Pre-electoral coalitions form across a variety of political contexts. Prevailing explanations suggest that ideological proximity between coalition partners and the size of the contribution that parties make explain the character of coalitions. These expectations hold only partially true in the case of Mexico. Rather, as this article suggests, parties also evaluate the degree to which they consider prospective coalition partners reliable and trustworthy. Some otherwise viable coalitions fail to form because of lack of trust or form despite ideological disparities, when a party’s contribution to defeating a common enemy is considered.

Las coaliciones preelectorales se forman en una variedad de contextos políticos. Las teorías más aceptadas sobre las coaliciones sugieren que la proximidad ideológica entre los componentes de la coalición, así como el tamaño de los partidos a los que representan, explican el carácter de dichas coaliciones. Sin embargo, estas teorías son sólo parcialmente ciertas en el caso de México. En este artículo, sugiero que los partidos políticos también evalúan el grado en que los posibles participantes de una coalición son considerados dignos de confianza. De ese modo, algunas coaliciones, que de otro modo serían viables, no se forman debido a la falta de confianza, mientras que otras se forman a pesar de las disparidades ideológicas cuando se considera la importancia de la contribución de un partido aliado para poder derrotar a un enemigo común.

In July 1999, the leftist Partido de la Revolución Democrática (PRD) proposed to the conservative Partido Acción Nacional (PAN) that the two parties form a national opposition alliance for the 2000 elections, with the express purpose of finally ending the long monopoly of the Partido Revolucionario Institucional (PRI) over the presidency. It was, at its core, an eminently pragmatic suggestion. The PRI had never conceded defeat in a presidential election, and although both the PAN and the PRD had reasonably attractive candidates to offer, there was no reason—at the time—to suspect that PRI president Ernesto Zedillo would pre-empt not only his own party but also the Federal Electoral Institute in conceding to the PAN’s Vicente Fox on election night. Nevertheless, the suggestion was met by the PAN with total rejection, followed by public criticism of the PAN’s reliability by the PRD (Zuñiga 2000). Neither party would concede the head of the ticket to the other, in part because of their ideological differences.

A decade later, PAN-PRD coalitions had become more common, though still a minority of all electoral coalitions. Between 2000 and 2015, 11% of the legislative coalitions in which the PRD participated included the PAN as a partner. These coalitions were effective, winning in 39.4% of the districts where they competed (Olmeda and Devoto 2019, 314). Coalitions between the PRD and the PAN also contested governorships in nineteen states and won eleven (58%). In 2018, the once unthinkable PRD-PAN presidential coalition formed behind Alberto Anaya (a panista) against the candidacy of former PRD party president Andrés Manuel López Obrador (AMLO). Meanwhile, AMLO’s new party, Movimiento de Regeneración Nacional (Morena), chose as one of its coalition partners the Partido Encuentro Social (PES), an evangelical party even farther to the right than the PAN. Finally, in the midterm legislative elections of 2021, a triple alliance, including the PRD, the PAN, and the once anathema PRI, formed to contest the aspirations of Morena to control a supermajority in the Mexican national legislature.

What accounts for these seemingly unnatural electoral coalitions? Does ideological compatibility influence the choice of coalition partner? If pragmatic concerns explain these ideologically diverse coalitions, why do such coalitions form in some cases and not in others? And what are the implications of pragmatic rather than programmatic alliances for the future of the Mexican party system?

This article assesses the motivations of political parties to form pre-electoral coalitions, using the case of Mexico. It examines two major hypotheses: that ideological proximity explains the likelihood of an electoral alliance; and that pragmatic minimum-winning coalitions form, despite ideological differences, when they seem likely to improve the odds of victory. This study does this primarily by looking at the records of gubernatorial races in Mexico between 1994 and 2018. These elections are good cases for examining patterns of alliance formation because they look at executive positions (which cannot be effectively subject to ticket balancing) and because there are more of them than presidential coalitions, allowing some quantitative exploration. I first determine the ideological positions of the parties and then code for ideological proximity in each electoral coalition. Second, I look at whether coalitions were more likely when, based on the prior gubernatorial election’s results, parties should have anticipated that these coalitions would have a good chance of winning—in other words, whether likely victory increases the probability of pragmatic coalition formation.

The article finds that ideology strongly influences the selection of coalition partners; indeed, roughly 75% of electoral coalitions that formed in gubernatorial elections in Mexico between 1994 and 2018 involved ideologically compatible coalition partners. Parties also consider the electability of a coalition versus running alone, and for all parties, coalitions became relatively more common. Yet neither approach fully explains the timing or geographical dispersion of these coalitions. Oversized coalitions occurred regularly, and many potential winning coalitions failed to form. I therefore turn to interviews about party motivations in the formation of pre-electoral coalitions in 2018 for additional clues.

One factor emerges from these interviews as potentially significant, and that is the role of trust. The departure of AMLO from the ranks of the PRD left the party in the hands of his most bitter (internal) rivals. He did not trust them to back him up; indeed, his suspicions about their reliability lay behind his decision to create his own party. He was unwilling even to consider an electoral coalition with the PRD, despite their ideological affinities. Likewise, the history of the PRI’s hegemony long ruled it out as a coalition partner for the PAN or the PRD, despite the fact that, as a centrist party, it was closer to either opposition party than they were to each other.

Just as important, however, is the opposite of trust: at first, suspicion of the PRI and later a growing suspicion of AMLO and Morena. The PRD’s reluctance to endorse AMLO and its support of a panista instead had an obvious and immediate cost: massive defections from its own base that turned it into a shell of a party. And the unprecedented coalition among all three of the (formerly main) parties—PAN, PRD, and PRI—in the 2021 legislative elections cannot be explained except in terms of the fears roused by the prospect of further constitutional reforms on the part of Morena’s volatile leader, with implications for the future of Mexican democracy itself. In fact, while it is still too early to confirm, we may be witnessing the replacement of the older PRI/anti-PRI split, with a new cleavage dividing the more established and pragmatic parties from the emerging populism of Morena.

One of the most consequential decisions that political parties make is the selection of their candidates. Running candidates allows parties to publicize their goals and principles, to build the party brand, and potentially to win access to office. Presumably, party members prefer their own party to any other and prioritize its success. Yet, under some circumstances, party leaders throw their support behind another party, forming a pre-electoral coalition and giving up the opportunities for party building that result from running their own candidate.

The most comprehensive work on the topic of pre-electoral coalitions comes from Sona N. Golder (2005, 2006a, 2006b), who argues that there is an institutional context that creates incentives for coalitions. Systems that reward larger parties (such as plurality or majoritarian systems) but have a significant number of parties (such as in federal systems) develop more pre-electoral coalitions.

When it comes to predicting which coalitions actually form, the literature suggests two major hypotheses based on slightly different assumptions about the primary goals of parties. First, if parties are mostly policy seeking, then ideological proximity between parties should predict coalition partners. Ian Budge and Hans Keman, for example, argue that parties will seek to reduce policy disagreements within the governing coalition “either by ensuring that ideologically contiguous parties form the government (a minimum connected winning coalition) or by reducing overall diversity—even though this may sometimes involve ‘jumping’ small neighboring parties” (1990, 17).

Focusing on pre-electoral coalitions, Golder suggests that “the ideological compatibility constraint facing potential coalitions is likely to be stronger prior to the election than afterwards…because voters might be unwilling to vote for electoral coalitions comprising parties with incompatible policy preferences” (2006b, 196) Thus, “pre-electoral coalitions are more likely to form when parties are ideologically compatible” (2006a, 9) Similarly, Marisa Kellam argues that “candidates and their parties are unlikely to form pre-electoral agreements with parties from the opposite end of the ideological spectrum” (2017, 397). Raimondas Ibenskas also finds that “electorally incompatible coalitions are less likely to form” (2016, 751)

Indeed, “coalitions between parties with extremely disparate policy platforms may well be subadditive. Even if the party leaders were willing to form a coalition, voters might reject it” (Golder 2006a, 41). Thomas Gschwend and Marc Hooghe find that “the more congruent the ideological positions of the coalition partners…the more likely it is that supporters of the constituent parties will support the PEC [pre-electoral coalition] at the polls” (2008, 560). Yan de Souza Carreirão and Fernanda Paula do Nascimento likewise argue that “candidates launched by ideologically inconsistent parties” may be punished by electors (2010, 75).1 And Marc Debus and Jochen Müller suggest that when governing coalitions form that do not match voter preferences, voters may punish the government in the next election (2013, 1008–9). Thus, there are incentives to form ideologically compatible coalitions and disincentives to form ideologically diverse ones.

The second hypothesis views parties primarily as office seekers and points to size as a key predictor of coalition participation. Parties participate in coalitions when they can contribute a meaningful number of seats to form a governing majority. Moreover, in William H. Riker’s famous formulation, “in n-person, zero sum games, where side-payments are permitted, where players are rational, and where they have perfect information, only minimum winning coalitions occur” (1962, 32); that is, since players expect coalitions to cost them something in terms of side payments (such as cabinet portfolios), they should try to minimize the cost by including only those parties that are necessary to construct a legislative majority.

Some scholars combine size with ideological criteria. Craig Volden and Clifford J. Carrubba (2004) argue that oversized minimum-connected coalitions form in order to facilitate policymaking through logrolling. Holger Döring and Johan Hellström, in contrast, found that while both policy and size mattered in the formation of Western European governments, only size and electoral performance mattered in Eastern Europe, with “little indication that policy positions have a direct effect on the probability of getting into office” (2013, 697). Looking specifically at pre-electoral coalitions in presidential systems, Aline Machado finds that parties “form more potential coalitions with those partners they considered crucial to win,” but they do so “at the expense of policy” (2009, 87).

Translating the size expectation into the context of pre-electoral coalitions is problematic because it is not clear in advance what will constitute a “minimum winning coalition.” Machado uses actual electoral results to determine post hoc whether a specific pre-electoral coalition was minimum winning. Parties cannot do this. Thus, Riker’s criteria of perfect information about the size of each party’s contribution will not be met. Before the election, parties may look at polls or previous electoral results as estimators for the probable electoral contribution of each coalition partner but cannot know with certainty how it will poll. We might expect more oversized coalitions among pre-electoral coalitions than among governing coalitions, in an effort to compensate for this uncertainty.

Analyses from Mexico find evidence for both policy seeking and office seeking but differ as to which predominates. Irma Méndez de Hoyos argues that ideological proximity is “a relevant factor in the process of strategic coordination between parties” (2012, 147), but Diego Reynoso (2010, 2011a, 2011b) argues that pragmatism rules and that parties cooperate whenever they smell a victory. In addition, he stresses the existence of an alternative anti-regime dimension to competition that explains some “unnatural” coalitions, particularly between the PAN and PRD, when these two parties see an opportunity to defeat the PRI. Lisandro Martín Devoto and Juan Cruz Olmeda (2017), meanwhile, suggest that parties are very nearly indiscriminate as they choose partners across states, time frames, and elections.

A third hypothesis derives from the long-held assumption in the study of parliamentary governments that coalition formation requires making side payments to subordinate parties—usually cabinet seats or policy concessions—in exchange for legislative support. Some scholars have argued that party competition in Mexico has become a nested game in which the smaller parties actively seek coalitions because of side payments. Reynoso (2011b) argues that parties participate in coalitions in order to remain viable for the purpose of receiving part of Mexico’s generous financial assistance to parties. Similarly, Ernesto Calvo, Fernando Guarnieri, and Fernando Limongi (2015) assert that small parties in Brazil are paid for their participation in nominations for legislative office to a greater extent than their contribution to the coalition vote would warrant. Euncheol Shin (2019) proposes yet another type of side payment: when vice-presidential nominations are possible, pre-electoral coalitions become more likely.

Finally, I hypothesize that coalitions are more likely to form when the participating parties have at least a minimal level of trust in their prospective partners. The dominant approach in political science evaluates citizen trust as a quality based on evaluation; thus, citizens “grant and withhold trust in politics based on an assessment of its merits” (Van der Meer and Hakhverdian 2017, 81). When a political system (regardless of its regime type) performs well in delivering the benefits that citizens expect, they reward it with greater trust. On the contrary, when political systems perform badly, we can expect citizens to trust it less. The discussion of which specific outcomes matter most to the development of citizen trust remains hotly contested. The concept of trust by political party leaders remains much less well developed. However, I propose here an essentially evaluative approach. Specifically, party leaders will evaluate the suitability of a coalition partner based upon how that party has previously behaved in two contexts: previous electoral coalitions (including attempts to form a coalition that fell apart) and legislative cooperation. Parties that have previously rejected efforts to form a coalition, or that have pursued legislative aims at odds with the efforts of the other potential partner, should be evaluated more negatively than parties that are either demonstrably willing to cooperate legislatively or have formed coalitions before that were successful. This is not the only factor that matters in the formation of electoral coalitions, but it is a relevant factor.

My hypotheses are thus the following:

  1. Parties seek ideologically similar partners to form coalitions.

  2. Parties will form pragmatic coalitions when the sum of two or more parties is needed to win elections.

    • Coalitions are more likely to form against the PRI when the total vote of the participating coalition parties in the previous election would have won.

    • Parties form minimum-winning coalitions (only as big as needed to win).

  3. Parties negotiate coalitions on the basis of side payments.

  4. Parties are more likely to form coalitions with partners with whom they have interacted successfully in the past and to avoid coalitions with partners with whom they have had previous conflicts.

To analyze these hypotheses, I draw primarily from evidence from 128 gubernatorial elections in Mexico, 4 per state, plus the Federal District/Mexico City. The earliest election took place in 1994 and the most recent in 2017. In all, 176 coalitions formed to compete in those elections. The quantitative analysis of these elections suggests that ideological considerations predominate. Some pragmatic coalitions form, but they do not form automatically even when the odds of victory seem promising, and many oversized coalitions also form. In order to better understand why such pragmatic coalitions formed in some cases but not in others, I then draw on interviews with high-ranking members of political parties to analyze how the parties see the negotiation of pre-electoral coalitions, with specific reference to the 2018 presidential election. Although the stakes are certainly much higher in a presidential election than in a governor’s race, both types of election focus on the executive office and therefore should have similar dynamics.

Alliances and Ideological Proximity

The first hypothesis expects alliances to form primarily between ideologically compatible parties. In order to test this hypothesis, I coded the ideological positions of each of the Mexican parties using the Comparative Manifestoes Project protocols. I secured electoral platforms for the intermediate congressional elections for 2015 and 2009 from the website of the Instituto Nacional Electoral (n.d.; INE). Earlier platforms were not available from INE, so I was unable to obtain platforms for 2003. However, I previously collected platforms from 1997 for all of the parties. We thus have relevant evidence from three separate moments in time when the parties produced independent platforms. Although subnational parties may be distinct from the national parties in at least some respects (as Reynoso [2015] suggests), using the platforms issued by subnational coalitions—even if available—would introduce an element of endogeneity into the analysis, as such platforms would be created in part in service of the coalition itself.2 The platforms I used are logically and empirically distinct from the gubernatorial elections and thus constitute an independent source of information about the parties’ relative ideological positioning across time, even if some information about subnational variation is lost. Indeed, to the degree that we find substantial ideological agreement in coalitions based on national platforms, we can be more confident that subnational ideological convergence in coalition construction really exists.

The next step in the Comparative Manifestoes approach is to break each platform into discrete “quasi sentences”—each expressing a particular pledge. Each quasi sentence is coded according to one of fifty-seven base categories, and the number of sentences in each category is converted into a percentage of the total sentences in the platform. The idea behind the approach is that parties emphasize policy areas in which they hold particular interest and credibility. Rather than a “great debate, or direct argument over a common range of problems…parties talk past each other, glossing over areas which might favor their rivals while emphasizing those on which they feel they have an advantage” (Budge, Robertson, and Hearl 1987, 24).3

Finally, the Comparative Manifestoes team devised a “left-right” score based upon adding the percentages of categories considered “right” and subtracting the percentages of categories considered “left.”4 Negative scores indicate that a party is located on the left and positive scores indicate that a party is located on the right. In this way, I obtained left-right scores for each of the national parties that participated regularly in alliances. I had no information about the platforms of a few subnational parties that occasionally joined coalitions in gubernatorial elections, so they are omitted.

Overall, these parties occupied similar positions between 1997 and 2015. The PAN is always on the right, the PRI in the center, and the PRD on the left. In addition, we find the Partido Verde Ecologista de Mexico (PVEM) and Partido Nueva Alianza (PANAL) in the center and the Partido del Trabajo (PT) and Convergencia/Movimiento Ciudadano (MC) on the left.5 Of the newer parties, the PES was located to the right of the PAN in 2015, and Morena to the left of the PRD. Figure 1 shows the positioning of these parties across the three points for which I have data. The PAN and PRI have moved to the right since 1997, as has the PRD (especially after the split that resulted in the formation of Morena), while the PT has moved left.

Figure 1.

Mexican parties’ ideological placement, 1997–2015.

Figure 1.

Mexican parties’ ideological placement, 1997–2015.

Close modal

In order to measure ideological proximity, I first arrayed the parties on a left-right scale and counted as contiguous parties those that were no more than two parties away from each other. This coding relaxes the assumption of many scholars (e.g., Axelrod 1970) that coalitions should be not only contiguous but connected—that is, they should not skip over parties. Alliances with more than three parties were necessarily coded as noncontiguous. For gubernatorial election years between 1997 and 2003, I used the 1997 platforms data; for gubernatorial years between 2004 and 2009, I used the closer 2009 platforms data. For elections from 2009 to the present, I used the immediately prior platform data.

The results of this operation demonstrate that, empirically, parties were substantially more likely to form alliances with parties in their ideological neighborhood. Of 163 gubernatorial alliances, 124, or 76%, occurred between ideologically compatible parties. The most common alliances occurred between the PRD-PT-MC on the left and the PRI-PVEM-PANAL in the center. One reason for the PVEM’s occasional alliances with the PRD now becomes clear: it is located between the PRI and PRD in two of three years. The PAN’s greater abstention from alliances is also understandable given that it is usually the only party to the right of the PRI.

Over time, the number of ideologically diverse alliances varies but does not follow a consistent trend. In the Zedillo administration (1994–2000), fewer alliances formed overall (only 16% of gubernatorial candidates ran within an electoral coalition), but 75% of them were ideologically compatible. In the Fox administration, after changes in electoral law made coalitions easier to form, nearly half of all gubernatorial candidates ran in an alliance, and only 54% of them were ideologically compatible. Yet by the administration of Enrique Peña Nieto, the trends had reversed: only 37% of gubernatorial candidates ran in an alliance, and 65% of these were ideologically compatible (fig. 2).

Figure 2.

Mexican parties’ alliance structure over presidential administrations, 1994–2012.

Figure 2.

Mexican parties’ alliance structure over presidential administrations, 1994–2012.

Close modal

During this period (prior to the victory of Morena in the 2018 presidential election), we never see alliances between the PRI and either the PAN or the PRD, despite the fact that, in every year, the PAN is closer to the PRI than to the PRD, and the PRD is closer to the PRI than to the PAN. In fact, the PAN would be classified as a contiguous party with the PRI in 1997 and 2015, and the PRD is contiguous with the PRI in 2009. The historical relationship between the PRI and its major rivals is behind this pattern of avoidance—foreshadowing in some respects my later arguments about trust. Jorge I. Domínguez and James A. McCann argue that voters in Mexico first decide what they think about the PRI. Then, if they oppose the PRI, they choose among opposition parties (1996, 11). In fact, they note the existence of a “large minority of sophisticated strategic voters.…These anti-PRI Mexicans wanted to defeat the governing party so much that they suppressed their ideological preferences in order to back the party most likely to beat the PRI” (12). The leadership of the PAN and the PRD behaved in similar fashion (at least, until 2021), allying with each other against the PRI. Moreover, a nontrivial number of alliances—nearly a quarter of the total—took place between parties not classified as ideologically contiguous, sometimes even at the opposite ends of the political spectrum. Most of these, however, were not PAN-PRD alliances; many, in fact, involved the PRI in alliance with the far-left PT. These ideologically diverse PRI-led alliances—a third of the total—display bandwagoning on the part of smaller parties, rather than a “beat the PRI” dynamic.

I next took the difference between the positions of the farthest-apart parties in each coalition as a percentage of the total difference across the entire spectrum. This analysis provides results similar to the contiguity measure. The average distance between coalition partners amounted to 34% of the available spectrum. A little over a third of all coalitions covered a maximum of 20% of the spectrum. Overall, 70% of cases occupied a third or less of the political spectrum. Again, however, a nontrivial percentage of cases—23%—spread out over more than half of the available space. This seems to stretch the meaning of ideological proximity.

The analysis confirms that there is an ideological basis to the formation of electoral alliances in Mexico. The distribution of cases is nonrandom. It is substantially more likely that a coalition will form between ideologically close parties. At the same time, ideological proximity clearly does not explain everything. In roughly a quarter of coalitions, parties are relatively distant from one another on the ideological spectrum.

Pragmatic Alliances

The second hypothesis suggests that parties form alliances in order to win plurality elections, such as governors’ races. There are several ways to examine this possibility. First, one can look for cases in which the given set of parties, summing their votes in the previous election, would have won. This is the most straightforward way in which parties might calculate whether a particular alliance would benefit them. Surprisingly, however, of 164 coalitions, less than 5% (a total of 8) fit this criterion. Parties may well be estimating the benefits of forming a coalition, but the previous election’s results do not seem to be their primary evidence.

Even more surprisingly, there are a significant number of alliances that would have constituted a supermajority in the previous election: for example, an alliance formed between at least two parties when one of the parties had won the previous election, or a party was added to an alliance when the previous alliance won the election on its own. This occurred in 48 of 166 cases, or 29% of the total. Twelve of the 38 cases of alliances that covered more than half of the ideological space were of this apparently unnecessary type. This finding poses considerable challenges to the argument that parties form incongruous alliances as a result of competitive pressures. Instead, such alliances frequently occurred among parties that already occupied the gubernatorial office.

A second way to assess the competitive conditions under which alliances took place requires looking at their effects. What we find is that alliances formed even when there was little reason to suspect they would make the difference between victory and defeat. Votes for alliances ranged from less than 2% to over 67%. Losing alliances outnumbered winning alliances. About 45% of alliances won the election; 41% of these winning alliances, however, were actually among the supermajority class—that is, one of the parties in the alliance (or a subset of these parties) won the previous election. Victory was therefore overdetermined. Nearly 75% of the time, forming an electoral alliance had no significant impact on winning or losing the election. Either the alliance would have won anyway, or it lost despite having formed a coalition.

A third approach to evaluating the impact of competition on alliance behavior looks at a specific subset of cases that pose particular difficulties from an ideological point of view. In 19 out of 176 cases of coalitions—about 11%—the PAN, on the right, and the PRD, on the left, joined forces. When a major party heads an alliance with one or more small parties, the risks to its own ideological brand and internal cohesion are minimized, even if one of the minor parties in the coalition is relatively distant. The head of the gubernatorial ticket will nearly always come from the dominant party, leaving it to shape the coalition’s platform and identity.

However, when the PAN and the PRD joined forces, only one could head up the ticket. Reconciling party platforms became more of an issue. Internally, such alliances may create tensions among party activists. Externally, as parties better known to the public, the PAN and PRD took a greater risk of diluting their party brands. For Noam Lupu (2016), party convergence between ideologically distinct parties is one of the key precipitators of party-system collapse, a point to which I return in the conclusion.

To analyze these “unnatural alliances,” I used negative binomial analysis, appropriate for rare events, with PAN-PRD alliance as the dependent variable. I included several independent variables. First, I calculated the electoral margin between the PAN and the PRI in the previous election. In cases where one party was closer to defeating the PRI, it might be more willing to engage in an alliance to tip the balance. Thus, narrower margins would result in more alliances. However, the reverse might also be true: where the PAN had reasonable expectations of defeating the PRI on its own, it would see no need to engage in an alliance with the PRD. This should yield more alliances, the larger the margin. Finally, the relationship might not be strictly linear. I therefore also tested a version of the model with the PAN-PRI margin squared.

Second, I calculated the electoral difference between the PAN and PRD. Discussions over which party should head the ticket may become difficult if both parties have roughly the same weight in the electorate. However, Golder suggests that, in fact, “pre-electoral coalitions are less likely to form if there is an asymmetric distribution of electoral strength among the potential coalition partners” (2006a, 87). If this is true, small margins between the PAN and PRD should yield more alliances, rather than fewer. I use the absolute difference between the vote shares of each party for this measure.

Third, I include a measure, winlag, to mark elections where, in the previous gubernatorial election year, a PAN-PRD alliance would have won.

Finally, I include a dummy variable for whether or not the PRI was in power at the time of the election. If either the PAN or the PRD already governs the state, I do not expect them to seek out the assistance of their ideological rival.

The analysis reveals that most of these variables had no significant impact on the likelihood of a PAN-PRD alliance. In fact, only the PRI in power significantly increased the probability of a PAN-PRD alliance.6 The PAN and PRD were not more likely to join forces when they would have won the previous election by allying. In fact, a PAN-PRD coalition formed on only five of the twenty-three occasions when such an alliance should have formed, based on the previous election’s results. Similarly, although a PAN-PRD coalition would have won thirty elections, assuming purely additive results (almost a quarter of all gubernatorial elections), such a coalition formed on only nineteen occasions.

This set of findings suggests, on the one hand, a strongly anti-PRI character to the PAN-PRD alliances. The fact that there were no alliances between the PRI and either the PAN or PRD—despite the fact that both were ideologically closer to the PRI than to each other—reveals an ongoing anti-regime cleavage stemming from the PRI’s long history as a hegemonic party that repressed and cheated both the PAN and PRD.

On the other hand, beating the PRI was not sufficient motivation. On nineteen occasions in the data set, a PAN-PRD alliance formed against the PRI. But on nineteen other occasions, even though summing the PAN-PRD vote from the previous election would have predicted victory for a PAN-PRD alliance, no such coalition was formed. The problem with the pragmatism explanation is that it seems to explain when odd coalitions do happen, but not when they do not. Why would the PAN and PRD behave pragmatically in some cases and not others? To be sure, there must be idiosyncratic factors occurring at the local level (such as the availability of an attractive candidate or greater/lesser antagonism between the local party branches), but we do not see systematic evidence that one of the major factors that should encourage pragmatism—competitiveness—actually has this effect. I come back to this in my comments about trust.

There is an interesting pattern that appears when we look at the distribution of PAN-PRD alliances across time. Just over half of these alliances took place in two years: 2010 and 2016—two years before a presidential election.7 Speculation in Mexico suggested that the PAN and PRD allied in order to deny the PRI the resources of those state governments in the upcoming presidential election. Yet they did not form a coalition to contest the presidency until 2018. Moreover, what changed between 2018 and 2021 to stimulate the unprecedented inclusion of the PRI in the PAN-PRD coalition?

Coalition Formation from the Perspective of the Parties

In order to gain further insight into how parties view coalition formation and the selection of coalition partners, I conducted interviews with seven high-ranking members of six of the eight registered parties (the PVEM and the PRD refused to participate) in 2018. Most interviewees referred to the context of the presidential coalitions in which they participated. There were three: Juntos Haremos Historia (Morena, PT, and PES, with Morena leader AMLO as the candidate); Por México al Frente (PAN, PRD, and MC, with PAN leader Ricardo Anaya as the candidate); and Todos por México (PRI, PVEM, and PANAL, with PRI leader José Antonio Meade as the candidate). Ultimately, Juntos Haremos Historia would win the election with 53% of the vote, against 22% for Anaya and 16% for Meade. AMLO of Morena led the election in the polls consistently throughout the campaign. The interviews took place in June 2018, two weeks prior to the presidential election.

All interviewees took pains to clarify the conjunctural nature of pre-electoral coalitions. In fact, they covered, at best, only about half of the positions in dispute.8 Party leaders uniformly highlighted pragmatic over ideological goals. All three coalitions submitted a common platform, according to electoral law, but these platforms tended to emphasize vague valence issues. Typical of this pattern was the response of leaders of the PES, who referred to the goal of the Morena coalition as simply “a real change” (confidential interview, June 2018).9 Likewise, the party’s president stated that “the great point of agreement [coincidencia] is the opportunity of an electoral victory that brings about a change of government” (Macedo Serna 2017). However, the PES seemed reluctant to commit to a postelectoral legislative coalition with Morena, conditioning their legislative support on “the positions they take.…We won’t accept positions with which we are not in agreement” (confidential interview, June 2018).

The Morena leader, meanwhile, classified the alliance with the PES as “circumstantial,” versus the “historical” alliance with the PT (which has supported AMLO since his first presidential campaign in 2006). Describing the PES as a “clearly clerical party,” he argued that the PES “took advantage of an opportunity” to support the probable winning presidential candidate. In explaining why his party had accepted the PES’s support, he noted bluntly, “We left ideology behind. Pragmatic reasons are worth more” (confidential interview, June 2018). Having lost two previous presidential elections, one by less than 1% of the official vote, AMLO’s strategists expected fraud again, and despite leading in the polls, they wanted a cushion of support to overcome expected manipulation of the vote. The entire coalition, indeed, was built by uniting dissident currents of all three of the major party oppositions to Morena—PRI, PAN, and PRD—in a most eclectic blend. Yet they did not seek the participation of either the leftist PRD or the progressive MC in their coalition. I return to this problem below.

The electoral coalition formed by the PAN, PRD, and MC created even more ideological contrasts than Juntos Haremos Historia. For the PRD, the 2018 election represented a critical juncture. From an ideological perspective, the most natural ally would have been Morena, whose presidential candidate was a founding member and former president of the PRD, and had twice been its presidential candidate. However, PRD leaders did not enjoy a good personal relationship with AMLO; AMLO, indeed, seems to have seen the leaders of the PRD in 2018 as bitter personal rivals. In 2006, PRD congressional candidates took office as usual, at a time when AMLO was alleging that the election had been stolen and proclaiming himself the legitimate president of Mexico. After the 2012 election, when AMLO again alleged fraud, PRD leaders went even further and participated in the Pacto por México, a legislative coalition including the PAN, PRI, and PRD, which enacted many reforms that were controversial among the party’s leftist base. The break between AMLO and the PRD became inevitable at this point; AMLO began to call for the transformation of Morena, formed as a civil-society organization to support his presidential candidacy, into a political party. After Morena became a party (in 2014), the PRD leadership continued to avoid collaboration with it. The most telling incident came in the critical 2017 gubernatorial elections in the state of Mexico. The Morena candidate was nearly even in the polls with the PRI candidate. The PRD had its own candidate, polling a distant third place. AMLO asked the PRD to resign its candidacy and support Morena to defeat the PRI. PRD leaders refused, and Morena lost. It would have won with PRD support. This incident further soured AMLO on cooperation with the PRD.

Despite these problems, the PRD remained ideologically much closer to Morena than to the PAN. Yet when it came to supporting a presidential candidate, the PRD chose to back a PAN candidate rather than Morena. Given widespread sympathy for AMLO among the party’s bases, this decision would have enormous consequences for the PRD. During the campaign, the party hemorrhaged members and activists who resigned in order to support AMLO. AMLO’s landslide victory left the PRD with only twenty-one seats in the new legislature versus fifty-four in the previous legislature. According to Morena, the PRD never asked for an alliance (confidential interview, June 2018), nor did Morena, however, seek out the PRD. Even if the PRD had wanted such a coalition, it is unlikely that AMLO would have accepted it, given his frequent and public condemnation of the PRD for “collaboration” with the PRI.

MC chose to coordinate with the PAN-PRD alliance in part because its leaders did not think an alliance with AMLO would be accepted. The MC leader I interviewed blamed AMLO for the lack of convergence, pointing to his criticism of MC and his tendency to lump them in with the “mafia of power” that he said ran the country. As a progressive party, MC was reluctant to support a PAN candidate at the top of the ticket but felt an alliance with the PRD alone “wouldn’t work”; only the sum of the three parties had a chance of victory (confidential interview, June 2018). The ideological goals of the coalition included getting the PRI out of Los Pinos (the presidential palace)—overlooking the fact that the main rival in 2018 was not the PRI at all, which struggled to achieve third place in the party’s worst showing ever.

Where party leaders downplayed ideological divergence, they highlighted pragmatic calculations. All of the party leaders agreed that electoral coalitions were more likely to form around executive (plurality) elections than legislative elections (which use a combination of plurality and proportional representation rules). The stakes in executive elections—including control of considerable resources—made it necessary, “for the relevance of the position,” to “sum votes” (PT leader, confidential interview, June 2018). Thus, according to an MC leader, “It’s about adding together forces to win” (confidential interview, June 2018).

The larger parties saw the addition of even small parties as potentially making the difference in a close race. The representative of the PRI argued cogently, “You make alliances to have a better number of votes. 2–3% per party is worth it in a close election” (confidential interview, June 2018). Yet interestingly, small parties, whose candidates rarely head the ticket, also saw coalitions as more advantageous in executive elections. A minor leader of the PES remarked that, in legislative years, “it is better [más conveniente] to go alone,” in part because legislative elections do not require as much money as executive elections (confidential interview, June 2018). Likewise, the PT leader suggested that in order to have its own legislative bench, it was better to follow the “party logic” in legislative elections (confidential interview, June 2018). In contrast, in executive elections, small parties calculate that, because they will lose anyway, they have little to lose and a lot to gain by throwing their support to one of the major parties.

What they have to gain comes in the form of side payments. Most crucially, small parties win support for other candidates, particularly legislative candidates. The negotiation of which joint candidacies will belong to which of the parties in the alliance is the most fraught part of coalition negotiation. Partial alliances were often explained as the result of failure to reach agreement on the division of candidacies in one or more states. For those candidacies and districts ceded to a specific coalition partner, the partner had near-complete autonomy to determine who would be the candidates. The PT leader was the only one who mentioned trading off positions with Morena (in order to fulfill legal gender requirements of a fifty-fifty split between men and women candidates). The PES, on the other hand, said that “our group is absolutely ours” (confidential interview, June 2018).

Negotiations reflected, to some extent, the degree to which the larger parties felt they needed the smaller ones to succeed. In the case of the coalition Por México al Frente (PAN, PRD, and MC), the relative equality of the larger parties (PAN and PRD) resulted in a straightforward assignation of percentages of candidacies depending on each party’s results in the 2015 national legislative elections. Of course, specific decisions about which districts would be given to which parties remained controversial. The prospect of reducing competition for a legislative seat is a significant attraction for the smaller parties.

The larger parties cede these spaces to secure victory for the candidate at the head of the ticket. In the case of Juntos Haremos Historia (Morena, PT, and PES), candidacies were split roughly 50% for Morena and 25% each for the PES and PT, a division regarded by the Morena leader as “a kind of generosity” (confidential interview, June 2018). The PRI leader also noted that the party had given more candidacies to the small parties than their electoral presence deserved (confidential interview, June 2018).

A second type of pay-off involved administrative positions in a potential administration. MC had worked out with the PAN and PRD which positions it would get in the event of coming in first, second, or third (confidential interview, June 2018). The PES merely hoped for some patronage positions in an AMLO administration. These positions go primarily to party leaders; thus leaders, who negotiate the electoral coalitions, also have the most to gain.

A third consideration for the small parties is the division of public subsidies. Mexico provides generous public finance to parties based in part on electoral results. Small parties expected to improve their access to public financing as a result of coalition participation. Mexican law establishes that, if a voter marks the ballot for all three members of a presidential coalition, the “vote” is divided in three, one-third for each party. Small parties can only expect to gain from this situation. They get their loyal voters (who mark the ballot only for one party), plus a third of any voters who support the presidential candidate in general. In contrast, the party whose candidate heads up the ticket could lose up to two-thirds of votes for the candidate that might otherwise have gone to the party in a forced-choice situation.

A final consideration is the need to avoid losing legal registry by getting less than the minimum required percentage of the vote (2%). With 2.99% of the vote in 2015, the PT came the closest to losing its registry, but the PES and PANAL also got less than 4% of the vote. In a presidential election, running one’s own candidate outside of a major party coalition could easily dip these smaller parties below the threshold necessary to retain registry, as strategic votes shift to the main presidential candidates with a chance to win. Participating in a coalition may thus be a form of life insurance for small parties.

Why do the major parties pay these costs? They may hope to gain enough votes from the small parties to put them in a position to win the executive position, but the metrics used to calculate whether small parties are necessary and/or sufficient for victory seem much less clear. On the one hand, as the previous section showed, many oversized coalitions form that cost the dominant party more than it needs to pay. On the other hand, sometimes coalitions do not form that “should” have been obvious winners. It could be that parties rely on polls rather than previous electoral results to calculate the odds of victory. However, in the Mexican context, polls are not considered reliable and impartial but are frequently discounted, depending on the source. Even where the polls agree, they do not seem to function as the principal guide to parties. In 2018, based only on polls, Morena should not have expected to need any help to win the presidential election, while the PAN, PRD, and PRI should not have expected coalitions to help them win. Yet all the main candidates ran in coalitions.

The reasons behind the failure of the PRD and Morena to form an alliance go well beyond any simple mathematical calculation of costs and benefits. The sense of betrayal (on both sides) and lack of trust made it much more difficult even to imagine negotiation of an alliance, despite the ideological similarities between the parties. Similar stories embedded in local politics probably lie behind at least some of the failures to form winning coalitions in gubernatorial races, as well as practical difficulties in negotiating the price to be paid. Ironically, the fact that Morena’s leader spent so much time within the PRD (and, previously, the PRI) may have made him more inclined to distrust the PRD than the new and untried PES, which Morena felt it could dominate. Likewise, while the PVEM is in some senses an “untrustworthy” party from an ideological standpoint—having run in coalition with every party in the Mexican party system, including, in 2021, Morena itself—it has proven trustworthy in the sense of supporting the legislative initiatives of the executive, if only out of mercenary calculations.

What are we to make, then, of the 2021 pre-electoral coalition among the PRI, PAN, and PRD at the national level? Is this the triumph of pragmatism over ideology? Why did the PRD and the PAN decide to trust their one-time nemesis? Three factors seem particularly relevant: First, ideologically, the new PRD (shorn of its populist and radical leftist branches by the split with Morena) has become closer to the center. There is less to divide the PAN, PRD, and PRI than there used to be in ideological terms. There is also evidence that all three parties (but particularly the PAN and PRD) have become more pragmatic over time, more professionalized and bureaucratic, and more inclined to compromise. The role of state funding for parties is crucial, as is the transition from a mostly volunteer-activist base who risked death by participating in opposition party activities to a paid professional staff and paid campaign teams. Whereas Kenneth F. Greene (2007) pointed out the tendency for parties that depend on ideological (versus material) motivations for recruitment to prefer ideological purity, Mexican parties, after fifteen years of state funding, were much less dependent on this activist core. This trend may also explain why coalitions between the PAN and PRD took time to emerge despite even more intense pragmatic pressures to cooperate against the PRI before it lost the presidency. Though at least two PAN-PRD coalitions occurred at the state level before 2000, such coalitions had become more acceptable by 2010.

Second, trust may have grown between the PAN, PRD, and PRI during the Pacto por México period (2012–18), as they cooperated on numerous agenda items in the legislature. Third, and perhaps most importantly, the emergence of Morena may be providing a new focal point for opposition resistance. As the PRI was once considered the enemy to beat, fear of Morena’s hegemony probably played a role in the decision of the more established parties to join forces in order to prevent Morena from capturing so many districts that it would control a supermajority in the legislature.

This article argues that none of the hypothesized explanations is sufficient, alone or taken together, to understand pre-electoral coalition formation in Mexico. The most consistent evidence suggests that ideological proximity frames the choice of coalition partners more often than not but fails to explain roughly a quarter of gubernatorial coalitions and presidential coalitions in 2018 or legislative coalitions at the national level in 2021.

Pragmatism is clearly playing a role, but it is not clear why pragmatism dominates in some situations but not in others. Party size does not predict which pre-electoral coalitions form, even though party leaders consistently referred to office-seeking goals in explaining why they participate in coalitions. Over 40% of all coalitions in gubernatorial races were supermajority coalitions, even when these entailed substantial ideological costs. Meanwhile, many potential pre-electoral coalitions did not form, even when adding together the votes of potential coalition partners in the previous election suggested that they would win. And while side payments occurred in all coalitions in 2018, they did not determine which coalitions would form. Instead, larger parties sought the coalitions that made the most sense to them and then negotiated what types of side payments would occur.

The additional factor suggested by this paper is that, beyond pragmatic considerations and ideological proximity, the choice of coalition partners also reflected levels of trust among the parties. Trust is distinguishable from ideological similarity. Thus, the PRD and Morena never formed a pre-electoral coalition despite ideological proximity, either in gubernatorial elections or in the 2018 presidential election. Party reputations at least partly constrain alliance options.

One final consideration deserves mention: the routine cooperation of parties in nonideologically driven alliances may pose risks to their ideological brand, partisan loyalty, and the future of the Mexican party system itself. Over the last two sexenios, as coalitions of all types—including ideologically diverse ones—became more common, the number of Mexicans identifying with any political party at all fell dramatically, from 49% in 2006 to less than 20% by 2018 (LAPOP, n.d.). Politics have become increasingly personalized and—most recently—polarized around those supporting or opposing the current president, AMLO. It becomes difficult to hold parties accountable when parties are merely vehicles for the personal ambition of politicians who cannot themselves be re-elected to the most important positions, such as the presidency. Programmatic mandates are difficult to discern: What did voters for the PAN-PRI-PRD alliance intend to signal, other than “Stop AMLO”? If party convergence in broad but ideologically incoherent alliances is a symptom of impending party collapse, Mexico’s most venerable parties may have more to worry about than the results of any specific election.

1.

All translations are mine unless otherwise noted.

2.

Reynoso’s article uses expert surveys in Mexico’s thirty-two states to rank the governor, the main opposition candidate, and the main officialist candidate. He uses this data primarily to look at ideological polarization by state, but the findings suggest that there is variation in the placement of candidates from the same party across the states (2015, 329). However, this data is not available for the entire set of governors’ races in my sample and is based on assessments of candidates rather than parties; it could not be used in this limited context to determine whether a particular set of parties should, or should not, have allied.

3.

Some “‘positional’ codings were included experimentally in the scheme (for or against defence expenditures, for example).” These sets remained mostly empty in the European data (and in my coding), because parties “rarely…want to be seen as against strong defence…so the negative category is usually thinly populated.” Nevertheless, “intentions are signalled by emphasizing other areas such as peace. Quite as sharp a gap thus opens up between the parties as if they had opposed each other directly on the same policy” (Budge 1994, 456).

4.

“Right” categories include support for the military; support for freedom/human rights; positive references to constitutionalism; support for effective authority of the state; support for free enterprise; support for the need for economic incentives to business; support for free trade; support for economic orthodoxy and banks, and limitation of social services; positive references to patriotism and the national way of life; positive references to traditional morality; support for law and order; and support for social harmony. “Left” categories include positive references to decolonization; negative references to the military; support for peaceful ways of resolving international disputes; support for internationalism and international institutions (such as the United Nations); support for democracy; positive references to regulation of capitalism; support for long-term economic planning; support for protectionism; support for state influence in and control of the economy (including price controls and subsidies); support for nationalization, expansion of social service, and expansion of education; and positive references to labor.

5.

Convergencia por la Democracia was founded in 1996 by Dante Delgado Rannauro. The party changed its name to Movimiento Ciudadano in 2011.

6.

Empirically, there are two cases of such an alliance forming with a non-PRI government in power. In one of these cases, a PAN-PRD alliance won the state in the previous election. In the other case, the PAN won the previous election in a different coalition.

7.

Each of these years involved the same subset of states (since gubernatorial elections take place every six years). There were also more elections in these years—twelve in 2010 and 2016—than in other years.

8.

Mexican law allows three types of coalitions: total (in which the parties in coalition support one common slate of candidates for all positions in dispute), partial (in which the parties support half of the potential candidates in common), and flexible (in which the parties share at least 25% of the candidacies). In the 2018 elections, parties built only partial and flexible coalitions.

9.

All interviews conducted in Spanish; English translations are mine.

Axelrod
,
R
.
1970
.
Conflict of Interest
.
Chicago
:
Markham
.
Budge
,
I
.
1994
. “
A New Spatial Theory of Party Competition: Uncertainty, Ideology and Policy Equilibria Viewed Comparatively and Temporally
.”
British Journal of Political Science
24
(
4
):
443
67
.
Budge
,
I.
, and
H.
Keman
.
1990
.
Parties and Democracy: Coalition Formation and Government Functioning in Twenty States
.
New York
:
Oxford University Press
.
Budge
,
I.
,
D.
Robertson
, and
D.
Hearl
, eds.
1987
.
Ideology, Strategy and Party Change: Spatial Analyses of Post-War Election Programmes in 19 Democracies
.
New York
:
Cambridge University Press
.
Calvo
,
E.
,
F.
Guarnieri
, and
F.
Limongi
.
2015
. “
Why Coalitions? Party System Fragmentation, Small Party Bias, and Preferential Vote in Brazil
.”
Electoral Studies
39
:
219
29
.
Debus
,
M.
, and
J.
Müller
.
2013
. “
Do Voters’ Coalition Preferences Affect Government Formation?
West European Politics
36
(
5
):
1007
28
.
De Souza Carreirão
,
Y.
, and
F. P.
do Nascimento
.
2010
. “
As coligações nas eleições para os cargos de governador, senador, deputado federal e deputado estadual no Brasil (1986–2006)
.”
Revista Brasileira de Ciência Política
4
:
75
104
.
Devoto
,
L.
, and
J. C.
Olmeda
.
2017
. “
‘Juntos pero revueltos’: Estrategias electorales y coaliciones partidarias para la elección de diputados locales en los estados mexicanos (2000–2016)
.”
Colombia International
90
:
157
87
. https://doi.org/10.7440/colombiaint90.2017.06.
Domínguez
,
J. I.
, and
J. A.
McCann
.
1996
.
Democratizing Mexico: Public Opinion and Electoral Choices
.
Baltimore
:
Johns Hopkins University Press
.
Döring
,
H.
, and
J.
Hellström
.
2013
. “
Who Gets into Government? Coalition Formation in European Democracies
.”
West European Politics
36
(
4
):
683
703
.
Golder
,
S. N
.
2005
. “
Pre-electoral Coalitions in Comparative Perspective: A Test of Existing Hypotheses
.”
Electoral Studies
24
(
4
):
643
63
.
Golder
,
S. N
.
2006
a.
The Logic of Pre-electoral Coalition Formation
.
Columbus
:
Ohio State University Press
.
Golder
,
S. N
.
2006
b. “
Pre-electoral Coalition Formation in Parliamentary Democracies
.”
British Journal of Political Science
36
(
2
):
193
212
.
Greene
,
K
.
2007
.
Why Dominant Parties Lose: Mexico’s Democratization in Comparative Perspective
.
New York
:
Cambridge University Press
.
Gschwend
,
T.
, and
M.
Hooghe
.
2008
. “
Should I Stay or Should I Go? An Experimental Study on Voter Responses to Pre-electoral Coalitions
.”
European Journal of Political Research
47
(
5
):
556
77
.
Ibenskas
,
R
.
2016
. “
Understanding Pre-electoral Coalitions in Central and Eastern Europe
.”
British Journal of Political Science
46
(
4
):
743
61
.
Instituto Nacional Electoral
.
n.d
. “
Plataformas electorales
.
Accessed March 2018
. http://portalanterior.ine.mx/archivos3/portal/historico/contenido/Plataformas_electorales/.
Kellam
,
M
.
2017
. “
Why Pre-electoral Coalitions in Presidential Systems?
British Journal of Political Science
47
(
2
):
391
411
.
LAPOP (Latin American Public Opinion Project)
.
n.d
.
Vanderbilt University (website)
.
Accessed September 2021
. https://www.vanderbilt.edu/lapop/.
Lupu
,
N
.
2016
.
Party Brands in Crisis: Partisanship, Brand Dilution, and the Breakdown of Political Parties in Latin America
.
New York
:
Cambridge University Press
.
Macedo Serna
,
Israel
.
2017
. “
¿Quién es Encuentro Social y qué busca al aliarse con AMLO?
Nación321
,
December 15, 2017
. https://www.nacion321.com/partidos/quien-es-encuentro-social-y-que-busca-al-aliarse-con-amlo.
Machado
,
A
.
2009
. “
Minimum Winning Electoral Coalitions under Presidentialism: Reality or Fiction? The Case of Brazil
.”
Latin American Politics and Society
51
(
3
):
87
110
.
Méndez de Hoyos
,
I
.
2012
. “
Coaliciones preelectorales y competencia partidista en México a nivel federal y local (1994–2011)
.”
Política y Gobierno
19
(
2
):
147
98
. https://www.nacion321.com/partidos/quien-es-encuentro-social-y-que-busca-al-aliarse-con-amlo.
Olmeda
,
J.
, and
L.
Devoto
.
2019
. “
¿Unidos o separados? Dimensionando el éxito de las coaliciones electorales a nivel subnacional en Mexico
.”
Revista Mexicana de Ciencias Políticas y Sociales
64
(
235
):
289
326
.
Reynoso
,
D
.
2010
. “
Alianzas electorales y contingentes legislativos en los estados mexicanos (1988–2006)
.”
Revista Mexicana de Sociología
72
(
1
):
113
39
.
Reynoso
,
D
.
2011
a. “
Aprendiendo a competir: Alianzas electorales y margen de victoria en los estados mexicanos 1988–2006
.”
Política y Gobierno
18
(
1
):
3
38
.
Reynoso
,
D
.
2011
b.
La estrategia dominante: Alianzas electorales en los estados mexicanos, 1988–2011
.
Buenos Aires
:
FLACSO / TESEO
.
Reynoso
,
D
.
2015
. “
El espacio ideológico subnacional mexicano según juicios de expertos
.”
Política y Gobierno
22
(
2
):
317
46
.
Riker
,
W
.
1962
.
The Theory of Political Coalitions
.
New Haven, CT
:
Yale University Press
.
Shin
,
E
.
2019
. “
A Model of Pre-electoral Coalition Formation
.”
Games and Economic Behavior
118
:
418
35
.
Van der Meer
,
T.
, and
A.
Hakhverdian
.
2017
. “
Political Trust as the Evaluation of Process and Performance: A Cross-National Study of 42 European Countries
.”
Political Studies
65
(
1
):
81
102
.
Volden
,
C.
, and
C.
Carrubba
.
2004
. “
The Formation of Oversized Coalitions in Parliamentary Democracies
.”
American Journal of Political Science
48
(
3
):
521
37
.
Zuñiga
,
J. A
.
2000
. “
Estrategia de simulación, la que impulse Fox, sostiene Cárdenas
.”
La Jornada
,
May 29, 2000
. https://www.jornada.com.mx/2000/05/29/cardenas.html.