Are stocks really a better bet than bonds?

"Do stocks always outperform in the long run?" asks the Psy-Fi blog. "No, they don't." Indeed, the idea that stocks always outperform other investments may be "just another market myth used to cajole unwary investors into parting with their hard-earned cash". So are stocks really only for mugs?

The shares sales pitch

The shares are better than bonds' story runs like this. Stocks are riskier than bonds (IOUs usually carrying a fixed return). Firms can cut dividend payments more easily than they can slash interest payments. That makes your income stream sensitive to short-term changes in a firm's profits and cash flow. And, if it goes bust, shareholders are further down the repayment queue. However, in return for the risk you can expect to be rewarded with a generous equity risk premium (ERP) an extra return. But the size of this ERP is a matter of huge debate. And so are the returns.

Reality bites back

In a 1985 paper, MIT professor Rajnish Mehra and Nobel prize-winning economist Edward Prescott noted that between 1889 and 1978 "the real (inflation-adjusted) yield on the Standard & Poor 500 index was 7%, while the average yield on short-term debt was less than 1%". The 6% gap is the ERP. Using these numbers, £10,000 invested in equities would be worth around £39,000 after 20 years earning 7% and just £12,000 at 1%. That's a gap worth the risk. But what if the ERP isn't anything like that large? Over the last decade, stock returns were miserable below that from US Treasuries. Have investors been kidding themselves about stock returns?

Subscribe to MoneyWeek

Subscribe to MoneyWeek today and get your first six magazine issues absolutely FREE

Get 6 issues free

Sign up to Money Morning

Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter

Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter

Sign up

There's a number for all seasons

The first problem with the ERP is that there are several ways to calculate it, each giving a different answer. The 6% quoted above is based on historical analysis. But so too is the 3% UK ERP, quoted by Barclays Capital's latest Equity Gilts Study of 1900 to 2011. Between 1926 to 2010 they calculate the US EPR as 4.54%. Such differences matter. Invest £10,000 over 20 years at 4% (1% over the Barclays UK ERP of 3%) and you end up with £21,911. At 1%, plus the Barclays US ERP of 4.54%, it's £29,400. But it's not just about picking the right historic time. As most studies use horizons of 20 years or more, investors get a rose-tinted view of returns because they aren't worrying about individual stock-level bankruptcy risk. For individual stocks, the risk/return trade-off may be very different.

The many faces of the ERP

And there's another problem. "The ERP is a concept that seems to mean different things to different people," notes Yale Professor Roger Ibbotson. Looking backwards is fine for a view of what investors have demanded for investing in stocks in the past. But the ERP is meant to help an investor make decisions now. So shouldn't you take a "consensus estimate of the opinions of all the participants in the marketplace"? They set current stock prices, not their historic counterparts.

"But there are a number of problems with this approach," says Ibbotson. Most investors aren't concerned with the long-term and "have very short-term horizons". A study by Anderson School of Management's Shlomo Benartzi and Richard Thaler of the University of Chicago found higher ERP estimates arise when investors check their portfolios more than once a year and sell due to temporary losses. So high ERPs are caused, says the Psy-Fi blog, by "investors narrowly framing their investment time horizons". Then there's "disappointment aversion" the tendency of investors to leave the market when their expectations aren't met (driving up the ERP). So using investor expectations to determine the ERP is also unreliable.

Option three, says Ibbotson, is to consider the "demand side of the equation" and estimate how much extra return an investor would demand for the risk of buying stocks. This takes a risk-free rate and adds a bit on for stock-specific risks (the capital asset-pricing model is a key tool here see page 45). But results usually fail "to match the historical data". Finally, the "supply side of the equation" considers "what the economy and corporations supply to the market in... earnings and cash flow".

So that's four methods of arriving at an ERP, which all too often give different answers. Considering that bonds have beaten shares hands down recently, as Morningstar's Peng Chen notes in his paper Will Bonds Outperform Stocks Over The Long Run, any investor might wonder, "why bother with stocks"? Luckily, there's one good reason.

The good news

"How likely are stocks to outperform bonds in the future?" asks Peng. He notes that, over the last 40 years, most of the annual return on bonds came from income. Stocks yielded around 3.2% in income with double that in capital gains. For bonds to repeat this, "their yields would have to move into negative territory", suggesting investors will pay issuers to hold their cash. Possible, but unlikely. Going forward, yields are likely to increase rather than decrease and the outlook for bond capital gains is poor, given that yields are at historic lows.

Combining the US long-term earnings growth rate of 4.65% with current dividend yields suggests an equity return of around 7%, compared to bonds at 3-4%. Whether the ERP gap of 3-4% is enough to entice you into equities is another matter. But "it's unlikely that stocks will underperform bonds over the next 40 years".