Tuesday, December 29, 2015

Do the Underpants Recession Gnomes have a case after all?


 - by New Deal democrat

A couple of months ago, I lampooned what I called the "Underpants Gnomes theory of importing recession," which went like this:

1.  China is undergoing a slowdown
2.  This has spread to China's suppliers, who are undergoing worse downturns.
3.  ??????
4.  This will bring about a U.S. recession 

I concluded that "Until someone comes up with a credible scenario for Step 3, all we have is the Doomer version of the Underpants Gnomes."  Since that time, however, there has been another deflationary pulse in commodities, and another run to new highs (since cooled off a little bit) in the US$.  In short, a renewed surge in the US$ is doing a pretty good imitation of whatever would need to fill in line #3. 


That has caused me to put some further thought into the matter, and to consider modifying my position. The starting point is that, generally speaking, as China in particular becomes an ever larger factor in the global economy, the US feels more of the global effects, in a way it has not done in the last century.  These will presumably become ever more important over time.  It seems natural that the identities and strengths of the various long and short leading indicators should gradually change as well.

With that in mind, let me next make two analogies.

First, the graphic analogy for the global slowdown is this graph from 3 years ago by ECRI:


In this graph, China occupies the middle line, because that is the epicenter of the commodity collapse.  China's "upstream" natural resource suppliers are the bottom line -- they get hit even harder.  The US is the top line, the "downstream" consumer.  While US exports suffer, the US gets cheaper consumer goods in return, more than offsetting the effects of the downturn. While there is a drag on the US economy, the line still points up, albeit at a subdued level.

Second, let's analogize to the Titanic.  The ship had watertight compartments sealed at the bottom, but not the top. She sank because the water overtipped one compartment at a time.  Had the ship's compartments been sealed at the top instead of the bottom, the Titanic would not have sunk.  But its bow would have been very low in the water.  So while it presumably could have limped into New York harbor, it would not have been a good idea to try to do so through a Nor'easter!

The new global deflationary pulse that began in October has caused the slope of all three lines in the ECRI graph to point lower than otherwise - even though the US's line is still positive.  This is  the equivalent to the US economy of causing it to sit lower in the water.  Domestic problems it could otherwise withstand might be enough to cause it to sink.

Let's quantify that a little bit.  Suppose 80% of the US economy is domestic, and 20% is directly exposed to global forces. If the domestic economy is growing 2%, and the globally exposed economy is declining 5%, the overall economy still grows: (.8 x 2%) + (.2 x -5%) = +0.6%.

But suppose the intensity of the globally exposed economy's downturn doubles.  Now we have (.8 x 2%) + (.2 X -10%) = -0.4%.  The overall economy contracts.  

As I wrote last week, the effects of the global slowdown have not entered the US economy through the traditional long leading indicators, but rather directly via coincident indicators like production of steel, and rail, truck, and shipping transport, and generally into sectors of industrial production.  The big harbinger was the surge in the trade weighted US$ in late 2014:



This makes me think that the leading sectors of the US economy, in particular housing and cars, may not have to decline for as long a time, or as much as normal, for such declines to signal an oncoming recession.  In this regard the 2001 recession, which was led by business and did not affect consumer spending very much, is a decent working template:



Housing permits only fell 175,000 at their worst, and vehicle sales declined 2.4 million on an annualized basis. Industrial production declined -2.3% before the onset of the 2001 recession; it is only down -1.3% from their peak now.

It's worth emphasizing, however, that every other long leading indicator did turn over before that recession began, and in particular real money supply and the yield curve have come nowhere near their recession signals at this point, while housing permits after going more or less sideways for 4 months, finally made a new high (ex-NY, the state responsible for the May-June upward distortions) in November.

It is usually wrong to think that "it's different this time," but I have no problem *modifying* the usual rules to take into account the growing secular trend of a more balanced globe.  Thus, *if* I see a decline in housing and cars that is about half of what would in the past have signaled a recession, and at the same time if there is a renewed surge in the US$, I do not think I will wait longer before going on recession patrol.