When I was looking at my "Can I retire?" question I did several things;
1) I looked at half a dozen free "percent of success" calculators to make sure they all gave me a general "OK to retire" result. I was more interested in seeing that they all said I was not doing something obviously dumb if I retired than if one said I had a 90% chance of success instead of a different calculator saying that it was 95%.
2) I did not take their X% chance of success very seriously. The problem was that when they say "failure" that sounds dire like you become broke and homeless. If you are not planning on a bare bones retirement what "failure" more likely means is that you might need to do something like reduce your spending by 10% when you are in your 70s if your portfolio is doing worse than hoped for. I was planning to have to have enough extras in retirement that I could cut my spending by 10% or more and still be more than comfortable.
3) I did an informal sensitivity analysis where I ran a retirement calculator with different assumptions. For example I might set the spending at $5,000 a month, then $6k, $7k, etc and then plot a graph and see how high I could go and still have something like an 80% success rate. I would have had to have increased my spending a lot to get down to an 80% success rate and there was no cliff where the success rate dropped sharply if I increased my spending by $500 a month. I also tried different asset allocation, length of retirement, etc to see how far I could push things without running into trouble if I set up my model differently.
4) Understand that retirement calculators based on historical data are very limited. How how far back they look varies but if one looks at the last 100 years that goes back to 1924. If you are looking at a 35 year retirement then 35 years ago was 1989 so when you run a retirement calculator using that data you are really asking it "How would I have done if I retired between 1924 and 1989?" That might not be very comparable to retiring in 2024 so you need to take any results with a huge grain of salt. A bad result would be a red flag which you should not ignore the difference between 90 and 95 percent success likely does not mean a lot since things have changed a lot since then. That is also only 65 different data sets starting in 1924 through 1989 with lots of overlapping data.
5) Almost all "percent of success" retirement calculator are wrong or at least misleading because they are asking the wrong question. Basically they are asking something like "What are the odd I was run out of money if I live to be 95?". The better question is "What are the odds my spouse or I will live long enough to run out of money?" Both are valid questions but it is important to keep in mind that if there is a 95% success rate if one of you lives to be 95 then there might be more like a 99% chance that you will both die before you run out of money.
This is the only retirement calculator which I have seen which tries to take your mortality risk into account. Even it is limited since it only looks at one person and not a couple.
"Rich, Broke, or Dead?"
https://engaging-data.com/will-money-last-retire-early/
1) I looked at half a dozen free "percent of success" calculators to make sure they all gave me a general "OK to retire" result. I was more interested in seeing that they all said I was not doing something obviously dumb if I retired than if one said I had a 90% chance of success instead of a different calculator saying that it was 95%.
2) I did not take their X% chance of success very seriously. The problem was that when they say "failure" that sounds dire like you become broke and homeless. If you are not planning on a bare bones retirement what "failure" more likely means is that you might need to do something like reduce your spending by 10% when you are in your 70s if your portfolio is doing worse than hoped for. I was planning to have to have enough extras in retirement that I could cut my spending by 10% or more and still be more than comfortable.
3) I did an informal sensitivity analysis where I ran a retirement calculator with different assumptions. For example I might set the spending at $5,000 a month, then $6k, $7k, etc and then plot a graph and see how high I could go and still have something like an 80% success rate. I would have had to have increased my spending a lot to get down to an 80% success rate and there was no cliff where the success rate dropped sharply if I increased my spending by $500 a month. I also tried different asset allocation, length of retirement, etc to see how far I could push things without running into trouble if I set up my model differently.
4) Understand that retirement calculators based on historical data are very limited. How how far back they look varies but if one looks at the last 100 years that goes back to 1924. If you are looking at a 35 year retirement then 35 years ago was 1989 so when you run a retirement calculator using that data you are really asking it "How would I have done if I retired between 1924 and 1989?" That might not be very comparable to retiring in 2024 so you need to take any results with a huge grain of salt. A bad result would be a red flag which you should not ignore the difference between 90 and 95 percent success likely does not mean a lot since things have changed a lot since then. That is also only 65 different data sets starting in 1924 through 1989 with lots of overlapping data.
5) Almost all "percent of success" retirement calculator are wrong or at least misleading because they are asking the wrong question. Basically they are asking something like "What are the odd I was run out of money if I live to be 95?". The better question is "What are the odds my spouse or I will live long enough to run out of money?" Both are valid questions but it is important to keep in mind that if there is a 95% success rate if one of you lives to be 95 then there might be more like a 99% chance that you will both die before you run out of money.
This is the only retirement calculator which I have seen which tries to take your mortality risk into account. Even it is limited since it only looks at one person and not a couple.
"Rich, Broke, or Dead?"
https://engaging-data.com/will-money-last-retire-early/
Statistics: Posted by Watty — Fri Jun 28, 2024 7:23 am — Replies 6 — Views 1025