Today's agenda

Every day the Uruguayan state publishes purchases, and once or twice a month the official agencies publish the numbers the economy is measured by. This page puts the two together, each with its date in plain sight.

What you will not find here is a daily press digest. Not because we would not like one: because we measured it and it does not hold.

We store press items by agency: one search for each of the 396 agencies that appear in the procurement record. The archive today holds 257 items.

Of those 257, none was published in the last 24 hours. None in the last 3 days. One in the last 7. Eight in the last 30. The median age is 666 days; almost two thirds are more than a year old and the oldest is from 2001.

Running the search more often does not fix that. The bottleneck is not the cadence: of 5,496 raw results the search returned, 257 survived the filter — a 4.7% pass rate — and the Uruguayan press simply does not write about public procurement every day. Sweeping all 396 agencies daily would multiply the load on a third-party service fivefold to obtain, as measured, roughly one new item a week across the whole site.

So press here is an occasional side dish, not the main course. And the section is designed so that being empty is its normal state.

There are two inflation-expectation surveys in Uruguay and they measure different universes: the central bank asks analysts and the statistics institute asks firms. Only one is here.

Firms expect 5.0% for 2026, 5.3% for the rolling year ending June 2027 and 5.4% for the next. The agency summarises with a 90% trimmed mean, which discards the extremes before averaging.

From the central bank's analyst survey we had noted seven figures. We do not publish them: the site is served with JavaScript and the expected files return errors, so a second verification could not reopen a single one. Publishing numbers that cannot be looked at again is exactly what this page holds against hurried coverage.

When the central bank publishes in a form that can be read and verified, both will sit here side by side — labelled, not averaged — which is what we do with opinion polls.

This is a search, not a curation: we query by the name of each agency in the record and store what comes back. That an item mentions an agency does not mean it is about a purchase, and that it does not appear does not mean it does not exist.

Each item carries a label for what kind of source it is, because a fifth of what the search returns is not press: of the 257 stored items, 53 come from portals of the state itself or of a party. Inside an eight-item panel that goes unnoticed; in a list ordered by date, a press release would sit at the top next to journalism, undistinguished.

And each item links to the agency in our record. If it does not cross with a record of ours, it does not enter: for loose news there are better places than this one.

From outlets we take headline, outlet, date and link, and nothing else. This is not a generous reading of a legal exception: it is the line that holds under any reading, because a headline with its link is a reference and not a reproduction of the work.

Two Uruguayan feeds deliver the full article inside the XML. Storing it would be trivial and would be a direct violation of this site's contract. Where an outlet publishes a news map we prefer it to the feed, because it contains exactly the four fields we want and none of the body.

One outlet expressly denied us access to its feed in its robots.txt and we do not consume it. Another publishes a machine-readable reuse policy that authorises headline and link and forbids use for training: we honour both directions.