Spain: growth in 2011 turned into a decline, after the fact
Spain: GDP growth in 2011
First estimate: in October 2012, after the year was over. The sign rarely changes: 126 of 3,426 cases.
More
A year’s GDP growth is first estimated from incomplete data soon after the year ends. Then statisticians receive tax returns, censuses and surveys, change methods, and the past gets recalculated. Forecasts are checked against the latest version of the past, while decisions were made on the first.
Blue: the first estimate of each year’s growth (from the first IMF edition in which the year was no longer a forecast); orange: the estimate of the same year now (April 2026). The shading is how much the past was rewritten.
- GDP growth 2011: first estimate
+0.4%
- GDP growth 2011: now
−0.6%
- Sign changes between first estimate and now, all countries
126 of 3,426
The first estimate is the year’s value in the first World Economic Outlook edition that marks it as actual; the latest, the April 2026 edition. In all, 126 of 3,426 “first versus now” pairs (2005–2024, all countries) changed sign; the average absolute revision is 1 pp. Cards cover countries of 3 million people or more with a change of sign or a revision of 2.5 pp or more.
Revisions come from new data, new methods or a new base year; the card does not establish which. Countries reporting by fiscal years, where the IMF may have shifted the year, and war-torn countries with sparse statistics are left out.
A question for this picture
If the past changes like this, how far can you trust the last year on any chart?