Opinion

Recognizing that an educated citizenry and workforce are vital to the preservation of the rights and liberties of the people of this state, California in 1879 included in its new Constitution a provision directing the state Legislature to encourage “by all suitable means the promotion of intellectual, scientific, moral, and agricultural improvement.” (Cal. Const., art. IX, § 1.) Since 1879, our state Constitution has also included a provision prohibiting state and local governments from granting anything “in aid of any . . . sectarian purpose, or helping] to support or sustain any school, college, university, hospital, or other institution controlled by any . . . sectarian denomination whatever . . . .” (Cal. Const., art. XVI, § 5; see id., former art. XIII, § 24, repealed Gen. Elec. (Nov. 5, 1974).)

*793Against that backdrop, this court in California Educational Facilities Authority v. Priest (1974) 12 Cal.3d 593 [116 Cal.Rptr. 361, 526 P.2d 513] (Priest), upheld a state bond program funding the construction of educational facilities at religiously affiliated colleges, which were expressly prohibited from using the bond proceeds (paid for by private purchasers of the bonds) for specified religious purposes. We concluded that neither the state nor the federal Constitution prohibited this form of indirect assistance to religiously affiliated colleges, a rule that for more than three decades has allowed California public entities to issue revenue bonds to raise private funds for campus improvements at religiously affiliated colleges. We declined in Priest, however, to decide whether that rule would also apply if a college were “pervasively sectarian,” a term the United States Supreme Court had used in Hunt v. McNair (1973) 413 U.S. 734 [37 L.Ed.2d 923, 93 S.Ct. 2868] (Hunt) to describe a religiously affiliated school that devotes a substantial portion of its functions to its religious mission. (Priest, supra, at p. 602, fn. 8.) Our decision in Priest is pivotal here.

This case involves bond financing agreements between a public entity and three religiously affiliated schools that, for purposes of this litigation, the parties have assumed to be pervasively sectarian. These schools are thus likely to include a religious perspective in their teachings. Each agreement, as in Priest, supra, 12 Cal.3d 593, expressly prohibits use of the bond proceeds for specified religious purposes. And, as in Priest, funds for the projects will not come from any government entity, but from private-sector purchasers of the bonds, and no public entity will have any obligation on the bonds in the event of default by the schools.

The trial court invalidated the agreements as violating the state Constitution’s article XVI, section 5. The Court of Appeal, in a two-to-one decision, upheld the trial court; the dissent would have validated the agreements.

As explained below, in resolving the state constitutional issue we conclude that the pertinent inquiry should center on the substance of the education provided by these three schools, not on their religious character. Therefore, whether the schools are pervasively sectarian (as the parties have assumed) is not a controlling factor in determining the validity of the bond funding program under our state Constitution. Rather, the program’s validity turns on two questions: (1) Does each of the recipient schools offer a broad curriculum in secular subjects? (2) Do the schools’ secular classes consist of information and coursework that is neutral with respect to religion? This test ensures that the state’s interest in promoting the intellectual improvement of its residents is advanced through the teaching of secular information and coursework, and that the expression of a religious viewpoint in otherwise secular classes will *794provide a benefit to religion that is merely incidental to the bond program’s primary purpose of promoting secular education.

Finally, we conclude that a public bond program satisfying our state Constitution would not violate the establishment clause of the First Amendment to the federal Constitution.

I

A. The Nature of This Action

Government Code section 6502 provides that “two or more public agencies by agreement may jointly exercise any power common to the contracting parties,” thus allowing for the creation of so-called joint powers authorities.1 A joint powers authority can issue tax-exempt revenue bonds to finance construction projects that provide a public benefit and are located within the geographical boundaries of its member public agencies. (§ 6588 et seq.)

As relevant here, some 350 California cities, counties, and special districts have entered into agreements to create a joint powers authority, known as the California Statewide Communities Development Authority (the Authority), plaintiff in this case. By issuing tax-exempt revenue bonds to finance industrial projects and residential units, as well as health care and educational facilities, the Authority promotes economic development for the benefit of its members. The Authority does not fund these projects or otherwise provide any financial subsidy in connection with the issuance of the bonds. It is involved in the financing transaction solely to provide a tax exemption to the private investors who purchase the bonds and thereby fund the private development. Because the government merely provides access to favorable tax treatment and does not itself finance the projects, this form of financing is commonly referred to as “pass through” or “conduit” financing; the government’s issuance of the bonds provides a “conduit” for private financing to “pass through” to the recipient of the bond proceeds. (See Note, Revenue Bonds and Religious Education: The Constitutionality of Conduit Financing Involving Pervasively Sectarian Institutions (2002) 100 Mich. L.Rev. 1108, 1111 (Revenue Bonds and Religious Education).) No public monies are expended in this type of arrangement, as the recipient of the bond proceeds bears responsibility for payments of principal and interest to the private bond purchaser, which has no recourse against the government. (Id. at pp. 1146, 1149-1150; see § 91535.) In addition, the recipient also reimburses the public entity for the costs of issuing the bonds. (Revenue Bonds and Religious Education, supra, at p. 1150.)

*795Under the statutory scheme, the Authority may issue tax-exempt bonds “whenever there are significant public benefits for taking that action.” (§ 6586.)2 In the case of educational institutions, the Authority requires that the beneficiary school be exempt from taxation under section 501(c)(3) of the Internal Revenue Code (26 U.S.C. § 501(c)(3)).3 The tax-exempt status of the bonds, however, does not flow from the tax-exempt status of the schools. Rather, the income that bondholders derive from bonds that the Authority has issued is exempt from taxation under Government Code section 6575 and section 103 of the Internal Revenue Code (26 U.S.C. § 103(a)), which exempt from taxation the income earned on state and local bonds. Thus, the program here is simply a mechanism by which the government extends favorable tax treatment to private individuals to encourage private financial support of development that will provide a public benefit to the community. The program encourages private support of certain activities and programs by way of a tax policy, just as income deductions provided in connection with private donations to tax-exempt organizations encourage private support of certain activities and organizations.

In May and July 2002, the Authority adopted resolutions approving agreements to issue revenue bonds to fund campus improvements at three private schools (Oaks Christian School, California Baptist University, and Azusa Pacific University), all operated by tax-exempt religious corporations. As to each school, the Authority found that the planned projects would produce one or more of the significant public benefits set out in section 6586. Thereafter, the Authority filed these validation actions, seeking superior court approval of the agreements. For purposes of these lawsuits, and apparently to create test cases, the Authority assumed, without conceding, that each of the three schools was “pervasively sectarian,” as the United States Supreme Court used that term in Hunt, supra, 413 U.S. at page 743, meaning that “a substantial portion of [each school’s] functions are subsumed” in its religious mission.

Each of the three lawsuits is a validation action, a form of “proceeding in rem” brought against “all persons interested” in a specified matter. (Code Civ. Proc., §§ 860, 861.1.) In a validation action, the party seeking court approval *796must publish notice of the lawsuit in a newspaper of general circulation and set a date for anyone “interested” to “appear and contest the legality or validity of the matter sought to be determined.” (Id., § 862.) The Authority did so. When no interested party responded, the Authority moved for default judgments in the three actions.

The trial court, on its own initiative, continued the matters to solicit the views of the California Department of Fair Employment and Housing as well as the California Attorney General. The former declined to participate. The latter responded by letter brief, expressing no view on the Authority’s entitlement to the default judgments. Instead, the Attorney General merely pointed out possible state and federal constitutional issues raised by the proposed bond funding agreements. The trial court found the agreements invalid under article XVI, section 5 of the California Constitution and under the establishment clause of the First Amendment to the federal Constitution. But in its order denying the Authority’s request for validation of the agreements with the three schools, the trial court relied exclusively on the state constitutional ground.

The Authority appealed. The Court of Appeal consolidated the three cases, and, in a two-to-one decision, affirmed the trial court. Relying solely on the California Constitution, as the trial court had done in its order denying validation of the bond funding agreements, the Court of Appeal held that issuing revenue bonds to raise capital in the private sector for improvements at pervasively sectarian schools would violate article XVI, section 5. But the dissenting justice concluded to the contrary. And, unlike the majority, the dissent went on to address the federal constitutional issue, concluding the bond funding agreements did not violate the establishment clause of the federal Constitution’s First Amendment.

The Authority sought review in this court. It contended that the Court of Appeal’s decision “casts doubt on the eligibility of all sectarian schools to obtain government services” and threatens to impede “educational opportunities for hundreds of thousands of students in California.” We granted review, and thereafter we granted the Attorney General’s motion to intervene as respondent in the case.

B. The Financing Agreements

Each of the three financing agreements provides for the appointment of an independent trustee to receive and handle the bond proceeds.4 Each school *797authorizes the trustee to disburse funds to pay for the facilities’ construction, and the school then pays the principal plus interest on the bonds to the trustee, which in turn pays the private purchasers of the government bonds.

Repaying the private-sector bond purchasers for the bond proceeds falls to the recipient schools and will “not constitute a charge against the general credit of [the Authority]” or be secured by any public property. The bond purchasers have no recourse for nonpayment against the State of California or any public agency. All costs incurred in the course of the bond issuance are reimbursed by the schools.

Each agreement includes a covenant by the recipient school that “no facility, place or building financed or refinanced with a portion of the proceeds of the Bonds will be used for . . . sectarian instruction or as a place for religious worship or in connection with any part of the programs of any school or department of divinity for the useful life of the Project.” And each agreement grants the trustee a right of access to ensure each school’s compliance with the covenant. The costs of these trustee inspections are to be paid from the trust funds (that is, from the bond proceeds).

In support of the validation actions it had filed in the superior court, the Authority submitted declarations describing each of the three schools, all located in Southern California. The Authority’s materials comprise the only evidence presented to the trial court. A summary follows.

Oaks Christian School, in the City of Westlake Village, provides a private, comprehensive, college preparatory, Christian education for students in the sixth through the 12th grades. The school intends to use the bond proceeds to build administrative offices, classrooms, and athletic facilities. The school seeks to foster student growth “in knowledge and wisdom through God’s grace,” and to encourage students “to dedicate [themselves] to the pursuit of academic excellence, athletic distinction and Christian values.” It admits students without regard to religion, and in the 2001-2002 school year about 35 percent of those enrolled were not Christian. Students and parents must agree, however, to support the school’s mission, statement of faith, and “biblical goals.” Faculty members must adhere to the Christian faith and must sign a statement of faith.

California Baptist University, located in Riverside, is an accredited Christian liberal arts institution that grants undergraduate and graduate degrees in a wide range of disciplines. The bond proceeds will fund the acquisition, construction, and improvement of classrooms, administrative offices, a student center, residence halls, and athletic facilities. The university’s undergraduate students are expected to accept and to live by “biblically based *798Christian principles,” and to attend services at a church of their choosing. In the 2000-2001 academic year, the school awarded some 450 student degrees, of which only about 5 percent were to students who had majored in Christian or ministry studies. The university requires its faculty members to adhere to a Christian faith, and it reserves 51 percent of faculty positions for members of the Baptist denomination. Faculty members must agree with the school’s theological and philosophical views but need not sign any statement of faith.

Azusa Pacific University, whose main campus is in Azusa, is an accredited Christian liberal arts school offering both undergraduate and graduate degrees. It intends to use the bond proceeds to improve and build campus facilities, “including but not limited to a residence facility, a dining facility, [and] a mail center.” The declaration of a campus administrator at Azusa Pacific University describes the school as “an evangelical Christian community of disciples and scholars who seek to advance the work of God in the world through academic excellence in liberal arts and professional programs of higher education that encourage students to develop a Christian perspective of truth and life.” Student applicants are expected to “exhibit moral character in harmony” with the University’s purpose. Applicants are chosen for their academic ability as well as their church involvement and participation in school and community activities. Undergraduates must complete “120 hours of student ministry assignments.” In the 2001-2002 academic year, less than 7 percent (39 of more than 600) of the undergraduate degrees awarded by Azusa Pacific University went to those who had completed majors in religious studies. All faculty members are Christian.

II

At issue is whether the Authority’s conduit financing agreements with Oaks Christian School, California Baptist University, and Azusa Pacific University violate our state Constitution’s article XVI, section 5. That provision states: “Neither the Legislature, nor any county, city and county, township, school district, or other municipal corporation, shall ever make an appropriation, or pay from any public fund whatever, or grant anything to or in aid of any religious sect, church, creed, or sectarian purpose, or help to support or sustain any school, college, university, hospital, or other institution controlled by any religious creed, church, or sectarian denomination whatever, nor shall any grant or donation of personal property or real estate ever be made by the State, or any city, city and county, town, or other municipal corporation for any religious creed, church, or sectarian purpose whatever; provided, that nothing in this section shall prevent the Legislature granting aid pursuant to Section 3 of Article XVI [allowing the use of state funds for the construction of hospitals and other charitable institutions when federal funds have been authorized for that purpose].” (Ibid., italics added.)

*799Under the bond funding agreements, the Authority, which is a municipal corporation, does not provide public money to the three private schools in question. Thus, there is no violation of the prohibition in the state Constitution’s article XVI, section 5, against making “an appropriation, or pay[ing] from any public fund” for the benefit of “any school, college, [or] university . . . controlled by any religious creed, church, or sectarian denomination.” Nor is there a violation of that provision’s prohibition against making “any grant or donation of personal property or real estate ... for any religious creed, church, or sectarian purpose.” (Ibid..)

At issue is whether the Authority’s proposed indirect assistance to the three schools, through its issuance of revenue bonds, would be “aid of any . . . sectarian purpose” or “help to support any school . . . controlled by any . . . sectarian denomination,” as prohibited by section 5 of article XVI of the state Constitution. On point here is Priest, supra, 12 Cal.3d 593, a unanimous decision authored in 1974 by Justice Stanley Mosk of this court. Priest held that issuing revenue bonds to fund capital improvements at religiously affiliated colleges did not violate former section 24 of article XIII, the identically phrased predecessor of the constitutional provision involved here.

Priest concerned the California Educational Facilities Authority Act (hereafter sometimes Act) (Ed. Code, former § 30301 et seq.), which is not involved here. That former statutory scheme authorized public entities to issue revenue bonds to assist private colleges in building classrooms and other campus facilities, but it prohibited the use of the monies to construct any facility for “ ‘sectarian instruction or as a place for religious worship or ... in connection with any part of the program of a school or department of divinity.’ ” (Priest, supra, 12 Cal.3d at p. 596.) In Priest, the school that was the subject of the financing agreement, University of the Pacific (UOP), had no religious affiliation.5 Nevertheless, this court addressed contentions by the State Treasurer (who was the respondent in the writ of mandate action) that the Act violated the religion clauses of both the state and federal Constitutions; we did so to avoid having “a cloud . . . remain over the Act,” because several other colleges with “denominational ties” had sought to participate in the state bond funding program. (Priest, supra, at p. 598, fn. 5.)

*800Of former article XIII, section 24 (the predecessor of current art. XVI, § 5 of the Cal. Const, at issue here), Priest stated:

“This section has been said to constitute ‘the definitive statement of the principle of government impartiality in the field of religion.’ (37 Ops.Cal.Atty.Gen. 105, 107 (1961).) An examination of the debates of the constitutional convention which drafted the Constitution of 1879 indicates that the provision was intended to insure the separation of church and state and to guarantee that the power, authority, and financial resources of the government shall never be devoted to the advancement or support of religious or sectarian purposes. . . .

“The section has never been interpreted, however, to require governmental hostility to religion, nor to prohibit a religious institution from receiving an indirect, remote, and incidental benefit from a statute which has a secular primary purpose. . . .

“The Act here challenged clearly provides a ‘benefit’ in that it enables sectarian institutions to borrow money through the use of a state instrumentality at a cost below that of the marketplace. Thus the crucial question is not whether the Act provides such a benefit, but whether that benefit is incidental to a primary public purpose. The framers of the [state] Constitution recognized the importance of education in our social fabric, and imposed a constitutional duty on the Legislature to ‘encourage by all suitable means the promotion of intellectual. . . improvement.’ (Art. IX, § 1.) The present law is responsive to that mandate. The Legislature has expressly determined that the Act, in supporting the maintenance and improvement of facilities for higher education, is in the public interest [citations], and that determination is entitled to great deference. [Citations.] The benefits of the Act are granted to sectarian and nonsectarian colleges on an equal basis; in both cases all aid for religious projects is strictly prohibited; and in no event is a financial burden imposed upon the state. In these circumstances the Act does not have a substantial effect of supporting religious activities. Rather, its primary purpose is to advance legitimate public ends, and it therefore does not violate article XIII, section 24.” (Priest, supra, 12 Cal.3d at pp. 604-606, italics added, fn. omitted.)6

*801From those statements in Priest we distill the following four-part test for determining whether the issuance of government bonds benefiting a religiously affiliated school violates the state constitutional provision in question: (1) The bond program must serve the public interest and provide no more than an incidental benefit to religion; (2) the program must be available to both secular and sectarian institutions on an equal basis; (3) the program must prohibit use of bond proceeds for “religious projects”; and (4) the program must not impose any financial burden on the government. (Priest, supra, 12 Cal.3d at pp. 605-606.) Because the last three requirements can be easily disposed of in this case, we address them first.

It is undisputed that the Authority has issued tax-exempt bonds to encourage private investment in a wide variety of private institutions, secular as well as sectarian. This satisfies Priest's second requirement—that the state bond program not discriminate between secular and sectarian institutions, treating both categories alike. (Priest, supra, 12 Cal.3d at p. 606.)

The bond agreements expressly prohibit each of the three schools from using the bond proceeds to construct or improve any facility for “religious projects,” that is “ ‘sectarian instruction or as a place for religious worship or in connection with any part of the programs, of any school or department of divinity for the useful life of the Project.’ ” (See p. 797, ante.) This satisfies Priest's third requirement. (Priest, supra, 12 Cal.3d at p. 606.)

Because of the utilization of conduit or passthrough financing, the capital for the construction projects at the three private schools is funded solely by private-sector purchasers of the bonds. The schools repay the advanced capital plus interest to an independent trustee, who then pays the private bondholders, who have no recourse for nonpayment against the Authority. All of the Authority’s costs of issuing the bonds are reimbursed by the schools. Thus, the bond funding places no financial burden on the Authority or any other public entity. This satisfies Priest's fourth requirement, that the program not impose a financial burden on the government. (Priest, supra, 12 Cal.3d at p. 606.)7

*802Having concluded that the second, third, and fourth requirements of Priest are satisfied, we still need to determine whether the bond program meets Priest's first requirement, that the program provide a public benefit and no more than incidentally benefit religion. Priest held that the state’s issuance of revenue bonds for purchase by private investors to fund construction or improvements of facilities at religiously affiliated colleges benefited the public at large by “ ‘encouraging] by all suitable means the promotion of intellectual . . . improvement’ ” in California, thus furthering the state constitutional mandate of article IX, section 1 (Priest, supra, 12 Cal.3d at p. 605), and that it “[did] not have a substantial effect of supporting religious activities” (id. at p. 606).

As mentioned earlier at pages 793 and 799, our decision in Priest concerned state bond financing to construct or improve facilities at schools that were “religiously affiliated”—schools connected to a religious organization—but not “pervasively sectarian”—schools “ ‘in which religion is so pervasive that a substantial portion of its functions are subsumed in the religious mission’ ” (Priest, supra, 12 Cal.3d at p. 601, quoting Hunt, supra, 413 U.S. at p. 743). Because the schools in Priest were not pervasively sectarian, Priest did not question that the religiously affiliated schools were providing education in secular subjects, thus furthering the state’s interest in promoting the intellectual improvement of its residents though secular education, and no more than incidentally benefiting religion.

Here, the proposed state bond program would benefit three schools that the Authority throughout this litigation has described as “pervasively sectarian.” Would that bond program satisfy Priest's first requirement, that the program serve the public interest and no more than incidentally benefit religion? To answer that question, we need to examine the bond program’s purpose and effects. The program must have a “primary purpose ... to advance legitimate public ends” (Priest, supra, 12 Cal.3d at p. 606), and its effects may not include a benefit to religious activity that is other than “indirect, remote, [or] incidental” to the primary secular purpose {id. at p. 605). When a court attempts to determine how, if at all, a bond program would have the effect of supporting religious activity at schools, the characterization of the schools as “pervasively sectarian” does not provide a reliable or satisfactory answer. A *803more useful and effective approach, we conclude, is to examine the substance of the education that each of these religious schools offers its students, as explained below.8

The goal of the Authority’s proposed issuance of revenue bonds here is to enhance the ability of private schools to improve their facilities through funding provided exclusively by private investors. By providing a tax exemption on interest earned on the bonds, the government makes the bonds more attractive to private investors and thereby enhances the ability of private institutions to expand their educational facilities. We pointed out in Priest: “The framers of [our state] Constitution recognized the importance of education in our social fabric, and imposed a constitutional duty on the Legislature to ‘encourage by all suitable means the promotion of intellectual . . . improvement.’ (Art. IX, § 1.)” (Priest, supra, 12 Cal.3d at p. 605.) In the circumstances of that case—a program available to sectarian and nonsectarian schools on an equal basis, in which all aid for religious projects was prohibited, and no financial burden was imposed upon the state—we concluded that the provision of tax-exempt bond financing did not violate article XVI, section 5, because “its primary purpose [was] to advance legitimate public ends” and it “d[id] not have a substantial effect of supporting religious activities.” {Priest, supra, at p. 606.) Can the same be said of the proposed bond funding in this case with respect to a school that includes a religious perspective in its curriculum? The answer is yes, if certain requirements are met.

First, the school that is the subject of the revenue bond financing arrangement must provide a broad curriculum in secular subjects. When it does, the bond program assists the religious school in providing educational opportunities to California residents, enhancing their employment prospects and deepening their understanding of critical political, social, scientific, and cultural issues. This broad curriculum requirement excludes from the bond funding program religious schools that offer classes in only a few secular subjects, because to provide bond funding for such schools would not sufficiently advance the program’s goal of expanding secular educational opportunities for Californians.

We are mindful of the concern that a school with a religious perspective may use the facilities built or improved with the revenue bond proceeds to substantially further its religious mission. Such use would *804provide more than an incidental benefit to religion, in violation of the principles we enunciated in Priest, supra, 12 Cal.3d 593. To ensure that the classes in secular subjects promote the state’s interest in secular education and no more than incidentally benefit religion, the religious school must meet a second requirement: the information and coursework used to teach secular subjects must be neutral with respect to religion. Of course, religion may be an object of study in classes such as history, social studies, and literature, just as in public schools, in a manner that neither promotes nor opposes any particular religion or religion in general. But a class that includes as part of the instruction information or coursework that promotes or opposes a particular religion or religious beliefs may not be taught in facilities financed through tax-exempt bond financing.9 On remand, in determining religious neutrality, the straightforward assessment for the trial court to make is whether the academic content of a religious school’s course in a secular subject such as math, chemistry, or Shakespeare’s writings is typical of that provided in nonreligious schools. When a school establishes, through its course descriptions or otherwise, that the academic content of its secular classes is typical of comparable courses at public or other nonreligious schools, it is not necessary to scrutinize the school’s day-to-day classroom communications. The circumstance that a teacher may, in addition to teaching a course’s religiously neutral content, express an idea or viewpoint that may be characterized as “religious” does not result in a benefit to religion that is *805more than incidental to the state’s primary purpose of enhancing secular education opportunities for California residents.

As we stated in Priest in addressing former section 24 of article XIII (the identically phrased predecessor of current § 5 of art. XVI of the Cal. Const.), this “section has never been interpreted ... to require governmental hostility to religion, nor to prohibit a religious institution from receiving an indirect, remote, and incidental benefit from a statute which has a secular primary purpose.” (Priest, supra, 12 Cal.3d at p. 605.) Priest also observed: “ ‘[M]any expenditures of public money give indirect and incidental benefit to denominational schools and institutions of higher learning. Sidewalks, streets, roads, highways, sewers are furnished for the use of all citizens regardless of religious belief. . . . Police and fire departments give the same protection to denominational institutions that they give to privately owned property and their expenses are paid from public funds.’ ” (Ibid.) Here, we have no expenditure of public money, and application of the standards we have set out will ensure that any benefit to religion from the bond funding program is merely incidental.

As we have pointed out, tax-exempt bond financing is a mechanism by which the government makes available to private investors a tax exemption on income earned on government bonds, thereby encouraging private investment in community development and enabling the recipients of these investments to borrow private funds at a lower interest rate. As the Virginia Supreme Court has observed, in concluding that issuing government bonds to benefit a religious college (the Reverend Pat Robertson’s Regent University, excluding its divinity school) did not violate the establishment clause of the federal Constitution’s First Amendment: “The nature of this aid is properly defined as the granting of tax-exempt status to the bonds which has the incidental result of permitting a qualifying institution to borrow funds at an interest rate lower than conventional private financing.” (Virginia College Bldg. Authority v. Lynn (2000) 260 Va. 608, 638 [538 S.E.2d 682, 698] (Virginia College).)

More recently, in 2002, the federal Court of Appeals for the Sixth Circuit in Steele v. Industrial Development Bd. of Metro. (6th Cir. 2002) 301 F.3d 401 (Steele), rejected an establishment clause challenge to a bond financing program benefiting Lipscomb University, a Tennessee college run by the Churches of Christ. Steele viewed the bond funding program as a form of tax policy, equating the issuance of revenue bonds to tax exemptions benefiting religious organizations that the high court has held do not violate the establishment clause. (See Hernandez v. Commissioner (1989) 490 U.S. 680, 688 [104 L.Ed.2d 766, 109 S.Ct. 2136] [deductions from taxable income for gifts to religious organizations]; Mueller v. Allen (1983) 463 U.S. 388, 396 *806[77 L.Ed.2d 721, 103 S.Ct. 3062] [state tax deductions for amounts paid for school tuition and textbooks even though such deductions substantially benefited religious schools]; Walz v. Tax Commission (1970) 397 U.S. 664, 675 [25 L.Ed.2d 697, 90 S.Ct. 1409] [city tax exemption for real estate owned by religious organizations].) In neither situation, the federal appellate court in Steele stressed, did the government transfer any public funds to the recipient institutions. (Steele, supra, at p. 409.) Steele concluded that the benefit to the recipient school from the issuance of the tax-exempt bonds was analogous “to an indirect financial benefit conferred by a religiously neutral tax or deduction.” (Id. at p. 413.) Steele explained: “The purchaser of a bond has recourse for repayment against Lipscomb University only,” and not against the issuing governmental entity. (Ibid.) Also: “No government funds are involved in the entire transaction. The interest paid to the bond holders by Lipscomb University is not subject to federal, state, or local income taxes. Since the bonds are tax-exempt, Lipscomb University reaps the benefit of a lower interest rate than that paid to a lender paying income taxes on the interest received. Only by the potential loss of tax revenue does the conduit financing involve any impact on public funds.” (Ibid.)10

The trial court here, in denying the Authority’s request to validate its bond funding agreements with the three religious schools, did not consider the substance of the education at those schools. Accordingly, we remand this case to the Court of Appeal, which in turn is to remand the matter to the trial court for that evaluation.

The trial court is to determine whether each religious school benefiting from the state bond program offers a sufficiently broad curriculum in secular subjects—e.g., English literature, history, math, sciences, professional or preprofessional training—that the school’s use of the educational facilities built or improved with the bond funding may be expected to promote the public interest in making secular education more available to California residents in general. The trial court’s inquiry should center on the school’s curriculum as a whole, but it should exclude theological or divinity programs because, under the terms of their agreements, the schools may not use the *807facilities built or improved with the state bond proceeds for those programs. If the school does offer this sort of broad secular curriculum, the trial court should consider whether the academic content of the classes in secular subjects is typical of such classes when taught in nonreligious schools and is thus neutral with respect to religion. In resolving this issue, the court may consider the school’s course descriptions or any other information submitted to establish the content or coursework of the secular classes. The circumstance that a religious viewpoint may also be expressed in these otherwise secular classes does not preclude a determination that providing the proposed tax-exempt bond financing to the school promotes the state’s interest in the intellectual development of its residents and only incidentally benefits religion.11

Having concluded that the proposed state bond funding, as discussed on pages 803 to 805, ante, would not violate our state Constitution if certain conditions are satisfied (a determination to be made by the trial court on remand), we now consider whether the proposed bond funding, on the conditions we have articulated, would pass muster under the establishment clause of the First Amendment to the federal Constitution. We discuss this issue below.

III

The First Amendment to the federal Constitution states that “Congress shall make no law respecting an establishment of religion . . . .” (Italics added.) This provision is incorporated in the due process clause of the Fourteenth Amendment, thus making it applicable not only to Congress but also to the states. (See Everson v. Board of Education (1947) 330 U.S. 1, 8 [91 L.Ed. 711, 67 S.Ct. 504].)

In Lemon v. Kurtzman (1971) 403 U.S. 602 [29 L.Ed.2d 745, 91 S.Ct. 2105] (Lemon), the United States Supreme Court adopted a three-part test to determine when a particular law or government practice would not constitute an establishment of religion: (1) The government program must have “a secular legislative purpose”; (2) the program’s “principal or primary effect must be one that neither advances nor inhibits religion”; and (3) the program “must not foster ‘an excessive government entanglement with religion.’ ” (Id. at pp. 612-613.) Failure to satisfy any one of Lemon’s three requirements *808would render a governmental program unconstitutional. (See id. at pp. 619-620, 622.) In Agostini v. Felton (1997) 521 U.S. 203 [138 L.Ed.2d 391, 117 S.Ct. 1997] (Agostini), the high court refined that three-part test.

Addressing the Lemon test’s third criterion—excessive entanglement between church and state—Agostini rejected the idea that “entanglement” should be treated as a separate inquiry. Rather, Agostini stated, the entanglement inquiry was an aspect of Lemon’s second inquiry, whether the government aid at issue has the impermissible “effect” of advancing religion. (Agostini, supra, 521 U.S. at pp. 232-233.) In determining such effect, Agostini explained, the pertinent inquiry is whether the government aid program “result[s] in governmental indoctrination; define[s] its recipients by reference to religion; or create[s] an excessive entanglement [between church and state].” (Id. at p. 234.) By folding Lemon’s “entanglement inquiry into the primary effect inquiry” (Zelman v. Simmons-Harris (2002) 536 U.S. 639, 668 [153 L.Ed.2d 604, 122 S.Ct. 2460]), the high court in Agostini has collapsed Lemon’s three-part test into just two parts.

The high court has acknowledged that it does not apply the Lemon test in every establishment clause case (see Van Orden v. Perry (2005) 545 U.S. 677, 686 [162 L.Ed.2d 607, 125 S.Ct. 2854, 2861] (plur. opn. of Rehnquist, C. J.) [“[m]any of our recent cases simply have not applied the Lemon test”]; id. at p. 699 [125 S.Ct. at p. 2868] (conc. opn. of Breyer, J.) [“the Court has found no single mechanical formula that can accurately draw the constitutional line in every case”]); and some members of the court have expressed disagreement with that test (e.g., Lamb’s Chapel v. Center Moriches Union Free School Dist. (1993) 508 U.S. 384, 398-400 [124 L.Ed.2d 352, 113 S.Ct. 2141] (conc. opn. of Scalia, J., joined by Thomas, J.)). But the high court has resorted to the Lemon test in dealing with issues involving government aid to sectarian educational institutions. (E.g., Mitchell v. Helms (2000) 530 U.S. 793, 807-808 [147 L.Ed.2d 660, 120 S.Ct. 2530] (plur. opn. of Thomas, J.) (Mitchell); Agostini, supra, 521 U.S. at pp. 222-223, 232-233; Hunt, supra, 413 U.S. 734.) Because Hunt is the case most closely on point here, we summarize it below.

At issue in Hunt was whether South Carolina’s issuance of revenue bonds benefiting the Baptist College of Charleston, a religiously affiliated institution, constituted an impermissible establishment of religion in violation of the First Amendment. Hunt was decided in 1973, just two years after Lemon and 24 years before Agostini (which collapsed Lemon’s three-part test into just two parts). Naturally, therefore, the high court in Hunt applied the original three-part Lemon test. Hunt held that the bond funding satisfied the first Lemon requirement because its purpose was “manifestly a secular one,” providing a funding mechanism for all South Carolina institutions of higher *809education “whether or not having a religious affiliation.” (Hunt, supra, 413 U.S. at p. 741.) And the program produced no “unconstitutional degree of entanglement between the State and the College,” thus satisfying Lemon’s third requirement, even though the agreement between the bond funding agency and the college allowed inspection of “the project to insure that it [was] not being used for religious purposes.” (Hunt, supra, at pp. 745-746.)

Most pertinent here, however, is the high court’s discussion in Hunt of Lemon’s second requirement, that the bond funding not have the primary result of advancing religion. Concluding that the second Lemon requirement was met with respect to the Baptist College of Charleston, the court observed that it had “no religious qualifications for faculty membership or student admission,” and that in its student body Baptists comprised only some 60 percent, “a percentage roughly equivalent to the percentage of Baptists in that area of South Carolina.” (Hunt, supra, 413 U.S. at pp. 743-744.) The court cautioned, however, that a similar bond program might have the impermissible effect of advancing religion if the proceeds “flow[] to an institution in which religion is so pervasive that a substantial portion of its functions are subsumed in the religious mission or when it funds a specifically religious activity in an otherwise substantially secular setting.” (Id. at p. 743.)

The high court in Hunt, supra, 413 U.S. 734, left open whether the state bond program in that case would have violated the establishment clause of the First Amendment if the college had been pervasively sectarian. To this date, the high court has yet to address that issue. (See Revenue Bonds and Religious Education, supra, 100 Mich. L.Rev. 1108.)

Certain observations in the high court’s more recent establishment clause cases suggest that even if state bond funding were to benefit a pervasively sectarian school, the program might still survive scrutiny under the federal Constitution. In Mitchell, supra, 530 U.S. 793, the high court upheld a federal aid program that provided library and media materials, as well as computer software, to private elementary and secondary schools, including religiously affiliated schools. Four of the court’s nine members who comprised the Mitchell plurality unequivocally stated: “[N]othing in the Establishment Clause requires the exclusion of pervasively sectarian schools from otherwise permissible aid programs, and other doctrines of this Court bar it.” (Id. at p. 829 (plur. opn. of Thomas, J., joined by Rehnquist, C. J., Kennedy, J., and Scalia, J.).) Moreover, the two justices who in a separate concurrence became part of the Mitchell majority did not consider whether the recipients of government aid in that case were pervasively sectarian. (Mitchell, supra, at pp. 836-867 (conc. opn. of O’Connor, J., joined by Breyer, J.).)

Furthermore, since 1985 the United States Supreme Court has not invalidated any government aid program on the ground that the recipients were *810pervasively sectarian. (See Mitchell, supra, 530 U.S. at p. 826.) In Agostini, supra, 521 U.S. 203, the high court repudiated the analysis of its two most recent cases that had done so—Aguilar v. Felton (1985) 473 U.S. 402 [87 L.Ed.2d 290, 105 S.Ct. 3232]; and Grand Rapids School District v. Ball (1985) 473 U.S. 373 [87 L.Ed.2d 267, 105 S.Ct. 3216], The Agostini court overruled Aguilar completely and overruled Ball in part. (Agostini, supra, at pp. 218, 225; see Mitchell, supra, at p. 826.) The high court’s overruling of Aguilar has been said to signal its “intent to weaken the Lemon test in favor of a more neutral stance toward sectarian education.” (Note, Educational Vouchers and the Religion-Clauses Under Agostini: Resurrection, Insurrection and a New Direction (1999) 49 Case W. Res. L.Rev. 747, 755.) It has also been observed that “[t]he case law on funding religious institutions has changed dramatically over the last twenty years” (Laycock, Theology Scholarships, the Pledge of Allegiance, and Religious Liberty: Avoiding the Extremes But Missing the Liberty (2004) 118 Harv. L.Rev. 155, 162), that “[f]ederal constitutional restrictions on funding religious institutions have collapsed” (id. at p. 156), and that decisions of the United States Supreme Court restricting government aid to sectarian schools “are confined to a remarkably brief period, from 1971 to 1985” (id. at p. 167).

Here, we need not decide whether extending the benefit of a government bond funding program to a pervasively sectarian school could ever violate the First Amendment’s establishment clause. As ¿xplained in part II, ante, we have concluded that, to avoid violating our state Constitution, a religious school participating in a conduit financing bond program must offer a broad curriculum in secular subjects, comprised of classes whose academic content must be neutral with respect to religion, and it must not use the facilities built or improved with the state bond proceeds for theological or divinity programs or as a place of worship. Accordingly, the issue we address here is whether state bond funding for a religious school, under those conditions, would violate the federal Constitution’s establishment clause.

As we have discussed in greater detail on pages 808 and 809, ante, in 1973 the high court in Hunt, supra, 413 U.S. 734, upheld a South Carolina revenue bond program benefiting a religiously affiliated college. In reaching that decision, Hunt applied the three-part test of Lemon, supra, 403 U.S. 602, which provides courts with a means of determining the constitutionality of government programs that aid religious institutions. We too apply the Lemon test here, but we do so using the test as refined in the high court’s 1997 decision in Agostini, supra, 521 U.S. at pages 232-233, which reduced Lemon's test to just two parts.

The first of these requirements is that the government program have a secular purpose. As we explained on pages 803, 804 and 805, ante, to *811comport with our state Constitution a religious school receiving the bond funding for building or improving its educational facilities must offer a broad curriculum in secular subjects, in which the academic content is typical of that at nonreligious schools and thus is religiously neutral. As we stated, under those conditions, the program serves the primary secular purpose of increasing secular educational opportunities for Californians in general. (See p. 803, ante.)

The second, and final, inquiry under Lemon, as refined by the high court in Agostini, is whether the government program will have the impermissible effect of advancing religion. Pertinent to this inquiry is whether the program will “result in governmental indoctrination; define its recipients by reference to religion; or create an excessive entanglement” between the government and the religiously affiliated beneficiaries of the program. (Agostini, supra, 521 U.S. at p. 234.) We first address governmental indoctrination below.

Government indoctrination of religion occurs when “any use of [the government] aid to indoctrinate religion [can] be attributed to the State.” (Agostini, supra, 521 U.S. at p. 230.) Here, the Authority’s issuance of revenue bonds to finance campus improvements at the three religious schools will not result in governmental indoctrination. As we have explained, a school cannot qualify for the bond funding unless it offers a broad curriculum in secular subjects and the academic content of its classes in secular subjects is similar to that offered in nonreligious schools. (See pp. 803-805, ante.) The mere incidental expression of a religious viewpoint while teaching a secular class does not constitute religious indoctrination. Moreover, the Authority’s financing agreements with the three schools expressly prohibit them from using the bond proceeds to build or improve facilities for “ ‘sectarian instruction or as a place for religious worship or in connection with any part of the programs of any school or department of divinity.’ ” (See p. 797, ante.)

We now consider whether the Authority here defines the recipients of its revenue bond funding “by reference to religion.” (Agostini, supra, 521 U.S. at p. 234.) As we have explained on page 794, ante, the Authority is by statute a joint powers authority, which is comprised of some 350 cities, counties and other public entities, and which has since its creation issued tax-exempt bonds to promote economic development within its members’ boundaries, based on principles of neutrality, without taking into account the recipient’s religious affiliation, if any.

Nor will the state bond funding here “create an excessive entanglement” (Agostini, supra, 521 U.S. at p. 234) between the state and the religious schools in question. As mentioned earlier, the Authority’s agreements with the three schools strictly prohibit them from using the bond-funded educational facilities for religious purposes, and we have conditioned *812a school’s participation in the bond program on its offering a broad secular curriculum comprised of classes with academic content similar to or typical of that provided in nonreligious schools. In the absence of contrary evidence, courts presume that recipients of government aid will comply with restrictions imposed. (Agostini, supra, 521 U.S. at pp. 223-224, 226-227; accord, Mitchell, supra, 530 U.S. at p. 847 (conc. opn. of O’Connor, J.).)

In addition, the state bond funding agreements here provide for the appointment of an independent trustee, with the right of access to the three schools to ascertain their compliance with the restrictions imposed. The trustee’s right of access could be exercised by occasional unannounced visits. We note that even when the monitoring visits are by the government entity, rather than as here by a trustee independent of the public entity issuing the bonds, such visits have been held insufficient to create an excessive entanglement between the monitoring agency and the sectarian schools. (See Agostini, supra, 521 U.S. at p. 233; see also Bowen v. Kendrick (1988) 487 U.S. 589, 615-617 [101 L.Ed.2d 520, 108 S.Ct. 2562] [government monitoring of educational materials used by grantee is not excessive entanglement with religion]; Roemer v. Maryland Public Works Bd. (1976) 426 U.S. 736, 764-765 [49 L.Ed.2d 179, 96 S.Ct. 2337] [no excessive entanglement with religion where state conducts annual audits of religious colleges to ensure state aid is not used for “sectarian purposes”].)

For all of these reasons we conclude that the Authority’s issuance of government bonds benefiting religious schools under the conditions we have here set forth (see pp. 803-805, ante) does not violate the establishment clause of the First Amendment to the federal Constitution. Our conclusion finds support in a 2000 decision by the Virginia Supreme Court and in a 2002 decision by the federal court of appeals for the Sixth Circuit, both of which upheld similar bond funding programs benefiting religious schools. (Virginia College, supra, 538 S.E.2d 682; Steele, supra, 301 F.3d 401.)

Disposition

The judgment of the Court of Appeal is reversed, and the matter is remanded to that court with directions to remand the case to the trial court so it can reconsider the Authority’s request for validation of its bond funding agreements with Oaks Christian School, California Baptist University, and Azusa Pacific University in light of our opinion here.

George, C. J., Baxter, J., and Corrigan, J., concurred.

In sweeping terms, article XVI, section 5 of the California Constitution (article XVI, section 5) provides: “Neither the Legislature, nor any county, city and county, township, school district, or other *813municipal corporation, shall ever make an appropriation, or pay from any public fund whatever, or grant anything to or in aid of any religious sect, church, creed, or sectarian purpose, or help to support or sustain any school, college, university, hospital, or other institution controlled by any religious creed, church, or sectarian denomination whatever; nor shall any grant or donation of personal property or real estate ever be made by the State, or any city, city and county, town, or other municipal corporation for any religious creed, church, or sectarian purpose whatever; provided, that nothing in this section shall prevent the Legislature granting aid pursuant to Section 3 of Article XVI.” (Italics added.) This section “constitute^] ‘the definitive statement of the principle of government impartiality in the field of religion.’ [Citation.]” (California Educational Facilities Authority v. Priest (1974) 12 Cal.3d 593, 604 [116 Cal.Rptr. 361, 526 P.2d 513] (Priest).) Its purpose, as revealed by the debates of the constitutional convention that drafted it, is “to insure the separation of church and state and to guarantee that the power, authority, and financial resources of the government shall never be devoted to the advancement or support of religious or sectarian purposes. [Citation.]” (Ibid.)

Given the trial court’s uncontested factual findings that “[r]eligion is both mandatory and integral to every aspect of student life” at the schools here at issue, and that the schools are “organized primarily or exclusively for religious purposes,” “restrict[] admission of students by religious criteria,” “discriminate[] on the basis of religion in hiring faculty,” and “integrated” “[r]eligion . . . into classroom instruction,” I conclude that the proposed bond financing agreements now before us are invalid under both the plain language and our judicial construction of article XVI, section 5. In my view, that provision simply does not permit a public entity to act as a fundraiser for schools of this nature in order, as the majority puts it, “to encourage” the “development” of such schools (maj. opn., ante, at p. 795) and to “enhanced” their “ability ... to expand . . . .” (Id. at p. 803.) I therefore dissent.

Factual Background

Because the majority completely ignores the trial court’s factual findings and glosses over the details of the proposed bond financings, I begin by discussing those matters.

In each of the three validation actions here at issue, the California Statewide Communities Development Authority (the Authority) moved for a default judgment. In its moving papers, the Authority “assumed,” but did not concede, that each school “would be considered” to be “pervasively sectarian,” meaning that “ ‘religion is so pervasive [at the school] that a substantial portion of its functions are subsumed in [its] religious mission.’ ”

*814The Authority also submitted declarations in support of its motions. Regarding Oaks Christian School, the Authority’s supporting declarations stated: (1) the school is a “Christian school for the education of children in the sixth through twelfth grades”; (2) its “mission statement is ‘to grow in knowledge and wisdom through God’s grace, and to dedicate [oneself] to the pursuit of academic excellence, athletic distinction and Christian values’ ”; (3) students and their parents “must agree” to “support the School’s mission, statement of faith and Biblical goals and objectives”; (4) “[f]acuity members must be Christian and . . . sign a statement of faith”; and (5) the school “seeks,” among other things, “to develop ‘each student’s mind, body and spirit . . . through . . . spiritual training by the finest Christian teachers and coaches in the nation,’ ” “foster an understanding of the sovereignty of God to provide a framework for the application of knowledge,” “refine the body and character through teamwork and in competition that honors God,” and “encourage a passion to love God.” In its supporting brief, the Authority added that the school “require[s] . . . that students attend assembly period twice a week during which prayers may be held.”

Regarding California Baptist University, the Authority’s supporting declarations stated: (1) the school is “a Christian liberal arts institution offering undergraduate and graduate programs of study”; (2) it “seeks” students who “believe in biblically-based Christian principles” and “expect[s]” students “to live in accordance with such principles”; (3) it “require[s]” students “to attend a church of their choosing” and to “complete a certain number of courses in . . . Christian studies”; (4) it “requires that [all] faculty members be Christian, and [that] at least 51% of the faculty members ... be Baptist”; and (5) it “expect[s]” all faculty members “to maintain a theological and philosophical position consistent with the University’s principles.”

Regarding Azusa Pacific University, the Authority’s supporting declarations stated: (1) the school “is an evangelical Christian community of disciples and scholars who seek to advance the work of God in the world through academic excellence in liberal arts and professional programs of higher education that encourage students to develop a Christian perspective of truth and life”; (2) applicants “must evidence appreciation for the standards and spirit of the University, and exhibit moral character in harmony with its purpose”; (3) an applicant’s “involvement in church” is “reviewed” during the admission process; (4) students “must complete . . . 120 hours of student ministry assignments”; and (5) all faculty members must “be Christian and are expected to maintain a theological and philosophical position consistent with the University’s principles.” The Authority’s supporting brief added that the school requires students to “attend chapel.”

In each action, the trial court denied the Authority’s motion for default judgment. In its orders, the court first set forth the facts regarding the schools *815as detailed in the Authority’s declarations. It then found as to each school that the proposed bond financing “fail[ed]” to pass muster under article XVI, section 5, explaining: “Based upon the facts presented, the educational institution is organized primarily or exclusively for religious purposes. It restricts admission of students by religious criteria and discriminates on the basis of religion in hiring faculty. Religion is both mandatory and integral to every aspect of student life. Religion is integrated into classroom instruction, [ft] Thus low cost financing for the school’s ‘acquisition, construction, improvement, renovation, remodeling, furnishing and equipping’ of classrooms and other facilities necessarily involves financing religious indoctrination.”1

In the Court of Appeal, the Authority did not contest the trial court’s factual findings. On the contrary, it argued that because “there [were] no contested factual issues,” the only issue was whether the trial court had applied the correct “legal analysis” in focusing “on the religious nature of the school[s]” instead of “the nature of the benefit being provided.” Moreover, “[f]or purposes of th[e] appeal,” the Authority did “not dispute that [the schools] could be characterized as ‘pervasively sectarian,’ ” and it explained that “an educational institution is considered ‘pervasively sectarian’ when a substantial portion of [its] function is subsumed in its religious mission and it is impossible to separate its religious aspects from its secular aspects.” Similarly, in this court, the Authority does not contest the trial court’s findings. Thus, we must accept the trial court’s factual findings regarding the nature of the schools.2

The details of the proposed bond financing are set forth in purchase and sale agreements between the schools and the Authority, the governing statutes, and declarations the Authority submitted in support of its motions for default judgment. Under these agreements, the Authority, which is a public entity, promises to issue, sell, and deliver the bonds and to “apply the proceeds received from the sale” to pay for the costs of the bond sales and the specified projects at the schools. By both statute and the agreements with the schools, the bonds constitute “special obligations” of the Authority. (Gov. Code, § 91535, subd. (a); see also id., § 91541, subd. (e).) The Authority’s obligation is special in the sense that the bonds are payable only from funds the Authority receives from the schools under the agreements. (Ibid.) In this regard, the Authority’s agreements with Oaks Christian School and California Baptist University state: “The Authority shall not be obligated to pay the *816principal (or redemption price) of or interest on the Bonds, except from Revenues and other moneys received by the Trustee on behalf of the Authority pursuant to this Sale Agreement. . . . [f] The [school] hereby acknowledges that the Authority’s sole source of moneys to repay the Bonds will be provided by the payments made by the [school] pursuant to this Sale Agreement, together with investment income on certain funds held by the Trustee under the Indenture.” The Authority’s agreement with Azusa Pacific University contains a substantively identical provision.

With respect to each school, the Authority simultaneously executes both a purchase agreement and a separate sale agreement. In the purchase agreements, the schools sell and transfer to the Authority all of their right, title and interest in and to the school property on which the improvements will be made. In the sale agreements, the Authority sells and transfers back to the schools all of the interest in the schools’ property it acquired under the purchase agreements. The purchase price for the school property the Authority sells back to the schools is essentially the amount needed to pay off the principal, interest, and any premium on the bonds. “As security for the payment of the [b]onds, the Authority” assigns to a trustee its right to receive the schools’ payments for repurchasing the property, and directs the schools to make these payments directly to the trustee. The trustee, as assignee of the Authority, uses this money to pay off the bondholders on behalf of the Authority and to satisfy the “special obligation[]... of the [Authority” to pay the bondholders. (Gov. Code, § 91535, subd. (a).) As this discussion makes clear, although the schools are to be the sole source of the funds to pay the bondholders, the legal obligation to pay the bondholders is the Authority’s.3

The purpose of these machinations is to save the schools considerable money in financing costs. In general, bond investors will accept a lower interest rate on investments where the interest is tax-exempt. However, interest on bonds cannot be tax-exempt unless a governmental entity like the Authority is the bond issuer. Thus, as the Authority explained in the Court of Appeal, the schools, through the Authority’s participation as bond issuer, will *817be able to borrow money to pay for their projects at a lower interest rate, thus enabling them “to finance their respective projects at a lower cost than is available through conventional private financing.” The savings would be considerable; for example, according to a declaration the Authority submitted in support of its motions for default judgment, the proposed bond financing would have saved Oaks Christian School “approximately $52,500 per month” in financing costs under interest rates prevailing in 2002. Each of the purchase and sale agreements acknowledges the importance of these savings to the schools, stating that “obtaining Tax-Exempt status for the financing of the Costs of the Project is a significant factor in maintaining the operations of the [schools] within the jurisdiction of the Project Program Participant.”

The agreements also describe the projects to be financed through the bond sales. Oaks Christian School intends to use the funds to “acqui[re], construct[], improve[], renovat[e], remodel[], fumish[] and equip[] . . . classrooms, laboratories, administration offices, dining facilities, athletic facilities, parking facilities, [and] a co-generation facility.” California Baptist University intends to make similar use of the funds for “residence facilities, parking facilities, classrooms, administration offices, the academic and student center complex, [and] athletic facilities.” Azusa Pacific University intends to use the funds for similar purposes regarding “certain educational facilities . . . including ... a new . . . residence facility, a dining facility, [and] a mail center.” In the agreements, the schools “agree[] that no facility, place or building financed or refinanced with a portion of the proceeds of the Bonds will be used ... for sectarian instruction or as a place for religious worship or in connection with any part of the programs of any school or department of divinity for the useful life of the Project.”4

Discussion

“The determination of [California’s] public policy . . . resides, first, with the people as expressed in their Constitution . . . .” (Jensen v. Traders & General Ins. Co. (1959) 52 Cal.2d 786, 794 [345 P.2d 1].) In other words, the California Constitution is both “the highest expression of the will of the people of the state” (Ex parte Braun (1903) 141 Cal. 204, 211 [74 P. 780]) and “the preeminent expression of California law.” (American Academy of Pediatrics v. Lungren (1997) 16 Cal.4th 307, 314 [66 Cal.Rptr.2d 210, 940 P.2d 797] (Lungren).) Accordingly, when construing a provision of the *818state Constitution, our “paramount consideration” is “the intent of those who enacted the provision at issue. [Citation.] To determine that intent, [we] look first to the language of the constitutional text, giving the words their ordinary meaning. [Citations.]” (Leone v. Medical Board (2000) 22 Cal.4th 660, 665 [94 Cal.Rptr.2d 61, 995 P.2d 191].) We give the words their ordinary meaning because we “ ‘presume[]’ ” that they were “ ‘so understood by the framers, and by the people who adopted [them].’ ” (Miller v. Dunn (1887) 72 Cal. 462, 465 [14 P. 27].)

In light of the trial court’s uncontested factual findings, the proposed bond agreements in this case unquestionably are invalid under the ordinary meaning of article XVI, section 5’s language. As noted above, among other things, that provision prohibits public entities from “granting] anything to or in aid of any religious sect, church, creed, or sectarian purpose, or helping] to support or sustain any school, college, university, hospital, or other institution controlled by any religious creed, church, or sectarian denomination whatever.” (Art. XVI, § 5.) As we have explained, a bond financing program like the one here at issue “clearly provides a ‘benefit’ ” to participating schools “through the use of a [governmental] instrumentality,” by “enabling]” them “to borrow money ... at a cost below that of the marketplace.” (Priest, supra, 12 Cal.3d at p. 605.) Indeed, as noted above, under interest rates prevailing in 2002, the proposed bond program would have saved Oaks Christian School “approximately $52,500 per month” in financing costs. As also noted above, both the Authority and the schools have acknowledged that the savings realized through the bond program, which would be unavailable without a public entity’s participation, are “a significant factor in maintaining the operations of the [schools] within the jurisdiction of the Project Program Participant.” Given the trial court’s uncontested factual findings that the schools are “organized primarily or exclusively for religious purposes,” “restrict[] admission of students by religious criteria,” “discriminate^ on the basis of religion in hiring faculty,” and “integrate[]” “[r]eligion . . . into classroom instruction,” the proposed bond agreements clearly violate the ordinary meaning of article XVI, section 5’s prohibition against “granting] anything to or in aid of any religious sect ... or sectarian purpose, or helping] to support or sustain any school, college, university, hospital, or other institution controlled by any religious creed, church, or sectarian denomination whatever.”

In several relevant cases, we have explored the reach of article XVI, section 5, and its materially identical predecessor. The more recent—and in my view more authoritative—is California Teachers Assn. v. Riles (1981) 29 Cal.3d 794 [176 Cal.Rptr. 300, 632 P.2d 953] (Riles). There, we held that statutes authorizing the Superintendent of Public Instruction to lend, without charge, textbooks used in the public schools to students attending private sectarian schools were invalid under article XVI, section 5. (Riles, supra, at *819pp. 797-798.) The religious schools at issue in Riles “offer[ed] instruction in secular subjects, but they also [had] as their purpose the teaching of the tenets of their faith. Some of the[] schools [gave] preference to enrolling Catholic pupils; more than 97 percent of the students attending [the] schools [were] Catholic. The schools ordinarily require[d] students to receive religious instruction, attend religious services during the school day, and participate in prayers and religious ceremonies. Sectarian symbols and pictures [were] distributed throughout the schools’ buildings. The teachers in these schools [were] for the most part members of the church.” (Id. at pp. 799-800.)

In Riles, to determine the constitutional validity of the textbook lending program, we first considered “whether it only indirectly benefited] parochial schools.” (Riles, supra, 29 Cal.3d at p. 809.) The trial court had concluded that the program had “a clearly secular . . . purpose” and that the benefit it provided the schools, “though substantial, [was] only indirect and incidental.” (Id. at p. 800.) We disagreed, holding that the benefit sectarian schools received from the program could not “be characterized as . . . only indirect, remote, and incidental.” (Id. at p. 809.) In reaching this conclusion, we focused principally on “the inseparability of the benefit to the pupil and the school, and the impossibility of characterizing the advantage to one as remote and to the other as direct.” (Id. at p. 810.) We explained: “The books are supplied for use in the school, and we are unable to perceive any significant distinction from a constitutional standpoint whether they are loaned to the students for use in the school, or to the school for use by the students. In either circumstance, both the child and the school benefit. The United States Supreme Court has recognized as much in characterizing textbook loan programs as a form of financial assistance to the school even though the loan is nominally made to the student. There is no rational reason for concluding that the school benefits only indirectly or remotely from the loan if the child is the nominal recipient, for it is an undeniable fact that books are a critical element in enabling the school to carry out its essential mission to teach the students.” (Ibid., fn. omitted.)

We next explained in Riles that our conclusion that the benefit was “neither indirect nor remote [did] not end our inquiry,” because not all aid that “directly” benefits sectarian schools is “prohibited (e.g., providing fire protection),” and not all aid that indirectly benefits sectarian schools is “valid (e.g., reimbursement for the purchase of religious articles by students in public and nonpublic schools).” (Riles, supra, 29 Cal.3d at p. 811.) Ultimately, we explained, the program’s validity under article XVI, section 5, turned on “whether the character of the benefit [the program] provided . . . resulted] in the ‘support of any sectarian . . . school.’ ” (Riles, supra, 29 Cal.3d at p. 811.) We concluded that the program did result in such support, and was therefore invalid, because textbooks have “a central place in the educational mission of *820a school” and the benefit being provided “advance[d]” the schools’ “educational function.” (Ibid.) “In this respect,” we explained, “the [textbook lending] program [was] distinguishable from ‘generalized services government might provide to schools in common with others’ [citation], such as fire and police protection, the maintenance of roads and sidewalks, and similar public services. These services, unlike education, have no doctrinal content, and they do not advance the essential objective of the sectarian school, which is the education of the child.” (Id. at pp. 811-812.)

Finally, we held in Riles that the textbook lending program was invalid under article XVI, section 5, even if the schools “used” the books “only for secular instruction.” (Riles, supra, 29 Cal.3d at p. 812.) We reasoned that article XVI, section 5, “do[es] not confine [its] prohibition against financing sectarian schools in whole or in part to support for their religious teaching function, as distinguished from secular instruction.” (Riles, supra, 29 Cal.3d at p. 812.) We also explained that “[¡jurisdictions with similar constitutional provisions [had] also refuse[d] to make such a distinction” (ibid.), and we quoted with approval the following discussion from a Nebraska Supreme Court’s decision interpreting a provision “virtually identical” to article XVI, section 5: “ ‘[T]he court must examine the character of the aided activity rather than the manner or the form in which aid is given. . . . [O]ne of the main purposes of the parent sending his child to a parochial school is to insure the early inculcation of religion. Assuming that textbooks promote the notion of an absolutely neutral and equal secular educational program, the reimbursement or the loan of textbooks to the students is for the purpose of augmenting the public school secular education with religious training. The state, by aiding the parents and the students by textbooks, secular though they may be, is providing a program for aiding the church and in advancing religious education.’ ” (Riles, supra, 29 Cal.3d at p. 812, fn. 15.)

Applying Riles and its analytical framework to this case, I conclude that the proposed bond financings at issue here are invalid under article XVI, section 5. In terms of their religious orientation, the schools at issue here are similar in important respects to the schools at issue in Riles', as the trial court found, “[r]eligion is both mandatory and integral to every aspect of student life” at the schools and the schools both “restrict[] admission of students by religious criteria” and “discriminate!] on the basis of religion in hiring faculty.” Indeed, religion is even more integral to the schools here than it was to the schools in Riles; whereas the trial court in Riles found unproven allegations that the schools there “conducted] their operations to fulfill religious purposes” or “blend[ed] secular and sectarian instruction” (Riles, supra, 29 Cal.3d at pp. 800-801), the trial court here found that the schools are “organized primarily or exclusively for religious purposes” and “integrate!]” “[r]eligion . . . into classroom instruction.” Moreover, the benefit to the schools in this case is every bit as direct as, and is probably far more *821substantial than, the benefit at issue in Riles. As explained above, in finding that the benefit of the textbook lending program was “neither indirect nor remote,” we emphasized in Riles that the program provided “a form of financial assistance to the school[s]” and that textbooks were “a critical element in enabling the school[s] to carry out [their] essential mission to teach the students.” (Riles, supra, 29 Cal.3d at pp. 811, 810.) Here, the proposed bond program clearly would provide a form of financial assistance to the schools; it would enable them to borrow money at below-market rates and significantly lower the cost of constructing, improving, furnishing, and equipping their facilities. Moreover, the facilities the proposed bond program would enable the schools to build or improve, and the furnishings and equipment it would enable them to buy, are critical elements in enabling them to carry out their essential mission to teach students. In this regard, the character of the benefit the proposed bond financings would provide would impermissibly result in the support of sectarian schools by “advancing]” their “educational function.” (Riles, supra, 29 Cal.3d at p. 811.) Finally, the schools’ promise not to use any “facility, place or building financed or refinanced with a portion of’ the bond proceeds “for sectarian instruction or as a place for religious worship or in connection with any part of the programs of any school or department of divinity” does not render the proposed bond financings constitutionally valid because, as Riles established, article XVI, section 5, “do[es] not confine [its] prohibition against financing sectarian schools in whole or in part to support for their religious teaching function, as distinguished from secular instruction.” (Riles, supra, 29 Cal.3d at p. 812.) Thus, under Riles, the proposed bond financings here violate article XVI, section 5.5

Several years before we decided Riles, in Priest, we discussed the scope of article XVI, section 5’s materially identical predecessor, California Constitution former article XIII, section 24, in connection with a statutory bond financing program that was similar in certain respects to the bond financing program now before us. The case arose when the State Treasurer, because of concerns about the statutory program’s constitutionality, refused to perform her duties regarding an approved bond financing. (Priest, supra, 12 Cal.3d at p. 598.) Regarding former article XIII, section 24, we began by explaining generally: “This section has been said to constitute ‘the definitive statement of the principle of government impartiality in the field of religion.’ [Citation.] An examination of the debates of the constitutional convention which drafted the Constitution of 1879 indicates that the provision was intended to insure the separation of church and state and to guarantee that the power, authority, *822and financial resources of the government shall never be devoted to the advancement or support of religious or sectarian purposes. [Citation.] Under this section, the fact that a statute has some identifiable secular objective will not immunize it from further analysis to ascertain whether it also has the direct, immediate, and substantial effect of advancing religion. . . . [f] The section has never been interpreted, however, to require governmental hostility to religion, nor to prohibit a religious institution from receiving an indirect, remote, and incidental benefit from a statute which has a secular primary purpose. [Citation.]” (Priest, supra, at pp. 604-605.)

After setting forth these general principles, we opined in Priest that the statutory bond program there at issue, although “appearing] to approach state involvement with religion [citation],” could not be said “in the abstract” to “cross[] the forbidden line.” (Priest, supra, 12 Cal.3d at p. 606.) We first explained that the statutory program “clearly provide[d] a ‘benefit’ ” by enabling schools “to borrow money through the use of a state instrumentality at a cost below that of the marketplace.” (Id. at p. 605.) We next noted, and deferred to, the Legislature’s determination that the statutory program, “in supporting the maintenance and improvement of facilities for higher education, [was] in the public interest [citations].” (Ibid.) Finally, we stated: “The benefits of the [program] are granted to sectarian and nonsectarian colleges on an equal basis; ... all aid for religious projects is strictly prohibited; and in no event is a financial burden imposed upon the state. In these circumstances the [statutory program] does not have a substantial effect of supporting religious activities.” (Id. at p. 606.)

Unlike the majority, which wholly embraces Priest without either analysis or hesitation, I believe there is good reason to question that decision’s authority and persuasiveness. To begin with, as Priest itself noted, the school seeking bond financing in that case was “not affiliated with any religious organization, and therefore [was] not directly concerned with a challenge to the [statutory program’s] validity . . . under” article XVI, section 5’s predecessor. (Priest, supra, 12 Cal.3d at p. 598, fn. 5.) For this reason, in Riles-—which, like Priest, Justice Mosk authored for the court—we explained that Priest was “dictum” insofar as it purported to address schools that were “affiliated with any religious denomination.” (Riles, supra, 29 Cal.3d at p. 806, fn. 10.) The majority ignores this statement in Riles in asserting that Priest “held that issuing revenue bonds to fund capital improvements at religiously affiliated colleges did not violate” article XVI, section 5’s predecessor. (Maj. opn., ante, at p. 799, italics added.)

Moreover, Priest’s assertion that article XVI, section 5’s predecessor had not been interpreted “to prohibit a religious institution from receiving an indirect, remote, and incidental benefit from a statute which has a secular *823primary purpose” (Priest, supra, 12 Cal.3d at p. 605) rested on very shaky ground. In support of this assertion, Priest cited our decision in Lundberg v. County of Alameda (1956) 46 Cal.2d 644 [298 P.2d 1] (Lundberg), and discussed the Court of Appeal’s decision in Bowker v. Baker (1946) 73 Cal.App.2d 653 [167 P.2d 256]. (Priest, supra, 12 Cal.3d at p. 605.) Lundberg does not in any way support Priest’s assertion; there, in rejecting a challenge to a tax exemption under another of article XVI, section 5’s materially identical predecessors, we tersely explained: “This section does not expressly mention tax exemptions, but, even if we assume that it prohibited them, it was superseded by the subsequent adoption of’ two other constitutional provisions. (Lundberg, supra, 46 Cal.2d at p. 653.) In Bowker, the Court of Appeal held that article XVI, section 5’s predecessor did not prohibit a school district from incurring “some slight added cost” (Bowker, supra, at p. 657) to permit parochial school students to occupy empty seats on public school buses, reasoning that “the direct benefit” of this expense “flowfed] to the children . . . , with only an indirect benefit to the private parochial school . . . .” (Id. at pp. 666-667.) Priest failed to mention Bowker’s narrow reasoning in making the much broader claim that article XVI, section 5’s predecessor did not “prohibit a religious institution from receiving an indirect, remote, and incidental benefit from a statute which has a secular primary purpose. [Citation.]” (Priest, supra, 12 Cal.3d at p. 605.) Moreover, in Riles, we rejected the “ ‘child benefit’ theory” on which Bowker was based, noting that it “ha[d] been criticized by courts and commentators” (Riles, supra, 29 Cal.3d at p. 807), had produced “dissonant decisions” that could not be “harmonize[d]” (id. at pp. 807, 808), and “in most instances . . . leads to results which are logically indefensible.” (Id. at p. 809.) Given these circumstances, it is little wonder that Riles, though written for the court by the same justice who wrote Priest, expressly declined to endorse or reaffirm either Priest or Bowker, and instead cast doubt on both by stating that because they were distinguishable, it was unnecessary to “consider whether [they] were correctly decided.”(Riles, supra, 29 Cal.3d at p. 813, fn. 16.)

These concerns aside, unlike the majority, I believe that the proposed bond financings in this case are invalid even under Priest’s dictum. As explained above, Priest stated that the purpose of article XVI, section 5, is “to guarantee that the power, authority, and financial resources of the government shall never be devoted to the advancement or support of religious or sectarian purposes,” and that even a statute with a “secular objective” is invalid if it “has the direct, immediate, and substantial effect of advancing religion.” (Priest, supra, 12 Cal.3d at p. 604.) For the reasons explained above, including and especially the trial court’s factual findings regarding the nature of the schools here at issue, the proposed bond financings would have the direct, immediate and substantial effect of advancing religion and would *824contravene the constitutional provision’s purpose, by devoting the government’s power and authority to raise money at below-market interest rates through the issuance and sale of tax-exempt bonds to the support and advancement of religious or sectarian purposes. This conclusion is fully consistent with Priest’s view that the bond financing program there at issue was not, “in the abstract” (id. at p. 606), invalid given the particular “circumstances” in that case. (Id. at p. 598, fn. 5.) As noted above, one of those “circumstances” was that “all aid for religious projects [was] strictly prohibited.” (Id. at p. 606.) The facts underlying this circumstance included not only statutory use limitations similar to the limitations in the restricted use covenants here at issue, but also the complete exclusion of schools that either “ ‘restricted] entry on . . . religious grounds’ ” or “ ‘require[d] all students gaining admission to receive instruction in the tenets of a particular faith.’ ” (Priest, supra, 12 Cal.3d at p. 596.) In Riles, we specifically noted this exclusion in distinguishing Priest. (Riles, supra, 29 Cal.3d at pp. 806, 813, fn. 16.) Here, given the trial court’s uncontested factual findings that “[religion is both mandatory and integral to every aspect of student life” at the schools, and that the schools are “organized primarily or exclusively for religious purposes,” “restrict[] admission of students by religious criteria,” “discriminate^ on the basis of religion in hiring faculty,” and “integrate^” “[r]eligion . . . into classroom instruction,” notwithstanding the restricted use covenants, the proposed bond financings would impermissibly provide aid for religious projects. Thus, even under Priest’s dictum, the proposed bond financing programs violate article XVI, section 5.6

In any event, Priest must be read in light of our subsequent decision in Riles. As explained above, the benefit the schools would receive in this case under the bond financing program would be “direct” as we interpreted that term in Riles. (Riles, supra, 29 Cal.3d at p. 810.) As also explained above, we held in Riles that the key question under article XVI, section 5, is not whether the benefit in question is direct or incidental to some public purpose— because not all direct benefits are invalid, and not all indirect benefits are permissible—but is “whether the character of the benefit . . . results in the ‘support of any sectarian . . . school’ ” by “advancing]” the school’s “essential objective ..., which is the education of the child.” (Riles, supra, 29 Cal.3d at pp. 811-812.) Notably, in Riles, notwithstanding the trial court’s conclusion that the program “ha[d] a clearly secular legislative purpose” (id. at *825p. 800), we invalidated the textbook lending program because textbooks “hav[e] a central place in the educational mission of a school” and are an “ ‘essential tool of education . . .’ [citation].” (Id. at p. 811.) Similarly, the classrooms, laboratories, residence and dining halls, and other facilities the schools here intend to build or improve with the bond proceeds, and the furnishings and equipment they intend to purchase with those proceeds, are also central to the schools’ educational mission and are essential tools of education.

The majority virtually ignores Riles, declaring in a footnote that it, unlike this case, involved “direct public aid!’ to parochial schools in the form of textbooks provided “at public expense.” (Maj. opn., ante, at p. 801, fn. 7.) By contrast, the majority asserts, because “no public funds are expended” in connection with the proposed bond financings here (ibid.), only “indirect assistance” is at issue in this case. (Id., at p. 799.)

For several reasons, the majority’s stated basis for disregarding Riles is invalid. Initially, the majority’s view rests on its unsupported assertion that the proposed bond financings would involve “no [expenditure of] public funds.” (Maj. opn., ante, at p. 801, fn. 7.) As noted above, the bonds here will be issued and sold by a public entity: the Authority. The majority does not explain why the proceeds from the sales, which will be used at the schools, are not public funds. Nor does the majority explain why the interest on those proceeds, which also will be used at the schools, are not public funds. Moreover, although the majority asserts that the obligation to pay the bondholders “falls to the recipient schools” and that the payment obligation will not be “secured by any public property” (maj. opn., ante, at p. 797), as noted above, payment of the bondholders is, by statute and by agreement, a special obligation of the Authority, and the Authority’s payment obligation is secured by its right to receive from the schools payments to repurchase school property from the Authority. The majority simply glosses over these aspects of the proposed bond financings.7

More broadly, the majority’s focus on whether the benefit would involve a direct expenditure of public funds is inconsistent with the language of article XVI, section 5, and our interpretation of that language in both Riles and *826Priest. The majority’s analysis indicates that only an expenditure of public money may constitute a direct benefit. However, in Priest, we expressly rejected that view, explaining that the “terms” of the constitutional provision “forbid granting ‘anything’ to or in aid of sectarian purposes, and prohibit public help to ‘support or sustain’ a sectarian-controlled school. The section thus forbids more than the appropriation or payment of public funds to support sectarian institutions. It bans any official involvement, whatever its form, which has the direct, immediate, and substantial effect of promoting religious purposes.” (Priest, supra, 12 Cal.3d at p. 605, fn. 12.) For reasons already explained, the bond financings proposed here would have such an effect, by enabling the schools “to obtain economic aid” they “would not otherwise be able to obtain without the government’s direct participation.” (Steele v. Industrial Development Bd. Of Metro (6th Cir. 2002) 301 F.3d 401, 438 (dis. opn. of Clay, J.).) Even if, as the majority asserts, there would be no expenditure of public funds, that circumstance would “not alter for one moment the fact that a direct economic benefit [would] accrue[] to [the schools] as a result of the government’s active participation in arranging for a low-cost loan that [would] enable[] the [schools] to advance [their] sectarian mission.” (Ibid.) Moreover, in Riles, as already noted, we held that even a so-called indirect benefit may violate article XVI, section 5, and that the key question is not whether the benefit in question is direct or incidental, but whether the “character of the benefit . . . results in the ‘support of any sectarian . . . school’ ” by “advancing]” the school’s “essential objective . . . , which is the education of the child.” (Riles, supra, 29 Cal.3d at pp. 811-812.) Thus, even the majority’s mistaken view that no “direct public aid’ is at issue here (maj. opn., ante, at p. 801, fn. 7) does not justify its complete disregard of our unanimous decision in Riles.

The majority’s analysis is at odds with Riles in another important respect. As noted above, we unanimously held in Riles that the textbook lending program there at issue was invalid under article XVI, section 5, even if the books were “used only for secular instruction.” (Riles, supra, 29 Cal.3d at p. 812.) We explained that article XVI, section 5, “do[es] not confine [its] prohibition against financing sectarian schools in whole or in part to support for their religious teaching function, as distinguished from secular instruction.” (Ibid.) Thus, under Riles, where a government program provides a benefit that “advance[s] the essential objective of [a] sectarian school,” the school’s ability to separate its religious and secular aspects and to use the benefit only for the latter does not save the program’s constitutionality. (Ibid.) Contrary to Riles, the majority’s test resurrects the inquiry into a school’s ability to separate its religious and secular aspects. Under that test, the constitutionality of aid given to a school that “provide[s] a broad curriculum in secular subjects” (maj. opn., ante, at p. 803), that “includes a religious perspective in its curriculum” (ibid.), and that “expresses]” its *827“religious viewpoint in [its] otherwise secular classes” (id. at p. 793), depends in part on whether the school uses the aid “for ‘religious projects’ ” (id. at p. 801), whether its “secular classes consist of information and coursework that is neutral with respect to religion” (id. at p. 793), and whether classes taught in facilities financed with bond proceeds “include[] as part of the instruction information or coursework that promotes or opposes a particular religion or religious beliefs.” (Id. at p. 804.) Under this analysis, the test of constitutionality depends on the extent to which a school separates its teaching of secular subjects from its teaching of religion. In this regard, the majority’s test is contrary to the language of article XVI, section 5, which, as we held in Riles, “do[es] not confine [its] prohibition against financing sectarian schools in whole or in part to support for their religious teaching function, as distinguished from secular instruction.” (Riles, supra, 29 Cal.3d at p. 812.)

Another problem with the majority’s newly minted test—which has not been proposed by any party to this case and which is not based on the language of article XVI, section 5, or on any existing case law—is that it is hopelessly vague in several respects. The majority states that a proposed bond financing is constitutional only if, among other things, the benefited school “offers a sufficiently broad curriculum in secular subjects . . . that [its] use of the educational facilities built or improved with the bond funding may be expected to promote the public interest in making secular education more available to California residents in general.” (Maj. opn., ante, at p. 806.) This formulation leaves the reader—and the trial court, which must apply the majority’s test on remand—asking: How broad is “sufficiently broad”? The majority states that offering classes in only “a few” secular subjects is not enough (maj. opn., ante, at p. 803), but exactly how many is “a few”—two, three, four?8 Do judges have discretion as to how many is “a few,” such that one judge may properly find that four secular subjects are not enough, while another may also properly find that four secular subjects are enough? Is any number more than “a few,” whatever that is, automatically enough to support the conclusion that the school’s “use of the educational facilities built or improved with the bond funding may be expected to promote the public interest in making secular education more available to California residents in general”? (Maj. opn., ante, at p. 806.) The majority provides no guidance on these questions.

The majority’s test is also vague in directing trial courts to determine whether “the academic content” of a religious school’s courses in secular subjects is typical of that provided in nonreligious schools. (Maj. opn., ante, at p. 804.) The majority fails to define what it means by the term “academic *828content.” The word “academic” simply means “of, belonging to, or associated with an academy or school.” (Webster’s 3d New Intenat. Dict. (2002) p. 9.) Thus, all content of a religious school’s courses in secular subjects constitutes “academic content.” Presumably, the majority means something else by its use of the term, but the majority does not say what. Insofar as the majority is using the term to mean nonreligious content, the majority’s test is tautological; the nonreligious content of a religious school’s course in a secular subject would seemingly have to be typical of the content provided in a nonreligious school’s course in the same subject.

The majority’s test is tautological in another sense. According to the majority, a school that includes a religious perspective in its curriculum may be the beneficiary of a bond financing only if two purportedly separate requirements are met: the school must offer “a sufficiently broad curriculum in secular subjects” (maj. opn., ante, at p. 806, italics added) and “the information and coursework used to teach [those] secular subjects must be neutral with respect to religion.” (Id., at p. 804.) In my view, a subject qualifies as “secular,” by definition, only if it consists of information and coursework that is neutral with respect to religion. In other words, any subject that qualifies as secular will necessarily consist of information and coursework that is neutral with respect to religion. Thus, these two separately stated requirements of the majority’s test are, in actuality, but one.

The majority’s test is also ambiguous in that it announces seemingly inconsistent standards. On the one hand, the majority states that a benefited school may not use the bond proceeds “for ‘religious projects,’ that is ‘sectarian instruction or as a place for religious worship or in connection with any part of the programs of any school or department of divinity for the useful life of the Project.’ ” (Maj. opn., ante, at p. 801.) On the other hand, the majority states that it is constitutionally permissible for a religious school, in those same facilities, to express its religious viewpoints and perspectives in the course of teaching secular subjects. (Maj. opn., ante, at p. 794.) The majority fails to explain why a school’s expression of its religious viewpoints and perspectives in secular classes does not constitute “sectarian instruction,” or why the building or improvement of a facility in which a school will express its religious viewpoints and perspectives does not constitute a “religious project.” In this regard, the majority’s analysis seems to be internally inconsistent; the first part of the majority’s four-part test appears to permit what the third part prohibits. More importantly, the majority’s analysis expressly permits the schools to impart their religious viewpoints and perspectives in secular courses, in classrooms built, improved, furnished and equipped with money raised by a public entity—the Authority—on the schools’ behalf through the Authority’s issuance and sale of bonds. In my view, this violates both the plain meaning and our judicial construction of article XVI, section 5.

*829A related ambiguity in the majority’s test is the distinction the majority appears to draw between religious instruction and the expression of religious viewpoints and perspectives. On the one hand, the majority declares that a religious school may not, in facilities financed with bond proceeds, teach a “class that includes as part of the instruction information or coursework that promotes or opposes a particular religion or religious beliefs.” (Maj. opn., ante, at p. 804.) On the other hand, the majority states, a school may “express[j” its “religious viewpoints] in [its] otherwise secular classes.” (Maj. opn., ante, at p. 793.) Thus, in the majority’s view, there is a difference between expressing a religious viewpoint, which is permissible, and “including] as part of the instruction information or coursework that promotes or opposes a particular religion or religious beliefs,” which is not permissible. (Maj. opn., ante, at p. 804.) Unfortunately, the majority fails to explain where the difference lies, or how a court is supposed to tell one from the other.

The majority’s test also gives rise to several other concerns. First, it appears to require significant and unworkable monitoring to determine whether the schools, in teaching secular courses, are impermissibly including information or coursework that promotes or opposes a particular religion or religious beliefs, or are merely expressing religious viewpoints and perspectives. In this regard, notwithstanding the majority’s claim to the contrary (maj. opn., ante, at p. 804), the majority’s test will require scrutiny of the schools’ day-to-day operations by the Authority and the courts.9 Second, the majority’s test appears to be too narrow to address the constitutional concerns the proposed bond financings raise. Although the majority acknowledges “concern that a school with a religious perspective may [impermissibly] use the facilities built or improved with the revenue bond proceeds to substantially further its religious mission,” the test the majority announces to address this concern focuses only on what happens inside the schools’ classrooms. (Maj. opn., ante, at p. 803.) Notably, as already explained, the schools here intend to use the bond proceeds not just for classrooms, but also for dining and residence halls and an academic and student center complex. The majority’s analysis appears to ignore the extent to which the schools may use those facilities to further their religious missions.10

*830Finally, for several reasons, I disagree with the majority’s view that the proposed bond financings here are valid because they are “in the nature of a tax policy rather than an aid program.” (Maj. opn., ante, at p. 804, fn. 9.) First, the majority’s discussion rests on the untested and unwarranted assumption that anything that may be characterized as being “in the nature of a tax policy” because it somehow involves a “tax exemption]” (ibid.) does not, for that reason, violate article XVI, section 5. Nothing supports this assumption; on the contrary, in Lundberg, in rejecting a constitutional challenge to a tax exemption for property used for school purposes and owned by religious entities, we relied on other constitutional provisions that expressly authorized the exemption and “superseded” any prohibition in article XVI, section 5’s predecessor. (Lundberg, supra, 46 Cal.2d at p. 653.) The majority cites no constitutional provision that expressly authorizes the proposed bond financings in this case.11 Second, the majority’s discussion rests on another untested and unwarranted assumption: that the terms “tax policy” and “aid program” (maj. opn., ante, at p. 804, fn. 9) are mutually exclusive, and that something that can be called a tax policy necessarily cannot also constitute an aid program. Third, I disagree with an express premise of the majority’s “tax policy” analysis: that enabling schools to borrow funds at below-market interest rates is merely an “ ‘incidental result’ ” of providing private investors with a tax exemption. (Maj. opn., ante, at p. 805.) It is clear that enabling schools to borrow funds at below-market interest rates is both the intended result and primary purpose of the bond financing program. Indeed, there seems to be no other reason for the program’s existence; plenty of other tax-exempt investments are available to private investors, and the only reason for providing this one is to enable the schools to finance construction and improvement projects at an interest rate lower than would be available through conventional private financing.

The majority’s “tax policy” discussion suffers from a fourth flaw: it is based on an analysis we long ago criticized and rejected in Riles. There, as noted above, we explained that the “ ‘child benefit’ theory”—which looks to whether a school is benefited only as an incidental result of a benefit provided directly to students—“has been criticized by courts and commentators” (Riles, supra, 29 Cal.3d at p. 807), produces “dissonant decisions” that cannot be “harmonize[d]” (id. at pp. 807-808), and “in most instances . . . leads to results which are logically indefensible.” (Id. at p. 809.) The majority here *831adopts a close variant of that discredited theory, by reasoning that the schools benefit only incidentally from a tax benefit being provided directly to private investors. In my view, having rejected that approach 25 years ago in Riles, we should not adopt it now. In any event, just as “[tjhere [was] no rational reason” in Riles “for concluding” that the schools there “benefited] only indirectly or remotely from the loan” of textbooks to students (id. at p. 810), there is no rational reason here for concluding that the schools would benefit only indirectly or remotely from the tax advantages being made available to private investors. On the contrary, given the availability of other tax-free investments, the primary beneficiary of the proposed bond financings here would be the schools, which would save millions of dollars in financing costs.

In the end, this case seems relatively simple. By issuing and selling bonds, making the proceeds available to the schools, and accepting the special obligation to pay the bondholders, the Authority—a public entity—would be raising money on behalf of the schools at below-market interest rates that the schools could not obtain on their own. As the majority explains, the Authority’s purpose in taking these actions would be “to encourage” the “development” of the schools (maj. opn., ante, at p. 795) and to “enhance[]” their “ability ... to expand . . . .” (Id. at p. 803.) By so interceding in the marketplace on the schools’ behalf, the Authority would save the schools millions of dollars in financing costs and would significantly enhance their ability to carry out their mission. Given the trial court’s factual findings regarding that mission and the religious nature of these schools, even if no transfer of public funds is involved, the proposed bond financings violate article XVI, section 5’s prohibition against “granting] anything to or in aid of any religious sect ... or sectarian purpose, or helping] to support or sustain any school, college, university, hospital, or other institution controlled by any religious creed, church, or sectarian denomination whatever.”

I emphasize that my conclusion does not reflect “ ‘hostility to religion’ ” (maj. opn., ante, at p. 805) any more than did our unanimous holding in Riles that lending secular textbooks to religious schools violates article XVI, section 5, even if the schools use the textbooks only for secular instruction. Rather, my conclusion reflects my fealty to the will of the people as expressed first and foremost in the language of the California Constitution, and my view of the judiciary’s limited role in applying that Constitution. “The question before us is not whether, as a matter of policy, the [proposed bond financings are] wise or beneficial, but instead whether [they are] constitutional. We [must] determine the validity of the [proposed bond financings] by applying

*832the relevant legal principles embodied in the California Constitution, the preeminent expression of California law enacted by the people.” (Lungren, supra, 16 Cal.4th at p. 314.) Applying those constitutional principles, I conclude that the proposed bond financings here at issue are invalid under article XVI, section 5.1 therefore dissent.

Werdegar, J., and Moreno, J., concurred.