The Farce Continues: ISM Still Has No Idea Why Its Computers Malfunctioned, And Much More...

Tyler Durden's picture

We are still in shock by yesterday's epic Snafu involving ISM manufacturing data was extensively documented (here, here, here and here). For those who missed it, the Institute of Supply Management somehow prepared an entire report, press release and prepared remarks by its chair Brad Holcombe, which minutes after its official, and extensively vetted release, turned out to be based on... completely flawed numbers. The reason: the ISM has somehow applied erroneous seasonal adjustments, supposedly adjusting unadjusted May data using April adjustment factors (we do hope the ISM makes a habit of disclosing to the public what all its data looks like, not just its seasonally-adjusted numbers).

As a result the ISM was forced to selectively adjust a subset of the final numbers, higher of course, not once but twice: after all we are at a very sensitive tapering time for the economy when every single data point simply has to boost confidence at all costs that the economy can sustain itself even with a decline in the Fed's liquiity injection and so nothing can possibly be allowed to show a growth slowdown.

The reason we bring it up is not because tomorrow the ISM is set to release its second for the week report, this time detailing the US non-manufacturing sector (fear not: it will be a smashing success, or else it too will promptly be adjusted higher), which following yesterday's farceworks (sic) is now a complete joke as far as "credible" surveys are concerned, but because the story simply refuses to go away.

As Bloomberg reports, the ISM said it will manually vet all data because it hasn’t identified why its computer software introduced an error into yesterday’s May manufacturing index. That's right: a day after its unprecedented humiliation, the ISM still has no idea how its data was "glitched" by a computer (note the use of the passive voice: there never is an actual person responsible for such glaring errors).

It gets better:

Analysts at the Tempe, Arizona-based purchasing managers association will verify the figures against what the computer spits out, Kristina Cahill, a spokeswoman for ISM said today. A glitch yesterday caused its manufacturing index to drop. That was later corrected to show factories expanded last month at the fastest pace of the year.

 

“We definitely are very serious about making sure that doesn’t happen again,” said Cahill, who added that the group received “a number of inquiries” about yesterday’s botched manufacturing report. “We’re incredibly concerned about credibility, and that’s why we’re taking these steps.”

You mean that an organization that is "incredibly concerned about credibility" never verified the figures that a computer "spits out" previously? As in nobody ever actually checks the numbers that go into the press release paragraph supposedly uttered by Brad Holcombe, who in addition to the quantitative data also lays out a qualitative assessment of the data?

As a reminder, this is what was released yesterday. Somehow this...

ISM'S HOLCOMB SAYS DROP MAY BE JUST A `SLIGHT PAUSE'

 

The report was issued today by Bradley J. Holcomb, CPSM, CPSD, chair of the Institute for Supply Management® (ISM®) Manufacturing Business Survey Committee. "The May PMI® registered 53.2 percent, a decrease of 1.7 percentage points from April's reading of 54.9 percent, indicating expansion in manufacturing for the 12th consecutive month. The New Orders Index registered 53.3 percent, a decrease of 1.8 percentage points from the 55.1 percent reading in April, indicating growth in new orders for the 12th consecutive month. The Production Index registered 55.2 percent, 0.5 percentage point below the April reading of 55.7 percent. Employment grew for the 11th consecutive month, registering 51.9 percent, a decrease of 2.8 percentage points below April's reading of 54.7 percent. Comments from the panel reflect generally steady growth, but note some areas of concern regarding raw materials pricing and supply tightness and shortages."

... became this:

“We apologize for this error. We have recalculated and confirmed that the actual index indicates that the economy is accelerating,” said Bradley J. Holcomb, CPSM, CPSD, chair of the Institute for Supply Management® (ISM®) Manufacturing Business Survey Committee. “Our research team is analyzing our internal processes to ensure that this doesn’t happen again,” he added.

 

“The May PMI® registered 55.4 percent, an increase of 0.5 percentage point from April’s reading of 54.9 percent, indicating expansion in manufacturing for the 12th consecutive month. The New Orders Index registered 56.9 percent, an increase of 1.8 percentage points from the 55.1 percent reading in April, indicating growth in new orders for the 12th consecutive month. The Production Index registered 61.0 percent, 5.3 percentage points above the April reading of 55.7 percent. Employment grew for the 11th consecutive month, registering 52.8 percent, a decrease of 1.9 percentage points below April’s reading of 54.7 percent. The Supplier Deliveries Index registered 53.2 percent, 2.7 percentage points below the April reading of 55.9 percent. Comments from the panel reflect generally steady growth, but note some areas of concern regarding raw materials pricing and supply tightness and shortages."

 

The error resulted when the software incorrectly used the seasonal adjustment factor from the previous month.

Needless to say, this is an absolutely idiotic explanation. So what is the ad libbed narrative we are expected to believe? According to Bloomberg, what happened? This:

Kenneth Kim, an economist at Stone & McCarthy Research Associates Inc. in Princeton, New Jersey, was the first to question the accuracy of the data. On a hunch that there was a discrepancy, he applied the April seasonal-adjustment factors to the raw data for May and came up with the number initially released by ISM.

 

In the wake of the discovery, “people I haven’t heard from for 10 or 15 years have messaged me out of the blue, saying, ‘great job,’” Kim said today in an interview.

 

The ISM wasn’t one of them. “I’m not expecting a call from them, which is fine,” he said. “I’ve caused them enough consternation.”

This is correct: we said just this yesterday:

... after we read the Stone McCarthy (SMRA) emails which, sent out at 10:22 am, or 22 minutes after the official ISM release, stated the following:

What follows is our first impression of the May ISM Manufacturing Index. In a later update, we will provide a more complete analysis.

 

The May ISM Manufacturing Index was reported at 53.2 from 54.9 in April. The declined wiped out the modest rebound in April from 53.7 in March and put the pace of expansion back to one consistent with only modest activity. Four of the five components that make up the index were down, and one was flat.

 

Market expectations centered at 55.5 within a range of 54.0 to 57.0 in a Bloomberg survey. Our expectation for the index was 55.3. The headline was below market expectations. This will be a disappointment in light of a hoped-for pickup during the spring months, although overall activity remains positive.

 

We believe the ISM may have misreported the data because our calculated results do not match up with published data. Our calculation would place the index at 55.4.

Which is ironic, because 55.4 is what the first ISM headline number revision of 56 (up from 53.2), was ultimately revised into for the second revision. Almost as if Stone McCarthy was not only telling the world what happened, but was also calculating the ISM's numbers for them (which probably would explain why it took ISM 6 hours to update its website with the "correct" report).

Stone McCarthy added this:

However, the ISM reported the May seasonally adjusted new orders index at 53.3. We have traced the error to utilizing the improper seasonal adjustment factor for May. Instead of utilizing the May factor of 1.064 for new orders, the ISM reused the same factor from April of 1.135, which erroneously produced a seasonally adjusted level of 53.3 using an unadjusted index of 60.5.

Which is great, because thanks to SMRA we finally get a glimpse of these mythical unadjusted numbers. Remember: it is the seasonal adjustment that is supposedly the culprit for everything.

So what actually happened in May based on real, unadjusted data?

Take a wild guess.

Below we share the table that SMRA sent out to subscribers which is the only place, to our knowledge, where unadjusted ISM data is available. And there you have it: the May unadjusted New Orders print was 60.5: the lowest since February while the % of respondents saying "Better" was actually tied for the lowest since January. Somehow, with the magic of seasonality this ended up being the best adjusted print of the year!

We must admit: we have no idea why one would seasonally adjust a survey's responses: after all when one responds "better", "same" or "worse" about how the economy is doing that is... well, their final answer, net of the seasons.

But we do have an idea that when the entire market moves due to a seasonal adjustment which makes the worst month since February into the best month of 2014, then something is very, very wrong, and certainly very manipulated.

We certainly have an idea why the ISM had zero desire to make these unadjusted numbers public in the first place.

All that said, we can't until tomorrow's Service ISM goldilocks number is released again, and again, and again, until the "just right" market reaction is evoked.