- The Big Picture - http://www.ritholtz.com/blog -

Debunking the ISM

Posted By Barry Ritholtz On June 3, 2004 @ 12:01 pm In Finance | Comments Disabled

A professional acquaintance I know (who abhors the limelight) has written the most thorough fisking of the ISM data [1] I have ever seen:

“Like the Philly FED [2], the ISM makes use of a diffusion index to arrive at some very misleading indicators. We also note that the PMI headline number is a composite index based on the seasonally adjusted diffusion indices for five of the components with varying weights:

New Orders, Production, Employment, Supplier Deliveries and Inventories. Probably the worse reference we can now give the ISM report is to announce that the seasonal adjustment criteria are courtesy of the Department of Commerce. (Do we have to repeat what we think of government stats?)

So if the ISM is relying on the government for seasonal adjustments, well. Whatever. That said, let’s take a closer look:

The following components scored a sentiment improvement:

Employment, Supplier Deliveries, Inventories, Import Orders.

The following components scored a lower sentiment:

New Orders, Production, Customer Inventories, Backlog, New Export Orders.

Of course, these stats are misleading, up and down, given the process they undergo before publication, but we thought that the piece de resistance of pure unadulterated enzyme-free donkey fazoo, the Prices Component, deserved special recognition: Prices were touted as 2 percentage points lower at 86 in the month of May. Ahem. The sentiment responses broke down as follows:

Higher: 74%
Unchanged: 24%
Lower 2%

So despite the fact that 74% of respondents said that they were paying higher prices, the actual read on Prices was lower.

How do they arrive at this fantasy? The Philly FED simply subtracts the “lower” from the “higher” responses and posts that figure. The ISM is a bit fancier. They take the “higher” read and to that, they add one-half of the “unched” responses. So 74 + 12 = 86 in this instance. (NB: This convoluted process in determining where the individual indices stand is why we use the term “scored” above, to describe the results of this sentiment survey.)

The upshot is that the printed numbers are grossly misleading as the red headlines make you think that Prices are actually lower. What a bunch o’ baloney! Apologies to the govvie dealers are in order.

And another side-splitter on Prices: at the end of the ISM report, they give a list of manufacturing commodities in short supply and also up/down in price. Clearly, the list of those things “up” in price is significantly extensive.

Down in price? Caustic soda, Copper, Eggs, Nickel and Steel Scrap. What’s funny about that list? Easy. Caustic soda and Copper were reported both up and down in the same month. Now think about some overzealous Purchasing Managers who got bagged both ways. Stuff happens, dig?

The Employment Index, which is always good for a laugh:

The tape touted this as the “7 th consecutive month of expansion, following a 37-month trend of contraction.” And: “last time Employment Index registered higher than May’s index was in April of 1973 (62.6%).”

Specifically, May advanced 4.1 percentage points over April to 61.9. This is despite the fact that 57% of respondents rated their sentiment on the employment index as UNCHANGED. This Index, then, follows in the footsteps of the worthless Philly FED’s almost-identical read on the hiring situation. 31-yr record this. Even bigger apologies now in order to the govvie dealers as the press has touted repeatedly overnight, their blowing out of US government securities on the heels of the “robust” Employment Index of the ISM. Geez, Louise. Next case.

Terrific stuff . . . I only wish I could convince this closet genius to go public.


Article printed from The Big Picture: http://www.ritholtz.com/blog

URL to article: http://www.ritholtz.com/blog/2004/06/debunking-the-ism/

URLs in this post:

[1] ISM data: http://www.ism.ws/

[2] Philly FED: http://www.phil.frb.org/

Copyright © 2008 The Big Picture. All rights reserved.