Investors cannot rely upon betas. This is one lesson of the recent stock market fall.
Betas are supposed to measure a share’s sensitivity to moves in the general market. Low-beta stocks should outperform when the market falls while high-beta ones should underperform.
But this isn’t entirely what we’ve seen. Since mid-January miners and banks – which are generally agreed to be high-beta stocks – have actually done better than telecoms and tobacco stocks, which are widely regarded as having low betas.
This reminds us that we should be wary of betas (as indeed of all statistics). They are not hard, immutable facts but rather the changing result of changing economic conditions.
One reason for this is simply that over short periods shares’ betas can change by accident. If a fall in the general market happens to coincide with good news for a particular share, that stock would rise and it would appear that it has a negative beta – if you measure beta in a narrow enough window. That wouldn’t, however, mean the stock’s beta were reliably negative and that we should expect it to fall if the market rises. Equally, if bad news for a particular stock coincides with a fall in the market, it might appear as if the stock has a high beta when in fact this is just an accident: this might be true of tobacco stocks in the last month.
There are, though, more substantive reasons why betas might change.
One is simply that companies themselves change. In the late 1990s telecoms were regarded as growth stocks. They therefore had high betas because growth stocks are more sensitive than others to changes in investors’ sentiment. In recent years, though, they are considered to have gone 'ex growth' and have become utility-type stocks. So their betas have fallen.
A similar thing is true of pharmaceuticals. In the 1980s and 1990s these were high-beta growth stocks. In the 2000s they became more defensive ex-growth stocks. In recent months, though, their betas have increased as they’ve become more 'growthy'.
To take another example, mining stocks were seen as dull 'old economy' shares in the late 1990s and so had low betas. However, as talk of a commodity super-cycle spread in the 2000s the stocks became more sensitive to investors’ sentiment and so their betas increased. And except for a brief period in 2014 they have stayed high. Tech stocks have had the opposite pattern. They had high betas in the late 1990s and early 2000s because the general market rose and fell as sentiment towards tech stocks waxed and waned. Since, then, however, tech stocks’ betas have been lower.
Betas can also change because investors’ perceptions shift. For example, in the mid-2000s banks were regarded as relatively safe because investors believed talk of a 'Great Moderation' and overlooked banks’ high gearing. As the financial crisis forced investors to rethink these views, banks’ betas soared. But they then declined as banks’ leverage fell, which caused them to become less risky.
Betas can also change because stock markets can fall or rise for different reasons which affect stocks differently.
Much of the recent fall in equities came because investors feared that higher US interest rates would reduce global growth. This hit some shares exposed to the global economy, such as oil stocks. But it wasn’t so bad for more domestically-oriented shares such as retailers. And banks have taken a little comfort from hopes that higher US rates might allow them to increase profit margins.
Other plausible reasons why the market might fall would, however, produce different patterns. Worries about domestic growth, for example, would hit retailers but not overseas stocks such as oil producers: that was what we saw in the immediate aftermath of the vote to leave the EU. An increase in global risk aversion, however, would hit shares that are sensitive to investors’ sentiment (such as miners) but not so much traditional defensives. And if (or when) markets worry about a possible Labour government, utilities would become high-beta shares as investors fear Labour would nationalise them on harsh terms.
None of this is to say that the concept of beta is entirely meaningless. We can measure betas over particular periods. Doing so often yields the result that lower-beta stocks outperform higher-beta ones; this is why I favour defensive stocks.
What’s not so reasonable, though, is to assume blindly that past betas are an infallible guide to the future. They are not, especially if we measure the past or future in terms of short periods. We must always ask: defensive with respect to what? Very few stocks indeed are defensive with respect to everything.
