From 31 May to 30 September the family portfolio made 5.01%. Over the same days the CDI made 4.59%; it is the rate Brazilian banks charge each other overnight, and close to what a plain bank deposit pays. October added 2.34% by the 19th. Then the run-off of 25 October went against us. We expected Flávio Bolsonaro to win, President Lula was re-elected, and Brazilian shares fell 13.9% on 26 October. At that day's prices we estimate the portfolio lost 5.9% in the session and stood 5.6% below its value of 19 October; on 27 October it was 5.2% below. We will sell the Brazilian shares, picking the moment by price, and hold none for one to two years. We have sold none yet. Fixed income and credit stay, in reais or in other currencies.
What we are doing
- Selling the Brazilian shares. We have sold none yet and will pick the moment by price. Our thesis was a Bolsonaro win, and it was wrong. Brazilian shares will sit below the range the Equities mandate sets, by our decision.
- Holding no Brazilian shares for one to two years. Fixed income and credit stay, in reais or in other currencies. Any shares we buy will trade abroad, and we are studying bonds and shares in Europe and the United States.
- Keeping the CDI as the yardstick. The family earns and spends in reais, whatever currencies the portfolio holds.
- Choosing a new mix by Friday. The next report gives it.
Four months ahead of the CDI, by 0.42 points
We report a time-weighted return. It counts what the investments earned and leaves out the money the family put in or took out, so a large deposit never shows up as a gain. On that measure the portfolio beat the CDI by 0.42 points from 31 May to 30 September. June lost 1.34% while the CDI paid 1.12%; September made 2.78% against 1.08% (Exhibit 1). With most of the money in fixed income, the portfolio should stay close to the CDI. The shares, the funds and crypto make the difference either way, and they carry the risk.
Exhibit 1. Ahead in July and September, behind in June and August
Monthly time-weighted return of the family portfolio and the CDI, June to October 2026, in %
Sources: Apoena Research, the family portfolio's books at the close of 19 October 2026; Banco Central do Brasil, CDI (SGS series 12). Notes: a month with a large deposit and no valuation on its day counts the deposit from the day it arrived (Modified Dietz). October runs to the 19th, and its CDI waits for the month's end.
Two-thirds of the money sat in fixed income
On 19 October 63.7% of the portfolio was in fixed income: bank notes that pay the CDI or a fixed rate, and funds that lend to companies (Exhibit 2). Government bonds added 6.3% and Brazilian shares 14.3%, after purchases on 13 October. One position is out of the appropriate range. By currency, 95.6% is in reais, 3.4% in crypto and 1.1% in assets that move with the US dollar. When an election shakes the market, fixed income that pays the CDI holds its value and the shares take the fall.
Exhibit 2. Two-thirds in fixed income before the vote
Share of the family portfolio's value by class, 19 October 2026, in %
Sources: Apoena Research, the family portfolio's books at the close of 19 October 2026. Notes: each holding at the value its own account statement gives; the shares may not add to 100% because of rounding.
The vote cost an estimated 5.9% in one session
Markets had priced a Bolsonaro victory, and the result reversed it in a day. On 26 October the Ibovespa, Brazil's main share index, fell 13.88%, the dollar rose 8.79% against the real, and the DI future for January 2029 jumped from 12.64% to 15.25% (Exhibit 4). Long rates set the price of every fixed-rate bond, so when they rise those bonds lose value. Between closes we estimate the portfolio every weekday by moving each holding with its own market price. That estimate fell 5.9% on 26 October and left the portfolio 5.6% below 19 October, because it had gained a little the week before. It recovered 0.4% on 27 October, to 5.2% below (Exhibit 3).
Exhibit 3. 100 invested on 31 May was worth 107.47 on 19 October and about 101.90 on 27 October
Value of 100 invested in the family portfolio and in the CDI on 31 May 2026; dashed, our estimate after 19 October
Sources: Apoena Research, the books at each close and the weekday estimates of 23, 26 and 27 October; Banco Central do Brasil, CDI to 10 October. Notes: the estimate keeps the holdings of 19 October, moves each listed holding with its market price and leaves bank notes at their last stated value.
Exhibit 4. The shares fell 13.9%; the portfolio, an estimated 5.9%
Change on 26 October 2026, in %
Sources: B3 and CNBC, closes of 26 October 2026; Valor, closing DI futures rates, 26 October 2026; Apoena Research, weekday estimate of 26 October. Notes: no share had been sold by 27 October, and the next close books each sale at its own price.
Too young for most risk measures
Most risk measures need more months than we have. Over four months the average month made 1.24% and the worst, June, lost 1.34%; that June loss is also the deepest fall from a high so far. Volatility (how much the monthly return swings), the Sharpe ratio (the return above the CDI for each unit of that swing) and value at risk (what a bad month could cost, one time in twenty) need twelve months, so they arrive on 31 May 2027 if we value every month-end. A range of good and bad years drawn from the portfolio's own history needs 24 months, which means 31 May 2028.
Income beat the costs we can see
In 2026 to 19 October the portfolio received income worth 0.75% of its value on 19 October: coupons, fund and real-estate fund payouts, dividends and interest. Fees and taxes on the statements took 0.64%. Read that as a floor, because funds charge fees inside their price that no statement shows. Sales closed at a net loss of 0.36%.
Next: the Fed today, 28 October; the ECB on 29 October; the three mandates from 2 November; the Focus survey early that week (2 November is a holiday); the Copom's decision on 4 November; and the October figure at our next close.
Sources
Apoena Research, the family portfolio's books at the close of 19 October 2026 and the weekday estimates of 23, 26 and 27 October; Banco Central do Brasil, CDI (SGS series 12), the Focus survey of 23 October 2026 and the Copom's calendar for 2026; B3 and CNBC, closes of 26 October 2026; Valor, "Juros futuros disparam após vitória de Lula e leva a intervenções do Tesouro", 26 October 2026; Tribunal Superior Eleitoral, result of the run-off of 25 October 2026.
Figures to the close of 19 October 2026; prices may have changed since. Apoena Research is a self-directed research project based on personal and family capital. This report is not investment advice, an offer of financial products, or an offer to manage third-party capital.