After dramatically rising last April and keeping above 10% for the following months, our indicator of global equity return expectations (iEMR) has dropped back to sub-4% over November and sub-3% in December (see screenshot above). Conversely, our indicator of global growth expectations (iGDP) is on the up again in November after a pause in October, beating the 6.6% peak of Feb-2018 and equalling the heights reached in Apr-2011 (albeit still far behind the 10% achieved in Aug-2009).
What to make of this divergence?
The short answer is that on balance it might not be good news for equity returns in general and equities outside of North America in particular.
To begin with, transitions from high to low iEMR levels tend to steer prospective equity returns from relatively clear waters into muddier ones. Based on the past 11 years (see graph below, starting from the right hand side), an iEMR level above 10% suggests little risk of materially negative returns over the following months (12% chance of absolute returns below 0% over 4 months) and good odds of materially positive returns (43% chance of annualised returns above 20% over 4 months). The odds rapidly shift when below that 10% level, with 3x more downside risk and nearly 2x less upside risk.
The picture doesn’t change when factoring in a supportive iGDP level. There has been two periods over the past 15 years when iEMR was similarly low and iGDP simialrly high to current: from Jul-09 to Jan-10 and from Feb to May-11 (see graph below). Both periods were marked by fast deteriorating returns, and although QE2 did eventually save the day in 2010, it was not so in 2011 (with QE3 still 18 months away).
What does iEMR represent and what goes into its computation?
A previous note on global macro nowcasting explained how the Butterwire engine used Principal Components Analysis to produce 3 global macro indicators from a list of equity-relevant highly-traded instruments (see extract in table below, e.g. a -0.2 value between the $US and iGDP suggests that a weaker dollar tends to correlate with higher global growth expectations), and how these indicators reflected market expectations with respect to global growth (iGDP), monetary policy (iEMC = implied Emerging Markets Capital inflows) and price stability (iLCI = implied Late Cycle Indicator).
These 3 macro indicators (their level and momentum) are used to generate iEMR (using multi-linear regression). iEMR is therefore an entirely top-down view of future global equity returns, as implied by the pricing of some of the most liquid securities traded globally (i.e. currencies, commodities, credit, etc.).
What is the predictive power of these macro indicators?
We call these indicators “nowcasts” as they merely distil “spot” market expectations as opposed to making predictions. An ability to forecast what the iGDP will be one year from now would in itself be enough to make pretty good predictions, for instance of future absolute equity returns (see graph below -- covers the last 10 years), but again we do not make any such attempt.
In turn, the predictive power of today’s iGDP level on absolute equity returns 1-year out is inexistant (see left graph below), even over shorter time horizons (e.g. 4 months out, see right graph below).
But when we combine these macro indicators we manage to squeeze out useful signals/baselines, such as the one described above (i.e. much higher chances of material absolute positive returns when iEMR stands above 10%), or as the graph below shows, that iEMR tends to peak before a sell-off period (where it typically picks up a roll-over in economic growth expectations) and to trough before the end of a sell-off period (where it typically picks up a build-up of Fed intervention expectations, which the Fed has been dutifully obliging, thereby vindicating “buying the dips”).
Of course, the graph also shows that iEMR managed to signal 10 of the last 8 sell-off periods… getting it wrong over 2Q13-3Q13 (during which poor EM and strong DM markets decoupled, when iEMR assumes that a fast rising US yields and exchange rates tends to starve EM of $ funding before spreading to DM) and 1Q17 (lag effects of fast deteriorating macro conditions over 4Q16 and time required for the signal to fully reflect the impact of the swift and extensive interventions by the Fed and the PBoC over January 2017). iEMR is therefore no silver bullet, as is the case with most signals distilled by Butterwire, but it does serve its aim of helping users think their way through market noise and tilt their odds of making the right decisions at the right time.
Can iEMR be calculated at regional level?
The same variables can be used to derive expectations of excess regional returns (relative to the global – ACWI – index). As can be seen in the first graph below, the relative levels of iEMR for North America and Europe currently strongly favour the former over the latter going forward (ca. next 3 months). Likewise for North America over Japan (2nd graph) and over Emerging Markets (3rd graph) whose sharp iEMR deteriorations of recent weeks are driven by, respectively for Japan, commodities (e.g. higher copper prices) and credit (e.g. lower global financial stress index), and for Emerging Markets, currencies (e.g. lower CNY, higher EUR) and composite indices (e.g. underperforming EM banks/S&P LowVol and Belgian/Swiss equity index ratios).
How else can the real-time nature of these indicators be used to get extra insights?
“Spot” iGDP can be used to perform real-time analysis that would normally require months of waiting for the actual (real and nominal) GDP figures to be released. For instance, computing the US Wicksellian spread (Real US Baa Bond Yields minus Real US GDP, see graph below) using iGDP instead of (delayed) US GDP data helps systematically flag risks of recession/sell-off/inflation 3-5 months ahead of official data publications:
- The sweet spot lies between -1% and 4% spread
- A spread above 4% points to higher recession or depression risk
- A spread below -1% points to higher inflation/price instability risk 3-4 quarters later, which, given the very negative values observed since 3Q20 would suggest a pick-up in inflation in 2H21.
How can these top-down insights influence bottom-up stock research priorities?
The explorer feature enables stock screening that mixes stock macro, fundamental, technical, and qualitative criteria. A search for, say, 10 interesting (or high octane) research candidates that are consistent with the above insights is illustrated by the screenshot below, where region was set as Europe, size set as XL, L or M, flag as green or none, alert as none, base score as > 3, and global macro profile set as either inflation hedge, or inflationary EM growth, or inflationary EM growth.