Lies, damned lies, and durables orders

There are days when you really, really wonder about the efficiency of the financial markets.

Apparently (according to the AP coverage), US shares have been sold off because "Traders were discouraged to see that orders for long-lasting manufactured goods fell in February for the third time in four months". The weak state of US durables orders appears to be having effects closer to home too, with the Sydney Morning Herald saying yesterday that "The [ASX] market was down from the opening bell as Wall Street stocks were sold off sharply after unexpectedly weak US durable goods orders".

Could everyone get a grip, please?

Here are the durables data they're all supposedly worried about (from the terrific, and free, FRED data resource that the St Louis Fed provides).

Over longer timeframes, the monthly changes in the durables orders series are pretty much useless as a cyclical guide. You did get a run of consecutive falls in the post-GFC recession (the darker shaded area in the graph), but that's it. Even in what is now a prolonged recovery, you don't get a corresponding clear string of good durables numbers: if there is one in there somewhere, it's been well hidden by the monthly volatility.

It's even worse if you're not taking the longer view. Here's the past couple of years on their own.

The volatility is very large: 4% or 5% moves up or down in a single month are quite common, with the occasional even larger humdinger to throw you completely like that 22.6% spike in July '14 (some huge order for transport equipment, as it transpired).

So the noise is immense, and the signal (if there is one) is inaudible. I know journalists have to write something to keep the ads apart, and economists and analysts have to do something in the office till the pubs open, but durable goods orders? Really?