It's annoying that two indicators as closely related as the dividend yield and PE ratio should produce such contradictory findings. But the key question is: which is likely to be the better harbinger, the cheap PE ratio or the unexceptional dividend yield?
Obviously, we're not talking about any old PE ratio, but a cyclically-adjusted version, which, supposedly, gives this ubiquitous tool special powers. The idea of smoothing 'E' to allow for the business cycle has been popularised by a Yale University academic, Robert Shiller, who is best known for his best-seller, Irrational Exuberance, whose subject matter is in its title and whose publication in 2000 coincided beautifully with the deflating of the technology bubble.
Cyclically adjusted PE ratio
Professor Shiller smoothes his earnings for the US market by taking the average of 'E' for the previous 10 years. You can download a spreadsheet of his data at http://www.econ.yale.edu/~shiller/data.htm. He also has the advantage of data going back to 1871. Bearbull's database for the UK only goes back to 1977, which is partly why I have smoothed earnings for the London market's cyclically-adjusted PE only over the previous five years. But clearly we could debate what is the best period for encapsulating a business cycle (10 years, too many? five years, too few?).
Equally clearly, a cyclically-adjusted PE is not comparable with a PE based on current or forecast earnings. The smoothed 'E' will usually be a small value compared with the current 'E', so the PE values it generates will tend to be high. But what's important is the shape of the chart and the current value of the adjusted PE ratio compared with its average value.
On that basis, the PE chart looks right. Ratings spiked in 1987 in the run-up to the October crash, then plummeted. They went on a long upwards march in the 1990s before subsiding as the technology bubble deflated. Last, the aftermath of the credit crunch drove ratings downwards as 'P' fell much faster than the smoothed 'E'. That might imply that 'E' has yet to fully reflect depressed company profits since 2008; in which case, 'E' might still be trending downwards so the cyclically-adjusted PE could rise without any change in share prices.
London's real dividend yield
That's another way of saying the market isn't as cheap as the chart suggests. Even so, when the adjustment has worked its way through, the smoothed PE will remain far below its average value for the previous 35 years and may still be around one standard deviation below the average. In other words, the smoothed PE ratio will continue to say that London share prices are cheap.
And it will still disagree with what the dividend yield chart tells us. The data for this chart, which only goes back to 1989, has been adjusted for the eroding effects of inflation (not that it makes much difference to the message). Its shape is also 'right' - spiking when the market was cheap (around 1990 and 2008) and hitting lows when it was dear (around 2000). Then the rally of 2009-11 took the London market from record cheap to expensive as it rebounded by about 67 per cent in quick time.
Since then, yields have risen, thanks slightly more to rising dividends than falling share prices. Even so, it's possible that share prices are cheaper than real yields suggest because cautious company bosses are distributing less in dividends than they could. Alternatively, record-low bond yields may be keeping equity yields lower - and therefore share prices higher - than they might be otherwise. But that notion does not really square with an especially low cyclical PE ratio.
But, if the signal is fuzzy, at least the message for investors should be clear. Which indicator is the better harbinger? Sensible investors will always choose the one that keeps them clear of trouble. Currently, real dividend yields speak louder than cyclically-adjusted PEs.
