Leading Indicators – As Buy/Sell Indicators for the Equities Markets

The OECD composite leading indicators for October were released late last week – they too seem to be confirming that the rate of the recovery has peaked:

The aggregate CLI for the OECD countries plus Brazil, China, India, Indonesia, the Russian Federation and South Africa continues to move higher just at a slower pace. (Note that the CLI’s are principally driven by changes in monthly reported industrial production.)

So to the concept that the leading indicators will signal when equities are scheduled for a correction (credit to Albert Edwards of SocGen via Zero Hedge). The inference from the analysis was that with the ECRI and US Conference Board starting to slow, they are flashing a warning signal for equities.

Okay, but then got to thinking about the difference between the growth rate peaking and an actual decline in the leading indicators. As Albert pointed out – the time to sell in Japan was when the leading indicators turned down. They haven’t turned down – they are still going up – just that the slope of their ascent has tapered off. Here is the SocGen leading indicator mapped against the Nikkei:

So to check the difference between the two measures following is the performance of the Nikkei mapped against the OECD’s composite leading indicator for Japan over the last 20 years both in absolute terms and as against the percentage change in the indicator:

Now I’ve only highlighted those 9 times when the CLI actually tipped into negative territory. As a simple measure of the statement “sell when the leading indicator turns down”, if you had done so you would have be in the money 5 out of 9 times three months later (for an average gain of 1613 points) and 7 out of 9 times six months later (for an average gain of 2131 points).

Using that same metric, if you had instead sold when the growth rate turned down, the result was 5/9 (+198 points) and 6/9 (+1479 points). (I assumed three consecutive months of declines to confirm the ‘turn’.)

Without trying to over engineer this, a general conclusion was that selling when the growth rate turned down tended to be early. While on average selling when the CLI went negative resulted in an entry ~640 points lower than when the growth rate turned (with some notably wild swings around the average), the former outperformed both in the 3 month and 6 month time frames.

Conclusion

Still like the theory, particularly given the view that a self-sustaining recovery is about as likely as Australia winning the world cup (or for that matter the US). Just that it might be wise to temper enthusiasm for the recent slowdown in growth as a sell signal.