Gerard O’Reilly serves as Co-Chief Executive Officer and Chief Investment Officer of Dimensional Fund Advisors LP and is a Dimensional Director. He was named Co-CEO in 2017 and has been with Dimensional since 2004. Along with Co-CEO Dave Butler, Gerard guides the firm’s vision and strategy through daily oversight of the company’s people and processes.
Since joining the firm, Gerard has been instrumental in the evolution of Dimensional’s systematic approach to creating and implementing investment solutions. He was formerly Head of Research, managing the firm’s rigorous, scientific approach to interpreting, testing, and applying research in portfolios. He is a member of the firm’s Investment Research Committee, which reviews potential investment strategy enhancements, and a member of the Investment Committee, which oversees portfolio and account management. He also serves on the Boards of Dimensional’s US mutual funds and exchange-traded funds.
Prior to joining Dimensional, Gerard obtained his PhD in aeronautics from the California Institute of Technology. He holds a master of science degree in high performance computing from Trinity College Dublin in his native country of Ireland, where he also received a first class honors degree in theoretical physics and was a recipient of the Foundation Scholarship.
You don’t need to be a rocket scientist to work at Dimensional Fund Advisors, but Gerard O’Reilly sees it as an asset, particularly when it comes to problem-solving. Now the Co-CEO and Chief Investment Officer of one of the fastest-growing US investment businesses, Gerard received a Ph.D. in aeronautics before entering fund management, attracted to Dimensional because of the opportunity it afforded him to learn from the world-leading economists at the company; including Eugene Fama, Myron Scholes, Merton Miller, Robert Merton, and Ken French. We recently sat down with Gerard to discuss the firm’s research-based culture and rules-based approach to investing. In this episode, we get into the nitty-gritty regarding Dimensional’s distinctive portfolio management decisions and the data sources they draw from and Gerard answers some technical questions regarding risk assessment, factor tilted portfolios, operating profitability, goodwill, and more. We also touch on the value of combining multiple metrics, why small-cap stocks deserve a place in your portfolio, and some of the biggest changes that Gerard has witnessed in Dimensional portfolios over the past decade, as well as how he applies his scientific learnings to make unique portfolio adjustments and some of the various benefits of Dimensional’s integrated approach. Make sure not to miss this informative, insightful, and in-depth conversation with Dimensional CIO and Co-CEO, Gerard O’Reilly!
Key Points From This Episode:
Market-cap-weighted passive strategies versus Dimensional’s rules-based higher expected return strategy. [0:03:45]
Assessing risk based on the Intertemporal Capital Asset Pricing Model (ICAPM). [0:07:07]
Diversification in a factor tilted portfolio versus a cap-weighted market portfolio. [0:10:57]
What criteria the variables that Dimensional uses need to meet before they’re considered dimensions of expected returns. [0:13:00]
Sources Dimensional draws from regarding portfolio decisions and implementation. [0:16:09]
How Gerard decides between underweighting or excluding securities in portfolios. [0:19:59]
Why Dimensional uses operating profitability rather than cash-based profitability. [0:22:38]
Gerard’s view on intangible assets, goodwill, and Dimensional’s investment strategy. [0:29:31]
The value of including internally developed intangibles in value and profitability metrics. [0:37:49]
Gerard reflects on the opinion that Fama and French’s findings are no longer valid. [0:42:58]
Whether or not it’s better to combine multiple metrics to measure relative price. [0:46:41]
How Dimensional targets value and profitability together (for large and small caps). [0:50:39]
How Gerard thinks about capacity for investment strategies in small and micro-cap stocks as Dimensional continues to grow. [0:54:03]
Understanding how entering into the ETF market has impacted his thinking. [0:57:24]
Expected premiums for owning smaller stocks over larger ones. [0:58:50]
The importance of security lending revenue for expected returns on Dimensional funds; improving the investor experience. [1:00:12]
How Dimensional deals with sector weights and the role that diversification plays. [1:04:12]
Why they decided to implement credit, despite research to suggest that it doesn’t add an independent source of expected returns. [1:06:08]
Some of the biggest changes in Dimensional portfolios over the past 10 years. [1:10:23]
How Gerard applies his scientific learnings to make unique portfolio adjustments. [1:12:52]
Comparing Dimensional’s core and vector strategies with a combined cap-weighted portfolio; from fees to the benefits of hindsight and more. [1:15:15]
Papers that seemed compelling but were deemed ineffective by their research team. [1:19:11]
Insight into Dimensional’s decision to make their internal research public. [1:21:42]
Why their rules-based approach is the hardest part of Dimensional to replicate. [1:25:55]
What to be aware of when comparing backtests: how data can be manipulated. [1:30:06]
Valuable lessons and perspectives Gerard has learned from their competitors. [1:32:22]
Commonalities between aeronautics and asset management, like problem-solving. [1:35:20]
Why Gerard believes that having his own financial advisor is invaluable. [1:36:55]
Gerard explains why we might expect factor premiums to persist in equilibrium. [1:38:28]
Read the Transcript:
Right off the top, most investors, I think it's safe to say, are pretty familiar with market-cap-weighted passive strategies. How do you articulate the difference between that and what Dimensional's products represent?
Yeah. I think that you're right. Most investors are familiar with the index approach. I'll take you back about 40 years, 1981, when Dimensional first began. David's idea was could you have a rules-based, broadly diversified, small cap strategy that would appeal to institutional investors. At that time, there were no small cap indexes on which to base the strategy. So he basically started with a blank sheet of paper. When you think about a small cap index, especially in the '80s, and you were going to be as rigid as an index, the consensus was you would get killed by trading costs, and you wouldn't be able to make money because the trading costs would be so high. When you start with a blank sheet of paper, really, you're not going to end with an index approach. You're going to end with something that is looking every day, implementing the rules on a daily basis, has some flexibility, is going to not have to buy every stock every day, or weight every stock perfectly relative to its weight in the market, but get that broad exposure. That's what started it all. Then you go through time. Academic research came out to identify areas of the market that offer higher returns, like small caps, like value, like profitability, momentum, so and so forth. So if you fast forward 40 years from that kind of beginning, some of those essence from the beginning are still in play today. The way I describe what we do is, first off, we start with what is it that the clients are looking for. Then how do we deliver rules-based, higher-expected-return strategies to meet those needs?
A few things that are important there. One is rules based. We work with mainly financial professionals, so intermediaries. We don't work with the end investor. We work with intermediaries. A rules-based approach is very good to work with intermediaries largely because we can communicate, "Here's what to expect," and then you can monitor, you got what you thought you were getting. So we think that's a good approach, but it has to have the right support, the right innovation, the right pricing. We can get into all that in the webinar.
The other two things, though, that I think are important to what we do and articulate our approach is, one, market prices are predictions of the future. You got to know how to use that information to manage risk and increase expected returns, but that leads you away from having a broad base market index. Number two, optionality has value. You got to be able to capture it for your clients. That also leads you away wanting a rigid index-based approach. Sensible ideas well implemented is how we often describe it. It has a lot of the benefits of what you mentioned, these index-based approaches, but it doesn't suffer from the drawbacks. The benefits are transparency, low cost. You know what you're getting. The drawbacks are lack of flexibility, rigidity. You're not reacting to the latest research. You're not building your rules for every type of market environment. Those types of drawbacks, we've left behind. So bring the benefits, but leave the drawbacks behind. Then you start to deviate from those market cap weightings to improve returns and manage risk, and I think that you end up with a good solution for investors.
I'm going to follow that with a pretty theoretical question, but it's a theoretical question that most of our listeners should be able to understand the basis of, because we've covered this a lot in recent episodes. Risk factors, as of course you know, Gerard, are based on the Intertemporal Capital Asset Pricing Model. That model suggests that the factors exist because lots of investors are unwilling or unable to take on the state variable sensitivity that they represent. How do you think investors should assess whether they are suited to tilt toward those risks?
Yeah. That's a big question and a technical one. Let me back up a little bit on two views on models. There are some models where the model is the hero, and there are some models where the data are the hero. I would call the Intertemporal Capital Asset Pricing Model where the model is the hero. You start off with assumptions and simplifications about how the world works and how the world looks. Then from that, you derive insights, and you try to understand, "Okay, if this model were reality, what would it imply about how risks and rewards are divvied up in the marketplace?" You hit it spot on, Ben, is just that there's these undiversifiable risks that everybody cares about. The securities that are more sensitive to those tend to have higher expected returns, and less sensitive, lower expected returns, because people want to hedge those risks and are willing to pay. That's the insight from the model.
The challenge with that theoretical model is that it doesn't tell you what the risks are. It doesn't tell you what the state variables are. It doesn't tell you what you should go and test to understand what those state variables are. So you fast forward and you go to the '90s. You get something like a three factor model, or then it goes to the five factor model, and so on so forth. That's a model where the data are the hero, because the model really exists to organize the data so you can gain insights from the data. So you're really interrogating the data, but the model is a framework to help you do that in a logical fashion. What that allows you to do is it allows you to identify the types of variables that might be picking up sensitivities to these state variables, but you can't really ever prove it. You can't prove that value stocks are riskier than the market. You can demonstrate and provide lots of evidence they have higher expected returns, but you can't prove the risk here. So my view on who a tilted portfolio is appropriate for, except in certain specific circumstances which I'll get to in a moment, is that it's really driven by your sensitivity to deviations from the market. If you are okay deviating from the market, then a factor-based or a tilted portfolio can be very appropriate for you, because there will be times when returns are disappointing in an absolute sense and a relative sense. Unless you can stay the course and be a long term investor through those disappointing times, you won't be able to be around for when the returns are strong and make you very, very happy.
The reason that I say that it can be appropriate for a lot of investors is because when you think about real risk, it's uncertainty of lifetime consumption. If you read a recent blog post by Ken, he has five things that he's learned in finance. That's one of them. It's uncertainly about lifetime consumption. Investors are risk averse, so for a given level of risk, they want more return. In a factor-based portfolio, I can't really tell you're getting more risks. You're getting similar turnover to the market. You're getting similar diversification of the market. You're getting a similar volatility to the market. But you're getting higher expected returns. That's a pretty good deal. Now, the question then becomes, what about those other circumstances. Like when you think about lifetime consumption, your labor wealth and your labor capital are also important. There may be instances where you work for a growth firm. Well then, maybe a stronger overweighted value is appropriate. Or you work for a value firm, and an overweighted growth is appropriate. Or you work in education, where your compensation is very, very stable and expects to be there for a long time. Then maybe you can take more exposure to equities than somebody who has more volatile compensation. So I think there's those types of examples that will drive it. But overall, I think that it can be appropriate for a lot of people, as long as they have the right support, which is where the financial professional comes in, for getting through those times when results are disappointing.
Is a factor tilted portfolio more diversified than the cap weighted market portfolio?
It's an interesting question, Cameron, because tell me how you measure diversification. Different people approach it in different ways. Some folks, and you probably are familiar with this, like a Herfindahl index, which basically looks at squared weights and sums that all up. If you only had one stock in the market, well, then that would equal one. If everything was equal weighted, depending on the number of stocks, you get a much smaller number. Some people look at Herfindahl indexes, the higher the number, the less diversified; the lower the number, the more diversified. In some respects, that gives you insights, but it's incomplete, because an equal weighted portfolio, in my view, is certainly not more diversified than a market cap weighted portfolio, because in an equal weighted portfolio, you're overweighting micro-cap stocks tremendously. So you're taking huge bets on tiny companies, which doesn't improve your diversification. The whole notion then of expected return versus expected volatility, that's another way people look on it, but there you're relying on the data, and it can be data specific.
My starting point for diversification is the market. Take a globally market cap weighted portfolio. That gives you a good starting point on what diversification is at the security level, at the country level, at the sector level. Because remember what we said, prices are predictions of the future. They're forward looking. Prices are forward looking. What that really means is that when people are buying and selling, and assessing whether they want to hold an investment, they're making a trade off, the expected return of that investment versus its contribution to their overall portfolio. There's expected returns, expected covariance matrices, all built into market cap weights. They update real time all the time, so they're a pretty good starting point, in my view, for diversification. Then if you're a little bit different than that, you're probably equally diversified. If you're a lot different than that, then that may mean you're giving up some diversification, however you want to measure diversification.
We've touched a little bit on factors, and this idea that there are differences in expected returns, and that prices are predictions of the future. Can you talk a little bit about what criteria the variables that Dimensional uses need to meet before they're considered dimensions of expected returns?
Yeah. One of the big ones for us, of course, is sensible. Would we expect this variable to be rated to differences in returns across stocks or across bonds before you even look at the data? That's an important one. When you think about prices, they are predictions of the future, that means they have discount rates built into them. What return do people demand for holding a security? That's its expected return. The price sets it to a level. So the demanded return equals the expected return. When you think about it in that way, you say, "Well, what tells me about lower price?" So people are willing to pay lower prices, lower market caps, lower price to book ratios, lower price to earnings ratios, whatever it may be. What variables predict the cash flows that you might expect? So higher profits, or less retained earnings, or less asset growth, whatever the case may be. All of those variables you would expect to be related to differences in returns, because they're telling you something about cash flows or prices, i.e. discount rates, the two of those. You basically have price equals cash flow discounted back to today. So you have those three things to play with. When you think about things like company size, value, profitability tells you something about future profitability. Investment or asset growth, that's how quickly a firm is growing its assets, tells you something about how much cash flows are leftover for shareholders.
For example, if a firm is retaining a lot of earnings to produce a certain amount of revenue, well then less of that revenue is available for shareholders. So that will predict lower cash flows, higher asset growth, lower expected cash flows. So all of those things we would say we'd expect them. Then when it comes to the data, we say, we want it to be robust in the data. So we look across different regions, across different sectors. We do the experiment in lots of different ways. We say, "Is this observation robust?" Because you want to find out, number one, is it a premium that you should worry about, because if it's a tiny difference, then maybe it's not worth considering in your overall investment portfolio. If it's a meaningful difference, like value stocks have outperformed growth stocks by three or four percentage points, historically, over the past 100 years in the US and elsewhere, then you say, "Oh, that's a difference that's worth considering in the portfolio." Then you want to understand volatility. That is part of the communication. How bad can it be? That's always what we do. Let's set expectations first. How bad can it be? Then let's talk about the upside. Then finally, you want to be assured that you can actually capture, in a well diversified, reasonably low turnover portfolio. There's things that you'll find in the data that maybe are very concentrated in particular parts of the market, or result in very high turnover. That has to be considered as well before you go from the computer to the real world simulation. So all of those things feed into how we think about what to include when we're making portfolio design decisions.
Yeah. That's exactly where I want to go next. When it comes to actual portfolio decisions and implementation, can you talk about the sources of information that Dimensional draws from?
Yeah. The way I often characterize it, Ben, is that when you look back in the past 30 years of academic research, there's been three main data sources that have been used, and academics have done two main things with those three data sources. You have market prices. You have income statement data, so revenues, cost of goods sold, selling and administrative and so on. You have balance sheet data: assets, liabilities. The two things that people have done with those three data sources is look at current values, so market cap. Take shares outstanding. Multiply it by price. That's a current price metric. Momentum. Look at changes in price over the past three months, six months, 12 months. So you're looking at changes. If you look at something like value, take current price divided by current book value. Okay. Now, you're using current values. Asset growth. Look at changes in balance sheet items. Look at changes in assets over time. Profitability. Look at current profits divided by current book value or current assets. Profitability growth are changes.
So there's basically six things. There's 400 plus factors out there, but they're basically six. So then the question becomes, how many of those six do we use. Currently, we use five out of the six, and we're looking at profitability growth right now. That may be something in the future. Once you to decide that you're going to use those, then you say, "Okay. How do I use them together?" We use market cap, company size, so large versus small; value versus growth; high profitability versus low profitability; high investment versus low investment. Then we'll use momentum: stocks that have outperformed the market recently, and stocks that have underperformed the market recently. You say, "How do you use all those together?" There, we think about time scales. Size, value, profitability, multi years. If you have a value portfolio, it's about 20% turnover per year. That means when you buy a stock, you expect to hold it for five years. That's something that tells you about expected returns over one, two, three, four five years hence. Momentum. When you buy a stock in a momentum portfolio, you expect to hold it for about three months, four months. So that tells you that information is information about the next few months.
We say, "Let's let the long term drivers drive the asset allocation." So we use size, value, profitability to drive the asset allocation. Then use the shorter term drivers, like asset growth tends to be a bit of a higher turnover one. Momentum, a bit of a higher turnover one. Use those to say, "How do I time to get to that asset allocation?" I want to own everything under that asset allocation, but how I build weights up and decrease weights will be driven by some of these short term drivers. For example, if I'm doing a little bit of portfolio turnover every day, I can say, "I only want to buy the stocks that are value, small cap, high profitability, and are in upward momentum today." I have that flexibility because tomorrow I'm going to look again, and the next day I'm going to look again. The one other area I'd mention, Ben, on that one is there are other markets with other prices that can be helpful, like the sec lending market, where you get prices on how much people are willing to pay you to borrow a security. That's another market with prices. That actually also tends to give you information over the very, very short term. So those are the big drivers, and how we consider them, and how we put them all together at a high level. It's long term, short term, intraday. Then as we're making decisions, what we're trying to do is increase our weight in the stocks, on that day, that look good under all of those metrics, and decrease our weight in stocks, on that day, that look poor under all those metrics. We do that every day. A little bit, five basis points, 10 basis points of turnover every day, so a small amount of turnover every day, to keep the strategy focused on where you want it to be.
How does Dimensional decide between underweighting or entirely excluding securities in portfolios?
There's a few inputs into the decision, but one of the big ones is it depends on how the return pattern looks when you sort stocks on a particular variable. Let me give you two examples. If I sort stocks on price to book, so a value sort, what you typically see is something that's about linear if I go, let's say, quartiles of growth over the value, quartiles of market cap. You get a pattern that's approximately linear. As you go from the growthiest quartile, to the next quartile, to the next quartile, to the value quartile, the returns gradually get bigger. So it's monotonic, and it looks somewhat linear. When you have a pattern like that in the data, well then, that tells you that's an overweight type of an approach. You can take gradually away from the growth side, and add gradually to the value side. You're going to exploit that linear pattern in the data. Now, when you think about something like asset growth, so that's firms that have grown their assets quite significantly over the past year. So how quickly are you growing your assets? You can grow assets by issuing stock, issuing debt, retaining earnings. There's all different ways that a firm can grow its assets.
But when you look at something like that, and you sort firms on asset growth, you don't actually find any spread in returns until you get to the really high asset growth where returns are way lower. So it's flat, boom, low returns for high asset growth. So for that type of an observation, that lends itself more to an exclusion, where you're not getting what ... All the spread comes from the high side underperforming. None of the spread comes from the low side outperforming. It all comes from the high side underperforming. So that would lead itself to be more of an exclusionary type approach.
The one extra wrinkle I'd put on there is something like momentum, where the way that we approach that is we say, "Let's generate a set of orders that we want to buy today." Or maybe it's for ... they're in the value side of the market. Well, let's assign a higher cost to purchasing those stocks that are in deep downward momentum, and a lower cost to purchase those stocks that are in upper momentum. So that's not really an in or out decision. It's a timing decision, but that's how we would use that. It's more day to day, and saying that we have a preference to buy stocks in upward momentum, but the fact that they're in upward momentum is not the reason that we're buying them. Size, value, profitability, investment characteristics are the reason that we're buying or selling. But then on top of that, let's consider the momentum characteristics.
Yeah. That's all really interesting. I want to get into another fairly technical implementation question. We've seen some research suggesting that cash-based profitability, which subtracts accruals from operating profits, outperforms operating profitability, which is what Dimensional uses to measure profitability. Why does Dimensional use operating and not cash-based profitability?
That's a technical question all right, Ben. Let's back it up a little bit, and let's start off with accruals and accrual-based accounting, and try to keep that simple. The way that you think about accrual-based accounting, and this is the way the accountants think about it, and the accounting rules are set up to do this, is that on the income statement, you're trying to match the revenues that you realize with the costs that were borne to realize those revenues. Let me give you an example. Let's suppose I run a shop, and I buy some inventory. So my cash account goes down, and my inventory account goes up on my balance sheet. But let's suppose I don't sell that inventory for a year. I hold onto the inventory for a year. In a accrual-based method, what would happen? So operating profitability. A year down the road, let's imagine the inventory cost you a hundred bucks, and you sold it for 150. You'd have revenues of 150. You'd have cost of goods sold of 100. You'd have an operating profit of 50. Very simple scenario.
Now, an accrual would say that as you increase revenue, so you grow your assets ... Accruals are highly related to asset growth, by the way. As you grow your assets, you would've a positive accrual. So in that first year, when I increased my inventory by $100, I'd have $100 accrual. Then when I sold that inventory, it will be a negative accrual. It will go down $100. Accruals often come from balance sheet changes, changes in accounts payable, accounts receivable, inventory, things like that. Let's start a little bit with accruals and the history of accruals. Sloan, he was an academic in the mid-'90s, '96, wrote a paper showing that if you sort firms on accruals, those that have grown their inventory or their accounts payable and so on by a lot, that those firms underperformed firms that had low accruals. So very similar to the asset growth. Remember firms that grew their assets by a lot underperformed firms that did not. That research has been around for about 30 years now. The magnitude of the phenomenon has declined over time. There was some papers in 2010, 2015, in and around there, to show the magnitude declining. When we look at Dimensional data, because we have global data sets that are very, very comprehensive, and we can run these types of experiments, the casual observation, if you will, is that in large caps, the data are mixed.
So you see high accrual firms underperforming low accrual firms in the US and emerging, but not in developed outside the US, and the spread is not huge. In small caps, you find it particularly pervasive in the US, non-US developed and emerging. High accrual firms tend to underperform low accrual firms. It's all that cliff-like pattern. It's flat when you sort on accruals, except till you get to very high accruals. We also find the drop off. Pre-1990, small caps with high accruals underperformed by about 6% a year. Post-1990, about 2% a year, give or take. So a kind of a change there. That's the accrual part of it. Now, you say cash profitability. So now we're getting even more complicated, so let's go back to our simple example. In the first example I gave you, in the first year, cash account went down, you bought $100 worth of inventory. You didn't sell any of it, so revenue is zero, but no cost of goods sold, profit zero, operating profit zero. In the second year, you sold it for 150, cost of goods sold 100, profits 50, operating profits 50.
In a cash based accounting what would happen, is year one, your accruals are plus 100, because your inventory went up by 100. So zero revenue, you subtract off accruals, because cash profitability is operating minus the accruals, so minus $100 worth of cash profits. In year two, you sell, 150, cost of goods sold is minus 100, but the accrual is now negative, so it adds plus 100, cash profitability 150. So it goes minus 100, plus 150, or zero 50. Now, our view is that the accrual based method is more informative of the true economic activities of a firm because they're already lining up the timing of when the costs and the revenues are being realized so that you don't have to do that extra lineup. And so, we looked at this, because the paper came out in 2014, 2015, and Ken and Gene, Fama and French, sent us the paper and said, "You guys should look at this." Because we had been using operating profits for about two or three years at that point, and Savina on the research team took the paper, and she went through it, and we went through the whole paper. We connected with the authors of the paper. We said "Hey, you've got a few little things that you probably should do differently because we used all your results."
And Ken and Gene actually wrote a paper in 2014, 2015, where they used cash based profitability instead of operating profitability in their five factor model, so it was a choosing factors paper. So that was in 2014, 2015. And what we arrived at that point was that operating profitability remained the way to go for a few different reasons. One is that when you look at the ability of operating profits to predict future operating profits, hands down, it beats cash profitability. And that's because cash profitability is much more volatile, because it's minus 100, plus 150, zero 50, right? So it's much more volatile, so it does a better job there.
When it comes to returns sorts, when you look at large caps, there's not much going on, but when you look at small caps, in the US, cash based profitability produces higher returns for a high profitability portfolio, in small caps, than operating profitability. Outside the US, no, it's a push. But in the US, as soon as you kick out the high investment firms, which we do, in our portfolios, all that goes away. So our viewpoint was, we like accrual based accounting methods, we think that it gives you a better representation, as is in FASBI. You can read FASBI's rules and they will tell you, this is why we like the accruals because it gives you a better economic view of the firm. When you're already considering size, value, profitability, all those types of things, it performs equally well. It's more stable, leads to lower turnover, and so at that time we said let's keep on with operating profitability. But we looked at that in 2015, and then more recently as we've gotten questions, we wrote some papers about it. We usually don't write papers about things that we find that we don't do and that doesn't work for us, but when we get a lot of questions about it, then we'll put something out there.
That's a pretty good answer. Let's shift to goodwill. So we've seen research suggesting that goodwill overstates book value when companies overpay for acquisitions. Or if they don't overpay, goodwill ends up being double counted in investment strategies that target value and profitability. So, in either case, based on this research, goodwill should be accounted for. Can you tell us how Dimensional deals with goodwill?
Yeah. And another technical question, Cameron, you guys are coming with the technical today. So I'm going to back up a little bit and talk about assets and goodwill, and I think that it's good to level set on some intuition. Every asset is worth something because it produces future cash flows. That's the value of an asset. If I have land, why is land worth something to a company? Because it can produce some future cash flows for shareholders. If I buy a piece of equipment, why is it worth something? Because it translates into future values for companies. So, when you think about all those assets, assets themselves, as do liabilities, as do prices, all have information about future cash flows. That's why they're related to returns, because they have information about future cash flows. All of them do, not just goodwill, all of them. Same with income statement variables. That's why they have information about differences in returns because they have information about future cash flows.
So when you think about goodwill, what kind of an asset is goodwill? Goodwill is an asset that gets generated as part of an M&A activity. So, one company buys another, and then they go through a very in depth process. It may be a competitive bidding process as well, and they assess, here are all the tangible assets, the land, the property, and all that, and here's the value of those. Here are all the intangible assets, the patents, the licenses, the trademarks, and they assign a value to those. And let's say that those all together are $100, and then we're going pay 110, that $10 are goodwill. And you say, "Well, why would you pay more than the 100?," which was the value of all the identifiable, tangible and intangible assets. And that's down to the whole synergy question.
So let me give you a couple examples. In the US, one company can buy another and there's a couple of tax elections they can make. One tax election is where the company that's purchasing gets all the tax benefit. The company that's getting sold gets a step up when they get shares of the company that it's acquiring. Those shareholders pay the taxes right there and then, and things move on. Another way is that the shareholders of the acquiring company, in certain circumstances, can not have to have the step up in basis. And so, the company that's getting acquired gets all the tax benefit, right? So there are two ways. What you find is that when the company that's doing the acquisition gets the tax benefit, it'll pay more. Why? Because the goodwill that it pays has a tax value. You can compute the tax value. It can be depreciated over the next seven years. It has value. It's a very demonstrable value. It's an asset, and therefore should be reflected with the rest of your assets. Another beautiful example is Disney, when Disney purchased Lucas Films. They paid 4 billion, 2 billion for the identifiable intangible assets and 2 billion in goodwill. And you say why? Well, Disney said in their release notes that they felt that the Disney brand and the Disney distribution, with the Lucas IP, would be very, very valuable to the company and could generate a lot of future revenue, and it did generate a lot of revenue. And so you say, is that reflected in the historical profits of either company? No. It may be reflected in future profits, but it's not reflected in the profits that you're using to predict future profits.
So that's kind of the concept of goodwill. Now, where is Dimensional's place in all this story? The first time I remember us doing stuff on goodwill was in 2009, 2010. And at that time, I'm still on the Investment Research Committee, but I was the note taker on the Investment Research Committee back then, so I used to write up the minutes. And Jim Davis, I don't know if you guys recall Jim Davis, but he was a long term-researcher, fantastic guy. For your audience, Jim joined Dimensional after he was a professor, and one of the amazing things that he did was, he hand-collected book value data from the twenties to the sixties, so Fama and French's original research could be extended to a new out of sample period, and then he joined Dimensional. So he was running the numbers for the goodwill. And actually, I looked at my notes, because Brad had said you guys were going to mention the question, so I went back and I got the minutes from that meeting in 2010.
And the data at that time, the reason that we were looking at it is that there was a change in accounting practices around the year 2000, 2001, where all companies had to include goodwill on their book value in an acquisition and they didn't get a choice, which we thought was a good thing. And the data didn't tell you anything. It was kind of like, in the US, the value premium was a little bit higher. If you took out goodwill, outside the US, it was a little bit lower. We've run those numbers many times since, and if you take out goodwill value, premiums tend to be a little bit lower with the more recent data. So our view was, it's an asset. You can't tell anything from the numbers. It might make for a nice marketing story to say, "Oh, we're adjusting book value like X, Y, Z."
We're like, hmm, sometimes if you want to help somebody, you tell them the truth, and if you want to help yourself, you tell them what they want to hear. And that's a good quote to keep in mind. And we're like, "No, this is not worth doing." But, the thing I would mention there, and sorry for going on for so long, is that when you look at what we do to financial variables, we adjust them when we think they ought to be adjusted. So let's take book value. We keep two book values for every company, because a lot of companies will have minority interest in other companies, and that won't be reflected in their price, their market cap, so it's taken off, the minority interest is taken off for the book value. But you add it back in because their operating profits will reflect all of the profits from the companies they have a monetary interest in, so we make that adjustment everywhere.
Or we make adjustments on a case-by-case basis, as we need to. We make adjustments on thousands of companies each year, we have a whole financial data working group, because there are cases where you'd say, "Yeah, I want to change that data a little bit. Somebody is saying that this is an extraordinary expense, but it's been on their extraordinary expense line item for the past two years, so maybe it's not an extraordinary expense. It's an ordinary expense, now we should treat it as such."
So there's things that we do change. So my view is that when you're accounting for five or six variables, you can't really tell anything from the historical data about which blend is better than the other blend if you're doing the experiments fairly. And then it really comes down to how do you have the expertise in implementation to catch those outliers, and those issues that you're seeing real time in the marketplace, and making adjustments real time in the marketplace rather than wholesale adjustments on some particular variable.
Interesting. So it sounds like with goodwill, empirically, if I understood correctly, empirically adjusting for it doesn't make much of a difference, therefore it's not worth it. Extending that, is there a downside to doing the adjustment systematically?
My view is that it's the wrong direction. Goodwill is an asset and therefore should be reflected as an asset. You mentioned, Cameron, double counting. I don't know what that means, but it doesn't make any sense to me, because all assets have information about future cash flows. Balance sheet items, income statement items, and prices, so I don't know really what that means to double count.
The challenge with subtracting goodwill, and actually the market has gone in the opposite direction, the challenge is that whenever you are computing a ratio, and either the numerator or the denominator gets close to zero, or goes negative in some cases, that ratio becomes less informative. And so in as much as you're subtracting off goodwill and it takes book value to be negative, then you've made that ratio much less informative about how you can use it, so that would be one potential downside. I don't think it makes sense, number one, but that would be a potential downside. I might say the market has gone in the opposite direction, what I mean by that is that, and it makes more sense to me, if you can figure out ways to include internally developed intangibles in your book value, which would increase the book value, that you will get a more accurate reflection of that company's assets and that company's fundamental value, so to speak.
So I've got a question about that too. I've also seen research suggesting that incorporating estimates of internally developed intangibles in value and profitability metrics for building strategies results in larger premiums. How does Dimensional address that?
Yeah, that research, the cash profitability and the goodwill, to me, go in the wrong direction. They don't really gel with what I would find as, here's the spirit of what we're trying to accomplish and here's how the accountants view this. We always look back at the accounting releases when they make a change here, the rationale for the release, does it make sense? And does it make this variable less useful for what we're trying to use it for? But when it comes to intangibles, that's an area where I'd say that it makes sense. Because, let's back up, so an intangible asset is something like a trademark, or a brand, or a patent, or a license, things of that nature. You can't grab them, they're intangible. And what happens when you develop intangible assets, there's two ways to get them. One is you buy them externally.
So for example, in a merger and acquisition, and sometimes it's put in goodwill, sometimes it's put in intangibles, you buy them externally, and then it's reflected as an asset. In fact, about 25% of the assets on the balance sheet of US companies are intangible assets that they've acquired through acquisition. But when you develop them internally, so you spend your research and development dollars, or whatever the case may be, you expense it, you don't capitalize it. And it would be lovely if you could say, "Well, it's those intangible assets that turned out to have value." R&D that was done that turned out to have value you could reflect as an asset. Outside the US, there's some provision for that in the accounting rules, but in the US, not so much. And you see a little bit of internally developed intangible assets outside the US on a company's balance sheets.
So then it goes back to, "Okay, how might I do that? And can I come up with a good estimate?" And the academic research that we've looked at and we've reproduced, the estimates are far too noisy. And what I mean by that is, that you're making these Herculean assumptions about research and development costs, about selling general administrative, and saying, "I'm not going to expense them, I'm going to capitalize them." And what that ends up doing is just making book value noisier, in fact.
And the way to demonstrate that, we wrote a paper recently, and a really interesting one, where we gathered data from 700 mergers and acquisitions, so about $2 trillion worth of mergers and acquisition. And in a merger and acquisition, if there's an intangible asset, it gets valued in a competitive bidding process, or it's already demonstrated that it has value, so you can assign a value to it much more clearly. And then we said, "Let's compute the internally developed intangible using the academic methods, and see how well it predicts what actually happens to the value of intangible in the M&A." And the short answer is, it's a lousy prediction.
And what I mean by that is, like 25% of the examples that we looked at, it understated it by about a half, 25% of the examples, it overstated it by, the internal one, overstated it by about 30 or 40% plus. And so it's a really bad prediction. So we'd like to include it if we could get a better assessment of it, but we're not sure that you can get a better assessment of it.
The final point I'd make on intangibles, because we've looked at this extensively, is that when you think about taking something off the income statement and putting it on the balance sheet, taking the expense and capitalizing it, you're changing profits and you're changing book. So you're not just changing the price to book ratio, you're also changing profitability, and you have to consider the two together. So when you look at the data, what you find is that the value premiums, when you add back in an estimate of intangibles, tend to be a bit higher, a little bit higher. But profitability premiums tend to be a little bit lower.
Now, the question is why? On the value side, it's all a sector effect. If you control for sectors, then the intangible adjustment doesn't make a difference. And when you consider value and profitability together, and you look at the difference between those growth stocks with low profitability and value stocks with high profitability, it doesn't matter whether you controlled or not for intangibles. Which I think is comforting because the data that we use to inform our expectations has had intangibles in it for hundreds of years.
Disney, let's go back to Disney, Mickey Mouse in the 1920s, an intangible asset for Disney. We use those data, the book value data, and all of those data, in doing our historical experiments, and it never had that. When you look at the estimates of intangible assets, actually to assets, it hasn't changed in the past 60 years, that fraction has been relatively constant. So again, it doesn't invalidate the old research, the fact that intangibles are around and so on and so forth, but that's an area that we keep on looking at, and if we can come up with a better estimate, I think then we will give serious consideration to making adjustments to book value with those better estimates.
Hmm. Very interesting. So it's a valid idea, but not currently implementable.
That's our view, you hit it perfectly, Ben. Yep.
All right. So we talked about goodwill and intangibles, but more generally, and this is a question that gets posed to us fairly often. So what do you think about the idea that the world has changed a lot since Fama and French's initial research and that their findings are no longer valid?
Yeah, I often answer that question taking in a few different angles. And the first angle is, what is their research all about? And the research is all about discount rate effects, that there are differences in expected returns across stocks, and what variables can you use to identify those stocks that either have been assigned a lower price by the market or a given level of expected cash flows? And unless you think that's gone out of style, and suddenly investors don't demand different expected returns to hold different stocks, so to hold a micro cap stock versus a mega cap stock, they don't demand the higher return to hold the micro cap stock, then you should expect those differences forever going forward, right? Because those differences are driven by uncertainty, there's always uncertainty about the outcome of the economy, the outcome of a particular security, and some stocks will have more uncertainty associated with them than others, and so people will demand those differences in returns.
And so I think, from that perspective, that's something that's, I would say somewhat kind of invariant or perennial or should be here, you should expect it for a very long time. Things do change, accounting practices change, market microstructures change, and so on, and you have to adapt to those changes, 100% I believe that. But the underlying premise of what they were identifying, I don't think, goes away.
The other example that I point people to, and I think this is actually, in my view at least, one of the most unique experiments in all of empirical finance. So if you go back to the nineties, and you go to 91, 92, and when Ken and Gene were first writing their paper, they took a data sample that went from the sixties to the nineties, and then they developed a methodology on that data sample on how do you form a factor? And how do you divide up the market? And that methodology, by the way, has been adopted by almost every academic since then and it's a very robust methodology.
But they developed it on that first data sample, US sixties and nineties. Then Jim Davis comes along, we talked a little bit about him. He extended the data to the twenties, to the sixties. So then Fama and French, Davis, took that methodology, the same methodology that they had tested on the first dataset, and applied it in almost identical fashion to the second dataset, and found a similar finding. Then develop market data came along, and they took that methodology, almost the same methodology, and applied it to a brand new dataset, found a very similar finding. Then emerging market data came along, we provided a lot of that to Ken and Gene, and they took the same methodology and applied it to that dataset and found a very similar finding. And another 30 years passed, and they took that same methodology and applied it to that dataset and found positive value premiums.
So now you have five independent data samples. Five. One was used to build the methodology, and then it was applied to four others with very little change. That's very unique when it comes to empirical finance, very unique experiment that took 30 years, basically, to make. And in four of those five data samples, you have value premiums that are reliably different than zero, so the T stats are above two. And in one of five, you still have positive premiums, but it's not reliably different from zero, that's the most recent one in the US. And then you can't tell the difference between the value premium, the realized one, between any of the data samples, like they're not statistically reliable from each other. And so I think that's a really unique experiment in all of finance, which tells you something about the robustness of their observations back from the early nineties, and why that's still basically the benchmark factor model today, is the Fama and French three factor, five factor model. It's still the benchmark, 30 years down the road.
Hmm. Dimensional uses book value, as we've been talking about, to measure relative price. One of the other things that we hear a lot is that it's better to combine multiple metrics to measure relative price. Why don't you guys do that?
There's a few different reasons. One is that when you're considering market cap, price to book, profitability, asset growth, momentum, maybe in the future profitability growth, adding in price to earnings, or price to sales, is not going to give you anything extra. So when you're already considering all of this stuff, adding in another couple of variables are really not going to give you anything extra.
Two other reasons. One, and I don't know if you guys recall this, you may recall it from 2010, 2011 timeframe, when we were using price to earnings and price to cash flow along with price to book. And so what we were doing in small caps at that time, was we were sorting stocks on each of those variables and saying, "Who looks like growth under each one of those variables?" And we called those extreme growth stocks, and we dropped them from our small cap portfolios.
It was around 2011, 2012, when, it was actually Gene that said, "You guys are kind of shooting from the hip on that one." Here's a pay paper that I'm looking at, and it's a paper by Professor Novy-Marx. He says, "You guys should take a look at it." And Professor Novy-Marx outlined the gross profitability, so, the other side of value paper, and we looked at that and said, "Okay, that's very interesting. Let's reproduce the results. Let's decide how we want to use that observation." And so then we dropped using price to earnings and price to cash flow, and started using profitability in its place, because we thought that was a better approach. We used it in the past and then we switched it to profitability when we found a better way. And then subsequently, we developed a relationship with Robert, and he's been on your show, that was a really good broadcast. So he's been wonderful to work with. And so that's another example.
The last example though, I would highlight is, when you look at turnover. So let's say you're not going to consider all the variables, you're going to consider them one at a time. We don't do that on our strategies, but let's just pretend we did. One at a time, you don't really see much of a difference in terms of the historical return premium generated by price to earnings versus price to book versus price to cash flow versus price to, in particular, when you control for sectors as we do. But what you do see are differences in turnover. And so you tend to have higher turnover from these other metrics than you do from the price to book. So all of those combined suggest that, with all the variables that we're considering today, we're well positioned to have a relatively complete view of differences in expected returns across stocks without needing to consider these other variables.
The last one, Ben, sorry to go on, this came to mind, is that one area where we're thinking that we might use it in the future as an active area of research, is that, remember, we talked about when either the top or the bottom goes to zero in a ratio or to negative? What you'd do with those firms? Well, that's where, let's imagine you're using price to book and the book goes negative. Well then, what we do with those firms today is we don't include them in the value strategy, we include them in core strategies at market weights.
But, let's say we did a price to earning sort, or a price to sales sort, and replaced its rank on price to book with a price to earnings rank. That might be a place that you could do that. And the reason that we've been looking at that recently is because there's been an increase of negative book values in the US, and so, it's historically been quite small, but if that grows in the future, then you want to have some way to account for a bigger percentage of market cap in the future. So that's one area where we have looked at, may look at again in the future.
You got to stop apologizing. This is the last podcast where you would apologize for talking too much about how to measure relative price. Now I got to say, I didn't know that Dimensional was using multiple metrics, pre-profitability, for the small cap growth exclusion. I knew there was an exclusion, I didn't know that it was using multiple metrics, that's interesting.
Yeah. Yeah. I mean, when we find a better way to do something, we do it. And if that means that we change a variable, or add a new variable in, that's what we do. But we have a very high bar about what's considered better.
Right.
But when we're confident, then that's what we do.
How do you target value and profitability together? And is it done the same way for large and small caps?
So, it's an interesting one, over the years, we've made them our focus, especially in core type strategies, much more similar, so a similar emphasis on value and profitability. And one way to illustrate that is just think about the market and split it into four buckets. So you have growth low profitability, growth high profitability, value low profitability, value high profitability. And so value high profitability, giving you some profitability premium, value premium. Growth low, some profitability, no value. Value low, some value, no prof. And growth low prof, neither.
And so the way that we think about equal means that if we're taking some weight away from the growth low profitability bucket, we give most of it to the value high profitability bucket, and then we give equal amounts to each of the value low prof, or growth high prof. Right? So that's how you can think about similar, that we're looking at them together, we look at them simultaneously in a core type strategy, and we're saying we're over weighting those firms that have good value and good profitability characteristics by the most, and that's how we do it in core strategies that are all cap.
In small caps, we might do something a little bit differently, if it's only small cap, so there's no large caps at all. And there, we often don't do this over and under weighting because that can lead to some turnover in areas of the market that can be costly to turn over the portfolio, there we might do an exclusion. So for example, if it's a small value strategy, we'll exclude some firms with the lowest profitability or if it's a small cap strategy, then we're sorting firms simultaneously on value and profitability and excluding some firms that are growth, low profitability type firms. So, they're the ways that we generally think about it but in a core type strategy, we think, you should have about an equal mix of both.
What are some of the other ways that they can be targeted together? And why has Dimensional decided not to implement those?
Well, there's different things that you could do, for example, you could combine the metric, so rather than do an independent sort on value and independent sort on profitability and then look at the intersection of those independent sorts. So, whose value and high prof and so on. You could then average the two ranks and come up with a single rank, that's another way to do it. And it has its merits. It has some pluses and some minuses, I would say on the plus side, it's simpler to illustrate because once you have value, company size and profitability, you're in a three dimensional grid and it's hard to show that on declines and becomes a little bit like, but what are you showing me here? So that can be a little bit challenging. So that's where you would have some benefits for doing something where you blend all those together.
On the flip side, I would say in my view, you get a little bit less control because you're really understanding as you push weight away from market cap weight because we're going to take the market and we're going to say, I'm going to overweight these stocks and underweight these other stocks, relatives to their weight in the market. You end up with a little bit less control about how you're pushing to both or three or four of those premiums at the one time. Both methods are reasonable. I would say that there's not a massive wedge between them and we've chosen that particular one because we feel that it just gives us a little bit more control.
Interesting. So Dimensional started in small and micro cap stocks back in 81, I guess. So as you continue to grow now, how do you think about capacity for the investment strategies in those original small cap, micro cap stocks?
Cameron, I'm sure you've heard this from Brad many times. It's all about the size of the room and the size of the door. When it comes to global equities and globals and bonds, the size of the room is massive. So for example, a global open stock market right now is about 70 to 80 trillion US dollars. It is a big, big room. And when you look at something like a core strategy, the size of that door is enormous. So why is it important to have a big room? Because as you grow in assets as an organization, you start to purchase more and more of companies and there's different types of restrictions that will apply when you become too big a shareholder of a particular company, so that puts an upper limit. And when you have a very, very big room, like 70 to 80 trillion, you can be a $560 billion manager and be a very small footprint in that very big room. So that's why people talk about the size of the room.
The size of the door is important because you want to keep a strategy focused on its asset category that you're targeting. So if you have a value strategy and you have so many assets in it that you have to let it drift to growth, well then when the value premium show up, you're not there to capture them and you're not well positioned to capture them. So the size of the door tells you something about how can I implement the strategy that I'm promising to deliver to our customers. And we have no issues with the size of the door because of the way that we design strategies, for a couple of reasons. One, we do a little bit of portfolio turnover every day. So that means that we're taking, let's say you have 20% turnover in the year, 10% turnover in the year, you're doing five basis points, 10 basis points of turnover a day.
That means each day you're taking a tiny part of the overall liquidity in the marketplace. So you might be 2%, 3% of the aggregate liquidity in the stocks that you want to trade. And you're not trading all the stocks in the marketplace. The way that we talk about it is participate, don't initiate. We don't want to be the ones initiating the trades, we want to be the ones participating the natural volume in the marketplace. And that's the way that we've designed the strategies. But when you look at years like last year or even this first quarter, that really tells you, 2021 and the first quarter of 2022 and you look at the relative performance of our strategies collectively. I mean, we blew the socks off indices, value indices, market indices, whatever you want. And that tells you that we've had no challenges implementing our strategies because we've been staying focused on value for value and high prof for high prof.
And when those premiums showed up, the investors in those portfolios got paid very, very handsomely. I mean the level of out performance was massive. We go to the first quarter of 2022. If you look at the major indices, US large, small non-US developed large, small emerging markets, all negative to the tune of five or 10%. If you look at our value strategies, all in the positive territory to the tune of a couple of percentage points because they were there to capture those value premiums. And you look at our trading cost analysis, our advantage over our market peers hasn't declined at all over time. So we worry about it. We think about it. We factor it into how we design our strategies but my viewpoint is that we have a long way to go before we need to adjust or close or any of those types of things that people sometimes do when capacity become the problem.
Has the entry into the ETF market changed any of that thinking or the way that you're thinking about it?
A little bit because in the ETF market, as you know, at least here in the US, what happens is when, unlike a mutual fund, in the mutual fund, the mutual fund deals directly with the end customer. So the end customer gives cash in and then takes cash out. So they exchange for cash and an ETF, the ETF deals with a limited set of what are called authorized participants that are institutional type shareholders and outside of some countries in emerging markets, what happens is they come in, in kind, what does that mean? They deliver stocks to get shares of the ETF and they go out in kind, IE they give shares of the ETF back in exchange for stocks inside the ETF. And it's the same on the bond side, they come in, in kind and go out in kind. And so what that means is that some of your turnover now in the ETF can be accomplished through that create and redeem.
That's what the process is called. And so that alleviates or reduces some of the trading that you have to do in the marketplace. You still trade in an ETF or at least our ETFs because we are in what they're called non index or active, transparent type ETFs. So we still trade a little bit every day in those ETFs but we also can use the create redeem where somebody else is doing the trading but we're getting closing prices. So just that's what it is. There's no real trading benefit or disadvantage from the create redeem process.
Interesting. So Gerard, is there an expected premium for owning smaller stocks over larger ones?
And my view is, yes but it's one of those things that is a little bit nuanced in the sense of the premiums interact. And that means that if you're a small cap stock and you're picking up some of the size premium but you have a very negative value premium or negative profitability premium or negative investment premium, then that will more than offset whatever you picked up from the size premium. And so my view is that when you controlled for those things, then small cap stocks on average have had higher returns than large cap stocks. The other way that I often think about that question, Cameron, is that when you look at different factor models, almost all of them work better when you include small cap factors or small cap stocks.
So when you're using factor models to explain the returns of very diversified portfolios, if you don't have a small cap factor in there and include small cap stocks in there and the portfolios that you're trying to explain include small cap stocks, they don't work nearly as well. So that is another piece of evidence that they're an important aspect of the overall portfolio. But regardless of what you feel about small cap premiums, they deserve a place in your portfolio. They improve your diversification and they help you pursue the other premiums because you can do it more diversified. And sometimes this spreads between value and growth and high prof and low prof and so on, have been larger in small caps than the large caps.
How important is sec lending, security lending revenue to the expected returns of dimensional funds?
I think it's important. And then it comes under a broader category because Ben, up till now, we've been talking about what's your buy, hold sell discipline more or less, how do you decide what to buy, what to hold, what to sell? But there's a second area of value add, which is how do you improve the investor experience when you're holding the security? And security lending is one way that you can do that, where if you have a process that's optimized and integrated with your portfolio management process, then you can loan out stocks, you get all the collateral back, so there is some risk but the risk can be well managed. You can invest that collateral in money market fund or something similar and then people will pay you or pay the investors in the fund money in exchange for taking that security out on loan.
And I think that's an important area of value add, it's not the first order, it's second or third order but it's an important area of value add and something that we spend a lot of time on trying to improve and increase the efficacy of. The other thing about it is, is that it gives you information about the securities lending market, which tends to be opaque. Can we use that information in other areas of how we manage the portfolio? The other aspects though, I would mention on that sometimes people worry about is, well, why is somebody borrowing your stock often to short it? Our view is that if you have enough information that you think that this stock is going to go down, that information is going to work its way into the price of the stock, whether we loan it to you or not.
So we figure let's get our investors paid while that information works its way into the price of the stock. And when people are willing to pay you a lot, the amount of under performance actually tots up to be about the same as the amount they're willing to pay you. So it tends to be somewhat of a push for the investors. The other item I'd mentioned though there, Ben is adding value, is stewardship. When we purchase securities in the portfolios that we manage for clients, we start to engage with the companies and our view is things like E S and G and all those risks are affected in company price. If through our engagement, we can improve governance, you should get a higher price. So how do you make the stocks work for you while you're in the portfolio? There's stewardship, there's securities lending, there's how you vote on different types of corporate actions. All of those things are in that broad category of, yeah, we have a great buy, hold, sell discipline but we can also add value for end investors while we hold the security.
Really interesting. So is sec lending pretty much table stakes or do you have some competitive advantage?
When we look and we benchmark our revenues and so on versus the industry on a stock by stock comparison, we often are able to get higher fees and leave it out for longer, so get better revenues. And that's in part because of what I mentioned. So we've written different tools and built different tools, such that when we're selling a security, we can time the pace of the recalls, such that we can leave it out for longer if it's generating revenue or we have the flexibility that if we see a stock go on loan at a high fee in the marketplace, we can jump into that because we have flexibility. We don't need to sell that stock today or we can hold onto that stock for a period of time. So we work with our lending agents to improve their revenue. The other peak then is that we've developed technologies and so on with our agents so that we can lend more efficiently in countries like Taiwan, where the lending revenue tends to be a lot higher.
So we get that, we just have a different process that manages the risk, that maybe not all the managers have that process to manage the risk can be comfortable. So I would say that we do better than the industry. And when you look at our Morningstar averages relative to the industry, we tend to outperform the typical industry by depending on the asset category, as much as 10 basis points and that's meaningful because that's the type of fee differentials that you'll see between us and index based approaches. So if you can add that back in with the securities lending, that's money in the pocket of the shareholders of the funds.
How does Dimensional deal with sector weights? And what's the, as you've been giving us all along, what's the thinking behind it?
Yeah. The thinking behind it is, goes back to your earlier question on diversification and the market provides you that snapshot of diversification at the security level, at the sector level, at the country level. So when we think about sector weights, we say, "Well, we're going to deviate from the market but we don't want to take such a massive bet versus the market that we're giving up some form of diversification that we could otherwise manage." And so what we do is we say, we look at a sector's weight in the market and we say, "We're not going to go over that weight by more than 10%." And so what that inherently does, it also is one of these tools we use, when people get into these arguments about different variables and so on. All variables, financial variables have warts when it comes to identifying stocks with higher expected returns, they all do, earnings does, book value, they all do?
And so the way that you deal with those warts is you say, well, "I'm going to use the variable but I'm not going to let it make me look so different in the market that I'm going to have an expectation, a bad outcome if this premium doesn't show up over the next 10 years, if it's zero over the next 10 years." And so in that way, what we end up doing in a lot of our portfolios is we start off with a cross sector comparisons on these various different financial variables. And if some sectors want to be too heavy in their portfolio, then we start doing within sector comparisons so that we can scale back the weights of those sectors. So we do a cross and within sector comparisons of securities on these various different fundamental variables, which also helps with the question that you had earlier on about, are you missing something by not having earnings price and all that thing in there? When you consider the variables that we do, plus how we control sector rates, there's no real improvement by putting those variables in a broad sense.
Very interesting. All right, I've got one more. We have seen research suggesting questions. We've seen research showing that there is not a credit premium, a premium for owning riskier bonds over safer bonds after you control for equity mark factors. So that would suggest that credit does not add an independent source of expected returns. How did Dimensional assess that and decide to implement credit?
Yeah, I'm not familiar with that research but it may be something like where people ran some factor models or things of that nature on the returns of credit bonds. There's a few, I think, simple ways to think about it. And here's a simple question. The correlation in the US of small cap stocks with large cap stocks has been about 0.9. So if I constructed a factor of large cap stocks only to explain the returns of small cap stocks, I would explain some of their returns. Does that mean I shouldn't own small cap stocks? No, I probably should they improve diversification. If I hold the stock of Apple, does that mean I shouldn't hold the bond of Apple? Of course, it can give me a different payoff or a different type of returns relative to the stock. When you look historically at the correlation, whether it's pair wise.
So you take the correlation, let's say, I control for issuers, I only include issuers that have both stock and bonds. So they're the only issuers that I'm looking at. And I look at the correlation of their stock and their bond returns. And then I average that across issuers, that tends to be close to zero for AAA bonds and about 0.2 or 0.3 for BBB bonds. So it goes up a little bit as you go down the credit spectrum but that's a very low correlation in terms of adding diversification to the portfolio. So when we looked at credit in the early days, so pre mid 2000s, the market infrastructure for us, wasn't quite there where we felt comfortable. And what I mean by that is the trading infrastructure and the pricing infrastructure. And in the early 2000s, what started to happen was we started to get a lot more transparency in where bonds with lower credit quality were trading and that was what was called Trace.
Well, recently we've added additional databases called Tracks and Emma and other types of databases that give us even more transparency when it comes to bond pricing. And that was big for us because then what we could do is come with enhanced credit monitoring. We could take information from the credit rating agencies. We could take information from market prices. We could take information from credit default swaps, and we put all that information together and we come up with a real time credit quality for each bond, 15,000 of them, updated every 15 minutes throughout the day. Now you're talking that we can do some real credit monitoring here. The second part of the development is in how they trade and over time, the efficiency by which they trade has really increased with peer to peer trading and all of these types of trading venues for this debt.
And so what that implies at least for us, is that we're able to get very broadly diversified exposure in a very systematic way to a set of bonds that can improve the overall return profile of a portfolio. And we think that's a good thing. The last part I'll mention, I don't know if you guys do it in Canada but down here in the US people look at the Fed and what the Fed's going to do all the time. What's the Fed Funds rate doing? Is it going up? Is it going down? And that's fine. But when you think about the return earns of a fixed income portfolio, especially one that includes credit across AAAS down to maybe BBS, instead of just the Fed Funds rate, driving the returns of that portfolio, you have returns across the entire maturity spectrum. You have returns across lots of different issuers, 10 to 15 different currencies and then you have returns across lots of different credit qualities.
So you have about 600 different interest rates driving the return of your portfolio. And that means that, that portfolio can have positive returns in time periods when the Fed Funds rates are zero or when the Fed Funds rate goes up a little bit or down a little bit. And that to me is another enormous advantage of having those lower credit quality bonds in a portfolio. In particular, when you have portfolios where the blend is heavy equity because then you don't really change the volatility characteristics at all, by having something that is all of investment grade. If it's very low equity, then putting in, moving from just AAA down to all of investment grade will increase volatility. But if it's very heavy equity, you hardly see a difference. You see a difference in expected returns but you don't see much of a difference in the volatility characteristics.
Wow. So what have been the biggest changes within Dimensional portfolios say over the past decade?
I think that we've touched on quite a few of them, Cameron, some of them, profitability has certainly been one, I think. Another one has been investments. So that asset growth, we've implemented that in the past few years. So that's been a big one. Another big one, I think has been the securities lending that we talked about earlier on, doing that globally. So that exclusion, when a stock goes loan at a high fee, we excluded for a short period of time from purchase. We don't sell up but we excluded from purchase. I think they're among some of the big ones. On the fixed income side. We've added to our fixed income lineup dramatically. We've added DDs here in the US, we've added some mortgage backed type securities. So we've added a lot on the fixed income side. And then the big elephant in the room, I would say.
And you mentioned it on Ben, is the ETFs and what we've done here in the SMAs over the past three, four years, we've been working real hard on that. So it looked effortless from the outside but I have to tell you, it was a labor of love from the inside. Where we were in the media all the time in part because we were doing something so brand new, so innovative and people couldn't understand how we made it look so simple. Something that was so complicated, which was taking six mutual funds with many billions of assets and converting them into ETFs. And then the flows that we've had into our ETFs have been just eye popping in the first 18 months. We're about to become a top 10 ETF manager here in the US about 18 months after starting. And in my view, we made it look effortless but it was really quite a lot of work going on behind the scenes.
And that was a big deal. And it was figuring out how do we take what we do in fund format and do it an ETF format without losing things along the way so we can deliver the same value add and that was a big deal. And the last one I mentioned is the SMAs and that may come global but we were pretty close now to having a relatively complete solution here in the US but we dropped our SMA minimum from 20 million to half a million here in the US and we built the whole FinTech solution to enable that.
And the feedback has been amazing from the advisors that we work with so far and that indeed, maybe something that we can extend more broadly but that was also quite a significant undertaking is, how do you take what we do with institutional size money and do it with a half a million dollars while giving the ability to customize on your values, customize on your tax situation, customize on your human capital. I work for this company, therefore I don't want to hold it in my portfolio. And that was a big deal too.
So I've heard David Booth in the past. Talk about applying learnings from science to basically pick up pennies. So I was wondering if you can give us some examples of perhaps some really small things, small adjustments you've made in portfolios that certainly people may not have heard of.
Yeah, there's a few in there, Cameron. One was the one I mentioned about the recall process of securities on loan. That was a little project where we were, when we wanted to sell a stock, we say, "What is it loan?" "Yes. Okay. In order to be safe that we can settle this trade, we're going to recall it all." Say, "Hey, that's leaving some money on the table." So can we set up a process by which as we're selling the stock, we have a very efficient way to recall just the amount that we need to settle the trades so that we manage that risk well. And that is incremental but additive to the sec lending revenue that we can get for our clients. Another one, Cameron, is netting of FX trades. So let's say one portfolio is buying euros and another portfolio is selling euros.
Well, you can do it individually and one buys and one sells or you can set up a process whereby you only do the net. So if one is buying a lot and one is selling a little, then you subtract and you only do the difference. And that saves us and the shareholder of many of our funds, a lot of trading in FX and paying potentially a bid offer spread because we can net that FX across and we've reduced the amount of FX trading as a result, quite significantly to tune of many billions over the past few years. Things like tracks. So that's a data set that we included recently, that can gives us even more information about interday pricing for bonds and where they're pricing the marketplace. Here's one Cameron, that I bet you... Let me ask you a question. You've been asking me all the questions. How many ESG data providers do you think we have?
I don't know, five.
Eight, good guess but eight. And we've been doing sustainability portfolios now for about 15 years, give or take, in the incoming side and in the separate account side for even longer. But even that those processes to improve how you identify, which one of the thousands of companies that we hold on behalf of clients to engage with, then how do we make ESG type decisions? All of those types of things are incremental improvements that people really don't see unless you're in the middle of it, working on it.
So we build Cameron and I, for our clients, portfolios that follow the dimensional core and vector strategies. And we combine those in a certain way to get tilts toward the factors that we've been talking about. Alternatively, and we've played with this, we've back tested it. We could take a market cap way to the index fund and combine that with a small cap value ETF for different geographic regions. And we can make that combined portfolio look very similar to the current core and vector portfolios that we use. Similar in terms of characteristics, in terms of regression coefficients, in terms of back test performance. So it looks very similar on the surface. How would you compare those two approaches? And I should add the combined cap weighted and small cap value portfolio costs a bit less. The fees are overall a little bit lower to do that. So I can make these look very similar through the lenses that I mentioned. How would you compare those two approaches and how would you explain why the cost of the dimensional fund is, the additional cost is worth it?
So let me take that question in two parts, Ben. One on the fees, Dimensional typically is in the lowest decile or second decile around the world broadly when it comes to fees, we're a little bit higher than indexing and we're much lower than traditional active management. So that's where we sit. So regardless of if it's a core portfolio or an asset category portfolio, we think that all the things that we do with respect to implementation more than covers the fees that we charge. And so you see that in our net returns, out performance, net of fees and expenses relative to an index. So that's one part of the question. The second part of the question, Ben, you're right, that when you have the benefit of hindsight, you can blend a small value and a market strategy to get a similar average return and a similar regression coefficients historically but you're not actually getting a similar asset allocation.
The asset allocation is different and it's different in a few meaning for ways. One meaningful way is the integration of the premiums. So the way that we think about excluding because of your asset growth characteristics or your value and profitability characteristics together in small caps, you can't reproduce that by blending a small value index in the market, you just can't. So you can't get that when you blend those two. The way that we blend value and profitability, you can't get that by taking a small value index and blending it with the market. So you can actually reproduce the asset allocation. And we think that asset allocation adds a lot of value. The other aspect of it that I would highlight is that when it comes down to managing risks, that an integrated approach is very much better. So if you look at the amount of the portfolio, that's overweight by a certain factor relative to the market. So let's say, I say that, how much of the portfolio is more than 10 times market cap weight? How much of the portfolios between five and ten, how much of the portfolios between one and five? Well, in a core approach, you'll find almost everything between one and five. When you do asset category plus market, you have this kind of everything at market, and then this huge spike for small value where it's 10X smart cap weight, and you have a big chunk of the overall portfolio at this very much overweight position relative to the market. And that's... I would say a less well diversified asset allocation.
Other things then would be things like turnover. If you take a core portfolio and you replicate it with asset category funds, you'll find the turnover is about, 10% and to 15% higher because the asset category funds are buying and selling between them and the core portfolio doesn't do that. So there's a lot of benefits we think of the integrated approach that you don't see from just a high level characteristics or the benefit of hindsight of if I take X% in this and Y% in that, I get a similar average return historically, or similar regression coefficients. There's a lot more going on under the hood that we think more than make up for the fees through better risk management or higher expected returns.
You mentioned earlier how professors far more French or others might send in academic papers and your research group will take a look at them. Do you have examples of papers that seemingly were pretty compelling that your research group has assessed and decided that it really wasn't for using in your portfolios?
Yeah. One that came in are a theme that came in that it's premise and if it were true would be very compelling, which is all the volatility work, where you sort stocks and volatility that you can have the same returns as the market, but a lot less volatility. That's pretty compelling research because at lower volatility than the same return, you have a higher compound return. So you grow your wealth more quickly. That's a compelling idea. And so we took that very seriously. We looked at it in depth, professor Novy-Marx also looked at it in depth. And what we found is that volatility predicts future volatility. But the reasons that the low volatility stocks historically came in with market like returns was because of their value and profitability characteristics and not to do with the volatility characteristics themselves. And so it was kind of one of those items where, should you expect those value and profitability characteristics from low vol stocks going forward?
And if you look back historically, well, on average, they had those. Sometimes they look like growth. Sometimes they look like value. Sometimes they look like high profs, sometimes low prof. And if you segmented them into time periods, when they were only growth and low prof versus value and high prof, then you see a big under performance in the times when they were growth and low prof and a big out performance in times when they're or value in high profit. So unless you really believe that low vol should always be the same as value and high profit, and that those two things work together, then it's not clear that you will get that same return pattern going forward, and you may have a lower return than the market. And so our view at that point was, "Well, you can get that by blending some fixed income with some equity, it's probably more robust. It's probably more reliable." So, that's where we arrived at from that research.
Hmm. That's a good one. I like that. When we were talking about accruals, you mentioned a similar kind of story where you took in the research and the capability that you guys have to do that is fascinating.
Yeah. We have about a hundred people on the research team, Ben and we've developed databases over the years that I think some academics will be envious of have access to because we can run stuff more broadly, more quickly, more comprehensively than many academics and also some of our competitors.
Now on that Dimensional only relatively recently started publishing internal research. I mean, we've as advisors been able to access it through the Dimensional website for a long time, but it was always behind a login screen. How is a decision to start making that research public made?
Well, Ben, we listen to folks like you and you've been telling us for a long time, "Get in a fight, Dimensional." So we took a lot of feedback from clients and this happened about five years ago. I would say Ben, where it wasn't just on the research side of things. It was more overall the public presence side of things, where the financial professions that we work with, whether their advisors or their institutions and so on, they all have their own constituencies. People that they're appealing to. So it's the mom and pop or the investment committee or whoever it may be the board of trustees. And the feedback that we were getting is that, "We know you, we love you, but the people that we're dealing with don't know you and they'll know anything about you. And we'd like them to know more about you, not just through us", so that when we go to work with them, they already have some understanding of Dimensional.
So at that time we started to put more effort into it. So we hired a person called Darcy Keller, and she's been fantastic in helping out with all different types of communications and she's built out a team there. And we also decided that it would be appropriate to put out our research on SSRN. We would do the research and it would be academic quality in my view, but we wouldn't write it up as an academic paper because we've been client and investment led forever. So if our clients don't care about it, we don't care about it. And if our clients know what we're doing, we're happy. We don't need other people to know what we're doing. If our clients know what we're doing, we're happy. And so our clients, weren't really wanting academic style papers from us, but then that kind of changed over time. So when we do a piece of research, we write it up as an academic paper, we write it up as a shorter white paper. We write it up as a blog. We may record a short video. And so we get better use out of all of those media on, how to translate those results into ways that clients find useful. So over the past three years, Savina has done a great job there. She's had about not just her, but the whole team, probably around 15 papers put on SSRN of lots of different topics, which I think that a lot of the advisors that we work with have appreciated and have enjoyed seeing those papers come up. Clients were telling us that, "We want more of you in the public domain." And so we said, "Okay, let's figure out how to do that."
Hmm. Can you share what the next big thing is that research is working on if there is next big thing?
There is a few and I don't know how big they will be, but I mentioned profitability growth. We've been working on that with professor Novy-Marx as well as the internal research team. And that kind of came from a paper that Robert did a few years ago. I think it was like fundamentally momentum is fundamental momentum or something like that, where he was looking at earning surprise and how it explained momentum and that translated into, well, should we look at how profits have been rolling and can that enhance your description of expected returns?
Another one is on the very short side of returns. So returns over the past week are few days. And we kind of take that into account with how we trade, but we're looking at are these very short term reversals, something that we can potentially add into as kind of that very short term effect. So in the timing of how we generate, buy and sell order. So, that's something that we're looking into. And then we've done a bit of work on asset allocation over the past number of years, whether it's goals based or whether it's kind of wealth based or whether it's kind of more risk based. And I think that you'll see more writeups and examples of asset allocation coming from Dimensional that hopefully financial professionals find helpful as they decide, which is that is useful to them to work with their clients on.
That'll be neat to read. And I do want to come back to my ICAPM question, but well, I'm going to save that. I'm going to say that for the end. I had a couple of follow up questions that came to mind as we are talking. So especially now that all of the research, not all the research, but a lot of the research is being published where anybody can go and read it, which is great. And even beyond that, a lot of the academic research that the thinking for Dimensional stems from has been in the public domain for years, if someone decided they wanted to set up a Dimensional competitor, given that all that information is out there, what do you think is the most difficult part of the firm to replicate?
Yeah, Ben, we've been often imitated, never replicated in my view. And that doesn't mean that other people can't come with good solutions, because if you look at something like strategic beta in Morningstar asset category, that's about a one and a half trillion dollar asset category now of investments when it was a hundred billion, about 10 years ago. So people have looked at what we've done over time and done variance of it. But if you truly want to create a new Dimensional, you have to take all of Dimensional. It's as simple as that, because when you think about, I mentioned rules based approach, that's what we try to employ. And all rules are down to somebody's judgment. When it's an index, some index committee applies some judgment and creates some rules. And then they try to give that to asset managers. If it's traditional active, that individual portfolio manager has some rules that they apply, they may not be able to communicate them well, and you may not know what to expect from them, but they have some rules. We have rules. And the judgment that goes into informing our rules is been done by financial professionals that are familiar with research, with portfolio management, with trading, with portfolio design over many decades, and those rules have evolved over decades. So they're institutionalized knowledge at this point that no one person at Dimensional has complete knowledge of all the rules that drive and govern our portfolios. But those rules have been battle tested. As we go through a financial crisis, like in 2008, 2009, the rules evolve, adapt and improve. So that for the next crisis we're better. And so when you think about that in itself, there's just so much knowledge encapsulated in the systems and the systems themselves are so evolved. Like I mentioned, the eight ESG data providers, that takes some serious effort and some serious expertise to feed that into your systems and then make informed decisions on it or all of the other data that we do where we're adjusting company financials across thousands of companies each year to make them better fit for purpose for what we're trying to accomplish.
So I think that there's an important aspect there just on the investment side. We remember at the start of the conversation, we talked about a 'rules based approach' is a good approach in the particular for when we're working with intermediaries like yourselves in financial professionals, with the right innovation, evolution, with the right support. And when you look at the support that Dimensional has built over the many years, that support is critical because it enables people to become long-term investors and go through time periods when returns are disappointing and strong and remain committed to the long term. And if you think about it in that sense, that support is also a part of our value add. We have over 110 client communities globally, and across those client communities, we had something like 500 events last year in 2021 with about 2,500 people.
And those client communities are emerging leader communities, study groups, women in wealth, all different types of communities where clients can come together and we're kind of the central point. And they all get better as a result of coming together. When you look at the conference... We did almost a hundred webinars last year, and we had almost 20,000 people come to those webinars. All timely topics and it helps them stay the course.
So it's not just the investment part. There's also this whole client support part that's integral. The last example I use, is the mutual fund conversion to ETFs. That took operations, legal, compliance, you name it, finance, and they all have to be familiar with what we do and know what we do to make that happen. The way I think about it we're one team, one dream, and everybody here at Dimensional understands our investment approach and there's only one of us. There's one investment approach that we've been trying to perfect and get better at for 40 years. We have that common language between us and that means that all the teams can coordinate, communicate, and that is really challenging to replicate anywhere else. You can do things that are similar and they can be fine and they can be great, but they're not Dimensional.
I have a back test question for you. So it's not hard through back testing to find a product that may have beaten one of your products, such a small cap value doesn't matter which one. So what do people need to be aware of when they're comparing back tests?
Yeah. You're never going to be shown a bad back test by a manager, Cameron. And that includes us. We're not going to show you a bad back test either, but I think that the tricky thing about back tests is that, I think that folks who don't actually work with the data and get hands really deep into the data, don't understand the games you can play with data. And that's really important to understand. I'll give you an example for small value. You bring up small value. I could take a US small value simulation, and I'm going to do a based same definition of the asset category. I'm just going to use price to book. I'm going to exclude some stocks and profitability. So same definition across all of the different simulations. And then I'm going to change the rebalancing month, or I'm going to add a momentum screen, or I'm going to lag the price and the price to book ratio, playing these little small games. What do you think the spread and returns I can generate by playing these small games are for this small value strategy are historically?
100 basis points.
100 and 200 basis points. The worst and the best I can get you almost 2%. Now, what the great thing is that every one of the simulations shows a strong value premium. So there's a reliable value premium. The value premium itself is robust to all the simulations, but by playing these little tweaks and these games, I can change the historical return by 1% to 2%. And so the way that I think about evaluating a back test is you don't compare back tests across managers, because there's just too many different things that each manager can do that makes them uncomparable.
But when you're doing things like showing an enhancement, and this is what we generally do, is we say, "Here's the existing approach. Here's if we add something like profitability or asset growth, and that's the only thing we're changing, everything else is identical. That's the only thing that we're changing." Then you can really assess the benefit of that one change. But if there's lots of things that are different, you can't assess the benefit of those changes because there's too much stuff going on to really have an informed view of if something is better or worse by comparing the back tests.
We talked a little bit about Dimensional competitors that may attempt to replicate. Is there anything you can tell me that you have learned from a competitor?
No, that's an interesting question. I would say not so much directly from looking at the competitor or from the outside, but more from maybe hiring some folks that used to work at that company. And so whether it's a BlackRock or a Vanguard or a State Street or a Schwab or whoever, they come in with a different perspective on how to do things, and they share that perspective about what they thought worked well, what they thought didn't work well.
And I think that helps us internally and it helps us improve. It helps us adapt and say, "Oh, that's an interesting way. We hadn't thought about doing it that way. And can we change a process to improve it?" And so that's where I would say that you get some benefits. If it's looking at them from the outside, it's very hard to know what a competitor is doing from the outside. You see their materials, but unless you really work there, you don't really know. And even if you have worked there, after a few years, your information becomes so stale, you can get a sense of what they're doing, but I don't think that you can really use that information to inform your strategies, all that well.
So you mentioned Vanguard, BlackRock, State Street. And when you put yourselves in that mix, do you worry about the increasing concentration of assets in these firms and shareholder votes in the hands of you guys?
No. Massively, and maybe I'm being a bit cavalier. I'm not sure. We have looked at the research and the team has. I'm not as familiar with that research as the research team is, but the research that we've seen so far... There's been papers that have come out over about the past five or 10 years, common ownership, and has it reduced competitiveness in certain industries? And I think that the research is, let's say, it's in its infancy to be kind to it. I don't think that you can draw some strong inferences. There's probably pluses and minuses. So I'll give you some of the pluses, like when Dimensional is purchasing securities on behalf of the portfolios that we manage for clients, well, then those clients are getting professional stewardship, professional engagement.
It is more of a one size fits all engagement. And we have a viewpoint that the purpose of corporate governance is to set the firm up well to maximize shareholder value. So we look at boards, we want certain board composition, certain management compensation approaches to maximize shareholder value, but is a professional viewpoint on how to engage with those companies, to improve shareholder value. And we don't believe that reduces competitiveness between companies, because there are plenty of laws and rules that prevent companies from colluding, collaborating and reducing competitiveness between companies. So I'm not sure that it does, but I think it's an ongoing area of research. And one that I know the team keeps an eye on.
It's good to know that you guys are keeping an eye. It because like Cameron said, you're right in there with those huge asset managers that some people worry about. So Gerard, you've got a PhD in aeronautics and applied mathematics, which is pretty cool, objectively, pretty cool. But you work in asset management. What are the commonalities between aeronautics and asset management?
I would say problem solving. And when you get a PhD, you're trying to solve something new. You're trying to bring some new piece of research, even if it's tiny, better understanding to a particular problem that hasn't been done before. That's kind of the criteria. This can't be something that's been done before. It has to be new. It has to be something innovative. And in doing that, you learn a way of thinking. You get experience in a set of tools that help you develop your knowledge and improve your knowledge and asset management is no different.
You have to have a way of thinking about a problem that's rational. That goes piece by piece. You have to have a way of learning from the data from your clients, from your employees, so that you can incrementally improve all of the time. And then those tools translate over very well. It's probably a bit fortunate in that the tools... The mathematics that I do now... Actually I don't really do much maths anymore as co-CEO, along with Dave, but that I was doing when I was here was not nearly as in depth or complicated as what I did during the PhD. So I would say that the tools that I had to develop here were more around communication, around teamwork, around collaboration. And then how do you take some of those models and boil them down for real world application that can impact people here and now.
Interesting. I'd like to ask you the same question. We asked Ken French when he was on. Do you have a financial advisor yourself?
Yes, I do. And they help me tremendously, not so much with my asset allocation. I generally take care of that, but there's so many other things that are part of a financial life and financial wellbeing, and that need to be done that I find a financial advisor invaluable. So I use one and they help us out with trust and grants and all different types of things. And today is my daughter's birthday. She's seven today. And so certainly after she was born, we took a different view of things, my wife and I, and they've helped us tremendously in being prepared and knowing that if something happens to us, that she'll be okay, but they even help us with our insurance needs, with our tax filings and reporting needs. They just take all that stuff off your hands. So you don't have to worry about it and focus on at least what my comparative advantage is, which is working in for an asset management firm.
All right. I want to come back to the ICAPM and this isn't going to be polished because I haven't really thought through the question that I want to ask, but I'll try and talk through it. So what I asked about that, I think that your answer was that, in some extreme cases, investors should consider things like their labor income in deciding whether they should tilt. But in general, people can just think about whether they're comfortable with tracking error. Is that an accurate representation of what you said?
That's a good representation, Ben. Yeah.
Okay. So given that, maybe I'm just thinking too theoretically here, but if that is the case, if that's how all investors think any equilibrium, for example, why would we expect the factor premiums to persist?
In equilibrium? Well, so here's the crux of the situation. You can go in and give a presentation and you can give a presentation about value and you can show the volatility of returns. And in that same presentation, you'll have one person come up to you. How could anybody stand? Then you have another person go up to you? That deal seems too good to be true. So I think that's the crux. I don't think that you will get all people to get into perfect agreement that they're able to tolerate those deviations from the marketplace. I don't think you'll get that. And it's probably another way of saying that, "I don't think that you can get to a point where all stocks have the same expected return" and people are totally indifferent as to which stocks they hold in their portfolio, because they're all the same expected return.
I just don't think that's a plausible state of the world. I think there will always be differences. Our disagreements in tastes and preferences and views on the world that will lead some stocks to have higher expected returns sucks lower. What I do think is that if you can't tolerate the tracking error, and if you work with a financial professional, then it's not a free launch. I'm not claiming there's a free launch here. What I'm saying is that, the tools that we have to really measure true risk are so imprecise that you can't really... Like it's lifetime consumption. If you have a slight overweight to size value of profitability in your equity portfolio, that's only a small fraction of all of your wealth, your human capital and so on and so forth. It would be hard for me to suggest that with respect to lifetime consumption, you've massively increased your risk across aspect. So I'm not saying it's a free lunch. Let me be clear. I'm not saying it's a free lunch, but I think there are opportunities for those investors that can stand and tolerate tracking error relative to the market.
Hmm. Interesting. Tracking error. And I mean, like you said, it's not a free lunch, but I think it's got to be more than tracking error because someone, for example, in a... Well, you mentioned in the extreme case of someone in a deep value industry, maybe doesn't want to tilt toward value in that case, they are taking some risk for the lifetime future consumption. But I think what you're also saying is that, and this is kind of where I've landed too. I think into theory, it makes a lot of sense. Why do the premiums exist? Why is there a multifactor structure of expected returns? Because people need to hedge outside income risks and other things like that. But practically speaking, if somebody is in a value industry and also owns a value portfolio, it's probably not going to make that much of a difference to their lifetime consumption. Is that kind of where you're going with it?
That's kind of where I'm going. And I don't know, value industry. It depends on what you expect the volatility of their human capital to be, but you have flexibility in life and that flexibility can be used to deal with uncertainty. So examples would include, I could decide to retire later. So if I don't get the draw that I want is not the end of the world, I will still be able to eat because I'll decide to retire three years later. So it's that flexibility that allows you to deal with the uncertainty. That means that it's not zero one, it's not a zero one outcome. If I have a value, slight overweight to value. And I'm not saying extreme, I'm saying in a core of a type of a strategy that flexibility allow you to bear that maybe additional risk, maybe additional uncertainty to capture those higher returns over time.
Yeah. Okay. That makes a lot of sense. I'm glad we kind of hashed that out because this is something that we've talked about ICAPM on the podcast quite a bit, and we had a recent guest that talked about a bunch of empirical work that he's done on asset pricing and how it looks in the individual account level. But this is something our podcast communities have been discussing a lot lately. Like what do you actually do with that? Should people be trying to reflect their hedging needs in their portfolios? And I think, 'maybe not' is probably a pretty good answer.
If you could measure it accurately, I think it all comes down to... You remember Jochie, he retired a couple years ago, but he had a saying, which is, what is it? "You measure with a micrometer and then you cut with an axe." And so I agree with you, which is that the theory is beautiful. It makes a lot of sense. I think it really describes reality quite well, but then how do you get the information to translate that so precisely into a portfolio, given all the noise and uncertainty and unexpected outcomes that happen in the world. I think that's where there's a kind of a gap.
Hmm. Yeah. Okay. No, that was great. That was a really valuable insight to finish with.
Good. I'm glad I could provide some insights.
There are more than just that some. All right. Well, this has been great, Gerard. We really appreciate you coming on the podcast and you've given us a ton of time, which has been fantastic.
Well, Cameron, Ben, appreciate all what you do. Have great respect for your podcast, but also everything that you've done up there with Brad and so on in Canada. So thank you for having me on and hope to see you again in person sometime soon.
You bet looking forward to it. Thanks, Gerard.
Links From Today’s Episode:
Rational Reminder on iTunes — https://itunes.apple.com/ca/podcast/the-rational-reminder-podcast/id1426530582.
Rational Reminder Website — https://rationalreminder.ca/
Shop Merch — https://shop.rationalreminder.ca/
Join the Community — https://community.rationalreminder.ca/
Follow us on Twitter — https://twitter.com/RationalRemind
Follow us on Instagram — @rationalreminder
Benjamin on Twitter — https://twitter.com/benjaminwfelix
Cameron on Twitter — https://twitter.com/CameronPassmore
22 in 22 Reading Challenge — Join the Rational Reminder’s 22 in 22 reading challenge!
Ben’s Reading Code (22 in 22 Challenge): 7XWESMK
Cameron’s Reading Code (22 in 22 Challenge): N62IPTX
'Five Things I Know about Investing'— https://www.dimensional.com/us-en/insights/five-things-i-know-about-investing
'Intangibles Are Noisier than You Think' — https://www.dimensional.com/us-en/insights/intangibles-are-noisier-than-you-think
'Fundamentally, Momentum is Fundamental Momentum' — https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2572143