Your Asset Allocation During Retirement

More Stocks, Please
Yesterday's column asked whether retirees should own more stocks over time. That is, perhaps mutual funds that are built for retirees should increase their equity allocations each year, thereby using a rising glide path, rather than follow either a flat glide path or the declining glide path used by most target-date funds. My source was "Reducing Retirement Risk with a Rising Equity Glidepath," a paper by Wade Pfau and Michael Kitces that evaluates income-withdrawal strategies across a variety of asset allocations and glide-path decisions.

The column evaluated two asset allocations, one with an average of 25% equities during retirement and the other with an average of 45% equities. I selected those two allocations because they typified the stock allocations found in the largest retirement-income funds. If the proposal is to change institutional investment habits, then the 25% and 45% portfolios are the appropriate ones to evaluate.

The findings did not fully support the authors' contention.

For the 25% portfolio, the traditional approach of reducing stock exposure over time looked better than the flat path, which in turn looked somewhat better than the proposal of increasing stock exposure. For the 45% portfolio, however, one could make the reverse argument. Not with great conviction, to be sure, as the results were similar for all three strategies, but there were advantages in some situations to adopting a rising glide path.

Today, we'll look at the results for higher stock allocations. In a sense, this analysis is mostly theoretical, as few fund companies are offering in-retirement funds that consist of mostly stocks. However, investors can certainly construct such a tactic on their own. In addition, the findings may stimulate fund companies to re-evaluate their current practices.

The results for a 55% average stock allocation, across each of the paper's three market-forecast scenarios (see yesterday's column for more details):

And for 75%:

Here are the takeaways:

1) At a 4% withdrawal rate, the results for different asset allocations and glide-path strategies are remarkable similar.

For the historical scenario, the worst showing was the traditional glide path at 75% stocks, which had a 90% success rate and a shortfall amount of seven years. The best was the rising glide path at 45% stocks, which had a 95% stock rate and no shortfall amount.

For the Evensky scenario, the worst showing was again the traditional glide path at 75% stocks, for a 69% success rate and a shortfall amount of 13 years. The best was, again, the rising glide path at 45% stocks, which had a 74% success rate and a shortfall amount of eight years.

Thus, varying the asset allocation by up to 50 percentage points, along with changing the shape of the glide path, meant the difference between failure and success in only one out of 20 simulations. I would not have guessed.

2) At a 5% withdrawal rate, a higher stock allocation is critical.

This one confirms common sense. Asking for 5% in real, inflation-adjusted money from a portfolio each year is asking a lot. There aren't risk-free securities offering such a return. Such a withdrawal amount is highly aggressive for those with time horizon of several decades (it perhaps may be defended as the starting point for a variable-withdrawal approach). Desperate times call for desperate measures--more stocks.

Indeed, for the conservative scenario, the highest success rate comes from a portfolio of 100% stocks. It starts at 100% stocks, finishes at 100% stocks, and carries 100% all along the way. Hey, that's an asset allocation, too. Of sorts.

3) At the lower withdrawal rate, the authors' suggestion of a rising glide path looks reasonable once stocks make up at least a large minority of assets.

It's neither a powerful effect nor one that carries across to the 5% withdrawal rate. Nor does it work with the 25% stock portfolio. However, for the 45%, 55%, and 75% portfolios, a rising glide path generally posts slightly better results than a declining glide path.

This is just the starting point of the analysis. Success rate and shortfall for the 95th-percentile simulation, as used in this column, are but two of many ways to judge a tactic's performance. (In their paper, the authors provide two more measures: median size of the asset pool after the conclusion of the 30-year test period, and the maximum sustainable withdrawal rate with a 90% probability of success.) Other assets could be modeled.

Nevertheless, it's a start. And that finding holds for all three sets of market forecasts.

4) There's certainly enough in the paper to suggest that current institutional practices should be re-examined.

Fund companies now structure the postretirement leg of their target-date funds by using declining or flat glide paths early in the retirement period, followed by flat glide paths later (that is, when retirement-income funds are used). Stock allocations are fairly low. The average stock allocation for the 10 leading retirement-income funds is 31.5%.

This paper suggests that fund companies should consider significantly higher stock percentages. Rising glide paths are also worth studying. At the very least, such paths don't seem to be worse than the alternatives for generating sustainable, moderate-level withdrawals--they might even be better.

John Rekenthaler has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own.