Back to articles

What rate of investment return can you expect?

June 2015

Investors often look to long-term rates of return for an insight into the future. However, it is a practice which has its limitations given the complex interaction of economic, global and general market conditions.

A classic study written near the peak of the ‘90s bull market, and reprinted in this year’s Financial Analysts Journal (Volume 71), attempted to extract at least one factor from the equation: short-term changes in investor sentiment which drive valuations. For example, price-earning multiple expansion distorts both absolute returns and measured risk premiums (the minimum amount of money a risky-asset, such as equities, must exceed the known return on a risk-free asset, such as government bonds).

The study’s author, Peter Bernstein, used decades of past return and valuation level data to calculate the arithmetic mean real annual “basic” return (stripping out P/E expansion) was just 5.7 per cent a year for US stocks compared to the actual annualized return of 9.6 per cent. Similarly, the “basic” return for bonds was 2.7 per cent, well below the actual return of 4.6 per cent.

The study cautiously suggests that investors can still expect their equity headline return of 8-10 per cent if the future resembles the past and that an investor’s investment timeframe stretches much further than generally imagined.

“Recall what Keynes had to say about the long run,” Bernstein wrote. “Even for institutions such as charitable foundations and educational endowments that aim to be around into perpetuity, the time required to assure them of 10 per cent (or even the old-fashioned 8 percent) requires more endurance than most human trustees can manage. Indeed, this kind of long run will exceed the life expectancies of most people mature enough to be invited to join such boards of trustees.”

Given many investors view the long-term as 3-5 year periods, another key question is where short-term fluctuations have positioned the market relative to its long-term average?
On this point, Bernstein was also sceptical given rising P/E ratios have propelled significant gains over time. For example, in the 14 years since 1981, P/Es climbed from 8x to 18x while dividend yields shrunk from 5.4 percent to less than 2.5 percent.

“Super rates of return are the consequence of changes in valuations, not the level of valuation,” he says. “Just as secular earnings growth rates in double digits have almost zero probability, only a rising trend in valuation can sustain super rates of return over the long run.”

However, while his share market analysis showed a clear basic return and an average to which short-term results tended to regress, the situation was not so clear with bonds.
“There is no mean, real or nominal, to which it seems willing to regress,” he wrote.

He concludes that equity markets are not as risky as generally suggested, estimating the equity-risk premium over bonds was about 3 per cent (a figure which tends to match long-term real economic growth) but far lower than many other authors have estimated.

The full paper can be found at: