Well, the stock market got what it wanted in terms of May's non-farm payrolls report: it was terrible.
With only 75,000 jobs created in May and, even more importantly, a net of 75,000 jobs revised out of the previously reported March and April figures, it is easy to surmise that the U.S. employment boom has ended. With no change from April to May in average hours worked, and a slight decline (3.1% versus 3.2%) in the average hourly wage, the employment market is showing clear signs of past-peak behavior.
There is an argument among economists -- and it is truly one of the most boring arguments that ever occur -- about the relative merits of the Household Survey (Table A in the Bureau of Labor Statistics data), which gives us the unemployment rate, and the Establishment Survey (Table B in the BLS data), which gives us the number for jobs created.
Although both are survey-based and borderline-innumerate, I have to side with the Table B folks; I pay much more attention to the Establishment Survey. I believe it is less irrelevant than the Household Survey, if that is even grammatically possible.
Anyway, there are some nuggets buried deep in Table B, and one of my favorites is the Diffusion Index. Diffusion indices, according to BLS methodology, "are the percent of industries with employment increasing plus one-half of the industries with unchanged employment, where 50 percent indicates an equal balance between industries with increasing and decreasing employment."
So, the BLS diffusion figures, which are divided into "Manufacturing" and "Total Private" segments, give an accurate reading on the health of the U.S. economy. Much more accurate, in my opinion, than some of the voodoo indicators -- any government-measured reading on inflation, for instance -- used by the bulls to justify their constant bullishness.
In analyzing the numbers (I use economics guru Ed Yardeni's charts for this) it is clear that the diffusion index for both manufacturing and the private economy as a whole have declined dramatically of late, especially when viewed on a year-on-year basis. Any reading above 50 indicates employment expansion, but at 54.8 and 52.0, respectively, the total and manufacturing sector diffusion indices are sharply below their May 2018 values of 67.1 and 63.8.
So, what does that tell us?
Googling "BLS diffusion index" brings up some interesting academic research. This piece from 1990, though horribly photocopied, shows that a declining diffusion index accurately called the 1980 and 1982 recessions, and though the author couldn't have known it at the time, the data in that piece was calling the 1991 recession, as well. That research also showed that it is possible for diffusion to remain above 50 for an extended period of time, which was certainly the case for the Reagan/H.W. Bush "Golden Era" of 1983-1990.
Back to Ed Yardeni's charts, we see the emergence of another golden era for private employment diffusion, as that index has not registered below 50 since the beginning of 2010.
Not surprisingly, manufacturing diffusion is more volatile than overall diffusion. After hitting 50.7 in April that index did recover slightly to 52.0 in May, but the downtrend is clear. Actually, 2018 is the only year in the current expansion that didn't include at least one monthly manufacturing reading below 50, and the downturn thus far in 2019 shows that 2018 was, indeed, as good as it gets.
But what about the markets?
A quick check of the charts shows that diffusion is a very good indicator of the near-term direction of rates on long-term Treasuries, less so on equities. By my back of the envelope calculations, If you had bought futures on the U.S. 10-year Treasury when the manufacturing diffusion index fell below 50 and sold those futures when the index rose though 60, you would have more than doubled the systemic returns from just holding the underlying bonds. Actually, it would be close to a 200% outperformance.
But what about stocks?
The equities markets seem much more focused on short rates -- the purview of the Fed -- than what the uber-liquid markets for longer-dated bonds are showing. The Fed began raising interest rates on Dec. 16, 2015, and for the next three years both manufacturing and total diffusion produced a steady uptrend. That changed in early 2019, however, and thus I believe any Fed actions to trim the front-end of the Treasury yield curve will be too little, too late.
I am in the process of raising funds for a new short-term, short-stocks trading vehicle. The stock market's Pavlovian Fed-bounce in the face of weak diffusion readings is just another indication that, even in June, winter is coming.
Shorting individual stocks is going to be a very lucrative summer trade and stay tuned for more from me on that topic in future Real Money columns.