Trends and tendencies

I typed employment in the FRED search box, went down the list that came up, and picked three different measures:

Graph #1: Three Measures of Employment

One's a little higher than the next, and the gap between lines varies a little. But the biggest difference I see, apart from the colors of the lines, is that one of them starts ten years or so after the others.

They all have humps in the same places. They all have dips in the same places. They accelerate and decelerate and respond to crisis in unison. So does it really matter which one you use to evaluate our economic past?

It might, if you are looking at some particular detail. It might. But if you're doing macro, it might not matter very much. And if you're just sticking a toe in the water, it would make very little difference.

You probably would want to check all of those lines, eventually, to see if your observations and conclusions apply in all cases. Of course. I'm not saying it is okay to be closed-minded about such things. But if you want to choose a data-set by throwing darts, I don't see anything wrong with that.

I usually pick the one that shows the most years.


When I change the graph to show not "Thousands of Persons" but instead "Percent Change from Year Ago", the three lines still move as one:

Graph #2: Growth Rates of the Three Measures of Employment

The green line wanders a bit from the other two, but the red and blue are so close that the one is almost always hidden by the other. And while the green one shows differences, it still shows very much the same pattern as the others.

That pattern reminds me of something: inflation-adjusted GDP. Here in black:

Graph #3: Growth Rates of the Three Measures of Employment and Real GDP

Again, pretty much the same pattern.

Shouldn't be surprising, really, the similarity between employment and output. It's Okun's law.