Remember when I was talking about economic journamalism on durable goods orders?

Well Bloomberg gets the story right, with U.S. Economy: Durable-Goods Orders Hover Near Lowest Since 1996, but Reuters (U.S. April durable goods orders post biggest gain in 16 months), and CNBC (Durable Goods in Surprise Jump; Jobless Claims Dip), both get it very, very wrong.

Let’s roll the Bloomberg story for what is going on:

Orders rose 1.9 percent in April after a 2.1 percent drop in March that was more than twice as large as previously estimated, the Commerce Department said in Washington. Meanwhile, the Labor Department said 6.79 million people are collecting jobless benefits, and another report showed new-home sales were lower than forecast in April.

(emphasis mine)

OK, this looks like a 1.9% bump, which is a significant bump, but Bloomberg covered this as low numbers.

Why is this correct, and Reuters and CNBC wrong?

That phrase, “more than twice as large as previously estimated,” is why it is wrong.

As Reuters notes a few ‘graphs down:

However, March orders were revised sharply lower, falling 2.1 percent from the previously reported 0.8 percent decline.

So, we are comparing initial estimates for April with revised numbers for March.

If you were to compare apples to apples, you would have 1.9%-2.1%-(-0.8%), or 1.9%-2.1%+0.8%, or a growth of 0.6% comparing apples to apples.

In fact, you get a lot of this out of the government statistic machine, and it’s faithfully echoed by the press.

You have a number that is an improvement over the last month’s (revised) numbers, and it’s really good news…..And then, a few weeks later it’s revised down, and the new figures for the next month come out, and they look really good compared to the downwardly revised previous figures.

Rinse, lather, repeat.