Recession Watch

Recessions are officially defined by the National Bureau of Economic Research; however, they will wait 1 to 2 years after events to see all the data before they make announcements.  The Economic Cycle Research Institute produces indexes weekly and monthly; however, they don't call recessions. 
By combining the two, we can see when a recession starts in almost real time.  The last three recessions were July 1990-March 1991, March 2001-Nov 2001, and Dec 2007-June 2009.  Looking at just these two periods, were there any data that signaled the start of the recession?

Yes, first the coincident monthly index growth rate.  When this drops below 0, we are in a recession.  This was below zero during Oct. 1990 to Jun. 1991, so this matched the recession with a 3 month lag.  It was also below zero during April 2001-Mar 2002, so it matched the recession is a 1 to 4 month lag.  It was below zero during May 2008 to Dec. 2009, so it had a lag of 6 months.  So it is safe to say that if the monthly growth rate is below zero then we are in a recession, with a possible lag of up to 6 months (i.e. the recession could have started up to 6 months earlier).

Looking at the weekly leading index growth rate, it was below 1 in April 1990 and stayed there until Feb 1991.  So this match the recession exactly except was early by 3 months.  It dropped below zero again in Dec 1991 (2 weeks), Nov 1994-April 1995 (20 weeks), Aug 1998 (23 weeks).  In May 2000, it dropped below 1.0 and stayed there until Jan 2002.  So again this matched except it was 11 months early on starting and 2 months early on ending.  It next dropped below 1 in Aug 2002 (37 weeks), July 2004 (20 weeks), Jan 2005 (5 weeks), May 2005 (6 weeks), June 2006 (21 weeks).  In Aug 2007 it dropped below 1 and stayed there until June 2009.  So this was 4 months early.  Since then the growth rate was below 1 in May 2010 (30 weeks), Aug 2011 (32 weeks) and April 2012 (9 weeks so far).

So I think any period where the weekly growth rate is below 1 for more than 6 weeks is what I call a mini-recession. A mini-recession may or may not turn into a recession.  Since January 1990 these are:

April 20, 1990-March 1, 1991
Nov 25, 1994-April 7, 1995 
Aug 21, 1998-Mar 12, 1999
May 12, 2000-Dec 28, 2001
Aug 2, 2002-Apr 4, 2003
July 16, 2004-Nov 26, 2004
Jun 23, 2006-Nov 10, 2006
Aug 7, 2007-June 12, 2009
May 28, 2010-Dec 17, 2010
Aug 19, 2011-Mar 23, 2012
April 20, 2012-current

So there were 11 mini-recessions, of which only 3 turned into recessions.

What separates a recession from a mini-recession?  If the coincedent growth rate is negative at the same time.  So what are we to conclude?  That we are in a mini-recession, but we are not in a recession at this point.

One more question is, what is the maximum negative weekly growth there has been without being in a recession?  On 7/23/2010, the weekly growth rate was -10.9, but we were not in a recession.

==================
Update 11/9/12: the last mini-recession ended on 8/24/12.