The Retirement Glass Is Half Full


Details, Details
When it comes to estimating retirement costs, the wind always blows conservatively. Asset-manager companies would prefer having more money than less; financial advisors desperately wish to avoid having their clients go broke; and plan sponsors providing 401(k) tools would rather see their employees overshoot the market than undershoot. That makes sense. Who would advocate better sorry than safe?

[Raises hand.]

Of course, I wouldn't phrase it that way. Rather, I would point out that the cautious approach is not without its faults. It can scare investors into avoiding the subject of retirement planning altogether. (The most common final screen for retirement-planning software is the gap-analysis page, when the user is shown the difference between what he or she is now doing and what is required. Upon seeing that, many exit the program, never to return.)

Also, conservatism generally implies simplified--and often unrealistic--models. One such example, which I've discussed in the past, is that retirees spend the same inflation-adjusted dollar amount each and every year, regardless of stock market behavior. It's possible, although peculiar, that somebody might wish for such an inflexible plan. More likely, though, the investor will be willing to accept a degree of flexibility in exchange for a lower target wealth goal at the time of retirement, as the withdrawal rate may be increased with flexible spending.

In a current working paper, "Estimating the True Cost of Retirement," Morningstar's David Blanchett addresses two additional items that may ease the burden of retirement savings: income replacement rate and the strategy of creating a "spending curve" for the retirement years rather than using constant real withdrawal rate. David didn't set out to deliver a happy pill--he's a spreadsheet junkie who goes where the numbers send him--but overall his message is a pleasant one.

Income-Replacement Ratio
Standard investment advice is that retirees should seek 70% to 80% of their pretax, preretirement income. That is, if a retired couple makes $80,000 combined before retirement, they would aim for $56,000 to $64,000 per year during retirement. Generally, this advice is given with the implication that the 80% mark (that is, $64,000 here) is the true goal, but, if the investor falls a bit short, that would be acceptable. Some even argue for higher amounts. David's paper cites a report by Aon Consulting that calls for income-replacement ratios of 78% to 94%.

David's research suggests that these numbers are high.

His analysis begins with the assumption that retirees will wish to maintain their preretirement level of take-home spending [my term, not his].

Take-home spending = Pretax income – taxes – preretirement expenses.

Retirees require less pretax income than workers to generate the same level of take-home income because, under normal conditions, retirees' taxes are significantly lower than they were before. They also need not pay certain expenses--most notably, savings for retirement, but also work-related costs such as transportation and clothes.

(It may be objected that, yes, retirees do have lower taxes and they eliminate some expenses, but they also face higher medical bills. David responds that as retirees' health declines, they tend to counterbalance their rising medical costs by reducing their spending on leisure items such as travel, entertainment, and restaurants.)

The required replacement ratio for fully matching the previous level of take-home spending varies by household, according to the tax situation and how many expenses are eliminated by the act of retirement. In general, higher-income households will have a somewhat lower required replacement ratio because they will benefit more from the tax change. The bigger effect, though, is that of expenses. Households that sharply cut their costs as of the retirement date need less replacement income than those that do not.

The short version of David's finding is that for medium to higher incomes, with a medium-level expense structure (for example, the household is contributing 6% of gross income into 401(k) and/or IRA plans, and it can save another 6% of gross income upon retirement), take-home income can be matched with roughly a 70% replacement ratio. That number climbs to about 75% for a lower-income household, as modeled in the paper $37,500 in collective wages.

Note that this range of 70% to 75% makes the conservative assumption that the retiree requires 100% of previous take-home monies. Relaxing that assumption, so that retirees change some of their lifestyle habits and accept lower take-home amounts, is outside the scope of David's paper, but it's a very reasonable consideration that would further trim the required replacement ratio. (My mother made such a choice three decades ago and was delighted with her decision. She regarded her first decade after retirement, when she was in her 50s, as the happiest decade of her life.)

Spending Curves
Although retirement-planning models typically use a constant real spending rate during retirement, investors don't necessarily behave that way. In the paper, David collects the average spending behavior of retirees over time, as taken from a household survey (RAND Health and Retirement Study). The findings are charted below. On average, retirees spend the most at the very beginning of their retirements. Spending then drops at an increasing rate through the mid 70s. Through the decade of the 80s, spending continues to fall (the line is rising, but it remains below zero); however, the rate of decline is not as steep as before because medical costs increase.

Applying the spending curves dramatically boosts the probability of success for withdrawal strategies. For example, David examines a 4% inflation-adjusted withdrawal rate, applied for 30 years. As has been widely reported, the 4% rate no longer looks like a sure thing when using realistic asset-class forecasts, rather than the higher market returns that were available in the past. Using Morningstar's forward-looking asset-class forecasts, David's model shows a 73.3% probability that the money will last the full 30 years. When a spending curve is substituted for the traditional flat withdrawal rate, the probability of success rises to 79.9% for a lower-income example, 86.0% at a middle income, and 91.1% for a relatively high income ($100,000 initial annual spending goal).

This article has mentioned four strategies for reducing the daunting, and sometimes seemingly impossible, accumulation goals that are shown to retirees.

1) Change lifestyle
2) Be flexible in annual spending
3) Find the appropriate income-replacement ratio
4) Adopt a spending curve

Those items have been covered separately so that their effects have not been combined into a single quantitative analysis. If they were, the results would be dramatic indeed. It's quite possible that with no great sacrifice, a household could finance a retirement that requires only a modest change in habits with an income-replacement ratio of less than 60%. (Indeed, for households that have unusually high retirement-savings rates and work-related costs, the third factor alone can do the trick.)

This isn't to minimize the importance of saving; a good retirement doesn't just happen. However, let's not maximize the importance of saving, either.

For another half-full argument, see this article by personal-finance veteran Scott Burns of Dallas News. It cites different research (from DFA) that approaches the topic from a different angle, but the general story remains the same: Things aren't quite as bad as they seem upon a closer look.

Finally, if you have thoughts on this column, or any other column, that you would like me to read, please email me directly at I check on the reader Comments section, but I do not catch them all. However, I do read every email.

John Rekenthaler has been researching the fund industry since 1988. He is now a columnist for 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.

View Comments