Mostly Voices [0:00]

A foolish consistency is the hobgoblin of little minds, 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 Mary [0:19]

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 [0:37]

Thank you, Mary, and welcome to episode 49 of Risk Parity Radio. Today on Risk Parity Radio, we are going to discuss the after effects of the recent inauguration. No, we're not. Man's got to know his limitations. Today on Risk Parity Radio, we are going to look into our mailbag and talk about a couple questions from our listeners. The first one comes from Chris and Chris writes, hi Frank, I've really been enjoying your podcast. I have a question for the risk parity concept. I have 200,000 inherited IRA which I have to pull all the money out of within 10 years. My plan is to remove $2,000 per month from the account, which is currently invested in a golden butterfly type portfolio. How can I evaluate this plan versus other risk parity portfolios with such a high drawdown? I want to be able to rely on the $2,000 per month and have the maximum amount left over after the 10-year period is complete. Thanks for consideration. Chris, well now that is an interesting question and it has kind of an interesting answer. It's not so much predicated on the exact portfolio you have. The key factor is actually where is the money going, believe it or not. Because for instance, and you can look at two points on either end of the spectrum. On one end of the spectrum, Chris will be taking the money, the 2000, and spending it, just using it as regular income or doing something else with it. And then on the other end of the spectrum, Chris could be putting this back into another portfolio, another long-term portfolio or a medium-term portfolio or any kind of portfolio. But that is the easy thing to answer. If you're simply going to put it into another portfolio that is similar to the golden butterfly portfolio you're talking about, you would simply, as you moved pieces of that portfolio out, rebuy those segments in another account, probably just a regular brokerage account. So you could take it out that way. The more difficult question is the question when you are drawing down and planning on spending the money and how do you preserve that best for the period. Knowing that the future, this 10-year period, is going to be unpredictable and it is not going to be clear which parts of this portfolio are going to perform the best at particular times. And I thought about it and basically I think there's two approaches to it. One could be where you simply draw from the best performing asset every month and you just go look at which is performing the best, sell a little piece of that and pull it out and there you go. And that would be similar to the way we are distributing from our Risk Parity Style Portfolios. Now it's possible, given the steepness of the drawdown, you would pull out of more than one fund, but the idea would be that you sort of keep them equal or in their allocations as they are being drawn down. Now that kind of preserves the general characteristics of that portfolio, the risk characteristics. The other, maybe a little bit more sophisticated way is to think about it in terms of we have this portfolio that has these risk characteristics right now. As we draw down though, we want to go to a portfolio that has lower and lower risk characteristics as we get towards that end so that that last money we're taking out isn't bouncing around a lot. And the way you'd probably want to look at that for this kind of portfolio, and we are looking, and I'm using the golden butterfly just to remind you what's in that, that is 20% long-term treasuries, 20% short-term treasuries, 20% total stock market, 20% short-term value stocks, and 20% gold. Now, in order to make this more conservative as you go down, you need to take a look at the volatilities of these various characteristics. And the idea is you would pull from the most volatile asset classes first, and then as you pull down, you get down to the asset classes that are the least volatile, and you are the most certain will remain the way they are. as you're getting to that end. And in this case, to look at the volatility, fortunately, if you go to the website www.riskparityradio. com and you go to the description of the Golden Butterfly in the portfolios page and click on the correlation analysis there, click on that correlation analysis, because in addition to a correlation analysis, that calculator also spits out standard deviations, which gives you a volatility metric to look at and compare between the various components of that. And if you do that, what you'll find is that four of the components have similar volatilities. They're all about 1% on the standard deviation, and those are everything except for the short-term bond fund. Now, the short-term bond fund is much less volatile. volatile than the rest of the portfolio, like 20 times less volatile than the rest of the portfolio. So in thinking through how you might do this, you could leave the short-term bond fund alone and look at those other four components and then pull those down equally, also pulling from the best performer. And so what that will result in is that you will basically empty those out in the first eight years, and then what you'll be left with for those last two years, approximately, will be that short-term bond fund, which will just be almost like cash and it'll be coming out. And that way you're going to have the most certainty in terms of bringing that down in an orderly manner and not having some weird volatility happen at the end. that messes this thing up. Now, I should say that this portfolio is not that volatile to begin with anyway. It typically has a, over the last 50 years, a maximum drawdown of about 20% over about a three-year period. So this is a 10-year period. It's long enough that just drawing it down in the way we first talked about may be just fine. if you don't want to go through the gyrations, but probably the most certain way is to draw down the more volatile components first and then use that short-term bond fund at the end of it. And I think it is also for that reason that it's not really possible to say whether the Golden Butterfly would be a better drawdown portfolio in this circumstance as compared to something like the golden ratio with similar risk characteristics. Obviously, if you wanted to be the safest, you would go with the least volatile portfolio, which is something like that All Seasons portfolio in our samples page. But I think the Golden Butterfly is a very useful portfolio for this kind of situation. And if you're comfortable with it now, I would feel comfortable with it going through this drawdown period. All right, the next question comes from Tommy V. And it's not a single question. It's asking for commentary on a video where they are interviewing the two gentlemen who run the RPAR risk parity ETF. I will be linking to this video in the show notes. It is worth listening to these gentlemen. have experience and they one of them comes out of Bridgewater and learned under Ray Dalio. Now we talked about their fund and their portfolio in episode 31, if you want to go back to that. But in this interview they talk a lot about risk parity principles and it's a long video. It's over an hour long. I still do recommend it. But Tommy Vee wanted me to focus on a couple of Comments that start around 119 in this video. And so we'll take a look at what he's got here in the email that he sent me. I should say this also follows up on episode 34 regarding correlations in a video from Larry Swedroe that we talked about then. It's also something you might want to go back to looking at. But taking a look at what Tommy V has pulled out of this transcript, there is a comment by Rodrigo, one of the interviewers, I believe, and he says, I think that people may have heard something or somebody say, Look, a good risk parity portfolio is some tips and some equities, and then they just hear bonds and equities. And then they say, well, revenues are correlated to each other. And he was saying that that was not an accurate impression. And I do agree that's not an accurate impression of risk parity style portfolios at all, simply because you do need to look at the types of bonds, and we've discussed those in particular in episodes 14 and 16, that some bonds are correlated with equities, particularly corporate bonds and high yield corporate bonds. And yes, those revenues are going to be correlated to each other because we're talking about either the equity portion or the debt portion of particular companies. And that is why you would want to look at bonds that are not correlated with stocks, and those are generally treasury bonds. And then let me read you the next pieces of transcript that Tommy V asked me to focus on as they're all related. Alex says, there's an over focus on correlation. So people will say, oh, the correlation of, you know, X and Y were high recently, and you're assuming that they're going to be low, and so you have a failed assumption there. Correlation is just a manifestation of the environment. There is an over focus on specific numbers. And then Mike says, you, don't even have to know the correlation in the short term. Just know that they're sort of generally not correlated in the long term. People will be very precise about, oh, well, they're correlated not over a decade or five years. And I'm just like hand to forehead. Alex says, you,'re much better off just looking at fundamental relationships that you know are true by trying to put all these inputs into an optimizer. You're trying to create precision out of something that's not precise. and that's very interesting. I mean, you might get the impression from hearing that that all of our discussion here about correlations is be thrown away, but that's not really what it's saying. What it's saying is that looking at short-term correlations, you need to take those with a grain of salt because some of these asset classes can be correlated for months or even years at a time. So if you're only looking at a small segment of data, you're not really getting an overall impression. If you look at decades worth of data, as they're suggesting though, then you'll have a good idea over the long term what the long-term correlations are. The other thing they are referring to there is that these correlations are a result of the different asset classes performing well in different economic environments or performing poorly in economic environments. And I'll link to something from Bridgewater that you can look at to see a little quadrant map in these sorts of things. And they're saying it's really those fundamental relationships that matter. And I agree with that. The only problem with those fundamental relationships on a little quadrant map is they're not precise enough even to know what's correlated with something else. So having a correlation number is a very useful thing to have. You just need to decide how meaningful it is. And if it's a correlation number over decades worth of data, you can rely pretty strongly upon it. If it is a correlation number that is only for a small segment of data, then you need to think about, well, is this reflective of long-term relationships or is this just a short-term anomaly? And they are also getting at the idea that you shouldn't market time correlations. You shouldn't look at what the correlations were between two assets were last year and think that, well, that's the environment we're in now. Therefore, I can adjust these allocations or rejigger something to take advantage of that. because that's not really the way it works. But what's more interesting about this is it gets you back to a core data science or modeling principle that applies not only in the area of personal finance and risk parity style portfolios, but in any sort of exercise where you are using past data to potentially model the future. And this is a corollary to what is we call the macro allocation principle. But the principle I'm getting at here is called the bias variance dilemma or the bias variance trade-off and it sounds complicated. I'm going to try and explain it in a simpler way. I will link to a Link in the show notes with some dart boards that may be more visceral to some of you to see how this works. Now the trade-off that is going on here is if you have very few data points you're looking at, that is called a biased data set. And so that in the future is likely to give you very similar results, but those results may not be anywhere near what you want them to be. So they're all going to be clustered on one part of the dartboard, but it may not be anywhere near the center where you want the darts to land. On the other end of the spectrum, and seems like we're always talking about spectrums here, that's what the background and science gets you. If you have too much data, or lots and lots of data points that you are trying to fit together, the danger then is you're going to have a high variance that in effect your model is not going to predict the future. It'll predict the past really, really well and really, really accurately, but the variance will be very high and its likelihood of predicting the future will be very low. So how does this play into creating portfolios? The way it plays in is it tells you that if you have just a few funds or asset classes, they might not perform that well, but at least they'll be predictable or more predictable in the future, especially if you have a lot of data that goes in the past. And so a good example of that would be a standard 60/40 portfolio where you're just taking a stock fund that covers the whole market and a bond fund that covers, that's just Treasuries or it's a total bond fund. But in either case, that gives you something that's relatively predictable in terms of what sort of performance you can expect in the future out of it, given all the data that there is in the past. But it's not going to be optimal. It's not going to be optimized in any real sense of the word. Now, on the other side of the spectrum, if you are going through and picking a whole bunch of different funds, particularly if they're in the same asset class, and you might be aware that, for instance, large cap growth funds have been particularly outperformers in the recent past. And if you decided, well, I'm going to ride with those, the chances are that you are optimizing for that period in the recent past and the variance will be high in terms of guessing whether that will work very well in the future. Typically it doesn't because what you end up is with reversion to the mean that things that have performed well in a recent past period are more likely to revert to average performances in the next period, just as a matter of statistical logic. It's very easy to over optimize any given portfolio, especially within a short time period. And this is the trap that most amateur investors fall into when they are given some kind of analyzer or back tester. They just keep putting different things in there to get the best performance in the past. Once you've done that with so many different things, you're really going out on the variance spectrum. So the likelihood of that doing as well in the past is very slim, slim to none, in fact, if you have many, many variables. Now, where does that leave us? in risk parity style investing. And they talk a little bit about this in that video that I'm going to link to. But the idea is picking more broad asset classes and just a few of them. So you are more optimized than say a simple 60/40 portfolio or something like that. But you are not going to the other end of the spectrum where you're over optimized with very specific funds or stocks or different kinds of precious metals, for example. And that's the balance. That is the trade-off between the bias and the variance, between the underfitting and the overfitting. And that's why we try to have portfolios that are relatively simple in construction. Most of them just have four, five, six funds. We do have the Risk Parity Ultimate just as an example of what goes out on one spectrum, but still that one, when you look at the asset classes, it's still just got four or five different asset classes put together there. And that also ties back into our principles, which we talk about in episode seven, and 37, those being the macro allocation principle, the holy grail principle, and the simplicity principle. The macro allocation principle and simplicity principle both go to this. Simplicity principle saying, don't put more stuff in there than you need to construct your risk parity style portfolio, because the more variables the more funds you put in there, the better the chances are that it will not perform as it did in past periods. You'll be going out on that variance spectrum. And the macro allocation principle says that what's really important here is those big allocations between stocks, treasuries, gold, et cetera. What's not really important is those allocations within the asset class. So fiddling around with lots and lots of different funds to get some kind of maximal performance is likely to get you too much variance in the future and it will not perform as it did in the past if your basis for picking it is some recent kind of performance. This also does remind me of Big Earn's critique of the Golden Butterfly portfolio, which we talked about back in episode 40. Basically, his critique was he thought that adding the small cap value to that portfolio was in fact an example of over optimization, because he did not believe that it would perform as well as it did in the past. So you can see how this all fits together when people are talking about these various variables or different funds or asset classes going in and out of these portfolios. A lot of it has to do with this bias variance trade-off issue because you have to gauge whether something is likely to give the same performance as it did in the past period and whether it is likely to have the same relationship with the other assets in the portfolio that it did in past periods. Now, hopefully that was not all too confusing for you and some of these links in the show notes will help you get your mind around that if you need to take a look at them. But I do thank Tommy V for submitting that email with those questions because we're really getting to these core principles and why Portfolio construction works the way it does and why these risk parity style portfolios fit pretty well in that bias variance trade-off. There isn't too much bias and there isn't too much variance. And by getting in that middle part of the spectrum, we're able to get portfolios that perform in a more optimized way than traditional portfolios. but not in such a way that we cannot believe that they will continue to perform that way in the future. But now I see our signal is beginning to fade. I wanted to thank Chris and Tommy V for their questions. I've got a few other ones floating around in here that will be making podcast episodes out in the future. But if you have questions or comments, please send them to me by email Frank@riskparityradio.com that's Frank@riskparityradio.com or you can go to the website www.riskparityradio.com and fill in the contact form there and I will get your message that way I wanted to thank all of my loyal listeners we have crossed 5,000 downloads which is pretty good for an unadvertised amateur podcasts like this one. Amateur in the best way, I hope. If you have not done so already, I'd appreciate it if you go to iTunes or Stitcher and leave me a five-star review so that other people can see this or get a handle on it if they're interested in these sorts of things. And we'll be picking up this weekend with our portfolio reviews of the sample portfolios that you can find at www.riskparadioradio.com. Thank you again for tuning in.

Mostly Mary [25:29]

This is Frank Vasquez with Risk Parity Radio signing off. The Risk Parity Radio Show is hosted by Frank Vasquez. The content provided is for entertainment and informational purposes only and does not constitute financial, investment, tax, or legal advice. 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.