On volatility risk

There are two reasons why rising volatility might be bad for shares.

One is that a change in volatility is a sign of a change in the investment environment. That itself creates uncertainty which is bad for shares.

The other is that higher volatility in itself can cause shares to fall. One increasingly common strategy is that of managed volatility portfolios. The idea here is that investors should target a particular level of volatility. This means that if equity volatility rises, you sell shares and buy safer assets. Andrew Lo at MIT and Tyler Muir and Alan Moreira have shown that such strategies have in the past produced higher returns with less risk because – contrary to economic theory – higher volatility often leads to lower returns.

If enough investors use managed volatility their selling as volatility rises will more than offset buying by value investors, causing the market to fall.

Some shares, however, are more sensitive than others to rising volatility.

To see this, let’s consider those months in which the Vix index has risen by a significant amount (which I define as five percentage points). There have been 25 such months since 1990. In 19 of these 25 months the All-Share index has fallen, giving an average move in all 25 months of minus 1.5 per cent.

My table shows the average sensitivity of some FTSE sectors to changes in the Vix in these 25 months. You might think it unsurprising that most of these are negative. After all, if the market falls we’d expect most shares to do so.

Here, then, are two surprises.

One is that some sectors are statistically significantly sensitive to spikes in volatility even if we control for changes in the All-Share index. Miners, construction and travel stocks all tend to fall when volatility rises, over and above what we’d expect from their co-movement with the index alone.

Secondly, a few sectors have a positive, though small, correlation with the Vix in these months – despite the fact that they have positive betas with respect to the All-Share index and so should fall when the market falls. These include tobacco, utilities and non-life insurance.

There’s a pattern here. When volatility surges investors rotate out of shares they consider risky and into more defensive ones. The shares that suffer most from this tend to be cyclical ones such as miners and construction whilst the beneficiaries tend to be non-cyclical stocks. This is consistent with the tendency for volatility to often rise in recessions – although there’s no sign of such a recession now.

This tells us that volatility risk – the danger that volatility will rise – is different from market risk. Some shares carry volatility risk on top of market risk, while others do not, and in fact do well when volatility rises.

Here, though, lies a surprise. You don’t get compensation for taking on volatility risk.

If you did, you’d expect to see stocks exposed to volatility risk do well in months when volatility falls or does not rise much. Generally speaking, though, this is not the case. If we look at average returns since 1990 in all months when volatility fell or rose less than five percentage points, there’s zero correlation across the 28 main FTSE sectors between those returns and volatility risk (defined as exposure to rises of more than five percentage points in the Vix).

Yes, miners, which carry volatility risk, have done well in months when volatility has been well-behaved. But so have tobacco stocks, which don’t carry such risk. Overall, volatility risk is not priced.

For me, this is yet another reason why we should favour defensive stocks. Not only do these do better than they should over time and on average, but also they do so by not exposing us to volatility risk.

There is, however, a deeper implication here. The fact that volatility risk does not pay off tells us that higher risk does not necessarily mean higher returns even in those periods when the risk does not materialise. This is yet more evidence against the idea that higher risk means higher returns.