Brazil's Bolsonaro is seen with 45.6% of votes vs Lula's 44.5% in the Presidential runoff, according to a CNT/MDA poll

A runoff poll this tight falls well within the margin of error that has historically rendered late-cycle Brazilian surveys unreliable as point estimates; polls in past Brazilian presidential contests have undercalled both turnout composition and late swings among lower-income voters, the cohort where the two candidates' bases diverge most sharply. For BRL the operative distinction is not the headline spread but what each outcome implies for fiscal architecture: the market has tended to price continuity candidates on spending-cap credibility and challengers on the size and financing of expanded social transfers, with the FX and the long end of the local curve moving on fiscal-rule signals rather than the result itself. Worth observing is whether the narrow gap shifts either campaign's stated commitments on the budget framework in the final stretch, since concession or transition rhetoric after a close result has in past episodes mattered for governability and congressional arithmetic. The near-term calendar points to any remaining debates, further poll releases for trend rather than level, and post-result signals on the economic team, which has typically been the first concrete repricing trigger for the currency and local rates. As a single poll, the signal is directional only.