Episode #411: Kai Wu, Sparkline Capital – Investing in Innovation, Intangible Value, & Web3
Guest: Kai Wu is the founder and Chief Investment Officer of Sparkline Capital, an investment management firm applying state-of-the-art machine learning and computing to uncover alpha in large, unstructured data sets. Previously, Kai worked at GMO, where he was a member of Jeremy Grantham’s $40 billion asset allocation team.
Date Recorded: 4/20/2022 | Run-Time: 1:33:10
Summary: In today’s episode, we’re talking about two topics that are important for investors to understand in 2022 – intangibles and innovation. Kai shares how he uses machine learning to track things like brand equity, human capital, network effects, and IP to measure the intangible value of each firm (and how he implements this through his ETF, ITAN). Then he shares why his research leads him to believe value is not dead.
Finally, we talk about his most recent paper about investing in innovation, a popular investment theme that’s under scrutiny as of late. Kai shares why he believes the current drawdown is not driven by pure innovation but by a selloff in expensive unprofitable stocks.
Comments or suggestions? Interested in sponsoring an episode? Email us Feedback@TheMebFaberShow.com
Links from the Episode:
- 1:12 – Intro
- 2:14 – Welcome to our guest, Kai Wu
- 4:45 – Starting his career at GMO before launching a crypto fund in 2014
- 11:40 – The origin story of Sparkline Capital and why focus on Intangible Value
- 17:22 – Kai’s intangible value framework
- 20:47 – Scraping social media to create personality profiles for companies; Episode #410: Chris Bloomstran, Semper Augustus
- 25:44 – Which companies have positive correlation between their brand and stock performance
- 29:00 – How Kai applies a value lens to his intangible framework
- 38:11 – Launching the Sparkline Intangible Value ETF
- 40:18 – Value Investor’s Guide to Web3 and opportunities a value investor can find in crypto
- 46:19 – Web3 is both a real industry and a bubble
- 50:19 – The strategy behind his newly launched private Web3 fund
- 56:51 – Overview of his newest paper, Investing in Innovation
- 1:13:01 – Why the drawdown in ARKK and other similar funds is not a result of a bear market in innovation
- 1:17:34 – Disruption at a Reasonable Price (DARP) and how that strategy manifests itself
- 1:23:47 – How challenging are these strategies are to replicate abroad?
- 1:26:52 – His most memorable investment
- 1:29:13 – Learn more about Kai; sparklinecapital.com
Transcript of Episode 411:
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: What’s up, my friends? We have a really fun show for you today. Our guest is Kai Wu, founder and CIO of Sparkline Capital. In today’s episode, we’re talking about two topics that are important for investors to understand in 2022: intangibles and innovation. Kai shares how he uses machine learning to track things like brand equity, human capital, network effects, and IP to measure the intangible value of each firm. Then he shares why his research leads him to believe that value is not dead. Finally, we talk about his most recent paper about investing in innovation, I love this paper, A popular investment theme that’s under scrutiny as of late.
Kai shares why he believes the current drawdown is not driven by pure innovation, but by a sell-off and expensive unprofitable stocks. Now, before we get to the episode, a favorite ask, I know there’s one person you know that would enjoy an episode like this and it’s a great one. So, take a second, share this podcast with someone you know. And based on the episodes we have coming up, they want to be sure to subscribe as well. Thanks for spreading the word. Now, please enjoy this episode with Sparkline Capital’s Kai Wu.
Meb: Kai, welcome to the show.
Kai: Hey, man. Thanks for having me on.
Meb: It’s good to see you again, buddy. You know I enjoy getting to meet you over the last few months, eating sushi, having a few beers. Where do we find you today?
Kai: I’m in the great city of Brooklyn, New York
Meb: What’s the vibe in Brooklyn like right now?
Kai: The vibe is good. We got good vibes down here. Everyone’s moving down in Brooklyn, it’s the place to be.
Meb: I used to make it in New York about once a quarter and pandemic hit, yadda-yadda, I’m due. This is like the best time of year, spring in New York is, like, my favorite. I’m currently displaced out of my house, I’m homeless. We’re renovating, which seems to be not resolving anytime soon. So, maybe I’ll wrangle the crew and get to New York in the next month or two, I’d love to. You’re going to have to host me.
Kai: Yeah, and I think I owe you some sushi.
Meb: Good, deal. I’ll take you up on that. New York and LA are two of the best sushi cities on the planet.
Kai: Yeah, we’re very spoiled.
Meb: Yeah. Cool, man. Well, I just saw you in Miami, holding court at the recent ETF conference. Was that a good trip for you? How’d you find Miami? Was it just full of VCs and crypto meetings for you or do you go to an ETF conference? You’re ETF manager now.
Kai: Yeah, well, I may have been one of like three people who did the doubleheader, did the Bitcoin conference and ETF conference.
Meb: Pull the thread for me between the two. There are 30,000 people at the first one and maybe like 3,000 at the latter one. What were the similarities, if any?
Kai: I was actually shocked by how much interest there is in crypto in the ETF community. If you tell me, “Hey, there’s an ETF conference,” I don’t think crypto. But you saw that at the allocator hall, there’s like half the booths for like Grayscale and other kind of disruptive blockchain style offerings.
Meb: Yeah, you know, I think the challenge with the crypto community on the ETF side is you’ve had this sort of waiting on the good dough for I don’t know how many years it goes back. I have an old tweet where people were getting super excited about crypto ETF spot coming to market and I was like, “Look, if anyone who wants to make a bet that this makes it out by the end of the year,” I was like, “Let’s do a dinner bet,” and no one would take me up on it. And I said, “I prefer sushi,” so I’ve been consistent over the past decade. I think that was in 2013. So, we’re going on like Year 8 of no spot ETF in the U.S., so I think the ETF crowd is sort of frustrated and helpless but hopeful at the same time. A lot of people doing some good work there.
Kai: Yeah, yeah. And each year that passes, we get one year closer to whenever it ultimately happens.
Meb: So, listeners, we’re going to go deep on a couple of topics, make sure you stick around for this entire episode because we’re going to be touching on a few of Kai’s papers. If Kai is new to you, you got to check out his company’s website, Sparkline Capital. But one of my favorite new…new to me, I should say, thinkers over the past few years, he’s got some really fun papers on an intangible value, which we’re going to talk about, investing in innovation, which by the time this podcast hits, should have dropped.
I got a sneak peek, it’s awesome. We’ll talk about it. But we got to start at the beginning with you because I think, and I’m not certain, my producer will have to check this, we’ve probably had more alumni on this show from GMO than probably anywhere else. Research Affiliates is up there, AQR may be up there. I’m trying to think. But you’re an alumni, right? You originally started out at GMO? We’re not talking Monsanto here.
Kai: No, no. My first job out of college was working for Jeremy Grantham.
Meb: And what was that, fetching coffee? How does somebody get a job right out of college? I mean, I guess you were local, but how does someone cozy on up to GMO?
Kai: I mean, it was more than just coincidence. So, I actually wrote my…so I studied economics at Harvard and wrote my senior thesis on financial crises and bubbles with Professor Kenneth Rogoff. It was kind of a natural segue to just hop across the river and start working with Jeremy and his team on that very topic, bubbles.
Meb: You start talking about bubbles, that’s catnip for Jeremy. So, you know, it’s like you got a warm intro. Give us a rewind, what were some of the ideas and conclusions that you guys were looking at back then?
Kai: On my thesis, the idea was…it was very, like, kind of Austrian, so it wasn’t a super popular favorite amongst the Harvard economic department at the time. But the idea is that imbalances build up in the economy, whether external credit, asset price, what-have-you, and these things eventually need to unwind.
And it’s impossible to predict what will be the straw that breaks the camel’s back, but you just know…and it turns out to be the case that we went through decades of data, basically pulling in every single crisis w could come up with, so Asia, Sweden, U.S., savings and loans. And what we found was that you could actually predict a higher probability of crisis when the imbalances had been built up prior to that. Very difficult to predict exactly how it all unfolds but there is that predictive power.
Meb: What was your role? So, you’re probably a 20-something, young 20-something, were you able to contribute at all at that point? Or was it, you know, “I’m here to learn and be a part of this team?” What were you actually doing? And what year would this have been? Give us the timeline.
Kai: Yeah, I was super lucky. I joined in 2009, so this was after GMO had its best year ever and made its clients so much money betting against GFC. And it was one of those situations where it was a small team. I mean, there were only a handful of investment professionals. Yet, AUM was just kind of growing every month, billion dollars and billion dollars and billion dollars. And I was lucky because I was the most useless person in the group, right? I was just, like, showed up out of college and I was like, “Hey, guys,” and everyone else, you know, had real responsibilities and I didn’t.
So, what they did was they sent me on this like world tour and I spent like three months in Australia, three months in London, three months in San Francisco, and then back to Boston, working with all the teams and our specialists and macro and EM and, you know, various quant stuff, coming back kind of bringing that knowledge to the mothership and serving as a liaison to those teams as we went on working on this project where we expanded our forecasts like from equities to other asset classes. So, you know, I kind of lucked into being somewhat useful merely because of my ignorance.
Meb: How fun, man, what a dream initial setup. I mean, granted, you top tick, the perfect time to join, but also at the same time, what a fun…it’s just like an absolute kid in a candy store opportunity. Okay, so at some point, you decided to say, “I’ve learned everything possible from GMO, it’s time to start my own adventures.” And then was this the entrepreneurship move out at this point? What was next?
Kai: Yeah. So, end of 2013 is when I left GMO, and I love the firm, I love my coworkers, and I had a great experience there. You know, working with Jeremy obviously set me on the path as a value guy and I’ve gone down the rest of my life, but, you know, I’m wanting something more entrepreneurial. So, what actually ended up happening, and this is kind of interesting because I’ll come full circle at the end of this conversation, is I actually got into crypto.
So, I started doing just like crypto trading strategy. This was like early 2014 when there were maybe seven things you could trade. And it was me and a few other people kind of sitting around, like, trying to come up with ways to make money. And the crazy thing was that, you know, you read those market research books and you have all these classical arbitrage that were just like lying around in like 1980. Well, it turns out you could do those in 2014 crypto, like, we’re doing these FX triangle trades where it was like, you know, Bitcoin and Litecoin like DoS or something, and you would look at the different payers and figure out if there was a mispricing.
Or even today, you know, liquidity is ultra-fragmented still across various exchanges. Well, back then, it was the case as well but also, there was no one really sitting there trying to arbitrage. So, it’s like, you know, gold in London versus gold in New York, that trade. So, you could do all this but what ended up happening was I kind of decided at some point that, like, we were still too early, that, look, the market cap of all crypto at the time was $8 billion and we were years away from like the serious institutions coming into the space.
And then I had the opportunity to join up with a guy who was leaving from my former team and we co-founded a firm in Boston called Kaleidoscope Capital. I was there for four years, setting up the firm, you know, from scratch to $350 million in AUM and good experience. And then in 2018, left to kind of start my solo project, which is Sparkline.
Meb: Awesome. That’s a fun journey. I remember back to those days of crypto, I mean, we had a crypto payment option. I was down in Mexico City at a conference and was chatting with some people that were early in crypto, and I said, “You know what? I’m going to start to dabble in this.” And we put a payment option on our idea farm research service, and no one used it. And my mistaken approach to it at that point was that my audience is the use case or at least the early adopter use case.
I was like, “You know, my audience is pretty tech-forward, I’m surprised they’re not using this,” and then we eventually took it down. I’m kind of sad because, for a long time, I was like, “I wonder if people actually paid in Bitcoin,” and I just never paid attention to my wallet and I was like really hopeful there was going to be like $2 million in there. To my knowledge, there isn’t. But humorously enough, that same Mexico City trip at the same hotel, I rode up in the elevator with Dwight Howard and James Harden, they were some exhibition game. It was the strangest trip of my life, but awesome as well. So, Sparkline, was the origin story for Sparkline similar at the time as it is today, or was it a crypto focus? What was kind of the inspiration on going solo?
Kai: The big thesis that we have in Sparkline…so crypto is something we kind of got into later and, you know, kind of came full circle. But really, the big thesis at Sparkline is the idea that the economy is becoming more and more intangible, right? Human capital, brands, intellectual property, and network effects, that these are the assets that now power the economy, yet most people don’t correctly value them and undervalued them even then because they’re just so hard to measure.
And so, that’s really been our focus. And what we realized is that accounting data is kind of insufficient in being able to measure these standings. And what you really need to do is to go beyond into artificial intelligence, unstructured data, social media, patterns, Glassdoor, LinkedIn, these sorts of things. And in order to do that, you really need to invest heavily in technology since it’s not so simple as taking a bunch of like 10-Ks in techs and, like, throwing them in linear regression.
We really need to build actual cutting-edge infrastructure. So, after I left Kaleidoscope, the first year and a half, maybe two years was 100% heads down, “Let’s build out the research platform that will allow us to do whatever we want.” Now, it’s crypto as well but, you know, it’s equities initially and still was equities. And the idea was, “Look, I’m so lucky, because I now, you know, a mid-tier, I have no employees, no clients, no portfolio, all I can do is I can sit here and, like, focus on building out the next generation research platform.” So, I did that and that’s kind of where we are now.
Meb: All right. Well, good, we’re going to spend a lot of time there. But was there an inspiration that really pushed you toward this concept of intangibles? Were you reading a paper? Was there just a day where you’re just sipping coffee at your local hipster Brooklyn coffee shop? I guess today that would be, what, like a matcha latte? Or was it something that just was a slow build up over time? What was the foundation where you built this kind of concept from? And then we can lead that into your paper on intangible value.
Kai: Yeah. So, the big problem that I tried to address is this question of the so-called death of value, which I’m sure you’re aware of. It’s this notion that it’s no secret that value strategies, and in particular, quantitative value strategies, have massively underperformed in the past 10 years, and you have people saying, “Oh, value is dead.” Now, look, I’ve been a value guy my entire career and there’s not a ton about this issue, and the conclusion I came to was that value is not dead.
And the idea of buying low and selling high compared to some measure of intrinsic value, that’s, like, by definition, true. The problem is that the metrics we use as to what is value, that needs to be adapted. So, you think back to like when Graham and Dodd wrote “Security Analysis” in 1930, the economy was totally different back then. It was industrial, the biggest companies were railroads and cement, and more importantly, value back then was tangible.
As a value investor, all you have to do…again, easier said than done, but all you had to do was find companies that were trading below net liquidation value and buy them, it’s pretty straightforward. Obviously, acquiring data was a lot more challenging back then, but at least conceptually not that hard. Now, fast forward 100 years to today and we live in the information age, the biggest companies are Google and Amazon and Apple. These companies don’t require net tangible assets to produce earnings, they rely on human capital, on IP, on branch and network effects.
And despite the rise of the intangible economy, right? Intangible capital is now over half of the capital stock of the S&P 500 and this trend is only going to keep increasing over time. Despite all this, accounting has basically done nothing. I already mentioned this, but, you know, accounting does not capitalize R&D or evidence. So, normally, when you create a factory, if you invest in buildings in tangible and physical assets, it goes on your balance sheet. Well, if you invest in R&D, it doesn’t go on your balance sheet, it comes out of your earnings, right? That seems inconsistent.
Or thinking about human capital, the only human capital disclosure in the accounting 10-Ks is headcount. We live in a world where CEOs are saying, “People are our greatest asset,” yet they can’t bother to disclose anything more than the number of people on their balance sheets. So, that just seems completely ludicrous. We kind of realized at some point that accounting data was a non-starter and we had to transcend that, which is why we went down this path of linking the kind of machine learning/AI skillset that we developed as clients back into the fundamental idea of trying to fix “value investing,” and trying to bring intangible assets into this definition of value. So, that really is the genesis of this concept.
Meb: You have a great quote on your website from Uncle Warren Buffett where it says, “The four largest companies today by market value do not need any net tangible assets, they’re not like AT&T, GM, or Exxon, requiring lots of capital to produce earnings. We’ve become an asset-light economy.” As you look back, is there a particular time or regime when this transition…I mean, obviously, it occurred over a period of years, but is there a certain kink in the data or time where you think it becomes more meaningful to incorporate this?
Kai: I think it really started accelerating in the ’80s and mid-90s. But what happened was there was this weird dislocation with a tech bubble where things just got so overdone and it had to come back and get released, that it’s kind of hard to look at the data and really draw conclusions because that’s such an outlier period. So then, if you put that aside, it’s really 2005 and on where once you start incorporating these metrics, you get a much better result than if you were to stay with the kind of tried-and-true book value, price-to-book value metrics and, you know, realize that it kind of had stopped working.
Meb: Yeah. So, let’s say one buys this argument conceptually, let’s walk through kind of the paper, how to think about it. What’s the framework for…as you say, “Challenge accepted,” in this paper, what’s the framework for how to incorporate and think about this?
Kai: Yep. So, we started with the first principle and that is, if you sat here with me and say, “Hey, Kai, let’s brainstorm, what are the intangible assets that might matter today?” And then tried to and say, “Can we collapse that list into the smallest number or the fewest pillars possible so that we can span the entire universe?” You come up with a list similar to what I did, which is there are four pillars of intangible value, the IP, brands, human capital, and network effects.
And that’s just first principles, I guess, without looking at any data, doing any data mining, etc. So now we say, “Okay, now that we know what matters today or we think we know what matters today, can we actually measure that?” Can we actually say, “All right, well, I want to figure out how strong human capital of Goldman Sachs is, how can I do that?” So, you know, each is so heterogeneous, and I’ve done a dozen papers now, each one is a deep dive on a specific thing.
Well, I think the best way is just through an example. So, yes, human capital, one of the papers I wrote focuses on the use of LinkedIn. And LinkedIn is super interesting because it gives you a record of everyone’s kind of employment today but also all the way back from time, so a time series, which allows you to track the flows of talent from company to company. And we use this to answer two questions. So, the first question we answer is, “How are companies able to attract and retain top talent?”
The idea being that if I am able to poach the top engineers from Apple and bring them into my company and then keep them happy so they don’t turn over, that should be a good thing. And how do you measure that? So, what we do is we actually go into LinkedIn and form these graphs where you can see, you know, where human capital is flowing, say, from Apple and Microsoft to Facebook, to Sparkline. And what we do is we use PageRank, which is the Google algorithm that’s used to do search.
So, PageRank is this idea and this is Larry Page’s original invention that is behind Google search, that the strength of a website is a function of its backlinks. And to the extent that it’s getting a backlink from like Wall Street Journal, that’s better than getting one from some random blog. We can do the exact same thing here where getting, say, a computer vision PhD who went through Google and before that was at Carnegie Mellon, that’s like really good, but getting somebody from like maybe your local IT support helpdesk is like less valuable.
So, that’s the first thing we do is look to figure out which firms are able to attract and retain top talent. And the second thing is trying to dispel this idea of like an AI feeder. So, this is the idea that every CEO…it’s becoming very, like, trendy for CEOs to get up there and say, “Hey, we’re doing a digital transformation, we’re investing in like a blockchain, you know, cloud computing, and Internet of Things,” so kind of throwing out all these buzzwords.
Like, if everyone does it, then we don’t actually know which firms are like BS-ing versus, like, truly are investing. So, the idea is can we figure out when CEOs are putting their money where their mouth is. And so, we’ll get into the patent data, like, later on, it’s obviously one way, the other way is to look at human capital. Because if you truly care about AI, you’re going to be hiring people with TensorFlow and PyTorch on their resumes.
If you truly care about blockchain, you’ll be investing in folks who like Solidity. So, these are all skills that are, you know, mapping to the space and help give a sense for which firms are truly, you know, willing to pay a premium to get, say, a top MIT-trained NLP engineer.
Meb: I like the phrase you used on your website of quantifying dark matter because, to me, a lot of this is things that people…like you say, they may talk about and discuss but they’re not necessarily actually quantifying. In some cases, it may not be…it may be misleading at best. So, do you want to talk a little bit about how you think about putting this together? And so, you start scraping some of these sources of information that most don’t, you have these four pillars, and you can expand on any of these pillars that you think are important to get deeper on as well.
Kai: So, let’s do one more example, just because why not?
Meb: Yeah. We can do all four. I love listening to this.
Kai: We can do all of them. I don’t know how much time we have here, do you want to do a three-hour podcast?
Meb: Well, I got plenty of time. We just recorded one with Chris Bloomstran and that was two hours-plus, but that just means you’re having a good chat. So, we’ll see if you can hit the two-hour mark. So, let’s go, tell me some more.
Kai: All right, so here’s another one is brand. Now, there is this…the way we frame this is if you think back to like “Mad Men,” right? We just sit around and watch our TVs and be told by the guys in the suits on ads what to buy. But today, the way that brand perception is being shaped is on social media. So, what we’re really doing now is listening to a third-degree connection who happens to be an influencer in matcha lattes, talking about why he or she is drinking, and then we go buy it.
So, as a investment manager, if you want to quantify the brand of a company, you need to go into the room where it happens, which in this case, is social media. So, what we do is we scrape Twitter and we say, “All right, let’s get like all the tweets, millions of tweets associated with all the brands that we have in reverse.” And then what we say is, “Can we create brand profiles, brand personality profiles for each brand?” So, for example, we say Christian Mingle is sincere, WD 40 is rugged, Tesla is exciting, right?
There are these five dimensions, kind of a Myers Briggs style, like, radar chart. And the insight is that it’s not so much what you’re known for, it’s more like being known for something that matters, right? Think about the apparel space, you know, Lulu, Nike, these brands all have, you know, really passionate followers because they are known for a specific thing. And the key is, you know, being known for something that puts you in kind of a brand map unique from your competitors.
So, we do all this work, kind of figure out the personalities, figure out how they look next to their competitors, and we find that the stocks with the strongest brands do tend to outperform. So, this is like a factor we can add to our models. So, that’s an example on the brand category. You know, you mentioned the notion of these things that everyone says are important but no one’s measured. The best example there is culture. Everyone quotes this notion that culture eats strategy for breakfast.
Everyone says, “Hey, look, if you have a good firm culture, that’s all that matters.” But the question is, like, are these CEOs just saying that or they just kind of ex-post trading a narrative for why they are successful when really they are, you know, insider trading or something? So, you know, we wanted to be the first starting to actually show a connection between firm culture and future stock prices. And the way we did that was we went into Glassdoor, right?
Because again, the same problem with AI is the CEOs love to talk about how amazing their culture is. No CEO is going to say, “We have a bad culture,” but that has no correlation with the actual experience of the rank and file, the day-to-day employees. So, we go to glassdoor.com. This, for those who don’t know, is like a forum where it’s like a Yelp, kind of, where people can leave reviews on their employers or former employers, and we scrape all the reviews.
And what we do is we first find…so we have a few findings. The first finding is that the star rating, so people can give a rating between one and five, is not actually that important, it doesn’t really predict that much. The second finding is that the text of the reviews is better and you can do sentiment models where you train supervised learning models to say, “Hey, is this text positive versus negative?”
That is somewhat productive, but what was the most productive was actually creating…it was similar to what we did with brand personalities, creating these profiles for firms on a handful of different dimensions. So, what we did is we said, “What are the dimensions? What are the values that companies care about?” And you go on all the websites and, like, tally up how many times they mentioned integrity, etc.
You find that, like, the first thing is integrity, then you have innovation, teamwork, all the way down. And so, we say, “Can we take each review and figure out on what dimensions these companies are strong?” And so, you might find that certain companies are really known for innovation, others are really known for being very team-oriented, etc. And what we find is that the firms that are strong on these main values actually do have it, that culture actually does matter in this case.
Meb: So, in your paper, which is a little outdated…not outdated, but it came out a year ago, you give some notable intangible companies. Are there any that you think stand out as being positive in this category? And/or if you want to say that there’s any that are particularly negative?
Kai: Yeah, so the point of this was to try to contextualize the model. Our goal here is to not create like a black box, like a quant black box, but actually what we’re trying to do is create a transparency and, you know, have intuition map to what we own. So, it’s like, “Hey, the idea was Nvidia, what is the reason why are they on this list? Why is Nvidia strong?” Well, it’s because their IP, their IP is what gives them a moat against, you know, other competitors in chip space.
Moderna obviously has the mRNA patent, things like that. On the brand side, we find, you know, Nike and Harley, human capital, Goldman, Regeneron, network effects, Amazon, Twitter. And, look, all we’re saying here is not so much, “Are these guys attractive and absolute?” But more what is the primary driver of each of these businesses without making any judgment as to whether or not that’s a long or short in general.
Meb: It’s interesting too because certainly, brand, as Buffett describes in some of these things that are a little squishier but that you’re trying to quantify, can be pretty time-dependent too. I would argue that this sentiment may be perhaps surrounding Robinhood is different today than it would have been, I guess, prior to them being public.
But a while back…and Facebook has certainly seen its ups and downs and thinking about scandals and one-off events, and also employee retention and getting fired. So, Robinhood was laying off a bunch of employees today. How often are these models updating? Is this something you’re looking at once a year or once a quarter? Is it daily? Like, what’s the frequency with sort of information is being ingested and spit out?
Kai: Yeah, we update these models every day. So, as new tweets come in, as new reviews are posted, as people change jobs, that information is in real-time and being fed into the models.
Meb: Yeah. As Elon is buying Twitter, by the time this is published, who knows what is even going to be happening with that, by the way, it seems to change on the dial. Okay, so as we’re cobbling together this concept that you’re talking about, about intangible value and some of these kind of four horsemen, intellectual property, brand equity, human capital, network effects, are there any more areas of this we haven’t covered before we talked about how the sausage gets made on incorporating these into sort of a composite concept?
Kai: Yeah, so on that front, look, the key here is that this is a value strategy. We’re not just going to go out and buy the companies with the most patents or the most PhDs. We’re going to look for companies that have a high number of PhDs or patents relative to their market cap. By normalizing everything by price, that’s what makes this a value strategy.
So, we’re not going to go out and buy Tesla. Why not? Well, Tesla is a very innovative company but the problem is it’s just so expensive that it’s very difficult for these models to get comfortable with that valuation. And so, by doing…it’s similar to how classical value investors look at, you know, different yields like dividend yield or price-to-book, we just replaced those fundamentals with intangible value metrics and then you end up with our portfolio here.
Meb: I was laughing because I was just reading your paper and you’re talking about sausage getting made and hopefully, Impossible Foods, they get a new sausage out. I haven’t tried it but I’m a shareholder, so I love pushing their…probably terrible for you, but not…
Kai: Let me know how that goes.
Meb: Yeah, their nuggets are amazing. Anyway, I posted on Twitter right before we started this, a funny picture where I was at Legoland, and so I probably have all five COVID variants now. But I was laughing because there was a buffet that had this big sign. Meanwhile, all the other buffet stations were…if you could come up with the least healthy possible things on the planet in a kitchen, there was Fruity Pebble pancakes, which I saw which was incredible.
But anyway, there was one station that was labeled “Healthy choices” or something, I was laughing because it was full of bagels and bread. So, this harkens back to the food pyramid of my childhood, where it’s, “Eat a bunch of pasta, cereal, and bagels and you’ll be healthy, just whatever you do, avoid fats.” It’s just funny how the views have changed over the years. And it seems like what you’re talking about is relevant.
The views of how to think about value, how to think about innovation and some of these concepts that others kind of subjectively talk about, you’re starting to quantify. So, okay, do you approach this where you look at each silo independently, and then you’re kind of looking at the unit as a whole? Are you coming out with composite metrics for each company? How do you start to rank order the entire universe of what it means to be intangible sort of value strategy versus the opposite?
Kai: Yep, we obviously have dozens of different metrics. I think we talked about a few here. But we only have three hours for this podcast, so I won’t go through all of them. But the idea is that we take all the metrics and for each pillar, kind of smush them all together to an average. And the reason you do that, by the way, is to deal with correlations. If I have a metric like the number of PhDs and another metric like the number of patents, those things might be correlated, and so you want to kind of deal with those sorts of issues.
So, you create these four pillars, so now you have four numbers, and then you do just kind of sum them up simply. The idea being that we’re not trying to make any judgments as to is IP more or less important than brand in the modern-day, but kind of like bottoms up, like, follow the fundamental values through the economy as, like, the world potentially changes or doesn’t, right? Like network effects are an interesting example because they become more and more powerful with the rise of the internet.
You know, shopping malls, I would say, are a form of network effects for a platform company. But, you know, you’re having a few thousand people go through a mall, whereas, like, now on Facebook, you have 6 billion people go on. So, that drastically accelerates the real network effects and you can see it in the data emerging increase over time, and you want to allow that to happen organically and not the constraint of being like, “Oh, well, I always like equal-weighted,” or, “I’m always going to put 10% into network effects,” because that would, you know, forego that opportunity.
Meb: So, the cool part about the paper is you start to show some strategies where you take this intangible value concept and you can take it back decades. So, my first question on this is have you reached out to Fama and French and see what they think about this? You pass this paper along? Are they open to this concept and evolution of some of their ideas? Or what’s the reception been?
Kai: I don’t actually know Fama and French. I know they had a three-factor model. Now, they have a five-factor model, so…
Meb: Soon to be six, the Kai Wu factor. Well, you got to muscle our buddy, Wes, and get an intro because he’s friendly with those peeps. Anyway, I was just saying, because, you know, so much work has been built upon some of their ideas. I mean, DFA built, whatever, a $400 billion business based on this concept of price-to-book alone. I feel like…correct me if I’m wrong, you probably know more about this than I do, I feel like they recently were writing about how they’re not just using price-to-book anymore as a value metric but considering other ideas. Does that sound familiar or did I just make that up?
Kai: I would love to hear that. I don’t know.
Meb: Well, I’ll Google around as you’re talking on this one. But anyway, so walk us through some of the conclusions on how this portfolio strategy has behaved. We got the conceptual idea but what’s the actual outcome for this type of strategy over the decades?
Kai: The point of this is to get outside of like the style box framework, and to get away from this idea that, “Oh, there’s like value stocks and then there’s growth stocks.” You know, this is what Warren Buffett said that values and growth are joined at the hip, we want to have a model that, again, follows the value through the economy and maybe value one day and maybe growth another day, just depending on where the fundamentals go.
So, right now, the portfolio is…it is heavily invested in what you might call like new economy sectors, so semiconductors, media, software. And that makes sense because that’s where economic activity, especially in the U.S. large-cap and mid-cap space, is concentrated. If you were able to backtest this 100 years, which I can, you would almost certainly see the sectoral composition and change. You know, remember, railroads were the technological marvel of the 1800s, right? They were kind of growth stocks.
And so, you would kind of see it move. But the really powerful thing here is that, yes, this portfolio does tend to own new economy sectors, but it does so in a way that’s still very cognizant of price. So, if you look at, say, valuation ratios, you find that the price-earnings and price-to-book ratios are similar to the markets. And more importantly, if you look at things like R&D divided by price or PhDs divided by price, these intangible value ratios, the portfolio is like 2X as attractive as either the S&P, the Russell 1000 value, or the Russell 1000 growth index.
And the key here is like, “Why is that? That seems weird.” Well, look, the value index doesn’t hold anything stocks, it doesn’t have any technology stocks. If you don’t give Google credit for its IP, you’re never going to hold Google, it’s always going to seem expensive. So, of course, there’s no intangible value. If you go to the growth section, then you’re like Tesla and you’re like Moderna.
And these companies, yes, they have a ton of IP and human capital, etc., but because the valuation is so high, the amount you get out for dollar put in is actually not that special, which is why we have it at 2x compared to all three of these metrics. And I think that’s really the crux of this portfolio is you’re getting exposure to new age, kind of disruptive, forward-looking companies, but doing so at a reasonable, if not attractive price.
Meb: So, for those listening, as they tend to think about the strategies historically beaten both, but as you think about the traditional framework of a value investor or growth investor, this is interesting because it has elements of each. It has companies that may be bucketed more traditionally as growth, these sort of sectors and industries, but also value characteristics. When you talk about the narrative of the strategy, do you tend to put it in a comparison of one or the other or it’s its own animal?
Kai: Yeah, I think it’s in the eye of the beholder. For those of us who are used to thinking about what value strategy is, this is a way of applying value into non-traditional areas, right? High intangible companies. For those of us who are used to investing in growth-like companies, this is a way of maintaining that exposure to the future while being a bit more cognizant of price. And for those of us who are DFA, kind of factor investors, this is potentially a sixth factor or maybe a ninth factor, let’s say, that folks can potentially tilt towards if they do buy the argument that these intangibles tend to be undervalued because they’re so hard to measure.
Meb: Well, it’s fun because you have in your paper, which we’ll link to on the show notes, on intangible value, you have fun charts of factor exposure breakdown, so S&P, Russell value growth, and then intangible value. But it’s fun because you see it in the price patterns, price PhDs, there’s some fun factors that most may not have seen before. Does the strategy constrain in any way as to industries or sectors or theoretically, it could be all in on railroads like you mentioned?
Kai: Well, I used to be a hedge fund guy so I’m very familiar with factor neutralization, long/short investing. We were very deliberate, though, when we build a strategy to not do any of that and, like, let the exposures grow bottoms up. And that’s for two reasons. So, first is like I don’t really buy the whole GIC classification, I think it’s obsolete. Of the five FAANG stocks, only one of them is in IT stock, and you have plenty of disruptive companies that are not classified as IT and then plenty of legacy tech companies.
So, I just don’t think it captures the factors we’re going after in the modern-day and I would much prefer if we were to classify companies into four buckets along with the four intangibles. So, that’s the first reason. And the second thing is that even let’s imagine we do buy the notion that these GIC classifications are the be-all and end-all of what is a sector, take the example of like green energy. So, like, imagine what is the energy exposure in this, it would be like 5% or something?
So, imagine we were to say there’s 5% exposure just to energy today, let’s just like fix that forever. And then let’s say a year from now or 10 years from now, green tech became, like, the biggest industry in the U.S. and it’s 40% of the market cap. Well, it’s too bad, you’re only going to own 5%. So, it just doesn’t make sense, I don’t think, to overly constrain this portfolio and instead let the notion of intangible value drive the allocations.
Meb: Yeah. You went the extra step, man. You eventually launched an ETF, the Sparkline intangible value ETF, listeners, ticker ITAN, great ticker. And for people who are interested in this sort of strategy and what it’s up to, what kind of positioning…I mean, here we are in 2022, things are getting weird just like they were in 2020 and 2021. I’m waiting for a break, I’m ready for a quarter just for it to be kind of mellow. But that’s the market, it’s always exciting. You decided to launch an actively managed fund with some friends of ours, do you want to tell us a little bit about the process? Was this as exciting and mind-numbing as you thought it might have been? More fun? Less fun? More headache-inducing? How’s the experience of launching a public fund been?
Kai: Definitely different. I never saw myself as an ETF manager. And this is how the story goes, actually. So, I posted a blog on our friend Dan Gardosh’s blog and this was like beginning of COVID I think. So, the world was about to go into a lockdown. And West reached out and he goes, “Hey, dude, do you want to start an ETF? Have you ever thought about doing an ETF?” And I was like, “Why would I do that? I’m a hedge fund guy.” You know, like, “What?”
But, you know, we started talking and, like, we did a bunch of calls and, you know, he was really persuasive when it comes down to like, “Hey, if you’re going to put money into a strategy yourself, don’t you want to wrap it in like a tax wrapper?” Right? This idea that you can kind of wash out capital gains and defer tax-free on those gains until you sell is just such a tremendous advantage over time.
I mean, it doesn’t make a difference if you’re going to hold for like six months as most people in ETFs…many people in ETFs do. But if you’re truly oriented of like a buy-and-hold investor in an active strategy you believe in and you plan on holding it for a long period of time, why wouldn’t you want to do ETF? If it’s almost a no brainer. And so, that’s really what persuaded us to go down this path
Meb: It’s exactly how Wes would have phrased it is, “Hey, dude.” We’re referencing Wes at Alpha Architect, former podcast alum as well as Dan at Verdot, another podcast alum. What’s funny, you know, I mean, look, man, the power of writing and research has been a great example. I can speak to it personally as can you, putting out awesome research like you guys have has led to a fund and hopefully many more.
So, as a quick segue, we got to talk a little bit about Web3, your paper there. Give us some the ideas and thoughts on that one and what’s going on in that world? What did you learn in Miami? And by the way, I think your choice of titling this paper of “Value Investors Guide to Web3,” that’s some good SEO because no one’s combining those two phrases, value investing and…
Kai: Because nobody’s doing it, it’s antithetical, right?
Meb: Thinking about that, there’s just going to be like the one result and then blank afterwards. All right, what’s the value guy doing in crypto land? Let’s talk about it.
Kai: Look, I came out with all this research on intangible value, mainly focused on companies. And the feedback I got from many, but not many people was, “Hey, that’s really cool, you know, this seems really cool and they work in equities, but what I’m really trying to figure out right now is like, what’s going on in crypto?” And the really powerful thing about the intangible value framework is that it can be used in non-traditional areas. I always call it value investing in weird places. So, the idea is that most traditional value investors are like, “Well, if it doesn’t have cash flows, if it doesn’t have book value, then I can’t invest in it.” Which is why like technology and biotech and all these more intangible-intensive sectors have generally been avoided by traditional value investors. And again, that’s why I can exist.
Now move to the crypto space, the same problem exists here, which is the crypto markets are dominated by trend followers, narrative-driven investors, and the value camp, the kind of safe folks, all my friends from Boston, they don’t go into crypto because they don’t feel comfortable that they can assess and ascertain an anchor of fundamental value in the space. But here’s the thing, which is a16z, all these VCs are pouring tons of money into space.
In what way is investing in Uniswap or any of these kinds of Web3 companies that much different from investing in the internet companies of the early ’90s? Yes, they’re early stage. Yes, they have yet to monetize. But at the end of the day, what you need to look for when you look for these companies is how good is their team? Do they have traction users? Have they built the brand? Have they managed to bootstrap network effects? So, it turns out that this framework of intangible value is powerful, if only because it’s a way of establishing a value framework without requiring cash flows or any traditional metrics. So, that’s kind of why I went down the path of saying, “Let’s extend the platform from just equities to also include cryptocurrencies.”
Meb: And next insights as you apply this new lens, what did you see?
Kai: I guess, there are the macro and the micro. So, the macro finding is, you know, as you might expect, which is 60% CAGR in fundamental value over the past several years. So, the number of developers, the number of folks with cryptocurrency wallets, amount of transactions going through the top protocols, Twitter followers. Tom Brady and Gisele are out there in the Bahamas right now talking about crypto. If you think back to like 2014 when I was doing this, it was like me and a bunch of degens and we are flipping seven different currencies amongst each other with no real-world use cases. Or you couldn’t even buy the idea farm with a Bitcoin and now here we are in 2022 and, like, everyone seems to be doing stuff in Web3. Snoop Dogg has an NFT and all the talent has poured into…all the developer talent has poured and it sticks.
I was at a wedding not too long ago for a friend who’s a tech CEO. And so, a lot of the other people there were, you know, highly-placed folks in tech and finance, and we were joking it’s kind of like the crypto conference in San Antonio because everyone was just talking about Web3 and, you know, potentially going into the space or they’re already there. So, it just definitely feels like a lot of talent is moving there. And as an investor, you need to be following the talent, that’s just like principle number one, follow the talents. And, you know, it’s just so impressive to see over the past six or seven years how much growth has been in this space. So, that’s the first finding, which is just like absolute…
Meb: Always follow the nerds, that’s for sure. And I say that lovingly. As an engineer, I can say that. Okay, follow the smart people, follow the nerds. All right, Finding 1.
Kai: Finding 1. Finding 2 is that these value metrics are actually useful, that they actually can be used to help us navigate what is otherwise a very treacherous asset class. So, why is crypto so treacherous? Well, there are like three challenges. And what you want to avoid is showing up in 1995 predicting the Internet and making no money. The equivalent could easily happen to somebody investing in crypto. There are three challenges.
The first is just the sheer number of projects, the opportunities have been very big, and being forced to pick winners. Because we live in this world now in digital markets, it’s the case that oftentimes, only a few winners will drive the return to the entire sector. How do we avoid missing Amazon? How do we avoid instead of buying Webvan? And that’s a big problem.
Unfortunately, a lot of investors own Bitcoin niche but not any of the long tail. So, what happens if Solana or something like that ends up becoming Webvans of the future? And a lot of VCs and other investors lock up in the best projects today, but you need to be cognizant of the fact that with the ecosystem evolving so quickly, you might be missing the new use cases as they arrive. You need to be able to kind of rotate. So, that’s the first challenge.
The second one is just the upward number of scams and the unevenness of the quality of projects. It’s the case, unfortunately, that setting up a $100 million market cap crypto is not that hard, like you and I can just do that right now. We’ll just like fork some other repo and then boom, there we go, buy some followers and whatever. And that’s really unfortunate, there are so many folks trying to cash in on the gold rush and starting straight up rug pulls or just hastily putting together projects that really have no use case.
So, that’s second challenge. And then the third challenge is just the volatility of this hype cycle. Any emerging technologies, not just crypto but the Internet and the railroads, they always go through these boom-and-bust cycles. So, what you want to avoid is buying into a project at such a high valuation that even if it ends up being like around in 10 years, you actually lost money, which took you years and years and years to make the money back.
Meb: You had some great quotes and we’ll add them, but I love the, “Web3 is both a real industry and a bubble.” And you kind of talked about the playbook.
Kai: There are four things and they address the problems aforementioned. So, the first one is diversification, this is the idea that you shouldn’t just buy Bitcoin or Eth, you need to spread your bets across all these competing protocols and all these other use cases that could easily become the killer app. And it’s not just about names, it’s about sectors, right?
You have funds that are focused on Metaverse or DeFi. Well, what if it turns out that DeFi is not the killer app of Web3? The second thing is look at data. So, I mentioned some of the data sources we use already and this is a way of weeding out the scams. A third of your 10,000 projects are just literally nothing? Well, that will show up because you can go on the blockchain and see that there’s nobody using this $1 billion market cap protocol.
You can go on GitHub and realize they just forked somebody else’s code and I mean no changes, and it’s just one dude in Singapore in his mom’s basement who is behind this, there’s no community around the developers. So, that’s the second thing. The third thing is to look at valuation. We don’t care about how many users you have, we care about how much you’re paying to obtain those users.
We don’t care about how many Twitter followers you have, we care about how much you’re paying. Like, imagine you’re a VC buying the enterprise value of this business to achieve these users. And in doing so, it allows us to filter out these firms, these projects that might be around in 10 years but are just so overpriced, it becomes very difficult for them to ever make their investment profit.
And then the final piece is to trade. So, the idea here is, look, I’m a public market investor and I oftentimes have been secretly jealous of my VC friends because they have the opportunity to invest in these world-changing, 100-bagger, power law companies, and I can’t. The thing is that crypto is super cool, especially small-cap crypto, because you get the best of both worlds.
You have the ability to get in early in these power law companies and also have liquidity. The problem is that most folks who approach it from the fundamental side are VCs, people who are not used to liquidity or how to use it. And so, we use it in two ways. The first is to course-correct, and this is the notion that the best projects today won’t be the best ones next year or in five years.
And we want to be able to stay nimble and rotate the portfolio as new use cases and projects arise and conversely fade as certain things start to, like rollover in their usage. And the second piece is on the price side, right?? So, price is fair value of the numerator, in this case. You know, you’re a value guy, you know that sell them to Mr. Market and buy into fear selling the greed.
Well, in this case, Mr. Market is like a raving lunatic. You have like a project that, you know, might be fair value and then tomorrow, some influencer tweets about it and it goes up like 200% and you know it’s going to come crashing back down. So, what you should do is you could sell and take profits, wait for the overcorrect, then buy the dip. So, trading around the fair value, in this case, is like a pretty nice source of returns.
When I was at GMO, we actually did a paper called “The Option Value of Cash” for our clients, and the idea was that the value of cash was kind of a Black-Scholes model. It was dependent therefore on the implied volatility of the opportunities that set in the future. Well, in crypto, you have thousands of tokens with 100% implied vol and some correlation that’s not less than 100%. That’s a pretty fertile ground to be harvesting this rebalancing premium, buying low/selling high in a value-oriented way.
Meb: So, this isn’t just a theoretical concept. And I definitely think it’s a little bit outside my wheelhouse, but I love, love listening to this area and ideas, and particularly anyone who comes to it with a value lens or…I’m a trend follower at heart too, so I’ll keep that in the room. Obviously, this isn’t going to be an ETF, at least not anytime soon. Maybe one day. Tell me about you launched a fund. Can we talk a little bit about it? What is the strategy? What are you guys doing? And how is it work?
Kai: The strategy is a value strategy, and so it trades on a longer horizon. And similar to the ETF, I am a quant and I do use AI and machine learning. But the focus here is not on trying to front-run the next guy, it’s on taking this fundamental intuition and scaling it as broadly as possible across thousands of cryptos or equities. So, the turnover is going to be medium, I would say.
And, you know, since I didn’t mention it yet, I mean, the key here is, you know, the data we use. So, for example, we use…so GitHub is kind of one of the key sources here, whereby you can see…because Web3 is being built in the open, you can see the status of the source code today, you can see it yesterday, all the way back to inception. And it allows you to form measures of IP that are, like, how many iterations? How many changes? How accurately are the developers changing the source code over some period of time?
Another angle we look at that is to look at the developers, who are then contributing to this code. You can actually see each time a change is made with who it is or who is the account name of the folks who are behind it. So, you can create metrics around developer community, how many developers are working on this project? Is it just one or two people? Or is it a massive, robust team of folks spread all over the world contributing their free time to this utopian vision?
And then on the public blockchain side is this idea of being able to look at the ledger…and by definition, you can see it now, and figure out how many people are interacting with this protocol. So, it measures like daily active users, monthly active users, how many unique wallets hold this cryptocurrency? What is the transaction volume? What is the dollar transaction volume going into any point in time? Similar to, like, if you could open up Visa’s 10-K on a daily basis and look to see an update in real-time, “Oh, how is this payment network being used?”
And then the third piece of data we look at is social media. Now, social media is important, not just in Web3, also Web2, obviously. But what makes it so powerful here is that because these companies are, by definition, borderless and decentralized, all of the coordination of the community occurs online in channels such as Twitter, Telegram, you know, Reddit, Discord. And to the extent that you can kind of acquire some of this data, you can see the amount of growth of each project developer community and…sorry, and just like online community, in general, its fans.
And that’s like a very powerful metric for brand and for network effects. So, you have these different sources that are, in many ways, very unique to Web3 that allow us to ascertain the value of the foreign tangible pillars. And that becomes the kind of the core of the strategy, that once we understand and have ways of quantifying intangible value, we’re really well-positioned to be able to run a strategy around it.
Meb: And so, what you end up owning, is it public securities? Do you own cryptocurrencies themselves? Do you own other things? How much does this differ from what, like, a market cap crypto sort of concept would be?
Kai: Yep. So, I thought a lot about how to get exposure to crypto, right? I used to be in GMO asset allocation. This is what we did. There are four ways of playing crypto. There is public equities, which you mentioned, so that’s like investing in Coinbase or Silverhead. There is private equities via VC, and that area has become extremely popular, … etc., are very crowded. There is the Bitcoin niche, it’s what I would call mega-cap crypto, and then there’s like a small-cap token space, they are going to longtail everything else.
And that is what we’ve chosen to focus on exclusively. So, we’re saying let’s not try to commingle too many different things, we want to be a puzzle piece. And of all these four things, this is the area we think is the single most attractive space. But think about it from an allocator standpoint, we can now invest as public investors in an asset class with power law upside. If you would buy Solana today, that’s not interesting. If you bought Solana two years ago, that was very interesting.
And the next Solana lives in the long tail, we have liquidity, we can kind of rotate in an evergreen way as the ecosystem evolves, so we’re not, like, kind of captive for 10 years into what the world looked like in 2022. And also, the point of alpha. So, the lesson of GMO was that being early in the frontier is a huge advantage. So, Jeremy founded GMO in the ’70s and was one of the first guys doing factor investing, not value investing, and was very successful doing that.
Then he did it in international and small-cap and EM, became one of the biggest…GMO is one of the largest EM managers in the early ’90s. So, the lesson there is you want to be on the frontier and what is more frontier than crypto and more specifically, small-cap, long-tailed crypto? So, it just stands to reason that there are a lot of inherent benefits of the beta itself but also what they offer.
Meb: So, I want to give you some money, what are the terms, man? Is this like $100 million minimum? Is this accredited only? How do you guys structure this?
Kai: Yes, this is a private fund. We would love to do it as an ETF but obviously, that’s not possible. And because of that low standards, if you’re interested, you would need to reach out to us to have a conversation.
Meb: I think it’s a super cool idea of investing. It’s odd talking a lot about the value accrue, old Charlie who’s knocking on a hundo, his old quote about fishing where the other fishermen aren’t, to me, this is a pretty thoughtful way to approach an asset class that’s growing. Where do… Like, we’re like at $1 trillion in assets at this point, or is it more than that?
Kai: It’s 2 trillion.
Meb: Two trill.
Kai: And the big things that’s happening is that as the market caps increase, so does the breadth of the market. So, remember, in 2009, it was just Bitcoin. In 2014, it was like seven things. And now it’s 10,000. So, what’s happened is you’ve seen like a flattening of the distribution, where a lot of the masses starting to move towards the tail, these smaller-cap things, where there’s just a proliferation of use cases to everything ranging from cloud computing to decentralized cloud computing, to Metaverse, NFTs, decentralized exchanges.
All that activity is in the long tail. The other half is the top five or so major cryptos and they’re the platforms, that’s Bitcoin, digital gold, Ethereum, world computer. But I think, as an investor, what I’m very interested in is the whole layer down here of this long tail.
Meb: Yeah. You know, I feel like each of these papers probably could have been and should have been an entire podcast. But I definitely want to reserve a nice chunk of time for your most recent paper, which I’m super stoked about. Anything else on Web3 we want to touch on before we mosey on?
Kai: No, let’s mosey.
Meb: All right, Kai. Listeners, if you made it this far, you get a free half-hour from Meb and Kai for sticking around. But to me, this is a timely paper. The word innovation has become pretty buzzword-y the last few years. As you mentioned, you can talk about this concept throughout the history of public markets. I’d love to joke, I can’t remember if it was railroads or utilities in the ’20s that got to a P/E ratio of 65. One of them did. Professor Shiller has a good paper on sector CAPE ratios. But just going to show that our grandparents, what today seems mundane to prior generations is a world-defining idea and concept. So, disruption and innovation, from a value guy, let’s talk about it. Walk me through the new idea of this paper.
Kai: So, the big idea of that paper is like, “What the hell is going on with disruptive innovation?” Obviously, everyone’s talking about inflation and crypto right now. But if you step back and, like, look at a longer timescale, the big story in markets over the past 10 years has been innovation. Software and technology are eating the world. Companies like Apple went from being a $1 trillion company a few years ago to a $2 trillion company today. Companies like Tesla went from basically being nothing to a $1 trillion company.
So, the world is clearly changing. And we’ve seen the entrance via SPAC IPO direct listing of a bunch of disruptive companies into the public markets and now there’s this big tug of war between what many investors view as a new guard compared to, like, the legacy companies. And then what’s happened is that fund managers have realized that it makes sense for them to launch products to package these innovative companies into funds. So, obviously, Cathie wood and Ark, they’re the OG in this space, they launched in 2014 their ETF.
But since then, many other players have come into the space trying to get a piece of the action. The problem is that none of this has really worked the past year, you know, Ark is down 60%, Zoom, Teladoc, all these guys are down 80%. And it’s ignited this massive debate, which is what should we do with our disruption stocks? Is this, as the defenders of innovation claim, a generational opportunity to buy world-changing companies at a deep discount?
Or is it, as many of the detractors have said, just to kind of prove that all innovation investing is a rebrand of growth investing? That this is something that folks have been trying to do for years seducing the innocent retail investor into buying these stories stocks and then getting destroyed when the bubble pops. Is that what’s happening here? Are we about to enter a dot-com style winter in innovation? The goal was to approach this with an objective and evidence-based approach.
The first thing I had to do was to reverse engineer what is it that is innovation. And, you know, you think about what people who are innovative investors say today, they say, “Hey, there’s like AI, there’s blockchain, there’s VR.” You can kind of get a sense for what they’re doing is they’re saying, “Let’s figure out what technologies are going to change the world, and then buy the companies that have exposure to that idea.” That’s what they’re doing.
So, what we need to do then is to figure out, “Can we get a historical record of the technologies that have been world-changing throughout history?” And the good news is that we can go to the patent’s data and have all that information sitting in front of us. So, the U.S. Patent and Trademark Office, they maintain records back to 1790. The first patent was actually signed by George Washington himself, which is a really fun piece of trivia. And since then, there’s been exponential growth in the number of patents available.
What we do is we say, “Can we look at these patents and then cluster them according to the technology that underlies it?” So, it might be the case that a patent on Lidar and the one on image recognition, they’re kind of related to this idea of autonomous vehicles, so we kind of put them together. And then we try to see through the historical record, can we identify trending technologies? So, my favorite finding is just this simple kind of nerdy thing, which is we can actually observe through the past 200 years, the rise and fall of key technologies.
We can see that railroad was really popular in the 1840s and then kind of diminished. Electricity started becoming a thing and then peaked in 1900, the automobile then grows and fall, followed by circuitry in computing, and then the internet. You can literally see the major technological revolutions that have powered human civilization in the patent record. And I thought that was super fun and also just kind of indicates that we’re on the right track here looking at this data.
Meb: Well, in the paper, you talk a little bit about how essentially some of these technologies tend to trend rather than necessarily mean revert. There’s like an echo of maybe three to five years, but there are some false starts, maybe electric vehicles, etc. Can you expand on that at all? What’s the way to think about the life of some of these? You know, because some of these concepts and trends probably will last a really long time and some will just wither away and turn into something else or die altogether.
Kai: Right. Yeah, and that’s very much the challenge, does technology trend or does it mean revert? I mean, that’s the fundamental question we’re trying to ask. The example of the car that was really cool, like, electric vehicles that…and I didn’t know this before I’m doing this research, was that the electric vehicle was actually the best-selling car in like 1900s. And it was competing with steam and the internal combustion engine for market share at the time.
And what ended up happening was the internal combustion engine improved fast enough that it eventually supplanted electrical vehicles and then became the mainstream car. And then EV made a comeback again in the late ’90s-2000s, and then faded out again. And then now with Tesla, I think it’s finally time for it to shine. And really, the problem with EV has been the gating technology of battery technology just hasn’t really been good enough to give any decent range until now. So, we’re now starting to finally solve these problems, which is opening up the potential of the asset.
Meb: And by the way, listeners, we’re not going to get into a lot of the fine details. But reading the paper, I think it’s really cool on some of the specifics where Kai is talking about, you know, the patent office, how they characterize technology groupings and classification, and very specifically how a huge percentage of categories are omitted and included. So, if you’re trying to do some work on this on your own, certainly read the paper because it gets deeper than what we’re going to talk about today.
He’s glossing over what must have been an enormous amount of work that the team put together. So, one of my favorite charts of this whole paper is sort of like the greatest hits by decade because I love looking back. It’s like the old movie with Dustin Hoffman, “The Graduate,” where he’s talking about plastics and looking back over time to seeing refrigerators and lasers, all sorts of fun stuff. What’s the big ones today? What’s popping up as some of the greatest hits of the 2020s?
Kai: The biggest hits today…so, actually, look at this. So, over the past decade, you can see that major technologies like cloud computing, social network, AI, one interesting finding here is that social networking became very powerful and very influential really early on, but it actually started fading in the past few years. So, that’s the one example of the modern technologies that is actually on the decline, everything else is still growing at least in some ways.
Now if you go down to the list, what you find is that the technology that has grown fastest is blockchain. And there are now 900 or so patterns over the past few years with a growth rate of about 400% over this period of time. The next highest is AI. Now, AI has been growing for a long time. It has been a very important technology for a long time. It’s a 20% growth rate from a higher base.
Meb: Yeah, AI is the granddaddy as far as the absolute level, if you look at the kind of chart, that sucker is growing in the Terminator T-2000s sort of way, that thing is getting scary.
Kai: Yeah, and it has the potential to cross-cutting across all technologies. That’s why it’s the biggest technology, it just has the most…you know, the biggest TAM, so to speak. And we have quantum computing, 3D printing, Internet of Things, VR, autonomous vehicles, robotics. And then the last one is actually fun and that’s cloud computing. And, look, it’s only growing at 20%. We think it’s a “mature” technology but it’s still pretty decent. So, even after all these years, you know, cloud has been around for a long time now, it’s still putting out pretty solid numbers.
Meb: You then kind of go a step further…and this is super fun. Obviously, a lot of these would be not that surprising to people. But others, it’s interesting to me to see…and you guys have more data than we would, but 3D printing, which may have had hype earlier but maybe still a lot of development and maturity ahead of it, perhaps, who knows? On and on and on.
But you then start to make the transition to connecting this with companies. And nothing in your paper, I think, is more hilarious to me than seeing the top blockchain patent holder being IBM and number three is Bank of America and four is Accenture, which is just fascinating to me. IBM, how this stock which is always the number one on…isn’t it the number one patents for like 50 years or something?
Kai: That’s basically a business model at this point, right?
Meb: It’s a giant patent troll database. But I think that’s so humorous to me when I saw that, I was like, “Okay, of course, it’s IBM.”
Kai: Yeah.
Meb: So, tell me the next chapter, which is investing in innovation. How do you start to link this to actual companies or investable ideas?
Kai: So, the key idea here is the…remember what we saw in the Dimson-Marsh study, GDP growth is not necessarily linked to stock returns? That was kind of my approach here where I was like, “If everyone knows that innovation is a good thing and that innovation generates progress and growth, it’s not clear to me that you as an investor in innovation have a God-given right to returns.”
Like, what would happen if the market just prices it? Or in the case of the growth investing argument, probably overriding it? Well, in that case, you’re not going to actually make any money doing it. So, it would have to be the case that the market is undervaluing the innovative companies in order for there to be a systematic return premium associated with innovation. I said, “We now know because we know technology is a trend that we can predict the future path of technology by extrapolating the past.”
So, for each point in time, we’re going to build a basket of, say, the top 10 technologies then, and what we’ll do is we’ll rotate over time into the next thing. So, this was like the chart I created on, like, the ladder of innovation where I show like S-curve, which is like the curve of adoption for technologies. You visualize climbing a series of escalating S curves where you kind of always want to own them at the steepest part of the S and then kind of grow into the next one as that technology matures.
So, that’s pretty much the idea of we have 10 technologies at any point in time and kind of that keeps recycling as the world evolves. So then, the next step becomes, “Can we determine which companies have exposure to each of these disruptive technologies?” That’s actually pretty easy to do because each patent has an assignee and you just need to map that back to its corporate owner. And so, for any given arbitrary basket, you can say, “All right, create a 3D printing ETF,” boom, one click of the mouse, I just go in and figure out 3D printing patents, figure out which guys own them, done.
So, like, we create these thematic portfolios, these 10 today, smush them together, and that’s what you own. So, there’s like 200 stocks currently, but that number has obviously changed over the course of time. And then what we do is we look at returns. So, this goes full circle to the question we initially asked, which is now that we have this backtest where we’re using machine learning to classify patents into technologies and then find the trending technologies, and then find the companies that are investing in said technologies, what does the return of that portfolio look like over time?
Well, it turns out that it does beat the market and it returns about 2.6% per year more, which is, you know, pretty solid outperformance. If you look at the relative performance, I mean, you find that it has a pretty steady uptrend with this big dislocation in the dot-com bubble where it went up and then back down full round trip. And that’s not altogether unsurprising, but does point out a key weakness in the strategy, which we’ll address, I guess, later.
Meb: Okay. And the cool part, listeners, this sucker goes back all the way to the 1970s, which is fun to look back on what’s going on. All right, so you got to put on your references earlier, like this possibility of inefficiency. What’s the explanation? You got two you propose as possible, what do you think is the best reality?
Kai: It’s got to be a bit of both. For context, there are two explanations as to why we had this outperformance. The first is just the market is inefficient. The simplest explanation, these assets are undervalued, they’re misvalued, they’re hard to measure, the Wall Street is very short term focused and they don’t value the world-changing impacts over long periods of time of technology. The second explanation is risk premium, that deep technological research is inherently risky, is inherently boom or bust.
You put in…Zuckerberg out there putting $10 billion a year in the metaverse, that could make Facebook into a $10 trillion company or into a zero, and investors are pushing his P/E ratio down as a result of that aggressive bet. So, it’s a bit of both, I think. And especially on the risk side, what I initially came to this thinking was, “Well, it’s because it’s exposure to higher volatility, higher growth, etc.” But you can strip all these things out. And the narrative today is all that innovation investing is just a rebrand of growth, that all these guys are doing is growth investing and like putting a fancy name on it.
Is that actually true? Well, we can test the beta or the factor now going back to the seven ways to growth, and the first thing you find actually is that it’s on average not been that hot. It’s only been 0.18. And moreover, you find that it fluctuates quite significantly through time. At GMO, we used to do these charts value of value, right? We look at the spread between value and growth stocks and then look at the output spread all through time to figure out is it cheaper/expensive.
So, you can do the same thing here with any basket of stocks. In this case, innovation companies. Are they cheaper/expensive? Well, what you find is that the beta peaked in 2000, basically, in a tech bubble. Because what happened was that innovative Internet stocks became expensive because people said, “This Internet thing is amazing, I want to buy all these stocks,” and then the valuations went up. And that’s probably what led to the drawdown that I pointed out earlier.
The interesting thing you also find is that this number has trended down over time, it actually looks like a local minimum now. And what does that mean? I think the reason why is that, at this point, the idea of using innovation and disruption is, like, mainstream. Every company is embracing innovation. And by the way, the most innovative companies by many measures are the biggest, most profitable ones.
Like, Google has more publications in AI than Stanford and MIT, the next two highest combined. They are outperforming, as a private institution, the leading academic organizations in this country and that’s a pretty big deal. So, I think the character is six of these companies has changed. Innovation is not the same as it was before, not with flaky dot-coms, and it will likely change again in the future.
Meb: So, I’m just upset that I see you think that Chipotle is non-innovative. Have you seen some of the new offerings they’ve been putting out lately?
Kai: What is your favorite menu item at Chipotle?
Meb: Well, actually, I’m joking because I actually don’t Chipotle anymore because they opened up near our office years ago and then we used to go there all the time. And then it just happened where there was just this total refractory period where no one could eat anymore because we just topped out on Chipotle. So, I don’t know if I’ve been back. They have amazing chips, though. Really great chips.
But anyway, okay, so we walked through, there are some sector French-Fama betas we could talk about, but you talk about really the most famous of the “disruptive innovative funds,” you talk about Ark and kind of where the factors lie. Are there any surprises there? Are they one and the same? And do all the factors play out in the way that most investors think it would play out?
Kai: No, actually, the results here are a bit counterintuitive. The first thing I said was, why is it that this innovation factor has not had a bigger drawdown the past year? Because the common narrative and common conception is that innovation stocks are in this massive drawdown as exemplified by Ark’s performance. So, I said, “Well, the only way to solve this puzzle is to go into Ark’s performance and kind of figure out what’s going on here.”
So, I do this on my French decomposition and, you know, break down its returns over the past, I guess, seven or eight years into five different buckets. It’s four different buckets. So, it’s market exposure, style factors, so that’s like valuing growth, and size and profitability and momentum. And then innovation, so its exposure to innovative stocks. And then finally is alpha. So, the first thing you find is that Ark has made 12.4% annualized from its beta, which makes sense, the market has gone up.
And you also find that it made about 4.2% annualized on innovation. You know, Cathie was very prescient in saying that, “This factor is something that, you know, makes sense to put into my portfolio.” This was many years ago. And even more impressive is that its alpha was 5.4% annualized. So, we’re saying that, “Look, they’re taking a lot of active risks but it’s paying off.” This is net of fees, so they’re actually doing quite well finding the winners from amongst the innovation universe, right? Tesla.
Now, here’s the problem is style factors. Style factors for the first five or six years of the sample were actually a positive contributor, but over the past year, it’s just been disastrous. I mean, it’s been 40%-50% drawdown. Like, let’s imagine that 80% of their losses in the past year has come from style factors. And then as you decompose that further, what you find is that it’s really two components of the Fama-French style factors that are really killing them. That is growth and junk or low profitability.
So, the way I frame this is, “Look, Ark is making a bet on innovation but they’re also making a bet on growth stocks and junk stocks.” Right? They’re buying these speculative companies at very high valuations without any profits or even revenue in some cases. And that factor tilt has been what has led to a lot of their gains the past year, not necessarily the innovation piece.
Meb: Yeah. If you were to look at the innovation basket, this might be a good time to segue into the next one, but we may be too early and out of order.
Kai: No, I mean, look, the only point I would make now is that…where are we now? Let’s take stock of where we are now in this paper. We have established that innovation has positive long-term returns, and moreover, that it is distinct from growth technology and other Fama-French factors. It is a unique asset class in and of itself, a unique factor, an undiscovered factor that, as you were saying earlier, could be like a sixth Fama-French factor, right? This is its own thing.
And that’s really cool. Well, the problem is that oftentimes, as you saw in that chart of the beta, oftentimes, innovation stocks are prone to being caught up in these bubbles. We saw that in the dot-com and if we were able to take the data back even further, we’d likely see that in the Industrial Revolution, the railroad revolution, you know, canals, Model T, etc.
Meb: This is coming full circle to your bubble days, man. You got this beautiful bubble chart that looks a lot like your old buddies on Canal companies, UK railways, and these are beautiful charts. Information Age. People go bananas every once awhile.
Kai: And it makes sense. I mean, the allure of innovation is very seductive, right? If we think that, you know, let’s say Web3 is going to be the new Internet and completely changed the technological paradigm, then we’re likely going to be very interested in throwing some money. That’s just the way humans are. And it doesn’t mean you shouldn’t do it because the opposite of doing that is to say, “I’m so afraid of potentially getting caught up in a bubble that I’m trying to sit out innovation in general, and I never buy any stocks that have any potential for being world-changing.” Right? So, you need to kind of balance those things and that’s where DARP comes into play because we think it’s a way you can thread this needle to stay invested in innovation, yet avoid the most overvalued speculative stocks using these valuation metrics.
Meb: So, do you do it where you skim off the really expensive, or you’re concentrating in the cheap? What’s the approach here for DARP? Great name, and what does that stand for?
Kai: So, DARP is Disruption at a Reasonable Price. Obviously, it’s a play on Warren Buffett’s claim, “It’s far better to buy a wonderful company at a fair price than a fair company at a wonderful price. So, it’s GARP, but DARP.
Meb: I like it. So, tell us what that means and how does that actually play out.
Kai: Yep. So, what we do is we say, “Let’s not just buy all the innovative companies, let’s look within that universe and rank them on their innovation yield.” Which, in this case, similar to what I described with the intangible value scores, we look at the number of innovative patents divided by market cap. So, this forms a score for, like, how much innovation are you getting for dollar invested against value yields. And then what we do is we take the most expensive stocks, the worst-performing stocks on that thing, and lop them off, and they’re done.
And then we just take the remaining ones, so that’s only disruptive companies at a reasonable price. And what we find is that the value characteristics of that portfolio increased dramatically. The basic innovation portfolio had an innovation yield of 13% and that goes to 49%. And then all the kind of traditional value metrics like price-to-earnings, sales yield, they also go up. Right? So, we’re able to get like a decent improvement simply by getting rid of the most overvalued companies within the innovation space.
Meb: And not surprisingly, given everything we know about markets. I’d love to see that free cash flow yield, baby. Does anyone else publish patent yield? Is that just you? I don’t know if I’ve ever seen that anywhere.
Kai: I’m not seeing it elsewhere.
Meb: It’s cool. All right. So, drumroll, I’m guessing getting rid of the crazy expensive helps?
Kai: It does help, it does increase returns, but the big reason…the big thing is that it helps with the drawdowns. And that’s the reason why, right? Sometimes the best offence is a good defense. In the dot-com bubble, there’s nowhere really to hide, let me just say that, but like it does help a lot. It helps reduce the drawdown, you get back to your high watermark two years earlier. And then in the current episode, this past year, there is no drawdown because you have managed to avoid the Zooms and, you know, Teladocs that are down 80%, and instead buy firms that are still engaged in innovation but are average price, like, normally priced.
Meb: You talk a little bit about…and it’s beautiful equity curve, listeners, you can check it out again in the paper, but I love this strategy, I love this idea, by the way. I think it’s really super cool. What’s the overlap with some of the concepts we talked about earlier? Is this something where there are a lot of philosophical commonalities to the intangible value ideas? Is that something where there’s a ton of overlap? Not much? They look kind of similar or different? What’s the story?
Kai: This concept is a subset of the intangible value framework. The way I would say is what I’ve created here is a innovative patterns yield, a way of identifying companies that are engaging in innovation in their patents. Now, patents, of course, are just one form of IP, right? Because as a company, I can choose to innovate but then keep it as a trade secret or reframe it in a different way outside of the patent legal system. So, I view this whole edifice that I’ve built in this paper as being one subset of IP, which, of course, is one subset of the four-pillar framework.
Meb: So, by definition, it’s going to have some definite commonalities.
Kai: Yes, so the correlation, I did run this of this factor, which, by the way, is part of the IP factor, so of course, it’s going to be high. It’s 81% with IP but it is close to zero with the other three pillars. Actually, that’s interesting, because these are all four distinct concepts. And so, brands, for instance, companies that have high brand are like your Nikes, your Polos, right? They have a good brand with Easter maybe, but they don’t necessarily have great IP.
And firms with, say, very strong human capital, like, you know, your McKinseys and Goldmans don’t necessarily have strong IP either. I view these four pillars as being kind of uncorrelated distinct concepts, which is part of the reason why it’s so important to not just have innovation as an asset class. Like, the whole point of this paper is to say, “You should do this.” But then at the very end I’m like, “But also don’t just do this, if you buy this argument, just go one step beyond, buy all four pillars because in doing so, you’ll now have a place to hide when innovation becomes expensive.”
In the dot-com, bubble, all innovative stocks were expensive. There’s just nothing you could do. You could use DARP to help a bit but you’re still kind of in a tough place. But if you own the other pillars, then you can just rotate your capital to the other sources of value today, so brand and human capital. I have this quote here that, “Look, like, genius alone has never been enough to drive financial success, history is littered with brilliant ideas that failed due to the inability to raise capital, attract talent, build grant, or bootstrap projects.” So, look, IPs are important but it’s still just one piece of a larger puzzle.
Meb: Yeah, you watch that old documentary on Tesla, the dude’s in the news every day, you know, where he talks about almost going bankrupt hours…I mean, it may have been days but it was certainly like hours away, back in the roads for 1.0 days. But fascinating is you look at a lot of these ideas and companies, and I look back, you know, even now to thinking about the ones that made it, the ones that didn’t.
I mean, here we are with Tik Tok and YouTube versus Vine. As we talked about Twitter, Vine being the early entrant there but they got put out to pasture but could have been a bigger market cap than Twitter if they just kept it open, some of these ideas. But that’s creative destruction, that’s part of all this fun game we play. So, the strategy, here we are, coming all full circle back to your original concept of which this is one piece. What’s the interpretation from where we sit here in Q2 2022?
Kai: The long-term story of innovation is you want to be long, and you do want to step out when things get super crazy as they did in dot-com. But other than that, just being strategically tilted towards innovation and trying to just avoid the most expensive companies, right, using this DARP approach. It has worked and it’s worked pretty well. And again, if you blend it with other pillars, you’re going to be even better off. So, that’s very much, like, been the focus of our kind of intangible value research.
Meb: How challenging…and it seems like you guys potentially include some ADRs in this, you can correct me if I’m wrong, but how challenging is this to replicate internationally if at all possible?
Kai: It should be quite possible because there’s the U.S. Patent and Trademark Office, which is the database I focused on here. And then there’s, you know, a bunch of other ones and then there’s these aggregators where all the data kind of sits and there’s just, you know, a lot of reciprocity. A lot of companies, for example, that patent in the U.S. are international companies. Because if you’re Samsung or Sony, if you want to sell into the U.S. market, you’re going to want to patent in the U.S. market too. So, it should be pretty doable to scale globally.
Meb: Yeah. Very cool, man. Well, we’ve been at this for a while, so let’s start to try to wrap this up so you can be released into the Brooklyn evening. When you look out to the horizon, man, you’ve been cranking out a lot of papers, a lot of ideas. This one obviously just hit the hopper. But as you think about putting pen to paper, next time you come out and we’re having sushi, what are you thinking about? What’s got you excited, confused? Obviously, you’re building an emerging money manager, that’s your day job. But is there any research ideas, things that are on the brain currently that you’re thinking about?
Kai: I think you hit on one very interesting point, which is, so far, a lot of my research has been very U.S.-centric, which makes sense because the U.S. is the single most intangible-rich market. Europe and Asia and South America and Africa have been less innovative than the U.S. over the past decade. But time only moves forward and it would make sense as an investor to also be able to figure out how to access intangible value abroad.
And one of the very interesting and cool things about my framework here is that it’s all based on statistical machine learning, statistical natural language processing. Which means if I want to go to Japan, let’s say, and start, like, looking at the filings there or whatever, I don’t need to go hire a bunch of, like, Japanese speakers or Japanese linguists. I can take my model, which is currently being trained at the base layer on all of English Wikipedia, and then just train it on the Japanese corpus, maybe it’s Wikipedia. And then you can do that in Thailand and Vietnam and all over the world.
And that makes what we’re doing in the U.S. so affordable, so easy to scale, right? That was why I spent two years building the technology because it becomes just very easy to then kind of scale it to the next year after we make that up-front investment. And by the way, if you do that, I think there’s a lot of alpha, especially in emerging markets where a lot of information is in English but if you’re in Thailand, a lot of the information will not be in English. And to the extent where you can comprehend information in a way that a lot of the bigger funds, it’s not really worth their while to put boots on the ground there to do, that’s a pretty big source of edge.
Meb: Yeah, that’s certainly an obvious extension of all the work you’re doing and the areas where it’s not as efficient. I mean, expanding this abroad, super cool idea, as particularly those markets evolve too. A lot of the emerging, in particular, is a little more traditional but things are changing to them fast that it’ll be fun to watch. What’s been your most memorable investment? You look back on your career, anything come to mind? Good, bad in between?
Kai: I got to get one for you, man.
Meb: Great.
Kai: All right, most memorable investment. You know, I got married a few years and as part of the whole ritual, I had to buy, like, a diamond for my wife to kind of signify my commitment. You know, but as a value guy, what I didn’t want to do was walk into the nearest jewelry store and, like, get my face ripped off. So, instead, as any good quants would, I went online and web scraped a bunch of databases. It turns out that a lot of these diamond vendors, like, have their entire inventory online.
So, I just pulled all these data, I had a spreadsheet with hundreds of different diamonds, you know, the 5Cs, I think it’s like color, clarity, cut, something like that, right? Carat. And then it’s built like a five-factor model, where you predict the price of the diamond as a function of the 5Cs and then what I did was look for residuals. Imagine you have a scatterplot and you’re looking for things that are below that line.
I found the few diamonds that were below, the most below, the most arbitragable or whatever, and the one that was in my price range and bought that. And what was so gratifying was I had to get it appraised for like insurance purposes and when I did that, it was like 50% higher. So, I literally found a diamond in the rough and my wife is going to be super thrilled that I mentioned this on air.
Meb: So, how long till the diamond arb private fund? A little too physically risky, you might get some guys coming after your kneecaps at some point if you publish all the secrets of the diamond industry trading. I’ve seen a lot of pitches in the last 5-10 years on sort of the artificial diamond production, whatever the right word for that would be, growing seems to be not the right word. But some of these online platforms…for a long time we did this…oh, my God, how many years ago was this?
We did a ETF contest where people would submit ETF ideas. Now, obviously, a lot of them are crazy and silly but many of the ideas have since launched. There used to be a lot of Swiss cheese hole in the ETF landscape, they get to be smaller and smaller by the day. But for a while, there was a time an ETF filed but it never came to market. I don’t know how they could possibly do it other than with swaps or something but…
Kai: It was like a GLD type thing?
Meb: People were trying. Yeah, I don’t know what SIG is but it was filed. So, who knows? We’ll see if that ever makes it to market but DIA is already taken. Kai, this has been a blast, man. I love reading your research. Where do people go if they want to keep up with what you’re doing? We’ll obviously post all these in the show notes links, but where can people find more about you and what you’re up to?
Kai: Just check me out on my website, that’s sparklinecapital.com.
Meb: Awesome, man. It’s been a blast, we’ll do this again soon. Thanks so much for joining us.
Kai: Good time. Thanks for having me, Meb.
Meb: Podcast listeners, we’ll post show notes to today’s conversation at mebfaber.com/podcast. If you love the show, if you hate it, shoot us feedback at feedback@themebfabershow.com. We love to read the reviews. Please review us on iTunes and subscribe to the show anywhere good podcasts are found. Thanks for listening, friends, and good investing.
