Jim
I know this is not the intent of the planner, but is there perhaps a workaround that will allow me to specify (for example) a $2m estate left when the planner period ends?
I have played around with the Goal Seek but it's not really accomplishing what I want to do, which is this:
Specify an amount left in the portfolio at death, and then work backwards to have the planner specify:
How much savings are needed each year, and
What the required rates of return are based on those savings that would:
create a portfolio with incomes streams and expenses that would have a 100% chance of success while leaving a specific legacy.
Right now, using Goal Seek, the planner will show the maximum that can be withdrawn, but some of those withdrawals seem ridiculous (8%, etc).
Goal seek will also show a minimum rate of return needed to fund the planbut not a RoR that will both fund the plan AND allow for a higher withdrawal that will still leave a specific amount at the end.
The problem with these choices is that they result in an ending value that is either way too high or way too low (for my purposes).
Force planner to leave specific estate?
Re: Force planner to leave specific estate?
Have you tried creating an "other expense" cash flow in the last year of the plan with the amount set to the amount you'd like to preserve (eg $2m)?

 Posts: 4
 Joined: Sat Mar 16, 2019 10:41 am
Re: Force planner to leave specific estate?
Great question and answer! But, I'm trying to model having an estate of zero when I die (I have no kids, etc). How do I do that?
Thanks!
Thanks!
Re: Force planner to leave specific estate?
Got it. This is not something that a Monte Carlo simulation can handle well. The reason is that the Monte Carlo method attempts to optimize a set of stochastically varying inputs to find a result that yields a very high probability of success (say 90% or more). Optimizing with a goal to have a zero ending balance while also having a widely varying sequence of portfolio returns is inconsistent.
The reason is that to have a high probability of success, the model must be so robust that it doesn't run out of money at least 90% of the time. Because returns vary widely across the 10,000 retirement paths thought he simulation, choosing a set of inputs that yields a 90% probability of success necessarily means that 90% of the time the portfolio will have unused funds at the end of the plan. Stated differently, with a 90% probability of success, there's a 10% chance you'll have close to a zero balance at the end of your plan and a 90% chance you'll have more than that, potentially much more than that.
The median ending portfolio balance that's reported in the output shows you the amount you'll have left over at the 50th percentile level. This is determined by sorting the ending portfolio balances for the 10,000 simulation paths and choosing the ending portfolio balance that's in the 5,000th "slot" of the list of 10,000 sorted values. It's not an average of the balances, but rather the one that is greater than 50% of the balances and less than 50% of the balances.
If you mouseover the graph on the main results page, you'll get a popup that shows the median total portfolio value (by year) as well as the top 90% portfolio value and the bottom 10% portfolio value. Those values "round out" the distribution of ending portfolio values. You'll likely see that the 90th percentile portfolio value is huge. This means that in 10% of the runs, the ending portfolio balance was very very large.
I hope that helps make sense of how the simulation model works and what the median ending portfolio value is telling you.
Finally, if you'd like you can "turn off" the Monte Carlo aspect of the tool by setting the "Rtn  Std Dev" input to zero. This causes the simulation to behave like a standard retirement calculator that takes a rate of return, but no standard deviation. Now the tool is acting like a calculator rather than a simulation. With this approach, you could try running goalseek on the withdrawal amount. With a standard deviation set to zero, all 10,000 simulation paths produce the exact same results and this means the goalseek can't converge and always show a zero probability of success (as if the goal seek failed). However, if you look at the detailed view, you'll see that the withdrawal rate shown is one that yields an ending portfolio balance of zero. One caveat is that the withdrawal amount might not be the amount of spending you get since you might have to pay taxes on part or all of the withdrawal. Check out the detailed view tab (with show all details radio selected) to see more. I realize this doesn't do exactly what you're looking for, but it may point you in the right direction.
The reason is that to have a high probability of success, the model must be so robust that it doesn't run out of money at least 90% of the time. Because returns vary widely across the 10,000 retirement paths thought he simulation, choosing a set of inputs that yields a 90% probability of success necessarily means that 90% of the time the portfolio will have unused funds at the end of the plan. Stated differently, with a 90% probability of success, there's a 10% chance you'll have close to a zero balance at the end of your plan and a 90% chance you'll have more than that, potentially much more than that.
The median ending portfolio balance that's reported in the output shows you the amount you'll have left over at the 50th percentile level. This is determined by sorting the ending portfolio balances for the 10,000 simulation paths and choosing the ending portfolio balance that's in the 5,000th "slot" of the list of 10,000 sorted values. It's not an average of the balances, but rather the one that is greater than 50% of the balances and less than 50% of the balances.
If you mouseover the graph on the main results page, you'll get a popup that shows the median total portfolio value (by year) as well as the top 90% portfolio value and the bottom 10% portfolio value. Those values "round out" the distribution of ending portfolio values. You'll likely see that the 90th percentile portfolio value is huge. This means that in 10% of the runs, the ending portfolio balance was very very large.
I hope that helps make sense of how the simulation model works and what the median ending portfolio value is telling you.
Finally, if you'd like you can "turn off" the Monte Carlo aspect of the tool by setting the "Rtn  Std Dev" input to zero. This causes the simulation to behave like a standard retirement calculator that takes a rate of return, but no standard deviation. Now the tool is acting like a calculator rather than a simulation. With this approach, you could try running goalseek on the withdrawal amount. With a standard deviation set to zero, all 10,000 simulation paths produce the exact same results and this means the goalseek can't converge and always show a zero probability of success (as if the goal seek failed). However, if you look at the detailed view, you'll see that the withdrawal rate shown is one that yields an ending portfolio balance of zero. One caveat is that the withdrawal amount might not be the amount of spending you get since you might have to pay taxes on part or all of the withdrawal. Check out the detailed view tab (with show all details radio selected) to see more. I realize this doesn't do exactly what you're looking for, but it may point you in the right direction.

 Posts: 4
 Joined: Sat Mar 16, 2019 10:41 am
Re: Force planner to leave specific estate?
Thank you for the incredibly thorough and quick answer. It cleared up a number of questions.
I LOVE this program!
I LOVE this program!
Who is online
Users browsing this forum: No registered users and 8 guests