Discussion

By Kurt M. Sanger on Jan 23, 2012

Dr Joe;

I don't understand why you chose 1992 to compare to 2010 as 1992 was a Presidential Election Year. I would expect the increase in 1992 October's and November's shipments to be correlated to the election. Wouldn't you?

By Joe Webb on Jan 23, 2012

As noted in the column the effects of presidential elections had little to do with the seasonal patterns from year to year. September through November were always big. When I did the research for the webinar that was pretty clear.

By Joe Webb on Jan 23, 2012

Here's the 1993 data for those who are interested in a non-presidential and a non-congressional year:

Jan -0.7%
Feb -0.1%
Mar 0.0%
Apr 0.0%
May -0.3%
Jun -0.3%
Jul -0.7%
Aug 0.0%
Sep 0.8%
Oct 1.0%
Nov 0.6%
Dec -0.3%

But is this the result of a different economy at that time? 1992 was a strong year for GDP, +4.3%, and 1993 was a weak year, +2.6% in real terms. At that time, macroeconomic influences played a bigger role in print volume. The fourth quarter of 1993 was up a whopping +5.3% annual rate. September and October, which are important months in elections, were a greater percentage of print's shipments in 1993, which was not a presidential year.

It's hard to discern these things, but when I averaged all of the years close to the years noted in the chart, the patterns in the chart were the same, and so are the conclusions. Digital media alternatives have both reduced the total demand for print and have changed the seasonal patterns for that demand.

By Kurt M. Sanger on Jan 23, 2012

DR. Joe;

I suspect there are many issues going on at the end of the year. First on my list would be Christmas. Second may be periodic elections with Major Presidential elections every fourth year and Minor elections in-between.

Then there may be a change in behaviour over time, plus a change in the overall economy on a yearly and within year basis.

I would create averages and standard deviations by month for Presidential Election Years, Minor Election Years, and Between Election Years to see if there were a statistical difference.

The print market may be another complication. An increase in election advertisements should create more demand for print thereby driving up the price. Might the increase in price reduce the demand for Christmas advertisements thereby masking the affect of the election?

By Joe Webb on Jan 23, 2012

For you statistical enjoyment, the data are here...
https://docs.google.com/spreadsheet/ccc?key=0Aiyk1RVInN8xdDVHUHcxTTdPMU5OcXFIYWtpQ1JWUGc