Direct Democracy: Can We Trust the Judgment of the People?
In a representative government, citizens elect officials who make laws and policy decisions on their behalf. Voters choose the decision-makers, but they usually do not vote on the decisions those officials make.
Direct democracy gives citizens the authority to vote on laws and policy questions themselves, often through initiatives and referendums. It does not necessarily eliminate elected representatives. The two systems can operate together, with representatives managing government and drafting laws while citizens retain the power to propose, approve, or reject specific measures.
One of the greatest frustrations with our country’s representative democracy is that decision-making at the top often does not reflect the will of the people, the very problem direct democracy addresses head on.
This investigation makes the case that direct democracy is superior to representative government as a way to align public policy with the will of the majority. That argument is not based merely on political theory or optimism about voters. Factual evidence involving voter participation, taxation, minority rights, legislative representation, and collective judgment supports the conclusion.
Good judgment. What is the most powerful objection to direct democracy in this country? People don’t trust others to have enough collective good judgment to make better decisions that affect us all.
The objections are familiar: inadequate intelligence to digest the complexity of legislative bills, lack of interest in participating, taxes going up, minority rights being disregarded, lemmings running over a cliff, Switzerland’s example not scaling to our country, and most fundamentally, the belief that elected representatives make wiser choices than average citizens do.
Let’s see about that.
1. Aren’t bills too hard for the average person to understand?
AI has been rightly hammered for its inaccuracies and confident fabrications, especially when asked open-ended questions. But one of its greatest strengths is its ability to take complicated written material and explain it in plain, understandable English.
It is now straightforward to give AI the actual text of a legislative bill and ask for a summary, including the strongest arguments for and against it, who benefits, who pays, and what the law would change.
Of course, AI can still make mistakes, even when summarizing documents. That’s why its explanations should include links to the actual bill, independent reviews, and ways to challenge anything that appears misleading or inaccurate.
But consider what this means. A citizen no longer needs a law degree or hours of free time to begin understanding legislation that affects their life.
We can make complex bills understandable without requiring every voter to become an expert. And we can present the strongest arguments on both sides without telling citizens what they should think.
See https://civox.us. Click the map or enter your zip code to see how your representatives voted recently and to see AI summaries of the legislative bills.
2. Motivation inspires itself.
It’s human nature to initially resist change. It’s also human nature to adapt to something that becomes a persistent part of life.
Why is there so much voter apathy in America? One possible explanation is that many people don’t see how voting makes a difference to them personally. They cast a ballot every few years, and then watch elected officials make thousands of decisions over which they have little direct control.
But what happens when citizens know their votes can actually decide the outcome of a particular law?
Switzerland offers an interesting answer.
Switzerland is not governed by public votes alone. It combines representative government with direct democracy: citizens elect a national parliament, parliament selects a seven-member Federal Council to administer the country, and voters retain direct authority over important laws and constitutional questions. Similar arrangements operate within the country’s cantons, which are roughly comparable to U.S. states, and its municipalities.
At the national level, citizens can propose a constitutional amendment by collecting 100,000 signatures within 18 months. They can also force a public vote on a law passed by parliament by collecting 50,000 signatures within 100 days. Constitutional amendments must be submitted to voters automatically and require approval from both a majority of voters nationwide and a majority of the cantons. These mechanisms bring Swiss citizens to the ballot several times a year and allow elected institutions and direct public decision-making to function together.
Swiss turnout for individual national votes averages around 45%. Critics love to point to that number.
But there’s more to the story.
Research examining participation across ten Swiss national votes found that about 20% of voters participated every time, 60% participated in some votes, and 20% didn’t participate at all.
In other words, about 80% participated at least once.
Most Swiss citizens are selective voters. They turn out when an issue interests them, affects them, or inspires them to participate. How refreshing.
A citizen who doesn’t care much about a particular transportation bill might care deeply about healthcare, taxes, education, or the environment.
Direct democracy gives people an opportunity to participate when the issue matters to them.
Once citizens see that their decisions can actually shape their circumstances, they will have all the motivation they need, well beyond choosing a politician every few years.
3. Won’t taxes go up?
Maybe. But when a tax proposal is put directly to voters, they get to decide whether it’s worth paying for.
Consider the country with a track record: Switzerland.
What’s particularly interesting is that Swiss voters have rejected tax breaks for corporations and approved minimum taxes on large multinational companies, while rejecting proposals to sharply increase taxes on the very wealthy.
In 2017, 59.1% rejected a corporate tax reform that included new tax relief for businesses.
In 2023, 78.5% approved a minimum tax of 15% on large multinational companies.
Then, in 2025, voters considered a proposal for a 50% federal inheritance tax on amounts above 50 million Swiss francs, with the money going toward climate programs.
They rejected it by 78.3%.
And in March 2026, 54.2% approved changes intended to treat married and unmarried taxpayers more equally.
Notice the pattern?
Swiss voters aren’t automatically voting for higher taxes, even when somebody else would pay them. Nor are they automatically voting to cut taxes.
They’re making judgments about individual proposals.
Switzerland doesn’t prove that voters always make fair tax decisions. It demonstrates that giving them direct authority over taxation does not automatically produce higher taxes or runaway spending.
4. Won’t the majority take away minority rights?
Consider that if a strong enough majority becomes determined to trample on the rights of a minority, there is no constitutional rule that can absolutely guarantee protection.
Even in our representative democracy, amendments to the Constitution can be proposed by two-thirds of both houses of Congress and ratified by three-fourths of the states.
So we inherently rely on the judgment and morality of the population to respect everyone’s freedom and civil liberties.
Hanging that responsibility entirely on institutional protections is only effective to the extent that the people operating those institutions exercise good judgment and respect the legal framework.
Constitutional safeguards don’t enforce themselves. People enforce them.
That’s not an argument against constitutional protections. It’s an acknowledgment that their effectiveness ultimately depends on human judgment.
The next question is how likely the majority is to restrict or eliminate the freedoms of particular groups.
History offers some troubling examples. In 1959, Swiss male voters rejected extending federal voting rights to women. It took until 1971 before they approved those rights.
But consider the broader evidence.
Researchers Bruno Frey and Lorenz Goette examined 20 Swiss national votes involving civil and minority rights. They found that 16 of those 20 votes, or 80%, produced outcomes favorable to minority interests.
Across 64 votes at federal, cantonal, and city levels, about 70% favored minority interests.
American polling also shows substantial majority support for specific minority protections. For example, a 2025 PRRI survey found that 72% supported laws protecting LGBTQ people from discrimination in employment, housing, and public accommodations.
Does that mean the majority will always protect minorities?
Of course not.
But here’s the underlying point: Democracy doesn’t depend on a majority that never makes a bad decision. It depends on a majority that generally exercises good judgment.
Nothing in life offers absolute guarantees. Constitutional protections can reduce the chances of injustice, but ultimately even those protections depend on people willing to enforce them.
The Swiss evidence shows that majorities can and often do respect minority rights.
5. What if the crowd makes a terrible mistake?
Politicians can be fooled, too. Or lie to us.
Even the sincere ones face pressure from parties, donors, lobbyists, and news outlets that can lead to mistakes. Voting for a representative does not make human weaknesses disappear.
And what makes a few hundred politicians inherently wiser than millions of citizens?
We don’t judge car safety by pointing to one crash. We look at many trips, crashes, and deaths over time.
Do we make good policy decisions based on single data points or on evidence-based statistical trends?
We should judge a voting system the same way.
The question isn’t whether the majority sometimes makes bad decisions. Of course it does. The question is how often, how serious those mistakes are, and whether elected representatives generally make better decisions.
That’s a question to be answered with evidence, not assumptions.
6. Can the Swiss example be scaled to our country?
Switzerland has about 9 million people. The United States has about 342 million, nearly 38 times the size.
California offers a useful example of how part of the Swiss model can operate within a much larger population. But there is an important distinction: blame California’s broader legal and regulatory system on its legislature and government agencies, not on citizen initiatives. Its high taxes, extensive regulations, costly approval processes, and other widely criticized policies are the result of legislative and bureaucrat decision-making, not the result of citizen initiatives in California.
Regarding citizen initiatives specifically, one survey found a roughly 70% satisfaction rate with successful California citizen initiatives. That finding covered these kind of initiatives:
2024 Restrict Drug-Program Revenue Spending
2024 Managed-Care Organization Tax
2024 Crime and Penalties
2022 K–12 Arts and Music Funding
2022 Uphold Flavored-Tobacco Ban
2020 Consumer Privacy Protections
2020 App-Based Driver Employment Classification
2020 Stem-Cell Research Bonds
2018 Farm-Animal Confinement Standards
2018 Ambulance Employee Work Breaks
2018 Children’s Hospital Construction Bonds
That limited scope puts the result in a different context. A 70% satisfaction rate seems reasonable for these initiatives.
What needs improvement is how costly and time consuming it is to get those initiatives on the ballot in the first place. But that’s a subject of process improvement.
California’s citizen-initiative process is therefore a successful direct-democracy subset of an otherwise onerous and legalistic system. It shows that direct public lawmaking can work at a scale substantially larger than Switzerland, even when the surrounding representative institutions remain deeply frustrating.
7. Does winning an election mean a representative speaks for most people?
Suppose one candidate wins an election with 55% of the vote. The other candidate gets 45%.
How likely are those 45% to agree with the winning candidate on future legislative votes?
And even among the 55% who elected that representative, how many are going to agree with every vote?
Winning an election tells us whom people preferred as their representative. It doesn’t tell us whether they agree with that person’s position on the next bill.
And now we have actual research measuring that gap.
Political scientist John Matsusaka compared 4,192 state legislative votes with the majority preferences of the lawmakers’ districts.
He found that legislators voted against their district’s majority about 34% of the time on the laws studied.
That’s approximately one out of every three legislative votes.
It doesn’t mean one-third of all constituents disagreed on every bill. It means that on about one-third of the legislative votes studied, the lawmaker took the opposite position from the district majority.
There is a substantial difference between electing someone to represent us and having that person’s decisions actually reflect our preferences.
And the difference can change from one bill to the next.
Switzerland demonstrates another benefit.
Swiss lawmakers know voters can challenge many laws.
That gives leaders a reason to listen and find common ground before a law is finished in the first place.
One research review found that only about 5% of parliamentary laws were challenged through referendums, and voters approved more than 70% of the challenged laws.
The possibility of a public vote hovers auspiciously over legislators even when no vote takes place, serving to keep them in line with the majority view.
Populism
Let’s not confuse populism with popular decision-making.
One is born from an ideology. The other is when the majority decides an outcome.
They’re not the same thing.
The wisdom of the crowd
Beyond addressing these concerns, there are some remarkable examples of the wisdom of crowds.
Get T. Greg O'Neil’s stories in your inbox
Join Medium for free to get updates from this writer.
Take the traditional county fair game of guessing the number of jelly beans in a jar.
Individual guesses tend to be wildly inaccurate. Yet when enough people make reasonably independent guesses, the average comes surprisingly close to the actual number.
Why?
Because some people guess too high, others guess too low, and their individual errors can cancel each other out. But there is an important caveat. If everybody shares the same mistaken assumption, simply adding more people won’t necessarily improve the result.
The strength comes from combining many reasonably independent judgments, not from blindly following the herd.
Another example is the modern internet search engine.
It’s remarkable how the top ten search results can come so close to addressing a question out of trillions of sources of information.
Google uses signals drawn from links between webpages, patterns in how people search and interact with results, and sophisticated algorithms that evaluate relevance and quality. Crowdsourced information is the key ingredient.
Search engines don’t prove that the majority is always right. But they demonstrate how powerful it can be to collect and combine signals from enormous numbers of sources.
And consider another finding from research on collective intelligence.
A study involving 5,180 participants found that combining the judgments of small, independent groups that discussed factual questions could produce better estimates than simply averaging thousands of individual guesses.
That doesn’t establish that crowds always choose good public policies. It does show that discussion, diverse perspectives, and combining information can improve collective judgment.
Which raises an interesting question.
If we recognize the value of collective intelligence in so many areas of life, why do we so readily dismiss its potential when ordinary citizens participate in making laws?
But here’s the most important question of all.
Does our current representative democracy actually deliver what the majority of Americans want?
We’ve examined the familiar objections to direct democracy. We’ve looked at voter competence, motivation, taxes, minority rights, the possibility of mistakes, and whether direct democracy can work in a large population.
But all those objections rest, at least in part, on an underlying assumption: that elected representatives make better decisions than ordinary citizens would.
So how well does our existing system represent the people?
Fortunately, researchers have measured that, too.
In a study published in 2025, political scientists Stephen Ansolabehere of Harvard and Shiro Kuriwaki of Yale examined 103 important congressional issues from 2006 through 2022.
They compared what the majority of Americans wanted with what Congress ultimately did.
The results were remarkable.
Congress made decisions consistent with the national majority’s wishes on just 55% of those issues.
On the remaining 45%, Congress produced an outcome contrary to what the majority of Americans wanted.
That’s nearly one out of every two decisions.
And here’s another revealing finding.
Most of those disagreements weren’t caused by Congress passing unpopular laws. They happened because Congress failed to pass proposals that a majority of Americans supported, with the Senate playing a particularly important role.
This doesn’t mean the public was necessarily right on every issue. Nor does it establish what would happen if citizens voted directly on every proposal.
But it does establish a serious gap between what most Americans wanted and what their elected government delivered on those major issues.
Which brings us right back to the central question.
We are repeatedly told that ordinary citizens can’t be trusted to make sound legislative decisions, that our elected representatives are better equipped to exercise good judgment on our behalf. The evidence does not support the claim that a representative government is superior to one based on direct democracy. In fact, the evidence supports the opposite.
We can continue pretending we are not sheep, or we can take the bull by the horns and demand that our voice counts.
References
1. AI and the Accuracy of Document Summaries
Bao, F. S., et al. (2025). FaithBench: A Diverse Hallucination Benchmark for Summarization by Modern LLMs. Proceedings of NAACL 2025.
Examines the accuracy of AI-generated summaries and identifies problems with unsupported or fabricated statements.
2. Swiss Voter Participation and Direct Democracy
Kübler, Daniel (2024). Citizen Participation Through Direct Legislation: A Road to Success? A Perspective from Switzerland. Global Public Policy and Governance, 4, 184–196.
Reviews Swiss direct democracy, including voter competence, selective participation, minority rights, government spending, and the influence of referendums on lawmakers. Includes the research showing that approximately 80% of Swiss voters participated in at least one of ten national votes examined.
3. Switzerland’s Tax Referendums
Swiss Federal Department of Finance. Federal Popular Votes on Taxation, 2017–2026.
Official results covering the rejection of corporate tax reforms in 2017, approval of a 15% minimum corporate tax in 2023, rejection of a proposed inheritance tax in 2025, and approval of individual taxation reform in 2026.
4. Direct Democracy and Government Spending
Feld, Lars P., and John G. Matsusaka (2003). Budget Referendums and Government Spending: Evidence from Swiss Cantons. Journal of Public Economics, 87(12), 2703–2724.
Found that mandatory spending referendums were associated with approximately 19% lower government spending in the median Swiss canton after accounting for other factors.
5. Constitutional Amendments and Minority Rights
U.S. National Archives. The Constitution of the United States, Article V.
Explains the procedures for amending the Constitution, including the requirements for congressional proposals and state ratification.
6. Swiss Women’s Voting Rights
Swiss Federal Archives. Women’s Suffrage in Switzerland.
Documents the rejection of women’s federal voting rights in 1959 and their approval in 1971.
7. Majority Rule and Minority Rights in Switzerland
Frey, Bruno S., and Lorenz Goette (1998). Does the Popular Vote Destroy Civil Rights? American Journal of Political Science, 42(4), 1343–1348.
Examined 64 Swiss popular votes involving civil and minority rights. Of 20 federal votes, 16 produced outcomes favorable to minority interests. Across all 64 votes, approximately 70% were classified as favorable.
8. American Public Opinion on Minority Rights
Public Religion Research Institute (2026). Mapping Support for LGBTQ Rights Across the 50 States: Insights from PRRI’s 2025 American Values Atlas.
Based on interviews with more than 22,000 American adults. Found 72% support for LGBTQ nondiscrimination protections, 71% agreement that transgender Americans deserve equal rights and protections, and 65% support for same-sex marriage.
9. Population Figures for Switzerland
Swiss Federal Statistical Office (2026). Population Growth Slowed Slightly in 2025.
10. California Voter Views of the Initiative Process
Public Policy Institute of California (2024). Californians and Their Government, September 2024.
Among California likely voters, 79% said allowing voters to make laws through initiatives was a good thing, 70% were satisfied with how the initiative process was working, 80% said it needed changes, and 87% said initiatives raised issues the governor and legislature had not adequately addressed.
Read the PPIC September 2024 survey
11. California Voter Control Over Taxes and Spending
Public Policy Institute of California (2026). Californians and Their Government, February 2026.
Among California likely voters, 77% preferred that voters make some state spending and tax decisions at the ballot box, while 22% preferred leaving all such decisions to the governor and legislature. Separately, 51% expressed trust in voters to make fiscal decisions at the ballot box.
Read the PPIC February 2026 survey
12. Trust in California State and Federal Government
Public Policy Institute of California (2026). Californians and Their Government, September 2026.
Among California adults, 40% trusted state government to do what was right just about always or most of the time, compared with 18% who said the same of the federal government. The difference does not establish that direct democracy caused the higher level of state-government trust.
Read the PPIC September 2026 survey
13. Direct Democracy in American States
National Conference of State Legislatures (2025). Initiative and Referendum Processes.
Explains how citizen initiatives and referendums operate in the United States, including the 24 states allowing citizen initiatives.
14. How Often Legislators Represent Their Constituents
Matsusaka, John G. (2026). When Do Legislators Represent Their Constituents? Evidence from Roll-Call and Referendum Votes. The Journal of Politics, 88(2), 854–866.
Analyzed 4,192 legislative votes on 33 laws across ten states. Found that legislators voted in agreement with their district’s majority preference 66.3% of the time, leaving a 33.7% mismatch.
15. Google Search and Aggregating Information
Google. How Search Works: Ranking Results.
Explains how Google combines numerous signals, including webpage content, links, relevance, quality, and aggregated user interaction data, to identify useful search results.
16. The Wisdom of Crowds
Navajas, Joaquin, et al. (2018). Aggregated Knowledge From a Small Number of Debates Outperforms the Wisdom of Large Crowds. Nature Human Behaviour, 2, 126–132.
A study involving 5,180 participants found that combining the judgments of small, independent groups that deliberated could produce more accurate answers to factual questions than averaging thousands of individual estimates.
17. Congress and the Will of the American People
Ansolabehere, Stephen, and Shiro Kuriwaki (2025). Collective Representation in Congress. Perspectives on Politics. Cambridge University Press.
Examined 103 important congressional issues from 2006 through 2022. Found that Congress reached outcomes consistent with national majority preferences on 55% of those issues and contrary outcomes on 45%. Most mismatches involved popular proposals that failed to become law.
