Are Open or Closed Lists Better for Women? Comparing Lima and the Provinces in Peru

Gregory D. Schmidt

University of Texas at El Paso

[email protected]

Abstract. This article reviews the longstanding debate in the literature on electoral systems regarding the relative merits of open and closed lists for the election of women, describes the very different systems used in congressional and municipal elections in Peru, and examines gendered disparities between Lima and the provinces. A series of empirical comparisons demonstrates that the type of list is not of great importance in the capital, but that female candidates fare better under closed lists in the provinces, where the socioeconomic context and political culture are less supportive of women.

Keywords: Women in politics; proportional representation; electoral system; elections; Peru.

Introduction

Peru provides an exceptional context for assessing the impact of ballot structure for the election of women for three major reasons. First, a variant of open-list proportional representation (PR) and a closed-list majoritarian system have been used contemporaneously to choose candidates for different levels of government over multiple electoral cycles. Second, comparing the impacts of different electoral rules within the common legal and institutional context of the same non-federal country largely controls for possible intervening variables. Third, Peru has a relatively high degree of socioeconomic inequality and broad cultural diversity, permitting comparisons of electoral rules in very different contexts within the same country.

The initial section of this article1 reviews the debate over the consequences of ballot structure for the election of women. Of particular relevance is the trailblazing work of Valdini (2013), who argues that the impact of ballot structure may vary with cultural attitudes regarding the roles of women in politics. The article then introduces an important intervening variable (district magnitude), provides background on open-list congressional and closed-list municipal elections in Peru, and highlights gendered disparities between Lima and the provinces. Subsequent sections assess the impact of ballot structure for the election of women in Lima and the provinces by (1) examining congressional elections held in a single national district in 1995 and 2000; (2) comparing the results of congressional elections carried out in departmental (regional) districts since 2001 and post-quota elections for district-level municipal councils; and (3) presenting counterfactual comparisons for congressional elections since 1978. A brief conclusion follows.

Before proceeding, it should be noted that the patterns outlined in this article have been remarkably consistent over time, despite Peru’s weak and volatile party system. After redemocratization in 1980, a tenuous party system began to emerge during the following decade, but it was undermined by severe economic decline and increasing political violence stemming from the Shining Path’s insurgency. The election of a political outsider, Alberto Fujimori, in 1990 and an authoritarian turn in 1992 accelerated the demise of the traditional parties (Tanaka 1998). Emulating Fujimori, since the 1990s various presidential contenders have created their own personalistic vehicles for winning elections, generating extreme fluidity in the party system (Levitsky 2018). Post-Fujimori Peru has been called a “democracy without parties” in which elections are contested by “anti-candidates” and “coalitions of independents” (Tanaka 2005; Meléndez 2011; Zavaleta 2014). The country has the most volatile legislative elections, as well as the greatest ideological volatility within parties, in Latin America; it ranks second in regard to volatility in presidential elections and third with respect to instability of membership of the party system (Mainwaring 2018: 45-47, 56).

At the time of this writing in May 2019, as many as 24 parties were officially registered at the national level, despite some attrition following the 2016 election (JNE 2019). In the last regional and municipal elections held in 2018, 136 parties and movements won offices, despite the use of majoritarian electoral rules.2 Discussion of specific political parties in Peru is beyond the scope of this article.

1.Literature Review: Ballot Structure and the Success of Female Candidates

The literature on electoral systems has seen a spirited debate on the implications of ballot structure for the success of female candidates. The discussion is sometimes framed in very general terms (candidate-centred vs. party-centred) across different types of electoral systems. At other times, the relative success of women running for office under closed and open variants of PR is a bone of contention. Under closed-list PR, voters choose only among alternative lists, not among individual candidates, and the seats won by each list are filled by the candidates in the order in which they appear. Under open-list PR, voters are permitted or required to choose candidates. Seats are allocated among the lists according to their respective shares of the vote, but ballots for individual candidates determine who fills these seats.

Several pioneering studies maintained that women are more likely to be elected when voters choose individual candidates under open or flexible3 formats (Rule and Shugart 1995; Shugart 1994; Taagepera 1994), but since the late 1990s scholars have generally viewed closed lists as more female-friendly (e.g. Ballington 2005: 116; Farrell 2011: 164; Htun 2005; Htun and Jones 2002: 167; Jones 1998; Jones and Navia 1999; Norris 2004: 197). This trend has coincided with the adoption of gender quotas, placement mandates,4 and—more recently—parity with alternating candidacies by gender, especially in Latin America. Whereas gender quotas increase the number of female candidates under both closed and open lists, placement mandates and measures to ensure parity are designed for closed lists; indeed, they make little sense under open lists. However, it is not at all clear that closed lists without placement mandates or parity measures are better for women than open lists.

Scholars also view closed lists more favorably because they are the most common format in PR systems, which are more female-friendly than first-past-the-post and other non-proportional systems. Indeed, “the literature implicitly has equated PR with closed lists” (Shugart 2005: 38; emphasis in the original). In contrast, open and flexible formats vary widely, have been understudied, and are poorly understood, even by many specialists on electoral systems (Shugart 2005: 39-44).

Some scholars have suggested that the impact of ballot structure may vary, or even not matter at all. According to Matland, “[t]he crucial question is whether it is easier to convince voters to actively vote for women candidates, or to convince party gatekeepers that including more women on the party lists in prominent positions is both fair and, more importantly, strategically wise” (2005: 104). Female candidates are more likely to be successful under open lists if the former is true, but closed lists are more likely to benefit women running for office if the latter holds. Matland believes that the comparative advantages of open and closed lists may vary from country to country. In a study of national-level systems of list PR around the world, the present author found no significant relationship between ballot structure and the election of women (Schmidt 2009).

In an innovative study, Valdini hypothesizes that the “personal vote” for specific candidates “has a conditional effect that depends on the level of bias against female leaders in a society. In certain cultural environments, the personal vote has no effect on the success of female candidates. In others, however, it has a powerful negative effect on women’s legislative representation” (2013: 80). Her analysis of public opinion data from 23 democracies provides important support for this hypothesis. However, she does not gauge the opinions of party gatekeepers, who play a decisive role in drawing up closed lists and who may be more or less biased against women than the public at large. Nevertheless, Valdini’s trailblazing work is especially relevant for this article. If the impact of ballot structure varies by culture, then different outcomes are likely not only across disparate national political cultures, but also within countries where cultural attitudes regarding female leadership diverge in significant ways.

Peru provides an ideal context for testing Valdini’s ideas within a single country, using a quasi-experimental design. As mentioned in the introduction, Peru has long used a variant of open-list PR and a closed-list majoritarian system for different types of elections. On the one hand, Peru is highly centralized with a common institutional and legal context that minimizes the impacts of such intervening variables as candidates’ access to resources and career patterns. On the other hand, the country has a high degree of socioeconomic inequality—like most of Latin America—and cultural diversity is especially rich in Peru and the other Andean countries, where the Incas and other sophisticated indigenous civilizations once flourished. The contrast between the relatively affluent and progressive capital city of Lima and the provinces—which are generally poorer and more traditional—is particularly striking.

If Valdini is correct, female candidates should fare better under open lists in Lima than in the provinces (Hypothesis I), and disparities in the election of women under alternative ballot structures (open vs. closed lists) should be greater in the provinces than in the capital (Hypothesis II). This article presents empirical evidence that strongly supports both hypotheses.

2.District Magnitude

District magnitude refers to the number of seats or offices filled in an electoral district5 in a given election. It has been a key variable in electoral studies since at least the seminal book by Rae (1967) more than a half century ago. Higher district magnitudes in list PR systems are thought to facilitate the election of women, who are assumed to occupy lower positions on closed lists or to be less competitive candidates on open lists. Thus, ceteris paribus, increasing the number of seats in an electoral district (i.e. higher district magnitude) should boost the chances of women winning office. However, the evidence that supports this argument comes overwhelmingly, if not exclusively, from closed-list systems. Female candidates may not benefit from higher district magnitudes in open-list competition, such as congressional elections in Peru. Moreover, as discussed below, this variable has a different logic under majoritarian lists, such as the system used in Peruvian municipal elections.

In Peru and other Latin American countries, high district magnitudes tend to be found in urban areas, such as Lima, which also have more favorable socioeconomic contexts for the election of women. Thus, district magnitude may provide an alternative explanation where women are elected at relatively high rates from areas that are culturally more supportive of female leaders.

A closely related concept is party magnitude—the number of members in a party’s delegation from an electoral district who win office in the same election. Higher district magnitudes are generally associated with larger party delegations, but party magnitude also depends on other factors, especially the distribution of the vote and rules for allocating seats among parties. Matland (1993) argues that party magnitude is even more directly associated with the success of female candidates than district magnitude. Once again, women are assumed to occupy lower positions on closed lists or to be less competitive under open lists.

3.Open-List PR in Congressional Elections

All Peruvian elections are conducted under virtual universal adult suffrage. Voting is obligatory for citizens who are at least 18 years old and younger than 70. Parties may nominate lists of candidates or form alliances to run joint lists in national, regional, and local elections. A 25% gender quota was adopted in 1997, and was raised to 30% in 2000.

Congressional elections are held every five years, concurrently with the first round of presidential elections, using a system of party-list PR with the D’Hondt rule, adopted in 1962. Since 1985 Peru has used a variant of open-list PR called the “double optional preferential vote” in legislative elections. Under this format, citizens vote for a list by marking the symbol of a party or electoral alliance. They then have the option of voting for one or two candidates on the list of their choice by entering the respective number(s) of the candidate(s) in boxes beside the symbol. Lists are awarded seats in proportion to their shares of the valid vote, but preferential votes (rather than list order) determine which candidates fill those seats. A precursor of this system with a single preferential vote had been used to elect a Constituent Assembly (CA) in 1978. Closed lists were used only in 1980.

Notwithstanding this basic continuity in open-list PR, legislative elections have been subject to major variations, due to changes in the structure of Congress and the electoral districts used to elect lists. During 1980–1992, Peru had a bicameral Congress with a 180-member Chamber of Deputies and a 60-member Senate. The Chamber of Deputies was elected by departments, with a special 40-member district for metropolitan Lima. The average magnitude of the other districts was only 5.7. In contrast, the Senate was elected by way of a single national district, which had first been used to elect the 100-member CA in 1978.

After Alberto Fujimori dissolved Congress via the 1992 presidential coup, an 80-member unicameral Democratic Constituent Congress (CCD) was elected through a single national district to serve both as the national legislature and a constitutional assembly. The 1993 Constitution, still in effect, established a unicameral Congress, initially with 120 members. Under Fujimori, the single national district continued to be used in the 1995 and 2000 congressional elections.

Since the collapse of the Fujimori regime in late 2000, the unicameral Congress has been elected by departments, now also called regions. In 2001 and 2006, the Department (Region) of Lima, which also includes Peruvians abroad, had 35 seats in the unicameral Congress, while the average magnitude of the remaining districts was a very low 3.4. A subsequent constitutional amendment modestly increased the size of Congress from 120 to 130 members and established a special electoral district for metropolitan Lima and Peruvians abroad. Beginning in 2011, this special district has had 36 seats in the slightly larger Congress, with a slightly higher 3.8 average district magnitude for the rest of the country.

4.Closed-List Majoritarian Elections for Municipal Governments

Municipal elections have been held every four years since 1998, and in conjunction with regional elections since 2002. There are two tiers of municipal government in Peru. Voters elect mayors and councilors in each of the country’s 196 provinces. These officials govern the province as a whole as well as its capital district, but voters elsewhere in the province also elect mayors and councils for their respective districts. There were 1,678 municipal governments at the district level in 2018, when local elections were last held.6 Voters cast separate ballots for each tier of municipal government, but the votes for executive and legislative offices are fused at each level. In other words, all voters cast one ballot for a single slate of candidates for mayor and councilors at the provincial level, and most also cast another ballot for a single slate of candidates for mayor and councilors at the district level.

A standard formula determines the number of municipal councilors, according to population. Over 90% of district councils have only five members, but a few have more: up to fifteen in Lima, thirteen in Callao, and eleven elsewhere. The 42 district councils in Lima have almost twice as many members, on average, as those in the rest of Peru (10.05 vs. 5.22 in 2018).

Municipal elections are closed list at both the provincial and district levels; thus, the council seats won by a party or electoral alliance are filled by candidates in the order that they appear on its list. Moreover, elections for both tiers of municipal government have been conducted under a majoritarian format since 1983. If a list wins a majority of the valid vote, council seats are allocated by D’Hondt PR, but this is seldom the case, given very high levels of party fragmentation. However, a list that wins even a small plurality receives “half plus one” of the seats, with the remainder allocated by D’Hondt to other lists that meet a 5% threshold (Law 23,671). Since the passage of the statute on municipal elections (Law 26,864) in 1997, a maximalist interpretation of the “half plus one” rule has prevailed: four of five seats, five of seven seats, six of nine seats, seven of eleven seats, and so on. This majoritarian formula produces higher party magnitudes than would be the case under PR, but it should be recalled that municipal elections are held in low magnitude districts outside of Lima and a few major cities. Thus, majoritarian municipal elections in Peru produce party magnitudes that are similar to those found elsewhere in closed-list PR systems with moderate district magnitudes.

5.Gendered Disparities between Lima and the Provinces

In 2017 Peru had a Human Development Index (HDI) score of .750—slightly less than the overall score of .758 for the Latin American and Caribbean (LAC) region—placing it 89th among 189 ranked countries. When HDI scores are adjusted for socioeconomic inequality, Peru falls slightly, to 91st place, but its adjusted score (.606) is somewhat higher than the adjusted LAC score (.593) because other major countries in the region have even more acute disparities. Moreover, Peru ranks 83rd in the world in terms of gender equity, with a score that is about 5% better than that for the LAC region.7 Thus, Peru is a rather typical Latin American country that has slightly less socioeconomic and gender inequality than the regional average.

These aggregate indicators help place Peru in global and regional perspective, but they mask glaring internal disparities, especially between the capital city of Lima—which accounted for 29.2 % of the population in the 2017 census—and other parts of the country. Women have many more opportunities in the capital, which has a markedly higher level of socioeconomic development and lower levels of gender inequality and discrimination than other areas of Peru. Although gendered socioeconomic contexts vary considerably across different regions and within the sprawling capital, comparisons between Lima and the rest of the country make it possible to use consistent socioeconomic, survey, and electoral data over a long period of time. Moreover, any differences in the mean indicators for the capital and the rest of Peru understate disparities between Lima and traditional areas. This is especially true given that the adjacent port city of Callao—which is very similar to Lima—is included as part of the rest of Peru in most of the examples subsequently cited in this section.8

Women in Lima are 12% more likely to use modern methods of contraception than those living elsewhere in the country and 24 % more likely to do so than women in rural areas. On average, limeñas (women from Lima) give birth to their first child at the age of 24.5, compared to 21.5 for women in other parts of Peru and 20.1 for those in rural areas. Doctors deliver 84.5% of the babies born in Lima, but only 59.2% of those born in the rest of the country and only 44.3% in rural areas. During 2015–2020, the life expectancy for women in the Department of Lima is 80.8 years, compared to 76.4 years elsewhere in Peru.

Disparities by gender and region are particularly pronounced with regard to education. In Lima only 2.9% of women and 1.0% of men who are at least 15 years old are illiterate, compared to 11.1% of women and 3.8% of men in the rest of Peru and 23.5% of women and 7.4% of men in rural areas. Among Peruvians at least 25 years old, 81.4% of women and 89.4% of men in Lima have completed secondary education, in contrast to 55.8% of women and 67.8% of men elsewhere. Dividing the respective figures for women by those for men produces a female/male ratio for secondary education of .91 in Lima and .82 in other areas of the country. In this and subsequent examples, higher ratios indicate greater progress for women. Attentive readers will notice that the female/male ratios are consistently highest in Lima or Lima-Callao.

Younger women are rapidly closing the gender gap in education, especially in Lima and Callao. In 2017, 78.8% of women but only 56.9% of men in Lima-Callao completed secondary school at the expected age of 16 (female/male ratio = 1.38), compared to 56.1% of women and 48.2% of men in other parts of Peru (ratio = 1.16), and 39.6% of women and 37% of men in rural areas (ratio = 1.07). Among Peruvians aged 17–24 in 2017, 36.9% of women and 32.2% of men in Lima-Callao were enrolled in post-secondary education (ratio = 1.15), in contrast to 30.6% of women and 27% of men elsewhere in the country (ratio = 1.13), and only 16.3% of women and 15.7% of men in rural areas (ratio = 1.04). Whereas 66.1% of women and 72.7% of men use the internet in Lima-Callao (ratio = .91), 35.9% of women and 41.9% of men are users in the rest of Peru (ratio= .86), with only 12.5% of women and 17.9% of men surfing the web in rural areas (ratio=.70).

Women employed in Lima earn 73% more than working women in the rest of Peru and more than three and a half times as much as women laboring in rural areas. Moreover, female earnings are 71.6% of male earnings in Lima-Callao, but only 65.9% of male earnings elsewhere. Compared to other women in Peru, limeñas are almost twice as likely to be covered by the pension system and 23% less likely to be employed in the informal sector. Single women in the capital are 20% less likely to lack their own income than those in other areas of the country. Data from 2010 show that women in Lima-Callao spent an average of 43 hours and 40 minutes each week on remunerated jobs and 35 hours and 44 minutes on uncompensated domestic labor. In contrast, women in other parts of Peru devoted 41 hours and 17 minutes each week to uncompensated domestic labor and 32 hours and 58 minutes to remunerated employment.

In recent years, Peruvian and international media have given prominent coverage to incidents of sexual harassment, physical abuse of women, and even femicide in Lima (e.g. Navarrete 2014 and Krumholtz 2018). These sorts of incidents in the capital are more likely to be reported by victims and covered by the media than those that take place in other parts of Peru; moreover, they often involve migrants from the provinces or their immediate descendants. Nevertheless, surveys and official records indicate that these types of incidents occur at even higher rates elsewhere in the country. In Lima 4.9% of women report that they have experienced sexual violence at the hands of a spouse or companion, in comparison to 7.2% of women in the rest of Peru, including 8.8% in the highlands. On a per capita basis, incidents of family and/or sexual violence reported by women and recorded by the Ministry of Women and Vulnerable Populations are 19.1% less frequent in Lima than in other parts of the country, while cases of femicide recorded by the Public Ministry for the 2009-16 period are 14.3% less frequent in the capital than elsewhere.

“Attitudes supportive of women in politics have been consistently strongest in Lima,” even though cultural differences between the capital and other areas of Peru have narrowed considerably in recent years (Schmidt 2017: 146-47). In rural areas, potential female candidates for public office have often faced rejection by their husbands and ridicule in their communities (Yáñez 1999: 108). Women elected to municipal councils in the provinces have frequently endured hostile work environments and even violence (Ruiz Bravo and Córdoba Huaytán 2010). In a national survey of female leaders, Lima accounted for 15.7% of reported incidents of sexual harassment, only about half its share of the population (Transparencia 2014). In a survey of female candidates participating in the 2014 regional and municipal elections, most of the reported incidents of sexual harassment were found to have taken place outside of the capital, especially in the southern highlands (Pinedo et al. 2017: 44-47).

6.Preferential Voting in the Single National District

As noted earlier, during the Fujimori era a single national district was used to elect the unicameral Congress established by the 1993 Constitution. An unexpected consequence of this format was that professional women from Lima-Callao could compete for seats throughout the country. The nationalization of congressional elections also centralized nominations in the hands of party leaders based in the capital, who were generally more sympathetic to female candidates than regional political brokers. Moreover, the single national district magnified the importance of Lima-Callao, because more affluent and educated citizens are more likely to cast one or two of their optional preferential votes. In 2000, Lima-Callao accounted for 34.2% of all voters in Peru, but 40.9% of all preferential votes and 44.9% of those cast for winning candidates (ONPE 2002).

The top rows of data in Table 1 display the average number of preferential votes for female and male candidates who won election in the single national district in 1995 and 2000. The national averages are disaggregated by Lima-Callao and the rest of Peru. In both years, Lima-Callao provided a majority of the preferential votes for the female candidates who were elected, even though most voters resided elsewhere. In contrast, winning male candidates harvested most of their preferential votes in the provinces. Furthermore, in both 1995 and 2000, there were significant differences between Lima-Callao and the rest of Peru in terms of the percentage of the vote received by winning female and male candidates (.01 level; two-tailed). These voting patterns provide strong support for Hypothesis I.

Table 1
Preferential Votes in the Single National District

	1995			2000
	Lima-Callao	Rest of Peru	All Peru	Lima-Callao	Rest of Peru	All Peru
Average for all winning candidates
Female	30,908	30,336	61,244**	29,445	20,710	50,155**
	(50.5%)	(49.5%)		(58.7%)	(41.3%)
Male	6,535	11,327	17,862**	22,840	32,305	55,145**
	(36.6%)	(63.4%)		(41.4%)	(58.6%)
Average for winning candidates who do not Head List
Female	10,043	11,799	21,842*	29,445	20,710	50,155**
	(46.0%)	(54.0%)		(58.7%)	(41.3%)
Male	4,774	9,872	14,646*	18,188	26,406	44,594**
	(32.6%)	(67.4%)		(40.8%)	(59.2%)

Calculated from JNE (1995b) and ONPE (2002).
*Significant differences at the .019 level (two-tailed) between the percentages of the vote received from Lima-Callao and the rest of Peru by female and male winners.
**Significant differences at the .01 level (two-tailed) between the percentages of the vote received from Lima-Callao and the rest of Peru by female and male winners.

Although Peruvian congressional elections are open list, parties and alliances typically place their “star” candidates at the top of their lists, hoping to attract support for the list as a whole. In 1995, Martha Chávez—a vehement supporter of President Alberto Fujimori—headed the list of his electoral alliance, which won a landslide victory and 56% of the seats in Congress. Largely for this reason, the average number of votes received by female winners in 1995 was almost three and a half times greater than the average for male winners, as shown in the top rows of Table 1. One may wonder if Hypothesis I would still hold if Chávez and other “star” candidates elected at the top of competing lists were excluded.

The bottom rows of data in Table 1 display the average number of preferential votes for female and male winners who did not head their respective lists, once again breaking down the national averages by Lima-Callao and the rest of Peru. Excluding Chávez and other candidates who headed lists in 1995, female winners drew 46% of their preferential votes from Lima-Callao that year, compared to only 32.6% for male winners. Differences in the bases of support for female and male winners are statistically significant at the .019 level (two-tailed).

Men headed all of the lists that competed in the 2000 election, when the bases of support for female and male winners were almost perfect mirror images of one another. Congresswomen elected in 2000 won 58.7% of their preferential votes in Lima-Callao and 41.3% in the rest of Peru, while congressmen elected in the same year won 58.6% of their preferential votes in the provinces and 41.4% in Lima-Callao. Virtually the same pattern of support was behind the victories of males who did not head their lists. Differences in the bases of support for female and male winners are statistically significant at the .01 level (two-tailed).

Looking back at the top rows of data in Table 1, the decline in the average number of preferential votes received by female winners in 2000 reflected the absence of any “star” female candidate, as well as the greater number of women running for office that year after the adoption of the 25% gender quota. The results for winning candidates who did not head their lists in the bottom rows of Table 1 are much more reassuring: female winners in 2000 continued to average more preferential votes than male winners, and they more than doubled the average number of preferential votes received in 1995. Indeed, the number of women elected to Congress doubled between 1995 and 2000.

In summary, the evidence in Table 1 provides strong support for Hypothesis I. When all members of Congress were elected by open-list voting in a single national district, female candidates fared much better in Lima than in the provinces. This pattern also holds if we exclude “star” candidates at the head of lists.

7.Open vs. Closed Lists: Elections for Congress and District Councils

Since Fujimori fled Peru in late 2000, the country has used departments (regions) to elect members of the unicameral Congress. The upper rows of data in Table 2 present the percentage of congressional seats filled by women since 2001. The lower rows of data in the table report the percentage of seats on district councils filled by women since the adoption of gender quotas in 1997. Table 2 disaggregates the data for both types of offices by Lima and the rest of Peru, displaying ratios of the latter to the former.9 The percentages and ratios in the table show that women in Lima have had greater success than those in the provinces in both open-list elections for Congress and closed-list elections for district councils.

Table 2
Seats in Congress and on District Councils
Percentage Female

		1998	2001	2002	2006	2010	2011	2014	2016	2018	Average
		25% Quota	30% Quota
	Unicameral Congress (Open-List PR)
Lima*		---	31.4	---	31.4	---	33.3	---	33.3	---	32.4
Rest of Peru		---	12.9	---	28.2	---	17.0	---	25.5	---	20.9
Rest of Peru/Lima (%)		---	41.1	---	89.8	---	51.1	---	76.6	---	64.5
	District Councils (Closed-List Majoritarian)
Metropolitan Lima (LM)		29.7	---	25.8	31.8	30.9	---	36.7	---	39.8	32.5
Rest of Peru		23.5	---	26.9	28.2	27.9	---	28.7	---	30.3	27.6
Rest of Peru/(LM) (%)		79.1	---	104.3	88.7	90.3	---	78.2	---	76.1	84.9

*Department of Lima and Peruvians abroad in 2001 and 2006; Metropolitan Lima (Province of Lima) and Peruvians abroad in 2011 and 2016.
El Comercio (1999), JNE (1999; 2001: 39-48; 2006a: 30, 36-38; 2006b; 2006c; 2006d; 2006e: 21-27; 2011a; 2016c), MMR (1998: 78-131), and data provided by Zoila I. Cruz Molina of International IDEA and Milagros Suito of the JNE.

The data on open-list congressional elections in the upper rows of Table 2 provide strong support for Hypothesis I. Since 2001 women have filled an average of 32.4% of congressional seats in Lima, compared to only 20.9% in the rest of Peru.10 The average ratio of the rest of Peru to Lima is 64.5%. These differences cannot be explained by district magnitude, which is much greater for congressional elections in Lima than in the rest of the country. With only one exception (2001), the correlations between district magnitude and the percentage of congressional seats filled by women—calculated from sources for Table 2—are weak or even negative. Moreover, district magnitude cannot explain the voting patterns in the single national district discussed in the preceding section or the results of the counterfactual comparisons undertaken in the next section.

Only in 2006, under quite unusual circumstances, did women running for Congress in the provinces come close to matching the success of female candidates in Lima (ratio = 89.8%). In that year, an outsider candidate, Ollanta Humala, won the first round of the presidential election but narrowly lost the runoff. Having failed to register his party, Humala ran as the candidate of the Union for Peru (UPP), a left-leaning party with a strong record of nominating female candidates (Schmidt 2011: 29). His first-round victory had long coattails for UPP-dominated lists, and women won a third of the 45 seats in the pro-Humala block. Moreover, Humala was much more popular in the provinces than in Lima, and only one of the 15 women elected on UPP lists represented the capital (JNE 2006a).

The data on closed-list majoritarian elections for district councils in the lower rows of Table 2 show that since the adoption of quotas, women have filled an average of 32.5% of the seats in Lima compared to 27.6% in the rest of Peru. Limeñas have won a higher percentage of seats in five of the six post-quota elections. The exception was 2002, after feminist activism in the capital had declined following the completion of PROMUJER: a joint project launched pior to the 1998 election by five Lima-based NGOs to recruit and train female candidates, while successfully pressuring party leaders to exceed the quota. A difference of means test, using data from sources in Table 2, shows that the decline in seats held by women in 2002 was significantly greater in districts where PROMUJER had been most active (.01 level; 2-tailed).

As in the case of congressional elections, the greater success of limeñas in municipal elections cannot be attributed to greater district magnitude. Indeed, in majoritarian electoral systems with gender quotas, lower district magnitudes can actually favor the election of women. Recall that under the maximalist interpretation of the half plus one rule, the winning list is awarded four of the seats on five member councils, which account for the overwhelming majority of district councils in Peru. Given that each list of five candidates must include at least two women in order to fulfill a gender quota of 25% or 30%, at least one female candidate will be elected, even if the winning party or alliance only minimally complies with the quota and places both women at the very bottom of its list. She alone will account for 25% of the list’s delegation and for 20% of the council’s total membership. This guaranteed female representation in five-member majoritarian districts is similar to placement mandates under closed-list PR11 and is far more prevalent in the provinces than in Lima. For example, in 2018, 92.9% of the district councils outside of Lima had only five members, compared to only 14.3% in the capital.12

Turning to Hypothesis II, the data presented in Table 2 show that ballot structure makes little, if any, difference in the capital. Women have held an average of 32.4% of the seats in the capital’s congressional delegation and 32.5% of the seats on its district councils. Female candidates are competitive in winning preferential votes in congressional elections, while party gatekeepers or activists are mindful of the electorate’s preferences when they determine closed lists for municipal elections. Lists in Lima frequently exceed the gender quota. Limeñas initially won a slightly higher percentage of seats in open-list congressional elections, except for 2006, but have fared somewhat better in closed-list municipal elections in recent years.

In contrast, disparities in the election of women under alternative ballot structures are much greater in the provinces, as predicted by Hypothesis II. Women have averaged 27.6% of the seats in closed-list elections for district councils, but only 20.9% in open-list elections for Congress. The average ratio of the rest of Peru to Lima is 84.9% for district councils, but only 64.5% in the case of Congress.

8.Counterfactual Comparisons in Congressional Elections

Although direct comparisons of open-list and closed-list PR in Lima and the provinces are not possible (see note 9), counterfactual comparisons can be constructed by reviewing the lists of candidates presented by the parties or alliances that won seats in Congress in order to determine which candidates would have been elected had the list order been followed. These comparisons are not perfect because a change in electoral rules might prompt shifts in strategic behavior. For example, when Peru used the single national district for congressional elections during the Fujimori era, several contenders of both genders requested the last position on the list of 80-120 candidates because it provides greater visibility than those in the middle. Nevertheless, a series of counterfactual comparisons is credible if it produces consistent results, especially when there are major changes in key factors, such as district magnitude and winning parties.

Table 3 displays the actual percentage of congressional seats won by female candidates and the counterfactual percentage that they would have won according to list order. When the actual and counterfactual results vary, the higher percentages are boldfaced. During the 1978-2000 period, Peru used a single national district to elect the CA, Senate, CCD, and unicameral Congress. Electoral districts based on departments (regions) have been used to elect the unicameral Congress since 2001. The Senate was elected by closed list in 1980, but for all other years the actual results in the table are from open-list elections.

Table 3
Women Elected to Congress: Actual and Counterfactual Results
Percentage Female

	Open List 1 Pref. Vote	Closed List	Open List (Double Optional Preferential Vote)
	Single National District							Departmental (Regional) Electoral Districts
	CA	Senate			CCD	Unicameral Congress
	No Quota						25% Quota	30% Quota
	1978	1980	1985	1990	1992	1995	2000	2001	2006	2011	2016
All Peru
Actual Results	2.0	3.3	5.0	6.7	8.8	10.8	21.7	18.3	29.2	21.5	27.7
By List Order	5.0	3.3	5.0	8.3	10.0	10.0	22.5	20.8	30.0	26.2	28.5
Lima*
Actual Results								31.4	31.4	33.3	33.3
By List Order								34.3	31.4	25.0	30.6
Rest of Peru
Actual Results								12.9	28.2	17.0	25.5
By List Order								15.3	29.4	26.6	27.7

*Department of Lima and Peruvians abroad in 2001 and 2006; Metropolitan Lima and Peruvians abroad in 2011 and 2016.
Calculated from data published in Blondet and Montero (1994: 122-128), Caretas (2001), JNE (1978; 1982; 1986: Vol I, 101-156; 1990; 1992; 1995a&b; 2000; 2001: 57-129; 2006a: 16, 30; 2011b: 95, 100; 2016a: 69-77; 2016b: 32, 36; 2016c: 4), ONPE (2002; 2005: 22), Tuesta Soldevilla (1994: 64-75; 2001: 66-68, 71-73, 76-88), Villanueva Flores (1998: 31), Webb and Fernández Baca (2000: 264), Yáñez and Guillén (2001: 20), and data sets on candidates in the 2006 and 2011 elections provided by the JNE.

Actual and counterfactual percentages for all of Peru are presented in the top two rows of data in Table 3. The numbers for 1980 are redundant because closed lists were used, but the counterfactual percentages are greater than the actual percentages in eight of the remaining ten years. The actual percentage is slightly greater only in 1995—when Martha Chávez headed Fujimori’s list in his landslide victory—and the numbers are identical in 1985. Given that most voters reside in the provinces, the consistent pattern of these counterfactual comparisons, like the analyses in the preceding two sections, strongly suggest that open lists work against women outside of Lima.

The remaining rows of Table 3 allow Hypothesis II to be explicitly tested by disaggregating data from congressional elections held in departmental (regional) electoral districts since 2001. Moreover, our confidence in these counterfactual comparisons for Lima and the rest of Peru should be even greater because the shorter lists used in post-2001 elections undercut the strategy of requesting the last position. The comparisons for Lima are ambiguous: a higher percentage of women would have been elected by list order in 2001, but open list was better for female candidates in 2011 and 2016, and the results were identical in 2006. In contrast, female candidates in the rest of Peru would have consistently won a higher percentage of seats under list order. Thus, these disaggregated counterfactual comparisons provide strong support for Hypothesis II: alternative ballot structures make little or no difference in Lima, but they clearly matter in the provinces.

9.Conclusion

The impact of ballot structure (candidate-centered vs. party-centered; open vs. closed list) for the election of women has been a longstanding issue in electoral studies. Various scholars have provided contradictory findings, while others have concluded that ballot structure may have a varying impact or may not even matter. By using a quasi-experimental research design to test hypotheses derived from Valdini (2013) that speak to this longstanding debate, this article makes an important contribution to the literature on electoral systems. The longitudinal perspective and “thick description” of this case study complement the cross-national approach in Valdini (2013).

In Peru, the socioeconomic context and political culture of Lima are more conducive to the success of female candidates than conditions in the rest of Peru. As anticipated by Valdini, women running for Congress in open-list PR elections have consistently fared better in Lima than in the provinces (Hypothesis I). However, female candidates also have been more successful in the capital than elsewhere in closed-list majoritarian elections for district councils.

It also follows from Valdini that disparities in the election of women under alternative ballot structures should be greater in the provinces than in Lima (Hypothesis II). Indeed, the average shares of seats filled by women in open-list PR elections for Congress and closed-list majoritarian elections for district councils are virtually identical in Lima (32.4% vs. 32.5%), but female candidates in the provinces have had notably greater success in elections for district councils (27.6% vs. 20.9% for congressional elections). The results of counterfactual comparisons provide further evidence for Hypothesis II: female congressional candidates in the provinces would clearly fare better in closed-list PR elections than in open-list PR elections, but the impact of ballot structure is ambiguous in Lima.

The two hypotheses hold up well over time in Peru, despite major variations in district magnitude and the party system. During the 1978-2000 period, various types of congressional elections were conducted in a single national district, with magnitudes ranging from 60 to 120, but since 2001 Peru’s unicameral Congress has been elected in much smaller departmental (regional) districts. Peru had a tenuous, declining party system during the 1980s, saw the rise of various personalistic parties supporting Alberto Fujimori and his competitors during the 1990s, and during the post-Fujimori era since 2001 has experienced unprecedented electoral volatility among competing “anti-candidates” and “coalitions of independents.” Nevertheless, the relationship between ballot structure and the election of women in Lima and the provinces, as outlined in this article, has remained remarkably consistent.

Looking forward, it will be interesting to see whether these electoral patterns persist, as cultural differences between Lima and the provinces continue to narrow. Recent proposals to adopt closed-list congressional elections with parity as well as alternating candidacies by gender would undoubtedly improve the descriptive representation of women, with a markedly greater impact in the provinces than in Lima.

References

Ballington, Julie. 2005. Ten years of progress: enhancing women’s political participation. In Ten Years of Supporting Democracy Worldwide. Stockholm: International IDEA, pp. 113-124.

Blondet, Cecilia and Carmen Montero. 1994. La situación de la mujer en el Perú 1980-1994. Lima: Instituto de Estudios Peruanos.

Caretas. 2001. Guía del elector, Supplement, April 5.

Cox, Gary W. 1997. Making Votes Count: Strategic Coordination in the World’s Electoral Systems. New York: Cambridge University Press.

El Comercio. 1999. Mujeres en el poder . . . y el poder de las mujeres, Special supplement on International Women’s Day, March 8.

Farrell, David M. 2011. Electoral Systems: A Comparative Introduction. Second edition. New York: Palgrave Macmillan.

Htun, Mala N. 2005. Women, political parties and electoral systems in Latin America. In Women in Parliament: Beyond Numbers, revised edition, edited by Azza Karam and Julie Ballington. Stockholm: International IDEA, pp. 112-121.

Htun, Mala N. and Mark P. Jones. 2002. Engendering the right to participate in decision-making: electoral quotas and women’s leadership in Latin America. In Gender and the Politics of Rights and Democracy in Latin America, edited by Nikki Craske and Maxine Molyneux. Houndmills, Basingstoke, Hampshire, England: Palgrave, pp. 32-56.

INEI (Instituto Nacional de Estadística e Informática). 2019. https://www.inei.gob.pe, last accessed May 24, 2019.

JNE (Jurado Nacional de Elecciones). 1978. Resultados de las elecciones para representantes a la Asamblea Constituyente. Lima.

JNE (Jurado Nacional de Elecciones). 1982. Resultados de las elecciones políticas generales de 1980. 2 Vols. Lima.

JNE (Jurado Nacional de Elecciones). 1986. Resultados de las elecciones políticas generales de 1985. 4 Vols. Lima.

JNE (Jurado Nacional de Elecciones). 1990. Aplicación de cifra repartidora de vot. senadores. Lima.

JNE (Jurado Nacional de Elecciones). 1992. Orden de candidatos por votación dentro de sus listas. Lima.

JNE (Jurado Nacional de Elecciones). 1995a. Autorizan la publicación de las listas de candidatos a congresistas para las elecciones políticas generales de 1995, Resolución No 141-95-JNE. El Peruano, March 9.

JNE (Jurado Nacional de Elecciones). 1995b. Elecciones políticas generales 1995: votación preferencial a nivel JPI. Lima.

JNE (Jurado Nacional de Elecciones). 1999. Elecciones municipales 1998 y complementarias 1999. Resultados generales. Lima.

JNE (Jurado Nacional de Elecciones). 2000. Elecciones generales 2000. Desarrollo y resultados. Lima.

JNE (Jurado Nacional de Elecciones). 2001. Elecciones generales 2001. Resultados generales. Lima.

JNE (Jurado Nacional de Elecciones). 2006a. Elecciones generales 2006. Elección presidencial, congresal y al Parlamento Andino. Resultados. Lima.

JNE (Jurado Nacional de Elecciones). 2006b. Elecciones municipales 2006: resultados y estadísticas. Lima.

JNE (Jurado Nacional de Elecciones). 2006c. Elecciones municipales en Lima y Callao 1998-2006. Lima.

JNE (Jurado Nacional de Elecciones). 2006d. Elecciones regionales y municipales 2002: estadísticas generales. Lima.

JNE (Jurado Nacional de Elecciones). 2006e. Resolución No 1163-2006-JNE. Lima.

JNE (Jurado Nacional de Elecciones). 2011a. Acta General de Proclamación de Resultados de la Elección de Congresistas de la República del Año 2011. Lima.

JNE (Jurado Nacional de Elecciones). 2011b. Compendio estadístico electoral: elecciones regionales y municipales 2010. Elecciones generales 2011. Lima.

JNE (Jurado Nacional de Elecciones). 2016a. Acta General de Proclamación de Resultados de la Elección de Congresistas de la República Realizada el Domingo 10 de Abril de 2016. Lima.

JNE (Jurado Nacional de Elecciones). 2016b. Estadísticas de las elecciones generales 2016. Lima.

JNE (Jurado Nacional de Elecciones). 2016c. Perfil del Congreso de la República 2016-2021. Reporte N.O 9, Serie EG 2016. Lima.

JNE (Jurado Nacional de Elecciones). 2019. Registro de Organizaciones Políticas. http://aplicaciones007.jne.gob.pe/srop_publico/Consulta/OrganizacionPolitica, last accessed May 30, 2019.

Jones, Mark P. 1998. Gender quotas, electoral laws, and the election of women: lessons from the Argentine provinces. Comparative Political Studies 31(1): 3-21.

Jones, Mark P. and Patricio Navia. 1999. Assessing the effectiveness of gender quotas in open-list proportional representation electoral systems. Social Science Quarterly 80 (June): 341-355.

Krumholtz, Michael. 2018. Lima ranks as the most dangerous city in the world for women. Peru Reports, October 2. https://perureports.com/lima-ranks-as-the-most-dangerous-city-in-the-world-for-women/8543/, last accessed March 21, 2019.

Levitsky, Steven. 2018. Peru: The Institutionalization of Politics without Parties. In Party Systems in Latin America: Institutionalization, Decay, and Collapse, edited by Scott Mainwaring. New York: Cambridge University Press, pp. 326-355.

Mainwaring, Scott. 2018. Party System Institutionalization in Contemporary Latin America. In Party Systems in Latin America: Institutionalization, Decay, and Collapse, edited by Scott Mainwaring. New York: Cambridge University Press, pp. 34-70.

Matland, Richard E. 1993. Institutional variables affecting female representation in national legislatures: The case of Norway. The Journal of Politics 55 (August): 737-755.

Matland, Richard E. 2005. Enhancing women’s political participation: legislative recruitment and electoral systems. In Women in Parliament: Beyond Numbers, revised edition, edited by Julie Ballington and Azza Karam. Stockholm: International IDEA, pp. 93-111.

Meléndez, Carlos (Compiler). 2011. Anti-candidatos: guía analítica para unas elecciones sin partidos. Lima: 50+1 Grupo de Análisis Político.

MMR (Movimiento Manuela Ramos) (ed). 1998. Poder político con perfume de mujer: las cuotas en el Perú. Lima: Instituto de Estudios Peruanos.

Navarrete, Priscila A. 2014. “Sílbale a tu madre,” el falso documental para frenar el acoso callejero en Perú. El País, December 7.

Norris, Pippa. 2004. Electoral Engineering: Voting Rules and Political Behavior. New York: Cambridge University Press.

ONPE (Oficina Nacional de Procesos Electorales). 2002. Web page. www.onpe.gob.pe

ONPE (Oficina Nacional de Procesos Electorales). 2005. Los procesos electorales en el Perú: 1978-1986. Problemas y lecciones. Lima: ONPE.

Pinedo, Enith, Carla Cueva, Rubi Rivas, and Diana Miloslavich. 2017. El acoso político en el Perú: una mirada desde los recientes procesos electorales. Lima: Jurado Nacional de Elecciones, Centro de la Mujer Peruana Flora Tristán, and IDEA Internacional.

Rae, Douglas W. 1967. The Political Consequences of Electoral Laws. New Haven: Yale University Press.

Ruiz Bravo, Patricia and Luciana Córdova Huaytán. 2010. Los retos del espacio público: fiscalización, violencia y acoso. El caso de las regidoras de San Martín, Puno y Piura – Perú. Trabajo presentado en el Congreso Internacional “Las políticas de equidad de género en prospectiva: nuevos escenarios, actores y articulaciones,” Buenos Aires.

Rule, Wilma and Matthew S. Shugart. 1995. The preference vote and election of women: women win more votes in open list PR. In Voting and Democracy Report 1995. Washington, D.C.: The Center for Voting and Democracy, pp. 177-78.

Schmidt, Gregory D. 2009. The election of women in list PR systems: testing the conventional wisdom. Electoral Studies 28 (2): 190-203.

Schmidt, Gregory D. 2011. Gender disparities in re-election: the 2006 and 2010-2011 electoral cycles in Peru. Report submitted to International IDEA. Lima.

Schmidt, Gregory D. 2017. Ballot Structure and Regional Context: A Comparison of Lima and the Provinces. In Gender Quotas in South America’s Big Three: National and Subnational Impacts, by Adriana Piatti-Crocker, Gregory D. Schmidt and Clara Araújo. Lanham, Maryland: Lexington Books, pp. 143-155.

Schmidt, Gregory D. and Kyle L. Saunders. 2004. Effective quotas, relative party magnitude, and the success of female candidates: Peruvian municipal elections in comparative perspective. Comparative Political Studies 37-6 (August): 704-734.

Shugart, Matthew S. 1994. Minorities represented and unrepresented. In Electoral Systems in Comparative Perspective: Their Impact on Women and Minorities, edited by Wilma Rule and Joseph F. Zimmerman. Westport, CT: Westview Press, pp. 31-41.

Shugart, Matthew S. 2005. Comparative electoral systems research: the maturation of a field and new challenges ahead. In The Politics of Electoral Systems, edited by Michael Gallagher and Paul Mitchell. New York: Oxford University Press, pp. 25-55.

Taagepera, Rein. 1994. Beating the law of minority attrition. In Electoral Systems in Comparative Perspective: Their Impact on Women and Minorities, edited by Wilma Rule and Joseph F. Zimmerman. Westport, CT: Westview Press, pp. 235-245.

Tanaka, Martín. 1998 Los espejismos de la democracia: el colapso del sistema de partidos en el Perú, 1980-1995, en perpectiva comparada. Lima: Instituto de Estudios Peruanos.

Tanaka, Martín. 2005. La democracia sin partidos. Perú 2000-2005: los problemas de representación y las propuestas de reforma política. Lima: Instituto de Estudios Peruanos.

Transparencia. 2014. Primer reportaje de acoso político hacia las mujeres, Febrero – Abril 2014. Lima.

Tuesta Soldevilla, Fernando. 1994. Perú político en cifras: elite política y elecciones. Second edition. Lima: Fundación Friedrich Ebert.

Tuesta Soldevilla, Fernando. 2001. Perú político en cifras, 1821-2001. Third edition. Lima: Friedrich Ebert.

Uchuypoma Soria, Diego and Freidenberg, Flavia. 2017. Con cuota o sin cuota: candidatas exitosas para el Congreso. In Perú: Elecciones 2016. Un país dividido y un resultado inesperado, edited by Fernando Tuesta Soldevilla. Lima: Pontificia Universidad Católica del Perú, pp. 119-153.

UNDP (United Nations Development Programme). 2018. Human Development Indices and Indicators: 2018 Statistical Update. New York: UNDP.

Valdini, Melody Ellis. 2013. Electoral institutions and the manifestation of bias: the effect of the personal vote on the representation of women. Politics and Gender 9: 76-92.

Villanueva Flores, Rocío. 1998. ¿Están justificadas las cuotas de mujeres en las listas electorales? In Poder político con perfume de mujer: Las cuotas en el Perú, edited by Movimiento Manuela Ramos. Lima: Instituto de Estudios Peruanos.

Webb, Richard and Graciela Fernández Baca. 2000. Perú en números 2000. Lima: Cuánto S.A.

Yáñez, Ana María. 1999. Participación política y liderazgo de la mujer. Socialismo y participación 85: 107-114.

Yáñez, Ana María and Lisbeth Guillén. 2001. Sección especial: elecciones 2001. El Cuarto Femenino No 10: 16-23.

Zavaleta, Mauricio. 2014. Coaliciones de independientes; las reglas no escritas de la política electoral en el Perú. Lima: Instituto de Estudios Peruanos.

---