A (Partially) New Approach to Fund Infrastructure



As has been extensively covered, California’s once-toothless housing law has grown some fangs — at long last. More housing looks very likely to get built around the state. This housing will put big demands on local infrastructure. In this short essay, we will consider one approach to fund some of this infrastructure. It is not wholly novel, but combines reasonable (and permissible) steps that we believe have not previously been combined. This approach is desirable because, ultimately, it requires the approval of a majority rather than a supermajority of the electorate.

Policy and History

Infrastructure (roads, bridges, sewers etc.) is expensive. As a practical matter, neither public nor private actors typically pay for infrastructure through saving up for it and then paying for it all at once. Rather, one borrows to build infrastructure. Borrowing is not just a practical choice, but it is theoretically appropriate. After all, infrastructure tends to be long-lived, and so through borrowing the cost of the infrastructure is spread to the future generations who will use it.

Of course, the fact that debt shifts costs into the future creates moral hazard in the present.[1] Politicians and current business leaders have an incentive to make future generations pay for benefits they can provide their constituents and shareholders now. There are, accordingly, safeguards to prevent this. One such safeguard is that we have sophisticated capital markets that either won’t lend at all to foolish projects or will charge high prices that signal the project’s problems.

Yet markets make mistakes, and it is possible for expensive borrowings to happen anyway. In the context of private borrowings, businesses fail when they undertake too much bad debt. That can happen to governments too, and this is rightly perceived as highly undesirable because governments are in the business of providing essential services. Thus in many cases the private markets would not protect future citizens from being burdened with paying taxes for foolish projects funded years ago.

Consequently, state constitutions typically have an additional layer of protection. State and local governments cannot borrow just because politicians want to; there must be an election and, in California, a supermajority must approve the borrowing. It has long been recognized in all states that have such provisions that they are overprotective.

Consider a borrowing for a water treatment plant that is to be wholly paid for by water charges. Such a project does not implicate the state’s taxing power: it is essentially a private borrowing for a commercial enterprise. Either the charges are sufficient and the project succeeds, or not and it fails. As a general matter, such borrowing secured by a “special fund” has been held to not require a special bond election. Ordinary political and marketplace protections are deemed sufficient. Accordingly — and reasonably — if a local government were to fund local infrastructure with a special fund, then broadly speaking that could be done by local legislative action.

But how can a local government create such a fund? Many constitutional rules restrict the ability to raise taxes. In general, the rule in California is that creating special taxes (taxes dedicated to a special fund, for example) requires a local supermajority. In other cases, the rule is that local taxes can be raised by a majority of the electorate. Recent court cases have permitted local voters to use the local initiative to impose special taxes by majority vote.[2] As a policy matter this rule makes sense. The election requirement still protects the electorate, but the lower threshold is reasonable because there is less reason to protect the electorate from itself in contrast to local politicians. Local voters could thus put a special tax on the ballot and that tax would create a special fund that could be used to build the infrastructure that new housing will require.


If a city or a county imposes taxes that are dedicated to a special use, then they are a “special tax” and require a two-thirds vote of the local electorate.[3] Because such parcel taxes are not based on the value of property, they are not property taxes and therefore not subject to the restrictions on property taxes in the California constitution.[4] The statutes authorizing parcel taxes for counties and school districts have been interpreted to require that the levied taxes be flat.[5] In contrast, counties, school districts, and charter cities have the inherent right to tax.[6] General law cities have been granted the same powers to tax by the legislature.[7] Thus, cities do not need special state authorization to levy a parcel tax, and the parcel tax can be tiered.[8] Specifically, a city can adopt a local parcel tax based on size and use that is roughly similar to what the legislature approved in the 1982 Mello-Roos Act.[9]

Recent cases hold that a voter-initiated special tax measure does not require a two-thirds supermajority.[10] If the electorate (not the city) initiates the tax measure, then it requires only a majority vote to pass.[11] And it is a longstanding rule that the local electorate steps in the shoes of the local government when it uses the initiative power.[12] Thus, if the city can create a tiered parcel tax then a city electorate can use the initiative to do the same.[13]

The Mello-Roos Act contains explicit authorization for the taxes it raises to secure an issuance of bonds.[14] Under California constitution article XVI, section 18 no city can incur debts exceeding revenue in any year without a two-thirds vote of the electorate.[15] But a borrowing secured by a dedicated stream of special revenue is not subject to the requirement that the local government hold a bond election and secure a two-thirds vote.[16] This special fund doctrine “exempts certain obligations that are not legally enforceable against the local government’s general fund or its tax revenues.”[17] The rationale “is that where the local government’s ‘credit [is] not involved in the incurring of the indebtedness’ and the debt will not effect an increase in property taxes or threaten foreclosure upon government property, the debt is outside the scope of [section 18].”[18] The special fund doctrine cannot apply where the bond is paid for by revenue sources, like general sales and property taxes, that would otherwise go into the city’s general fund.[19]


Under current law, local city (or county) electorates can fund new infrastructure by means of a majority vote. This provides an opportunity for voter-initiated infrastructure financing measures as the 2024 election approaches. By pairing a parcel tax with a “special fund” bond measure, a vote of 50% plus 1 of a local electorate can authorize taking on debt and establish the siloed revenue stream to finance it. Yet this opportunity is at risk: a qualified ballot measure on the 2024 ballot would retroactively raise the threshold for voter-initiated tax measures from majority vote to two-thirds.[20] That would undermine the ability of localities to fund the infrastructure needed to serve all the new housing California needs, and roll back the local electorate’s power to decide its own taxes.


Darien Shanske is a Professor of Law at UC Davis School of Law. Grayson Peters is a research fellow at the California Constitution Center at Berkeley Law.

  1. For further discussion of some of these issues, see Darien Shanske, The (Now Urgent) Case for State-Level Monitoring of Local Government Finances (Or, One Way to Protect Localities from Trump’s “Potemkin Villages of Nothing”) (2017) 20 NYU Journal of Legislation and Public Policy 773, 820–24.

  2. See, e.g., City & County of S.F. v. All Persons Interested in the Matter of Prop. G (2021) 66 Cal.App.5th 1058, 1070, as modified on denial of reh’g (Aug. 17, 2021), review denied (Nov. 17, 2021); see also Cal. Cannabis Coalition v. City of Upland (2017) 3 Cal.5th 924, 948, as modified on denial of reh’g (Nov. 1, 2017). See also Darien Shanske and David A. Carrillo, Opinion Analysis: S234148 California Cannabis Coalition v. City of Upland (arguing that City of Upland compels the conclusion that local voters can increase special taxes by initiative with a simple majority because the supermajority limitation does not apply to initiatives any more than the general election requirement applies to initiatives).

  3. Cal. Const., art. XIIIA, § 4.

  4. “We conclude that an ad valorem tax, unlike the [graduated parcel] tax here involved, is any source of revenue derived from applying a property tax rate to the assessed value of property.” Heckendorn v. City of San Marino (1986) 42 Cal.3d 481, 483.

  5. See Borikas v. Alameda Unified School District (2013) 214 Cal.App.4th 135. Though it is beyond our scope here, we do not mean to say that we think this case was rightly decided, as we think the court did not need to interpret “uniformly” as it did. The legislature could (and should) also just amend the statute. For more, see Shanske & Shatara, A Proposal to Increase California School District Autonomy and Funding Through Parcel and Mello-Roos Tax Reforms (2019) 28 Ca. Tax Lawyer 1.

  6. Cal. Federal Savings and Loan Assn. v. City of L.A. (1991) 54 Cal. 3d 1, 12 (“[T]he power of taxation is a power appropriate for a municipality to possess’ […] that proposition [is] ‘too obvious to merit discussion’”) (citing Ex parte Braun (1903) 141 Cal. 204, 209).

  7. Cal. Gov. Code § 37100.5 (“[T]he legislative body of any city may levy any tax which may be levied by any charter city, subject to the voters’ approval pursuant to Article XIII A of the Constitution of California.”)

  8. County electorates also have the power of initiative and counties also have the power to tax, but counties could not craft their own tax: the voters would need to impose a tax authorized by the legislature, such as a traditional parcel tax.

  9. Cal. Gov. Code § 53311 et seq.

  10. “In Matter of Prop. C, this court held that the supermajority vote requirements of article XIII A, section 4 and article XIII C, section 2(d) constrain only local government entities such as the Board of Supervisors, and do not displace the people’s power to enact initiatives by majority vote. (citation omitted) We affirm that holding here and extend it to include article XIII D, section 3(a), rejecting Nowak’s theory that these provisions require a citizens’ initiative enacting a special tax to command a supermajority vote.” City & County of S.F. v. All Persons Interested in the Matter of Prop. G (2021) 66 Cal.App.5th 1058, 1070, as modified on denial of reh’g (Aug. 17, 2021), review denied (Nov. 17, 2021); see also Cal. Cannabis Coalition v. City of Upland (2017) 3 Cal.5th 924, 948, as modified on denial of reh’g (Nov. 1, 2017).

  11. City & County of S.F. v. All Persons Interested in the Matter of Prop. G (2021) 66 Cal.App.5th 1058, 1070, as modified on denial of reh’g (Aug. 17, 2021), review denied (Nov. 17, 2021).

  12. “When the Legislature enacts a statute pertaining to local government, it does so against the background of the electorate’s right of local initiative, and the procedures it prescribes for the local governing body are presumed to parallel, rather than prohibit, the initiative process, absent clear indications to the contrary.” DeVita v. County of Napa (1995) 9 Cal.4th 763, 786.

  13. This happened recently in San Francisco. See City and County of San Francisco Department of Elections, San Francisco Workforce Education and Reinvestment in Community Success Act (May 12, 2022). One of the authors of this post (Shanske) proposed this structure.

  14. Cal. Gov. Code § 53345 et seq.

  15. “No . . . city . . . shall incur any indebtedness or liability in any manner or for any purpose exceeding in any year the income and revenue provided for such year, without the assent of two-thirds of the voters of the public entity voting at an election to be held for that purpose.” Cal. Const., art. XVI, § 18.

  16. “It is settled in California . . . that, as a general rule, a constitutional provision such as section 18 of article XI is not violated by revenue bonds or other obligations which are payable solely from a special fund, provided the governmental body is not liable to maintain the special fund out of its general funds, or by tax levies, should the special fund prove insufficient.” City of Oxnard v. Dale (1955) 45 Cal.2d 729, 733.

    However, the two-thirds vote requirement applies when “the asserted relationship of such funds to the . . . project during the entire life of the bonds is too indirect and intangible to effectively remove what are ordinarily general funds from that category.” City of Redondo Beach v. Taxpayers, Property Owners, Citizens and Electors of City of Redondo Beach (1960) 54 Cal.2d 126, 133. In City of Redondo Beach, “special fund” bonds were to be paid off by general sales and use tax revenues, not by a dedicated funding stream separate from the general city coffers.

  17. Harris, California Constitutional Debt Limits and Municipal Lease Financing (2002) League of California Cities, 3.

  18. 67 Ops. Cal. Atty. Gen. 349, 353 (1984) (citation omitted).

  19. See City of Palm Springs v. Ringwald (1959) 52 Cal.2d 620, 627 (Section 18 “clearly encompasses not alone property taxes but also the other types of ‘income and revenue’ taxes that produce revenue for the general fund.”); id. at 626 (“The sales and use taxes which petitioner proposes to pledge are excise taxes and will constitute general funds of petitioner unrelated to the parking district. Therefore, the fund established by the ordinance here in question does not meet the requirements of the special fund doctrine. Accordingly, the sales and use taxes may not be diverted from petitioner’s general funds for the purpose proposed for a period beyond the year in which received unless there has been compliance with section 18 of article XI of the Constitution.”)

  20. See California Secretary of State, Eligible Statewide Initiatives (Mar. 2023).