Monte Carlo Holding Living Standard Fixed
If this has been answered, please just direct me to the answer.
I understand and agree on the logic of consumption smoothing over a lifetime. The problem of course is the future is unknown. So, I'd like to take the suggested spending and living standard from a basic setup without Monte Carlo (say a 2% or 3% real rate of return on the retirement assets) and see if we stick to that, what are the very low, low, median, etc values of the retirement assets at the end of the planning horizon, or what year they run out. In other words, I'd like to be able to see trajectories of retirement assets (or the whole estate for that matter) from very low to very high Monte Carlo simulations without being forced to be profligate in the last few years of life in the positive trajectories.
In other words, we'd be content living with the mean expected values of income over time, and are curious about how much will be left in the estate if we consume just that level of income, and what are the risks of outliving the assets following that spending strategy.
Work arounds I've considered are: 1. set the age to death to 120 instead of 100 and look at the output for 100, but that might mess up the smoothing. 2. Take the excess or shortage of income each year in the income trajectory from the base value, and run it into a separate portfolio (not in the model). The trouble with this approach is knowing what the Monte Carlo driven annual portfolio returns are.Income Trajectories Categorized by Lifetime Living Standard
Very Low Low Median High Very High
Subscribe to RSS feed
Comments
dan royer
Mon, 03/21/2016 - 09:44
Permalink
The program is not designed
The program is not designed to show you the unspent retirement accounts directly as a result of MC analysis, but it would seem that you can can get at what you are interested in by using Economics Mode instead of MC mode and then indicating in the Retirement Account settings that you want to spend some said percentage of your retirement assets. It defaults at 100% (spend them all) but you can instead change that to 0% and spend RMD only. Then, given say a 2% or 3% nominal return, you can see what your mean lifetime consumption level is given that you leave some amount of retirement assets unspent for consumption (they will be left as part of the estate). You can also run such a scenario as MC mode, but you'll have less control over the assumed rate of return since instead you will be assuming the historical averages and variance implicit in the combination of asset class you use. Of course there may be other ways to model this, but this is what comes to mind for me. I personally prefer to set my assumed rate of return in economics mode across three or four runs ranging from 0% real return to 3% or 4% real return and get a sense that way of the range of lifetime consumption levels and then seek to adjust my asset allocation to target my return assumption (instead of trying to match my return assumption to my desired living standard). Or put differently, I try to see what is the most conservative asset allocation I can get away with or be happy with.
BrianVezza
Tue, 03/22/2016 - 11:35
Permalink
An option that I've used is
An option that I've used is to set the "std of living" (in assumptions tab), high for the last year of life or the last few years.
This forces the program to reduce consumption during earlier years which saves assets (presumably growing over time) for late in life spending or, to your point, as a cushion just in case. Compare the results with your "baseline" scenario to see how much extra you have in assets during those last few years. The difference is the amount you "saved" by reducing your consumption to roughly match the mean expected values you are trying to model.
You can use this approach with MC (or with MC turned off). It won't be an exact match for what you are asking for, but it's a nice approximation and preserves all of the year-by-year ESPlanner tax calculations, asset values, etc.
Best,
Brian
bill1hampel
Wed, 03/23/2016 - 22:10
Permalink
Thanks to both Dan and Brian.
Thanks to both Dan and Brian. I'll play with those suggestions.