COMMENTARY NUMBER 529
Retail Sales Benchmark Revision

May 31, 2013

__________

Annual Retail Sales Revised Lower by 0.43% in 2011 and 0.22% in 2012


__________

 

PLEASE NOTE: The next regular Commentary is scheduled for Tuesday, June 4th, covering the April trade deficit and trade benchmark revisions, and April construction spending.  A subsequent Commentary on June 7th will cover the May employment and unemployment data.

Best wishes to all — John Williams

OPENING COMMENTS AND RETAIL SALES REVISION

 

Retail Sales Are Not Quite What They Used to Be.  As discussed in the May 29th No. 527: Special Commentary, the broad U.S. economy remains in serious trouble.  Reflecting that, the regular stream of annual benchmark revisions to major economic series generally has shown that the post-2009 “recovery” was weaker than previously believed.  A common thread between today’s (May 31st) benchmark revision of retail sales, and the recent benchmark revisions to industrial production and new orders for durable goods orders, has become obvious.  Where better-quality data were available (usually 2010 and 2011), such as with annual surveys or censuses, the pre-existing reporting was revised lower.  Where the revisions included the ongoing, overly-optimistic underlying assumptions that had generated the initial overstatement of activity, the overstatement continues, boosting the revised current levels of activity back up into range of recent reporting.

The annual benchmark revision to the retail sales series included new samples, revamped seasonal factors and data from the 2011 Annual Retail Trade Survey.  Largely due to the recasting of the concurrent seasonal-adjustment factors on a consistent basis (for one month), the revisions went back to the 1992, the beginning of the current series.  In terms of annual average growth, sales were revised lower by 0.43% in 2011 and by 0.22% in 2012—reasonably small amounts—with some smoothing of previous month-to-month sales volatility.  At worst, the year-to-year growth was revised lower by 0.82% in June 2011.

Despite the revisions, the general outlook for the actual U.S. economy remains the same.  In conjunction with benchmarks to other series, the ongoing pattern suggests downside revisions to post-2009 activity in real (inflation-adjusted) GDP activity, come the comprehensive GDP revision of July 31st.  

Revised plots of the seasonally-adjust retail sales series are shown the following seven graphs.  Except for the first graph, which shows broad, post-World War II real (inflation-adjusted) retail sales activity—updated for the latest revisions—all the other plots show both the revised and prior reporting.  The background thicker red line in those graphs reflects the previous estimates.  Several of the graphs just show closer detail of the graph before.  All the graphs are adjusted for inflation, with Graphs 1 to 5 reflecting sales deflated by the CPI-U, as published regularly by the St. Louis Federal Reserve.  Graphs 6 and 7 are corrected for the understatement of the CPI-U inflation using, instead, the ShadowStats 1990 Alternate measure. 

Graph 1: Total Real Retail Sales (Two Series), 1947 to Date, Shows Only the Current Revised Data,
In Billions of 1982-1984 CPI-U Dollars.

Graph 2: Total Real Retail Sales, 2000 to Date, Shows Revised (Blue) and Prior Reporting (Red),
Series Is Indexed to January 2000 = 100, Scale Is Consistent with Graph 6.

 

Graph 3: Total Real Retail Sales, 2010 to Date, Shows Revised (Blue) and Prior Reporting (Red),
In Greater Detail, Series Is Indexed to January 2000 = 100.

Graph 4: Year-to-Year Percent Change in Total Real Retail Sales, 2006 to Date,
Shows Revised (Blue) and Prior Reporting (Red).

Graph 5: Year-to-Year Percent Change in Total Real Retail Sales, 2010 to Date,
In Greater Detail, Shows Revised (Blue) and Prior Reporting (Red).

Corrected Retail Sales.  Graph 2 reflects revised real retail sales as should be reported by the St. Louis Fed, deflated by the CPI-U (ShadowStats did the CPI-U adjustments for these graphs).  The shown series is indexed to January 2000 = 100.  The CPI-U, however, understates inflation (see the Public Comment on Inflation), with the effect of overstating inflation-adjusted growth.

Instead of being deflated by the CPI-U, the “corrected” and revised real retail numbers, reflected in Graphs 6 and 7, use the ShadowStats-Alternate Inflation Measure (1990-Base) for deflation.  As discussed in No. 527: Special Commentary, Hyperinflation 2012 and No. 485: Special Commentary, with the higher inflation of the ShadowStats measure, the revamped numbers show a pattern of plunge and stagnation.  The revised, recent topping-out process still has reverted to renewed decline, as of second-quarter 2012, in a series that had been bottom-bouncing along a low-level plateau of economic activity, following the unofficial economic collapse from 2006 into 2009.  Graphs 2 and 6 are indexed to a consistent scale.  Graph 7 encompasses a shorter timeframe, allowing for greater visual detail of the corrected-series revisions.

 

 

Graph 6: Corrected Total Real Retail Sales, 2000 to Date, Adjusted for Understatement of Inflation in CPI-U,
Shows Revised (Blue) and Prior Reporting (Red), Series Is Indexed to January 2000 = 100, Scale Is Consistent with Graph 2.

Graph 7: Corrected Total Real Retail Sales, 2010 to Date, in Greater Detail, Shows Revised (Blue) and Prior Reporting (Red), Series Is Indexed to January 2000 = 100.

__________


 

HYPERINFLATION WATCH

 

Hyperinflation Outlook.  The current hyperinflation outlook was revised and updated with new detail, May 29th, in No. 527: Special Commentary.  Reflecting the content of that Special Report, the regular synopsis of the general outlook will be revamped with the first regular Commentary in the week of June 3rd (on June 4th).  For the current broad outlook, please read or link back to No. 527, or to Commentary No. 525 for the prior synopsis.

 

 

__________

 

 

WEEK AHEAD

 

Weaker Economic and Inflation Data Are Likely for Data Published in June.  As seen in the last two months of consumer inflation reporting, May 2013 consumer inflation also should be muted by seasonally-adjustment constraints on oil and gasoline prices.  That said, the highly irregular, unadjusted oil and gasoline price movements turned somewhat higher in May.  Distortions from increasingly irrelevant, shifting and severely-negative gasoline and oil price seasonal adjustments should flip to positive-side distortions with June and July’s adjusted CPI reporting, and to neutral in May and then to positive in June for the PPI.  Going forward, reflecting the still-likely negative impact on the U.S. dollar in the currency markets from continuing QE3 and the still-festering fiscal crisis/debt-ceiling debacle, reporting in the ensuing months and year generally should reflect much higher-than-expected inflation (see No. 527: Special Commentary).

Where expectations for economic data in the months and year ahead should tend to soften, weaker-than-expected economic results still remain likely, given intensifying structural liquidity constraints on the consumer.  Increasingly, previous estimates of economic activity should revise lower, particularly in upcoming annual benchmark revisions, as has been seen already in industrial production, new orders for durable goods and retail sales, and as pending for the trade deficit (June 4th), construction spending (July 1st) and the GDP (July 31st—comprehensive overhaul and redefinition back to 1929).

Reporting Quality Issues and Systemic Reporting Biases.  Significant reporting-quality problems remain with most major economic series.  Headline reporting issues are tied largely to systemic distortions of seasonal adjustments.  The data instabilities were induced by the still-ongoing economic turmoil of the last six-to-seven years, which has been without precedent in the post-World War II era of modern economic reporting.  These impaired reporting methodologies provide particularly unstable headline economic results, where concurrent seasonal adjustments are used (as with retail sales, durable goods orders, employment and unemployment data), and they have thrown into question the statistical-significance of the headline month-to-month reporting for many popular economic series.

With an increasing trend towards downside surprises in near-term economic reporting, recognition of an intensifying double-dip recession should continue to gain.  Nascent concerns of a mounting inflation threat, though muted, increasingly have been rekindled by the Fed’s monetary policies.  Again, though, significant inflation shocks are looming in response to the fiscal crisis and a likely, severely-negative response in the global currency markets against the U.S. dollar.

The political system and Wall Street would like to see the issues disappear, and the popular media do their best to avoid publicizing unhappy economic news, putting out happy analyses on otherwise negative numbers.  Pushing the politicians and media, the financial markets and their related spinmeisters do their best to hype anything that can be given a positive spin, to avoid recognition of serious problems for as long as possible.  Those imbedded, structural problems, though, have horrendous implications for the markets and for systemic stability, as discussed in Hyperinflation 2012, No. 485: Special Commentary and No. 527: Special Commentary.

Construction Spending (April 2013).  Due for release on Monday, June 3rd, by the Commerce Department, April construction spending should continue its recent trend of month-to-month stagnation, particularly after adjustment for inflation.  As usually is the case, reported month-to-month seasonally-adjusted changes are likely to be statistically-insignificant.

U.S. Trade Balance (April 2013, Annual Benchmark Revisions).  The April trade deficit is scheduled for release on Tuesday, June 4th, in conjunction with the publication of annual benchmark revisions of the series.  The benchmark data likely will show larger historical trade deficits than have been reported previously.  In turn, those numbers should be reflected in downside revisions to historical GDP growth, which will be revised back to 1929 in the pending July 31st “comprehensive” revision (see Commentary No. 528).

With the U.S. trade deficit continuing in fundamental deterioration, the April reporting is at serious risk of showing a meaningful deterioration—worse than the developing consensus of minor monthly widening—versus the existing March deficit estimate.  Any significant deficit-widening or deficit-narrowing revision to the March data would have relative negative or positive impact on the third estimate of first-quarter 2013 GDP, due for release on June 26th.

Employment and Unemployment (May 2013).  The May labor data are due for release on Friday, June 7th, from the Bureau of Labor Statistics (BLS).  Most commonly, the consensus jobs estimate settles around the trend estimate from the BLS seasonal-adjustment models.  The May 2013 payroll trend number is for a 204,000 jobs gain, versus April reporting of 165,000 (see Commentary No. 521).  The developing early consensus appears to be about twenty-percent below the trend, at the moment.  Separately, the markets appear to be expecting the May unemployment rate to hold at the 7.5% headline U.3 level reported in April.

Reflecting underlying fundamental economic activity that is much weaker than consensus expectations, reporting risks continue to the downside of expectations for payrolls and to the upside for the unemployment rate.

Although the unemployment rate should move higher, there is a persistent reporting problem that has been discussed frequently with this series (see Commentary No. 451 and Commentary No. 487, for example).  Month-to-month comparisons of the headline unemployment data cannot be made legitimately.  The headline change in the unemployment rate is of no meaning, other than in misguided-media and market reactions.  Specifically, all the recent historical unemployment rates are re-calculated each month as part of the concurrent-seasonal-adjustment process, but where the BLS publishes the new headline unemployment rate, it does not publish, and it does not make available, the revised number from the month before, which would be consistent with the new number.

 

 

__________