An analysis of legislator preferences as expressed by roll-call votes as well as public statements finds that Juan Guaidó would have received 86 votes against Luis Parra’s 71 in a full session to elect the President of the Venezuelan National Assembly. While still maintaining the support of a plurality of legislators, these calculations show that Guaidó’s ability to achieve the needed quorum for a valid session is now dependent on the votes of Amazonas legislators, whose validity is contested by Maduro’s forces. We estimate that the opposition has lost 29 legislator votes since the 2015 elections through a combination of judicial actions against it and its own errors in coalition management.
At around noon on Sunday, January 5th, a group of dissident opposition legislators voted with representatives of Nicolás Maduro’s socialist party to elect a new President of the National Assembly of Venezuela. The legality of the vote was immediately contested by the country’s opposition, which convened an alternate session of the Assembly to re-elect Juan Guaidó for another term as President of the same Assembly. As a result, Venezuela now has two boards that claim to be the legally elected leaders of its Legislative Power.
Importantly, maintaining the Assembly’s presidency is key for Guaidó’s claim to being the legitimate interim president of Venezuela, which relies on a constitutional provision according to which the head of the Legislature will hold the interim office of the presidency in the absence of an elected president at the start of the constitutional term. The country’s opposition, as well as a significant part of the international community, does not recognize the 2018 presidential elections as free and fair and thus claims that there is no legitimately elected president for the 2019-25 presidential term in Venezuela.
However, there is still considerable uncertainty regarding exactly what happened on that day and what it tells us about the balance of power within the legislative branch. Both groups claim to have had a majority of legislators present in the sessions at the time of the vote. Parra claims to have obtained 81 votes in a session attended by 150 legislators but has yet to produce attendance records. Since no roll-call vote was taken (voting was conducted by a show of hands), there is no direct way to verify the number of legislators present or the number of votes obtained by Parra.
On the other hand, focusing on the balance of votes within the chamber is not necessarily meaningful given the evidence that some legislators were blocked from entering the chamber. Because of the way in which the National Assembly’s rules of debate (originally penned by Chavista-controlled legislatures) are written, it is technically possible to elect a National Assembly President with the support of as few as 43 legislators. This is because while 84 legislators (more than half of the chamber’s 167 representatives) must be present for there to be a valid quorum, the board of directors of the Assembly is chosen by a simple majority of those legislators present. Thus, it is not difficult for a group that controls access to the chamber to assure itself of a majority of legislator votes.
VENEZUELAN ARITHMETIC 101
Analysis of the roll-call vote given in the evening session is more informative. Since the session was broadcast live and is publicly available there is no doubt that the legislators referred to actually cast their votes. Therefore, we can use the roll call vote to infer legislators’ support for Guaidó at present and at the time of the vote.
Our primary focus of interest is estimating the size of the coalition supporting Guaidó, for which the January 5 roll-call gives us the best direct evidence. Understanding how many votes Guaidó commands is only part of the picture; in order to know the overall balance of forces we would also need to estimate Parra’s votes. We attempt to estimate Parra’s support further below, though the absence of a roll-call vote makes that exercise more tentative. However, there is one very relevant sense in which estimating the size of the Guaidó coalition in itself is relevant: to assess whether he commands the majority of votes necessary for the required quorum for the Assembly to session and take valid decisions, which is 84 legislators.
In total, 100 legislators took part in the vote held during the alternate session which re-elected Guaidó held on the evening of January 5th in the headquarters of the pro-opposition newspaper El Nacional. All of these voted for the slate of candidates headed by Juan Guaidó. It is crucial to understand that this does not imply that Guaidó would have counted with 100 votes in a regular session of the National Assembly (which is what we seek to estimate). Several of the legislators present at the session re-electing Guaidó were substitutes of principals whose votes may have differed from theirs. In fact, precisely because of the regime’s efforts to co-opt legislators has been focused on principal legislators, there are several cases in which we know that the principal sided with Parra (or was unwilling to support Guaidó) while the substitute supported Guaidó. In a regular session of the Legislature, the substitute’s vote intention would be rendered irrelevant if the principal were to show up.
In fact, it is technically possible for two legislatures to function in parallel and for both to hold sessions that satisfy the valid attendance quorum of more than half of legislators even if they have no participants in common. This will happen if they are both partially stacked with substitutes whose principals participate in the alternate session. In normal conditions, it would correspond to the judiciary to decide which of the two sessions has been convened constitutionally. In Venezuela, where the sides to the conflict do not recognize the same judiciary, there is no straightforward way to resolve this conflict.
Table 1 shows the breakdown of the 100 votes obtained by Guaidó in the evening session by the legislator’s condition (principal or substitute) as well as, in the case of substitutes, by whether the principal would have voted for Guaidó. We also provide a subdivision by party or legislative group of the legislator who voted in the session. Of the 100 legislators, 14 were substitutes of legislators who did not support Guaidó. These are either legislators who explicitly voiced their support for Parra (11) or who belonged to parties that had announced that they would not back Guaidó’s re-election (3). This means that in a regular Assembly session in which legislators not supporting Guaidó were present, Guaidó would have counted with the support of only 86 legislators, two votes more than the minimum majority of 84.
Table 1: Breakdown of January 5 evening vote and hypothetical plenary votes
However, 3 of those 86 legislators represent the state of Amazonas. Their suspension by the Supreme Court’s Electoral Chamber shortly after the 2015 election on fraud allegations is precisely what set off the protracted conflict of powers between the judicial and the legislative branches that continues to this day. If we exclude those legislators, then Guaidó’s support would have fallen to 83, below the 84-vote threshold for a simple majority and valid quorum. Note that this does not necessarily mean that Guaidó would have lost the vote for National Assembly President, as these 83 votes may still have exceeded the number of votes in favor of Parra (who himself claims to have obtained only 81 votes). But it does mean that Guaidó would not have had enough votes to sustain a valid quorum. This is likely the reason why Guaidó spent so much effort on an unsuccessful attempt at trying to get all legislators, including the Amazonas deputies, into the chamber on the morning of January 5th.
How many votes is Parra likely to have mustered? In Table 2 we estimate the vote of a hypothetical plenary session in which all principals in the country joined the session. We assume that e-vote is not permitted (we return to this below). We also assume that all PSUV legislators would have voted for Parra, as well as all opposition dissidents who have been expelled from their parties for corruption investigations or have voiced their support for Parra. However, we assume that centrist minority parties (AP and Cambiemos) would have abstained.
Our results are summarized in Table 2. If the Amazonas deputies had been allowed to vote, Parra would have gotten 71 votes and lost to Guaidó’s 86. Without the Amazonas legislators, the difference would fall to 83-70. Note that there are four empty seats and six legislators who would have abstained. Therefore, whatever the scenario, it appears that Guaidó would have won the vote by a comfortable margin of 13-15 votes.
Table 2: Hypothetical balance of forces in plenary vote (without e-vote)
Given uncertainties about legislator loyalty, it is possible that the actual number of votes could have been different if the session had actually been held on Sunday morning. For example, the hardline 16-J faction, which voted for Guaidó in the evening session, had previously threatened not to vote for Guaidó, whom they charge with being too soft on Maduro. Recently, they had refused to vote for a proposal to change the rules of debate to permit electronic voting sought by Guaidó and approved on December 17. If their threat had effectively materialized, then Guaidó’s 86 potential votes would have fallen to 83, and to 80 without the Amazonas legislators. Alternatively, there are several principal legislators who are in principle pro-Guaidó and are not in exile but nevertheless did not show up to the January 5th evening session despite having no apparent physical impediment to attending. Their failure to show at the vote introduces uncertainty as to how they would vote in a full session of the Legislature.
On the other hand, if Guaidó had fallen below the required threshold he would have likely invoked the e-vote provision approved on December 17 allowing legislators to cast virtual votes. We estimate that this would have given him 9 additional votes, taking his majority to 95 votes (92 without Amazonas).
WHERE DID ALL THE OPPOSITION VOTES GO?
The bottom line is that while Guaidó still would in all likelihood have comfortably won a fair vote for the presidency of the Assembly, he commands the support of only around half of all legislators, making his coalition vulnerable. This is perhaps the more notable fact that surfaced out of the tumultuous January 5 sessions. Over the course of the past four years, the opposition’s majority has fallen steeply from the 112 legislators it won in 2015 to 86 legislators (83 without the Amazonas deputies). How did this happen?
The answer is a combination of persecution, persuasion, exhaustion and mistakes. While some legislative seats have been lost as a result of overt and explicit actions by the judiciary, others reflect voluntary defections by legislators who were originally elected in an opposition slate and who have decided not to back Guaidó, and others reveal genuine dissent within opposition ranks.
Table 3 lists the 29 legislative seats lost by the opposition during the past four years. Seven of these were lost as a result of judicial decisions. This includes the 3 Amazonas legislators whose election was invalidated by the Supreme Court’s electoral chamber and 4 seats which have been rendered vacant as a result of political persecution as both principal and substitutes are in exile, jail, or otherwise impeded from attending.
Table 3: Reasons for individual legislator losses by Guaidó coalition
But this only explains around one-fourth of the loss of seats. The remaining loss of votes shows a more conventional pattern of defections and dissent. Six votes were lost as a result of the expulsion of legislators that took place in November of last year after investigative news site Armandoinfo published a report containing allegations of wrongdoing. Another six legislators explicitly defected, announcing that they would back Parra despite still belonging to parties that support Guaidó. Four legislators belong to minority centrist parties that strongly disagree with Guaidó on key policy and strategic issues such as economic sanctions or the boycotting of elections (both of which are supported by Guaidó but rejected by large segments of voters) and have decided not to back him. Five legislative seats were lost by a combination of reasons, where the principal has gone into exile and the substitute switched sides. One additional legislator (Biagio Pilieri) showed up on Sunday evening’s session but did not vote for unknown reasons.
WAS THE LOSS OF LEGISLATORS INEVITABLE?
It is not unusual for authoritarian governments to be able to co-opt large parts of their opposition. Because autocracies have unbridled control over the use of force and significant economic levers, it is easy for them to generate powerful incentives to sway some elected politicians. To a certain extent, what is surprising is that Maduro was not able to use these tools more effectively in the past, and that it is only recently that he has proven able to achieve large defections from the opposition.
But it is not enough to simply look at Maduro’s actions to explain the large decline in opposition support among legislators. If we seek to understand the magnitude of the parliamentary losses experienced in 2019, it is also important to consider some of the strategic decisions made by Guaidó and his governing team. Holding together coalitions is a complex task, and there is evidence to suggest that some of the strategic choices made by the Guaidó administration may have contributed to accelerating the rate of legislator attrition.
Coalitions are typically held together by a combination of policy concessions and appointments. It is common for coalitions to include actors with different viewpoints on key policy and strategic decisions. These actors are often persuaded to continue to support the coalition through the deployment of selective incentives. Because it is difficult to maintain everyone happy on central policy issues (such as whether or not to negotiate with the regime), positions on subsidiary policy dimensions can play an important role in holding together a coalition. So will appointments to key public offices, which are often remarkably effective in getting politicians to change their policy views.
During his administration, Guaidó has made relatively few appointments to public office, presumably out of concern with creating positions with no effective power. He also has not tapped the funds in bank accounts of the Republic or PDVSA that were transferred to his management as a result of his recognition by the United States (except for very limited purposes). Both of these decisions have diminished his capacity to offer the necessary selective incentives to hold the coalition together.
When the Guaidó government has made appointments, it has been much more likely to choose technocrats than political figures. To take one example, when a Debt Restructuring Advisory Commission was appointed in July 2019 and tasked with issuing general guidelines for dealing with the nation’s debt, it was integrated by two academics and a former Wall Street analyst. Notably, it did not have any representatives from the National Assembly’s Finance Commission, which had been overseeing and investigating debt issuance decisions since 2016.
If we consider the list of opposition legislators that openly backed Parra, it is striking that of the 18 legislators (14 of which are principals), none is from the capital region. In stark contrast, the two highest ranking parliamentary appointees of Guaidó (Foreign Minister Julio Borges and UN representative Miguel Pizarro) are legislators who represent districts of the country’s capital. Greater geographic balance and emphasis on regional issues could have helped Guaidó deal with the discontent among legislators from the provinces.
There is also the issue of how to deal with dissidence. When in September of 2019, a number of centrist minority opposition parties announced that they had accepted to participate in negotiations with the Maduro regime, the reaction of the Guaidó team was to strongly attack them for entering into a “false dialogue,” claiming that the government was trying to use them to create a “tailor-made opposition.”  It is not surprising that these parties, while refusing to back Parra’s candidacy, also refused to vote for Guaidó on the evening of January 5th.
Guaidó’s dissidence problems are not limited to the centrist parties. They have also had to do with reining in hardliners. As we noted above, the 16-J faction had threatened not to vote for Guaidó on January 5th and only seems to have changed its mind after that morning’s events. Relations have been tense with this hardline faction for some time; for example, the group has systematically complained that Guaidó violated debate rules by impeding discussion of topics on which 16-J dissented from the majority.
The large attrition in legislative support should also lead to a reconsideration of the effectiveness of individual sanctions in spurring regime change. The selectiveness and conditionality of individual sanctions has often been touted as one of their advantages. When on April 30th, the commander of the National Intelligence Service sided with a failed military rebellion against Maduro, the Treasury Department promptly removed sanctions on him, citing the case as an example “that U.S. sanctions need not be permanent and are intended to bring about a positive change of behavior” However, the fact that the opposition has lost at least 17 congressional seats through defections (and gained none) over the past year precisely as these sanctions intensified suggests that this tool is at best ineffective – and at worst counterproductive – in weakening the governing coalition.
It may well be that the events of the past few days will strengthen Guaidó – at least for the time being – and lead opposition groups to rally around him in the defense of the last bastion of democratic institutionality. Yet unless the opposition leadership revises its approach to dealing with intra-coalition differences and its international allies reconsider their approach to engaging with the regime, there is a risk that the problems that generated this large attrition in legislative support will continue to weigh on the prospects for real democratic change in Venezuela.
 Luis Eduardo Parra R (@LuisEParra78). “Como cada #5Ene, el día de ayer, la Asamblea Nacional eligió una nueva Junta Directiva de conformidad con la Constitución y el Reglamento de Interior y de Debates, donde obtuvimos 81 votos de los 150 diputados presentes.” [As in every January 5, yesterday, the National Assembly chose a new board in compliance with the Constitution and the Internal Debate Rules of the Assembly, were we obtained 81 votes from 150 lawmakers present]. January 6, 2020 3:42pm. Tweet. However, the government TV station Telesur had originally reported that 140 legislators were present. teleSur English (@telesurenglish) Luis Eduardo Parra has been elected president of #Venezuela’s National Assembly with 140 legitimate votes. January 5, 2020 3:08pm. Tweet. Constitutional Convention President Diosdado Cabello, in turn, claimed that an opposition legislator had admitted there were 127 legislators in the session. Cabello respalda “legitimidad de la directiva autojuramentada de la AN” [Cabello backs the self-proclaimed AN’s legitimacy]. TalCual, January 6, 2020.
 A show of hands vote is customary by National Assembly rules and can only be replaced by a roll-call vote if a lawmaker requests it. (Article 92 of the AN internal debate rules states that all votes are public, while article 94 stipulates that public votes are initially by show of hands, unless a lawmaker asks that it be done through roll-call). By Parra’s own count, however, the difference would have been of approximately 10 legislators, suggesting that the vote was close enough so as to make it difficult to ascertain who had the majority without a roll-call.
 There is significant confusion as to whose access was being restricted by the National Guard. Government spokespersons claim that access was restricted only to the Amazonas legislators and to others who had arrest warrants issued against them(GNB y PNB se guían por lista de diputados «inhabilitados» para permitirles ingresar al Parlamento [GNB and PNB use an “disqualified” lawmaker list to decide if they can enter Parliament]. NoticiaAlDía, January 5, 2020. Also see GNB impide acceso a Juan Guaidó a la sede de la Asamblea Nacional [GNB impedes Juan Guaidó from accessing National Assembly palace], El Nacional, January 5 2020.), while the opposition claims that more legislators, including Guaidó, were restricted from access to the chamber. As we will show, an actual vote would have been tight enough that even just the Amazonas restrictions would have been enough to tilt the balance.
 In principle, opposition legislators could have left and broken quorum, as by Parra’s own admission he has the support of less than 84 legislators. It is likely that the reason why the show of hands vote was taken at an unexpected moment was precisely to make it difficult for the opposition to try to break quorum.
 EN VIVO – Elección de la nueva directiva de la Asamblea Nacional 2020. [LIVE – Election of the new AN 2020 board]. Youtube, January 5, 2020.
 Since legislator preferences may have changes as a result of the political events on the past few days, the January 5 roll-call is most informative about preferences at the time of that vote; nevertheless, unless preference changes have been too great, it should still serve as a reasonable proxy for support at the present time.
 One legislator, Biagio Pilieri, was at the session but left before his turn to vote. If we include him, attendance would be 101.
 One of those legislators, Romel Guzamana, represents the indigenous population of the Southern region of the country, which includes Amazonas and Apure. Indigenous peoples have separate congressional representation as per article 125 of the Constitution.
 The National Assembly’s rules of debate do give the presidency enough latitude to incorporate substitutes when the principal is not present. This means that if Parra’s supporters had tried to filibuster the vote, Guaidó could have incorporated the substitutes and in principle obtained the same 100 votes as in the evening session. Nevertheless, while Guaidó could have managed to get re-elected legally, it would have been as a result of his authority to decide on the incorporation of substitutes, illustrating the fragility of the arrangements. The threshold of 84 legislators in a full vote is meaningful because it implies that the result of the election is not conditional on the control by the Assembly’s presidency of the process of incorporation.
 Is it possible that Parra obtained the 81 votes he claims in the morning session? To do so, he would have had to gain an additional 11 votes. We have not been able to identify the positions of opposition substitutes, but it is not impossible for there to be 11 substitutes from opposition, centrist or Chavista dissident votes that may have also been swayed by the government. Nevertheless, we underscore again that a vote held while limiting access to the chamber has relatively little legal or normative significance.
 Pilieri has later claimed that he fully supports Guaidó (Bloque Parlamentario 16 de Julio (@fraccionAN16J) “2/2 #5Ene @omargonzalez6: Aunque la Fracción 16J tenia acordado abstenerse el día de hoy, decidimos de manera patriota apoyar a esta Junta Directiva” [2/2 #Jan5 @omargonzalez6: Though the 16J coaltion had agreed to not vote today, we have patriotically decided to support this Board], January 5, 2020 6:17pm) so he would potentially raise the total of votes to 87 (84 without Amazonas). Nevertheless, we prefer to hold to the strict criterion of observable votes cast rather than expressed voting intentions. The rationale for this criterion is that legislators could have multiple incentives to dissemble, making it remarkably hard to gauge voting intentions from public statements when these contradict the vote cast. Our methodological choice should not be taken as a judgment of this specific legislator’s loyalty to the opposition cause, on which we think there is a reasonable case that he can be expected to continue siding with the opposition.
 Although an expanded commission was created on August 13 to include some legislators, our discussions with market participants indicate that the role of these legislators in debt talks has been nonexistent.
 We count 18 legislators instead of the 17 in Table 3 to reflect the case of Lucila Pacheco, Zulia legislator originally elected as a substitute PSUV candidate who then joined the opposition yet now supports Parra. This vote is not a net loss because it originally belonged to PSUV. Rather, it is a vote that was initially gained and then lost.
 Julio Borges: aquellos que se presten al falso diálogo no representan a Venezuela. [Julio Borges: those who partake in the fake dialogue do not represent Venezuela]. America Digital news, September 19, 2019.
 Treasury Removes Sanctions Imposed on Former High-Ranking Venezuelan Intelligence Official After Public Break with Maduro and Dismissal. U.S. Department of Treasury, May 7, 2019.