NBER: Double Dip or Banana Split?

by the Consumer Metrics Institute

We founded the Consumer Metrics Institute precisely because we felt that the economic bureaucrats in Washington were out of touch with the economy that most of us live in. They remind us of those patients sitting in wheelchairs in the "memory impaired" wards at nursing homes: with crystal clear recall of 1937 but no clue about what they ate for breakfast. Thank you, NBER, for making our case.

In contrast, we measure what consumers are actually doing on a daily basis. If, for the sake of argument, we accept that we are not experiencing just "one big scoop," but rather a "double dip" (thereby making the 1930’s a "banana split"), we can show evidence that the first dip ended early in 2009. Arguably, we’ve been monitoring in real-time what could be viewed as two independent consumer demand contraction events that were separated by a stimulus induced "sugar high" last summer. If so, the first dip is ancient history. What is important now is future course of the second dip — which may just now be revealing itself.

We are far enough "up-stream" in the economic cycle that we can measure changes in consumer demand for discretionary durable goods long before those changes flow "down-stream" to the factories and the GDP. From our up-stream vantage point the "double dip" is not hypothetical, but rather something that we have been watching unfold on a daily basis since January. Now, for the first time, we may have measured what will be the worst of the second dip when it eventually hits the factories — all because, ironically, our data has started to improve.

Over the 45 days from August 1 to September 15, our Weighted Composite Index has improved substantially, rising from recording a year-over-year contraction rate in excess of 9% to recently registering a contraction rate much nearer to 3%. This is the largest positive movement that we have seen since late 2009. That said, it is important to remember that consumer demand for discretionary durable goods is still contracting, albeit at a slower rate. But the improvement has stopped (at least temporarily) the decline of our 91-day trailing quarter average (our Daily Growth Index):

(Click on chart for fuller resolution)

Our Daily Growth Index reached a -5.86% contraction rate on September 12, which was fully 97% as bad as the worst contraction rate reached during the "Great Recession of 2008" (-6.02% on August 29, 2008). A calendar quarter of comparable GDP growth has occurred among only 1.29% of all quarters of U.S. GDP growth recorded by the Bureau of Economic Analysis of the U.S. Department of Commerce, since the spring of 1947. This corresponds to level of contraction that should be expected only once in 19.4 years, and it comes close on the heels of the 2008 contraction that should occur only once in every 21.4 years.

One of the tools that we have used to monitor the 2010 contraction event is a chart that we call our "Contraction Watch," which overlays graphically the day-by-day progression of the current 2010 contraction onto the "Great Recession of 2008":

(Click on chart for fuller resolution)

In the above chart the two contractions are aligned on the left margin at the first day during each event that our Daily Growth Index went negative, and they progress day-by-day to the right, tracing out the daily rate of contraction. This chart conveys important information about the 2010 event, in particular how it differs in profile from the "Great Recession of 2008." It has now lasted three weeks longer than the "Great Recession" and is perhaps only just now forming a bottom. Furthermore, that bottom is very nearly as low as the one experienced in 2008. Even if the 2010 contraction immediately starts to retrace the recovery pattern seen in 2008, we should expect at least another 120 days or so of net contraction before the consumer portion of the economy is growing once again.

We have previously pointed out that the true severity of any contraction event is the area between the "zero" axis in the above chart and the line being traced out by the daily contraction values. By that measure the "Great Recession of 2008" had a total of 793 percentage-days of contraction.

Similarly, the current 2010 contraction has just reached 592 percentage-days, and its severity can be visualized as the amount of area covered by blue in this chart:

(Click on chart for fuller resolution)

The blue area above already covers about 75% of the area covered by the 2008 "Great Recession", and the curve has only just begun to start back up. Looking ahead, should the 2010 event recover from its bottom exactly like the 2008 event did, it would still experience another 466 percentage-days of contraction before ending — resulting in a grand total of 1058 percentage-days of contraction for the 2010 event, fully 33% more severe than the "Great Recession of 2008."

That probably bears repeating: if the 2010 contraction we are now monitoring in consumer demand for discretionary durable goods scales to the full economy as faithfully as the "Great Recession" did, the second dip will, at minimum, be 33% more painful than the first dip and will extend at least half again as long. This, of course, assumes that stimuli comparable to those seen in 2008-2009 will be available to cause such a recovery during 2010-2011. Furthermore, the upturn that we measured in 2008 started when unemployment was still at a 6.1% rate, substantially better than we are observing now. Absent fresh consumer stimuli and dropping unemployment rates, the consumer demand contraction we are witnessing now could very well linger even longer.

Supporting that concern is the shape of the 2010 contraction in the above charts, which is significantly different from that of the "Great Recession of 2008." Of particular interest is the fact that in 2010 consumer demand plateaued for some time in a zone between 1% and 3% contraction from about day-25 through about day-180, before falling off the plateau. Since our data is always reflecting year-over-year changes in consumer demand, we had anticipated a sharp dip in our index as an inverted reflection of the stimuli-induced "green shoots" of late last summer. The long plateau described above, however, is not a reflection of any such now lapsed stimuli — and as such it may be a new normal baseline for a lingering consumer contraction. Before we get too excited about a new recovery we will wait until our Daily Growth Index breaks significantly above the plateau levels visible in the 2010 line within our "Contraction Watch."

We are monitoring the behavior of internet shopping consumers on a daily basis. Those "up-stream" consumer activities will flow "down-stream" to factories and the GDP over the course of weeks or quarters. It’s really not unlike being far up a great river and watching a water-level gauge predict that communities further down the river will be flooding catastrophically in a few days or weeks. Although our flood-gage may have just peaked, unfortunately the damage further downstream remains inevitable — it simply hasn’t arrived yet.