September Housing Starts

October 17, 2012



September Housing Starts Surge Was Marginally Significant
But Not Credible



PLEASE NOTE: The next regular Commentary is scheduled for Friday, October 26th, covering third-quarter GDP, September new orders for durable goods and new and existing home sales.

A Special Commentary reviewing economic, inflation and systemic conditions is planned for Monday, October 29th.  It will include significant September-economic and third-quarter-GDP reporting, as well as detail on fiscal-2012 operations of the federal government.  Any updates to timing of this missive will be advised in the schedule box on

Best wishes to all — John Williams


Opening Comments and Executive Summary.  This morning’s (October 17th) statistically-significant 15.0% upside bounce in the September 2012 housing starts reflected the highest level of reported monthly activity since July 2008, when the series was collapsing.  As reported, though, September’s level of housing starts remained 62% below the January 2006 series high. 

The monthly surge is not credible in this extremely volatile series, and it likely will be muted by revisions or downside reporting in the months ahead.  The underlying fundamentals to support a sustainable upturn in housing starts simply are not in place.  In particular, consumer liquidity remains severely impaired by negative real (inflation-adjusted) income growth, as well as by restricted ability to expand borrowing (see discussions, for example, in yesterday’s Commentary No. 476 and the earlier Commentary No. 469). 

The best estimate I can come up with for a seasonally-adjusted monthly increase in September 2012 residential-mortgage lending by commercial banks is 0.4%.  That is in line with combined expectations for changes in existing- and new-home sales for the month (there would be some mortgage replacement issues with existing home sales), but it certainly is well shy of something that would sustain the September 2012 housing-starts growth rate.

The housing starts series is particularly volatile and tends to suffer large monthly revisions.  Extreme monthly variations can be generated by exogenous factors such as requirements on builders to maintain lines of credit, unusual weather patterns, regulatory shifts in major markets, as well as poor-quality seasonal adjustments.  Unusual factors in those areas for September, however, were not immediately obvious.

The monthly gain in September housing starts of 15.0% was in the context of a government-estimated 95% confidence interval of +/- 14.2%.  That means that once in twenty months, the error in the month-to-month reporting can be could be expected to exceed +/- 14.2%.  The aggregate monthly 15.0% change in the series was dominated by a 25.0% surge in apartment starts, although single-unit housing still had a decent 11.0% monthly gain.  Yet, the monthly increases in apartment starts and single-unit starts were not statistically significant.

As reported, though, third-quarter 2012 housing starts were up at an annualized pace of 30.4% versus second-quarter 2012, due primarily to the September report.  In turn, second-quarter housing starts were up by 12.3% from first-quarter 2012.  Accordingly, the residential-investment category, in the pending October 26th “advance” estimate of third-quarter 2012 GDP, should pick up some growth from the latest housing-starts estimates.

As shown in the two graphs that follow, current monthly housing starts activity remains off the record monthly low seen for the present series in April 2009.  Recent series detail is seen more easily in the first graph, which covers reporting since 2000. 

The second graph shows that current activity—in a broader historical context—has just notched above the levels of activity seen at the troughs of the most-severe post-World War II declines in housing starts, as shown in both the current housing starts series and the predecessor nonfarm housing starts series, which was introduced in 1946.

[More complete information on the housing starts report is found in the Reporting Detail section.]




Hyperinflation Watch—Outlook Unchanged, Broad Outlook to Be Revised in Upcoming Special Commentary.  The following summary of the broad outlook has not changed since yesterday’s October 16th Commentary, but it is included here for those who may not be familiar with it, including new subscribers.  The nature and implications of QE3—announced recently by the FOMC of the Federal Reserve Board—were discussed in the Opening Comments of Commentary No. 470.  Specifically, while general circumstances have continued to advance towards the ultimate demise of the dollar, the general outlook is unchanged.  While QE3 is an enabling action for the onset of massive inflation, the outside timing of 2014 for the hyperinflation forecast remains in place.  The hyperinflation outlook will be reviewed and updated fully in the Special Commentary scheduled for October 29th. 

The detail in Special Commentary No. 445 (June 12th) updated the hyperinflation outlook and the outlook for U.S. economic, U.S. dollar, and systemic-solvency conditions.  That Special Commentary supplemented Hyperinflation 2012 (January 25th), which remains the primary missive detailing the hyperinflation story.  Those reports are suggested as background reading for new subscribers.

Official GDP reporting shows plunging economic activity from fourth-quarter 2007 to second-quarter 2009, with an ensuing upturn in activity that has led to a full recovery as of fourth-quarter 2011, and that “recovery” has continued through second-quarter 2012 GDP reporting. 

In contrast to the GDP reporting—and in line with patterns seen in better-quality economic series—I contend that the economy began turning down in 2006, plunging in 2008 into 2009 and subsequently stagnating—bottom-bouncing—at a low level of activity, ever since.  There has been no recovery since mid-2009, and the economic downturn now is intensifying once again.  The renewed slowdown is evident in the current reporting of nearly all major economic series.  Not one of those series shows a pattern of activity that confirms the full recovery shown in the GDP series.

Federal Reserve Chairman Ben Bernanke has observed that broad aggregate measures of the U.S. economy, such as GDP, do not appear to be reflecting the common experience of the general public.  Indeed, common experience suggests that the economy has not recovered.  The official recovery simply is a statistical illusion created by the government’s use of understated inflation in deflating the GDP, which overstates deflated economic growth, as discussed in Commentary No. 467, Special Commentary No. 445, and Public Comment on Inflation.

The long-term fiscal solvency issues of the United States—where GAAP-based accounting shows annual deficits running in the $5 trillion range—are not being addressed, and the politicians currently running the government lack the political will to address those issues.  That circumstance initially suggested a hyperinflation crisis by the end of this decade, but federal government and Federal Reserve actions—in response to the systemic-solvency crisis of 2008—accelerated the process, suggesting a hyperinflation problem by no later than the end of 2014.  The continuing economic downturn is intensifying the fiscal- and systemic-solvency problems, and public awareness of this should grow rapidly in the months ahead.

Neither economic nor systemic-solvency issues have been resolved by U.S. government or Federal Reserve actions, and the most recent readings on income variance suggest that the worst is yet to be seen, as discussed in Commentary No. 469

With the economy weak enough to provide political cover for further Federal Reserve accommodation to the still-struggling banking system, QE3 was introduced on September 13th.  That action effectively provided for open-ended monetization of U.S. Treasury debt at the Fed’s discretion.  The mechanism for eventual full debasement of the dollar now is in place, and it likely will come into full play, as needed to support the banking system and as needed to assure “successful” auctions of Treasury debt.

QE3 likely will lead to a massive dollar-selling crisis, and that will begin the process of a rapid upturn in domestic consumer inflation.  A near-term dollar-selling crisis is now of a much greater risk, post-QE3.  Separately, though, a dollar-selling crisis could begin at any time, triggered by various economic, sovereign-solvency or political issues.  With the guidelines set for QE3, even negative employment reports could trigger massive dollar selling.









September Housing Starts Surge Was Unusually Strong, But Not Highly Significant.  The strong upside, statistically-significant bounce seen with the September 2012 housing starts, showed the highest monthly level of housing starts since July 2008, when the series was in free-all.  As reported, the September number still remained 62% below the January 2006 series high.  This series remains particularly volatile and tends to suffer large monthly revisions, as well as irregular monthly distortions from a host of other factors, as discussed in the Opening Comments and Executive Summary.

Protracted stagnation in housing starts at historically-low levels of activity has continued well into its fourth year of activity, still averaging 73% below 2006’s record-high construction level.  Still within the normal scope of volatility for the series in the last four years, a slightly higher plateau of activity appears to have developed in the eleven months through September that is 68% below the 2006 high.  Given the underlying economic fundamentals, there is no recovery or relief in sight (see Special Report No. 445), and today’s strong report likely will revise sharply lower or be offset by monthly declines in the months ahead

September 2012 Reporting. The Census Bureau reported today (October 17th) a statistically-significant—albeit marginally-significant—month-to-month headline jump in seasonally-adjusted September 2012 housing starts of 15.0% (a gain of 16.3% before prior-period revisions) +/-14.2% (all confidence intervals are at the 95% level).  The monthly change in August starts was revised to a 4.1% gain from initial reporting of a 2.3% monthly increase. 

The aggregate housing starts number was the highest monthly level reported since July 2008, when the series was in free-fall.  September’s monthly gain was dominated by apartment starts, although single-unit housing also had a decent gain for the month.  The monthly increases in the sub-series, however, were not statistically significant.

Activity in starts for apartment buildings remained extremely volatile, yet it indeed was no more than statistical noise, month-to-month, jumping by a statistically-insignificant 25.0% +/- 45.5% in September, after a revised 1.4% (previously 2.8%) decline in August.  The “one-unit” category for September showed a statistically-insignificant 11.0% gain +/- 13.0%, after a revised 7.3% (previously 5.5%) increase in August.

The year-to-year change in aggregate September 2012 housing starts was a statistically-significant increase of 34.8% +/- 21.3%, following a revised 30.5% (previously 29.1%) annual gain in August, but, again, that remains in the continued context of what has been a protracted period of low-level bottom-bouncing.

For the last 46 months, the pattern of housing starts generally has remained one of stagnation at an historically low-level plateau of activity.  Since December 2008, housing starts have averaged a seasonally-adjusted monthly rate of 51,300.  In that period, monthly readings have been within the normal range of monthly variability for the aggregate series, around that average, except for the September 2012 reading of 72,700, which is beyond the upper end of that range.  Nonetheless, activity appears to have moved to a somewhat higher plateau since November 2011, with the average of the monthly housing starts at 61,500 for the eleven months through September 2012.

The official reporting of monthly housing starts is expressed at an annualized monthly pace, which was 872,000 for September 2012.  Particularly in the event of extreme monthly volatility, as seen in September, it is my preference to use the actual, non-annualized monthly number.  The graphed patterns are the exactly same, it is just that the monthly levels tend to be a little more realistic.

The regular graphs for this series are shown in the Opening Comments and Executive Summary section.  Current monthly housing starts activity remains off the record monthly low seen for the present series in April 2009.  September 2012 activity has just notched above the levels of activity seen at the troughs of the most severe post-World War II declines in housing starts shown in the current series, or in the predecessor nonfarm housing starts series.


Week Ahead.  As noted in Commentary No. 474, public speculation as to political manipulation of labor-related data, produced by the Department of Labor, has hit a level of activity that is unprecedented. Whether or not manipulations have taken place with current reporting, supporting evidence either way likely will be forthcoming in the year ahead.  Upcoming, initial reporting of third-quarter GDP on October 26th and the October labor report on November 2nd—in the two weeks before the election—would be likely manipulation targets, if the data indeed were being massaged.  In any event, numbers that are far removed from common experience will tend to be viewed by the public with extreme skepticism.

Nonetheless, market recognition of an intensifying double-dip recession has taken stronger hold, while recognition of a mounting inflation threat has been rekindled a bit by recent Fed monetary policy announcements and rising headline inflation numbers.  The incumbent political system would like to see the issues disappear until after the election; the media does its best to avoid publicizing unhappy economic news or, otherwise, it puts a happy spin on the numbers; and the financial markets do their best to avoid recognition of the problems for as long as possible, problems that have horrendous implications for the markets and for systemic stability, as discussed in the Hyperinflation Watch section.

Until such time as financial-market expectations catch up fully with underlying reality, or underlying reality catches up with the markets—shy of manipulated data—reporting generally will continue to show higher-than-expected inflation and weaker-than-expected economic results in the months and year ahead.  Increasingly, previously unreported economic weakness should continue to show up in prior-period revisions.    

New and Existing Home Sales (September 2012).  The September existing home sales release is scheduled for Friday, October 19th, from the National Association of Realtors, while the new home sales report from the Census Bureau is scheduled for Wednesday, October 24th.  Entrenched patterns of stagnation likely have continued in both series, and the pending monthly results are not likely to be statistically-significant.

New Orders for Durables Goods (September 2012).  The initial estimate for September 2012 new orders for durable goods is scheduled for release on Thursday, October 25th, by the Commerce Department.  As has been the case for months—except for August’s sharp decline in areas other than commercial aircraft orders—whatever month-to-month change is reported for this regularly volatile series, it likely will not be outside the bounds of normal variability.

In terms of the potential inflation contribution to new orders growth, the seasonally-adjusted, month-to-month change in the September 2012 PPI finished goods capital equipment index was a decrease of 0.1%, with year-to-year adjusted inflation at 1.7%.  Due to hedonic-quality adjustments to this portion of the PPI series, however, as with the GDP numbers, those inflation data understate inflation reality and correspondingly overstate inflation-adjusted growth.

Gross Domestic Product—GDP (Third-Quarter 2012, First or “Advance” Estimate).  The first-estimate of third-quarter 2012 GDP is due for release by the Bureau of Economic Analysis (BEA) on Friday, October 26th.  A consensus appears to be developing for stronger growth than the 1.3% annualized, real (inflation-adjusted) headline growth last reported for second-quarter 2012.

Reality is in the direction of much weaker growth, with mixed signals from industrial production (negative), real retail sales (positive), the trade deficit (negative) and housing starts (positive) pressuring quarterly change to below consensus.  With the BEA usually trying to target its “advance” growth estimate at the consensus, however, and with some potential for other massaging of numbers the week before the election, any stronger growth that is reported likely will revise to the downside in later months.

In any event, the new numbers most certainly will not be statistically significant, with the GDP series remaining the most worthless, the most heavily politicized and the most misleading of any major government economic statistical release.