The comfort of stagnation

Polarization, revealed preference and the absence of a national project

On the eve of the 2026 election, Brazilian voters seem to be saying two things at once. In opinion polls, they declare themselves dissatisfied: in April this year, 68% said the country was heading in the wrong direction.¹ At the ballot box, however, they reaffirm the same old choice, split between two poles: supporters of President Lula and supporters of the Bolsonaro family. This essay is about that contradiction, and it argues that it is not an accident but the key to understanding the country.

Economics has long distinguished between stated preferences and revealed preferences. The first are what people say they want; the second, what their choices actually show. When the two diverge, it is prudent to give more weight to the second. Applied to Brazilian politics, the concept produces an uncomfortable diagnosis: Brazilians say they want change, but their votes reveal a preference for permanence.

The numbers support this reading. A BTG/Nexus survey from August 2026 classifies 25% of the electorate as committed Lula supporters and 30% as committed Bolsonaro supporters.² More than half the country has made up its mind before the campaign even begins. Brazil’s two-round system should, in theory, free voters to back alternatives in the first round. Even so, the vote concentrates: in the Datafolha poll released on September 24, the two leading candidates together held 76% of first-round voting intentions.³

It pays, however, to be exact. The vote does not reveal that voters are satisfied with the country. It reveals that the fear of the other side winning weighs more than the desire for change. The practical result is the same: each pole gives the other its reason to exist, and the equilibrium holds.

One might object that voters stick with the poles for lack of viable alternatives, not out of attachment to permanence. That objection, however, describes the problem rather than refuting it. What it points to is a coordination failure: a candidacy’s viability is not given in advance; it is built precisely by the votes it receives. Voters who wait for an alternative to become viable before supporting it guarantee that it never will. With two rounds, the cost of coordinating in the first one is low, and yet coordination does not happen. In that sense, the absence of options does not refute the revealed preference. It is one of its consequences.

The same pattern repeats outside the ballot box. Over the past decade, Brazil has passed significant reforms under governments of opposite orientations: a constitutional spending cap, pension reform and tax reform. In almost all of them, however, the original impulse was diluted or reversed. The spending cap was abandoned a few years after it was created, and the tax reform came out full of special regimes won by organized sectors. Reform is accepted in principle, as long as its cost falls on someone else.

The results reflect that choice. GDP per capita took eleven years to surpass its 2013 peak, which only happened in 2024.⁴ The investment rate, at 16.8% of GDP in 2025,⁵ remains below the world average of close to 27%, and far from the more than 30% that South Korea sustained for decades.⁶ The high interest rates that hold back that investment, in turn, reflect a public debt that neither side is willing to contain. This is not collapse, but accommodation. The developmental ambition of the twentieth century has given way to a logic of preserving benefits, special regimes and structures won over decades.

Development, however, requires sacrifice. Its costs are immediate and concentrated; its benefits diffuse and in the future. No government can impose it alone: it is a collective decision, one that requires accepting unpleasant consequences in the short term.

A national project, in that sense, does not need to be complicated. It would start with a credible fiscal rule that stabilizes the debt and makes room for lower interest rates; lower rates unlock investment, today the main constraint on growth. It would require revisiting the special regimes and sectoral benefits accumulated over decades, and focusing public effort on what raises productivity: basic education, infrastructure and a simpler business environment. None of this is new. The difficulty has never been knowing what to do, but accepting who pays the bill.

History shows that societies rarely make this decision on their own initiative; it is usually forced by a crisis. Israel stabilized its currency in 1985, with inflation above 400% a year; Poland adopted the Balcerowicz Plan in 1990, after the collapse of its planned economy; Peru and India reformed in the early 1990s, facing hyperinflation and a balance-of-payments crisis, respectively. More recently, Javier Milei’s Argentina began its adjustment after decades of chronic inflation. Even apparent exceptions, such as Paul Volcker’s monetary discipline in the United States or Ireland’s 1987 social pact, responded to situations that had already become unsustainable; the difference was acting before the collapse. The most lasting transformations, unlike those imposed through concentrated power, were the ones society accepted democratically, because it understood their benefit.

Brazil itself knows this path. The Real Plan of 1994 demanded discipline and patience from a population exhausted by inflation, and it worked, because the cost of doing nothing had become visible to everyone. Today that cost exists, but it is diffuse. It does not show up in a monthly index; it shows up in modest growth, low investment and generations that prosper less than they could.

Hence the dilemma. If society does not decide to change while the choice is still in its hands, the decision will likely end up being made by circumstances, under far less favorable conditions. History suggests that is the more common path. It does not have to be ours. In 2026, voters have the chance to align what they declare with what they reveal, and the vote can stop being a choice between sides and become a statement about which country we want to build.

Footnotes

1. Ipsos, What Worries the World, Brazil report, April 2026. ↩

2. BTG Pactual/Nexus, 2026 Election Survey, round of August 24, 2026. ↩

3. Datafolha, presidential voting-intention poll released on September 24, 2026. ↩

4. FGV IBRE, National Accounts Unit, estimates of GDP per capita at constant prices, 2024. ↩

5. IBGE, Quarterly National Accounts, release of March 3, 2026. ↩

6. International Monetary Fund, World Economic Outlook, gross fixed capital formation series. ↩