JOHN WILLIAMS’ SHADOW GOVERNMENT STATISTICS

COMMENTARY NUMBER 312

Durable Goods, Home Sales, Upcoming GDP Revisions

Restated Downturn Should Be More Severe

Nonsense Home-Sales Reporting

Roughly 144,000 Census Jobs Lost in July

PLEASE NOTE: The next regular Commentary is scheduled for Friday, July 30th, following the release of the "advance" estimate of second-quarter 2010 GDP and the annual GDP revisions, which will restate economic history back to the first-quarter 2007.

U.S. Economy Slows Anew. With a weakening labor market, softer consumer confidence, softening housing activity and retail sales, and an intensifying trade deficit, the early stages of a renewed economic decline — what likely will be popularized as a "double-dip" recession — have begun to unfold. Accordingly, most business reporting in the months ahead will tend to show patterns of accelerating contraction. A detailed economic review will be included the Commentary on Friday (July 30th), following the "advance" estimate of second-quarter GDP and annual GDP revisions.

GDP Revisions Should Show More Severe Contraction. Based on a review of benchmark revisions that have been published this year for various series, including payrolls, retail sales, trade balance and industrial production, the pending annual GDP revisions should tend to show a deeper and possibly longer recession than previously reported, with the contraction most noticeably more-severe in 2008 and early-2009. There appears to be an effort to show more of a recovery, beginning in the second-half of 2009, but any relative gains that appear there likely will disappear in the next round of annual revisions in 2011.  

New Orders for Durable Goods Stalled Again. The Census Bureau reported this morning (July 28th) that the regularly volatile, seasonally-adjusted new orders for durable goods fell by 1.0% (down by 0.8% net of revisions) in June, after a revised 0.8% (previously 1.1%) decline in May. Unadjusted, year-to-year change in June 2010 new orders was a gain of 15.9%, following a revised 15.9% (previously 15.3%) increase in May. The recent benchmark revision appears to have built in something of an upside sampling bias to the monthly numbers. Nonetheless, the two most recent reports have shown monthly pullbacks in new orders. In this extremely variable series, however, next month’s reporting likely will see a heavy boost from long-term, multi-year airplane orders. 

The widely followed nondefense capital goods declined by 1.6%, seasonally-adjusted, for the month in June, after a revised 0.6% (previously 2.8%) decline in May. For June, the unadjusted year-to-year change in the series was up by 18.8%, versus a revised May annual gain of 21.7% (previously reported at 18.9%).

Meaningless Home Sales Numbers Serve as Fodder for Irrational Stock Market. The monthly new home sales data simply cannot be taken at face value.  The numbers are of such poor quality and of such minimal statistical significance that one has to wonder why the Census Bureau even bothers to report them. Census might consider holding back the data six months, or so, until the regular and volatile revisions stabilize. The only value for the June new home sales numbers was in the realm of Wall Street hypesters trying to support irrationally high stock prices for a couple of hours.

Monday’s (July 26th) release of May new home sales (counted based on contract signings, Census Bureau) showed sales soaring by 23.6% (up just 10.0% net of prior-period revisions) +/- 17.9% (95% confidence interval). Such followed a revised 36.7% collapse (a 40.1% decline before prior-period revisions, and, as initially reported, a 32.7% drop). 

Aside from rarely having any statistical significance at the 95% confidence level, the monthly new homes sales numbers tend to suffer extreme revisions. Consider that April 2010’s sales were revised downward from initial reporting of 504,000, to 446,000 with May’s report, and to 422,000 with the June report. May’s sales initially were reported at 300,000 (a record low) and were revised to 267,000 (a new record low) with the June report. June’s monthly "surge" was off the downwardly-revised May data. The June level of 330,000 annualized sales is the lowest on record, except for the current reporting on May.

The July 22nd report of June existing home sales (counted based on actual closings, National Association of Realtors [NAR]) was more stable than the new home sales reporting, but not too much more meaningful. June sales fell by a seasonally-adjusted 5.1%, versus an unrevised 2.2% monthly contraction in May.

Foreclosure activity remained a major distortion in these numbers, with NAR estimating 32% of new home sales for June in the "distressed" category (the May estimate was 31%, but the number of foreclosures still has been fairly steady). That percentage should jump sharply in the next couple of months, as overall sales activity starts to slow markedly. The Census Bureau acknowledges that a portion of new home sales is from foreclosure activity but offers no estimates. Purportedly, foreclosure activity is on the rise, and some in the construction trade have difficulty competing with the pricing of foreclosed properties. Until the foreclosure problem works itself out, monthly changes in these home sales numbers cannot be taken as meaningful indicators of trends in underlying activity in homeowner real estate, as it relates to general economic activity. 

Both the new and existing home sales series have seen recent volume vary, tied to the April 30th expiry of the home-buyer tax credit, where a sales contract had to be in place by April 30th.  Beyond the unstable reporting of new home sales and heavy distortions from foreclosure activity, and consistent with June housing starts (see Commentary No. 311), the pattern of activity in residential real estate appears to have moved beyond the impact of that short-lived stimulus, with renewed or re-intensifying economic contraction beginning to unfold.

The following updated graphs reflect different measures of home sales activity since February 2009. The numbers, through June 2010, reflect the seasonally-adjusted level of monthly sales, rather than the annual rate usually published. 

Week Ahead. Given the unfolding reality of a weaker economy (or re-intensifying downturn) and more serious inflation problems than generally are expected by the financial markets, risks to reporting will tend towards higher-than-expected inflation and weaker-than-expected economic reporting in the months ahead. Increasingly, that will be seen in economic reporting net of prior-period revisions.

Gross Domestic Product — GDP (Second-Quarter 2010 "Advance Estimate" and Annual Revisions). What had been an early consensus for some relative pick-up in annualized quarterly real (inflation-adjusted) growth in the largely guesstimated "advance" second-quarter 2010 GDP — due for release on Friday, July 30th — has softened to some slight relative slowing growth. Briefing.com is showing a consensus estimate of 2.5% for the second-quarter, down from the 2.7% as currently reported for the first-quarter. As discussed in Commentary No. 310, a slower growth rate would be consistent with recent underlying economic reporting. Even though consensus estimates often are targeted for setting the "advance" GDP growth rate, risks are fairly high that the reported growth will surprise the consensus on the downside.

As discussed in early comments, a pattern of generally weaker-than-previously-reported economic growth is a good bet for the annual revisions back through first-quarter 2007. The 2007 recession should appear to be more severe, post revisions.

Payroll Employment and Unemployment Rate (July 2010). Due for release on Friday, August 6th, the July labor numbers are a good bet to be more negative than an already soft consensus. Briefing.com reports expectations of a 116,000 decline in monthly payrolls. That number includes layoffs of temporary and intermittent census workers in July, which will be roughly 144,000 (per Census reporting), implying a minimal monthly gain expected for nonfarm payrolls of about 28,000, ex-census workers. In June, payrolls declined by 125,000, gaining 100,000 ex-census. Briefing.com also reports a consensus July U.3 unemployment rate at 9.6%, up from June’s 9.5%. 

I expect nonfarm payrolls to show an outright monthly contraction, ex-census workers, with the unemployment rate jumping more than expected. Further detail will follow in the July 30th Commentary. 

__________