Redefining Value Investing in a Magnificent Seven Dominated World | Jacob Pozharny
What this covers
In this episode of Excess Returns, hosts Jack Forehand and Justin Carbonneau sit down with Jacob Pozharny, partner at Bridgeway Capital Management, to explore the increasingly important role of intangible assets in modern investing.
Jacob breaks down what intangible assets are - from intellectual property and proprietary algorithms to brand value and customer relationships - and explains how these harder-to-measure assets are changing traditional investment approaches. He discusses Bridgeway's pioneering research on "intangible intensity" and how it affects their investment strategy, particularly for high vs. low intangible companies.
Key topics covered:
How intangible assets complicate traditional valuation metrics Why sentiment analysis matters more for high-intangible companies The implications of AI for intangible asset valuation Bridgeway's approach to long-short investing International investing opportunities and market efficiency The importance of understanding model assumptions and staying humble as an investor
Whether you're interested in quantitative investing, understanding modern valuation frameworks, or keeping up with evolving market dynamics, this conversation offers valuable insights into how one of the industry's leading firms approaches these challenges.
00:00:00 Introduction and podcast overview 00:03:17 What are intangible assets? Basic definitions 00:06:04 Origins of intangible intensity research 00:14:06 Price to book as a risk factor rather than alpha factor 00:16:25 Different approaches for high vs low intangible companies 00:20:36 Impact of AI on intangible assets valuation 00:27:52 Overview of absolute return strategy approach 00:32:01 Examples of when systematic models need human oversight (meme stocks, M&A) 00:42:16 U.S. vs International markets perspective 00:47:25 Discussion of unique datasets like short availability data 00:51:12 Career insights and firm culture at Bridgeway 00:53:22 Final advice: The importance of staying humble as an investor 00:54:20 Closing remarks and contact information
#investing #quantitative #finance #valueinvesting #intangibleassets
SEE LATEST EPISODES https://www.excessreturnspod.com
FIND OUT MORE ABOUT VALIDEA https://www.validea.com
FIND OUT MORE ABOUT VALIDEA CAPITAL https://www.valideacapital.com
FOLLOW JACK Twitter: https://twitter.com/practicalquant LinkedIn: https://www.linkedin.com/in/jack-forehand-8015094
FOLLOW JUSTIN Twitter: https://twitter.com/jjcarbonneau LinkedIn: https://www.linkedin.com/in/jcarbonneau
Source description (no synthesized summary yet).
Jacob Pizor argues that classical valuation metrics have degraded as predictive tools for stock selection due to the rising importance of intangible assets, requiring differentiated investment approaches: fundamentals-based analysis for low-intangible companies and sentiment-based analysis for high-intangible companies.
- Intangible assets (IP, brands, human capital, R&D) now represent a much larger portion of company value than traditional balance sheets capture, doubling as a proportion of assets from 1994-2018
- Classical metrics like Price-to-Book and earnings-based ratios have lost explanatory power for high-intangible firms, functioning as risk factors rather than alpha signals
- High-intangible companies require sentiment and forward-looking analysis (analyst revisions, earnings forecasts, growth expectations) rather than backward-looking fundamentals
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Unlike pure quantitative strategies where backtesting uses out-of-sample testing and established historical testing methodologies, LLM-based systems know all historical data simultaneously, potentially negating traditional out-of-sample validation techniques that have been foundational to quantitative strategy development.
“in theory these things know everything historically you know we've been taught there's a certain way we test things you know we test things you know out of sample and and we do all this other stuff but I I wonder like do these things negate that because they know everything it's not like you can go back and create an llm and say only know what you would have known in 2014 or something like that or in 1972 or anything like that so it's an interesting balance like how do use these things and couple it with our the way we've tested things historically”
Capitalized intangible asset intensity (excluding goodwill) as a proportion of total assets doubled from 1994 to 2018; R&D expenses as a proportion of total revenue increased by 50% over the same period; while SG&A expenses as a proportion of total revenue remained quite stable.
“with the capitalized and tangible intensity um from 1994 to T to 2018 um that measurement as a proportion of assets actually um doubled um so it was a substantial increase obviously it is varies from country to Country industry to Industry talking about generalities R&D expenses um also um increased quite substantially they increased by 50% um on average from 1994 to 2018 as a proportion of total revenue as GNA expenses actually remain quite stable”
An absolute return strategy aims to build a return stream that is agnostic to overall market direction, often called market neutral, though true market neutrality is not predictive going forward—zero beta historically doesn't guarantee zero beta forward.
“what we've done uh with our absolute strategy absolute return strategy portfolios um is that we're we we try to build a return stream that is agnostic to Market Direction um a lot of folks call this a market neutral strategy I I I I'm very wary of calling strategies like these Market neutral because even when you uh uh build a strategy that has a zero beta um it is zero beta going backwards it's not zero beta going forwards beta isn't necessarily predictive going forwards so I I like to think of our return streams as agnostic to overall um Market Direction”
Intangible assets are company assets that are difficult to measure but very important in explaining the market value of a company; companies develop them to enhance operational efficiency, reduce costs, and drive revenue growth, but these investments decrease net income and negatively impact book value despite being designed to be additive to growth.
“intangible assets um are just like company assets but they're assets that are difficult to measure but very important in explaining the market value of a company so um companies develop intangible assets to enhance their operational efficiency to reduce costs to drive their revenue growth but these Investments actually decrease um the N income and they have a tendency to negatively impact Book value although they're designed to be additive to uh to growth”
Bridgeway is very careful evaluating model assumption failures on the short side of portfolios because shorts have asymmetric loss potential compared to longs, with potential losses that can be quite substantial.
“this is particularly important when you're running a long short strategy because the return asymmetry between Longs and shorts there it's it's it's quite substantial we're very careful um especially looking at the assumptions of the stock FES on the short side because the the the potential loss from shorts could be quite quite sizable um so we're we're constantly evaluating it um less concerned about model assumption failures on the long side than I am on the short side”
When trading, investors should assume that their counterparty has more information and greater skill than they do, and only execute trades that make substantial sense given that assumption.
“I I would say it's to stay humble um I I think that investors forget often that when they trade they trade with somebody on the other side who may have substant stally more information and substantially more skills than they do whenever I trade I assume that the counterparty has more information than I do um and is smarter than I am um so I want to make sure that the trade makes a lot of sense”
Jacob emphasizes that new hires to Bridgeway should focus intensively on understanding data quality, complexity, and nuances (especially international data) and on reading academic literature before attempting optimization or portfolio management, because foundational knowledge is critical to later success.
“it was very important for me to understand the complexities especially with international data um understanding how to process that understanding the financial literature um this is something that at the beginning of my career I spent a lot of time with and this is something that I encourage um folks when we bring them on to focus on as they develop their research and development techniques as they develop their optimization techniques”
Price-to-Book should be viewed as a risk factor rather than an alpha factor in portfolio construction, particularly for high-intangible companies, which has changed the way Bridgeway builds portfolios since this pattern became overwhelmingly true after 2007.
“the implication of the research that uh that that we've uh applied to our portfolios is we're actually seeing Price to Book as more of a risk factor in our portfolio construction uh rather than an alpha Factor”
Price-to-Book has shifted from being an alpha factor to functioning as a risk factor in portfolio construction, particularly after 2007, and this shift is consistent with the growth in intangibles observed in their research.
“the implication of the research that uh that that we've uh applied to our portfolios is we're actually seeing Price to Book as more of a risk factor in our portfolio construction uh rather than an alpha Factor so it's changed our thinking quite dramatically in terms of how we actually build out the portfolios”
For low-intangible industries, fundamentals (ROE, book value, earnings) continue to explain price action quite well and have stable explanatory power, but for high-intangible industries, fundamentals have much lower explanatory power in predicting future price action.
“we find that there the measurements of Roe the measurements of Book value the measurements of earnings are uh very unstable and what's most important um and we put out this page paper in financial analyst Journal I probably should um very important to mention that this is a collaboration um between myself amitab dugar and Andre Burkin um we put uh This research together in this publication and what we find is that um the explanatory power of fundamentals in terms of price action um hasn't really been affected for the low and tangible Industries so for the low and tangible Industries fundamentals continue to explain price action quite well but for the high and tangible Industries we find that fundamentals quality measurements valuation measurements are affected and they have a much lower explanatory power um in terms of explaining future price action”
The primary value of human judgment in systematic quantitative strategies is understanding model assumptions and questioning whether those assumptions continue to hold when new evidence or external events emerge.
“systematic process every systematic process has underlying assumptions the human needs to understand what those assumptions are and question um the stock selection preferences do the assumptions hold of the model you know what those assumptions are because you built the model um so it's absolutely critical in my opinion to apply quantitative strategy effectively to constantly keep in mind what the assumptions of your models are um and make sure you throw out the stock picks that fail um uh when assumptions fail”
A major challenge with using large language models (LLMs) like ChatGPT for historical analysis is look-ahead bias: an LLM trained in 2024 knows what happened in 2020-2022, making it impossible to trust what it tells you about what it would have done in 2020 without contamination from future knowledge.
“my biggest concern in applying that type of technology is that there's element of a look ahead bias so when you're looking at uh trying uh to assess what the your stimulation would have done in 2020 and you're using an llm that's fitted in 2024 that llm knows um what happened in in 2020 2021 2022 so um how do you really trust what it is that it's telling you”
An absolute return strategy differs from a market-neutral strategy in that market-neutral aims for zero beta backward-looking, but beta is not predictive forward-looking, so absolute return strategies are better characterized as agnostic to market direction.
“what we've done uh with our absolute strategy absolute return strategy portfolios um is that we're we try to build a return stream that is agnostic to Market Direction um a lot of folks call this a market neutral strategy I I I I'm very wary of calling strategies like these Market neutral because even when you uh uh build a strategy that has a zero beta um it is zero beta going backwards it's not zero beta going forwards beta isn't necessarily predictive going forwards so I I like to think of our return streams as agnostic to overall um Market Direction”
Classical measurements for valuation and quality have degraded significantly in their ability to be used for stock picking because intangible assets now comprise a much larger portion of company value that traditional accounting frameworks fail to capture.
“the classical um stock screens the classical measurements for valuation for Quality they've degraded um a lot in terms of um the ability to use them for stock picking and we got to thinking what's what's happening”
When running a long-short strategy, assumption failures and external risks are more critical on the short side than the long side because the potential loss asymmetry from shorts is substantially larger than from longs, requiring more careful monitoring of short-side model assumptions.
“this is particularly important when you're running a long short strategy because the return asymmetry between Longs and shorts there it's it's it's quite substantial we're very careful um especially looking at the assumptions of the stock FES on the short side because the the the potential loss from shorts could be quite quite sizable um so we're we're constantly evaluating it um less concerned about model assumption failures on the long side than I am on the short side”
Long-short strategies allow for more effective mitigation of systematic exposures compared to long-only strategies because they eliminate index underweight constraints and enable direct control of country exposures, sector exposures, size exposures, and price-to-book exposures to near-zero levels.
“long short strategies allow you to mitigate systematic exposures um much more effectively um with a long only strategy the underweights you con are limited by um your your your index underweights relative to the index obviously are limited by the index uh index value in a long short strategy you don't have that type of restriction so we're we're able to mitigate um country exposures by making sure that the net exposures across countries um is around zero”
In 2008-2009, a Canadian mutual fund sub-advised strategy holding US stocks benefited from dollar strength in 2008 but suffered from dollar weakness in 2009, demonstrating currency movement's substantial impact on international portfolio returns.
“during the financial crisis because the dollar was so strong and they were the Canadian Bank was holding US Stocks like our strategy actually did very well like relatively speaking because the dollar helped prop up yeah so basically you're right the the dollar like for those types of strategies was a huge boost in 2008 but it was a huge headwind in 2009 so effectively like the returns of that type of strategy you had significantly less losses in 2008 but you had significantly less gains in 2009 because the dollar went one way on you and then went the other way”
Government regulation, such as price caps on utility rates, represents an externality that violates fundamental assumptions about earnings predictability and requires questioning the underlying assumptions of investment models.
“there's government regulations um government regulations like the Brazilian government loves to cap uh the amount different utilities can charge obviously that has an effect on earnings um whenever there's government regulation we really question the underlying assumptions of our investment process”
Meme stocks are a key example of an externality where fundamentals are nearly irrelevant and assumptions about earnings-price relationships break down, requiring removal from the systematic selection process.
“sure it happens all the time I would say that uh more than half per half of the time of portfolio management time is actually devoted to studying externalities to the process meme stocks are an example of that for meme stocks fundamentals are almost irrelevant”
M&A activity creates significant uncertainty and unpredictability in earnings and cash flow forecasts for the acquiring and acquired companies, requiring Bridgeway to mute stock selection signals for M&A-involved securities.
“when ever there's m&a activity um the uh earnings of the company that is buying and being bought um it becomes totally unpredict it's not predicted what's going to happen with earnings what's going to happen to cash flow in in the following year or two so we tend to mute meme stocks we tend to mute anything that's going through m&a activity”
Capitalized intangible assets as a proportion of total assets doubled from 1994 to 2018, while R&D expenses increased 50% as a proportion of total revenue, but SG&A expenses remained relatively stable over the period.
“with with the capitalized and tangible intensity um from 1994 to T to 2018 um that measurement as a proportion of assets actually um doubled um so it was a substantial increase obviously it is varies from country to Country industry to Industry talking about generalities R&D expenses um also um increased quite substantially they increased by 50% um on average from 1994 to 2018 as a proportion of total revenue as GNA expenses actually remain quite stable um obviously variations existed from industry and from across countries but that that's an element that actually remained quite stable overall research period”
During the 2008-2009 financial crisis, hedging currency exposure provided significant protection in 2008 (strong dollar) but became a drag in 2009 (weak dollar), illustrating the cost of currency hedging and the temporal uncertainty of its benefits.
“during the financial crisis because the dollar was so strong and they were the Canadian Bank was holding US Stocks like our strategy actually did very well like relatively speaking because the dollar helped prop up yeah so basically you're right the the dollar like for those types of strategies was a huge boost in 2008 but it was a huge headwind in 2009 so effectively like the returns of that type of strategy you had significantly less losses in 2008 but you had significantly less gains in 2009 because the dollar went one way on you and then went the other way”
Long-short strategies allow portfolio managers to mitigate systematic exposures (country, sector, beta, size) much more effectively than long-only strategies because shorts are not constrained by index weights; managers can short out unwanted sector or country exposures rather than merely underweighting them.
“long short strategies allow you to mitigate systematic exposures um much more effectively um with a long only strategy the underweights you con are limited by um your your your index underweights relative to the index obviously are limited by the index uh index value in a long short strategy you don't have that type of restriction so we're we're able to mitigate um country exposures by making sure that the net exposures across countries um is around zero we don't like taking large sector exposures so we mitigate that type of systematic risk uh by making sure the sum of our Longs and sum of our shorts in each of the sectors is uh roughly equal”
Classical measurements for valuation and quality metrics have degraded significantly in their ability to be used for stock picking, particularly in high-intangible industries, starting around 2016-2017.
“the classical measurements for valuation for Quality they've degraded um a lot in terms of the ability to use them for stock picking”
Bridgeway uses machine learning with singular value decomposition (SVD) or similar techniques to identify latent systematic risks in higher-frequency price data across 35 countries, enabling detection of geopolitical risks and trade-related exposures that classical risk models miss.
“with our risk modeling what we also tend to do is identify systematic drivers so what we tend to do is we look at higher frequency um uh data elements higher frequency prices um and we're based on how Global prices move um across the 35 countries that we invest in on a daily basis we're able to pick up latent risks um using um we have our own version of this but the classical uh uh Quant technique is using singular value decomposition to identify um igen vectors that are drivers of overall systematic risks um we have our own flavor of this but what it allows us to do is identify latent risks in the process it's particularly useful in our long short strategies because we're able to pick up more fluid less known um risks so classical risk models aren't necessarily going to be able to pick up geopolitical risks or risks to stocks coming from um cars that that could be in place but as soon as there's for example activity um that's geopolitical in nature between Taiwan and China there's certain stocks that are tied to trade between China and and Taiwan that that lied up as red um in in in our own latent risk models”
When using sentiment-based analysis for high-intangible stocks, it is critical to identify which sentiment is already priced into the stock and which is not, because investment opportunities exist only in unprice sentiment.
“the key here in using sentiment-based analysis is to make sure that when you're using sentiment based analysis you are looking an information that isn't priced in and we have another set of processes to um understand what sentiment is priced in to the stock and what sentiment is not priced into the stock so naturally the the opportunities in sentiment that isn't priced in”
Stock selection opportunities are greater in less efficient markets: the US stock selection opportunity is less in mega cap names than in small caps, and internationally and in emerging markets the stock selection opportunity is even greater due to less market efficiency.
“the United States present a lot more stock selection opportunity um for example in the US I think most plants agree the stock selection opportunity um in small caps is far greater than it is in mega cap names information is just much more faster moving in in the larger names it doesn't give you an opportunity to really make money um based on your Alpha International that gets even more um Emerging Markets the stock selection opportunities is even greater”
Prime brokers have a massive daily short availability dataset showing what stocks are available to short and at what cost, but this dataset is heavily guarded and underleveraged in academic research despite having significant practical value for portfolio managers.
“Prime Brokers have a massive short availability data set it's it it's daily it tells you what the availability is across every stock um in in the investable Universe it tells you what the price of shorting a particular stock is I I've been really surprised by the lack of academic research in this area I've spent a lot of time studying this um I want to put something out but Prime Brokers uh are very guarded of their short availability data set so I I think that might explain why there um isn't as uh more Publications in in that space”
Short availability data does not follow classical patterns that quants would expect, requiring creative approaches and understanding the mindset of actual shorters to extract alpha from the dataset.
“the classical things that you would expect to look at in that data set they don't pan out so you have to get pretty creative in uh you have to get into the mindset of the shorter uh the person who does take on the shorts to um to to understand how to use that data set effectively”
Capitalized intangible assets excluding goodwill, innovation capital proxied by R&D expenses, and organizational capital proxied by SG&A expenses are the three measurable elements Bridgeway uses to quantify intangible capital intensity because they are available on balance sheets and income statements across 15 countries including developed and emerging markets.
“we focused on three items as I mentioned one is capitalized intangible asset um that excluding Goodwill and I can get into why we excluded Goodwill a bit later um the second is innovation Capital um which is proxied by R&D expenses um and the third is organizational capital and this is uh proxied by selling General and administrative expenses s GNA a expenses so these three items actually available in the balance sheet they're available in the income statement um and we were able to pull this for a fairly substantial substantial Universe of uh about 15 countries us developed economies as well as Emerging Markets”
Intangible asset investments have higher uncertainty and variability in outcomes compared to tangible investments, making it challenging to predict how any given investment will work out, which parallels the uncertainty in tangible capital investments.
“when when a company makes a tangible investment sometimes those workout exceptionally well sometimes those work out exceptionally poorly and it's the same thing in the intangible world right like I mean we don't know it's hard you know just to just because somebody makes an investment you don't know how it's going to work out so it can be challenging”
Human beings can be better than computers in systematic quantitative investing by understanding and questioning the assumptions underlying their models, and by recognizing when those assumptions fail in real-world conditions.
“systematic process every systematic process has underlying assumptions the human needs to understand what those assumptions are and question um the stock selection preferences do the assumptions hold of the model you know what those assumptions are because you built the model um so it's absolutely critical in my opinion to apply quantitative strategy effectively to constantly keep in mind what the assumptions of your models are um and make sure you throw out the stock picks that fail um uh when assumptions fail”
Tangible investments sometimes work out exceptionally well and sometimes work out exceptionally poorly, and the same is true of intangible investments—the uncertainty in outcomes is inherent to both types of investments.
“just because somebody makes an investment you don't know how it's going to work out so it can be challenging”
Intangible intensity has grown more for international and emerging market companies compared to the US, despite the US having a higher starting point, which was a surprising finding.
“we actually found um uh intangible intensity to grow more for um International companies compared to the US I was a bit surprised by this result but I think it's all about um where you start right I mean if you're starting at a higher level you would expect a bit less growth if you're starting from a lower level you would expect a bit higher growth um but that was a bit of a surprising result for us was that”
Bridgeway does not hedge out currency exposures in their long-only international strategies because hedging is expensive and if investors want to diversify, they should diversify into both currencies and international equities.
“because we are typically uh more or less country neutral um we don't take on too much currency um exposure in our long only strategies my preference is not to hedge out currencies hedging out currencies are very expensive things to do um and I think that if an investor really wants to diversify they want to diversify into currencies and into International equities so my preference is not to hedge currencies but that's just my opinion”
The US represents approximately 60% of global market value across developed and emerging markets, but produces only about one-third of total revenue in MSCI ACWI IMI, raising the question of how much further outperformance the US can deliver relative to international markets.
“the US is about 60% of the market value globally um across developed and emerging economies um and the US produces about a third of total revenue of companies in msci aqu IMI for example so um I mean my question is how much more growth how much more outperformance can you expect from from the US uh relative to International with that type of uh Revenue market value disparity”
Quants must balance using AI/ML tools to improve processes with maintaining reliance on empirically validated factors that have explanatory foundations, rather than going 'crazy' with AI by developing novel strategies without theoretical grounding.
“we're quants too it's an interesting time to be a Quant right now because we have all these tools available to us but by the same token we want to rely on what we know is worked over time like if if we believe that Factor should have an explanation for it to work then and some people believe that now and some people don't but if we believe that then we want to start with that as a core and we don't want us going crazy with AI developing things but by the same token it's also a technology that's going to make us better at what we do so there's an interesting balance there”
AI expenses in technology firms reduce net income and distort PE ratios, making these companies appear more expensive on earnings-based metrics despite the fact that those expenses are creating valuable intangible assets.
“I think that a lot of uh the value of intangibles especially for um the the technology firms explain a lot of the difference between market value and Book value um you rece huge appreciations of market value they look much more expensive based on earnings but these AI expenses are actually in negative to net income naturally that distorts the the PE ratios of these companies”
When model assumptions fail or externalities occur, Bridgeway mutes stock selection signals by downgrading stocks: for example, meme stocks, M&A activity, and government regulation all cause them to downgrade selections because fundamental analysis becomes unreliable in these contexts.
“it happens all the time I would say that uh more than half per half of the time of portfolio management time is actually devoted to studying externalities to the process meme stocks are an example of that for meme stocks fundamentals are almost irrelevant when ever there's m&a activity um the uh earnings of the company that is buying and being bought um it becomes totally unpredict it's not predicted what's going to happen with earnings what's going to happen to cash flow in in the following year or two so we tend to mute meme stocks we tend to mute anything that's going through m&a activity there's government regulations um government regulations like the Brazilian government loves to cap uh the amount different utilities can charge obviously that has an effect on earnings um whenever there's government regulation we really question the underlying assumptions of our investment process”
Intangible intensity has been increasing across most knowledge-based industries, not just in technology, including biotech, pharmaceuticals, financials, software, telecom, and semiconductors.
“how much of this is a technology thing and how much of this is intangible intensity has actually been increasing in other Industries where people might not think it has been um we're seeing this consistently increasing across most of the knowledge-based industry so it's not just technology um it's uh you have software you have biotech um there is a substantial increases um in in terms of financials um the technology the algorithms that are being used were quite influential so it's not just in the technology space it's across different Industries we actually put together a a a table of different Industries um ranked by intangible Capital intensity from highest to lowest so uh Pharma software Telecom uh were some of the some of the highest um semiconductors obviously another one”
The research on intangible capital intensity was conducted as a collaboration between Jacob Pizar, Amitab Dugar, and Andre Burkin and was published in the Financial Analyst Journal.
“we put out this page paper in financial analyst Journal I probably should um very important to mention that this is a collaboration um between myself amitab dugar and Andre Burkin um we put uh This research together in this publication”
A solution to the look-ahead bias problem with LLMs is to develop a specific accounting taxonomy and apply NLP techniques carefully to maintain temporal integrity in backtesting.
“I think there needs to be a a specific accounting taxonomy and applying NLP techniques and this is something we've been thinking about a lot and I'm sure a lot of quants have been”
Intangible intensity has increased more substantially in international companies compared to US companies, which was a surprising finding even after adjusting for industry composition differences.
“we actually found um uh intangible intensity to grow more for um International companies compared to the US I was a bit surprised by this result”
Prime brokers have a massive daily short availability dataset that reveals what shorts are available at what prices across the entire investable universe, but this dataset has received surprisingly little academic research attention, likely because prime brokers guard it closely, despite its significant value for long-short stock selection.
“Prime Brokers have a massive short availability data set it's it it's daily it tells you what the availability is across every stock um in in in the investable Universe it tells you what the price of shorting a particular stock is I I've been really surprised by the lack of academic research in this area I've spent a lot of time studying this um I want to put something out but Prime Brokers uh are very guarded of their short availability data set so I I think that might explain why there um isn't as uh more Publications in in that space”
Bridgeway uses machine learning and singular value decomposition on daily price data across 35 countries to identify latent systematic risks (such as geopolitical tensions affecting Taiwan-China trade stocks or tariff-related trade exposure following Trump's election) that classical risk models cannot detect.
“we look at higher frequency um uh data elements higher frequency prices um and we're based on how Global prices move um across the 35 countries that we invest in on a daily basis we're able to pick up latent risks um using um we have our own version of this but the classical uh uh Quant technique is using singular value decomposition to identify um igen vectors that are drivers of overall systematic risks”
In industrials, AI optimizes manufacturing and supply chains through patents, algorithms, and training programs that complicate profitability metrics; in media, AI enhances content personalization and ad placement; in software, AI drives heavy R&D investment that distorts short-term profits while pursuing long-term growth; in finance, AI improves fraud detection and risk management through proprietary algorithms.
“in in Industrials for example AI optimizes manufacturing Supply chains but the the these intangibles like patents um in in in different types of algorithms training programs it complicate profitability metrics because it's very hard to assess the value of of these AI related Assets in the media space AI enh enhances um enhances uh content personalization it enhances ad placement ability but again how do you value um the firm's ability to to Value these AI related expenses AI related assets it's it's it's really tough to do in in the software space there's heavy R&D Investments that are AI driven um it distorts short-term profits with the objective of growing um long-term growth um so it it we definitely see it as particularly important um to look at right now in the finance space um AI improves fraud detect risk management but these proprietary algorithms are very difficult to Value”
Jacob's primary investment recommendation for average investors is to stay humble about trading, assuming the counterparty has more information and skills, and ensuring every trade makes strong sense; for quants specifically, it is critical to always question the assumptions of models rather than blindly following stock screens.
“I I would say it's to stay humble um I I think that investors forget often that when they trade they trade with somebody on the other side who may have substant stally more information and substantially more skills than they do whenever I trade I assume that the counterparty has more information than I do um and is smarter than I am um so I want to make sure that the trade makes a lot of sense”
There is no such thing as a purely quantitative strategy because ultimately a human being created the strategy and decides whether to change it, making human decision-making and emotion impossible to completely remove from the process.
“there's no such thing as a purely quantitative strategy because at the end of the day there's a who created that strategy there's a person who's deciding whether to change that strategy like ultimately you can't completely get like human decision- making and human emotion out of this process”
Examples of intangible assets include intellectual property (Google's PageRank algorithm), proprietary algorithms (TikTok's recommendation engine), brand equity (Coca-Cola, Nike, Disney), and customer relationships (Amazon, Costco), all of which are material to business value but very difficult to value precisely.
“examples of um intangibles would be intellectual property um um proprietary algorithms like Google's page rank algorithm is an example of a of an intangible asset it's something that's very useful very productive but it's hard to Value Tik Tock uh for you recommendation engine is a proprietary algorithm very hard to Value very important to the business brand Equity is another example of an intangible asset think Coca-Cola Nike Disney customer relationships are another element of intangible assets um think of the relationship between the client um and Amazon The Client and Costco”
Natural language processing and textual analysis of quarterly earnings call Q&A sections can provide valuable information about company AI investments and strategies, and this represents a major research focus for Bridgeway in 2024-2025, though they must avoid lookahead bias with LLMs trained on future data.
“NLP techniques uh natural language processing textual processing um the Q&A that uh you you hear on um quarterly calls I think it does give you a lot of information of uh what companies are doing in terms of AI that can be used um in terms of stock selection again this is an area of research this is a actually a big Focus for us in 2025 and a bit in 2024”
Approximately 85% of Bridgeway's portfolio management process is systematic while 15% is discretionary, with the discretionary portion focused on identifying assumption failures and externalities (like meme stocks, M&A activity, or government regulations) that require muting the systematic signals.
“I would say 85% of our process to be um somewhat systematic 15% is discretionary so we constantly are looking at the assumptions um that our models have and whenever there's an assumption failure whenever there's an externality to the process we tend to um uh mute our stock selection preferences”
Sentiment-based analysis requires understanding what sentiment is already priced into a stock and what sentiment is not priced in, with opportunities existing only in unprice sentiments.
“the key here in using sentiment-based analysis is to make sure that when you're using sentiment based analysis you are looking an information that isn't priced in and we have another set of processes to um understand what sentiment is priced in to the stock and what sentiment is not priced into the stock so naturally the the opportunities in sentiment that isn't priced in”
Standard reported short interest data has funky distributions that make it difficult to extract alpha, whereas prime broker daily short availability data has superior distribution characteristics that make alpha extraction more feasible.
“it's the the data set is distributed in a fairly funky way um so whenever you have funky distributions in data sets it's hard to create um Alpha out of them what I'm talking about is getting to the prime broker getting them to share that short availability data um the hedge funds out there that have access to that have an edge over the long only investors who don't have access to that data set um and that's uh something that I would encourage getting and getting creative with”
Jacob would not typically re-evaluate an entire strategy more than once every couple of years, which would be very unusual, preferring to focus team effort on understanding when specific model assumptions fail.
“I I don't like to change strategies very often um it would be unusual for me to re-evaluate uh an entire strategy more than once every couple of years that that would be very unusual”
Bridgeway uses completely different gearing (weighting/emphasis) for stock selection between high-intangible and low-intangible stocks, applied across global strategies in developed markets and emerging markets, which differs from competitors who primarily make adjustments to different measurements of intangibles.
“they make adjustments um to different measurements of intangibles and we certainly um do make these types of adjustments on an industry by industry basis but I think what we do a bit differently is we use completely different gearing um for our stock selection for high intangible and low intangible stocks um and this is applied across um our Global strategies across develop markets and the cross Emerging Markets”
The US represents approximately 60% of global market value across developed and emerging economies but produces only about one-third of total revenue in MSCI ACWI IMI, raising the question of how much additional growth and outperformance should be expected from the US relative to international markets given this valuation disparity.
“the US is about 60% of the market value globally um across developed and emerging economies um and the US produces about a third of total revenue of companies in msci aqu IMI for example so um I mean my question is how much more growth how much more outperformance can you expect from from the US uh relative to International with that type of uh Revenue market value disparity”
Portfolio managers at Bridgeway are hired for the ability to do their own data analytics, read academic literature, and apply research directly in portfolio construction rather than being siloed into separate research or implementation roles.
“my preference in hiring is not to pigeon hole people into uh different types of uh operational tasks or research tasks my preference in hiring is to have uh portfolio managers have the capability of doing um data analysis Financial research as well as implementing the resch and being accountable for the research that they Implement in portfolios”
Portfolios combining low-intangible and high-intangible stock selections become fairly balanced in aggregate because the two categories respond to different value drivers that work with and against each other, creating diversification of alpha sources.
“when we combine um our stock selection to low and tangibles and high in tangibles the portfolio in aggregate becomes fairly balanced but we're across all Industries they different metrics of valuations and quality that we do look at”
High-intangible companies requiring sentiment-based analysis produce portfolios with higher turnover than low-intangible companies, because sentiment is a faster-moving and more dynamic information source than historical fundamentals.
“we actually noticed that because of that the the portfolios are much more Nimble they have higher turnover expectations for the high and tangible Industries compared to the low and tangible Industries because sment is something that is much faster moving information is much more Dynamic”
Intangible capital intensity is measured using three components: capitalized intangible assets excluding Goodwill (measured relative to total assets), R&D expenses (measured relative to total revenue), and Selling, General, and Administrative expenses (measured relative to total revenue), which are aggregated into a ranking system to classify industries by intangible intensity.
“we focused on three items as I mentioned one is capitalized intangible asset um that excluding Goodwill and I can get into why we excluded Goodwill a bit later um the second is innovation Capital um which is proxied by R&D expenses um and the third is organizational capital and this is uh proxied by selling General and administrative expenses s GNA a expenses”
Magnitude 7 (the largest technology companies) are spending outrageous amounts of money on AI with a very wide range of potential intangible asset values being created, making it very difficult to value these intangible assets, particularly with a new technology not previously seen.
“like the mag 7 firms are spending outrageous amounts of money right now on AI and it's like you think about how do I value that in terms of the intangible asset they're Crea creating that seems like that's a very there could be a very wide range there and that's a very difficult thing to figure out particularly with a new technology that we are really seeing for the first time”
Bridgeway uses completely different gearing and stock selection metrics for high-intangible versus low-intangible stocks across global and emerging market strategies, rather than using the same metrics with industry adjustments.
“what we do a bit differently is we use completely different gearing um for our stock selection for high intangible and low intangible stocks um and this is applied across um our Global strategies across develop markets and the cross Emerging Markets”
Currency hedging is expensive and investors wanting to diversify should diversify into both currencies and international equities together, not eliminate currency exposure through hedging.
“hedging out currencies are very expensive things to do um and I think that if an investor really wants to diversify they want to diversify into currencies and into International equities so my preference is not to hedge currencies but that's just my opinion”
Bridgeway's absolute return strategy targets about 10% annualized volatility over a market cycle, maintains equal long and short exposure with 200% gross exposure (100% long, 100% short), and scales gross exposure proportionally to stock selection opportunity across 35 countries, 11 sectors, and 250-350 long and short positions.
“we're targeting um about a 10% annualized volatility um over our Market cycle and that seems to be a marketable strategy um what we typically do in our strategies is we invest in a very Diversified uh way we're invested in 35 different countries 11 different sectors typically um all sectors so 200 50 to 300 Longs um versus about 300 to 350 shorts”
Bridgeway's intangible capital intensity research drew from data spanning 1994 to 2018 across approximately 15 countries including the US, developed economies, and emerging markets, with industry rankings created on a month-end basis using average values and rankings that appeared very consistent over time.
“we were able to pull this for a fairly substantial substantial Universe of uh about 15 countries us developed economies as well as Emerging Markets um so we created a measurement based on this we looked at the um average value of companies and industries of each of these measurements we rank these um on a on a month-end basis and then we looked at the average ranking and that naturally tended to uh differentiate uh certain industries and that differentiation was very consistent over time I think we did this research from 1994 to 2018 in terms of data”
Pharma, software, telecom, and semiconductors are among the highest intangible-intensity industries, but high intangible intensity spans across various knowledge-based and new-economy industries, not just technology.
“so uh Pharma software Telecom uh were some of the some of the highest um semiconductors obviously another one but it's it's not just the technology it's anything that's new economy that's knowledge based”
Junior people at Bridgeway focus heavily on understanding data complexities, particularly with international data, and reading financial literature, while senior people focus more on idea generation and literature review and ensuring weekly discussions are informative.
“for the more Junior people there's a lot more handson work with data um and for the more senior people there's more of idea generation and literature review um and uh just uh making sure that the discussions that we have on a weekly basis are informative and productive”
At the portfolio construction level, latent systematic risks identified through ML are managed by attempting to mitigate correlated risk exposures, preventing concentrated bets in emerging risk clusters.
“so we are identifying these types of systematic risks we're trying to understand those systematic risks and at the portfolio Construction level we're trying to mitigate these latent systematic risks using a risk model so it's particularly valuable to do that especially when you're dealing with the long short strategies”
Portfolio managers at Bridgeway are encouraged to have capability in data analysis, financial research, and implementation, with junior staff focusing more on data understanding and senior staff focusing more on idea generation and literature review.
“my preference in hiring is not to pigeon hole people into uh different types of uh operational tasks or research tasks my preference in hiring is to have uh portfolio managers have the capability of doing um data analysis Financial research as well as implementing the resch and being accountable for the research that they Implement in portfolios”
Because sentiment changes faster than fundamental changes, high-intangible stock portfolios have higher expected turnover compared to low-intangible portfolios.
“that's why we spend um a lot more time focused on uh sentiment analysis for the high and tangible stocks we actually noticed that because of that the the portfolios are much more Nimble they have higher turnover expectations for the high and tangible Industries compared to the low and tangible Industries because sment is something that is much faster moving information is much more Dynamic”
Jacob does not favor frequent or drastic strategy re-evaluation; it would be highly unusual to re-evaluate an entire strategy more than once every couple of years; the focus is instead on identifying when model assumptions fail and adjusting positions accordingly, not wholesale strategy changes.
“I I don't like to change strategies very often um it would be unusual for me to re-evaluate uh an entire strategy more than once every couple of years that that would be very unusual where um the team is really focused on is making sure they understand when the uh model assumptions fail”
Bridgeway is not currently using machine learning or AI for stock selection, but is using machine learning for risk modeling, with natural language processing on quarterly earnings calls as an active area of research for 2024-2025.
“in our current process um we're not using machine learning or AI for stock selection we are using machine learning for risk modeling um so we can talk about that a bit later we're not ready to use Ai and machine learning for actual stock selection this is an area of research for us um NLP techniques uh natural language processing textual processing um the Q&A that uh you you hear on um quarterly calls I think it does give you a lot of information of uh what companies are doing in terms of AI that can be used um in terms of stock selection again this is an area of research this is a actually a big Focus for us in 2025 and a bit in 2024”
Bridgeway's absolute return strategy maintains equal amounts of longs and shorts, typically net-zero at rebalancing with 100% longs and 100% shorts (gross 200%), and gross exposure is proportional to stock selection efficacy, varying by geography and market cap.
“our strategies have equal amounts of Longs and shorts um they typically at the time where re balance are Net Zero gross 200% so 100 Longs by 100 shorts I think where we might be different from other folks is that our gross exposure is proportional to our stock selection efficacy um so we're finding a lot more opportunities in small caps and midcaps we're finding a lot of opportunity uh in terms of stock selection efficacy outside of the US and emerging economies so our gross exposure is proportional to that opportunity so in our um uh Global strategy we can have as little as 15% of our gross exposure in the US it varies based on opportunity”
For long-only strategies, emerging market small caps represent the highest alpha opportunity due to market inefficiency, which is why Bridgeway focuses long-only international strategy on that segment rather than diversifying across all market caps.
“for me um and it's really an expression of our long short strategy for me what's interesting is the opportunity for investing in markets that are less efficient the United States present a lot more stock selection opportunity um for example in the US I think most plants agree the stock selection opportunity um in small caps is far greater than it is in mega cap names information is just much more faster moving in in the larger names it doesn't give you an opportunity to really make money um based on your Alpha International that gets even more um Emerging Markets the stock selection opportunities is even greater um so I had a choice of what type of strategy I wanted to run we do have a long only International strategy and our choice my choice was to actually focus on Emerging Market small caps why is that for a long only strategy because I see the greatest Alpha opportunity to exist in the least efficient markets”
Baruch Lev's book 'End of Accounting' was published and argued that traditional accounting was becoming obsolete for valuation purposes due to the rise of intangible assets.
“I came across back then a book by baru flb titled end of accounting”
The book 'End of Accounting' by Baru Flb was very influential in Jacob's thinking and motivated their research on intangible capital intensity, which expands on chapters Flb presented in that work.
“I came across back then a book by baru flb titled end of accounting that was very influential in in in my thinking um our research on intangible Capital intensity expands on a couple of the chapters that um he presented um in in in that book”
Jack Forehand is a principal at Validia Capital Management and Justin Carbono is a managing director at Life and Liberty Indexes; they are co-hosts of the Excess Returns podcast.
“Jack forehand is a principal at Validia Capital Management Justin Carbono is a managing director at life and Liberty indexes”
The challenge of integrating LLMs into quantitative strategy development will be one of the most interesting and important problems for quants over the next decade.
“I think yes I think it's going to be one of the most interesting challenges for Quant in the next decade um we're we're it's a very exciting area research”