On Sunday, Peru will hold a presidential runoff to choose who will govern the country for the next five years. This is the fourth democratic presidential election since Alberto Fujimori’s authoritarian regime ended in 2000, when he fled the country to escape prosecution for charges of corruption, abuse of power and human rights abuses.

This year, however, Fujimorismo may return to power. Keiko Fujimori, daughter of the imprisoned former president, won the first round of presidential voting on April 10 with just under 40 percent of the vote. The second-place finisher, Pedro Pablo Kuczynski, took only 21 percent — inching ahead of Verónika Mendoza, a left-wing congresswoman who speaks the indigenous language Quechua and took 18.8 percent of the vote.

Fujimori is a populist who gains from some parts of her father’s reputation for rescuing Peru’s economy and defeating the terrorists of the Shining Path. But she’s simultaneously feared for her association with the rest of his reputation, including accusations that he dissolved democratic checks on his power, deployed death squads, took bribes and rigged votes.

Kuczynski, by contrast, is a former World Bank economist and Wall Street trader with a highly cosmopolitan and managerial background. But he is also considered a poor campaigner with few political skills. His proposed policies don’t differ much from those of Fujimori, although his style does. Going into Sunday’s final vote, Fujimori and Kuczynski are statistically tied, according to the latest polls. Both Fujimori and PPK, as Kuczynski is known in Peru, will be struggling to win by picking up volatile southern voters and specific voting blocs.

Voters are choosing a successor to President Ollanta Humala, a left-wing nationalist who squeaked past Fujimori to win in 2011 and who governed on a center-right platform over his five-year term in office. Under Peruvian law, Humala cannot seek reelection until the 2021 presidential race.

This year’s elections have been highly controversial because two major contenders — César Acuña and Julio Guzmán — were excluded from the race. That’s especially significant because Guzmán was seen as Fujimori’s primary rival. Both were removed by Jurado Nacional de Elecciones, the national electoral board, for minor infractions of electoral rules, in decisions that local pundits believe were overzealous interpretations of electoral law. Even the secretary general of the Organization of American States, Luis Almagro, publicly dubbed the elections “semi-democratic” on April 1.

Let’s look at what happened during the first round of voting.

First, who are Fujimori’s voters?

Fujimori was the front-runner during most of the campaign, so the fact that she won the first round of voting wasn’t a surprise. But the margin by which she won, almost 40 percent of all valid votes, beat even the most optimistic poll’s estimates of under 38 percent. What’s more, her support is the most evenly distributed nationwide of the top three candidates, as you can see in the map below. She was the candidate with the most votes in 113 out of the 196 provinces and had relatively homogenous results territorially, with the exception of some of the poor southern provinces, where Mendoza’s electorate was concentrated. Fujimori also did well in the most populated provinces, including Lima, Callao, Trujillo, Chiclayo, Huancayo and Ica.

We measured this nationalization of the vote with the standardized Party Nationalization Score (sPNS) proposed by Daniel Bochsler, taking into account the relative weight and number of territorial unit. This index ranges from zero to 1, with 1 representing the most nationalized support. Keiko’s sPNS reached 0.83, a fairly high score.

How did she do this, considering she had a much narrower electoral base in 2011? First, she has presented herself as a moderate, attempting to distance her eventual government from her father’s authoritarian regime and legacy. Second, and significantly, during the past few years Keiko Fujimori invested in building a party organization (as documented by Adriana Urrutia or Paolo Sosa). She looked for new supporters and strengthened existing organizational structures.

At the same time, Fujimori managed to negotiate alliances with regional political leaders. In the first-round voting itself, she probably benefited from Acuña’s disallowed candidacy and the gradual weakening of the traditional support for Alianza Popular Revolucionaria Americana, the party of the former president Alan García (2006 to 2011), in the north. As a result, Fujimori expanded her electoral base, as you can see.

Third, Fujimori’s supporters come from almost every socioeconomic level in Peru. To be technical, at the aggregate level, there is a rather weak positive correlation between the Human Development Index (HDI multiplied by 100) and percentage of votes in each province. In other words, Fujimori’s voters came from both the more and less well-off provinces, as did her father’s in his time. That support is especially strong from poor rural and possibly also lower-class urban voters, often acquired using clientelistic strategies.

Second, who are PPK’s supporters?

Kuczynski and Mendoza ran a close race for second place — although PPK, in the end, pulled ahead more than polls predicted, with 21 percent of the valid votes compared with Mendoza’s 18.8 percent. PPK’s support is far more regionalized and less uniform territorially than Fujimori’s. His votes came especially from the southern of Arequipa region (mainly the capital), northern Loreto, central Junín, Huánaco, and districts of Lima Province (as you can see in the map 2 below). Thus, his sPNS index reaches only 0.72.

Kuczynski’s support comes mainly from urban voters and the more developed provinces. His voters back the country’s current economic model, which is based on foreign investment, privatization, and a limited state role. They praise his experience and ability in managing the economy.

PPK’s voters also reject the authoritarian legacy of Fujimori’s father. They fear a return of corruption, lack of transparency, organized crime, drug trafficking and possibly even Fujimori’s “mafia” and illegal business activities related to the state.

To be technical once again, we can see a strong correlation between the Human Development Index (HDI) by province and the pro-Kuczynski vote, as you can see in the figure below. Pearson’s r measuring the association between both indicators reaches 0.633. PPK fares well in provinces like Lima, Arequipa, Callao, Mariscal Nieto and Yauli (all with HDI higher than 0.55) and the richest districts of Lima Capital. Still, his support is rather fragmented and limited; only in a few places did he get more than 30 percent of the valid vote.

Who were Mendoza’s voters?

Kuczynski’s main competition for the second round, the leftist Verónika Mendoza, also had fragmented support, though with different (and to a certain degree inverted) patterns of support than PKK’s (Map 3). Her vote was highly concentrated in the southern departments of Apurímac, Huancavelica, Ayacucho, Cuzco, Puno, Arequipa (though to a lesser extent, due to the presence of PPK), and Tacna, where she won more than 50 percent of the valid vote.

Mendoza also won voters from Ancash and to a lesser degree Huánaco, although their provinces make up a smaller part of the national electorate. This territorially marked support also gives her a lower nationalization score, with a sPNS of only 0.66.

The southern provinces have been most left behind by the recent economic boom. With persistently high levels of poverty, the region has been mobilizing against the transnational corporations that control the local mining industry.

This socioeconomic position shows up in the correlation between HDI by province and Mendoza’s vote, with a negative relationship (Pearson’s r of -0.356). Her voters represent mainly the so called “anti-system vote,” defined rather in economic anti-authoritarian terms. These were the regions that helped greatly Ollanta Humala into the presidency in 2011, and had been his principal supporters in 2006.

What does all this mean for the final vote on Sunday?

On Sunday, the candidate with the most votes will become the new president of Peru. There is still an important sector of mainly left-wing voters whose absence at the voting booths or spoiling of the ballot could potentially favor Fujimori. She did well across the country, although she struggles somewhat for support in the south.

PKK needs more support from the north (especially in Cajamarca to obtain Gregorio Santos’s votes from the first round), as he has been expected to win most of Mendoza’s voters in the South. These two macro-regions have the largest number of available votes, or votes that neither Fujimori nor PKK won in the first round, as you can see in the map 4 below. These available votes’ socioeconomic profile, as you can see in the figure below, has a correlation with HDI by province that’s negative with Pearson’s r of -0.435. In other words, these voters are from poorer regions, which may make them less likely to support PPK. Mendoza’s supporters in the South have for several election cycles been explicitly critical of the economic model. While southern voters are considered essentially anti-Fujimori and expected to treat PPK as the lesser evil, that can’t be taken for granted.

With a very close result expected, every second continues to count in this campaign, and any detail could be decisive. PPK has been strongly criticized for taking it for granted that he would win anti-Fujimori support, without really working for it, and not searching for support and alliances with other political leaders. To win, he’ll need to bring in lower socioeconomic strata votes than he’s picked off before, and to gain voters outside Lima and Arequipa, particularly in the North. To do that, many believe he has to confront Fujimori head on, which he has not done so far.

However, PPK has gotten some significant endorsements, including those of Acuña and Acción Popular, the party of whose leader Alfredo Barnechea finished fourth in the first round (but who personally did not express support to Kuczynski); from many different civic movements, including the most important like “No a Keiko;” and more or less explicitly from his former rival Mendoza herself.

PPK may be calculating that if he wins, an alliance with the Fujimori majority in Congress would be convenient and probably necessary for governing the country. PPK’s party (Peruanos por el Kambio) has only 18 deputies in Congress; if he wins, he will have to work with a Congress where Fujimori’s party (Fuerza Popular) controls 73 of the 130 members.

Fujimori’s run-off campaign has been criticized for distancing itself from some former promises, including recognizing same-sex civil unions and decriminalizing abortion for rape victims. Moreover, the attempt to negotiate with informal transport sectors or with illegal miners to formalize their work and allow them to continue extracting from the Amazon region with negative environmental impact brought some backlash against her candidacy.

Her image has also been damaged by accusations that her party’s general secretary (and a member of Congress), Joaquín Ramírez, laundered money to help fund her run for president.

Despite the scandals and criticism, Fujimori has campaigned steadily, reaching out incessantly to the regions and to diverse sectors of society she needs to win. Before the runoff, she campaigned strategically in the southern provinces especially in the districts where she did not win in April. If she wins on Sunday, it will be because of her nationwide electoral strategy, her backing by local leaders and activists and sectorial alliances, and her presence and popularity among the poor.

Tomas Dosek is a PhD student in political science at the Instituto de Ciencia Política, Pontificia Universidad Católica de Chile.

Maritza Paredes is assistant professor at the Departamento de Ciencias Sociales, Pontificia Universidad Católica del Perú.