Saturday, April 2, 2016

You Can Do Better Than the IRS Required Minimum Distribution (RMD) Rules for Your Withdrawal Strategy

A number of financial experts tout the benefits of using the same IRS Publication 590 life expectancy factors required for determining required minimum distributions from qualified plans for the purpose of determining annual spending withdrawals from accumulated savings in retirement.  They argue that these factors are readily available, reflect remaining life expectancy, generally must be applied (for RMD purposes) in each year beginning with the year a retiree reaches age 70.5 and are conservative.

In his March 31 article, Henry K. “Bud” Hebeler sets forth seven essentials in retirement planning that do-it-yourself planners should not ignore.  I am in total agreement with the first 6 essentials and the first part of the seventh.  In this post, however, I’m going to take issue with the last few sentences of Bud’s article.

Even casual readers of this blog will know that I’m not a big fan of any withdrawal strategy that doesn’t properly coordinate with other sources of retirement income to produce a reasonable spending budget.  As I have discussed in prior posts, the RMD approach will generally not work very well for a retiree who has other fixed dollar sources of retirement income [pensions, immediate or deferred annuities] and who desires to have relatively constant dollar spending throughout retirement.  I did, however, give kudos to Henry “Bud” Hebeler in our March 26 post for developing an adjustment to apply to fixed dollar pension benefits/annuities so that he could add withdrawals determined the RMD approach and still develop a reasonable spending budget.  Notwithstanding the existence of Bud’s fixed dollar adjustment (which would be appropriate to use with any rule of thumb approach), I believe the RMD approach is just too conservative for most retirees. 

The table below shows rates of withdrawal by age for the RMD approach vs. the Actuarial Approach using recommended assumptions for discount rate, inflation rate and expected period of retirement.  Note that the withdrawal rates shown for the Actuarial Approach were developed by assuming no fixed dollar pensions/annuities exist and assuming no bequest motive to facilitate an “apples to apples” comparison of withdrawal rates. 

(click to enlarge)

The first column shows withdrawal rates by age under the RMD approach.  These rates are obtained by dividing 1 by the life expectancy from the IRS Publication 590 tables.   The second column shows withdrawal rates under the Actuarial Approach under our current recommended assumptions (4.5% discount rate, 2.5% inflation and period of retirement equal to 95-attained age or life expectancy if greater).  The withdrawal rates in the second column are designed to produce constant real-dollar withdrawals at least until about age 90 (after which time real-dollar withdrawals would be expected to decline somewhat from year to year) if the recommended assumptions are exactly realized.  As can be seen in the table above, the RMD withdrawal rates are consistently significantly lower than the rates using the Actuarial Approach.  The primary reason for this is that instead of assuming a 2% real discount rate (4.5% - 2.5%), the RMD approach effectively assumes a 0% real discount rate.   The net effect, then, of using RMD withdrawal rates would be to back-load real dollar spending to later years of retirement under the recommended assumptions.

The final two columns of this table, show withdrawal rates for males and females using the 2% real discount rate but instead of assuming death occurs at 95, these columns assume death occurs at the end of life expectancy determined using the 2012 SoA Individual Annuity Mortality Table with 1% mortality improvement (a link to which may be found in our Other Calculators and Tools section).  Also shown in these columns in parentheses are the remaining life expectancies for applicable ages. 

While Mr. Hebeler indicates that every year you live your life expectancy grows, this is technically not an accurate statement.  As can be seen from the table, your life expectancy decreases as you age, but your expected age at death (i.e., the sum of your age and your life expectancy) does increase.  This is probably what Bud intended to say.  Using your life expectancy as your expected period of retirement produces higher withdrawal rates than the age 95 approach, all things being equal, but as you age this results in declining real dollar withdrawals if all other assumptions are realized.  Therefore, unless you consciously want to front-load real dollar spending earlier in your retirement, we recommend planning to live until age 95.  We also recommend that you avoid using the RMD approach (with possible necessary adjustments of fixed dollar pensions, etc.) and just stick with the Actuarial Approach.