Voices [0:00]

A foolish consistency is the hobgoblin of little mind, adored by little statesmen and philosophers and divines. If a man does not keep pace with his companions, perhaps it is because he hears a different drummer. A different drummer.

Mostly Queen Mary [0:18]

And now, coming to you from Dead Center on your dial, welcome to Risk Parity Radio, where we explore alternatives and asset allocations for the do-it-yourself investor. Broadcasting to you now from the comfort of his easy chair, here is your host, Frank Vasquez.

Mostly Uncle Frank [0:36]

Thank you, Mary, and welcome to Risk Parity Radio. If you are new here and wonder what we are talking about, you may wish to go back and listen to some of the foundational episodes for this program. And the basic foundational episodes are episodes 1, 3, 5, 7, and 9. Yes, it is still in my memory banks. We have also created an additional resource, a collection of additional foundational episodes and other popular episodes.

Voices [1:07]

We have top men working on it right now.

Mostly Uncle Frank [1:14]

Top men. And you can find those on the episode guide page at www.riskparty radio.com. Inconceivable. And all thanks to our friend Luke, our volunteer in Quebec. Sacosh. We'd be helpless without him.

Voices [1:35]

I have always depended on the kindness of strangers.

Mostly Uncle Frank [1:41]

Because other than him, it's just me and Marion here. I'll give you the moon, right?

Voices [1:46]

I'll take it.

Mostly Uncle Frank [1:48]

We have no sponsors, we have no guests, and we have no expansion plans.

Voices [1:52]

I don't think I'd like another job.

Mostly Uncle Frank [1:55]

Over the years, our podcast has become very audienced focused. And I must say we do have the finest podcast audience available.

Voices [2:04]

Really top drawer.

Mostly Uncle Frank [2:07]

Along with a host named after a hot dog.

Voices [2:10]

Light in the French.

Mostly Uncle Frank [2:13]

But now onward, episode 485. Today on Risk Party Radio, we're just gonna do what we do best here. Which is attend to your email.

Voices [2:24]

I want you to be nice.

Mostly Uncle Frank [2:27]

And so without further ado.

Voices [2:30]

Here I go once again with the email.

Mostly Uncle Frank [2:33]

And first off. First off, we have an email from Ben. And Ben Wright.

Mostly Queen Mary [2:50]

Thanks as always for the great podcast. I was wondering if you would be able to help me differentiate my understanding between managed futures versus momentum. Or are they the same? That's not how it works. My understanding is that the five-factor premium model, momentum can also explain some of the premium in stocks, but is not included in the five-factor model. My understanding of managed futures is that they adhere to a trend following strategy, but is that the same as momentum? In that case, how is something like DBMF different from QMOM? Thanks as always for your insights.

Mostly Uncle Frank [3:38]

Well, the short answer is no, they are not the same.

Voices [3:41]

That's not how any of this works.

Mostly Uncle Frank [3:44]

Well, they share some superficial characteristics. So the fund QMOM or QMOM is the Alpha Architect Quantitative Momentum Strategy, and it's a fund of almost all US stocks. It might have some ADRs in it. But it is generally tracking companies with positive momentum characteristics over a relatively long period of time. In terms of its other characteristics, it's essentially a mid-cap growth fund, is how you would describe it. Now, in a sense, that's kind of like a trend following system, but that's not really what true trend following systems look like and what is going on in a broad managed futures fund like DBMF. A typical trend following system will either be on a short time frame of a few weeks or up to a much longer time frame. I would say up to a year. Between a few weeks and a year is probably where you would find just about every trend following system. But the way those funds work is they're not just tracking one kind of asset. And really the point of them in many respects is to apply a similar strategy to a whole broad category of assets. So we're talking not only stock indices, both domestic and international, that would only be one small component. You're also talking about tracking interest rates, you're also talking about tracking commodities, you're also talking about tracking currencies.

Voices [5:15]

That was weird, wild stuff. I did not know that.

Mostly Uncle Frank [5:20]

So a typical managed futures fund will follow trends in anywhere from about 12 to 50 or more markets. And it can go both long and short, whereas a fund like QMOM is just a long-only fund. So for instance, if the stock market is going down, a fund like DBMF will actually go short the stock market at some point and follow that trend down. So in terms of performance characteristics, they're vastly different and they're uncorrelated.

Voices [5:51]

That's the fact jack! That's the fact jack!

Mostly Uncle Frank [5:55]

If you want to see that, just go to somewhere like test folio and put in the two funds and then go look at the correlation matrix, and you'll see the correlation is I think it's between zero and zero point two, which is typical for a trend following system against just about any other asset class. It has about a zero correlation. So you would treat something like managed futures as its separate asset class and part of the alternatives asset class, whereas you'd treat something like QMOM as just part of your equities. So we have discussed what are managed futures and how do they work in many episodes in the past. I would go back to episode 453, I think was the most recent one where we really went through that and there's a nice article there explaining the whole history and how the whole thing works. But in particular, if you search the podcast page and look for episodes where managed futures is actually in the title of the podcast, those are probably going to be the most relevant to what you're talking about here. So while you could make an argument that QMOM is essentially following a trend following system or something like a trend following system, it doesn't really bear much of any relationship to a true managed futures fund or managed futures strategy.

Voices [7:15]

Forget about it.

Mostly Uncle Frank [7:17]

And that's why you can use managed futures as a diversifying asset, because it's not correlated with either stocks or bonds. So hopefully that helps. Check out episode 453 and thank you for your email. Second enough we have an email from Todd.

Voices [7:54]

Now take Todd, for instance. When that perfectly nice young man face his eyes on you and that get up, his hormones are going to go berserk. Oh Bob, I don't like Todd in that way.

Mostly Uncle Frank [8:10]

And Todd writes.

Mostly Queen Mary [8:12]

Frank, what is the best way to run a Monte Carlo simulation for a risk parity style portfolio? I know Portfolio Visualizer has this capability to some degree, but I'd really like to be able to use other financial planning tools like Portfolio Lab to test more complicated withdrawal and tax strategies. Since none of the tools, as far as I know, use detailed investment returns in their calculations, is the best approach to use the mean and standard deviation for my portfolio and portfolio charts portfolio as the inputs for the Monte Carlo inputs, i.e. average real returns of 6% with an 8.2% standard deviation for the golden butterfly portfolio? Thanks.

Voices [8:56]

Hello, Mrs. New Brother.

Mostly Uncle Frank [9:02]

All right, interesting question. And this is a point of a lot of confusion because even people who think they understand what a Monte Carlo simulation is often do not and do not understand how their calculator actually works because it is different whether you are using parameterized data or parameterized returns, and whether you are using historical returns for the basis for your simulation. And if you go to a place like Portfolio Visualizer, it will give you the choice to use either historical returns or parameterized returns. So let's talk about a Monte Carlo simulation using historical returns first, because I think it's easier to understand. Obviously, if you just look at historical data, it goes in a particular sequence, and that's the only sequence it goes into. Now, if you wanted to use that to test other kinds of sequences, you can basically break that data up, and usually it's broken up annually, although it could be broken up by month or by multiple years, in fact. And when you hear somebody talk about the bootstrapping method, that is dealing with how is the data broken up into packets. Because after you break up the data into packets, what a Monte Carlo simulation does is scramble that, usually in about 10,000 different ways, and then comes out with a range of outcomes from the worst outcomes to the best outcomes based on these different sequences of the same data. Okay, so you can see how that works for historical data, because you know what it is. You put it in there for each year and you're using particular assets. You list the asset classes in your portfolio and it looks for, say, the year 1970, and it takes all of the data for the asset classes you picked and puts that into one packet. Now, what if you're using parameterized returns? Now, what are parameterized returns? It's when you're talking about using an average or median return and a standard deviation to develop a data set, essentially. And this is done around what is known as a normalized or Gaussian distribution, the bell curve. And that is what makes it somewhat problematic and less accurate than a Monte Carlo done on historical data.

Voices [11:21]

Just clap!

Mostly Uncle Frank [11:22]

Because essentially the data is not real, it's constructed.

Voices [11:26]

My name's Sonia. I'm going to be showing you the crystal ball and how to use it or how I use it.

Mostly Uncle Frank [11:33]

And a normal distribution is actually not how financial markets work.

Voices [11:40]

That's not how any of this works.

Mostly Uncle Frank [11:42]

Financial markets have what are called fat tails. They are more disorderly than a normalized distribution would suggest. So the shape of the curve or data set suggested by a parameterized data set is not the same shape as it would be for historical data or real data. Another problem with the parameterized data sets is that they treat each year as random and the performance of each asset as random within that year. But that is also not how markets work. Because in reality, the performance of various assets depends on the macroeconomic environment they're in. So you never have a situation where bonds are performing like they're in an inflationary environment while stocks are performing like they're in a recession at the same time. That just can't happen in the real world because you only have one macroeconomic environment at a time. A parameterized Monte Carlo does not recognize that. Surely you can't be serious. I am serious. And don't call me Shirley. So this then begs the question: why do lots of financial calculators only have parameterized returns for their Monte Carlo simulations? And why does that seem to be the more popular way of doing Monte Carlo simulations? And the answer is really just ease and laziness. It's much easier to construct a calculator that way. It's much easier to come up with parameters for various assets or portfolios and effectively use that as a kind of crystal ball that you stick in there.

Voices [13:24]

A crystal ball can help you.

Mostly Uncle Frank [13:29]

But you should always be very suspicious of any Monte Carlo simulation that is only run on parameterized data. It is essentially a kind of crystal ball forecast.

Voices [13:39]

Now the crystal ball has been used since ancient times. It's used for scrying, healing, and meditation.

Mostly Uncle Frank [13:47]

And for that reason, any good financial advisor who is using a calculator that relies on parameterized data is also doing another test, a historical stress test on the worst case scenarios historically, and or a separate Monte Carlo simulation based on known historical data. Because if you're only doing the parameterized data, that is not good for your forecasting. And I know a lot of people are doing that. If you're using Bolden, that's what you're doing, and it's not very good.

Voices [14:20]

Rex Quando, we use the Buddy system. No more flying solo.

Mostly Uncle Frank [14:24]

So you need to also be doing historical analysis of worst-case scenarios for your portfolio, like 1968, 1999, as start dates for withdrawing, for example. And I would also be doing a Monte Carlo simulation of historical data, which you can do at someplace like Portfolio Visualizer. And some of the other portfolios do have that feature. If you're looking around for calculators, I would be asking which one allows for Monte Carlo simulations of historical data. Because that's the one you want, even if you're also doing parameterized Monte Carlo simulations. Okay, but the specific question I think you're asking is where does one find this kind of parameterized data about various portfolios to put into a parameterized data Monte Carlo simulation? And yes, one of the places you can find that is at portfolio charts. If you bring up any of those portfolios, or just use the annual returns chart and put in whatever portfolio you want, it will give you the average returns over the data set it's got since 1970 and also a standard deviation. And that's not bad, but it's really more useful for comparing portfolios than it is for using as the basis for your analysis. Now, the one other thing about the data there, it is real data, as you noted, which means it's already incorporated inflation into it. This is another mistake that amateurs frequently make, is essentially counting inflation twice. So you would not take this data and then go put it in a calculator which had another inflation adjustment embedded in it as well, because this one already has the data with inflation accounted for in it. So if you're using parameterized data from portfolio charts in some other calculator, you would need to set the inflation rate to zero. Now, another way of getting this data to use that is on nominal returns, not accounting for inflation, would be to go to testfolio because you can put in any portfolio there and it will give you a compounded annual growth rate, average return, standard deviation, a whole bunch of other data if you'd like it. So if you look at the golden butterfly portfolio over at testfolio, what you'll see is it's got a compounded annual growth rate of 9.97 over the past 57 years or whatever it is, and a standard deviation that is also about eight. The standard deviation shouldn't change that much. It's in volatility, is where it appears in testfolio. But you can also take those parameters and then put them into a calculator that uses parameterized metrics for its analyses. But again, I think the best use of this is not to get absolute forecasting out of it, but simply to compare one portfolio to another. Because that's really about all you can do with this kind of parameterized data that is not actually reflecting the real world but a facsimile thereof.

Voices [17:32]

Now you can also use the bull to connect to the spirit world.

Mostly Uncle Frank [17:36]

But this is why I always say that you really want to use more than one calculator for whatever you're doing, because ultimately what you're trying to do is decide between a portfolio A and a portfolio B for a particular purpose. And you want to be able to do that comparison in a number of different calculators and get similar results, because then you have confidence that what you're looking at is not idiosyncratic to some calculator and whatever assumptions that are embedded inside that calculator. When I first started doing this work 15 years ago, we didn't really have any of these fancy calculators.

Voices [18:15]

In my day, we had radio and you couldn't see anything, and it was primitive and lousy, and we liked it.

Mostly Uncle Frank [18:21]

So it's much better that we do now, but I do think they are subject to a lot of misuse these days.

Voices [18:28]

We didn't have this technology. Yeah, look at these ditty bitty microphones. I hate them. So tiny and efficient. Imagine we had giant microphones the size of a watermelon, and they were cumbersome and they broke your face so no one could see you, and the only sound that came out was a statically globbly gloop. And that's the way it was, and we liked it!

Mostly Uncle Frank [18:55]

Because people either don't understand the assumptions they're putting into them, or they're just using erroneous assumptions. This will actually be the topic of my presentation at Economy this year on proper forecasting techniques using base rates and reference classes. But I digress.

Voices [19:52]

You are talking about the nonsensical ravings of a lunatic mind.

Mostly Uncle Frank [19:58]

Hopefully that helps on this interesting topic. And thank you for your email.

Voices [20:07]

I wish you come with us. Come on with us. Come on, email. I would love to sneak a peek at those pun decorations your kids work so hard on.

Mostly Uncle Frank [20:25]

Last of Last of Up, we have an email from Tom. And Tom writes.

Mostly Queen Mary [20:44]

Hi, Uncle Frank. First off, thank you for all your education and fun with the podcast. What do you mean, funny? Funny how? How am I funny? I first heard you on the Bigger Pockets Money podcast earlier this year, and I am about halfway through them at this point and skipping around to some of the later ones that pique my interest. You're a real blessing to the finance community, and I can't thank you enough for all your efforts. I hope it continues to be fun for you.

Voices [21:11]

Yeah, I just stare at my desk, but it looks like I'm working.

Mostly Queen Mary [21:16]

I've been doing a deep dive into the sample portfolios, and from what I gather, you are not withdrawing in a classic inflation-adjusted manner, where you increase the amount withdrawn each year by the rate of inflation from the previous year, as was shown in the original Trinity study. I know the average retiree isn't increasing spending each year by the rate of inflation, but I believe you've even said yourself that they increase some based on the rate of inflation. Is there a reason you are not increasing spending each year? If I'm misreading the web page and the withdrawal schedule and you are increasing by the rate of inflation each year, please disregard this email. Wishing you and Mary all the best, Tom.

Mostly Uncle Frank [22:21]

Well, Tom, I'm glad you're enjoying the podcast. Now, as for the method of withdrawals on the sample portfolios, what we're using there is a constant percentage withdrawal method, and so it does go up as the value of the portfolios increases, and goes down as they decrease. Now, why did I decide to do it that way? Well, mostly out of laziness. Isn't that the way I usually do things around here? It's not that I'm lazy, it's that I just don't care. Yeah, that's much easier to calculate on a month-to-month basis. And because in reality, what you were doing is matching this to actual expenses. Using the CPI generally makes no sense for what people do in the real world. I will tell you that if I was using our actual experience in the past five years, we'd be reducing our spending because our spending has, in fact, gone down. Although I think we're bumping it up this year for some renovations.

Voices [23:26]

Looks like a tricky job, Bob. Not when you have a good team, Mr. Bentley.

Mostly Uncle Frank [23:30]

Okay, team, let's get to work. Can we fix it? Yes, we can. But the other reason I'm really not concerned about it is because you can run these simulations at any of the places like portfolio charts or portfolio visualizer or testfolio and just put in a withdrawal rate and say account for inflation, and there you go. You can compare these portfolios to your heart's content, and you can change the dates to match the dates we started here and whatever else you want to do with them. And you will see some surprising things, particularly if you do really stress test these things and start them at the worst date possible. I'll link to in the show notes an analysis from test folio of the golden butterfly golden ratio, a 6040 portfolio, and a three-fund portfolio. If you were to start them all at the year 2000 and subject them all to a 5% safe withdrawal rate, adjusting for inflation. And what you'll see is the golden butterfly and golden ratio portfolios did just fine. The three fund portfolio ran out of money, and the 6040 is on the way to running out of money. Uh what? It's gone. It's all gone.

Voices [24:41]

What's all gone?

Mostly Uncle Frank [24:42]

The money in your account. It didn't do too well, it's gone. But what you'll find is regardless of what kind of withdrawal mechanism you are using, that is not going to change these kind of outcomes. That the best portfolios for withdrawing down are going to be the best portfolios for withdrawing down, almost regardless of the particular withdrawal mechanism. But what those withdrawal mechanisms actually do is essentially make adjustments to the overall safe withdrawal rate. And if you look at the most recent Morningstar State of Retirement report, there's a nice section, second half of it really, where they are taking one basic portfolio and doing different kinds of withdrawal mechanisms. And you can see how that affects the overall safe withdrawal rate of the portfolio. So if you use something like average retiree inflation or actual spending, which is for most retirees, CPI minus 1% or CPI minus 2%, that tends to raise the safe withdrawal rate by about a percentage point just by itself. And then if you use more aggressive guardrail strategies or other kinds of strategies, you'll see different outcomes with effectively higher safe withdrawal rates, the more flexible that you are willing to be. What you should always keep in mind is those original Bengen and Trinity studies used the CPI just for convenience because it was there and some you needed some kind of mechanism for increasing withdrawals. But that was never based on any real data about retirees that we didn't have for 20 more years. It has only been developed in the past ten years. I should say 15 years. But it did provide a consistent basis for comparing portfolios, just as picking a 30-year time frame also provided a consistent basis for comparing portfolios. And that's how you should take those assumptions because when you get to real life, nobody only plans for 30 years unless they're all already quite old, in which case they might be planning for less.

Voices [26:51]

Death stocks you at every turn. Grandpa? Well it does.

Mostly Uncle Frank [27:01]

And nobody actually withdraws on the basis of increasing by this exactly the CPI every year. Because in fact, for retirees, the only thing that inflates at more than the CPI is health care and everything else underinflates the CPI, which is why the overall retiree experience of inflation is between CPI minus 1% and CPI minus 2% on average. And there are very good studies by JP Morgan, T. Row Price, and the RAN Corporation that show that. So hopefully that explains everything to your satisfaction. And this is all written up on the website for each portfolio, exactly how we are rebalancing the portfolios and withdrawing from them.

Voices [27:45]

And you can read it for yourself in this photostatic copy, fax, mendis, incendium, gloria calpum, etcetera, etc. Memo bis punitor delicatum. It's all there, black and white, clear as crystal.

Mostly Uncle Frank [28:00]

And so hopefully that helps. And thank you for your email. But now I see our signal is beginning to fade. If you have comments or questions for me, please send them to Frank at RiskPartyWriter.com. Email is Frank at RiskPartyWriter.com. Or you can go to the website www.riskpartyradio.com. Put your message into the contact form and I'll get it that way. If you haven't had a chance to do it, please go to your favorite podcast provider and like, subscribe, maybe some stars, a follow, a review. That would be great. Okay. Thank you once again for tuning in. This is Frank Vasquez with Risk Party Radio. Signing off.

Mostly Queen Mary [30:05]

Please consult with your own advisors before taking any actions based on any information you have heard here, making sure to take into account your own personal circumstances.