How SWR is impacted by key inputs

[Finance Dave]

I'm sorry if this has been posted elsewhere, I read it in the Journal of Financial Planning and thought others may be interested.

Retirement Researcher Blog: Capital Market Expectations, Asset Allocation, and Safe Withdrawal Rates: Additional Results

For some variables, the impact is minor....I can safely withdraw 10% if I only want to plan for a 10 year retirement.

NOT!!

 

[Onward]

Interesting that, for a 40-yr horizon and a 10% failure rate, the maximum withdrawal rate is achieved with a stock allocation less than 60%.

 

[landover]

Interesting that, for a 40-yr horizon and a 10% failure rate, the maximum withdrawal rate is achieved with a stock allocation less than 60%.

It seems from the chart that there is a very small difference in SWR (almost flat line) between 35% to 70% stock allocation when time horizon is 25+ years. There is a clear indication that longer the time horizon, higher stock allocation is better. With current situation, stock yielding more that treasury bonds, higher stocks allocation may be better now intuitively.

 

[jdstein]

The author of the article clearly put a lot of time running various scenarios.

Here is where I struggle with this type of analysis.

1. Retirees only retire once. You only get one return stream, not the multiple return streams derived in a Monte Carlo simulation. Consequently, it makes no sense to base your SWR on the analysis of multiple return series. Rather, your SWR should be decided based on the expected return stream today - which is highly dependent on starting market valuations and the current level of interest rates.

2. Why would anyone base their SWR on a 10% or even a 5% failure rate. Again, you only get one return stream so if you run out of money in your late 80s your failure rate is 100%.

3. The Monte Carlo Simulations used in the analysis assume a normal (more specifically a lognormal distribution). Unfortunately, market returns don't follow a normal return pattern. Severity, frequency and the clustering of losses are much higher than what is assumed in a typical simulation study. See Benoit Mandelbrot's book on the Misbehavior of Markets: A Fractal View of Financial Turbulence

 

[Animorph]

What he doesn't show here, but does elsewhere, is that the higher stock portfolios also have significantly higher average ending balances. That has its own contribution to risk mitigation.