dissenting:
I respectfully dissent. Although I agree fully with the majority’s analysis at step one of the inquiry, I would hold, at step two, that the differential fees survive summary judgment. Further evidentiary development is necessary to determine whether the nonresident fees “merely compensate the State for any added enforcement burden [nonresidents] may impose or for any conservation expenditures from taxes which only residents pay.” Toomer v. Witsell, 334 U.S. 385, 399, 68 S.Ct. 1156, 92 L.Ed. 1460 (1948).
We have little guidance to assist us in determining what the United States Supreme Court meant in the foregoing passage from Toomer. Only twice since Toomer has the Court quoted the phrase “taxes which only residents pay” in a privileges and immunities context, and in neither case did it explain the meaning of those words. Baldwin v. Fish & Game Comm’n, 436 U.S. 371, 401, 98 S.Ct. 1852, 56 L.Ed.2d 354 (1978); Mullaney v. Anderson, 342 U.S. 415, 417, 72 S.Ct. 428, 96 L.Ed. 458 (1952). As I explain in more detail below, two state supreme courts have reached different conclusions about the proper interpretation of that phrase. But we do not know which (if either) of those courts got it right, because the Supreme Court denied certiorari in both cases. Carlson v. Alaska Commercial Fisheries Entry Comm’n, 519 U.S. 1101, 117 S.Ct. 789, 136 L.Ed.2d 730 (1997); Glaser v. Salorio, 449 U.S. 874, 101 S.Ct. 215, 66 L.Ed.2d 94 (1980). Further complicating our interpretive task, the Privileges and Immunities Clause of Article IV “is not one the contours of which have been precisely shaped by the process and wear of constant litigation and judicial interpretation over the years since 1789.” Baldwin, 436 U.S. at 379, 98 S.Ct. 1852.
Acknowledging those limitations, we must decide how to interpret the phrase “taxes which only residents pay.” Toomer, 334 U.S. at 399, 68 S.Ct. 1156. On the one hand, as the State urges, the phrase could be read to refer to residents’ aggregate tax contribution to -commercial fishing. Under that reading, California per*967missibly could charge differential fees to nonresidents so long as those fees do not exceed the nonresidents’ fair share of the portion of commercial fisheries management costs that California residents’ tax dollars fund.
The New Jersey Supreme Court has interpreted Toomer in this way. In Salorio v. Glaser, 82 N.J. 482, 414 A.2d 943 (1980), the plaintiffs challenged New Jersey’s imposition of an Emergency Transportation Tax, which applied only to nonresident users of the state highway system. Although it found the record insufficiently developed to render a final decision, the New Jersey Supreme Court held that a tax that applied only to nonresidents could, in theory, pass constitutional muster, because “[t]he Constitution does not entitle nonresident commuters to a ‘free ride.’ The State may exact from them a fair share of the cost of adequate transportation facilities without violating the Privileges and Immunities Clause.” Id. at 954. The court read Toomer and other Supreme Court cases to authorize a state y to “impose upon non-residents the additional expenses occasioned by their activities within the state, or the reasonable costs of benefits which they receive from the state.” Id. at 953.
Applying the Salorio court’s reasoning here, nonresidents are on “equal footing” with residents so long as they are not charged more than their “fair share” of commercial fisheries management expenses that residents’ tax dollars fund. California introduced evidence that nonresidents purchased 11% of commercial fishing licenses, while the differential fees for out-of-state licenses equaled only 3% of the net general fund contributions to the Department of Fish and Wildlife (“DFW”) budget. The State asserts that it constitutionally could charge differential fees that total up to 11% of the DFW’s general fund-supported commercial fishing expenditures, so the smaller fee that California actually charges is — a fortiori — permissible.
On the other hand, Plaintiffs read “taxes which only residents pay,” Toomer, 334 U.S. at 399, 68 S.Ct. 1156, very differently. They contend that the phrase requires a per capita calculation of a California resident’s tax burden related to DFW’s commercial fishing budget. The Alaska Supreme Court adopted this alternative interpretation in Carlson v. State, 798 P.2d 1269 (Alaska 1990). There, the plaintiffs challenged Alaska’s commercial fishing fees, which were three times higher for nonresidents than for residents. The state urged the court to follow Salo-rio. But the Alaska Supreme Court rejected the New Jersey Supreme Court’s interpretation of Toomer:
Implicit in Salorio is the notion that it is permissible to require nonresidents to pay up to 100% of their pro rata share of expenditures regardless of what percentage of their pro rata share residents are in fact pajdng. In other words, Sa-lorio, as applied to this case, seems to add up to a general proposition that the state may subsidize its own residents in the pursuit of their business activities and not similarly situated nonresidents, even though this results in substantial inequality of treatment.
Carlson, 798 P.2d at 1278. The court held that the proper inquiry was “whether all fees and taxes which must be paid to the state by a nonresident to enjoy the state-provided benefit are substantially equal to those which must be paid by similarly situated residents when the residents’ pro rata shares of state revenues to which nonresidents make no contributions are taken into account.” Id.
Under the Carlson court’s approach, the state would have to divide general fund *968expenditures for commercial fishing management by the total number of California taxpayers; the quotient would represent the maximum permissible differential fee. The State introduced evidence that net annual general fund outlays for commercial fisheries management total at least $12 million. Thus, for instance, if there were 12 million taxpayers in California, the per capita formula would limit the permissible differential fee to $1 per nonresident fisher.1 According to Plaintiffs, this formula puts residents and nonresidents on “equal footing” because their out-of-pocket costs to support commercial fisheries are the same.
I would reject the per capita formula. The purpose of the Privileges and Immunities Clause is to “place the citizens of each State upon the same footing with citizens of other States, so far as the advantages resulting from citizenship in those States are concerned.” Paul v. Virginia, 75 U.S. (8 Wall.) 168, 180, 19 L.Ed. 357 (1868). In my view, the per capita approach does not advance that goal. The per capita formula -attributes to each resident a pro rata contribution to every program and activity supported by a state’s general fund expenditures. But that sort of rigid across-the-board calculation does not accurately re.flect the real benefit that a taxpayer obtains through his or her tax dollars. Taxpayer dollars support a large number of state-funded programs. Education, natural resources management, healthcare services, corrections and rehabilitation, infrastructure, and transportation all are at least partially funded with state tax revenues in California. In a given year, an individual taxpayer likely receives no direct benefit from some of those programs, but a benefit that far exceeds his or her pro rata contribution from others. This is the deal that we make when we pay taxes: We all put a portion of our income into a big pot and it is spent in a variety of ways, some of which benefit us directly and some of which do not.
California residents subsidize each other with their taxes. For example, suppose that each taxpayer’s share of state support for secondary schools is $1 per year. A certain California taxpayer has a teenager who attends public high school. That taxpayer’s per capita “payment” for the educational benefit is $1, but the benefit to the taxpaying parent is worth much more than that. The parent agrees to subsidize a number of other activities in the state, including commercial fishing. In exchange, taxpayers without school-age children subsidize public education.2 The per capita formuja permits a nonresident fisher to obtain the same benefit as a resident fisher, but the nonresident does not have to subsidize any other programs or activities in California in exchange. The per capita formula thus systematically disadvantages the resident vis-a-vis the nonresident.
Instead of using a per capita formula, I would adopt the Salorio court’s “fair *969share” approach. At step two of the privileges and immunities inquiry, the state must show that the discrimination against nonresidents is “closely related to the advancement of a substantial state interest” Supreme Court of Va. v. Friedman, 487 U.S. 59, 65, 108 S.Ct. 2260, 101 L.Ed.2d 56 (1988). We recently reiterated that “[a] ‘substantial reason’ for discrimination does not exist ‘unless there is something to indicate that non-citizens constitute a peculiar source of the evil at which the statute is aimed.’ ” Council of Ins. Agents & Brokers v. Molasky-Arman, 522 F.3d 925, 934 (9th Cir.2008) (quoting Toomer, 334 U.S. at 398, 68 S.Ct. 1156)). Nonresidents increase the amount of commercial fishing activity in California’s coastal waters. That increased activity, in turn, requires the state to spend more money than it otherwise would spend on commercial fisheries management, including enforcement and conservation. Because nonresidents are a “peculiar source” of those additional costs, I would hold that not subsidizing nonresident participation in an activity funded with residents’ tax dollars is a substantial reason for discrimination. See Tangier Sound Waterman’s Ass’n v. Pruitt, 4 F.3d 264, 268 (4th Cir.1993) (assuming, without deciding, that such an interest is permissible under the Privileges and Immunities Clause).
Turning to the “close relationship” requirement, I would hold that the State has the burden to show three things. First, it must isolate the state expenditures that benefit only the licensees.3 See id. (rejecting Virginia’s differential license fee in part because it unfairly charged nonresident commercial fishers “for programs funded by all taxpayers to benefit all fishermen, whether commercial or sport fishermen”). Second, it must determine what portion of those expenditures fairly may be characterized as deriving “from taxes which only residents pay.” Toomer, 334 U.S. at 399, 68 S.Ct. 1156; see Tangier, 4 F.3d at 267 (striking down Virginia’s differential commercial fishing license fees in part because the state calculated the fee without considering nonresident fishers’ payment of state sales and use taxes); Salorio, 414 A.2d at 955 (discussing whether property and sales taxes are “taxes imposed upon residents alone” in light of the fact that some nonresidents pay them). And third, it must assess what portion of qualifying expenditures' is fairly allocable to the nonresidents as “the additional expenses occasioned by their activities within the state.”4 Salorio, 414 A.2d at 953; see also Tangier, 4 F.3d at 267 (holding that “the record does not disclose that the Commonwealth of Virginia has shown that it created any credible method of allocat*970ing costs as between residents and nonresidents”).
I would hold that the “close relationship” requirement of step two is satisfied so long as the state charges a differential fee that, in the aggregate, does not exceeds 5 the amount that the state spends that (1) benefits only licensees, (2) derives from taxes that only residents pay, and (3) is fairly allocable to nonresidents.6 This test puts residents and nonresidents on “substantially equal footing” with respect to commercial fishing: Residents reap the benefit of the tax dollars that they alone pay, and nonresidents cannot be required to pay more than their “fair share” of the benefits they enjoy that are subsidized by those resident-paid tax dollars.
This fair share approach accurately reflects the relative benefit that residents and nonresidents obtain from a state’s general fund expenditures. Suppose that a state charges a $50 license fee to resident commercial fishers. Over and above the revenue collected from those fees, the state spends $1 million in tax-supported funds on commercial fisheries management. If 10,000 people per year obtain licenses, the benefit of the $1 million subsidy to each fisher is $100. Thus, a nonresident may be charged the $50 fee that residents pay, plus a $100 differential. If only 5,000 people obtain licenses, each nonresident may be charged a $200 differential. This variance makes sense, because the benefit to each fisher of the tax-supported outlay decreases as more people use the resource. The per capita approach makes less sense because it is unresponsive to such changes; so long as a state’s tax rate and general fund outlay on the commercial fisheries program remain unchanged, the permissible differential is fixed. It is the same whether 10 or 10,000 people obtain licenses and use the resource.
Plaintiffs raise the specter of a year in which only one nonresident purchases a commercial fishing license. They argue that the state’s approach would permit California to collect hundreds of thousands of dollars from that single licensee. Not so. The fair share formula accounts for this possibility. Assuming the scenario described above, in a year in which a single nonresident and 4,999 residents obtain licenses the permissible differential for that nonresident would remain $200.7
Finally, Plaintiffs challenge the “fair share” approach because, using it, the *971state could set nonresident license fees ten, twenty, or even a hundred times higher than resident license fees. They point out that the Supreme Court has rejected nonresident fees at such ratios before. See Mullaney, 342 U.S. at 418, 72 S.Ct. 428 (invalidating nonresident fees ten times higher than resident fees); Toomer, 334 U.S. at 389, 68 S.Ct. 1156 (striking down nonresident fees one hundred times higher than resident fees). And they urge us to rely on the ratio of nonresident to resident fees here (roughly three to one)8 to reject the “fair share” analysis. ' Plaintiffs’ argument is flawed for two reasons.
First, the Supreme Court did not reject the differential fees because of the size of the ratio. Rather, it rejected the nonresident fees because Alaska and South Carolina had failed to show any connection between the differential and state spending on services to the nonresidents. See Mullaney, 342 U.S. at 418 & n. 1, 72 S.Ct. 428 (rejecting the state’s argument that the differential fees merely compensated the state for enforcement against nonresidents because the state had not calculated the cost of that enforcement and the total amount of differential fees collected “may easily have exceeded the entire amount available for administration” of the office in charge of enforcement); Toomer, 334 U.S. at 398, 68 S.Ct. 1156 (noting that “[n]othing in the record indicate[d] ... that any substantial amount of the State’s general funds [was] devoted to shrimp conservation” and that, even if there had been such evidence, it “would not necessarily support a remedy so drastic as to be a near equivalent of total exclusion”).
Second, focusing on the size of the ratio requires consideration of fees in a vacuum. That isolation makes little sense in light of the Supreme Court’s statement that a state may charge a fee designed to “compensate [it] for any added enforcement burden [nonresidents] may impose or for any conservation expenditures from taxes which only residents pay.” Toomer, 334 U.S. at 399, 68 S.Ct. 1156. In Salario, the tax at issue applied only to nonresidents. On appeal after remand, the New Jersey Supreme Court ultimately invalidated the tax — but not because the nonresident-to-resident ratio was too high. The problem was that, during a period of two decades, the revenues collected by the state through the tax had exceeded the costs attributable to nonresidents by a factor of more than two. Salorio v. Glaser, 93 N.J. 447, 461 A.2d 1100, 1107 (1983). Focus on the size of the ratio per se is misplaced; the privileges and immunities inquiry requires consideration of all taxes and fees paid by residents and nonresidents in support of commercial fishing.
Because it applied a different test, the district court did not address whether the net general fund outlay benefits only licensees, whether that outlay derives solely from taxes that only residents pay, or what portion of qualifying costs is properly allo-cable to nonresident fishers. Thus, on the current record, I would hold that we cannot determine whether the differential fees are permissible under the Privileges and Immunities Clause. Accordingly, I would reverse the summary judgment of the district court and remand for further proceedings.
. This illustrative example likely is a generous estimate, as the population of California was nearly 39 million in 2014. U.S. Census Bureau, State & County QuickFacts, http:// quickfacts.census.gov/qfd/states/06000.html. Thus, the permissible differential likely would be less than $ 1 under the per capita formula, even though a substantial number of California residents — for example, minor children— are not taxpayers.
. Of course, some commercial fishers are parents whose children attend public school. But that fact just demonstrates that each taxpayer benefits directly from a different set of state programs supported by his or her tax dollars. The Value of the taxpayer-funded investment in a given program to each individual taxpayer who benefits from that program varies. The value is less than the taxpayer’s total tax bill, but more — generally, significantly more' — than the taxpayer’s strict pro rata contribution to the program.
. These expenditures would include any costs associated with programs or activities in which only licensees participate — for example, the cost of enforcing rules such as size of fish or season limits. They also would include conservation expenditures made necessary by licensees' activities. If the state engages in conservation activities designed to keep fish stocks at a certain level, some of those activities benefit only licensees. To count those costs, the state must separate general conservation activities from conservation activities directed to the effect of commercial fishing.
. It may be, as the State asserts, that multiplying the qualifying expenditures by the percentage of commercial fishers who are nonresidents is the appropriate way to calculate those nonresidents’ fair share, but that is not necessarily the case. See Salorio v. Glaser, 93 N.J. 447, 461 A.2d 1100, 1106 (1983) ("Although the State has not shown that New York commuters cause higher average costs per commuter than New Jersey commuters, the New York commuter does exacerbate the peak load. Accordingly, both incremental and average costs are pertinent factors in determining the costs attributable to the New York commuter.”); Salorio, 414 A.2d at 955 (questioning a "strict percentage computation” that assumed equal transportation costs for nonresident and resident commuters).
. Because the Privileges and Immunities Clause neither bars the residents of a state from deciding to use their tax dollars to subsidize the activities of nonresidents nor precludes a state from providing a greater benefit to nonresidents than it provides to residents, it is permissible for a state to charge less than the maximum allowable differential.
. The test here is one of "substantial equality of treatment,” not absolute equality. Austin v. New Hampshire, 420 U.S. 656, 665, 95 S.Ct. 1191, 43 L.Ed.2d 530 (1975). So long as the state "fairly attempts to distribute the burdens and costs of government to those receiving its benefits” pursuant to a reasonable methodology, I would hold that the requirements of the Privileges and Immunities Clause are met. Salorio, 414 A.2d at 952; see also Travelers’ Ins. Co. v. Connecticut, 185 U.S. 364, 371, 22 S.Ct. 673, 46 L.Ed. 949 (1902) ("It is enough that the state has secured a reasonably fair distribution of burdens, and that no intentional discrimination has been made against nonresidents.”).
.The only way the permissible differential charged to a nonresident would skyrocket is if the overall number of fishers obtaining licenses plunged to single digits. But if that happened, the state likely would slash its commercial fisheries management spending. And if it did not cut spending, it is hard to see how the State could prove that the full $1 million in my example benefitted just a handful of fishers, because it is not reasonable to attribute hundreds of thousands of dollars in enforcement and conservation costs to a single fisher.
. The ratio of nonresident fees to resident fees for commercial fishing licenses and commercial boat registrations is three to one. For dungeness crab vessel permits, the ratio is lower (two to one); and for herring gill net permits, the ratio is higher (nearly four to one).