Thursday, May 23, 2013

Achieving a Higher Safe Withdrawal Rate with the Target Percentage Adjustment 
David M. Zolt (Journal of Financial Planning)

Nice article by Mr. Zolt, who is a financial planner and another member of the Society of Actuaries.

"A much higher initial withdrawal rate than previously thought possible can be achieved without increasing the probability of failure as long as the retiree reduces or eliminates the inflation increase for years indicated by the Target Percentage™.   The Target Percentage is developed and used to determine whether the portfolio is ahead of or behind target at any point during retirement. If the portfolio is ahead of target, the full inflation increase is taken in that year. If the portfolio is behind target, the inflation increase for that year is reduced or eliminated."

I like the approach suggested by Mr. Zolt because it is not as static ("set and forget" as defined by Wade Pfau) as the traditional safe withdrawal rate method.  Adjustments to withdrawals are made (as frequently as annually) to take into account "good" and "bad" years and to keep the spending plan from veering off the tracks.

Note that Mr. Zolt's approach (or something similar) can easily be accomplished using the suggested process and spreadsheet found on this website.  As an example, let's assume that a retiree would like to have a higher initial withdrawal rate and is comfortable with future increases of CPI minus 1% rather than full CPI increases.  Let's further assume that she believes the best estimate assumptions for future experience are 5% annual investment return, 3% per year inflation and a 30-year withdrawal period.  Also assume no annuity income and no bequest motive.    The retiree runs the New and Improved Spending Calculator on this site with her best estimate assumptions which determines an initial withdrawal rate of 4.34%.  She doesn't like that rate and determines that she can live with lower inflation protection (1% per year less), so she inputs 2% annual desired increases in the spreadsheet (but retains the 3% inflation assumption to measure the potential effect on future inflation-adjusted withdrawals).  This yields an initial withdrawal rate of 4.92%, which is much more to her liking (about 13% higher compared with Mr. Zolt's 10%).  She also looks at the inflation-adjusted runout tab on the spreadsheet and sees that if experience is exactly as assumed, her withdrawals will decrease in inflation adjusted dollars (by almost 25% in year 30).

As discussed in the original March, 2010 article in this website, our hypothetical retiree needs to employ an algorythm (rules) to adjust for actual experience and changes in assumptions and other input items each year.   She likes Mr. Zolt's basic approach so she decides that she will use the following rules to determine subsequent year's withdrawals:
  1. If the preliminary withdrawal rate falls inside the "corridor", she will increase her withdrawal amount for the previous year by CPI-1%
  2. If the preliminary withdrawal rate falls above the high end of the corridor, she will increase her withdrawal amount for the previous year by the full CPI.
  3. If the preliminary withdrawal rate falls below the low end of the corridor, she will withdraw the greater of i) the average of the preliminary withdrawal rate and the expected withdrawal rate or ii) the same dollar amount withdrawn for the previous year (i.e., no CPI increase).
For this purpose, the preliminary withdrawal rate is the rate produced by running the spreadsheet at the beginning of the year based on assumptions and new asset data as of that date (and presumably continuing with desired increases of CPI minus 1%), the expected withdrawal rate is the rate for year two shown in Column M of the previous year's run-out tab and the corridor could be something like 95% to 105% of the expected withdrawal rate.

Note that I am not necessarily advocating this approach.  I'm only illustrating that something similar to what Mr. Zolt suggests can be accomplished with the tools set forth in this website.

Mr. Zolt has graciously provided the following spreadsheet for those who would like to build their own target percentages. [Target_Percentage_Calc_2013_05_24.xls]