The Better Good News Indicator - Aggregate Hours Worked Leveled Off
Forget the unemployment numbers-- as lots of commentators note (including Robert Reich), it's hard to take seriously a number that drops because people give up looking for work. And employment numbers don't usually reflect well whether people are working full-time or having their hours (and income) cut back due to involuntary part-time hours.
So I've long believed that the best number to track in recessions and recoveries is the aggregate number of hours worked across the whole economy. The Federal Reserve of St. Louis regularly tracks this number here and as the graph shows after the jump, the hours worked did seem to level off last month, after over a year of catastropic declines:
This could obviously just be a one-month blip so any hope is only contingent and celebrations should only tentatively begin once the hours worked begins to climb. But there is a reality that once recovery starts, the first actions by employers is more likely to be restoring hours to workers currently working only part-time, so the aggregate number of hours worked could very likely be the first place to see longer-lasting signs of recovery.
It appears, to me, to make a difference if this index is just total hours worked, or if it bears some relationship to working population. The latter changes over time, so I would think that something like total-hours-worked-per-1000-number-of-working-people would be more meaningful.
Can someone clarify this?
Thank you.
August 8, 2009 6:54 PM | Reply | Permalink
This is an absolute number of hours worked, so in normal growth periods it goes up rather than remaining level with population growth.
So yes, until the aggregate hours are growing significantly to include not only restoring work to the unemployed but also pacing population growth, we will not be in real recovery.
But the fact that the aggregate hours may have leveled off may be a sign of the end of the crash, if not the beginning of the recovery.
August 8, 2009 8:44 PM | Reply | Permalink
But somewhere around July of 2008 there is an actual INCREASE over a month.
August 8, 2009 9:41 PM | Reply | Permalink
Go to the source of the graph, and click on the "edit graph" button on the lower left. Then, switch to the other tab (there are only two). Change the measure from an index to percent change from a year earlier.
Congratulations. This indicator is leveling off at the worst year-on-year contraction in history.
.
This number of hours measure has a good history and can be used to gauge this recession against previous ones, and perhaps give us an idea of where we are in the cycle and how bad it really is.
There were six other extended declines in this indicator(excluding this one) since it was first produced in 1965 (1970-71, 1974-75, 1980-81, 1981-83, 1990-92 and 2001-04).
In those six contractions, this measure was negative for an average of 18.8 months, and fell to a depth of -4.3%. The lowest point was typically in the 8th month.
On that basis, and given that we already have 15 month on contraction (to -6.86% in June, the low), we would assume – if the pattern holds – that this indicator will remain negative until end-2010, a total of 32 months.
That’s some recovery, folks!
August 9, 2009 9:30 PM | Reply | Permalink
Thanks for the data. Your conclusion is a reasonable but not necessarily an indisputable one.
Even assuming that five recesssions is an adequate sample,the recovery in a recession which bottomed in the 15th month might differ in kind from ones that reached that point in half the time.
Either way.
On the rosy scenario side, this longer delay in consumer gratification indicates greater unsatisfied demand leading to a quicker snap back.And I can equally imagine qualitative arguments for the opposite position- for a more extended recovery.
My gut feeling is that you're right but I think we need to take your conclusion cum granum saltum.
August 11, 2009 6:10 AM | Reply | Permalink