Flipping Withdrawal Rates and Success Probabilities Upsidedown

A recent article at Forbes.com by Wade Pfau on WhatDo Market Expectations Have To Do With Safe Withdrawal Rates?  uses Monte Carlo Simulation to flip withdrawal rates and success probabilities upside down.  He calculates the highest sustainable withdrawal rate linked to a particular probability of success, overlays an efficient frontier, and then presents it in a graph and table for use by retirees or about-to-be retirees.  The article is short and simple and does not require much effort to read and certainly does not require me to digest and analyze it…yet here I am.  Just for fun, here are a few take-aways -- none of which are new or surprising -- if one were to look at the article strictly thru the lens of an early retiree with a 40 year planning horizon: 

• "The exhibit further shows, unsurprisingly, that sustainable withdrawal rates are higher for shorter retirement durations" 

• "the optimal stock allocation tends to increase both for longer retirement durations and higher allowed failure probabilities."

• "the upside potential from stocks is increasingly important to sustaining a longer retirement" 

• "an acceptance of greater lifestyle risk (the risk of portfolio depletion) allows for more aggressive behavior in relation to a higher spending rate and stock allocation." 

• "clear evidence that lower stock allocations can compete with higher stock allocations in retiree portfolios"

Taking his Exhibit 2 and distilling it down for a 40 year retirement with fail rates less than or equal to 5% (what most early retirees probably think they want before they realize they don't have sufficient capital) we get: 

            - Suggested Withdrawal Rates          2.8 - 3.4%

            - Optimal Equity Allocation              30%-47%

            - Optimal Bond Allocation                53-70%

            - Range of Stock Allocations             18-65% for WR within .1% of max

            - Range of Bond Allocations              35-82% for WR within .1% of max

(I am combining the rows for 1% and 5% fail rates for a 40 year duration from Exhibit 2)

Thoughts?

- Withdrawal rates are similar to what I've seen in other research and my modeling

- Optimal allocations look a little counterintuitive or low from what I've seen in the past.

- The range of allocations does not surprise me too much; I have seen similar before

- Allowing a higher fail rate, say 20%, permits 79% equity but one might ask "why do this?"