from 2019paying particular attention to CoRev’s statistical analysis (at the end):
reader JBH wrote:
“Under this president, the employment situation is very good.”
I laughed and laughed and laughed when I read this. Why? Take a look at:
figure 1: Nonfarm payrolls (dark blue) and stochastic trends (red). Random trends estimated using 2010-2016 data, and regression of the first log difference on a constant. Source: U.S. Bureau of Labor Statistics, March 2019 Employment Scenario, and author’s calculations.
Even after the massive tax cuts and trillions of dollars in spending restrictions ended, we were as good as we were from 2010 to 2016.
But thanks to JBH for living in a fantasy world…it sure is more enjoyable than mine. Thanks, Drumpf!
Updated April 12, 2019 4:15pm PT: reader CoRev request:
Can you provide the raw data used?
here is data set Used, downloaded directly from FRED.
Update, April 16, 2019, 9AM PT: Reader CoRev provided his review of my trend analysis. No comments, here it is:
About three years after CoRev provided his “anomaly analysis”, I still don’t fully understand what he’s doing. Looks like a deviation from the average. If anyone can tell me why it validates his worldview (i.e. my choice to use random trends and time sample bias against seeing a boom in Trump jobs) please let me know. I’m (still) eager to understand.
Meanwhile, here’s a graph showing the problem applied to a linear (deterministic) trend in nonfarm payrolls (reproduced from “Why Friends Don’t Let Friends Apply Deterministic Time Trends to Nonfarm Payrolls”):
figure 1: Subsamples of nonfarm payrolls, 000s, sa (black), and linear certainty trend estimates over 20 years. Recession dates as defined by NBER are shaded in gray. Sources: Bureau of Labor Statistics, May Employment Release, NBER, and author’s calculations.
