February 1, 2000

Nest Eggheads

Taking money out of a retirement portfolio can be

After all these years of writing about everything related to the planning world, I've found that some subjects generate a lot more interest and feedback than others. The most feedback I get, interestingly, comes whenever I publish something on a somewhat esoteric topic: how best to liquidate investment portfolios during retirement. Indeed, the planning profession is still evolving ways to answer a deceptively simple question: What's the safest, most reasonable way for clients to take living expenses out of their retirement portfolios, and avoid having to change their diet to cat food at, say, age 75 or 80?

The discussion tends to take place on many levels. Let's start with the simplest. Here, advisors are mostly looking for ways to show their clients that the easiest approach is also, alas, the least effective. After years of reading Money magazine, people have been trained to think that they should, upon retirement, immediately convert their portfolios to income-producing securities, and then clip coupons for the rest of their lives. Leaving aside the hideous tax consequences of selling everything and probably paying commissions to buy something else at age 65, are bonds an effective way to preserve the real (after-inflation) value of the "corpus" of the portfolio?

Taking money out of a retirement portfolio can be trickier than putting it in. Here's the cutting-edge thinking about how much to withdraw, and when

The answer, most of us realize, is no. But how do you show that to a client? One illustration was offered to me by Bart Boyer, who practices in Asheville, North Carolina. Suppose, Boyer proposes, that Client A starts with $1 million in Treasury bills in 1970, spends all the income every year, and in 1999 has done a wonderful job of avoiding what we traditionally think of as losses; that is, he still has the $1 million in principal. Boyer's data show rather drastic changes in income as interest rates have bounced around, but the client could have purchased 30-year securities and smoothed out the income stream considerably. So, for the sake of argument, I'm going to suppose that the 6.5% rate that prevailed in 1970 determined the client's income for the next three decades. The client has been able to spend $65,000 (pretax) each year since then. He lived like a king in the 1970s, and earns a decent retirement income now. In all, he received $1,950,000 in income during the 30-year period, and is now, as the 30-year issues finally come due, buying new Treasuries, maybe in the 10-year range, which, at today's rates, will generate a little under $50,000 a year. He's taken a long-delayed hit to his income, but he's still far from eating Little Friskies.

Now consider Client B, the gambler. In 1970, he puts $1 million into the S&P 500, and spends 6% of the portfolio per year. His income bounces around a bit, dropping to just under $40,000 in 1975 after the bear market we all turn back to if we want to scare clients and small children. But as the market recovers, so does his income, and by the end of last year, 6% of the portfolio came to $390,000 a year, and the portfolio itself was worth a little less than $8 million. In all, Person B has received $3.7 million in income.

Investors of the future may not get quite this same bull market result, but the long-term argument in favor of equities is nearly as strong as this illustration makes it out to be. However, the alert reader is already thinking of some other issues that have shown up in various ways in my e-mail. The argument for equities, they point out, is actually much stronger for the person who starts out with a $500,000 investment portfolio and $30,000 a year in retirement income in 1970. Our $1 million investor then - the rough equivalent of a person who has $4 million in her retirement account today - would have been able to live comfortably on a 3% distribution in 1970, and really doesn't need - or, at least, not desperately - the growth that the equities portfolio provided. Not so the person who retired on $250,000 or $500,000. For those investors, keeping ahead of inflation is a little like keeping one step ahead of a pack of wolves; survival depends on it.

This suggests that the portfolio advice you give to clients might be different for people who have accumulated different amounts by their date of retirement. Perhaps the best discussion I've heard on the subject comes from Mike Martin, who practices in the Washington, D.C., area. He has worked out a way to talk to clients about their envy for the 20% a week returns that day traders are said to be scoring with Internet companies. (For the record, no matter how many breathless articles I read in the consumer press, I don't believe a word of it.)

Martin's clients tend to have enough put away that they can live on a projected 7% rate of return on their investments. So he projects a 7% return out into retirement, and shows clients that if they live to a ripe old age, they will not (under that assumption, at least) have to live on peanut butter in the foreseeable future.

Then he asks two questions that help control his clients' Internet envy. Ques-tion one: How would your life change if you could have an 11% annual return during the time of your retirement? Most of the time the answer is, I don't think my life would change. In that case, does it make sense to take meaningful risk to try to get from 7% to 11%?

Question number two is the flip side of question number one: Suppose you went through the next 10 years and experienced 0% real returns, the way stock investors did in the 1970s? The answer usually is that this would change clients' lives. Gifting to the children would have to stop. They wouldn't be able to take care of their grandchildren's college. Their standard of living would go down. The point is made.

Most clients fit somewhere between Boyer's reluctant equity investor and Martin's greedy geezer, so the two illustrations offer a way to help clients realize what their options are.

One step deeper into the complexity of this discussion, advisors are looking for how much they can assume a client will be able to liquidate from the portfolio in retirement. Perhaps the best series of studies on the subject, by El Cajon, California, planner Bill Bengen, suggests that a totally safe yearly distribution on a portfolio that is between 55% and 75% invested in equities is 4.08% of the initial portfolio, held constant throughout retirement. This is based on historical data, including the Depression, and uses quarterly returns, where you can run into four consecutive quarters with very different returns than the yearly numbers we've all grown accustomed to. For example, during the six months ending June 30, 1932, large-cap stocks lost 68% of their value - considerably above the worst year on the Ibbotson charts - i.e., -43% in 1931.

Bengen then looked at whether the numbers were more stable when you assumed that each year in retirement, you reduced the equity percentage of the portfolio by one percentage point; that is, if you started with 75% allocated to equities in your first year of retirement, then the second year you would be 74%, and so forth. The simple answer to a rather lengthy analysis is, no, even for clients with 30-year retirement horizons, these incremental reductions don't mean you can, with absolute safety, take out a higher percentage of your portfolio each year.

But, interestingly, the phasedown does slightly increase the odds that, if you take out a higher percentage of your portfolio each year, you'll get away with it without running out of money. If you set your distributions at 5% of the total value of your portfolio when you retire, Bengen calculates that, with the 1% a year phasedown of equity exposure, you'll have a 71% chance that the portfolio will last 30 years.

The big problem with equity exposure is that most of the time the historical numbers show you coming out just fine - indeed, much better holding stocks than any other investment. But the data includes some real bummers, including, as we've mentioned, the Great Depres-sion and 1974 to 1975 years. Planners are understandably reluctant to simply assume those awful return years away, and so every spending plan has to have built-in defenses against liquidating too much of the portfolio when stocks have dropped significantly in value.

The Bengen defense is either to self-annuitize the portfolio at the relatively low rate of 4.08%, or to systematically reduce stock ownership over time. But there are other ways to handle the same issue. One is to construct a portfolio invested aggressively in equities, but that maintains a cash reserve equal to two or three years' worth of client income. The idea is that, if the market goes into some awful nosedive, you can simply take another year of income out of the cash reserve, without having to sell out at a loss and invade the corpus of the investment portfolio. If the market slide continues another year, you take another year's distribution out of cash. This at least reduces the amount of liquidation that takes place during market bottoms. During good return years, you keep replenishing what you take out.

I haven't yet seen an analysis that looks at the opportunity costs of this approach, and relates these costs to the (presumably greater) chance that the client's income will hold up over 30 years. But a similar analysis, which is being conducted as you read this, is looking at whether it makes sense to buy an immediate annuity with a portion of the client's initial retirement assets, in order to place an iron-clad guarantee over some portion of the client's income. Once again, you have an opportunity cost - this time borne by the heirs, since their inheritance is reduced if the client dies before the annuity has paid back its value. Does it buy commensurate peace of mind for the retiree? How would you quantify something like that?

Boyer, meanwhile, proposes another, simpler defense against overliquidating client portfolios: Simply cut back on how much you take out of the portfolio during the lean years. He starts by assuming that a diversified equity portfolio will generate an average of 12% a year through returns that are going to be all over the lot. You can argue with this assumption, but hold that thought for a moment and look at the real point, which is how he self-annuitizes a client's portfolio. He proposes that the client spend 6% of the portfolio in the first year - in other words, exactly half of what he expects the average yearly long-term return to be. If the market is up and the portfolio has experienced positive returns, then the client will continue taking 6% out of a higher number. If the market is down, then the client simply takes the same dollar amount as she took last year. If there is a sustained downturn, then he adds another limit: The client will never spend more than 12% in any given year.

Boyer doesn't expect to have to invoke this second rule, however, since it would require more than a 50% drop in the market, which hasn't happened since the Great Depression. He adds that if there is such a significant downturn, then clients should be looking at ways to reduce their living expenses anyway.

How do the numbers look in real life?

Consider Client C, starting with a $1 million portfolio at her retirement in 1969, spending 6% a year, and receiving a return equal to the S&P 500's performance. In the first year, the return is positive, but not enough to offset the distribution from the portfolio, so the spending is frozen at the original $60,000. Over the next two years, the returns more than make up for the living expenses, and so the distributions grow to almost $72,000.

The mid-1970s' decline hits the portfolio hard. For the next 12 years, Client C has frozen spending at $72,000 a year; in 1975, the system comes perilously close to the dreaded 12% distribution. But then a series of generally positive years, turning really positive in the 1980s, takes the portfolio away from the edge of danger.

Over the 29 years, Client C never has a decrease in spending, and her income in 1999 would be about $368,000. The principal, which was originally $1 million, would now be $6 million. After weathering the worst decade financially since the Depression, she still has a six-fold increase in spending and in principal, and never a negative year in spending.

No, those are not inflation-adjusted numbers, and the decline in living standards is somewhat more severe than "no decline in spending" might indicate. But Boyer argues, plausibly, that there are no guarantees in life, and clients should be prepared to tighten their belts during downturn periods.

After reading through all this feedback, and reporting a lot of it back through to my readers, I created a wish list that would help all of us who are wrestling with these issues. Analyzing the opportunity cost of keeping various numbers of years of income in cash has been mentioned. Doing the Bill Bengen research using some kind of Monte Carlo analysis, which would liberate the conclusions from depending on the actual sequence of investment returns we have already experienced, is another. Suppose we have six consecutive years of ugly market results; what would happen to these self-liquidation systems then?

The last item on my wish list has already been answered, sort of. I was going to ask for a quarterly reporting system that puts each quarter's return numbers into a format that answers the client's real concerns: not how much she did or did not beat the market by, but whether she is still on course to be able to take the retirement distributions that she needs in retirement. Or, if she's already retired, something that makes clear whether a course correction is needed since the last statement. That way, the focus is forward, not backward - a change I think would make all the difference in the world.

I was sort of grumbling about this issue when I was made aware of a Web site called fplanauditors.com, which allows advisors to develop retirement projections online, and to create online performance statements that update the retirement projections. The site even incorporates a form of Monte Carlo technology, which illustrates the uncertainty of future numbers and invites clients to address those uncertainties by moderating spending if something awful happens to the stock market. This is a huge step in the right direction, and further steps would include updating how the distributions would look under one or more of the various systems, with probabilities assigned to the various future scenarios.

Right now, there appears to be more hard thinking about these portfolio liquidation issues than anything else we do in the planning profession. For those of you who think financial planning is stuck in an asset allocation, retirement planning rut, here's an example of 21st century thinking that could offer some real benefits to the consuming public.

Reprints Discuss this story
This is where the comments go.