# International Monetary Fund · Hindsight | Envisioning

Source: https://www.envisioning.com/hindsight/imf

---

---

1.  [Hindsight](/hindsight)
2.  ›International Monetary Fund

---

# International Monetary Fund

Every publication series from International Monetary Fund in Hindsight, and how its predictions turned out.

---
[Publishers](/hindsight)[Forecast vs actual](/hindsight/forecasts)[Side by side](/hindsight/comparisons)
---

[World Economic Outlook](/hindsight/imf/world-economic-outlook)Numeric forecasts

Validated: graded by rule against the published actuals, with an agent audit of the matching ([D19](https://github.com/envisioning/hindsight/blob/main/docs/DECISIONS.md#d19-numeric-forecasts-are-graded-by-a-deterministic-command)).

73 publications, 1990 to 2026 · 1,560 predictions

41%

050%100%

hit rate, 576 of 1420, 95% interval 38% to 43% with the audit error applied ([D28](https://github.com/envisioning/hindsight/blob/main/docs/DECISIONS.md#d28-the-audit-error-rate-widens-every-published-interval-d11-issue-61))

Light band: with the audit error. Dark bar: count-only interval 38% to 43%.

-   hit576
-   partial293
-   miss551
-   ungradable96
-   open44

1.  1,560forecasts
    
    44 not due yet, 96 with no actual on the same definition
    
2.  →
    
    1,420graded by rule
    
    code compares each forecast with the published actual ([D19](https://github.com/envisioning/hindsight/blob/main/docs/DECISIONS.md#d19-numeric-forecasts-are-graded-by-a-deterministic-command))
    
3.  →
    
    152audited
    
    an agent checked the matching: 11% errors, 16 fixed in code for every row, 0% left ([D11](https://github.com/envisioning/hindsight/blob/main/docs/DECISIONS.md#d11-validation-thresholds))
    
4.  →
    
    1,420published
    
    with the remaining audit error added to the interval ([D28](https://github.com/envisioning/hindsight/blob/main/docs/DECISIONS.md#d28-the-audit-error-rate-widens-every-published-interval-d11-issue-61))
    

Rates, in percentage points: forecasts ran 0.05 points low on average; a typical forecast was off by 1.17 points (n 1,420).

By how far ahead the forecast looked

Same year

050%100%

49% of 722

Next year

050%100%

31% of 698

---
Data, code and grading rules are public at [github.com/envisioning/hindsight](https://github.com/envisioning/hindsight). Data under CC BY 4.0 (the BCB Focus survey under ODbL), code under MIT. Every rule is in [the decision log](https://github.com/envisioning/hindsight/blob/main/docs/DECISIONS.md). Think a verdict is wrong? [Open a dispute](https://github.com/envisioning/hindsight/issues/new?template=dispute-verdict.yml).
---
