Many equity and bond investors seeking regular insights into economic performance will check out economists’ views on the Institute for Supply Management’s (ISM) monthly business reports on the manufacturing and non-manufacturing industries. The reports include the purchasing managers’ index (PMI), which in the US the ISM produces in tandem with the Department of Commerce (DOC).
The long-established regularity of these data releases is such that a) how they’re put together, b) what they comprise and c) why they’re important may not necessarily be all that clear, so I thought it worth a quick look at the background of how it all works, at least in the US.
The ISM Manufacturing Report on Business®, the composite index of which is the PMI (on which more later), is based on data supplied by more than 300 US company purchasing and supply executives. These respond anonymously and confidentially to a monthly questionnaire designed to elicit facts, not opinions, about changes from one month to the next in;
- New orders
- New export orders
- Lead times
- Timeliness of supplier deliveries.
The questionnaire also seeks data on commodities (whether prices or supplies are up or down) and general remarks on business conditions.
The Non-Manufacturing ISM Report On Business® is based on monthly surveys of more than 375 purchasing executives in US non-manufacturing industries. The report covers;
- Business activity
- New orders
- Backlog of orders
- New export orders
- Inventory change
- Inventory sentiment
- Supplier deliveries
A composite index is then calculated as an indicator of the overall economic condition of the sector. The NMI is based on the indices for four indicators: business activity (seasonally adjusted), new orders (seasonally adjusted), employment (seasonally adjusted) and supplier deliveries.
The PMI is also a composite index, but one made up of five “sub-indicators” – for production, new orders, supplier deliveries, inventories and employment – based on surveys of more than 400 US purchasing managers, chosen for their geographic and industry diversification.
Through analysis by the DOC, these five components are adjusted for normal seasonal variations and calculated as a single monthly index number. The survey gives respondents just three options for each question: are conditions in their industry, as they see it, “better”, “the same”, or “worse”?
The resulting PMI figure (from 0 to 100) is calculated from the percentage of respondents reporting “better” conditions than the previous month and then adding half of the percentage of respondents that reported “no change”. So a PMI reading of 50 would show an equal number of respondents reporting “better conditions” and “worse conditions”.
What PMI means for investors
The PMI is a very important sentiment reading of manufacturing and the economy as a whole. The magic number is 50. A reading of 50 or higher generally indicates expansion. If manufacturing is expanding, the general economy should be doing likewise, so it is considered a good indicator of future GDP growth. The month-on-month rate of change is also important. A reading of 51 (i.e. still in expansion territory) after a month with a reading of 56 would not be viewed well, especially if the economy had recently been showing solid growth.
For bond investors, the growth in supplier deliveries and prices paid can be of greater importance as these have been historical pivot points for inflationary concerns.
Headlines or components?
When the ISM or other PMIs are released, economists (including myself I have to admit) often go into details about the components. However, the growth components (Production, New Orders, Order Backlog, Supplier Deliveries) all have high correlations with the overall index, so in broad terms the additional information from these components is limited.
Sometimes the new orders minus inventories indices are said to have leading aspects. Indeed, the correlation between this created sub-index and the overall index increases from 0.56 with no lead or lag to 0.61 with a two-month lag. That looks promising, but doing a simple linear regression from 1989 to 2012 and then forecasting 2013 out of sample (taking the actual orders minus inventories to forecast the overall index two months ahead) gives disappointing results. It only had the direction of change correct in 10 out of 18 cases.
My own conclusion is that if these data releases are being viewed to get an indication of overall growth, it is enough to look at the headline index. If your area of interest is more particular, then of course some of the sub-indices (such as employment) will prove useful.
 See also www.ism.ws and for instance http://www.investopedia.com/terms/i/ism-nonmfg.asp
 PMIs are also produced by organisations such as Markit, which produces indices for more than 30 countries, and regional sources such as ISM-Chicago. More on http://www.markit.com/product/pmi and https://www.ism-chicago.org/