December 2011 ISM Manufacturing Shows Improving Economy

The ISM Manufacturing survey index (PMI) improved from 52.7 to 53.9 in December 2011 (50 separates manufacturing contraction and expansion).   The improvement was better than expected;  the market expected between 53.2 and 53.4.

Econintersect sees the new orders sub-index as having a higher and more precise correlation to recessions and the economy then the PMI – and the new orders sub-index improved strongly this month.

“The PMI registered 53.9 percent, an increase of 1.2 percentage points from November’s reading of 52.7 percent, indicating expansion in the manufacturing sector for the 29th consecutive month. The New Orders Index increased 0.9 percentage point from November to 57.6 percent, reflecting the third consecutive month of growth after three months of contraction. Prices of raw materials continued to decrease for the third consecutive month, with the Prices Index registering 47.5 percent, which is 2.5 percentage points higher than the November reading of 45 percent. Manufacturing is finishing out the year on a positive note, with new orders, production and employment all growing in December at faster rates than in November, and with an optimistic view toward the beginning of 2012 as reflected by the panel in this month’s survey.”

PERFORMANCE BY INDUSTRY: Of the 18 manufacturing industries, nine are reporting growth in December, in the following order: Apparel, Leather & Allied Products; Printing & Related Support Activities; Textile Mills; Petroleum & Coal Products; Machinery; Food, Beverage & Tobacco Products; Computer & Electronic Products; Primary Metals; and Paper Products. The nine industries reporting contraction in December — listed in order — are: Plastics & Rubber Products; Nonmetallic Mineral Products; Furniture & Related Products; Chemical Products; Wood Products; Miscellaneous Manufacturing; Fabricated Metal Products; Transportation Equipment; and Electrical Equipment, Appliances & Components.

It is interesting to note that ISM Manufacturing represents less than 10% of USA employment, and approximately 20% of the business economy. Historically, there is little linkage between real employment and the ISM employment index.

New orders have direct economic consequences – IF the opinions are accurate.  Expanding new orders is not a reliable sign a recession is imminent;  New Orders have given false recession warnings twice since 2000.

Based on the new orders component of ISM manufacturing an improving economy is indicated.

Caveats on the use of ISM Manufacturing Index:

This is a survey, a quantification of opinion – not facts and data.  However, as pointed out above, certain elements of this survey have good to excellent correlation to the economy.  Surveys lead hard data by weeks to months, and can provide early insight into changing conditions.

Econintersect has run correlation coefficients for the ISM manufacturing employment and the BLS manufacturing employment data series above going back to 1988, using quarterly data. The coincident correlations are actually negative, but poor (r = -0.2 to -0.4 for various time periods examined).  See here for definitions.

Before 2000 the ISM employment data had a weak positive correlation to the BLS data 4 to 7 quarters later (r values above 0.6).   Since 2000 the correlations for ISM manufacturing employment as a leading indicator for the BLS manufacturing employment have been between 0 and 0.3 for r (correlation coefficient).  These values define correlations as none to poor.

In other words, ISM employment index is not useful in understanding manufacturing jobs growth.

Addendum: The graph below shows BLS private non-farm employment month-over-month gains against the ISM Manufacturing employment index.

Related Articles

All Articles on Institute of Supply Management Surveys

3 replies on “December 2011 ISM Manufacturing Shows Improving Economy”

  1. The repeated complaints about surveys seem a bit doctrinaire. Does anyone on the GEI site have any experience in doing surveys? Anyone that even took a class on the subject? Are any of the criticisms based upon any experience?

    Just wondering.

    Meanwhile — the ISM survey asks about facts that purchasing managers know. Do you think that they are unaware of hiring? Inventories? Prices paid?

    Are they lying?

    Meanwhile, my own research has found a nice contribution from this survey to employment growth. Dean Baker tells me that the employment sub-index is even better on this front.

    My own approach involves getting as much information as possible from each source. GEI seems to be very receptive to many sources that might graciously be viewed as “new” and “not established” and “speculative.”

    So why the tough line on sources like the ISM?

  2. Good day Jeff,

    even if you are criticizing, i personally appreciate your perspective. the “truth” is never simple – and it forces me to re-look at my analysis critically.

    to answer your question, early in my other life i was a purchasing manager – and was involved in this survey. the issue is not whether they are lying, but in a position in that organization whether they are getting the information first hand, or if it is scuttlebutt. every organization is different – and many purchasing managers are a subunit of accounting.

    further, i will admit i was too busy to do the survey directly – and i gave it to my secretary or subordinate to complete. my performance evaluation never rated my ability to correctly fill out the survey. my bosses never saw the survey to correct my misconceptions or errors.

    yet, this is not the basis of the warnings. i look for correlations to real data. if i could correlate ground hog sightings to employment – i would use it. it does not matter whether the source is logical or not – track record counts.

    i have found good correlation between new orders portion of ism manufacturing and the economy – and this was pointed out in the post. i do not find reasonable correlation to employment (i have just added a graph to the post showing correlation between the ism survey and non-farm private employment).

    from my perspective, most elements of surveys and hard data roughly parallel economic activity. this issue is whether in real time, there is enough precision to forecast.

    overall, i find this survey economically positive. but i accept your criticism which could be interpreted as trashing ISM (and i am not). i will work on toning down the heat.

  3. Steven — Thanks for your reply.

    Your story about delegating the survey to an assistant reminds me of how some coaches do the weekly football poll!

    My success in using the ISM data comes from a multivariate model for “explaining” monthly payroll job changes. I use the final series — after revisions and benchmarking. The data covered a long period of time. The relationship might not be obvious in a chart covering a shorter period, no revisions, and with only two variables.

    My research approach to employment changes, one of the most extensive areas of discussion on my blog, includes looking at both job creation and job destruction. You probably need at least two variables to do this. I considered the ISM services index, but the series is a lot shorter. I have not yet gone back to test Dean Baker’s suggestion.

    Just as your work reflects your background and skills, my approach is what you would expect from a social scientist who got a PhD from Michigan. Nearly all of us spent time at the Institute for Social Research, where “survey data” is not an oxymoron! I also have conducted surveys and taught that course for undergraduates.

    Sometimes a survey is the best, and maybe the only way to get the data you need in a timely fashion. In the first two weeks of a class, you learn all of the standard objections and the best techniques for dealing with them. Sample design and question wording are obviously important. Given that hundreds of very intelligent people have been working on these problems for decades, it is not surprising that techniques have gotten better.

    Most obviously you can see the improved results in election surveys.

    My conclusion is that the survey is another data source, with strengths and weaknesses like any other. It is usually coincident or only slightly leading.

    Thanks again for your gracious response.


Comments are closed.