Using goal seek but with a consistent withdrawal rate

[FRPJunkie]

The goal seek feature is really useful. But one issue I have is with the "Use Goal Seek to Find retirement portfolio value needed to sustain withdrawals."

While the feature "works" in the general sense, it allows the planner to withdraw huge, non-real world percentages of the portfolio in some years. For example in my case it allows for withdrawing 15% in some early years, because it sees that in later, post-social security years expenses are covered and the WR is 1%. I don't think I, or most people, would do this.

So my question is this:

Is there a workaround for this feature where we can make the planner run this goal seek function and to ALSO withdraw at a reasonable, realistic rate?

For example, say I want to know the portfolio value I need in order to retire and have a constant (or flexible) withdrawal rate of around 4%.

How can this be accomplished?

[jimr]

I assume the reason the withdrawal rate gets to 15% is because you specified expenses that require that high of a withdrawal in those years. If that high of withdrawal rate is unrealistic, what would you like the simulation to do differently in those years and how should it decide what's reasonable?

Would it be ok if only 50% or 25% of your desired expenses were funded or is there something else the simulation could do to preserve the portfolio while still modeling realistic spending behavior?

[FRPJunkie]

I assume the reason the withdrawal rate gets to 15% is because you specified expenses that require that high of a withdrawal in those years. If that high of withdrawal rate is unrealistic, what would you like the simulation to do differently in those years and how should it decide what's reasonable?

Would it be ok if only 50% or 25% of your desired expenses were funded or is there something else the simulation could do to preserve the portfolio while still modeling realistic spending behavior?

Reasonable question. I guess I would say that to correct for this issue (let's not call it a problem) in this situation, the portfolio value needed to sustain withdrawals needs to be much higher.

So, for example, using goal seek the planner currently reports that I would need, upon retirement, a portfolio value of $1m. The planner then draws down this number, significantly, until it reaches $523,000 right before social security begins. Then the value creeps back up since most expenses are covered.

In reality (or, I should say, in the non-goal seek runs) the portfolio value is $1.8m upon retirement (based on all the inputted assumptions).

So, there must be some kind of happy, realistic medium where the goal seek function is more conservative with what it calculates to be the needed value of the portfolio upon retirement. Somewhere between $1m and $1.8m where the WR is something people might actually use. Maybe this could be solved by telling the planner: "Tell me the portfolio value needed to sustain withdrawals, but using a maximum 6% withdrawal rate."

I just don't think most people would use a 15% WR in the real world even if they knew, with 100% certainty, that SS would cover their predicted expenses.

[jimr]

One suggestion is to play around with the other settings to see anything gets you closer to what you're looking for.

A few things you might try:
1) set a higher probability of success for goal seek (maybe 95%?)
2) Try adjusting the spending policy floor and ceiling values (maybe 90% floor, 100% ceiling)
3) Set the spending policy to stable

Although I know exactly what the code does, I've found Goal Seek to be a bit of a crap shoot. To be fair, I've had the same experience with Excel's goal seek. Sometimes it works really well, but sometimes it produces crazy or meaningless results.

When you get down to it, goal seek isn't a super smart feature. All it does is iterate while adjusting one or two variables. The simulation code itself doesn't have special decision logic in it to support goal seek. Goal seek just tweaks inputs and calls the standard simulation code multiple times.

Personally, I've had the best luck by using goal-seek just to get in the ballpark, then iterating manually without goal seek to fine tune things and play around with more specifics.