Reply to
@AKAFlack:
Hi AKAFlack,
It’s good that you still participate on this forum. Your views are always respected.
You are spot on-target that accumulated portfolio value is pathway dependent. We would all hate to experience a significant down year immediately after our
retirement date. Depending on the magnitude of the downturn, immediate action might be necessary. More on this later.
However, your Monte Carlo calculations are deeply flawed. It is not that the simulations themselves are wrong; it is that you corrupted the analyses by conflating a single probability analysis event into a double probabilistic sequence of events.
Your starting conditions were distorted by the combination of events that you postulated. In essence, you unwittingly created a Conditional Probability problem and contrasted it against a single Monte Carlo series.
A simple analogy that illustrates this error is that you initially generated the likelihood of the birth of a girl and then contrasted it with the probability of a blond girl being born. Adding constraints always reduces the combined probability.
If you propose to estimate the likelihood of Event A and Event B both happening, the probabilities of each must be multiplied together to get an overall probability. You failed to do so.
To again illustrate, let’s examine your T. Rowe Price Monte Carlo-based analysis in a little more detail.
By assuming a first year 30 % loss in the
retirement portfolio, you dramatically changed the initial conditions of the problem Instead of retiring with a portfolio nest egg at the $750,000 level, the actual probabilistic assessment started a year later at the 0.7 X 750000. = $ 525,000 level. That’s not a fair comparison of equals.
By assuming a 30 % downdraft, you postulated a very unlikely event A. How unlikely?
Scanning the S&P 500 data from 1928 onward, only 3 equity annual return losses exceeded that horrendous performance. Using the historical data to establish a Black Swan Base Rate, that magnitude drop has about a 3.6 % likelihood of happening. Therefore, your scenario of a first year loss of 30 % followed by a conventional Monte Carlo simulation has a combined likelihood of under 3.6 %. The final result is dominated by the high, rare loss that you postulated as a given.
Personally, I will not develop white knuckles worrying over such an improbable investment series. By definition, Black Swans are unpredictable. I recall that you teach finance/investing at the Junior College level. I do worry about this type of faulty analysis finding its way onto the curriculum; it is a common mistake. Please do not make it in the classroom.
Now, there are simple but not always easy steps to protect against unhealthy equity surprises. A diversified portfolio mix of equities and bonds dampens the impact of a large equity downfall.
Proper asset allocation can reduce portfolio volatility (standard deviation) by about a factor of two without compromising expected annual returns. The lower portfolio standard deviation reduces the frequency of negative annual returns while mitigating the overall impact of a negative equity period. A 30 % equity loss might only mean a 12 % portfolio value reduction with bonds serving to cushion the whirlwind.
Finally, a retiree must always be flexible to adjust his withdrawal schedule. If returns fall short of expectations, simply pass on the inflation increase usually included in any competent
retirement withdrawal plan.
If the downturn persists, the retiree has the option to again pass on the inflation adjustment and, if needed, modestly reduce the basic withdrawal rate. Sometimes hard times demand hard measures and a little sacrifice. Old soldiers understand the need for sacrifice.
I have done numerous Monte Carlo simulations that conclusively demonstrate that these modest drawdown devices greatly enhance the survival prospects of a portfolio during
retirement.
I recognize that you know most of what I said. In haste, sometimes there is a disconnect between the brain and the keyboard. I am well aware of that disconnect.
Best Wishes.