The initiative and referendum movement: how American states let voters write law

The initiative and referendum movement was a late nineteenth and early twentieth century American reform campaign that gave voters the power to propose laws by petition (the initiative) and to approve or reject laws passed by legislatures (the referendum). South Dakota adopted the initiative first, in 1898, and according to the National Conference of State Legislatures 24 states, the District of Columbia, and the U.S. Virgin Islands now have a citizen initiative, with Mississippi’s process found void in 2021.⁠³ Reformers modeled the tools on Swiss practice and promoted them as a check on legislatures they considered corrupt;⁠¹ the Supreme Court refused to strike them down in 1912,⁠² and later disputes turned on paid petitioning and interest group money.⁠^4,6

In plain words

In the late 1800s, many Americans believed state legislatures were controlled by party bosses and big companies. Reformers had an idea borrowed from Switzerland: let voters write laws themselves.

The idea had two parts. With an initiative, citizens collect enough signatures to put a proposed law on the ballot. With a referendum, citizens can force a public vote on a law the legislature already passed.

South Dakota adopted the idea first, in 1898. Oregon followed in 1902, and many western states joined in the next twenty years. A phone company argued in court that this kind of lawmaking was unconstitutional. The Supreme Court said that question was for Congress, not the courts, and the system stayed.

Over time a new problem appeared. Gathering hundreds of thousands of signatures costs money. Companies and interest groups began paying people to collect them. Some states tried to ban that, and the Supreme Court said the bans violated free speech.

Researchers still disagree about whether the initiative serves ordinary voters or the groups that can afford to use it. The evidence on that question is mixed, and this chapter lays it out.

Defined terms

• Initiative. A process by which citizens propose a statute or constitutional amendment by petition. See the glossary entry on initiative.
• Direct initiative. A qualifying measure goes straight to the ballot.⁠³
• Indirect initiative. A proposal goes first to the legislature; if the legislature does not adopt it or a substantially similar measure, it goes to voters, sometimes with a competing legislative measure.⁠³
• Popular referendum. A petition process to force a public vote on a law already passed by the legislature. NCSL counts 23 states plus D.C. and the U.S. Virgin Islands.⁠³ See referendum.
• Legislative referral. A measure the legislature itself places on the ballot; all 50 states have some form.⁠³
• Recall. A petition and vote to remove an official before the end of a term. See recall.
• Guarantee Clause. Article IV, Section 4 of the U.S. Constitution, guaranteeing each state “a republican form of government,” the basis of the 1912 challenge.⁠²

Evidence

Primary source: the reformers’ case

James W. Sullivan’s Direct Legislation by the Citizenship through the Initiative and Referendum, copyrighted in 1892 and issued in New York in 1893, was an early American argument for the tools. Sullivan wrote that under representative government “the citizens empower legislators and executive officers to make the law,” producing “rule by a succession of quasi-oligarchies,” and he argued that direct legislation would let majorities dismantle the “ring,” “boss,” and “heeler” system.⁠¹ His model was Switzerland. He reported that fourteen of the twenty-two cantons had a cantonal initiative and described the obligatory referendum in Zurich as making the citizenry “a deliberative body in perpetual session”.⁠¹ Modern Swiss practice is described in participation between elections and in the Switzerland case.

History: the spread, 1898 to 1918

NCSL’s table of adoption years shows how concentrated the movement was in time and region:⁠³

• First wave: South Dakota (1898), Utah (1900), Oregon (1902), Montana (1904 initiative, 1906 referendum), Nevada (1904 referendum, 1905 initiative), Oklahoma (1907), and Maine, Missouri, and Michigan (1908, Michigan for constitutional initiatives).
• Peak years: Arkansas (1910), Arizona and California (1911), Colorado, Idaho, Nebraska, Ohio, and Washington (1912), North Dakota and Mississippi (1914), Maryland (referendum only, 1915), and Massachusetts (1918 initiative).
• Later additions: Alaska (1956), Wyoming (1968), Illinois (1970, constitutional only), and Florida (1972, constitutional only).

Apart from Maine and Massachusetts, the Northeast is absent from the list, and few Southern states appear.⁠³ The movement was largely a western and midwestern phenomenon. California’s experience is covered in depth in the California initiatives case.

Law: the constitutional challenge

Opponents argued that lawmaking by voters violated the Guarantee Clause, because a “republican” government is one that legislates through representatives. The test case came from Oregon, which had adopted the initiative and referendum through a 1902 amendment to its constitution. In 1906 Oregon voters used the initiative to impose a 2 percent annual license tax on the gross revenue of telephone and telegraph companies.⁠²

In Pacific States Telephone and Telegraph Co. v. Oregon, decided February 19, 1912, the Supreme Court dismissed the case for want of jurisdiction. The Court observed that “the assault which the contention here advanced makes is not on the tax as a tax, but on the state as a state,” and held that whether a state government is republican in form is a question “solely committed by the Constitution to the judgment of Congress”.⁠² The ruling did not declare direct legislation constitutional on the merits; it declined to decide. In practice, that left the initiative in place.

Law: paying people to gather signatures

As qualifying thresholds grew with population, signature gathering became an industry. Colorado banned paying circulators. In Meyer v. Grant, decided unanimously on June 6, 1988, the Court struck down the ban, describing petition circulation as “core political speech”.⁠⁴

A decade later, in Buckley v. American Constitutional Law Foundation, decided January 12, 1999, the Court struck down Colorado’s requirements that circulators be registered voters, wear name badges, and report each paid circulator’s name, address, and pay. The registration rule, the Court said, “cuts down the number of message carriers in the ballot-access arena without impelling cause,” and the added benefit of individual payment disclosure was “hardly apparent.” Disclosure of proponents’ names and total spending was left in place.⁠⁵ Justice O’Connor, joined by Justice Breyer, and Chief Justice Rehnquist would have upheld the registration and disclosure rules.⁠⁵

These cases mean that states can regulate the initiative process but cannot simply keep money out of signature gathering.

Law: process rules still decide outcomes

Mississippi shows that technical drafting can switch the process off. NCSL counts Mississippi among initiative states but notes that its process was found void in 2021, leaving 23 states with a working citizen initiative.⁠³ The episode illustrates how a procedural requirement, written for one set of conditions, can disable a right when conditions change.

Research results: who benefits

• Gerber. Elisabeth Gerber’s The Populist Paradox (Princeton University Press) studied campaign finance records from 168 ballot campaigns in eight states and surveys of interest groups. She found that well-funded economic interests rarely succeed in passing new laws against voters’ existing preferences, but they do succeed in blocking others’ initiatives and pressuring politicians, while citizen groups with broad support and strong organization are effective at passing new laws.⁠⁶
• Matsusaka. John Matsusaka’s For the Many or the Few (2004) treats the initiative mainly as a tool of fiscal policy. In John Gastil’s review in Public Opinion Quarterly, Matsusaka concludes that the evidence supports initiatives making policy “responsive to public opinion”.⁠⁷ Daniel Smith’s review in Perspectives on Politics summarizes the opposing view in the debate, that critics see the process as “out of control” and dominated by special interests.⁠⁷

Competing interpretations

Direct legislation as a corrective. Sullivan’s case was that representatives drift from voters and are captured by organized interests.⁠¹ Matsusaka’s findings, as reviewers summarize them, support the claim that the initiative brings policy closer to majority opinion.⁠⁷

Direct legislation as a new arena for organized money. Critics argue that the initiative did not remove interest groups but gave them a new venue. Gerber’s evidence supports a narrower version of this concern: money is better at blocking than at passing.⁠⁶ The Meyer and Buckley decisions mean regulation cannot easily close that venue.⁠^4,5

Representation versus plebiscite. The Guarantee Clause challenge raised a principle that the Court left to Congress: whether lawmaking by voters is consistent with representative government.⁠² Defenders of representation argue that legislatures can bargain, amend, and weigh tradeoffs across many issues, while a ballot measure is a single yes or no. Defenders of the initiative reply that legislatures may refuse to act at all. Problems such as lock-in and unclear meaning are documented in referendum failures.

Rights and majorities. A process that lets majorities write law directly also lets them limit minorities’ rights. That concern is treated in tyranny of the majority and does a majority decide rights.

Interpretation, this site’s reading. The movement’s founders correctly saw that whoever controls the agenda controls outcomes. They moved part of agenda control from legislators to whoever could gather signatures. That changed who holds the agenda; it did not remove the question.

Application to this framework

These are design implications, not claims about the framework’s performance.

1. Agenda power is the real power. The initiative moved the right to propose; it did not abolish gatekeeping. In My Vote Forward, who adds a proposal or an item to the priority list is a governance question; see who chooses the ten and agenda capture.
2. State legal effect up front. Initiatives bind because state constitutions say so.⁠^2,3 A priority signal from this framework binds no one, and that should be said on every result.
3. Money follows thresholds. Signature thresholds created a paid circulation industry.⁠⁴ Any future threshold in a participation tool, such as a number of supporters needed to add an item, will attract organized effort; see can wealthy groups dominate.
4. Disclose who is behind an item. Buckley left aggregate disclosure of proponents and spending in place.⁠⁵ A comparable practice for a participation tool is to show who proposed an item and whether it came from an organized campaign.
5. Procedural rules need review. The Mississippi example shows how a technical rule can disable a process.⁠³ Framework rules should be dated, reviewed, and revisable through a published process; see governance.

Hypothetical example. A city lets any resident add an item to its public priority list once 500 others endorse it. Within a month, a developer-funded group adds three items using paid outreach. The rule was neutral, but the threshold rewarded organization. Labeling items by origin and reporting how support was gathered would let other residents weigh them knowingly.

Unresolved questions

• Does the initiative make policy more responsive to the median voter, as Matsusaka’s work is read to show,⁠⁷ or more responsive to organized interests in the specific ways Gerber found⁠⁶? Both can be partly true in different states and decades.
• How should states regulate paid signature gathering after Meyer and Buckley without suppressing speech⁠^4,5?
• Should the Guarantee Clause question ever be revisited, and by whom, given that the Court left it to Congress⁠²?
• Would indirect initiatives, which give legislatures a chance to act first, reduce lock-in and drafting failures? The comparative evidence is limited.