Against 'income stocks'

To see this, consider why any stock should have an above-average dividend yield.

One reason is that it is seen as unusually risky, so investors need a high income to compensate them for the extra risk. This is the case for banks and miners.

The other reason is that investors might believe the stock offers less growth than others, so they need a high income to offset lower future growth.

Here, though, lies a crucial distinction: sometimes, those expectations of weak growth aren't pessimistic enough - in which case the income stock is over-priced - but in other cases the expectations are too pessimistic and the income stock is a bargain.

This distinction is not random, because investors make systematic errors about predicting growth.

One error is that they underreact to bad news, so although prices fall, they don't fall enough. For example, in 2014 supermarkets offered decent yields but their prices subsequently fell. And in early 2015 miners offered fat yields, but their prices subsequently halved.

A very different error, though, is to under estimate the importance of what Warren Buffett calls "moats": barriers to entry such as high capital requirements or big brands which prevent rivals from stealing market share. Pharmaceuticals, utilities and tobacco stocks have offered decent yields and good returns because investors have under-priced "moats".

In this context, you must look for momentum. A high yield combined with negative momentum - a price fall in the last 12 months or a price below its 200-day moving average - is often dangerous. Otherwise, it can be attractive.

This is why I say "income stocks" describes very different assets. They include risky stocks; defensives; and ones with negative momentum. These have little in common other than being shares.

My table quantifies this for some currently high-yielding sectors.

The first point to note here is that these vary enormously in their risk. Banks, for example, have a beta of 1.36 while pharmaceuticals have one of just 0.5. And miners are almost three times as volatile as utilities: this implies that a fall of 10 per cent in a month is around a one-in-six chance for miners, but only around one in 130 for utilities. In this sense, we have very different assets - which makes it odd to give them the same name.

Correlations of relative returns corroborate this point. Now, absolute returns are all positively correlated simply because these stocks all tend to rise and fall (albeit at different rates) as the market rises and falls. If we look at returns relative to the All-Share index, however, we see low and often negative correlations. For example, utilities are likely to beat the market when miners under-perform (and vice versa), and banks are likely to out-perform when food retailers or pharmaceuticals under-perform (and vice versa).

The final row of the table is also important. This shows that defensive stocks have highish alphas - that is, returns controlling for market risk - while banks have a negative alpha, as do food retailers. These numbers are significant. A monthly alpha of 0.32 implies an annual price rise of 3.9 per cent even in a flat market.

These alphas are not merely a quirk of the past 10 years. They fit into a longstanding fact - that defensive stocks have tended to outperform while high-beta ones have underperformed.

Investing for income, therefore, can be a good idea if it leads you into defensive stocks. But it can be a bad one, if it puts you into over-priced high-beta or negative momentum stocks.

Luckily, lots of income funds do now have a defensive bias, with big weightings in pharmaceutical and tobacco stocks - though the latter actually now have a slightly lower yield than the All-Share index. The problem with discretionary fund management in a diverse asset class, however, is that this could change - albeit slowly because it's difficult to turn around big portfolios quickly.

Personally, I would rather we ditched talk of income stocks and income funds. The important difference is between defensives and less defensive stocks, and talk of 'income' overlooks this crucial distinction.