How much growth this year? Will next year improve? A number promises a little clarity. Then another report changes it, and we wonder what the first one was worth.
Evaluate what was known at the time and what changed afterward. A revision can carry information and still deserve scrutiny. Uncertainty does not excuse missing evidence.
Put the versions side by side
The World Bank's January 13, 2026 press release forecast global growth of 2.6% for that year. Its June 11 release put the corresponding forecast at 2.5%. Both refer to 2026, not actual growth in two completed years. January release, June release
The change is 0.1 percentage points. That differs from “down 0.1%,” and does not mean economic output itself fell 0.1%. One compares growth rates; the other would require a different calculation.
Align dates, target, and units first. This article compares historical versions as of October 24, 2026. June is not presented as final confirmation of the year's outcome.
More changed than the last decimal
The releases faced different information. January discusses trade, policy uncertainty, and supply-chain adjustment. June gives greater prominence in its assessment to energy, inflation, financing costs, and conflict shocks.
Ask which information entered the analysis, which assumptions became unsuitable, and how effects were expected to reach demand, production, and prices.
A prominent event does not establish that it caused the entire 0.1-point revision. If the report does not decompose the change, stop where the evidence stops.
A baseline is a set of conditions
Forecasts rely on assumptions about such matters as disruption duration, financing conditions, and policy paths. Significant changes can require a new calculation.
Think of a conditional travel plan: departure time, connections, and expected waits. This analogy explains dependence on conditions; it does not imply economic models and travel plans have the same precision.
Look beyond the central number for assumptions and risks. What better or worse scenarios are discussed? How do they differ from the baseline? A dramatic risk scenario is not automatically the institution's sole forecast.
A small global change can hide large differences
Regions and sectors contribute differently. Upward revisions in some places and downward revisions elsewhere can leave a modest global change alongside substantial internal differences.
A macroeconomic number cannot by itself decide whether a shop should expand or predict a household's income. Location, industry, and individual conditions are additional evidence.
This preserves the value of the broader report: it helps describe an environment and its risks, on a different scale from a personal budget.
Keep forecasts available for later checking
Save the publication date, indicator, period, and original assumptions. When actual results become clearer, compare what was judged well, where errors were large, and whether the explanation withstands inspection.
Overwriting old numbers with the latest erases the revision process. Selecting one error alone can also miss whether direction and risks were identified well. Fair evaluation preserves the original conditions and checks for explanations that quietly change afterward.
An institution's reputation cannot replace this work. Neither can liking or disliking the eventual result.
Keep four questions
Who made the judgment, and when? Which indicator and period? Under which conditions? What new information should prompt reassessment?
If an article supplies only the number, return to the original release. If the release remains unclear, leave that part unknown rather than completing it with your imagination.
The future will change again. Give each judgment a date, object, and basis. The next report then has something definite to be compared with.