On The Ever Increasing Inconsistencies In Reported Economic Data
Ever get the feeling that the Bureau of Truth is not being completely truthful? Feel like the ADP is to the NFP like the ISM to the regional Fed Surveys, and as the surging Mfg ISM employment diffusion index is to the plunging Service ISM employment diffusion index (i.e., both can not possibly be correct)? You are not alone. David Rosenberg summarizes which recent data releases are so blatantly incomprehensible, one wonder when the government will announce an AXA Rosenberg-like computer glitch and say all its data for the past 12 months has been compromised. Either that, or we await the introduction of the Birth/Death adjustment to every single data series released in America imminently.
Here is David with much more on the topic (from Gluskin Sheff):
The latest batch of data has been highly confusing, to say the least. The chain store sales data were skewed by one-offs, such as retroactive jobless benefit checks that were mailed out in early August and the growing number (17 this year) of States offering sales tax holidays. We estimate that absent these influences, year-on-year sales growth would have been closer to 1% than 3%.
The spending data also belied the information contained in the Conference Board’s consumer confidence survey, as the facts-on-the ground ‘present situation’ index sagged to 24.9 in August from 26.4 in July — only 5% of the time in the past has it been so low. The ISM manufacturing index, which really got the ball rolling on this ‘take out the double-dip’ trade, managed to spike even though the three leading sub-indices — new orders, backlogs and vendor performance — all declined in what was a 1-in-100 event.
Not only that, but the employment component of the ISM surged to its highest level since December 1983, and yet the manufacturing employment segment of the payroll survey fell 27,000 — the first decline this year and the sharpest falloff since last October. Furthermore, the manufacturing diffusion index slumped to a seven-month low of 47 from 53 — in other words, fewer than half of the industrial sector was adding to staff requirements last month. It begs the question as to what exactly the ISM is measuring.
The list of inconsistencies in the data didn’t stop there. The entire increase in private sector employment in August was in the service sector — mostly health and education, which says little about the cyclical state of the economy. Yet 90 minutes after the jobs number was released, we got the ISM non-manufacturing survey and it flashed a contraction in services employment to a seven-month low of 48.2 from 50.9 in July.
Just a tad confusing, but the newly found bullish view of the economy is sort of corroborating evidence.
The employment report did not detract from the view that the economy is losing steam. The fourth quarter of a recovery typically sees real GDP growth of over 6% at an annual rate, but in this post-bubble credit collapse, what we got this time was 1.6% at an annual rate in Q2.
Moreover, there is nothing in the data to suggest anything but a further slowing in Q3, and the only reason why there is no contraction this quarter is because it looks as though we are getting another lift from inventories — though now the buildup looks involuntary, which will cast a cloud on fourth-quarter GDP barring a sudden reversal in the declining trend in real final sales.
Private payrolls were +247,000 when the equity market peaked in April, it slowed to +107,000 by July and was +67,000 last month. What does that suggest about the trend? Ditto for goods-producing employment, which was +67,000 in April, subsequently softened to +37,000 by July, and in August was the grand total of zero.
One can easily draw the conclusion from the data that we have dodged a bullet. But that does not mean we are out of the woods. Employment is a coincident indicator. Leading indicators, such as the ECRI, continue to deteriorate and to levels still consistent with nontrivial double-dip risks. Keep this in mind — private payrolls came in at +97,000 in November 2007 and the “Great Recession” began the next month. In other words, the +67,000 tally we saw today basically tells you nothing about how the pace of economic activity is going to unfold as we move into the fall.