Thursday, August 21, 2014

The Actuarial Approach vs. the "Bucket System"

This post welcomes yet another challenger to my claim that the Actuarial Approach is a better withdrawal strategy for you to use to develop a reasonable annual spending budget in retirement.  The challenger in this post is called the "Bucket System" and an example of its use is set forth in this Bankrate article.

Under the Bucket System, Jason Flurry proposes to establish three separate investment buckets (with increasing levels of investment risk) and a strategy for draining those buckets over time to provide retirement income to a hypothetical 55-year old couple with $2 million in accumulated assets.  His proposal would be expected to provide the couple with income of $60,000 per year for the first ten years of retirement.  In order to facilitate a comparison of the approaches, I am going to assume the following additional facts:  

  1. the couple has earned combined Social Security benefits of $4,000 per month payable at age 65 (before any increases in inflation),
  2. investment return per annum for the Bucket System equal to the percentages specified in the example (0% for Bucket 1, 4% for Bucket 2 and 6% for Bucket 3
  3. 5% per annum investment return for the Actuarial Approach
  4. 3% per annum inflation
  5. no other sources of retirement income, no desire to leave significant amounts to heirs, expected payout period under the Actuarial Approach of 40 years at initial retirement, and
  6. the couple expects to commence their Social Security benefits at age 65.  They expect the total annual amount of their Social Security benefits at that time (with 10 years of inflation at 3% per annum) will be $64,508 ($4,000 X 12 X 1.3439).
Under the Actuarial Approach, the couple uses the Social Security Bridge spreadsheet from this website, enters $2,000,000 in assets, expected Social Security of $64,508, number of years until commencement of Social Security: 10, zero to be left to heirs and the other recommended assumptions for the spreadsheet.  The resulting withdrawal for the first year is $103,340.  And, if the assumptions are exactly realized each year (and the couple spends exactly the total budget amount each year), each future year's total budget (Social Security plus withdrawals) until the couple reaches approximately age 91 is expected to remain at $103,340 per year in real (inflation-adjusted) dollars. 

By comparison, if all assumptions are realized under the "Bucket System" set forth in the article (and the couple spends the total budget amount each year--withdrawal plus Social Security), the total budget (in real dollar terms) is expected to increase significantly when the couple starts commencement of Social Security and again after 15 years when income from the third bucket kicks in.  The figure below shows a comparison of the expected total budget amounts under the two approaches from the couple's age 55 until they reach age 90.  I calculate that there will be in excess of $3.2 million in assets (including about 871,000 in Bucket 2) remaining at age 91 under the Bucket System vs. $749,489 at age 91 under the Actuarial Approach.  Note that this amount ($749,489) is shown in the run-out tab of the Social Security Bridge spreadsheet.  If you can't see the run-out tab, you need to maximize the spreadsheet window. 

Of course, actual future experience will not exactly follow assumed experience.  That is why it is important for a systematic withdrawal approach to automatically adjust for actual experience.  Under the Actuarial Approach this is easily accomplished:  As long as the previous year's budget amount increased by the increase in inflation for the previous year falls inside of a 10% corridor around the actuarial value, you stay with the previous year's value increased by actual inflation.  Automatic adjustment occurs under the Bucket System after the 15 year when the couple uses the IRS Required Minimum Distribution (RMD) rules to determine distributions.  As mentioned in previous posts, budget amounts can vary significantly from year to year under RMD approach as there is no smoothing of the actual experience.

The three major problems with the Bucket System (based on the facts of this example) from my perspective are: 1) it does not coordinate well with expected commencement of Social Security benefits, 2) it does not anticipate future inflation and 3) it uses the RMD rules to determine distributions from Bucket 3.  As a result of the first two problems, the matching of retirement income to the energy level of the hypothetical couple may not be optimal.  During the first ten years of their retirement, when they may wish to enjoy a more active life style including travel and adventure, their retirement income is significantly restricted relative to expected income in later years.  In addition, RMD is more likely to leave more money to heirs than desired.

Note that the Actuarial Approach produces an expected budget pattern under these assumptions that is constant in real dollar terms.  Some argue that retirees don't need constant real dollar income in retirement, but can accept some degree of decrease as they age, perhaps followed by an increased need much later in life.  This "smile" pattern can also be accomplished under the Actuarial Approach by inputting a lower percentage for desired increase in payments than the percentage anticipated for inflation and by inputting the expected increase in cost (for assisted living, etc.) as amounts desired to be left to heirs.  I believe it is better to factor these needs directly into the calculation than it is to use a method (such as RMD) that can simply result in unintended and undesirable deferral of spending. Of course, different spending patterns can also be achieved by consciously spending more or less than the spending budget produced under whatever approach is used.


 (click to enlarge)