ReportGem ReportGem 中文

GLOBAL RESEARCH ARCHIVE

Brazil Election Monitor #2 Mixed signals – diverging voter intention polls

Published: 2026-07-21Institution: HSBC Global Investment ResearchPages: 7Original language: 英语Evidence page: 1

Research evidence excerpt

Brazil Election Monitor #2 Mixed signals – diverging voter intention polls

21 July 2026

Brazil Election Monitor #2 EconomicsBrazil

Mixed signals – diverging voter intention polls

◆ Last week, a stream of contradictory voter intention polls left Daniel Lavarda

local market participants confused Head of Brazil Economics Research

Banco HSBC S.A.

daniel.lavarda@hsbc.com

◆ Meanwhile, Lula’s net approval ratings appear to have reached +55 11 2802 2640

positive territory for the first time since December 2024

◆ Four more relevant polls to be released in the next few days

likely to be closely followed by market participants

Last week’s polls delivered mixed signals, prompting shifting interpretations

among local market participants. On Monday, Nexus showed broadly stable second-

round voting intentions, with incumbent Lula da Silva ahead of Flávio Bolsonaro. On

Tuesday, however, a Futura poll (7-11 July) indicated rising support for Bolsonaro in

second-round scenarios, putting him level with Lula. On Wednesday, Quaest (10-13

July) showed Lula leading Flávio, with a wider advantage than last month. Overall,

last week’s polling seems to have covered the full range of plausible outcomes, and

their lack of consistency left local market participants confused.

What do the second-round simulations suggest? Based on last week’s polling,

they show a broadly stable Lula vs. Bolsonaro match-up, with no material change in

relative positioning (Exhibit 5). By contrast, the spreads in Lula vs. Ronaldo Caiado

and Lula vs. Romeu Zema narrowed slightly, pointing to modest gap compression

rather than a meaningful shift in voter intentions (Exhibit 6-7).

Positive news for Lula (PT): Notably, our estimates of the government’s net approval

rating improved versus the prior week (Exhibit 2).

The English excerpt is extracted automatically from the cited source page and may contain layout or recognition errors. It is never batch translated.

Open report viewer