Random Eyes turned up this graph:
 |
| Graph #1: Employee Compensation, Personal Income, and GDP All so Close Together it's Hard to Find Three Lines on the Graph |
And then I looked at the left border. It says (Index). That explains it. That's why the lines are so close together. That, and the pattern they share.
Anyway, I tweaked the graph to show the actual numbers -- billions of dollars, rather than some index.
 |
| Graph #2: The same three data series, in Billions of Dollars. The similarity of pattern persists. |
I thought it was interesting that the one graph shows the three series running close together, and the other shows them spreading apart. And I got wondering whether employee compensation is always about half of GDP. That first graph suggests an extraordinary similarity among all three series. I wanted to see whether that similarity says employee compensation and personal income are consistent, invariant portions of GDP.
 |
| Graph #3: Employee Compensation (blue) gradually falls as a Percent of GDP Personal Income (red) gradually rises as a Percent of GDP |
 |
| Graph #4: Employee Compensation falls as a Percent of GDP |
It looks like a substantial decline to me. So now I have to wonder why all the lines looked so close together in the first graph. They were indexed, which means the three lines were pinned together at one point.
That one point, in the first graph, was 2007. So the lines were pinned together toward the end, toward the right, when (as Graph #3 shows) this is where the biggest differences occur. So I'm thinking, if I pin the lines together earlier, say at the start of the 1970 recession, we will see them spread apart in the later years.
 |
| Graph #5: Same as Graph #1, but with the lines pinned together in at the 1970 recession |
 |
| Graph #6: Same as Graph #5, but showing only up to 1980 |
3 comments:
In all the three-line graphs, the blue employee compensation line is the lowest.
The red and green lines overlay in graph 1, but green is higher in Graph 2 and red is higher in graph 5.
What I get from this is that a dishonest person has a chance to show different results and draw different conclusions, depending on how the data is presented.
This bothers me a lot.
Cheers!
JzB
Two questions.
1) Why don't you just use a log y-axis? That would solve most of your problems.
2) For the compensation, you need to use "Compensation of Employees: Paid". I think the series you are using doesn't include supplements to wages and salaries, which includes pension and insurance payments that have taken on a larger portion of compensation over time. You'll still see a drop in compensation after the 1970's, but not as steep.
Hi Kevin.
Random Eyes lets me pull up random FRED graphs that people have created links for. Sometimes I just browse random graphs, looking for something interesting. In this case, the pattern similarity was interesting, and the late-year indexing is what made the pattern similarity obvious. In the back of my mind was something like "how the data is presented" as Jazzbumpa says.
Yeah, there are so many data series it is hard to know which one to use. I appreciate the feedback on that. Thanks.
Post a Comment