Choosing a risk premium

The number in question is the 'equity risk premium' - the expected excess return that investors get for putting their capital into companies' shares as opposed to the so-called riskless assets such as government bonds. It's vital because it's a key component of the discount rate - ie, the interest rate - that any investor uses to value any asset.

In financial markets, you can't have a value without having a discount rate. Say you are happy to buy a company's shares if they are rated on a PE ratio of 12 times, but not 14 times - where's the discount rate in that? Turn the earnings multiple upside down and you'll find it in the earnings yield. What you're really saying is that you're happy to buy if the earnings yield - the discount rate - is 8 per cent or so, but not if it's 7 per cent.

And, helpfully, this discount rate puts a number on the price that you should be willing to pay. If 8 per cent is the return - the interest rate - that you want, then for, say, 12p-worth of earnings you would pay 150p, but no more. Granted, a gung-ho investor who thinks the shares are a bargain on 14 times earnings would pay 170p. That's his decision. Maybe he knows something you don't; maybe he's a fool. It doesn't matter. What matters is that using a discount rate sets limits on what you will pay for an asset and brings discipline to your portfolio management.

This ubiquitous discount rate has two parts - the risk-free rate and the equity risk premium, which is the vital bit, especially as it's veiled in mystery.

True, it's no mystery what the risk premium has been. It's easily measured by taking total returns from equities and subtracting the returns from a chosen risk-free asset. That calculation tells us that depressingly often in the 21st century the risk premium has actually been a discount, as government bills or bonds performed better than equities.

Yet, over the long haul, the risk premium stands tall; indeed, strangely so. We can pick from a variety of data, but the ball-park figure is that equity markets in the UK and the US generated an average annual risk premium of between 5 and 6 per cent in the 20th century. Various explanations are offered as to why the 'ex-post' premium was so generous. More relevant is to ask whether the historical excess returns can tell us something about the future. Once, many experts assumed that somehow there was an immutable law that said the risk premium must be 6 per cent a year or so. That's no longer acceptable. However, rule-of-thumb ways of calculating the expected risk premium often indicate oddly generous returns from doing something as undemanding as buying a portfolio of equities.

There are variations to the calculation. But the basic way is to say that the total return from holding equities will equal the expected dividend yield over the coming year plus the long-term rate at which dividends will grow; that's sensible since it assumes that - short-term fluctuations aside - share prices can't rise faster than dividends (and, implicitly, profits) can grow. Then you deduct the risk-free return to find the risk premium.

The table shows the wide disparity in risk premia this method can produce. Sure, the forthcoming dividend yield is pretty well known, but the rate at which the payout can grow is much trickier and, for what it's worth, the risk-free return depends on which asset you choose. Assume optimistic rates and you might get a risk premium of 8.6 per cent (upper row of the table); assume cautious rates and you end up with a risk premium of just 2.2 per cent.

Which would you choose? Before you answer, remember the paradox of investment - that the higher your expected return, the lower the price you must pay. In other words, cautious investors such as Bearbull would choose the 8.6 per cent risk premium.

My concern in addressing this issue was that the discount rate I have used since the late 1990s - 8.5 per cent - to drive my valuation models is outdated by low interest rates, low risk-free returns and low equity risk premia. Responding by lowering my own discount rate would mean all sorts of company shares that previously looked expensive would suddenly pass muster. But do I want that? The success of the Bearbull Income Portfolio has been based on being fussy. Becoming less fussy would be an odd way to respond to deeply uncertain times. If the risk premium really is the most important number in investing, then better to err on the side of caution - high ball the risk premium and, in so doing, low ball the risks that your portfolio will face.