Massive GDP Revision Implications

by Rick Davis

Until several quarters ago our website featured a headline graph comparing our Daily Growth Index with the Bureau of Economic Analysis‘ (BEA) quarterly GDP growth. As recently as a week ago that chart looked something like this:

[click to enlarge]

The left side of the above chart shows a strong correlation between the two sets of data, with the behavior of our Daily Growth Index leading the subsequent movements in the annualized GDP growth — thereby providing a good leading indicator for the GDP. However, when GDP data for third quarter of 2010 was published in late October 2010 that relationship began to unravel, and until we better understood what was going on we moved the chart to our “History” page (where it still resides as the third chart down).

We have been asked any number of times what happened to cause our data and the BEA’s GDP growth rates to diverge in such a striking manner, starting late in 2010. In our FAQs section we have written quite a bit about the many possible sources for the divergence. In a nutshell (and in order of likely contribution) the factors probably are:

— Federal fiscal stimuli supporting (or levitating) the non-consumer components of the GDP.

— Federal Reserve monetary stimuli promoting the further “financialization” of the economy, again missing consumers and most of “Main Street” America.

(Note that both of the above factors constitute variations of “kicking the can down the road,” since neither is economically “organic” or sustainable. Without support from (or for) consumers any fiscal stimulation is likely to end badly, when the expiration (or ultimate repayment) of stimulation monies sucks any of the transitory levitation out of the system.)

— Demographics causing the younger “well educated” consumers (that are disproportionately represented in our on-line data) to be especially hammered by the “Great Recession.” They were the shoppers most likely to have ended up “upside-down” in their mortgages, and the ones with the lowest seniority at the job-site.

New and Not-So-Improved

Then suddenly on July 29, 2011 the BEA massively revised their historic data, and the chart changed from the above to this:


Although the divergence is certainly still present, many parts of the line plotting the BEA’s “new historic” GDP have been shifted downward. The most dramatic changes made the “Great Recession” far deeper and the subsequent recovery far milder than previously reported. Among the “highlights” from the revisions for the 10 quarters spanning 4Q-2008 through 1Q-2011 are:

— 6 of the 10 quarters were revised materially downward.

— A total of $376 billion in “real” GDP just evaporated over the 10 quarters, representing the equivalent of 2.6% of a full year’s GDP (and an amount offsetting nearly 70% of the spending portion of the “American Recovery and Reinvestment Act of 2009”).

— An admission that “real” GDP has not yet (through 2Q-2011) recovered to pre-recession (4Q-2007) levels.

Imagine …

The key for us is not so much the new numbers, but the extent to which the old numbers were found to be wrong. Imagine the CFO of an S&P 500 corporation issuing an earnings statement some 30 days after the end of a quarter, reporting earnings of $1.75 billion, calling that earnings statement an “Advance” estimate and providing no guidance for the next quarter. Suppose then that the same CFO issued a revised “Second Estimate” a month later with quarterly earnings raised to $1.84 billion, and yet another “Third and Final” estimate one more month later, now raising the prior quarter’s earnings figure to $1.92 billion, but still providing no guidance.

Now suppose that a mere 35 days after the “Final” quarterly report that same CFO said, “Oh by the way, the first quarter earnings were actually only $0.36 billion.”

What would happen to the share price of that CFO’s corporate equity offering? What would happen to the credibility of the CFO, his accounting team and the corporation’s earnings reports? And would the CFO (or the entire management team) keep their jobs? Or would angry (and misled) investors launch class action lawsuits?

That, however, is almost exactly what the BEA did for 1Q-2011.

The BEA is supposed to provide “hard” data, in contrast to the econometric models or “real-time” measurements provided elsewhere (including here at the Consumer Metrics Institute). But what is “hard” about the data from over two years ago, when 4Q-2008 is suddenly dropped by and additional 2.12%. And that quarter is hardly alone:

Selected Quarterly GDP Revision Amounts

Quarter Revision Amount
4Q-2008 -2.12%
1Q-2009 -1.78%
1Q-2011 -1.41%
4Q-2009 -1.21%
4Q-2007 -1.20%
1Q-2008 -1.05%
4Q-2010 -0.75%

Our Bottom Line

At minimum the recent massive revisions to historic BEA data suggest a fourth possible cause for the “divergence” visible in the charts above:

— The BEA has become severely “challenged” in its efforts to measure what has been happening in this economy (or indeed any economy experiencing dynamic change).

It would be nice to be able to tell you that the BEA is particularly poor at measuring certain segments of the economy, but that apparently is not the case. There was little rhyme or reason to the detailed items involved in the revisions, which variously were the result of seemingly random line-item adjustments or reworked “deflaters.” It would also be nice to tell you that the revisions are merely the result of some data arriving at the BEA headquarters a little late. But it would be a stretch to apply that particular rationale to the substantial revisions to the fourth quarter of 2007.

Cynics, on the other hand, will argue that the BEA publishes numbers close to “consensus” expectations and then revises them when they actually know what happened. But that still doesn’t explain 2007. Alternately, conspiracy theorists will suggest that the BEA spins numbers that are favorable to the political oligarchy, and then quietly revises downward each July hoping that nobody will notice. If that indeed was their plan, then the “nobody will notice” part failed miserably for the data from the first quarter of 2011.

We generously suspect that the BEA is simply a nice place for economic academicians to be employed, particularly given the lack of comparably cushy jobs (and benefit programs) in the real world. And the culture rewards those who don’t ruffle any feathers — particularly by challenging the methodologies pioneered by Wesley Mitchell in 1937 or compromising the consistency of the data series that they have been collecting since late into the 1930’s.

Our bottom line is that we now have even greater concerns about both the accuracy and the veracity of the BEA data in real time. One hopes that Mr. Bernanke shares our concerns before taking his next turn at the monetary controls.

Related Articles

Analysis Blog articles by Rick Davis

Are We Headed for Great Recession 2.0? by Caroline Corbett and Lance Roberts

2Q2011 Advance GDP:  1.3% Economic Growth is Not Growth by Steven Hansen and Doug Short

CFNAI: Confirms Economic Situation is Worsening in June 2011 by Steven Hansen

Econintersect’s Economic Forecasts