The 3% WR may scrimp your lifestyle quite a bit compared to 4.25%, yet during bad years, you are not helped that much. Looks how you do not go broke with either 4.25% or 3%, but with the 3% you are always spending less, in good times and also in bad times.
Unless I find something wrong with the data, I would vote for 4.25% of remaining portfolio.
It turns out that this is not completely true. While this is generally true on the way down and at the bottom of a bad run, with a lower withdrawal rate, the portfolio recovers more quickly from the bad times. So the income will increase more quickly with a lower withdrawal rate, and it can surpass the higher rate income in $ terms. You’d have to look at specific runs and compare rates to see how this works.
It also turns out that I didn’t quite have the worst years in my prior tables. While it’s true that the worst year after 1916 is 1966 dropping to the lowest portfolio in 16 years, there is a cluster of several cases from 1892 to 1916 with worse performance than 1966. [Probably no coincidence that the Federal Reserve was founded at the end of 1913.]
For the cases under 4% withdrawal rate, the 1906 run reaches the lowest portfolio of all runs at year 15 before recovering. For the cases 4% and higher, the 1899 run contains the lowest portfolio of all runs at year 22, before recovering. For the 6% case, the 1892 run is the worst at year 29. It took quite a bit of staring at charts and experimenting to tease out the exact years, but I needed to figure it out so I could see the actual run and get the value of the lowest portfolio year during any run.
So I have updated my tables with these worse scenarios. These lowest portfolio values are around 5% lower than the 1966 case.
Inputs - % remaining portfolio, 30 year run, 50/50 total stock market and 5 year treasury. Runs from 1871. All other values are default.
I added a few more withdrawal rates to see how things behave. 4.35% is pretty close to the “break even” which could be inferred by the average ending portfolio value of 100% and a few other results. This also happens to have a worst ending portfolio of around 50% of the starting portfolio. As you go much above 4.35%, more and more bad runs show up in the later years indicating that the portfolio is eroding more than recovering in more cases. A 6% rate ends up with 30% of the starting portfolio in the worst case after 30 years.
% withdrawal remaining portfolio | average ending portfolio real | lowest ending portfolio real | highest ending portfolio real | Worst case real income reduction (1906, 1899 or 1892 run) | starting year for worst case run | |
| | | | | | |
6.00% | $60,000 | $16,248 | $593,418 | $35,605 | $298,371 | $17,902 |
5.00% | $50,000 | $17,294 | $815,142 | $40,757 | $409,854 | $20,493 |
4.50% | $45,000 | $17,470 | $954,171 | $42,938 | $479,758 | $21,589 |
4.35% | $43,500 | $17,481 | $1,000,171 | $43,507 | $512,306 | $22,285 |
4.25% | $42,500 | $17,476 | $1,032,020 | $43,861 | $518,901 | $22,053 |
4.00% | $40,000 | $17,419 | $1,115,994 | $44,640 | $561,123 | $22,445 |
3.50% | $35,000 | $16,593 | $1,304,200 | $45,647 | $655,754 | $22,951 |
3.33% | $33,300 | $16,210 | $1,374,917 | $45,785 | $691,310 | $23,021 |
3.25% | $32,500 | $16,018 | $1,409,464 | $45,808 | $708,681 | $23,032 |
3.00% | $30,000 | $15,369 | $1,522,919 | $45,688 | $765,726 | $22,972 |
%remaining portfolio is a pretty robust scheme and recovers well, but over many years income could possibly (although not likely) drop as much as 50% (real), even with a modest withdrawal rate of 3%. So you’d have to figure out how to deal with that income variation. On the other hand, portfolio drop didn't get worse than 40%, unless you start going above 4.35%. Higher withdrawal rates caused smaller remaining portfolios as expected.
And here is a specific example using the $1M portfolio where you can compare the income from each of the starting and lowest year cases. I included a couple more columns - income from lowest ending portfolio and average ending portfolios - that illustrate how the higher rates can eventually cause lower income as they shrink the remaining portfolio more aggressively and that can overcome the higher income at some point.
% withdrawal remaining portfolio | $1M starting portfolio income | $1M starting portfolio lowest income | average ending portfolio | income from average ending portfolio | lowest ending portfolio | income from lowest ending portfolio |
| | | | | | |
| | | | | | |
6.00% | $60,000 | $16,248 | $593,418 | $35,605 | $298,371 | $17,902 |
5.00% | $50,000 | $17,294 | $815,142 | $40,757 | $409,854 | $20,493 |
4.50% | $45,000 | $17,470 | $954,171 | $42,938 | $479,758 | $21,589 |
4.35% | $43,500 | $17,481 | $1,000,171 | $43,507 | $512,306 | $22,285 |
4.25% | $42,500 | $17,476 | $1,032,020 | $43,861 | $518,901 | $22,053 |
4.00% | $40,000 | $17,419 | $1,115,994 | $44,640 | $561,123 | $22,445 |
3.50% | $35,000 | $16,593 | $1,304,200 | $45,647 | $655,754 | $22,951 |
3.33% | $33,300 | $16,210 | $1,374,917 | $45,785 | $691,310 | $23,021 |
3.25% | $32,500 | $16,018 | $1,409,464 | $45,808 | $708,681 | $23,032 |
3.00% | $30,000 | $15,369 | $1,522,919 | $45,688 | $765,726 | $22,972 |
Lowest income peaks for the 4.35% case, then starts to drop off. For the lowest portfolio value after 30 years, the highest income wouldn't come from the higher withdrawals, but actually from the lower withdrawals, illustrating that these lower rates recovered so much more than the higher rates they were producing higher income. In the average ending case we see the same thing. For some reason 3.25% happens to beat the higher withdrawal rates in both cases.