WEBVTT - Measuring Corporate ‘Dark Matter’ With an ETF 0:00:06.000 --> 0:00:14.280 Welcomer trains. I'm Joel Webber and I'm Eric Belchunas. Eric, 0:00:14.360 --> 0:00:15.400 this was your idea. 0:00:15.720 --> 0:00:18.440 I'm really curious about where this conversation goes today because 0:00:19.320 --> 0:00:21.280 all of this stuff was like candy as I've prepared 0:00:21.320 --> 0:00:21.600 for it. 0:00:21.920 --> 0:00:24.360 Yeah, you know, we're getting back to the basics today. 0:00:24.440 --> 0:00:25.920 You know, how do you value a stock? 0:00:26.280 --> 0:00:26.440 Now? 0:00:26.440 --> 0:00:30.760 That's his stock conversation. But stocks make up index funds, 0:00:30.840 --> 0:00:33.320 which make up ETFs, and a big wing of ETFs 0:00:33.360 --> 0:00:36.600 are called smart beta, which are ETFs that sort of 0:00:36.840 --> 0:00:40.200 use metrics that an active manager would use, price to earnings, price, 0:00:40.280 --> 0:00:43.599 the book, dividends, et cetera. And within there there's all 0:00:43.680 --> 0:00:46.080 kinds of variations. So in a way it's active even 0:00:46.080 --> 0:00:48.560 though it uses an index. And so there's an ETF 0:00:48.600 --> 0:00:52.080 that launched recently called the Sparkline Intangible Value ETF. I 0:00:52.200 --> 0:00:56.760 TAN is the ticker, and this uses the intangible value, 0:00:57.200 --> 0:01:00.920 which this person claims is a new factor. The reason 0:01:00.920 --> 0:01:02.920 to call my attention is I was at the Democratized 0:01:03.000 --> 0:01:06.240 quant event I don't know six months ago, which Wes 0:01:06.319 --> 0:01:09.240 Gray from alf Architect puts on, and I saw this 0:01:09.280 --> 0:01:13.160 guy Kai Wu debate Cliff Asnas who is like giant 0:01:13.200 --> 0:01:16.240 in the quant world. I mean, he's like heavyweight, a lister, right, 0:01:16.720 --> 0:01:19.560 and Kai has this ETF. It's indy and it was 0:01:19.600 --> 0:01:21.920 sort of like a David and Goliath debate, no offense, 0:01:22.160 --> 0:01:26.319 but Cliff had met his match here. I thought Kai 0:01:26.480 --> 0:01:29.080 made some very good arguments. I was more on his 0:01:29.200 --> 0:01:30.760 side by the end of the debate. Cliff was a 0:01:30.760 --> 0:01:33.240 good sport. It was a great discussion. I love the quants. 0:01:33.440 --> 0:01:35.800 They do a very academic type of rigorous debate when 0:01:35.800 --> 0:01:37.679 they have events, and I like that. It was two 0:01:37.760 --> 0:01:39.640 sides presented, and I thought we got to get this 0:01:39.680 --> 0:01:42.240 guy on because not only is intangible value interesting and 0:01:42.280 --> 0:01:45.039 people should know what that means. But when you think 0:01:45.040 --> 0:01:47.960 of smart baita ETFs, like a value ETF, many of 0:01:47.960 --> 0:01:50.320 them use price the book. Well what does that mean? 0:01:50.320 --> 0:01:52.280 What is book value? Well, a lot of the book 0:01:52.320 --> 0:01:55.040 values are old. They don't use things like the brand. 0:01:55.400 --> 0:01:59.200 They use like how much actual literal capital goods the 0:01:59.280 --> 0:02:02.320 company owns, and so they don't use real estate. So 0:02:02.320 --> 0:02:04.240 there's this huge debate in the quant world on how 0:02:04.280 --> 0:02:07.400 to actually define price the book, and that is a 0:02:07.560 --> 0:02:10.120 major pillar of how you define value. So if you're 0:02:10.160 --> 0:02:13.000 shopping for a value etf or value manager. This stuff 0:02:13.040 --> 0:02:13.800 is important to know. 0:02:14.440 --> 0:02:18.000 Joining us Kai Wu who's the founder chief investment officer 0:02:18.280 --> 0:02:21.120 of Sparkline Capital, as well as Chris Kane, who's a 0:02:21.160 --> 0:02:31.720 quant analyst at Bloomberg Intelligence, this time on Trillions the intangibles, Kai, Chris, 0:02:31.840 --> 0:02:32.800 Welcome to trillions. 0:02:32.919 --> 0:02:34.600 Thanks for having me, Thanks for having me. 0:02:35.000 --> 0:02:38.399 Okay, Kai, what is intangible value? 0:02:38.960 --> 0:02:42.000 So an intangible asset is, you know, as Eric was saying, 0:02:42.280 --> 0:02:45.640 anything that's not your kind of factories, your cash, your property. 0:02:45.919 --> 0:02:49.360 It's at Sparkline we have four pillars of intangibles. We 0:02:49.400 --> 0:02:53.800 talk about intellectual property, brand equity, human capital, and network effects. 0:02:53.919 --> 0:02:57.560 So again IP, human capital, brand, network effects, and these 0:02:57.600 --> 0:03:00.920 assets are becoming more and more important for companies today. 0:03:01.360 --> 0:03:04.600 Why is value investing not capturing this? 0:03:05.240 --> 0:03:07.200 So when you think about value investing, it was really 0:03:07.200 --> 0:03:11.240 popularized in the nineteen thirties with bang Gram's Security Analysis book. 0:03:11.760 --> 0:03:14.160 And you go back to the thirties, right, the economy 0:03:14.200 --> 0:03:17.680 was fully industrial. The big companies were railroads and textile mills, 0:03:18.000 --> 0:03:20.440 and as a result, you know, these sorts of intangils 0:03:20.480 --> 0:03:23.600 didn't really matter too much. But today the biggest companies 0:03:23.600 --> 0:03:26.640 are you know, Apple, Facebook, firms for whom their book 0:03:26.680 --> 0:03:28.839 value does not actually is not actually. 0:03:28.560 --> 0:03:29.800 Required to produce earnings. 0:03:30.160 --> 0:03:32.960 And so you know, think about you know Apple for example, right, 0:03:32.960 --> 0:03:35.400 it's their brand, it's their the network effects around the 0:03:35.440 --> 0:03:38.800 iPhone iOS ecosystem, it's the human capital and IP around 0:03:38.840 --> 0:03:41.520 you know, their internal processors and such that allow them 0:03:41.560 --> 0:03:43.080 to earn such fat profit margins. 0:03:43.720 --> 0:03:47.040 Why hadn't this been captured in an ETF until yours? 0:03:48.200 --> 0:03:49.240 You know, I don't know. 0:03:50.200 --> 0:03:55.040 There have been attempts to to kind of capture some 0:03:55.120 --> 0:03:59.880 of the intangible assets using accounting based metrics. So what 0:04:00.120 --> 0:04:03.480 the weird anomalies within the way accounting works that gap 0:04:03.480 --> 0:04:08.000 accounting is that they allow for the capitalization of physical 0:04:08.040 --> 0:04:11.240 capex but not intangial assets. Right, so there are they 0:04:11.240 --> 0:04:13.960 have been attempts to reverse that by saying, all right, 0:04:13.960 --> 0:04:16.680 you're gonna spend one hundred million dollars building a building. 0:04:16.320 --> 0:04:17.840 A factory that gets capitalized. 0:04:17.880 --> 0:04:19.560 You're gonna spende hundred million dollars doing R and D 0:04:19.640 --> 0:04:22.440 to develop a patent that gets capitalized, and doing so 0:04:22.480 --> 0:04:24.360 you can kind of create a more holistic version of 0:04:24.360 --> 0:04:26.599 book value is a little bit better. But what we 0:04:26.720 --> 0:04:29.279 found here at sparkline is that you know that only 0:04:29.279 --> 0:04:31.840 takes you so far, because again, like the link between 0:04:31.880 --> 0:04:34.360 the money you put into R and D and what 0:04:34.400 --> 0:04:37.279 you get out is super wide. So what we like 0:04:37.320 --> 0:04:40.680 to do here is to focus instead on the actual products, 0:04:40.680 --> 0:04:43.120 the actual outputs. You know, what patents do you actually get, 0:04:43.160 --> 0:04:44.839 how strong is the brand you actually build through your 0:04:44.839 --> 0:04:47.640 advertising efforts? And I think that's kind of a novel 0:04:47.640 --> 0:04:50.320 and unique approach that really only became available with the 0:04:50.360 --> 0:04:52.960 advent of you know, instructured data and natural language processing, 0:04:52.960 --> 0:04:53.960 which we'll get into in a bit. 0:04:54.240 --> 0:04:57.160 So it is your is your fund. 0:04:57.680 --> 0:05:00.560 It isn't a value ETF that actually just uses this 0:05:00.600 --> 0:05:03.680 one tweaks this one part. It's more of let's go 0:05:03.760 --> 0:05:06.880 after the companies with the highest in tangible value. Again, 0:05:06.920 --> 0:05:09.120 that's different then let's do a traditional value ETF. But 0:05:09.200 --> 0:05:12.360 let's correct how we define price the book, right yours 0:05:12.400 --> 0:05:14.480 is let's go after these stocks that are high and 0:05:14.560 --> 0:05:16.359 intangible value, right yeah. 0:05:16.400 --> 0:05:18.680 And I think one important thing to mention is that 0:05:18.839 --> 0:05:20.440 it's not like we're going to go after the companies 0:05:20.600 --> 0:05:23.839 with simply the most total overall innovative patents, let's say, 0:05:24.040 --> 0:05:26.440 because then that just map do large cap names. Right, 0:05:26.480 --> 0:05:28.440 what we carry about is how much as a shareholder 0:05:28.480 --> 0:05:30.880 you get per dollar invested. Right, So it's very similar 0:05:30.960 --> 0:05:33.480 like a dividend yield or like an earning yield. So 0:05:33.480 --> 0:05:35.719 for each dollar invest how many you know PhDs do 0:05:35.760 --> 0:05:37.960 I get? As an investor? How many you know Twitter 0:05:38.000 --> 0:05:40.760 followers do I get? And so these are kind of 0:05:40.760 --> 0:05:43.440 proxies for intangible assets. But again the key just being 0:05:43.440 --> 0:05:46.200 that they're price based, very similar to price to book, 0:05:46.200 --> 0:05:48.400 but you know, using kind of a more expansive set 0:05:48.839 --> 0:05:49.880 of variables. 0:05:50.960 --> 0:05:53.520 And let's bring in Chris Kane here, because Chris spends 0:05:53.560 --> 0:05:58.120 all day looking at this quantitative data he has builds indices, 0:05:58.839 --> 0:06:01.240 And I was curious, Chris, you know, sort of your 0:06:01.279 --> 0:06:03.560 take on this. And you have to have a price 0:06:03.560 --> 0:06:06.560 to book in your metrics and how you define that, 0:06:06.680 --> 0:06:09.880 and I'm just curious to get your you know, how 0:06:09.920 --> 0:06:11.359 you've considered this for your work. 0:06:11.480 --> 0:06:13.599 Sure, I mean, I you know, I love this concept. 0:06:13.600 --> 0:06:16.200 You know when I go to you know, customers and 0:06:16.240 --> 0:06:18.000 speak to them about value investing, you know a lot 0:06:18.000 --> 0:06:20.080 of the feedback I get it as well. These are 0:06:20.160 --> 0:06:22.080 old companies. This is an old way to do it. 0:06:22.160 --> 0:06:24.440 You know, this is kind of like anti innovation, and 0:06:24.960 --> 0:06:27.000 you know, I don't have to tell everyone, you know 0:06:27.160 --> 0:06:29.320 that we're kind of living through a gold native innovation 0:06:29.360 --> 0:06:31.280 in many ways. You look at the cues, you look 0:06:31.320 --> 0:06:35.200 at r KK, et cetera. So you know, do you 0:06:35.360 --> 0:06:36.640 do you? 0:06:36.640 --> 0:06:38.680 But what really always helped me back. 0:06:38.600 --> 0:06:40.920 From those type of funds is like they're anti factor, right, 0:06:40.920 --> 0:06:43.760 they're very high vaulved, they're very expensive. So do you 0:06:43.920 --> 0:06:47.480 view you know, your your fund more like a value 0:06:47.560 --> 0:06:51.200 tilt or you know, an innovation tilt, but without those 0:06:51.240 --> 0:06:53.240 like bad factor waitings. 0:06:53.440 --> 0:06:55.960 Yeah, I think that's a very fair way of characterizing 0:06:55.960 --> 0:06:59.719 the strategy, right, It's it's in an innovation fund without 0:06:59.839 --> 0:07:02.000 the kind of baggage, where as an investor, you don't 0:07:02.000 --> 0:07:05.080 have to sacrifice your value exposure or your quality exposure 0:07:05.520 --> 0:07:06.280 by going into it. 0:07:07.200 --> 0:07:08.440 Yeah, so interesting. 0:07:08.600 --> 0:07:11.200 So would you consider this like, I mean, would you 0:07:11.200 --> 0:07:13.720 consider more like a growth fund or like a traditional 0:07:13.800 --> 0:07:16.720 value fund, or would you consider it completely different and 0:07:16.760 --> 0:07:17.480 separate in the stins? 0:07:17.560 --> 0:07:17.720 Yeah? 0:07:17.760 --> 0:07:19.600 Look, I mean I don't love the whole value versus 0:07:19.640 --> 0:07:21.560 growth economy. I don't think it's it's fair to say 0:07:21.600 --> 0:07:23.720 you have to be either one or the other. You know, 0:07:23.760 --> 0:07:25.600 Warren Buffett has talked about this as well, as you know, 0:07:25.640 --> 0:07:28.440 this is kind of being a false construction. The way 0:07:28.480 --> 0:07:31.160 I would think about it is a traditional value ETF. Right, 0:07:31.160 --> 0:07:33.240 what are they trying to do. They're trying to look 0:07:33.280 --> 0:07:36.400 for stocks with low price to book ratios. In other words, 0:07:36.520 --> 0:07:39.080 book value is a proxy for tangible capital. So they're 0:07:39.120 --> 0:07:41.360 going to look within the tangible economy, the old economy 0:07:41.360 --> 0:07:45.120 as you point out, industrials, banks, energy materials, and find 0:07:45.120 --> 0:07:47.600 the cheapest names, which is a totally valid thing to do. 0:07:47.880 --> 0:07:49.720 But you know, obviously, you know, as we move forward 0:07:49.720 --> 0:07:52.360 in time with innovation, AI, et cetera, this is becoming 0:07:52.360 --> 0:07:54.840 a vanishingly small part of the stock market. So what 0:07:54.880 --> 0:07:57.040 we're trying to do with the intangible value ETF is 0:07:57.080 --> 0:07:59.640 the same exact thing. We're looking for cheap stocks, but 0:07:59.680 --> 0:08:03.080 relive not to tangible but intangible capital, which ends up 0:08:03.120 --> 0:08:07.200 mapping to consumer brands to you know, tech platforms, you 0:08:07.200 --> 0:08:12.040 know life sciences companies, and you know services businesses. So 0:08:12.120 --> 0:08:14.200 it's kind of the same concept, but apply to the 0:08:14.200 --> 0:08:16.160 other half, so to speak, of the stock market. 0:08:22.960 --> 0:08:25.600 So is what you're saying part of the reason that 0:08:25.880 --> 0:08:29.000 traditional value investing just sort of gets punched in the 0:08:29.040 --> 0:08:33.200 face all the time and just lags for with fifteen 0:08:33.280 --> 0:08:34.840 years at this point. I had a nice little run 0:08:35.280 --> 0:08:38.199 in twenty twenty two, I believe, but now it's kind 0:08:38.200 --> 0:08:41.880 of back in the gutter. Is that why traditional value 0:08:42.400 --> 0:08:45.960 doesn't ever seem to have like a true regime takeover. 0:08:46.280 --> 0:08:48.800 And at the same time, every time you think something 0:08:48.840 --> 0:08:51.440 is coming back, like small caps or international, the CUES 0:08:51.600 --> 0:08:53.599 just wakes up and says, uh uh, I'm going to 0:08:53.679 --> 0:08:54.360 run over you. 0:08:54.280 --> 0:08:56.800 Again and again, run away and again. 0:08:56.800 --> 0:08:59.240 Like Marshall Lynch when we was talking about running people over, 0:08:59.280 --> 0:09:01.640 He's like, I'm to smash you in the mouth again 0:09:01.880 --> 0:09:04.720 and again and again and again and then you finally 0:09:04.760 --> 0:09:06.160 am gonna run over you and then like you just 0:09:06.200 --> 0:09:08.320 talk about how we scores touch downs anyway, piece mode. 0:09:08.679 --> 0:09:11.480 The Cues is in constant bast mode mode. But is 0:09:11.480 --> 0:09:15.560 that is intangible value? The reason that that phenomenon exists 0:09:15.600 --> 0:09:16.280 again and again. 0:09:16.640 --> 0:09:19.320 Yeah, We've actually done some analysis on both the CUES 0:09:19.320 --> 0:09:21.800 and on ARKK and what we did was we said, 0:09:21.840 --> 0:09:24.200 let's look at a factory based framework, right, think about 0:09:24.200 --> 0:09:27.040 the Fama French model, which is, you know, there's the market, 0:09:27.040 --> 0:09:29.040 there's a small cap there's value so on and so forth. 0:09:29.280 --> 0:09:31.280 And we added a sixth factor, which is the intangible 0:09:31.320 --> 0:09:33.600 value factor. And we looked at the holdings of both 0:09:33.600 --> 0:09:36.400 of these funds and then decompose the return, say, can 0:09:36.440 --> 0:09:40.360 we retrospectively explain its performance by allocating to the six 0:09:40.440 --> 0:09:44.240 factors and then idiosyncretic risk their alpha right, And what 0:09:44.280 --> 0:09:46.839 was quite interesting was both of these funds actually had 0:09:46.880 --> 0:09:49.559 a very positive loading on intangible value and in fact, 0:09:49.760 --> 0:09:51.960 a lot of their outperformance relative to Yes and people 0:09:51.960 --> 0:09:55.319 one hundred the traditional stock market has been due to this, 0:09:56.200 --> 0:09:59.280 you know, this this exposure to innovative companies. That being said, 0:09:59.320 --> 0:10:01.839 there's also a lot of volatility around that, as you 0:10:01.880 --> 0:10:04.960 point out, Chris, due to say, exposure to you know, 0:10:05.040 --> 0:10:07.160 cheap press to bookstocks, which you know did really well 0:10:07.160 --> 0:10:09.360 and then did really poorly, and you know kind of 0:10:09.840 --> 0:10:13.079 cycles in these really wide gyrations. And also, you know, 0:10:13.200 --> 0:10:16.240 especially in the case of the ar KAK, the exposures 0:10:16.640 --> 0:10:20.240 earlier stage unprofitable tech companies has been you know, kind 0:10:20.240 --> 0:10:23.400 of a negative contributor to their returns. Just given that 0:10:23.480 --> 0:10:25.280 quality as a factor has just done so well the 0:10:25.320 --> 0:10:26.160 past two decades. 0:10:26.600 --> 0:10:30.520 Curious where the idea for the for for it came 0:10:30.520 --> 0:10:32.920 from was did you have the idea for the ETF 0:10:33.080 --> 0:10:34.880 or did you see a company? And we're like, that 0:10:35.000 --> 0:10:37.640 is the poster child for intangible value. I'm going to 0:10:37.679 --> 0:10:38.600 build a product around it. 0:10:38.640 --> 0:10:40.640 Well kind of both, right, I mean you look at 0:10:40.640 --> 0:10:42.560 the stock market, you look at companies like you know, 0:10:42.679 --> 0:10:45.600 McDonald's or Coca Cola, you know, for whom brands are obviously. 0:10:45.320 --> 0:10:48.160 Critical, Apple, Google. 0:10:47.880 --> 0:10:49.439 Right, and it just kind of makes sense that these 0:10:49.440 --> 0:10:52.200 are the things that should matter today. And it's shocking that, 0:10:52.440 --> 0:10:54.760 you know, the quantitative metrics that we've used for many 0:10:54.800 --> 0:10:55.599 many years. 0:10:55.520 --> 0:10:57.760 Are have not really evolved to do that. 0:10:58.640 --> 0:11:00.760 You know, I used to work for GMO Jeremy Grantham, 0:11:00.800 --> 0:11:02.800 who was a pioneer in developing a lot of systematic 0:11:02.840 --> 0:11:06.040 value strategies in the seventies and eighties, and so I've 0:11:06.040 --> 0:11:08.840 always been thinking about this this problem. And you know, 0:11:09.040 --> 0:11:11.480 we're talking on an ETF podcast value ets or like 0:11:11.480 --> 0:11:14.800 a multie hundred billion dollar if not trillion dollar category, 0:11:14.840 --> 0:11:17.120 if you you know, expand that to also include active 0:11:17.120 --> 0:11:19.880 managers hollow value strategies. So this is a huge question 0:11:20.080 --> 0:11:22.880 and one which I feel like up until you know, 0:11:22.960 --> 0:11:25.599 now you know, just hasn't really been kind of satisfactory, 0:11:26.240 --> 0:11:29.000 literally like addressed. You know, we need more research, more 0:11:29.000 --> 0:11:31.640 and more work to understand the valuation of these names. 0:11:31.840 --> 0:11:33.680 And what problem did you have to solve in order 0:11:33.720 --> 0:11:34.920 to make this thing a reality? 0:11:35.160 --> 0:11:37.080 Well, this goes back to your question about like timing, 0:11:37.120 --> 0:11:40.400 like why now you know, the big problem is that 0:11:40.480 --> 0:11:45.840 accounting accounting statements don't really contain enough insight into intangible assets, 0:11:46.400 --> 0:11:48.360 and so you really need to go to unstructured data 0:11:48.440 --> 0:11:50.680 or alternative data. Right, We're lucky that we live in 0:11:50.679 --> 0:11:53.439 an air now. It's just been exponential growth in big data. 0:11:53.720 --> 0:11:57.960 We have everything from we use LinkedIn glassdoor, you know, 0:11:58.080 --> 0:12:02.240 job postings, patents, mars, all this information you know, obviously 0:12:02.280 --> 0:12:05.520 just by first principles contains insight into intangible value. The 0:12:05.600 --> 0:12:07.400 challenge being that, like the information is kind of locked 0:12:07.400 --> 0:12:09.640 in there because you can't, you know, as a quant 0:12:09.679 --> 0:12:11.840 just take a linear aggression running over it at twenty 0:12:11.840 --> 0:12:13.679 thousand more document and get anything meaningful out. 0:12:13.679 --> 0:12:14.440 It's all just noise. 0:12:14.760 --> 0:12:17.240 And so that's why the advent of the transformer natural 0:12:17.280 --> 0:12:19.600 language processing. You know, we were actually talking about this 0:12:19.640 --> 0:12:21.760 in twenty twenty. We've wrote a paper saying, you know, 0:12:21.800 --> 0:12:26.160 the killer app of AI within investing is then natural 0:12:26.200 --> 0:12:29.880 processing language and NLP, you know toolkit, which allows us 0:12:29.880 --> 0:12:32.960 to take unstructured data and kind of create structured factors 0:12:32.960 --> 0:12:36.400 which can then be used as inputs into traditional valuation models. 0:12:36.840 --> 0:12:38.200 You know what this reminds me of, Joe. I'm going 0:12:38.240 --> 0:12:40.120 to go full metaphor here. Dark matter. 0:12:40.920 --> 0:12:42.880 You know it's out there, you just can't see it, 0:12:42.920 --> 0:12:45.560 and it is. It kind of explains some most of 0:12:45.600 --> 0:12:48.720 the universes comprised of U. Yes, this is why the 0:12:48.800 --> 0:12:51.360 cues are the cues. It's this dark matter of intangible 0:12:51.400 --> 0:12:57.359 value because I'm looking at the holdings here. You know, Amazon, Meta, Google, Cisco, Intel, 0:12:57.880 --> 0:13:00.480 those are some of the firms driving the cues. Chris, 0:13:00.520 --> 0:13:03.559 you know in your world again this concept of dark matter, 0:13:03.640 --> 0:13:07.880 you have to correctly capture factors, track them. How do 0:13:07.920 --> 0:13:09.320 you work this in so? 0:13:09.480 --> 0:13:11.480 I you know, I read the white paper and a 0:13:11.559 --> 0:13:13.680 big fan. You know, I do view this as a 0:13:13.720 --> 0:13:16.320 different type of factor. You know, I don't think as 0:13:16.320 --> 0:13:18.000 you did with your six factor model. I don't think 0:13:18.000 --> 0:13:21.280 you throw out per se traditional value as you showed 0:13:21.280 --> 0:13:23.840 in the paper. You know the correlation between traditional value 0:13:24.320 --> 0:13:26.880 and tangible value is pretty low. If I remember, actually 0:13:26.920 --> 0:13:29.960 the correlation was higher to quality with intangible value. So 0:13:30.720 --> 0:13:33.200 you know, to me, that's a value add I think, 0:13:33.480 --> 0:13:37.200 you know, it's you know, the economy has changed. I 0:13:37.240 --> 0:13:39.240 mean no one would say not right. I mean, it's 0:13:39.280 --> 0:13:42.480 not plants anymore, it's not those tangible things. So this 0:13:42.559 --> 0:13:44.720 is very logical. I view it as, you know, a 0:13:44.760 --> 0:13:47.360 separate factor at least somewhat, and it can certainly add 0:13:47.440 --> 0:13:49.160 value to a multi factor process. 0:13:49.280 --> 0:13:53.040 Yeah, but why why not just forget traditional value? Like 0:13:53.360 --> 0:13:55.880 why even use old Price the Book? Why isn't the 0:13:55.960 --> 0:14:01.760 quant world much more adjusting things for this? Because it 0:14:01.800 --> 0:14:03.840 does explain so much, and it just seems like if 0:14:03.840 --> 0:14:07.000 you're doing value investing using Price the Book, it's like 0:14:07.080 --> 0:14:09.560 using like a rotary phone or something. I don't understand, Like, 0:14:09.559 --> 0:14:10.720 why isn't this a bigger deal? 0:14:12.120 --> 0:14:13.880 You know, that's a great, great question, and I ask 0:14:13.960 --> 0:14:17.320 myself that each day. But no, But look, we're all 0:14:17.440 --> 0:14:19.480 as researchers kind of building on the edifice of what's 0:14:19.720 --> 0:14:23.000 what's come before us, and you know, Fama French in 0:14:23.000 --> 0:14:26.080 the mid nineties and Germany beforehand, you know, popularize this 0:14:26.120 --> 0:14:28.520 idea of this book to market factor, which is important. 0:14:28.560 --> 0:14:30.560 It's not that it doesn't matter, right to take the 0:14:30.600 --> 0:14:33.640 converse to companies with a lot of IP, but one 0:14:33.680 --> 0:14:35.680 has a huge real estate portfolio and a huge cash 0:14:35.720 --> 0:14:36.880 hoard and the other doesn't. 0:14:36.920 --> 0:14:38.880 Of course, that company should be worth more than the 0:14:38.880 --> 0:14:41.040 other one. So you don't want to not use this. 0:14:41.200 --> 0:14:43.120 It's just that you know, we can maybe do better 0:14:43.560 --> 0:14:47.480 by adding additional dimensions of risk and dimensions of corporate 0:14:47.480 --> 0:14:50.160 performance to our kind of mulo of factors. 0:14:50.800 --> 0:14:54.160 When you think about this and what you've created is 0:14:54.600 --> 0:14:58.040 your model just the model, and it finds the companies 0:14:58.120 --> 0:15:01.240 and then you just you know, balance rebalance quarterly like 0:15:01.520 --> 0:15:04.360 a smart beta fund or are you are you putting 0:15:04.360 --> 0:15:06.120 a little bit of finger on the scale. 0:15:05.880 --> 0:15:06.840 No finger on the scale. 0:15:06.920 --> 0:15:09.160 So I mean my involvements only as a researcher kind 0:15:09.160 --> 0:15:11.400 of setting up the parameters the model, figuring out what 0:15:11.480 --> 0:15:13.400 data sets to look at, and how to build the 0:15:13.480 --> 0:15:16.640 machine learning uh, you know infrastructure, but you know it's 0:15:16.680 --> 0:15:19.520 it's all systematic, it's all data driven, right. Every day, 0:15:19.640 --> 0:15:22.320 you know, new information comes in about you know, employee turnover, 0:15:22.520 --> 0:15:26.360 about you know, cultures, corporate culture increasing, decreasing, you know, 0:15:26.520 --> 0:15:28.640 scandals in the media or all all the good stuff 0:15:28.680 --> 0:15:31.000 new patents, new trademarks, and that kind of feeds into 0:15:31.000 --> 0:15:34.960 the models and it automatically adjusts the relative rankings of stocks. 0:15:35.000 --> 0:15:37.080 And how big of a universe are you able to 0:15:37.200 --> 0:15:38.840 come through right now? And where do you want to 0:15:38.840 --> 0:15:39.080 get to? 0:15:39.320 --> 0:15:41.080 Well, we'll start with the where I want to get to. 0:15:41.640 --> 0:15:43.360 You know, I've actually just been working on a super 0:15:43.360 --> 0:15:47.920 interesting project expanding the universe of stocks to global so 0:15:48.040 --> 0:15:50.400 you know, effectively MSCI all country world. 0:15:50.640 --> 0:15:50.920 I am. 0:15:50.960 --> 0:15:54.440 I so like the nine thousand stocks or so right 0:15:54.480 --> 0:15:56.760 now when you know, in terms of launching products, the 0:15:56.840 --> 0:15:59.640 itn ETF is focused on the top one thousand largest 0:15:59.720 --> 0:16:02.240 us ST so used larger medcap stocks. But obviously that 0:16:02.280 --> 0:16:05.080 if it's not there's no kind of technological reason why 0:16:05.120 --> 0:16:06.720 that was the case. We just wanted to start with 0:16:06.720 --> 0:16:08.440 a product that you know, most people could kind of 0:16:08.440 --> 0:16:09.640 get their heads around. 0:16:09.920 --> 0:16:11.920 You know, one thing I wanted to ask you it It 0:16:11.920 --> 0:16:14.200 was more kind of like the methodology of intangible value. 0:16:14.240 --> 0:16:15.880 You know, you don't have to share secret sauce here 0:16:15.960 --> 0:16:16.560 or anything, but. 0:16:16.800 --> 0:16:18.800 You know, or feel free to or if you want to. 0:16:18.920 --> 0:16:20.760 Yeah, it's probably in the perspective. 0:16:22.280 --> 0:16:24.720 But I you know, you mentioned that you use alternative data. 0:16:24.800 --> 0:16:28.720 I'm guessing as higher frequency data NLP techniques to to 0:16:28.760 --> 0:16:31.440 put some you know, context around it. 0:16:31.560 --> 0:16:34.040 So do you need to use alternative data? 0:16:34.120 --> 0:16:38.000 Could you you know, substitute more traditional like balance sheet 0:16:38.080 --> 0:16:41.840 data or financial statement data for that? How far would 0:16:41.880 --> 0:16:43.520 you get if you did do that? Or is the 0:16:44.200 --> 0:16:46.680 is there really the value add the NLP and the 0:16:46.720 --> 0:16:47.520 alternative data. 0:16:47.600 --> 0:16:51.800 So we use both traditional accounting based information and alternative data, 0:16:51.880 --> 0:16:53.600 and we actually I can give you a very clear answer. 0:16:53.880 --> 0:16:56.200 So if you look at like the performance historically of 0:16:56.240 --> 0:16:59.520 say the MSCI, you know value index right relatively in 0:16:59.520 --> 0:17:01.360 the markets in pretty bad. You know, as you point 0:17:01.360 --> 0:17:04.280 out the past fifteen years. If instead you say, well, 0:17:04.359 --> 0:17:07.800 well let's now allow the capitalization of intangible investment so 0:17:07.960 --> 0:17:09.520 R and D. You know, as you kind of invest 0:17:09.600 --> 0:17:11.240 R and D, you build up a balance sheet asset 0:17:11.280 --> 0:17:13.400 for that and then you appreciate it over time. Same 0:17:13.480 --> 0:17:16.320 for sales and marketing expenditures. Well you get a line 0:17:16.320 --> 0:17:18.760 that's a little bit less bad, but still no panacea, right, 0:17:18.800 --> 0:17:21.560 it still goes down. And then when we said well 0:17:21.840 --> 0:17:24.080 let's start adding you know, more sources of data like 0:17:24.359 --> 0:17:27.040 I mentioned patents, I mentioned LinkedIn, you know, to measure 0:17:27.040 --> 0:17:29.360 each of the pillars using unstructured data. And that's when 0:17:29.359 --> 0:17:32.280 the line starts to look pretty interesting. Right And if 0:17:32.280 --> 0:17:33.639 you look at just the name, so put us out 0:17:33.640 --> 0:17:35.800 even the historical performance, because that's just a back test. 0:17:36.200 --> 0:17:38.160 Is the names you know, changed dramatically as you kind 0:17:38.160 --> 0:17:40.600 of continually iterate and add more and more data sources 0:17:40.840 --> 0:17:42.800 to a portfolio that just looks more like what it 0:17:42.840 --> 0:17:44.720 should look like. Right Like if you if I said, 0:17:44.720 --> 0:17:47.760 like first principles, build me a portfolio companies that are 0:17:47.920 --> 0:17:51.240 you know, attractively valued relative to prodigious and tangibles, right 0:17:51.280 --> 0:17:53.760 that that portfolio looks a lot more like the result 0:17:53.880 --> 0:17:57.280 of having added alternative data than just making this simple 0:17:57.320 --> 0:17:58.840 accounting based changes. 0:17:59.440 --> 0:18:02.560 It seems to me that you know, most people would 0:18:02.560 --> 0:18:04.560 hear this and go, I get it. It's kind of 0:18:04.560 --> 0:18:06.880 like tech stocks, right They they don't have a lot 0:18:06.920 --> 0:18:10.760 of machinery lying around, they're mostly intangible value. But there 0:18:10.760 --> 0:18:13.440 are some companies here that aren't tech. Right, So just 0:18:13.560 --> 0:18:15.440 let's just go over how are these are intangible value? 0:18:15.640 --> 0:18:20.480 Wells Fargo and General Electric those almost seem more traditional value. 0:18:21.000 --> 0:18:23.120