The CVM report behind a Brazilian REIT's numbers

Every real-estate-fund multiple — P/BV, net asset value, vacancy — begins in a public document that almost no aggregator shows: the report the fund files with the CVM. Understanding that document is understanding where the number comes from, and why it can (or can't) be trusted.

What the monthly report is

Every listed FII is required, under CVM Instruction 472, to file a structured monthly report (the informe mensal). It isn't a press release or a manager's opinion: it's a standardized file, with fixed fields, that becomes public data. Among the fields that matter most to anyone doing analysis:

Net assets (patrimônio líquido) — what the fund is worth after liabilities.
Number of shares (cotas) — how many parts the net assets are split into.
Net asset value per share (NAV/share) — net assets divided by the number of shares; the fund's "book".
Number of shareholders (cotistas) — how many investors hold the fund.
Segment — corporate offices, logistics, paper (CRI), malls, and so on.
ISIN — the international code that identifies the share unambiguously.
Data_Entrega — the date the report actually became public. It's the most underrated field, and the most important.

The quarterly report and vacancy

The monthly report gives the fund's aggregate snapshot. The quarterly report (informe trimestral) drops to the property level: it details, building by building, vacancy, delinquency and area. This is where the vacancy metric comes from — not an estimate, but what the fund itself declared per asset. When you see "physical vacancy of 12%", this is its source.

Why Data_Entrega changes everything

The report describes a reference month. But the market only learns those numbers on the day they are filed — and that date is almost always days or weeks after the month ends. Ignoring this is the silent error of nearly every aggregator.

paired price = first trading session on or after Data_Entrega

At Dados B3, the price we pair with a fund's NAV per share is that of the first trading session on or after Data_Entrega — never the price at the end of the reference month, which the market couldn't yet have known. This removes look-ahead: the multiple only uses information that was already public when it was computed. It's the same point-in-time discipline we apply to stocks.

The black-box problem. Most aggregators show a finished "P/BV", without saying which price, from which day, paired with which report. If you can't see how and when the number was built, you can't audit it — and a multiple no one can reconstruct is faith, not data.

Why this makes the data auditable

None of this is secret. The raw report lives at dados.cvm.gov.br: public, free and structured. Anyone can download the file and redo the math. Dados B3's edge isn't holding data no one else has — it's the disciplined, methodology-open processing of that same public source: point-in-time pairing, and invariant tests that flag the number when net assets divided by shares doesn't match the reported NAV per share. When it doesn't reconcile, we flag it rather than hide it.

From the document to the number you read

It helps to see the whole path: the fund files the report with the CVM → the file becomes public data with its Data_Entrega → we pair NAV per share with the price of the first trading session from that date → the result passes the invariant tests → it becomes the multiple shown on the fund's page. We detail every step at /metodologia/fiis, and show how to read the result in REIT P/BV and How to analyze a FII.

See the FIIs

Honest limitations

The report is only as good as what the fund declared; property appraisals are periodic and can lag.
Data_Entrega aligns the data in time, but doesn't guarantee the appraisal behind net assets is current.
Nothing here is advice. The report explains where the number comes from — it doesn't say whether the fund is a good investment.