Journal of Asian and African Studies http://jas.sagepub.com/ Identifying the Institutional Effects of Mixed Systems in New Democracies: The Case of Lesotho Timothy S Rich, Vasabjit Banerjee and Sterling Recker Journal of Asian and African Studies published online 12 November 2013 DOI: 10.1177/0021909613507233 The online version of this article can be found at: http://jas.sagepub.com/content/early/2013/10/28/0021909613507233 Published by: http://www.sagepublications.com Additional services and information for Journal of Asian and African Studies can be found at: Email Alerts: http://jas.sagepub.com/cgi/alerts Subscriptions: http://jas.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav >> OnlineFirst Version of Record - Nov 12, 2013 What is This? Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 507233 JAS0010.1177/0021909613507233Journal of Asian and African StudiesRich et al. research-article2013 JAAS Article Identifying the Institutional Effects of Mixed Systems in New Democracies: The Case of Lesotho Journal of Asian and African Studies 0(0) 1–17 © The Author(s) 2013 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0021909613507233 jas.sagepub.com Timothy S Rich Political Science Department, Western Kentucky University, USA Vasabjit Banerjee Department of International Studies, Indiana University, USA and Department of Sociology, University of Pretoria, South Africa Sterling Recker University of Missouri – St. Louis, USA Abstract This paper addresses the effects of the mixed system used for the last three elections in Lesotho (2002, 2007 and 2012), a small African country with a turbulent history regarding opposition acceptance of elections. The decision to implement a mixed system was in part to encourage democratic stability, yet whether the electoral system has become more conducive to democratic competition is unclear. Through an analysis of national and district-level results, this paper addresses the following questions. First, at the district level, is competition consistent with Duverger’s law or the contamination thesis and is a progression over time evident? Second, does the population size of a district influence the number of candidates and the concentration of votes? Finally, following recent research on detecting electoral fraud, this paper tackles whether the reports of district results suggest extra-institutional manipulation. Keywords Africa, democratization, Duverger, electoral institutions, mixed-member majoritarian, Lesotho, postconflict Introduction Since independence, establishing a stable democracy in Lesotho has been riddled with challenges. Yet with the adoption of a new electoral system for the National Assembly in 2002, multiparty competition and democratic stability seem to have emerged in the small African country. Moving Corresponding author: Vasabjit Banerjee, Department of International Studies, Indiana University, 1100 East 7th St., Woodburn Hall Rm. 316, Bloomington, Indiana 47401, USA. Email: vasabjit@gmail.com Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 2 Journal of Asian and African Studies 0(0) away from a first-past-the-post majoritarian system, Lesotho adopted a mixed legislative system, increasingly common across the globe. Mixed member legislative systems employ two systems to allocate seats within the same legislative house, usually in the form of single member districts (SMDs) as well as seats allocated through proportional representation (PR). However, most of the literature on the institutional effects of mixed systems is based on evidence from older democracies (e.g. Germany, New Zealand and Japan). This paper addresses the effects of the mixed system used for the last three elections in Lesotho (2002, 2007 and 2012). The decision to implement a mixed system was in part to encourage democratic stability and a more broadly representative National Assembly (Elklit, 2002), yet whether the electoral system has become more conducive to democratic competition is unclear. Through an analysis of national and district-level results, this paper intends to address the following questions. First, at the district level, is district competition consistent with Duverger’s law (two-party competition) or the contamination thesis (multiparty competition) and is a progression over time evident? Second, do the district population and the number of candidates influence vote concentration and turnout? Finally, following recent research on detecting electoral fraud, this paper attempts to identify whether the reports of district results suggest extra-institutional manipulation. There are several valid reasons for using Lesotho as a case study on mixed legislative systems. Mixed systems have proliferated globally in the past 20 years, but often fail to live up to the original promise of providing local accountability and representation of diverse interests. Among postcommunist countries, for example, several opted for a mixed system for early elections (e.g. Russia, Ukraine and Albania) before moving to another formula. Furthermore, because much of the literature relies on how mixed system reforms fared in established democracies, the question remains whether the electoral formula is effective in poor developing countries with thin democratic roots, especially ones that are geographically small. Lesotho thus provides a means to test how mixed systems perform in what appears as a less than ideal setting. Sartori (1976) argued that revolutions of decolonization could realign cleavages in a manner favorable to the rise of a dominant party. Labeling Lesotho’s political history as a dominant party system is problematic in that there has not been a consistent party in power, but some similarities exist. From the disputed 1970 election until Lesotho’s first mixed election in 2002, the majority party won almost every seat, with the opposition holding only one seat in 1998. As such, the ruling parties could modify electoral rules to suit their needs. Similarly, like dominant party systems, ruling parties failed to be monolithic and instead often balanced interests and factions.1 While not matching the efforts of pork distribution and promises of future policy benefits of typical dominant party examples like the Liberal Democratic Party (LDP) of Japan (Pempel, 1990: 352), Lesotho party politics showed similar pathologies. Rather than evaluate Lesotho’s democracy on poorly defined standards of consolidation or rely on impressionistic accounts, we instead focus on the electoral institutions and whether voters in Lesotho act in a manner consistent with theoretical expectations. We also acknowledge that, even while electoral institutions shape voting behavior, the salience of institutions cannot be properly evaluated in the presence of electoral manipulation, for example electoral fraud. Thus, besides just evaluating theoretical expectations regarding electoral institutions, we attempt to measure the presence or absence of electoral fraud through indirect means. This paper will first introduce the literature on mixed systems and what we should expect in Lesotho. Second, an introduction to Lesotho politics follows. We then present our research design before empirically analyzing the national and district-level election results in order to compare with expectations from mixed systems. The subsequent statistical analysis of patterns in electoral returns gives little cause to assume significant electoral fraud. In total, this analysis suggests that Lesotho’s mixed system is producing outcomes in line with democratic expectations. Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 3 Rich et al. Evaluating Mixed Systems An expansive literature in political science analyzes how electoral institutions shape party systems (e.g. King, 1990; Lijphart, 1994; Norris, 2004; Rae, 1971, 1995; Rokkan, 1970; Taagepara and Shugart, 1989). Perhaps the most consistent finding has been Duverger’s law (Duverger, 1954), that the mechanical and psychological effects of single-member districts (SMDs) with plurality rule lead to two-party competition whereas larger district magnitude, usually associated with proportional representation (PR), produces a multiparty system. As there can only be one winner in plurality SMDs, smaller parties tend to drop out or align with larger parties with greater chances of victory. Similarly, voters are unlikely to support candidates with little chance in winning (Cox, 1997; Golder, 2006). While outliers exist usually due to regional parties2 and newer democracies often do not immediately conform, as a whole Duverger’s law has held in a vast array of systems comprised solely of SMDs (Amorim-Neto and Cox, 1997; Jones, 1999; Ordeshook and Shvetsova, 1994; Taagepera and Shugart, 1993). Mixed member systems provide a new institutional context to test Duverger. Proponents see these systems as ‘the best of both worlds’ (Shugart and Wattenberg, 2001), combining the accountability and local constituency focus of SMDs with the diverse representation and national policy focus of PR. Analytically, proponents praise these systems for allowing a form of natural experiment, holding all but seat type as constant (Lancaster and Patterson, 1990; Moser, 1995, 1997; Stratmann and Baur, 2002). Moser and Scheiner (2004: 2) specifically refer to these systems as a ‘social laboratory in which effects of different types of electoral systems can be studies in isolation from influence of the social context.’ Accordingly, such approaches need not consider additional factors, especially in single-country studies. Each tier is expected to separately conform to the Duvergerian institutional expectations, with two-party competition in district seats and multiparty competition in the list tier. As a result, strategic voting and interparty coordination is expected to be limited to the district tier while voters remain sincere in their list vote, encouraging parties to go it alone. In essence, legislator actions should fall within the rationalist tradition. While natural experiments are appealing, the two sets of electoral rules for the same legislature may affect one another creating interdependence and non-Duvergerian outcomes (contamination thesis). Bawn (1999) argues that at its base there must be some voters who use information from the district tier in calculating their party vote. Duverger (1986) himself suggested that district competition could spill over into the PR tier. Most scholars focus on additional factors that cross tiers (such as party interests) and may increase the number of district candidates (Ferrara et al., 2005; Herron and Nishikawa, 2001). Sartori (1997: 75), for example, warned that such institutional arrangements potentially result in a ‘bastard-producing hybrid, which combines their defects.’ Whether mixed systems produce positive or negative outcomes compared to other systems is debatable, but theoretical and empirical evidence of tier interaction is pervasive. For example, since the proportional tier dictates the overall results, determining the utility of an individual vote under the mixed member proportional (MMP) system is complicated, likely reducing strategic voting (Reynolds et al., 2005). Overall, the expectation in mixed systems is a deviation from ‘Duvergerian gravity’ and a greater dispersion of votes from the top two competitors in district competition (Cox and Schoppa, 2002; Ferrara et al., 2005; Herron and Nishikawa, 2001). The psychological and mechanical effects of Duverger’s law are products of district competition and are open to distortion with aggregation to the national level. Therefore, the district level is a more appropriate level of analysis for testing Duvergerian expectations. Variation in competition between mixed systems has been analyzed at the national level, usually through a measure of the Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 4 Journal of Asian and African Studies 0(0) effective number of parties. In contrast, measures of district-level competition remain rare (e.g. Ferrara and Herron, 2005; Moser and Scheiner, 2004) and cover only a limited number of mixed systems, none including evidence from Lesotho. While the literature suggests that MMP systems produce greater incentives for additional entries at the district level, there appears little discussion in regards to the competitiveness of district races under MMP versus the mixed member majoritarian system (MMM). Unlike the MMP system, the MMM system encourages parties to move towards the median district vote, potentially increasing the closeness of such elections. Similarly, as voters increasingly understand the system, Duvergerian pressures promote a greater concentration of voters around the top two candidates (Bartolini, 2002). However, the increased number of candidates in districts under MMP potentially lowers the bar in elections in terms of the total percentage of the vote necessary for victory, making these districts more competitive as well. Lastly, the literature on mixed systems remains quiet in regards to its effects on turnout. While implementation was often predicated on increasing citizen participation, little research to date tackles how mixed systems influence showing up to the ballot box (Rich, 2013). Still, the broader electoral systems literature gives us a few expectations. While an increased number of parties means voters should be able to find a party close to their preferences, this does not necessarily lead to a boost in turnout. Rather, increases in the number of parties appear to depress turnout, in part due to confusion over party stances and the likelihood of less decisive coalitional governments (Blais and Dobrzynska, 1998; Brockington, 2004; Jackman, 1987). Meanwhile, mixed systems are not immune to the 30-year global trend in declining turnout in elections in new and older democracies (Franklin, 2004; Franklin et al., 2004; Kostadinova, 2003; Niemi and Weisberg, 2001; Wattenberg, 2002). Thus, to understand Lesotho’s results compared to other democracies, we must consider both forces that appear to curb turnout. The Political Trajectory of Lesotho: Understanding the Promise of Electoral Institutions The evolution of Lesotho’s electoral institutions informs our understanding of democratic transition and consolidation in several ways. First, Lesotho has a recent history of political instability and military rule, as have many post-Cold War African countries. However, Lesotho seems to have implemented electoral reform as a means to end violence as opposed to a way to consolidate the power of a specific party or ethnic group. Second, the recent 2012 elections, characterized by a free and fair outcome, suggest the reforms have led to a period of democratic consolidation wherein parties and voters now accept the outcomes and refrain from violence. In order to understand the central role of electoral institutions, this section presents a brief history of the factors that led to political instability and violence before explaining how electoral institutions were reformed to resolve these problems. Lesotho is a small kingdom within South Africa with a history of strong state control and influence in political and economic institutions. After decolonization in 1966 from the British, Makoa (2004) argues that Lesotho was left without a clearly defined political ideology, but was expected to foster a form of constitutionalism resembling the Westminster system. This left the newly independent people with limited understanding of democracy in practice, even though they had the formal institutions in place for such a system. As a result, political conflicts originating in British colonial policies of dividing and ruling Lesotho became institutionalized at the time of independence, creating what Turkon (2008) refers to as ‘colonial politics.’ Aside from the colonial legacy, elections in Lesotho and the pendulum swings towards and away from democratic consolidation have been influenced by its economic structure and Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 5 Rich et al. the politics of its powerful neighbor, South Africa. Amin (1972) classifies Lesotho as part of a macro-region that supplied labor to colonial territories ruled by the British. Economic growth in Lesotho has long depended on agriculture and remittances from migrant workers in the South African mining industry (Johnston, 1996), making the majority of the population reliant on lowwage labor and subsistence farming. The early 1990s also saw extensive labor unrest due to foreign investment and the government’s lack of industrial and labor standards in regards to wages and working conditions (Tangri, 1993). Given these economic characteristics, Lesotho remains a fragile economy with a surplus labor pool as the private sector is largely non-existent. Thus, citizens frequently look to their South African neighbor for temporary employment. This economic fragility also encourages parties and candidates to view elections as a means to access state resources. Lesotho’s transition coincided with the end of the Cold War and Apartheid in South Africa, thus Lesotho’s political development has been heavily influenced and shaped by external factors (Kabemba, 2003; Olaleye, 2004). The Apartheid regime in South Africa directly and indirectly intervened in Lesotho politics by supporting the Lesotho Liberation Army, the armed wing of the Basutoland Congress Party that was created after the party’s suppression in 1970 (Pherudi, 2001). Even in 1998, South African armed forces intervened to stop post-electoral violence. By understanding Lesotho’s development from both the colonial and post-colonial perspective, we can better understand the formation of parties and how party politics has shaped the current electoral system. Matlosa (2006) identifies six periods in Lesotho’s post-independence politics: an embryonic democracy (1966–1970); one party authoritarianism (1970–1986); military authoritarianism (1986–1993); the return of multiparty democratic competition (1993–1998); a fragile and less violent democracy (1998–2002); and, finally, a relatively stable democracy since 2002. Lesotho had two dominant parties between the periods of independence from Great Britain in 1966 to the fall of the military regime in 1993: the Basutoland Congress Party (BCP) and the Basotho National Party (BNP). The nationalist BCP was formed in 1952 and the BNP was established soon after in 1957. After winning the 1965 elections, the BNP suspended the constitution, thus laying the groundwork for decades of political rivalry between these two dominant parties (Olaleye, 2004). Unlike many Sub-Saharan African countries, Lesotho is relatively homogenous; thus, pre-independence elections were defined early on by other means of party mobilization, not traditional conceptions of ethnic divisions. Admittedly lineage (e.g. Bakoena, Bafokeng, etc.) as well as religious divisions (e.g. Catholicism, Anglican, Lesotho Evangelical) still remain an undercurrent in Lesotho politics, but do not appear as deep as cleavages traditionally associated with ethnicity. The lack of ethnic cleavages is key to understanding the evolution of party politics in subsequent decades, as voters made decisions based on factors other than ethnic identification. This alone distinguishes Lesotho’s electoral development from many other African nations, making it a unique country from which to understand the importance of post-conflict democratization via electoral reform. Regardless of these early experiences with proto-democratic elections, the BNP contested the first post-independence election in 1970 by refusing to step down, after the BCP was unofficially projected to win 33 out of the total 57 seats (McCartney, 1973: 485). This stolen victory began several decades of conflict between the parties that led to periods of authoritarian rule until 1993. This era of political conflict disenfranchised many smaller parties who felt they had limited opportunities under military rule (Cho and Bratton, 2006). The return of multiparty elections in 1993 created additional impetus for conflict as the BCP won all 65 seats with only 74.7% of the vote. In 1998 the Lesotho Congress of Democrats (LCD), a breakaway party from the BCP, won 79 of 80 seats in parliament, once again marginalizing smaller parties under the first-past-the-post (FPTP) electoral system. Due to the history of one-party dominance, Lesotho adopted the current MMP system for the 2002 electoral cycle to encourage smaller parties to run for office, thus Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 6 Journal of Asian and African Studies 0(0) Table 1. National Assembly election results (1965–1970). Party 1965 No. of seats Party % of votes BNP BCP MFP MTP Independents 41.5 39.7 16.5 2.2 0.03 1970 No. of seats % of votes 31 25 4 0 0 BNP BCP MFP Other 42.2 49.8 7.3 0.7 23 36 1 0 ensuring broader representation. The 2002 elections saw seven additional parties win seats in parliament, while the majority of seats were distributed between the BNP, the BCP and the LCD. Unlike previous attempts at electoral democracy, the 2002 elections were the first in Lesotho’s history that were not marked by violence and corruption, suggesting that Lesotho, like a number of other Sub-Saharan African countries, was consolidating its democratic gains after the violence and instability of the 1990s. Although the opposition challenged the results of the 2002 results, protests remained largely non-violent (Makoa, 2004), marking the first peaceful election without an opposition boycott since 1965 and the first since independence. This pattern has endured post-reform. Whereas the remnants of monarchy and Lesotho’s political and economic history would suggest an inhospitable environment for democratization, electoral reforms appear to have ushered in a stable and what appears to be a largely fair electoral process. This era of democratic competition is largely credited to the enactment of a mixed legislative system. As seen elsewhere, Lesotho’s mixed system incorporated single-member districts (SMDs) that had been used in previous elections in the country, as well as seats allocated by proportional representation (PR). In most mixed systems the results of the two seat types (or tiers) do not directly influence the other and can be viewed as independent. Such systems are often called parallel systems or mixed member majoritarian (MMM). In contrast, Lesotho follows Germany and New Zealand in enacting a mixed member proportional (MMP) system in which the overall results are proportional. Whereas under MMM the total results may be very disproportional if one party wins more district seats than their PR vote share, under MMP party list seats are allocated only after adjusting for district victories. As such, MMP guarantees proportionality beyond that of MMM, which is of particular importance for Lesotho because prior elections using just SMDs often failed to produce a representative assembly. The first two National Assembly elections (1965 and 1970) produced relatively proportional outcomes as seen in Table 1; both cases led to violent conflicts with opposition forces challenging the legitimacy of the elections. The return to multiparty competition in 1993 and 1998 created a heavily disproportional outcome leading to continued questions about the legitimacy of the country’s elections. Moving to a system which ensured proportionality, even while maintaining local districts, thus potentially provided a mechanism to encourage parties to view the results as legitimate. Interviews with representatives of the IEC reaffirmed that the broader goal of enacting the MMP system was to stop one-party dominance and increase representation. After the 1998 political violence and South African intervention, public debates held by the Interim Political Authority (IPA) in conjunction with the United Nations Development Program (UNDP) showed broad support for a mixed system as it would guarantee greater representation to opposition parties. The first election under MMP indeed produced a more representative legislature. The LCD obtained 77 of 118 seats with 54.8% of the vote, while the BNP captured 21 seats with 22.4% of the vote. Eight other parties received up to five seats. Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 7 Rich et al. The 2007 election continued this trend, with the LCD garnering 62 of the 120 seats. While the BNP collapsed winning only three seats, three other parties won 10 or more seats: the National Independent Party, All Basotho Convention (ABC) and Lesotho Workers’ Party. The 2012 elections included an additional institutional reform, passed under the National Assembly Elections Act of 2011, which replaced the two-vote system with a one-vote system. In such systems, also used in Mexico and previously in South Korea from 1996–2000, the results of district elections are aggregated to determine the party list, creating a contamination effect in contrast to Duvegerian expectations. Supporters of small parties, instead of strategically defecting in districts, cast a sincere vote in order to capture proportional seats. In addition, 2012 saw a split in the LCD, effectively resulting in three sizable parties: the new Democratic Congress (DC), the ABC and the LCD. While a cursory view of National Assembly elections certainly suggest a more representative parliament, it gives only limited insight into patterns of electoral competition. The new electoral institutions also remain devoid of a mechanism to resolve election-related conflict in a neutral and systematic way. The Independent Electoral Commission (IEC) since 2002 has been more transparent than previous elections (Kabemba, 2003; Venter, 2003) through a restructuring of electoral laws and regulations. Still, the IEC maintains a monopoly on dispute resolution and has prevented access to ballots and other voting materials necessary for effective challenges by domestic groups as well as international observers (Makoa, 2004). The investigations surrounding voter fraud allegations in 1998 highlighted this lack of transparency. This lack of transparent oversight coupled with a continuous registration process that potentially allows for inaccurate rolls provides several causes for concern. However, instead of delving into broad systemic problems in Lesotho, we opted to focus on electoral institutional impacts on voting behavior for three reasons. First, such a method requires fewer assumptions on beliefs and expectations of actors. Second, it makes our studies replicable. Third, given ease of replicability, it facilitates future comparisons with other countries using the MMP system. Research Design To better grasp whether elections in Lesotho are matching expectations in more mature democracies, we analyze national and district-level results from each of Lesotho’s mixed elections. Election results were culled from the website of Lesotho’s Independent Election Commission. This two-level approach allows us to identify the extent of competition in Lesotho, while providing additional evidence to match against theoretical expectations. National-level results, for example, often mask a very different district-level dynamic. Where a multitude of parties gain seats nationally, district competition may remain more heavily restricted. District-level analysis provides a better means to identify the degree of party fractionalization and to the extent that Lesotho’s mixed system behaves like those elsewhere. Furthermore, district-level analysis allows an indirect means to identify electoral fraud that it is simply not possible to capture at the national level. Based on the existing literature, this analysis tests the following hypotheses: Hypothesis 1: in line with the contamination thesis, district-level competition will not converge towards two-party competition. Hypothesis 2: district population will not correlate with vote concentration around the two top candidates. Hypothesis 3: district population and the number of candidates will both negatively correlate with turnout. Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 8 Journal of Asian and African Studies 0(0) Table 2. The effective number of parties in votes (ENPV) and seats (ENPV). 1965 1970 1985 1993 1998 2002 2007 2012 ENPV ENPS 2.79 2.32 Missing 1.64 2.27 2.78 Missing 3.69 2.14 1.97 1 1 1.03 2.16 3.15 3.67 Empirical Analysis A common measure for party-system competition is the effective number of parties, a weighted measure of all parties based on either vote shares or seat shares. While other factors beyond the electoral institution influence the number of parties (Mozaffar et al., 2003; Patterson and FadigaStewart, 2005), the literature on electoral institutions expects two-party competition in systems using just SMDs and multiparty competition when using PR. Table 2 summarizes the effective number of parties by votes (ENPV) and by seats (ENPS).3 Lesotho’s first two post-independence elections confirm Duvegerian expectations: the effective number of parties by vote share are under three, with a decline after the first election indicative of a learning curve. Seat share in contrast is much closer to Duvergerian expectations from the start. The following elections show sizable support for other parties, but this fails to convert to seats. Only with the move to a mixed system does the discrepancy between vote and seat share diminish, consistent with the proportionality inherent in MMP systems. With few exceptions, most countries see an increase in the number of parties when moving from a majoritarian system to a mixed system (e.g. New Zealand). Lesotho is consistent with these general trends. Moving to district-level competition, Nagayama triangles provide an effective means to graph competition and evaluate Duvergerian expectations. Nagayama (1997) noticed that graphing the vote percentage for winning candidates in single-member districts on the x-axis by the percentage received by the runner-up (which by definition cannot be more than 50%) created a triangle shaped distribution (also see Nishikawa, 2002). Regions within this triangle correspond with general trends in party competition (Diwakar, 2007; Reed, 2001, 2007; Taagepera and Allik, 2006). The top point of the triangle corresponds with competitive two-party competition where smaller parties receive little support, consistent with Duvergerian expectations in SMDs (e.g. 51% vs. 47%). Districts with multiparty competition and thus a larger percentage of the vote going to smaller parties are found as one approaches the left corner of the triangle (e.g. 47% vs. 35%). Districts where two parties receive nearly all of the vote, but where one party dominates, are located towards the right corner of the triangle (e.g. 85% vs. 15%). Figure 1 presents a Nagayama triangle comprised of district election results for 2002, 2007 and 2012. This includes the top two candidates regardless of party identification, including independents where applicable. Duverger would expect district results to concentrate around the top tip of the triangle if close two-party races and along the far right if one party dominates in two-party contests. Evidence from Lesotho shows greater dispersion, with a majority of districts having viable candidates beyond the top two. In 2002 a plurality of districts include a candidate that received a supermajority of the district vote, with greater dispersion by 2007. By 2012 multiparty Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 9 Rich et al. 2007 0 0 10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 2012 0 Runner_up 10 20 30 40 50 2002 0 10 20 30 40 50 60 70 80 90 100 Winner Figure 1. Nagayama triangles of Lesotho’s district elections by year competition with more than two effective parties was the norm. Every district had at least four candidates (including independents), with as many as 16 in 2002 and 21 in 2012. Many of these received only a small fraction of the vote, yet third parties performed better than would be expected in systems comprised solely of SMDs, suggestive of a contamination between the two tiers in the mixed system. The top two candidates received on average 80.26% of the district vote in 2002, with 81.8% and 72.31% in following elections. The move to a one-vote system in 2012 encouraged greater multiparty competition (towards the left side) as parties vying for party list seats had to run often non-viable district candidates. Yet this does not explain the dispersion under a two-vote system where strategic defection would be expected at the district level if Duvergerian pressures were dominant. The results suggest that parties and/or voters fail to understand the mechanical effects of district competition in the first two elections, relying more on personal connections to candidates rather than analysis of candidate viability. Meanwhile, the move to a one-vote system may have simply reinforced personal connections. Another measurement to identify strategic voting associated with Duverger is the Second Loser ratio, the percentage of the vote for third place divided by second place. If strategic voting is high (voters defect from non-viable candidates), then the ratio should approach zero whereas rates over .50 suggest limited or no strategic voting. The reasons for non-strategic voting under mechanical pressures to do so are several, notably ethnic loyalties or a lack of information regarding candidate viability. Due to data availability restrictions, this ratio could only be generated for 2002 and 2012. This produced non-intuitive results: the ratio increased from an average of .39 to .64. Instead of Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 10 Journal of Asian and African Studies 0(0) Table 3. Regression on vote concentration for the top two candidates (2002 and 2012). Model 1 District population Number of candidates 2012 elections Constant N Adjusted R2 Model 2 −0.0004 −1.8370*** 0.0002 0.1978 102.1093*** 158 0.4254 2.9269 0.0000 −1.7424*** −3.5370*** 97.9840*** 158 0.4398 0.0003 0.1998 1.5837 3.4299 ***p < .05; **p < .01; *p < .001. viewing these results as irrational, Lesotho voters may also be responding to an increase in the number of candidates and parties with difficulty in determining viability without extensive polling data common in Western democracies.4 In sum, evidence from Nagayama triangles and Second Loser ratios are consistent with the contamination thesis. While initial empirical evidence is consistent with the contamination thesis and similar to results in other MMP systems (e.g. Germany and New Zealand), it gives only limited insight into the district-level election dynamics. For example, to what extent does the size of the district population and the number of district candidates (including independents) influence the concentration of the vote around the top two candidates? Table 3 presents ordinary least squares (OLS) regression results. Unfortunately limited district-level data precludes additional variables to capture aggregate economic and social influences. In model 1 each additional candidate correlates with a near 2% drop in vote concentration, significant at the .001 level. This pattern endures after controlling for the most recent election (model 2), while district population fails to reach significance in either model.5 District-level turnout in mixed systems remains largely unexplored despite its importance in constructing meaningful representation. What determines ‘good’ turnout remains debatable (Blais and Aarts, 2006; Powell, 2000; Teixeira, 1992; Wattenberg, 2002), but the general assumption remains that increasing turnout positively influences perceptions of the legitimacy of election results. For our purposes, we wish to focus on how the number of parties or candidates influences turnout, especially since multiparty competition is the norm in Lesotho politics and increased in 2012. The conventional wisdom remains that additional candidates increase general interest in elections, yet this also increases the likelihood of the winner having less than a majority of the vote. Unfortunately we have limited data on turnout for the 2007 election and thus again must focus on the first and last elections under a mixed system. Table 4 presents OLS regression results on the percentage of registered turnout. Models 3 and 4 use the same independent variables as the previous models. In the first, each additional candidate decreases turnout by a paltry 0.5%, significant at the .01 level. Once a control for the 2012 election is included (capturing in part the tendency for lower turnout over time), additional candidate entries fail to reach statistical significance. Meanwhile, in both models population size negatively correlates with turnout, significant at .01 or better. Overall, these results are consistent with the general trend in mixed systems of a decline in turnout. Whether due to disenchantment with the process as campaign pledges fail to materialize into concrete policy shifts or a more manipulative origin such as voter intimidation, Lesotho’s district-level results do not appear to be substantively distinct from those from similar electoral institutions elsewhere. Although Lesotho’s results may appear largely consistent with expectations from other mixed systems, the above analysis on its own does not allay concerns regarding whether the elections are Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 11 Rich et al. Table 4. Regression on registered turnout (2002 and 2012). Model 3 District population Number of candidates 2012 elections Constant N Adjusted R2 Model 4 −0.0022*** −0.4764** 0.0002 0.1886 91.8421*** 158 0.4873 2.7911 −0.0005** −0.0644 −15.4083*** 73.8710*** 158 0.8218 0.0002 0.1138 0.9017 1.9528 ***p < .05; **p < .01; *p < .001. free and fair. Freedom House labeled Lesotho’s political system as ‘partially free’ from 1991 through 2001 and upgraded the country to ‘free’ in 2002–2008. However, since 2009 Lesotho has been labeled ‘partially free’ again. Admittedly much more goes into Freedom House’s measure than electoral transparency, but this does give us some concern about whether Lesotho is engaged in democratic backsliding. To tackle this concern, we return to district-level electoral data to identify electoral fraud. A growing literature tackles identifying electoral irregularity by analyzing patterns in the digits of subnational results (e.g. precincts; see Mebane 2006; Mebane and Sekhon 2004; Wand et al. 2001). Beber and Scacco (2012) present several methods to use subdistrict-level data to identify fraud, but the clearest means is to evaluate the ending digit in election results. In an election free of manipulation, there should be no systematic difference in the rates of ending digits. In other words, a record of precinct-level results should have roughly the same amount of precincts ending in 0s as 1s and 2s and so on. Existing evidence suggests human biases in generating numbers intended to appear random (Mosimann et al., 1995, 2002; Mosimann and Ratnaparkhi, 1996). Results that were beyond a standard deviation and especially those beyond two standard deviations would imply electoral fraud, although this does not identify the source of such manipulation. Beber and Scacco use this to suggest Sweden’s 2002 election shows no evidence of fraud (all rates are within one standard deviation), whereas 2003 data from Nigeria find 0s overrepresented and 2s underrepresented at similar rates, beyond two standard deviations, suggestive of electoral fraud. Precinct-level data was not available for Lesotho, which potentially would allow us to identify concentration of fraud within in a district. The IEC states that roughly 3000 polling stations blanket the country, with each designed to handle roughly 600 voters. Furthermore, electoral disputes are resolved at this level by IEC officials.6 While this level of analysis would be most appropriate in identifying electoral fraud as well as variation within a district, such data were not provided by the IEC after our requests. However, we did have district-level returns for 2002 and 2012, including the total number of ballots as well as a breakdown of valid versus invalid ballots.7 The same logic of random numbers employed by Beber and Scacco should apply here as well. Figures 2 and 3 display the frequencies of ending digits of valid, invalid, and total votes for district elections for 2002 and 2012. If randomly distributed, we would expect each number, 0–9, to occur roughly eight times (as there are 80 districts). While some variation would be expected, results beyond one and especially two standard deviations from the mean would be cause for concern (one standard deviation from the mean is demarcated with dark lines). Starting with 2002, we find that, among total votes, only one ending digit (5) is counter to expectations (N = 16), two standard deviations from the mean. When disaggregated by valid and invalid votes, several reach one standard deviation (1 and 2 for valid; 2, 3, and 7 for invalid). Moving onto 2012, among total votes two ending digits are beyond one standard deviation, 4 (N = 13) and 9 (N = 2), but neither reaches two standard Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 12 Journal of Asian and African Studies 0(0) 18 16 14 12 Valid votes 10 8 Invalid votes 6 Total votes 4 2 0 0 1 2 3 4 5 6 7 8 9 Figure 2. Rates of ending digits in 2002 district elections Note: The black horizontal lines denote the range of one standard deviation from the mean. 14 12 10 8 Valid votes Invalid votes 6 Total votes 4 2 0 0 1 2 3 4 5 6 7 8 9 Figure 3. Rates of ending digits in 2012 district elections Note: The black horizontal lines denote the range of one standard deviation from the mean. deviations. Among valid votes, half of the digits are beyond one deviation (1, 5, 6, 8, and 9). Meanwhile, among invalid votes, three ending digits are beyond one deviation (1, 5 and 7), while one falls beyond two deviations (6, N = 12). These results fail to provide a smoking gun for electoral fraud similar to what Beber and Scacco find for Nigeria. The variation may be a statistical anomaly in part related to the small sample size or due to poor data collection by Lesotho’s election commission. Variation within one standard deviation is to be expected and the evidence beyond this in total votes cast is minimal. As such, our results are consistent with those of the Electoral Institute for Sustainable Democracy in Africa (EISA), whose observers declared Lesotho’s 2012 election as consistent Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 13 Rich et al. with legal requirements.8 Thus, despite a history of disputed elections and an election commission lacking greater transparency, indirect evidence suggests a relatively free and fair election process. A second measure to uncover fraud is to use regression analysis of the reported district turnout of registered voters (as a percentage) on the last digit reported at the district level. As the two should not have any statistical relationship, evidence of any systematic pattern would indicate manipulation. Separate regressions for 2002 and 2012 using the last digit for total votes casts, valid votes, and invalid votes find no evidence of such manipulation. Only one model, using valid votes from 2002, even approaches statistical significance (p = .103). Taken as a whole, the empirical analysis fails to find evidence of electoral fraud. While admittedly electoral fraud is difficult to identify from a distance and experienced fraudsters consistently find new ways to manipulate the system, these initial findings should be viewed as confirmation of Lesotho’s evolution to a more stable democratic system. Conclusion A large literature on electoral systems provides broad expectations on how electoral institutions shape outcomes. Yet much of this literature is based on older Western democracies, giving limited insight into institutional reforms in younger cases. An analysis of Lesotho provides additional evidence consistent with the contamination thesis, as district-level competition has failed to whittle down to only two electorally viable parties as expected by Duverger. While Duvergerian equilibrium may not occur immediately as parties and candidates learn the new system, evidence from Lesotho suggests either an absence of the theoretically expected pressures or that additional factors override these concerns. Why additional parties and candidates endure is unclear. A fused ballot would encourage what normally appears as irrational voting in districts, but this would only apply to the latest election.9 Cultural ties or the lack of pre-election data of public support levels would also be consistent with this behavior. Each additional entry pulls votes away from the top two candidates, with limited evidence of this addition depressing turnout. Furthermore, additional analyses of district-level results in 2002 and 2012 present minimal evidence of potential electoral manipulation in terms of falsifying electoral returns. While democracy appears to be growing roots in Lesotho, several challenges remain. First, party system instability endures. The move to a mixed system has done little to curb party switching. For example, Matlosa and Shale (2006) found that 35 seats shifted after the 2002 election. Similarly, public trust in parties remains low and has not improved over time, declining in fact with the 2007 election. Second, the impact of moving to a one-vote mixed system cannot be understood entirely after only one election, especially one which occurs simultaneously with the split of Lesotho’s ruling party. Still, the mixed system appears to provide both winners and losers with enough influence to forego extralegal means to gain power. Although the ‘best of both worlds’ literature suggests a welcome balance of national and local representation, an analysis of Lesotho suggests that enacting mixed systems in weakly institutionalized democracies encourages parties to accept electoral outcomes while providing a degree of proportionality. Notes 1. Based on the experience of South Africa’s pacted transition from apartheid and the African National Congress’ (ANC) subsequent rise to power, some argued that dominant parties represent class coalitions that sustain power by balancing between the redistribution required from their proletarian supporters and the property rights required by the bourgeoisie to make wealth for this redistribution (Giliomee, 1998; Simkins and Giliomee, 1999: 338–341). Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 14 2. 3. 4. 5. 6. 7. 8. 9. Journal of Asian and African Studies 0(0) Chhibber and Kollman (2004) found that India, Britain, and Canada traditionally have over two effective parties at the district level. In addition, Papua New Guinea consistently has five or more effective parties per district in large part due to the relative isolation of villages within each district. Unfortunately the precise number of votes received by party in 2007 is unavailable. Rather than estimate this from existing records, we opted to simply note this data as missing. Since all other parties but the Basotholand National Party boycotted the 1985 election, ENPV cannot be reliably measured as well. The average number of district candidates increased from 2002 to 2012 from 10.4 to 13. An additional test using the Second Loser ratio as the dependent variable finds that both population and the number of candidates positively correlate with ratio at .05 or stronger, but fail to reach significance once a control for 2012 is included. According to an interview with Mrs. Mamatlere Matete, spokesperson for the IEC, there were no significant cases of electoral fraud, with such accusations resolved at the voting station. Admittedly, though, electoral fraud may not be patently obvious at the polling station (interview, 14 September 2012). We would expect a high rate of invalid votes if widespread voter ignorance about how to cast a ballot existed. As a percentage of all votes casts, invalid votes were low in both 2002 (3.23%) and 2012 (1.68%).This suggests a basic understanding of the system. EISA’s post-election report does suggest some concern regarding the secrecy of ballot, observing that the placement of voting booths potentially allow parties and other observers to identify how individuals voted. One-vote systems incentivize voting for non-viable district candidates in order to secure party list seats for one’s preferred party. As such, this institutional reform potentially benefits parties that can run in every district. References Amin Samir (1972) Underdevelopment and dependence in Black Africa: Origins and contemporary forms. Journal of Modern African Studies 10(4): 503–524. Amorim-Neto O and Cox GW (1997) Electoral institutions, cleavage structures, and the number of parties. American Journal of Political Science 41(1): 149–174. Bartolini S (2002) The political consequences of the Italian mixed electoral system. Paper presented at the Conference on Elections and Democracy, Social Science Institute, University of Lisbon, 1–2 February. Bawn K (1999) Voter responses to electoral complexity: Ticket splitting, rational voters, and representation in the Federal Republic of Germany. British Journal of Political Science 29: 487–505. Beber B and Scacco A (2012) What the numbers say: A digit-based test for election fraud. Political Analysis 20(2): 211–234. Blais A and Aarts K (2006) Electoral systems and turnout. Acta Politica 41: 180–196. Blais A and Dobrzynska A (1998) Turnout in electoral democracies. European Journal of Political Research 33: 239–261. Brockington D (2004) The paradox of proportional representation: The effect of party systems and coalitions on individuals’ electoral participation. Political Studies 52(3): 469–490. Chhibber P and Kollman K (2004) The Formation of National Party Systems: Federalism and Party Competition in Canada, Great Britain, India, and the United States. Princeton, NJ: Princeton University Press. Cho W and Bratton M (2006) Electoral institutions, partisan status, and political support in Lesotho. Electoral Studies 25: 731–750. Cox GW (1997) Making Votes Count: Strategic Co-ordination in the World’s Electoral Systems. Cambridge: Cambridge University Press. Cox KE and Schoppa LJ (2002) Interaction effects in mixed-member electoral systems. Comparative Political Studies 35(9): 1027–1053. Diwakar R (2007) Duverger’s law and the size of the Indian party system. Party Politics 13: 539–561. Duverger M (1954) Political Parties: Their Organization and Activity in the Modern State. London: Methuen and Company. Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 15 Rich et al. Duverger M (1986) Duverger’s law: Forty years later. In: Grofman B and Lijphart A (eds) Electoral Laws and Their Political Consequences. New York: Agathon Press, 69–84. Elklit J (2002) Lesotho 2002: Africa’s first MMP elections. Journal of African Elections 1(2): 1–10. Ferrara Federico and Herron Erik S (2005) Going it alone? Strategic entry under mixed electoral rules. American Journal of Political Science 49(1): 16–31. Ferrara F, Herron ES and Nishikawa M (2005) Mixed Electoral Systems: Combination and its Consequences. New York: Palgrave. Franklin MN (2004) Voter Turnout and the Dynamics of Electoral Competition in Established Democracies since 1945. Cambridge: Cambridge University Press. Franklin M, Lyons P and Michael M (2004) Generational basis of turnout decline in established democracies. Acta Politica 39(2): 115–151. Giliomee H (1998) South Africa’s emerging dominant party regime. Journal of Democracy 9(4): 128–142. Golder SN (2006) Pre-electoral coalition formation in parliamentary democracies. British Journal of Political Science 36(2): 193–212. Herron ES and Nishikawa M (2001) Contamination effects and the number of parties in mixed superposition electoral systems. Electoral Studies 20(1): 63–86. Jackman RW (1987) Political institutions and voter turnout in industrial democracies. American Political Science Review 81: 405–424. Johnston Deborah (1996) The state and development: An analysis of agricultural policy in Lesotho, 1970– 1993. Journal of Southern African Studies 22(1): 119–137. Jones MP (1999) Electoral laws and the effective number of candidates in presidential elections. Journal of Politics 61(1): 171–184. Kabemba Claude (2003) From military rule to multiparty democracy: Political reforms and challenges in Lesotho. Electoral Institute of Southern Africa: Research Report 2: 1–87. King G (1990) Electoral responsiveness and partisan bias in multiparty democracies. Legislative Studies Quarterly 15(2): 159–181. Kostadinova T (2003) Voter turnout dynamics in post communist Europe. European Journal of Political Research 42(6): 741–759. Lancaster TD and Patterson WD (1990) Comparative pork barrel politics: Perceptions from the West German Bundestag. Comparative Political Studies 22(4): 458–477. Lijphart A (1994) Electoral Systems and Party Systems. Oxford: Oxford University Press. Makoa FK (2004) Electoral reform and political stability in Lesotho. African Journal on Conflict Resolution 4(2): 79–95. Matlosa K (2006) Electoral system design and conflict mitigation: The case of Lesotho. In: Institute for Democracy and Electoral Assistance (IDEA) (ed.) Democracy, Conflict and Human Security: Pursuing Peace in the 21st Century, Vol. 2. Pretoria: Institute for Democracy and Electoral Assistance (IDEA), 94–110. Matlosa K and Shale V (2006) Impact of Floor Crossing on Party Systems and Representative Democracy: The Case of Lesotho. Paper presented for a Workshop on ‘Impact of Floor Crossing on Party Systems and Representative Democracy in Southern Africa,’ Co-hosted by EISA and KAS, The Vineyard Hotel, Cape Town, South Africa, 15 November. Revised 29 November. McCartney WJA (1973) The Lesotho general election of 1970. Government and Opposition 8(4): 473–494. Mebane WR Jr (2006) Election Forensics: Vote Counts and Benford’s Law (working paper). Mebane WR Jr and Sekhon JS (2004) Robust estimation and outlier detection for overdispersed multinomial models of count data. American Journal of Political Science 48(2): 392–411. Moser RG (1995) The impact of the electoral system on post-communist party development: The case of the 1993 Russian parliamentary elections. Electoral Studies 14(4): 377–398. Moser RG (1997) The impact of parliamentary electoral systems in Russia. Post-Soviet Affairs 13(3): 284– 302. Moser RG and Scheiner E (2004) Mixed electoral systems and electoral system effects: Controlled comparison and cross-national analysis. Electoral Studies 23: 575–599. Mosimann JE, Dahlberg JE, Davidian NM and Krueger JW (2002) Terminal digits and the examination of questioned data. Accountability in Research 9(2): 75–92. Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 16 Journal of Asian and African Studies 0(0) Mosimann JE and Ratnaparkhi MV (1996) Uniform occurrence of digits for folded and mixture distributions on finite intervals. Communications in Statistics – Simulation and Computation 25(2): 481–506. Mosimann JE, Wiseman CV and Edelman RE (1995) Data fabrication: Can people generate random digits? Accountability in Research 4(1): 31–55. Mozaffar S, Scarritt JR and Galaich G (2003) Electoral institutions, ethnopolitical cleavages, and party systems in Africa’s emerging democracies. American Political Science Review 97(3): 379–390. Nagayama M (1997) Shousenkyoku no Kako to Genzai [The past and present of SMDs]. Paper presented at the annual conference of the Japan Political Science Association, Tokyo. Niemi RG and Weisberg HF (eds) (2001) Controversies in Voting Behavior. Washington, DC: CQ Press. Nishikawa M (2002) Does the Duverger’s law really work? Public Choice Studies (Japanese) 38: 19–30. Norris P (2004) Electoral Engineering: Voting Rules and Political Behavior. New York: Cambridge University Press. Olaleye W (2004) Democratic consolidation and political parties in Lesotho. The Electoral Institute of Southern Africa (EISA) Occasional Paper, No. 15, December. Ordeshook P and Shvetsova O (1994) Ethnic heterogeneity, district magnitude, and the number of parties. American Journal of Political Science 38: 100–123. Patterson D and Fadiga-Stewart L (2005) The Strategy of Dominant Party Politics: Electoral Institutions and Outcomes in Africa. Paper presented at the annual meeting of the American Political Science Association, Marriott Wardman Park, Omni Shoreham, Washington Hilton, Washington, DC, 1 September. Available at: http://www.allacademic.com/meta/p41406_index.html (accessed 10 September 2008). Pempel TJ (1990) Conclusion. In: Pempel TJ (ed.) Uncommon Democracies: The One Party Dominant Regimes. Ithaca, NJ: Cornell University Press, 333–360. Pherudi ML (2001) The Lesotho Liberation Army (LLA): Formation, mission and schisms. South African History Journal 45(1): 266–277. Powell GB (2000) Elections as Instruments of Democracy: Majoritarian and Proportional Visions. New Haven, CT: Yale University Press. Rae Douglas (1971) The Political Consequences of Electoral Laws, 2nd edn. New Haven, CT: Yale University Press. Rae Douglas (1995) Using district magnitude to regulate political party competition. Journal of Economic Perspectives 9(1): 63–75. Reed SR (2001) Duverger’s law is working in Italy. Comparative Political Studies 34: 312–327. Reed SR (2007) Duverger’s law is working in Japan. Japanese Journal of Electoral Studies 22: 96–106. Reynolds A, Reilly B and Ellis A (eds) (2005) Electoral System Design: The New International IDEA Handbook. Stockholm: International Institute for Democracy and Electoral Assistance. Rich T (2013) Institutional Influences on Turnout in Mixed Member Legislative Systems: An Exploratory Analysis. Unpublished paper. Rokkan S (1970) Citizens, Elections, Parties. Oslo: Universiteteforlaget. Sartori G (1976) Parties and Party Systems: A Framework for Analysis. Cambridge: Cambridge University Press. Sartori G (1997) Comparative Constitutional Engineering: An Inquiry into Structures, Incentives and Outcomes. New York: New York University Press. Shugart M and Wattenberg MP (eds) (2001) Mixed-Member Electoral Systems: The Best of Both Worlds? Oxford: Oxford University Press. Simkins C and Giliomee H (1999) Conclusion. In: Simkins C and Giliomee H (eds) The Awkward Embrace: One Party-Domination and Democracy. Cape Town: Tafelberg Publishers, 338–341. Stratmann T and Baur M (2002) Plurality rule, proportional representation, and the German Bundestag: How incentives to pork barrel differ across electoral systems. American Journal of Political Science 46(3): 506–514. Taagepera R and Allik M (2006) Seat share distribution of parties: Models and empirical patterns. Electoral Systems 25: 696–713. Taagepera R and Shugart M (1989) Seats and Votes. New Haven, CT: Yale University Press. Taagepera R and Shugart M (1993) Predicting the number of parties: A quantitative model of Duverger’s mechanical effect. American Political Science Review 87(2): 455–464. Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013 17 Rich et al. Tangri Roger (1993.) Foreign business and political unrest in Lesotho. African Affairs 92(367): 223–238. Teixeira R (1992) The Disappearing American Voter. Washington, DC: Brookings. Turkon D (2008) Commoners, kings, and subaltern: Political factionalism and structured inequality in Lesotho. POLAR: Political and Legal Anthropology Review 31(2): 203–223. Venter Denis (2003) Democracy and multiparty politics in Africa: Recent elections in Zambia, Zimbabwe, and Lesotho. Eastern Africa Social Science Research Review 19(1): 1–39. Wand JN, Shotts KW, Sekhon JS, Mebane WR Jr, Herron MC and Brady HE (2001) The butterfly did it: The aberrant vote for Buchanan in Palm Beach County, Florida. American Political Science Review 95(4): 793–810. Wattenberg MP (2002) Where Have All the Voters Gone? Cambridge, MA: Harvard University Press. Author biographies Timothy Rich is an Assistant Professor of Political Science at Western Kentucky University. His research focuses on electoral behavior in mixed-member legislative systems with a focus on East Asia. Vasabjit Banerjee is a Post-Doctoral Fellow and Visiting Assistant Professor in the Department of International Studies at Indiana University, Bloomington and a Research Associate at the Department of Sociology at the University of Pretoria. His research covers social movements, political violence, regime formation, and democratization in developing countries with a focus on South Asia, Latin America, and southern Africa. Sterling Recker is a Doctoral Candidate at the University of Missouri, St. Louis and Adjunct Professor in the Department of Political Science at Southern Illinois University. His research focuses on the politics of rural development and natural resources in Africa and Latin America, with a specific interest in the Great Lakes region of Africa and the Andean region of Ecuador. Downloaded from jas.sagepub.com at INDIANA UNIV on November 12, 2013