Don’t dump your dividends

Hindsight is the most useful skill in investment. That’s why in the aftermath of the collapse in dividends and the intractable problem this poses for income investors, we can see a growing number of fund managers telling us that the real mistake was investing for income in the first place. Rather than relying on the dividends or interest that investments naturally pay, investors should instead focus on generating capital gains that can regularly be sold to provide an income. 

The timing of this advice is, of course, not especially helpful. But the advice itself is also suspect. Yes, regularly harvesting gains is a perfectly workable way of producing an income. Indeed, since the average retiree will not build up enough savings that they can live solely on dividends and interest, even running down capital is likely to be part of most people’s plans for later in life. Yet the idea that most investors would do better to abandon dividends leaves me distinctly unimpressed, for three main reasons.

It’s all about the cash

First, the fundamental value of an asset depends on the cash that it will produce and return to investors. In the case of a stock, this will usually be through dividends (in the US, share buybacks are equally common, but i) buybacks are increasingly funded with debt rather than sustainable cash flows and ii) dividends remain much more important in the rest of the world). So the value of a stock and the value of future dividends are linked (see below). If we don’t expect companies to deliver healthy growth in dividends, we cannot expect the continued rise in share prices needed to create a sustainable income from capital gains. If we do, it undercuts the idea that dividend investing is dead.

Second, this crisis has been peculiarly favourable to capital growth. Share prices have rebounded quickly, while dividends and earnings have collapsed. That made growth-based strategies look very good by comparison – but this market is not normal (see right). In 2000-2003 and 2007-2009, dividends fell by much less than share prices. If you relied on capital gains in those markets, you had to slash the income you took, just as dividend investors have been hit this time. The alternative was to risk irreparable long-term damage to your finances.

Finally, both strategies need careful planning. Dividend investors should build a reserve to cope with a drop in income. Capital-gain investors must draw sustainably. But the latter is trickier to do well, mathematically and psychologically. The quirks of this unprecedented crisis don’t change the principle that relying on dividends is simpler – and probably safer – for most investors.

I wish I knew what the Gordon growth model was, but I’m too embarrassed to ask

The Gordon growth model is a simple but powerful way of valuing shares based on the dividends that the company is expected to pay in future. It gets its name from Myron Gordon, an economist who originally published the method in the 1950s, and is the most common and straightforward example of a class of valuation methods called dividend discount models. The theory behind these is that the value of a company is equal to the value of all the dividends that it will ever pay to investors, discounted back to their present value (so a dividend paid in five years’ time is worth less than a dividend paid tomorrow).

To apply the Gordon growth model, you need an estimate for next year’s dividend per share (D), the long-run expected stable growth rate of dividends (g), and the investor’s required return (r). The model says that the price (P) of a particular share should be next year’s expected dividend per share, divided by the investor’s required return less the long-term dividend growth rate. Expressed as a formula, this is P = D ÷ ( r − g ).

Let’s assume that a company will pay a dividend of 10p per share, the long-term growth rate is 4%, and the investor wants an 8% return. The value of the share to this investor is 10 ÷ ( 0.08 − 0.04 ) = 250p. 

In practice, we can almost never know what dividends a company will pay far into the future, but we can try out different scenarios to get a range of estimates for its fair value. We can also invert the model and work out what assumptions are needed to justify the current share price of a company and whether they seem reasonable.

The Gordon growth model will not work for stocks that do not currently pay any dividends or those where dividend growth is so high that g exceeds r. These problems can be solved by using a multi-stage model, where dividends start in the future, grow at higher rates for a finite period of time and then settle down to a steady, lower long-term rate.