Tuesday, September 10, 2013

Life Expected Income Breakeven Comparison Between SPIAs and Managed Portfolios

by Larry Frank Sr., John B. Mitchell and Wade Pfau 
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2317857 

A tip of my hat to Messrs. Frank Sr., Mitchell and Pfau for publishing this fine (and very thorough) paper.  The authors use a sophisticated Monte Carlo simulation approach to conclude that many retirees may find it financially beneficial to delay purchase of a single premium life annuity until a later age and self-manage their retirement assets until such age.

"The paper provides insight and guidance for the retiree decision making between whether to annuitize or manage their retirement savings."  While the authors' analysis examined this decision on an "either or" basis, it would be interesting to see their analysis for partial SPIA annuitization/partial self-management strategies (which could affect investment allocation decisions) or strategies that involve purchase of single premium deferred annuities (sometimes referred to as longevity annuities).

If you have read even a few of my prior blog posts, you will know that I am not a big fan of "set and forget" Safe Withdrawal Rates determined using Monte Carlo simulations.  As indicated in the authors' paper, the authors advocate a dynamic approach, described as follows:

"The newer camp is more dynamic with annually recalculated, serially connected, simulations to arrive at a Prudent Withdrawal Rate (PWR) that is sustainable given current conditions. Client annual reviews include annual updates to the simulation data to reflect 1) period life table changes and changes in personal health, 2) current portfolio value, 3) latest market data series, and 4) current year feasible spending needs. The dynamic school provides an ongoing method to address how often and by what method "revisiting" the PWR and making corrections to it recognizing that markets affect the safety of withdrawals and that time allows the PWR to increase."

Thus, while the authors use Monte Carlo simulations, they are using an approach that is similar to the actuarial approach advocated in this website.  In discussions with Mr. Frank Sr., he even refers to his approach as an "actuarial" approach.  Since it is so similar, he makes it very difficult for me to find fault with it.

This paper is a practical application of the previous work done by Messrs.  Frank Sr., Mitchell and Blanchett in three papers that describe in more detail their dynamic approach.  Links to these excellent papers may be found in Mr. Frank Sr.'s website and blog "Better Financial Education.com"