seasonal adjustment makes some data jump around
Posted: Fri Oct 30, 2026 8:53 am
Raw economic data has patterns. Retail spending rises in December, construction slows in winter, hiring jumps in spring. Statistical agencies remove that regular pattern so that a figure can be compared month to month. That's seasonal adjustment.
It isn't perfect. When the calendar shifts, say an early Easter or a harsh winter, the adjustment can overcorrect and a figure looks stronger or weaker than reality. Revisions later often partially undo it.
Weather is the best-known culprit. A cold month that cuts hiring can be followed by a rebound the next month, and the pair reacts twice to what's really one event.
Also, the adjustment factors themselves get updated, which can alter past numbers without anything changing in the economy.
For a scalper, the lesson is modest. A surprise in a figure known to be noisy with seasonal quirks is less reliable than one in a cleaner series.
Anyone noticed a release where a seasonal explanation was offered afterwards, and did price take that explanation seriously?
It isn't perfect. When the calendar shifts, say an early Easter or a harsh winter, the adjustment can overcorrect and a figure looks stronger or weaker than reality. Revisions later often partially undo it.
Weather is the best-known culprit. A cold month that cuts hiring can be followed by a rebound the next month, and the pair reacts twice to what's really one event.
Also, the adjustment factors themselves get updated, which can alter past numbers without anything changing in the economy.
For a scalper, the lesson is modest. A surprise in a figure known to be noisy with seasonal quirks is less reliable than one in a cleaner series.
Anyone noticed a release where a seasonal explanation was offered afterwards, and did price take that explanation seriously?