Monday, July 15, 2013

Very early indications of macroeconomic headwinds

I have been a macro optimist since 2008.  Even now, indicators generally seem to be headed in the right direction, and housing is finally settling into a cyclical recovery period, which I think still will have positive forward effects in the labor market.

But, some very early indicators are starting to cause concern.

The lack of inflation in the face of QE3 could be a sign that the Fed has reached a point that the expectations channel is the only source of influence they have left and that the small inflationary effect of exchanging treasuries for cash that eventually ends up as excess reserves is dwindling.  Sober look suggests that we are seeing the effect of some ECB tightening.  There are also fiscal reasons why inflation might have dipped.  Expectations appear to be for a rebound, but further drops in inflation would be very bearish.

Here is a graph of bank credit (blue), and commercial loans (red).  Both are showing some flattening.  The green line is the inverted unemployment rate.  Normally, we would see unemployment bottoming as these measures flatten, and we aren't seeing that yet.


However, the labor market is a bit of a mystery:

The blue line is total unemployment.  The green line is the rate of unemployment if we subtract out unemployed workers who are on extended unemployment insurance.  Most of the excess unemployment in this cycle consisted of long term unemployed with extended unemployment insurance.  Much has been said about how high this number was and how slowly the number of long term unemployed workers has declined.  Interestingly, all of the decline in unemployment has been in long term unemployed workers with extended insurance (blue line - UEI):

Both short term unemployment and long term uninsured unemployment have flatlined since 2010.  Oddly, initial and continuing unemployment claims have continued to drop.  But, I would have expected this to show up in lower numbers for short term unemployment (less than 26 weeks).  But, again, oddly, an unemployment forecast based on initial and continuing claims (shown in the red "UE trend" line above) would seem to have given a fairly tight correlation to total unemployment.

(Admittedly, this is an in-sample regression.  However, in the other two most recent recessions, we see the same pattern:  At the initial recessionary unemployment surge, the unemployment rate rises more quickly than unemployment claims would predict.  This is because, as shown in the JOLTS data, the initial cyclical issue in the labor market is from anemia, not from too much turnover.  After the initial surge, the relationship between continuing unemployment claims and the unemployment rate becomes surprisingly linear.  The peak of the recovery is signaled by a breakdown in this linear relationship, as the unemployment rate now proceeds to a lower rate than would be predicted by unemployment claims.  This is the result of healthy new turnover in the labor market.  This is mitigated by new entrants into the labor force, reflected by an increasing labor force participation rate.  In summary, an unemployment forecast from continuing claims that overstates unemployment and an unusual rise in labor force participation both should be early signals of a cycle peak.  These signals are not currently active, although labor force participation will be difficult to read, because demographic factors are creating a declining secular trend line, so that the LFP rate may be flat in absolute terms when it begins to hit unsustainable levels.  These interpretations might be somewhat novel, which I am hoping means that I can use them to create profitable positions leading into the next cyclical downturn.  But, the these signals would predate recessionary markets by even 2 to 3 years, so the fact that they are not active probably means that my other concerns about the labor market are premature.)

The most optimistic reading I can give of this is that there is still a considerable amount of competition between short term unemployed and long term unemployed workers, and that eventually short term unemployed workers will re-enter the workforce more quickly as the long term unemployed are brought back into the workforce.

This is where the JOLTS data bothers me (12 month moving averages in black):

What JOLTS data makes clear is that recessions are related to sclerosis in the labor market as much as they are related to excessive layoffs.  JOLTS data is fairly young, but what we can see from the start of the bear cycle in 2007 is that the first signs of the recession came from flatlining quits (aqua) and hires (blue), which can be seen as early as 2006.  Job openings peaked in 2007 (red).  Layoffs (green) are a lagging effect, and they only show unusual activity in 2008 & 2009.  This was a peculiar cycle, so this process may not be universal.  I believe Scott Sumner's account that we had a small and manageable contraction related to the housing bust, which was developing in 2006 and 2007, and that the worst part of the recession was due to a liquidity crisis created by the Fed in 2008.  If that is the case, then the large bump in layoffs late in the cycle is not typical.

But, the current situation is a setup for bad news.  Hires have been flat for a year, even as Job Openings has increased.  This could be a very early bearish indicator.  And, in the past 3 months, the moving average for Quits is up only 9 thousand.  If Quits confirms a cyclical peak over the next few months, I am afraid that we will be in a very dangerous situation, where we will be lucky if unemployment drops below 7% before we hit the next downturn, and we could have a Fed that could be unwilling or unable to counter deflationary forces.  If we continue to be saddled with federal policies that are hamstringing employers, the downside risks in that scenario would be high.

A resumption in the fall of initial claims and a continuation of the fall in continuing claims would be encouraging, but I consider the Quits rate in the JOLTS data to be a key very-early indicator of potential problems to come - at least as important as Hires.

Corporate profits have leveled out, which deserves watching, and the number of unemployed per opening is starting to level out.  This is a bearish indicator, although the number of unemployed persons per opening is still much higher than in previous recoveries, so it is another indicator that is odd in the current context.

On the positive side, bank loan spreads, demand & standards all continue to look strong, housing is strong, and currency in circulation is strong.  It would be unlikely to have a fresh downturn before these indicators turn sour.  And, possibly most bullish of all, the yield curve is relatively steep, and has recently steepened significantly.  I would expect the yield curve to flatten coincidentally with these other leading indicators.  On the other hand, this is a unique situation, where the Fed has had little success in creating inflation, and has built up a tremendous balance sheet in the process.  We have short term rates pegged at zero, even though the Fed hasn't bought a short term bond in years.  Normally, this would come about partially from rising short term rates, and there might be some combination of a movement of demand for credit from the long term to the short term out of concern for the nominal economy, or overtightening of monetary policy by the Fed.  But, either demographics, foreign politics, or fear, have put so much money into short term risk free securities, I am not sure anything could pull short term rates from zero.  That means a low term spread would have to come entirely from lower long term rates.  I would consider a sharp decline in long term yields to be very bearish, but I wonder if we could see a recession without such low rates.  Maybe this is the cycle where that indicator breaks down.

Normally, downturns wouldn't happen in the face of an ostensibly accommodating Fed.  But, I'm afraid that we could find ourselves in a position where the Fed, having failed to inflate the currency, watches us descend into a long stagnation because they don't want to risk hyperinflation with all of the duration risk they now have.  I suspect that if they were able to work more outside the box, four years ago they could have bought a few hundred billion dollars worth of S&P500 index funds, and we would have seen an immediate inflationary boost without all this hand wringing about the risks of QE.  Or, for that matter, buying floating rate bonds would have helped.

The awful kinds of fiscal responses we could see in the possible bearish context would be likely to make things worse.

This is all still speculative, though.  I consider recent rate moves to be a bullish sign.  What I am describing above would be the first signs of a problem among the earliest indicators I follow, and at this point, the concerns haven't been confirmed.

No comments:

Post a Comment