Many of you like stocks with high dividends. But many of these carry a big danger. Back in 2007-08 housebuilders and mortgage lenders such as Northern Rock and Bradford & Bingley were on great yields. But they subsequently collapsed, with investors in some cases being wiped out completely. That episode warns us that a high yield can be a sign of high risk – in particular the danger that a stock might do horribly badly in recession. Good returns on value stocks in normal times are often just a reward for taking on this cyclical risk.
Which poses the question: can we predict when these risks will materialise and when not, so we can have some ideas as to when to be a value investor and when not?
You might think the answer is: no. Prakash Loungani, an economist at the International Monetary Fund (IMF), has shown that economists are terrible at forecasting recessions, which means they cannot tell when cyclical risk will materialise.
Such pessimism, however, might be unwarranted. We have some evidence to suggest that we can at least partially predict returns on value stocks.
To see this, I looked at annual changes in my no-thought value portfolio (which has comprised the 20 highest-yielding stocks with a market capitalisation of over £500m, rebalanced every quarter), and asked: what factors can predict these changes?
Of course, none of you own this actual portfolio. But I suspect that it serves as a guide to what happens to deep-value stocks generally – those on exceptionally high yields – so what follows might be useful in spotting the performance of these generally.
My chart summarises how a handful of such factors have predicted the performance of deep-value stocks relative to the FTSE 350. Except for under-predicting value’s performance in 2016, the predictive power has been pretty good.
The important factors here are:
- The dividend yield on the FTSE 250. The higher this is, the better value stocks do in the following 12 months. The 250’s yield is a measure of sentiment towards cyclical stocks: a high yield means investors are avoiding cyclical risk. This means that stocks that offer such risk – such as deep-value ones – are underpriced and so will subsequently do well, other things being equal.
- The yield on Aim stocks. When this is high, value subsequently underperforms. This is because a high Aim yield betokens low sentiment towards speculative growth stocks. As such sentiment recovers, these stocks do well. But this means other stocks must underperform – and these others include value stocks.
- Three-month interest rates. The higher these are, the worse value stocks subsequently do.
- The gilt yield curve. An inverted curve (with two-year yields above 10-year ones) leads to value stocks doing well, other things equal. This suggests that value stocks price in the bad news about future growth which is contained in an inverted curve, and so subsequently do well.
- Sterling. A strong pound (on its trade-weighted index) leads to value stocks doing badly.
- Oil and shipping costs. The higher these are, the worse value stocks subsequently do. This could be because both predict slower economic growth, which is when value stocks do badly.
In fact, shipping costs are a remarkably good predictor of value stocks. This isn’t just because they were high in 2007-08 just before value slumped and low in 2008-09 before value recovered. Low levels of the Baltic Dry index in 2013 and 2015 led to value stocks doing well, and higher levels in 2010 and 2017 led to them doing badly.
Right now, these factors point to value stocks doing well over the next 12 months. This is largely because the yield on the FTSE 250 is above average, which signals that sentiment towards cyclical stocks is low, which suggests such stocks are cheap. The economic damage of Brexit might by now be fully priced in, unless we get a messy no-deal exit. Low interest rates, shipping costs and oil prices are also helpful. These bullish signs are, however, mitigated by the fact that the Aim yield is above average, which predicts growth stocks doing well at the expense of value.
All this said, there are three caveats here. One is that all this applies only to bundles of value stocks, not to each specific one. A high yield is sometimes a sign not just of especial systemic risk such as exposure to the danger of recession, but of company-specific dangers.
Secondly, some stocks are on high yields because of political risk – the danger that a Labour government might nationalise utilities or regulate them more severely. Our lead indicators are silent on whether these risks will increase or recede.
Thirdly, there’s a statistical problem. So far we have only in-sample evidence that these predictors work. The prediction that value will outperform should perhaps be seen as a first out-of-sample test of the hypothesis that value stocks are predictable. I can’t be certain that they are. But I am certain the question deserves more attention that it has usually had.
