[Maxifi Users] What rate of return are you using for simulations?
[Maxifi Users] What rate of return are you using for simulations?
I recently subscribed to Maxifi and was shocked when it told me I'd need to work another 11yrs to have the amount of discretionary spending I anticipate needing in retirement. I looked at the assumptions and it uses a negative 25bp real return to perform its calculations. This seems overly conservative to me for a 30-40yr retirement. I'm 60/40 AA with only US Total Market and US Bond Total Market. I thought I was being conservative with my calculations and using 2% real as my return number over the course of my lifetime.
FWIW, we're currently sitting at a NW of ~3.5M and anticipate expenses in the 100-110K range for retirement. I figured I was pretty close to being able to retire (maybe next year?) until I ran through the tool.
Really curious how others are using this and would love the forums thoughts overall on your best guess for a 60/40 portfolio in real terms to see if I'm overly optimistic (and yes, I realize these are all guesses).
FWIW, we're currently sitting at a NW of ~3.5M and anticipate expenses in the 100-110K range for retirement. I figured I was pretty close to being able to retire (maybe next year?) until I ran through the tool.
Really curious how others are using this and would love the forums thoughts overall on your best guess for a 60/40 portfolio in real terms to see if I'm overly optimistic (and yes, I realize these are all guesses).
Re: [Maxifi Users] What rate of return are you using for simulations?
Are you saying it is assuming the real return for a 60/40 asset allocation is -0.25%? That can't be reasonable. It is probably a good number for the current return on bonds.
There is a problem with all simulations that require a guess regarding future expected returns. That problem should be dealt with by testing the sensitivity to the assumptions. You can make a table of a range of input assumptions and the output of how many more years you have to work and see how much the one depends on the other. If the dependence is strong, then you (or they) are going to have to do some hard work to figure out the right assumption.
Another way to formulate the projection is to use a long period of past actual data and the assumption is that ex ante the next 30-40 years are not going to be drastically different in history of returns than the last hundred years or more. That would be the approach in FireCalc, for example. If you enter that model with expenses today of $120k and a 60/40 portfolio of $3.5M (invested assets or net worth????) then you project 0% failures. You have to raise the expected spending to $130k to get 1 failure out of 120 experiments.
There are other forecasters out there. Some of them are MC simulators that use historical data for input; some are MC simulators that ask for or make assumptions; some use actual historical data method such as FireCalc, CFireSim, Retirement Optimizer, and (I think) VPW.
There is a problem with all simulations that require a guess regarding future expected returns. That problem should be dealt with by testing the sensitivity to the assumptions. You can make a table of a range of input assumptions and the output of how many more years you have to work and see how much the one depends on the other. If the dependence is strong, then you (or they) are going to have to do some hard work to figure out the right assumption.
Another way to formulate the projection is to use a long period of past actual data and the assumption is that ex ante the next 30-40 years are not going to be drastically different in history of returns than the last hundred years or more. That would be the approach in FireCalc, for example. If you enter that model with expenses today of $120k and a 60/40 portfolio of $3.5M (invested assets or net worth????) then you project 0% failures. You have to raise the expected spending to $130k to get 1 failure out of 120 experiments.
There are other forecasters out there. Some of them are MC simulators that use historical data for input; some are MC simulators that ask for or make assumptions; some use actual historical data method such as FireCalc, CFireSim, Retirement Optimizer, and (I think) VPW.
Re: [Maxifi Users] What rate of return are you using for simulations?
From what I can tell so far, the way Maxifi works is that it takes your total account value and then uses a base return assumption across all accounts. It does not ask when you input your information what your asset mix is, just the account value. From there it applies what they call safe and reliable asset returns, which they default to -.25%. This seems to be the core value it uses to determine maximum discretionary income, which is the overall purpose of the tool. When I change this rate to a real 1.75%, which is what I've been thinking makes sense, my income comes back into the range I was expecting (~9300/mo). I guess I just need to spend more time looking at their help and figuring out how to use it, its definitely not what I was expecting.
FWIW, I paid for the Monte Carlo version of their software and that says a 60/40 portfolio of Large Caps + Total Bond will average 6.47% over the next 38yrs (I'm 47 and have myself being dead at 85). Given they set interest rates at 2%, thats a real return of 4.47% which doesnt seem unreasonable to me.
FWIW, I paid for the Monte Carlo version of their software and that says a 60/40 portfolio of Large Caps + Total Bond will average 6.47% over the next 38yrs (I'm 47 and have myself being dead at 85). Given they set interest rates at 2%, thats a real return of 4.47% which doesnt seem unreasonable to me.
Re: [Maxifi Users] What rate of return are you using for simulations?
Right, so that default kind of sounds like the bottom range of all possibilities rather than actual advice regarding what to use. So you are on your own. Your estimates are probably supportable by the range of what is out there for people trying to guess at future returns. They could be more helpful and make some money by suggesting people obtain the return input from their MC simulation of the investors actual portfolio. Of course the MC uses as an input the mean and standard deviation of the annual returns by asset class, so you could see if you can find out what numbers they are using. (Most MC tools think of the returns being from a normal distribution specified by mean and SD, but some may be more sophisticated than that.) A subtlety is that the compound average return projection is not the same thing as inputs that are arithmetic averages of annual returns.
Re: [Maxifi Users] What rate of return are you using for simulations?
I have used Maxifi or its predecessors for years. No question it is the most comprehensive estimator i have tried. But like each and every estimator, itvis in one way or another an attempt to predict a future that cannot be predicted. So as i thought long and hard i came to the conclusion that it was a waste of time to try to use these programs for anything more than a “best case - worst case” estimator. So i have set my Maxifi parameters so that my total portfolio returns 1% real. I then look at the output and see that the annual income that my portfolio will support is one that we can assuredly live on.
Maxifi is a slick (if pricy)platform that allows one to get down in the weeds on so many aspects of our financial lives, and I find that a great exercise in understanding our financial status. But it can’t predict the future. If one uses it, or any similar program, with that in mind, you should be OK,
And recall that when you enter estimates on future returns with anything other than a jaundiced eye, the only one you’re fooling is yourself.
Maxifi is a slick (if pricy)platform that allows one to get down in the weeds on so many aspects of our financial lives, and I find that a great exercise in understanding our financial status. But it can’t predict the future. If one uses it, or any similar program, with that in mind, you should be OK,
And recall that when you enter estimates on future returns with anything other than a jaundiced eye, the only one you’re fooling is yourself.
Re: [Maxifi Users] What rate of return are you using for simulations?
I have not used MAXFI, but two basic math tests tells me that this tool is overly pessimistic. I don’t know if a value is not entered correctly or the bond rate is incorrect for the long term.
3.5M / 100,000 = 35 years without any growth considered
3.5M * .04 = $140,000 based on 4% annual withdrawal rate
Haven’t you contacted MAXFI support?
3.5M / 100,000 = 35 years without any growth considered
3.5M * .04 = $140,000 based on 4% annual withdrawal rate
Haven’t you contacted MAXFI support?
"I started with nothing and I still have most of it left."
Re: [Maxifi Users] What rate of return are you using for simulations?
Thanks everyone. Sounds like I'll just need to keep playing with it and see if its worth the money. It definitely lets you model all sorts of outcomes and has a lot more flexibility than other calculators. Just not sure if its worth $150. I've been playing around with different scenarios of anywhere from 0 real returns to 2% to see how things look.
@Wiggums: Its not a bug, its a feature . I think they're very intentionally trying to be overly conservative w.r.t. long term planning to give you the best estimate of a "fail safe" number. At least thats how I've interpreted all the data so far.
@Wiggums: Its not a bug, its a feature . I think they're very intentionally trying to be overly conservative w.r.t. long term planning to give you the best estimate of a "fail safe" number. At least thats how I've interpreted all the data so far.
Re: [Maxifi Users] What rate of return are you using for simulations?
Well, one does need to consider the consequences. It would be a tragedy to keep working another ten years at a job one does not want to stay in because some model used a worse than worst case fail safe to convince someone they needed to do that.jjunk wrote: ↑Wed May 12, 2021 11:03 am Thanks everyone. Sounds like I'll just need to keep playing with it and see if its worth the money. It definitely lets you model all sorts of outcomes and has a lot more flexibility than other calculators. Just not sure if its worth $150. I've been playing around with different scenarios of anywhere from 0 real returns to 2% to see how things look.
@Wiggums: Its not a bug, its a feature . I think they're very intentionally trying to be overly conservative w.r.t. long term planning to give you the best estimate of a "fail safe" number. At least thats how I've interpreted all the data so far.
But there is also common sense. What is implied about the economy and the future course of history to suggest that investors should estimate the expected real returns of stocks and bonds to be less than zero for the next thirty years?
I think the suggestion to call support and ask them what one is to make of this is a good idea.
I went to Maxifi and viewed a video on how to run the model. Someplace along they point out the settings options which include entries for inflation, returns, etc., and a statement to adjust those as desired. In the example in the video the default return for investments is 3.5%, so I don't know where your negative return for investments is coming from, but it should obviously be adjusted and that can be done in the program.
Re: [Maxifi Users] What rate of return are you using for simulations?
I went to the program, and it is indeed set so that the default future inflation metric is 2%, and the default total portfolio return metric is set at 1.75%. So you are correct, the implied real rate of return does pencil out at -.25%. In the section on return assumptions, the first line in the section reads : "Rates of return can be highly volatile. Please be very cautious setting your long-term average rates of return below. Being very cautious with this and all other MaxiFi inputs will lead you to spend less and save more, helping protect you against adverse outcomes.". If you follow Larry Kotlikoff at all, you will understand that he is closer to Nouriel Roubini than to Jeremy Siegel in his financial world view. I am not surprised at all that his presets are such that it would stop one in one's tracks. I wouldn't at all be surprised if that was done intentionally. Indeed, if you go to the users manual for the program, and type in "rates of return," he explains in depth why the presets are as such. One thing for sure, you can't accuse him of being overly optimistic.jjunk wrote: ↑Wed May 12, 2021 8:59 am I recently subscribed to Maxifi and was shocked when it told me I'd need to work another 11yrs to have the amount of discretionary spending I anticipate needing in retirement. I looked at the assumptions and it uses a negative 25bp real return to perform its calculations. This seems overly conservative to me for a 30-40yr retirement. I'm 60/40 AA with only US Total Market and US Bond Total Market. I thought I was being conservative with my calculations and using 2% real as my return number over the course of my lifetime.
FWIW, we're currently sitting at a NW of ~3.5M and anticipate expenses in the 100-110K range for retirement. I figured I was pretty close to being able to retire (maybe next year?) until I ran through the tool.
Really curious how others are using this and would love the forums thoughts overall on your best guess for a 60/40 portfolio in real terms to see if I'm overly optimistic (and yes, I realize these are all guesses).
All that aside, this is a very powerful program. I return to it a few times yearly, and learn more and more.
Re: [Maxifi Users] What rate of return are you using for simulations?
I have found their support staff to be very responsive and helpful. I would not hesitate to contact them.
Re: [Maxifi Users] What rate of return are you using for simulations?
Support staff are responsive and very helpful. If you search in the support section the include information on how they come up with their estimates.
I have been using 1% real for several years now. With the run up in assets valuations, I should probably reduce it in my standard model.
I do run simulations a negative 1 and negative 2 percents as worst case for long term planning.
I don't think anyone knows what are reasonable projections from here. We have never been here before.
Negative 1 and 2 along with plus 1 and 2 would bracket a conservative range pretty well.
I have been using 1% real for several years now. With the run up in assets valuations, I should probably reduce it in my standard model.
I do run simulations a negative 1 and negative 2 percents as worst case for long term planning.
I don't think anyone knows what are reasonable projections from here. We have never been here before.
Negative 1 and 2 along with plus 1 and 2 would bracket a conservative range pretty well.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either |
--Swedroe |
We assume that markets are efficient, that prices are right |
--Fama
Re: [Maxifi Users] What rate of return are you using for simulations?
Nice approach.afan wrote: ↑Wed May 12, 2021 12:34 pm Support staff are responsive and very helpful. If you search in the support section the include information on how they come up with their estimates.
I have been using 1% real for several years now. With the run up in assets valuations, I should probably reduce it in my standard model.
I do run simulations a negative 1 and negative 2 percents as worst case for long term planning.
I don't think anyone knows what are reasonable projections from here. We have never been here before.
Negative 1 and 2 along with plus 1 and 2 would bracket a conservative range pretty well.
Re: [Maxifi Users] What rate of return are you using for simulations?
I would think that if a person wants to assess a worst case that along with the outcome there should also be an estimate of the likelihood of getting that case. In fact the best answer is a probability distribution of what the outcomes could be. That is pretty much the method used in most retirement planners. Usually that is derived by entering the estimated return as a statistical distribution rather than as a fixed single parameter.
I do need to be careful here because I am not familiar with this particular program. But a program that asks for a value for expected return should also be asking for a value for variability in annual returns in order to generate a probabilistic output. That is one reason the asset allocation matters, because different asset allocations have different variability as well as different average return.
I do need to be careful here because I am not familiar with this particular program. But a program that asks for a value for expected return should also be asking for a value for variability in annual returns in order to generate a probabilistic output. That is one reason the asset allocation matters, because different asset allocations have different variability as well as different average return.
Re: [Maxifi Users] What rate of return are you using for simulations?
The concern that I have with probability distributions and Monte Carlo simulations is that, in the end, I will have one personal retirement distribution that I will have to live with, and that there is absolutely no way to "a priori" know what it is. Probability distributions and Monte Carlo simulations are great exercises, but in the end they provide precious little reassurance as to what my personal outcome will be. I am as likely to be a winner as I am to be a loser, and it is out of my control what the future will bring. You are correct, probability distributions will tell me what my chances are for success or failure in achieving my desired outcome. But rest assured that no broker, investment advisor, or program designer will give me a money back guarantee. It's up to me to roll those dice and hope for the best.dbr wrote: ↑Wed May 12, 2021 1:13 pm I would think that if a person wants to assess a worst case that along with the outcome there should also be an estimate of the likelihood of getting that case. In fact the best answer is a probability distribution of what the outcomes could be. That is pretty much the method used in most retirement planners. Usually that is derived by entering the estimated return as a statistical distribution rather than as a fixed single parameter.
I do need to be careful here because I am not familiar with this particular program. But a program that asks for a value for expected return should also be asking for a value for variability in annual returns in order to generate a probabilistic output. That is one reason the asset allocation matters, because different asset allocations have different variability as well as different average return.
It is for those reasons that I have chosen a safety first approach (TIPS ladder, longevity annuity, SS for the basics) with the majority of the remainder in Vanguard Total World Stock Market. I accept mediocrity and, as close as I can possibly make, security over the possibility of a greater return.
My best friend once told me that the secret to a happy life is low expectations. The older I get, the truer that becomes.
Re: [Maxifi Users] What rate of return are you using for simulations?
Yep, that is exactly right and a lot of people don't appreciate how that works. But it is also the basis for the whole concept of something like a safe withdrawal rate, which is the surviving scenario in the worst case.bigskyguy wrote: ↑Wed May 12, 2021 1:41 pmThe concern that I have with probability distributions and Monte Carlo simulations is that, in the end, I will have one personal retirement distribution that I will have to live with, and that there is absolutely no way to "a priori" know what it is. Probability distributions and Monte Carlo simulations are great exercises, but in the end they provide precious little reassurance as to what my personal outcome will be. I am as likely to be a winner as I am to be a loser, and it is out of my control what the future will bring. You are correct, probability distributions will tell me what my chances are for success or failure in achieving my desired outcome. But rest assured that no broker, investment advisor, or program designer will give me a money back guarantee. It's up to me to roll those dice and hope for the best.dbr wrote: ↑Wed May 12, 2021 1:13 pm I would think that if a person wants to assess a worst case that along with the outcome there should also be an estimate of the likelihood of getting that case. In fact the best answer is a probability distribution of what the outcomes could be. That is pretty much the method used in most retirement planners. Usually that is derived by entering the estimated return as a statistical distribution rather than as a fixed single parameter.
I do need to be careful here because I am not familiar with this particular program. But a program that asks for a value for expected return should also be asking for a value for variability in annual returns in order to generate a probabilistic output. That is one reason the asset allocation matters, because different asset allocations have different variability as well as different average return.
It is for those reasons that I have chosen a safety first approach (TIPS ladder, longevity annuity, SS for the basics) with the majority of the remainder in Vanguard Total World Stock Market. I accept mediocrity and, as close as I can possibly make, security over the possibility of a greater return.
My best friend once told me that the secret to a happy life is low expectations. The older I get, the truer that becomes.
Taking variability out of the problem is a well regarded solution. Writers in academic finance have sometimes assumed that buying an annuity is the default application for retirement savings. Financial Engines originally didn't even have a retirement spending portfolio analysis because it was assumed that at retirement one buys an annuity.
Exactly what Maxifi does with that problem is not quite clear to me. It seems like people are saying that Kotlikoff or others are suggesting to just assume the worst course of returns and plan on that. It has no personal consequence to me, but I was wondering how users of the program are relating to that.
Re: [Maxifi Users] What rate of return are you using for simulations?
It’s been a while since I have used the tool, but IIRC you can adjust the rate of return. I believe it attempts to derive the present value of future spending based on real risk-free rates (i.e., TIPS). This may be fine for basic living expenses, but is far too conservative for discretionary expenses. Can you just update the results to use your 1% real assumption?
80% global equities (faith-based tilt) + 20% TIPS (LDI)
Re: [Maxifi Users] What rate of return are you using for simulations?
I like the 1% real return mentioned above. Plenty of other retirement calculators out there to test against.
I just tend to use the personal capital one and that is telling me I can spend $10K a month when I only really need $6K a month...
I just tend to use the personal capital one and that is telling me I can spend $10K a month when I only really need $6K a month...
-
- Posts: 992
- Joined: Sat Feb 08, 2020 4:25 pm
Re: [Maxifi Users] What rate of return are you using for simulations?
I am a Maxifi subscriber. When I emailed them about this topic this is essentially what they said. In short, you can set the rates to whatever you want but being "happy" with the "worst course" is the least risky model to plan for spending on retirement.
They also offer the monte carlo simulation where you can easily model "better" returns from the "worse course."
I have just be using the Maxifi defaults for my baseplan and then using the simulation to model other outcomes.
PS - Their customer service has been great w/ quick detailed email responses and their webinars are informative.
Re: [Maxifi Users] What rate of return are you using for simulations?
Thanks for the response. That’s good to know. I read through all the responses. I think investors should expect real returns of stocks and bonds over the next thirty years. We are quite conservative with our retirement projections, but the program default setting seems overly pessimistic to me. I retired at 56, based on our sources of retirement income / savings and no real reduction to annual expenses. So same lifestyle. We had detailed expense records in Quicken, but added a buffer to expenses anyway. So far, we learned that insurance costs rise and the coverage has been reduced by our former employer. Glad we added a buffer to our expenses for these little unknowns.jjunk wrote: ↑Wed May 12, 2021 11:03 am Thanks everyone. Sounds like I'll just need to keep playing with it and see if its worth the money. It definitely lets you model all sorts of outcomes and has a lot more flexibility than other calculators. Just not sure if its worth $150. I've been playing around with different scenarios of anywhere from 0 real returns to 2% to see how things look.
@Wiggums: Its not a bug, its a feature . I think they're very intentionally trying to be overly conservative w.r.t. long term planning to give you the best estimate of a "fail safe" number. At least thats how I've interpreted all the data so far.
"I started with nothing and I still have most of it left."
Re: [Maxifi Users] What rate of return are you using for simulations?
The reason why MaxiFi does this is outlined in this article - Two Approaches and Two Purposes to Planning (why be so cautious?). Now, I would say this is too conservative and I would like a way to distinguish different kinds of regular assets but this is their point of view.
Re: [Maxifi Users] What rate of return are you using for simulations?
Specifically, they describe the process and justification as [Emphasis mine]:vieques wrote: ↑Thu May 13, 2021 1:50 am The reason why MaxiFi does this is outlined in this article - Two Approaches and Two Purposes to Planning (why be so cautious?). Now, I would say this is too conservative and I would like a way to distinguish different kinds of regular assets but this is their point of view.
(https://support.maxifi.com/support/solu ... 5000159069:
"We base our default Inflation rate and nominal rate of return on the 30-year nominal Treasury bond and the 30-year Treasury Inflation-Protected Securities (TIPS).
As of January 2021, we are using the following:
Inflation = 2.00% and Nominal Rate of Return = 1.75%, which implies a -0.245% real rate of return.
We determine inflation as follows (see column E in the attached spreadsheet):
Inflation = (1 + the 30-year nominal Treasury bond yield) / (1 + the 30-year Treasury Inflation-Protected Securities (TIPS) yield) – 1. We then round the calculated inflation and the 30-year nominal Treasury bond yield to the nearest ¼% to determine our default Inflation Rate and Nominal Safe Rate of Return for Regular Assets. We adjust these values if needed in early January and July of each year based on average daily yields on 30-year nominal Treasury bonds as well as 30-year TIPS in the prior months of December and June, respectively to reflect the average values of these 30-year returns over the course of the prior month. Our default assumptions assume that you are either investing exclusively in 30-year inflation-protected/indexed/adjusted bonds or that you are planning based on the conservative assumption that your assets won't earn a higher real (after-inflation) return than this long-term investment vehicle. The 30-year real return is generally positive. But it can also go negative. The returns on 30-year regular and inflation-indexed bonds are set by their market-determined prices. Assuming any other default values than those we are assuming would entail assuming bond prices that differ from those that are actually prevailing.
We urge you to adopt our default values for the rate of return and inflation. Entries that produce much higher or much lower safe real returns are not appropriately cautious and may lead you to spend, save and insure more or less than you should."
Re: [Maxifi Users] What rate of return are you using for simulations?
Appreciate all the suggestions everyone. I spent most of last night playing with the tool. Things look better at 1% real returns (obviously). I also noticed that the tax assumptions are also set to the most conservative values (effectively everything looks non-qualified and short term from a tax perspective). Modifying those to reflect my reality where most dividends are qualified and all of my gains are long term also made things seem more realistic. It's a little more fine tuned than the simple formula I've been using of annual expenses x33.
The thing I'm trying to figure out how is how to handle the discretionary number. As a renter, the housing piece looks to be the only place where a user can define fixed expenses. So I've been playing around with that value and placing all my fixed costs there and then parsing the significantly lower discretionary number into a monthly value to see how it maps up to our overall planning. Its an interesting tool, still not sure if its worth $150 but its interesting. Wish they'd add support for HSA modeling as well.
The thing I'm trying to figure out how is how to handle the discretionary number. As a renter, the housing piece looks to be the only place where a user can define fixed expenses. So I've been playing around with that value and placing all my fixed costs there and then parsing the significantly lower discretionary number into a monthly value to see how it maps up to our overall planning. Its an interesting tool, still not sure if its worth $150 but its interesting. Wish they'd add support for HSA modeling as well.
Re: [Maxifi Users] What rate of return are you using for simulations?
Don't know if this (my) approach will answer your question, but here goes.jjunk wrote: ↑Thu May 13, 2021 11:10 am
The thing I'm trying to figure out how is how to handle the discretionary number. As a renter, the housing piece looks to be the only place where a user can define fixed expenses. So I've been playing around with that value and placing all my fixed costs there and then parsing the significantly lower discretionary number into a monthly value to see how it maps up to our overall planning. Its an interesting tool, still not sure if its worth $150 but its interesting. Wish they'd add support for HSA modeling as well.
Under Household, there is a place to put "special expenses." We calculated our annual base expenses (minus housing, Medicare Part B and Taxes which each have a separate line item in the expense ledger) and entered that number in "real" terms under "special expenses." The result after running the program leaves true "discretionary spending" again as a separate category. That discretionary spending column does then represent our uncommitted money for excess/unplanned/unpredicted/fun expenses.
I'm not sure this is the nexus of your question, but it's what we have done.
Re: [Maxifi Users] What rate of return are you using for simulations?
This is a very interesting way to split the two. Thanks for the suggestion!bigskyguy wrote: ↑Thu May 13, 2021 12:16 pmDon't know if this (my) approach will answer your question, but here goes.jjunk wrote: ↑Thu May 13, 2021 11:10 am
The thing I'm trying to figure out how is how to handle the discretionary number. As a renter, the housing piece looks to be the only place where a user can define fixed expenses. So I've been playing around with that value and placing all my fixed costs there and then parsing the significantly lower discretionary number into a monthly value to see how it maps up to our overall planning. Its an interesting tool, still not sure if its worth $150 but its interesting. Wish they'd add support for HSA modeling as well.
Under Household, there is a place to put "special expenses." We calculated our annual base expenses (minus housing, Medicare Part B and Taxes which each have a separate line item in the expense ledger) and entered that number in "real" terms under "special expenses." The result after running the program leaves true "discretionary spending" again as a separate category. That discretionary spending column does then represent our uncommitted money for excess/unplanned/unpredicted/fun expenses.
I'm not sure this is the nexus of your question, but it's what we have done.
-
- Posts: 992
- Joined: Sat Feb 08, 2020 4:25 pm
Re: [Maxifi Users] What rate of return are you using for simulations?
This software definitely has opinions. One of which can be summarized as what makes people happy is to spend discretionary money at the same level over a long period of time (retirement until death). It does not however tell you what you can spend your discretionary money on - it could be a boat, restaurants, gifts, whatever.
To model an HSA, enter it as an IRA.
- TimeRunner
- Posts: 1939
- Joined: Sat Dec 29, 2012 8:23 pm
- Location: Beach-side, CA
Re: [Maxifi Users] What rate of return are you using for simulations?
Also, under "Settings and Assumptions" you can modify your living standard to, for example, provide more discretionary money in early retirement, reduce the amount over time when elderly, and then kick it back up the last few years of assumed life to cover increased caregiving.
I agree support is very good. I think development could be faster. It bothers me that the numbers are based on January 1 account balances, etc. As the year progresses, those numbers become less realistic. You should be able to start the model with anticipated NEXT January 1 account balances as the year progresses, rather than still look back at that year's January 1 balances. This function existed in ESPLANNER, which was the desktop version that preceded MaxifiPlanner. They intend to add it, but they haven't done that yet.
I agree support is very good. I think development could be faster. It bothers me that the numbers are based on January 1 account balances, etc. As the year progresses, those numbers become less realistic. You should be able to start the model with anticipated NEXT January 1 account balances as the year progresses, rather than still look back at that year's January 1 balances. This function existed in ESPLANNER, which was the desktop version that preceded MaxifiPlanner. They intend to add it, but they haven't done that yet.
One cannot enlighten the unconscious. | "All I need are some tasty waves, a cool buzz, and I'm fine." -Jeff Spicoli
Re: [Maxifi Users] What rate of return are you using for simulations?
I have found that it takes a fair amount of time to input all of the relevant info, but once that's done, the output is quite impressive. There are certain work arounds (HSA entered as an IRA for instance), but if you can deal with them, the result is quite a treasure trove of info.TimeRunner wrote: ↑Thu May 13, 2021 7:44 pm Also, under "Settings and Assumptions" you can modify your living standard to, for example, provide more discretionary money in early retirement, reduce the amount over time when elderly, and then kick it back up the last few years of assumed life to cover increased caregiving.
I agree support is very good. I think development could be faster. It bothers me that the numbers are based on January 1 account balances, etc. As the year progresses, those numbers become less realistic. You should be able to start the model with anticipated NEXT January 1 account balances as the year progresses, rather than still look back at that year's January 1 balances. This function existed in ESPLANNER, which was the desktop version that preceded MaxifiPlanner. They intend to add it, but they haven't done that yet.
-
- Posts: 992
- Joined: Sat Feb 08, 2020 4:25 pm
Re: [Maxifi Users] What rate of return are you using for simulations?
I hear you on this one but I think over the lifetime of retirement the inaccuracy caused by using a Jan 1 balance vs say a June 1 balance would be insignificant. And if anything using the Jan 1 balance later in the year would lead to more conservative (ie better than worst case) outcomes. I use Quicken so it is easy to go back to a January 1st account value.TimeRunner wrote: ↑Thu May 13, 2021 7:44 pm It bothers me that the numbers are based on January 1 account balances, etc.
Also, for other readers of this thread, in my opinion, Maxifi isn’t a day-to-day budgeting tool (like Quicken) or a portfolio modeling tool (like Portfolio Visualizer).
-
- Posts: 6560
- Joined: Tue Jul 26, 2011 1:35 pm
Re: [Maxifi Users] What rate of return are you using for simulations?
I use 4% real after assuming:
50% drop in equities immediately upon retirement
only 75% of estimated SS benefits
I also plan assuming that I need $100K from my Roth
when I turn 80, just in case .
50% drop in equities immediately upon retirement
only 75% of estimated SS benefits
I also plan assuming that I need $100K from my Roth
when I turn 80, just in case .
Re: [Maxifi Users] What rate of return are you using for simulations?
I find this modeling flexibility one of the strongest features of the program, along with the ability to put alternative plans side by side and see what the effects of changing various variables might be. The findings can be counterintuitive and often raise issues that might not have previously been considered.MathWizard wrote: ↑Fri May 14, 2021 9:11 am I use 4% real after assuming:
50% drop in equities immediately upon retirement
only 75% of estimated SS benefits
I also plan assuming that I need $100K from my Roth
when I turn 80, just in case .
This is not a beginners tool, nor is it a set it and forget it tool. If you are like me, and often wonder "what if?" this program often provides the ability to see what the effects might be.
-
- Posts: 6560
- Joined: Tue Jul 26, 2011 1:35 pm
Re: [Maxifi Users] What rate of return are you using for simulations?
I actually wrote my own tool (actually tools). The programs are the most flexible, but more complicated to use.bigskyguy wrote: ↑Fri May 14, 2021 9:28 amI find this modeling flexibility one of the strongest features of the program, along with the ability to put alternative plans side by side and see what the effects of changing various variables might be. The findings can be counterintuitive and often raise issues that might not have previously been considered.MathWizard wrote: ↑Fri May 14, 2021 9:11 am I use 4% real after assuming:
50% drop in equities immediately upon retirement
only 75% of estimated SS benefits
I also plan assuming that I need $100K from my Roth
when I turn 80, just in case .
This is not a beginners tool, nor is it a set it and forget it tool. If you are like me, and often wonder "what if?" this program often provides the ability to see what the effects might be.
The last tool is just a spreadsheet where some of the assumptions are baked-in, like when we claim SS benefits,
and when we each die. (85 and 100). The last has quite an effect on taxes, and I found from my own tools that
if I do Roth conversions until age 76, that the RMDs would not push my widow into a higher tax bracket.
I did it as a spreadsheet with withdrawals/income from each source by year, with a balance that can be adjusted each year,
and with the tax brackets that can be adjusted as needed.
This for my wife and my son who may need to take over if I am no longer able.
If I make it past 76, hopefully things will be pretty much automated, and I won't have to fiddle with finances
much after that.
Re: [Maxifi Users] What rate of return are you using for simulations?
Based on the info in Maxifi help, the main program default return & inflation is intended to result in a "for sure" discretionary spending result. The defaults are based on 30 year nomial Treasury bond & 30 year TIPs. My thinking is step 1 is in the main program compare different profiles to see what you like best. Then go to the Monte Carlo Risk Analysis part of the program where you can enter assets. Then run the simulation and see how the the Discretionary Spending Range changes during retirement. Use this to decide if you are comfortable adjusting the default return nos. (called "spending behavior" in Monte Carlo). Took me a while to absorb this process, but I like it. The top spending trend in this link, might be an example where you'd want to increase your return or "spending behavoir" (because the trend tips upward): https://support.maxifi.com/support/solu ... onte-carlo
-
- Posts: 992
- Joined: Sat Feb 08, 2020 4:25 pm
Re: [Maxifi Users] What rate of return are you using for simulations?
The field being discussing is labeled “safe rate of return” and they define it at length in the help documentation. They also define why they set the defaults to the rates they do for inflation and rate of return. The defaults can all be overridden.dbr wrote: ↑Wed May 12, 2021 11:27 am What is implied about the economy and the future course of history to suggest that investors should estimate the expected real returns of stocks and bonds to be less than zero for the next thirty years?
I think the suggestion to call support and ask them what one is to make of this is a good idea.
I reached out to tech support multiple times and they are very responsive.
In short, if you use the defaults and are happy with the default/worst/conservative/low estimates than a better outcome should be a welcome situation. Overriding the defaults and using the Monte Carlo will model other rates of return. They will even show you whether or not your chosen strategy has more reward for a similar risk level.
Re: [Maxifi Users] What rate of return are you using for simulations?
I logged into MaxiFi today and saw that right now they have updated their default values:
Inflation = 2.25% and Nominal Rate of Return = 2.25%, which implies a 0% real rate of return.
Inflation = 2.25% and Nominal Rate of Return = 2.25%, which implies a 0% real rate of return.
67/12/21 US stock/international stock/bonds. Bonds capped at 10x annual spending. Semi-retired as of 2022.
-
- Posts: 2481
- Joined: Wed Apr 27, 2016 9:42 pm
Re: [Maxifi Users] What rate of return are you using for simulations?
In MaxiFi, use of the term discretionary spending is more inclusive than may be more commonly used. It means everything except Medicare Part B, taxes, housing, life insurance premiums (if any) and retirement contributions (if you're still making them.)tonyclifton wrote: ↑Thu May 13, 2021 7:05 pmThis software definitely has opinions. One of which can be summarized as what makes people happy is to spend discretionary money at the same level over a long period of time (retirement until death). It does not however tell you what you can spend your discretionary money on - it could be a boat, restaurants, gifts, whatever. ...
Re: [Maxifi Users] What rate of return are you using for simulations?
MaxFi definitely leans to a more conservative results. Even when I email their customer service (they are very responsive!) they say to assume lower returns. However you can adjust many of the investing/ spending parameters to what you want. The component I don't like is their Progress Tracker for spending.
Last edited by ROIGuy on Mon Jul 12, 2021 3:19 pm, edited 1 time in total.
Re: [Maxifi Users] What rate of return are you using for simulations?
Thanks for this info.
If the software doesn’t track the AA for each account, then it makes sense to use conservative values. I created a spreadsheet and noticed that a 1/4% makes a big difference over a long retirement. You definitely can’t use a 6% return if this value also applies to fixed income.
"I started with nothing and I still have most of it left."
- TimeRunner
- Posts: 1939
- Joined: Sat Dec 29, 2012 8:23 pm
- Location: Beach-side, CA
Re: [Maxifi Users] What rate of return are you using for simulations?
Yeah, it's really lame. Much easier using a spreadsheet or Quicken for that. I think people were asking for it, and they just plugged it in to check a features box. I'd just ignore it.
One cannot enlighten the unconscious. | "All I need are some tasty waves, a cool buzz, and I'm fine." -Jeff Spicoli
Re: [Maxifi Users] What rate of return are you using for simulations?
I use a spreadsheet for expenses. Much more detailed and accurate.TimeRunner wrote: ↑Mon Jul 12, 2021 11:05 pmYeah, it's really lame. Much easier using a spreadsheet or Quicken for that. I think people were asking for it, and they just plugged it in to check a features box. I'd just ignore it.
-
- Posts: 992
- Joined: Sat Feb 08, 2020 4:25 pm
Re: [Maxifi Users] What rate of return are you using for simulations?
It doesn’t really need to track AA as you are setting the total return for all your retirement accounts combined. In other words, you tell it what your total portfolio performance will be.
In the Monte Carlo feature you can set asset allocation and expected returns (different feature / different purpose).