Swiss direct democracy by the numbers: initiatives, referendums, money and minorities

Switzerland holds more national popular votes than any other country, and its experience is the main evidence base for arguments both for and against direct democracy. This study reviews what 178 years of federal votes show. Popular initiatives — constitutional amendments proposed by citizens — almost always fail: of 244 put to the vote by September 2026, 26 were accepted, a success rate of about 11%. Yet the instrument shapes policy indirectly, through counter-proposals, negotiated withdrawals and the threat of a vote. The optional referendum, by which 50,000 citizens can force a vote on a law, is used against only a small share of legislation but defeats a substantial minority of the laws it targets. Econometric studies of cantons and municipalities associate stronger fiscal referendums with lower spending and debt. The record on minorities is more troubling: the 2009 minaret ban, the 2010 expulsion initiative, the 2014 vote against "mass immigration" and the 2021 ban on face coverings all passed, and research on municipal naturalisation votes found clear discrimination by national origin, which fell when elected councils took over the decisions. Comparison with US states suggests that institutional detail — who can trigger votes, how courts review results, how much campaign money matters — explains as much as the principle of direct democracy itself.

Key findings

• Initiatives rarely pass. Of 244 federal popular initiatives voted on between 1893 and 27 September 2026, 26 were accepted (about 11%). The average Yes share was 35%. Fourteen of the 26 successes came in the 21st century.
• Referendums often do. Of 244 mandatory referendums and direct counter-proposals, 178 were accepted (about 73%). Of 218 optional referendums, the challenged law was upheld in 127 (58%) — so voters struck down about two in five of the laws that opponents chose to challenge.
• The double majority bites. Nine proposals since 1955 won a popular majority but failed on the cantons, including the 2020 Responsible Business initiative (50.73% Yes, 8.5 of 23 cantonal votes) and the 1994 proposal to ease naturalisation for young foreigners (52.84% Yes, 10 of 23).
• Turnout is moderate and varies with the issue. Mean turnout across voting days, computed from Swissvotes data, was around 40–45% from the 1970s to the 2010s and 48.4% in 2024; salient votes push it far higher.
• Fiscal referendums are associated with lower spending. Feld and Matsusaka estimated that mandatory budget referendums reduced cantonal spending by 19% for the median canton (1980–1998); Feld and Kirchgässner found dampening effects on spending, revenue and, locally, on debt.
• Minority rights are the strongest objection. Several initiatives targeting Muslims or foreigners passed. Where naturalisation was decided by popular vote, applicants from (former) Yugoslavia and Turkey faced about 40% more "no" votes than similar applicants from northern and western Europe; naturalisation rates rose by about 60% when elected councils took over.
• Implementation can soften results. Parliament implemented the 2014 immigration initiative without the quotas its text required, relying instead on a job-vacancy reporting duty for occupations with high unemployment.

Research question and scope

The question is: what are the measurable outcomes of Swiss federal direct democracy — participation, the success of initiatives and referendums, effects on policy and public finance, and effects on minorities — and how do they compare with direct democracy in US states? The study focuses on federal votes, drawing on cantonal and municipal evidence where the research literature does (fiscal effects and naturalisation). It does not cover the Landsgemeinde, the open-air assembly still used in Glarus and Appenzell Innerrhoden, which the atlas treats separately.

Background

The instruments

Swiss direct democracy grew in stages. The federal constitution of 1848 required a mandatory referendum on constitutional amendments; today it also applies to joining certain international organisations. The optional referendum came in 1874: if 50,000 eligible voters sign within 100 days, a law passed by parliament goes to a popular vote. The popular initiative for partial revision of the constitution arrived in 1891: 100,000 signatures collected within 18 months force a vote on a citizen-drafted amendment. Initiatives and mandatory referendums need a double majority — of voters nationally and of the cantons (with six "half-cantons" counting half) — while optional referendums need only a popular majority. Parliament may set a direct counter-proposal against an initiative, in which case voters choose between them and answer a tie-break question, or pass an indirect counter-proposal at the level of law that can lead sponsors to withdraw.

The Federal Chancellery checks titles and counts signatures and declares an initiative valid if 100,000 valid signatures have been collected within 18 months. Federal votes are held up to four times a year on fixed Sundays, usually bundling several questions.

Data

This study uses the Swissvotes dataset maintained by Année Politique Suisse at the University of Bern, which records every federal vote since 1848 with results, cantonal votes, turnout and official positions. Counts were computed by the editors for votes up to 27 September 2026; four questions due on 29 November 2026 are excluded. The counts match the Federal Statistical Office's published totals for June 2026 (242 initiatives, 26 accepted) once the two initiatives voted in September 2026 are added.

Method and sources

Quantitative findings on vote outcomes were computed from Swissvotes and cross-checked against official summaries. Effects on policy, finance and minorities come from peer-reviewed econometric studies; we report their designs and caveats. The comparison with US states relies on the comparative literature rather than on new data, because US state ballot-measure data are not compiled to the same standard. Three limitations apply. First, vote outcomes are not policy outcomes: an initiative can fail and still change policy. Second, much of the fiscal evidence is cantonal and dates from the 1980s–1990s. Third, causal claims about direct democracy face the problem that cantons with more direct democracy may differ in culture and history.

Finding 1: who votes and how often

Turnout is the first objection critics raise. Mean turnout per voting day, computed from Swissvotes, has been:

The fall in the 1970s partly reflects the enlargement of the electorate when women won the federal vote: men rejected women's suffrage in 1959 (33.08% Yes) and approved it on 7 February 1971 (65.73% Yes). Average figures conceal large swings. The 2020 Responsible Business initiative drew 47.04%; the 2021 nursing initiative 65.31%; the March 2026 votes on cash and public-broadcasting fees about 55–56%; and the September 2026 neutrality initiative 47.06%. Because votes are frequent, a large share of citizens vote selectively: participation over several votes is much higher than turnout in any one, though that cumulative figure is not in the Swissvotes data and is not estimated here.

Finding 2: initiatives — a tool that rarely wins but often matters

Success rates over time

The first initiative ever accepted, in 1893, banned slaughter without prior stunning — a measure widely understood at the time as aimed at Jewish ritual slaughter. Other early successes introduced proportional representation for the National Council (1918) and banned absinthe (1908). No initiative passed between 1949 and 1982. The rise in use since the 1970s — more than 40 per decade — has been accompanied by a modest rise in success: 14 of the 26 accepted initiatives were voted in the 21st century, including the 2013 "rip-off" initiative on executive pay (67.96% Yes, all cantons), the 2021 nursing initiative (60.98%) and the March 2024 initiative for a 13th monthly old-age pension payment (58.25%).

Why low success is not low impact

Initiatives work through several channels besides winning. Parliament can offer a counter-proposal: of 43 direct counter-proposals put to the vote, 27 were accepted, often against the original initiative. Indirect counter-proposals at the level of law lead many committees to withdraw; the 2021 federal rules on political-finance transparency, for example, were adopted as an indirect counter-proposal to a transparency initiative. Even a defeated initiative can move the agenda, and parties use initiatives to mobilise supporters before elections. These channels are hard to quantify, which is why the 11% success rate understates the instrument's influence.

The double majority and the cantons

Because small, mostly rural cantons carry the same weight as large urban ones, the cantonal majority acts as a conservative brake. Proposals that won a popular majority but failed on the cantons include:

The reverse also occurs: the 2002 initiative "against abuse of asylum law" won a majority of cantons but lost the popular vote by 49.91% to 50.09%. The cantonal requirement thus matters most in close votes, where it systematically favours the preferences of smaller cantons.

Finding 3: referendums — vetoes on legislation

The optional referendum is the instrument parliament anticipates most. Only around 6% of legislative proposals were challenged by referendum during the 20th century, according to Swissinfo's summary, but when a referendum is called the outcome is uncertain: of 218 optional referendums voted to September 2026, 127 upheld the law and 91 rejected it. Since 2000 the law has been upheld in 59 of 81 cases. Recent defeats include the motorway expansion and two tenancy-law changes in November 2024 and the occupational-pension reform in September 2024; the e-ID Act of 2025 survived with 50.39%. Political scientists have long argued that this veto threat pushes parliament towards broad, inclusive coalitions — the "consensus" character of Swiss government — because any large group able to collect 50,000 signatures can threaten a law.

Mandatory referendums are approved more often: of 244 mandatory referendums and direct counter-proposals, 178 were accepted. The sponsors here are the government and parliament, which normally propose only amendments with broad support.

Finding 4: effects on policy and public finance

Fiscal referendums and spending

The best-known Swiss evidence concerns public finance. Feld and Kirchgässner (2001), using data for the 26 cantons from 1986 to 1997 and for 134 municipalities in 1990, found that mandatory referendums on fiscal issues have a dampening effect on spending and revenue, and at the local level also on public debt. Feld and Matsusaka (2003), with a cantonal panel for 1980–1998, estimated that mandatory budget referendums on new spending reduced government spending by 19% for the median canton after controlling for demographics and other determinants. Kirchgässner, Feld and Savioz's 1999 book drew these findings together under the title Die direkte Demokratie: modern, erfolgreich, entwicklungs- und exportfähig ("modern, successful, capable of development and export").

How to read the fiscal results

These studies are often cited as evidence that direct democracy "works". Three caveats temper that reading. First, lower spending is good or bad depending on one's view of the right size of government; the studies show that fiscal referendums bring outcomes closer to what voters approve, not that the resulting policies are better. Second, the evidence is mostly cross-sectional or panel variation between cantons, which differ in many ways; the authors use controls, but historical and cultural differences cannot be fully removed. Third, the findings concern fiscal referendums, a specific institution, not the initiative or direct democracy in general. Matsusaka's review of the US and Swiss literature summarises the "emerging view" that direct democracy "often seems to improve the performance of government", while noting that many questions remain unanswered.

Finding 5: minority rights

Initiatives targeting minorities

Several accepted initiatives have targeted religious or national minorities:

The minaret initiative was launched by the "Egerkinger Komitee" of figures from the Swiss People's Party (SVP) and the Federal Democratic Union, and collected 113,540 signatures. The Federal Council and parliament opposed it; the National Council voted 132 to 51 against. It passed, adding a ban on building minarets to the constitution; four minarets existed at the time and were not affected. Two challenges at the European Court of Human Rights failed at the admissibility stage: on 28 June 2011 the Court declared Ouardiri v. Switzerland (no. 65840/09) and Ligue des musulmans de Suisse and Others v. Switzerland (no. 66274/09) inadmissible because the applicants did not claim that the ban had been applied to them or had any practical effect on them, and so were not victims under Article 34 of the Convention. The 2014 initiative required quotas and caps on immigration and the renegotiation of conflicting treaties within three years; in its final vote of 16 December 2016 parliament instead adopted a duty for employers in occupations with high unemployment to report vacancies to employment offices — applied from 1 July 2018 to occupations with unemployment above 8%, and from January 2020 above 5% — without quotas.

At the same time, many proposals hostile to minorities have failed: the 1994 naturalisation reform for young foreigners lost only on the cantons, the 2002 asylum initiative lost narrowly, and the June 2026 "sustainability" initiative to limit the resident population was rejected with 45.21% Yes. Frey and Goette (1998), responding to evidence from the United States, argued that Swiss federal, cantonal and Zurich city votes on civil-rights issues had on balance protected rather than eroded minority rights. The record is mixed rather than uniformly hostile.

Naturalisation by popular vote

The sharpest evidence comes from municipal naturalisation decisions, which until the 2000s were made in some Swiss municipalities by referendum on individual applicants. Hainmueller and Hangartner (2013), using the official applicant descriptions that voters received, found that country of origin determined success more than any other characteristic, including language skills, integration and economic credentials: the average share of "no" votes was about 40% higher for applicants from (former) Yugoslavia and Turkey than for observably similar applicants from richer northern and western European countries. In a follow-up study of about 1,400 municipalities from 1991 to 2009, they exploited Federal Court rulings that forced municipalities to move decisions to elected councils, and found that naturalisation rates rose by about 60% once politicians rather than citizens decided. The increase was larger for more marginalised groups and in more xenophobic areas, consistent with the explanation that councils must justify rejections and are subject to judicial review, while anonymous voters are not.

Finding 6: comparison with US states

US states offer the main comparison. South Dakota adopted statewide initiative and referendum in 1898 and Oregon first used the initiative in 1904; today around half the states have some form of citizen initiative or popular referendum, but there is no national initiative. Key institutional differences shape outcomes:

On minorities, Gamble (1997) analysed three decades of US civil-rights initiatives and referendums in five areas and found that voters approved more than three-quarters of measures restricting civil rights, while approving only about a third of all initiatives and referendums. Hajnal, Gerber and Louch (2002), studying California propositions since 1978, found that racial and ethnic minorities lost regularly on the small number of racially targeted propositions — under 5% of the total — but were on the winning side of most propositions overall. Matsusaka's For the Many or the Few (2004) argued from US state data that the initiative generally moves fiscal policy towards majority preferences rather than towards narrow interests. Taken together, the US evidence suggests that judicial review in the United States corrects some of the most direct threats to minority rights that, at the Swiss federal level, are checked mainly by parliament, the cantonal majority and international law.

What the evidence does not show

• It does not show that direct democracy produces better policy in general. The fiscal studies show lower spending and debt where fiscal referendums are stronger; whether that is better is a value judgement.
• It does not show that the initiative's low success rate means low influence. Counter-proposals, withdrawals and anticipatory bargaining are central but are not captured by win rates.
• It does not show that Swiss voters are systematically hostile to minorities. Several minority-targeting initiatives passed and others failed; the naturalisation studies concern individual decisions by municipal voters, a different setting from federal votes on general rules.
• It does not establish causation cleanly. Cantonal comparisons cannot fully separate institutions from culture and history.
• It does not show how campaign money shapes outcomes. Disclosure rules took effect only in October 2022, so systematic data on campaign spending are recent.

Implications for institution design

1. Thresholds and timelines matter more than principle. Low signature thresholds make vetoes easy to call and encourage broad parliamentary coalitions; high thresholds would weaken the "consensus" effect.
2. Counter-proposals improve outcomes. The ability of parliament to propose an alternative turns an all-or-nothing vote into a negotiation, and explains many withdrawals.
3. Protect individual rights from decision by plebiscite. Swiss municipalities' shift from popular votes to council decisions on naturalisation, with reasons and judicial review, sharply reduced discrimination. Decisions about individuals are poorly suited to anonymous majority votes.
4. Give courts or treaties a role in reviewing measures that target minorities. Switzerland's lack of constitutional review of federal laws leaves international law and implementing legislation as the main checks, as the 2014 case shows.
5. Make campaign finance transparent. The disclosure rules in force since 2022 are a minimum; spending data should be analysed vote by vote.
6. Expect implementation politics. Initiatives written as constitutional principles leave room for parliament to interpret them; designers should decide in advance how far that discretion should go.

Open questions

• How large is the indirect effect of initiatives through counter-proposals and withdrawals, measured systematically?
• Does frequent voting produce informed or fatigued electorates — and what is cumulative participation over several votes?
• Would a federal constitutional court improve minority protection, or weaken popular sovereignty in ways voters would reject?
• Do the fiscal effects found for the 1980s and 1990s persist after the federal debt brake and cantonal fiscal rules?
• How will disclosure of campaign finance change campaigns and outcomes?
• What lessons transfer to countries without Switzerland's federalism, small size and consensus tradition?

Sources

1. Swissvotes / Année Politique Suisse, University of Bern. Swissvotes dataset of federal popular votes since 1848 (download of October 2026; counts computed by the editors).
2. Swissvotes. Vote 547: Volksinitiative «Gegen den Bau von Minaretten» (29 November 2009).
3. Swiss Federal Chancellery. Volksinitiativen (procedure and signature requirements).
4. Swissinfo. How Swiss direct democracy works.
5. Feld, L. P. & Kirchgässner, G. (2001). The political economy of direct legislation: direct democracy and local decision-making. Economic Policy 16(33), 329–367.
6. Feld, L. P. & Matsusaka, J. G. (2003). Budget referendums and government spending: evidence from Swiss cantons00140-8). Journal of Public Economics 87(12), 2703–2724.
7. Kirchgässner, G., Feld, L. P. & Savioz, M. R. (1999). Die direkte Demokratie: modern, erfolgreich, entwicklungs- und exportfähig. Basel: Helbing & Lichtenhahn; Munich: Vahlen. ISBN 3719018377.
8. Matsusaka, J. G. (2005). Direct democracy works. Journal of Economic Perspectives 19(2), 185–206.
9. Matsusaka, J. G. (2004). For the Many or the Few: The Initiative, Public Policy, and American Democracy. Chicago: University of Chicago Press. ISBN 9780226510811.
10. Hainmueller, J. & Hangartner, D. (2013). Who gets a Swiss passport? A natural experiment in immigrant discrimination. American Political Science Review 107(1), 159–187.
11. Hainmueller, J. & Hangartner, D. (2019). Does direct democracy hurt immigrant minorities? Evidence from naturalization decisions in Switzerland. American Journal of Political Science 63(3), 530–547.
12. Gamble, B. S. (1997). Putting civil rights to a popular vote. American Journal of Political Science 41(1), 245–269.
13. Frey, B. S. & Goette, L. (1998). Does the popular vote destroy civil rights? American Journal of Political Science 42(4), 1343–1348.
14. Hajnal, Z. L., Gerber, E. R. & Louch, H. (2002). Minorities and direct legislation: evidence from California ballot proposition elections. Journal of Politics 64(1), 154–177.
15. European Court of Human Rights (2011). Ouardiri v. Switzerland (dec.), no. 65840/09, and Ligue des musulmans de Suisse and Others v. Switzerland (dec.), no. 66274/09, 28 June 2011. Information Note on the Court's case-law No. 142.
16. Swiss Parliament. Curia Vista 16.027: Ausländergesetz. Steuerung der Zuwanderung und Vollzugsverbesserungen bei den Freizügigkeitsabkommen (final vote 16 December 2016).
17. Altenburger Ltd (2018). Job vacancy reporting duty for Swiss employers. Legal 500 Intelligence, 11 July 2018.