Those trading, investing, making policy and otherwise betting on an upturn based off of employment data are betting on the wrong horse. Although buoyed by the initial claims data this morning, this group was obviously disappointed by the June jobs data and reacted by issuing policy mea culpas, widening credit spreads, dumping equities and buying Treasurys ( see post "Ten-Year Treasury Yields – What's Coming Next" ).
Although the NBER's standards for setting a trough mean no job growth no recovery, employment data do not offer an early indication of an upturn. First off, month to month job changes are essentially random and so there are no steady monthly progressions during up or down cycles. And then there are the sizable revisions (May was revised up and upside revisions are usually signals of economic growth).
One big source of revision comes when the BLS re-benchmarks the survey and fixes what the birth/death adjustment estimates. Inexplicably, for the 12 month period ending in June the BLS added 831,000 jobs from net new enterprises against 5.8 million jobs lost. In May the 12 month add was 811,000 and in March it was 717,000. Yes, during the worst recession since at least the 1980-82 period if not the post-war the BLS job birth/death job add is increasing.
Beyond the month-to-month data, using labor data to gauge the upturn is less reliable because upturns don't need as much labor as they used and they don't need new workers for a increasingly long period of time. As the table below illustrates, after the post-war recessions up to and including the 1981-82 event it took a median 14 months until the jobs total matched the peak prior to that recession's start.
After the 1990-91 recession it took 27 months for jobs to fully recover. After the 2001 downturn, it took 43 months. And that is just the absolute number of jobs – I am not adjusting for the increase in working age population which would lengthen the job recovery period. As for why it is now taking longer for a recovery to boost jobs my view is, first, that the economy is more service oriented and its firing/rehiring patterns are much less dynamic than manufacturing. Second, increased consumer demand flows to jobs overseas as so much more consumer product is satisfied by imports.
All of which is to say the coming recovery will be even worse when it comes to hiring people. This is much more a jobs recession in the service industry and financial services in particular. In addition, imports are an even greater percentage of consumer items than in the previous two recessions. Lastly, housing construction often leads the economy out of recession and that will not be the case this time around. Perhaps most important in delaying job growth is the structural shift from the highly leveraged economy that boosted household and business spending and dropped unemployment to 4.4%. The coming upturn will offer nothing of the sort, nor should policy aim the economy in that direction if there is going to be balanced sustainable growth.
There are early signs of recovery and they are in indicators of business activity. But recovery isn't coming from labor data and, as an aside, low future jobs growth and high unemployment rates means those inflation worriers should find something else to fill their time. More on early signs of recovery in the next post.
Recovering Lost Private Employment
Months To Return To Prior Peak*
(Dated From Trough)
*Median from 1948 to 1982 was 14 months