So what I did was went to EH.net and downloaded RGDP for the US going back to 1790 and calculated the 5, 10 and 20 year variances in RGDP growth rates, resulting in a chart that I left at my other office. Of course, it doesn't matter too much because official GDP data was not consistently compiled before the 1930's, so the estimates are somewhat questionable. Regardless, the chart did confirm what I saw, the volatility in economic growth has been decreasing.
I wanted to use some more consistent data and so I went to the FRED database at the St. Louis Fed and ran the same calculations except since 1947 we've been recording RGDP quarterly, allowing for more datapoints. This resulted in 5, 10 and 20 QUARTER variances shown below;


Several ideas come to mind. First and foremost my hero, Alan Greenspan. Notice variance all but ceases to exist once Al came in, implemented wise monetary policy, and destroyed inflation. Stable prices I'm sure contribute greatly to this decrease in volatility.

Regardless, the primary benefit to such consistency is that it would eliminate the riskiness of long term investments (which may in part explain why long term interest rates are so low and we're approaching an inverted yield curve).
I'm busy here, so feel free to do the writing for me and tell me what say you.
No comments:
Post a Comment