Firecalc improvement idea: Guyton/Klinger

First off, many many thanks for an amazing tool. This is proving crucial to my own decision-making, and I cannot thank enough whoever worked so hard in creating such a great tool.

Now I have a suggestion. On this thread, we're discussing what we dubbed the Guyton/Klinger withdrawal model, where the withdrawal rate becomes more adaptive (essentially going up & down depending on the market girations).

A formal and very clear definition of the model can be found in this paper.

This G-K model seems quite sensible, fairly easy to implement, and many of us would probably apply (possibly in a coarser and more improvised manner) a similar adaptive logic anyway at some point during our retirement.

It does seem different from the 95% model, if only because of the all-important 'prosperity' rule.

Being able to model such new spending model based on all the typical Firecalc input parameters would be fantastic. For sure, there would be key questions like:
1. can I *really* go up to 5% withdrawal rate to start with and make it?
2. in a set of scenarios, how often would we reach a predefined floor in spending (WR getting too low for your strictly minimal set of expenses)
3. if I set a floor (lower cap) and/or a ceiling (upper cap), how does this impact the outcome?

I am sure this is non-negligible work (if only to add historical inflation data to the model), but this still seems like an incremental change to Firecalc, and something that would be extremely useful. Yes, I know, easy to say...

Thoughts?

I hope you noticed that rayvt has done some excellent work on this back in the earlier thread.

Independent said:

I hope you noticed that rayvt has done some excellent work on this back in the earlier thread.

Click to expand...

Yup, saw it, very useful work.

I still would love to see Firecalc incorporate such model, so that we can much more easily appreciate the distribution of outcomes across past historical results, and enter more complex input parameters, but for sure, great work from Rayvt.

FireCalc was a major support for my decision to retire (stopped work) in 2004. I had been playing around with it since the WSJ article first appeared. I took the output as an Excel spreadsheet and added information from mortality tables. After all, what I was interesting was, "would the money run out while I was alive". As it has turned out, we have not kept a constant spending rate. The economic turndown sort of automatically cut us back without our even thinking about it. We did not buy a new car, we vacationed on our boat, we did not make any home improvements, and we did not buy much of the stuff that the TV said we had to have. We just had a good time. Without trying, we are up 35% June 2004 to June 2013. Now it is time to buy a car, paint the house, re-wallpaper dwonstairs...

My guess is that the added complexity is not really needed, especially given the uncertainity of the future. You will do the right thing automatically. Why worry!!!

If I'm reading that correctly, you're looking for a lower limit AND an upper limit on withdrawal rate. If you go to Crowdsourced Financial Independence and Early Retirement Simulator/Calculator and select "Adjust Spending > Percent of Portoflio" a couple of options will appear. You can set a lower limit (like the 95% rule) and an upper limit that tracks either CPI or a constant set of inflation.

There's a good example of how the lower/upper limits affect the spending amount in the Instructions for the Basic Inputs section.

wsmurdoch said:

FireCalc was a major support for my decision to retire (stopped work) in 2004. I had been playing around with it since the WSJ article first appeared. I took the output as an Excel spreadsheet and added information from mortality tables. After all, what I was interesting was, "would the money run out while I was alive". As it has turned out, we have not kept a constant spending rate. The economic turndown sort of automatically cut us back without our even thinking about it. We did not buy a new car, we vacationed on our boat, we did not make any home improvements, and we did not buy much of the stuff that the TV said we had to have. We just had a good time. Without trying, we are up 35% June 2004 to June 2013. Now it is time to buy a car, paint the house, re-wallpaper dwonstairs...

My guess is that the added complexity is not really needed, especially given the uncertainity of the future. You will do the right thing automatically. Why worry!!!

Click to expand...

I totally agree with you that real-life is different from theoretical models, and for sure, we human beings are very adaptive.

This being said, the G-K model seems to imply that it is perfectly ok to start at least one point above the usual 4% rule, as long as you are fairly thorough about using a proper variable withdrawal rate. It also suggests that you can withdraw a bit more to start with (early phase of retirement) and be ok, while withdrawing a little less over time, which maps well to my goals. Finally, it seems to manage the gyrations of the market much more effectively.

So... this all sounds quite attractive, but also... a tad anguishing... will this work... will this not imply undue variations of my withdrawals (notably too low)... Hence the value of extending Firecalc towards such model, to be able to test scenarios at will, and be more serene about whatever decision you take.

And then yes, life will likely throw you in a slightly different direction than planned!

siamond said:

Hence the value of extending Firecalc towards such model, to be able to test scenarios at will, and be more serene about whatever decision you make.

Click to expand...

Be more serene? Wow. You must be using a different FireCalc than I use. When I test scenarios, even the ones with a 100% success rate frequently have scary dips and dives with some starting years ending with near zero results. Or, just as bad, I look and see I might have an enormous ending portfolio after years of not doing things I would have enjoyed doing.......... All these wide ranging outcomes driven by the same inputs.

Doing many, many FireCalc runs has taught me that my plan (I'm 7 years into FIRE) can have a wide range of outcomes. If history is a reasonable predictor of the future, I'll likely not run out of money. But I might almost run out or I might end with a huge surplus. A huge variation. Variation over which I have little control seldom makes me feel "serene."

youbet said:

... I might almost run out or I might end with a huge surplus. A huge variation. Variation over which I have little control seldom makes me feel "serene."

Click to expand...

That would not be so bad.

Why do I have the feeling that stock market return of the decades ahead may give us long restless nights of nightmarish sleep, interspersed by brief days of fleeting [-]Wh***[/-] glee? Short interruptions of dark despair, just to give one false hopes, which turn out to be even more cruel?

But never mind my pessimism.

I look at those wiggly lines of FIRECalc, uncertain as they are, as an inspiration. Yes, yes, yes, there's still some hope. That the future is like this decade here. And here.

Or at the worst, like these decades here, where the downslope was slow enough to kindly allow me to expire of a natural cause, before my stash becomes completely depleted, and me being sent to a poorhouse.

I am curious. Is anybody using the Guyton method to calculate their yearly withdrawal amount?