The Great Recession ended in July 2009 (give or take)

One might reasonably think the process for mapping the beginning and ending of recessions is a fairly transparent and straightforward process.  After all, the economic cycle is integral to the lives of most everyone I know (and to the remainder, I envy you) but the truth is that these calls come months, if not a year or more after the fact.

Just who makes these calls anyway?  Well despite our massive investments in the public sector and related data collection activities, there is in fact no one in the DC apparatus... no presidential appointee, no congressional committee, nay, not even an intern, charged with this rather important task. 

Most tune in to the reports issued by an entity called the National Bureau of Economic Research (NBER).  The NBER is a private, non-profit group that does quantitative and fundamental analyses on the economy out of the brain trust cities of Cambridge, Palo Alto and New York.  And it really is the only one that declares when recessions begin and end.  Typically these declarations come months (if not a year or more) after the fact. For our current Great Recession, they didn't identify the start date of December 2007 until November 27 2008!

This sort of strikes me like an ump who waffles on calling the pitch, when it's kinda sorta needed right there.  According to the Rulebook of the Economy however, the umps at the NBER have the option of heading home, and watching the replay from five different angles before making the call.  Meanwhile the rest of us hold our collective breath, hope our jobs don't go poof, and all the while policy makers waste time arguing whether or not there is a recession and what should (or should not) be done about it.

So what goes on inside the NBER?  Well as a card-carrying member of the Public I don't know either.  Now in reality I like the organization and I poke fun at them in jest, but seriously, in the age of Now the wait just seems damn slow.  I can't understand why it takes so long to "declare" a beginning of a recession or its end.

I have a superabundance of enthusiasm, a trait inversely proportional to patience.  In layman's terms this means that for me, spending three long years to see the entire Lord of the Rings trilogy practically drove me batshit.  Lesson learnt, I chose to wait and watch seasons 1-4 of "Lost" all in one Epic DVD blast.  I was much happier.

Now as a Quant, when I see gigabytes of data (in many cases floating free on those internets)  I want to wrestle with it, pound it, spank it, and above all make it give me answers.  The statistical methods are a means to enlightenment, a way for me to cut through the bullspin and identify where we are now and what conditions will likely prevail tomorrow. 

 And when I see a pattern emerge from columns of dummy variables I screech, buckle down and compute.

I spent some time poring over different public data sets in order to find a better real-time indicator of the economic cycle.   In the matter of recessionary starts and stops I discovered one that appears to be uniquely suited to the task, and I have been following it monthly since the first term of the second Bush.

I found the keystone within the Bureau of Labor Statistics, which among other series discloses US Non Farm Payroll data. Most media report the same data, which is monthly or weekly estimated changes but there is one iteration of the raw data which is routinely overlooked and must be computed manually.

1.  Take the raw data series (curiously, better delivery of BLS from the St. Louis Fed Reserve) and import to Excel. 

2.  Text-to-column the series

3.  Compute simple Year-over-Year percentages, subtracting integer '1' from the result.

4.  Overlay NBER Cycle Dates.

5.  Multi-line plot to get the following (click to enlarge).

This series covers monthly data for 70 years, from 1940 - Feb 2010.

The blue line shows the YOY change in Non Farm Payroll data (seasonally-adjusted, though NSA is comparable).

The vertical red lines are the NBER-declared beginning and end dates for officially-declared recessions in the United States.  The 0% YOY horizontal line is set about the middle of the chart, to better clarify the positive and negative deltas in relation to the recessionary periods.

Note that nearly every declared end to a recession (no matter when it was declared, often much later) occurs around an absolute trough bottom in the YOY series.  This doesn't mean the pain is over, but rather that the rate of job losses is slowing down instead of accelerating.  The economy grinds its way back towards the growth line (0% is the inflection) even though Main St is still shrinking, until it reaches positive growth again.

The precise beginnings of recessions are a little more nebulous using this method/series but generally coincide with:

• A steep declination in positive job YoY that reaches the ~0.75-1.30% zone, and
• Continues to fall through the 0% floor into negative territory.  

When these happen a recessionary period is generally declared.  The initial model I built using these momentum conditionals had pegged the start of our wonderful Great Recession as around Oct-Nov 2007, which is not far from the NBER-declared date of December 2007, a call they issued in November 2008.

Let's zoom in closer to the Naughty decade ('00+):

The telecom-dot com recession occupies a relatively short period at the beginning of the decade.  Our current recession is pegged as starting in December 2007 and no end has yet been declared.  However notice the trough was reached around July 2009 and job losses have been slowing on a YOY basis ever since.  If whatever methodology used by the NBER holds true for this cycle, then I believe this supports the notion that eventually they will declare the end of the current recession as being somewhere in the summer of 2009.

The reason I think this pattern holds through many different cycles (and they all had varying causes) is that non farm payroll data in absolute form is the best indicator of growth and contraction, period.  These data estimate how many people are actively collecting paychecks, and represent creative destruction cycles as one industry (or cheap money cycle, asset class, etc) rises, plateaus, fades and is replaced by other growing segments of the economy.  

This to me is an elegant series.

The purpose of this indicator is not to displace careful study that is needed on the business cycle.  Rather the idea is to provide early warning indicators to policy makers, bankers, business leaders and academia so that we can take actions to keep the economy growing in a sustainable way, and be aware well in advance of a severe decline so we can mitigate the impacts on ordinary Americans.  If the exact dates end up being revised a bit back and forth later, so be it!  Just as doctors help us monitor our health in a preventive sense, we can do better to monitor our overall economy and communicate these findings to the public in a sensible way.  

While many in our country were bearing the early brunt of the recession in 2008-2009, there were too many at the controls who were in denial that a contraction was underway.  Bond measures, social programs and war spending all amped up as if we were just taking a breather.  As we have seen, those assumptions proved baseless and ended up doing more harm in the long run.  

Forewarned, forearmed.