Purchasing Managers Index
Purchasing Managers' Indexes (PMI) are economic indicators derived from monthly surveys of private sector companies. The two principal producers of PMIs are Markit Group, which conducts PMIs for over 30 countries worldwide, and the Institute for Supply Management (ISM), which conducts PMIs for the US.
Markit Group and the Institute for Supply Management separately compile The Purchasing Managers' Index (PMI) surveys on a monthly basis by polling businesses that represent the makeup of the respective sector. The surveys cover private sector companies, but not the public sector.
The current Purchasing Managers' Index (PMI) survey can be found under "ISM Reports On Business" on the ISM home page, or directly on this page. The surveys are released shortly after the end of the reference period. The actual release dates depend on the sector covered by the survey. Manufacturing data are generally released on the 1st working day of the month, followed by construction on the 2nd working day and services on the 3rd working day.
The Institute for Supply Management (ISM) in Tempe, Arizona, began to produce the report for the US in 1948. The data for the index are collected through a survey of 400 purchasing managers in the manufacturing sector on five different fields, namely, production level, new orders from customers, speed of supplier deliveries, inventories and employment level. Respondents can report either better, same or worse conditions than previous months. For all these fields the percentage of respondents that reported better conditions than the previous months is calculated. The five percentages are multiplied by a weighing factor (the factors adding to 1) and are added.
The Chicago-PMI survey registers manufacturing and non-manufacturing activity in the Chicago region. Investors care about this indicator because the Chicago region somewhat mirrors the nation in its distribution of manufacturing and non-manufacturing activity.
The Markit and ISM Purchasing Managers Indices include additional sub indices for manufacturing surveys such as new orders, employment, exports, stocks of raw materials and finished goods, prices of inputs and finished goods.
Other similar purchasing managers indices are operated by the IFO in Germany, The Bank of Japan in Japan (tankan), the China PMI operated by the Chinese Government, and the Swedish PMI run by private bank Swedbank.
Formula and calculation
PMI data are presented in the form of diffusion indexes and are calculated as follows:
INDEX = (P1*1) + (P2*0.5) + (P3*0)
- P1 = Percentage number of answers that reported an improvement.
- P2 = Percentage number of answers that reported no change.
- P3 = Percentage number of answers that reported a deterioration.
Thus, if 100% of the panel reported an improvement the index would be 100.0. If 100% reported a deterioration the index would be zero. If 100% of the panel saw no change the index would be 50.0 (P2 * 0.5).
Therefore, an index reading of 50.0 means that the variable is unchanged, a number over 50.0 indicates an improvement while anything below 50.0 suggests a decline. The further away from 50.0 the index is, the stronger the change over the month. E.g. a reading of 55.0 points to a stronger increase in a variable than a reading of 52.5.
Headline Manufacturing PMI
The headline manufacturing PMI is a composite of five of the survey indices. These are New orders, Output, Employment, Suppliers' delivery times (inverted) and Stocks of purchases. The ISM attributes each of these variables the same weighting when calculating the overall PMI, whereas Markit uses the following weights: New orders (0.3), Output (0.25), Employment (0.2), Suppliers' Delivery Times (0.15), Stocks of purchases (0.1).
Markit Economics' PMI surveys
Purchasing managers form a near ideal survey sample base, having access to information often denied to many other managers. Due to the nature of their job function, it is important that purchasing managers are among the first to know when trading conditions, and therefore company performance, change for the better or worse. Markit therefore uses such executives to produce data on business conditions.
In each country, a panel of purchasing managers is carefully selected by Markit, designed to accurately represent the true structure of the chosen sector of the economy as determined by official data. Generally, value added data are used at two-digit SIC level, with a further breakdown by company size analysis where possible. The survey panels therefore replicate in miniature the actual economy. A weighting system is also incorporated into the survey database that weights each response by company size and the relative importance of the sector in which that company operates. Particular effort is made to achieve monthly survey response rates of around 80%, ensuring that an accurate picture of business conditions is recorded over time.
Data are collected in the second half of each month via mail, email, web, fax and phone.
A key feature of the PMI surveys is that they ask only for factual information. They are not surveys of opinions, intentions or expectations and the data therefore represent the closest one can get to “hard data” without asking for actual figures from companies.
Questions asked relate to key variables such as output, new orders, prices and employment. Questions take the form of up/down/same replies. For example, “Is your company’s output higher, the same or lower than one month ago?”
Respondents are asked to take expected seasonal influences into account when considering their replies.
For each main survey question, respondents are asked to provide a reason for any change on the previous month, if known. This assists not only the understanding of variable movement but also in the seasonal process when X12 cannot be used.
The seasonal adjustment of PMI survey data is usually calculated using the X12 statistical programme of adjustment, as used by governmental statistical bodies in many developed countries. However, the X12 programme only produces satisfactory data if five years' historical data are available. In the absence of such a history of data, the PMI survey data are seasonally adjusted using an alternative method (see next paragraph), developed by Markit Economics.
This method was initially designed to provide analysts with a guide to the underlying trend in the survey data and should be recognised as a second-best approach to X12. However, past experience in other countries suggests that Markit’s method of seasonal adjustment goes beyond this initial purpose and in fact in many cases outperforms X12 as a guide to comparable official data.
Markit's method involves using reasons cited by responding survey panel member companies for changes in variables, which are then used to ascertain whether a reported increase or decrease in each variable reflects an underlying change in economic conditions or simply a seasonal variation. Seasonal variations may include changes in demand arising from Christmas, Easter or other public holidays. Normal, expected changes in weather are not included. The net balance of companies reporting an improvement in a variable less those reporting a deterioration is then adjusted to allow for the percentages of companies reporting seasonal induced increases or decreases in the variable.
- See release dates section
- Markit PMI
- 'Inköpschefsindex' (Swedish) http://www.swedbank.se/om-swedbank/analyser/samhallsekonomi/inkopschefsindex/ Retrieved 2013-04-02