Wednesday, May 16, 2018

Actuarial Budget Benchmark (ABB)—A Much More Robust Spending Algorithm

In our post of May 1, we talked about how the ABB differs from Monte Carlo models, and we sort of promised that we weren’t going to talk about this subject again for a while.  Subsequent to our post, our friend Dirk Cotton over at The Retirement Cafe also discussed this subject in his post of May 14, 2018.  Dirk did a very good job of explaining the different uses of these models (better than we did), so if you are still interested in this subject, we encourage you to read Dirk’s post.  We also recommend that you read it just because Dirk said some nice things about the ABB.  Because of our promise, however, we are going to keep this post very brief.  l

If you do read his post, you will note that Dirk refers to the ABB (and other Structured Withdrawal Plans) as “Spending Rules.”  Unlike Dirk (and others), we draw a distinction between withdrawal rules and spending algorithms.  The ABB is not a Structured Withdrawal Plan (SWP) designed to determine annual withdrawals from an investment portfolio after retirement, but rather is a variable spending algorithm that is used to develop a total annual recurring spending budget from all sources of income (either before or after retirement and either as a stand-alone spending budget calculator or as a spending algorithm in a Monte Carlo model).  As such, it is much more robust than a SWP.  Unlike SWPs, the ABB:

  • Considers all sources of assets and all spending liabilities (both recurring and non-recurring)
  • Considers expenses that may not last a lifetime
  • Considers sources of income that may not last a lifetime or may start or stop at different times
  • Can be used both before and after retirement
  • May be smoothed from year to year
  • Can be used in combination with other approaches
Under the ABB, withdrawals from accumulated savings for a particular year are equal to the calculated total annual recurring spending budget for the year minus income from other sources for the year.  This will also be true for a spending algorithm in a good Monte Carlo model but will generally not be the case when using a SWP.   Thus, the ABB coordinates income from all sources to develop a true spending budget rather than simply developing a way to “tap one’s savings” in retirement.