In my last post, Moving Averages, I made a bold comment that seasonally adjusted economic quantities used to intrigue me until I learned that they are merely moving averages. I said this only semi-seriously, and Rick Williams commented that seasonally adjusted averages really are seasonally adjusted, taking into account day of the week, month of the years, and other factors. This makes sense; otherwise month-to-month variability due to fluctuating numbers of weekends in each month would disrupt a year-over-year analysis by month. Month-by-month analyses would suffer because some months (e.g., November and especially December) show strong effects of holidays.
Rick has provided a link to FAQs on Seasonal Adjustment from the US Census Bureau. I will follow up with this and other sources.
I’ve been using my blog stats, courtesy, to discuss various charting techniques. I’ll use this same database to explore approaches to seasonal adjustment. The raw stats with a 7-day moving average are shown here.