By Pater Tenebrarum
Are Recent Small Improvements Meaningful?
In recent weeks, evidence has emerged that the weakness in the manufacturing sector has begun to spread to the services sector as well (see e.g. Mish's summary on the Markit Services PMI as well as the worrisome state of services activity on a global basis). This is not particularly surprising: as we have frequently pointed out, economic slowdowns and busts always tend to hit the capital goods industries first.
At the same time, a few data points in manufacturing have actually slightly improved - which is to say, that the pace of the declines has slowed and was in some cases not as bad as expected. This always seems to happen: as soon as economists adapt their previously overoptimistic expectations to a recent worsening trend, they will promptly be proven wrong again, as the pace of the trend slows.
Few things in this world are more certain than the fact that the consensus forecasts of mainstream economists are likely to be wrong. One can probably easily improve on their hit rate by merely flipping a coin. It is quite eerie actually - it is as if they were all Gartman clones.
However, the question that interests us is this: do these improvements mean anything? Our friend Michael Pollaro has provided us with updated charts on the situation, several of which we present below. As we have repeatedly said in previous updates, the probability of a recession has clearly increased - but the evidence that one has already begun or is about to begin shortly is not yet definitive. The following chart depicts the year-on-year change rate in the value of new factory orders, unfilled orders and inventories for non-defense capital goods excl. aircraft. Focus on the areas highlighted by the black circles and green rectangles.
Non-defense capital goods excl. aircraft: y/y percentage change in new orders (red line), unfilled orders (blue line) and inventories (gray area).
There have been two occasions when gyrations similar to the recent ones have preceded recessions - shown in the first two black circles (2000/2001 and 2007/2008). The important thing here is that the short term ups and downs just prior to the beginning of the recessions were actually meaningless. In both cases, there were one or two occasions when the data improved in the short term, but it turned out in retrospect that these signals were misleading. A much sharper downturn followed shortly thereafter.
The green rectangles show "false signals" - in these two cases, capital goods orders and associated data (such as industrial production, manufacturing sales, etc.) also declined close to recessionary thresholds, but no recession ensued. Let us consider these two cases: the first occasion was in 1997-1998. This was during the Asian, resp. Russian crisis. At the time, the US economy was in the middle of a major credit expansion and in the grip of the greatest stock market mania in history.
Japan's bubble economy could not be slowed down by the crash of 1987, and the same happened with the US bubble economy in the late 1990s: a little crisis overseas could not stop it from resuming its expansion, especially with Alan Greenspan's Fed immediately easing monetary policy further and the GSEs vastly expanding their mortgage purchases so as to pump additional liquidity into the banking system.
The second occasion was in 2011/2012. This was at the height of the euro area debt crisis. What happened? The Fed embarked on Operation Twist, and shortly thereafter on "QE3" (the ECB concurrently began to implement various expansionary measures, such as its LTRO program). Bubble activities resumed with a vengeance.
So there were very good reasons why the beginning corrections were aborted on these occasions: in both instances, the Fed loosened monetary policy significantly. However, this time, the Fed is reportedly eager to tighten rather than loosen monetary policy (for the time being, anyway). And although this "tightening" has so far only had a small effect on credit and money supply expansion (although the latter has begun to slow down), it makes it less likely that the correction can be arrested and reversed by reigniting yet another round of bubble activities.
Regional Data vs. the ISM
It is also worth noting that the improvement in data was confined to the national ISM. The same cannot be said of any other relevant data series. Here is a comparison of the combined regional Fed manufacturing survey data with the data reported by ISM. First the headline numbers:
Headline numbers: combined regional Fed surveys vs. ISM - a huge gap has opened up. The last time this happened was in 2007 - a recession followed shortly.
As can be seen, the regional surveys have actually been leading the ISM numbers in the last recession. It is definitely possible that the same thing is happening again. Here is a comparison of new orders:
Combined regional Fed surveys vs. ISM, new orders. Once again, the regional surveys are far weaker.
Next up the employment figures - here the correlation is much closer, but the regional survey numbers are still weaker than the ISM data at present:
Combined regional Fed surveys vs. ISM, employment.
Finally, here is a chart comparing ISM new orders to total business sales (manufacturing sales). Business sales have only been updated to December, but even if the change rate in their decline should have slowed a bit in the meantime, there would still be a large gap between the two series - and in this case too, business sales have actually been leading ISM new orders in the last two recessions:
ISM new orders vs. business sales (y/y change).
All these data continue to suggest that a recession remains highly likely. What will it take for final confirmation? Among other things, an even stronger decline in ISM new orders (likely below the 48 level as Steve Saville has suggested), a downturn in real non-residential private investment and a rise in initial unemployment claims above the most recent interim peak (according to John Hussman, a rise above 320,000 would do the trick).
To this it must be kept in mind though that the fundamental backdrop of the current cycle is slightly different from that of previous cycles, as the Fed has directly boosted the money supply during much of the recovery period, with commercial bank credit expansion playing second fiddle. Moreover, a lot of the credit that has been extended to businesses by banks was used for financial engineering purposes (share buybacks, LBOs, and even dividend payments!) rather than capital investment.
However, quite a bit of malinvestment has taken place anyway, as evidenced by the boom-bust sequence in the oil patch and the absurd valuations attained by tech "unicorns" (we believe many of these will crash and burn and take the earnings of a great many tech companies down with them).
Effects on the Production Structure
Lastly, below is our own boom-bust indicator, which compares the industrial production indexes for capital goods and consumer goods. As can be seen here, during boom periods, investment and factors of production are drawn towards the capital goods industries, and the production of capital goods rises strongly vs. that of consumer goods. Since booms driven by credit created from thin air cannot be properly funded by the economy's pool of real savings, they prove unsustainable in the long term.
Capital vs. consumer goods production - a rough indicator of the changes experienced by the economy's capital structure during booms and busts.
The economy will over time tie up too many final goods (or the economy's "free capital" in the terminology of Richard von Strigl) relative to the amount it releases. Saving, investment and consumption schedules are no longer properly aligned, and the production structure becomes distorted: too many goods are produced for which demand is lower than expected (see oil), while not enough is invested in the production of goods that would have been demanded more urgently.
Generally, the structure of production will be lengthened beyond the point that can be sustained with existing real savings. This is the result of the distortion of relative prices caused by the expansion of credit and money ex nihilo, which falsifies economic calculation and leads businessmen astray.
As consumption hasn't been sufficiently lowered - as it normally would have been if sufficient savings had actually been accumulated - bottlenecks will eventually appear. If investment and consumption don't "add up," the over-consumption must still be funded somehow - and it is ultimately funded by the consumption of capital. Often this will greatly impair the middle stages of the production structure, in which capital maintenance will tend to be neglected while malinvestment in higher stages and over-consumption take place concurrently.
However, this cannot be kept up indefinitely, although the day of reckoning can certainly be greatly delayed by continuing the expansion of credit. Eventually though, consumer goods will become scarce relative to demand, and the price movements that have characterized the boom will begin to reverse, while upward pressure on previously suppressed market interest rates will begin to emerge (see junk bonds). The concrete sequence and the details of the phenomena attending the boom and bust periods differ from case to case, depending on contingent circumstances. However, what remains a feature of all types of historical business cycles is that in the end, the production structure has to be realigned to properly reflect reality.
Malinvested capital must be re-purposed and/or liquidated, which usually involves a shortening of production time by eliminating stages of production that cannot be sustained at the given level of savings, with factors of production re-directed towards the lower stages (consumer goods production), away from the higher stages (capital goods production). This rearrangement takes time and certainly feels uncomfortable - it's not as if one could simply flip a switch after all - and is known as the bust period, or recession.
While we can still not proclaim with certainty that a recession is imminent, recent data releases have so far done nothing to lower the relatively high probability of a recession emerging in the US. Depending on upcoming developments, this assessment could change again, but for the time being, we remain on red alert.
Keep in mind that the growth rate of the broad true money supply, while much lower than at the peaks of 2011-2012, remains at a historically high level (more than 8% in terms of TMS-2). This tends to support bubble activities, but there is a limit to what loose monetary policy can achieve with respect to misdirecting scarce resources (a misdirection that masquerades as "economic growth"). The problem is precisely that capital is scarce - no new capital can be conjured into being by the actions of a central bank.
Charts by Michael Pollaro, St. Louis Federal Reserve Research