Should you let your stocks ride?

From riskless to risky

For this column, assume three possible investments: 1) stocks, 2) bonds, 3) cash. The future returns on these assets will be modest because fixed-income yields are historically low. Stocks will gain an annualised 7 per cent, bonds 3 per cent, and cash 2 per cent.

Naively, one would expect a portfolio that invests equally in each asset to make 4 per cent annualised. That is indeed what would occur if the investments were riskless; if the assets were rebalanced at the end of each year; and if the portfolio carried no expenses. A $10,000 initial investment would grow to $10,400 after one year, would expand to $10,816 the following year, and so forth. Like clockwork.

Let’s add the first layer of reality. No investments are static. What’s more, stocks are riskier than bonds, and bonds riskier than cash. To reflect those facts, I created random 30-year annual sequences for each of the three assets. The 30-year averages remained at 7 per cent, 3 per cent, and 2 per cent, but the yearly performances fluctuated, strongly for stocks, moderately for bonds, and modestly for cash.

The diversification premium

Adding volatility affected the outcome. With yearly rebalancing, the risky portfolio’s annualised return improved to 4.20 per cent. The gain from the previous 4 per cent isn’t huge, but it does become meaningful in time. Over 30 years, an initial $10,000 investment into the riskless portfolio would appreciate to $32,434, while the same purchase of the risky, rebalanced portfolio would reach $34,367.

The benefit from diversification across risky assets varies, becoming larger as the correlations between the investments shrink and as volatility increases. In this case, the correlations between the randomly generated numbers were low and asset-class volatility was modelled at historical levels. I created another example that had higher correlations and lower volatilities. The return on the rebalanced portfolio dropped to 4.08 per cent.

As the exercise shows, there is no single “diversification premium.” That number has sometimes been estimated to be as high as 2 per cent, when using heroic assumptions about future returns, as well as the superiority of alternative investments. However, in these days of minimal yields, and with alternatives not what they once were (as well as largely unavailable to retail investors), the true gain from diversification looks to be far smaller: 10 to 20 basis points.

Expense reductions

On to the second reality, expenses. The effect of costs on both the riskless and the risky, rebalanced portfolios is easy to understand. Subtracting the year’s expenses from the gross return leaves the net return. Thus, if the three investments of stocks, bonds, and cash could only be owned through mutual funds, and all mutual funds charged 0.40 per cent per year, the annualised return for the riskless portfolio would be 3.60 per cent, while that of the rebalanced portfolio would be 3.80 per cent.

So far, the lessons have been straightforward. That investments don’t return the same amount each year is regrettable, as it prevents investors from placing all their monies into the single best future performer, but at least they can improve their results by diversifying and rebalancing. However, unless expenses are kept unusually low, they will consume the benefit that comes from diversification.

Removing the constraints

Now things get complicated. Rebalancing is not an unmitigated good. On the bright side, under normal conditions (meaning that the portfolio holdings are imperfectly correlated), the practice does improve returns, albeit but not dramatically. Less happily, rebalancing means the ongoing habit of trading the portfolio’s best-returning asset for its worst.

That tactic helps investors when asset-class returns gyrate, as generally occurs over short- to intermediate-term time periods, but it reduces returns over the long haul when one asset substantially outgains the others. Such is the case with my three-investment exercise, as stocks compensate for their extra risk by providing extra returns. Consequently, investors who let stocks ride, rather than selling them to rebalance, would have posted the highest portfolio returns.

An annualised 4.68 per cent over the 30-year period, to be exact. Those who divided their $10,000 among the three assets and then forgot about their portfolios would have outgained the ants who rebalanced diligently. At an annual 48 basis points, the incremental improvement that came from letting the portfolio ride, rather than rebalancing annually, was larger than the effect of expenses.

Since stocks enjoyed higher returns during their early days (with, admittedly, some unpleasant setbacks; for example, they dropped 15 per cent in Year 1), they gradually became a larger part of the portfolio, thereby magnifying the effect of their future performances. More begat more. By the end of the time period, stocks made up almost two thirds of the portfolio, bonds one fifth, and cash the final 15 per cent.

Counterarguments

There are, of course, major problems with letting stocks ride. One is that although my exercise neatly requires that investments match their expected returns for each 30-year period, the real world does not. On this planet, as opposed to a contrived spreadsheet, bonds or cash may beat stocks over the next three decades. (Cash has never accomplished the feat in the US, but bonds did from 1981 through 2011, although stocks also performed well through that period.)

Another rejoinder is that in not rebalancing, investors will also be permitting risk to increase, markedly. Possessing a 33 per cent stock position is quite different from holding double that amount! Presumably, those who were willing to accept such volatility would have done so deliberately, when making their initial investments, rather than accidentally wandering into the position.

In short, the idea of letting the portfolio ride reeks of hindsight bias, while also being unrealistic about investors’ risk tolerance. The spreadsheet presupposes that what did occur with US assets will, very broadly, continue to occur (with an adjustment to all assumptions for today’s lower yields). Historically, that has generally been the case, but this time could be different, in which case those who let stocks ride will be heading down a blind and dangerous alley.

Just wondering

For that reason, I positioned this column as a question, not an answer. The suggestion is perilous. I will say this, though: There is more reason to let stocks ride today than in the past. At the start of the '90s, both bonds and cash yielded about 8 per cent. There was good reason to believe that those assets might rival stocks’ returns over the ensuing decades.

Today, there is much less reason. If equities perform only adequately, never mind well, they will surely outgain fixed-income securities.