Question of Timing: Pre-Electoral Coalitions in Multiparty Presidential Regimes

This paper focuses on the timing of coalition formation under presidential systems. While elections in parliamentary regimes have recently been characterized by a high degree of uncertainty both in the results and the formation of cabinets, this is not the case for presidential politics. As a result, we argue that coalition cycles differ across systems of government. As a matter of fact, we argue that the process of coalition formation under presidentialism is more complicated. To test our claim, we look at forty-four multiparty governments in Latin America and Asia. We find that presidentialism — with its fixed mandate and its specific institutional mechanism for selecting the president — directly affects the ‘coalition life cycle’ and fosters a stronger propensity to pre-electoral agreements.

lthough the study of coalition governments was initially restricted to parliamentary settings, research on coalition cabinets in presidential regimes has been mushrooming over the last few years.By and large, the literature can be summarized in the following manner.The first generation of studies was marked by an effort to prove the existence and viability of coalition governments in Latin American presidential polities (CHASQUETTI, 2001;CHEIBUB et al., 2004;DEHEZA, 1998).This was an attempt to respond to the 'presidentialism versus parliamentarism' debate (ELGIE, 2005), in which many scholars stated that coalitions would be unstable or undesirable under a multiparty presidential system (LINZ, 1990;MAINWARING and SHUGART, 1999).
The findings of this first wave showed that coalition governments had made up more than half of Latin American governments since the mid-twentieth century.Having done that, the first generation laid the foundations to bring to the fore that some presidential countries have mainly been governed through coalition governments, as has been the case for Brazil and Chile (ALBALA, 2016b).Following closely, a second wave emerged and explored the structural differences within coalition cabinets under various presidential regimes in light of the accumulated knowledge about their counterparts under parliamentary regimes (ALTMAN, 2001;AMORIM NETO, 2006;CHASQUETTI, 2008;FREUDENREICH, 2016;MARTÍNEZ-GALLARDO, 2012;MEIRELLES, 2016).Finally, the latest generation saw the scope of the analysis spread to a study of coalition formation and cabinet appointments, more specifically focusing on topics such as coalition governance (ALEMÁN and TSEBELIS, 2011;KELLAM, 2015), conflict management (HIROI and RENNÓ, 2014;MARTÍNEZ-GALLARDO, 2014;RAILE et al., 2011), and ministerial turnover (CAMERLO and PÉREZ-LIÑÁN, 2015).
Notwithstanding the progress made thus far, it is surprising that only a handful of works have raised the 'when' question.This issue is quite relevant as the presidential system seems to bring high visibility and predictability to the election results, as the winner of the presidential election automatically becomes the president.The when question becomes even more salient as recent works have shed light on the fact that coalition governance is highly influenced by pre-electoral bargaining: the formation of preelectoral agreements explains why some coalitions are enduring and others not (ALBALA, 2021), who gets into the cabinet (FREUDENREICH, 2016), and which status is achieved by the president in the post-electoral scenario (BORGES et al., 2021).Moreover, the relative shortage of studies dealing with pre-electoral coalitions in presidential countries runs counter to the current state-of-the-art on coalition theories.Research on coalitions under parliamentarism, for instance, has long paid attention to the issues related to pre-electoral coalitions (ALLERN and AYLOTT, 2009;DEBUS, 2009;GOLDER, 2006).Nonetheless, when A Adrián Albala & Lucas Couto (2023) 17 (1) e0001 -3/28 explaining the phenomenon of coalition cabinets in presidentialism, the literature implies, more often than not, that coalitions are a product of post-electoral agreements whose principal goal is to grant the president a majority that could not be obtained in the election.
The main evidence used to support this argument focused on emphasizing the minority status of the president-elect's party.However, this contention is potentially out of kilter.
This is because the focus on post-electoral multiparty negotiations overlooks that presidential candidates anticipate the bargaining process and coordinate electoral strategies with other parties through the formation of pre-electoral agreements.In exchange for their electoral support in the presidential contest, the coalesced partners receive electoral support in electoral disputes in other arenas (BORGES et al., 2017).In this context, even if a president's party had no majority in parliament, this situation could have derived merely from a bargaining strategy.Put differently, commitments made prior to the elections may be accountable for the presidential status after the results of the polls are set and done.A typical example of this electoral electoral strategy is found in the behavior of the 'Concertácion' in Chile, a coalition composed of several center-left and left parties in the first elections following the return of democracy in the country.Such a coalition frequently resorted to the launching of joint party lists and even to the complete withdrawal of candidates from certain parties in some districts in order to sustain the coalitional pact amidst electoral races (ALBALA, 2016a).
Thus, the first objective of this work is to draw attention to the existence of preelectoral coalitions and to their impact on coalition formation in presidential countries.The research question here is 'how does presidentialism shape the formation of coalition cabinets in the post-electoral scenario?'.With that in mind, we aim to show how the temporal bound institutionally imposed by presidential regimes affects the timing of coalition formation.In so doing, we contribute to the literature by reinforcing that there are different levels of 'earliness' in the timing of coalition formation, ranging from fully preelectoral to fully post-electoral coalitional pacts (ALBALA, 2021).Overall, our findings show that pre-electoral coalitions are far from being the exception in presidentialism.In fact, they tend to be the precursor of coalition governments.
Our second goal in this paper is to address the selection bias that remarkably afflicts most comparative research on coalitional presidentialism.Although pivotal to our understanding of coalition cabinets under presidential settings, the excessive focus on the Latin American cases casts a shadow on presidential coalition cabinets elsewhere.The bright side, though, is that some studies have started to fill this gap by taking into account coalition governments in presidential countries other than those located in Latin America (ARIOTTI and GOLDER, 2018;CHAISTY et al., 2014; HANAN, The remainder of this work proceeds as follows.The first section discusses how presidentialism introduces new elements to the study of coalitions.This discussion feeds into the second section, in which we highlight that presidential regimes have a coalition life cycle of their own, as opposed to the circular cycle found in coalition governments in parliamentary regimes.In particular, we focus on the inception of coalition agreements.
More specifically, we outline different stages of 'earliness' in the timing of coalition agreements, with different binding impacts among the coalition players.The next section engages in a more empirical perspective and probes how coalition governments are spread across the coalitional cycle based on a sample of forty-four coalition cabinets in Asia and Latin America formed as soon as the presidents were sworn into office.The results point out that the vast majority of coalition governments are tailored prior to the elections.In Set-Theoretic terminology, the findings allow us to argue that pre-electoral commitments could even be labeled as a necessary condition for the formation of coalition governments in presidentialism.The paper concludes by advancing the importance of the presidential coalition life cycle to future studies on coalitional presidentialism.

The presidential factor
Apart from some recent works (FREUDENREICH, 2016;KELLAM, 2017;SPOON and WEST, 2015), studies on coalition formation are commonly based on the rational assumption that players have perfect information about everyone's strength and position, both of which would become evident the day after the election.This approach to coalition formation does not work well for presidential systems.
According to the presidential vs. parliamentary debate initiated by Linz (1990), there are three main differences between these two regimes.First, in presidential systems, the direct election of the head of government (who is at the same time the head of state) by popular vote1 is seen as a zero-sum game.leading to a bipolarisation of the competition between winners and potential winners vs.
To illustrate this concept, we use a triad game (Figure 01), in which a president-elect from party ' A' may form a government with party 'B' or party 'C'.In very rare and improbable cases, we can find a 'Grand Coalition' unifying A, B, and C altogether.It can be seen that a BC government excluding A is logically impossible once the president-elect cannot be precluded from her own government unless the legislature initiates an impeachment process and successfully removes the president from her post.Secondly, the election of the president (either simultaneous with legislative elections or not) does not depend on parliamentary bargaining (CHEIBUB et al., 2014).The president's legitimacy derives from her own election.This does not mean that the president rules alone, however.Where there is a case of a coalition cabinet, the president is thus constrained by a coalition agreement in order to keep her legislative basis.
Finally, yet importantly, under presidentialism, the mandate of the head of the executive is temporally fixed.This means that even if the president loses her majority in congress, she remains in office until the next election, which takes place following a constitutionally-mandated time schedule 2 .This presidential feature may potentially diminish the bargaining strength of coalition partners by reducing their 'walk away value' (LUPIA and STRØM, 2008).Since the departure of any member of the coalition does not lead to the fall of the government immediately, parties may think twice before stepping back and ultimately getting out empty-handed from the cabinet, losing access to pork and influence on policies (BORGES et al., 2017;MARTI NEZ-GALLARDO, 2012).
At the same time, the presidential mandate's principle of fixity concerns not only the termination but also the beginning of the mandate.This is tantamount to saying that the process of cabinet formation under a presidential system is limited in time, running from the proclamation of the result of the election to the president assuming office, which is generally fixed by the constitution.As such, the cabinet must be formed by the day the president assumes office.This is quite different from the procedure found in parliamentary regimes, where, broadly speaking, the bargaining rounds for government formation begin after the election and are (theoretically) unlimited (CHEIBUB et al., 2015;GOLDER 2015GOLDER , 2010)).As an unintended consequence of this trait, the absence of a temporal bound in government formation under parliamentarism contributes to blurring the lines between clear 'winners' and hinders the responsiveness of 'unexpected' coalitions 3 ( GLASGOW et al., 2012).
Hence, the separation of powers and the fixity of the mandate (for both its beginning and its conclusion) constitute definite constraints that have a significant impact on the formation, development, and conclusion of coalitions.Furthermore, since the president is, by definition, the 'formateur', the coalition options are more limited in presidential regimes.This feature leads to a quite restricted dimension of ______________________________________________________________________________________________ 2 According to Linz (1990), temporally-fixed mandates are the main reason why presidentialism is prone to instability, in addition to being a constraint on coalition formation since the president does not 'need' a majority to remain in office. 3Di Rupo Government in Belgium is a typical example of how a bargaining delay can lead to an 'unexpected' coalition.In 2010-2011, negotiations to form a coalition in Belgium lasted more than a year, during which the country was ruled by a caretaker government.This struggle in forming a government is not a Belgium idiosyncrasy, as demonstrated recently by the UK (2010), Italy (2018, 2013), Ireland (2020, 2016), Spain (2016), and Israel (2020-2019); moreover, the later composition of these governments raises the question of the responsiveness of these governments.
In 2020, for instance, would a Fine Gael voter have been pleased to see her party forming a government led again by Sinn Féin in Ireland?This question gains more prominence as the coalition options for the competing parties matter to voters when they are casting their vote (BLAIS et al., 2016).(2023) 17 (1)  e0001 -7/28 'willingness' to form coalitions (CHEIBUB et al., 2015;KELLAM, 2015;STRØM, 1990) around the president-elect or prominent presidential candidates.

Adrián Albala & Lucas Couto
In addition, majoritarian electoral systems are supposed to be more prone to electoral coalitions due to the narrowing of viable contestants (BLAIS et al., 2016;GLASGOW et al., 2012;GOLDER, 2015GOLDER, , 2006)).As presidential elections are by nature majoritarian, with only one possible winner in a single constituency, we can expect this form of election to have an impact on coalition formations.By the same token, scholars have stressed that 'ballotage' is a strong coalition constraint, since, after selecting two finalists among all the candidates, the first-round losers must 'choose' among the finalists in order to maximize their possibilities to be part of the next government (CHASQUETTI, 2008;MAINWARING and SHUGART, 1999;McCLINTOCK, 2018).
It is fair to say that presidentialism is a widespread form of government in Latin America.Given this, the concern for the region should not come as surprising.
After all, it seems natural to conduct research on a region marked by similar institutional traits.However, much of the literature on presidential coalition cabinets has dealt 'only' with Latin American countries.In these circumstances, our current knowledge of coalitional presidentialism suffers from a severe regional case selection bias.
The lack of studies on Asia has wrongly suggested that coalition governments under presidential settings are not found in this region.Nevertheless, in stark contrast to this notion, multiparty governments have been found in Indonesia, Maldives, the Philippines, and South Korea 4 .As a matter of fact, since the re-establishment of presidential competitive and direct elections, Indonesia has been governed solely by coalition cabinets.
The interest in Asian governing coalitions is reinforced since there are both similarities and differences between Asian presidential regimes and those in Latin America (on which much knowledge has been accumulated).As an example of the former, let us take the finding that the formation of pre-electoral agreements is Finally, a further consequence of bridging the existing gap between Asian and Latin American literature is bringing novel elements to Asian-centered studies.
To the best of our knowledge, the bulk of studies on Asian presidential coalition cabinets disregards comparative research designs in preference to case studies.As a result, Asian scholarship on coalitional presidentialism also suffers from a certain degree of provincialism.
Having said that, Figure 02 illustrates the presidential timing constraint and shows the constitutional lapse between election day and the day the president assumes office, for Asian and Latin American polities that are under a presidential system and experienced coalition cabinets.Although the average lapse ranges from 1.4 to 2.9 months, this value does not reflect the disparity among the cases.Indeed, the Uruguayan presidents-elect have five months to form their cabinet before taking office (four if the election came after a 'ballotage').At the opposite end of the scale, Argentina's and Bolivia's presidents-elect have the shortest lapse to form their cabinet: from 0.5 to 1.5 months, again depending on the need for a run-off.
Note that run-off conditions vary according to the respective Constitution.
In our dataset, nine out of twelve polities include it in their electoral system.The extent of the lapse between the first and the second round of the election is worth noting: from a couple of weeks (Maldives), a month (Argentina, Bolivia, Brazil, Colombia, and Uruguay) to a month and a half (Chile, Ecuador, and Indonesia).On the other hand, only three countries resort to the first-past-the-post electoral system for presidential elections: Mexico, the Philippines, and South Korea.

The framing of a coalition cycle
A coalition government presupposes the presence of distinct political forces within the cabinet, each counting with the corresponding support of all their respective members in congress (STRØM et al., 2008).This participation must derive from an inter-party agreement.That is, a coalition government is first and foremost the result of a negotiation between two or more parties, a negotiation that requires sufficient strength and mutual commitment on a broad list of topics at different levels (mostly at the executive and legislative levels).The collectivization of these outcomes derives from particular goals expressed as shared positive or negative values 7 , common ambitions of power, policy orientations, and/or aim to achieve re-election.All this engenders a higher exposition to both ______________________________________________________________________________________________ political and electoral accountability.On these grounds, considering presidential features, coalesced members under presidentialism are thus accountable to both the president and their partisan basis in the legislature (ALBALA, 2016b).
Using this narrow definition, we shall then not consider as 'coalition cabinets' 01.governments whose allies consist of co-opted or independent ministers8 ; 02.governments of a single party comprising several internal factions, irrespective of their degree of autonomy9 ; 03.governments of a single party as a result of a previous party merger10 , and 04.governments formed by a dominant party and 'satellite' parties11 .
Following Strøm et al. (2008), we aim to outline a "coalition life cycle", which contains all the phases of the coalition process, "from the formalization of the alliance until its termination, shrinking, and/or enlargement" (STRØM et al., 2008, p. 09), as exposed in Figure 03.Thus, any change in the original composition of the coalition (e.g., the entry/exit of a partner) entails the beginning of a new coalition cycle.Moreover, since the focus here is on coalition governments, we do not consider 'sporadic' legislative agreements (CHAISTY et al., 2014;RAILE et al., 2011).The formation of a coalition government -and the beginning of a coalition cycle -is intrinsically connected to the occurrence of recent or forthcoming elections, noted as 't 0' in Figure 04.Hence, a coalition government is considered as 'post-electoral' (t +1') when its formation and formalization occur as a result of interparty bargaining in a lapse between the final results of the election and the presidential inauguration.This temporal conception, although obvious and classic, must not disregard the study of informal and unpublicized agreements or the historical and 'familiar' links of collaboration (FRANKLIN and MACKIE, 1983;GIBSON, 1999;MOURY, 2011).In these cases, 'ex post' coalitions are seen as predictable, natural, or 'inertial' (BLAIS et al., 2016;DUCH et al., 2010).Similarly, 'ad hoc' agreements (noted t +2, t +3 ...)

______________________________________________________________________________________________
are the result of endogenous events such as the exit of one or more new members of an existing coalition, and/or voluntary and 'afterthought' enlargement of an existing coalition, with no real connection to immediate electoral matters.
On the other hand, 'ex ante coalitions' are those whose formalization has taken place 'before' election day.This type of pre-electoral coalition is thus the product of a public agreement between two or more parties that have consented to share their economic, organizational, and logistical resources to run together in an upcoming election (GOLDER, 2006, p. 12).This kind of coordination may encompass various aspects ('united electoral front', 'common candidate', etc), according to greater or lesser binding levels, and noted in Figure 04 as a 't -1' stage.The most conventional form of this kind of agreement is the merger of electoral lists (GOLDER, 2006), as a way to maximize the potential of smaller parties.
As a matter of fact, since every election implies previous organization and publicity during the campaign period, a dynamic approach for the study of such alliances should adopt a broader temporal perspective that sets as its starting point not the day before the election but the day before the electoral 'campaign' kicks off (CARROLL and COX, 2007, p. 301).
Run-off agreements (noted as 't-0.5' in Figure 04) are another typical example of these bargains: they consist of a public call for electoral campaigning between the two rounds of an electoral contest and include a transfer of resources from the eliminated candidate(s) to a contender who qualified for the 'ballottage'.
Let us also mention two additional pre-electoral types of agreement which suppose a higher commitment.First, the celebration of common primaries for the selection of a single coalition candidate.This requires the concerted organization of a campaign before the electoral campaign and a higher degree of convergence in the organizational and programmatic fields.Secondly, we mention the renewal of the electoral alliance of an incumbent coalition.Because these kinds of commitments occur at an even earlier stage, both are noted as a 't -2' position in Figure 04.
Finally, pre-electoral coalitions should not be considered as purely static and exclusive.Indeed, some coalition agreements may have had pre-electoral inception (either t-1 or t-2) but for some reason had to redesign their initial agreements to fit new members in post-electoral bargaining, at t+1.The incorporation of new members into an existing 'core' alliance can arise, for instance, when the original alliance could not reach a legislative majority in one or both houses, for bicameral cases (ALBALA, 2017).Thus, this scenario points toward a reopening of the bargaining process, the length and difficulty of which would depend on the political strength of the newcomers (SPOON and WEST, 2015).
As a consequence, whereas coalition cycles under parliamentary regimes resemble a perfect circular process, as can be seen in Figure 03, for presidential regimes, it seems that the coalition cycle is a little more complex, as shown in Figure 05.To sum up, we have set out that coalition formation should not be considered in a merely binary fashion (i.e., either pre-or post-electoral).We should, rather, consider that the coalition cycle may follow previous negotiations (that is the familiarity of the members) and early bind pre-electoral agreements (that is the precocity of the agreements).Indeed, even if pre-electoral agreements are not automatically binding (KELLAM, 2017), political parties and presidents-elect have strong incentives to maintain their electoral promises once they assume office.As a matter of fact, following MOURY (2013MOURY ( , 2011)), we may logically assume that explicit and formal pre-electoral agreements are politically and electorally more constraining than discrete 'ad hoc' informal ones.In other words, coalition cabinets should be more responsive and accountable when their formation is clearly identifiable.
Therefore, we propose a classification scheme for all the possible 'scenarii' of coalition formation, following their level of earliness and consequent potential binding effect.First, no electoral alliance, merely post-electoral formation (t+1), with a '-' direction.Secondly, partially constraining coalitions where the 'core' group of the cabinet members ran together in the election (t -1 or t -2), but where some new partners joined the cabinet after the election at t+1 (the constraining effect seems limited, '+/-').Third, run-off agreements, passed at t -0.5, where all the government partners ran separately in the first 'round' of the election but joined forces for the run-off (positive expected constraint, '+').Fourth, electoral coalitions, when all the partners of the cabinet ran together during the elections, at t-1, a strong constraining agreement ('++').Finally, full pre -electoral agreements, with stronger constraining expectations (+++), when all the partners in the coalition ran together during the elections and celebrated common primaries to select a common candidate, or when an incumbent president ran for re-election with the same partners with whom she was governing (t -2).This scheme is described in Table 01, below.Now that both the political science literature and the timing patterns for the formation of coalition cabinets have been discussed, we go on to explore the existing empirical data on the timing of coalition formation in Asia and Latin America.

Data on the timing of coalition formation in Asia and Latin America
In light of these arguments, we can assume that the more straightforward the election for 'selecting' the head of the government is, the more predictable and identifiable the coalition options are.Furthermore, t he more identifiable and recognizable the political options are, the earlier inter-party agreements are reached.As a matter of fact, this argument is not completely original.Indeed, Freudenreich (2016) has already shown that pre-electoral commitments tend to foster posterior government formation.The proposal here reverses the question, as we argue that every or most post-electoral coalition cabinets in presidential regimes derive 'from' pre-electoral agreements12 .
For the demonstration, we ranked all forty-four coalition cabinets13 formed in Asia and Latin America since the return of democracy in the late 1980s (Table 02).In this sense, our unit of analysis is presidential coalition governments 'per se', as defined in the last section.Importantly, we focus on the first cabinets formed 'after' presidential elections since the temporal aspect of coalition cabinets is of special interest to us.Thus, coalitions formed in-between presidents' tenures are mostly out of the scope of this article, unless they served as the spearhead of a future government.In the final column, we indicated whether these governments followed electoral agreements or not and, consequently, their expected binding level as exposed in Table 01, above.
Data come from several sources.For Latin American countries, to check the partisan composition of governments and thus whether they could be correctly labelled as coalition governments 14 , we relied primarily on Amorim Neto (2019Neto ( , 2006) ) and Camerlo and Martí nez-Gallardo (2017).The same information for their Asian counterparts was garnered through access to several editions of the Political Handbook of the World 15 .In addition, we retrieved data relative to Latin American pre-electoral coalitions from Borges et al. (2021), Freudenreich (2016), Kellam (2015), and Lopes (2022).In a complementary fashion, evidence on Asian preelectoral coalitions came from case studies, namely Hanan (2012), Kim (2011Kim ( , 2008)), and Teehankee (2020), and WikiLeaks 16 .In case of conflicting information across the different sources, we crosschecked it with data obtained directly from the countries' respective electoral bodies to ascertain which direction to follow.
Table 02 provides strong empirical evidence that most post-electoral presidential coalition governments are the product of (pre-)electoral coalitions.
Indeed, 40/44 (90%) of the post-electoral coalition governments in Asia and Latin America were derived, completely or partially, from pre-electoral agreements.This score is even more substantial when compared to the rate of coalition cabinets deriving from pre-electoral coalitions under parliamentary regimes (CHIRU, 2015;GOLDER, 2006;IBENSKAS, 2016).More surprisingly, this proportion is practically identical to the 91% ratio of presidential coalition governments resulting from (pre-) electoral commitments when the sample is restricted to Latin American countries.(ALBALA, 2021).
Therefore, we find only four cases of fully post-electoral agreements (10%): Paz Estenssorro, Ramos, Lacalle, and Sanguinetti.However, among those, the two Uruguayan cases may be considered as 'least-likely cases', as pre-electoral coalitions were not allowed in Uruguay before the 1997 constitutional reform.Also, the two Uruguayan presidents-elect, Lacalle and Sanguinetti, had already publicly announced during the campaign their intent to form a coalition government with the two main Uruguayan parties (the Colorado Party and the National Party, respectively).These cases are a sort of "inertial but informal agreement" (FRANKLIN and MACKIE, 1983), following a long tradition of political and government ______________________________________________________________________________________________ 14 Case studies, such as Albala (2016a), Amorim Neto and Pimenta (2020), and Kim (2008), provided a theoretical basis to rule out a handful of multiparty cabinets because they were not actual coalition cabinets. 15Available at ˂https://sk.sagepub.com/cqpress/series/political-handbook-of-the-world˃. 16Available at ˂https://wikileaks.org/˃.cooperation between the two parties, or factions of parties (ALBALA, 2016a).
Indeed, even where the agreement was formally made in the lapse between the election and the president taking office -thus officially at a t+1 timing, due to the impossibility of merging at a pre-electoral stage -the path had already been established between the partners before the election, at a t-1 timing.
Hence, there are only two 'genuine' occurrences of fully post-electoral formation of presidential coalition cabinets: the bizarre case of Paz Estenssoro's coalition in Bolivia (1985Bolivia ( -1989)), and the rather strategic Ramos' coalition in the Philippines (1995)(1996)(1997)(1998)  Altogether, only a minority of post-electoral coalition governments have not had a sort of pre-electoral conception.If we were to adopt a Set-Theoretic terminology, following which a 0.9 cut-off point is advisable for establishing necessity between conditions (RAGIN, 2008), we could go as far as saying that commitments made before elections are a necessary condition to form coalition governments in presidential regimes 18 .In other words, barring some exceptions, coalition governments following presidential elections cannot be found in the absence of some kind of pre-electoral multiparty commitment in presidential democracies.
The records also allow us to review the idea of 'ballotage' as a strong coalition-maker.In fact, only seven cabinets (17.5%, 7/40) resulted from run-off agreements19 .It is also worth mentioning that under the legal framework of the Bolivian Constitution, in a run-off, Bolivian members of parliament were responsible for electing the president.This situation led to the above-mentioned election of Paz Estenssoro and to the election of Paz Zamora, who -despite coming in third in the national election -was elected due to an agreement with the rightist party led by Hugo Banzer.Either way, we can argue that run-offs do not play such a decisive role in the coalition formation process as was usually argued in the literature.In brief, a multiparty cabinet was maintained even if a coalition was unnecessary to achieve a majority status in the legislature.

______________________________________________________________________________________________4
It is worth noting that the accurate definition of the South Korean form of government is an open issue in the literature(KIM, 2008).While some scholars assign South Korea to the set of presidential countries(BORGES and RIBEIRO, 2021; CHEIBUB, 2007;SEDELIUS and LINDE, 2018), others prefer to label South Korea as a semi-presidential regime(KIM, 2011(KIM, , 2008)).Question of Timing: Pre-Electoral Coalitions in Multiparty Presidential Regimes (2023) 17 (1) e0001 -8/28 relevant to explaining enduring coalitions (ALBALA, 2021).To a certain extent, Kim Dae-jung's government in South Korea confirms such a claim.Prior to the 1997 presidential election, the NCNP (National Congress for New Politics) had made an early electoral agreement with the ULD (United Liberal Democrats).Indeed, the preelectoral coalition, comprised of the NCNP and the ULD, turned out to be the winner of the presidential election in 1997; also, even though the new government was plagued by extensive deadlocks in the legislature, the initial coalition pact remained intact 5 until September 2001, covering approximately 70% of Kim Dae-jung's presidential term (KIM, 2011).By contrast, other patterns in the region are at loggerheads with the current presidential coalition theories.For instance, previous scholarship has suggested that presidential cabinets are unlikely to include parties that presented a strong candidate in the previous presidential election (FREUDENREICH, 2016).Much to the contrary, though, this has been a deep trend in Indonesian politics (HANAN, 2012).For illustrative purposes, the PD (Democratic Party) won the 2004 and 2009 presidential elections.In both, the Golkar Party had received a fair share of votes and finished with the third most voted candidate.Notwithstanding its electoral appeal, the Golkar Party was still invited to take a seat in both multiparty cabinets.In a similar vein, the 2019 Indonesian presidential election consisted of a dispute between two pre-electoral coalitions.On the one hand, there was a coalition led by the PDI-P (Indonesian Democratic Party of Struggle); on the other hand, the one advanced by the Gerindra (Great Indonesia Movement Party).In the end, Widodo (PDI-P) came out victorious and was sworn into office soon afterward.More curiously, though, the Gerindra made part of the new cabinet despite arguably being an electoral threat to the president.In examining both Asian and Latin American coalition cabinets together, we also improve the variability among the cases typically used to study multiparty presidential systems.This can be exemplified by the fact that all four Asian presidential regimes mentioned above have intermediate legislative elections, whereas this very same institutional feature is only found in Argentina and Mexico in Latin America.Moreover, not all presidential countries found in Asia allow presidents to run for reelection -the Philippines and South Korea being examples ______________________________________________________________________________________________ where the president is ineligible for immediate reelection.In contrast, once we subset our sample only to those countries that have had multiparty cabinets, only Chile, Mexico, and Uruguay do not allow presidents to seek reelection in Latin America 6 .

Figure 02 .
Figure 02.Lapse between election day/ Inauguration day of the president

Figure 04 .
Figure 04.Timeline of coalition formation

Figure 05 .
Figure 05.The coalition cycle under presidentialism more careful thought, despite not appearing in Table 02.To start with, not all pre-electoral coalitions convert themselves into coalition governments.This is perfectly demonstrated by the Roh Moo-hyun singleparty government.Even though Roh Moo-hyun (MDP) had formed a pre-electoral coalition with the PP 21 (People's Power 21) prior to the 2002 presidential election, the agreement between the MDP and PP 21 fell short even before the election took its place(KIM, 2011).To date, the literature has only tangentially approached this theme asBorges et al. (2021)  have contended that coalition governments are often derived from pre-electoral agreements, but they have not ruled out completely the possibility that the commitments could break and lead to single-party cabinets.A similar story occurred in Mexico, where the PAN (National Action Party) did not stick to its end of the bargain and decided not to allocate a single portfolio to its thenelectoral partner, the PVE (Ecologist Green Party of Mexico).In this way, we wish to draw attention to the counterargument that pre-electoral coalitions can still be dismantled, notwithstanding the fact that they demand much convergence among the signatory parties.Closely related, the scholarly literature ought not to disregard discovering which factors lay the foundations for maintaining coalition cabinets, especially those preceded by earlier agreements that could look needless after the distribution of seats in parliament.As an example, in order to win the 2018 Maldivian presidential election, Solih (MDP, Maldivian Democratic Party) had built a comprehensive pre-electoral agreement, including the AP (Adhaalath Party), the JP (Jumhoory Party), and the MRM (Maumoon Reform Movement), in addition, of course, to his own party.However, the president's party obtained a large majority by itself following an intermediate legislative election to the only Legislative House in the Maldives 20 .Yet, the coalition cabinet remained the same, not losing a single member despite the majority the 'formateur' party gained from the legislative elections.

Table 01 .
Classification scheme of the degree of earliness of coalition cabinets and expected binding levels Source: Authors' elaboration.Question of Timing: Pre-Electoral Coalitions in Multiparty Presidential Regimes (2023) 17 (1) e0001 -16/28 . Firstly, after running in an electoral alliance with the leftist party MIR (Revolutionary Left Movement), once he had been elected, Paz Estenssoro (MNR, Revolutionary Nationalist Movement) rescinded his electoral agreement with MIR and decided to form a coalition cabinet with his main rival during the electoral campaign -the conservative ADN (Nationalist Democratic Action).The other instance is much more ordinary, though.At the beginning of his mandate, Ramos (Lakas, Lakas ng Bansa) opted not to allocate portfolios to any other political parties than his own.However, he changed his mind in order to establish an electoral coalition with the LDP (Laban ng Demokratikong Pilipino) for the 1995 intermediate legislative election(TEEHANKEE, 2020).
Albala & Lucas Couto (2023)___________________________________________________________________17An example was the second cabinet of Brazilian President Lula in his first mandate(2003)(2004)(2005)(2006):the late inclusion of the PMDB party a year after the election (that is at a t+2 timing -an example of 'ad hoc' coalition) led to a fracture among the 'core members' and thus to the exit of the PDT(FIGUEIREDO, 2007).AdriánAlbala & Lucas Couto (2023)17 (1) e0001 -19/28

Table 2 .
Electoral and government coalitions in Asia and Latin America