Two years ago, Chileans voted overwhelmingly to replace their constitution. Tonight, they have overwhelmingly rejected the constitution written to replace it.
The margin is not close. With almost all polling stations counted, 61.86% have voted to reject the proposed constitution, while only 38.14% voted to approve it. Nearly 7.9 million Chileans voted Rechazo against roughly 4.9 million voting Apruebo.
Those numbers become much more interesting when placed beside the referendum that began this process. In October 2020, 78.28% of voters supported writing a new constitution, while only 21.73% wanted to retain the existing one. An elected Constitutional Convention was subsequently created to draft the replacement.
At first, the two results appear almost contradictory. How does a country move from nearly four in five voters demanding a new constitution to more than three in five rejecting the constitution eventually produced?
The answer may be that Chileans have not necessarily reversed their position as dramatically as the percentages suggest. In 2020, voters answered one question: should Chile write a new constitution? Today, they answered another: should this particular document become that constitution? Wanting institutional change does not require accepting every version of institutional change offered afterwards.
That distinction matters because political movements frequently interpret mandates too broadly. A population demands reform, a reforming coalition wins power, and politicians begin treating the original demand as permission for their preferred solution. Yet electoral mandates rarely contain that level of detail. Millions of people can agree that the existing system is inadequate while disagreeing profoundly about what should replace it.
Chile has just demonstrated the difference numerically.
The constitutional process originated in the enormous protests of 2019, when anger over a metro fare increase developed into a much broader challenge concerning pensions, inequality, healthcare, education and the distribution of economic opportunity. The constitution inherited from the Pinochet era, despite having been amended repeatedly since democracy returned, became a symbol of a political and economic settlement that many Chileans believed needed deeper reform.
The 2020 referendum appeared to confirm that conclusion overwhelmingly. Chileans did not merely elect another government promising incremental policy changes. They authorised a constituent process intended to reconsider some of the fundamental rules governing the state.
That is an unusually powerful democratic mandate.
But a mandate to write a constitution is not a mandate to ratify whatever the writers produce.
The distinction may seem obvious tonight. It was politically easier to overlook while the document was being drafted.
Chile’s proposed constitution was extraordinarily ambitious. It would have replaced the existing constitutional framework with a substantially more expansive conception of social rights, environmental protections, gender equality and indigenous recognition. It proposed changes to political institutions and recognised Chile as a plurinational state, acknowledging distinct indigenous nations within the country.
Supporters saw many of these provisions as an attempt to construct a more inclusive social state after decades in which market-oriented institutions had produced growth but also persistent dissatisfaction with inequality and insecurity.
Critics saw a document that attempted to change too much simultaneously, contained institutional uncertainties and moved beyond the consensus implied by the original referendum.
Those interpretations do not map perfectly onto left and right.
That may be the most important feature of today’s result.
If every voter who wanted constitutional change had supported the draft, Rechazo could not have reached anything close to 61.86%. The arithmetic itself tells us that large numbers of people who supported replacing the constitution in 2020 have rejected this replacement in 2022.
That should make reformers cautious about dismissing the result as simple conservative resistance.
Some voters undoubtedly oppose the constitutional project’s progressive objectives.
Others may support many of those objectives while opposing particular mechanisms.
Those are politically different constituencies, even though both appear tonight under one word: Rechazo.
Referendums compress complicated political preferences particularly aggressively. A voter receives one binary choice over a document containing hundreds of provisions. They may strongly support environmental rights, moderately support pension reform, oppose plurinationality, dislike the proposed political structure and remain uncertain about judicial reforms.
The ballot cannot record that combination.
It records reject.
Another voter may oppose the entire ideological direction of the document.
The ballot records exactly the same thing.
This makes interpreting referendum mandates unusually difficult. The result tells us conclusively that the document did not obtain public approval. It does not tell us which clauses caused each voter to reject it or which reforms those voters would accept instead.
The political temptation will be to interpret the result maximally. Conservatives can say Chile rejected the progressive transformation associated with the constitutional convention. Supporters of reform can say misinformation or fear defeated necessary change. President Gabriel Boric’s opponents can interpret the result as rejection of his government.
Each explanation may capture part of the electorate.
None should automatically be treated as the electorate’s single collective intention.
There is no single Chilean voter hiding inside the 61.86%.
This is why the next stage matters more than assigning ideological ownership of tonight’s result. If Chileans still broadly support replacing the existing constitution, political leaders need to determine which elements of the failed draft prevented that support from translating into approval.
That requires listening to the rejection rather than explaining it away.
Political movements are often much better at listening to citizens while trying to gain power than after they acquire it. Before victory, every dissatisfied voter is evidence that institutions need reform. After victory, criticism can begin looking like obstruction by people who do not understand the project.
That is how reform movements become establishments remarkably quickly.
Gabriel Boric should be particularly careful about this. His presidential victory last December was closely connected to the political generation that emerged from Chile’s student movements and the broader demand for structural change. Although the Constitutional Convention was independent from his government, Boric supported the proposed constitution and its defeat will inevitably weaken his political position.
The worst response would be to conclude that Chileans simply failed to understand what was good for them.
That would transform a democratic reform project into a technocratic one.
Constitutional legitimacy cannot be produced through expert confidence alone.
A constitution is not an ordinary policy document. Governments can pass tax reforms with narrow parliamentary majorities and reverse them after the next election. Constitutions are supposed to establish rules under which competing political movements are willing to govern and lose for decades.
That requires a broader kind of legitimacy.
A successful constitution should ideally be tolerable even to people who dislike the government operating under it.
This is why constitutional design differs from ordinary legislation. If a left-wing majority writes institutions principally to implement left-wing objectives, the right has incentives to rewrite them when it eventually gains power. If the right does the same, the left responds similarly.
The constitution then becomes another prize of electoral competition rather than the framework within which competition occurs.
Durability therefore requires some degree of loser’s consent before anyone has actually lost.
The people writing the rules need to ask whether their opponents could live under them.
That does not mean constitutions should contain no substantive values. Every constitution embeds values through rights, institutional structures and distributions of authority.
It means constitutional ambition needs enough consensus that political opponents still recognise the framework as theirs.
Chile’s convention may have struggled with this partly because of how the constituent moment developed. The 2019 protests generated enormous demand for transformation. The 2020 referendum produced an overwhelming vote for a new constitution. The Constitutional Convention election then produced a body containing many independents and left-wing representatives while traditional political parties performed relatively poorly.
Each stage could reasonably be interpreted as evidence that Chileans wanted a significant break with the existing order.
But political momentum can create its own information problem.
If every successive institution is populated primarily by people most enthusiastic about reform, the process may gradually become less representative of citizens who supported change more cautiously.
This resembles selection effects.
The people willing to spend enormous time campaigning for a constitutional convention are not necessarily representative of the median voter who simply wants better pensions and healthcare.
Political activists generally possess stronger ideological preferences than ordinary voters.
That is not a criticism of activism.
It is a structural feature of participation.
People who care most intensely participate most.
The result can be a gap between the preferences of a movement’s most active members and the broader coalition that originally supported the movement.
Chile’s referendum may be evidence that such a gap became too large.
This problem appears across democracies. Party activists can be more ideological than general-election voters. Social-media political communities can appear much more polarised than the wider population. Protest movements can mistake the views of people attending demonstrations for the views of everyone sympathetic to the cause.
Intensity and breadth are different political resources.
The convention possessed enormous intensity.
Today’s referendum measured breadth.
The second did not confirm the first.
There is also a broader lesson about political expectations. The 2019 protests placed enormous pressure on Chile’s political system because many citizens felt the economic model was producing unacceptable insecurity or inequality. A new constitution became one mechanism through which those frustrations could be addressed.
But constitutions cannot solve every economic problem.
A constitution can recognise rights to healthcare, housing, education or social security. It cannot itself create doctors, houses, teachers or pension revenue.
Those require fiscal resources, administrative capacity and economic growth.
This distinction between constitutional rights and state capacity matters because writing a right into law changes the government’s obligation without automatically changing its ability to fulfil it.
That does not make social rights meaningless. Constitutional recognition can shape legislation, budgets and judicial interpretation.
But political systems should be cautious about promising outcomes institutionally without explaining the resources required to provide them.
A right becomes politically fragile if citizens repeatedly discover that its formal existence does not improve actual access.
Again, government cannot legislate away scarcity.
Chile is wealthy by Latin American standards, but resources remain finite. Expanding social provision requires choices over taxation, spending, private-sector participation and economic growth.
Those arguments do not disappear when the constitution changes.
They become the politics conducted underneath it.
The proposed recognition of Chile as plurinational illustrates another difficulty: institutional recognition of identity can correct historical exclusion while simultaneously creating fears about political fragmentation.
Chile’s indigenous peoples, particularly the Mapuche, have legitimate historical grievances concerning land, representation and state treatment. A political system that simply insists everyone is formally equal can reproduce inequality when historical institutions were not equal.
Recognition therefore matters.
But institutionalising distinct national identities can also make some citizens worry that common citizenship is being weakened.
The question is not whether either concern is irrational.
It is how institutions can acknowledge historical difference without creating political incentives that permanently divide citizens into competing constitutional categories.
Northern Ireland demonstrated a related problem in May. Power-sharing institutions successfully protect communal identities but can also preserve those identities as the central organising principle of politics.
Constitutional design always creates incentives.
Recognition can empower.
It can also entrench.
The task is deciding which trade-off is justified in a particular society.
Chile’s convention attempted to answer many of these questions at once: indigenous recognition, social rights, environmental governance, gender representation, political institutions and decentralisation.
Ambition can be admirable.
It also increases the number of reasons a voter might reject the package.
This is a basic coalition problem.
Imagine a constitution contains ten major reforms. Each individual reform has 60% support, but different 40% minorities oppose each one. If voters must accept or reject all ten together, the coalition supporting the entire package can be much smaller than the majority supporting any individual provision.
Bundling changes the politics.
This is exactly the problem I discussed after Australia’s election, where political parties package policies into bundles.
Constitutions are the largest political bundles imaginable.
That means constitution-writing requires unusually careful coalition construction.
The objective is not maximising the number of reforms that can be inserted while your side controls the convention.
It is maximising the durability of the institutional settlement that emerges.
Those are different optimisation problems.
Today’s result suggests the convention may have optimised too heavily for transformation and insufficiently for ratification.
That conclusion should remain provisional because understanding voter motivations will require more evidence than the aggregate result provides.
But the margin is large enough that blaming a small communications failure would be implausible.
A proposal rejected by nearly 62% of voters has a coalition problem.
The responsible question is why.
This does not mean Chile should return complacently to the pre-2019 political settlement.
That would misread the earlier referendum just as badly.
In 2020, nearly four in five voters supported beginning a new constitutional process. Those votes do not disappear because today’s draft failed.
The two referendums need to be interpreted together.
Chile wants change, but it has not accepted this version of change.
That is a more complicated democratic instruction than either side would probably prefer.
It is also potentially productive.
Political leaders now know that constitutional reform needs a broader coalition than the convention achieved.
The centre-right cannot plausibly claim the old constitution has suddenly acquired overwhelming positive legitimacy.
The left cannot plausibly claim its preferred transformation already possesses a popular mandate.
Both need one another more than they did yesterday.
That could create a better constitutional process.
Or it could create paralysis.
The outcome depends on whether political actors interpret defeat as information.
This may be one of the most useful definitions of democratic maturity: losing should teach you something.
Not every electoral defeat means your principles are wrong. Sometimes the candidate was weak, the campaign failed or external circumstances changed.
But a large defeat is evidence.
Political movements that treat every victory as a mandate and every defeat as misunderstanding eventually stop learning from voters.
Democracy becomes useful precisely because it provides feedback politicians cannot completely control.
Markets provide price signals.
Science provides experimental results.
Democracy provides votes.
All three mechanisms become useless if participants explain away every unfavourable signal.
Boric’s government therefore faces a difficult but valuable moment.
Its supporters believed Chile was ready for profound institutional transformation.
Tonight, voters have rejected the document associated with that transformation by a very large margin.
The mature response is neither abandoning reform nor repeating the same proposal more loudly.
It is updating.
Which reforms retain broad support?
Which provisions frightened moderate voters?
Which institutional designs appeared unclear?
How much change can command durable consensus?
How should indigenous rights be protected?
How should social rights be financed?
How should the constitution distribute power?
Those questions are less emotionally satisfying than declaring victory or betrayal.
They are constitution-building.
Chile has done something democratically impressive today even though the process itself has failed.
A country emerging from enormous protests attempted to rewrite its fundamental political rules through an elected convention and then gave citizens the power to reject the result.
They rejected it.
That is not democracy malfunctioning.
That is the ratification mechanism doing exactly what it was designed to do.
The constitutional project now has to become better at persuading people who were not already convinced.
Perhaps that is the difference between political transformation and democratic transformation.
Political transformation asks whether you have enough power to change the rules.
Democratic transformation asks whether enough of the people who disagree with you can still accept the rules after you change them.
Chile achieved the first opportunity.
Tonight, voters said it had not yet achieved the second.
