Moshe Milevsky -When Monte Carlo analysis meets a black Swan

Post questions about how to use the planner, user inputs, how the planner works, and comments and suggestions.
Post Reply
jimr
Posts: 848
Joined: Thu Feb 28, 2008 6:48 pm

Moshe Milevsky -When Monte Carlo analysis meets a black Swan

Post by jimr »

There's an interesting article that discusses the problem that Monte Carlo Simulations that use the normal distribution to model portfolio returns may underestimate the likelihood of extreme market declines. Like most Monte Carlo retirement tools, flexibleRetirementPlanner uses the normal distribution to model portfolio returns, so this article is especially relevant.

https://www.investmentnews.com/when-mon ... swan-21541 (updated link on 06-07-21)

In the article, Prof. Milevsky introduces the idea of using something he calls a sustainability ratio to capture the level of risk associated with a given probability of success that pops out of a Monte Carlo retirement planner. This sustainability ratio is computed by simulating a "worst case" one-in-a-hundred-years portfolio return early in retirement and incorporating the results into the analysis.

The sustainability ratio or Sordex (sequence-of-returns downside exposure) as coined by Prof. Milevsky, can be computed using flexibleRetirementPlanner, with a few minor tweaks.

Here are the steps:

1) Enter the inputs for your baseline retirement plan including a return and standard deviation that matches your intended portfolio.
2) Run the simulation and record the probability of success - call it P1
3) Click on the Additional Inputs window and create an entry for "Portfolio Return" in the top panel. Enter your retirement age+3 for BOTH the start and end year. Enter -50% for the return (or whatever you consider your 1 in 100 year worst case return) then enter 0 for std deviation.
4) Run the simulation a second time and record the second probability of success as P2

Now you should have 2 probabilities of success that you can use in the formula below:

Sordex = (P1/P2) - 1

If sordex > 1 -> be afraid
If sordex > 2 -> be very afraid

Happy Computing!

Jim
Post Reply

Who is online

Users browsing this forum: Ahrefs [Bot] and 2 guests