Jaime Molera v. reagan/invest in Education

428 P.3d 490

Opinions

PER CURIAM:

*492¶ 1 In this opinion, we explain the reasons for our prior order disqualifying the Invest in Education Act initiative from the November 2018 general election ballot. We greatly respect the initiative process, including the civic activism required to collect the signatures necessary to qualify a ballot measure, and we do not lightly disturb the fruits of such efforts. However, we must do so, as the Court has done in various prior circumstances, when essential requirements necessary to qualify a measure are not adequately followed. We hold here that the initiative's proponents did not comply with the requirements of A.R.S. § 19-102(A) because their description of the initiative's principal provisions omitted material provisions and created a significant danger of confusion or unfairness to those who signed petitions to place the measure on the ballot.

I.

¶ 2 On April 30, 2018, the Invest in Education Committee ("Committee") filed with the Secretary of State a proposed initiative called the "Invest in Education Act," which would increase K-12 education funding and raise certain income tax rates to support it. As required by § 19-102(A), the Committee prepared a 100-word initiative description for placement on the petitions to qualify the measure for the ballot. The description provided:

The Invest in Education Act increases the classroom site fund by raising the income tax rate by 3.46% on individual incomes over a quarter million dollars (or household incomes over half a million dollars), and by 4.46% on individual incomes over half a million dollars (or household incomes over a million dollars); designates 60% of new funds for teacher salaries and 40% for operations; adds full day kindergarten and pay raises for student support services personnel as permitted fund uses; requires governing boards seek teacher and personnel input on fund use plans; defines teacher and student support services personnel.

¶ 3 The petitions also contained the following language required by § 19-102(A) :

Notice: This is only a description of the proposed measure (or constitutional amendment) prepared by the sponsor of the measure. It may not include every provision contained in the measure. Before signing, make sure the title and text of the measure are attached. You have the right to read or examine the title and text before signing.

¶ 4 On July 5, the Committee submitted approximately 270,000 signatures to the Secretary of State in support of the initiative. Although the Secretary invalidated some petition sheets, she determined that the Committee filed a sufficient number of valid signatures to qualify the initiative for the ballot.

¶ 5 Petitioners Jaime Molera and Jennifer Henricks ("Petitioners") filed a special action in Maricopa County Superior Court seeking to invalidate the initiative because (1) the 100-word initiative description was misleading in that it mischaracterized the size of the proposed tax increase and omitted a change in income tax indexing; and (2) although § 19-102(D) requires a circulator to check a box on petition sheets to indicate whether he or she is paid, a third party pre-marked the boxes on most petition sheets. The Committee filed a cross-complaint challenging the constitutionality of A.R.S. § 19-102.01(A), which requires strict compliance with constitutional and statutory requirements for statewide initiatives. The Committee also sought to restore some of the petition sheets invalidated by the Secretary of State.

¶ 6 The superior court ruled that § 19-102.01 is unconstitutional, that both the 100-word description and the pre-checked circulator boxes satisfied statutory requirements, and that the Secretary of State erroneously excluded some petition sheets. The court thus concluded that the initiative was eligible for the ballot.

¶ 7 Pursuant to A.R.S. § 19-122(C), Petitioners filed an expedited appeal in this Court contesting all but the last of those rulings. Following our review, we issued an order determining that the 100-word initiative description created a significant danger of confusion or unfairness, thus invalidating the petition. As a result, we do not decide the *493other issues raised in the appeal. We set forth the reasoning for our conclusion below.

¶ 8 The only issue before us involves interpretation and application of constitutional and statutory provisions regarding initiatives, which we review de novo. See Pedersen v. Bennett , 230 Ariz. 556, 558 ¶ 6, 288 P.3d 760, 762 (2012). We have jurisdiction over this matter pursuant to article 6, section 5(3) of the Arizona Constitution.

II.

¶ 9 The Arizona Constitution reserves to this state's citizens the power to propose and enact laws by initiative. Ariz. Const. art. 4, pt. 1, § 1 (1)-(2). Under our constitutional separation of powers, the courts must not intrude upon the people's power to legislate, subject to constitutional and proper statutory requirements. See Kromko v. Superior Court , 168 Ariz. 51, 57-58, 811 P.2d 12, 18-19 (1991). This Court has observed that the citizens' legislative authority "is as great as the power of the Legislature to legislate." State ex rel. Bullard v. Osborn , 16 Ariz. 247, 250, 143 P. 117 (1914) ; accord Cave Creek Unified Sch. Dist. v. Ducey , 233 Ariz. 1, 4 ¶ 8, 308 P.3d 1152, 1155 (2013). Indeed, with the enactment through initiative of the Voter Protection Act, legislation enacted by the voters is even more consequential, such that the legislature cannot repeal an initiative-enacted law and may only modify it by a three-fourths vote when the changes further the law's purposes. See Ariz. Const. art. 4, pt. 1, § 1 (6)(C); see, e.g. , State v. Maestas , 244 Ariz. 9, 13-14 ¶¶ 19-20, 417 P.3d 774, 778-79 (2018) (striking down legislation restricting the possession of marijuana on college campuses because it did not further the purposes of the Arizona Medical Marijuana Act); Cave Creek , 233 Ariz. at 4, 7-8 ¶¶ 9, 25, 308 P.3d at 1155, 1158-59 (concluding that legislative adjustments to voter-approved funding scheme for public education violated the Voter Protection Act).

¶ 10 Just as the legislature must comply with restrictions on its lawmaking powers, see, e.g. , Ariz. Const. art. 4, pt. 2, § 19 (prohibiting the legislature from enacting local or special laws); Ariz. Const. art. 21, § 1 (requiring the legislature to refer constitutional amendments to voters separately), so too must the people comply with appropriate regulation of the initiative process. Article 4, part 1, section 1(14) of the Arizona Constitution provides that the initiative power "shall not be construed to deprive the legislature of the right to enact any measure except that the legislature shall not have the power to adopt any measure that supersedes" an enacted initiative. Further, article 7, section 12 directs the legislature to enact "registration and other laws to secure the purity of elections and guard against abuses of the elective franchise."

¶ 11 Thus, although our decisions safeguard the voters' legislative power, this Court in many cases has invalidated citizen initiatives and referenda that did not comply with applicable requirements. See, e.g. , Transp. Infrastructure Moving Ariz.'s Econ. v. Brewer , 219 Ariz. 207, 211-14 ¶¶ 17-36, 196 P.3d 229, 233-36 (2008) (upholding dismissal of challenge to Secretary of State's invalidation of ballot measure signatures as time-barred by statute); Taxpayer Prot. All. v. Arizonans Against Unfair Tax Schemes , 199 Ariz. 180, 181-82 ¶¶ 1-8, 16 P.3d 207, 208-209 (2001) (invalidating initiative for violation of constitutional single-subject rule); McDowell Mountain Ranch Land Coal. v. Vizcaino , 190 Ariz. 1, 3-5, 945 P.2d 312, 314-16 (1997) (disqualifying petition signatures gathered by referendum circulators who were not qualified electors); Perini Land & Dev. Co. v. Pima County , 170 Ariz. 380, 382-84, 825 P.2d 1, 3-5 (1992) (holding referendum would not appear on ballot for failure to comply with signature requirement); W. Devcor, Inc. v. City of Scottsdale , 168 Ariz. 426, 428-32, 814 P.2d 767, 769-773 (1991) (determining referendum petitions were invalid because they did not contain required circulators' statements); Saggio v. Connelly , 147 Ariz. 240, 241-42, 709 P.2d 874, 875-76 (1985) (holding initiative invalid for failing to propose a law or ordinance); Cottonwood Dev. v. Foothills Area Coal. of Tucson, Inc. , 134 Ariz. 46, 48-50, 653 P.2d 694, 696-98 (1982) (invalidating referendum petitions that did not attach resolution); Direct Sellers Ass'n v. McBrayer , 109 Ariz. 3, 5-6, 503 P.2d 951, 953-54 (1972) (holding referendum petitions invalid where *494amendments were not made within time limits); Kerby v. Griffin , 48 Ariz. 434, 446-56, 62 P.2d 1131, 1143-53 (1936) (holding that initiative failed to comply with publication requirement).

¶ 12 Challengers are also required to conform to statutory requirements. Two years ago, for example, in the context of an initiative that proposed raising the state's minimum wage, the trial court found the measure lacked sufficient valid signatures to qualify for the ballot. We vacated the judgment, reasoning that the opponents failed to meet the statutory deadline to file the challenge, thus allowing the initiative to proceed. Hitzeman v. Reagan , No. CV-16-0204-AP/EL, slip op. at 1-2 (Ariz. Aug. 30, 2016) (decision order). We explain below how the proponents here failed to meet the statutory requirements to qualify this measure for the ballot.

III.

¶ 13 The statutory provision pertinent to our analysis is § 19-102(A), which requires an initiative's sponsors to provide on the petition "a description of no more than one hundred words of the principal provisions of the proposed measure or constitutional amendment." The description need not be impartial. See Save Our Vote, Opposing C-03-2012 v. Bennett , 231 Ariz. 145, 152 ¶ 28, 291 P.3d 342, 349 (2013). Nor must the description detail every provision, as the statutorily required disclaimer acknowledges. § 19-102(A). However, the description will require us to invalidate the petition if "it is fraudulent or creates a significant danger of confusion or unfairness." Save Our Vote , 231 Ariz. at 152 ¶ 26, 291 P.3d at 349 (citation omitted).

¶ 14 Petitioners assert that the description of the proposed Invest in Education Act violates the applicable requirements in two ways. First, it fails to mention that the measure modifies the inflation indexing of income tax rates that was adopted in 2015, thus exposing most taxpayers to tax increases. Currently, tax brackets and rates are set by A.R.S. § 43-1011(A)(5). The income dollar amounts for each tax bracket are indexed for inflation by § 43-1011(C). In other words, the tax brackets are adjusted by the rate of inflation so that as incomes rise, they are not subject to higher rates of taxation simply because of inflation. The proposed initiative replaces those brackets with new tax rates and brackets in § 43-1011(A)(6), restoring the pre-indexed brackets for individual taxpayers making less than $250,000 and for married taxpayers filing jointly making less than $500,000, and adding new brackets for taxpayers earning more than $250,000. The proposed initiative makes subsection (A)(6) "[s]ubject to" subsection C, the indexing provision, but that provision applies only to the old tax brackets set forth in § 43-1011(A)(5). By restoring the old pre-indexing tax brackets, Petitioners argue, the initiative would reverse indexing back to 2015 and place taxpayers in higher tax brackets; and by failing to apply the indexing provision to the new tax brackets in § 43-1011(A)(6), the initiative would repeal indexing going forward. This, they assert, would lead to higher taxes for most taxpayers, not just those earning more than $250,000 as stated in the initiative description.

¶ 15 Petitioners argue that the description's omission is deceptive, or confusing at best, in that petition signers were led to believe that only very wealthy taxpayers' rates would be increased. The Committee responds, and the Chief Justice's dissent asserts, that the initiative does not affect indexing as the new proposed tax rates and brackets in § 43-1011(A)(6) are "[s]ubject to" the indexing provision of § 43-1011(C). If the measure does change indexing, the Committee asserts, it does so inadvertently, not intentionally.

¶ 16 Second, Petitioners argue that the description of the funding source-"raising the income tax rate by 3.46% ... and by 4.46%" on specified taxpayers-is misleading and confusing in that it suggests a modest tax increase. In reality, Petitioners assert, the measure would raise the applicable rates by seventy-six percent and ninety-eight percent, respectively. The Committee responds that the description is accurate, and that any confusion could be alleviated by reading the initiative text, as the notice on the petition provides.

¶ 17 We address these two issues in turn.

*495A.

¶ 18 The dispute over income tax indexing did not arise in the first instance in this lawsuit. Rather, it arose in the context of an analysis by the Legislative Council, whose nonpartisan staff reviews proposed bills for legislators of both parties. Although the Legislative Council opinion is not binding, had the challengers here sought such an analysis, the question might never have become a judicial one and the measure might well be before the voters.

¶ 19 Apparently recognizing the hazards inherent in initiative drafting, the legislature enacted A.R.S. § 19-111.01, which provides that upon filing an application for an initiative petition and a statement of organization, a political committee may submit a copy of the proposed text of the measure to the Legislative Council. Within thirty days, the Legislative Council staff must review the measure, limiting its consideration (as relevant here) to "errors in the drafting of the measure" and "confusing, conflicting or inconsistent provisions within the measure." § 19-111.01(B). The initiative's proponents may then "accept, modify or reject any recommendations made by the legislative council staff regarding the text of the measure solely in [their] discretion." § 19-111.01(C). In other words, if the proponents had requested an analysis as soon as the application was filed, they would have received an analysis within thirty days, well in advance of the filing deadline, and they could either have modified the text or description or taken the risk of having it invalidated. But the Committee failed to seek the Legislative Council's review.

¶ 20 The measure's opponents, in contrast, did avail themselves of this process while petitions were circulating. On June 20, Legislative Council staff issued its analysis, noting that it is "similar to the review that our office would conduct for a legislator on any draft of proposed legislation in that it identifies potential issues." The analysis noted multiple concerns, including the measure's effect on income tax indexing. Specifically, it explained that the measure would create new income tax brackets in § 43-1011(A)(6) and that the paragraph is subject to § 43-1011(C). However, the analysis concluded that the reference to subsection C is "meaningless" because subsection C "require[s] inflation adjustments only to the tax brackets prescribed in A.R.S. section 43-1011, subsection A, paragraph 5." Because subsection C "do[es] not apply to the new A.R.S. section 43-1011, subsection A, paragraph 6 ... the prefatory language does nothing." The analysis went on to observe that "[i]t might be that the drafters of the initiative intended that the dollar amounts in the new tax brackets be adjusted for inflation. The language of the initiative does not accomplish this purpose, however."

¶ 21 The Legislative Council staff was not alone in this analysis. Pursuant to A.R.S. § 19-123(E), the Joint Legislative Budget Committee ("JLBC") staff is required to prepare a fiscal-impact summary for voter-initiated ballot measures. The JLBC concluded the initiative "reinstates the individual income tax brackets for incomes up to $250,000 in effect in 2014." In its accompanying analysis, the JLBC staff explained that as a result, "most taxpayers would have a small portion of their income taxed in a higher bracket, resulting in a small increase on most taxpayers." By way of illustration, as a result of indexing, married taxpayers filing jointly and earning $108,617 in 2019 would have the same tax rates as married taxpayers filing jointly and earning $100,000 in 2014, according to the JLBC. But because the initiative would restore the brackets to 2014 levels, such taxpayers would be subject to a higher tax rate on the amount earned above $100,000. Eliminating indexing would affect tax rates at all income levels. And because indexing would not apply at all to the new brackets in § 43-1011(A)(6), the relatively modest initial tax impact would expand over time, resulting in a tax increase to most taxpayers.

¶ 22 We agree with Petitioners and the Legislative Council that the initiative's "subject to" language, even if intended to do so, does not preserve tax indexing. Section 43-1011(C) provides indexing "for each rate bracket prescribed by subsection A, paragraph 5." The proposed new subsection (A)(6) establishes new tax brackets after *4962018, supplanting the tax brackets in subsection (A)(5). In other words, the "subject to" language circles us back to a subsection that is no longer operational. Even were we to credit intent rather than text, the new subsection (A)(6) establishes tax brackets "for taxable years beginning from and after December 31, 2018." Those brackets, on their face, erase the effects of indexing since 2015. And because both the "subject to" and indexing provision relate only to the tax rates in subsection (A)(5), which would no longer be active, the result is that the new tax brackets would not be indexed going forward.

¶ 23 That conclusion requires us to determine whether the indexing changes are "principal" provisions whose omission from the initiative description must disqualify the measure from the ballot. We conclude they are.

¶ 24 Because the statute does not define "principal provisions," we apply the term's common meaning. In Sklar v. Town of Fountain Hills , the court of appeals disqualified a citizen referendum because, as here, the petition failed to adequately describe the measure's principal provisions. 220 Ariz. 449, 453-55 ¶¶ 12-22, 207 P.3d 702, 706-08 (App. 2008). The court consulted dictionaries to determine the meaning of "principal" and found that the plain meaning includes "most important, consequential, or influential," "chief," and "a matter or thing of primary importance." Id. at 453 ¶ 13, 207 P.3d at 706 (internal quotation marks and citations omitted).

¶ 25 The change in indexing is a primary, consequential provision because it imposes tax increases on most Arizona taxpayers rather than only the state's wealthiest taxpayers, as the description clearly suggests. Indeed, identifying the source of new revenue was so significant to the initiative proponents that it is set forth in some detail in the opening part of the description. A description indicating that other people's taxes will be raised, but not the taxes of most of those signing the petition, creates a significant risk of confusion or unfairness and could certainly materially impact whether a person would sign the petition. Thus, the failure to disclose the measure's impact on tax indexing constitutes the omission of a principal provision that renders the initiative invalid.

¶ 26 The Chief Justice's dissent observes that the Committee disputes this reading of the measure's effects on income tax indexing, contending that we should extend the "subject to" language to the new tax brackets in subsection (A)(6) to avoid those effects, and that the proper place to resolve this dispute is at the ballot box rather than the courtroom. See infra ¶ 43. Although we try to give effect to all of a statute's words in order to resolve ambiguity, in this instance doing so would require us not merely to construe those words but to rewrite the proposed statute. That invitation must be rejected on separation-of-powers grounds. Just as we cannot rewrite statutes to smooth their rough edges, see, e.g. , City of Phoenix v. Butler , 110 Ariz. 160, 162, 515 P.2d 1180, 1182 (1973), so may we not rewrite proposed initiatives to arrest unintended consequences. We note that the measure's drafters could have corrected the errors themselves had they availed themselves of the Legislative Council's expertise. Their failure to do so does not empower us to do so now. Moreover, rewriting the proposed statute to preserve income tax indexing would necessarily substantially decrease the amount of revenue that the measure would generate. We cannot substitute our judgment on such a consequential matter for the plain words chosen by the measure's drafters.

¶ 27 Relatedly, Justice Timmer asserts in dissent that § 19-102(A) only requires the petition description to describe "known " principal provisions and not unintended consequences, so that the measure should proceed to the ballot. See infra ¶ 51. We hold § 19-102(A) requires an objective standard for evaluating the description of the actual provisions rather than crediting the drafters' subjective intent. Section 19-102(A) does not qualify "principal provisions" with either "known" or "intended." The purpose of the petition description is to inform prospective signers of the measure's principal provisions so they may determine whether to endorse it for the ballot. This Court has long held that the proper remedy for failure to satisfy statutory prerequisites is to enjoin the *497measure from appearing on the ballot. See, e.g. , Kerby , 48 Ariz. at 445, 62 P.2d 1131 ; cf. City & County of Honolulu v. State of Hawai'i , No. SCPW-18-0000733, slip op. at 1-2, 2018 WL 5117002 (Haw. Oct. 19, 2018) (order) (removing education funding measure from the ballot because it did not satisfy statutory requirement that the language must be "clear [and] neither misleading nor deceptive"). Our failure to determine whether the description omits a principal provision before the measure appears on the ballot would reward sloppy or even deceptive drafting, and would render the statutory transparency requirement meaningless because it would allow a measure to proceed even if voters signing the petition were not made aware of principal provisions.

¶ 28 Moreover, recourse here to the measure's text to correct any uncertainty is unavailing because that text is the source of the problem. Because the omission of the principal provision violates § 19-102(A) and creates a substantial danger of confusion or unfairness, the proper remedy is removal from the ballot. See Sklar , 220 Ariz. at 455 ¶ 22, 207 P.3d at 708. Indeed, were the measure to proceed and win voter approval, the legislature's authority to restore income tax indexing, as the proponents insist they intended, would be greatly circumscribed by the Voter Protection Act, so that a substantive fix might well require a second initiative. All of that underscores the important purposes served by the statutory requirement to describe an initiative's principal provisions in the petition.

B.

¶ 29 The petition's description of the magnitude of the tax increase on wealthy taxpayers also "creates a significant danger of confusion." Save Our Vote , 231 Ariz. at 152 ¶ 26, 291 P.3d at 349. The petition description stated that the measure would increase taxes on wealthy Arizonans by 3.46% and 4.46%, which on its face seems modest. However, the affected tax rates would actually increase by seventy-six percent and ninety-eight percent, respectively. This difference is so significant that it could materially affect whether a person would sign the petition, as it is one thing to increase someone's taxes by between three and four percent and another to nearly double them. Indeed, as the JLBC observed, the changes would make Arizona's top tax rates the fifth highest in the country.

¶ 30 Given the description's required brevity, the initiative sponsors need not have described this change in great detail, but they may not describe it in a confusing way. Had they simply changed the wording, saying that the rates applicable to the two high-income categories would be increased by 3.46 and 4.46 percentage points, the description would have been much clearer. Indeed, the Chief Justice's dissent acknowledges that this wording "could have resolved the ambiguity." See infra ¶ 40. Instead, by choosing to describe the increase in percentage terms, the initiative proponents made it appear more likely that the magnitude of the increase was slight rather than substantial. Applying basic mathematics principles, "[i]f a quantity is increased by a percentage, then that percentage of the quantity is added to the original." Vassilis C. Mavron & Timothy N. Phillips, Elements of Mathematics for Economics and Finance 11 (2007). Thus, increasing the prior rate "by 3.46 percent" would change it only from 4.54% to 4.7%, and "by 4.46 percent" from 4.54% to 4.74%-quite a difference from the actual new rates of eight percent and nine percent in the initiative.

¶ 31 At best, the two possible interpretations of the "by [x] percent" language yield a significant danger of confusion. The Chief Justice's dissent characterizes the wording as ambiguous rather than confusing. See infra ¶ 40. Ambiguity is the root of confusion. Where the description lends itself to two sharply divergent interpretations with very different and significant ramifications, the danger of confusion is sufficiently great that it undermines any assurance that the voters received adequate notice of what they were signing.

¶ 32 The Chief Justice's dissent further asserts that voters who were confused could read the actual language of the text to clear up the ambiguity. See infra ¶ 41. But the description itself must be adequate in its description of the principal provisions to avoid confusion. The obvious purpose of *498§ 19-102(A)"is to ensure that the public has immediate and full disclosure" of the initiative's principal provisions. Sklar , 220 Ariz. at 454 ¶ 17, 207 P.3d at 707. The description need not encompass minor provisions and may be presented in a biased manner, but it may not create a substantial danger of confusion or unfairness. To hold that such a confusing description is permissible because the truth may be discovered in the many pages of the initiative, or that the proponents actually intended something different from what the words they chose to use indicate, is to eviscerate the description requirement and its important purposes of transparency, fairness, and disclosure.

IV.

¶ 33 The omission of the change in tax indexing paired with the confusing language about the magnitude of tax increases makes it clear that petition signers were not adequately informed about what they were signing, as the requisite description failed to provide adequate notice of the measure's principal provisions as required by § 19-102(A).

¶ 34 Accordingly, we reverse the superior court's decision.

BALES, C.J., joined by TIMMER, J., dissenting.

¶ 35 I agree with the majority that better drafting of the 100-word description and Proposition 207 itself could have avoided the issues addressed in today's opinion. But we have never required perfection in drafting as a condition for the valid exercise of legislative authority, and doing so with initiatives would infringe upon the people's constitutional right to enact laws independently of the legislature. Ariz. Const. art. 4, pt. 1, § 1 (1). Because I do not believe that the 100-word description presents a substantial danger of fraud, confusion, or unfairness sufficient to invalidate the initiative petitions, I respectfully dissent from my colleagues' decision to strike Proposition 207 from the ballot.

¶ 36 The description for Proposition 207 states:

The Invest in Education Act increases the classroom site fund by raising the income tax rate by 3.46% on individual incomes over a quarter million dollars (or household incomes over half a million dollars), and by 4.46% on individual incomes over half a million dollars (or household incomes over a million dollars); designates 60% of new funds for teacher salaries and 40% for operations; adds full day kindergarten and pay raises for student support services personnel as permitted fund uses; requires governing boards seek teacher and personnel input on fund use plans; defines teacher and student support services personnel.

¶ 37 As required by A.R.S. § 19-102, the description appeared on the petition signature sheets along with this disclaimer:

Notice: This is only a description of the proposed measure (or constitutional amendment) prepared by the sponsor of the measure. It may not include every provision contained in the measure. Before signing, make sure the title and text of the measure are attached. You have the right to read or examine the title and text before signing.

¶ 38 The majority accepts the challengers' arguments that the description poses a danger of confusion or unfairness by its "description of the change in the tax rate combined with the omission of any discussion of changes in indexing for inflation."

¶ 39 If passed, Proposition 207 would have increased the tax rate from 4.54% to 8.00% on individual incomes from $250,001 to $500,000 and from 4.54% to 9.00% for individual incomes over $500,000 (with corresponding increases for joint filers). Thus, two new tax brackets would have been created. The tax rates for amounts up to $250,000 would remain the same regardless of a person's total income. The proposed new tax rates are "marginal tax rates" in that they only apply to amounts of income in excess of the first $250,000.

¶ 40 The issue regarding the tax rate reflects that changes in percentage rates can refer to either absolute or relative changes. For example, if a percentage rate increases from 6% to 8%, the difference in absolute terms is 2%. The relative change, in contrast, *499is an increase of 33% (8% minus 6%, divided by 6%). In this respect, the 100-word description is not inaccurate but ambiguous. In terms of absolute change, the Proposition would-as the description stated-raise the tax rates by 3.46% (8/100 - 4.54/100 = 3.46/100 = 3.46%) and 4.46%. But in terms of relative change, as the majority correctly notes, the increase in the marginal rates would be approximately 76% (3.46/4.54) and 98% (4.46/4.54). To avoid ambiguity in describing changes in percentage rates, the commonly accepted convention is to use the term "percentage points" to refer to absolute changes. For example, here the 100-word description could have resolved the ambiguity by saying that Proposition 207 would raise the income tax rate by 3.46 percentage points instead of "by 3.46%." But this does not mean that the description, as worded, created a substantial danger of confusion or unfairness. See Save Our Vote, Opposing C-03-2012 v. Bennett, 231 Ariz. 145, 152 ¶ 27, 291 P.3d 342, 349 (2013).

¶ 41 In assessing the adequacy of a 100-word description, this Court considers not only the description itself, but also the fact that it is accompanied by a notice stating it is only a description prepared by the sponsor and alerting petition signers to their right to review the attached initiative's title and text before signing. See id. ; Wilhelm v. Brewer , 219 Ariz. 45, 48-49 ¶¶ 13-15, 192 P.3d 404, 407-08 (2008). The description clearly states that Proposition 207 would increase the tax rates on individual incomes greater than $250,000 (or twice that amount for joint filers). If-as the majority supposes-a prospective signer was confused about the exact quantitative change, the accompanying text clearly reflects in section 3 that Proposition 207 would add two new income tax brackets to A.R.S. § 43-1011 that would increase the marginal tax rates from 4.54% to 8% for individuals earning $250,001 to $500,000 and 9% for those earning more than $500,000 (or joint filers earning twice those amounts). Notably, even if the description had used the phrase "percentage points" instead of saying that the rates would be raised by 3.46% and 4.46%, a signer still would have had to refer to the text itself to identify the resulting tax rates. Although the majority is concerned that the description understated the proposed tax increases, the description also arguably overstated them by not explaining they were increases in only marginal rates rather than rates applicable to total income.

¶ 42 Nor is the 100-word description flawed for omitting any discussion of changes in the indexing of taxes for inflation. The 100-word description is not a complete description-it need only describe the principal provisions of a measure. A.R.S. § 19-102. See also Save Our Vote, 231 Ariz. at 152 ¶ 27, 291 P.3d at 349. Because Proposition 207 does not purport to eliminate inflation indexing, it is not surprising that the description does not discuss this matter. Even if the Proposition were ambiguous in its possible effects on indexing, that would not provide grounds for striking the measure from the ballot, but instead would properly be a subject of debate between proponents and opponents in their seeking to persuade the voters. Cf. Winkle v. City of Tucson , 190 Ariz. 413, 418, 949 P.2d 502, 507 (1997) ("... this court should not create an impediment to the exercise of one of our state government's bedrock institutions ... In a democracy, the process itself is often as valuable as the result."). For example, had the measure appeared on the ballot, the Secretary of State's Publicity Pamphlet for the 2018 Election would have included a fiscal impact summary prepared by the JLBC discussing the effects of eliminating indexing-an effect disputed by the proponents of Proposition 207.

¶ 43 The proponents of Proposition 207 disclaim any intent to affect inflation-based tax indexing. Rather than resolve the interpretative issue now (which, after all, concerns the effect of a proposed rather than enacted law), we should instead allow the voters to consider the competing arguments about inflationary indexing in deciding how to cast their votes. But if we must decide the issue, the proponents' reading of Proposition 207 is the one better supported by the measure's text. Currently, A.R.S. § 43-1011(A)(5) identifies five income tax brackets for taxable years beginning after December 1, 2006, "subject to subsections B and C" of § 43-1011. Subsections B and C in turn provide for the indexing of the brackets identified in *500subsection (A)(5). Proposition 207 would not repeal subsections (A)(5), (B), or (C). Instead, it would add a new (A)(6) for taxable years beginning after December 31, 2018 that restates the five (A)(5) brackets for incomes up to $250,000 and adds two new brackets for individual incomes exceeding $250,000 and $500,000 (or twice those amounts for joint filers). The new (A)(6), like (A)(5), is expressly subject to subsections (B) and (C) of § 43-1011 -the subsections that provide for inflation-based indexing.

¶ 44 On its face, Proposition 207 would not repeal the indexing of the tax brackets for incomes of $250,000 or less. The measure leaves in place the existing brackets for those income levels and it expressly states that it is subject to the subsections providing for indexing. While subsections (B) and (C) refer to indexing the brackets identified in (A)(5), those same brackets are echoed in proposed (A)(6). The only way to conclude that (A)(6) eliminates indexing is to assume that its reference to subsections (B) and (C) is meaningless and that it implicitly repeals rather than coexists with (A)(5) for tax years after 2018. Our usual approach, however, is to interpret statutes to harmonize and give effect to all their provisions. See David C. v. Alexis S. , 240 Ariz. 53, 55 ¶ 9, 375 P.3d 945, 947 (2016). And even if Proposition 207 would eliminate indexing for the new brackets for incomes over $250,000 and $500,000, such a change could hardly be called a "principal provision" of the measure that needed to be described in the description. Cf. Save Our Vote , 231 Ariz. at 152 ¶ 27, 291 P.3d at 349 (noting that A.R.S. § 19-102(A)"requires only a description of the principal provisions, not a complete description").

¶ 45 Finally, if the majority were correct that Proposition 207 would eliminate tax indexing for all brackets, JLBC's analysis shows that the change would result in only a "small increase" in taxes for most taxpayers. For instance, JLBC estimated that a married couple with a taxable income of $101,000 would pay approximately $34 more in income tax due to this change, and that tax filers with adjusted gross incomes ranging from $0 to $200,000 (nearly 90% of all tax filers) would on average pay only $12 more annually. None of the additional revenues from eliminating indexing on the lower tax brackets would be dedicated to the Classroom Site Fund by Proposition 207. That Proposition 207 conceivably could result in "small increases" in taxes going to the general fund hardly makes that prospect a "principal provision" of a ballot measure aimed at increasing school funding by raising the marginal tax rates for higher income earners.

¶ 46 At bottom, the majority concludes that Arizona's voters should not be allowed to consider a legislative proposal supported by hundreds of thousands of petition-signers because the description used the % symbol instead of the words "percentage points" and the measure did not add "or 6" to A.R.S. § 43-1011(C). Even if one agrees with that analysis, it should give one pause that the conclusion only comes after untold hours of volunteer efforts over many months to place the measure on the ballot and then accelerated litigation in the trial court and this Court. The initiative provisions of the constitution are self-executing and do not require legislation to be effective. Ariz. Const. art. 4, pt. 1, § 1 (1).

¶ 47 Our state could be well served, however, by legislation affording some pre-circulation review of ballot measures and their summaries and a resulting safe-harbor that could avoid measures being struck from the ballot only after their circulation and with substantial support by qualified electors. Cf. A.R.S. § 19-111.01 (allowing for non-binding review and recommendations on ballot measures by legislative council). Although the majority notes that the proponents could, and the challengers did, seek review and comment by the legislative council regarding Proposition 207 under A.R.S. § 19-111.01, that review only came late in the process-about two weeks before the signature filing deadline-and it has no binding effect in terms of a measure's validity.

¶ 48 Because I do not think that Proposition 207's description is fatally flawed, and I would otherwise affirm the trial court's judgment leaving the measure on the ballot, I respectfully dissent.