Will Bolivian president Rodrigo Paz Pereira remain in power through the end of 2026?

TL;DR

I assign an 81% chance that Rodrigo Paz Pereira remains continuously president through December 31, 2026. He has already survived the larger May–June siege, still commands the military and a newly installed police leadership, and his likely successor has not built an elite coalition (OAS, Los Tiempos, Visión 360). (vision360.bo) The main NO path is a renewed national blockade after the September 18 emergency-rule deadline that combines with authenticated scandal evidence and a police or military break (La Razón).

Context

Paz remained Bolivia's legal and effective president at the September 5 cutoff. The official presidency site lists him in office since November 8, 2025, and he was publicly defending his economic program before business leaders in Santa Cruz on September 4 (Presidency of Bolivia, El Deber). (eldeber.com.bo) About 118 days remain.

He is still unusually weak for a first-year president. His cabinet has been heavily reshuffled, his vice president is an open adversary, fuel reform is generating fresh protests, and he lacks a dependable legislative majority. Yet the roads were open at the cutoff, the main labor federation was not leading a new resignation campaign, and no major opposition or business bloc had endorsed succession.

Evidence

The historical baseline starts low. From Bolivia's democratic restoration in October 1982 through the current cutoff, I count 12 completed presidential spells over roughly 43.9 president-years. Five ended prematurely under a broad definition, but only three—Gonzalo Sánchez de Lozada in 2003, Carlos Mesa in 2005 and Evo Morales in 2019—were clear protest- or crisis-driven forced exits (Bolivian historical chronology, U.S. State Department account of 2003–2005, CRS account of 2019). A constant-rate calculation gives about 4% risk of any premature exit and 2% risk of a politically forced exit over a random 118-day interval.

Paz is not in a random interval, so that baseline must be raised sharply. A comparative study of presidential governments from 1978–2005 found 17 challenges among 44 elected South American terms and 11 failed presidencies. Street protests became most dangerous when scandal, weak congressional support and elite abandonment converged (Hochstetler and Edwards). A later study found resignation-demanding protests in 86.7% of Latin American presidential interruptions from 1992–2016 (Polis México). Paz has nearly all the background risk factors. What he does not yet have is the decisive last step: a security-force break or an agreed elite succession.

The full monthly protest series shows both the danger and the de-escalation. These are national monthly counts of ACLED-coded events retrieved on September 5; recent months are provisional. Demonstrations combine protests and riots, while political violence covers directly violent events. The data do not count deaths caused indirectly by blocked medical access (HDX HAPI, ACLED).

The May–June crisis lasted 53 days, produced explicit demands for Paz's resignation and caused severe shortages. A later tally recorded 22 conflict-related deaths, including indirect deaths and an aircraft accident, while ACLED records only direct violent fatalities (OAS, AP). Paz survived, split the protest coalition through an agreement with the COB, and reopened the road network. Demonstration events subsequently fell 89% from May to August. Survival of two large event-count peaks—270 events in January and 216 in May—is real evidence of institutional resilience.

The September restart is serious but not yet comparable. EFE's September 4 report described dozens of cocaleros, campesinos and interculturales marching in Cochabamba, with the immediate demand focused on repealing the diesel decree rather than a unified transition plan (EFE). Four Santa Cruz blockades installed on September 2 were lifted by September 3 after fuel-allocation concessions, although protesters retained their demand that the decree be repealed (Los Tiempos, El Deber). The danger is concentrated around September 18: the government says organizations are waiting for emergency rule to expire, the COB says an extension could be counterproductive, and Evo-aligned organizers have demanded elections or threatened mobilization (La Razón, La Razón on the COB, La Razón on the ultimatum). (larazon.bo)

Security-force behavior remains the strongest survival signal. On August 7, armed-forces commander Víctor Hugo Balderrama pledged to defend the democratically elected government and said the voters' sovereign will was not negotiable (Los Tiempos). Police commander Mirko Sokol and possibly other senior officers offered their resignations on August 31 amid the leaked-audio controversy, but Paz installed a new interim command the same day without a reported mutiny or refusal of authority (El Deber, Unitel). (eju.tv) I read the episode as institutional turbulence but a small net positive for Paz's immediate control. Public loyalty statements can fail under stress, so this is not proof of future cohesion.

Paz is weaker in Congress than several headline seat counts suggest. The voting legislature contains 130 deputies and 36 senators. PDC has 49 deputies and 16 senators, or 65 of 166; Libre has 51 and Unidad 33. Figures of 70, 53 and 35 include nine supranational representatives who do not vote in the Assembly (2025 election observation report). Congress demonstrated its power by censuring the economy minister 91–34, with four abstentions among 129 legislators present (Radio Splendid). Even so, presidential removal is harder: recall is unavailable until half the term has elapsed, and a criminal trial requires a prosecutor-general request, authorization from two-thirds of legislators present, a Supreme Tribunal trial and a final conviction (Bolivian Constitution, Articles 169, 170, 184 and 240).

The elite succession problem also favors Paz. Unidad National said on September 1 that it opposed overthrow and early elections and wanted Paz to complete his term, even while acting as a constructive opposition (Visión 360). (vision360.bo) Vice President Edmand Lara has openly suggested that Paz resign and let him take over, so his incentive is clear (Oxígeno). His capacity is weaker. An August 19–31 online urban survey of 400 respondents found 10% approval and 75% disapproval for Lara, versus 21% and 65% for Paz, with a referential margin of error of ±4.9 points (Ipsos CIESMORI reporting). No reliable evidence showed a Lara transition cabinet, military recognition or a stable opposition pact.

The Cerimedo affair is a live threat, but the evidentiary chain is incomplete. Lara released purported recordings involving Paz's circle, yet Nadia Beller—Cerimedo's accuser and former partner—said one recording did not contain Cerimedo's or the first lady's voice (Visión 360). The request to initiate a presidential responsibility trial remained a private filing rather than a prosecutor-general accusation or legislative authorization (Bolivia Informe). The scandal's main risk is therefore indirect: authenticated evidence could unite Congress, protesters and security elites around resignation.

Economic stress keeps that indirect pathway open. The IMF and Bolivia reached a staff-level agreement on July 29 for a 36-month facility of about US$1.9 billion, but it remained subject to prior actions, legislative handling and IMF Executive Board approval (IMF). The inflation data do not reconcile cleanly. EFE reported cumulative 2026 inflation of 14.18% through April, while INE's September 3 release reported only 3.01% through August and 5.02% over twelve months (EFE, INE). (efe.com) Without a documented revision or methodology break, I do not treat the latest CPI as a clean stabilization signal.

My quantitative synthesis uses three views. A stressed historical-hazard model raises the random-period exit risk by roughly fivefold and gives about 82% YES. A scenario model assigns a 43.0% chance to managed disputes with 2% exit risk, 34.0% to serious but fragmented conflict with 13% exit risk, 18% to a renewed unified national crisis with 52.0% exit risk, and 5% to an abrupt security, scandal or health rupture with 65.0% exit risk; it produces 82% YES. A mechanism-by-mechanism check gives about 81% YES. Weighting these while allowing for model error gives the final 81% forecast.

What's non-obvious

The key issue is not whether Paz is unpopular. It is whether actors who dislike him prefer Lara and disorder to a weakened Paz. At the cutoff, Unidad explicitly favored completing the presidential term, business leaders were bargaining over reforms rather than endorsing succession, and the military publicly backed constitutional continuity. This blocks the coalition that ended the presidencies of Sánchez de Lozada, Mesa and Morales.

The main datasets can mislead in both directions. ACLED shows a sharp fall in demonstrations, but its fatality measure omits people who died because blockades prevented medical care. The official inflation release looks reassuring, but it conflicts with the April figure reported earlier in the year. Crowd-size reporting for the September 4 march also ranged from dozens to thousands. I therefore give more weight to road control, coalition membership, legislative declarations and security-force behavior than to any single protest, inflation or polling number.

Uncertainties

• The largest gap is the post-September 18 mobilization plan. Public threats exist, but the real commitments of the COB, transport unions, miners, Santa Cruz civics and agricultural groups are unknown (La Razón).
• Public military statements do not reveal rank-and-file sentiment. Internal readiness reports, disciplinary incidents and regional-command communications would materially improve the estimate.
• The Cerimedo recordings lacked an authoritative forensic result at the cutoff. Authentication that directly implicated Paz in criminal conduct or illegal security decisions would cut the forecast sharply (Visión 360).
• There is no reliable current whip count for Paz, Lara and neutral factions inside the PDC. The August ministerial censure proves that a large anti-government majority can form, but not that it supports presidential succession.
• A representative nationwide poll on resignation, early elections and acceptance of Lara would be more useful than the available 400-person online urban survey. Reconciled inflation data and daily fuel-supply figures would also clarify whether economic pressure is easing or merely shifting form.