Tuesday, January 20, 2009

No posts until this weekend

Unfortunately (no, really), I am too busy to write until this weekend. I have put off revisions on an article for entirely too long, and now it's down to the wire - must finish!

If I haven't replied to your email just yet, I will this weekend. Same for comments.

Thank you all for reading,



  1. Yes, they are pretty kitties. Look for you later....aj

  2. This comment has been removed by the author.


    Holiday purchases comprise a large percentage of the entire year’s sales for many of the nation’s retailers.

    When the Census Bureau reports December sales data in late January, the data are “seasonally adjusted.”

    That is, accounting for the seasonal component of December’s sales (Christmas, Hanukkah, and other traditions)creates a more accurate, overall picture of the increase or decrease in sales.

    Employment is also seasonal: Retail
    stores hire temporary workers for the holidayrush, and schools close in late spring, pushing students from the classrooms into summer

    According to the Bureau of Labor
    Statistics (BLS), seasonal variation mayaccount for as much as 95 percent of the month-to-month change in employment. In
    fact, between September and October 2008, the economy created a total of 303,000 jobs; after accounting for seasonal factors, the economy actually lost 240,000 jobs.

    Most economic data series can be thought of as consisting of a trend component, a seasonal component,and a business cycle component.1 The trend component tends to be easy to identify; it is the long-run movement
    in the data, either upward or downward.

    Many economists remove the trend and seasonal components from data; this is called seasonally adjusting the data.

    As the chart shows, employment data that are not seasonally adjusted are quite volatile. It is difficult to tell if one month’s change is due to seasonal factors
    (such as post-holiday layoffs) or the business cycle.

    A policymaker who uses only non-seasonally adjusted data runs the risk of making a mistake when trying to determine the direction of the economy.

    Although monthly seasonal patterns are typically the most significant in economic data, other subtle seasonal factors are also important. For example, the dates of Hanukkah (and the Super Bowl for that matter) do not occur in the same month every year. Other seasonal factors may not even occur every year (e.g., a leap year).

    To remove these seasonal patterns, the Census Bureau has created a statistical procedure called
    X-12-ARIMA, which can be thought of as a “black box” with information on monthly effects, as well as information on when, and if, these other types of events will occur.

    Seasonal adjustment, then, is a statistical method of trying to adjust for predictable events. Of course, by adjusting for these types of occurrences, the seasonally adjusted numbers are not “real.” Rather, they are
    artificial numbers generated by a statistical method.

    These generated series, nevertheless, are considered by many to be superior economic indicators and influence financial markets and policymakers.

    By Charles S. Gascon, Senior Research Associate, Federal Reserve Bank of St. Louis


Note: Only a member of this blog may post a comment.