Wednesday, August 3, 2016

GDP is not a good predictor of recession


Just for the hell of it, I downloaded quarterly data from FRED for the same period we looked at yesterday -- the start of 2009 to the most recent data. I put a Hodrick-Prescott on it in Excel, using a constant of 1600. That's the constant I'd ordinarily use for quarterly data. Here is the result:

Graph #1: Quarterly RGDP since Q1 2009 with an Unresponsive H-P Constant

As with yesterday's post, I did the graph over using a more responsive constant:

Graph #2: Quarterly RGDP since Q1 2009 using a More Responsive H-P Constant
This is similar to the second graph yesterday. The highs and lows follow the same basic pattern. And the third peak is still high and long, relative to the others. But the blue line has been "all downhill" since the start of 2015. And the red line mimics that downhill trend.

The second graph here and the second graph yesterday are similar except for the RGDP data for the last year or so. The difference seems to be in the data itself, not in the H-P calculation.

So I grabbed the GDPnow data from FRED, for comparison:

Graph #3: GDPnow thru Q2 2016
Well, this data goes up at the end. Not down. I guess we'll have to wait and see which one's right.

2 comments:

The Arthurian said...

I think the title of this post stands on its own without argument: GDP is not a good predictor of recession.

The Arthurian said...

Jacob Assa:
"... the ‘value-added’ of financial services is imputed based on banks’ revenues and costs, and the inclusion of such income in GDP has caused a deterioration in its correlation with measures of employment and median income, as well as in its performance as a leading indicator."