Episode #180: Rodrigo Gordillo, “This Craftsmanship Perspective Is About Identifying The Difference Between Complex Versus Robust”

Episode #180: Rodrigo Gordillo, “This Craftsmanship Perspective Is About Identifying The Difference Between Complex Versus Robust”






Guest: Rodrigo Gordillo is a Co-Founder, Managing Partner & Portfolio Manager of ReSolve Asset Management. He has co-authored the book Adaptive Asset Allocation: Dynamic Global Portfolios to Profit in Good Times – and Bad as well several whitepapers and research focused on adding new insights to the quantitative global asset allocation space. Rodrigo began his career on the institutional side with John Hancock before transitioning to the ultra-high net worth space at a boutique wealth management firm. Subsequently, Rodrigo, along with his partners, Mike and Adam, continued to evolve their quantitatively focused investment methodology as Portfolio Managers at Macquarie Private Wealth and Dundee Goodman Private wealth before launching ReSolve Asset Management in 2015.

Date Recorded: 08/01/19

Run-Time: 1:16:41

To listen to Episode #180 on iTunes, click here

To listen to Episode #180 on Stitcher, click here

To listen to Episode #180 on Pocket Casts, click here

To listen to Episode #180 on Google Play, click here

To stream Episode #180, click here

Comments or suggestions? Email us Feedback@TheMebFaberShow.com or call us to leave a voicemail at 323 834 9159

Interested in sponsoring an episode? Email Justin at jb@cambriainvestments.com

Summary: In episode 180 we welcome back our guest, Rodrigo Gordillo. Meb and Rodrigo start the conversation with a walk through Rodrigo’s background and his experience growing up in Peru. Rodrigo then gets into his framework for thinking about investing and how that evolved into what he and his team is doing at ReSolve.

Rodrigo then spends some time on the knowledge gained by studying and backtesting investment strategies. He stresses the use of “ensembles” rather than isolating single parameters for more robust investment processes.

Meb shifts the conversation and asks Rodrigo to talk about ReSolve’s machine learning project. Rodrigo discusses applying machine learning to finance, and how it is a tool, and another element of the ReSolve team’s process.

Meb and Rodrigo chat about risk parity, and some of the common misunderstandings that exist, as well as the basic functions of how the strategy works.

As the conversation winds down, Rodrigo gets into some research projects on the horizon for ReSolve.

All this and more in episode 180.

Links from the Episode:


Transcript of Episode 180:

Welcome Message: Welcome, to “The Meb Faber Show” where the focus is on helping you grow and preserve your wealth. Join us as we discuss the craft of investing and uncover new and profitable ideas, all to help you grow wealthier and wiser. Better investing starts here.

Disclaimer: Meb Faber is the co-founder and chief investment officer at Cambria Investment Management. Due to industry regulations, he will not discuss any of Cambria’s funds on this podcast. All opinions expressed by podcast participants are solely their own opinions and do not reflect the opinion of Cambria Investment Management or its affiliates. For more information, visit cambriainvestments.com.

Meb: Welcome, podcast listeners. We have an awesome show for you today in person, after having our guest and his entire team with us for episode 17. It was our first and only menage a quatro episode. And if that doesn’t get you a hot and bothered, we’re gonna be talking about quant finance, which should really get you going. Our guest is co-founder, managing partner, portfolio manager of ReSolve Asset Management, co-author of the book “Adaptive Asset Allocation” as well as all sorts of white papers, including one we worked on together. Began his career on the institutional side, John Hancock, and then subsequently with the rest of his crew, which you guys will remember worked at Macquarie, Dundee Goodman before launching ReSolve, 2015. Welcome back to the show, Rodrigo.

Rodrigo: Thanks for having me, Meb.

Meb: It’s good to see you in person, man. How is the eFoil? Will you tell our listeners what eFoil is, by the way?

Rodrigo: Oh, my God, it is God’s gift to man. I think it’s the most exciting piece of machinery to come out for anybody that pretends to be a surfer, likes the ocean, or even a lake. I’m born and raised in Peru and my adolescence was spent surfing the waves. I was literally five minutes from the ocean for years. And when I was forced to move back to Canada, and everybody told me about the Great Lakes, I was like, “What have I gotten myself into?” I haven’t been able to do anything in the ocean for years and the eFoil comes along.

Meb: Explain to the audience.

Rodrigo: It’s like a surfboard that then has kind of like this foil that goes two to three feet underwater, and it looks like the wing of a plane. It’s an eFoil because it’s purely electronic, you have a trigger in your right hand or left hand or whatever, and it has a little motor that you press and depending on how much you press, you’re gonna go faster or slower, it starts getting you off the ground. If you can get up on your feet, you’ll be on the board, two or three feet off the water and doing all types of fun tricks and go as much as I think 40 miles an hour.

Meb: So I’ve seen Laird do it non-electrical power surfing.

Rodrigo: He’s the Godfather.

Meb: Okay. And that’s pretty awesome and fun to watch. You know, I actually saw him shoot the pier once in Malibu when we were out there. I was on my like Costco board playing in the shore but watched him shoot the pier. Which listeners, if you’re not a surfer, means you go through the pylons on the pier on a wave, which I would 100% just faceplant into a bunch of barnacles. But I’ve also seen him surf in Hawaii too. So I can claim I’ve surfed with Laird twice.

Rodrigo: Have you seen him surf the foil?

Meb: I think they had some whole crew out doing it when I was there. And watching some guy just fail over and over again trying to do it, which was made me feel better about myself. But it looks like a lot of fun.

Rodrigo: Oh, it’s so much fun. Honestly, you can get up on it really quick if you have any experience on a board. It took me about I don’t know, half an hour to get it and then after that, it’s a matter of…

Meb: What do you do? Do you start on your stomach like you would like a normal…?

Rodrigo: You start on your stomach, you go a little slow, then you get on your knees, and then pick up the speed a little bit. Then you get up on your feet. Try that for about 10 minutes and then you start pulling the trigger high enough, the board gets off. And then it’s a matter of centimeters, inches as you like to say in America. You don’t wanna move around too much. It’s a delicate balance. And it’s like the coyote in cartoons. When you’re feeling it, when you’re in the zone, you’re going, everything is perfect. The moment you feel like, “Oh, my god, I’m flying,” you look down, you’re gonna faceplant.

Meb: Is the main thing where you just kind of dip your nose and then faceplant, is that usually…?

Rodrigo: You dip your nose and faceplant and again, it’s one of those like your eyes point down, you’re done. You gotta go. You gotta feel it.

Meb: That lesson took me a long time learning how to wakeboard, the looking down where you end up going. This is a really dumb question, but its one that seems like you’re going 40 miles an hour theoretically and the bottom of that foil must be like a missile. Have you ever, like, crushed and murdered any poor fish, any dolphins? I mean…

Rodrigo: No, I killed a coconut one though. We were with…so a part-owner of the firm of Lift, which is this eFoil company is Jerry Szilagyi, president of Catalyst Funds and Rational Funds, we run one of their funds. And we went to the Context Conference in January. I get there, he calls me up and he says, “Are you in the conference?” We didn’t know we were going to the same place I’m like, “Yeah.” “You wanna skip a couple of meetings and go eFoli?” So we went to the…there’s the ocean side, which is a bit too choppy, and there was a bay side and the bay…

Meb: This is Miami?

Rodrigo: This is in Miami, right on, I think it was the W Hotel. And there’s a lot of palm trees, a lot of coconuts that are falling into the water. So I did faceplant while a coconut got lodged into the propeller.

Meb: That’s kind of the successor conference to the old AlphaMetrix and there’s one other, the Future Focus.

Rodrigo: I think SALT is the other.

Meb: Okay. SALT’s in Vegas. All right, enough about that. I wanna hear a little bit about your background. I think the listeners would appreciate it because you grew up in Peru.

Rodrigo: Yeah.

Meb: Good surf breaks in Peru. It’s a lot of lefts, isn’t it? Are you regular or goofy foot?

Rodrigo: I’m regular.

Meb: Okay.

Rodrigo: There’s the longest left in the world.

Meb: I had a bunch of friends going and I said, “Look, I have a hard enough time riding regular backside for me, I don’t need the longest left in the world. Like I’d love to go down and hang out but…” So I missed that trip. Tell me about Peru. Tell me about growing up. You were just a quant out of the womb?

Rodrigo: That’s right. Actually an aspiring quant. My father was a math professor. He was a naval officer in the Peruvian Navy. He was also a professor at the University of Lima. Me and my three brothers, whether we liked it or not, were gonna be forced to become mathematicians. It was fun being at the dinner table after dinner, then the math equations would come out and we’d compete against each other.

The thing about Peru at the time, though, was I was there from the time I was born and left in 1988. So I was there for my first eight years and saw a lot of crazy stuff. ’88, like Lima, it was the first time we’d heard about the terrorism, you guys might have heard of Shining Path, Sendero Luminoso. They were mostly out of the provinces but when they made it to the city, it got really, really dodgy. A lot of bombs going off, constantly being in blackouts. And I was a kid, I didn’t mind that, thought it was fantastic. Family would congregate. Nobody else had anything to do but play board games. It was amazing. But that combined with some of the economic issues, which was the Peruvian currency was pegged against the U.S. dollar. We were reneging on IMF funds, inflation got crazy out of control. And so the combination of those two things made us emigrate. We had two options, we were either gonna go to Australia or Toronto and sadly, my parents chose Toronto. I’m just kidding. I love Toronto, but…

Meb: I’ve never been, I’m gonna be there this fall. So you’re in charge of showing me around if you’re there, if you’re not at your co-worker’s new compound relocating to the Caymans.

Rodrigo: Exactly, yes, the Cayman Islands. We’re trying to get back to the beach. Toronto has been great to us but it’s a little too cold.

Meb: Right. Well, Australia, it seems like that would have been…Australia, to me, feels very similar to California in many ways.

Rodrigo: I’ve never been.

Meb: Never been. It’s wonderful. Lovely place. Big fan, love going there, been a few times for work, oddly enough. All right, so you guys decide to go to Toronto.

Rodrigo: Yeah, I mean, the reason we left was because it was a crazy time. And I don’t know how much of this is things that I remember and things that I pieced together with time interviewing family members. But it was the first time I saw the power of macro-economics could really affect your life. A few of the things that happened, which was kind of crazy, was that inflation was always high. It’s like Argentina right now, you’re kind of used to have 20%, 25% annualized inflation. You go to a bank, it’s kind of fairly stable the rate of inflation, you get a 20%, 23% rate of return on your GIC or your interest-bearing bonds.

Meb: You get a pretty good lesson on the difference between nominal and real…

Rodrigo: That’s right.

Meb: …returns at a very early age.

Rodrigo: That’s right. But in 1988, inflation in Peru, after we reneged on the IMF loans, went from 23% to 7200%. And some of the things that I remember was my grandfather had recently retired, and he’d send a letter to every family member saying, “Look…” he was an accountant his whole life. He had said, “You guys will never have to worry about me, I have a million dollars, equivalent U.S. dollars in Peruvian soles,” is I think it’s what it was at the time. “So I’m taken care of for life.” And after that year, his million dollars turned into the equivalent of zero U.S. dollars. That was the first I remember having the discussion around the family members that we had to go back to work, we all had to kind of help him out.

And then the contrast to that was my next-door neighbour. He had reneged on his mortgage. He owed this huge mortgage, like a year earlier, they were about to evict him. He had a few U.S. dollars, actual U.S. bills under his blanket. After that same year, he was able to pay off his mortgage, which was hundreds of thousands of Peruvian intis with like a couple of hundred dollars U.S. dollars. So the debtors won and the savers lost in a massive way. And that led to a mass exodus of anybody who had any sort of education into everywhere around the world, Australia included, not us though.

Meb: All right, so Toronto, you bought a coat, winter gear, what was the career arc then, you wanted to be a mathematician?

Rodrigo: It was more of the same observations. I remember getting there in 1989-’90. And my father bought our house with zero money down. It was a massive bull market in housing. And so we bought it zero money down and then the housing crisis hit. The housing crisis that every Canadian has forgotten about. But housing prices in Toronto were down as much as 50%, 55%. And after a few years there where I saw my father struggle through putting his own money in the markets in 1990. He lost so much money in Peru that putting money to work in small-cap equities in Canada right before the recession meant that he lost more money.

And then we ended up moving back to Peru in ’94 where we had to sell the house at a 50% loss with money that we didn’t have. It’s just like, all these things are happening. And my dad, being a math professor, I would have thought would be smarter about the dollars and cents side of things but he wasn’t. But it was certainly a formative experience for me. I went back to Lima, did my high school years there, nothing to really talk about except for it was quite an exciting time. It was really fun down there. You can do a lot more things in Peru as a teenager than you can in Toronto.

And then I came back to the University of Toronto, I went to Rotman, did my commerce degree there. And when I got out, I didn’t see any other way to attack the market but from a macro perspective. I didn’t understand why everybody needs to focus on individual stocks, everybody just trying to beat the S&P. My experience was such that I wanted to kind of mitigate against global disaster. And then the background in math and quant, I mean I can’t be a 26-year old that can gather 25 years of experience ala Warren Buffett, I needed to figure it out mathematically. So I did. So I focused on asset allocation quantitatively from the get-go.

Meb: All right, so you started out in really finance from the beginning. So hopped around to a few different places, what time did you pick up with the three amigos?

Rodrigo: So 2004, I get into the business, John Hancock briefly, then I worked for a family office, we were running $2.5 billion. Trying to find the best managers on the planet to put together, more like an analyst role of funded funds. I wanted to get more hands-on, I was already running some strategy myself. So I decided to go off on my own and gather some assets as a lowly Peru immigrant that knew nobody. I had to smile and dial, grab as many people, mostly attorneys, as I knew they had money. And I went around and talked about quantitative investing and systematic, rules-driven, not letting emotions get in the way. I told them my whole life story.

Meb: What I know about Canadians, that sounds like a horrible strategy. I feel like you should show up and just be like, “I have a few junior miners, I’ve got some…”

Rodrigo: You have no idea. It was at a time too, like it was 2006 when I decided to do that on my own. And I was in the middle of the massive commodity bill. I remember going out and my buddy was a geologist, an Argentine that went to the University of Toronto with me, he would take me around places and all he did was those mining deals. And it was so crazy. When you go to a bar, everybody was doing deals left, right, and center. They’d look at you and they’d say, “You know, who’s this guy?” “Oh, he’s a buddy of mine, Rodrigo.” “Well, okay, I’ll give him some free shares for now, but you got to talk it up.” And I was like…people were just giving you free shares left, right, and center. I was terrified of the whole thing. I didn’t understand it so I refrained from participating. But every time I refrained from participating and I saw the numbers, the Eric Sprott commodity fund.

Mb: You were finding warrants all over the floor, just somebody just dropped these warrants.

Rodrigo: Always pitching was, “I don’t want the tech crisis to happen to us again. And why do you want to have 25% annualized return when you can get a steady 8%?” It was tough but managed to gather enough assets to make a go at it warning of the possible scenario similar to the tech crisis. And lo and behold, 2008 came, it all happened. And that’s where my career really went through the roof, made some money for clients that year, and then everybody came to me at the wrong time.

Meb: After the fact, right?

Rodrigo: I think it was the single-digit positive potential clients. And then the next year, everybody who had thought about investing with me invested with me right before my strategy does 7% the following year and Canada is up 35%.

Meb: People wish…

Rodrigo: So immediate.

Meb: People buy what they wish they had bought.

Rodrigo: Exactly.

Meb: And vice versa.

Rodrigo: And then so you asked me when I met Mike. I was 28, so I was on my own for all those years. And in 2011, Mike and Adam come to Macquarie. I remember the night they came it was the night and they moved over from another bank. Adam comes up to me and says, “So what do you do?” And I looked at him and I said, “You know what? It’s a bit too complicated. Why don’t you come back to my office tomorrow, we can talk about it.” He looks at me in the eyes and said, “Why don’t you try me?” That night we spent like six hours until 4 a.m. whiteboarding. It was like in a room full of horses, you know where the zebras are. It was the first time I found two people that were thinking the same way I was. That night, we partnered up and I moved over to their area, and the rest is history.

Meb: That’s so funny. So what was the original inspiration for you guys to cruise out on your own?

Rodrigo: Well, Macquarie, we were managing money internally at the bank, we were already managing money for other advisors. It was an employee managed account, it was kind of this overlay program that we called Adaptive Asset Allocation. They were gonna create a separate Macquarie arm for asset management, we were gonna run the asset management arm and get all the distribution for our product. Took a couple of years to get that through to Sydney, Sydney approves it. And then I think it was late 2014, they come out and say, “We couldn’t do it in Canada, we’re out.” And so they left Canada. And it gave us the opportunity to figure out whether we wanted to be part of another bank or have a go at it. And we’ve been publishing enough content by then that we recognized that the demand really came from Americans. And Canada was still if you weren’t talking about gold and mining, and today, marijuana stocks…

Meb: Cannabis and real estate.

Rodrigo: That’s right.

Meb: Real estate started to cool off, hasn’t it?

Rodrigo: It’s starting to very, very slowly, you can actually bid on a house and possibly get it versus bidding on 14 houses and then paying 150% more than you expected. We published enough and realized who our audience was. And we said none of the Canadian banks were ready to get into the U.S. We decided to go off on our own, get our licensing and build what we’re building.

Meb: All right, so you guys do all sorts stuff now, private fund, mutual fund, wanna talk a little bit about the investing framework because you definitely do a few different things. We touched on some in the first episode, number 17, which we will link in the show notes. But I love this paper it’s titled “Global Equity Momentum: A Craftsman’s Perspective.” What’s the thesis? Talk to me about this paper.

Rodrigo: It’s grabbing a popular strategy from a guy that we love and respect, Gary Antonacci, he wrote a book.

Meb: He’s been on the pod.

Rodrigo: He has. And he basically put out a very simple approach to extracting returns or minimizing your drawdowns from being invested in U.S. equities, global equities, and then getting out when things are not kind of in a downtrend and getting into bonds. And it’s very simple, you’re basically executing I believe it’s over a 12 month period, if the S&P 500’s above over the last 12 months, then you keep investing in it. If the global markets are above the S&P 500, then you invest in those. If both of them are below zero, then you go to bonds. Fairly simple, very popular across the planet actually. When I speak to Peruvian advisors and I ask them what they do, they use Gary’s process.

Meb: Low turnover, probably.

Rodrigo: Very low turnover, simple to implement, this word simple, simplicity. And then what we do is on the other side of the spectrum, if you were to see how it is that we trade our products. If we were to have a separately managed account product that is similar to this, I think it includes emerging markets, we trade more often, we’re actually looking at multiple signals. And people see it as, they say, “Well, that’s complex, that complexity leads to fragility.” And they couldn’t be more wrong when you look at the data. This craftsmanship perspective is about identifying the difference between complex versus robust. And really, it starts with this idea that one specification could really hurt you in your lifetime.

The original specification of 12 months using 2 asset classes for equities and 1 asset class for bonds will work over 100 years. It’ll extract momentum, it’ll get you out most of the time, it’s a perfectly viable solution. But if you’re hired as an asset management firm, we wanted to take that trend approach to its limit in order to maximize the chances that we’re not gonna get unlucky. And so what really matters to most investors is how we perform over a 2 to 3-year period, not over a 100-year period. And with the simple specifications, you can go, you can get chopped up in a single year and get into a lot of trouble.

Meb: Here’s a very real-world, simple example that I think it’s probably one of my favourite examples of this where if you are a trend follower using something similar call it, you know, long-term simple moving average, you say, “You know what, I’m gonna invest in stocks when they’re above the 200-day, gonna be out when they’re below 200-day, it’s worked kind of forever,” you roll into the 1987 crash. And if you were in, I think the 200-day or shorter, you would have been in cash during the crash, I could get this specifically wrong. But if you were in the 200-day moving average or longer, you would still be invested in stocks during the crash. And so it is a very binary outcome if you have clients or you have a fund that one day you’re either 0% return or probably actually up a bunch because you would have been in cash and bonds or down 20%. When you have a strategy that across probably all the parameters, 10-day, 50-day, 100-day, 200-day, 300-day, you would have had a more blended outcome.

Rodrigo: That’s right. A portion of your strategy might have been in, a portion might have been out. The likelihood that that one outcome might have hurt you badly and your clients badly or everybody where your getting fired and your clients are getting pissed off at you is actually mitigated. So I like to use this analogy of the black hole image that we took recently as humanity. We’ve finally seen an actual picture of a black hole. And when you read the headlines, it says that the image was captured by the Event Horizon Telescope.

Now, if you read the headlines you thinking, “This single telescope captured that image?” Well, it’s actually not a single telescope. It’s hundreds of telescopes in eight global sites, that each individual telescope captured a very, very poor image of the black hole. If you really squinted your eyes, you would have kind of seen it, looked like the black hole that we imagined in Hollywood or what mathematicians kind of said what an event horizon should look like. But it’s very faint, it’s a very faint signal. So again, it works but not as precisely as it should. What they needed to do is they need to aggregate the data of these hundreds of telescopes into a site. Took them a couple of months to put it together, all the error terms were eliminated. And then once they had aggregated all the data, they were able to provide a high-fidelity image of that black hole.

Well, this whole papers about that. We start with, here’s a single specification, 12 months, that’s a single telescope. It does okay, you can kind of get momentum from that. But what we did is we actually iterated how we define momentum. So we said, “What if it’s not 12 months, but rather 1 month, or 2 months, or all the way up 18 months?” We also did moving averages, so the 1-month crossing the 18-month, the 2-month crossing 12-month, and so on. And so we just randomly created a wide variety of specifications.

And instead of looking at each individual backtest and saying, “Oh, I like the 10th percentile backtest, let’s go with those.” We said, “Let’s just put them all together.” Because we don’t know, we can’t tell whether it’s gonna be the six-and-a-half-month lookback that’s gonna be the best performing market over the next five years. That’s a really tough gig. If you just use the ensemble, if you use them all, similar to the Event Horizon Telescope, what you end up capturing is a true more high-fidelity signal of that trend strategy.

Meb: By the way, I 100% agree with you and love this. And I think it applies to not just momentum, but really any approach, value certainly is one where if you’ve been using price-to-book, you’ve really been sucking it up for going on decades now. But if you use enterprise value, you’ve been just fine. But going forward, which one’s gonna work better, who knows? And so I don’t know why anyone would read this paper, or along the same theory, any of the research on value and decide not. I mean, I have a theory that we can debate as to why. Why would anyone not use an ensemble? What’s the pushback you hear?

Rodrigo: There’s a couple.

Meb: Okay.

Rodrigo: So the first one is that it’s too complex. So for people who have read “A Simple Strategy” and apply it for their business, they won’t be able to do this. It requires a lot more trading, it requires some programming. It’s just too complex, and I understand that. But as an asset manager, if you’re a quant and you’re saying, “This is too complex,” then it’s just a matter of what is it that you’re saying no to? Are you saying no to it because it kind of takes away the value of the strategy that you gave a name to? So there are a lot of quants out there that have established a specific parameter set and have given it a name or have done enough marketing on it to say, “Hey, that is the way to do it,” rather than…it’s really you’re going against the ego of everybody in our industry. We’re trying to say that we’d rather be broadly correct about capturing that specific signal than specifically wrong.

And when it comes to value investing, there have been careers and multi-billion dollar businesses built on a specific parameter, price-to-book, for example. But as you mentioned, I think FactorResearch came out with a fascinating research report that had the U.S. market, it had the backtest performance of price-to-book, enterprise value, price-to-sales, price-to-earnings. And what you saw is price-to-book is at the bottom of the heap, slightly outperforming over the last 20 years against the S&P 500. So still extracting value. But I think the best performing one was enterprise value, don’t quote me on that. But when you go to the European market, price-to-book is at the bottom again. And then you have these other ones really outperforming these other parameters outperforming price-to-book.

And so you would think, and this is kind of a meme right now, is that price-to-book is dead, that it doesn’t…Well, if you go to Japan, people to take offence to that because price-to-book is the best performing backtest. And all the other specifications are at the bottom of the heap. What’s interesting across the board is that by combining all of them in equal way, the multi-parameter approach does better than all of them.

Meb: It just feels like a non-compensated concentration risk. I mean, the example we give when we talk about valuation is you say a lot of times there’s a structural…you mentioned some of the different geographies…structural reason why some of them may be biased. So for example, we talked about Australia earlier, their tax code incentivizes companies pay high dividends. So if you’re just a sword [SP] on dividends, you’re always gonna get a bunch of Aussie stocks because they have typically a different structural reason.

And my theory, because we believe in composite ensemble, whatever you wanna call it, across everything we do, I think…laziness is not the right word, but I think people like to gamble, and having one parameter…everyone thinks in binary terms about everything. It’s like, “Hey, I have this stock, I have an investment in gold, should I sell it, should I stay in?” And if you have a binary indicator, like the 200-day moving average, you’re either in or you’re out. And then you have something to cheer for. But if you’re only 25% allocated, what do you cheer for? Are you cheer for adding more or selling it? Like it’s like, people love certainty.

Rodrigo: They like binary outcomes. So should I or shouldn’t I? If your answer is always ah, a little bit, maybe just take a tad and then kind of ease into it, nobody wants that. The things that sell are the ones that are certain that you’re going out and saying, “Here’s the specific things you need to do.” And anything that’s nuanced becomes much more problematic. This is super-nuanced. This does not guarantee that you will be the best performing strategy. In the paper, we show how I think it’s like 1,200 different specifications so the idea of 1,200 different strategies. The ensemble isn’t the best one, its 90th percentile Sharpe ratio. There’s a dream that, “Well, my specification is 95th percentile. I’m gonna create an index based on that, I’m gonna market that one.” It’s easy to explain. So I’ll add another thing is marketability. I can tell you firsthand, this is not easy to articulate. It’s a lot easier to be a big index provider and say, “Here are the two rules. Here are the rules, this is what you’re gonna need to do.” They can regurgitate it to their clients or clients can buy-in and you’re in, so there’s value there.

When you’re saying we have 1,200 different ways of looking at this problem, and you’re in the 90th percentile Sharpe ratio, but what about that 95th percentile one? Well, that is luck, we can’t be sure that that one over the next 5 years is gonna be in the 95th percentile again. But we can be more sure that using the ensemble is gonna provide us a smoother ride in the time period that we need it to matter.

And the paper goes through the backtest that was done for the book, which I think was 1970 to 2012. And it had that specification, the 12-month had a return of 16% low drawdown and whatnot. What we did is we extended the backtest to go from 1950 to 1970 and then from 2012 to now and saw what that specification did during that period using a bootstrap method. And all of a sudden, the return went back down to the median of all strategies. As expected, it wasn’t the 95th percentile outcome, it was the same statistically speaking. So the other benefits of using ensemble is that you minimize those periods of bad luck. The drawdown is in the 99th percentile, best drawdown is in like the smallest drawdown of any single specification. The same thing for the worst five-year average period.

And again, what’s beautiful about the ensemble is that if you don’t use ensembles and you’re faced with 1,200 different virtual managers that you need to choose from, you need to choose a manager in hopes that five years from now when you look back, it’s at the very least better than the 50th percentile, at least better than a coin toss. How difficult is that? Using ensembles means you don’t have to make a choice upfront at all, you just get the 90th percentile or more without having to be biased in any way.

Meb: And then on top of that, theoretically, if the parameters really are no better than the others or you don’t know to time which ones are, if you’re using a manager that only uses one, for example, that manager will survive, in which case the money will flow to that manager because they look brilliant. But often is the case the chances of that single parameter doing well in the future, like you just mentioned, could be again a coin flip. And all of a sudden, like, you probably would have been better, who knows, going with the short parameter.

Rodrigo: What’s tough is let’s say we do much more than momentum. But let’s say in this particular example, you’re a momentum manager, you’re using the ensemble methods. There’s always gonna be another momentum manager that people are pointing to say, “Well, you guys are both momentum, he’s killing you. I’m gonna do what he’s doing.” How do you bridge that gap of understanding? The paper’s trying to really discern robustness, anti-fragility versus possible luck risk.

Meb: Here’s another real-world example. I mean, our very first paper was written with a single parameter and have received oh, I don’t know, 10,000 e-mails since. Well, people say, “Meb, so in your paper when you’re using…” In the paper, we used the 10-month moving average. In the paper, we also showed it was irrelevant which one you used, you could use 6-month, 12-month, doesn’t matter. Over the years, received about 10,000 e-mails about, “So, read your paper about the 200-day moving average.” I’m like, already that’s wrong. They’re like, “So when you said…” Like literally, the description of the paper was so funny, because it says, “You buy when the price is above and you sell when it’s below.” That was like the entire…could not have been any simpler. Every e-mail says, “So do you mean you sell when it’s like 1% below?” Because everyone starts, they wanna introduce…they’re like, “It’s too black and white.” They wanna introduce their own…

Rodrigo: Their own twist to it or the magic for it.

Meb: Their own discretion, yeah, their own twist, their own whatever. And I said, “No, like, what are you talking about? The rules could not be more clear.” But I said also, “It doesn’t matter which parameter you use over time. What you should be doing is using a blend or using whatever else.” But that’s no fun. You can’t cheer for that again.

Rodrigo: It’s also not only that, that everybody wants to have their own stories. Like, “Well, that was a good paper, but I wait an extra day and make sure that there’s volume before I make the trade,” whatever it is. They’re also trading a single asset class more than…people love to trade the S&P 500. And so this goes into the idea of like Corey Hoffstein talks about the ingredients and recipes. The ingredients is what is the strategy or the signal you’re trying to extract, in this case, momentum. The recipe is what you put into it, asset classes, how often you rebalance, what your lookback is.

The guys that are doing a single asset class, again, over time likely to work. But it’s only happened a handful of times in history if you’re talking about the 200-day moving average. Adding more asset classes increases the breath, improves the recipe. I know that you do a lot of different asset classes for your approach, that minimizes the chances you’re gonna be specifically wrong. And then beyond that, looking at different lookbacks, that increases the chances that you’re gonna be broadly correct, and so on. So you can improve the recipe by simply increasing asset classes and doing different iterations of that particular signal. And few people like to do that. they like to…it’s just too tough to explain. They like the parsimonious approach.

Meb: It’s funny you mentioned because we spend so much time thinking about the challenges of getting investors to comply. And it’s actually so many advisors or investors we talked to, they want that simple answer. You mentioned it like they wanna be able to say, “Yeah, this fund does X.” And there’s been some very successful product providers that put out products that just have one rule, a switch, and people can explain it despite the fact it’s probably very suboptimal. But it potentially aligns better with careerists. So who knows?

Rodrigo: We’ll see. After we launched this paper, Corey Hoffstein was also writing a bunch on ensembles at the same time independently. And he also talked about the global equity momentum stuff. And we got independently demands for somebody to create rules-based strategy in what we were publishing. So we actually got together with Corey from Newfound, and us and we created an index, selective index called the Newfound/ReSolve Robust Equity Momentum Index, just kind of went out. We’ll see what type of reception we get if people really embrace the ensemble or prefer the 200-day moving average.

Meb: What’s the big overview of the theory or composition of the index? How’s it work?

Rodrigo: Again, it’s really people that liked global equity momentum and that approach just said, “I want that but better.” People who just really like trend, really just like to focus on two to three asset classes. So it’ll be, the index includes developed global equities, developed U.S. equities, a little bit of emerging market so I think it’s capped at 25%. And then if nothing’s working, you go to treasuries, that’s it. But the recipes included are multiple. So half of them came from Corey’s thought process, half of them came from ours. Again, like you say, it doesn’t really matter, we just put a bunch of things in the pot. They’re all virtual momentum managers that are trying to find momentum in these four asset classes. We’re giving the audience what they wanted, what they asked for. This is a fun project. And we’re actually working with an ETF company that’s contemplating launching this as an ETF, so.

Meb: Good, I’d love to see it.

Rodrigo: We’ll see if it beats out the 200-day moving average companies.

Meb: Funny, I love it. And does it end up going all in one asset class? Or does it mix?

Rodrigo: So what you’ll find, if you look at the transition map in the paper, the original specification…again, it’s 100% certain. You’re either 100% of the S&P, or 100% and global equities, or 100% in bonds. When you look at the transition map of the multi-ensemble, you never find or you rarely find 100% of the single asset class. When it comes to investing, when have you ever been 100% certain of anything?

Meb: I’m always certain and almost always wrong anytime I’m certain. That’s why I’m a quant. I was like, anytime I feel myself, the emotions creeping in, it’s like the best contrary signal.

Rodrigo: Oh, I know.

Meb: When you feel the urge of whatever the pull of crypto in January 2018, whatever it may be.

Rodrigo: Oh, yeah, not 100%. And the key thing here is that what this does automatically is it infuses a sense of humility to the process. You see the transition. You see that the systems together, all the aggregate boats never agrees or rarely agrees 100% that you should be holding one of the four asset classes. And then infusion gives it a little bit of humility and makes sure that it’s not a complete blow up at any single time. Now, this is a fun project and it is for people that like this process. It’s a better version of that. But of course, there’s ton more you can do, as you mentioned earlier, the ingredient, there’s ingredients beyond momentum, and then there is asset classes beyond these four. Most of our work goes in that direction.

Meb: Right, infinite variants. Yeah, we’d start to marinate on some tickers for you guys. ROB robust, that wouldn’t be a good one. ROB. ROBU for robust, ROBU.

Rodrigo: We spent an inordinate amount of time thinking about future tickers. I can’t, I’m not even gonna talk about it.

Meb: We could go on for hours. We love thinking about that and the world of reserving and launching funds. One more comment before we kind of skip on. It’s funny going back to this concept of binary decision making. So often…I was having this conversation with a friend the other night, and she has a lot of our net worth in one stock. She’s like, “Meb, what I do with it?” I’m like, “Do you want my honest advice? Or just do you wanna be told what you want to hear?” She said, “Give me your honest advice,” I said, “You should sell it all and diversify or sell most of it.” And it’s one of the most famous high flyers in the U.S. So obviously, they wanna keep it because they’ve made a bunch of money. And I said, “Wow, this is going on.” You know, I said, “Look, best advice, I’ll tell you what, sell half, go halfsies.”

And I’ve given that advice to hundreds of friends over the years, I guarantee you zero have ever done that. You have too much allocate or something like, sell a quarter of it, sleep at night. That way you don’t have the hindsight bias. This also applies to, I think, dollar cost averaging or investing all your money today, people stress so much about that. I’m like, “Just spread it out over a year then.” But they don’t want to do that. They wanna cheer for one or the other.

Rodrigo: We like clarity. And I think on the stock side, I give the same advice. I said, “Look, you’re part of a company, they’re giving you stock options and stock in that company. They’re allowing you to sell. Anything that you can sell, you should sell right away. Because do you believe in diversification? Is that a thing that you think makes sense? Imagine what happens to the stock when the company starts doing poorly? What if it does really poorly, what’s gonna happen to your career and your job? You’re gonna get fired, you’re gonna look at your savings, it’s gonna be wiped out, and you’re gonna have nothing to lean back on.”

And no matter how often I have that conversation, the answer is inevitably, “Yeah, but you know, I really understand, it’s not gonna happen to me. It’s not gonna happen.” I had this with my brother who was at Lehman Brothers right before the crash. I sent him an e-mail, I have an e-mail in ’07 saying, “Look, I’m looking at the data here. I’m looking at the technicals. I know you have a lot of stock options in Lehman.” He was a director. He’s gonna kill me for even putting this on your podcast. But I said, “You got to get out of this or at least start selling out.” He sent me back like a five-page e-mail dissertation on how long Lehman Brothers had been around, that the risk department is showing these statistics, there’s zero chance that it’s gonna go under, we don’t need governmental support. And it just went away for him.

Meb: You would think what’s interesting about this is that the lessons you guys learned early in life with concentration, in this case, to a particular currency. Say, “Look, it applies to any part of your life, diversification on so many things.” Whether the vast majority of listeners put all their money into one house, that’s probably 75% of listeners’ net worth, and then add on all the other problems that leverage and everything else makes it even worse. But that challenge is the old Chinese proverb, “Fish see the bait, not the hook.” They see the gains, they see the upside but never think in terms of walking through, like, what’s your life like if this goes to zero?

Rodrigo: Cryptocurrency. I mean, I could not get the 17-year-old kids that were worth millions.

Meb: You seem like you would be a huge crypto proponent. No?

Rodrigo: Look, I love the idea. I love the theory. I could never pull the trigger. I just bought some now because Philbrick, Mike Philbrick, he finally found like…he’s been a technical analyst forever. He wasn’t touching it until it lost 90% then they bought it and I’m like, “All right, buy some at this point.” But this is probably totally luck.

Meb: By the way, trend ensemble were, I assume…I mean trend, in general, looks like it worked fantastic on crypto, but you may need just a little shorter timeframe.

Rodrigo: We had actually a longer timeframe. We explored launching something like that a couple years…not even, a year-and-a-half ago. The issue ended up being we had $100 million ready to go, Canadian investors. But the issue was custody. Mike’s brother, he actually does security detail for…what’s the guy’s name from Ethereum, the founder? I can’t remember. And he was saying, “You don’t want to do that,” because you’re gonna have to store it in these…and the moment they know where you’re at, the amount of people kidnapped, or they kidnap your family to try to get that coin, like you cannot have direct access to the coinage. And so a custody solution needed to exist first before we felt comfortable and went forward with it. [inaudible 00:39:42] a Canadian company called 3iQ has launched a fund that passively holds the cryptocurrencies, they figured out the custody approach. So that’s where we bought ours, but diversified ensemble of currency.

Meb: There’s a lot of index funds that have popped up, I always laugh because they’re all like 3% per year fee. I’m like, I get that there’s probably extra fees involved in custody and everything else, but I love the at least the theory that crypto…And I’m a huge cheerleader, I’ve never owned any I’m not particularly interested in crypto, but as the sort of libertarian-leaning side of me, I love the concept. But I also love the idea of Wall Street taking every chance to just creep in where you’re like crypto is a decentralized revolution and then Wall Street is like, “You know what, let’s put on some 3%, 3 and 20 management fees on top of this.”

Rodrigo: We’ll take care of the custody, let’s make it institutional. But you know, for third-world immigrants like myself, it’s a great tool. I mean, you can bypass a lot of the 10% fees to transfer money internationally. A lot of third-world nations are exchanging goods based on that crypto.

Meb: That was always the challenge why I missed the boat on the first place is I’m not the use case. But the people that are in some despotic regime that can’t get their money out like it makes 100% total sense and then I’m sure there’ll be a thousand use cases anyway. Crypto guys, don’t out me after this and hate on me for not being a cheerleader. I was tweeting about it the other day, I’m like, I’ve been a cheerleader for many years, but they always get really angry. We still have the huddle [SP] ticker, I think. So that was a good one, no intentions of a launch product.

But anyway, all right, off-track as usual. All right, so ensemble makes a lot of sense, if not too sensible as an approach. And Jim O’Shaughnessy was writing about this and what works on Wall Street, he went through his every single factor. He’s like, “Here’s price-to-book, back to the ’60s. Here’s enterprise value, da, da, da,” and he put together composites and the composites almost always blew away everything else.

Rodrigo: Didn’t he also say the nine-month…I think in one of the early iterations, he actually said, “Out of all of these, I think the nine-month evaluation period was the best one, we should use that one.”

Meb: I don’t know.

Rodrigo: I don’t know.

Meb: It doesn’t ring a bell. Anyway.

Rodrigo: Anyway, it’s common sense. This idea of the black hole is using informational sciences forever. My father did his master’s degree for the Navy in Monterey, California. Actually, he did two years here. And he always said that operations research was answering a question poorly that would otherwise be answered even worse using statistical methods. And basically, even when you see the black hole image, it’s kind of hazy, even when you’re using ensembles.

Meb: It doesn’t just look like Vladimir Putin’s soul. What’s it a picture of, Trump’s hairpiece? I don’t know. Wasn’t it a young girl who wrote the algorithm for it? I feel like it was a young…

Rodrigo: I wonder how much of that is PR. But yeah, apparently, she’s one of the first to kind of put all the data together. But the point being that you have…this has been around forever for people that actually are trying to win the game of extracting the signal from noise. Our game, this game of finance, the ego is so big. And the story that you build around that ego is so powerful that to say, “I have no view on what parameters,” is…

Meb: Well, let’s be even more sceptical. If you were a devious money manager and you said, “You know what, I want the fastest path to managing $10 million.” You launch one or a series of funds that have a single parameter because you know you’re gonna get an outsized signal one way or the other if something happens. So you’re either gonna look amazing or you’re gonna look terrible. If you look terrible, you got out of business, whatever, thousands of funds close every year. If you look amazing, they say, “Yes, we have rigorous proprietary quantitative algorithm that does XYZ,” and you raise a ton of money. Whereas if you’re average, you’ll survive, but you won’t look as good as so and so and been anointed. I mean, that story of…

Rodrigo: Well, this is happening all the time.

Meb: …investment management forever.

Rodrigo: Not for a specific factor.

Meb: It’s stock-picking, concentrated stock picking. If you picked Amazon or your VC and Uber, you forever look brilliant. If you put all your money into GoPro, you’re in the dustbin of history. Anyway, diversify, people. Let’s talk about some more simple topics like machine learning. You guys have been a bunch of math nerds up there in Toronto and made a push into this world. It sounds like something that you guys have kind of been interested in doing for a while, but it’s becoming much more of the vernacular today. You wanna give us a little overview of what’s going on there?

Rodrigo: Our head quant came from the machine learning world. So actually trying to dispel some of the myths around machine learning, it’s not a silver bullet, it’s simply more of the same. We’re using traditional statistical tools to figure out the world of finance. Machine learning is just more of that. Our head quant came from that world. He was using machine learning to identify the optimal way of extracting oil from reservoirs out in the West Coast of Canada. And so when he came into the world of finance and tried to apply those simple machine learning techniques to the world of finance, it was worse than if he just used simple heuristics. This has mostly to do with the fact that when you’re dealing with Manitoba oil sands or oil reserves, most of the parameters set that you’re dealing with are fairly static. Geology doesn’t move around that much.

So if the world that you’re dealing with or the data that you’re dealing with is fairly static, you can apply a machine learning technique that’s trying to find really complex patterns. Things that we humans can’t do, using our eyeball or simple regressions. Finds complex patterns, really curve-fits the data and comes out with a much better output algorithm. When you bring that to the world of finance, it just doesn’t work the same. I like to use the analogy that when the common machine learning tool is trying to…you feed it a bunch of pictures of cats, and then learns from that database of cats, figures out all the curves and structures. And then all of a sudden, you have this awesome database that you can or machine learning algo that is able to identify pictures of cats. You type in cats, cats come out 99% of the time certainty.

Well, the equivalent in the world of finance is you do that, except that the cat that you identified in your database, all of a sudden the cat morphs into a cat with one ear then it morphs into a pug and then it morphs into an alligator. And all of a sudden, whatever you captured, that data has completely shifted. So it’s a lot tougher to use machine learning in the world of finance, but not impossible. There is a faint signal out there in a room full of noise.

And it took our head quant, Andrew Butler, a few years to kind of get some hands-on experience about what financial markets are. There’s a big difference between academic and what actually works in real life and real trading, to then merge the two in a way that you could build on top of the whole area of factor investing. And try to identify very, very small edges in a world of noise. And the big difference between applying machine learning and just data sets, market data sets versus trying to do the old-fashioned identify the long-term factor premium value momentum trend is that you’re extracting a lot of signals that may be spurious, that may work for a year or two or six months and then go away.

So it may be a very real arbitrage opportunity or a very real signal that happened, for whatever reason, structural reason like an institution decided to do certain trades for six months [inaudible 00:47:11] and then turn it off. The key behind applying machine learning to finance is that you have to be aware that you’re finding a lot of highly curve-fit strategies. And so it’s great for finding a ton of different strategies. But the key thing here is now who’s the governor? Who’s gonna let those strategies into the market? And that is an area that’s not really talked about in machine learning. It is the ability to exclude strategies that are simply not viable in live trading.

Meb: How does it kind of come across your desk, you’re just gonna apply common sense filter? Is it a set of like rules you have about how it works?

Rodrigo: It’s interesting because it is the merger between human ingenuity and using tools like machine learning, and this is not a silver bullet. You can have all the tools that you need to apply machine learning out there for everybody to grab like anybody can get access to the same things that we’re using. Where it really comes down to is how are you using your imagination, your ingenuity, your understanding of the market to identify patterns? So yeah, the first one is identifying a feature set. And the feature set, when we first started, were the things that we kind of knew, moving averages, breakout systems, Bollinger bands, you create hundreds of thousands of these different parameters.

This is, by the way, one way, one tool of using. What you do want to initiate the process by saying, “Don’t just go out and find the patterns, but rather, these are time tested techniques.” Like we did in the “Gem” paper, let’s mix them all up. Let’s allow the machine to use these tools to identify patterns. And let it identify patterns that we wouldn’t have time to do on our own. You can do it the old-fashioned way, program every one of these rule sets, or there are interesting machine learning techniques where you can just say, “Look, find me any patterns using strategies of moving averages between 1-month and 18-month, randomize it, come back to me.” Every week, we get fed a series of strategies that may or may not be viable, the ones that really stand out based on some statistical measure that we want, whether we want to optimize based on Sharpe ratio or number of percentage wins, those go into the what we call the sentinel.

The sentinel does not allow…and this is where the magic really is. It’s not in the machine learning side, it’s not an identifying these things that’s mostly producing garbage. But out of that lot of signals, you’re able to use a filtering or validation process that only lets through those things that make sense from a statistical perspective. And then once you have those in live performance, the other process that is, again, nothing to do with machine learning is just ingenuity, understanding what works is being able to prune those systems that all of a sudden stop working.

Meb: Sentinal, is that an “X-Men” reference or is that something else random, you just come up with a random name?

Rodrigo: The sentinel, so the sentinel is actually a person who stands guard, there is no magic machine learning. The magic is in what opportunity sets it gives you and then how do you filter those out? How do you filter the noise? And oftentimes, when you look at academic papers on machine learning, all these strategies are coming out. That sentinel part that I’m discussing doesn’t even come to play. What I found, I applied this machine learning technique, I’m trying to identify when the S&P was gonna go up or down. They publish, they show phenomenal results, but there’s no validation process, we have no confidence that’s a real thing. The machine just found, possibly, a curve-fit strategy. The magic needs to be in how you’re gonna filter these out. And then what you got to recognize is that unlike identifying factors that are based on behavioral patterns, that human beings are constantly gonna likely continue to do these behavioural errors are you gonna harvest, you just gonna leave value on there forever, leave momentum, leave trend, lead quality strategies out there because you believe in the economic reasoning behind it.

What you got to understand when you apply machine learning, and you get through the sentinel and you have these things that are working, is that you’re probably extracting signals that have no economic explanation. I can’t tell you in hindsight why they’re working, and that they might go away. It’s like Peru, having a fixed exchange rate to the U.S. dollar, a trader that knew that trade, knew how to front-run, the times that the Peruvian central government was gonna buy or sell. But of course, until it didn’t, until it reneged, and then it goes away, and all of a sudden that edge goes caput. So the whole series of the discussion goes into this more deeply. The point we’re trying to make here is that machine learning is just another element of everything that we built up to so far and that you got to be careful what it finds. And you got to be very, very discerning as to what you actually implement, and then prune quickly,

Meb: Has it started to make its presence known in your process of managing the funds?

Rodrigo: It has. So I mean, right now, all of our research goes into if you think about that spectrum of recipes, so you got the ingredients. The ingredient we talked about was momentum, but again, like its momentum, trends, skewness, quality, carry, things that we can identify. And then through this other process that I just described, we’re identifying things that have strategy one, strategy two, strategy 2.1, all those are ingredients. Then you have the recipe. How do you put those things together? We haven’t even talked about weighting schemes and how we can maximize that. But we go through the gauntlet of as many edges as we can get our hands-on, we wanna maximize the breadth of those things, the diversification benefits of all of them. And we wanna be able to go long-short.

So that is our focus right now, it’s futures based only. So it’s just easy to execute program, it’s called the Evolution Multi-Strategy Futures Program, been running for a couple of years now. And every month it gets better. It’s called the Evolution Program for a reason, we started with a couple of strategies and, you know, now we have dozens of strategies.

Meb: I just had this very fond remembrance of not safe for work, somewhat inappropriate, it was on public television, so it’s not that inappropriate, where you mentioned like, it just keeps getting better. There was an old “Saturday Night Live” skit, and it was Vince Vaughn and Will Ferrell and others, tell me if you’ve seen it, we’ll post it to the show notes. But Vince was getting ready to get married and he’s sitting down with his buddies at the bar, Will Ferrell and someone else, can’t remember who it was. And he’s like, “You know, I don’t know. I don’t know if I wanna do it.” And then he’s like, “I’m worried once I get married, my wife is gonna…we’re gonna fight all the time and she’s gonna be nagging at me,” or something. And Will Ferrell was like, “No, after you get married it’s amazing, they get nicer.” Vince Vaughn is like, “I don’t know, you know, a lot of people, once they get married, they kind of mail it in, they stop exercising, they get…” Will Ferrell was like, “No, it’s amazing, my wife’s underwear just keeps getting bigger and better.” So it’s the best clip. Really funny. But so anyway, getting better, cause an improvement.

Rodrigo: Well, the key is that the research really is focused on an arena where we’re not restricted by any sort of regulations and whatnot. We run a 40 Act fund.

Meb: And that’s more of the plain vanilla.

Rodrigo: That’s more like what we started with back in the day, which is a long, flat, tactical, global, it’s like risk parity on steroids, having asset classes that benefit in different economic chains.

Meb: By the way, it’s funny because I have no dog in this fight. But the media, what is it about risk parity that the media and the journalist just…at least you have Cliff out there trying to fight the good fight. But 2018, media is like all over risk parity, and then it’s got to be romping and stomping this year, I assume.

Rodrigo: It’s doing really well.

Meb: Right. So risk parity 2019 is crushing it and you don’t hear anything about risk parity this year.

Rodrigo: Crickets.

Mb: The dumbest is like risk parity is somehow gonna create like a snowball effect. I’m like, “That is the dumbest possible…” Anyway.

Rodrigo: It’s fascinating. We actually did a rebuttal. You know, Real Vision TV, they were just destroying risk parity.

Meb: It’s a Cayman-based.

Rodrigo: It is a Cayman-based thing, we love the guys, and they gave us the opportunity to do a rebuttal and I put together like an hour-and-a-half-long presentation on the myths behind risk parity. And one of them, the first one is the one that bothers me the most is this idea that it’s just a Libor bond portfolio.

Meb: Wait, before we keep going…

Rodrigo: Yeah.

Meb: For those who are not familiar with risk parity…

Rodrigo: Yeah.

Meb: …just give us a quick description.

Rodrigo: So risk parity is this concept that…it was originally…the concept came from Ray Dalio. He runs a lot of alpha strategies. He has Pure Alpha Fund. And he was contemplating what was going to happen the day that he passed away and his trust. And he wanted to put together a way of managing money that didn’t depend on the expertise of his team that he didn’t trust after he died, I guess. But the point is, how do you put together a portfolio where it will do well regardless of the economic regime that we’re in?

The way you got to think about it is that the economic environment is split between two dynamics, inflation dynamics and growth dynamics. High inflation, low inflation, low growth, high growth, and depending on the intersection of those two dynamics, certain asset classes are expected to do well. This makes intuitive sense. If you’re gonna have a period of high inflation and low growth like we saw in the ’70s, well, we know upfront categorically that TIPS are going to perform well because they’re designed structurally to make money in high-inflation environments, and in low growth environments, treasuries tend to do well, so it’s a double whammy.

And in periods of high growth and high inflation like we saw from ’99 to 2007, you’re gonna see commodities and real estate and emerging market bonds do really, really well. 1982 to 1990, when you have reducing or lowering interest rates and high growth, you’re gonna see developed equities and bonds do really well. In periods like [inaudible 00:56:42] or the tech crisis or December of last year when there’s a growth shock and there’s reducing inflation, then you’re gonna have treasuries and gold do really…You wanna have exposure to all these components but the key thing is that you don’t want to equal weight them.

Meb: I think you talk about the…what is it called, the Brown portfolio, the…

Rodrigo: Permanent portfolio.

Meb: The permanent portfolio.

Rodrigo: Right. Harry Brown, perfectly viable solution to the same concept, except that it’s equal dollar weight. The key thing about risk parity is this idea that equities have two to three times the amount of risk or volatility as bonds. But we kind of, in empirical testing, when you look at each one of these asset classes over 100 years, they all have the same Sharpe ratio, different volatilities with the same Sharpe ratio. So if they’re all returning the same Sharpe, then all we need to do is make sure that the maniacs aren’t taking over the asylum, that equities aren’t dominating the risk of the portfolio. So risk parity says, “I’m gonna give more dollars to the low-volatility asset class until it hits the same volatility is a riskier one.” So you weight them in such a way where each one of those asset classes contribute the same amount of risk to the portfolio, where not one of them can dominate the risk of the portfolio. And this is a great concept, except that when you do it, you end up getting that very low volatility. High Sharpe, low volatility.

Meb: I feel like if you could go back, it should have come up with a better marketing phrase than risk parity. It needs something like the whole low vol that has raised a gazillion dollars. Like that’s great marketing, but risk parity, put the word risk in anything, like are you kidding me? But this is a really good description you clearly have done that before.

Rodrigo: I have done this once or twice.

Meb: The one additional point I’d like to make is that if you think about asset classes, and you mentioned stocks versus bonds, there’s really no reason to accept them pre-packaged. Like you don’t have to accept stocks at their level of volatility, they trade out, if you just buy a total nominal position, you could put half your money in stocks and a half and in T bills, in which case, you’re just de-levering stocks. And the good example too is that stocks also have debt so they’re inherently leverage. And there’s no reason except bonds…so this goes back to the old-school capital market line and everything else.

And the thing I would always add to the Bridgewater side is that a lot of other people also kind of came to this conclusion through different routes, the old-school commodity trading advisors in the ’70s. Absolutely, volatility weighted their positions, they had to, otherwise, you’re trading Eurodollars in wheat and lean hogs, how to equalize those positions. And if you even go back further in the research, you’ll find people that have come up with all sorts of different ones. The one we joke about the most is the Talmud, which is like 2000 years old, which dollar weight, it’s three asset classes, but it’s stocks, real estate, bonds, not what they call it but essentially the same thing.

Rodrigo: So here’s the biggest complaint against risk parity is that, so what happens is you get this well-balanced portfolio that has a low volatility and a lower absolute return than the 60-40 U.S. equity portfolio. And so people say, “Well, why would I do that complex equation when I can just do 60-40?” The Sharpe ratio is higher. So this idea of you can’t eat Sharpe, well, that’s…you can’t eat Sharpe if you’re not willing to use leverage, which is a whole other conversation. Well, what they do, what he came up with, he said, “You tell me the risk that you’re willing to take and I will lever that portfolio to the point where we hit that risk target or that return expectation.” And at that point, what you’re doing is you’re levering the whole portfolio, you’re not levering bonds. But because bonds represent the largest portion of the portfolio because of it’s all volatility, the concept is that you’re levering a bond and not levering all of it together.

Meb: Right. And people, their brain just sort of just totally malfunctions when they hear this risk parity. It’s up there with buybacks for me on journalists. I’m not signaling out journalists, just people, in general, commenting on it and this kind of makes me shake my head. I used to own risk parity, I think I still own risk parity and risk party.

Rodrigo: That’s right, yeah.

Meb: The oldest domains you are welcome to have.

Rodrigo: The domain, yeah.

Meb: But okay, so this concept of putting together a portfolio, like once you hear that description, it’s hard to look at portfolio…it’s like, you know, the matrix, taking the pill. Look at it the same way after and go back to another way of thinking, it just makes so much sense.

Rodrigo: This is why I like to tell people the 12 days of investing wisdom. It’s like the “Game of Thrones.” Once you start, you can’t stop because you start…and I went through this myself, right? You start recognizing these gems of wisdom, these ideas that kind of give you a whole paradigm shift of portfolio construction, and you can never go back to anything else. I remember, one of my pet peeves about risk parity was a paper that came out from just a large asset manager that is mainly focused on equities and valuations on equities. And they went after risk parity saying it was just a leader bond portfolio, look what happened to bonds. And I think you mentioned this once or twice, but from 1940 to 1981, bonds from a real return perspective lost 68%, bonds lost 68% in real terms during that 40 years, it was a massive bear market.

And so when you think about if you’ve been told that risk parity is a lever bond portfolio, then you were terrified of that happen happening again. But because risk parity is not a lever bond portfolio, you have an inflation component with commodities and tips, you have a growth component with equities, and you have a kind of deflationary component with treasuries. Those may have different dollar amounts. If you’re levering 200%, you might have 120% in bonds, you might have 40% in equities and 20% in commodities. But during that 40-year period, equities crushed and commodities did really well.

And in fact, the risk parity portfolio is what I show the “Real Vision” podcast, the risk parity component during that period was an upward-sloping line and actually outperformed in 60-40. Moreover, if you were okay…and what they were saying in this paper is like you should be 100% equities because if this happens, that’s the thing that did the best. Well, if you’re okay with taking equity-like risks or volatility around 15%, so U.S. equities, while from 1940 to 1981, it was around 15%. If you were to lever the risk parity portfolio upfront to 15% vol without knowing which asset class was gonna perform best, it outperformed drastically the U.S. equity markets with lower drawdowns and a pretty straight line. So, see what I mean without having to make a decision upfront, you’re doing better. This is the whole concept of risk parity. You say what risk you can take and your chances of being right, in your lifetime, are significantly higher because you’re covering in those regimes.

Meb: There’s one incredible benefit of the risk parity strategy in my mind is it gets investors thinking asset class agnostic. So you mentioned this earlier, where a 60-40 stocks-bond portfolio is 60-4- dollar-weighted but it’s like 90%-plus volatility dominated by stocks just because they’re so much more volatile than bonds. The problem, again, was a marketing decision where you have to barbell problems and risk parity strategy. One, the name, risk parity, and the second being the fact that it often involves something that people don’t understand, which is leverage up or down. And then the third would probably be that if you’re levering it up, it often involves derivatives. And then people, their brain totally…the three of those together, derivatives, leverage, risk parity, because we all know those are bad, those three phrases are bad for long-term capital. But you could risk parity target 4% vol if you wanted.

Rodrigo: Exactly. So you can do not even a non-levered but a risk parity plus cash portfolio. And you’re still more likely to do well when it matters in your two to three to five-year timeframe.

Meb: Bridgewater was smart because they named theirs intelligently. They call it All-Weather. And by the way, we found this great book, I love reading old investment books. And we found one that was like a decade before All-Weather launched. I think the name of the book was called like, “Diversify.” And I never even heard of it. It was probably just randomly picked it up either on Amazon or bookstore. And they outline a portfolio in the book that’s risk parity-like that’s called All Weather. Not saying Ray stole it, guys, I’m just saying.

Rodrigo: What’s hilarious is that AQR launched their risk parity fund, and was called the Risk Parity Fund. And now I think they changed it to Balanced [SP] or something.

Meb: They changed the name, which is brilliant.

Rodrigo: We have an ETF that we run in Canada, a non-levered risk parity ETF. And the biggest feedback from the wholesalers is like, “Please, God change that name.”

Meb: Yeah.

Rodrigo: We just haven’t gotten around to it.

Meb: Put our heads together just like this so mellow, chill portfolio.

Rodrigo: You were asking about the 40 Act, the adaptive asset allocation strategy, it’s risk parity on steroids because that’s kind of just measuring correlations of volatility and doing everything in equal risk contribution and investing in all those asset classes. You and I both know that there are times to be invested and those times not to. So this, the adaptive asset allocation approach, the one that we really launched our careers on, our ReSolve careers on, was a tactical asset allocation strategy that applies risk parity to the asset classes that make it into the portfolio. So we will, at times, completely exclude asset classes and this really takes it, smooths out the light even further, increases the returns that much better. So that from then is when we then evolved into the multi-strategy, long-short, everything portfolio.

Meb: And so it’s August, could you give us any indication of what they sort of look like at all today? Is it just a big amalgamation of all sorts of different stuff? I assume.

Rodrigo: We’ve had a kind of a romping, stomping year. You know, tactical, especially that long, flat tactical in 2018 was terrible. I’ve been writing about this for a while that what’s the kryptonite? I think that’s what’s gonna make this not work. And I always said negative hawkish fed policy is the kryptonite. If you go back to 1994, it is terrible for multi-asset strategies, and why ’94? Because Greenspan came out behind the curve inflation and he started raising rates. I think he raised them from 3% to 8% in 13 months, 3& to 8.75% and would raise 50 basis points [inaudible 01:06:56] without letting the market know. And whenever you hit the market without them expecting it that poorly, every single asset class goes down at the same time, cash is king, all of a sudden, money is sucked away from risk and put towards cash.

It also if you wanna think about it from an academic perspective, the present value is discounted by the cost of money, which is the interest rate. As interest rates go up, your present value goes down. So everything goes down together. And when you depend on diversification and you’re a long-only strategy, your risk parity, it kills you. So if you go back to ’94, brutal for risk parity, if you go back to 2004, the first-rate hike Ben Bernanke’s…no, so the first rate hike by Greenspan after the tech crisis, that month was brutal for every single asset class together high correlation. 2006, the last rate hike from Bernanke, 2013, the taper tantrum in 2018. It was, I think, the first year since ’94 where every single asset class over the last 12 months underperformed cash.

So that is poor for that. But then all of a sudden, so you have this policy shock that affects every single asset class. That bled into a growth shock, so go back to risk parity where you saw in December not everything going down, but just risk on going down and treasuries making a killing. And I wrote in October, I’m like, so generally, if we’re gonna follow history here, I am a quant, I don’t wanna hit my head on this, but this idea that the moment that the hawkish governor realizes he screwed up, everybody’s on him, he’s gonna go dovish, and if you see that in newspapers in ’94, you see it in 2004, you see it in ’06. And sure enough, by the end of December, it starts going easy money and it didn’t matter over the last eight months. It doesn’t matter what you invested in, you made money, treasuries are up, commodities are up, equities are up. So if you’re a risk parity strategy, you’re having a good year. If you’re a momentum-driven risk parity strategy, like adaptive asset allocation, you’re having a great year.

Meb: It’s like a mirror image of Q4 2018 that this year’s been good for a lot of things, risk parity in particular. All right, man, so we covered a lot of stuff today. What else is on the horizon? You guys look out, I know you’re potentially launching some funds. Are there any studies you’re thinking about? You guys also now have a podcast?

Rodrigo: Yeah, we’re finally launching the traditional podcast.

Meb: We jabbered on a while on there.

Rodrigo: Yeah, we had you on, we’ve had five guests so far, mainly quanty, nerdy people with very complex topics, but they’re fun nonetheless for the right audience. What’s interesting about the podcast, like a month before you launch your podcast, Mike, Adam, and I, when we weren’t even ReSolve, we were still part of the bank, we launched our first episode. We didn’t launch it, we recorded our first episode. I took a picture of the Deep Blue microphone and put it on my Twitter feed, and said, “The Butler-Philbrick-[inaudible 01:09:40] first podcast.” And then we just got busy and never launched it. And then like a month later, you launch yours and like, well, we’re too late to the game now. Fast forward three, four years, and it’s like, My God, did we miss that boat?

Meb: Well, it’s great. I think it’s still in its infancy. It’s funny, we struggled for a long time doing the podcast because we wanted to do video, we thought people wanted to consume in video, which turned out to be totally wrong. But then the world’s come full circle where everyone loves video because now it’s so easy to produce and consume. We started out putting out some videos, which humorously like the number one comment is just about that I don’t have a beard in the videos. And then I was like, “This is a really weird style of recording, but we’re gonna try it, experiment for a while.”

Rodrigo: I like it. I think isn’t like YouTube one of the top search engines, when people search for stuff they go for video search first and foremost?

Meb: Yes. Well, welcome to 2019, soon to be 2020, that’s gonna be weird to say.

Rodrigo: I was gonna record…so we have the machine learning podcast with Adam and I. We were going to record and I was gonna put a video on my quant said, “There’s no way I’m getting on cameras.” I wonder how that’s gonna go.

Meb: That’s funny. So okay, so y’all got that going on. It’s a lot of fun at the very least like you’re just chatting with buds about who knows what. You guys working on any new research projects, papers, ideas? I mean, you’re pretty prolific. I mean, I’ll post a shown a link to the “Global Equity Momentum” paper, that must have taken a while.

Rodrigo: Yeah, you have two or three projects on the go. That’s Adam’s baby. And I think thee…I don’t know if I published this one or the portfolio optimization three-part series for practitioners, that was a pretty popular one as well. That one goes into…here’s a key aspect that I think is missed by a lot of practitioners is this idea, going back to the recipes, of everybody wants to improve their edge. So how do I get a better momentum signal? How do I get a better value signal? And it’s really just about which asset classes securities am I gonna hold and which ones am I gonna exclude? And then it ends there. How will you weight them as an afterthought?

We talked a little bit about risk parity. But this is where like, almost 50% of the magic is in terms. If you care about maximizing your return per unit of risk, ignoring the weighting scheme is a big mistake. And that’s what this optimization series is about is saying portfolio construction, what do you think you can estimate? Can you estimate volatility? Can estimate correlation? Can you estimate returns? And depending on what you think is estimable, then you have an optimization technique to use.

But what’s fascinating about this, if you think about it from…let’s just simplify this for a second. If you believe in the equity risk premium, let’s say you believe in the U.S. equity risk premium, you can buy the S&P 500. And that’ll be the signal, your ingredient. You also believe in the term premium, you buy the 10-year Treasury, so you have a 50-50 bond-equity portfolio. That portfolio, same as these two asset classes, 50-50 from ’91 to now, analyzes something like 0.5% a year, 25% maximum drawdown, it’s not a bad strategy. It captures what people believe to be the real reason you should invest in the market. But merely changing the weight, so instead of doing 50-50, you go to risk parity weight, 75-25. Just doing that takes the…I think the Sharpe ratio for the 50-50 is 0.94% during that period. The Sharpe ratio for the 75-25 is 1.23%. So you’re increasing your Sharpe ratio by 30% merely from changing the weighting scheme.

So that’s what this paper is really focused on is you have your skill, how do you maximize that skill, that whatever that skill is? But then instead of doing just equal weight or [inaudible 01:13:10] volatility, be more thoughtful about how you weight things. And by being thoughtful about the weighting scheme, you’re able to increase your Sharpe ratio by as much as 30%. And it has nothing to do with your ability to pick better stocks or pick better asset classes. It has everything to do with just being more thoughtful about how you weight things. Nobody talks about it, it’s a boring subject. When people ask us, “What are you guys proficient in?” Breadth. That’s not sexy. Anyway, we’re trying to get the word out with on that side of things, hopefully, push the envelope a little bit more for practitioners. We have…all the formulas are there for people to take a look at and improve there. If you have an edge and you have a good Sharpe ratio, give it to us, we’ll read our papers and we’ll deliver back the same edge with better performance simply by weighting.

Meb: You hit the road quite a fair amount, where you going to be the rest of the year? Just Toronto and the Caymans?

Rodrigo: Well, I’m going to be in Toronto, hopefully opening up a Cayman office, so Mike Philbrick will take the bullet for the rest of the team. He’s gonna go out there, build the office.

Meb: Such a hero.

Rodrigo: He’s gung ho about building the Cayman compound similar to Joe Rogan, the Joe Rogan compound, the ReSolve Asset Management compound. So if anybody’s out in Cayman and wants to have a good time, and by good time, we mean like cold baths, and eFoiling, and 100-degree saunas and…

Meb: A lot of squats and snatches and clean and jerks.

Rodrigo: That’s how we [inaudible 01:14:27].

Meb: Deadlifting.

Rodrigo: Adam would take offense to that being fun, but for myself and Mike and a couple of other guys in the firm.

Meb: You’re only allowed to eat keto and in between the hours of 5 p.m. and 7 p.m.

Rodrigo: That is correct. A little bit of intermittent fasting, you know, we want to live to 150 so we can then transfer our brains to the matrix.

Meb: You still get to Peru ever? You know, I’ve actually never been. Lima is now like a world culinary capital.

Rodrigo: I go back every year. I got two of my brothers, you know, the Lehman Brothers back in Lima. My younger brother’s there. I got a brother in San Francisco. So we all go once a year and get the family together and eat some amazing food. We have 2 of the top restaurants on the planet are in Lima right now, one is in the top 10.

Meb: That sounds like a great spot for a quant nerd conference.

Rodrigo: Yeah, quant has nothing to do…

Meb: Sponsored by ReSolve Asset Management.

Rodrigo: I’ve tried to go hitting those investor conferences in Lima, and if it’s not private equity or private real estate backed by the government guarantee, there is…you mean you can’t get me a 20% rate of return with no volatility and no risk? No, thank you.

Meb: Yeah. Why so low? People want to find you, they wanna see what you’re up to, they wanna follow in y’all’s writings, ruminations, everything else, where they go?

Rodrigo: Yeah, so we have our blog, which is investorshub.com/blog is where you can find that. Just go to investorshub.com, go to media, we have the two podcasts, one is just the mini-series and the other one is the interview style. And we have our blog post where we post just kind of thoughts as well as summaries of our more like 35, 40-page white papers and, of course, Twitter. I’m in a FinTwit community. I should be more active than I am and have been, but Adam, our CIO, is the most active at gestaltu.com. He doesn’t mince words, likes to get aggressive in there. So it’s a fun stream to watch. Mike’s pretty fun too putting all his videos of his snatches and deadlifts.

Meb: I love it.

Rodrigo: And myself @RodGordilloP.

Meb: Rodrigo, it’s been a lot of fun. Thanks for joining us.

Rodrigo: No, thank you. It’s been great.

Meb: Investors, we will post the show note links to all these, mebfaber.com/podcast. Subscribe to the show on iTunes, anywhere else good podcasts are found. Leave us a review, shoot us feedback@mebfabershow.com. Thanks for listening, friends, and good investing.