PMIs are not a good month-on-month gauge

It’s confession time for IHS Markit, the firm behind the purchasing managers’ indices.

The firm asks thousands of purchasing managers whether activity over the current month has declined, improved, or stayed about the same compared with the previous month. IHS Markit then tallies responses up to produce a net figure where 50 signals no change in activity from month to month. Therefore, anything below 50 is seen as a contraction, and anything above 50 an expansion.

Recently, however, the figures – which central bankers rely on as a bellwether for growth – have been increasingly looking out of whack with what’s going on in the real world.

As we’ve noted before, when read at face value, they show that activity levels in many economies in Europe continued to decline in May and June despite an easing of lockdowns.

This morning IHS Markit added an important disclaimer to that effect to its press release recording the final eurozone PMIs. It said – in essence – that the indicators should not be read as an accurate reflection of changes in output from month to month when conditions are as volatile as they are right now: (h/t to our favourite twitter unicorn @econhedge)

To be fair, this is a point Chris Williamson, IHS Markit’s chief business economist, has made to us in the past. (IHS Markit has today also published an article going into a great deal of depth about what the European manufacturing PMIs do – and don’t – show right now. It’s here and it’s worth reading.)*

The problem is that the reason why the PMIs are so influential is because of their timeliness and their ease of use – the 50 level viewed by many as some magic marker that separates an expansion in activity from a contraction from month to month.

See, for instance, this excerpt from a speech delivered by the Bank of England’ s chief economist Andy Haldane yesterday:

Globally, fast indicators point towards the trough in activity in the UK’s main trading partners having come during, or just after the end of, the first three months of the year: in China in February and in the US and euro area in early April. That means UK-weighted global GDP has perhaps been recovering for around three months . . . 

...Although the PMIs are a less good guide than usual to GDP outturns, they have bounced back sharply, as have some other business surveys.

Haldane’s reading of the PMIs is odd in that, rather than showing a “bounce back”, most of the indices he quotes remain below 50 – which suggests activity levels are still weakening despite fewer restrictions on movement. That speaks to a bigger point – that the PMIs are now becoming so hard to read that people are using them to make whatever point they feel like. In Haldane’s case, that we are heading for a v-shaped recovery.

This is the wrong approach. It’s an unsettling truth, but the distortions in the PMIs and other economic data, including inflation, show that policymakers are facing even more uncertainty about where the economy is at right now than they might of thought.

At least IHS Market has done the honourable thing and stepped up to admit the issue.

*We added this mention of the IHS Markit report on Wednesday evening, after the original post was published.

Related links
PMIs: perpetually misleading indicators? – FT Alphaville
The European PMIs look troubling – FT Alphaville
When the economy re-opens, will anyone turn up? – FT Alphaville
Is the virus, not lockdowns, doing most of the economic damage? – FT Alphaville
Inflation figures are about to get fuzzier – FT Alphaville