The Wall Street Journal claims federal employment, excluding postal workers, is the “lowest total in seven years” and includes the following chart.

Because this is a surprising factoid, I posted it on Facebook. After receiving a comment that 1.6% is probably off by a decimal, I started to question the data. I headed off to the Bureau of Labor Statistics’ website to see if I could duplicate the WSJ’s chart.

Here is the BLS chart of federal employment excluding postal workers.

This doesn’t match the Wall Street Journal’s chart, not even close. Yeah, there’s the huge bump for World War II, but the trends don’t match. This chart shows federal employment steadily increasing until the early ’90s instead of generally decreasing since the mid-’50s. That’s when I noticed that the Wall Street Journal plotted federal workers as a percentage of total non-farm employment, not as an absolute number.

So, what is total non-farm employment? Here is the BLS chart of non-farm employment, which shows a steady rise until the turn of the millennium.

Now, all I need is an easy way to divide the first chart by the second chart…

Getting both data series using the BLS series report tool wasn’t that hard. Nor was creating a spreadsheet that contained both and then dividing government employment by total non-farm employment. Surprise — the Wall Street Journal did not slip a decimal.¹At least, not in the chart. The WSJ article’s first paragraph reads, “21.9 million: The number of government workers in the U.S. in January, the lowest total in seven years.” That first number is off by a factor of ten. It should be 2.19 million.

Now for the chart I created:

Ha! It looks pretty close to the WSJ’s.²The blips every ten years in my chart are probably census employees, which the WSJ eliminated from its chart…somehow. Good to know. :-)

• 1
  At least, not in the chart. The WSJ article’s first paragraph reads, “21.9 million: The number of government workers in the U.S. in January, the lowest total in seven years.” That first number is off by a factor of ten. It should be 2.19 million.
• 2
  The blips every ten years in my chart are probably census employees, which the WSJ eliminated from its chart…somehow.