in question reflects an allocation decision authorized under article VIII, section 4 of the state constitution which the Board must necessarily make between users involved in different fisheries.”20 On this basis, Gilbert held that the regulation’s provisions were therefore subject to conventional rational-basis review: “Such decisions are within the power of the Board, so long as they are not arbitrary and unreasonable and are ‘consistent with and reasonably necessary to the conservation and development of Alaska fishery resources.’ ”21

In effect, then, Gilbert recognized that, even when a fish and game regulation cre*1228ates an openly non-uniform classification, the conventional rational-basis standard applies if the regulation deals with an issue of resource allocation that article VIII, section 4 allows the board to address and if the board finds that existing circumstances make it necessary to decide the issue. The situation here is functionally identical to the one in Gilbert: because this case involves a Tier II determination, the disputed regulation addresses an allocation decision that article VIII authorizes and that “the Board must necessarily make”22 in order to preserve scarce game resources. Under Gilbert, it follows that the classifications at issue here identify competing resource users who are not similarly situated and that the regulation must accordingly be reviewed under the conventional rational basis test.

But even if Gilbert did not dictate this conclusion, Justice Moore’s concurring opinion in McDowell would require no more than the moderately heightened equal protection review we applied in Enserch, which demands a close but not perfect means-to-end fit. Under either approach — Gilbert’s or Justice Moore’s — there would be no reason to declare the game ratio invalid.

Today’s opinion implicitly acknowledges that the challenged regulation is a facially neutral provision and seems to agree that the game ratio serves an important purpose: to assess the availability of alternative game resources in a given community.23 The opinion identifies no reason to suspect that the game ratio’s neutral requirements might conceal an invidious discriminatory intent. And it points to no record evidence showing that it generally produces arbitrary results or actually resulted in unfairness here. Instead, based solely on the theoretical possibility that the ratio’s formula “may be ... problematic in smaller communities” or that its ratcheting effect might reflect “questionable assumptions,” the court broadly declares the game ratio “structurally infirm and ultimately inaccurate,”24 and categorically invalidates it as unconstitutional in all applications.

This proclamation turns the usual presumption of constitutionality on its head by requiring the state to demonstrate that a seemingly reasonable (though perhaps not invariably perfect) regulation actually works in practice.25 More fundamentally, it misapplies the McDowell concurrence’s moderately heightened equal protection standard: instead of inquiring whether the game ratio is reasonably suited to achieving its goal, and thereby results in a reasonably close means-to-end fit, the court’s approach leaves no room for slippage or error. By proclaiming the ratio to be “structurally infirm” and “ultimately inaccurate” as a matter of law merely because it theoretically might produce an anomalous result in a marginal ease, today’s opinion effectively applies a test even more demanding than the leasUrestrictive-alterna-tive test discussed by the McDowell plurality-

In addition to straying from the settled analysis under McDowell’s concurrence, moreover, the opinion breaks sharply from the established standard for declaring a law to be facially unconstitutional — that is, invalid in all applications rather than just in the way it applies to the particular case at hand. As this court has recently recognized, “even under a relaxed standard of facial review it would be improper to declare [a law] invalid on its face if it has a ‘plainly legitimate sweep.’ ”26 Accordingly, there would be no basis for declaring the game ratio to be facially invalid unless Manning showed either that it arbitrarily discriminated on the basis of residency in many or most situations, and thus had no plainly legitimate sweep.

Here, Manning produced no evidence to support a finding of general invalidity, and today’s opinion makes no such finding. The *1229assumptions and structural defects that the opinion identifies as potentially flawed would at most create a risk of appreciable error only in exceptional situations.27 And even if exceptional circumstances suggested the presence of a significant risk of error with respect to a particular community or Tier II permit, these circumstances would not necessarily establish that the game ratio had inaccurately measured available alternative sources of game; instead, they would merely indicate a possible inaccuracy that would warrant further investigation before deciding whether to grant or deny the disputed permit.

Admittedly, Manning’s case presents some circumstances suggesting that the game ratio might have produced a statistical anomaly in his case. To this extent, today’s opinion correctly observes that the state’s decision to deny Manning’s application might reflect potentially “questionable assumptions.” But whether this potential for inaccuracy actually produced an inaccurate result is subject to reasonable dispute. The state, for instance, insists that in Manning’s situation, the game ratio actually did what it was supposed to do: it accurately determined that Manning had alternative sources of game within easy reach.

In my view, this factual dispute calls for a case-specific resolution. On the one hand, if Manning shows that the game ratio probably did produce an anomalous result as applied in his case, then the state should be barred from relying on subsection .070(b)(1) to deny permits to Manning and similarly situated Girdwood residents. On the other hand, absent proof that the game ratio inaccurately reflected Manning’s ready access to alternative sources of game, I see no basis for barring the state from applying the ratio to him — or anyone else. As today’s opinion recognizes, Manning is similarly situated to applicants who were granted a Nelehina permit only if “he is correct” in asserting “that the regulations in question are unconstitutional because they are an inaccurate means of determining which applicants have access to alternative food sources.”28 Absent proof that the game ratio lacks a plainly legitimate sweep, then, it follows that Manning should at a minimum have been required to show a likelihood that the regulation actually produced inaccuracy as applied to his situation.

Yet the superior court did not decide, or even consider, whether applying subsection .070(b)(1) to Manning’s case would actually cause injustice in his situation. Instead, the court summarily struck the regulation as unconstitutional in all applications based on a speculative risk of inaccuracy in seemingly marginal applications. Because this ruling converted the ratio’s potential but unproved inaccuracy in exceptional situations into an unwarranted facial disqualification in all applications, I would reverse and remand for an evidentiary hearing to determine whether the regulation actually did unfairly deny Manning his Tier II permit.

I therefore dissent from today’s opinion upholding the superior court’s judgment.

dissenting in part.

The main question posed by this case is whether 5 AAC 92.070(b)(1), (2), and (3), which enumerate criteria for determining who qualifies for Tier II subsistence hunt permits, are valid. I conclude that none of the challenged criteria contribute to the statutorily required individualized determination of which applicants most need a subsistence permit to meet their food needs. The regulations are therefore inconsistent with the subsistence use statute and should be invalidated.

This appeal involves a challenge to regulations 5 AAC 92.070(b)(1), (2), and (3). Normally, when reviewing an administrative regulation, this court limits its examination to “(1) whether the regulation is reasonable and not arbitrary; and (2) whether the regulation is consistent with the statute and reasonably *1230necessary to its purposes.”1 While a rational basis standard would apply if the “issue involve[d] agency expertise or the determination of fundamental policy questions on subjects committed to an agency,” generally, as here, determining whether a regulation is consistent with a statute is a question “of statutory interpretation to which we should apply our independent judgment.”2 We reach constitutional issues only when a case “cannot be fairly decided on statutory or other grounds.”3 Since in my view none of the challenged regulations are consistent with the subsistence use statute, we need not reach the issue of whether or not the regulations are constitutional.

The subsistence use statute requires that when determining which subsistence uses to allow, the Board “distinguish among subsistence users”4 in part by making a determination about “the ability of the subsistence user to obtain food if subsistence use is restricted or eliminated.”5 As this language indicates, the statute requires the Board to make individualized, user-by-user determinations of who is best able to obtain food from non-subsistence sources.

None of the challenged regulations assist the Board in making the statutorily required individualized determinations. The “game ratio” in 5 AAC 92.070(b)(1) does not advance an individualized determination of who can best obtain alternate food in part because, as Judge Tan pointed out, it is not an accurate method of measuring access to other game.6 The other two challenged portions of the regulation, 5 AAC 92.070(b)(2) and (3), which award points that are capped according to the cost of food and gasoline in the community nearest the. applicant’s residence, presume that an individual’s ability to secure food is governed by community-based costs. While food and fuel costs undoubtedly have a slight impact on an applicant’s ability to obtain food, differentials in prices are much less determinative of a subsistence applicant’s ability to access other sources of food than are differentials in income.7

*1231In McDowell v. State, we acknowledged that the subsistence use statute then in effect had the important purpose of “ensuring] that those Alaskans who need to engage in subsistence hunting and fishing in order to provide for their basic necessities are able to do so.”8 We concluded that the urban/rural distinction drawn by the statute was an “extremely crude” means “to accomplish this purpose.”9 While the subsistence use statute has changed somewhat in the intervening years, its purpose of ensuring that Alaskans who need to are allowed to engage in subsistence activities remains the same. The food and gas criteria, just like the rural/urban criterion, are extremely crude means of determining who most needs to engage in subsistence activities. In my opinion they are of such slight relevance to the question of individual need as to be inconsistent with the purpose of the statute. Because the criteria fail to withstand the standard of review normally applied to regulations, they should be invalidated.

In order to advance the statute’s purpose of making an individualized determination of who among subsistence applicants is least able to access alternative sources of food, the Board needs to take into account applicants’ income. While it is the province of the Board to draft new regulations, one way for the Board to obtain income information would be to ask applicants to list their adjusted gross income as reported on their most recent tax return.10

The state and subsistence users have raised concerns about the fairness and feasibility of using income as a criterion for determining which applicants receive Tier II permits. An income-based criterion can be structured to address many of the concerns raised. For instance, since the cost of living does affect (although it does not determine) a person’s ability to obtain food, the Board could decide to adjust each applicant’s income to account for cost-of-living differentials. As for the point that members of wealthier rural households are often the primary subsistence hunters who then share with poorer members of the community, the Board can create regulations that take into account the income of those who rely upon the applicant for food. If the Board is worried about applicants manipulating the process by having poorer household members apply for permits, it can order occasional audits to make sure that those people doing the hunting are the ones in whose names the permits were issued.

Many of the objections made to an income criterion are also overstated. For instance, an income criterion is not uniquely manipulable. Under a portion of the subsistence regulations not challenged in this appeal, the Board already asks applicants to report how long they have hunted or eaten meat from a *1232specific game population. This presents a ready opportunity for the unscrupulous to stretch the truth. An income criterion would serve as an improvement, since applicants will be less able to mislead when their answers are easily verifiable by reference to their tax returns. When the Board considered an income criterion, it also expressed concerns that asking for income information would compromise the privacy interests of applicants. However, the Tier II Subsistence Hunting Permit Application already asks applicants to provide their social security numbers, which are also sensitive personal information. Presumably the Board has privacy safeguards in place that could be applied to income information.

Finally, the Board has objected to the administrative burden it believes an income criterion would create. In the words of one Department of Fish and Game analyst, “[w]e don’t want to be in the business of collecting income tax return forms for 20,000 Alaskans and — and process that every year.” However, an income criterion would not necessarily result in the Board being inundated by tax returns from all applicants. Having asked applicants both to include their adjusted gross income on the permit application form and to provide the Board with permission to obtain a copy of the applicant’s tax return, the Board could limit its acquisition of tax returns to those circumstances in which it wished to verify an applicant’s income response.

For the reasons stated, the challenged regulations are not consistent with the subsistence use statute. They therefore should be invalidated.