The traditional approach for managing the imbedded tax liability in retirement assets is fairly straightforward:

1) Access taxable accounts (such as a taxable savings account).

2) Access partially tax-deferred assets (e.g., stock held outside a qualified plan).

3) Access tax deferred accounts (an IRA, annuity or qualified retirement plan).

This approach is based on the premise that continued income tax deferral creates greater economic benefit for the individual.

But is the traditional approach always best? And, does the optimal tax distribution plan change if the client wants to maximize assets available to future generations?

Let’s take a hypothetical case. Assume a husband, age 65, and wife, age 63, have $25,000 in taxable investments, $250,000 in capital or equity investments (with a tax basis of $50,000), and $450,000 in an individual retirement plan. Both are retired, but the wife plans to work 2 more years for $25,000 a year. The couple seeks $75,000 of after-tax income (including Social Security) in year one of retirement, to increase each year by the rate of inflation. For ease of comparison, assume a constant 7% investment return for all asset classes. All current tax law rules, including scheduled tax rates changes and required minimum distributions for the IRA, apply. The target retirement income and Social Security benefits will grow by 2.8% a year.

The couple wants to maximize the assets left for their 2 children, and so is seeking the optimal income distribution strategy. The tables show 3 scenarios for the couple, where Social Security for the older spouse commences at 1) age 62; 2) age 67; and 3) age 70. The comparison focuses on the traditional distribution strategy described above versus a pro rata distribution strategy, where the couple receives proportionate distributions from all accounts, regardless of tax status.

Table 1, the age 62 analysis, assumes Social Security payments begin when the husband reaches age 62 (or 3 years before the start of the distribution modeling). Here, pro-rata withdrawals from currently taxable and tax-deferred assets result in a slightly longer asset duration. This shows that it is possible for an alternative withdrawal strategy to result in equal or greater asset longevity relative to that produced using the traditional approach.

Would the traditional strategy be optimal if the couple died prematurely? If both die within 20 years of retirement, the pro-rata withdrawal strategy maximizes the net after-tax amount for heirs. Given that this strategy is also slightly better for asset longevity than the traditional strategy, the financial planner should suggest a pro-rata withdrawal from all asset classes rather than the traditional distribution approach.

Note: Once the couple’s assets have been depleted, Social Security provides only 35% of the target income. Of course, it is not optimal to be left with an asset that only provides for 35% of expected retirement needs; we will look at ways to increase this percentage later.

Table 2 shows what happens if the start of Social Security payments is deferred until age 67. Here, the traditional distribution strategy is optimal.

The delay in start date also extends the duration of the assets, from approximately 19 years to 23 years. This is likely because the delay in benefits both increases the amount of Social Security benefit and reduces the income tax paid on those benefits because the wife continues to work (Social Security benefits are taxable over certain income thresholds). It is also possible that the delay forces use of assets that are temporarily subject to lower tax rates (i.e., the lower rates on capital gains which are in effect through 2009) to meet the required annual target income.

When considering assets available for heirs in event that both die within 20 years of retirement (or when the husband is age 85), the traditional distribution leaves the greatest after-tax amount for the couple’s heirs. However, if the couple dies within 15 years (when the husband is age 80), the pro-rata strategy is slightly preferable.

This makes the right choice for the couple less clear. If the goal is to maximize asset duration without regard to whether the approach will maximize after-tax assets for heirs, then the traditional strategy is optimal. However, if the couple is concerned about leaving the maximum amount for heirs and also about their own health, more modeling may be required to determine the best strategy.

Delaying the start of Social Security payments until age 67 also leaves the couple in a better position relative to their income target, once the assets are depleted.

Consider: When opting for early Social Security (commencing at age 62), their Social Security benefit only provides 35% of target income. By waiting 5 years, Social Security provides them 41% of target.

Note: The couple does not adversely impact their standard of living by tapping personal assets earlier. In fact, they extended the time the assets lasted, since the increased Social Security payments resulting from the delayed start more than made up for the early use of personal assets.

Table 3 presents the situation where Social Security benefits do not start until the older spouse is age 70. Here again, the best modeling result for asset longevity is achieved with the traditional distribution approach.

However, these results seem less advantageous than in Table 2, where Social Security benefits start at age 67. This reflects the fact that the assets only last until the husband is age 85 (or 20 years following start of retirement).

This model certainly compares favorably with the 19-year asset duration when Social Security begins at age 62, but unfavorably with the 23-plus years that the assets last when the traditional theory is used and Social Security commences at age 67.

However, when Social Security payments start at age 70, the couple is left with Social Security payments equaling 46% of target income after their assets run out. By comparison, when the start date is age 67, the remaining Social Security benefit is only 41% of target income. Simply put, the couple must choose between a 23-year payout and then living on 41% of target income, or a 20-year payout and then living on 46% of target income.

In the event the couple dies within 20 years of retirement (when the husband is age 85), no assets remain for distribution to heirs. Again, this compares unfavorably to the results when Social Security payments starts at age 67. In the event the couple dies within 15 years of retirement, the traditional approach is the best choice for maximizing after-tax funds to heirs.

Do the projected results for the couple’s heirs inform the choice as to the best distribution methodology in the age 70 Social Security situation?

Yes and no. If the couple has no health issues and reasonably expects that one or both will live beyond age 85, then the traditional strategy should yield the best results from both an asset duration and wealth transfer perspective. However, if the couple has health issues or a family history suggesting that neither will live long in retirement, then more modeling would be required to determine whether another strategy would yield a better result under premature death scenarios.

These income distribution models suggest that determining an appropriate retirement income approach can be fairly complicated.

For example, will the client be working in retirement, and what impact will those earnings have on Social Security benefits and the taxation of other income? What is the health situation and life expectancy of the client? What are the projected retirement costs (including the difficult yet critical projection about healthcare and drug costs)? Does the client value asset duration or wealth transfer more highly? What is the expected investment return of the client’s investments and how does that compare to the increase in Social Security benefits afforded by a delay in commencing those benefits? How comfortable is the client relying on an expected investment return given market volatility? Does the client have strong personal views about when to start Social Security payments?

A financial plan is vulnerable to the risk that the investment performance will under perform, and that the client will live longer than assumed. A financial planner may want to consider an annuity or annuity type payout to help manage the investment and longevity risks associated with distributing retirement assets.

In general, assuming the client has no unusual health issues, delaying the commencement of Social Security benefits past age 62, and locking in a guaranteed payout for life, will better position the client to manage the investment and longevity risks associated with retirement’s distribution phase. Combining delayed Social Security with annuity purchase can be a powerful planning combination.

The traditional retirement distribution approach, which maximizes tax deferral, is generally sound; however, in certain circumstances, a non-traditional distribution strategy may yield better after-tax results.

Depending upon whether the client’s goals are asset longevity, wealth transfer, or some balance between the two, the financial planner will need to consider which of the many possible distribution strategies best meet those needs. Ultimately, this analysis reinforces the notion that there is no one-size-fits-all solution. Individuals will need professional support to determine the optimal distribution plan to minimize tax liabilities and maximize their retirement assets for themselves and their heirs.

Robert Fishbein is vice president and corporate counsel in the tax department of Prudential Financial, Newark, N.J. His e-mail address is