U of M, AP Falsely Blame Consumer Sentiment Drop on Stock Market Plunge

August 28th, 2015 10:22 PM

At the Associated Press today, Christopher Rugaber appears to have played along with a game of make-believe in his coverage of the August release of the University of Michigan's Survey of Consumers.

The index (dowloadable PDF is at link) dropped for the second straight month, this time from 93.1 to 91.9, a point below August's prelimnary reading of 92.9. That trailed expectations that it would come in at 93.0. The survey's director, Richard Curtin, claimed that the drop occurred "mainly due to the recent volatility in stock prices." Whatever his reason for making that claim, it doesn't pass the smell test, and Rugaber had all the information needed to figure that out (which he may have) and report it (which he didn't).

Here are the key paragraphs from Rugaber's writeup (bolds are mine):

SURVEY: PLUNGING STOCK PRICES HIT US CONSUMER SENTIMENT

Plummeting stock prices have taken a toll on U.S. consumer confidence, though there are signs the setback may be temporary.

The University of Michigan says its consumer sentiment index fell to 91.9 this month from 93.1 in July. The index is still up 11.4 percent from a year ago.

The figures provide an early read of the impact on consumers from the 1,900 point drop in the Dow Jones industrial average over six days through Tuesday. Stock prices have since recovered some of those losses.

The University of Michigan surveys consumers throughout the month and so some of the responses were tallied as stock prices plunged. The University said it extended its interviews until Aug. 23.

Many economists were reassured by the limited drop in the index.

The big problem with blaming the drop on the stock market is that it didn't drop below 17,000 until Thursday, August 20, the second-last business day before the extended survey cutoff of August 23, as seen in the following table for the Dow Jones Industrial Average:

DJIA073115to082415

The DJIA's August 3 through August 18 closes ranged from a high of 17,615 on August 10 to a low of 17,349 on August 19. Though that August 19 close represented a 163-point, one-day drop, it was still fewer than 60 points below three of the previous week's closes. So "plummeting stock prices" really didn't begin until August 20. At most, survey respondents could have been taking that plunge into account for four calendar days — and a large plurality if not a majority of Americans don't follow the stock market's daily ups and downs unless and until they break into the topline national headlines.

Investopedia, in describing the U of M survey, claims that "The preliminary report, which includes about 60% of total survey results, is released around the 10th of each month." So the remaining 40 percent of the survey — or even 50 percent if one assumes that the the survey's extension picked up more respondents than usual — looks like it was spread over about 15 days. It would therefore appear that the highest possible percentage of respondents who could (emphasis could) have taken "plummeting stock prices" into account was about 13 percent (50 percent times 4 days divided by 15). Since not everyone follows daily stock prices closely, it's seems very safe to say that the number of respondents taking the stock plunge into account was well below 10 percent.

Rugaber's reference to the "1,900-point drop" in the Dow "through Tuesday" is way too clever, and gives those not reading his work in audit mode the impression that the entire drop might have been within the survey's scope. Nearly 800 of those points were lost after the survey's cutoff (including August 25, not shown in the table above), and the markets didn't really get widespread headline-level general-news attention until the Dow plunged over 1,000 points in the opening minutes of trading on Monday — the day after the final survey interviews.

The contrarian blog Zero Hedge also noticed something the U of M wasn't particularly interested in telling readers in its narrative, and that Rugaber didn't report:

(The Expectations Index fell) to its lowest since 2014 (biggest 7mo decline in 2 years).

The "since 2014" portion of that claim is confirmed here.

U of M's narrative dressed it up as follows:

The Expectations Index posted a much smaller loss, falling only slightly to 83.4 in August from July’s 84.1 and remaining significantly higher than the 71.3 recorded last August.

The year-ago comparison shouldn't be presumptively more important than the December peak, because as far as I can tell the results are not seasonally adjusted — nor should they be. So it's at least as relevant as the year-ago value that the Expecations Index is down from its January 2015 high of 91.0.

I think it's more than fair to ask if the U of M extended the survey by a few days so it could trot out a ready-made (but bogus) "external" excuse for a decline in sentiment other than the one which makes far more sense, namely that consumers believe — based on the fundmentals of the economy, not the stock market — that the U.S. economy and especially their own personal financial prospects aren't all that strong.

If so, note for the record that the AP's Rugaber played right along.

Cross-posted at BizzyBlog.com.