July 25, 2010
Update on the bumpy recovery
I was curious to take a look at how Mike Dueker's Business Cycle Index and other measures assess the current situation.
From Russell Investments:
Current forecasts show a considerably deeper setback than last month. The stumble in the economy is real and reflects expected lackluster demand for exports, capital expenditures and consumer durables.
The most likely path for the BCI calls for improvement in the cyclical state of the economy this autumn rather than a double-dip recession.
Bill McBride, with whose assessments I've found it's wise not to disagree, has this take:
I agree with Krugman that a further slowdown-- following the below trend first half of 2010-- will definitely feel like a recession-- and it will probably lead to an unemployment rate "double dip".... The 2nd half slowdown is here. I still think we will avoid a technical double-dip recession, but it will probably feel like one.
Federal Reserve Chair Ben Bernanke last week reported that the Federal Reserve has an outlook that has also darkened slightly in the last few months but nevertheless remains brighter than some private analysts:
In conjunction with the June FOMC meeting, Board members and Reserve Bank presidents prepared forecasts of economic growth, unemployment, and inflation for the years 2010 through 2012 and over the longer run. The forecasts are qualitatively similar to those we released in February and May, although progress in reducing unemployment is now expected to be somewhat slower than we previously projected, and near-term inflation now looks likely to be a little lower. Most FOMC participants expect real GDP growth of 3 to 3-1/2 percent in 2010, and roughly 3-1/2 to 4-1/2 percent in 2011 and 2012. The unemployment rate is expected to decline to between 7 and 7-1/2 percent by the end of 2012. Most participants viewed uncertainty about the outlook for growth and unemployment as greater than normal, and the majority saw the risks to growth as weighted to the downside.
Elsewhere, the Aruoba-Diebold-Scotti Business Conditions Index, which was thrown into a deep funk by the lost temporary Census jobs (orange plot in the figure below), cheered up some at the modest increase in industrial production reported July 15 (magenta), and slid back down with the continued stagnation in the new unemployment claims numbers released July 22 (red).
Or, if you prefer a steadier bottom line that is less prone to bobbing about with each new data release, you can always turn to the Econbrowser Emoticon, which has been stuck in neutral (signaling sluggish growth) since August 2009. Which, by the way, remains my call today.
Posted by James Hamilton at July 25, 2010 06:40 AMdigg this | reddit
JDH: ...the Aruoba-Diebold-Scotti Business Conditions Index, which was thrown into a deep funk by the lost temporary Census jobs
I don't understand why the lost Census jobs should have thrown the index into a "deep funk." The loss of the Census jobs was not a stochastic event; it was a known quantity and I would have expected any forward looking index would have anticipated that. According to your link the index does make some kinds of data adjustments so I would have expected it to adjust for something as predictable as expiring Census jobs.
I have no idea why FOMC members expect 2010 growth to be in the 3 to 3.5 percent range. That's looney tunes. And goofier still is to expect growth in 2011 to average 4 percent. What will drive this growth? Are they expecting a big Democratic victory in November that would give Harry Reid 62 votes? I don't get it.
Posted by: 2slugbaits at July 25, 2010 07:34 AM
2slugbaits: ADS is simply a mechanical data-processing algorithm rather than a human being. ADS looks at the monthly decline in total employment and signaled an alarm based on the historical correlations.
I review some assessments of both humans and computers here and you can take your pick.
Posted by: JDH at July 25, 2010 07:50 AM
The question is whether "stuck in neutral" can be a likely state for the economy following a debt crisis.
First, the fiscal stimulus turns into a sequential headwind in 3q. Presumably, the economy requires some jobs (income) or business investment momentum to overcome this drag.
Second, a number of debt service projections more or less require strong nominal growth. A period of slow growth, for instance, may produce non-linear effects on fiscal deficit forecasts. Improvement in debt service implies polar-opposite paths of retrenchment in the numerator or growth in the denominator. The "animal spirits" tendency towards reducing debt outstanding is deflationary and requires, again, a great deal of nominal GDP momentum to overcome.
Third, the world economy is still structurally dependent on robust U.S. household consumption growth. In its absence, it is not clear that many emerging markets (China, Brazil) can sustain the trajectory of their own recoveries.
Fourth, and finally, the downside to a double dip is rather severe: escalating l.t. unemployment, rising consumer credit delinquencies, further need for banking system rescues, etc. In other words, many of the structural problems that caused the 2008 crisis are still present, and would re-emerge were growth to slow to a crawl. There is a tendency for asset markets to hedge against this outcome, and such hedging is, again, deflationary.
I can understand (even if I don't agree) why some argue that the solution to the above is for the Fed to adopt a high nominal growth target. What I can't understand is how we can have a solution that assumes slow growth ahead.
Posted by: David Pearson at July 25, 2010 10:42 AM
I have no idea why FOMC members expect 2010 growth to be in the 3 to 3.5 percent range. That's looney tunes. And goofier still is to expect growth in 2011 to average 4 percent. What will drive this growth? Are they expecting a big Democratic victory in November that would give Harry Reid 62 votes?
Democratic supermajority = GDP growth?
Posted by: W.C. Varones at July 25, 2010 11:09 AM
One big difference between the Russell index and the Conference Board's leading indicators, both which show a moderate slowing, and the ECRI WLI, which is showing a much more pronounced slowdown, is that both use the slope of the yield curve, while the ECRI uses the spread between Treasuries and commercial bonds.
Anything using the Treasury yield curve is going to show a marked reduction in the degree of the downturn. That's what the idiot who claimed there was "0% chance of a slowdown" was using.
But surely, the reason the leading indices use the yield curve is that it almost always inverts before a recession. So it's used as a signal of what the market is seeing.
But it clearly is useless in a period when the Fed rate is at 0%. Even after the surge in Treasuries over the last few months, the yield curve is the same area as it was in 1985, 1995 and 2005. Are the credit situations in those years really analogous to what we're seeing now?
Someone should run those indices leaving out the yield curve and see what they say.
Shouldn't the ADS index also been thrown by the previous addition of the census jobs? Dosen't that make it a wash? or is the trend too short?
Posted by: Bob_in_MA at July 25, 2010 11:49 AM
W.C. Varones: I'm not arguing that a Democratic supermajority is a sufficient condition for GDP growth, but I do think it's a necessary condition. It's really hard to see how we have anything like strong GDP growth as long as the GOP has the votes to filibuster any and every stimulus plan that comes along. Or worse yet, God forbid a GOP majority in either House. I think the numbers are telling us that we're pretty much headed for a near worst case scenario; one in which we continue to run large deficits year after year, but not large enough to really shock the economy out of a low output equilibrium position. At the end of the day we will end up with much higher debt levels and much lower growth than we would have had if we had done the right thing in the beginning.
Meanwhile we can watch a real world experiment because Britain is about to throw itself over the cliff. If the Tory govt's policies work, then we ought to give austerity and Hoover economics a second look. But the smart money says that the Tory policy fails and by this time next year the govt collapses and Britain has to go back to fiscal stimulus albeit with much less room to manuveur than they have today.
Posted by: 2slugbaits at July 25, 2010 12:13 PM
I've developed a new model using the predictive powers of the yield curve, but adding changes in bank lending.
It pretty much parallels the yield curve previous to recessions, dropping to zero or so. It also explains the slow recoveries after the 1991 and 2001 recessions.
It dipped below zero in Sept 2009, and right now is about where it was on the eve of the 2007 recession.
Unfortunately, it has two fudge factors, but who's perfect. ;-)
The blue line is the normal 10y-3mo, the orange line is my new version.
Posted by: Bob_in_MA at July 25, 2010 01:54 PM
A couple of quibbles, a reply comment, and a request:
1. Druecker's model, after a spot-on performance during the recession, didn't predict the last 9 months too well. It had strong employment growth in 4Q 2009 followed by a slowdown to near 0 in 1Q 2010, followed by mild growth.
2. The "period of census hiring" overlaid by Russell Investments is not accurate. In February and March, only about 50,000 census jobs were created. The only month really explained by census on the upside was April, with almost 400,000 census jobs created, which is going to be totally unwound between June's -200,000 and July's -175,000.
3. Bob in MA: I think your model "proves too much." It seems to also predict "jobless recoveries" following almost all recessions since 1948, and seems to lag payrolls as often as lead.
4. One theme here is "deflation is different." Some indicators useful in periods of inflation (like the yield curve) seem to lose their usefulness during deflation, or operate on a different time scale. Has any research been done specifically as to leading indicators during times of deflation, or transition from inflation to deflation, e.g., Japan in the last 20 years? Or commercial vs. government paper during the 1930s? If anyone is interested, if you google "Economic Indicators during the Roaring Twenties and Great Depression", you will find a 5 part series in which I examined the behavior of the yield curve and "real" money supply during that time period.
Posted by: New Deal democrat at July 26, 2010 04:59 AM
I still like the ADS Index, although I'm a bit perplexed as to why it's shaped like clouds, little mounds of parabolas. I'm not sure I experience economic activity in that way.
Posted by: Steven Kopits at July 26, 2010 05:08 AM
The chart from Russell shows a long term prjection of +200,000 jobs a month. At that rate, how long would it take to get employment levels back to a point as if the recession never happened.
In my view, the economy has not "snapped back" and we are now on a permanently lower trajectory.
Posted by: MikeR at July 26, 2010 07:23 AM
commercial vs. government paper during the 1930s
Throughout my readings a review is coming very often as a reference,to the economic situation in the USA during the 30s. This review was going by the name of:
(I have been unable to trace this review through Google)
Posted by: ppcm at July 26, 2010 07:26 AM
Maybe rather than the yield curve at the zero bound you should use quality spreads.
Historically, quality spreads are a very good leading to concurrent indicator of economic weakness.
Posted by: spencer at July 26, 2010 11:06 AM
Population growth in the US is around 1% a year and has been for a long while (more than 4 decades). For a labor force of around 150 million, this means that each year there's about 1.5 million more workers, or about 125000 per month. Actual job growth has to exceed this number before the percentage of employed people goes down.
In reality, it takes much more than that. Over the Bush years the labor force participation rate has continuously dropped (in total about 1% between 2001 and 2008), indicating that more and more people that should be able to work, did not find work or did not look for it. This represents a net loss of about 1.5 million jobs during the "good" times. The Great Recession that capped Bush's stint in office took out another 1.5 million jobs (and reduced the quality of a lot of the others, no doubt, but participation rates are only concerned whether you have some kind of job, not whether it's a good one).
The participation rate is now back at the point where it was in 1986. Maybe it will take another Clinton boom to get the participation rate to go up again, but it is more likely that the high participation rates won't come back completely, i.e. less people of working age will be employed from now on. The internet and real estate booms were just a phase...
Posted by: endorendil at July 28, 2010 06:08 AM