Seasonal adjustment
Nearly every headline economic series is 'seasonally adjusted' — raw counts like January layoffs or December retail sales are dominated by calendar effects that would drown the underlying signal. Adjustment strips the predictable annual pattern to reveal the actual trend.
Why it matters for signals
Seasonal factors are estimated, not fixed — they're recalculated every year, which means the 'adjusted' number for last January gets quietly revised for years after. It's a hidden source of revisions layered on top of the headline revisions this site already tracks.
The pandemic's lesson
COVID broke seasonal models: the April 2020 collapse was so far outside the historical pattern that adjustment factors swung violently. Several series were re-estimated under 'outlier' treatment to keep the calendar from being recalibrated by a one-off catastrophe — a reminder that the preprocessing is a model with assumptions, not a mechanical correction.