17 January 2011

Seasonal Adjustment for a Bad Mood

In my search for articles about seasonal adjustment, I often find articles about seasonal adjustment disorder, also known as seasonal affective disorder. But aside from seasonal affective disorder, there is some truth about moods being seasonal, and I always wish that I could seasonally adjust my moods because December and January can be difficult times.

In the article, "Feeling Blue? Some Say Jan. 17 is the Worst Day of the Year" by Shelley Emling at http://livingston.patch.com/articles/jan-17-the-worst-day-of-the-year. She says that several years ago, a British psychologist named Dr. Cliff Arnall devised a formula that combined "various ingredients" in a "complicated equation" to calculate that the third Monday in January is the "gloomiest" day of the year. The variables include weather, debt, monthly salary, time since Christmas, and low motivational levels.

When I first read about this formula, I thought that maybe he had done some kind of regression analysis, but after more reading (http://www.canada.com/edmontonjournal/story.html?id=6637dce1-14f7-403b-9b46-a347bd3c8f23&k=53047 for example), it seems that the formula isn't so scientific. He says that most of his data comes from talking to people, seeing when they feel down, and working out a common denominator. The original sponsor of the study was an airline magazine, and the purpose of the study was to encourage people to book travel.

Dr. Arnall has since admitted that knowing the worst day of the year could make it be the worst day of the year. And the worst day of the year would depend on the person.

I was thinking that I should write an application that would run a regression on the amount of sunshine in a day, temperatures, debt, etc., and find the worst week. I wouldn't even need trading day because it's always good to blame Monday as the worst day of the week.

In the meantime, I will try not to think about this being the worst day of the year.

14 January 2011

Bad Seasonal Adjustments

Several web sites have referred to "bad seasonal adjustment" or "bad seasonal adjustment combined with inclement weather in late December" as the cause for the increase in unemployment claims or problems with the payroll numbers.

Sometimes I take the phrase "bad seasonal adjustment" as a criticism of the seasonal adjustment software or the people who set the options in the seasonal adjustment software. I have to remind myself, though, that most people think about any kind of seasonal adjustment like it's a black box. So I suppose that a bad seasonal adjustment means that there is an extreme seasonal factor for that particular month so that the changes to the seasonally adjusted series from the raw series will also be extreme.

And if there was unusual weather in December that was causing something unusual to happen in the time series, then the seasonal adjustment software will take this into account.

Maybe I take it too personally. I looked at the seasonal adjustments for payroll, and all the factors and all the diagnostics, and it all looked really good to me.