If you'd put £1,000 into the All-Share index in January 1986 and spent the dividends, you'd now have just under £2,000, adjusting for inflation but ignoring tax. If, however, you had reinvested the dividends you would now have £5,787. Dividends, therefore, have accounted for four-fifths of the total returns on equities since the mid-80s.
Even if you'd bought the index at its low-point in February 2009 and so enjoyed a big price rise, dividends would still have contributed a lot to returns. A £1,000 investment in the All-Share index back then would have grown to £1,905 now with dividends reinvested but to £1,553 if they had been spent. Even in a period of unusually big price gains, therefore, dividends still have contributed almost two-fifths of total returns.
If these numbers surprise you (and they did me), it's because of a very common mistake. "People underestimate the magnitude of compounding interest" says Matthew Levy at the London School of Economics. A yield of just over 3 per cent doesn't sound like much. But 3 per cent per year compounds very nicely over time to contribute the bulk of equity returns.
This might well remain the case over the longer-run. The least bad way to think about long-term returns goes something like this. Let's assume real GDP grows by around 2 per cent per year - its long-run average in the past - and that profits grow at the same rate and that the dividend pay-out ratio doesn't change. This implies growth in real dividends of two per cent per year. Let's also assume that the dividend yield doesn't change; this is reasonable given that, at 3.2 per cent, it is not far from its post-1986 average of 3.5 per cent. These assumptions imply that we'll get two percentage points of price appreciation per year and 3.2 percentage points of yield, giving a total return of 5.2 per cent. This is after inflation.
Dividends, then, are likely to contribute around three-fifths of annualised returns.
That word "annualised" is doing a lot of work because 3.2 per cent compounds better over time than 2 per cent. Over 20 years, these numbers imply that £1,000 will grow to £2,756 if we reinvest dividends but only £1,486 if we spend them. Dividends, then, will account for around three-quarters of total returns over the next 20 years.
Of course, this calculation rests on several assumptions. But I reckon that, if anything, I have understated dividends' likely contribution. One big danger is that the economy will grow by less than two percent in coming years. If so, shares' capital gains will be smaller and hence the contribution of dividends will be bigger.
The only way in which dividends wouldn't contribute a hefty part of returns would be if we get a big rise in equity valuations, with prices rising faster than dividends. Whilst this is possible, at least in the short-term, it would be rash for a longer-term investor to bet upon it.
There's something here which I'm not saying. None of this is a case for buying higher-yielding shares. Sure, these might be attractive because the market is undervaluing them or because a high yield is a good reward for their extra risk. But that's a different story entirely. My point is that for a long-term investor dividends matter enormously - probably more than you think. The simplest way to exploit this fact is simply to hold a tracker fund in which dividends are automatically reinvested.
