Town & City

Earlier this month, Brazilians went to the polls in an election billed as the most momentous since democratization in 1985. Far-right president Jair Bolsonaro faced off against former two-term president Luiz Inácio “Lula” da Silva. Though Lula did win the first-round election by more than 5 percentage points, or 6 million votes, it was not enough to clear the 50 percent threshold needed for first-round victory. The opponents will face a polarizing run-off on October 30. 

In Brazil, the national and state-level executives are elected by direct proportional representation for four-year terms. This is also the case with the Senate, comprising eighty-one seats, each serving eight-year terms. The lower-house national legislature, the Chamber of Deputies (comprising 513 seats, each serving four-year terms) is elected via open list proportional representation through state-based lists, with a 2 percent threshold and a rather complicated alliance system between parties. As a result, there has systematically been a disjuncture between the executive and the legislative in the country, and presidents need to cobble together shifting alliances to try and form the temporary majorities they need to legislate. At best this has led to weak governments, and at its worst it has been the breeding ground for semi-legal practices and various corruption scandals over the years.

The results from Brazil’s first round show a sharp turn to the right in the legislature—for which there are no run-offs—as well as in gubernatorial races. In both upper and lower houses, the number of right-wing members is set to reach a historic high. In the Senate, Bolsonaro’s Liberal Party won eight of the twenty-seven Senate seats up for grabs, making it the largest party in the chamber. Other right-wing parties also made substantial gains, such as União Brasil which took five more seats, bringing its total to eleven. Meanwhile, Lula’s Workers’ Party (PT) made only modest inroads, gaining two seats in the Senate to take its total to nine. Other left-wing parties failed to increase their representation. 

A similar pattern can be seen in the lower house. There, too, the right made gains and the leadership will remain in the hands of the Centrão, the amorphous, right-leaning group of Deputies best known for their skill in pork-barrel politics and for selling, at eye-gouging prices, their support to the government of the day. Nevertheless, there were some developments that went against the grain: in particular, there was a significant increase (from seventy-seven to ninety-one) in the number of women deputies elected, and a small rise in the number of self-declared black and brown deputies, from 124 to 135, even if they continue to be woefully under-represented. 

There are two main implications from these results. First, whoever is elected in the run-off will have his work cut out to legislate—all the more so for a Lula presidency. One can expect more of the horse-trading politics that have characterized executive-legislative relations in Brazil for decades. Second, while the executive remains up for grabs, the results demonstrate the strength of the new right. If Bolsonaro’s sweeping victory in 2018 could in part be read as a widespread rejection of established politics, the same cannot be said for 2022. In a deeply polarized country, this novel right-wing politics—centered around a punitive approach to public security, conservative values, and an economic agenda that is unclear beyond its unflinching support to agribusiness—is now a consolidated force, which encompasses but exceeds Bolsonaro himself. Understanding this force and its supporters is perhaps the key challenge over the medium term. 

The standard map

A standard narrative about the Brazilian constituency in recent years is that the poorer north and northeast tends to vote for the Workers’ Party, while the wealthier south(east) throws its support behind the right. Standard visual representations of the Brazilian vote—spanning the first two rounds of 2018, plus the first round of 2022—seem to confirm this reading. The north and northeast of the country are depicted as broad swathes of red, albeit with some exceptions—most prominently the blue colored state of Roraima in the far north. The south, by contrast, is shown to be overwhelmingly colored blue, representing Bolsonaro’s sweeping victory there.

Figure 1: Map of presidential election results in Brazil, 2018 (first and second round) and 2022 

This north-south divide has socio-economic roots, as the north has historically tended to be poorer and less developed. Whilst there are complex reasons for this pattern of uneven geographic development, it can be traced to political and economic dynamics dating back to the nineteenth century. The coffee economy, which kicked-off in the early nineteenth century and, later, industrialization in the twentieth century, were concentrated in the southeast of the country, especially in Brazil’s most populous city, São Paulo. Meanwhile, the state apparatus was housed in Rio de Janeiro, the national capital until 1960. These two mega cities drew internal migrants from other regions of the country and concentrated resources, political power, and economic growth in the southeast. 

Nevertheless, as André Singer has shown, the PT’s popularity in the north and northeast is a relatively recent occurrence. Before it came to power in 2002, the party had drawn its main support from better-educated, higher-income voters in the big cities, largely in the south; after 2006 this changed, and the PT became most popular among lower-income groups in smaller municipalities.¹ This was in large part thanks to the PT’s redistributive policies which lifted the minimum wage and introduced conditional cash transfers like the famous Bolsa Familia. The PT’s policies were squarely aimed at the lowest (and the very highest) end of the income spectrum, rather than expanding the high-paid labor market in such a way that would have attracted the middle class, and so the shift was consolidated. Ongoing corruption scandals—by no means unique to the Workers’ Party, but nevertheless a recurring feature of its administrations—further corroded middle-class support. The PT’s expansion of finance and higher education to previously excluded groups also brought the latter closer to the party—and especially to the figure of Lula, who himself had been a migrant from the northeast to São Paulo. As redistributive and growth-enhancing policies had a greater impact in the poorer, northern parts of the country, a regional pattern gradually consolidated itself from 2006 onwards and has, so far, remained in place.²

If the northeast of the country has been loyal to Lula and the Workers’ Party since then, the north has been a more complicated case since Bolsonaro came to power. The President and his government have ramped up deforestation and expanded agribusiness across the Amazon—which covers most of Brazil’s northwest—and elsewhere. It’s a policy that has attracted global consternation but loyal support among the logging, mining, and farming industries. Deforestation in the Amazon was reduced by roughly two thirds during PT rule, but has since then almost doubled again to cover 13,000 square kilometers in 2021—and continues to accelerate.³ The effects of Bolsonaro’s policies have been fatal, but the PT’s policies towards the region are themselves not without controversy. This is perhaps best encapsulated by the example of PT plans to build the Belo Monte dam in the state of Pará, which was bitterly resisted by affected indigenous and riverine communities. Electorally, this means the state has shifted from supporting the PT in 2010 and 2014 to supporting Bolsonaro in 2018, and it is currently split across its municipalities. As such, the Amazon and the wider north are very much at the crossroads of different pressures.

New terrain

If these standard maps of Brazil reveal a major cleft between north and south, they nonetheless skip over much of the texture and detail of Brazil’s political geography. Most significantly, their mode of representation ignores the differences that occur within regions and within states, as well as distorting population densities. Brazil’s election is won based on the percentage of the total vote rather than state-by-state victories, and analysis that distinguishes between larger and smaller conurbations is necessary. In Brazil, fourteen cities account for 20 percent of the vote—but just one percent of the territory. Similarly, 50 percent of voters are concentrated in just 8 percent of the territory. Looking at particular states, São Paulo accounts for about 3 percent of the country’s area, but 22 percent of the votes. The northern state of Amazonas, by contrast, spreads over 18 percent of the country’s area but yields just 2 percent of the vote.

The cartograms below give a richer picture. Figure 2 maps the same two rounds of the 2018 election, plus the first round from 2022. It reveals a more accurate representation of the electoral map in so far as the mapped area is proportional to votes cast in each election.

Figure 2: Electoral Brazil, cartogram of total votes for presidential candidates in Brazil, 2018 (first and second round) and 2022

The main cities in each state are now clearly visible, with São Paulo in particular standing out. In 2022, it is the large, light pink area in the southeast—marginally won by the Workers’ Party—with a cordon of deeper red municipalities amidst a blueish sea. 

What these finer-grained cartograms highlight is that there are significant intrastate divisions, both in the north and the south. The bellwether state of Minas Gerais is a clear example of this, reflecting its geographical position between the generally pro-Bolsonaro south and center-west and the pro-PT northeast. Capital cities in the north and northeast, made visible through their larger size in the cartograms, also stand out for often having a different color—and hence voting pattern—to the rest of their states. This calls attention to the political geography of the country’s towns and cities.

Unlike in much of the global North in recent years, however, this is not simply a story of the biggest cities—the beneficiaries of globalization and with access to high-quality public services—voting for progressive candidates, and towns—often left behind and “deindustrialized”—backing the right. Brazil bucks this trend. Most large cities supported Bolsonaro in 2018, including a few in the PT-leaning northeast, both in the first and second rounds. Results for the first round this year were more mixed, and as Laura Carvalho and Pedro Abramovay have indicated, the PT did have a better showing in large cities this time.⁴ Nevertheless, with Bolsonaro holding on to most large cities of the south and Lula holding those of the north, the picture remains complicated. “Bolsonarismo,” then, is not a preserve of the hinterlands, nor is “Petismo” a metropolitan trend. Rather, large cities in Brazil are sites of polarization and tight margins, whilst Bolsonarismo is predominantly a medium-sized city phenomenon and Petismo is strongest in small towns. This plays out across the country, as the  following cartograms show.

Reading the first round

The Workers’ Party gained ground in the first-round election, albeit not to the extent that the polls anticipated. It has advanced in virtually all of the south, southeast and center-west, often to the tune of 30 percentage points (as in the city of São Paulo). At the same time, Bolsonaro has made substantial inroads in segments of the north and northeast. This is highlighted in Figure 3, which, comparing the first round of 2022 to the second round of 2018, shows changes in the vote gap.

Figure 3: How 2022 is different: cartogram of changes to the vote gap between the PT and Bolsonaro, 2022 compared to second round of 2018

The broad pattern that emerges is that the PT has gained ground where it was weakest and ceded terrain where it was strongest. Regionally, of the 10 million additional votes the PT received (compared to the second round of 2018), 6 million came from the southeast and about 2 million from the south and center-west combined. As a result, the party has narrowed its losing gaps by between 18 and 26 percentage points in regions it is currently losing, and ceded some ground in the northeast, where it is winning.

In general, the PT advanced in cities, particularly large ones, and stalled in towns. Across the country, Lula’s party moved ahead 3 percentage points in small towns, where it is currently at its strongest and leads by 24 points. It has trimmed its losses by 19 points in medium cities, where it holds its worst results, trailing by 9 points. The swing was even stronger in large cities, though, where the party went from a negative gap of 21 points in 2018 to a slightly positive lead of 2 points in the latest round. This pattern holds across the country. The PT’s losses (or rather the narrowing of their lead) have been concentrated almost exclusively in towns in the north and northeast. In the other three regions, the party has made similar inroads in large cities, hovering around a gain of 25 percentage points, but advanced much less in the towns and medium-sized cities of the south and center-west.

Further research is needed to reveal why these shifts have occurred, but three hypotheses offer themselves. First, income-supporting benefits were recently hiked, and as they have a greater impact in small cities this might have drawn voters to Bolsonaro. Second, the PT might have taken for granted its regional stronghold of the northeast as it sought votes in the southeast, and Bolsonaro’s support for agribusiness may well have made him more resilient in the center-west, the country’s main agricultural frontier. Third, the pandemic. Under a denialist (mis)management of the disease, ripe with Bolsonaro militating against the use of masks, casting unfounded doubts on the vaccines, and even proffering scoffing remarks towards people’s suffering, the pandemic has claimed nearly 700,000 lives in Brazil and led to some of the highest mortality rates in the world, at about 320 per 100,000 inhabitants. Arguably this helped tilt the overall balance of votes towards the PT, but perhaps particularly so in large cities that might have borne the greatest brunt of the disease. 

Despite these shifts in 2022, there is an enduring divide between town and city, explored through the cartograms below. They demonstrate three salient points:

1. Bolsonarismo is fundamentally a medium-sized city phenomenon

The maps in figure 4 below plot the proportion of Bolsonaro’s votes: the darker the blue, the higher the vote share. Immediately apparent is the lighter tones for the northeast of the country; the area constitutes 28 percent of the electorate but just 16 percent of Bolsonaro’s vote. Large cities also have a lighter hue, reflecting their polarization. In the southeast, for example, the three major cities—São Paulo, Rio de Janeiro to its east, and the capital of Minas Gerais, Belo Horizonte, slightly to the north of the two—have lighter hues than their surroundings.

Figure 4: Bolsonaro’s Brazil, cartogram of votes for Bolsonaro in Brazil, 2018 (first and second round) and 2022

The fine lines on these cartograms have the value of distinguishing between small, medium and large conurbations. Focusing on the darkest shades of blue, it is apparent that Bolsonaro’s support is greatest in medium-sized cities, which represent about 20 percent of the total electorate. This has been the case for all three elections depicted above; in every region but the northeast, Bolsonaro’s share of the vote has exceeded the PT’s by double digits, often with a gap of 20 or 30 percentage points. 

2. The PT’s stronghold is small towns

Figure 5 maps the PT’s share of the vote since 2018. The deep red of the northeast represents Lula’s stronghold there, as it represents between 38 and 47 percent of total votes for the party. Nevertheless, the state of São Paulo has been the largest absolute source of votes for the PT, both in the second round of 2018 and in the first of 2022—despite Bolsonaro ultimately winning there. Of the 57.3 million votes Lula received in 2022, 10.5 million came from the state of São Paulo, followed by 5.9 million from Bahia in the northeast, and 5.8 million from Minas Gerais, in the southeast but bordering the Bahia. In the city of São Paulo itself, the PT went from trailing by twenty-one points in 2018 to leading by nine in 2022, with Lula ultimately receiving 3.3 million votes there, compared to Bolsonaro’s 2.6 million.

Figure 5: The PT’s Brazil, cartogram of votes for PT presidential candidates in Brazil, 2018 (first and second round) and 2022

The novel element here is that these maps show that it is neither big nor medium-sized cities that constitute the PT’s most reliable base, but small towns. Indeed, every time the PT has gone to an election against Bolsonaro, these small towns across the country have been its electoral bastion, where it has held the highest leads or sustained the smallest losses. In the first round of 2022, this meant a total of excess votes upwards of 50 percentage points in the northeast and a minor lead of one point in the small towns of the southeast.

3. Major cities are polarized and occasionally flip the state-wide vote

If large towns are principally the domain of Bolsonaro and smaller towns overall vote for Lula, what of the biggest cities in Brazil? 35 percent of the electorate live in big cities, the largest of which by far is São Paulo, followed by Rio de Janeiro, Brasilia, and then Salvador in the state of Bahia and Belo Horizonte in Minas Gerais. 

What is noticeable from these maps is that these cities sometimes vote against the rest of the state that surrounds them. This can be seen in Figure 2 above, where in all three rounds, small islands of blue can be identified amid seas of red, or vice-versa. The state of Amazonas in the Northwest of the country is striking: a large, strongly pro-Bolsonaro city—the capital of Manaus—stands out in an otherwise consistent patch of red. The intra-state division revealed here may well be due to the different policies of the PT and Bolsonaro towards the Amazon, which sprawls across most of northwest Brazil. It is nonetheless surprising that Manaus, the largest city in the Amazon, favored Bolsonaro despite having twice suffered the collapse of its healthcare system during the pandemic when hospitals were above capacity, oxygen supplies ran out, and mass graves were required to bury the dead. 

This means that in states that generally vote for Bolsonaro, medium cities often exacerbate region-wide tendencies and small towns attenuate them; in pro-PT areas, by contrast, the reverse is true. When the voting difference between small towns and medium cities is large enough to make them straddle the political divide, large cities might fall on either side—eventually breaking the broader regional or state-wide tendency. Which is to say, the town and city divide in Brazil is not simply a matter of metropolises versus villages; rather, it works in and through regional dynamics and the urban network.

Reading the regional divide

On the eve of the second-round election that will decide the country’s fate, it is worth asking: what accounts for Brazil’s recent voting patterns and why does the country tend to buck the global trend with regard to electoral politics, insofar as its biggest cities can’t be relied upon to vote progressively? 

An obvious factor contributing to this polarization is the extreme inequality of Brazilian cities. Where small towns in Brazil are certainly marked by deprivation, the very rich and the very poor there less commonly live cheek by jowl. In São Paulo, by contrast, lavish apartment blocks abut shantytowns; in Rio, crumbling favelas with little access to formal services overlook some of the wealthiest neighborhoods. Inequality is by no means unique to Brazil, but it is unlike European cities, for example, where public services are more evenly distributed among urban areas. Cities in the US, by contrast, are more commonly subject to the postcode lottery experienced in Brazil, but sanitation and basic infrastructure are nevertheless generally fixtures throughout the country. This cannot be said for the favelas of shantytowns of Brazil, even in its largest cities, where paved roads, modern sewage, and formal access to electricity remain an aspiration rather than a reality. 

Large cities in Brazil also benefit from substantial migratory flows but, unlike in the global North, migration is mostly internal, comprising Brazilians from small towns moving across the country. Concerns over security add further fuel to the urban conflict at the heart of these spaces. In this highly contested social space, what emerges is not the open cosmopolitanism that often characterizes the biggest cities of the global North, but rather an entrenched polarization gridlocking individuals into a bitter struggle over deeply constrained processes of social mobility. 

Overhauling this situation is as urgent as it is likely to be difficult. Bolsonaro’s platform is divisive and exclusionary to its core, so should he hold onto power, there is little hope from that quarter. Lula’s campaign has meanwhile focussed on reviving the conciliatory tone of his former government, but has been lacking in concrete policy proposals. Should he win, he will face a right-wing legislature, no commodities boom to relax spending constraints, and tougher domestic economic conditions with high inflation and public and private debt levels. Bringing towns and cities together in an inclusive development strategy, under these circumstances, will be no mean feat.