Thursday, July 22, 2010

Beware the Bureau of Economic Analysis 1937-Era Focus on Measuring Factory Activity Measures

As a part of Fed Chairman Bernanke's remarks yesterday before the Senate Banking Committee, he mentioned that "real consumer spending appears to have expanded at about a 2-1/2 percent annual rate in the first half of this year, with purchases of durable goods increasing especially rapidly."

I sent a link of Bernanke's testimony to Rick Davis of Consumer Metrics Institute to get his take on Bernanke's comment. As long-term EPJ readers know, I consider Davis' macro-tracking of the economy (in real-time) to be the best in the business. Here's Rick's response to me:


Thank you for the link to Mr. Bernanke's testimony.

Yesterday (even as Mr. Bernanke was speaking) I was writing commentary for our website addressing what we expected the BEA's numbers to be when the 2nd quarter GDP release comes out on July 30th. My posting is below, but the essence is that although the numbers should in theory be even softer than the 1st quarter numbers (which surprised most observers to the down side), I would ultimately not be surprised by anything that they publish --
for all of the reasons listed below.

That said, I can't imagine Ben Bernanke speaking before a Congressional committee without first gaining privileged access to the preliminary BEA data. Based on his testimony, and presuming that he had an advanced peek at the data, we should expect the numbers to come in somewhere between 2% and 2.5% if his projection of 2.5% for the first half is correct (although it's not clear exactly what he expects to be growing at 2.5%, since there are about a half dozen different ways to calculate "real consumer spending").

However, at the Consumer Metrics Institute we see no evidence that "real consumer spending" was expanding at that rate, nor signs of "purchases of durable goods increasing especially rapidly."

Furthermore, our readings of the previous Q1 2010 GDP release certain don't support numbers that high. As even Federal Reserve Bank of Boston President Eric Rosengren observed: "final sales — which is just taking inventories out of GDP — only grew by 0.8% in the first quarter". If this is correct, we're not sure how "real consumer spending" could have a first half growth rate of 2.5% without some extraordinary numbers for the second quarter.

But, anyway, here is what I was saying yesterday before I had Mr. Bernanke's testimony in hand:


July 21, 2010 - What Will the July 30th GDP Report Say?:

On July 30th the BEA will release their preliminary estimate for GDP growth during the 2nd quarter of 2010. What does our data predict about the numbers that will be in that report? The short answer is:

► While we have a very good idea what consumers were really doing during the 2nd quarter -- and by inference what factories should have been doing in response -- we have have no faith at all that the BEA's 1937 inspired methodologies will come remotely close to capturing the reality of the consumer economy.

We have been surprised by some other projections about the new GDP numbers. The Association of American Railroads has extrapolated their rail-car loadings report to forecast that if "the 88% correlation (between their numbers and the GDP) that held from Q1 2000 to Q1 2010 held for another quarter, Q2 2010 GDP would be 5.0% higher than Q2 2009 GDP, which translates to 9.3% higher than Q1 2010 GDP." (Rail Time Indicators, July 2010, page 17). Most other opinions have reflected more modest growth, ranging from 3% to 4%, with the full year ultimately growing at about a 3% rate. In general, most economists have viewed the halving of economic growth from Q4 2009's 5.7% to Q1 2010's 2.7% as a return to normalcy after a couple of quarters of heated recovery.

Others have been less optimistic. Federal Reserve Bank of Boston President Eric Rosengren said in an interview with The Wall Street Journal that if "you look at final sales — which is just taking inventories out of GDP — final sales only grew by 0.8% in the first quarter and that is after two previous quarters that averaged below 2% as well. We're getting to the point in the recovery where we wouldn't expect as much support coming from the inventory side. If inventories start to ebb it becomes really essential for some of the other components of GDP to start to pick up at this time and there is some reason to believe that we may not get as much of a pick up as some had been anticipating earlier this year."

Mr. Rosengren went on to explain that one reason "would be the labor market weakness. The employment to population ratio has dropped in the past two months. It is now 58.5% and is well down from where we were before the recession ... We're not seeing significant signs of improvement yet."

His conclusion?

"When I put all of that together I become concerned that the second half is going to be a little weaker than we might have anticipated a couple of months ago and I'm not expecting to see that much progress on the unemployment rate over the course of the second half of this year. Ideally we'd be seeing growth north of 4% in order to be really pushing the unemployment rate down from its very elevated levels and we're not seeing growth at nearly 4% at least for the second half this year. Unfortunately, it looks like it will be a good bit slower than that. It is quite possible we'll be in the 2% to 3% range."

Even Mr. Rosengren's modest 2% to 3% range might seem high given his observation that first quarter "final sales only grew by 0.8%" and if "inventories start to ebb it becomes really essential for some of the other components of GDP to start to pick up". Maybe housing starts?

So, what do we expect the BEA will report on July 30th?

Who knows? We have noted before a number of reasons why the BEA's numbers can differ from what we measure:

► Their original (and ongoing) 1937-era focus on measuring factory activity levels. Factories are way downstream economically from where today's real economic action is -- probably 4 or 5 months. We understand why factories were emphasized back in 1937 (given FDR's constituency and 1937 jobs demographics), but the economy is much more than 'factories' in
2010. How many iTunes sales are captured in the factory reports? Or in 'Rail Car Loadings', for that matter?

► Their questionnaire approach leads to survivor and large firm biases. Not to mention the time lags and revisions when the data does finally come in.

► Their inclusion of inventory adjustments creates wild swings, which is the ultimate consequence when trying to reconcile their measured factory production levels to the actual consumer demand that we measure.

► Their inclusion of non-consumer stimuli (e.g., 747's, aircraft carriers & interstate highways), which accounts for about 30% of the economy -- but which we don't measure at all. Federal stimuli run amuck could increase their numbers substantially.

► Our numbers are strictly year-over-year growth, and require no seasonal adjustments. Their numbers represent surveyed quarterly growth, which are then "annualized" and seasonally adjusted.

But, if the BEA (through dumb luck) was accurately reflecting actual consumer demand, they would report a contracting 2nd quarter.

What the BEA's freeze-action snapshot of the first quarter missed was the dynamic nature of demand during the quarter. The static picture raised the "glass half full" vs "glass half empty" conundrum when economic growth was suddenly cut in half. We might ask: was the plunge arrested, with the economy subsequently bottoming at a "normal" 2.7% growth rate? Or was the plunge sustained?

Our data would suggest the sustained drop scenario. And by now the daily year-over-year contraction in consumer demand has deteriorated to near where 2008's "Great Recession" was 180 days after that earlier contraction began.

The bottom line?

We don't know what the BEA will report on July 30th. Maybe stimulus spending will push the number up. Maybe factories continued to build up inventories in anticipation of a strengthening recovery -- and if that happens, watch for factories to over-correct the other way during the third quarter (with numbers that will be released 4 days before the U.S. mid-term election).

Or maybe the BEA will accurately reflect what has actually happened to consumer demand. In any event, it should be interesting.


Bob, thanks again for the link to the testimony. As I said above, next Friday should be interesting ...

Rick Davis
Consumer Metrics Institute


  1. Fascinating. Thank you Wenzel and Rick Davis!

  2. RW - no good can come of this fascination with "unadjusted" data as if it were somehow more revealing. Richard's own description shows its uselessness:

    As Richard explains,

    "We simply report what consumers have been doing on a day by day basis by mining on-line U.S. consumer tracking data for purchases of discretionary durable goods. We only look for discretionary durable goods transactions because we believe the discretionary durable goods segment of the consumer economy is the most volatile and stimulative portion of the economy." Stimulative?

    So if consumer spending as he measures it continues to decline we will suffer a "...double dip in the demand side of the economy." And if it doesn't decline but instead inclines we will "enjoy" stimulation of the demand side? How is this narrow redundant statement useful? Richard seems to think that because his graphs show it anticipating moves by the GDP it is validated. But why do we care what the GDP is doing?

    Macroeconomics was created by politicians for politicians. It is useless except for deceiving and subtrefuge.

  3. Efinancial,

    What do you think of Rothbard or Reisman's ideas about constructing a GPP (gross private product) statistic? Rothbard mentions this concept in America's Great Depression, amongst other publications.

    I see what you're saying about macro econ generally, however.

  4. @eFinancial

    Forget Davis' theories and linking to GDP. As I posted in my earlier comments about the Consumer Metrics Institute, what Davis is tracking is what I consider a very important indicator as far as where we are in the business cycle. He measures "durable goods" spending, what I consider an important part of "capital goods."

    If this is in a downtrend (as it is)it is very much confirming the slowdown in the money supply and the impact it will have in the economy adjusting away from a manipulated money growth period.