COMMENTARY NUMBER 330
September Production and Housing Starts

October 19, 2010

__________

Third-Quarter Production Growth Slowed and Housing Contracted

Third-Quarter GDP Should Have Slowed
But the Heavily-Politicized and Guesstimated Series Virtually Is Worthless

__________

 

PLEASE NOTE: The next regular Commentary is scheduled for Friday, October 29th, following the release of the "advance" estimate of third-quarter GDP.  Analysis also will cover September New Orders for Durable Goods and Home Sales data.  Intervening Commentaries will be published as needed to address unusual market or other circumstances.

– Best wishes to all, John Williams

 

Third-Quarter GDP Growth At Least Should Have Slowed.  As with third-quarter real retail sales, annualized quarterly growth in third-quarter industrial production slowed versus annualized second-quarter growth.  Third-quarter housing starts contracted again on a quarterly basis.  The third-quarter trade deficit was a drain on broad economic activity, and inventory build-up, which accounted for the bulk of second-quarter GDP growth, should have seen some work-off, not further building in the third-quarter.

In combination, these factors suggest some slowing — if not an outright contraction — in third-quarter GDP growth, versus the second-quarter’s real (inflation-adjusted) 1.7% quarterly growth rate (see Week Ahead).  Yet, consensus expectations are for a stronger third-quarter GDP result, and the Bureau of Economic Analysis (BEA) tends to target consensus expectations for their "advance" releases.  Keep in mind that the "advance" third-quarter GDP release will be published less than a week before the election.  It seems increasingly likely that this most-heavily politicized of major economic releases once again also will be the most meaningless of major series in terms of reflecting underlying economic activity.

September Production Contraction Partially Reflected Upside Prior-Period Revisions.  In September reporting, monthly levels of seasonally-adjusted industrial production were notched higher in revision for April through August.  Before those revisions, the reported level of September production would have been on the downside of flat, month-to-month, still below consensus expectations of ongoing production growth.  The Federal Reserve Board reported Monday (October 18th) that seasonally-adjusted September 2010 industrial production fell by 0.22% (down 0.02% before revisions to August’s initial reporting, versus a revised 0.18% (previously 0.16%) gain in August.

For the third-quarter, seasonally-adjusted industrial production reportedly grew at annualized pace of 4.77%, down from a 6.99% pace in the second-quarter.

Year-to-year change in September production continued to ease back, to 5.42%, from a revised 6.43% (previously 6.21%) in August, and down from the near-term peak of 8.27% in June.  The relatively high volatility in annual growth is due largely to the pattern of collapsing activity a year ago (a May 2009 trough), followed by cash-for-clunkers boosted production.  The year-to-year contraction of 12.86% seen in May 2009 was the steepest annual decline in production growth since the shutdown of war-time production following World War II. 

 

The first graph preceding shows recent monthly detail, with a fair chance that the August number indeed marked the turning point of renewed decline in the production series.

The "recovery" in production is shown in the second graph, where month-to-month volatility is smoothed using a six-month moving average.  For the last 21 months, the production index has averaged 89.57, around which the series has been fluctuating, with September’s six-month moving average reading at 92.73, versus 93.19 for the single month.  Production activity had leveled off at a low-level plateau of activity that effectively wiped out the last eight years of growth in industrial production; the current rally has topped the pre-2001 recession high, reclaiming levels last seen in 2004 (six-month moving average). 

Despite the near-term upside gains generated by short-lived stimulus and seasonal-factor distortions, the series likely still is bottom-bouncing and should begin to soften anew, significantly, in the next several months.  The six-month moving-average series remains 7.62% below its December 2007 historic high.

September Housing Starts in Third-Quarter Contraction.  The Census Bureau reported a statistically-insignificant monthly gain for September 2010 housing starts, this morning, October 19th, locking in another quarterly contraction for the housing starts component of residential construction.

As reported, September housing starts were on the plus-side of flat, up for the month by a statistically-insignificant 0.3% (up by 2.0% before prior-period revisions) +/- 12.0% (95% confidence interval).  August’s starts were up an unrevised 10.5% for the month. 

For the third-quarter, seasonally-adjusted housing starts contracted at an annualized pace of 8.16%, versus a 9.38% annualized contraction in the second-quarter.

Year-to-year change in September starts was a statistically-insignificant gain of 4.1% +/-14.0% (95% confidence interval), following an upwardly revised annual gain of 3.9% (previously 2.2%) in August.

Activity has started to slow anew, although the pattern of housing starts generally remains one of bottom-bouncing at an historically low-level plateau of activity, as seen over the last 22 months.  Much weaker monthly numbers remain likely in the months ahead.  Since December 2008 housing starts have been bottom-bouncing, averaging a seasonally-adjusted annual rate of 574,400.  In those 22 months, all monthly readings have been within the normal range of monthly volatility for the series around that average, including September 2010’s reading of 608,000. 

The "recovery" in housing is shown in these graphs.  The first graph reflects the level of monthly activity in terms of the annualized monthly rate of housing starts, which is randomly volatile and subject to large revisions on a monthly basis.  The second graph shows the six-month moving average, which tends to smooth out that monthly volatility.  

 


 

Week Ahead.  Given the unfolding reality of an intensifying double-dip recession 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, previously unreported economic weakness will show up in prior-period revisions.

New and Existing Homes Sales (September 2010).  September existing home sales (National Association of Realtors) are due for release on Monday, October 25th, followed by new home sales (Census Bureau) on Wednesday, October 27th.  As with housing starts, theses volatile series should remain extremely weak, with any monthly change (new homes) again lacking statistical significance.  Briefing.com is showing consensus estimates of gains in both series.  Yet, with extreme volatility, reporting error and revisions in these series, with poor-quality seasonal adjustments, with extremely negative anecdotal evidence, and with massive distortions from what should be volatile foreclosure activity in an increasingly difficult legal environment, monthly movements in these numbers are particularly meaningless.

New Orders for Durable Goods (September 2010).  Due for release on Wednesday, October 27th, September new orders for durable goods are expected to show a 0.8% monthly gain (Briefing.com) versus a 1.3% monthly decline reported in August.  This irregularly volatile series likely will not show meaningful monthly change, despite upside biases built into these numbers in the last benchmark revision.

Gross Domestic Product — GDP ("Advance Estimate" Third-Quarter 2010).  Due for release on Friday, October 29th, the "advance" estimate of third-quarter 2010 GDP is expected to reflect 2.4% annualized inflation-adjusted quarterly growth, up from 1.7% in the first-quarter.  As discussed in the opening comments, key underlying series are more consistent with slowing or contracting growth.  The Bureau of Economic Analysis will tend to target consensus expectations as they stand on Friday, October 22nd.  Reporting risk remains for something of a downside surprise to those expectations.

 

__________