Are your dividends under threat?

Chart: Number of 10%+ dividend cuts per Merton model quintile

One of the first lessons that any income investor learns is that a high dividend yield is not enough of a reason to buy a share. Exceptionally high yields can often be a sign of distress – a warning that the market expects the dividend to be cut. So it’s equally important to pay attention to whether the dividend is likely to be sustainable.

The usual way that investors are taught to look at this is a metric called dividend cover. This is the company’s earning per share (EPS) divided by the dividend per share (DPS) – so if EPS is 50p and DPS is 25p, the dividend cover is 50÷25=2. Some investors prefer to invert the formula (making it DPS÷EPS), in which case it’s known as the payout ratio. But both amount to the same thing (dividend cover of two is equal to a payout ratio of 50%).

The greater the gap between the dividend and earnings, the more scope the company should have to maintain the dividend if earnings fall. There’s a widely accepted rule of thumb that dividend cover of two or more is fairly safe, while you should start getting nervous if it’s less than 1.5. However, certain sectors with exceptionally stable earnings – such as utilities – are assumed to be safe with much lower levels of cover.

Does dividend cover work?

Dividend cover is a very logical idea. Yet the theory that it’s a good predictor of dividend cuts doesn’t seem to have been rigorously tested. There are some signs that payout ratios may be loosely related to returns, according to James O’Shaughnessy in What Works on Wall Street (a mammoth tome that analyses the track record of hundreds of investment strategies in the US market). He found that the 10% of stocks with the highest payout ratios (lowest dividend cover) underperformed the market on average over the period from 1964 to 2009.

He also finds that the 20% of stocks with the lowest payout ratios also underperformed, while those with middle-ranking payout ratios did best. However, it’s not clear that this relationship is very strong or persistent: I tested the same strategy on the FTSE 350 using the period from 2002 to 2014 and the results were weak.

More importantly, these results provide no real insight into whether the payout ratio can indicate whether any given firm’s dividend is under immediate threat. Indeed, research by analysts at Societe Generale following the global financial crisis – when dividend cuts were common – suggests it’s probably not that useful for this purpose.

They found that in 2009 around 50% of stocks that dropped their dividend altogether were in the top 40% of the market on free cashflow cover. (Free cash flow uses the amount of cash a company takes in after allowing for investment – it’s a more conservative figure than reported income, so you’d expect it to be a stronger indicator of safety.) Around a third of stocks that decreased their dividend were in the top 40%. The same pattern applied across the UK and the US, and slightly more weakly in Europe and Japan.

Watch the balance sheet

So if dividend cover doesn’t work that well, what might be more effective? One obvious candidate is the state of a company’s balance sheet (in other words, how much debt or cash it has). That’s because most firms don’t need to fund their dividends solely out of current earnings but may be able to draw down their cash reserves to keep paying when earnings go through a soft patch. Given that management are very aware that investors dislike dividend cuts, they will usually try to do this if possible.

Societe Generale’s research supports this idea. They use a fairly complicated measure of balance sheet strength called the Merton model or distance-to-default. Due to lack of space, I’m not going to try to explain exactly how this is calculated, but it works on the basis that shareholders have an option on what’s left over if all the creditors are paid off. The stronger the firm’s finances, the more that option is worth; thus firms where the dividend is weakest should have the lowest score.

The analysts looked at the relationship between significant dividend cuts (defined as a dividend of 10% or more lower than previously expected) and the results of the Merton model. You can see the results in the chart above: to summarise, from 2002 to date, 75% of dividend cuts in their database came from firms that fell in the lowest or second-lowest group under the Merton model. That implies it’s useful, though not infallible.

Unfortunately I’m not aware of a site where private investors can check how a stock scores on this metric, but the red flag is a firm with a high debt load whose shares have recently been volatile. So if you’re hunting for income bargains, it may be worth bearing that in mind.

Below, we look at how ten popular stocks score – although it is important to bear in mind that low scorers are by no means certain to cut and those with high scores are not guaranteed to be safe.

Popular stocks scored

The table below shows how ten large UK income stocks score on the Merton model. All data is from Societe Generale and is as of 1 April.

Company Dividend yield Merton quintile
High dividend risk
J Sainsbury 4.10% 1
Glencore 4.30% 1
Anglo American 5.80% 1
Centrica 5.10% 2
BP 6.30% 2
Low dividend risk
AstraZeneca 4.10% 4
National Grid 5.10% 4
GlaxoSmithKline 5.20% 4
British American Tobacco 4.50% 5
Next 4.30% 5