Radar Perene / Archive / science

IBGE SIDRA: how a price-index table becomes research input

◦ Index methodology v2.2 (working papers with DOI). See the methodology.

Science

The inflation number that opens the evening news has a house of origin almost no one visits. Before becoming a headline, the IPCA is one row in a public IBGE table — surrounded by hundreds of other rows the headline never mentions: the index opened by expenditure group, by item, by metropolitan region, with the weights each component carries. Whoever knows only the aggregate number knows the building's roof. The building has an address.

SIDRA — the IBGE's automatic retrieval system — is the public platform where the IBGE publishes the complete tables of its surveys: price indices, national accounts, employment, industrial output, trade, agriculture. Free consultation, no registration, through a web interface or an API, with the data at the level of detail at which it was produced.

The decisive word in the definition is "tables". SIDRA does not deliver loose series; it delivers the complete statistical object, with its dimensions — period, territory, classification — open to the user's choice. That structure is what turns a newspaper number into research material.

The same IPCA, three depths

Readers of this section already know the Central Bank's SGS, treated in this trail's first article. The comparison illuminates both: in the SGS, the monthly IPCA is series 433 — one number per month, ideal for whoever needs the aggregate's long history. In SIDRA, the same IPCA lives in tables such as 1737, which holds the index's historical series since 1979 — index number, monthly change and accumulations — and in the disaggregation tables, which open each month by groups, sub-items and regions, with their weights.

The difference is not bureaucratic; it is a difference in the questions that become possible. With the aggregate, one asks "how much did it rise?". With the open table, one asks "what rose, where, and with what weight in the arithmetic?" — food inflation versus services, the item that drove the month, the gap between metropolitan regions. Research questions live almost always in the second layer. And the origin matters in the reverse direction too: the SGS mirrors the IPCA because the index is produced by the IBGE; the primary source, with all the detail and the methodology, is the house described here.

The path to an essay in the archive

In this house's archive, the IPCA appears as a decisive supporting character. The essay on March 2020 — the month fear dominated asset prices — records a contrast that carries the entire reading: a market in panic and a price index standing still. In numbers: the IPCA that month came in at 0.07%, practically zero, while the dollar was flagged as a statistical anomaly — the episode is in "The Selic at 3.75% at the peak of fear". That 0.07% is one row of SIDRA table 1737; any reader reaches it in a minute, with no password.

The route the datum travels to the page is the same described for the Central Bank series in the trail's first article: collection with source and date stamped, contrast against the series' own history, editorial judgment on what deserves the record. What the SIDRA origin adds is the ability to go one floor down when the question demands it — when an aggregate number behaves strangely, the disaggregated table is where one checks whether the strangeness is general or lives in a single component. Research that sees only aggregates routinely mistakes a specific shock for a broad phenomenon.

What SIDRA asks in return

Nothing in money — and some patience with vocabulary. The tables carry the terminology of the surveys that generate them ("variable", "classification", "territory"), and the first visit tends to be slower than a first visit to the SGS. The API has its own documented syntax, which builds queries by dimensions. The investment pays off at the first question the aggregate cannot answer. The division of labor between the public portals is also worth recording: for an indicator produced by the IBGE, the primary source is SIDRA; for a financial or monetary indicator, the home is the SGS — and the two cite each other without embarrassment.

Frequently asked questions

Do SIDRA and the SGS hold the same data?

There is overlap in the most used aggregates (IPCA, GDP), but the houses differ: SIDRA is the primary source of IBGE statistics, with all the disaggregation; the SGS is the Central Bank's repository, which mirrors some IBGE aggregates and is the primary source for monetary and financial series.

Do I need to code to use SIDRA?

No. The web interface builds tables and exports spreadsheets without code. The API exists for automated collection — useful, not mandatory.

What is a SIDRA table number?

A stable identifier — table 1737, for example, holds the IPCA's historical series. Citing the table number in a study serves the same role as an SGS series code: two researchers are certain to be looking at the same object.

Are SIDRA data revised?

Some surveys revise earlier releases, and the IBGE documents those policies. Careful research records the consultation date — the same discipline described for the other sources on this trail.

---

Continue the trail: from the index that measures the cost of living to the data that measures everyone's debt — Tesouro Transparente: public debt data, open and ready for reproduction →

House reading: these series flow into today's reading, in the Diário; the episodes they illuminate, in the Atlas.

Choosing the right table for a specific question is the silent half of the work — the kind the house does on commission.