Even with variable withdrawals, you can start as conservatively as you want. If you start conservatively, it reduces the probability of low spending later on. You can do this in TPAW by selecting a high spending tilt.It seems that variable spending can leave someone in a significant reduction of income in some or many years. Not good. SWR may leave someone in a significant excess but isn't that better than a lengthy reduction of income?
The key distinction between variable and fixed withdrawals is not how conservative you start out, but whether or not you adjust spending over time in response to portfolio performance. Not adjusting spending over time in response to portfolio performance leads to overly low withdrawals if the market does well and overly high withdrawals and running out of money when the market does poorly. Both outcomes are suboptimal and easily avoidable. There is only a narrow range of circumstances where SWR would turn out to be just right, and usually only in in hindsight. It's not a strategy that people can or will actually follow, especially if the markets perform badly over a long time and portfolios are seriously depleted. It's just instinctively wrong, and nobody would do it outside of SWR modeling. It's simply a lazy assumption for withdrawal modeling that leads to nonsensical results.
TPAW uses Shiller's data for US stock and bond returns from Jan 1871 to Jul 2023. Average returns during this period have been high. While there can be differences in a historical simulation (data source and range, monthly vs annual modeling, etc.), you should broadly speaking be seeing high numbers. I suspect that you are not doing historical simulations in TPAW correctly. Keep in mind that you need to do two things to do a historical simulation:By the way when I run the SWR in historical the outcomes are significantly different than firecalc. In your model I would have gone broke or had to reduce my spending in my older age. The 5% worst case. In firecalc I always have a significant amount left over at 100. Not sure why historical outcomes would differ in TPAW from firecalc. It should be the same exact historical returns. There is another calculator that does the same thing. You put in the same data and get something totally different out from a historical calculator.
(1) Set "Expected Returns and Volatility" to "historical", and
(2) Set "Simulation" to "historical sequence"
If you are getting pessimistic results using these settings, please share a link to a plan with made up numbers that shows these pessimistic results, and I can look at what's going on there.
Statistics: Posted by Ben Mathew — Wed Jun 26, 2024 9:04 pm