Market Risk and the "4% Solution"

The next thing that was done was to decide what methodology to use to evaluate ‘risk’ and the requirement for mid-course corrections. As I said previously, it was not clear to me how the following should be handled (I’ll be starting with the assumption that there is no fixed annuity in the equation).

Assume that you have decided that a 4.4% withdrawal rate is proper and in the first year you start with $1M, withdraw $44,000 for living expenses, and the markets have a very bad year resulting total portfolio loss of 20%. At the time for withdrawal at the beginning of year #2 your $1M in assets are now worth

(1,000,000-44,000) x (1-.2) = $764,800.

Even assuming no inflation in the first year, another $44,000 withdrawal is 5.8% of the remaining portfolio. Granted the analysis done says that there has never before been a combination of years where this would have presented a problem. But I was very uncomfortable with the concept of withdrawing 5.8% under these circumstances. So the question is what is the probability of this happening, what is the correct withdrawal strategy for handling this, and how does this impact the planned income stream?

Again, for reasons of consistency with previous studies, I stayed with a 30 year time horizon. I also decided that the ‘proper’ strategy was to limit the annual withdrawal to be the smaller amount of:

The initial withdrawal amount adjusted for inflation

A fixed percentage of your current portfolio

After thinking about this for a while I modified the second limit to recognize that shorter timeframes should be able to accommodate more aggressive withdrawal percentages. So I modified #2 to linearly scale the percentage from the initial withdrawal percentage in the first year to 7% in the 30^th (final) year.  I started the analysis with an initial withdrawal amount of 4.4% based on earlier analysis. The choice of 7% was basically arbitrary and I have not yet done any further analysis to try to optimize this choice. Note that I did NOT assume that your withdrawals would increase in those cases where your portfolio value grows faster than inflation.

In these terms the withdrawal rules for the Jarrett/Trinity studies were the smaller of:

The initial withdrawal amount adjusted for inflation

100 % of your current portfolio

Presenting the results of this analysis is more complex than the previous cases. I decided to analyze this situation by looking at the following data for each of the 54 different retirement starting years. The starting conditions were a $1,000,000 initial portfolio value and initial withdrawal of $44,000.

The number of times that the annual withdrawal had to be reduced by more 10% of the original withdrawal amount (after inflation adjustment) in each of the 30 year retirement scenarios. Click Here for Income Volatility.

The largest income drop encountered in each of the 30 year retirement scenarios. Click Here for Worst Year.

The average inflation adjusted withdrawal amount for each of the 30 year retirement scenarios. ClickHere for Average Income.

The ending inflation adjusted portfolio value for each of the 30 year retirement scenarios. Note that this withdrawal methodology means that the ending value will never be zero, although it can approach zero in the limit in some cases. ClickHere Ending Value.

In this chapter the data is presented for one initial withdrawal target (4.4%).

I ended up drawing the following conclusions. 

It is very likely that even using a ‘conservative’ 4.4% withdrawal strategy, you will encounter years with an income below the target of 4.4% of the initial portfolio value (adjusted for inflation).

Seeing the data in this form pushes me toward a more aggressive asset allocation target for retirement.  I found the ‘they all worked at 40/60 and some failed at 60/40’ view compelling using the Trinity/Jarret withdrawal methodology. But from the data that  I'm seeing using this withdrawal methodology, I find a conservative 40/60 asset allocation to be less compelling. The opportunity for a higher ending portfolio value using a 60/40 asset allocation (vs. 40/60) is clear (as is the opportunity for a larger worst case income drop if you look carefully at the graphs). Even 80/20 looks less than reckless when viewed from this perspective.