For most investors, goals-based investing is the appropriate way to manage their money (see “Why most investors need goals-based investing”). That’s because most investors must save and plan for life’s events, from college for the kids to vacations and retirement. These events present a contingent liability problem to investment professionals—not one of return maximization.
What’s the difference?
The truly wealthy have long and occasionally perpetual time horizons (thanks to trust, foundation and/or corporate structures), as well as sufficient wealth to address intergenerational needs. This patient wealth can afford to rebalance to a fixed strategic asset mix (e.g., 60% equities, 40% bonds), making up short-term losses over future cycles. Preserving capital and maximizing long-term returns are appropriate objectives for this group.
For everyone else, time is rarely a friend. If the stock market falls precipitously just before Sarah starts grad school or mere months before Kelly’s retirement, priorities may have to be revisited. Most investors should focus on achieving goals, a different objective than maximizing returns.
Why maximizing returns isn’t the most direct way to achieve goals
Consider the activity of delivering packages. If the objective is to take the shortest route (i.e., maximize returns), drivers might sacrifice time for a direct route shown by Google Maps. In contrast, UPS drivers’ delivery routes are programmed to minimize left-hand turns, which cause delays and increase collision risk. With 90% right-hand turns, UPS makes more deliveries while reducing fuel consumption and fleet size.
Advisors are trained to maximize risk-adjusted returns by rebalancing to the riskiest portfolio indicated by a client’s risk profile, regardless of market conditions. This approach might sound good to investors and be appropriate when time horizons are very long. But it may be hazardous when they’re not.
Most investment advisors build portfolios using modern portfolio theory (MPT). To maximize returns and minimize risk, MPT defines the relationship between mean returns and risk, or variance, to establish an optimal asset allocation. But MPT can’t solve for liabilities, and is vague about time. Most investors need to accumulate a defined amount of money to offset an obligation or goal by a defined date, so they need a strategy that considers both.
Another MPT shortcoming is its reliance on expected returns. The mean-variance model requires estimates for returns, variance and covariance (correlations). Variance and covariance change over time but are relatively stable compared to expected returns. As a result, small input errors for returns can lead to magnified output goofs.
The investment industry’s primary risk management tool is diversification. But professionals know that diversification fails (in other words, correlations rise) when protection is most needed: during major market downturns. Making losses back during the next cycle is the excuse given for rebalancing to a fixed mix. But, except for their long-term goals, investors can’t afford the negative impact of sharp short-term losses because they have to meet financial obligations spread over their entire lives.
Chart 1 illustrates the correlations between U.S. industries, divided into return quantiles from low to high. During periods of crisis (left) and rebounding from crisis (right), the risk mitigation properties of diversification are diminished as correlations rise. Higher correlations mean assets are moving in the same direction at the same time and providing less protection. Eliminating both tails of this series would improve a portfolio’s results from diversification.
Table 1 shows the compound annual growth of the same U.S. equity series with the best (highest-return) and worst (lowest-return) days removed, as well as the best and worst five days and 39 days removed. As more best and worst days are removed, the compound annual growth rate and total wealth increase. (Since there are more really bad days than really good days, returns are negatively skewed.)
Table 1: Returns with best and worst days removed U.S. stocks, 1973-2011
|Annual return||Difference||Total wealth if US$1 was invested 1973-2011|
|U.S. market index (Datastream)||9.82%||—||$38.61|
|Exclude worst and best days||10.10%||0.28%||$42.57|
|Exclude five worst and best days||10.23%||0.41%||$44.59|
|Exclude 39 worst and best days||10.90%||1.08%||$56.50|
Source: Woo Chang Kim, KAIST
This result complements the previous example that suggests reducing exposure to return extremes helps to avoid periods during which diversification benefits are diminished.
By eliminating periods of extreme uncertainty, both during up and down markets, more stable inputs for building goals-based portfolios can be achieved. To do this effectively, create a target risk level for portfolios, not a target asset mix. Do that, and you’ll reduce portfolio risk when market volatility increases, and increase portfolio risk when market volatility decreases.
Next time we’ll explore building portfolios to incorporate time and probability of loss. Advisors who build portfolios addressing time have a material advantage over the majority who can’t do that because they’re slaves to MPT’s shortcomings.