California Constitutional Law: Direct Democracy – Article by David A. Carrillo, Stephen M. Duvernay, Benjamin Gevercer & Meghan Fenzel


From Volume 92, Number 3 (March 2019)


California Constitutional Law: Direct Democracy

David A. Carrillo,[*] Stephen M. Duvernay,[†] Benjamin Gevercer[‡] & Meghan Fenzel[*][*]

The California electorate amended the state constitution in 1911 to reserve to itself the powers of initiative, referendum, and recall. Most research on direct democracy in California focuses on its political science effects. We consider the substantive constitutional issues the electorate’s powers create and present a defense of direct democracy as a net positive force in California government.

We review every California constitutional amendment to date, distinguishing between legislatively proposed amendments and initiative amendments. We solve the enduring mystery of how many times the California constitution has been amended. We prove that the initiative process does not have a disproportionate effect on the amendment rate of the California constitution, and that the state legislature (not the electorate) is responsible for the vast majority of California’s constitutional changes. We also debunk the myths that California’s is the longest constitution in the world and that the state uses the initiative more than any other.

Next, we discuss the substantive constitutional issues the electorate’s direct democracy powers can raise. Critics frequently blame the initiative for many of the state’s woes, but we argue that direct democracy in California is a net social good. We show that while direct democracy’s cumulative quantitative and individual qualitative effects are indeed significant, they are not so severe that structural change is warranted. We identify one flaw in the initiative process that merits a solution. Recognizing, however, that any change is an unlikely prospect, we argue that the existing checks on the electorate are capable. Because direct democracy’s harms are adequately mitigated, there is no urgent need for fundamental change.



A. Overview

B. Direct Democracy’s Design Considerations


A. 1849–1911: No Direct Democracy

B. 1911: Direct Democracy Is Added to the State

C. California’s Direct Democracy Tools Described

D. Constitutional Amendments Before 1912

E. Less Initiative Activity 1912–1959, More Initiative
Activity 1960–2017

F. Initiative Activity Quantitative Analysis


A. Complaints About Direct Democracy

B. Population Size

C. Money’s Impact

D. Effect on Turnout

E. California’s Lack of a True Majority Requirement
Harms Legitimacy

F. A Proposed Solution: The Dual Approval Quorum


A. Single-Subject Rule

B. Revision and Amendment

C. Separation of Powers

D. Individual Rights





A.  Overview

This Article addresses an oft-debated question in California—just how problematic is direct democracy?—by challenging the premise. We quantitatively analyze how the electorate acts in California’s hybrid republic and show how that system prevents the electorate from unbalancing it. We reviewed all California constitutional amendments, parsing them between legislative and initiative. Our data show that the legislature is primarily responsible for constitutional change in California, not the electorate. We analyzed the initiative’s effects on the amendment rate, turnout, and other practical effects, and our results contradict the conventional wisdom that the initiative has disproportionate effects. Our substantive analysis similarly concludes that despite some notable outliers and one fixable problem, overall the existing checks on direct democracy are suited to the task. As a result, this Article stands apart from most scholarly work on California’s direct democracy tools: this is a defense of California’s hybrid republic.[1]

First, some conceptual definitions. Popular sovereignty and direct democracy are related but not synonymous terms; so too are “people” and “electorate” related but distinct. The people is the collective body of persons who constitute the state. The electorate is the subset of the people who can vote. We use popular sovereignty to describe the idea that in California, the people hold ultimate political power and delegate it to a government that persists only with their consent.[2] Direct democracy is any mechanism for an electorate to exercise political power without an intervening representative.[3]

California’s constitution has four direct democracy tools: initiative statute, initiative constitutional amendment, recall, and referendum. We focus on the initiative power to amend the state constitution. We divide California’s experience with popular constitutional change into three distinct periods. From the state’s creation in 1849 until 1911, the state constitution had no provision for any popular legislating. From 1911 (when the state’s direct democracy tools were instituted) to 1959, there was some direct constitutional change, but less compared with the period from 1960 to the present.

The other states vary widely in their constitutional change mechanisms.[4] As Appendix Table 1 (initiative states) illustrates, California is one of 24 states with the initiative (18 of which permit initiative constitutional amendments); every state has a legislative process for the government to place issues on the ballot; and every state except Delaware requires a popular vote to approve constitutional amendments.[5] This means that today most Americans live in the kind of hybrid republic that exists in California, where the state government includes both representative and direct democracy.[6]

B.  Direct Democracy’s Design Considerations

Direct democracy presents value-set tradeoffs between more public participation in lawmaking (which effectuates majority preferences but can be inefficient) and more government control (which may be more efficient but could compromise individual liberty). Overvaluing either principle (participation or efficiency) encourages extreme forms of government: mob rule or dictatorship. Avoiding either extreme requires adjusting the value set to achieve whatever the society finds is the most workable balance between direct popular participation and representative republicanism. Finding that balance is a process rather than a one-time event, and because the balance may change as a society evolves, the system needs a change mechanism to adjust as necessary.

Perspectives on how much direct popular control is best vary by time and location. For example, early American political thinking held that a political system’s successful functioning depended on striking and maintaining a proper balance between the government’s power and the people’s liberty.[7] This view divides the people and their government into two distinct groups with opposing interests that must be balanced to prevent either anarchy or tyranny.[8] The designers of the federal government intentionally eliminated direct popular participation almost entirely.[9] The representative republic designed by the 1787 convention excluded any direct popular involvement in lawmaking other than electing representatives, and the checks and balances in the divided-powers structure of that government were primarily aimed at controlling the government’s power, not permitting public participation.[10] In contrast, the early states experimented with incorporating direct popular lawmaking in their state constitutions.[11] California itself is a miniature example of this variation: its original 1849 constitution had no direct democracy features, and the state rebalanced its value-set choices in 1911, when it incorporated direct democracy tools into its current state constitution. These differences between the federal and state governments, and between early and current versions, do not necessarily indicate progressive thinking or show that one variant is superior; they are different charters for different purposes.

Direct democracy is not an inherent good and adding it to a government requires proper integration. Like any other power in a government, it may evolve beyond its limits and come to dominate the others.[12] Indeed, any divided-government system suffers from an inherent design problem: it is necessary to balance the risk that government gains too much power against the risk that containing its power prevents government from functioning at all. Direct democracy is no different from any other government design feature—for direct democracy in California to work well, it must function as a part of the state government, not as an outside actor. It must be included in the checks and balances to maintain both the optimal balance of internally separated powers and the external balance between the government and the governed.[13]

This is because the electorate is no less given to abusing its power than any other political actor.[14] Changing a government’s design to include a new legislative actor, as California did, requires either applying existing means of evaluating power disputes, or creating new methods specifically for the new actor.[15] And if governance is a social contract where the people cede their sovereignty to representatives so long as the government promotes the public interest, the contract still requires a self-regulatory feature when the people are their own representatives.[16] Though they ultimately are sovereign when acting as the people, when exercising legislative power (as California’s electorate does) the voters are a legislative branch of government that must be restrained to prevent the voters from oppressing themselves.[17]

These design concerns inform the questions we consider here: how the electorate acts in California’s hybrid republic, and how well that system prevents the electorate from unbalancing it. Our analysis does not support the common themes that California uses the initiative more than any other state, or that the state’s electorate amends the state constitution excessively.[18] We find instead that, rather than acting as an outside disruptor, the electorate is adequately incorporated into the California system and that there are functional checks on the electorate. This rebuts the charges that the initiative needs structural reform or that the electorate needs additional checks. The electorate is not the great disruptor of California government—on the contrary, it generally functions well as part of a balanced system. Most importantly, the initiative has served its intended purpose: overcoming legislative inaction to solve several major public policy problems.[19] Its negative effect on the California constitution is often overstated, and the existing checks on the electorate are suited to the task.


A.  1849–1911: No Direct Democracy

Direct democracy was not included in the federal constitutional design. In revolutionary America, popular sovereignty was a core concept: the idea that ultimate power rested with the people themselves collectively.[20] But the federal framers considered and rejected direct democracy as the model for the federal government.[21] Consequently, there are no direct democracy tools in the federal constitution.[22] And although direct democracy was a significant factor in the colonial, Confederation, and early federal periods, the initiative was largely absent nationwide during the 1800s until the Populist and Progressive movements revived it around 1900 as a political reform measure to limit special interest influence on government.[23]

Similarly, direct democracy was not part of California’s original constitutional design. Delegates discussed the general concept of popular sovereignty in the first week of California’s 1848 constitutional convention: “The declaration of the sovereignty of the people, emanates from the foundation of our Republic. It has been adhered to ever since, and . . . would be adhered to in all time to come.”[24] Article 2, section 1 of the state constitution incorporates that principle: “All political power is inherent in the people. Government is instituted for their protection, security, and benefit, and they have the right to alter or reform it when the public good may require.”[25] Yet that sentiment remained conceptual until the Progressive reforms in 1911.

B.  1911: Direct Democracy Is Added to the State Constitution

California’s direct democracy mechanisms were created during the Progressive era as a comprehensive package of voter reforms that resulted from popular dissatisfaction with corruption and influence in the state legislature.[26] The Progressives argued that the cure for the ills of democracy was more democracy.[27] During that period, South Dakota was the first state to adopt the initiative and referendum in 1898, and between 1898 and 1918, twenty-two states adopted direct democracy constitutional provisions.[28]

Hiram Johnson was elected California’s governor in 1910 on a reform campaign platform aimed at influential special interests, particularly the Southern Pacific Railroad.[29] In his inaugural address, Johnson declared his intent to add direct democracy tools to the state constitution.[30] At the time, Article XVIII, section 1 provided that amendments could only be proposed by the legislature with popular ratification.[31] In 1911, the legislature proposed amending the state constitution to add four new electorate powers: initiative statute, initiative amendment, referendum, and recall.[32] The voters approved those reforms in a special election on October 10, 1911.[33] Given its substantial powers, some observers call the electorate the state’s “fourth branch” of government.[34] But as the ranking in Table 1 shows, California voters do not use the initiative the most: the state ranks second in total initiative use, behind market leader Oregon.[35]

California is one of 18 states that permit citizen-initiated amendments, and one of 16 states where those amendments go directly on the ballot.[36]

C.  California’s Direct Democracy Tools Described

California has specific terms for each power the state electorate can exercise on its own: recall, referendum, and initiative.[37]

Recall is the electorate’s power to remove an elected official in a special election before the official’s regular term expires: “Recall is the power of the electors to remove an elective officer.”[38] The California electorate has only once used its recall power against a high state officer: the voters recalled Governor Gray Davis in 2003.[39] That was the first (and so far only) successful gubernatorial recall in California; at the time it was only the second in U.S. history (North Dakota’s governor was recalled in 1921), and the third (unsuccessful) attempt occurred in 2012 in Wisconsin. Nineteen states and the District of Columbia permit recalls.[40]

Outside the United States, the general term “referendum” is commonly used to describe any non-candidate election matter the electorate votes on.[41] Not so in California. The referendum is the electorate’s power to veto statutes passed by the legislature: “The referendum is the power of the electors to approve or reject statutes or parts of statutes except urgency statutes, statutes calling elections, and statutes providing for tax levies or appropriations for usual current expenses of the State.”[42] The referendum is not much used.[43] Between 1912 and 2016, a total of 89 referenda were titled and summarized for circulation. Of those, 39 (43.82%) failed to qualify for the ballot, and 50 (56.18%) qualified for the ballot. Of the 50 that qualified, voters approved the law in 21 instances (42%) and rejected the law in the remaining 29 referenda (58%).[44]

The initiative is a means for the electorate to place a legislative act (a statute or a constitutional amendment) on the ballot by signature petition and to enact such proposals by majority vote: “The initiative is the power of the electors to propose statutes and amendments to the Constitution and to adopt or reject them.”[45] Only the statewide electorate holds this power; a local community, for example, cannot use the initiative to enact statewide legislation.[46] Only the electorate can amend the California constitution.[47] California constitution article XVIII provides two amendment procedures: The legislature may propose amendments for voter approval, or the electorate may amend the state constitution through the initiative process. Revisions may be made only by convening a constitutional convention or by legislative referral to the electorate.[48] (We parse the distinctions between an amendment and a revision in Section III.B.) Once passed, the legislature cannot alter initiative measures without the electorate’s consent, and there is no executive veto.[49]

D.  Constitutional Amendments Before 1912

Before 1912, the state constitution was amended 85 times total: the 1849 constitution was amended just three times, and the 1879 constitution was amended 82 times.[50] The sole amendment procedure during this period (under both constitutions) required the legislature to propose each amendment for voter approval. Comparing the pre-1912 amendment numbers for the 1849 and 1879 constitutions shows that between these roughly similar thirty-year periods (18491878 and 18791912) the amendment ratio is 1:27.3. We suspect that the disparity stems from the fact that the 1849 constitution was shorter and simpler; as discussed below, some research shows that a long constitution invites more frequent amendment. Environmental factors such as California’s smaller population, simpler economy, and overall lower government activity before 1878 may also contribute to the disparity.

E.  Less Initiative Activity 1912–1959, More Initiative Activity 1960–2017

The pre- and post-1960 periods have distinct levels of initiative activity: less activity before 1960 and more after. (We define “activity” here as the number of initiative amendments approved by the voters in a given period.) As the data in Table 2 (initiative amendments by period) show, although initiative activity increased after 1960, the number of legislative amendments and the total number of amendments decreased during that period. This is particularly interesting given that the later period is ten years longer than the earlier period.

The total number of amendments made in these time periods is similar: 225 amendments from 1912 to 1959 (47 years), and 208 amendments from 1960 to 2017 (57 years). That is only a 7.85% difference, or a 7.56% decrease for all amendments. And the total amendments averages-per-year are not grossly divergent: 4.79 before 1960 compared with 3.65 after 1960 (a 27% difference, or a 23.8% decrease). But from 1912 to 1959, just 17 amendments were initiatives—the remaining 208 were legislatively referred.[51] That ratio is 1:12.2. Note that the tally in each period is affected by the fact that before the 1960s, bond measures were constitutional amendments—19 of the amendments during this period were bond issuances. In 1962 the electorate adopted Proposition 6, amending the constitution to permit bond measures to go on the ballot as statutes instead of constitutional amendments. Because of this change any bond measures after 1962 drop out of the amendment tally. This added to the obvious contemporaneous rate change (see supra Figure 1) justifies the pre-and-post-1960 division. It also affects the initiative to legislative amendment ratio: removing the 19 bond amendments changes the ratio slightly to 1:11.1, which does not significantly alter the comparison between the two periods.

From 1960 to 2017, California voters approved 208 constitutional amendments: 36 initiative constitutional amendments and 172 legislative constitutional amendments in this 57-year period (0.63 per year), which is twice as many initiative amendments (36 versus 17); the average-yearly-adoption-rate doubled (0.63 per year versus 0.36 per year); and the initiative-to-legislative amendment ratio (1:4.8) is approximately twice the pre-1960 ratio (1:12.2 with bond amendments and 1:11.1 without them). Together, the change in averages and ratios indicates relatively greater initiative amendment activity after 1960. The electorate also enacted 15 constitutional revisions during the second period; given the distinction between amendments and revisions (discussed in section III.B), we do not include these in the amendments tally.[52] But the main conclusionthat the legislature is the primary constitutional change initiatorremains unchanged: in this period the legislature initiated approximately five times as many constitutional amendments as the electorate did.

It is difficult to determine what sparked the increase in initiative constitutional amendments since 1960. Several constitutional changes could be contributing factors. The 15 constitutional revisions between 1962 and 1974 made significant changes and deletions. Like most observers, we note that the adoption of Proposition 13 in 1978 (a change in the state’s residential real property taxation) caused a wave of similar tax reform nationwide,[53] which occurred during this period of higher initiative amendment activity. We note that the electorate enacted 19 initiative constitutional amendments before Proposition 13, and 33 since. And during this period, Proposition 1A in 1966 created a full-time legislature; that measure is credited with professionalizing the legislature and providing it significantly more authority and resources.[54] It is possible that a full-time legislature is more active than a part-time legislature and that greater legislative activity prompts more initiative use to check the legislature. Finally, Figure 7 shows that over time legislative and initiative constitutional amendments have converged. There is no doubt that California saw more initiative activity in the period following 1960 relative to the preceding period, but we decline to speculate about what combination of social, political, and economic factors caused the increase.

F.  Initiative Activity Quantitative Analysis

Our research shows that from 1849 to 2017, the California constitution was amended 518 times.[55] Since the initiative became available in 1912, the state constitution was amended 433 times total: Of those 53 were voter initiatives (12% of all amendments since 1912) and the remaining 380 (88%) were legislative proposals. And including constitutional revisions, the California constitution was changed 539 times.

California ranks second in its overall use of the initiative, and while the California constitution has been amended more than most other states, it is not the most-amended state constitution (Alabama’s is).[56] Part of the reason California’s constitution has a higher number of initiative amendments than some other states is the fact that compared to them, California has the least onerous procedural requirements for the electorate to enact constitutional amendments.[57] The usual conclusion is that the initiative process has a substantial effect on the rate of constitutional amendment in the state.[58]

Our data show that the opposite is true: the initiative process does not have a disproportionate effect on the amendment rate of the California constitution.[59] Initiative amendments (53) make up just 12% of the total ballot measures (433) amending the constitution from 1912 to 2017.[60] Nor is it surprising that California’s constitution is longer or has more amendments than the federal charter. State constitutions tend to be longer than the federal government’s because they design different political systems: the state is a general government, while the federal government has limited powers. And state constitutions (including California’s) generally have more accessible amendment procedures than the onerous process provided in the federal constitution; as a result, “every state constitution is amended more frequently than the U.S. Constitution.”[61] As of 2017, the aggregate state constitutional amendments number 7,586—over 150 amendments per state on average, which is over ten times the federal amendment rate.[62] These differences in kind, not degree, mean that the state and federal amendment rates are not comparable.

California’s constitution does have a relatively high number of amendments compared with other states.[63] Some scholars explain this with practical features: its age, length, and complexity.[64] Our analysis does not support that theory, but neither do we think the initiative is to blame.[65] California’s 1879 constitution is the twentieth oldest state constitution overall.[66] Of the 16 states with initiative amendments that go directly on the ballot, California has the seventholdest constitution, the highest number of amendments, and the highest amendment rate.[67] California’s constitution is the seventh longest U.S. state constitution.[68] But comparing California to other similar states shows at most weak evidence that the initiative is responsible for California’s relatively high amendment number.

As Table 3 (all states ranked) shows, the 20 oldest constitutions divide evenly by length, with ten in the top 50% and ten in the bottom 50%. The 20 most-amended constitutions also do not show a strong length correlation: of the 20 most amended constitutions 13 are in the top 50% and 7 are in the bottom 50%.[69] And most telling: only 9 of the 20 most-amended constitutions are initiative amendment states. California is the only one of the topfive mostamended states with initiative amendments, and neither of the two closest statesSouth Carolina (500 amendments) and Texas (491 amendments)has initiative amendments. This shows that availability of citizen initiative amendments is at most a contributing factor to a relatively high amendment number.[70] And it counters the oft-made claim that California is at the “radical end” of the direct democracy spectrum.[71]

 We compiled data on initiative measures generally and initiative constitutional amendments specifically from 1912 (the firstyear initiatives appeared on the ballot) to 2017, as follows.[72]

 These data permit several plausible conclusions. Most importantly, initiative amendments have nearly the same success rate as initiatives generally, which shows that an electorate attempt to amend the state constitution is about as likely to pass or fail as any other initiative measure.

From 1912 to 2017:

  • 19% of all titled initiatives qualified for the ballot (376/1952).
  • 6.8% of all titled initiatives passed (132/1952).
  • 35% of all qualified initiatives passed (132/376).
  • 34% of all qualified initiative amendments passed (53/154).

These results also show that as more initiatives are proposed, there is no corresponding increase of the rate of qualifying. Figure 1 shows that while the number of initiatives being titled as ballot measures increased significantly over time, there is at most a modest increase in the number of initiatives qualifying for the ballot. Nor do they rise and fall in tandem over time. Figure 2 shows that while the number of qualifying and approved initiatives did increase, neither did so to the same degree as the number of titled measures. Interestingly, Figure 3 shows that the change rates for qualified and approved initiatives correspond; the fact that they rise and fall roughly in tandem may indicate that there is a maximum possible success rate for initiatives on any given ballot, regardless how many initiatives are qualified.

This potential “change tolerance” figure contradicts the down-ballot falloff and voter fatigue arguments that are commonly leveled against the initiative. Instead, our data show that no matter how many proposals are circulated, only a consistently low number of them will qualify, and of those qualified only a consistently low percentage will pass.[73] Whether comparing titled to qualified measures (Figure 1), or qualified to approved (Figure 2), the takeaway is the same: most proposals fail, either at the qualification or the approval stage. The most dramatic discrepancy is between titled and approved measures (Figure 3), which shows an overall titled-to-approved ratio of 14:1—just a 7% chance of any given measure succeeding.

Increasing the number of initiatives and amendments on the ballot does not produce a concurrent increase in the success rate of those proposals. More proposals mean more will pass, but the likelihood of success remains low. While the total number of qualified initiatives and amendments has increased since 1960, the qualifying and approval rates have remained consistently low.[74] As Figure 5 shows, the disparity between initiatives qualified and amendments qualified remains fairly consistent over time, and a significant rise in the number of qualified initiatives corresponds with only a modest increase in qualified amendments.[75]

Initiative amendments occur less frequently than statutory measures. The simplest explanation is that, as Tables 4 and 5 show, fewer initiative amendments qualify for the ballot. Since 1912, of the 376 initiatives qualified, fewer than half (154) were amendments; and of the 132 initiatives approved fewer than half (53) were amendments. Comparing Figure 4 (initiative amendments approved) and Figure 8 (all initiatives approved) shows that the respective approval rates for all initiatives and initiative amendments are similarly variable and generally under a 50% passage rate; this is consistent with the overall passage average of around 35% for each. Figure 6 shows that amendments are qualified and approved in lower proportions than initiative measures generally, and while the quantity of initiative and amendment approvals are both increasing over time, the number of approved amendments per decade has remained in the single digits until this decade, while the total number of approved initiatives has increased significantly.

The electorate has been far less impactful with its constitutional change power than the legislature, which contradicts the conventional wisdom that California’s electorate overuses its amendment power.[76] As Table 6 shows, from 1912 to 2017, the electorate approved 433 amendments, but the overwhelming majority (380, or 87.8%) were legislatively proposed; only 53 amendments (12.2% of all amendments since 1912) have been initiative measures. This shows that the effect of initiative constitutional amendments is not as dramatic as the conventional wisdom holds. Instead, the state legislature has initiated the clear majority of constitutional change in California, even after 1912.[77] While overall initiative process use is rising, the increase is slow, and its impact on the constitution remains at a consistently low level.[78] And because court challenges to approved initiatives are common, some are partly invalidated or never take effect at all.[79]

But this may be changing: as Figure 7 shows, the trend lines for legislative and initiative constitutional amendments recently converged, as over the past forty years legislative action declined sharply and electorate action increased slightly.[80] And Figures 4 and 6 may indicate a possible recent upward trend in initiative amendment approval rates. Because we do not have complete data for this decade these possible indications should be viewed with caution.



A.  Complaints About Direct Democracy

With its hybrid government, California could benefit from the best aspects of both representative government and direct democracy or be paralyzed by the worst features of each. In the first scenario, the state can moderate direct democracy’s negative effects with its representative institutions, while its direct democracy institutions can mitigate a republic’s undesirable tendencies. Or California’s system may permit a small and unrepresentative segment of the electorate to make binding policy decisions for the state, reducing elected representative efficiency and devaluing minority interests.[81] We considered which scenario best describes the state and how successful the state is at balancing these competing dynamics. We conclude that California’s direct democracy tools are a net positive. California now has 105 years of experience with popular constitutional change. Its experience shows that direct democracy institutions can be as effective as traditional governmental institutions, particularly when direct democracy is combined with designed structural checks. For this state, the wisdom of crowds is real, albeit imperfect.[82]

Having direct democracy in a state constitution can be a net good, in theory, for several reasons: it is a check on the institutional branches of government; it encourages citizen participation in policy debates and governance; and it permits the governmentgoverned relationship to adapt to changed circumstances. All that assumes adequate institutional checks on the electorate’s power. In practice, California proves the theory: after a century of initiatives, California thrives.[83] The initiative does not supplant representative government, it supplements it.[84] Judicial review and the future electorate’s power to reverse past acts provide adequate safeguards.[85] And while the electorate sometimes creates problems for itself, the electorate more commonly uses the initiative to solve major institutional problems.

For example, in 1990, Proposition 140 imposed legislative terms limits, ending an era of lifetime legislative service.[86] In 2010, Proposition 20 created the California Citizens Redistricting Commission to stop partisan fights over drawing electoral districts, and Proposition 25 ended the required two-thirds majority budget vote requirement that caused chronically late budgets.[87] All were initiative amendments that tackled problems the legislature was unable or unwilling to address—exactly the initiative’s intended use. And as our quantitative analysis shows, the initiative is more commonly deployed cooperatively, with the legislature and the electorate working together to solve policy problems.[88] When it does act alone, the electorate is fairly conservative: the average success rate is under 40% for all metrics we evaluated, and the approval rate remains fairly constant almost independent of how many measures are proposed. That data and history do not support the conclusion that California’s electorate is a destructive political actor. But there are counterarguments, which we now consider.

Researchers have shown that direct democracy as a government institution can have both intended and unintended effects. Counterintuitively, the intended effects can be negatives, while the unintended effects can be positives.

The intended effects have negative consequences. Direct democracy was intended to (and does) increase participation and make government more responsive to electorate views on some issues, but it also makes government less efficient and less effective.[89] Unsurprisingly, the legislature suffers from the same inefficiency, which is a known and intended consequence of representative government.[90] A presently good solution for the proponent interest group may prove unworkable when applied to the population at large going forward.[91] Direct democracy has similar process inefficiencies to legislative action (enacting laws is costly), and it cannot adjust a proposal either before or after enactment without repeating the entire initiative process (again, costly).

Direct democracy’s indirect effects can be net positives. The single-subject nature of initiatives necessarily concentrates voter power on an individual issue, as with a single exercise against one representative in a recall.[92] Yet having the initiative available can improve elected official performance on issues that are not the subject of initiative action, because the electorate “saves” its limited resources for votes on the highest-interest issues, which in turn improves outcomes by focusing representative attention on those issues while also allowing them to devote more resources to other issues.[93]

The charge that initiative states are more poorly governed than non-initiative states[94] is a chicken-and-egg argument: do the legislature’s shortcomings encourage initiative use, or does using the initiative prevent the legislature from being effective?[95] And the answer depends on how one defines “effectiveness.” Direct democracy improves achievement of electorate preferences, and government responsiveness to voter preferences is itself a performance index. In other words, voters are more likely to get what they want, and the government they deserve, which may imply a difference between what scholars think is a measure of effective government and what that concept means to the electorate.[96] Viewing direct democracy from an economic perspective provides the same result: democracy is competitive government, and the alternative is monopoly government. From that perspective, electoral competition is the best guarantee that government will provide the voters with their preferred results, so democracy is the best method of satisfying voter preferences.[97]

Some scholars argue that the initiative’s potential negative effects outweigh its potential positives.[98] The initiative has been criticized for its disorganizing and bloating effects on the state constitution since the Progressive reforms were enacted in 1911.[99] Citizens may be too uninformed to make good decisions on public policy issues.[100] Initiatives force voters into a binary choice on an issue, and so fail to encourage debate and consensus.[101] Initiatives cannot weigh the intensity of interest group views. Initiatives forgo the legislative process of translating community preferences into policy through deliberation.[102] A legislature has lower transaction costs than the initiative process, and by reducing the transaction costs of bargaining, the legislative process increases the probability that political factions will cooperate and reach consensus.[103]

Even with democracy it is possible to have too much of a good thing.[104] Direct democracy was originally conceived as a necessary brake on the influence of wealthy corporate interests, but it is now criticized as having outgrown its initial purpose and as a vehicle for an excess of democracy.[105] It is further criticized as creating conflicting policy mandates that cripple the state government, ultimately encouraging more initiative activity to address government dysfunction in a process of diminishing returns.[106] California voters complain about the sheer number of ballot propositions and their confusing wording,[107] which can hinder educated consideration of ballot measures.[108] Voters favor improvements to the initiative process that increase opportunities for informed deliberation.[109] One scholar argues that the initiative:

  • Creates worse outcomes and weakens the democratic process,
  • Makes suboptimal outcomes more likely because the issues are too complex for the electors to understand, and
  • Prevents debate because issues are presented in final form or at most as competing alternatives.[110]

Yet these arguments against direct democracy institutions are equally valid against representative systems.[111] An argument against direct democracy is one against having any democracy at all.[112] Initiative measures are not limited to presenting a single set of alternatives; nothing prevents competing solutions or paired initiative measures from appearing on the same ballot.[113] The legislative filtering effect has a direct democracy equivalent: the large gap between titled, qualified, and approved measures shows that the electorate engages in a similar filtering process in that not all ideas make it to the ballot and not all are approved. While voters are more likely to feel somewhat ambivalent about the initiative process in general (consistent with voter dissatisfaction with government overall), voters are comfortable with their ability to properly evaluate individual ballot propositions.[114] And the ignorant-electorate hypothesis proves too much: an electorate unable to make good decisions on initiative measures is equally incapable of choosing good representatives—a hypothesis that undermines the very foundation of a representative republic.[115] California’s experience shows that voters are capable of understanding electoral issues and becoming sufficiently informed to make decisions.[116]

The practical reality of California’s direct democracy institutions is they are neither as bad as their critics believe nor as good as their supporters believe. The presence of initiatives on a ballot has only a small turnout-increasing effect in presidential elections; the same is true when initiatives are present on midterm ballots.[117] But in general, initiative propositions do increase voter turnout, which translates to a more informed and involved electorate.[118] With some variation, the available statistical evidence shows that the part of the electorate that actually votes on initiative ballot propositions is relatively wellinformed, conscientious, and cautious.[119] And there is evidence that, rather than discouraging participation in representative government, or causing interest groups to substitute action in one arena for another, the initiative increases opportunities for political involvement and action overall.[120] Overall, in direct democracy systems there is little to show that initiative outcomes are inferior to legislative outcomes.[121]

With that overview, we now consider several related direct democracy effects: population size, money, turnout, and majority approval. Of those, we conclude that the one problem that calls for a solution is majority approval.

B.  Population Size

In theory, direct democracy’s effectiveness is inversely related to the community’s size: the smaller the community, the more effective direct democracy is at achieving the goals of government.[122] Direct democracy, in its earliest conception, could only operate in small communities—a larger community where the people could not conveniently meet to personally discuss public matters required another system.[123] Representative government is the usual solution to the more complex needs of a larger community.[124] Indeed, the experience of the ancient Greeks suggests that the maximum population for a successful direct democracy is 5,000 to 10,000 citizens.[125] In the American revolutionary period there was significant experimentation with direct democracy, both before and after the 1789 constitution.[126] The founding generation’s experience resulted in a profound suspicion of undiluted direct democracy.[127] This may explain the fact that modern pure direct democracy primarily exists only in town-size communities with populations comparable to the ancient Greek city-states.[128] This evidence, historical and modern, of experiments with direct democracy suggests that significant use of direct democracy is effective only in small communities and for limited issues.[129]

California currently has a population of 39.5 million, including 24.8 million eligible voters, 19.4 million registered voters, and 14.6 million who voted in the 2016 presidential election.[130] So California should be too large to benefit from direct democracy. It should be both impractical and undesirable to use direct democracy in a community California’s size. Impractical, because even with modern electronic communication means it is impossible to fully engage such a large electorate. Undesirable, because the relatively low percentage of voters needed to qualify and pass measures risks majority tyranny.[131] The chronically low voter turnout and cost of initiatives could be symptoms of the over-large population using direct democracy in California. But as we discuss in Section III.D, voter turnout is low and declining nationwide, regardless of state size or initiative availability.[132] The evidence we review there does not show a correlation between low turnout and the initiative. And as we discussed in Section III.A, the evidence for the initiative compelling suboptimal governing outcomes is weak. Rather than indicating a basic incompatibility between direct democracy and larger populations, the core turnout issue is the “slim majority” problem we discuss in Section III.E.

C.  Money’s Impact

Currently, an initiative statute requires 365,880 signatures, and a constitutional amendment requires 585,407;[133] at a rate of two to three dollars (or more) per signature, any interest group lacking funds in the million-dollar range will be excluded from the process.[134] And the cost of qualifying an initiative measure for the ballot has increased dramatically over time, from a median of approximately $45,000 in 1976 to nearly $3 million by 2006.[135] Consequently, the very issues that are up for debate during any given election are largely dependent on choices made by interest groups with sufficient funds to qualify initiative measures for the ballot.[136] We think the explanation here is correlation, not causation. Money’s effect on campaigns has proved to be less than suspected: well-funded corporate interest campaigns succeed at a lower rate than initiatives generally, and the best success rate of particularly well-funded campaigns is in securing a “no” vote, which is also the most common voter reaction to initiative measures.[137]

D.  Effect on Turnout

In theory, direct democracy should foster voter engagement. According to the Condorcet Jury Theorem, where right answers exist and the voting group has average competence, the majority will arrive at the right answer as the size of the voting population increases.[138] In practice, this means majority voting rules work best when there is high turnout. But voting nationwide has been declining for decades, across all ballots.[139] So does direct democracy increase turnout in practice? As with the other empirical studies we reviewed, the results on this point are mixed, with a small net positive effect: ballot initiatives are more likely than not to increase voter turnout. Ballot propositions do not increase turnout in presidential election years, when voters are most engaged with the presidential campaign, but they do increase turnout during midterm elections.[140] And initiatives increase turnout in off-year elections.[141] Citizen-initiative races attract the most attention and have the greatest effect on turnout, while uncompetitive legislative initiatives and referenda have little effect.[142]

Turnout effects can be self-sustaining: because they are known effects, proponents may factor them into their timing strategy to best target their voters, and so compound the initiative’s turnout effects. For ballot measure proponents who seek to appeal to an intense minority of voters, waiting for a low turnout gubernatorial election may be the best path to approval. Because ballot measures pass with a simple majority of votes cast on that measure, propositions become law in California regardless of turnout levels.[143] Low turnout reduces the signature requirements to qualify for the ballot. Qualifying with a lower threshold, the proponents could then target a low-interest election.[144]

To curb this practice and its effects, the California legislature took action in 2011 with Senate Bill 202 (“S.B. 202”), requiring any measure approved after July 1, 2011 to go on general election ballots only.[145] The bill’s sponsors were concerned that “special interests” would “game the system” in low turnout elections, justifying the move to consolidate to general elections.[146] While S.B. 202 largely solved the turnout problem, it created another: by consolidating ballot measures to general elections only, general election voters are now overwhelmed with lengthy ballots.[147] Voters faced with a long ballot tend to opt out of educating themselves on all the issues, harming both participation and deliberation levels, and benefiting the status quo by making abstentions and “no” votes more likely.[148]

Overall, there is no reason to believe that California’s low voter turnout is an adaptation to the higher process burden of achieving consensus in a larger polity. If that were true, the 24 states with initiative powers would have correspondingly lower voter turnout rates than the other 26 states. Instead, as Table 7 (turnout) shows, voter turnout rates are consistently low nationwide.[149] And there is a positive correlation between turnout and citizen initiatives during non-presidential election years. We conclude that California’s low voter turnout reflects broader turnout trends and is not a reaction to direct democracy. Direct democracy does not deter turnout, but lengthy ballots do result in greater voter abstention on down-ballot propositions and races.

E.  California’s Lack of a True Majority Requirement Harms Legitimacy

Having discounted population, money, and turnout, we turn to the one problem we see in the state’s direct democracy system that needs addressing: California ballot measures rarely receive approval from a true majority of the electorate. The available current voter data supports the conclusion that approximately 18% of the state population is the controlling “majority” deciding any given initiative measure, which is an unrepresentative sample of the community that does not reflect the population’s diversity on a variety of factors.[150] For example, one proposition became law with approval from less than 15% of registered voters.[151] We call this the slim majority problem.[152]

This problem has two contributing factors: registration and turnout. A significant proportion of eligible voters (about 25%) is not registered.[153] This is lower than in other states.[154] And some registered voters do not vote; even fewer vote consistently.[155] Calculating turnout based on eligible voters better captures the true gap between potential voters and actual voters. Since 1990, on average just under 40% of eligible voters participated in gubernatorial elections, 33% participated in presidential primaries, 24% participated in statewide primaries, and 31% of eligible voters participated in special elections.[156] Only in general presidential elections do a majority of eligible voters regularly vote.[157] This decline in voter turnout mirrors a decreased participation trend, and California is below the national average.[158]

A slim majority of registered voters regularly participate in California elections, and they are not a representative sample. On average, 50% of registered voters voted between 1990 and 2016.[159] Among eligible voters, turnout during the same period averaged less than 37%. These voters are not representative of California’s electorate: despite California’s demographic diversity, the laws end up reflecting the preferences of the regular voter, who tends to be older, whiter, and more conservative.[160] The participating electorate is the same for initiatives as voters generally: they trend towards the upper end of the income and economic scale regardless of political affiliation.[161]

The slim majority problem applies in nearly all California initiative contests. Only four propositions since 1990 received approval from a registered majority: Proposition 1A in 2004, protecting local government revenue from statewide use; Proposition 59 in 2004, providing the right of public access to government meetings and records; Proposition 35 in 2012, increasing penalties on human traffickers;[162] and Proposition 58 in 2016, restoring bilingual education in California public schools. Each passed with a resounding margin in a presidential election year.[163] No proposition since 1990 earned an eligible majority.[164]

Turnout figures for any given election do not reflect participation levels for ballot propositions. Consistently, segments of the participating electorate abstain from voting on ballot propositions. In each election from 1990 to 2016, an average of 8.1% of participating voters declined to mark a choice on each ballot proposition. Table 9 (average abstention rates) below reflects overall abstention rates in recent elections. Currently, these abstentions do not factor into a proposition’s approval because the California constitution only requires “a majority of votes thereon” for a proposition to become law.[165]

Factoring in these abstentions, even fewer ballot propositions receive approval of a majority of voters in that election. On average, in any given election, nearly a third of ballot propositions fail to win approval of a majority of that election’s voters.[166] In other words, the number of approving votes for the proposition does not exceed 50% of the number of voters participating in that election. This deficit in voter approval occurs consistently across all election types, regardless of the length of the ballot. (See below Table 9 majority approval.)

Because we view the slim majority problem as direct democracy’s chief defect in California, in the next Section we propose a solution.

F.  A Proposed Solution: The Dual Approval Quorum

Is there a workable solution to the slim majority problem? If not, can the system adequately self-maintain without a fix? Solving this problem is important because direct democracy’s legitimacy depends on a minimum level of popular interest (the electorate must opt in and collectively decide), which in turn requires a minimum level of voter participation. Yet the state’s initiative system permits an unrepresentativelysmall electoral majority to approve laws and constitutional amendments. If direct democracy relies on collective consent, something close to a true majority should be required for an initiative to pass. To correct this flaw, we suggest a quorum requirement.

While the legislature and electorate wield equivalent legislative power, their respective quorum requirements are entirely distinct. Legislative and electorate acts, whether statutes or constitutional amendments, are substantively equivalent. But the processes vary substantially: the legislature requires quorum and several stages of deliberation and committee review. The initiative requires none of those; the electorate votes, and that is all.

Quorum provides deliberative bodies the authority to act. If a body functions through collective deciding, a threshold of members must be present to take action.[167] This principle runs throughout California common law,[168] procedural rules,[169] and governing statutes.[170] At all levels of government, deliberative bodies in California face quorum requirements.[171] This is true for the state legislature and for local city councils. While the electorate legislates and functions as a deliberative body, it currently lacks a quorum requirement. This means that unlike the state’s other legislative body, the electorate can pass laws without a threshold of member approval. Adding a quorum requirement will address the representation problems with California’s current direct democracy framework and solve the slim majority problem.

There are two kinds of quorum requirements that could be added to an electoral process: a participation quorum and an approval quorum. A participation quorum requires that a minimum portion of the voting population considers the ballot measure at the polling station. That would mean a measure is not enacted unless a certain percentage of registered voters turns out to vote.[172] But this requirement tends to induce those who oppose the ballot measure to abstain from the vote entirely.[173] Because of that opportunity to game the system, we do not suggest adopting a participation quorum in California.

An approval quorum ensures that a sufficient portion of registered voters—or voters in that election, depending on the framework applied—votes in favor of the ballot measure. This sets a baseline threshold to reflect the popular will. In this system, abstentions count: voter abstention on an individual ballot measure factors into the approval calculation. A quorum requirement may also serve as a safeguard against “false majorities,” a small but intense minority supporting a particular policy goal.[174] While more common abroad, other U.S. states have adopted quorum rules for citizen initiatives. The thresholds vary: measures can only pass when voter turnout reaches 30% in Massachusetts, 35% in Nebraska, and 40% in Mississippi.[175] Oregon has a 50% participation quorum requirement for local-level property tax ballot measures.[176] Wyoming has adopted a “this election” approval quorum, where the measure will only pass when it receives approval from a majority of voters who turned out in that particular election.[177]

Adopting a dual-approval quorum framework would improve California’s direct democracy system by solving the slim majority problem and requiring a true electoral majority to enact initiative measures. The dual approval quorum solution would look like this:

An initiative could amend California constitution Article II, section 10 to change “by a majority of votes thereon” to “by a majority of votes out of all voters in that election.” The same measure could repeal SB 202’s changes to the state’s Elections Code section 9016 and establish a 25% approval quorum requirement for registered voters in all elections. It could also amend Article II, section 10 to add: “No initiative statute or referendum may take effect without approval votes from a minimum of 25% of the registered voter population.”

The measure would include a legislative ratification process for any initiative that passed the first approval threshold but not the second. This would amend Article II, section 10 to add:

Any initiative receiving approval from a majority of voters in that election, but failing to meet the registered voter threshold, is automatically referred to the legislature for consideration and possible ratification. Each house of the Legislature must hold a vote on any such initiative within 90 days of the Secretary of State’s certification of the result of the official canvass of the returns of the election. If the measure receives majority approval from each house, the measure must be presented to the Governor. It will take effect immediately if it is signed by the Governor.

These proposed reforms align with voters’ express desire to enact reforms that would reengage citizens in the initiative process.[178] The first proposal would address two key concerns with the initiative process: inadequate deliberation and lengthy ballots. Currently, when voters choose to abstain from voting on certain ballot measures, either due to a lack of knowledge, opinions on the proposition, or simple voter fatigue, those abstentions do not affect the outcome of the vote. Proponents know this and have no incentive to limit themselves to serious issues that would galvanize the public. They need only convince a determined minority of active voters. But as seen in Table 9, many voters abstain.[179] Voters armed with the knowledge that their abstentions count could then focus their own voter education on the issues that matter to them. This will provide an opportunity for greater deliberation and results that better reflect the electorate’s will.[180]

The 25% approval quorum will address direct democracy’s core legitimacy in a low turnout environment. The threshold is the equivalent of requiring at least half of registered voters to turn out to vote, with at least half of those voters approving the ballot measure. Because only general elections tend to see turnout over 50%, it may seem that the effect of a 25% approval quorum will not differ significantly from the effects of SB 202.[181] Not so—there are flaws in SB 202 that the 25% approval threshold will correct. For example, if a groundswell of voters chose to support a reform in a primary or special election, they should not have to wait another year to pass the reform. Currently, SB 202 would block such a move by limiting propositions to general elections. The 25% approval quorum will both maintain the positive elements of SB 202—preventing proposition gamesmanship in low turnout elections—and restore balance where it is too draconian by reviving the option to propose initiatives in primary and special elections. Most importantly, voters will know that no measure can pass through the ballot box that did not reflect the will of a true majority of registered voters.

The legislative ratification proposal will provide an avenue for voter-approved initiatives to become law even when low turnout bars fulfillment of the 25% quorum. Given the trend of low voter turnout overall, even a measure that earned support from 62% of voters could fail to pass the dual quorum framework if only 40% of voters turned out. Forty percent is the average turnout for recent gubernatorial elections, so this could apply to a significant number of initiatives on the ballot. But rather than permitting a measure to pass without broad voter support, the legislative ratification mechanism would ensure that the legislature considers the proposal before enacting it. And by requiring the legislature to hold a vote, the ratification mechanism ensures that the electorate is not thwarted by legislative inaction on a measure that received majority support. This would recognize voters’ expressed interest while protecting consent of the governed from minority rule.

These reforms will not diminish the initiative power. On the contrary, they will enhance voter legislative power by increasing its perceived and actual legitimacy. Intense minorities gaming the system in low turnout elections threaten that legitimacy; these reforms will prevent an unrepresentative interest group from hijacking the process. Still, we should not overstate the effects. Solving the slim majority problem does not address broader trends in voter disengagement. A quorum requirement will only mitigate the consequences of low turnout.[182] Various reforms have solved some of the legislature’s problems.[183] We see no reason why the state’s other legislators could not also tolerate some improvements.

We recognize that changes to the initiative process are extremely unlikely to pass. And there is a counterargument to this proposal:

In many states, the requirement that a proposed amendment receive a majority of all persons voting in the election, rather than just on the amendment, frustrated constitutional change. This requirement frustrated change because “political experience shows that there is a consistently smaller proportion of the total vote in a general election cast for constitutional proposals than for live candidates for office.”[184]

While we are confident that the initiative can be improved, as discussed above, the status quo still provides net benefits to the people of California. This is partly due to the effective systemic checks on the electorate that we discuss in the next Part.


Several serviceable checks on the electorate currently exist. Adequate checks on the electorate are necessary because any branch of government can become a tyrant if it accumulates enough power.[185] Wielding legislative powers, the electorate is no different. Maintaining both a balanced government and an equal society when direct democracy is added to a representative republic requires institutional means for maintaining the relationship between the electorate’s and representatives’ powers. California has two system-maintaining features, which on the whole are adequate to the task of managing direct democracy: the future electorate and judicial review. Those features could be improved if the initiative process itself could accept some changes (like our quorum proposal above).

Because even a meritorious proposal to change the electorate’s direct democracy tools is so unlikely to succeed, we analyze the adequacy of the existing checks on the electorate. There are several checks on the electorate’s legislative power, including some constitutional limitations:

The only express constitutional limitations on the electorate’s exercise of the statewide initiative power are those in sections 8 and 12 of article II. Section 8, subdivision (d) of article II bars initiative measures “embracing more than one subject,” and section 12 of that article bars constitutional amendments and statutes which name[] any individual to hold any office, or name[] any private corporation to perform any function or to have any power or duty . . . .”[186]

The electorate can check itself: a future electorate can always correct or undo the errors of a past electorate. And the judiciary is an effective brake on the excesses of popular sovereignty, as it is with the other state political actors. We conclude that these checks have proven to be adequate, and we expect them to continue to be so absent some changed condition.

Think of the electorate on a continuum: past, present, and future. The past electorate enacted various procedural and substantive provisions when it adopted the state constitution. That past electorate’s acts cannot prevent the present electorate from changing the substance of the constitution; nor can the present electorate stop the future electorate from doing the same.[187] Procedural limits in a constitution are similarly at the present and future electorate’s mercy.[188] Thus, whatever wrong the past and present electorate does, the future electorate can always right. Obviously, the reverse is also true: the good acts of the past and present electorates can also be undone. The point is that the electorate owns its mistakes and has the power to correct them if it wishes. The present electorate legislates knowing that the future electorate is always just around the corner, with complete power to alter the present’s enactments at will.

The judiciary’s ability to check the electorate is based on the power of judicial review. In California, the power to legislate is shared between the legislature and the electorate through the initiative process.[189] “As direct democracy has become an increasingly prevalent force in state policy making, it has shifted power away from elected representatives and toward the ‘parallel legislature’ of governing by initiative.”[190] Because the California constitution divides the state’s legislative power between the electorate and the representatives, and because the electorate acts autonomously in discretionary exercises of its powers, we argue that (for separation-of-powers purposes) the electorate should be considered an independent branch of the state government with legislative power.[191]

Armed with this general power of judicial review over initiative measures and the power to resolve separation-of-powers disputes, California courts have the mandate and ability to police the electorate when necessary. This is just and proper. Judicial review of ballot propositions fosters direct democracy—preserving direct democracy by curbing its abuses and increasing participation incentives through the appearance of legitimacy created by enforcing process fairness.[192] And judicial review is the answer to a common criticism of direct democracy—that the majority of citizens will vote to undermine the rights of the minority.[193] Ordinarily, concerns about overconcentration of power would counsel a more restrained judicial role, but in California the ultimate check on judicial authority lies with the electorate, which has used its power to remove state high court justices.[194]

Next, we review the substantive constitutional issues the initiative potentially can create for the courts to resolve, evaluate the judicial tools appropriate to each problem, and show that judicial review is generally adequate to address them.[195] We first discuss the textual limits on the initiative power (singlesubject, revision–amendment, and separation of powers), and then turn to secondary effects caused by a textually valid initiative on individual rights.

A.  SingleSubject Rule

The single-subject rule provides that an initiative measure embracing more than one subject may not be submitted to the electors or have any effect.”[196] The rule’s main purpose is “to avoid confusion . . . and to prevent the subversion of the electorate’s will.”[197] This provision was added to the California constitution in 1948, in “response to a lengthy, multifaceted initiative provision that recently had been the source of considerable controversy.”[198] The rule “is a constitutional safeguard adopted to protect against multifaceted measures of undue scope” that “serves an important role in preserving the integrity and efficacy of the initiative process.”[199]

Notwithstanding the strict language of the provision, the California Supreme Court has adopted an “accommodating and lenient” legal standard “so as not to unduly restrict . . . the people’s right to package provisions in a single bill or initiative.”[200] The court has explained:

The single-subject provision does not require that each of the provisions of a measure effectively interlock in a functional relationship. It is enough that the various provisions are reasonably related to a common theme or purpose. . . . The governing principle is that an initiative measure does not violate the single-subject requirement if, despite its varied collateral effects, all of its parts are reasonably germane to each other, and to the general purpose or object of the initiative.[201]

The “reasonably germane” standard reflects the California Supreme Court’s “liberal interpretative tradition . . . of sustaining statutes and initiatives which fairly disclose a reasonable and commonsense relationship among their various components in furtherance of a common purpose.”[202] Accordingly, the state high court has

upheld a variety of initiative measures in the face of a single-subject challenge, emphasizing that the initiative process occupies an important and favored status in the California constitutional scheme and that the single-subject requirement should not be interpreted in an unduly narrow or restrictive fashion that would preclude the use of the initiative process to accomplish comprehensive, broad-based reform in a particular area of public concern.[203]              

On the other hand, “[t]he common purpose to which the initiative’s various provisions relate, however, cannot be ‘so broad that a virtually unlimited array of provisions could be considered germane thereto and joined in this proposition, essentially obliterating the constitutional requirement.’”[204]

This leaves California courts with a broadly deferential standardone that rarely requires striking down an initiative measure, to the extent that some question the singlesubject rule’s effectiveness.[205] As one pair of commentators put it, the “single subject rule in California has devolved into a virtual nullity; it is a rule with few, if any, teeth.”[206] So while it is an important structural protection, it rarely provides a sturdy basis for judicial intervention.[207]

The criticism of the single-subject rule as a paper tiger, however, is somewhat overblown. We found 69 cases where the California Supreme Court considered a singlesubject rule challenge (see Appendix Table 3, SingleSubject Rule Cases) including both legislative acts and popular initiatives. Of those 69 cases, the court used the rule to invalidate an act 8 times (11.6%). Of the 69 results, 57 dealt with legislative acts (82.6%); the remaining 12 concerned the initiative (17.4%). In the twelve cases where the California Supreme Court expressly considered a single-subject challenge to an initiative, it relied on the rule to invalidate all or part of an initiative twice (16.7%). The rule applied to invalidate a legislative act 6 times (10.5%). Several factors explain the higher numbers for legislative versus initiative acts. Obviously, the legislature enacts more legislation than the electorate does. The singlesubject rule has applied to initiatives for only seventy of the initiative’s 105 years, while the rule has limited the legislature for all of its 167 years.[208] And there is a variant of the singlesubject rule that applies only to legislative appropriations.[209] Some of the 69 cases concern appropriations; no equivalent rule applies to the initiative. And still the rule applied to a higher percentage of initiative than legislative acts.

We recognize that the single-subject rule does not often apply. Still, the threat of a pre- or post-election single-subject challenge is an active deterrent to proponents who may otherwise push the envelope. As a practical matter, an initiative measure that has the financial and political backing to make it to the ballot is unlikely to run afoul of the single-subject rule. Well-heeled proponents are generally unwilling to risk placing an initiative on the ballot that could be vulnerable to a constitutional challenge. Proposed initiative measures are commonly prepared with the assistance of attorneys (if not drafted by them outright), and then vetted through a public review process where proponents have the opportunity to amend the proposed initiative.[210] Only then is the final proposed initiative submitted to the Attorney General to prepare the circulating title and summary.[211] This process provides proponents time to identify and address potential defects in the form of the proposed initiative measure long before it reaches the voters. That explains the singlesubject rule’s infrequent application better than the rule’s claimed weakness.

B.  Revision and Amendment

The principal limitation on the initiative is the constitutional constraint against using the initiative power to enact sweeping or fundamental changes to the state’s governmental framework through constitutional revisions. Specifically, the initiative power can be used to amend, but not revise, the California constitution.[212] An amendment is any law that effects a more modest addition or change to the state’s constitution. Revisions are laws that “fundamental[ly] change . . . the basic governmental plan or framework” set forth in the state constit ution.[213] This distinction means that far-reaching changes in the state constitution can only be accomplished through a deliberative process with the state’s legislature and electorate acting together.

Although “amendment” and “revision” are not defined in the state constitution, the text makes clear that distinct procedures apply to each act. As the California Supreme Court has put it, the concept of a revision as a higher-level exercise of constitutional power “is based on the principle that ‘comprehensive changes’ to the Constitution require more formality, discussion and deliberation than is available through the initiative process.”[214] The California Supreme Court has developed the following standard to distinguish between them:

A “revision” denotes a change that is qualitatively or quantitatively extensive, affecting the “underlying principles upon which [the Constitution] rests” or the “substantial entirety of the instrument.” By contrast, an “amendment” denotes a change that is qualitatively and quantitatively limited, making a modification “within the lines of the original instrument as will effect an improvement, or better carry out the purpose for which it was framed.”[215]

The state constitution imposes a much higher procedural barrier to enacting revisions than it does for amendments. Specifically, voters can propose and adopt constitutional amendments directly through the initiative process, while revisions may only be accomplished by the state legislature and electorate acting together.[216] As discussed above, the legislature can propose specific revisions directly for ratification by popular vote, or propose a convention to revise the constitution.[217] Prohibiting direct adoption of revisions therefore provides a critical structural safeguard against electoral overreach: it ensures that broad changes to the state constitution can only be made when the legislature and the electorate act in concert. Yet the bar is not so high that it prohibits effectively using the revision power: the legislature and electorate have together revised the state constitution 21 times (see Table 6).

While the revision–amendment distinction provides a critical structural check on the electorate’s ability to change the state constitution, when called upon to enforce this constitutional limitation, the California Supreme Court has produced mixed and arguably inconsistent results.[218] As the authors have explained:

In Strauss v. Horton, for example, the California Supreme Court held that abolishing the state right of marriage by initiative constitutional amendment was not a qualitative revision of the state constitution—reasoning that the measure did not have a substantial or, indeed, even a minimal effect on the governmental plan or framework of California. In Raven v. Deukmejian, on the other hand, the court found a qualitative violation where an initiative constitutional amendment abolished state substantive rights for criminal defendants because it altered the authority of state courts to independently interpret criminal law. Divergent results like those in Strauss and Raven invite charges of hypocrisy. Worse, comparing the results in Legislature v. Eu (upholding legislative term limits and a forty percent reduction of the legislature’s budget) with Raven (rejecting an initiative that only affected judicial discretion) invites the conclusion that the courts will protect their interests but not those of other state government branches.[219]

Strauss, in particular, highlights another critical shortcoming of the revisionamendment test. Because its primary focus is preserving the structure of California’s government, it is ill-suited to resolving disputes over initiative measures that do not significantly change to the state’s “governmental plan or framework”—even where an initiative takes aim at fundamental constitutional rights.[220]

In the following Sections we discuss doctrinal solutions to these two shortcomings.

C.  Separation of Powers

While the revisionamendment distinction provides a critical structural check on the electorate’s ability to change the state constitution, it provides an incomplete solution when courts confront an initiative that does not amount to a revision, but nevertheless infringes on the core powers of the state government’s branches.[221] These critical disputes have arisen on multiple occasions in the past and will surely arise again.[222] We have argued that this doctrine can be improved by the judiciary treating the electorate in this scenario as a co-equal branch of state government and relying on existing separation-of-powers principles to police inter-branch disputes.[223]

Applying the separation of powers doctrine to the electorate when it acts in its legislative capacity addresses the largest gap in the revisionamendment analysis. California’s separation of powers doctrine “recognizes that the three branches of government are interdependent, and it permits actions of one branch that may ‘significantly affect those of another branch.’”[224] “[A]lthough the state constitution ostensibly requires a system of three largely separate powers, the state separation of powers doctrine does not create an absolute or rigid division of functions; instead, the California view assumes that there will be some mutual oversight and influence between the branches.”[225]

Policing separation-of-powers disputes is the judiciary’s province. Courts “have not hesitated to strike down provisions of law that either accrete to a single Branch powers more appropriately diffused among separate Branches or that undermine the authority and independence of one or another coordinate Branch.”[226] While a branch “may not use its powers to ‘defeat or materially impair’ the exercise of its fellow branches’ constitutional functions, nor ‘intrude upon a core zone’ of another branch’s authority,”[227] the doctrine does not “prohibit one branch from taking action properly within its sphere that has the incidental effect of duplicating a function or procedure delegated to another branch.”[228]

Vesting shared legislative power in the state electorate, as California’s constitution does, changes the tripartite power dynamic typical of modern republican government. Article III, section 3 of the state constitution provides that “[t]he powers of state government are legislative, executive, and judicial. Persons charged with the exercise of one power may not exercise either of the others except as permitted by this Constitution.” The “primary purpose of the separation-of-powers doctrine is to prevent the combination in the hands of a single person or group of the basic or fundamental powers of government.”[229] As we have explained, the direct democracy provisions in the California constitution require including the electorate among the “persons charged with the exercise” of the state’s legislative power, which means the existing separation of powers analysis must adapt to include the electorate.[230] California’s direct democracy tools reduce the executive and legislative powers relative to the electorate, and increase the governor’s power relative to the legislature.[231] This increased diffusion of power ultimately benefits individual liberty.[232]

But that additional dispersion of power requires its own separation of powers analysis. The legislature is the creative element of government.[233] Like the legislature, with which it shares the state’s legislative power, the electorate can create separation of powers problems.[234] Judicial review is adequate to manage that problem, especially since the California Supreme Court has made it clear that the core powers analysis applies to the electorate.[235] Armed with judicial review and the revisionamendment rule (particularly with this modification), the courts are well-versed in handling separation-of-powers disputes involving the electorate.

D.  Individual Rights

What happens when the electorate passes an initiative that potentially infringes on individual rights secured by the state or federal constitution? One of the fundamental purposes of a constitution is to provide protection for individual rights.[236] Rights in a constitution are countermajoritarian.[237] Direct democracy potentially has a significant effect on individual rights, and it presents a risk for minority groups.[238] This is because the countermajoritarian individual rights necessarily conflict with the majoritarian power of the initiative: any temporary majority can effect a permanent change to individual rights that disadvantages the minority. Similarly, the principle of equal protection requires protecting minority rights against the majority,[239] while the initiative tends to preserve majority preferences.[240] And because the state’s median voter controls the final outcome of any initiative, any constitutional change will necessarily have a majoritarian bias.[241] This characteristic of the initiative favors stability over expansion of individual rights, causing a slower rate of adopting constitutional rights for minority groups.[242]

The federal constitution was designed as a representative republic, on the principle that the checks and balances inherent in the government’s design would prevent tyranny by any of the federal government’s branches, and the lack of direct democracy would prevent tyranny by the people.[243] But Congress and the President have overcome those restrictions.[244] Even the judiciary, the least dangerous branch, has been guilty of such sins.[245] State legislatures have been no less despotic at times.[246] Similarly, electoral majorities have both the ability and tendency to use the initiative process to deprive unpopular minorities of rights or to prevent such groups from gaining rights.[247] This has occurred many times in California history.[248] The takeaway here is that the electorate is no different from any other branch of government regarding the risk of tyrannical behavior.

California’s experience with same-sex marriage illustrates this point. The state constitution provides for the equal protection of individual rights.[249] In 2008, the California Supreme Court held that limiting the definition of “marriage” to opposite-sex couples violated the constitutional guarantee of equal protection.[250] But the voters then passed an initiative constitutional amendment restricting the right of marriage to only opposite-sex couples.[251] This was a difficult issue for the courts to resolve. The California high court decided that equal protection did not apply; the U.S. Supreme Court held that it did.[252] This problem is not specific to the debate over same-sex marriage, and we use that issue here only as an example of the risk the initiative can present to individual rights.

The same-sex marriage issue illustrates a significant structural limitation of the initiative. Although there are procedural hurdles to passing an initiative measure, there are few constraints on the subject matter that can be placed on the ballot. Say, for example, a group proposed an initiative measure stating that only women could vote and revoked male suffrage. Even though such a measure would be patently unconstitutional, there are no direct constitutional constraints to prevent voters from considering and approving the initiative: the Attorney General has a constitutional duty to prepare a circulating title and summary for the measure, and the proponents are then free to gather signatures to qualify it for the ballot and then campaign for its passage.[253]

In such a case, judicial intervention is the only means to prevent an unconstitutional initiative measure from reaching the ballot. As a general matter, “it is usually more appropriate to review constitutional and other challenges to ballot propositions or initiative measures after an election rather than to disrupt the electoral process by preventing the exercise of the people’s franchise, in the absence of some clear showing of invalidity.”[254] But the California Supreme Court has recognized that “the principles of popular sovereignty which led to the establishment of the initiative and referendum in California . . . do not disclose any value in putting before the people a measure which they have no power to enact.”[255] Accordingly, pre-election review of a proposed initiative is appropriate to challenge “the power of the electorate to adopt the proposal in the first instance.”[256] This can be accomplished in two primary ways. The Attorney General can seek judicial relief from its duty to prepare a circulating title and summary,[257] or citizens can bring a petition for writ of mandate to prevent the Secretary of State from acting on a proposed initiative measure.[258] In either event, judicial intervention is available to prevent a patently unconstitutional measure from reaching the ballot.[259]

Two recent examples show how this process works in practice. In 2015, a proponent submitted a proposed initiative titled the “Sodomite Suppression Act,” which sought to amend California’s criminal code to penalize what the proponent described as “sodomy” or “buggery” by requiring “that any person who willingly touches another person of the same gender for purposes of sexual gratification be put to death,” and by barring from public employment any person “who is a sodomite or who espouses sodomistic propaganda or who belongs to any group that does.”[260] The Attorney General filed a complaint for declaratory relief from its duty to prepare a circulating title and summary of the initiative on that grounds that the proposed measure was “patently unconstitutional on its face,” and that “[r]equiring the Attorney General to prepare a circulating title and summary would be inappropriate, waste public resources, generate unnecessary divisions among the public, and mislead the electorate.”[261] The proponent did not respond to the complaint, and the trial court entered a default judgment in the Attorney General’s favor, relieving it of “any obligation to prepare a title and summary of the Act.”[262] And in Planning & Conservation League v. Padilla,[263] the California Supreme Court directed the Secretary of State to refrain from placing on the ballot a proposed initiative measure to split California into three states, holding that such relief was warranted because “significant questions have been raised regarding the proposition’s validity, and because . . . the potential harm in permitting the measure to remain on the ballot outweighs the potential harm in delaying the proposition to a future election.”[264]

Accordingly, judicial review is an essential tool to police the initiative power and to ensure that it is not used to violate fundamental individual rights secured by the California and U.S. constitutions. The judiciary is adequately equipped in this area because the courts have a well-developed role and clear guidelines for policing initiative excesses to ensure the electorate remains within the lines drawn by the state and federal constitutions. In combination, these process and substantive limits on the electorate’s legislative power have on the whole proved to be capable at keeping the electorate in its lane. Given that, and the results of our data analysis, other than an incremental improvement (like our quorum idea) we see no need for major structural reforms to the initiative.


Direct democracy in California government is a net social good.[265] Rather than weakening the democratic process by removing decisions from elected representatives (thereby reducing their authority, removing incentives to act, and degrading the legitimacy of their acts), direct democracy can strengthen the democratic process by checking the legislature and contributing to legislative results that more closely conform to community views. Combining direct democracy and representative republicanism moots the debate over which system better produces optimal results. California’s experience belies the conventional wisdom: the legislature, not the electorate, is the primary constitutional change actor; the electorate is reliably reticent to pass initiatives; and the initiative is not to blame for the length and mutability of California’s constitution. Consequently, it is difficult to argue that the state is the fifth largest economy in the world despite the initiative.[266]

Direct democracy remains a popular institution in California, albeit one colored by the pervasive voter frustration with state government as a whole.[267] A significant majority of the electorate believes voters should have a direct say in making law and public policy through the initiative process, while a similarly large majority believes that the initiative process needs reformwith some of the most favored changes potentially making the initiative a more powerful political force.[268] Accordingly, despite its defects, the electorate is highly unlikely to approve any limits on its powers, and direct democracy will remain a powerful state governmental institution.[269] All things considered, that’s not so bad.




Between 1912 and 2016:

  • 89 referenda were titled and summarized for circulation.
  • 39 referenda (43.82%) failed to qualify for the ballot, and 50 referenda (56.18%) qualified for the ballot.
  • Of the 50 which qualified and have been voted on, 21 referenda (42%) were approved by the voters.*
  • 29 referenda (58%) were rejected by the voters.*

* Once a referendum is on the ballot, the law is repealed only if voters cast more NO votes than YES votes on the referendum in question. Accordingly, research regarding how many referendum campaigns are successful in repealing a law, should consider a referendum that was “rejected” by the voters (which thereby strikes down an existing law) as agreement by the majority of voters that the law should be repealed. Therefore, as of the end of 2016, 58% of the referenda that qualified for the ballot were successful in repealing a law.



[*] *. Lecturer in Residence and Executive Director of the California Constitution Center at the University of California, Berkeley School of Law; B.A. (1991), J.D. (1995), LL.M. (2007), and J.S.D. (2011), University of California, Berkeley.

[†] †. Senior Research Fellow, California Constitution Center; B.A., University of California, Berkeley; J.D., University of Notre Dame Law School. Mr. Duvernay is an attorney in private practice.

[‡] ‡. Senior Research Fellow, California Constitution Center; B.A., University of California, Santa Cruz; J.D., University of California, Davis. Mr. Gevercer is an attorney in private practice.

[*] **. Senior Research Fellow, California Constitution Center; B.A., Tufts University; M.A., New York University; J.D., University of California, Berkeley. Ms. Fenzel is an attorney in private practice.

[*]  The views expressed herein are solely those of the authors. The authors presented an early draft of this Article at the Korea Legislation Research Institute conference on February 2, 2018 in Seoul, South Korea, as a primer on California constitutional law and the state’s experience with direct democracy. Portions of this Article draw from the authors’ previous work on these subjects, including: David A. Carrillo & Stephen M. Duvernay, California Constitutional Law: The Guarantee Clause and California’s Republican Form of Government, 62 UCLA L. Rev. Disc. 104 (2014); David A. Carrillo, Stephen M. Duvernay & Brandon V. Stracener, California Constitutional Law: Popular Sovereignty, 68 Hastings L.J. 731 (2017); David A. Carrillo & Danny Y. Chou, California Constitutional Law: Separation of Powers, 45 U.S.F. L. Rev. 655 (2011); and David A. Carrillo & Stephen M. Duvernay, The California Judiciary, 7 Cal. J. Pol. & Pol’y, no. 4, 2015, at 1. Thanks are due to our research assistants, the brothers Belcher, for their dedicated assistance on this project.

 [1]. It takes little work to find scholarly and popular press criticism of California’s direct democracy tools. See generally, e.g., Arne R. Leonard, In Search of the Deliberative Initiative: A Proposal for a New Method of Constitutional Change, 69 Temp. L. Rev. 1203 (1996); Note, California’s Constitutional Amendomania, 1 Stan. L. Rev. 279 (1949); Harry N. Scheiber, Foreword: The Direct Ballot and State Constitutionalism, 28 Rutgers L.J. 787 (1997); Rachel A. Van Cleave, A Constitution In Conflict: The Doctrine of Independent State Grounds and the Voter Initiative in California, 21 Hastings Const. L.Q. 95 (1993); Direct Democracy: Origin of the Species, Economist (Apr. 20, 2011),; Power from the People, Economist (July 6, 2013), (“Direct democracy is often blamed for making California ungovernable.”); Proposition 13: War By Initiative, Economist (Apr. 20, 2011),
/20/war-by-initiative; California’s Legislature: The Withering Branch, Economist (Apr. 20, 2011), We note that the criticism is not universal and that California’s direct democracy has other defenders. See, e.g., Zev Yaroslavsky, Can Californians Handle Direct Democracy?, L.A. Times (Nov. 6, 2016),

 [2]. Eastlake v. Forest City Enters., Inc., 426 U.S. 668, 672–73 (1976) (citing The Federalist, No. 39 (James Madison)) (noting that the power to govern comes entirely from the people, who can delegate powers to their representatives and reserve powers to themselves); Brosnahan v. Brown, 651 P.2d 274, 277 (Cal. 1982); C&C Construction, Inc. v. Sacramento Mun. Util. Dist., 18 Cal. Rptr. 3d 715, 727 (Ct. App. 2004) (“In California, the people are sovereign, whose power may be exercised by initiative.”). For the distinction in California law between the people and the electorate, see People v. Lynch, 51 Cal. 15, 27–28 (1875):

But the “sovereignty of the people” is more than a meaningless phrase. The people of California created the State government, and it was for this people to place (in the State Constitution) as many checks upon, and conditions and limitations of the general grant of legislative, executive or judicial power as they deemed proper or expedient. The people of the State alone possess and can exercise supreme and absolute authority; the Legislature, and the other departments of government, are but the depositaries of delegated powers more or less limited—according to the terms of the Constitution.

Id. (internal citations and quotations omitted); see also Karl Manheim & Edward P. Howard, A Structural Theory of the Initiative Power in California, 31 Loy. L.A. L. Rev. 1165, 1191–92 (1998) (“California’s constitution thus gives a name to the power of self-governance. The ability of individuals to ‘create’ and regulate government institutions is dubbed the ‘political power.’ This is the organic power of a sovereign polity. It has been invoked twice in California, in the 1849 and 1879 conventions.”); Herman Belz, Popular Sovereignty, the Right of Revolution, and California Statehood, 6 Nexus 3, 11 (2001) (noting that popular sovereignty is the right of self-government inherent in any community, the right of internal legislation in a community).

 [3]. Other than the United States, only Switzerland makes substantial use of direct democracy. Ronald Steiner, Understanding the Prop 8 Litigation: The Scope of Direct Democracy and Role of Judicial Scrutiny, 14 Nexus 81, 83 (2009). But see Dennis C. Mueller, Constitutional Democracy 100, n.5 (1996) (noting that modern Japan, Poland, Iceland, Turkey, the former West Germany, England, and Wales have used popular assemblies on a small scale). In the United States, it is primarily an artifact of Progressive politics in the central and western states. Robert F. Williams, State Constitutional Law Processes, 24 Wm. & Mary L. Rev. 169, 205 (1983); Steiner, supra, at 84.

 [4]. For an excellent contemporary overview of state constitutional change mechanisms, see John Dinan, State Constitutional Politics: Governing by Amendment in the American States 11–23 (2018). Note that there is some variation in the various tabulations of how many states have which initiative procedures (for example, in the authorities cited infra note 5), and for consistency, we employ John Dinan’s numbers because they are the most recent.

 [5]. Mark Baldassare & Cheryl Katz, The Coming Age of Direct Democracy 9–11 (2008) (noting the most recent state to adopt the initiative was Mississippi in 1992); Shaun Bowler & Amihai Glazer, Direct Democracy’s Impact on American Political Institutions 2, 35 (Palgrave Macmillan eds., 2008); Bruce E. Cain & Roger G. Noll, Constitutional Reform in California 265 (1995); Thomas E. Cronin, Direct Democracy: The Politics of Initiative, Referendum, and Recall 47, 51, tbl. 3.1 (1999); Dinan, supra note 4, at 16–17; Lawrence LeDuc, The Politics of Direct Democracy 137 (2003) (thirty-one states have some kind of referendum process, twenty-four have the initiative specifically); Tracy M. Gordon, Pub. Policy Inst. of Cal., The Local Initiative in California 3 (2004); see also Initiative and Referendum Institute, Univ. S. Cal., (last visited Apr. 9, 2019).

 [6]. Bowler & Glazer, supra note 5, at 1.

 [7]. See Gordon S. Wood, The Creation of the American Republic, 1776–1787, at 18–19 (1998).

 [8]. Id. at 19.

 [9]. Jonathan Zasloff, Taking Politics Seriously: A Theory of California’s Separation of Powers, 51 UCLA L. Rev. 1079, 1122–23 (2004).

 [10]. Mueller, supra note 3, at 56, 83, 85.

 [11]. Wood, supra note 7, at 363–72.

 [12]. See Leroy A. Wright, Reasons Why Senate Constitutional Amendment No. 22 Should Not Be Adopted, in Proposed Amendments to the Constitution of the State of California, with Legislative Reasons for and Against the Adoption Thereof 8, 8 (1911) (“[The initiative’s] tendency is to change the republican form of our government and head it towards democracy, and history teaches that democracies have universally ended in turbulence and disaster.”).

 [13]. David A. Carrillo, Stephen M. Duvernay & Brandon V. Stracener, California Constitutional Law: Popular Sovereignty, 68 Hastings L.J. 731, 747–51 (2017).

 [14]. See Wood, supra note 7, at 21.

 [15]. Carrillo et al., supra note 13, at 751–62.

 [16]. Wood, supra note 7, at 20.

 [17]. Id.; Perry v. Brown, 265 P.3d 1002, 1027 (Cal. 2011) (noting the electorate’s “authority to propose and adopt state constitutional amendments or statutes embodied in the initiative provisions of the California Constitution is essentially a legislative authority”); Carrillo et al., supra note 13, at 747–50; see also Adrian Vermeule, Mechanisms of Democracy 67 (2007) (arguing that self-interested enactments by a popular majority, even if temporary, “systemically tend to enjoy a protection against subsequent appeal that impartial ones do not possess” because such enactments will have a “core group of intensely interested defenders around to defend them from repeal” that impartial enactments lack).

 [18]. See, e.g., John M. Allswang, The Initiative and Referendum in California, 1898–1998, at 3–4 (2000) (“[California] has used these mechanisms almost constantly and with accelerating frequency throughout the twentieth century—more so than any other state.”).

 [19]. Consider, for example, Proposition 140 (Cal. 1990) (imposing term limits and solving the problem of effectively lifetime legislative seats); Proposition 11 (Cal. 2008) and Proposition 20 (Cal. 2010) (creating the California Citizens Redistricting Commission and solving the problem of the legislature being unable to agree on redistricting); and Proposition 25 (Cal. 2010) (solving the problem of the perennially late state budget by removing the two-thirds vote requirement for a revenue-neutral budget and docking legislator pay after the budget deadline).

 [20]. Mueller, supra note 3, at 60.

 [21]. The Federalist No. 10, at 74–79 (James Madison); Gordon, supra note 5, at 7.

 [22]. In fact, a popular vote was disfavored at the time of the nation’s founding. The U.S. Constitution was ratified by the states, not by plebiscite, and only a few early state constitutions were popularly approved. Charles A. Beard & Birl E. Shultz, Documents on the State-Wide Initiative, Referendum and Recall 15, 28–29 (1912).

 [23]. Gordon, supra note 5, at 7. But see Akhil Reed Amar, The Consent of the Governed: Constitutional Amendment Outside Article V, 94 Colum. L. Rev. 457 (1994) (arguing for an unenumerated right of a majority of voters to amend the federal constitution); Akhil Reed Amar, Philadelphia Revisited: Amending the Constitution Outside Article V, 55 U. Chi. L. Rev. 1043 (1988) (arguing for the unenumerated rights of voters to amend the Constitution).

 [24]. John Ross Browne, Report of the Debates of the Convention of California, on the Formation of the State Constitution, in September and October, 1849, at 34 (statement by Mr. Norton).

 [25]. Cal. Const. art II, § 1. Compare this with the Swiss concept of popular sovereignty, where the people are the supreme authority. Swiss cantons began experimenting with direct democracy in the 1830s, and the Swiss constitution has contained the initiative power since 1848. Gordon, supra note 5, at 7, n.1. Under the Swiss constitution, the Swiss people are sovereign and ultimately the supreme political authority; the concept includes all Swiss adults who are eligible to vote—approximately 4.8 million citizens, or 60% of the population. Bundesverfassung [BV] [Constitution] Apr. 18, 1999, tit. 5, ch. 2, art. 148 (Switz.), translated at

 [26]. Joseph R. Grodin et al., The California State Constitution 28–29 (Oxford Univ. Press, 2d ed., 2016); Beard & Shultz, supra note 22; Stephen H. Sutro, Interpretation of Initiatives by Reference to Similar Statutes: Canons of Construction Do Not Adequately Measure Voter Intent, 34 Santa Clara L. Rev. 945, 948 (1994).

The initiative process has been characterized as a “legislative battering ram”—a tool for the populace to enact legislation ignored by elected representatives. Lobbyist control of Sacramento at the turn of the century prompted California professionals and small businessmen to push the initiative process as a means to give power back to the people. Accordingly, the initiative process was designed to allow grassroots access to law-making. Structurally, the process is relatively unchanged from its original form of 1911.

Sutro, supra, at 948.

 [27]. Beard & Shultz, supra note 22 passim; Bowler & Glazer, supra note 5, at 6.

 [28]. Baldassare & Katz, supra note 5, at 7; Grodin et al., supra note 26, at 29; Gordon, supra note 5, at 8; Williams, supra note 3, at 205.

 [29]. As the California Supreme Court explained

[i]n California, a principal target of the [progressive] movement’s ire was the Southern Pacific Railroad, which the movement’s supporters believed not only controlled local public officials and state legislators but also had inordinate influence on the state’s judges, who—in the view of the progressive movement—at times improperly had interpreted the law in a manner unduly favorable to the railroad’s interest.

Strauss v. Horton, 207 P.3d 48, 84 (Cal. 2009); see also Cronin, supra note 5, at 56–57 (noting that the direct democracy reforms were not the “‘panacea for all our ills,’ said California governor Hiram Johnson, ‘yet they do give the electorate the power of action when desired, and they do place in the hands of the people the means by which they may protect themselves’”); Baldassare & Katz, supra note 5, at 9; Gordon, supra note 5, at 1. Senate Constitutional Amendment 22 was proposed by the legislature under the procedure provided by Article 18 section 1, which does not distinguish between the procedure for the legislature to propose amendments or revisions. The version of Article 18 section 1 from the 1879 constitution, in effect in 1911, provided:

Any amendment or amendments to this Constitution may be proposed in the Senate or Assembly, and if two-thirds of all the members elected to each of the two Houses shall vote in favor thereof, such proposed amendment or amendments shall be entered in their Journals, with the yeas and nays taken thereon; and it shall be the duty of the Legislature to submit such proposed amendment or amendments to the people in such manner, and at such time, and after such publication as may be deemed expedient. Should more amendments than one be submitted at the same election they shall be so prepared and distinguished, by numbers or otherwise, that each can be voted on separately. If the people shall approve and ratify such amendment or amendments, or any of them, by a majority of the qualified electors voting thereon, such amendment or amendments shall become a part of this Constitution.

See also Robert Desty, The Constitution of the State of California 362 (Sumner Whitney & Co., 1879).

 [30]. Governor Johnson put it this way:

How best can we arm the people to protect themselves hereafter? If we can give to the people the means by which they may accomplish such other reforms as they desire, the means as well by which they may prevent the misuse of the power temporarily centralized in the Legislature . . . then all that lies in our power will have been done in the direction of safeguarding the
future. . . . And while I do not by any means believe the initiative, the referendum, and the recall are the panacea for all our political ills, yet they do give to the electorate the power of action when desired, and they do place in the hands of the people the means by which they may protect themselves. . . . The opponents of direct legislation and the recall, however they may phrase their opposition, in reality believe the people can not be trusted. On the other hand, those of us who espouse these measures do so because of our deep-rooted belief in popular government, and not only in the right of the people to govern, but in their ability to govern.

Inaugural Address of Governor Hiram Johnson (Jan. 3, 1911), in Franklin Hichborn, Story of the Session of the California Legislature of 1911, at iv–v (James H. Barry Co., 1911).

 [31]. Cal. Const. art. XVIII, § 1 (1879).

Any amendment or amendments to this Constitution may be proposed in the Senate or Assembly, and if two-thirds of all the members elected to each of the two Houses shall vote in favor thereof . . . it shall be the duty of the Legislature to submit such proposed amendment or amendments to the people. . . . If the people shall approve and ratify such amendment or amendments . . . by a majority of the qualified electors voting thereon, such amendment or amendments shall become a part of this Constitution.

 [32]. Initiative and Referendum. California Proposition 7 (1911), U. Cal. Hastings C.L., (last visited Apr. 9, 2019) (Senate Constitutional Amendment 22); Recall by the Electors of Public Officials, Proposition 8 (1911), U. Cal. Hastings C.L., (last visited Apr. 9, 2019) (Senate Constitutional Amendment 23).

 [33]. The direct democracy provisions were approved as Proposition 7 (initiative and referendum) and Proposition 8 (recall). There is an argument that adding the direct democracy improperly revised the state constitution in 1911, see Manheim & Howard, supra note 2, at 1230–31, 1235 (concluding “[s]o what! Given the ethereal ill-understood nature of how popular sovereigns gain widespread legitimacy, is not the foregoing analysis mere formalism?”). We agree. To the extent it was a revision, that process requires a legislative proposal and popular vote, which is what happened. After more than a century of judicial and political acceptance, this is at most an interesting academic argument.

 [34]. Baldassare & Katz, supra note 5, at 13; Gordon, supra note 5, at 23; see also, e.g., Allswang, supra note 18, at 1. Regardless of which state is number one, at least one commentator argues that California has set the standard for direct democracy. LeDuc, supra note 5, at 149.

 [35]. Statewide Initiatives Since 1904–2000, Initiative & Referendum Inst., Univ. S. Cal., (last visited Apr. 9, 2019) (individually describing and tabulating every initiative measure on each state’s ballot by year in the given period). The California Secretary of State calculates a different number of ballot-qualified initiatives for this period (1904–2000): 286. California would rank second with either figure. History of California Initiatives 1912–2017, Cal. Sec’y State, (last visited Apr. 9, 2019)

 [36]. Dinan, supra note 4, at 16–17.

 [37]. Cal. Const. art. II, § 13 (recall), Cal. Const. art. II, § 9(a) (referendum), Cal. Const. art. II, § 8(a) (initiative).

 [38]. Cal. Const. art. II, § 13; accord Gordon, supra note 5, at 1. The electorate attempted to recall sitting governors thirty-two times between 1911 and 2003, but the recall of Governor Gray Davis was the first successful attempt in the state, and only the second time that the governor of any state had ever been recalled (the first was North Dakota Governor Lynn Frazier in 1921). Baldassare & Katz, supra note 5, at 11.

 [39]. Baldassare & Katz, supra note 5, at 1.

 [40]. Ann Bowman & Richard C. Kearney, State and Local Government 98 (Wadsworth Publishing, 10th ed. 2016). The election in 1986 when three California Supreme Court justices (including the Chief Justice) were removed from the bench is sometimes mentioned in this context. This is incorrect; those justices were voted out in a regular retention election, rather than through a recall. Recalls of judges are exceedingly rare, in California and in general. See Cal. Constitution Ctr., What Does California’s Experience with Recall of Judges Teach Us?, SCOCABlog (Nov. 10, 2016),

 [41]. See Direct Democracy Database, Int’l Inst. Democracy & Electoral Assistance, (last visited Apr. 10, 2019) (defining in its glossary a referendum as “[a] direct democracy procedure consisting of a vote of the electorate on an issue of public policy such as a constitutional amendment or a draft law. Also known as popular consultation or a plebiscite”).

 [42]. Cal. Const. art. II, § 9(a); Gordon, supra note 5, at 1; Mueller, supra note 3, at 177–78.

 [43]. See, e.g., California Ballot Measures, Berkeley L. Libr.,
/library/dynamic/guide.php?id=29 (last updated Sept. 18, 2012) (“Despite a recent uptick in use of this device (9 referenda filed in 2011 alone, compared to less than 15 per decade since inception, and only 1 or 2 per decade in the 40s, 50s, 60s, 70s and 90s), the history of referenda in California can still be summarized in less than five pages.” (emphasis removed)).

 [44]. Summary of Data, Cal. Sec’y of State, (last visited Apr. 10, 2019) (providing the California Secretary of State’s summary of California referendum results). As the Secretary of State’s summary notes, a law is repealed by referendum

only if voters cast more NO votes than YES votes on the referendum in question. Accordingly, research regarding how many referendum campaigns are successful in repealing a law, should consider a referendum that was “rejected” by the voters (which thereby strikes down an existing law) as agreement by the majority of voters that the law should be repealed. Therefore, as of the end of 2016, 58% of the referenda that qualified for the ballot were successful in repealing a law.


 [45]. Cal. Const. art. II, § 8(a); Gordon, supra note 5, at 1; Mueller, supra note 3, at 178. California had both a direct citizens’ initiative and an indirect legislative initiative until 1966, when the electorate abolished the indirect process, in part due to its lengthy pre-election circulation period. Baldassare & Katz, supra note 5, at 10. In 1965, the Constitution Revision Commission recommended that the indirect initiative process be eliminated due to disuse. Cal. Sec’y of State, A History of California Initiatives 9 (2002).

 [46]. City of Malibu v. Cal. Coastal Comm’n, 18 Cal. Rptr. 3d 40, 48 (Ct. App. 2004) (“Good governance cannot permit local voters to override a state decision with a local referendum. . . . [W]hether legislative or administrative . . . to permit local voters to overturn state enactments would upend our governmental structure and invite chaos.”); see also Jahr v. Casebeer, 83 Cal. Rptr. 2d 172, 176–77 (Ct. App. 1999) (discussing state preemption and limits on local referenda). The voters in cities and counties have local initiative and referendum powers. Cal. Const. art. II, §11. It is generally co-extensive with the legislative power of the local governing body. DeVita v. County of Napa, 889 P.2d 1019, 1026 (Cal. 1995); Simpson v. Hite, 222 P.2d 225, 228 (Cal. 1950). It may even be broader than the statewide initiative power. Rossi v. Brown, 889 P.2d 557, 561 (Cal. 1995).

 [47]. The California constitution grants amendment power only to the electorate. Cal. Const. art. II, § 8; art. XVIII, §§ 3, 4; Strauss v. Horton, 207 P.3d 48, 79–80 (Cal. 2009) (noting that a proposed amendment or a proposed revision of the Constitution must be submitted to the voters, and becomes effective if approved by a majority of votes cast thereon at the election); Rossi v. Brown, 889 P.2d 557, 561 n.3 (Cal. 1995). The initiative is not a right granted to the electorate, it is a power reserved by them. Associated Home Builders of the Greater Eastbay, Inc. v. City of Livermore, 557 P.2d 473, 477 (Cal. 1976).

 [48]. Cain & Noll, supra note 5, at 279 (explaining the distinctions between the people’s political power and the electorate’s initiative power); see also Carrillo et al., supra note 13, at 743–47; Manheim & Howard, supra note 2, at 1194–96.

 [49]. Cal. Const. art. II, § 10(c) (“The Legislature may amend or repeal an initiative statute by another statute that becomes effective only when approved by the electors unless the initiative statute permits amendment or repeal without the electors’ approval.”); see also Sutro, supra note 26, at 949. The Governor’s veto power applies only to bills passed by the Legislature. Cal. Const. art. IV, § 10(a).

 [50]. See infra Table 6 (all amendments by type).

 [51]. The California legislature’s ability to propose constitutional amendments remained unchanged after the 1911 amendments that introduced the electorate’s ability to do the same by itself. So going forward from 1911, we distinguish between legislative constitutional amendments (those placed on the ballot by the legislature) and initiative constitutional amendments (those placed on the ballot by the electorate).

 [52]. Consistent with the constitutional and doctrinal distinction between amendments and revisions, we count them separately.

 [53]. Baldassare & Katz, supra note 5, at 3.

 [54]. Sherry Bebitch Jeffe, A History Lesson on Part-Time Lawmaking, L.A. Times (Aug. 8, 2004),

 [55]. See infra Table 6 (all amendments by type). For other estimates, see, for example, Grodin, et al., supra note 26, at 29 (120 initiatives approved from 1914–2012); California Research In-Depth: Constitution, Georgetown L. Libr., (last updated Dec. 10, 2018) (“California’s current constitution was ratified on May 7, 1879 and has been amended over 480 times.”).

 [56]. Dinan, supra note 4, at 25–26 tbl.1.3 (showing that Alabama has 926 amendments, nearly double California’s); see also Bowler & Glazer, supra note 5, at 172; Cain & Noll, supra note 5, at 265.

 [57]. Cain & Noll, supra note 5, at 265. But see Baldassare & Katz, supra note 5, at 10 (“Most states are in the range of 5 to 8 percent of voters participating in the last gubernatorial election.”).

 [58]. See, e.g., Lessons from California: The Perils of Extreme Democracy, Economist (Apr. 20, 2011),

 [59]. Cain & Noll, supra note 5, at 267.

 [60]. See infra Table 6.

 [61]. Dinan, supra note 4, at 11. For example, every state constitution permits its legislature to generate amendments. Id. at 11, 13.

 [62]. Id. at 23. California’s amendment rate ranks fourth among the states, after Alabama, Louisiana, and South Carolina. Id. at 25–26 tbl.1.3.

 [63]. Using John Dinan’s average of 150 amendments per state and 1.3 amendments per year, id. at 23, California exceeds both figures with 518 total amendments and three amendments on average per year 1850–2017, see infra Table 6.

 [64]. See Cain & Noll, supra note 5, at 275–77 (“[I]t would be wrong to blame the policy orientation of the California Constitution per se or its high rate of amendability on the initiative.”). The authors argue that the age and complexity factors contribute to California’s amendment rate and point out that constitutions (like California’s) adopted during the late 1800s are populist documents, and California adopted its direct democracy tools at the height of the Progressive era. Id. at 276 (“[T]he main causes of California’s constitutional hyper-amendability are the era in which it was adopted and the influence that the populist and Progressive movements had on its contents.”).

 [65]. California’s constitution has featured significant constitutional legislation since the original 1849 constitution was debated. Brown, supra note 24, at 33, (“The proposed bill is objectionable. It embraces legislative enactments. . . . When a Convention assumes to pass laws and impose them upon the people, it constitutes itself an oligarchy.”) (statement by Mr. Botts); id. at 41 (arguing for no legislative enactments in a bill of rights as that subject belongs in statute books) (statements by Mr. McCarver, Mr. Ord, and Mr. Jones); id. at 42 (“While taking the first step in the first movement to form the first fundamental law of the new State, it would be improper to insert legislative enactments for her government five, ten, or twenty years hence.”) (statement by Mr. Shannon).

 [66]. Dinan, supra note 4, at 25–26 tbl.1.3.

 [67]. Dinan, supra note 4, at 16–17 and at 17 tbl.1.2. We exclude Massachusetts and Mississippi because the legislatures in those states can either block or change initiative amendments.

 [68]. Cal. Constitution Ctr., California’s Constitution Is Not the Longest, SCOCABlog (June 24, 2017), And California’s is not the longest constitution in the world: it is the eighth longest constitution worldwide. Id. Cain and Noll argue that the more topics covered by a constitution, the greater the likely perceived need for amendment over time, and that California’s constitution covers the widest range of topics with the greatest degree of specificity compared with the other states. Cain & Noll, supra note 5, at 273, 276. Note that others reach different results on this issue. See, e.g., Grodin, et al., supra note 26, at 23 (noting California has the world’s third-longest constitution after India and Louisiana) (citing Brian P. Janiskee & Ken Masugi, Democracy in California: Politics and Government in the Golden State (Rowman & Littlefield, 3d ed. 2011)).

 [69]. But see Dinan, supra note 4, at 28 (“Every major study has concluded that the longer and more detailed state constitutions are amended more frequently than short and spare constitutions.”).

 [70]. There is no scholarly consensus on why some state constitutions are amended more or less than others, and in particular there is disagreement about the citizen initiative amendment’s effects. Dinan, supra note 4, at 24–30.

 [71]. See, e.g., Manheim & Howard, supra note 2, at 1173.

 [72]. The source data from Tables 3, 4, 5, and 6 and Figures 1–8, is taken from the California Secretary of State study, History of California Initiatives 1912–2002, Cal. Sec’y of State,
-initiatives (last visited Apr. 10, 2019), along with an updated 2011 version of the same table provided directly to us by Secretary of State staff, data on titled initiatives provided by the initiative coordinator at the California Office of the Attorney General, the Initiative and Referendum Institute at the University of Southern California Historical Database, and the Hastings College of the Law California Ballot Pamphlet, Propositions and Initiatives databases. The Secretary of State numbers are only current to 2017; as of this article’s publication, the 2018 numbers were not available. And note that the California Supreme Court occasionally removes measures from the ballot; these few instances are included as rejected. See Baldassare & Katz, supra note 5, at 6; Cain & Noll, supra note 5, at 268; Ctr. for Governmental Studies, Democracy by Initiative: Shaping California’s Fourth Branch of Government 2 tbl.1, 6 tbl.2, 12 tbl.5 (2d ed. 2008).

 [73]. There are conflicting study results about whether voter participation and approval are related to ballot position. See Cronin, supra note 5, at 68­–69. We think this supports the idea that there is a maximum effective use limit for the initiative that is independent of how many proposals are on the ballot.

 [74]. See Wyn Grant, Direct Democracy in California: Example or Warning?, in Democracy and North America 133, 137–38 (Alan Ware ed., Frank Cass & Co., 1996) (arguing that while the number of circulated initiatives shows a strong upward trend, the number qualified does not rise as quickly, and the gap between circulated and approved initiatives is much wider than that between qualifying and approved initiatives.). But see Baldassare & Katz, supra note 5, at 17 (arguing that the overall rate of initiative passage has increased from an average of 35% in the 1900’s to 53% from 2000 to 2006). Note that Baldassare and Katz worked from partial data. With the benefit of data for the whole period of 2000–2010, the research here shows that passage rates during that period are within the normal range.

 [75]. The increase in initiatives and amendments also coincides with the anti-tax movement that caused the passage of the property tax reform initiative Proposition 13 in 1978. Baldassare & Katz, supra note 5, at 75; Gordon, supra note 5, at 2 fig.1.1. Doubtless there are sociological contributing factors to explain the cyclical rise, fall, and rise of initiative usage, such as distraction during and disinterest following World War II. Baldassare & Katz, supra note 5, at 12; Bowler & Glazer, supra note 5, at 5 (showing initiatives declined during and after World War II to a low in the 1960’s, increasing beginning in the late 1970’s and continuing to the present).

 [76]. See, e.g., Allswang, supra note 18, at 248 (“[T]here is no evidence that voters make much distinction between an initiative that is a statute and one that is a constitutional amendment.”).

 [77]. Our results for California are consistent with conditions in other states: “Legislature-referred amendments make up the vast majority of recent amendments, generally about 90 percent of all amendments in each election cycle.” Dinan, supra note 4, at 34.

 [78]. This analysis does not cover the nature of the initiatives tabulated here, in the sense of whether they advanced the aims of a particular political party or interest group, or more generally the distribution of conservative or liberal principles in initiatives. Note, however, that there is some support for the conclusion that California initiatives overall do not indicate any bias in favor of liberal or conservative causes. See Robert D. Cooter, The Strategic Constitution 144 (2000).

 [79]. Allswang, supra note 18, at 247 (“[T]he proportion of initiatives that ends up in the court system has greatly increased in recent years.”); Bowler & Glazer, supra note 5, at 152 (“[A]bout half of the initiatives passed in California . . . between 1960 and 1999 faced legal challenges . . . and many had significant portions of their content invalidated . . . .”).

 [80]. This convergence provides some support for Allswang’s conclusion that “the direct legislation process is having a greater-than-ever effect on current California and even national affairs,” Allswang, supra note 18, at 245, but given our other findings, we would not characterize this evidence so strongly.

 [81]. See Baldassare & Katz, supra note 5, at 221.

 [82]. See generally James Surowiecki, The Wisdom of Crowds: Why the Many Are Smarter than the Few and How Collective Wisdom Shapes Business, Economies, Societies, and Nations (2004).

 [83]. See Thomas Fuller, The Pleasure and Pain of Being California, the World’s 5th-Largest Economy, N.Y. Times, (May 7, 2018),; Gross State Product, Cal. Dep’t. of Fin.,
/Economics/Indicators/Gross_State_Product (last visited Apr. 10, 2019); Regional Data: GDP and Personal Data, Bureau Econ. Analysis,
=70&step=10&isuri=1&7003=200&7004=naics&7035=-1&7005=1&7006=xx&7001=1200&7036=-1&7002=1&7090=70&7007=-1&7093=levels (last visited Apr. 10, 2019).

 [84]. See Cal. Dep’t of State, Reasons Why Senate Constitutional Amendment No. 22 Should Be Adopted, in Proposed Amendments to the Constitution of the State of California, with Legislative Reasons for and Against the Adoption Thereof 5, 5–6 (1911) (“It is not intended and will not be a substitute for legislation, but will constitute that safeguard which the people should retain for themselves, to supplement the work of the legislature by initiating those measures which the legislature either viciously or negligently fails or refuses to enact; and to hold the legislature in check, and veto or negative such measures as it may viciously or negligently enact.”).

Moreover, a study of the history of the initiative and referendum in those states where they have been in vogue shows that representative government is not destroyed. In most states the system has scarcely been applied at all, and remains in abeyance to be used whenever any considerable portion of the voters think that the legislature has failed to do its duty; and even in Oregon, where the system has been most extensively used, the legislature has been by no means abolished, or even set on the way to destruction.

Beard & Shultz, supra note 22, at 22–23, 37 (discussing “the advantages which the representative system affords in initiation may be combined with those of popular initiative”).

 [85]. See generally David A. Carrillo & Danny Y. Chou, California Constitutional Law: Separation of Powers, 45 U.S.F. L. Rev. 655 (2011).

 [86]. California Term Limits, Proposition 140 (1990), Ballotpedia,
/California_Term_Limits,_Proposition_140_(1990) (last visited Apr. 10, 2019). This proposition modified and added to the California Constitution, see Cal. Const. art. IV, §§ 1.5, 2, 4.5, 7.5; art. V, § 11; art. VII, § 11(d); art. IX, § 2; art. XIII, § 17; art. XX, § 7).

 [87]. California Proposition 20, Congressional Redistricting (2010), Ballotpedia,,_Congressional_Redistricting_(2010) (last visited Apr. 10, 2019) (modifying and adding to the California Constitution, see Cal. Const. art. XXI, §§ 1–3); California Proposition 25, Majority Vote for Legislature to Pass the Budget (2010), Ballotpedia,,_Majority_Vote_for_Legislature_to_Pass_the_Budget_(2010) (last visited Apr. 10, 2019) (modifying and adding to the California Constitution, see Cal. Const. art. IV, § 12).

 [88]. For example, the legislatively-referred constitutional amendment Proposition 14 in 2010 abolished the party primary system, replacing it with a single open primary with a top-two finish regardless of party.

 [89]. Bowler & Glazer, supra note 5, at 1–2, 5 (discussing how the initiative process lacks critical legislative process elements and has intended and unintended effects on ability of representative government to develop comprehensive policy).

 [90]. One Hundred and Thirty-First Day, in 3 Debates and Proceedings of the Constitutional Convention of the State of California 1268, 1277 (Sacramento, J. D. Young, Supt. State Prtg. 1881) (statement of Mr. Ayers) (“It is true that large bodies are unwieldy and move slowly, but they move surely and justly, and they are representative in their character. They take in and represent all the diversified interests of the State, and every measure is thoroughly and exhaustively discussed before it is acted upon.”).

 [91]. Wright, supra note 12, at 8.

It may be easy to determine what the effect of a given law will be upon a certain trade or a particular community, but its ramifications often extend beyond the vision of the wisest. Well-meaning laws not infrequently bring about results not contemplated. . . . (b) . . . No law should be enacted without a systematic study of its necessity, and the injury it may inflict as well as the evil it is intended to correct. . . . (c) Any ill-considered law is dangerous to the public good. . . . (g) Every law before being enacted should be submitted to some forum in which it is subject to deliberation and amendment. Under the proposed initiative and referendum no amendment is possible, even though a law should be proposed containing a provision which is palpably unjust and vicious.


 [92]. See Cooter, supra note 78, at 28 (presented with several choices, citizens vote strategically); id. at 214–15 (raising transaction costs decreases demand for enacting legislation, causing increased total expenditures on legislation focused on laws considered necessary not merely desirable, privileging the status quo).

 [93]. Bowler & Glazer, supra note 5, at 128; Cooter, supra note 78, at 214–15 (raising transaction costs decreases demand for enacting legislation, causing increased total expenditures on legislation focused on laws considered necessary not merely desirable, privileging the status quo).

 [94]. Bowler & Glazer, supra note 5, at 16.

 [95]. Allswang, supra note 18, at 248 (“This chicken-and-egg argument has been around for a long time, indeed since Progressivism.”); Grant, supra note 74, at 142; The Withering Branch, supra note 1 (“California’s legislature must therefore have undergone a stunning decline in the past three decades. What role the initiative process had in this deterioration is a chicken-and-egg question. In Hiram Johnson’s day initiatives seemed to be needed as a check on a venal legislature. Now perhaps a dysfunctional legislature is triggering a plethora of initiatives as citizens take matters into their own hands.”).

 [96]. Allswang, supra note 18, at 249; Bowler & Glazer, supra note 5, at 133; Joseph de Maistre, 1 Lettres et Opuscules Inédits, no. 53 Letter of 15 August 1811 (1851) (“Toute nation a le gouvernement qu’elle mérite”), reprinted in The Yale Book of Quotations 485 (Fred R. Shapiro ed., 2006).

 [97]. Cooter, supra note 78, at 4. Note that competition in government is not an unqualified good—democracy simply is the best overall at minimizing the maximum potential harm from such competition by harnessing it to achieve greater citizen satisfaction with government. See id. at 128–29 (“Increasing political competition carries the hope of improving alignment between the interests of politicians and the preferences of citizens.”); id. at 360 (“Competition does not produce good results as predictably in politics as it does in economics.”).

 [98]. There is a wealth of scholarly and popular debate over direct democracy’s process and result quality. See, e.g., Bowler & Glazer, supra note 5, at 5 (“[T]he accumulating effects of 25 years of initiatives . . . have so hamstrung both state and local governments that elected legislators, county supervisors and school board members have become the handmaidens, not the leaders, of policymaking in California. Because of it they’ve become increasingly unable (and sometimes unwilling) to set priorities and respond to problems when they occur.”); Cronin, supra note 5, at 60–62.

 [99]. R. Jeffrey Lustig, A People’s Convention for California, in Remaking California: Reclaiming the Public Good 195 (R. Jeffrey Lustig ed., 2010) (“In 1930 Governor Young was already complaining that initiative amendments had produced a constitution ‘bad in form, inconstant in particulars, loaded with unnecessary detail, encumbered with provisions of no permanent value, and replete with matter which might more properly be contained in the statute law of the state.’”).

 [100]. See, e.g., id. at 195 (“[The initiative’s] narrow, single-shot focus and insulation from information about their possible consequences at the drafting stage, initiatives are also most conducive to incoherence and disorganization in the political system as a whole.”); Jesse H. Choper, Judicial Review and the National Political Process 14–15 (Univ. of Chicago Press, 1980) (stating that studies show that “a distressingly large percentage of voters is almost totally uninformed” and many voters know little and care less about candidates and issues).

 [101]. See, e.g., Allswang, supra note 18, at 247 (“[I]nitiatives still present voters with a ‘take it or leave it’ situation, where there is no room for compromise.”).

 [102]. Bowler & Glazer, supra note 5, at 12; Vermeule, supra note 17, at 80–81 (discussing “the deliberative virtues of forcing lawmaking to proceed through the hurdles of the legislative process”).

 [103]. Cooter, supra note 78, at 53 (discussing the Coase Theorem, which posits that bargaining tends to succeed as transaction costs approach zero).

 [104]. Bowler & Glazer, supra note 5, at 6–7.

 [105]. Grodin, et al., supra note 26, at 16–19.

 [106]. Id. at 3.

 [107]. LeDuc, supra note 5, at 43.

 [108]. Mark Baldassare, Pub. Policy Inst. of Cal., Reforming California’s Initiative Process 5 (2013), (finding consistently that approximately three-quarters of California voters find there are both too many propositions on the ballots and that proposition wording is too confusing, causing them to question what would happen if an initiative passed).

 [109]. See Just the Facts: The Initiative Process in California, Pub. Policy Inst. Cal. (Oct. 2013),

 [110]. Mueller, supra note 3, at 179.

 [111]. Beard & Shultz, supra note 22, at 34–35.

 [112]. Mueller, supra note 3, at 187–90. One modern problem in particular, the influence of money on campaigns, has similar effects on outcomes in initiative campaigns as it does in representative and legislative issue elections. Id. at 190. See also Beard & Shultz, supra note 22, at 38; One Hundred and Twelfth Day, in 2 Debates and Proceedings of the Constitutional Convention of the State of California 1053, 1063 (Sacramento, J. D. Young, Supt. State Prtg. 1881) (statement of Mr. Hager) (“If we cannot trust the people themselves, how can we trust a Legislature elected by the people? Legislatures have disappointed the people, will the people prove unfaithful to themselves?”).

 [113]. See Cooter, supra note 78, at 145 (“A practical reason compels restricting each ballot initiative to a single issue. Logrolling, which combines issues in a single vote, requires bargaining. Bargaining among different groups requires representation. Ballot initiatives bypass elected representatives. Thus a multiple-purpose ballot initiative invites bargaining without bargaining agents.”). This argument, that multiple-issue voting inevitably fails, is true as far as it goes—it does not prevent competing propositions from qualifying for the ballot, but it may indicate that this feature of California direct democracy is a flaw that invites cycling.

 [114]. Cronin, supra note 5, at 74–75 tbls.4.2 & 4.3.

 [115]. Mueller, supra note 3, at 189 (assuming that people can evaluate both candidates and issues, the question is what set of institutions leads to the optimal outcomes representing the people’s consensus views); see also Beard & Shultz, supra note 22, at 34–35.

 [116]. See Allswang, supra note 18, at 239. One commentator argues that the ballot pamphlet is so important to the electorate’s thought process that judicial review should limit evidence of voter intent to the ballot pamphlet’s contents. Sutro, supra note 26, at 947, 968 (“Voter exposure to initiatives is limited solely to official materials presented in the ballot pamphlet, and judicial review should reflect this.”); see also Sutro, supra note 26 at 973 (“[T]he only reliable source for interpretation of initiative language, other than its common meaning, is the material presented to all voters in the voter pamphlet prior to the election.”).

 [117]. See Bowler & Glazer, supra note 5, at 35–36.

 [118]. Id. at 37, 50–51 (discussing studies that show ballot propositions increase voter turnout especially in low-information election contexts, and so may motivate the population segment least likely to vote); LeDuc, supra note 5, at 151 (stating that voters can and do use various sources of information to learn what they need to know).

 [119]. See Cronin, supra note 5, at 85; see also Grant, supra note 74, at 140–41 (“[V]oters do have enough knowledge and judgment to detect attempts by business interests to use the initiative process to serve their own interests.”).

 [120]. Bowler & Glazer, supra note 5, at 15. But see Cooter, supra note 78, at 144–45 (arguing that, although ballot initiatives cost more than legislative lobbying, California voters “apparently pursue the more costly alternative because they believe that ballot initiatives mostly create laws that the legislature would not enact”).

 [121]. Bowler & Glazer, supra note 5, at 7 (arguing that voters can figure out how to vote their preferences and scholarly disagreement results from the true preferences of voters striking scholars as unpleasant, shortsighted, narrow, or all three); Bowler & Glazer, supra note 5, at 36 (noting Swiss cantons with initiatives show increased levels of participation); Mueller, supra note 3, at 189 (citing evidence that voter turnout is uniformly higher in elections with initiative measures on the ballot); Mueller, supra note 3, at 190 (finding the historical record suggests that voter initiatives are “useful addition” to democratic institutions in most countries where they exist); Mueller, supra note 3, at 191 n.13.

 [122]. See Mueller, supra note 3, at 95; see also Wright, supra note 12 (“The voter should remember that though the initiative and referendum may work satisfactorily in small communities, or in cities where the population is compact, it does not necessarily follow that it will be a success when applied to a commonwealth in which the interests are as varied and the population as large and the needs of the people as multifarious as they are in California”).

 [123]. Allswang, supra note 18, at 1–3; Wood, supra note 7, at 58 (“[T]he republican state necessarily had to be small in territory and generally similar in interests.”).

 [124]. Mueller, supra note 3, at 102.

 [125]. Id. at 97–98 (“Even in ancient Greece membership to the assembly had to be restricted once the size of the polity grew beyond these limits.”).

 [126]. See Wood, supra note 7, at 364–68; see also Zasloff, supra note 9, at 1122–23 (discussing how federal government was intentionally designed to eliminate direct democracy).

 [127]. John C. Yoo, Crisis and Command 29 (2009).

 [128]. Mueller, supra note 3, at 97–98 (“In Switzerland and some New England towns, direct democracy is practiced in communities of 10,000 or even 20,000 or more . . . but the most successful direct democracies are likely to be smaller communities.” (citation omitted)). In the modern era, popular assemblies have been used on a smaller scale (“a few hundred citizens”) in Japan, Poland, Iceland, Turkey, the former West Germany, England, and Wales. Mueller, supra note 3, at 100 n.5.

 [129]. Mueller, supra note 3, at 102.

 [130]. Cal. Dep’t of Fin., New State Population Report: California Grew by 335,000 Residents in 2016 1 (2016),
/documents/E-1_2017PressRelease.pdf; Cal. Sec’y of State, Statement of the Vote 1–3 (2016),; see also The Council of State Gov’ts, The Book of the States, 346 tbl.6.8, 574 (2010) (providing figures for 2008 presidential election); Baldassare & Katz, supra note 5, at 33–36; Mueller, supra note 3, at 97 (“[E]ven the smallest nation-states today are too large to make collective decisions using procedures in which citizens actively debate and decide issues in open meetings.”).

 [131]. For example, the “majority” that enacted the initiative measure Proposition 8 (banning same sex marriage) was only 7 million votes. That figure is only 41% of the state’s registered voters (17 million), 30% of eligible voters (23 million), and only 19% of the total state population (37 million). Statement of the Vote, supra note 130; see also Baldassare & Katz, supra note 5, at 33–36; The Book of the States, supra note 130, at 346, 574 (figures for 2008 presidential election).

 [132]. Allswang, supra note 18, at 246.

 [133]. How to Qualify an Initiative, Cal. Sec’y State, (last visited Apr. 11, 2019).

 [134]. Cal. Const. art. II, § 8(b); Cal. Elec. Code § 9035 (West 2018); Baldassare & Katz, supra note 5, at 76; see also Beard & Shultz, supra note 22, at 36 (“Wherever the initiative is in force, a new trade, that of getting signatures, develops.”); Grant, supra note 74, at 138–39; LeDuc, supra note 5, at 150 (describing the professional initiative industry).

 [135]. Ctr. for Governmental Stud., supra note 72, at 11 tbl.4, 15 tbl.6 (2nd ed., 2008); John Wildermuth, Costs Soar to Qualify Initiatives for Ballot, S.F. Chronicle (Jan. 2, 2019),

 [136]. Cooter, supra note 78, at 144; see also David L. Callies, Nancy C. Neuffer & Carlito P. Caliboso, Ballot Box Zoning: Initiative, Referendum and the Law, 39 Wash. U. J. Urb. & Contemp. L. 53, 61–62 (1991).

In reality, however, the initiative process may not be a tool for the politically powerless, but a tool for the well-financed and politically connected. The cost, the time, and the energy required to place an initiative on the ballot are impractical for local grassroots movements. Petition circulation has become a multi-million dollar business in California, with costs per signature gathered for the 1990 campaign estimated at $1.21. Not surprisingly, the high cost of seeing an initiative to the ballot affects who sponsors initiatives. Well-financed individuals, lobbyists, and special interest groups proposed most of the initiatives for recent elections. Such a result is ironic, given the original goals of the initiative process.

Sutro, supra note 26, at 949–50 (footnotes omitted).

 [137]. Cronin, supra note 5, at 85, 109 (concluding that with a 25% success rate promoting “yes” campaigns compared to 75% success rate promoting “no” campaigns, Cronin concludes that “money counts the most” in opposing a ballot measure); see also Grant, supra note 74, at 140 (arguing while voters may simply vote “no” out of “cussedness,” high spending on the “no” side of an initiative heavily favors rejection).

 [138]. See Vermeule, supra note 17, at 170.

 [139]. Allswang, supra note 18, at 246.

 [140]. See id.; Matt Childers & Mike Binder, The Differential Effects of Initiatives and Referenda on Voter Turnout in the United States, 1890–2008, 19 Chapman L. Rev. 35, 41 (2016).

 [141]. See Childers & Binder, supra note 140.

 [142]. See Vermeule, supra note 17; Childers & Binder, supra note 140, at 35.

 [143]. See Cal. Const. art. II, § 10(a).

 [144]. See Jeremy B. White, Why Californians Have to Vote on 17 Ballot Measures, Sacramento Bee (Nov. 4, 2016, 04:14 PM),

 [145]. See S.B. 202, 2011 Leg., Reg. Sess. (Cal. 2011).

 [146]. White, supra note 145.

 [147]. See generally Ned Augenblick & Scott Nicholson, Ballot Position, Choice Fatigue, and Voter Behavior, 83 Rev. Econ. Stud. 460 (Apr. 2016); Simon Hedlin, Do Long Ballots Offer Too Much Democracy?, Atlantic (Nov. 3, 2015),

 [148]. Augenblick & Nicholson, supra note 1478, at 478. We discuss S.B. 202’s other effects in Section IV.B. See also Helios Herrera & Andrea Mattozzi, Quorum and Turnout in Referenda, 8 J. European Econ. Ass’n. 838, 853 (2010). In a sense, consciously nonvoting citizens are by default encouraging an alternative de facto representative system, where the nonvoters are represented by the voting population. See Cronin, supra note 5, at 77. As with elected representatives, presumably the nonvoters are at least somewhat satisfied with the results, and the nonvoters always retain the option of flocking to the polls to elect different representatives or to vote for different propositions.

 [149]. See Allswang, supra note 18, at 246.

 [150]. Of the approximately 39.5 million people in California, 24.8 million are eligible to vote, 19.4 million are registered to vote, and 14.6 million voted in the November 2016 general election—considering that most initiatives pass with approximately 50% of the votes cast, that means that a majority of 7.3 million (or 18% of the state population) sets policy for the state. Statement of the Vote, supra note 130; see also Baldassare & Katz, supra note 5, at 33–36; The Book of the States, supra note 130, at 346, 574 (figures for 2008 presidential election). But see Cooter, supra note 78, at 20 (arguing that economists find general voter participation rates to be surprisingly high: given the negligible probability that a single vote will change the outcome in a large election, the cost-benefit analysis for a self-interested citizen should result in the effort required to vote exceeding the expected benefit.).

 [151]. Proposition 42 in 2014 on public records and open meetings passed with approval from just 13.92% of registered voters. We calculated the final majority vote percentage from the official California Secretary of State registration and turnout figures. Cal. Sec’y of State, Historical Voter Registration and Participation in Statewide General Elections 1910–2016 (2016),

 [152]. Note that this problem has been a known bug since Progressive times. Beard & Shultz, supra note 22, at 37–38. And those authors proposed a similar solution to ours. See id. at 41.

 [153]. “As of May 2018, 19 million of California’s 25.1 million eligible adults were registered to vote. At 75.7% of eligible adults, this is a slight increase from the registration rate in 2014 (73.3%), the year of the last gubernatorial election.” California’s Likely Voters, Pub. Policy Inst. Cal., (last visited Apr. 11, 2019). Note that the number of registered voters (while remaining low) also remains consistent: “The share of eligible adults who are registered—currently 73%—has not varied much in recent years.” Voter Participation in California, Pub. Policy Inst. Cal., (analyzing turnout data 2000–2014),
/publication/voter-participation-in-california (last visited Apr. 11, 2019); see also McGhee et al., Pub. Policy Inst. Cal., California’s Future: Political Landscape (2018),

 [154]. See McGhee et al., supra note 153.

 [155]. As California Secretary of State voter data reveals, participation varies across election types. See supra Table 8.

 [156]. The data for these calculations (and those in Table 8) is derived from the California Secretary of State’s official participation and election summary data, see Cal. Sec’y of State, supra note 151.

 [157]. See supra Table 8 (California voter turnout by election type).

 [158]. See Abdurashid Solijonov, Voter Turnout Trends Around the World, Int’l Inst. Democracy & Electoral Assistance 8 (2016),; McGhee et al., supra note 153, at 2. As noted in the previous Section, we found no evidence that California’s low turnout rate is caused by the initiative’s existence. See supra Section III.D.

 [159]. Calculations derived from Cal. Sec’y of State, supra note 151.

 [160]. See McGhee et al., supra note 153, at 2. There are contrary findings. See, e.g., Allswang, supra note 18, at 145. But note that Allswang ultimately concurs with our point: “Not only is the number of people actually deciding these propositions quite small—it is also . . . hardly a representative cross-section. The wealthier, better-educated, older, and white vote in considerably larger numbers than the poor, ill-schooled, young, and minority group members.” Id. at 246.

 [161]. Cronin, supra note 5, at 76, tbl.4.4.

 [162]. Proposition 35 never took effect. See Doe v. Harris, 772 F.3d 563, 563 (9th Cir. 2014).

 [163]. Proposition 1A received 83.7% approval; Proposition 59 received 83.3% approval; Proposition 35 received 81.4% approval; and Proposition 58 received 73.5% approval. Calculations derived from Statewide Election Results, Cal. Sec’y State,
/statewide-election-results (last visited Apr. 11, 2019) (using 1990–2016 results).

 [164]. Proposition 1A in 2004 received the highest percentage approval from eligible voters: 42.6%. Calculations derived from Statewide Election Results, supra note 163.

 [165]. Cal. Const. art. II, § 10(a).

 [166]. Out of 156 approved ballot propositions between 1990 and 2016, 37 failed to pass this majority threshold. See infra Table 10.

 [167]. The California Government Code requires quorum, “which is a majority of the five members,” “before the council has legal authority to act.” Malathy Subramanian, Voting Requirements: Absences, Vacancies, Abstentions, and Disqualifications 1 (2006) (citing Cal. Gov. Code § 36810 (West 2018)).

 [168]. See id. at 1 (citing People v. Harrington, 63 Cal. 257, 260 (1883) (“We . . . regard the law as well settled that . . . the action of a quorum is the action of the board, and that a majority of the quorum present could do any act which a majority of the board if present might do.”).

 [169]. See, e.g., Henry M. Robert, Roberts’ Rules of Order Newly Revised 347 (11th ed. 2011) (“In a committee of the whole or its variations, the quorum is the same as in the assembly unless the rules of the assembly or the organization (that is, either its bylaws or its rules of order) specify otherwise.”).

 [170]. See Cal. Civ. Code § 12 (West 2018); Cal. Civ. Code Pro. § 15 (West 2018).

 [171]. See, e.g., Cal. Gov. Code § 36810 (West 2018).

 [172]. See Elliot Bulmer, Int’l Inst. for Democracy & Electoral Assistance, Direct Democracy 17­–18 (2014),

 [173]. Luís Aguiar-Conraria & Pedro C. Magalhães, Referendum Design, Quorum Rules and Turnout, 144 Pub. Choice 63, 64–65 (2010).

 [174]. Herrera & Mattozzi, supra note 148, at 858.

 [175]. Id.

 [176]. Id. at 839.

 [177]. See id. at 858.

 [178]. See Baldassare, supra note 108.

 [179]. For average abstention rates, see supra Table 9.

 [180]. Studies have shown that the longer the ballot, the more voter choices deviate from their expressed ideology. See Peter Selb, Supersized Votes: Ballot Length, Uncertainty, and Choice in Direct Legislation Elections, 135 Pub. Choice 319, 332 (2008).

 [181]. S.B. 202 limits ballot propositions to general elections.

 [182]. See Let the People Fail to Decide, Economist (May 19, 2016),
/leaders/2016/05/19/let-the-people-fail-to-decide (“These dangers can be mitigated. Requiring minimum turnouts can guard against the tyranny of the few. Italy’s 50% threshold is about right.”).

 [183]. See Baldassare, supra note 108, at 8–9.

 [184]. Williams, supra note 3, at 225.

 [185]. See Wood, supra note 7, at 21.

 [186]. Rossi v. Brown, 889 P.2d 557, 560–61 (Cal. 1995).

 [187]. See Williams, supra note 3, at 225.

 [188]. We noted elsewhere that there are difficult questions at the outer limits of this principle: “[A]lthough the provisions of the constitution are binding on future legislatures and electorates alike, the electorate cannot restrict its own future initiative power through the initiative process.” Carrillo, supra note 13, at 746; see also County of Los Angeles v. State, 729 P.2d 202, 209 n.9 (Cal. 1987) (“Whether a constitutional provision which requires a supermajority vote to enact substantive legislation, as opposed to funding the program, may be validly enacted as a Constitutional amendment rather than through revision of the Constitution is an open question.” (citing Amador Valley Joint Union High Sch. Dist. v. State Bd. of Equalization, 583 P.2d 1281, 1289 (Cal. 1978))).

 [189]. Prof’l Eng’rs in Cal. Gov’t v. Kempton, 155 P.3d 226, 239–40 (Cal. 2007) (citing Cal. Const. art. IV, § 1).

 [190]. Baldassare & Katz, supra note 5, at 13.

 [191]. See Carrillo, supra note 13, at 731.

 [192]. See Choper, supra note 100, at 64–65.

 [193]. See Cooter, supra note 78, at 146–47. Another response to the pro-majoritarian criticism is Professor Cooter’s argument that direct democracy factors issues, which does not necessarily harm minorities more than the spliced voting that would occur in the legislature. In factored voting, the minority on one dimension of choice is not necessarily the same group across all issues, with the result that any one person may win on some issues and lose on others. Thus, only some minorities will lose, and only sometimes; under those conditions, majorities will not exploit minorities more under direct than under indirect democracy. Cooter, supra note 78, at 146.

 [194]. Frank Clifford, Voters Repudiate 3 of Court’s Liberal Justices, L.A. Times (Nov. 5, 1986), (detailing voters’ rejection of California Supreme Court Chief Justice Rose Bird and Justices Joseph Grodin and Cruz Reynoso, who were on the November 1986 general election ballot for retention). Scholars debate how strictly courts should review electorate acts. In his seminal article on that subject, Professor Eule argued that courts should scrutinize plebiscites more aggressively than legislative acts. Julian Eule, Judicial Review of Direct Democracy, 99 Yale L.J. 1503 passim (1990). But he cautioned that not only are his arguments inapplicable to the states, states (like California) whose constitutions give the voters direct lawmaking power are the strongest case for greater judicial deference to electorate acts. Id. at 1547–48.

 [195]. In this Section, we briefly explain several procedural constraints and substantive rules that California courts have developed to define and limit the process of constitutional change and regulate the exercise of the electorate’s power. We note, but do not discuss, the various procedural issues that commonly arise, related to such things as signature gathering, title, and summary.

 [196]. Cal. Const. art. II, § 8(d).

  [197]. Senate of Cal. v. Jones, 988 P.2d 1089, 1098 (Cal. 1999); Sutro, supra note 26, at 961–62.

The primary purpose of the legislative single-subject rule is recognized as the prevention of log-rolling, the practice of several minorities combining their legislative proposals as different provisions of a single bill and thus consolidating their votes so that a majority is obtained. Additional purposes of the legislative single-subject rule are the preservation of an orderly legislative process and the prevention of deception of the legislature and the public. Single-subject legislation promotes clarity in the legislative process and ensures there will be little confusion due to multi-subject bills.

Sutro, supra note 26, at 961–62 (quotations and footnotes omitted); see also Steven W. Ray, The California Initiative Process: The Demise of the Single-Subject Rule, 14 Pac. L.J. 1095, 1096–98 (1983).

 [198]. Jones, 988 P.2d at 1098; Sutro, supra note 26, at 963–64.

 [199]. Jones, 988 P.2d at 1098–99 (quoting Brosnahan v. Brown, 651 P.2d 274, 284 (Cal. 1982)).

 [200]. Briggs v. Brown, 400 P.3d 29, 38 (Cal. 2017) (quoting Californians for an Open Primary v. McPherson, 134 P.3d 299, 318 (Cal. 2006)); see also Perry v. Jordan, 207 P.2d 47, 50 (Cal. 1949); Evans v. Super. Ct., 8 P.2d 467, 469 (Cal. 1932).

 [201]. Briggs, 400 P.3d at 38 (emphasis added) (internal quotation marks omitted) (citations omitted).

 [202]. Brown v. Super. Ct., 371 P.3d 223, 232 (Cal. 2016) (ellipsis omitted) (quoting Legislature v. Eu, 816 P.2d, 1309, 1321 (Cal. 1991)). The California Supreme Court interprets legislative and initiative acts with the same test. The cardinal rule of statutory interpretation in California is that the statute is to be construed so as to give effect to the intent of the lawmakers. Mercer v. Perez, 436 P.2d 315, 320 (Cal. 1968). In construing constitutional and statutory provisions, “whether enacted by the legislature or by initiative, the intent of the enacting body is the paramount consideration.” In re Lance W., 694 P.2d 744, 754 (Cal. 1985).

 [203]. Jones, 988 P.2d at 1098.

 [204]. Manduley v. Super. Ct., 41 P.3d 3, 28–29 (Cal. 2002) (quoting Jones, 988 P.2d at 1162).

 [205]. Grodin et al., supra note 26, at 70 (calling the single-subject rule “a toothless tiger”); see also Ray, supra note 197, at 1096 (“[T]he court should adopt a stricter interpretation of the single-subject rule where initiatives are concerned to prevent those proposals from ever being presented to the electorate.”). See generally Robert D. Cooter & Michael D. Gilbert, A Theory of Direct Democracy and the Single-Subject Rule, 110 Colum. L. Rev. 687 (2010) (proposing a democratic process theory of the single-subject rule).

 [206]. Manheim & Howard, supra note 2, at 1207 (internal quotation marks omitted).

 [207]. Two commentators argue that the distinctions between legislative and initiative acts require distinct single-subject rules. Ray, supra note 197, at 1101 (“The two processes here in question, the initiative and the legislative, are not the same. In fact, the vast differences between the two compel a change in the current application of the single-subject rule to initiatives.”); Sutro, supra note 26, at 966 (using canons to interpret initiatives wrongly assumes voter knowledge of existing law and an intent for uniformity and consistency, ignoring limited voter knowledge).

 [208]. Strauss v. Horton, 207 P.3d 48, 86 n.19 (Cal. 2009) (noting that when McFadden was decided, there was no California constitutional provision applying the single-subject rule to initiative measures).

 [209]. Cal. Const. art. IV, § 12(d).

 [210]. Cal. Elec. Code §§ 9001–02 (West 2018).

 [211]. Id. § 9004.

 [212]. See Strauss, 207 P.3d at 132.

 [213]. See id. at 61.

 [214]. Id. at 97 (quoting Raven v. Deukmejian, 801 P.2d 1077, 1085 (Cal. 1990).

 [215]. Legislature v. Eu, 816 P.2d 1309, 1340 (Cal. 1991) (alterations in original) (quoting Livermore v. Waite, 36 P. 424, 426 (Cal. 1894)).

 [216]. An initiative constitutional amendment may be placed on the ballot after collecting a number of elector signatures equal to 8% of the votes for all candidates for Governor in the last gubernatorial election. Cal. Const. art. XVIII, § 8(b). By contrast, only the state legislature is empowered to propose revisions. Id. § 1 (“The Legislature . . . may propose an amendment or revision of the Constitution . . . .”); id. § 2 (“The Legislature . . . may submit at a general election the question whether to call a convention to revise the Constitution.”); id. § 4 (“A proposed amendment or revision shall be submitted to the electors . . . .”).

 [217]. Cal. Const. art. XVIII.

 [218]. See Carrillo, supra note 13, at 738–40.

 [219]. Id. at 740.

 [220]. The authors have explored Strauss’ revision–amendment analysis in a related context. See David A. Carrillo & Stephen M. Duvernay, California Constitutional Law: The Guarantee Clause and California’s Republican Form of Government, 62 UCLA. L. Rev. Disc. 103, 120–22 (2014).

 [221]. Carrillo, supra note 13, at 738–40.

 [222]. See id. at 733 n.4 (identifying Howard Jarvis Taxpayers Ass’n v. Padilla, 363 P.3d 628 (Cal. 2016); Strauss v. Horton, 207 P.3d 48 (Cal. 2009); Legislature v. Eu, 816 P.2d 1309 (Cal. 1991); Raven v. Deukmejian, 801 P.2d 1077 (Cal. 1990), as four cases where an initiative measure has created a separation-of-powers issue).

 [223]. The authors advanced this proposal in a recent law review article. See Carrillo, supra note 13, at 751–64. For another perspective, see Manheim & Howard, supra note 2, at 1203–06 (arguing that the initiative does not invade the legislature’s core powers).

 [224]. Carrillo, supra note 13, at 751. For much the same reasons, the state’s separation-of-powers doctrine differs from its federal analogue. See Carrillo & Chou, supra note 85, at 665–73 (discussing the differences between the separation of powers doctrines embodied in the California and federal Constitutions); see also Marine Forests Soc’y v. Cal. Coastal Comm’n, 113 P.3d 1062, 1076–78 (Cal. 2005).

 [225]. Carrillo & Chou, supra note 85, at 678–79. Put another way, “the state constitution vests each branch with certain core powers that cannot be usurped by another branch.” Id. at 679.

 [226]. Carmel Valley Fire Prot. Dist. v. State, 20 P.3d 533, 538 (Cal. 2001) (quoting Kasler v. Lockyer, 2 P.3d 581, 594 (Cal. 2000)).

 [227]. Howard Jarvis Taxpayers Ass’n v. Padilla, 363 P.3d 628, 634 (Cal. 2016) (quoting Marine Forests Soc’y, 113 P.3d at 1087).

 [228]. Younger v. Super. Ct., 577 P.2d 1014, 1024 (Cal. 1978) (emphasis in original) (citing Parker v. Riley, 113 P.2d 873, 873 (Cal. 1941)).

 [229]. Carmel Valley, 20 P.3d at 538 (internal quotation marks omitted).

 [230]. Carrillo, supra note 13, at 738–40.

 [231]. Bowler & Glazer, supra note 5, at 119 (“[T]he primary effect of the initiative is power transfer from both branches of government to the median voter.”); Bowler & Glazer, supra note 5, at 116–17 (“The evidence indicates that direct democracy brings about material changes in the functioning of the executive branch”); Bowler & Glazer, supra note 5, at 118–19 (discussing how “some practical considerations suggest that the governor will usually benefit” from direct democracy by allowing the governor to take proposals directly to the voters).

 [232]. Id. at 118 (“[I]t is clear that the voter is never worse off when the initiative is available.”); id. at 119 (“The political actor that always wins (never loses) from having the initiative available is the median voter.”).

 [233]. See, e.g., Myers v. English, 9 Cal. 341, 349 (1858).

 [234]. Briggs v. Brown, 400 P.3d 29, 50–61 (Cal. 2017) (analyzing separation-of-powers issues created by the passage of Proposition 66, the Death Penalty Reform and Savings Act of 2016).

 [235]. Id.; see also Carrillo, supra note 13; Carrillo & Chou, supra note 85.

 [236]. Wood, supra note 7, at 20; see also Cooter, supra note 78, at 245 (stating that the purpose of individual rights is to provide the legal basis of autonomy).

 [237]. Zasloff, supra note 9, at 1125.

 [238]. Steiner, supra note 3, at 86. One California study showed that “[o]n . . . minority-targeted initiatives, Latinos consistently lose out,” and that “Latinos, indeed, have much to worry about when issues that target their rights are decided via direct democracy.” Zoltan Hajnal et al., Minorities and Direct Legislation: Evidence from California Ballot Proposition Elections, 64 J. Pol. 154, 171 (2002); see also Zoltan Hajnal & Hugh Louch, Pub. Pol’y Inst. of Cal., Are There Winners and Losers? Race, Ethnicity, and California’s Initiative Process (2001). A nationwide study concluded that initiatives to restrict civil rights pass more regularly than other types of initiatives. Barbara S. Gamble, Putting Civil Rights to a Popular Vote, 41 Am. J. Pol. Sci. 245 passim (1997); LeDuc, supra note 5, at 41; Steiner, supra note 3, at 86 (noting the “substantial body of academic literature offering cautions about California’s practice of ballot propositions” based on initiatives being used by powerful special interest groups to capture the powers of the state in self-interested ways, and to threaten the civil rights of vulnerable minorities or exploit and increase racial or ethnic tensions) (citing David S. Broder, Democracy Derailed: Initiative Campaigns and the Power of Money 43 (2000); Ian Budge, The New Challenge of Direct Democracy (1996); Richard J. Ellis, Democratic Delusions: The Initiative Process in America 77 (2002); John Haskell, Direct Democracy or Representative Government? (2001); Bruce A. Larson, Dangerous Democracy (Larry J. Sabato, Bruce A. Larson & Howard R. Ernst eds., 2001); Giovanni Sartori, The Theory of Democracy Revisited (1987); Peter Schrag, Paradise Lost: California’s Experience, America’s Future (1998)).

 [239]. Cronin, supra note 5, at 98 (“If we are to give occasional free rein to majority rule at the ballot box, we shall have to give additional consideration to protecting the rights of minorities.”); LeDuc, supra note 5, at 151 (using the initiative to target vulnerable minorities is a modern example of Madison’s tyranny of the majority).

 [240]. Bowler & Glazer, supra note 5, at 119, 139 (“The evidence is fairly strong that the initiative does in fact bring about policies favored by the majority.”); Bowler & Glazer, supra note 5, at 147 (“The initiative works as a form of veto point, forcing policy to the position of the median voter on each dimension, and preventing the construction of logrolling coalitions that can challenge the status quo.”).

 [241]. Bruce E. Cain, Constitutional Revision in California, in State Constitutions for the Twenty-First Century 69 (G. Alan Tarr & Robert F. Williams eds., State Univ. of N.Y. Press 2006) (“[T]he eighteenth-century concept of a constitution that balances the rights of the minority against those of the majority simply makes no sense at the state level. Measures that would protect or favor a minority against the majority’s will cannot make it through the constitutional approval process.”).

 [242]. Bowler & Glazer, supra note 5, at 139. Similar to the current slow rate of adoption of individual rights for same sex persons as a group, Switzerland denied suffrage to women until 1972; in non-initiative systems, the franchise could be extended in a legislative solution as part of a broader political compromise, while in the initiative system, the change required approval from a majority of male voters to reduce their political power by expanding the electorate. The result is similar to the low rate of adoption of legislative term limits in non-initiative states in America, as both situations are governed by the principle that interest groups rarely vote to reduce their power voluntarily.

 [243]. Cronin, supra note 5, at 90–91.

 [244]. See generally, e.g., Boumediene v. Bush, 553 U.S. 723 (2008); Hamdan v. Rumsfeld, 548 U.S. 557 (2006); Hamdan v. Rumsfeld, 542 U.S. 507 (2004); Rasul v. Bush, 542 U.S. 466 (2004); Korematsu v. U.S., 323 U.S. 214 (1944).

 [245]. See generally, e.g., Korematsu, 323 U.S. at 214; Plessy v. Ferguson, 163 U.S. 537 (1896); Dred Scott v. Sandford, 60 U.S. 393 (1856).

 [246]. Cronin, supra note 5, at 91–92 (collecting examples).

 [247]. LeDuc, supra note 5, at 41, 150–51; Callies et al., supra note 136, at 94–97; Julia Anne Guizan, Is the California Civil Rights Initiative a Wolf in Sheep’s Clothing? Distinguishing Constitutional Amendment from Revision in California’s Initiative Process, 31 Loy. L.A. L. Rev. 261 passim (1997).

 [248]. See, e.g., Seventy-Third Day, in 1 Debates and Proceedings of the Constitutional Convention of the State of California 627 et seq. (Sacramento, J. D. Young, Supt. State Prtg. 1881) and Seventy-Seventh Day, in 2 Debates and Proceedings of the Constitutional Convention of the State of California 700 et seq. (Sacramento, J. D. Young, Supt. State Prtg. 1881) (anti-coolie provision); Seventy-Seventh Day, supra, at 801 (English-only provision); David A. Kaiser & David A. Carrillo, California Constitutional Law: Reanimating Criminal Procedural Rights After The “Other” Proposition 8, 56 Santa Clara L. Rev. 33 (2016); Proposition 1, Alien Land Law (Cal. 1920), (anti-Japanese initiative amending state’s alien land law); Proposition 14, Right to Decline to Sell or Rent Residential Real Estate (Cal. 1964), (initiative amendment overturning statute prohibiting racial discrimination in housing), invalidated by Reitman v. Mulkey, 387 U.S. 369, 375–76 (1967); Proposition 63, English Is the Official Language Amendment (Cal. 1986), (enacted at Cal. Const. art. IV, § VI) (initiative amendment making English official state language); Proposition 187, Illegal Aliens Ineligibile for Public Benefits (Cal. 1994), (initiative amendment denying public benefits to illegal immigrants); Proposition 209, California Affirmative Action (1996), (enacted at Cal. Const. art. 1, § XXXI) (initiative amendment prohibiting affirmative action); Proposition 227, “English Language in Public Schools” Initiative (Cal. 1998), initiative statute enforcing English-only education); Proposition 8, “Eliminates Right of Same-Sex Couples to Marry” Initiative (Cal. 2008), (2008 initiative amendment restricting marriage to opposite-sex couples, invalidated by Hollingsworth v. Perry, 570 U.S. 693, 693 (2013). But see Proposition 6, the Briggs Initiative (Cal. 1978),
/ca_ballot_props/838 (rejected initiative limiting gay teachers’ rights); Proposition 64, Mandatory Reporting of AIDS (Cal. 1986), (rejected initiative permitting quarantine of AIDS patients). Of course, discriminatory state governmental actions are not limited to the electorate. See Baldassare & Katz, supra note 5, at 22; Lustig, supra note 99, at 9 (noting that some 1849 delegates wanted California to be a “white man’s republic,” that the state denied Native Americans, blacks, and Chinese the right to vote, testify, or serve on a jury, and that California did not ratify the Fifteenth Amendment until 1962).

 [249]. Cal. Const. art. I, § 7(a) (providing that “[a] person may not be deprived of life, liberty, or property without due process of law or denied equal protection of the laws.”).

 [250]. In re Marriage Cases, 183 P.3d 384 (Cal. 2008).

 [251]. Proposition 8, “Eliminates Right of Same-Sex Couples to Marry” Initiative (Cal. 2008),, invalidated by Hollingsworth v. Perry, 570 U.S. 693, 693 (2013) (invalidating an initiative measure approved by a majority of voters at the November 4, 2008 election that added a new section—section 7.5—to California constitution article I: “Only marriage between a man and a woman is valid or recognized in California.”).

 [252]. Obergefell v. Hodges, 135 S. Ct. 2584, 2588 (2015) (holding that the right of same-sex couples to marry is protected by the Due Process and Equal Protection Clauses of the U.S. Constitution’s Fourteenth Amendment); Hollingsworth v. Perry, 570 U.S. 693 (2013); Perry v. Brown, 671 F.3d 1052, 1095 (9th Cir. 2012) (holding that Proposition 8 violated Equal Protection Clause); Strauss v. Horton, 207 P.3d 48 (Cal. 2009) (rejecting argument that Proposition 8 violated constitutional guarantee of equal protection).

 [253]. Cal. Const. art. II, § 10(d); Cal. Elec. Code § 9002 (West 2018) (“The duty of the Attorney General to prepare title and summary for a proposed initiative measure is a ministerial one, and, . . . mandate will lie to compel him to act when the proposal is in proper form and complies with statutory and constitutional procedural requirements.”); Schmitz v. Younger, 577 P.2d 652, 653 (Cal. 1978) (“[W]ithout prior judicial authorization [the Attorney General] may not delay or impede the initiative process while claims of the measure’s invalidity are determined.”).

 [254]. Brosnahan v. Eu, 641 P.2d 200, 201 (Cal. 1982).

 [255]. Am. Fed’n of Labor v. Eu, 686 P.2d 609, 615 (Cal. 1984).

 [256]. Id. at 614 (quoting Legislature v. Deukmejian, 669 P.2d 17, 21 (Cal. 1983)). There is some tension on whether the electorate’s “power” in this regard refers only to their procedural power, not to their ability to enact laws that substantively violate the constitution. As Justice Mosk explained in his concurring and dissenting opinion in Brosnahan:

The principle is firmly established that unless it is clear that a proposed initiative is unconstitutional, the courts should not interfere with the right of the people to vote on the measure. In the service of this precept, courts have frequently declined to strike an initiative from the ballot despite a claim that its adoption would be a futile act because the measure offends the Constitution. . . . But this rule applies only to the contention that an initiative is unconstitutional because of its substance. If it is determined that the electorate does not have the power to adopt the proposal in the first instance or that it fails to comply with the procedures required by law to qualify for the ballot, the measure must be excluded from the ballot.

Brosnahan, 641 P.2d at 202–03 (Mosk, J., concurring and dissenting). The Court held a similar line in Legislature v. Deukmejian, where it allowed a pre-election challenge that “[went] to the power of the electorate to adopt the proposal in the first instance. This challenge does not require even a cursory examination of the substance of the initiative itself. The question raised is, in a sense, jurisdictional.” Legislature v. Deukmejian, 669 P.2d 17, 21 (Cal. 1983). There is little reason to doubt that the Court would reach the same conclusion, however, when considering an initiative that clearly violates enumerated constitutional rights. The underlying rationale for permitting pre-election review of an invalid initiative applies with equal force in such circumstances. “The presence of an invalid measure on the ballot steals attention, time and money from the numerous valid propositions on the same ballot. It will confuse some voters and frustrate others, and an ultimate decision that the measure is invalid, coming after the voters have voted in favor of the measure, tends to denigrate the legitimate use of the initiative procedure.” Am. Fed’n of Labor, 686 P.2d at 615.

 [257]. See Younger, P.2d at 653.

 [258]. Am. Fed’n of Labor, 686 P.2d 611, 629; see also Planning & Conservation League v. Padilla, No. S249859, 2018 Cal. LEXIS 6817, *1–2 (2018).

 [259]. See, e.g., Howard Jarvis Taxpayers Ass’n. v. Padilla, 363 P.3d 628, 631 (Cal. 2016).

In response to a petition for writ of mandate urging the unconstitutionality of the Legislature’s action, we issued an order to show cause and directed the Secretary of State to refrain from taking further action in connection with placement of Proposition 49 on the ballot. Our action did not rest on a final determination of Proposition 49’s lawfulness. Instead, we concluded “the proposition’s validity is uncertain” and the balance of hardships from permitting an invalid measure to remain on the ballot, as against delaying a proposition to a future election, weighed in favor of immediate relief.


 [260]. Memorandum of Points and Authorities in Support of Request for Entry of Default Judgement at Ex. A, Harris v. McLaughlin, No. 34-2015-00176996 (Super. Ct. Feb. 24, 2015). Further background on this issue can be found in an unpublished appeal from a related lawsuit filed by the proponent. McLaughlin v. Becerra, No. B280529, 2018 Cal. App. Unpub. LEXIS 739 (2018) (appeal from Los Angeles City Super. Ct. Case No. BC622687).

 [261]. Memorandum of Points and Authorities in Support of Request for Entry of Default Judgement ¶¶ 13–15, Harris v. McLaughlin, No. 34-2015-00176996 (Super. Ct. June 16, 2015).

 [262]. Default Judgment by Court in Favor of Plaintiff, Harris v. McLaughlin, No. 34-2015-00176996 (Super. Ct. June 22, 2015).

 [263]. Planning & Conservation League, 2018 Cal. LEXIS 6817.

 [264]. Planning & Conservation League v. Padilla, No. S249859, 2018 Cal. LEXIS 5200, at *1–2. The court explained its rationale:

Although our past decisions establish that it is usually more appropriate to review challenges to ballot propositions or initiative measures after an election, we have also made clear that in some instances, when a substantial question has been raised regarding the proposition’s validity and the “hardships from permitting an invalid measure to remain on the ballot” outweigh the harm potentially posed by “delaying a proposition to a future election,” it may be appropriate to review a proposed measure before it is placed on the ballot.

Id. (citations omitted).

 [265]. This is not a universally-held view. See, e.g., Lustig, supra note 99, at 65–69, 72.

The initiative theoretically counteracts the federalist model and is a majoritarian tool. . . . One can certainly make the argument that supermajority requirements and the stripping of legislative discretion over spending and taxing are good things in the abstract, but it is difficult to see how those have been good for California in practice. In fact, initiative governance has caused legislative failure on many issues facing the state.


 [266]. Fuller, supra note 83. This is the relative size of the California economy according to the California Department of Finance and the Bureau of Economic Analysis at the U.S. Department of Commerce as of May 4, 2018. See Gross State Product, Cal. Dep’t. Finance, (last visited Apr. 13, 2019); GDP and Personal Data, Bureau Econ. Analysis,
=1200&7036=-1&7002=1&7090=70&7007=-1&7093=levels (last visited Apr. 13, 2019).

 [267]. Sheila James Kuehl, Either Way You Get Sausages: One Legislator’s View of the Initiative Process, 31 Loy. L.A. L. Rev. 1327, 1329–30 (1998).

Californians love their initiatives. They do not like reading the long ones. They do not like it when the courts strike them down for their constitutional defects. They do not like finding out later that they were wrong or misled about the contents. But generally, the people of California jealously guard their ability to make and shape the law independent of the legislature. For the most part, the people feel excluded from the long and arduous process of legislation. They read about the new laws on January 1 of each year and shake their heads or wonder at the omissions. The initiative process provides the people with a way to remedy the paralysis and inaction they perceive in the legislature.

Id. (footnotes omitted).

 [268]. Allswang, supra note 18, at 245; Baldassare & Katz, supra note 5, at 23, 31, 217, tbl.1.2; Cronin, supra note 5, at 78–80 and tbls.4.5 & 4.6, 199, 234 tbl.9.3; Ctr. for Governmental Stud., supra note 72, at 17–27; Grant, supra note 74, at 139; Gordon, supra note 5, at 1.

 [269]. Cronin, supra note 5, at 232 (“Initiatives and referenda are here to stay.”); Cain, supra note 242, at 69 (“[T]o change the initiative process, one would have to ask the voters whom the process has served well to give up their control over policy outcomes. This is unlikely to happen.”); Manheim & Howard, supra note 2, at 1237 (“[O]ne wonders at this point whether Californians would ever accept a government as legitimate if it did not provide for some form of direct democracy.”).


Like this article?

Share on Facebook
Share on Twitter
Share on Linkdin
Share on Pinterest