Demonstrated by James Please, the designer and developer of the FI Calc Retirement planning calculator.
Hosted by the Chicago virtual Bogleheads chapter. Recorded on May 5, 2021.
Chapter meetings are listed in the Bogleheads blog calendar. You can add this calendar to your Google account. Notifications are also sent from an email subscription list. See the blog for more information.
Transcript
Bogleheads® Chapter Series - FI Calc Demonstration
[Music]
Host: Welcome to the Bogleheads Chapter Series. This episode was hosted by the Chicago Virtual Chapter and recorded May 5, 2021. It features James Please, the developer of the FiCalc product. Bogleheads are investors who follow John Bogle’s investing philosophy for attaining financial independence. This recording is for informational purposes only and should not be construed as investment advice.
James Please: Perfect. Yeah, so thanks for taking the time again on such short notice to join this chat with me and hear a little bit about FiCalc. I guess a little bit of history on the app is that I've been a big fan of FIRECalc and cFIREsim and other retirement calculators for some time and I had three goals going into building this app.
One was I wanted it to work on a mobile phone because I use my mobile phone a lot and just in my experience in the industry. I know more and more people are using their mobile phones so that was one goal that I wanted. And then another goal that I wanted was--you know when I first went to those calculators I was a little bit overwhelmed because they seemed so complex--so it was really important to me that I create a lot of guides that document how the calculator works. And really that the goal of the guides is if anyone has any questions about it they should be able to find it here.
I'm not sure how successful I was. Maybe you guys can let me know. But yeah, so that was the goal behind the guides. And then the third thing that I wanted was for there to be at least the same amount of features as FIRECalc and cFIREsim, and ideally a little bit more. Just so that there's also some added functionality to the calculator as well.
I don't think FIRECalc has been updated in some time. But I know cFIREsim just had a rewrite. I'm not too sure, I haven't had a chance to look at it too closely, so I'm not too sure, there may be things that are in cFIREsim that aren't in FiCalc, and vice versa. I'm not too sure but what I can do is I can speak to what is in FiCalc.
We don't know that. But what we do know, what we can look at, is historical data. So we can say, “Hey, what if you retired in 1970 and had a 30-year retirement, would you have run out of money or not.” But then you can say well about 1971. You can check all of the historical data that we have, and that's represented here in the interface by the number of simulations. This is the number of times…
[ Musical interference]
…retirement length. And then we'll really stress test an idea. Maybe that would give me more confidence in my plan. Well if you click “learn more” that's going to open up this guide. And there's actually a section here that describes some of the problems if you go with a retirement that's too short or retirement that's too long. So I won't get too much into the details of that right now, but that's just one of the benefits that I think the guides provide. It can give you some guidance on the range of parameters and a way to think about these different inputs, and really the effect that they can have on the algorithm that's being run.
So that's a little bit of a tangent there. But I'll move on, and we can talk about the portfolio. So in FiCalc you can put your initial portfolio value in a dollar amount and then what you get to do is specify the allocation that you want. So for instance, 80% equities/ 15% bonds/ and 5% cash. That's the default, and then also the default is that it rebalances annually.
Now if you click into this you can see that there are a number of different configuration choices that you can make. So maybe you're not interested in rebalancing. Maybe you just don't think that it's worth the effort, or you just want to see the impact of not doing that work would have had historically. Where you can turn that off, or you can do something like rebalancing every five years if you want.
Another thing that you can do is you can specify what's commonly called a glide path where the allocations change over time So maybe at the start you want to be pretty aggressive and have this 80% equities portfolio toward the end of your retirement you might not be as interested in larger swings in the value of your portfolio, so maybe you would cut that down to 50% equities and 45% bonds. And then maybe you want to see the impact of that kind of slow progression over time.
And then the last thing that you can change about your portfolio is how quickly you reach that new value. So the default is that it's just evenly. So halfway through--so this is a 15-year retirement; so I’m sorry, 30-year retirement--so at year 15 you're halfway to your new values. Well you can change that to be instead quickly, which would get you there to that final value a little bit quicker. Or you could do it slowly. So in this situation you'd be hanging around 80% stocks/15% bonds, you'd be hanging around that for quite a bit longer. And then toward the end it would speed up and get you to that 50%. So just a few different options to configure how you might want your glide path to look. And you can see the results that that might have on the results.
Host: James there's a question about breaking up the stock bond value into small value sectors and things like that. Is it basically what you just showed us? Yeah. Let me see if I can open up the chat window just so I can see that. Okay here we go. Is there any ability to break into say US small value?
James Please: Yes,great question. And the short answer is no, not right now. It's all equities and it's all basically the S&P 500. And the reason for that is because the data source that I'm using for FiCalc comes from an economist named Robert Shiller. He's a guy you may have heard of, something called the Cape, well he's the guy who came up with that, and he also publishes market data all the way back to 1871. Unfortunately his stock data tracks the S&P 500. And if you go to his website and actually click the info button on historical data that will actually include a link--it's his name Robert Schiller--and that will take you to the data source. And then also on this page he describes a little bit about what the data is. So the short answer is I would love to be able to split that up to have small cap, large cap, things like that. But just due to the data source that I'm using right now that's not possible.
There is another data source and it's the one that's used by the Trinity Study and Bengen's original paper. They both use the same data set and I do think that that one breaks things down in way more detail. But it's $500 per year to get that data set and it's just not tenable for me to foot that bill.
Bonds great question. And I’ll also answer the other one by Robert Strauss. But first I'll do the bonds since that's also about the data source. So let's see here, what he [Shilller] has about bonds because I know it's listed here [Shiller site]. Maybe I have a description here [FiCalc site].It may be tracking 10-year treasury notes, but I would need to look that up. Unfortunately the information on this page [Shiller] isn't structured in a way that makes it really easy to skim and get that out. But I'm going to say for now I think it's 10-year treasury notes.
And then, does retirement assume starting at age 65, and then how do you set the age? So here's an interesting thing. Starting, there's no start age in FiCalc. It's all based on how long you think your retirement might last so if you're 65 years old, well maybe you might think that a 30-year retirement seems appropriate. So you might go for 30 years here. If you're 50 years old maybe you might bump that up to 45 or something like that.
So yeah, I know other calculators have that start year and then maybe also an estimated age or something like that. But FiCalc is all based on that length of your retirement, if that makes sense. I hope that answers your question Robert, and thank you LadyGeek for posting that link to the data source, cool.
Host: So unless there are any other questions about the portfolio--back on that age thing. So it really wouldn't know when RMDs kick in, right or would it?
James Please: Correct, it would not. That's something that you would need to factor in manually through something like additional income, or additional withdrawals. Which we can talk about once we get there.
Host: Cool.
James Please: So for withdrawal strategy, another thing that I wanted in FiCalc,or something that I thought could differentiate it from other calculators, was just adding more withdrawal strategies. So yeah, I think there are potentially more withdrawal strategies in FiCalc than some of the competitors. They might not all interest you, but they might be worth checking out. Maybe you'll learn about the things that you value in your withdrawal strategy by seeing what these different strategies prioritized.
And they might not all be familiar to you, so maybe sensible withdrawals, maybe that's something you haven't heard of. Well for each one there is a button that says learn about this strategy, and if you click it there's a little bit of information describing who came up with the strategy, what the goal of the strategy is, some of the math behind it, the equation that it uses. And also some examples. And then I tried, in the last paragraph, to compare it to a strategy that you might understand.
And maybe mention some downfalls- like for instance, sensible withdrawals suffer from these high withdrawal fluctuations from year to year--so if a steady withdrawal amount like what you might get with constant dollar is important to you, then you might want to consider other strategies.
So for each of these different strategies you can then configure it based on that particular equation. So for instance, for constant dollar you can specify how many dollars you want to withdraw each year. And then you can choose to adjust that for inflation, or not. Whereas for say, custom VPW [Variable Portfolio Withdrawals] you can adjust the pmt [payment] values that are part of the custom VPW algorithm.
And so another one would be Guyton-Klinger. So this one is the most complicated withdrawal strategy that I'm aware of. It just has these three rules. They're a little bit complicated and it creates this larger form, but you're still able to fully configure Guyton-Klinger based on the paper where they introduce and describe this method.
So one thing that may come in a future update is there's a difference between the intention behind these withdrawal strategies and there's a difference between your ability to configure things and what the creators of these strategies want you to be able to do.
So from what I could read on the Bogleheads forum, one of the original authors of VPW, long invest, he really did not want you to to change the the pmt formula values of VPW, so that's why custom VPW is separate from that because it allows you to modify those things.
And another one is in the endowment strategy. So these values here that the Yale University endowment team published, there's no flexibility, from what I could read, in adjusting these values. So for that reason it's not adjustable here. But I'm thinking of adding a future update where there would be an advanced option setting, and it would say, hey adjusting these values may change the behavior of this strategy in such a way that goes against the intention of the author. But I mean you're allowed to do that, if that's something that you're comfortable doing.
Host: Tom I see you noticed here that you may have noticed a potential error here in the description of Vanguard Dynamic Spending.
James Please: I’d be curious to hear more about that if you're able to elaborate. Do you want me to comment now or wait till the end.
Host: I’m fine if you want to comment now.
James Please: I think that the 5% is if your portfolio changes. If your default withdrawal is 4%, but then your portfolio drops a lot, it might be that the 4% is actually looking like a far larger percentage of your current portfolio balance. More so then it caps it at only 5%. So it's 5% of the current amount I thought, but I maybe just re-read their white paper on it.
Same thing with 2.5%. I think if your portfolio grows rapidly it won't cap he current value of the portfolio because the standard withdrawal rates based on this Bengen style approach allows, I thought, depending on the current value, you're spending based on your current value portfolio. But cool, that's super helpful. I'll have to give that another read.
Host: You're right. It doesn't mean, by the way, I'm not sure it means your program is incorrect. But I think at least the description, maybe re-read it right.
James Please: Yeah I'll have to give that another read. And yeah I'll be sure to--well I don't know how I can contact you all--but yeah, I’ll take another look at that for sure. Thanks for calling that out.
So there are a couple of different withdrawal strategies there. And another thing that you can do is specify these minimum annual withdrawals and maximum annual withdrawals and some of these strategies, they're pretty untenable if you don't specify any mins or maxes, like percent of portfolio. I mean the available spend can go down way ,way low if maybe you're trying to shoot for around $40,000 a year and you do percent of portfolio, and it only withdraws $15,000. That might be a pretty tough year for you.
So this allows you to say, hey most of the time, hopefully, that 4% hovers maybe around $40,000, but I can flex down to $35,000, and then that ensures that your available spend always stays at $35,000, at that minimum. And then you can also do the opposite way as well. You can see here with this percent of portfolio shoots way up to $160,000. You may not need that, so you can restrict that as well, and see the impact of that on the results.
So that there is a withdrawal strategy. And I'll just hop on over to additional income. So the way additional income works is you might get a side job and it's going to bring in $10,000 a year. Maybe you're expecting a raise, or just an increase in that amount that's tracking inflation, so you adjust that amount for inflation. And maybe you're going to do that for the first 10 years. So that's the flexibility provided by this idea of additional income.
And maybe you plan to work forever. So you can just do income repeats indefinitely, if you don't want to put that end year into it. Or you might say I'm going to take 10 years off, and then maybe I'll get that side job, and then I'll work indefinitely. So that's what additional income allows you to do.
And if we open this up, well okay, we have it here. So this shows a few of the different use cases that additional income is intended to help you model. Some of these--there may be room for a specific feature to allow it to be a little bit more accurate, or more powerful for the use case that you have in mind--but for now it all needs to be entered as additional income. So one example is rental property cash flow. If you have a pension, Social Security, a part-time job, project inheritance, any of those things that you might get as money for a given year.
Host: Is this where dividend income would be entered as well?
James Please: Dividend income is factored in through these assets that you get here. So you would not need to do a separate dividend income as additional income. This would be--these are mainly just -- these are all things that are separate from the portfolio that's defined above if that makes sense.
Host: Cool.
James Please: So there are things that I would like to add to additional income based on feedback people have messaged me. So I know that it doesn't support all of the features that everyone needs for everything but it's a start. And the way that additional income works is it reduces the withdrawal that you actually make. So one thing that's a more complicated idea in FiCalc is this difference between the target spend that you have and the actual withdrawal that you make.
So just to show an example. If we go over to constant dollar and we do $40,000 per year and we adjust it for inflation, and we add in a job and that gives us $40,000 and we adjust it for inflation. It starts immediately and it goes forever. We will never actually withdraw from our investments because this additional income that we've got fully covers the amount that the equation determined that we should withdraw, if that makes sense. And if this were say $20,000, then we would spend that money that year, and then we would withdraw the remainder, which would be $20,000, as determined from the constant dollar strategy.
Now if you pull in $60,000 a year you're going to take that first $40,000 and this calculator assumes that you spend all of that $40,000 that the withdrawal strategy calculates, but then that additional $20,000 gets then reinvested into your portfolio.
Host: I assume everything is in after tax dollars for additional income.
James Please: Yes, you would definitely want to be doing after tax dollars for this. So I'll move on to additional withdrawals. So sometimes you might know that you're spending $40,000 a year, but also you know that you need a car, and maybe you want to factor in buying a new car or a used car every ten years, or something like that. Or maybe you're saving up for a call at someone's college, or just just other things that might come up that you need to factor in.
Well you can do that through additional withdrawals. So you could do college and you could say, okay we're going to spend up to $30,000, we'll adjust that for inflation that's going to happen in four years, and it's going to last, sorry, it's going to start in 10 years, and it's going to repeat for four years. So that's the additional withdrawal, and the way that that works is that this is just tacked on top of whatever the withdrawal strategy calculates. So if the withdrawal strategy calculates $40,000 then when this year occurs that's going to be a $70,000 withdrawal.
Host: And then Rob says, when you say after tax, anything to deal with Roth versus taxable versus tax deferred.
James Please: So for additional income, these are just assuming that they're just dollars that you're receiving, they would be after tax dollars. There' snot really any way to specify things like a Roth or taxable or tax deferred. FiCalc doesn't get into the details of the specifics of where you're in, where your money is. For instance, just looking at the portfolio, it's just equities, bonds and cash. This isn't saying half of this is in a tax deferred account and the other half is taxable. It's not getting into that level of detail. It really does track the map that Bengen and the Trinity Study used, which to my knowledge, doesn't make that distinction between where the accounts are. But that'd be a great feature to add for sure down the road.
Host: And then cs69 says how can everything be an after tax value if you don't specify the type of account the assets are in?
James Please: Well I hope what I just said matches, that again, if you're familiar with the Bengen study, the Trinity Study, the way that this calculator works is very similar to that. It does just add a few additional ideas, like additional income and additional withdrawals, that those studies didn't factor in, but it's still the same idea though. So the returns on equity, bonds, and cash are pre-tax returns. The cash withdrawals are after-tax, correct. So taxes are currently not taken into account when it's calculating things like returns and growth and things like that.
Host : And that is something that people have asked and then cash withdrawals are after tax.
James Please: Do you mean additional withdrawals here?
Host: Yes, correct that would be after tax.
James Please: So moving on down to historical data, the full data set goes back to 1871. But you might want to restrict your data for different reasons. So, for example, one reason that you might want to is that the S&P 500 didn't exist before 1926. So if you look, Robert Schiller describes how he creates this data for 1871 through 1926. You might read that and say, hey, that sounds kind of weird to me. I don't want to trust that. I don't want to include that data. Also Bengen's analysis and the Trinity Study limited their data set, using 1926 as the start date, because they didn't do any of that sort of computation that Shiller did to create that earlier data set. So that's one reason why you might want to limit your data. And that's also why the default value, that the data is limited to 1926 for that reason.
So that's just another configuration option provided by the calculator. And then for user preferences, two little quick things that I just mentioned. If you look there's a little bit of motion when these pop-ups appear, and you might not like that. Well, if you turn this off then it's going to get rid of that, so that it's a little bit less annoying. And also if you have if you're colorblind you can choose one of these options, and it might make some of these charts over here look a little bit more easy to read. So definitely worth checking those out if either of those things appeal to you.
So I thought I'd mention that. So that's all the configuration stuff. And I'm going to hop on over to the questions.
Tom: I thought the Bengen study assumed you paid your taxes with the money you withdrew, that is the withdrawals were pre-tax. Can anyone confirm?
James Please: So Tom just to clarify, that is my understanding as well. And that is how this calculator works as well. So if you're withdrawing $40,000 a year here, you're going to get a tax bill. FiCalc does not include that tax bill into this so you would need to include that into your spending if that makes sense.
And then LadyGeek what statistical distribution do you use to simulate over that data. Are you referring to these charts here, or something else, LadyGeek.
Lady Geek: Yeah I was just trying to understand when you get those nice charts there and there's an underlying--when you're making 121 runs--what are you varying, and by how much.
That's what, that's where the magic hits, where everybody calls it a Monte Carlo. Monte Carlo is just another name for running statistical distribution, but nobody really describes the underlying models or the distributions used to create those curves. So just a little interest in the model itself.
James Please: Yeah so I'm going to--I think I'll answer that as I describe what is in these results--so how about I start describing the results and then once we get down to this chart and let me describe it, and then let me know if you still have questions, how does that sound?
Lady Geek: Okay,
James Please: So one thing, if you read the Bengen study, if you read the Trinity Study, they talk a lot about this idea of success rates. Well if you click this, what is a success rate, what does it mean to be successful or not, well if you click this you can see what the definition is of success for the strategy that you've chosen. So for a category of withdrawal strategies that I call longevity, success is defined as the portfolio never running out of money. So in the Bengen study, if on that 30th year, the portfolio had one dollar left, Bengen would say that that's a success.
Now you might disagree, but that's the definition that Bengen used, and that's also the definition used here in the success rate number. Now that doesn't tell the full story, and that's why I introduced these two other values that might be of interest to people. So the first is this idea called “nearly failed”. So if you click that, you can see what that means. That means that the end portfolio value is less than 35% of the initial portfolio value. Now the reason I added that is because I was toward the end of my retirement and my portfolio value was down as almost a third of what I started with, I might be a little bit concerned. So that's why I added that there because I think that's a piece of information that people would find meaningful. But that was not captured in those studies.
And we're seeing here that it's zero percent, and that's because I added some of these additional incomes. So I'm actually going to refresh the page to wipe all that data out. And you can see here that with this sort of default 1 million dollar portfolio with this $40,000 constant dollar strategy, we're getting 5% of these simulations as being nearly failed.
And then another thing that people are interested in, is this idea of a large portfolio where in FiCalc that's defined as the end portfolio being $300,000, sorry 300 percent, larger than the initial portfolio value. So some people want to maximize how much they spend in their life. They don't want to have lots and lots of money left over. So you can actually see, with the constant dollar strategy, you get a third of your simulations with this large end portfolio.
Now if you go down to something like the endowment strategy, you're going to see here that it does a little bit of a better job at not getting to that nearly failed state or to that large and portfolio state. So that might be something that might be interesting.
Host: All right. Percent of portfolio, all right, cool. So why does percent of portfolio use longevity?. It seems it should never be able to run out of money.
James Please: So I’m going to answer that, but first let me get into a strategy that is not longevity. So an example is VPW. So if you think about it, one of the qualities of VPW is that in the final year your portfolio is zero dollars, right? So if you look here at this chart it's zero dollars. That is by definition the goal of the strategy and all three of the strategies in this maximized spend bucket have that in common. So if you use that Bengen or Trinity Study definition of success, then you're always going to get a 0% success rate.
So instead, success is defined as when two things are met. One is that the portfolio is exhausted and the final year withdrawal is greater than or equal to your minimum specified withdrawal. So if you say that you always want to be able to withdraw $20,000 and you specify VPW, and in that final year you're only able to withdraw $15,000, that's considered a fail for VPW. So it's a little bit of a different definition. It's unique to FiCalc. You may disagree with that, you might agree with it-- definitely open to feedback about that.
And then for percent of portfolio. Okay. So why is this one measured as longevity? It seems it should never be able to run out of money. Well the strategies that tend to not run out of money are ones that are a good fit for longevity, right. The point of longevity is that it shouldn't run out of money. So if you devise an algorithm that literally can never run out of money it's still good to evaluate it as a longevity one, because the alternative, which is the maximum spend, what they have in common is that that final year always drops you to zero dollars. Percent of portfolio doesn't have that quality, so that's why it's longevity. I hope that makes sense, David. But I'm definitely open to hearing alternative perspectives on that.
So moving on down, there's a little bit of analysis that--oh well actually one thing I want to say. An additional something I'd like to add in the future to FiCalc would be the ability to adjust these values. You might say less than 35%, I'd be pretty worried if it was less than 45% or 50%. Well I think the user should be able to configure that or you might say, David, this is a bad definition of success, four percent of portfolio. Well you should be able to modify that. So that's a feature that I'd like to add. I don't have it yet.
And then also the ability to add additional analysis blocks that might be something that appeals to you. So beginning to blur the line a little bit between this calculator app which has more restricted functionality and a spreadsheet app which would really give you kind of limitless flexibility with how you modify that data. And on that note I will mention that you can download all the data. And Ficalc has a CSV, and then load that into a spreadsheet app if you want to do your own analysis.
Host: So you can do that, cool.
James Please: So portfolio at the end of the retirement. This just tracks information about that final portfolio value. So one thing I would like to add would be a way to specify, say you know what's the portfolio like at year five, what about your ten, what about halfway through. I don't have that yet. You can only look at the final year. But I think it'd be pretty cool if you could also specify the time scale that you'd like to see.
If you click these three blocks over here you can actually see a sorted list of all of the end portfolio values. If you click largest, then they'll be sorted from biggest to small. If you click smallest they'll be sorted smallest to large. And then if you click zero dollar portfolios it's going to show--oh whoa! We got a bug here. Well I'll have to fix that bug--it's not showing the zero dollar ones, but it should be showing the zero dollar ones.
So yeah, and then here we go. We’ve got these charts. So this chart is meant to show the quantiles of the data. So it's supposed to show the median value. It's then supposed to show the values that are within 50% of that median, so 25% larger, 25% smaller. And then the 90% quantile and then the 100% as well. And then it also shows the initial portfolio value so you can see about where these values sit.
So one thing that to keep in mind is that these charts they're a V1 version of the chart. The mode should actually be what's at the peak, but right now it's the median. So that's just something to keep in mind, that the actual distribution curve would look a little bit different. So just keep that in mind it's meant more to communicate the idea.
Okay cool. So we have a question from Mosh. And I'm sorry if I'm mispronouncing that. Constant dollar minus dollars, additional income plus dollars. So the net result is what the calculator uses, is a portfolio withdrawal ,assuming constant dollar is greater than additional income plus portfolio return. Give me one second, I'm going to read that again to make sure I understand it before I respond. Mosh, I'm going to ask maybe, if you could elaborate a bit more on that final paragraph I'd be happy to answer it, but I'm having trouble following along with what you're asking there.
Mosh: I totally understand.
James Please: Hopefully you can hear me.
Mosh: I can hear you. Yeah awesome, thanks James, terrific-looking tool. So what I think I'm asking and I think you've confirmed it, but I wanted to just be sure because I'm not sure--I suspect other people aren't sure as well--when we're looking at the calculator, right. You have a constant dollar withdrawal, let's say for sake of discussion forty thousand dollars $40,000, right. Let's say at some point, ten years, so I retire early, right, and after ten years I get Social Security. And let's say my Social Security is $20,000.So obviously the debit of those two numbers is $20,000, right. So logic would tell me that that $20,000 is what should come out of the portfolio, not the $40,000.
And I just wanted to verify that. That's all but obviously plus any portfolio return. So that's all I was trying to say in that convoluted last step paragraph. So is that accurate, what I said.
James Please: That's accurate, and thank you so much for explaining that. Yes, that is an accurate description of what happens in this calculator.
Mosh: That's what I assumed and I appreciate that James. And I'm assuming if I download the CSV file I can then look at that and actually see it in the data. To see it behaving that way.
James Please: You should be able to. Yeah and I think once we click into some of these years we can see that as well. And on that note that is a good segue down here into this available spend. So this is called available spend, and not available, and not withdrawal amount. So although constant dollar is saying-- you know people talk a lot about withdrawing- this is saying that you have that $40,000 a year. It is not saying that you actually withdrew $40,000 that year. So that's an interesting thing, right. Because this withdrawal strategy is a little bit different from maybe what you would want to spend, but it does actually determine, based on your additional income, what you actually do have available to spend.
And what you can see here is that these charts aren't optimized for every single use, so you it's kind of weird, it's like a straight line, but this how it renders when-- well look at that, it says average initial withdrawal amount, I need to update that label there to say average spend amount. So yeah, because it's saying all withdrawals are $40,000, but it's actually all available spend is $40, 000.
And then if we even turn this off for a minute you can see again this is another example where the chart hasn't been optimized for every use case or every situation that you might find yourself in. And the reason that this chart is displaying this way is because you're pretty much always withdrawing $40,000 unless you run out of money, and then you withdraw zero dollars.
So yeah, okay. So I'll move on down to simulations by start year. So this is another kind of distinguishing feature that I wanted to include in FiCalc, which was the ability to really drill down into a particular year and see information about that year. So just at a high level there are three kinds of three color coded values for each year. A red year that means it ran out of money entirely. These yellow years or orange years means it almost ran out of money. It's less than about a third. And then these blue years that's when you end with a lot of money, with that over $300,000 of the initial portfolio.
So just at a glance you can see, okay look at that the ‘60s start of the ‘70s that was a rough time. Maybe around the turn of the century, that was also a bit of a rough time. And then you can see that there were these more booming times in between those and then when you say ran out of money you mean according to the planning available amount for that year.
Agree, what I mean is that if you say, well really what I mean is when you say ran out of money it means the end result is that you have zero dollars in that portfolio. So if I click into one of these and I look at this portfolio value, boom, we hit zero dollars in 1990, several years before the portfolio ended, So any red means that there's zero dollars in that portfolio either in the final year or in any year during the retirement, really.
Host: And then are those end portfolio numbers considered in constant or inflation-adjusted dollars.
James Please: Every dollar that you see on this page is in first year dollars. And actually if you just give me one sec, I will show you. So if you come to this page and it's your first time visiting the page, you'll see this message here that explains that, cs69. So all dollar amounts displayed have been adjusted for inflation to be in first year dollars. So I hope that's clear enough to help allow you to understand what's on the page here.
All right. So what I'll do is I'll hop on in. I'm actually going to turn back on that--okay I must have deleted it--but I'm going to turn on that Social Security for $20,000, starting in 10 years and last forever, adjusting it for inflation. So I'm going to turn that on, and we're just going to hop on in here so you can see that has a pretty, pretty cool result on our retirement here. And what I'll do is I'll hop into 1972 and we'll see what happens.
So I think one of the highest impact features I could update to FiCalc would be showing the withdrawal that you made on the same chart as the available spend. I think it'd be pretty cool to see that you're actually withdrawing $40,000 for these first 10 years, and then once Social Security kicks in, that would show you drop that down to $20,000. And then you would see that $20,000 a year withdrawal. You can't see that right now. You have to download this as a CSV and dig into that.
Rob: And then, you're just using the exterior blocks. Is there any option to randomly change the sequence of returns?
James Please: Great question, Rob. In the works, not done yet. So I actually have it pretty much working but the problem is that right now it could run 120 sims pretty quickly in the browser but I had it run like 100,000 or like 250,000 to try to be a legitimate Monte Carlo simulator, and a competitor to some of the other ones that are out there. But right now the algorithm is too slow to do all of that in the browser. So I either need to improve the efficiency of the algorithm or move it to the server.
And on that note, one thing I'd like to mention is I care a lot about privacy. FiCalc doesn't store any of your data anywhere. No information that you type into this calculator leaves your device. I don't ever want that to change. And also if you click privacy you can see our privacy policy which is we don't store your data, we don't sell your data. That's not what FiCalc is about.
I do have google analytics on the site, and that's because I just like to see who's using it and things like that, and where they're using it from. I'd love to move off google, honestly and maybe go with a more privacy focused alternative. But it's just a time commitment and I think that it would also cost me a little bit of money. So I just haven't been able to to justify that.
And then on that note, there is actually a single ad, and if I calculate it right now, there it is. I hate this ad. Probably remove it. But I was just trying to think of ways to justify--there are a lot of features that I'd like to add to this calculator--but it's really hard to justify spending potentially hundreds of hours. When I'm doing the big time investment, it's not currently earning me any money.
Host: Cool. Talk about the programming tool.
James Please: Yeah It's built in a framework cloud. Well, so the programming language that
I'm using are html and css, which you have to use for the web, and javascript which I use for some of the interactive behaviors. And then within javascript there's a framework called react that makes it a little bit easier to build interactive web applications like FiCalc. So FiCalc is a react front end application. And other than that everything that you see is custom. So that's pretty much the TLDR [Too long; didn’t read] of the tech stack there.
And I'm trying to think of what else. I actually want to head back a little bit and go back to a question that LadyGeek asked a little bit ago about the statistical distribution. I'm curious if that question has been answered, LadyGeek, or if you still have an open question about that.
LadyGeek: Yeah it's just, it's more of a fundamental question. Like you run 121 simulations. Simulation number one, you get a data point. What do you vary on simulation number two? What parameters are varied? And that's where I'm trying to understand the statistics of what is exact, what parameters are you varying in with what distribution? I guess it's just more of a curiosity than anything.
James Please: That makes sense. Yeah. So this is not a Monte Carlo simulator that has things like small variation, like small perturbations on data. It's using Shiller's data set, unmodified. Simulation one starts at 1871. Simulation two starts at 1872. That's the only difference between the first sim and the second sim. So there are 121 sims being run here because that's how many 30-year intervals there are between 1871 and 2020.
And Tom mentions a good point. That if you change the number of years you'll get more or less simulation. So if I bump this up to say 120, you can see only 31 sims are run. Whereas if you know knock this down to 5 we're going to see 146, because they're just more five-year intervals that fit in the data set.
LadyGeek: So it's more like a sliding window.
James Please: Yeah, I think sliding window is an accurate way to describe it.
LadyGeek: Yes, that's what I said because that's very serious. That's why I ask very basic questions. Well I learned a lot certainly, but because it's also a use of terminology. Because when you say Monte Carlo I'm thinking normal log, all the classicals. But now you're talking about variation of the actual data set and time, where your start and end points are. So I think that's a different perspective of what people (me) are assuming of the underlying mechanics here.
James Please: That makes sense.
LadyGeek: Yeah and there's a lot of proprietary information on the commercial tools. Everybody hides exactly how they do it and they all claim accuracy. This is very open, very clear but so that's good.
James Please: So one thing that you might be interested in is there's a section in the guides called “how it works” and it really is intended to be a very clear description of everything that happened. So you can even go into what happens in one simulation year and that kind of stuff. So that might also interest you. So definitely check that out as well.
LadyGeek: Yeah, because the other nice thing I heard, that I say I'm an engineer, but very detailed on getting the basics right. Especially when I did a lot of things for the wiki, terminology is so important because as soon as you said everything is brought back to year one dollars it made perfect sense. Because in finance--that's the main difference actually between engineers and finance people, as engineers think forward in time, finance brings everything backward in time. So that's why they talk about net present value. You talk, you reference things in the present time and it's just a different mindset than engineers. I think of things about 20 years out. How's my product going to work, and of course I have perfect data. But they say that it's a matter of defining the basics, fundamentals. I see you threw a few equations in your explanations, which is good. So thank you.
James Please: Yeah, absolutely. I'm 100% in agreement about the importance of making sure that the terminologies, that the correct terminology is used and that the explanation of what's going on is as clear as possible. I would love for FiCalc to be completely transparent. It may not be there yet, but if anyone sees something that seems like an unanswered question, just let me know, and I would absolutely add something to the guide that explains that.
And I see here we have a question from Jaina or Jonna--sorry if I'm mispronouncing that--but it's just asking a little bit about my background. Yeah, I’d love to share that. So I'm kind of a hybrid role designer engineer. I work at Netflix. I’ve been interested in FIRE [Financial Independence Retire Early] and that kind of thing for a little over 10 years now, I guess. And yeah, I just always thought how it would be pretty cool to give back to the community. I've learned so much from blogs, from forums, just from all the resources and tools people have made. So I just always thought it'd be really cool if I was able to build something that people found valuable. So that motivated me to put FiCalc together. I hope at least some people find it useful. Yeah, so that's a little bit about me.
Host: I guess Barnes is asking if FiCalc uses first year dollars. So after say 10 years, I'll be removing $40,000 plus whatever inflation was over the intervening 10 years, correct.
James Please: It's a great question Barnes. Let me try to clarify. Every dollar that's displayed on this results page will be adjusted for inflation to be in first year dollars, but that does not mean that the underlying calculation is keeping everything in first year dollars. So your question about it would be removing $40,000 plus whatever inflation was, that's something that you specify at the withdrawal strategy level.
So for constant dollar you specify that you want $40,000 per year and then you specify I do want to adjust this amount for inflation. Now that's because that is specified. That's why when we come down here and we see this available spend we see that everything's $40,000 because you've specified in the withdrawal strategy that it's going to adjust for inflation. So over time it's actually increasing that amount that you're withdrawing each year, the actual number of dollars. And then when we actually display the results, we show it in current year dollars because I think that that's a little bit easier for you and just users and myself to evaluate that.
Now if we turn this off we can see things are going to start going crazy because we're no longer adjusting it for inflation. So in first year dollars it's really going to track, check that, track inflation, and we're going to see a more variation in actual purchasing power that you have.
Host: And then Rob says what inflation projection do you use?
James Please: So this is using the CPI data from Robert Shiller. and I think--I'm not sure, I know there are different kinds of CPI --so it looks like he says Consumer Price All Urban Consumers. That may be the one he's using. You'd definitely want to give this paragraph a read just to make sure that that is the actual one. But yeah, it's CPI US-based and potentially the all urban consumers CPI value.
Host: Okay cool. So we have another question from David, I think. There are two basic categories of sims, historical aka back testing, and Monte Carlo.
James Please: Cool, yeah David. That lines up with my understanding as well. I think there's value in both. I definitely appreciate the back testing approach and I also like the Monte Carlo approach as you've described it. And I did start work on a monte carlo version of FiCalc. It's not done yet but people have expressed to me that they find value in that as well. And I think it'd be pretty cool if that were an option available.
So I'm trying to think if there's anything else worth touching on. I think one thing I want to mention again is this thing works great on your phone. And most users of the app do browse it on their phones. So definitely if you have a phone check it out.
Host: If you're just at the grocery store and just wanted to run a quick calculation perhaps, and the word back test simulation of footnote, it's a great, great, great point.
James Please:I definitely agree. Maybe adding the word backtesting or historical or something like that right here on the calculator could make that a little bit more clear. Plus one to that. Thank you.
I’m going to just hop into the guides real fast, just to see if anything comes to mind that we might want to call out. So I will call out one thing. So when you click learn about this strategy, you're going to get this description here. Well if you go into the guides there's another page, there's another section for the withdrawal strategies, and right now they don't link up. So there's a separate description here then there is in the calculator. You might be thinking, what the heck, why would you do that? Why not just have one.
I definitely want one .I'm just not there yet. I plan to--really the goal was to really go really detailed here on the guides, and they are a little bit more detailed than what's in the actual calculator. For instance it has this strengths and weaknesses section for each of the strategies. And then again, that's based on my opinion. so you might have a different opinion, which is fine. But if you do I'd love it if you shot me an email because I’d love to hear more about it.
And then yeah, I guess one other thing I want to mention is if you want to save a calculation. So here we go, we've got this Guyton-Klinger calculation. Let's do a 35-year retirement and let's rebalance every five years. You can hit save--you can, there's this button over here, save or share--and it gives you this url. And if you copy that url, you can then paste it into a new tab, and that's going to load that back up. So as you can see, we've got Guyton-Klinger here, we’ve got that five year rebalance and we’ve got that 35 years. So that's how you can share a calculation if you want to send it to a friend or family member. And then you can also bookmark that. And that's how saving your calculations works as well.
And then just briefly a few high impact things that I think could be interesting to add. One would be this idea called similar results. So what if you run this calculation for 35 years and a million dollars. Well what if, when you scrolled down to the bottom here, it said well what would the results be if you had done 40 years and 1.1 million, and just a brief summary of that. And then there'd be a link and you could click in and you could view that.
So it'd just be a way to get a quick glance at some minor tweaks that might interest you. And then another one would be a way to set up two calculations side by side and do a side by side comparison. Right now you can do it by having two browser windows side by side, but some people expressed to me that they would find value in there being just one in the app.
[Music]