Larry McDonald commented on a paper called Dynamic Lifecycle Strategies for Target Date Retirement Funds, and of course Michael James hit the nail on the head with his response: if you guess too high on the expected market return, you will stay 100% in stocks until retirement, which is unwise.
A more sensible approach occurred to me while reading the paper. Suppose you are invested in a traditional Lifecycle fund, and then suddenly, the market takes a nosedive. You should move your money into another Lifecycle fund that is aimed at the target date at which you could expect to retire considering the market losses.
To explain what I mean, consider someone with $1 million invested, who wants $2 million in order to retire. Suppose they're in a lifecycle fund that unwinds linearly over 20 years, going from 100% stocks (at say 10% nominal return) to 100% fixed-income (at say 4% return). Suppose that he is contributing $20k per year. This person could expect to reach his $2M goal in 12 years.
Now, suppose the market crashes, losing half its value. Since our investor's target date was 12 years away, he was still 60% in stocks, so his portfolio is now worth $700k.
Going on the same assumptions as before, he would simply move his money into a fund with a 17-year target date. This move has the effect of increasing his stock exposure from 60% to 85%, which is almost as aggressive as the dynamic lifecycle strategy. If the market recovers, he can move his money back where it was.
This approach is much less sensitive to errors in the expected rate of return. If you overestimate returns by 1%, this scheme may have you making minor adjustments to your portfolio every so often as you find your actual returns diverging from your plan. In contrast, the dynamic lifecycle approach would have you 100% in stocks until the day you retire, which is clearly unwise.
It also seems more realistic to admit that your target retirement date has been impacted by the market crash and to adjust your plans accordingly, rather than try to go double-or-nothing with the market and risk your retirement entirely.
The statistical analysis of this strategy is left as an exercise for the reader.