Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Holy God Citi’s done a deep-dive on PMIs.
Everybody loves PMIs right? How could you not at least appreciate a major leading indicator that is so prone to doing weird and wonderful things as soon as the going gets tough. Plus, its idiosyncratic problems are fun!
[If you’ve somehow wandered this far into an article about PMIs and don’t know what PMIs are, they’re measures of business activity based on firms’ responses to question about things like orders received, price inflation, hiring conditions etc. They are a type of diffusion index, where a score above 50 indicates expansion compared with the previous month, and below 50 indicates a contraction. This causes endless problems for journalists, who often think something slowing down less means it is speeding up.]
Here are two colourful charts showing how this pattern of expansion and contraction has looked over recent years (we’re assuming this is the US PMIs given the comparator):
Citi analysts, led by Chris Montagu, write:
While PMIs are frequently regarded as a leading indicator for economic trends, we see weak causality between PMIs and equity market returns, instead markets may actually lead PMIs. Nevertheless, our long-run analysis shows causality relations are evident in a small subset of industries. Furthermore, we find PMI indicators effective when used within the context of economic state models, helping to not only explain markets trends but also to support tactical style allocations.
Their analysis (which focuses on manufacturing PMIs, but also takes some the parallel of services activity into accounts) draws some slightly fuzzy conclusions, as indicated above, but there are some interesting nuggets.
Citi’s analysis suggests that it is turning points in the PMIs — the peaks and troughs of the index — that seem to bear the most notable relation with sustained adjustments in market prices. The problem? How markets adjust appears to vary greatly by region, and then by sector within each region:
Market returns 12 months following PMI throughs have generally been positive, and the spread between Contraction and Recovery is considerably more pronounced and consistent for both Europe and US, when compared to PMI peak turning points.
A distinct characteristic of PMI peaks in the US is the tendency for markets to continue rising, albeit at a slower pace following the peak.
In contrast, Europe has experienced more pronounced and frequent market underperformance following PMI peaks.
So is it as simple as waiting for the inflection points and buying US equities? Citi did a lot of work on expected risk/return characteristics across different markets, and concluded… kinda yes but you’re a scrub if you try it:
Our historical observations suggest that selectively investing in factors based on different states of the PMI economic cycle may yield favorable outcomes. A naïve implementation of this would be to allocate/overweight styles that have historically been successful in a particular state, however much of the analysis so far relies on generalizations drawn from past observations made in hindsight, which may not translate to an effective real-world implementation.
The dangers of generalisations are pretty well demonstrated by the accompanying chart, which looks a bit like a Chelsea FC player were liquidated and dripped, Jackson Pollock-style, onto some axes for maximum messiness:
The fundamental conclusion, ultimately, isn’t super pretty:
There is limited evidence of causality effects between changes in manufacturing PMIs and market performance across regions, rather our results suggest the opposite, that there is stronger evidence that markets often lead PMIs.
But the analysts add:
That said, causal analysis is one approach to how investors can infer the relationship between markets and PMIs. We find it more revealing to view PMIs within the context of a state model which provides a natural economically intuitive partition between different economic states.
In this setting, we see strong evidence of diverging risk adjusted performance across PMI states. We find consistently strong risk-adjusted performance in PMI Expansionary phases. Emerging Markets and China exhibited periods on accelerated performance during PMI Recovery phases, but these growth spurts are relatively brief.
Comparing the sensitivities of equity markets to PMI changes, we find Emerging Markets the most sensitive to PMI changes, and that markets in Europe appear more sensitive to shifts in manufacturing PMIs when compared to the US.
Alphaville’s conclusion: it seems like PMIs are going to continue to be bothersome. What, were you expecting something simpler?
