GFO

McKinsey Global Institute, 2021

From The Future of Work after COVID-19

20 to 25 percent of workers in advanced economies could work from home three to five days a week on a long-term basis after the pandemic, which is four to five times the pre-pandemic level.

Credibility?Composite credibility score, weighted blend of Specificity, Accuracy, and Calibration. Higher means more credible.

41/ 100

Evaluated

Specificity?Was the claim falsifiable? 100 means a precise, dated, quantitative prediction. 0 means an unfalsifiable platitude.

72

Accuracy?Did the predicted thing happen by today? 100 means clearly yes, 0 means clearly no, 50 means mixed or partial.

35

Calibration?Was the magnitude and timing right? 100 means right number and date. 0 means off by an order of magnitude or many years.

30

Reasoning

The McKinsey prediction claimed that 20–25% of workers in advanced economies could work from home 3–5 days a week on a long-term basis — representing a 4–5x increase over the pre-pandemic level of roughly 5%. The evidence as of late 2024/early 2025 tells a more nuanced story. Remote work has indeed structurally increased above pre-pandemic levels, but the magnitude falls well short of the prediction's specific claim. The Global Survey of Working Arrangements (G-SWA, PNAS 2025), covering 40 countries, finds that college-educated workers globally average only ~1.23 WFH days per week (about 25% of workdays), and even in the most remote-friendly English-speaking economies (US, UK, Canada, Ireland), the average is only 1.5–1.9 days per week — far below the 3–5 days threshold McKinsey specified. BLS data for November 2024 shows 23.3% of US workers did some work from home, but only 10.9% worked entirely from home. Stanford/SIEPR data shows WFH accounts for roughly 21% of paid workdays in the US. The share of workers doing 3–5 days WFH is a much smaller subset: most hybrid workers are in the office 3 days a week (i.e., WFH only 2 days), and fully remote workers represent only ~11–14% of the workforce. The prediction's directional claim — that WFH would be structurally elevated — is correct, and the 4–5x multiplier over pre-pandemic levels is roughly right for the share of workdays spent at home. However, the specific claim that 20–25% of all workers would be working from home 3–5 days per week is significantly overstated: the actual share doing so is closer to 10–15% in the most advanced economies, not 20–25%. The calibration is therefore poor on the frequency threshold (3–5 days), even if the directional shift is real.

Sources

Trajectory

2022-12-31: 49/1002023-12-31: 55/1002024-12-31: 63/1002025-12-31: 43/1002026-05-31: 40/1002026-12-31: 41/1000501002022-12-312025-12-312026-12-31

Last evaluated 5/31/2026, 12:32:53 PM, claude-sonnet-4-6