Here is what I found:
_26389_image002.gif)
It would appear there is -- or has been until now -- a very high correlation between PMI Manufacturing (Blue, Left Hand Scale) and the Year-Over-Year Percentage change in Nonfarm Payrolls (Red, Right Hand Scale). That relationship has disintigrated in this cycle, as PMI turned up while employment continues to fall. This is a very troublesome chart. What it shows us, I think, is that the government can exercise some influence over production (e.g. Cash for Clunkers, homebuyer tax credit, both of which supported their respective industries), but it simply cannot exert sufficient influence over the labor market to make any meaningful difference. And, again, what happens as the life-support is removed?
5 comments:
Eh? Look at the last couple of months of all previous recessions. You get exactly the same shape -- the red line (payrolls) continues to drop, while the blue line (manufacturing) reverses and grows. It's just a lag. Your correlation will definitely bump up if you add a 3 or 6 month lag parameter. Nothing either new or necessarily troublesome.
What I think you are noticing is the lag between ISM growth and payroll growth. I've checked via the St. Louis Fed, and what your graph clearly shows is the historical ~6 month lag between when ISM turns up and YoY payroll turns up in % turns.
ISM manufacturing bottomed in December 2008. In this case the YoY% in payrolls bottom was almost certainly 8 months later in August 2009.
Here's what the numbers look like beginning in July, and continuing forward until next March under two scenarios: (1) a loss of 300,000 jobs every month, and (2) a loss of 100,000 jobs every month - but no gain under either scenario.
First percentage is "pessimistic" scenario, second is "optimistic":
7/09 -4.2 -4.2
8/09 -4.3 -4.3
9/09 -4.2 -4.2
10/09 -4.2 -4.0
11/09 -4.0 -3.7
12/09 -3.7 -3.3
1/10 -3.4 -2.8
2/10 -3.1 -2.4
3/10 -3.0 -2.0
In either event, the YoY% payroll graph turns up, just as it has done in the past.
If you check this relationship ISM
to monthly change in nonfarm payrolls it fits fairly well..positive job growth is coming
soon....
I'll concede the lag. Let's call it six months, because I'm told that six months produces the highest correlation. I'll re-run the chart and, without already having done so, will guess that it will still show a staggeringly wide, unprecedented gap. Either ISM has got to come down (and how much higher is it going to go?), or jobs have got some serious catching up to do. I simply don't see this chart in a "business as usual" way.
Danske Research has a very good
predictive model on ISM and it has
ISM hitting 60.
Post a Comment