We looked last week at Simpson’s Paradox to explain why individual investors can outperform or underperform the investment funds that they own. This causes some real cognitive dissonance, because we expect statistical samples to be formed from randomly selected, equally weighted groups. In fact, investment funds cannot control the timing of inflows or outflows and studies have shown that retail investor fund flows are a contrary indicator, ie. higher deposits take place when the market level is high and likely to fall and higher withdrawals take place when the market value is low and likely to rise. It becomes apparent then the investment vehicle chosen is somewhat less important than ordinarily assumed in comparison to other aspects of investing which tend to be dismissed as unimportant.
The timing of deposits and withdrawals has a material effect on performance. Due to the resulting weighting of returns, deposits and withdrawals have the result of placing additional weight on certain market periods. Most professional investors agree that it’s next to impossible to time the market. Maybe this is the reason that they place little emphasis on the timing of fund flows. Besides this difficulty, mutual fund managers have no control over cash flows into and out of their fund. These typically poorly-timed retail fund flows may also explain the difficulty that mutual funds have in matching the performance of their benchmarks: their performance is fine, but the retail fund flows overweight “toppy” and falling markets and underweight “cheap” and recovering markets.
The sequence of returns, normally believed to have no impact during saving years, actually matters. In a presentation about the importance of safety for portfolios producing retirement income, the idea of the “sequence of returns” was described as follows. During the saving years, it was explained, the sequence of returns don’t matter because the average of: 2, 7, 15, -8, 20, 8 is the same as the average of: -8, 8, 20, 7, 15, 2. (This is true if we assume an unweighted average.) Withdrawals place far more importance on the sequence of returns, however. When a retiree begins taking withdrawals of 7% (let’s suppose), there is less money left to experience the recovery if early years are negative. The weighted average should be: (100% x -8 + 86% x 8 + 79% x 20 + 72% x 7 + 65% x 15 + 58% x 2)/6. (This is true.) However, what was ignored is that people make deposits during saving years. These deposits weight the returns during this period in the same way, especially when lump sum deposits are made. Suppose an extreme example of a young professional who wants to retire at age 35. He works 8 years, over the course of which he receives bonuses. In the early years, there is little income and next to no deposits due to mortgage payments and the cost of living for a young family. But in years 4, 5, 6 and 7, there are large deposits. Finally, in year 8, there is a very large deposit worth 50% of the account value due to taking a package. The weighted average may look like this: ($5k x 8% + $30k x 10% + $55k x -40%, + $80k x 41% + $150k x 15% + $500k x ??%)/6. These performance numbers are based on the TSX over the past five years. It should be clear why the final year’s performance would be far more important to this investor than any year up to now.
Most investors diversify their holdings in some way. But with diversification comes rebalancing. Besides the cognitive bias that causes investors to cut their winners and let their losses run, rebalancing is another act that impacts the weighted average of returns. An investor who owns 50% stocks and 50% bonds might rebalance annually. When moving funds from stocks to bonds, the investor reduces their exposure to the outperforming asset class (eg. stocks) and increases their exposure to the underperformer (eg. bonds). If and when performance reverts to the mean (eg. stocks crash), this will appear wise, because the poor performance will now affect a smaller proportion of the portfolio. However, if outperformance comes in runs of longer than a year (it has been suggested that four years is more realistic, but this isn’t consistent over time), then increasing the weighting of the underperforming asset class (eg. bonds) will reduce the final dollar-weighted return. The sequence of returns, then, also plays a role in final performance.
How much value can professional investment advice bring to this dilemma? It is a fact that no one knows which investment or asset class will perform best in the future. advisors are no better than individual investors at timing the market. What advisors do well, however, is talk investors out of selling at the bottom. An especially good advisor will caution an investor against overoptimism at the top. A good advisor should also encourage investors to make dollar-weighted returns work for them by recommending deposits soon after market crashes. While we can’t divine future performance, it’s relatively easy to recognize a market crash after-the-fact. Advisors aren’t all worth their fees, but the ones who help investors resist the impulses of fear and greed can add value that far outstrips their cost.
Opposed to conventional wisdom, the sequence of returns is as important, or more important, depending on the size of deposits, for savers as compared to income-seekers in determining the dollar-weighted (real life) returns of their investments. This affects not only the timing of deposits, but also the choice to rebalance between assets and asset classes. Without the ability to predict the future, investors may want to rely on a third party who can help avoid emotional errors, such as reducing risk or withdrawals after market crashes, and instead encourage the opposite behaviour.