[00:00.000 --> 00:10.640] Welcome to Excess Returns where we focus on what works over the long term in the markets. Join us as we talk about the strategies and tactics that can help you become a better long term investor.
[00:10.640 --> 00:21.100] Justin Carbonneau and Jack Forehand are principles at the Lydia Capital Management. The opinions expressed in this podcast do not necessarily reflect the opinions of a Lydia Capital. No information on this podcast should be construed as investment advice.
[00:21.100 --> 00:24.700] Securities discussed in the podcast may be holdings of clients at the Lydia Capital.
[00:24.700 --> 00:30.240] Hey guys, this is Justin. In this episode, Jack and I sit down with our friend Wes Gray from Alpha Architect and ETF Architect.
[00:30.240 --> 00:37.540] ETF Architect, a firm that helps advisors and managers launch ETFs or convert existing funds to ETFs, is seeing massive demand and growth.
[00:37.540 --> 00:44.860] We talk to Wes about the developments in that business. In addition, we talk factor investing, value, momentum, quantitative model tweaks and more.
[00:44.860 --> 00:50.620] Wes shares with us some of the interesting new strategies there running in an ETF wrapper that may be attractive to some investors.
[00:50.620 --> 00:56.460] As always, thank you for listening. Please enjoy this discussion with Wes Gray of Alpha Architect and ETF Architect.
[00:56.460 --> 01:00.460] Hey, Wes, thank you very much for coming back on this day.
[01:00.460 --> 01:04.140] Yeah, thanks for having me. Always honored to be here, guys.
[01:04.140 --> 01:11.340] Good stuff. I think this is the third time on and these are always fun.
[01:11.740 --> 01:24.540] It's fun, interesting, lively discussions. I think today we wanted to kind of try to chunk this conversation out into three different but related to some extent topics.
[01:24.540 --> 01:32.940] Get your view on the ETF business and the landscape and what you guys are building over at ETF Architect.
[01:32.940 --> 01:40.140] We'll kind of step into the factor world a little bit with you and talk about value and momentum.
[01:40.860 --> 01:48.540] Two factors that you guys have spent a lot of time studying and running actual investment strategies through the ETF wrapper on.
[01:48.540 --> 01:57.740] Then I think discuss or talk to you about some of the newer strategies that you guys are launching there.
[01:57.740 --> 02:09.740] What you guys have done is really impressive and amazing. I was looking before we jumped on over a million and a half dollars in actual factor ETF.
[02:10.460 --> 02:16.700] Strategy AUM. Then in addition to that, which is through the separate company ETF Architect,
[02:17.340 --> 02:26.860] $5.6 billion across 47 funds. That's a relatively new piece of business for you in the last couple of years.
[02:26.860 --> 02:30.380] Congratulations on all your success and what you guys are building on there. It's pretty awesome.
[02:31.420 --> 02:39.100] Yeah, appreciate it. I think you said $1.5 million in our ETF complex, which would be also awesome.
[02:39.100 --> 02:44.300] Did I say a million? Yeah, an action. We have a few more zeros behind it now.
[02:44.300 --> 02:52.300] There's a new getting you having a little bit of luck, but it's all good. We get your point. We're growing a lot, which is awesome.
[02:54.140 --> 02:57.580] I think it was, I don't know, maybe like five years ago, like Jack and I came down.
[02:59.100 --> 03:04.140] You were living and working out of Pennsylvania. You now live and work out of Puerto Rico.
[03:04.620 --> 03:09.500] But the company has grown a lot. The offices and you've added a lot of people.
[03:09.500 --> 03:14.460] But one of the things I just wanted to ask you out of the gate is if you could go back and get
[03:14.460 --> 03:21.980] yourself a lesson 10 years ago and getting into this business, what do you think you would
[03:21.980 --> 03:29.260] say to yourself now with your experience? I mean, like all the fundamental things I think we've
[03:29.260 --> 03:34.460] always had just because we learn from the service, right? So do more with less complacency kills,
[03:34.460 --> 03:39.820] grind it out every day and no excuses. So I would keep doing that part of the advice.
[03:39.820 --> 03:48.380] The one area where I think we've had our most scrups and potential issues is in lack of focus.
[03:49.500 --> 03:53.020] And usually that happens when you're smaller and you don't have any money.
[03:53.020 --> 03:57.180] If anyone shows up with an idea or concept, they'll pay you. You're like, I'm in.
[03:58.140 --> 04:03.340] But with the problem with that is that distracts you. It destroys your capability to scale, make
[04:03.340 --> 04:09.580] things efficient. And so every single time we've been faced with a situation where a shiny rock
[04:09.580 --> 04:17.100] showed up and we jumped towards it, that never ended well. And so now that we've kind of decided
[04:17.100 --> 04:22.780] what we want to do when we grow up, at least to at a certain point, we feel pretty focused in the
[04:22.780 --> 04:27.580] business lines. Like we know what we do, we know what we're focused on, and we just execute and
[04:27.580 --> 04:34.620] get more efficient, faster and better at it, that we always win when we do that. It's focused creep
[04:34.620 --> 04:42.380] has always been a problem for us. It's funny, Justin, he just basically outlined our biggest
[04:42.380 --> 04:46.460] weakness of us as a firm. It's like, anybody comes to us and we're like, yeah, let's do that.
[04:46.460 --> 04:50.940] That sounds great. Yeah, let's do that too. Yeah, so we've already learned something we have
[04:50.940 --> 04:56.700] to get started. Yeah, now that said, I'm not really sure it's a solvable, I don't know if it's
[04:56.700 --> 05:02.460] good advice, right? Because in the end, the other problem is you need to make money to stay alive.
[05:02.460 --> 05:08.060] So in, you know, so the other alternative is like, okay, great. So I focus, that's good advice,
[05:08.060 --> 05:14.540] and I'll avoid a lot of issues and problems. But at that point in time, if I didn't take that deal,
[05:14.540 --> 05:20.220] I don't get to eat food or pay for my rent. Well, it's almost like an enigma. It's an
[05:20.220 --> 05:27.180] unsolvable puzzle, but you know, if I could in a vacuum, you know, create a new world and eliminate
[05:27.180 --> 05:32.540] some of the issues that we've had in the past would be focus problems. But at the same time,
[05:32.540 --> 05:37.740] you know, maybe we needed to do that to like, you know, get a few extra shekels in the door just
[05:37.740 --> 05:43.340] to make sure everyone's, you know, surviving. So I don't know. No, I think that's, you know,
[05:43.340 --> 05:49.100] that's very fair, valid. But then how, let's talk about the ETF sort of white label business,
[05:49.100 --> 05:53.580] ETF architect, you know, you guys in the last few years have gone from, you know,
[05:54.300 --> 05:59.260] not having that business to it looks like, you know, it's a very successful business. And like,
[05:59.260 --> 06:06.620] the runway of that looks, the potential of that is huge. Like, how did you, that was a focus quite,
[06:06.620 --> 06:10.860] at some point you had to, you know, you guys had to sit down and say, yeah, let's, let's attack
[06:10.860 --> 06:15.820] that market. Walk us through that, I guess that decision-making process.
[06:16.860 --> 06:23.820] Yeah. So, so actually the reason we fought that for so long was because I was leaning on my lesson
[06:23.820 --> 06:29.820] learned to stop focusing on every opportunity that comes in the door, right? As you guys know,
[06:29.820 --> 06:34.540] you just run an ETF like, you know, everyone asked you like, Hey, can you run my ETF do this?
[06:34.540 --> 06:40.940] And we said, no, no, no, no, no, because we were focusing on alf architect, you know, asset management
[06:40.940 --> 06:47.260] business. And then eventually at some point, you know, Perth kind of convinced us with the first one,
[06:47.260 --> 06:51.900] and we're at a tipping point where we actually thought we knew what we were doing. And it turned
[06:51.900 --> 06:59.900] out we probably did. And the demand and the vision, especially after the new active ETF rule came about,
[06:59.900 --> 07:03.900] which kind of leveled the playing field and opened up a lot more opportunity, which, you know,
[07:03.900 --> 07:10.620] that's around four years ago. That's when we said, okay, this will decided to be a focus creep,
[07:10.620 --> 07:16.940] but we're going to eyes wide open approach it. And we're going to crawl walk around like we do
[07:16.940 --> 07:21.820] everything. We're going to start slow. And then, you know, as we get more successful, we'll layer in
[07:21.820 --> 07:26.700] additional resources additional time. You know, of course, four or five years later, you know,
[07:26.700 --> 07:31.420] the thing's five and a half billion dollars, it'll probably double again next year. And it'll probably
[07:31.420 --> 07:36.300] double again after that. And it'll be like the biggest business we've ever had by a landslide.
[07:36.300 --> 07:42.540] And so that's how we got into that is the deliberate decision to distract from focus,
[07:43.740 --> 07:48.140] you know, because we know that's usually the worst thing we can do. But then, and then once we
[07:48.140 --> 07:52.860] decide to do something, just tactically do what they call like a crawl walk run approach where,
[07:52.860 --> 07:57.820] hey, we're going to crawl, do it right, fine tune it, and then eventually get to a running stage.
[07:57.900 --> 08:01.020] And that now we're at the sprinting stage in that business I'd say.
[08:02.060 --> 08:04.700] Yeah, one of the cool things for me about that story is this idea that like,
[08:05.420 --> 08:09.180] when you have a business and you have an asset and you think about a completely different way to
[08:09.180 --> 08:13.900] use that asset, like that can be a big part of a lot of success stories. Like AWS or Amazon is one
[08:13.900 --> 08:18.220] that people give all the time. And you did that. I mean, you basically realized we built our own ETFs,
[08:18.220 --> 08:22.540] we have this entire infrastructure to build ETFs, we can pivot that and use it to launch other
[08:22.540 --> 08:24.940] people's ETFs. So that's a really cool part to me about the story.
[08:25.740 --> 08:32.460] Yes. And also, it's all about timing as well. Like, and also it's about getting lucky because
[08:32.460 --> 08:37.100] just through the, and this is something we kind of foresaw before everybody, just because I'm so
[08:37.100 --> 08:43.180] tax sensitive, is you know, once the active ETF rule came up, the 60 11 that leveled the playing
[08:43.180 --> 08:51.020] field, automatically index based funds, active based funds, everybody can use ETF rebalancing
[08:51.020 --> 08:57.260] to minimize tax burdens. Well, that's interesting because in the end, we all know anyone that has any
[08:57.260 --> 09:05.020] money in taxable capital should be managing their capital inside an ETF. So this whole market for
[09:05.020 --> 09:11.180] people that want to treat their capital better and more efficiently, it just the demand is going to
[09:11.180 --> 09:16.620] be insane. But Vanguard, I share state street, they're not going to launch ETFs for other people.
[09:16.620 --> 09:23.020] Someone has to be lower the barrier to entry to the ETF marketplace business. And that's, you know,
[09:23.020 --> 09:30.060] essentially what we did. And then it turned out to be true. And so we got lucky that the actual thesis
[09:30.060 --> 09:35.500] is here, you know, could change, but it seems like this is going to continue for a long time.
[09:36.220 --> 09:39.740] Yeah, on that idea of all the use cases, I remember when we launched our ETF in 2014,
[09:39.740 --> 09:43.340] it was basically pretty straightforward. I mean, people were launching their own ETFs from
[09:43.340 --> 09:47.660] scratch. It was all pretty standard strategies. And if you think it to now,
[09:47.660 --> 09:52.220] like, I mean, you guys have done mutual fund, I think conversions, hedge fund conversions,
[09:52.220 --> 09:56.060] SMA conversions. Can you talk about all that in some of the interesting use cases that are out
[09:56.060 --> 10:01.100] there for ETS now? Yeah, so so what kind of got opened up that's different also, it all revolves
[10:01.100 --> 10:07.180] back to the 60 11, which is like the new ETF role that even the playing field is, it's one thing
[10:07.180 --> 10:12.780] to start up a new ETF. And as you guys know, that's the most painful, risky, hard thing to do on a
[10:12.780 --> 10:19.660] planet Earth, right? However, if I already have assets and they're in a non tax efficient rapper,
[10:20.380 --> 10:26.620] man, I would love to be able to get this inside an ETF. The problem is how do I do that without
[10:26.620 --> 10:31.660] paying taxes in the transition? That's what we've solved. And that's what we're focused on. They're
[10:31.660 --> 10:38.780] called 351 transactions were how do I go from SMA? How do I go from hedge fund? How do I go from
[10:38.780 --> 10:44.380] mutual fund, which is not technically through 51, but same with concept and move these assets
[10:44.380 --> 10:50.700] into a better wrapper that's more efficient without causing friction along the way? And as you
[10:50.700 --> 10:55.980] guys know, there's trillions of dollars that are not being treated well in these other wrappers
[10:55.980 --> 11:01.500] that would be treated way better in the ETF tax structure. And so that's what our business is
[11:01.500 --> 11:06.140] about. How do we find these people and then efficiently and with low frictional costs,
[11:06.140 --> 11:11.660] transition them into the more efficient way to manage money in not all the time, but in many
[11:11.660 --> 11:17.660] respects, like equity, tactical, fund, upons, all the stuff that a lot of people use.
[11:18.780 --> 11:22.940] And do you think that'll get like more efficient and less costly over time converting all this
[11:22.940 --> 11:28.700] stuff as you guys gain experience doing it? Yeah, I mean, real time, if you go look at our press release
[11:28.700 --> 11:33.820] on ETF architect every every six months, we have a new press release that says, hey, we're doing this
[11:33.820 --> 11:40.140] to lower cost. And the rack rate used to be, you know, way up here, and it keeps coming lower,
[11:40.140 --> 11:45.660] we've kind of squeezed out a lot of the low hanging fruit. But you know, as you guys know, like every
[11:45.660 --> 11:53.260] thousand dollars here or $2,000, there are matters. And you know, if there's so much regulatory issues,
[11:53.260 --> 11:58.300] it's going to be hard to get it free, obviously. But you know, let's say it used to be like 300k
[11:58.300 --> 12:03.980] soup the nuts and with the 250, you know, now we got it to just under 200 for plain vanilla,
[12:03.980 --> 12:09.340] you know, maybe you could get it down to 150 someday for everything, like literally just show up,
[12:09.340 --> 12:15.340] send the tickers and sell. You know, we might be able to get there in five years, I don't know,
[12:15.340 --> 12:20.460] it's it's ever going to be cheap. But you know, let's say, miraculously, you got the whole cost
[12:20.460 --> 12:26.700] structure for everything I'm talking soup the nuts to like 150k. Well, you know, if you charge 50
[12:26.700 --> 12:31.580] dips, you know, your break even is pretty low. You know, maybe you're 30 mil, right? If you charge
[12:31.580 --> 12:37.260] more, you know, you could do the math, but it's just that it's just it's becomes more accessible
[12:37.900 --> 12:43.020] to a lot more people. And what I usually do is I is I compare like, Hey, ETF, what is it? It's
[12:43.020 --> 12:47.500] basically a tax deferral. Let's be honest, what are some alternatives? Well, there's things called
[12:47.500 --> 12:53.580] insurance reps, like it's called PPLI, PPBA. And they usually charge just to get the deferral benefit,
[12:53.580 --> 12:59.500] you know, they were from 25 to 100 bips. Well, if I have a hundred million dollars, and it
[12:59.500 --> 13:05.580] costs me 200k to do it, I'm way cheaper than the insurance variety. And there's way less constraints,
[13:05.580 --> 13:11.980] right? Even at 50 million dollars, if it's 200k, well, it's 40 bips. That's still probably better
[13:11.980 --> 13:17.260] than the insurance equivalent. So we've already got it cost competitive with alternative ways to
[13:17.260 --> 13:21.500] get, you know, basically deferral rappers, but you still need a lot of money. It's not like,
[13:22.140 --> 13:27.820] unfortunately, like, if you're not rich, you know, you probably can't do your own personal ETF,
[13:27.820 --> 13:33.500] but you can syndicate it, right? Like RIA's, you know, if you have 10 rich buddies, we all have a
[13:33.500 --> 13:39.660] million dollars or $2 million or $5 million. Okay, now we can do a deal, even though any one of us
[13:39.660 --> 13:44.300] couldn't do a deal, it'd be too expensive. Because the good news is you can syndicate these deals
[13:44.300 --> 13:49.740] and then everyone wins by, you know, making a bigger pool of capital and everyone benefits as
[13:49.740 --> 13:55.980] well. So I just see so many opportunities and so much growth. And there's not enough
[13:55.980 --> 14:01.740] people building the shovels to deliver it. Like we're basically the only ones I know that are
[14:01.740 --> 14:06.300] exclusively focused on this. Yeah, it's really cool. You know, like, we're like the cheapest
[14:06.300 --> 14:10.380] people on the face of the earth. And so when we did our own ETF, we basically did every piece of
[14:10.380 --> 14:15.020] work you could conceivably do yourself and like got our cost down to the absolute minimum. And I
[14:15.020 --> 14:19.420] think you're below that now and you're doing all the work. So that's a pretty cool transition from
[14:19.420 --> 14:25.660] where we were back then. Yeah, yeah, exactly. And the whole point is to empower you guys who
[14:25.660 --> 14:30.460] were like the, you know, intellectual property developers, relationship folks like you, your value
[14:30.460 --> 14:37.420] ad is not doing compliance and monkey work and like, you know, smashing buttons. That's just
[14:37.420 --> 14:42.140] crazy. You guys are doing the intellectual property and like the distribution and marketing of the
[14:42.140 --> 14:48.540] product. That's what you should be focused on. And so that's what we're trying to do as best we can.
[14:49.020 --> 14:53.500] So I want to pivot and talk to you. We can't have you on the podcast without talking about a factor
[14:53.500 --> 14:58.860] investing some. And I got to use my PhD at some point. We still have like you were one of our I
[14:58.860 --> 15:02.460] think you were maybe our third interview we ever did. And we still have clips like there were so many
[15:02.460 --> 15:05.660] great clips from that about factor investing that I think we still have clips we're sharing like to
[15:05.660 --> 15:11.340] this day from that interview. So hopefully hopefully we'll get some more again. Yeah, sure. I want
[15:11.340 --> 15:15.260] to ask you about value first because it's actually done a little bit better in recent years, but like
[15:15.260 --> 15:19.820] value at a really long period where it's troubled. And I just want to ask you like if you could talk
[15:19.820 --> 15:22.940] about putting that in context, like thinking about, you know, a lot of people when it was
[15:22.940 --> 15:26.700] struggling, we're saying, well, for this reason or that reason value doesn't work anymore,
[15:26.700 --> 15:30.460] you know, no longer makes sense. Too many people are following it. Technology killed value,
[15:30.460 --> 15:34.380] whatever it is, like there were all kinds of things out there. Like, how do you think about that period
[15:34.380 --> 15:37.580] with value and how do you think about like evaluating a strategy like that when it was
[15:37.580 --> 15:45.100] struggling? I don't think about a lot of things. I just go get the data as far back as I can. And
[15:45.100 --> 15:51.820] just look at it and tell and see what it says, right? And so we always do as best we can try to get
[15:51.820 --> 15:57.660] as much data going back 100 years or 200 years if it's possible nowadays. And just look at say,
[15:57.660 --> 16:03.500] for example, five year rolling performance against XYZ value strategy in the market.
[16:04.060 --> 16:09.980] And you can look at the rolling chart and identify how many times or what percentage of the time
[16:09.980 --> 16:15.420] can I be, you know, looking like an idiot over five year time periods, which is long enough to
[16:15.420 --> 16:21.660] definitively get fired. And in the context of pretty much any factor that I say quote unquote
[16:21.660 --> 16:29.580] works over long haul, every single one of them is it's very, very challenging to stick with,
[16:29.580 --> 16:35.260] because five year underperformance, 10 year underperformance, that's not rare. It's, it's,
[16:35.340 --> 16:40.780] you know, it happens all the time. So when you, when all of a sudden you're in a, in a episode like
[16:40.780 --> 16:48.940] we were, I'm like, well, this sucks. It's painful, but it's not objectively mind-blowingly different
[16:48.940 --> 16:54.460] than the other 30 times that's happened in history. Right. So for me, it's like, whatever, we used to
[16:54.460 --> 17:00.300] write out these articles like dying, investing sucks, you should avoid this because everyone's
[17:00.300 --> 17:04.940] saying it's going to come back. Well, guess what? They always say that it could take another five years
[17:04.940 --> 17:09.740] before it comes back. And just now, finally, that is happening. But you know, I can, I'm not
[17:09.740 --> 17:14.140] going to name names, but there's people that are very famous. They were talking about value making
[17:14.140 --> 17:20.220] a comeback, you know, five, six years ago. And they kept saying it every single year. And, you know,
[17:20.220 --> 17:24.300] finally, this year or last year, they might be right. And that's not because they're not
[17:24.300 --> 17:28.540] cognizant. That's just also there's pressures to kind of like, you know, portray that out there.
[17:29.260 --> 17:35.580] But yeah, I just think it's all these things. You have to go in knowing that underperforming by
[17:35.580 --> 17:41.340] five or 10 year periods is just going to happen, probably, at some point. Yeah, one of the things I
[17:41.340 --> 17:44.940] think about a lot, you know, you mentioned the hundred years of data is for like these backers
[17:44.940 --> 17:49.900] like value and momentum that we follow, if let's say in the future, somehow they did stop working,
[17:49.900 --> 17:53.500] like, how would we know it? Because, you know, we wouldn't have enough like a 10 year period of
[17:53.500 --> 17:56.860] struggle doesn't tell us really anything about it. So like, how would we think about that? Like,
[17:56.940 --> 18:00.460] if there was a day that came that one of these factors didn't work, how would we figure it out?
[18:01.100 --> 18:06.700] So, so the technical way you would do that, rationally would be through like base analysis,
[18:06.700 --> 18:10.540] right? Where you're like, you have, I can't remember the exact terms because I haven't used
[18:10.540 --> 18:14.860] this in a while, but I know the concept. So you have like, you have a prior, right? Well, it's 50,
[18:14.860 --> 18:21.100] 50, the value works value doesn't work. And then I get like tons of data. And then I update my prior,
[18:21.100 --> 18:25.900] I think it's called a posterior. And then it'll shift the distribution given my prior, given the
[18:25.900 --> 18:32.220] additional data. The problem is we have so many samples, we have so much data, so much time.
[18:32.780 --> 18:38.700] And now I have this new posterior. And it's, it's, it's so well established that that's my new,
[18:38.700 --> 18:44.540] like prior. We need to live another, like statistically, if you're using Bayes rational
[18:44.540 --> 18:50.940] updating, we need at least 50 years out of sample data to ever make any sort of like empirical
[18:50.940 --> 18:56.140] conclusion that like value doesn't work, which is, you know, for all intents and purposes worthless,
[18:56.140 --> 18:59.260] right? Like, you know, it let what are we going to do? Like run an experiment,
[18:59.260 --> 19:04.780] well, 50 years, I'm going to be dead prize. So who gives a shit? But so what I'm saying is the
[19:04.780 --> 19:09.580] better way to think about it is like, let's use base there. Right? Okay, all this data,
[19:09.580 --> 19:17.260] this be on a shadow of a doubt that it probably likely that value delivers like some sort of excess
[19:17.260 --> 19:21.740] premium over generic market, right? Can I prove that? Of course not. But that's just what Bayes
[19:22.380 --> 19:28.140] theorem suggests to us. And now at this point, okay, I also know I need to live 50 years out of
[19:28.140 --> 19:33.580] sample, you know, to ever justify if I was right or not, that's like going to happen.
[19:33.580 --> 19:39.580] Well, I'm just going to have to bet on the odds. And then also, if you I also kind of now back
[19:39.580 --> 19:43.980] it into a story as well, just to make sure I stick with it, is I fundamentally believe that
[19:43.980 --> 19:51.020] there's two things that dominate the market fear and grief value is the fear trait, right?
[19:51.020 --> 19:57.980] People do not light ugly nasty may not work, you know, no good throw the baby out the bathwater.
[19:57.980 --> 20:04.540] And I think fundamentally that is value in a nutshell. And I think fundamentally in 10,000 years
[20:04.540 --> 20:10.380] from now, it's also going to be buried in the cake. I think it's always going to be hard to do
[20:10.380 --> 20:16.860] and it's going to go through long, you know, bowels of bad performance. But I just fundamentally
[20:16.860 --> 20:23.260] have faith. Weirdly, I'm an empirical super data science driven person. But at this point,
[20:23.260 --> 20:29.020] I'm faith based, because I need 50 years of out of sample now to convince myself otherwise.
[20:29.020 --> 20:34.860] And so I just what am I going to do? I believe in the value God now, I admit it, you know.
[20:35.740 --> 20:38.940] Yeah, it's interesting. Like, I don't love using like as a factor, like a quant guy,
[20:38.940 --> 20:43.260] I don't love using intuition. But like you said, if you need 50 years of data to disprove it,
[20:43.260 --> 20:47.500] you have to, to some degree, use your intuition and say like, it's something changed here.
[20:47.500 --> 20:50.940] Is there something that I'm seeing that's, you know, is different than the past or, you almost
[20:50.940 --> 20:55.740] have to use that. You can't be like 100% pure quant no matter what. You have to think it through
[20:55.740 --> 21:01.580] some 100%. And one thing that a lot of people that are really smart bring up, which is a great point
[21:01.580 --> 21:07.500] is, well, that all based there and all the rational updating make a lot of sense, assuming you have
[21:07.580 --> 21:13.500] independent samples. But like, let's say I did tons of data mining and, you know, cherry picking
[21:13.500 --> 21:18.940] what have you, you know, well, maybe it's all there's just noise that we're betting on. But I really
[21:18.940 --> 21:23.180] am convinced, you know, from reading basically every paper that's ever been written and doing every
[21:23.180 --> 21:30.380] research study that's ever been done on this myself, that value is real. Like, we can argue
[21:30.380 --> 21:36.060] about the exact way, but buying cheap stuff that everybody hates, I believe fundamentally is not a
[21:36.060 --> 21:41.020] data mining phenomenon. I think it is a real phenomenon. We can argue over it's risk based,
[21:41.020 --> 21:47.260] mispricing, whatever. But but I really do believe fundamentally that it's it's real. It's not a
[21:47.260 --> 21:54.300] statistical fluke. It's not data mined. And therefore I do believe that if we're a Bayesian updater,
[21:54.300 --> 21:59.100] you have to wait 50 years to get disproved. Otherwise, other other things I agree could
[21:59.100 --> 22:03.900] probably be data mined and they're they're overthinking it. So maybe like looking at numbers
[22:03.980 --> 22:10.060] is not really tiny thing. But I do think value and momentum are the two like just truisms of the
[22:10.060 --> 22:15.580] world when it comes to factors. Yeah, assuming value still works, you know, another thing you have
[22:15.580 --> 22:19.740] to think about is sort of evolving your strategy over time, which becomes another balance between
[22:19.740 --> 22:24.220] looking at long term data, but also thinking about things with intuition. And I know your book
[22:24.220 --> 22:27.420] quantitative value is, you know, one of my favorite books and one of the books I learned the most
[22:27.420 --> 22:31.580] about in terms of how to build a strategy value strategy. I was just wondering, like, since you
[22:31.660 --> 22:35.820] wrote that book, comparing that to what you do now on the value side, like, how much is your
[22:35.820 --> 22:38.940] strategy evolved over time? And how much is it very similar to what was in the book?
[22:40.060 --> 22:47.580] Yeah, so so weirdly, like I'm a big fan of Occam's razor, like the simple solution if it's equal is
[22:47.580 --> 22:54.220] probably better. So the book was the best effort at the time. And there was a lot of things that
[22:54.220 --> 22:59.340] were actually pretty complex in there, right? And that's because it made sense. But you know, for
[22:59.420 --> 23:03.500] example, we used to have like all kinds of like distress models and like use like logistic
[23:03.500 --> 23:09.420] regressions, which now apparently they call AI and machine learning. It's like the early, early
[23:09.420 --> 23:14.700] model for that kind of stuff. But then in the end, we said, okay, we have all this complexity.
[23:15.660 --> 23:21.740] What if we just started doing simpler negative screens and can we achieve the same outcome
[23:21.740 --> 23:26.380] but with less brain damage, right? So we moved from a lot of these like financial distress models
[23:26.380 --> 23:30.700] that were really complicated, require a lot of data to like, what if we just bomb out like the
[23:30.700 --> 23:37.340] lowest momenta? What if we just bomb out the highest beta, you know, just stupid simple ideas?
[23:37.340 --> 23:44.300] Guess what? I achieved the same in state with way less complexity. Boom, I'm a buyer, right? So
[23:44.300 --> 23:49.820] so we are negative screening went from really, really complex to much more simpler.
[23:50.940 --> 23:56.060] Just just because it works just as well, if not better, and it's easier to understand and
[23:56.060 --> 24:02.780] simpler to communicate. So that was one key change. The other thing is auto auto like the quality
[24:02.780 --> 24:10.620] spectrum. Same issue, like we used to have like this insanely complicated like quality metric system
[24:10.620 --> 24:15.420] with like hundreds of these damn things. And in any in which you realize is, you know what,
[24:15.420 --> 24:21.580] you know, all this stuff is basically the same. Let's just stick with like a brain dead piotroski
[24:21.580 --> 24:27.420] type score. Let's augment it to make it like even he calls it ad hoc. Let's make it a little bit smarter.
[24:27.420 --> 24:31.900] Like like for example, like one example would be like like net repurchases in the original
[24:31.900 --> 24:38.060] piotroski f score, which I know you guys are familiar with. It's not net repurchases. It's just
[24:38.060 --> 24:44.140] like repurchases. Well, what if you issued like 10 million shares of options? You didn't net buy
[24:44.140 --> 24:49.260] back anything, right? You actually net issue. That's terrible. So we should look at net issue. It's
[24:49.340 --> 24:54.300] not just the piotroski f score version. So there was a few tweaks to it. But fundamentally,
[24:54.300 --> 25:00.140] like like a 10 point checklist that good enough for government work that achieves the same in
[25:00.140 --> 25:07.820] state, but simpler. And so we've actually spent most of our time simplifying our models, not making
[25:07.820 --> 25:12.540] them more complex because it's harder to do that. And you know, so that's what we've done. And
[25:12.540 --> 25:17.420] those are the key differences is simplicity, not more complicated. It's funny. I don't know if
[25:17.420 --> 25:20.940] you remember this, Justin, but on the complexity issue, you know, we one of the things we do is we
[25:20.940 --> 25:25.980] capture publicly available quantitative strategies. And I had like an intern at one point, like make
[25:25.980 --> 25:29.980] a word document with me with everything from quantitative value. And he came back with like it was like
[25:29.980 --> 25:33.900] 10 pages of, you know, some of these like, it was like the bankruptcy score type stuff. Some of
[25:33.900 --> 25:39.660] these scores were like incredibly complicated to implement. So I do like the you you simplified a
[25:39.660 --> 25:43.740] little bit of make work a little easier. Well, yeah. And that's what we always want to ask the
[25:43.740 --> 25:50.220] question like, am I getting am I getting tricked by data mining? Or can this be done more simply
[25:50.220 --> 25:55.900] and achieve the same in state and basically doing the same objective? And more importantly,
[25:55.900 --> 26:00.780] can I communicate it more effectively? Because as you guys know, it's not just West managed in
[26:00.780 --> 26:06.620] his own personal capital. It's we're managing capital on behalf of other people. And if the other
[26:06.620 --> 26:13.580] people in the end don't understand at the fundamental level as deep as like someone who has a PhD
[26:13.660 --> 26:17.740] does, it's worthless, right? Because I got to stick with it where is if we can make something
[26:17.740 --> 26:22.940] simpler, achieve the same objective, but it's easier to understand and easier to communicate.
[26:22.940 --> 26:28.460] So therefore other people can internalize like their ability to stick with the program and understand
[26:28.460 --> 26:35.340] it. Well, that's better for everybody, right? Why not? Like, strategy is only as good as to the
[26:35.340 --> 26:40.700] extent that you can stick with the thing through hard times. So I've always been one of these
[26:40.700 --> 26:44.620] guys who said, well, there's there's no size premium. But when you look at all the other premiums,
[26:44.620 --> 26:48.700] they're magnified greatly, like in the small cap space. And you guys, and I think Jack wrote it,
[26:48.700 --> 26:52.860] you guys wrote something recently that basically said that I'm completely wrong about that. So I'm
[26:52.860 --> 26:56.220] wondering if you could just talk about that a little bit about this idea that, you know, in the small
[26:56.220 --> 27:02.460] cap space, things like value are magnified. Yeah, so so unfortunately, we can't make the claim that
[27:02.460 --> 27:08.620] we've said this because people like Robico, AQR, there's so many smart people who said the same thing.
[27:08.620 --> 27:13.260] But the problem is they set it through factor regressions and a lot of gobbledygook of like,
[27:13.260 --> 27:18.940] well, controlling for this, it doesn't matter blah, blah, blah. So so again, going back to communication
[27:18.940 --> 27:25.340] is how do we communicate this this idea to normal people that are also really smart, but don't have
[27:25.340 --> 27:31.260] times like forensically investigate factor regressions all day? Let's do a simple brain dead study
[27:31.260 --> 27:38.140] then like, like, let's just go look at say the top 1000 universe, the biggest 1000 and the next
[27:38.140 --> 27:44.220] 2000, which is what people actually trade in, right? And let's just compare an equal weight of the top
[27:44.220 --> 27:52.220] 1000, which is still 20 times bigger than the next 2000, but it's just equated. So it's not dominated
[27:52.220 --> 27:58.060] by like Apple, who's 50% of the portfolio. And how does that do against, say, for example, an
[27:58.060 --> 28:04.620] equiweight small or an equiweight evaluated? Well, it's the same. I eat what and then when you do
[28:04.620 --> 28:09.820] the forensic investigation, well, why is that? Because when all of a sudden we go from evaluating
[28:09.820 --> 28:15.820] to equiating the top 1000, which still has an average market cap that's at least a magnitude
[28:15.820 --> 28:20.540] higher than the small cap portfolio, why is the performance the same? Well, there's really two
[28:20.540 --> 28:26.140] answers. One is you basically have a miss specification and evaluate portfolio because you put a lot of
[28:26.220 --> 28:33.340] money like way too much money in a few stocks. And so really it's a mega cap, mega mega mega cap value
[28:33.340 --> 28:38.780] portfolio against a small cap portfolio. That's not real. Let's do the equiweight, right? That's
[28:38.780 --> 28:43.980] more realistic. Second thing is, is then when you look in those portfolios, why are they equal?
[28:43.980 --> 28:50.700] It's because the value characteristic is the same. Because what drives the value effect?
[28:51.660 --> 28:59.500] Value-ness, not size, right? If you say West, here's a $10 trillion company with the PE of five,
[28:59.500 --> 29:06.460] here's a $100 million company with the PE of five, which one will outperform all else equal? I'd say
[29:06.460 --> 29:12.380] they're going to be the same. Why? Because that size does not matter. The only reason that size
[29:12.380 --> 29:19.180] might matter is to the extent that it was accidentally proxying for more cheapness. So whenever you buy
[29:19.260 --> 29:23.900] the reason it looks like it's adding more is like a lot of times you form portfolios that are
[29:23.900 --> 29:29.020] quote unquote smaller. Really what's driving their excess performance expectation is the fact that
[29:29.020 --> 29:35.900] they're cheaper, right? Because the value weighted large cap portfolio has a PE of 30. The value weight,
[29:35.900 --> 29:40.700] you know, small cap portfolio has a PE of five. It has not to do with size. It has to do with the
[29:40.700 --> 29:46.700] fact that PE is different. And once you control for PE, size does not matter, which is what Jack
[29:46.700 --> 29:51.900] showed easily to people and what already way smarter people have shown via factor aggressions
[29:51.900 --> 29:58.460] and long short blah, blah, blah. Or it's just more beta too. That's what like Clif and those guys
[29:58.460 --> 30:04.220] show pretty definitively is, you know, controlling for value-ness. All size is doing is giving you
[30:04.220 --> 30:13.100] juiced beta. But who cares? It's not giving you more value juice. That's not the intuition you
[30:13.180 --> 30:17.340] want to take away. Yeah, it's interesting on this idea of equal weight versus value weight. Like
[30:17.340 --> 30:20.540] this year has been interesting from that perspective because everybody's talking about, oh, values have
[30:20.540 --> 30:24.300] an affordable year this year. But a lot of that is size. I mean, you correct me if I'm wrong about
[30:24.300 --> 30:28.380] this. But it's just interesting, like size is really driving a lot of that this year. And people
[30:28.380 --> 30:31.500] maybe are drawing the wrong conclusion about how value is actually performing this year.
[30:31.500 --> 30:36.460] Well, it's even more confusing than that though, because there's also quality, right? So for example,
[30:36.460 --> 30:43.020] like we use like operating income, EV, EBIT TV, the key thing is any ratio that involves income
[30:43.820 --> 30:49.500] necessarily is high quality because you have to have income. If you look at generic book-to-market
[30:49.500 --> 30:56.220] portfolios, you know, 30% of the companies don't even make money. And yet they're in the value bucket.
[30:56.220 --> 31:01.820] So there's also a big quality component where if I'm looking at like a book-to-market value ratio,
[31:01.900 --> 31:07.900] well, we got to also remember that that's really junk fundamentally. Whereas I'm looking at a real
[31:07.900 --> 31:13.420] value portfolio, a la Ben Graham, like you actually make money and profits, but you're selling for
[31:13.420 --> 31:20.060] cheap on that multiple of those profits, you know, again, it's all about comparing the
[31:20.060 --> 31:24.380] value-ness of what you bought. Like, you know, our quantity value strategy, for example,
[31:24.380 --> 31:29.020] if you look at the overall PE on that, and you compare that to the overall PE of say a
[31:29.020 --> 31:33.740] quote-unquote small cap version, to the extent that the ratios were equivalent,
[31:33.740 --> 31:38.540] the performance is probably going to be similar, all is equal. To the extent that you're looking at,
[31:38.540 --> 31:44.460] again, size that's co-founding this value-ness effect, it may look like it's outperforming,
[31:44.460 --> 31:50.060] but that's just because all the large caps are way overpriced, right? Like if I'm running a value cap
[31:50.060 --> 31:56.860] or value weighted or market cap weighted large cap for the last two years, and I'm forcing myself
[31:56.860 --> 32:05.420] to be in big stocks, you can't get a PE ratio cheap, right? So anything that's not mega cap value
[32:06.060 --> 32:11.500] is going to outperform, it's not really because of size, it's because the value-ness access is
[32:11.500 --> 32:19.580] constrained. So it's the same effect, it's not size, it's mega cap constrains you to not be able to
[32:19.660 --> 32:29.180] access greater depth in the value pool. It's like being a sector-constrained value fund in 99.
[32:29.180 --> 32:33.500] I don't know if you guys remember that, but in 99 people are like, well, sector-constrained,
[32:33.500 --> 32:40.140] I got to have 50% in tech. Oh, the minimum PE ratio is 10,000. How is that value, right?
[32:40.140 --> 32:46.140] It's constrained, you can't buy any cheap stocks. So it's fundamentally a terrible idea,
[32:46.140 --> 32:51.580] it's not value investing. It has nothing to do with like value-ness, it has to deal with
[32:52.220 --> 32:59.180] the weird construction problem. So again, size only matters to the extent that it extends the
[32:59.180 --> 33:07.180] fishing pond and allows you to access cheap names. Then it can outperform, but it's not because of the
[33:07.180 --> 33:14.380] size, it's because you got to buy stocks at a three PE ratio by messing around in the penny stocks.
[33:14.940 --> 33:21.980] That's why, at least what the data says, of why it works. It's that size is itself as a factor.
[33:21.980 --> 33:26.300] If you find pets.com in your value portfolio, you might want to rethink your strategy.
[33:26.860 --> 33:32.300] Well, we do that all the time. Beyond meat is in the vanguard, well, I don't know what it was now,
[33:32.300 --> 33:39.260] but it's super funny. If you go look at certain stocks where any cave person with half a brain would
[33:39.260 --> 33:44.060] be like, well, that's not a value stock. And you go type it in and we have a little tool that
[33:44.060 --> 33:50.140] does that, you'll find all kinds of stocks where you're like, how is Tesla in a value fund? Or how
[33:50.140 --> 33:56.060] is beyond meat a value fund? And they'll be in all these value funds, right? That's because they're
[33:56.060 --> 34:03.660] not value funds, they're just marketing bullshit. And again, if you want to assess a value fund and
[34:03.660 --> 34:09.660] you care about capturing the value factor, it's what we focus on is we want to buy the cheapest,
[34:09.740 --> 34:15.340] comma, highest quality stocks, i.e. focus on the value characteristic. And then to the
[34:15.340 --> 34:19.820] extent you have some leftover brainpower, you know, you all else equally want to buy like higher
[34:19.820 --> 34:26.940] quality cheap versus lower quality cheap. It's that simple. On your point about sectors in 1999,
[34:26.940 --> 34:30.220] that kind of gets at a bigger issue with value. Like, how do you think about this idea about,
[34:30.220 --> 34:34.860] you know, sector neutral value versus overweighting the sectors and maybe you're cheaper or
[34:34.860 --> 34:38.300] excluding things like technology when there are no cheat stocks? How do you think about those two
[34:38.300 --> 34:43.020] different implementations? Well, there's a trade off, right? And so, and it depends what you're
[34:43.020 --> 34:47.980] trying to achieve. So if I'm in the con, if I'm Cliff, as this running long short portfolios,
[34:47.980 --> 34:55.820] you know, I have to sector neutralize. I can't be long energy in short, you know, tech,
[34:56.380 --> 35:01.340] you, you will die. It's impossible from a risk management perspective. So you need to control
[35:01.340 --> 35:07.500] all else equal there. And it's much more important to synthesize kind of your alpha. However, if I'm
[35:07.500 --> 35:13.260] a long only investor, and I'm not worried about trying to capture like after eliminating all this
[35:13.260 --> 35:18.300] other stuff, this synthesized, like, you know, little alpha antidote, and I can be long only.
[35:18.300 --> 35:24.140] Well, now I have a serious problem, right? Because I could go, I could go totally sector
[35:24.140 --> 35:30.860] unconstrained. And it does, it never happens empirically, but you could be 100% the energy patch,
[35:30.860 --> 35:37.740] or I could go totally sector constrained and disable my ability to access cheap stocks,
[35:37.740 --> 35:43.340] because cheap stocks are highly correlated with sectors, right? Like all the happiness is,
[35:43.340 --> 35:47.900] is going to, is going to be in one sector, typically, and all the hate and discontent is using another
[35:47.900 --> 35:52.220] one just on a fundamentally absolute level. And so then, then there's just like a balance you got
[35:52.220 --> 35:57.580] to make. And there's people that have obviously studied this, but in a long only context, I think
[35:57.580 --> 36:03.020] it's an empirically fair statement that if you care about achieving excess returns over a long
[36:03.020 --> 36:08.620] whole, and you, you know, on a just even on a volatility adjusted basis, but you do not care
[36:08.620 --> 36:15.500] about career risk, that's the answer. If however, you tracking error is a major concern,
[36:16.060 --> 36:19.580] and you can't just look at the five year rolling charts and be like, yeah, whatever
[36:19.580 --> 36:24.300] roll with it's my money, I just want to compound my face off, then yeah, so we would need to have
[36:24.300 --> 36:29.500] a lot more sector constraints. But it's fundamentally not improving your investment results on a,
[36:30.140 --> 36:36.220] unexpected return, even sharper just a ratio, in my opinion, in a long only context, long short
[36:36.220 --> 36:42.220] is very different. And a lot of people get confused by these comments. Like we don't say don't ever
[36:42.220 --> 36:48.060] sector neutralize. It's just we say in the context of long only back to investing,
[36:48.060 --> 36:53.020] the bang for the buck by being concentrated and being loose and being allowing yourself to buy
[36:53.020 --> 36:58.460] more cheap characteristic or momentum characteristic is going to make you better off than constraining
[36:58.460 --> 37:03.980] yourself from accessing those securities. I want to ask you about momentum and I want to ask the
[37:03.980 --> 37:07.980] same question I asked about value at the beginning. When you originally wrote quantitative momentum,
[37:07.980 --> 37:11.740] like, how was your strategy evolved? Or is that strategy more similar, maybe to what's in the book
[37:11.740 --> 37:16.380] and has changed less? Or how's that strategy compared to what you were? Momentum, it's easy.
[37:16.380 --> 37:23.500] Buy winners, right? That one, it was already simple. And I don't even know if we've tweaked
[37:23.500 --> 37:28.220] that one that much, to be honest, because there's nothing ever complicated about momentum. It's all
[37:28.220 --> 37:35.740] just price action. Whereas on the value side, it is almost like the curse of knowledge. We geeked
[37:35.740 --> 37:41.900] out too much because we happen to know a lot with, it almost got like blinded by our own complexity.
[37:41.900 --> 37:46.300] Like, well, this has to be valuable because you remember that financial distress model. Like,
[37:46.300 --> 37:50.940] you got to know a lot of stuff to be able to understand what the hell is going on in that thing.
[37:50.940 --> 37:54.540] But then, you know, with a little bit of time and distance, you're like, wait a second.
[37:55.100 --> 38:01.020] Do we really need it to be that complicated to achieve the same goal? No. That's just, I was too
[38:01.020 --> 38:07.020] much of a value. We're momentum with some, you guys probably know, like, I came to that religion
[38:07.020 --> 38:14.460] later on in life. I was always a value person, like 99%. And then I came to momentum, like,
[38:14.460 --> 38:20.460] just through resistance over many years, to just be like, you know what? You're right. That's real.
[38:20.460 --> 38:25.580] We should do it. And so I didn't, I didn't, wasn't that smart about momentum to have to,
[38:25.580 --> 38:29.980] like, alter geek out and come up with something that was overly complex to start out with. It
[38:29.980 --> 38:35.100] always was simple. And so it's still simple. So there really has been any changes there.
[38:35.980 --> 38:42.140] By winning stocks. It's just that simple. I mean, it's just that easy in the end.
[38:42.940 --> 38:46.460] I'm just curious. Do you find momentum, like, we always find momentum harder to explain to people.
[38:46.460 --> 38:50.620] Not because it's not easy. It's actually easier to explain, but it's harder to explain, like,
[38:50.620 --> 38:54.220] why it works. So people like, you know, I'm buying, I'm paying less for a dollar of earnings
[38:54.220 --> 38:58.220] or I'm buying high quality companies or, you know, investors love that kind of stuff. But it's like,
[38:58.220 --> 39:01.580] when you're like, we're just going to buy this stock as it's gone up, it becomes much harder to
[39:01.660 --> 39:07.020] explain. And do you find that as well? Yeah, I've kind of come to understand. I don't like
[39:07.020 --> 39:11.020] anyone knows how any of this stuff works. And actually, there's this brand new paper.
[39:12.700 --> 39:17.580] I'm blanking on the guy's last name is first name is Andrew, where he has this paper that's
[39:17.580 --> 39:23.420] kind of like debilitating because my whole life, you know, as a Chicago school PhD,
[39:23.420 --> 39:29.260] you're learning like, hey, theory, narrative, like financial economics, you need first,
[39:29.260 --> 39:33.500] and then you do empirical studies because you don't, you want to avoid data mining, right? And just
[39:33.500 --> 39:37.820] making something up because it sounds good. And then this guy basically says, listen,
[39:38.620 --> 39:43.420] we looked at all the peer review papers and some of them have super convoluted theories
[39:43.420 --> 39:48.860] backing up the thing. Other ones have mispricing, mispricing theories backing up the thing.
[39:48.860 --> 39:55.580] You could also just data mine this stuff. And guess what? They're all the same, right? He basically
[39:55.740 --> 40:00.860] what he results says is like, I could either data mine what works and add a sample. It's going to
[40:00.860 --> 40:05.980] work a little bit worse than it did beforehand, but it'll continue working. Or I could have this
[40:05.980 --> 40:12.220] like 50 page like physics envy paper backing up the theory of why it works and look at the results.
[40:12.220 --> 40:17.660] He's like, it's all BS in the end. Data mining achieves the same victory,
[40:17.660 --> 40:22.940] which I don't know what to think about that, honestly, but it, you know, it goes going to this
[40:23.020 --> 40:29.180] point, like does it even matter why it works? All we know is it's, it's empirically in that data
[40:29.180 --> 40:31.660] and it's real. So let's just do that.
[40:33.580 --> 40:36.940] Yeah, it's created Justin. I were just talking about the paper like before we got on with you.
[40:36.940 --> 40:40.940] It's Andrew Chen and it's Alejandro Lopez-Leira. I think, I think those are the,
[40:40.940 --> 40:44.380] yeah, please listen to this. I can't remember his last name, but we were actually talking
[40:44.380 --> 40:48.140] about it because we want to have them on because we're like, we have to, this paper is like so crazy.
[40:48.140 --> 40:51.740] We have to like have them on and like dig into it. Maybe we'll, maybe we'll actually come back
[40:51.740 --> 40:54.620] and join us. You can help us interview him because you're a lot smarter about this than we are.
[40:54.620 --> 41:00.700] Yeah, I just think it's very interesting because I, you know, I read papers every single day. I've
[41:00.700 --> 41:05.100] read every paper that's ever been written. It feels like sometimes. And honestly, I've gotten
[41:05.100 --> 41:10.460] bored. It's very rare to find like a piece of new research where I'm like, oh, that's just not a
[41:10.460 --> 41:15.500] proxy or misspecification or shitty empirical analysis to try to tell a story that's fake.
[41:16.300 --> 41:21.500] Like, I like stuff like this because it's like, wow, like he's straight up throwing a grenade in my
[41:21.500 --> 41:27.580] face and telling me to deal with this. I got to think about this. And I like that, right?
[41:27.580 --> 41:34.220] Because now I'm like, oh, that's actually kind of interesting. And it's fundamentally, you know,
[41:34.220 --> 41:39.900] making me think about how I should think about this. So I don't know, you know, what the answer is,
[41:39.900 --> 41:45.020] I don't know if my thoughts will change, but I do find that kind of research compelling.
[41:45.980 --> 41:52.300] But my prior is still, hey, there's greed, there's fear value does greed or sorry,
[41:52.300 --> 41:57.020] does fear momentum does greed. And you can measure in a million different ways. But as long as you're
[41:57.020 --> 42:00.380] approximately correct, and you stick with it, it'll probably continue working.
[42:00.380 --> 42:04.940] Just one more for me. We're handed back to Justin. You know, one of the things you see, like in the
[42:04.940 --> 42:08.060] quad and the factor space, you'll see a lot of different ways to rebalance portfolios. You'll
[42:08.060 --> 42:11.740] see it done monthly quarterly. I mean, sometimes you'll even see it done longer than that. You'll
[42:11.740 --> 42:15.340] see, and then the way it's done on those rebalances is often very different too. And I know you guys
[42:15.340 --> 42:19.340] just changed the way you rebalanced. So I'm wondering if you could just talk in general about how you
[42:19.340 --> 42:23.980] think about creating an optimal rebalancing strategy and the change you made recently.
[42:24.620 --> 42:29.580] Sure. So we've done lots of studies over the years where you do the same factor and just
[42:30.140 --> 42:36.140] rejigger two elements, the frequency of turnover and the number of securities. And intuitively,
[42:36.140 --> 42:41.900] it should make sense. To the extent you have a real factor, if you concentrate more in it,
[42:41.900 --> 42:47.500] it should do better. And to the extent you freshen it up more often, it should do better, right? And
[42:47.500 --> 42:52.540] that's exactly what you find. So the more concentrated and the faster you rebalance it,
[42:52.540 --> 42:56.620] because that you're more in like the value juice. So you're more in the momentum juice,
[42:57.260 --> 43:03.660] gross of fees, gross of operational issues, gross of all that stuff in a vacuum, you would want to do
[43:03.660 --> 43:10.460] a daily rebalance value in momentum fund, right? Now, of course, you don't get to do a daily
[43:10.460 --> 43:16.460] rebalance value fund or a daily rebalance momentum fund, because you would get annihilated in all
[43:16.460 --> 43:23.260] kinds of issues, right? And so now it's an optimization trade off of like, okay, sure,
[43:23.260 --> 43:30.220] I'd love to do a monthly rebalance, like 50 stock portfolio. But if I can't avoid taxes for
[43:30.300 --> 43:35.740] a reason, if the frictional costs are so insane, they overwhelm any net benefit, then I'm actually
[43:35.740 --> 43:41.820] better off maybe doing a lower frequency, higher number security portfolio, right? And so that's
[43:41.820 --> 43:49.900] like the framework for how this all works. And to answer your question, the reason we've adapted
[43:49.900 --> 43:56.780] to higher frequency is we over the years feel like we have way better standing of like frictional
[43:56.860 --> 44:03.100] costs areas. We also have more importantly, we have an ability to handle taxes a lot better
[44:03.100 --> 44:10.060] at higher frequency with the ETF structure, not talking about any ETFs here. But the ETF
[44:10.060 --> 44:14.700] structures, you guys probably know like the ability to use like custom rebalancing, that imposes a
[44:14.700 --> 44:19.820] lot of cost on the ecosystem. And you can't just go ask the street like, Hey, we want to do a weekly
[44:19.820 --> 44:26.540] rebalance custom without without having leakage somewhere else, right? And so if you can find a
[44:26.540 --> 44:33.100] way to get faster frequency of customs, which allows you to ensure like tax efficiency. And it
[44:33.100 --> 44:39.580] doesn't come at like a huge cost. Well, now maybe that's opened our eyes. Well, maybe we could do a
[44:39.580 --> 44:46.620] strategy more frequent and get and have the excess benefits outweigh the excess cost. And so those
[44:46.620 --> 44:53.580] are the kind of things that we've concluded and why we've adapted some of our strategies recently.
[44:54.540 --> 45:01.180] I think a lot of people that are listening to this sort of know you and alpha architect for,
[45:01.180 --> 45:05.820] you know, mostly the value and momentum sort of factors that you're deploying. But you know,
[45:05.820 --> 45:11.420] you guys have some, you know, you've expanded your strategy lineup. And so I just want to kind of
[45:11.420 --> 45:15.660] work through just a couple of these other newer strategies that are now out there and available.
[45:15.660 --> 45:23.580] And the first is this, you know, could you explain what a box spread is and why you guys
[45:23.580 --> 45:30.140] decided to sort of launch a strategy that uses this technique? Yeah, so stepping back,
[45:30.140 --> 45:35.180] like the firm mission is empower investors for education and the firm value proposition is
[45:35.180 --> 45:43.100] affordable alpha. What does that mean? How do we deliver after tax, after fee, after fictional
[45:43.100 --> 45:50.540] costs, weird, boutique, crap, at affordable prices? How do we do Vanguard, but do it profitably,
[45:50.540 --> 45:55.580] right? That's always been our objective. Like we're never going to be $100 trillion firm like
[45:55.580 --> 46:01.020] Vanguard. But how do we do weird niche things and figure out how to do it cheap, tax efficiently
[46:01.020 --> 46:04.860] and affordably? We're not going to be priced at a Vanguard fund because it's too much brain
[46:04.860 --> 46:09.580] damage, but we can make it still a win-win where the net benefits of this thing outweigh the net
[46:09.580 --> 46:15.020] cost of our higher fees than a Vanguard fund, right? So that's high level. We start at factors
[46:15.020 --> 46:21.100] because we knew that. But then obviously over time, you come across other asset classes in areas
[46:21.100 --> 46:27.420] where we can come up with ideas. We're like, how can we deliver that exposure cheaper and more
[46:27.420 --> 46:35.420] tax efficient to folks? And so we explored two venues. One is we've always been very interested in
[46:35.420 --> 46:41.260] insurance type assets, right? Like things that deliver tail benefit. So they're kind of shitty
[46:41.260 --> 46:46.300] most of the time. But then when the world blows up, they make money. Gosh, it would really be nice
[46:46.300 --> 46:51.100] if we could figure out how to do that affordably just on straight up fees and then also figure out
[46:51.100 --> 46:55.900] how to do that without having to pay a lot of taxes because that's part of affordability.
[46:55.900 --> 47:01.820] And so we did a lot of work and teaming up with some folks to figure out how to do the tail risk
[47:01.820 --> 47:07.340] exposure and different strategies and what we think is an affordable after-tax way.
[47:08.060 --> 47:17.500] The other issue is in short duration credit risk, i.e. treasury bill concepts. How do we deliver that
[47:17.500 --> 47:23.660] kind of risk return but do it with the value proposition of affordable alpha? So after fee,
[47:23.660 --> 47:29.180] after taxes, the whole package, can we create a value proposition that other people can't?
[47:29.180 --> 47:37.500] And the answer was maybe. And maybe we can do this because one can financial engineer effectively
[47:37.500 --> 47:43.900] a treasury bill via a thing called a box spread, which is basically, you know, it's like book
[47:43.900 --> 47:48.220] call parity. But essentially you're creating a synthetic long and a synthetic short position
[47:48.220 --> 47:55.100] via option markets. And you're essentially extracting like the borrow and lending rates that are embedded
[47:55.100 --> 48:01.580] inside of the option market. And then we figured out how to, you know, turn that into,
[48:02.220 --> 48:06.940] you know, a strategy that we thought we could deliver with affordability on both after-fee
[48:06.940 --> 48:13.900] and after-tax basis. Not factors, but what we saw an opportunity that took it.
[48:15.180 --> 48:23.580] Yeah, it's kind of, I think, a big opportunity. What are the advantages of like a box spread
[48:23.580 --> 48:29.580] strategy versus T-bells? Are the taxes like different treated differently? How does that work?
[48:29.580 --> 48:36.220] Well, there's a few advantages. I mean, there's a lot of disadvantages and we don't like to talk
[48:36.220 --> 48:40.220] about tax because there's a lot of compliance people out there. And you know, I'm not supposed
[48:40.220 --> 48:45.820] to give tax advice, but I'll just talk high level. So one of the things about box spreads is
[48:46.780 --> 48:50.940] there, and this, you know, you can read the paper. There's a paper called Ristory Rates that was
[48:50.940 --> 48:57.180] published in the JFE where they actually use box spreads as the real risk-free rate,
[48:57.180 --> 49:02.380] and they consider treasury markets as risk-free, but there's a convenience yield because there's a
[49:02.380 --> 49:08.380] lot of forced buyers of treasury bills because the U.S. government has a thing called Basel,
[49:08.380 --> 49:14.220] which forces insurance companies and banks, they have to own that debt to fulfill some sort of
[49:14.220 --> 49:19.820] regulatory requirement. And there's certain times where demand for that kind of stuff goes up.
[49:19.820 --> 49:23.900] And so what they do this really interesting study where they're like, well, is there any
[49:23.900 --> 49:29.580] risk-free table instrument out there that's not affected by regulatory creep and convenience yield?
[49:29.580 --> 49:36.060] Oh, there actually is. It's called a box spread. And sure enough, box spreads actually over time
[49:36.060 --> 49:40.700] for a lot of reasons, mainly because they don't have to deal with forced buyers all the time,
[49:40.700 --> 49:46.060] is they actually earn, in general, excess returns for equivalent duration at the same,
[49:46.060 --> 49:51.820] basically the same risk. It's like an arbitrage in plain sight, in some sense, right? That doesn't
[49:51.820 --> 49:59.820] mean it's arbitrage a bull, but for over time and historically, at any given point, assuming
[49:59.820 --> 50:05.500] there's no frictional cost, the treasury bill pays X, the box spread will pay you more than X,
[50:05.500 --> 50:11.100] and it's going to have roughly equivalent risk and return. It just is what it is. So that's one benefit
[50:11.100 --> 50:18.300] of box spread is the potential for higher returns for a given level of risk. The second benefit is
[50:18.300 --> 50:24.460] they're not affected by the US government as much. So for example, when the government's going to
[50:24.460 --> 50:30.220] shut down, OCC is still clearing trades and you're still getting paid on your box spread.
[50:30.220 --> 50:34.300] The T bill, you may not get paid. And we actually saw that earlier this year,
[50:34.300 --> 50:40.300] where all of a sudden, like treasury bills, they actually had higher yields because the three months
[50:40.300 --> 50:44.380] yield wasn't three months because there was a big risk, you would actually get paid for six
[50:44.380 --> 50:50.700] months from now, right? And so there's another risk that is tied to like politics of, you know,
[50:50.700 --> 50:55.820] treasury and how they pay out with government shutdowns, that just doesn't affect box spreads,
[50:55.820 --> 50:59.420] you know, assuming those those entities are also up and running, which they generally are
[50:59.420 --> 51:05.020] even during the government shutdown. And then so then you get to the tax. Well, again,
[51:05.980 --> 51:11.340] I'm not a tax lawyer, and this is disclaiming everything. You have to do your own work. But in general,
[51:12.780 --> 51:18.780] box spreads are usually done via what they call 1256 contracts, right? So they're under section
[51:18.780 --> 51:25.980] 1256, which is futures tax and index based options, like SPX, right? SPX is typically what
[51:25.980 --> 51:33.980] you do a box spread on. Those are going to be 1256 contracts. So for example, if you may say 5%
[51:34.060 --> 51:41.340] on a T bill and you made 5% on a box spread, there's two key fundamental differences. The T bill,
[51:41.340 --> 51:46.140] one of the good news about it is it's not going to be subject to state tax, which is awesome.
[51:46.140 --> 51:53.100] But obviously it's subject to income tax on the 5% yield. The box spread is going to be 60%
[51:53.100 --> 51:59.580] long term capital gain, 40% short term capital gain, and it's all capital gain in nature,
[51:59.580 --> 52:07.100] which is key. However, it doesn't have state deductibility. So why might that be useful
[52:07.100 --> 52:12.620] at the outset? And there's some additional more advanced things with box spreads that are tied
[52:12.620 --> 52:19.020] to ETF structures that we won't talk about here. But you can go do the math or reach out, we can
[52:19.020 --> 52:23.580] discuss it. But I'm just talking high level box spread, just not talking about ETF or anything.
[52:23.580 --> 52:27.500] And so what could be interesting there, right? Well, one of the really interesting things
[52:28.380 --> 52:34.540] is let's say you have a million dollars in tax loss harvesting shields, because I got destroyed
[52:34.540 --> 52:41.420] in my tech stock in 21 or something like that. Well, the problem is I'm subject to a cap of
[52:41.420 --> 52:51.260] $3,000. I can use my fricking huge loss. Oh, great, guys, let me shield $3,000 of income. Thanks.
[52:51.820 --> 52:57.580] That's all but worthless, if I'm a high net worth. Whereas if I generate box spread,
[52:57.580 --> 53:04.540] it's all capital gain nature, all 5% of it, I can use the whole book of tax loss harvesting
[53:04.540 --> 53:10.940] shields against that effectively creating tax-free income in some sense. So I'm not saying you should
[53:10.940 --> 53:16.700] do that. I'm just saying a lot of financial planners have highlighted that there's just always
[53:16.700 --> 53:22.460] a benefit to the extent that you have a certain nature changed in whatever asset you're doing.
[53:23.740 --> 53:28.300] And it could be interesting for people when they use it. And this is why box spreads and generally
[53:28.300 --> 53:32.620] have historically been sold by private wealth and large institutions to their ultra high net
[53:32.620 --> 53:38.940] work because they use them to facilitate this kind of tax efficiency through the box spread,
[53:38.940 --> 53:45.660] which is it's not exactly a T-bill. But in many respects, it kind of looks and feels a lot like it.
[53:46.780 --> 53:50.780] And for certain people in certain circumstances, there could be some tax efficiency there.
[53:51.820 --> 53:57.020] Interesting, yeah. Just to go back to the tail risk strategy, one of the
[53:57.900 --> 54:04.460] issues some of those types of strategies can have is you can just have a return drag.
[54:05.660 --> 54:10.220] The market's doing really good or flat. You're kind of down or losing money, but then you
[54:10.220 --> 54:16.380] hope that it really works and pays off in the event of a big down market. But how is the
[54:16.380 --> 54:24.060] strategy sort of managing that potential drag? Yeah, so if I could write up like the ultimate
[54:24.060 --> 54:31.260] insurance ever, it would be, okay, how do I get something that makes a lot of money in really
[54:31.260 --> 54:37.420] bad times and on average, it doesn't cost me a thing to hold it? That would be amazing if
[54:37.420 --> 54:43.020] someone could somehow cook that up, right? And also, by the way, even if they cook that up,
[54:43.020 --> 54:48.300] if I got to pay two and 20, and then I got to pay 50% of the winnings back to the government
[54:48.300 --> 54:53.900] anyways, that's also effectively worthless, right? So if I have to pay big fees, well, it's not like
[54:53.900 --> 54:59.020] makes a lot of money during a chaos, and then in good times, it doesn't have a drag because
[54:59.020 --> 55:04.460] I got to pay two and 20. Or let's say it makes 50% in a total blowout, but I got to pay half of
[55:04.460 --> 55:08.060] that back to the government. Well, it wasn't really a great insurance instrument, right?
[55:08.060 --> 55:15.100] So ideal situation is how can I create a product that delivers insurance in total
[55:15.100 --> 55:25.020] chaos zone of like negative 50% death world? On average has no carry, and I can shield it
[55:25.020 --> 55:30.140] from tax drag and I can keep the fees down. That would be the ultimate thing that I would
[55:30.140 --> 55:35.660] personally want to own. Well, that is what I personally was on a hunt and a quest to find
[55:35.660 --> 55:40.860] my whole life. And I finally found it. And then I don't want to talk about ETFs on here,
[55:40.860 --> 55:46.220] but we partnered with a group that I was personally using because I was like, wow, you guys are doing
[55:46.220 --> 55:51.500] exactly what I want to do, but I need to get the fees down and I need to get the taxes solved to
[55:51.500 --> 55:56.940] make this viable for me. And if you guys can solve that problem and we can jointly come up with a
[55:56.940 --> 56:04.220] solution, I love this product. And as you guys know, we only make products and strategies that I
[56:04.220 --> 56:09.500] would personally want to invest in, because we just have a philosophy of skin in the game and
[56:09.500 --> 56:13.740] never asking other people to put their heart and wealth into something that I wouldn't personally
[56:13.740 --> 56:20.460] also put my heart and wealth into. And so that's how that came about. So again, anytime we have an
[56:20.460 --> 56:26.700] opportunity to deliver affordable alpha, where that's just overarching, like how do we deliver
[56:26.700 --> 56:31.500] an exposure after tax, after fee cheaper, better than anyone else out there in the marketplace,
[56:31.500 --> 56:36.620] we're going to go after whether it's factors, fixed income, tail risk, whatever you got,
[56:37.660 --> 56:42.780] and if anyone's got ideas, let me know. We've been trying to do it in managed
[56:42.780 --> 56:48.940] futures, a lot of other things, we haven't solved it, but we're always trying to find areas where
[56:48.940 --> 56:55.020] there's high margin and make them low margin and make them more like vanguardize these things,
[56:55.020 --> 57:00.860] but it's challenging, as you can imagine. I'm just curious with managed futures like solving that,
[57:00.860 --> 57:07.100] and is the issue there like a tax issue in terms of like, so fundamentally, so managed futures is
[57:07.100 --> 57:12.300] also something that I don't want to talk about it too much. Like, I'm going to save off our product
[57:12.300 --> 57:16.700] just to disclaim, we don't even have a product, we don't have a strategy, so don't even call me
[57:16.700 --> 57:21.420] an NFA, I'm not marketing anything right now. I'm just going to talk high level about managed
[57:21.420 --> 57:26.860] futures and tax. So fundamentally, managed futures are usually way overpriced, you've got to buy
[57:26.940 --> 57:32.220] them through an LP, they're usually just too expensive. They're awesome in the sense that if you do
[57:32.220 --> 57:37.740] trendfall of managed futures, they're generally low carry and usually when the world bombs out,
[57:37.740 --> 57:44.700] they provide an assurance like asset. The key is the fees and the taxes. Well, the fees can be solved
[57:44.700 --> 57:51.500] by delivering it cheaper, but the issue is taxes, how do we solve the taxes? Well, the ETF has a
[57:51.580 --> 57:57.980] fundamental problem because if you put managed futures in a 40-act structure, commodity futures
[57:57.980 --> 58:02.860] generally what they call bad income, which means you now have to take something that was 60-40 capital
[58:02.860 --> 58:09.900] gain and create a Cayman entity to block the bad income, and that turns capital gain into income.
[58:09.900 --> 58:15.660] So you actually make the taxes worse by delivering it through a 40-act structure. So that's why we
[58:15.660 --> 58:19.580] have always thought about doing it for the last 10 years, but have never done it,
[58:19.660 --> 58:24.780] because that's not creating affordable alpha. That's creating negative alpha. It's making the
[58:24.780 --> 58:31.660] tax worse, right? So we just can't solve it because usually I use the ETF structure to figure
[58:31.660 --> 58:37.100] out how to avoid taxes. I can't use it, it makes it worse. And so now I'm back to SMA,
[58:38.300 --> 58:44.780] we haven't been able to solve the tax component and fee component you can do by just trying to get
[58:44.780 --> 58:52.220] scale in technology, but you also need to solve tax. And so there's just structural regulatory
[58:52.220 --> 58:58.300] tax reasons why managed futures are fundamentally never going to be solved, in my opinion.
[58:59.420 --> 59:03.740] One of the things that, and I seem to, I don't know how many times you guys have done it,
[59:03.740 --> 59:09.180] but I know just recently you cut fees on some of your ETFs, so you continue to try to drive down
[59:09.180 --> 59:13.980] the cost, make it affordable to investors. And that's ultimately a really good thing.
[59:15.980 --> 59:19.980] Now, I'm not an everyone's email distribution list, I don't know what these other ETF companies
[59:19.980 --> 59:27.020] are doing, but I would say to me, you guys seem to kind of be leading leaders in terms of trying
[59:27.020 --> 59:36.460] to reduce fees when possible. Yeah, it's simple, man. And I'm using this
[59:36.540 --> 59:41.500] vanguardizing as just like a verb, but because everyone understands it, I'm not like trademarking
[59:41.500 --> 59:46.620] everything. I'm just saying we have always been about affordable alpha and vanguardizing weird
[59:46.620 --> 59:54.300] boutiquey stuff that is typically too expensive and crappy taxes, right? And so anytime we have
[59:54.300 --> 01:00:01.020] opportunity where we have excess efficiency, excess profits, excess benefit that the firm,
[01:00:01] if I can reinvest that in lower fees, that's almost a guaranteed way to help people
[01:00:08] have an expectation of better performance, right? And so that's just how we philosophically operate.
[01:00:14] And God bless whoever wants to try to compete against us in concentrated,
[01:00:20] high octane, you know, active strategies at the same prices, because I think it's going to be
[01:00:27] impossible to beat us because we're just on the flywheel right now, right? So good luck.
[01:00:33] I'm happy to see someone try, but we're just bold face in your face, and we will continue to
[01:00:40] extend that we have capacity, scale, and ability. We'll keep lower in the cost as long as we can do
[01:00:46] it profitably and robustly. You know, that's what we do. We're delivering affordable alpha to the
[01:00:51] marketplace or trying to seek to deliver just for compliance purposes. Make sure you get that
[01:00:58] seek to. Yeah, and for context, for the audience, the reason I'm so sensitive to compliance is I
[01:01:03] am now the new CCO of an alpha architect entity. So, you know, we're always disclaiming or using
[01:01:11] email him if you have any compliance issues. Yeah, I'm sorry, language in here. And I don't want to
[01:01:17] talk about any service or strategy I offer. I want to talk about any ETFs. We're just talking
[01:01:22] academic materials as best as possible. Right. Wes, this has been great. Thank you very much. We
[01:01:29] really always appreciate you coming on and sharing your knowledge, being totally honest and fourth
[01:01:33] rate, common sense, first principles, all that stuff that you're known for. So, if people want to
[01:01:39] find out more about launching an ETF or the strategies you run, where can they go to learn more?
[01:01:46] Yeah, just hit up ETFarchitect.com and then obviously anything on our investment side,
[01:01:51] just hit up Balfarctech.com and, you know, click the buttons to contact us and eventually you'll find
[01:01:58] us. And you'll see that they are in the holiday spirit over at the firm over there because they
[01:02:03] have the Christmas hat, Santa's hat on the logo. So, Mary Christmas. That's right. Yeah,
[01:02:09] architect. We can still do goofy, boutiquey things. You know, ETFarchitects too professional,
[01:02:15] institutional. So, we don't, we got to keep it high, bro. But yeah, now for our tech side,
[01:02:20] we always have a little fun. So, we continue to do so.
[01:02:24] Thanks. Love it. Thanks, Wes. Appreciate it. Yeah, you got it, guys. Appreciate your time and
[01:02:28] look for our next conversation. This is Justin again. Thanks so much for tuning into this
[01:02:34] episode of excess returns. You can follow Jack on Twitter at at practical quant and follow me on
[01:02:39] Twitter at JJ Carboneau. If you found this discussion interesting and valuable, please subscribe in
[01:02:45] either iTunes or on YouTube or leave a review or a comment. We appreciate it. Justin Carboneau
[01:02:50] and Jack Forehand are principles, ability of capital management. The opinions expressed in this
[01:02:54] podcast do not necessarily reflect the opinions of the ability of capital. No information on this
[01:02:58] podcast should be construed as investment advice. Securities discussed in the podcast may be
[01:03:02] holdings of clients ability.
Transcription results written to '/home/forge/transcribe.sonicengage.com/releases/20240210005325' directory
