In this case, among other issues which we need not reach, we are asked to decide whether a Texas tax statute, Article 5248 of the Revised Civil Statutes of Texas, as amended in 1950,1 discriminates unconstitutionally against the United States and those with whom it deals. We hold that it does.

Appellant, Phillips Chemical Company, engages in the commercial manufacture of ammonia on valuable industrial property leased from the Federal Government in Moore County, Texas. The lease, executed in 1948 pursuant to the Military Leasing Act of 1947, 61 Stat. 774, is for a primary term of 15 years and calls for an annual rental of over $1,000,000. However, it reserves to the Government the right to terminate upon 30 days' notice in the event of a national emergency and upon 90 days' notice in the event of a sale of the property.

In 1954, appellee, Dumas Independent School District, assessed a tax against Phillips for the years 1949 through 1954. The tax, measured by the estimated full value of the leased premises, was assessed in accordance with the District's ordinary ad valorem tax procedures.

When the District assessed the tax, Phillips commenced the present action in the state courts to enjoin its collection. Phillips contested both the District's right to levy the tax and the valuation figure upon which the amount of the tax was calculated. The latter issue was severed by the trial court for later decision and is not involved in this appeal. The lower state courts denied relief for the years subsequent to the effective date of the 1950 amendment to Article 5248, and on writ of error the Supreme Court of Texas, by a divided court, affirmed. 316 S.W.2d 382. Phillips appealed from the decision, and we noted probable jurisdiction. 359 U.S. 987, 79 S.Ct. 1118, 3 L.Ed.2d 977.

The District's power to levy the tax was found to lie in amended Article 5248. Before 1950, Article 5248 provided a general tax exemption for land and improvements 'held, owned, used and occupied by the United States' for public purposes. In 1950, the Texas Legislature added two provisions to Article 5248, one providing for taxation of privately owned personal property located on federal lands, and the other reading as follows:

'(P)rovided, further, that any portion of said lands and improvements which is used and occupied by any person, firm, association of persons or corporation in its private capacity, or which is being used or occupied in the conduct of any private business or enterprise, shall be subject to taxation by this State and its political subdivisions.'

As construed by a majority of the Texas court, this provision is an affirmative grant of authority to the State and its political subdivisions to tax private users of government realty. While the subject of the tax is the right to the use of the property, i.e., the leasehold, its measure is apparently the value of the fee.2 The constitutionality of the provision, thus construed, depended upon the court's interpretation of our decisions in the Michigan cases two Terms ago, where we held that a State might levy a tax on the private use of government property, measured by the full value of the property. United States v. City of Detroit, 355 U.S. 466, 78 S.Ct. 474, 2 L.Ed.2d 424; United States v. Township of Muskegon, 355 U.S. 484, 78 S.Ct. 483, 2 L.Ed.2d 436; cf. City of Detroit v. Murray Corp., 355 U.S. 489, 78 S.Ct. 458, 2 L.Ed.2d 441.

However, three members of the Texas court, joined by a fourth on petition for rehearing, were of the opinion that under the majority's construction the statute discriminates unconstitutionally against the United States and its lessees. Their conclusion rested on the fact that Article 7173 of the Revised Civil Statutes of Texas3 imposes a distinctly lesser burden on similarly situated lessees of exempt property owned by the State and its political subdivisions. We agree with the dissenters' conclusion.

Article 7173 is the only Texas statute other than Article 5248 which authorizes a tax on lessees. It provides in part that:

'Property held under a lease for a term of three years or more, or held under a contract for the purchase thereof, belonging to this State, or that is exempt by law from taxation in the hands of the owner thereof, shall be considered for all the purposes of taxation, as the property of the person so holding the same, except as otherwise specially provided by law.'

As construed by the Texas courts, Article 7173 is less burdensome than Article 5248 in three respects. First, the measure of a tax under Article 7173 is not the full value of leased tax-exempt premises, as it apparently is under Article 5248, but only the price the taxable leasehold would bring at a fair voluntary sale for cash—the value of the leasehold itself.4 Second, by its very terms, Article 7173 imposes no tax on a lessee whose lease is for a term of less than three years. Finally, and crucial here, a lease for three years or longer but subject—like Phillips'—to termination at the lessor's option in the event of a sale is not 'a lease for a term of three years or more' for purposes of Article 7173. Trammell v. Faught, 74 Tex. 557, 12 S.W. 317. Therefore, because of the termination provisions in its lease, Phillips could not be taxed under Article 7173.

Although Article 7173 is, in terms, applicable to all lessees who hold tax-exempt property under a lease for a term of three years or more, it appears that only lessees of public property fall within this class in Texas. Tax exemptions for real property owned by private organizations—charities, churches, and similar entities—do not survive a lease to a business lessee.5 The full value of the leased property becomes taxable to the owner, and the lessee's indirect burden consequently is as heavy as the burden imposed directly on federal lessees by Article 5248. Under these circumstances, there appears to be no discrimination between the Government's lessees and lessees of private property.

However, all lessees of exempt public lands would appear to belong to the class defined by Article 7173.6 In view of the fact that lessees in this class are taxed because they use exempt property for a nonexempt purpose, they appear to be similarly situated and presumably should be taxed alike. Yet by the amendment of Article 5248, the Texas Legislature segregated federal lessees and imposed on them a heavier tax burden than is imposed on the other members of the class by Article 7173. In this case the resulting difference in tax, attendant upon the identity of Phillips' lessor, is extreme; the State and the School District concede that Phillips would not be taxed at all if its lessor were the State or one of its political subdivisions instead of the Federal Government. The discrimination against the United States and its lessee seems apparent. The question, however, is whether it can be justified.

Phillips argues that because Article 5248 applies only to private users of federal property, it is invalid for that reason, without more. For this argument, it relies on Miller v. Milwaukee, 272 U.S. 713, 47 S.Ct. 280, 71 L.Ed. 487; see also Macallen Co. v. Massachusetts, 279 U.S. 620, 49 S.Ct. 432, 73 L.Ed. 874. Macallen might be deemed to support the argument, but to the extent that it does, it no longer has precedential value. See United States v. City of Detroit, supra, 355 U.S. at page 472, note 2, 78 S.Ct. at page 477. Miller was a rather different case. In Miller it was thought that a State had attempted indirectly to levy a tax on exempt income from government bonds. Phillips' use of the Government's property, by way of contrast, is not exempt. 10 U.S.C. § 2667(e), 10 U.S.C.A. § 2667(e);7 United States v. City of Detroit, supra. It is true that in Miller the ostensible incidence of the tax—shareholders' income from corporate dividends—was not itself exempt, but the measure of the tax excluded all income not attributable to federal bonds owned by the corporation; that was the defect in the tax.

See Pacific Co. v. Johnson, 285 U.S. 480, 493, 52 S.Ct. 424, 427, 76 L.Ed. 893. Therefore, in practical operation, the tax was either an indirect tax on the exempt income, or a discriminatory tax on shareholders of corporations which as bondholders, dealt with the Government. Thus, if Miller has any relevance here, it is only to the extent that it may support the proposition that a State may not single out those who deal with the Government, in one capacity or another, for a tax burden not imposed on others similarly situated.

A determination that Article 5248 is invalid, under this test, cannot rest merely on an examination of that article. It does not operate in a vacuum. First, it is necessary to determine how other taxpayers similarly situated are treated. Such a determination requires 'an examination of the whole tax structure of the state.' Cf. Tradesmens' National Bank v. Oklahoma Tax Comm., 309 U.S. 560, 568, 60 S.Ct. 688, 693, 84 L.Ed. 947. Although Macallen may have departed somewhat from this rule, nothing in Miller, at least as it has been interpreted in later cases, should be read as indicating that less is required. Cf. Educational Films Corp. v. Ward, 282 U.S. 379, 51 S.Ct. 170, 75 L.Ed. 400; Pacific Co. v. Johnson, 285 U.S. 480, 52 S.Ct. 424.

Therefore, we must focus on the nature of the classification erected by Articles 5248 and 7173. The imposition of a heavier tax burden on lessees of federal property than is imposed on lessees of other exempt public property must be justified by significant differences between the two classes. The School District addresses this problem, essentially, as one of equal protection, and argues that we must uphold the classification, though apparently discriminatory, 'if any state of facts reasonably can be conceived that would sustain it.' Allied Stores of Ohio, Inc., v. Bowers, 358 U.S. 522, 528, 79 S.Ct. 437, 441, 3 L.Ed.2d 480. The argument, in this context, turns on three supposed differences between the two classes. First, the School District and the State say that the State can collect in rent what it loses in taxes from its own lessees—something it cannot do, of course, with the Federal Government's lessees. Second, they argue that the State may legitimately foster its own interests by adopting measures which facilitate the leasing of its property. Finally, they claim that because of its allegedly greater magnitude, federal leasing of exempt land has a more serious impact on the finances and operations of local government than does the State's own leasing activities.

None of these considerations provides solid support for the classification. It is undoubtedly true, as a general proposition, that the State is free to adopt measures reasonably designed to facilitate the leasing of its own land. But if the incentive which it provides is in the form of a reduction in tax which discriminates against the Government's lessees, the question remains, is it permissible?

Likewise, it is not enough to say that the State can make up in rent what it loses in taxes from its lessees. What the State's political subdivisions lose in taxes from the State's lessees cannot be made up in this fashion. Other local taxpayers—including the Government's lessees—must make up the difference.

Nor is the classification here supported by the allegedly serious impact of federal leasing, as contrasted with state leasing, on the operations of local government. It is claimed, in this respect, that neither the State nor its subdivisions lease property exactly comparable—in size, value, or number of employees involved—to the ordnance works leased by Phillips from the Government. However, the classification erected by Article 5248 is not based on such factors. Article 5248 imposes its burdens on all lessees of federal property. It is conceded that the State and its subdivisions lease valuable property to commercial and business enterprises, as does the Federal Government. Warehouse facilities are an example.8 But the identity of the exempt lessor bears no relation to the impact on local government of otherwise identical leasing activities. Still, the variant tax consequences to the lessee, under Article 7173 on the one hand and Article 5248 on the other, differ widely.

It is true that perfection is by no means required under the equal protection test of permissible classification. But we have made it clear, in the equal protection cases, that our decisions in that field are not necessarily controlling where problems of intergovernmental tax immunity are involved. An Allied Stores of Ohio, Inc., v. Bowers, supra, for example, we noted that the State was 'dealing with (its) proper domestic concerns, and not trenching upon the prerogatives of the National Government.' 358 U.S. at page 526, 79 S.Ct. at page 440. When such is the case, the State's power to classify is, indeed, extremely broad, and its discretion is limited only by constitutional rights and by the doctrine that a classification may not be palpably arbitrary. Id., 358 U.S. at pages 526—528, 79 S.Ct. at pages 440—441. But where taxation of the private use of the Government's property is concerned, the Government's interests must be weighed in the balance. Accordingly, it does not seem too much to require that the State treat those who deal with the Government as well as it treats those with whom it deals itself. Compare Esso Standard Oil Co. v. Evans, 345 U.S. 495, 500, 73 S.Ct. 800, 802, 97 L.Ed. 1174.

Nevertheless, it is claimed that the classification here is supported by our decision in United States v. City of Detroit, supra, because of the assertedly similar nature of the classification created by the statute involved in that case.9 The Michigan statute, although applicable generally to lessees of exempt property,10 contained an exception for property owned by state-supported educational institutions. Appellee's argument, essentially, is that the exemption of lessees of school-owned property from the Michigan statute supports the imposition here of a heavier tax on federal lessees than is imposed on lessees of other exempt public property, in general

This argument misconceives the scope of the Michigan decisions. In those cases we did not decide—in fact, we were not asked to decide—whether the exemption of school-owned property rendered the statute discriminatory. Neither the Government nor its lessees, to whom the statute was applicable, claimed discrimination of this character.11 Since the issue was not raised, the basis for the separate classification of property owned by schools was not examined. Therefore, the Michigan cases shed no light on the classification problem here.12

None of these arguments, urged in support of the Texas classification, seems adequate to justify what appears to be so substantial and transparent a discrimination against the Government and its lessees. Here, Phillips is taxed under Article 5248 on the full value of the real property which it leases from the Federal Government, while businesses with similar leases, using exempt property owned by the State and its political subdivisions, are not taxed on their leaseholds at all. The differences between the two classes, at least when the Government's interests are weighed in the balance, seem too impalpable to warrant such a gross differentiation. It follows that Article 5248, as applied in this case, discriminates unconstitutionally against the United States and its lessee. As we had occasion to state, quite recently, it still remains true, as it has from the time of M'Culloch v. Maryland, 4 Wheat. 316, 4 L.Ed. 579, that a state tax may not discriminate against the Government or those with whom it deals. See United States v. City of Detroit, supra, 355 U.S. at page 473, 78 S.Ct. at page 478. Therefore, this tax may not be exacted.

Reversed.

Mr. Justice FRANKFURTER concurs in the result.