Luiz Inácio Lula da Silva will return to the Brazilian presidency twelve years after leaving office. In last Sunday’s runoff, Lula received 60,345,999 votes, or 50.90% of valid ballots, against Jair Bolsonaro’s 58,206,354 votes, or 49.10%. More than 118 million Brazilians cast valid presidential votes, and only a little over two million separated the candidates.
The result is decisive in the constitutional sense and extraordinarily narrow in the political one. Brazil does not require a president to win by ten percentage points before the result becomes legitimate. Lula received more votes, the electoral authorities declared him elected, and the constitutional process now moves towards a transfer of power. But a democracy can produce a completely unambiguous winner while simultaneously revealing a society divided almost exactly in half.
That distinction should shape how Lula interprets his victory.
He has won the authority to govern Brazil. He has not won evidence that Bolsonaro’s political movement has disappeared, that the country has shifted decisively leftwards or that 50.90% of the electorate supports every part of his programme. Some Lula voters enthusiastically support the Workers’ Party. Others may have voted primarily to remove Bolsonaro. Centrist politicians who once opposed Lula supported him because they considered Bolsonaro a greater threat to democratic institutions. Lula’s own running mate, Geraldo Alckmin, was previously one of his major centre-right political opponents.
The coalition that won the election is therefore broader than Lula’s ideological coalition.
That can be a source of strength. It can also become a governing constraint.
Negative coalitions are easier to maintain before victory than afterwards. Several groups can agree that Bolsonaro should not receive another term while disagreeing substantially about taxation, public spending, privatisation, environmental regulation and the role of the state. Once the common opponent loses, those disagreements return.
This is why winning a polarised election can create a misleading impression of political unity. The ballot compresses a complicated coalition into one number.
Government decompresses it again.
Brazil’s election also matters because of everything Bolsonaro said before the vote. For years, he repeatedly questioned the reliability of Brazil’s electronic voting system without producing evidence demonstrating the large-scale fraud he alleged was possible. That rhetoric created concern that a narrow defeat could be followed by an attempt to reject the electoral result itself.
The similarity with Donald Trump’s behaviour before and after the 2020 American election is difficult to ignore, but the comparison should be made carefully. Bolsonaro is not Trump, Brazil is not the United States, and institutional structures differ. We should analyse what Brazilian actors actually do rather than assume they will reproduce another country’s crisis.
Still, the underlying democratic mechanism is familiar.
A politician tells supporters repeatedly that an electoral system may be fraudulent. The politician then loses. Supporters have already been given an explanation for the defeat that does not require accepting that more people voted for the opponent.
This is why attacks on electoral legitimacy before an election can be more dangerous than complaints afterwards. They construct the interpretive framework through which defeat will later be understood.
Once the votes arrive, the politician does not need to prove fraud from the beginning.
The suspicion has already been planted.
Brazil’s electoral institutions therefore faced a problem extending beyond counting ballots correctly. They needed citizens to believe that the count deserved authority.
Those are different tasks.
An election can be technically accurate and politically unstable if large numbers of people reject the institution producing the result.
This is why trust and verifiability matter together. Brazil’s electronic voting system has been used for decades, and electoral authorities have repeatedly defended its security. Political parties and institutions have mechanisms for auditing and scrutinising the process. Bolsonaro’s allegations did not establish fraud capable of invalidating the election.
But simply repeating that institutions are trustworthy will never be sufficient once distrust becomes partisan identity.
Institutions need procedures capable of demonstrating reliability even to sceptics.
That is difficult because some political claims become effectively unfalsifiable. If an audit finds no fraud, someone already convinced that the entire system is corrupt can claim the auditors are corrupt too. If courts reject the allegation, courts become part of the conspiracy. Evidence against the claim becomes evidence of institutional coordination.
At that point, the dispute is no longer primarily factual.
It concerns who possesses epistemic authority.
Who gets to establish what happened?
Election officials?
Courts?
Political leaders?
Social-media networks?
The answer cannot simply be whoever has the largest audience.
Democratic systems need institutions whose factual determinations retain enough legitimacy that political disagreement can proceed from a common baseline.
Lula won.
A Bolsonaro supporter can believe Lula’s economic programme will be disastrous, that his Workers’ Party governed badly previously or that Bolsonaro was a better president.
Those are legitimate political disagreements.
The democratic threshold is accepting that those disagreements continue under a president elected by the rules.
This is the distinction between opposition and delegitimation.
Opposition says: you won, and I will spend the next four years trying to defeat your policies and remove you electorally.
Delegitimation says: because I believe you should not have won, your victory itself cannot have been legitimate.
The first is democracy functioning.
The second eventually makes democracy impossible.
Bolsonaro’s behaviour immediately after the result was therefore closely watched. He remained publicly silent for nearly two days rather than immediately conceding. When he eventually spoke, he did not explicitly concede defeat, although his administration began permitting the presidential transition to proceed.
That ambiguity matters, but actions may matter more than the absence of a conventional concession speech.
A concession is a democratic norm rather than the legal mechanism transferring power. Lula does not require Bolsonaro’s personal permission to have won.
The more important question is whether state institutions cooperate with the transition and whether Bolsonaro uses his political authority to obstruct the constitutional transfer.
This distinction between symbolic concession and institutional compliance is useful.
Ideally, democracies receive both.
If forced to choose, the second matters more.
A defeated politician can remain personally furious while still instructing officials to obey the law.
Democracy does not require losers to become emotionally reconciled to defeat.
It requires behavioural compliance with institutional rules.
That may be a more realistic standard.
Bolsonaro’s supporters also deserve to be treated as political citizens rather than an undifferentiated anti-democratic bloc. More than 58 million Brazilians voted for him. They did so for different reasons. Some support his conservative social politics. Some prefer his economic agenda. Some distrust Lula and the Workers’ Party. Some value his positions on crime or religion. Some undoubtedly share his distrust of electoral institutions.
Reducing all 58 million to one motivation would make understanding Brazil impossible.
This is where Lula’s language after victory becomes important. He has emphasised that he intends to govern all Brazilians rather than only those who voted for him.
Every winning politician says some version of this.
The difficult part is translating it into political behaviour.
National unity cannot mean pretending substantive disagreements disappeared on election night. Nor can it mean asking Bolsonaro voters to abandon their political beliefs.
A democratic version of unity is much thinner.
It means agreeing that political opponents remain legitimate members of the same political community.
That may be enough.
Polarised democracies often search for emotional reconciliation when procedural reconciliation is more realistic. Lula voters and Bolsonaro voters do not need to like one another. They need to accept that each has the same right to participate and that neither side owns Brazil.
This is particularly important because Brazil’s political divisions contain geographical, religious, class and cultural dimensions. Lula performed particularly strongly in the poorer Northeast, while Bolsonaro retained substantial support across parts of the wealthier South, Southeast and Centre-West. Evangelical voters became especially important to Bolsonaro’s coalition, while Lula’s economic legacy among lower-income voters remains central to his support.
These patterns create competing experiences of the same country.
A voter in Pernambuco may remember Lula’s earlier presidencies through rising incomes, social programmes and expanded opportunity.
A conservative evangelical voter elsewhere may experience Bolsonaro’s politics through defence of religious and cultural values.
An agribusiness producer may prioritise regulation and property rights.
An environmental activist may prioritise Amazon protection.
Each constructs a different hierarchy of political concerns.
Polarisation becomes particularly intense when people stop believing the other side merely prioritises different interests and begin believing it threatens the country itself.
Bolsonaro frequently framed the Workers’ Party through fears of socialism, corruption and moral decline. Lula’s coalition framed Bolsonaro partly through threats to democracy, environmental protection and social welfare.
When both sides believe the election determines whether the country survives, losing becomes psychologically harder.
This is why democratic politicians need to be cautious about existential language even when the stakes genuinely are high.
If every election is described as the last chance to save the nation, what exactly are supporters supposed to believe when their side loses?
Democracy requires tomorrow.
That may be its most important psychological assumption.
You lose today and continue organising because another contest will come.
The constitutional system remains available.
Your political identity survives defeat.
This is the principle I wrote about during the January 6 hearings: democratic systems need to make losing survivable.
Brazil is testing the proposition on an enormous scale.
There is another reason Lula’s victory is politically remarkable. His career itself contains almost every contradiction in modern Brazilian democracy. A former metalworker and trade-union leader, he governed from 2003 to 2010 and left office extraordinarily popular. He was later convicted on corruption charges and imprisoned, preventing him from contesting the 2018 election. Those convictions were subsequently annulled on procedural and jurisdictional grounds, allowing him to return to electoral politics.
Now he has defeated the president whose political rise occurred partly during the collapse of public trust in the Workers’ Party.
That does not mean the corruption controversies surrounding Lula’s political era disappear.
Nor does annulment of convictions mean every criticism of the Workers’ Party was invented.
Brazil’s Operation Car Wash investigations revealed genuine and extensive corruption across political and business networks, even as the conduct and impartiality of parts of the investigation later became deeply controversial.
Again, several things can be true simultaneously.
Corruption can be real.
An anti-corruption investigation can itself become procedurally problematic.
A politician can have serious political baggage and still win a legitimate election.
Democracy does not require voters to choose morally uncomplicated people.
It requires voters to possess enough information and institutional freedom to make the choice themselves.
Lula’s return therefore demonstrates something about democratic rehabilitation. Political defeat, imprisonment or scandal does not necessarily end a political career permanently when institutions allow competition to reopen.
That can feel uncomfortable.
But the alternative is allowing institutions outside elections to decide permanently which political movements citizens are permitted to reconsider.
Courts should determine legal guilt.
Voters determine political forgiveness.
Those are different forms of judgement.
Lula now inherits an economy and political system very different from those he governed during the commodity boom of the 2000s. His earlier administrations benefited from strong global demand for Brazilian exports and used increased revenues partly to expand social programmes and reduce poverty.
The external environment is less forgiving now.
Global growth is weakening, inflation remains a problem internationally, interest rates are high, and Brazil faces substantial fiscal constraints.
This means nostalgia for Lula’s first presidency cannot itself reproduce its economic outcomes.
Political leaders frequently return to office promising to restore a successful previous period.
Economic conditions rarely return with them.
This creates a problem of expectations.
Voters may remember outcomes without remembering the external conditions that helped produce them.
A competent second Lula era therefore requires adaptation rather than restoration.
The Amazon will be another major test. Bolsonaro’s environmental record drew substantial international criticism as deforestation increased during much of his presidency. Lula has promised stronger environmental protection.
That creates a familiar conflict between global public goods and domestic distribution.
The Amazon provides enormous climate and biodiversity benefits to the world.
The economic costs of restricting land use are concentrated locally among Brazilian farmers, miners, landowners and communities.
Foreign governments therefore have an interest in Brazilian conservation policies whose domestic costs they do not fully bear.
This gives Brazil legitimate grounds to demand international financing for preservation.
Climate politics again becomes distributional.
The world wants Brazil to protect a globally important ecosystem.
Brazil can reasonably ask what the world is willing to contribute towards that protection.
Lula’s environmental diplomacy may therefore become part of a broader attempt to restore Brazil’s international influence.
Under Bolsonaro, relations with several Western governments became strained, while his foreign policy often emphasised sovereignty and conservative ideological alignment.
Lula historically preferred a more multilateral strategy, emphasising Brazil’s role among developing countries and institutions such as BRICS while maintaining relationships across ideological blocs.
That approach may become particularly important in the increasingly multipolar environment I discussed after the G7 summit.
Brazil is precisely the kind of middle power capable of refusing simple alignment.
It has strong relationships with the United States and Europe, enormous economic ties with China, membership in BRICS alongside Russia, India, China and South Africa, and ambitions to represent developing-country interests.
Lula’s return may therefore strengthen the politics of strategic autonomy rather than simply move Brazil into a Western camp.
This is another reason the election matters internationally.
Brazil is not merely Latin America’s largest country.
It is one of the states capable of influencing how a more multipolar international order develops.
Yet none of those foreign-policy ambitions will matter if domestic democratic legitimacy deteriorates.
The immediate priority remains the transfer of power.
Road blockades by some Bolsonaro supporters appeared after the result, including demonstrations calling for intervention. Those actions should be treated carefully. Protest against Lula is legitimate. Demanding that the military overturn an election is not ordinary democratic opposition.
The distinction should be enforced consistently.
Citizens can demonstrate.
They can criticise electoral authorities.
They can demand lawful investigations.
They can campaign against the incoming government before it even takes office.
What they cannot democratically demand is that an unelected institution cancel verified votes because they dislike the outcome.
The military’s role is therefore crucial precisely through not acting politically.
Brazil experienced military dictatorship from 1964 to 1985. The armed forces consequently occupy a complicated place in the country’s political memory. Bolsonaro, a former army captain, has frequently praised aspects of the military regime and maintained close relationships with military figures.
That makes institutional neutrality especially important now.
A democratic military demonstrates professionalism by refusing invitations to become an electoral referee.
This is another example of power exercised through restraint.
The armed forces possess enormous coercive capacity.
Their democratic contribution is often choosing not to use it politically.
The next several weeks will therefore tell us much about Brazil’s institutions. Bolsonaro does not need to endorse Lula’s programme or disappear from political life. His movement can become a powerful opposition force. More than 58 million votes give Brazilian conservatism an enormous electoral base.
The healthiest outcome is not the destruction of Bolsonarismo.
It is its institutionalisation.
Compete in Congress.
Win governorships.
Challenge policies in court.
Organise for the next election.
Persuade voters.
That converts an intensely personalised political movement into ordinary democratic opposition.
Lula faces the mirror-image responsibility. Winning by 1.8 percentage points does not give him permission to treat Bolsonaro’s electorate as an obstacle to be overcome rather than citizens to be governed.
A narrow victory should produce humility about mandate.
It should not produce timidity about constitutional authority.
Lula has the right to govern.
He also has strong reasons to govern as though 49.10% exists.
Those ideas are perfectly compatible.
Perhaps this is the most important lesson from Brazil’s election. Democracy does not require consensus before producing a winner. Eventually someone must receive authority even when society is divided almost perfectly in half.
The genius and danger of elections is that they turn tiny numerical differences into enormous institutional consequences.
50.90% receives the presidency.
49.10% does not.
That discontinuity is unavoidable. Somebody has to govern.
The political system survives when the losing 49.10% believes that losing the office does not mean losing its place in the country.
That requires restraint from the losers.
It also requires restraint from the winners.
Brazil has successfully counted the votes.
The harder democratic task begins after the counting stops.
