Dory - any interest in another 'little' Mod?..................

[perinova]

Originally Posted by dory36

Well yeah, but remember that it takes the higher of 4% of the CURRENT portfolio or 95% of last year's withdrawal.

Since 4% of current will always succeed, a failure only happens when the market is in freefall and the 95% of previous year w/d puts you into the hold -- but the $$ available for spending in a 9% withdrawal starts getting down around to cat food levels by the end of 30 years.

That is the BIG difference between the 95% rule and the default FIRECalc behavior -- by default, your spending ability is preserved exactly as it was when you retired. In the 95% rule, there's no protection of your spending level. This can be fine, as long as you understand it and can live with fluctuations.

In Bob's book, he suggests withdrawals around the same 4% range to create a situation in which you are preserving your portfolio and keeping your spending in a reasonable range.

Make sure you run the model with the default "The success rate of your portfolio ..." and look at the spending by the end. I think in a default case except for 9% spending and using the 95% rule, there are ballpark 1/3 of the years when you'll be down around 1/4 to 1/3 of your original spending.

* I am not sure I understand. Are the withdrawal limited to x% of port. value (which then gives you 9% as possible)- Or are they limited to 4% (always 4%). But then how do you get 9% possible?

** Also - even though possible. I am not sure that the x%SWR w/ 95%-scheme rule applies to current portfolio value in this case but x% of initial port. value adj. for inflation. Drawing less in good time allows you to draw more in bad time (in this case 95% of previous year when the current port. value adj. for inflation is below the initial port. value.


During a stretch of bad time your withdrawal will decrease by only 5% over previous year, then 10% etc

[dory36]

Sorry -- the 4% was in the original definition from his book -- but the actual withdrawal you use is whatever you enter. I was flashing back to his writing.

The 95% rule does not adjust withdrawals for inflation, and it uses the current portfolio (see pages 187-188*), whereas the behavior of FIRECalc (except with the 95% rule checkbox selected) is to use inflation-adjusted withdrawals based on the starting portfolio.

Quite a different model. Default in FIRECalc assumes you want to be able to spend just as much each year. His rule assumes you'd prefer to tighten your belt when the market is in the tank.


(* - and to be sure, I asked him)

[donheff]

I think this option is a useful one to retain but it may need an explanatory blurb or pop up bubble to caution users. You can get 100% SWR rates on some pretty high initial withdrawals, but only because you are protected by increasing reductions in down cycles. The best way to use this calculator is to display the inflation corrected spending pattern under some of the bad cycle scenarios and think about whether you could live with the impact (e.g. a drop to 33% of original spending). Looking at the standard calculator, this one, and the reality approach gives you some very useful data to chew on -- but it also takes some serious thinking.

With a few good early years you could sock away some extra savings in nice secure CDs or MMs and use those funds to compensate for reductions during downturns. If Dory decides to model that the whole simulator may explode.

[dory36]

OK, I changed "success" in the 95% rule to match the book definition (ending portfolio at least as big in inflation adjusted dollars as the starting portfolio.) Lots or wording changes associated with that.

[dory36]

Originally Posted by MasterBlaster

Dory:

You are doing a great job on FIRECALC.

If I could make but one little suggestion, it would be to put text labels on all of the graphs. If it isn't too difficult I would label each of the x and y axis. I would also suggest putting a title on each graph.

That's just good style for any graph.

I am still debating this with myself. (Not sure who is winning... )

I agree in general. The tradeoff is the layout. To allow two charts side by side, and still allow people to print the page without chopping off a lot of the right edge of the text and charts, they have to be fairly small. Text labels will make the charting area even smaller, so I might have to switch to a stack of charts instead of side by side.

I have been focusing on a lot of other aspects so far -- will probably address this soon.

Thanks --

dory36

[samclem]

dory36,
Thanks a lot for adding the ESRBob 4%/95% feature. It is very useful.
- I found the "old" method (% of starting portfolio adjusted for annual inflation) useful when I was considering ER/ESR. It gave me the big picture: At a nest egg of $XXX I'd be able to quit and build myself an inflation adjusted income stream of $XXX per month. That was the critical question then.
- Now that I'm "semi-retired" the option of modeling variable withdrawals based on portfolio value is of greater use to me.


Since everyone gets to "pile on" requests, here's' mine: Would it be possible to characterize in some way the withdrawals taken under the ESRBob method? It would be handy to know if the purchasing power of the annual withdrawals decreased over time (the best indication that the withdrawal % is too big), and some indication of variabilty of the amounts over the time window considered (e.g maybe how many times the annual withdrawals hit the 95% (or other) "safety stop," or some expression of the standard deviation of the returns, etc)

If it's impractical todo this, any suggested "user work arounds" to get the same info would be appreciated
P.S. This suggestion is really for a function that is "above and beyond" as the program is super as-is.

[dory36]

Originally Posted by samclem

Would it be possible to characterize in some way the withdrawals taken under the ESRBob method?* It would be handy to know if the purchasing power of the annual withdrawals decreased over time (the best indication that the withdrawal % is too big), and some indication of variabilty of the amounts over the time window considered (e.g maybe how many times the annual withdrawals hit the 95% (or other) "safety stop,"* or some expression of the standard deviation of the returns, etc)

* If it's impractical todo this, any suggested "user work arounds" to get the same info would be appreciated

I see the question, and I think the results would be easy to collect, but I have not quite figured out how to report the results.

My first impression is that the important part of this question is already answered. The chart on the left shows what sort of income, in today's dollars, you'd be living on in the final year of your plan.

While this doesn't show the variability from year to year during the ~30 years, the problem with a retirement cycle generally depletes it enough that it never quite (or at all) recovers. There are not normal situations where a sequence would have you eating cat food 15 years into retirement, then caviar a few years later. So I suspect that if the ending years are good, so are the intervening years.

In the default settings, but using the 95% rule, the 1950s look pretty lame compared with the other years -- and all of these would have been impacted by the Great Depression. Retirements well established before then, and those starting right after then, all seemed to end up in pretty good shape.

If the internal data is really of interest, I can perhaps generate a text file with the yearly withdrawals, the next time I am writing code to extract data. But I'd need a set of starting conditions to make that useful -- so use the bug report link and check the include data box if you want me to try to get the data.

dory36