Friday, February 24, 2006
Do The Math
I read this line this morning: “every one percentage-point reduction in GDP growth cuts printing industry revenues by more than $1 billion.”
Since 1994, print has averaged -0.5% decline. GDP has averaged +3.3%. The faster GDP grows the more we lose in volume. So perhaps we should actually be rooting for GDP to go down so that fewer people have the ability to invest in print-avoiding technologies like broadband, networking, wi-fi, and who knows what else.
Hmmm... there is no statistical relationship.... the r-squared is 2.2%. I think that's a lot less than the 75% we look for in a decent forecasting model.
I live by rules of thumb; when you handle statistics all of the time and are pressed for instant answers quite often, you have to. But I try to revisit them as often as possible.
We know that the link to GDP has been lost in the mid-1990s. It still is. There are numerous other trends that are more important. As far as a slowing GDP and print? Print is a discretionary expense for the most part, and where it is not, there are more e-alternatives and they are growing in acceptance. Do we really think that a decline or rise in GDP really matters in the new media world?
But let's say that the spirit of the statement is true. If we take the fourth quarter printing shipments, we lost -$2.5 billion in inflation-adjusted shipments compared to 2004 in that quarter alone when the preliminary estimate of real GDP was +1.1%. If we therefore lost 2.2% GDP points compared to the average, that's on an annual basis not $1 billion in shipments, but about $5 billion per GDP point, or 5x more that the rule of thumb would indicate. The statement, aside from not having strong statistical support, is actually optimistic.
Since 1994, print has averaged -0.5% decline. GDP has averaged +3.3%. The faster GDP grows the more we lose in volume. So perhaps we should actually be rooting for GDP to go down so that fewer people have the ability to invest in print-avoiding technologies like broadband, networking, wi-fi, and who knows what else.
Hmmm... there is no statistical relationship.... the r-squared is 2.2%. I think that's a lot less than the 75% we look for in a decent forecasting model.
I live by rules of thumb; when you handle statistics all of the time and are pressed for instant answers quite often, you have to. But I try to revisit them as often as possible.
We know that the link to GDP has been lost in the mid-1990s. It still is. There are numerous other trends that are more important. As far as a slowing GDP and print? Print is a discretionary expense for the most part, and where it is not, there are more e-alternatives and they are growing in acceptance. Do we really think that a decline or rise in GDP really matters in the new media world?
But let's say that the spirit of the statement is true. If we take the fourth quarter printing shipments, we lost -$2.5 billion in inflation-adjusted shipments compared to 2004 in that quarter alone when the preliminary estimate of real GDP was +1.1%. If we therefore lost 2.2% GDP points compared to the average, that's on an annual basis not $1 billion in shipments, but about $5 billion per GDP point, or 5x more that the rule of thumb would indicate. The statement, aside from not having strong statistical support, is actually optimistic.
