CIPS Markit index – more questions?

I raised this a while ago here - whether the CIPS/Markit Purchasing Managers’ Index  (PMI) was still as good an indicator as it certainly was for some considerable time?  Now others are making the same comments in slightly more illustrious publications; here's today's Times. and their Business Editor, David Wighton.

"It appears there may be something awry with the UK index findings. Last month’s reading was higher than that for China. As Colin Ellis at Daiwa says, anyone who thinks that British manufacturing is expanding faster than China needs their heads examined."

31.5% of my degree was statistics, so I got thinking about what might have caused this if indeed there has been a change in accuracy. The obvious question is how large the sample is (is it big enough to be statistically reliable), but just as importantly, is the sample statistically representative of the whole population?  How is it chosen? Or could something have happened recently to make it less representative?  What about if struggling manufacturers have dropped out - perhaps they no longer can afford to employ a Purchasing Manager so they don't submit data any more?  Or companies doing particularly badly just don't want to put in bad news?  Or firms are skewing the data to paint a rosy picture in the hope of it becoming a self-fulfilling prophecy?

Anyway, if I was a subscriber,  I would be asking some questions about the sample; perhaps CIPS or Markit would like to reply here?  Or to the Times?

Discuss this:

Your email address will not be published. Required fields are marked *