Back in March I posted a FRED chart that Bill McBride over at Calculated Risk shared tracking a set of data that pretty reliably coincides with recessions. Even better is that in almost fifty years of data there have been only two false positives which brings us to a very interesting point. First, here’s the chart as it appeared when I posted back in March:
Next let’s look in more detail at those false positives:
This is what I like about this data series: Even if we set the bar as low as 5%, there have only been two instances since 1967 where a reading was given above that level when the economy didn’t go into recession within a year. Also interesting to note that there are no false negatives, as in times when the data said all systems normal and a recession occurred. So two false positives and no false negatives in nine instances means that false readings only happen 22% of the time or conversely 78% the time this data accurately predicts recessions. Sometimes without much warning, see 1969-70, 1973, 2007 for instance, but in economics being right nearly 80% of the time is pretty much hitting it out of the park.
All sort of interesting in a geeky and non-important way you might be thinking, so why bring it up now? Well because almost as soon as the ink dried (metaphorically speaking) on that blog post the chart went all jaggedy (yes I’m using the technical term 🙂 Have a look:
Now I’m not an economist and I don’t even play on on TV but it looks like there’s a 78% chance of recession starting in the next twelve months. With Q1 GDP having been downwardly revised (by 25%!) and Q2 expectations being downgraded (See Two more Q2 GDP Downgrades over on Calculated Risk) there’s something in the air that tells me to have the foul weather gear handy.