Diffusion indices like the PMI data can be tricky. And have been particularly so since COVID, because activity trends have been so divergent across sectors. Japan is a particular case in point this month: PMI readings have been weak over the summer but actual output has been super-hot. Autos are the main reason.
The PMIs surveys they tell you about the net proportion of firms experiencing a rise or fall in – for example – output or new orders. When activity is highly correlated across sectors they are a pretty good guide to actual output growth; when it is very divergent across sectors, as has often been the case since COVID, they are a poor guide to actual output.
To see why, imagine a situation when 80% of output is produced by just 20 firms and 20% of output by 80 firms. In an unweighted diffusion index you would get a very low PMI New Orders number if most of the small firms had declining orders and most of the large firms had rising orders. So output could in principle be very strong even though the PMI number was very weak.
Japan is the current case in point. In May, PMI new orders were 50.4 but industrial production fell by over 7.5% on the month. In June, PMI new orders was 50.2 but industrial production jumped by 9.2%. In both cases auto and machinery production were the big swing factors, and output trends in most other sectors were pretty similar.
Meanwhile, the METI projections for production in July and August are for further rise of 6% or so, even though manufacturing new orders slipped to 47.6 in today’s flash report for August. METI’s surveys are usually over-optimistic: on average actual output is around 60% of the survey number, but even so we expect decent further gains in Japanese IP for July and August. As a result IP momentum in Japan will be super-hot at a time when PMI new orders were a) close to 50 and b) trending down - Figure 1.
Simplifying a little most of that is explained by autos, where sales have been improving as COVID cases receded, and output is now catching up because the chip shortage has eased - Figure 2.
Japan is not alone. US auto output has also surged in recent months as chip shortages eased despite the downtrend in PMI new orders, while German auto production jumped about 20% in Q2 this year (Figure 3) at the same time as PMI new orders cratered by 10 percentage points.
As a result, global IP has strongly outperformed new orders data recently - Figure 4. Looking forward from here, there is still pent up demand for (lower priced or electric) vehicles. But consumers are becoming more cautious about buying bigger ticket items and the potential for further gains in the global auto sector has diminished.
Bottom line is that we think the current depressed level of new orders will be echoed much more closely in IP momentum over the next few months.