Episode 384: Mamdouh Medhat - A Profitability Retrospective, and Private Fund Performance

25:51So we're going to diversify across signals. Sure, it's not wrong, but there is a more efficient way to allocate your capital. There is a way to go directly to the source. Right. Once you know about profitability, that is true. Prior to that, maybe we didn't know. So maybe it did make sense to diversify. You mentioned values of recent underperformance, particularly in the US earlier. How much of that is explained by profitability? The underperformance that we've seen for value is very much a US-centric phenomenon. It has not been to the same extent negative in developed markets outside the US. And it's even been positive in emerging markets over the same period.

26:30But in the US, you saw an underperformance for value since its latest peak, which was in 2007. And so we argue that around half of that negative performance, around half of that underperformance can be explained by profitability, or at least can be linked to profitability. We're not sort of arguing for a causation here, right? This is all empirical asset pricing is almost always about correlations. The thing is that what you had was that it was a period of strong underperformance for the standard value factor, but it was coupled with a period of strong outperformance for the profitability factor during that same period of time.

27:11And they were strongly negatively correlated. What we argue is that by just accounting for the negative correlation that standard value had to profitability, it accounts for a headwind of about half of that underperformance that you see over that sample. The remainder, so the remaining half of the underperformance, we provide some evidence that it has something to do with industry tilts. The standard value factor tends to take quite a bit of industry tilts beyond sort of being short profitability. it's also betting on some industries and so we show in the paper that if you construct a standard value strategy but you put in some sensible controls for that negative that short profitability exposure and also you try and rein in a bit its industry tilts then you actually get something that looks a lot more reasonable it has lower volatility and it actually ends up having a small but positive average return since 2007.

28:11It's not significant, but it is small and positive, which is a long way to go from the very negative one that you see for just unadjusted value. So if you go and build a standard academic value portfolio, it's going to be tilted away from profitability. It's going to have some big industry tilts. And if you go and adjust for those things or take those things away, take the profitability, negative profitability, tilt away by adding profitability exposure to the value portfolio and constraining the industries, then all of a sudden that negative gap over that period shrinks is what I'm hearing. Yes.

28:42The negative profitability exposure, we understand why. Value stocks, they tend to be small and unprofitable. Growth stocks, they tend to be large and highly profitable. So there's kind of a mechanical shorting of profitability in just the pure sort of academic value strategy like that. But when you also look at it, you tend to see that certain industries show up disproportionately often on the value side and on the growth side of a standard value strategy, right? So on the value side, you tend to have things like energy and financials. On the growth side, you tend to have, you know, tech and healthcare and things like that.

29:19What we show in the paper is that those industry tilts have driven more of value's volatility than it has driven its average returns. So it accounts for a big portion of the volatility of the value premium, but over the long sample, it's actually been flat in terms of adding anything to its return. So if you manage those industry bets a bit in the construction of your value strategy, you tend to get something that is less volatile and also performs better, especially when you also control for profitability. There isn't a one-to-one mapping between what you do in the sort of academic factor space and what you do in practice, but it's generally speaking what we try and do in dimensional with our value strategies.

30:02We also consider profitability when we construct value strategies and we also manage the industry and sector tilt so they don't sort of run haywire and end up driving a lot of uncompensated risk, if you will. I think you mentioned earlier that profitability has sort of a strong theoretical story. Can you talk about at a fundamental theoretical level, what makes the combination of traditional value like price book and profitability a sensible combination for explaining differences in returns? To me, the fundamental story really starts with this basic rule that we have in asset pricing that price is equal expected discounted cash flows.

30:40The price of any investment today has to be my expectation of what I will get out of it in terms of cash flows. And then I discounted back to present value using some discount factor, some notion of the cost of capital, if you will, that reflects the riskiness that reflects my expected return. If I observe a low price, then that can mean two things. It can mean that the discount factor is high, which is sort of what I want. That's a high expected return. But it could also mean that the expected cash flows are low. So prices can be low because of low expected cash flows or because of high discount rates.

31:17I'm mostly interested in the latter. And so by controlling for profitability, what you end up with is making the price signal more informative. Because now if I compare two stocks that have the same profitability, that's my proxy for the expected cash flow channel. And then I know that the one with the lower price, oh, that's because of the discount rate channel, right? So now I've made the price signal more informative. Coming to what we actually do in practice, book to market and profitability work really well together because book to market is the cleanest pure valuation measure that you can get.

31:53It has very little to do with profitability. So book equity, yes, there is a mechanical link to pass earnings and things like that, but very little in terms of the actual profitability channel. And so it becomes a very clean valuation metric. Profitability, on the other hand, because we avoid scaling it by price or something like that, it becomes a very clean expected cash flow channel. And so you have two measures that are cleanly separating these two effects. And so they work really well. And beyond sort of, you know, the academic jibber jabber in practice, price to book and profitability work really well because they have hardly any overlap.

32:31Price to book tends to go for these small companies that don't have a lot of profits. Operating profitability tends to go for large companies that have growth-like valuations. And so they sit on opposite sides of the market. So that means that very little overlap, lots of diversification, lots of flexibility to try and blend them together in a sensible way. In practice, they also work really well. Man, that makes a lot of sense. So it's like instead of taking one of the other value metrics we talked about earlier that has some blend of value and profitability in one metric. These are like two very distinct ways to measure the two different things.

33:06And then you can blend them together however you want, but you can be precise about it because there's such distinct metrics. Our portfolio managers here, they tend to talk about that they give them a lot of control. There's little overlap. And so they can clearly see why they're holding this particular stock due to the valuation measure and why they're holding this particular stock due to the profitability measure. And so that's really nice in constructing portfolios and managing them. So it matters in practice as well. What are the main academic implications of the findings documented in this paper?

33:36I think the paper for academics, first of all, it provides a unifying framework that connects a lot of seemingly unrelated things. It connects a lot of these different quality notions to basically just profitability. It connects the low risk anomaly, the defensive equity strategies, it connects it also to profitability. It cleans up a little bit in the factor zoo and it provides a unifying framework for understanding many of these things. That's kind of empirically, but theoretically, it also has some implications because it basically says that you really need to think about value and profitability together, even at a theoretical level.

34:17So lots of academics, they will write down theoretical models for why we should, for instance, see a value premium, right? This is one mechanism that will lead to a value premium. It could be related to capital investments, or it could be related to macroeconomic risks or some combination. They're trying to explain why you should see a value premium and derive some testable predictions from that. I think our paper is pointing towards that that's no longer enough. A model has to be able to produce a positive value premium, yes, but it should probably also be able to produce a positive profitability premium.

34:49And really importantly, those two should be negatively correlated within the model. If the model cannot produce those sort of empirical facts, then it's missing something. And then that means that whatever you're getting out of it might be incomplete. Listening to this are many practitioners who I'm sure are wondering, what are the main insights for them? I would say that I think this shows that parsimony goes a long way. You can be parsimonious about the way that you think about quality investing. You don't need a whole list of different quality metrics. Profitability really is the one that tends to be the strongest and tends to explain them all.

35:24And you don't need a separate consideration for defensive equity. That's what we talked about before. So again, it shrinks kind of what you need to think about. It brings it down to something more of a unifying framework. Many investors that I talk to, they're constrained in, let's say, how much tracking error they can take or how much active share they're willing to take in their asset allocations, how much they can deviate from the market. And so they want to think about how to efficiently allocate, how to efficiently deviate from market weights. And you just don't get a lot more efficient than just by having a clean cash flow channel, which is profitability and a clean valuation channel, which is price to book, I would say.

36:01I'm in a lot of conversations right now with investors who are maybe looking at the size of the US and global markets and thinking, should I do something differently. If I am going to hold the US at market weight and given those really high valuations at the top and given that concentration that we're seeing, maybe I should go for quality, right? Maybe I should at least get some quality for those high valuations. Well, I think our paper sort of points towards, okay, first of all, how should you look at valuations and how should you try and identify quality for that valuation? And so this is why I think profitability right now is very appealing, especially when you combine it with price to book.

36:38We've been talking about profitability. You mentioned it once, I think. We haven't talked much about how you actually measure profitability. Can you talk about what metric Dimensional uses to implement profitability in portfolios and maybe also why that metric instead of some of the alternatives? I've kind of purposefully not defined what I mean by profitability, and that's because it's surprisingly robust. You get by and large the same results depending on irrespective of which profitability measure we use. But we do quite a bit of work to try and identify which one we should use. We've done that here at the firm and we also continue to do that in the paper.

37:11So the one that we use in the firm and in the paper is operating profitability, which is you start with Robert's gross profitability from 2013, right? So revenues minus cost of goods sold. Then you subtract SG &A and you also subtract interest expense and you scale that by book equity. The reason why you have these two extra elements, SG &A and interest expense, is because those two items, they account for differences in operating expenses across industries, and they also account for differences in leverage across industries. And those two things are important. It just makes the profitability measure comparable and easier to interpret.

37:51In the literature, and this is something we review in the paper, there's talk about sort of, you know, one particular alternative to operating profitability, which is called cash profitability. And the idea is you take operating profitability, just like I was talking about it before, but then you subtract one additional item, which is accruals. And so accounting accruals are sort of supposed to, it's an accounting tool to align the timing of expenses with their revenues. So you can book the expenses when the revenues come in, you can book the revenues when the expenses are occurred. It basically means your profitability metric becomes comparable across firms and over time.

38:26But when you subtract that out, you get down to the actual cash that the company has at hand. What are the actual cash flows that the company has? And so we compare those in the paper to try and figure out which one corresponds better to theory. We think that's probably the best way to differentiate between them. And theory basically calls for a profitability measure that predicts future cash flows. This is why we use current profitability, right? We're not interested in the past in terms of what the company has done. We're interested in what is it expected to deliver in terms of cash flows in the future.

38:59And so you should really think about the profitability measure that is the best predictor of future cash flows. Operating profitability, we show in the paper that it predicts its own future growth. So companies with high operating profitability today tend to grow faster in terms of operating profitability in the future, but it also predicts the growth in earnings. So they also tend to grow their earnings faster in the future, which we think is a desirable property for a profitability measure. Cash-based operating profitability is a lot less persistent. So it actually exhibits mean reversion in the sense that companies that look high on cash profitability today, they actually tend to grow negatively or sort of, you know, stagnate in terms of their cash profitability going forward.

39:43And it doesn't predict the growth in earnings or anything like that. This is true in the U.S. It's true outside the U.S. and developed and in emerging markets over the shorter samples as well. And we're talking about earnings growth over long periods of time, five years, seven years, things like that. So we think that's a pretty compelling argument for sticking with operating profitability. So not a huge, huge difference, but theoretically you want something that's persistent and operating tends to be better than cash-based for that purpose. I would say so, yes. And also one thing that we've observed in the data is that the accruals effect has gotten weaker over time.

40:20And so that's the only difference between them, really. the accruals have generally speaking the accruals effect is very much a u.s phenomenon and it's gotten a lot weaker post-2000 there's no theoretical kind of argument for why it should get weaker but that's kind of what we see in the data and so the difference between them has also sort of gone down in terms of their ability to predict returns and the accruals effect doesn't really seem to be adding that much over the last 20-25 years you mentioned earlier you're just talking about the idea of people worrying about their u.s equity allocations right now based on where valuations are and the size of the US market and all that stuff.

40:55You've got a paper that looks at that concept. Can you talk about the downsides of tilting toward or away from a country based on its aggregate characteristics, like in the US example, tilting away from the US because it looks expensive? A bit tongue in cheek, we call the paper few and far between. And that's both because when you start sorting sort of countries or industries on these factor characteristics, they're few and far between, but also the benefits are few and far between. You don't actually get that much from it. So the premiums have historically not been driven by countries. They've been driven by stocks within countries.

41:31You know, when you sort countries on something like their valuation or their size or their profitability, you tend to see that the spreads kind of go in the right direction. They have the right signs, right? You expect to see something like a size premium, a value premium, a profitability premium, maybe even a momentum premium, but they tend to be quite weak and they tend to produce unreliable spreads and returns and no alphas relative to the factors when you construct them from individual stocks. So this is why we don't really see much of a benefit there. The intuition is that there is some informational loss, right?

42:08There's a lot of cross-sectional variation in these characteristics. And so by sort of aggregating up into these little components that are sort of, you know, countries, you lose a lot of that variation. And so you don't really get much of a benefit in terms of trying over the long term, at least trying to sort of tilt away from countries or industries based on these characteristics. There might be other reasons for, you know, tilting away from certain countries. You might think about a home bias. I live in Europe. I might want a home bias in my portfolio for other reasons, but it would not be driven by, let's say that I'm expecting a premium for that because Europe is on average smaller or cheaper than the US.

42:48I just want to dive into industries a bit more. I know you mentioned it, but what about over or underweighting industries based on their aggregate characteristics? The story is relatively consistent. You tend to see that these industries don't really drive the premiums. Again, the spreads kind of go in the right direction, but they tend to be fairly weak. The only nuance to that is when you look at industry momentum. There's a famous JF paper by Moskowitz and Grimblad from 1999 that looks at industry momentum and shows that it's actually strongest at the one-month horizon, which is kind of interesting because for individual stocks, at the one-month horizon, you tend to see reversals.

43:28I've actually written a paper with Robert and some of my colleagues here at Dimensional on that. But at the industry level, you actually tend to see momentum. And that tends to be quite strong, and it is significant, and it does produce an alpha, but it just has four to 500 % turnover per year. So it's very difficult to capture sort of in your standard practical portfolio. And so this is why we kind of reached somewhat the same conclusion with the footnote that industry momentum exists and tends to be quite strong in the US and outside. Other than the industry momentum piece and the previously mentioned results, we were saying that tilting toward industries and countries doesn't have much of an effect.

44:09How important to that finding is it that the investor is tilting at the security level? That is really the nail in the coffin in terms of trying to pursue these premiums at the country level or at the industry level. Because if you weren't aware of how these premiums behave at the individual stock level, you would again sort of be just looking at sorting entities on characteristics that should be related to expected returns. And then maybe that's the only thing you can do. But again, that ignores that we know that these factors work really well at the stock level. And so once you know that, once you look at whether you're adding anything by taking country bets and industry bets on top of having the security level exposures, then you see that these country bets and industry bets don't add anything.

44:54They don't have any alpha. They're spanned by the security level factors, if you sort of want to talk in the wonky academic way. One thing I really do want to mention also is that once you start aggregating individual stocks up into specific countries and industries, you have fewer building blocks to play around with. And so that means that it's quite difficult to take into account premium interactions. So imagine just doing a simple two by two, two groups on price to book, two groups on profitability. Over that grid, you would want to emphasize the ones that have a low valuation but a high profitability, and you would want to underweight, de-emphasize the ones that have a high valuation and no profitability.

45:37Just a simple two-by-two grid can lead to very empty and under-diversified portfolios if you do that at the country level, and something very similar also happens at the industry level. So there are a lot of drawbacks to sort of going away from just doing things at the security level. If you, for some reason, couldn't tilt at the individual security level, maybe there's something there trying to tilt at the country or industry level, but it's a lot noisier. You've probably got more idiosyncratic risk. And because we can tilt at the individual security level, it's like there's really no point in trying to tilt at the country or industry level.

46:09Completely agree, yes. So how does dimensional deal with country and industry weights? Based on that research, but also a lot of other work that we've done, we generally tend to sort within countries, right? So we go within each country and then we sort on the characteristics that we care about, size, relative price, profitability. And then we aggregate up across countries and we tend to generally hold countries at their market weight. Because these borders that we've drawn in the sand don't really tend to be something that differentiates high from low expected returns. So we tend to hold things at market weight.

46:46There's also that whole discussion around, are some of these big companies, are they truly sort of based in one country or are they really multinationals or they're getting sort of, you know, their revenues from many regions and so on. The thing about doing things at the individual country level and then aggregating up is that you will naturally get some industry bets from that. So tilting towards value, as we talked about before, you tend to get sort of more exposure in energy and financial tilting towards profitability. you get more exposure in tech and healthcare, things like that. So you will naturally in the pursuit of the premiums get some industry bets.

47:20If you try and neutralize those completely, then it tends to hamper your ability to capture these premiums because they want to go into a little bit of deviation at the sector level. And having a completely sector neutral portfolio often requires a lot of turnover, a lot of trading, which is again, inefficient use of capital. On the other hand, letting them go without constraining the industry deviations at all will create some of that volatility and some of those concentrations that we talked about before. So we balance that trade-off by introducing sector caps. So we limit the amount that you can deviate from the market weight in that given sector when you are pursuing these premiums, right?

48:03And that can vary a bit from strategy to strategy, but we think that's a good way of sort of having enough flexibility to capture the premiums, but not introducing concentration at the sector level. Last few questions here on private markets. You've got a really interesting paper looking at that. Can you talk about how much of the global investable universe in stocks, bonds, and real estate consists of private funds? Yes. So we wrote a paper last year. We call it understanding private fund performance because that's basically what we tried to do. We were getting a lot of questions about private markets and about private funds.

48:39And we like to look at the data. And so we got some data on private market vintages going back to the, I think the earliest start is in the eighties for VCs and then buyouts, then private credit and private real estate somewhere in the mid nineties. We have about 6 ,000 funds in the sample. So it's a pretty sizable data set. And we try to understand different things that might be of interest to investors. But one thing that we did look at sort of as a snapshot at the end of the sample was what's the size of the investable universe here? And there's a lot of discussion around that. What we found is that about 10 % of the total global investable universe is in alternative funds.

49:21So that includes private funds and hedge funds. You often hear a much larger number. You often hear that private markets are much larger than public markets. But the thing is that not everything that is unlisted is investable. You cannot go and buy shares in every single private company. You can't take a stake in a private infrastructure or a private real estate project. And so we really have to think about what's the AUM? What is actually investable today? Where can you go out and get some exposure to this in a fund? And there it turns out that it's about 10 % of the total investable universe.

49:55The remainder is about 45 % in global equities and REITs and 45 % in global fixed income. And so that kind of sort of market breakdown, we think is a pretty decent starting point. If you are actually going to venture out into private investing, think about those market weights when you're allocating. We definitely see institutions and other advisors with like 25, 30 % on alternatives. Yeah. For what reason? And so if you wanted to say, well, hey, I'm going to build a market cap weighted exposure to private asset classes, it's going to be like 10%, not 30. Something in that ballpark, yes. That's at least how it looked at the end of last year, I think, when we last updated those numbers.

50:35And it's been fairly stable. It hasn't moved around that much. We've even seen some this week, Ben, where they're 100 % in private assets because public markets aren't working anymore. It's incredible. So how important is benchmark choice when analyzing the performance of private funds? I would say that it's crucial. Based on the results in the paper, your choice of benchmark can completely change your conclusion. We often see private funds being benchmarked to style neutral indices, sort of market wide indices. You can, for instance, think about US buyout funds. Their favorite benchmark is the S &P 500, which granted has performed exceptionally well over the recent past.

51:19But generally speaking, you don't tend to see that US buyouts look like the S &P 500. They are not this broad, kind of relatively style neutral type of investment. They tend to take style exposure. They tend to take certain sector deviations compared to the market. And those tend to not be constant from vintage to vintage. They tend to vary. And so I think you get an incomplete picture of opportunity cost if you just benchmark to something that is market wide. The same is true in VCs, in private credit and in private real estate. When you benchmark against a style appropriate benchmark, how do private funds tend to perform?

51:59This is where you need some specialized methods to be able to actually talk about relative performance in the private fund space. An internal rate of return, an IRR, is not just comparable to an annualized compound return. I know that you've had Ludovic Falipu on the podcast. He's brilliant. And this is one of his pet peeves that he writes a lot about in his papers and in his book, all the issues with IRRs. The main one sort of right now, at least for me, is that it's not comparable to sort of the annualized compound return of any index. And so you need some specialized methodology. The consensus in the literature is to look at something called the Kaplan-Schwar public market equivalent.

52:37It's named after this really phenomenal paper in the JF by Steve Kaplan and Etienne Etchouar from 2005. And it's a money multiple. It's a benchmark adjusted money multiple. It has several interpretations. You can sort of think about it as in you finance your capital calls by selling shares in the index or the benchmark, and you reinvest any distributions back into the benchmark. That's one way to form. You form a money multiple of that, then you get the KSPME. Another interpretation is that it is the present value of distributions and NAV relative to the present value of contributions, where the present value is calculated by using the benchmark's cumulative return as the discount rate.

53:24And that last interpretation is really important because it links it to valuation theory, and it means that the KSPME is a proper risk adjusted performance measure. And it can also not be fooled by leverage. This is often one of the concerns with IRRs is that if you use leverage to pay out early, then you can sort of, you know, get a higher IRR early on. And because long out into the future, the cash flows matter less and less for the IRR, you can basically have a high IRR by paying out early. So there are some ways to fool the IRR that's not possible with the KSPME. And so this is the measure that we think of and that we think that people should use.

54:04And so by using that, what you see is on average, buyouts and VCs, they look like they outperform the S &P 500. They have a KSPME above one on average, you know, going over the long sample relative to the S &P 500. But when you calculate that same KSPME relative to a small cap value index, or even a small cap growth index that is tilted towards the highly profitable small cap growth ones, then you tend to see that they on average perform just about in line with that. So your choice of benchmark can really change your conclusion around relative performance. For private credit, you see that again, according to the KSPME, they on average outperform the Bloomberg US credit index, but they perform just about in line with the Bloomberg US high yield index.

54:56So on average, you've gotten the same return by being in US high yield bonds. Private real estate tends to actually underperform the sector portfolio of listed real estate companies. So this is not REITs, but listed real estate companies, real estate operating companies, and they tend to underperform REITs based on the KSPME. So again, you can get wildly different conclusions depending on the benchmark. What I want to stress here is that this is for the average fund. There's lots of dispersion around that, but for the average fund, you can basically get the same performance by going into the public market.

55:32That's a pretty wild finding, isn't it? We find it quite striking as well. We've had people like Ludo that you mentioned on the podcast. It's not that surprising, I guess, because we've seen research like this before, but you guys looked at more asset classes, more data, I think, than a lot of other papers have looked at. That's pretty wild. Yeah, we're quite happy with that paper. And it's about a year old, if not a little bit more. And we're still presenting it at conferences and for a lot of institutional investors. Yeah, it's been really useful sort of digging into the private markets, even though we don't as a firm invest in that.

56:05So what do private assets look like through the lens of public market factor exposures? Yeah, I mean, this is kind of a natural analysis to do, right? Whenever you talk to anybody that likes to play with systematic factors or asset pricing models, then you want to figure out what's the exposure, right? When you look at private funds and you look at their periodic returns, often you can only get them quarterly. And so the quarterly private fund returns, they tend to be a function of the NAV, of the NAV. And the NAV is an accounting appraisal of the ongoing investments, the ones that haven't actually been exited yet.

56:42And so because it's based on an accounting appraisal, they tend to not change that much and they tend to not change that often. And so what that leads to is lagging and famously smoothing. So periodic private fund returns, they have an element of lagging and smoothing. And so that means that you need to adjust for that whenever you're measuring factor exposures, because they can screw up your factor regressions. They can screw up your factor exposures. By the way, this is no different from when people started looking at real estate returns back in the 1970s. We have methodology to do that. It's really not rocket science, but it's important to do that.

57:21Another caveat and important fact to mention is that we saw a structural shift in 2007 because there was a change in the accounting rules. And in particular, there was the adoption of fair value accounting. This happened in Europe and it also happened in the US. And fair value accounting basically means that you as a private fund manager cannot get your books audited unless you can show that there's some mark-to-market accounting. You have to show somehow that you're marking your current investments to deals that have already happened or comparable investments from the public market. There needs to be something like that.

57:56And so you asked me about what do the factor exposures of the private funds look like? Well, pre-2007, pre-fair value accounting, you mostly have negligible factor exposures, even when you adjust for lagging and smoothing. There's very little explained variation. After the adoption of fair value accounting, you see that the factor exposures of these private funds in aggregate can explain up to 70 or 80 % of the variation. It's really a material sort of, you know, shift that we saw after the introduction of fair value accounting. And so if you look at the different asset classes, buyouts in aggregate, they mostly look like highly profitable companies, perhaps in the mid-cap space, something like that, which is kind of intuitive.

58:40They tend to go for sort of your cash cow type of investment. And in aggregate, they tend to look like that. And that explains north of 70 % of the variation in quarterly buyout returns. VCs in aggregate, they tend to look like small cap growth, but high profitability. So not these short end of the defensive equity strategies, they tend to be small growth, but high profitability. Private credit tends to look like high yield and they tend to not take material term exposure. So they're not really differing that much on whether they take long duration, short duration. They tend to look fairly market-like in their duration, but have high yield.

59:18And private real estate, it does correlate with REITs and it also does correlate in aggregate with the public equity market, but not very much. Private real estate is the asset class that we have a hardest time explaining by public market factors. It's the one that looks the most different, I would say, from the public market factors. Perhaps this is why it is somewhat popular as the quote-unquote diversifier. Do you have any idea why that is? I just think the composition of what you have in private real estate funds is quite different from REITs in the sense that a lot of private real estate funds, they will get these real estate investments and just basically get the capital from them by renting them out.

1:00:00Let's say it's slightly different from what you see in real estate investment trusts. And so that basically means that while you have some connection, because they're part of the same economy and they're largely speaking, investing in real estate, it's very different assets that they're actually holding. So we've been talking about average fund returns, which has been super interesting. So they've got PMEs, pretty close to style appropriate public market benchmarks, which is super interesting, but does make sense. And then we have public market factor exposures that will largely explain returns, aggregate returns, post-2007 once we have fair value accounting.

1:00:34But again, those are average funds. In public markets, there's not huge dispersion between the worst and best cohorts of mutual funds, for example. Can you talk about how much dispersion there is in private market fund performance? I think there's a little bit of a wrinkle to what you just said around public sort of funds. Let me get to that in the end. But what we see in the paper for private funds is that there is actually quite a lot of dispersion. So for instance, the top 5 % in any of the four asset classes that we consider, they often double or triple your invested capital after management fees and carry.

1:01:09So the top 5 % of funds there, they really have quite impressive returns. The bottom 5%, on the other hand, they tend to lose half of your money after fees, maybe even more than that. So quite a lot of dispersion between the top performers and the bottom performers. And we see that in all the asset classes, it is sort of the dispersion is widest in VCs and it's lowest in private credit, which is kind of what you would expect. That's sort of quite intuitive. It's quite difficult to compare that to what you see in public markets because of survivorship bias. So in public markets, the funds that live for a considerable period of time tend to be the ones that haven't closed down, kind of a tautology.

1:01:50But funds close down because of bad performance, because of outflows, the funds that do badly, they tend to not survive over long periods of time. I think over the past 20 years, when you look at U.S. equity funds, so US domiciled equity funds, the survival rate, I think, is only like 45%. And the outperformance rate is close to 15 % or something like that. That survivorship bias means that it's difficult to compare that dispersion because private funds, they tend to be, you know, five, seven, 10, maybe even longer now year investments. If you look at the dispersion that you see in returns over relatively short periods of time where you have less of that survivorship issue in public markets, you tend to see as much dispersion in the private space as you do in the public space.

1:02:37That is a really interesting point. There's a chart that I've seen that shows the increasing dispersion when you go from public to private markets. And I always just took that as, yeah, that makes sense. I always accepted that as true. But what you just said also makes a ton of sense that when you account for survivorship bias, you can actually have a lot more dispersion than it would look like if you just look at the sample of existing funds. And so you're saying that there's as much dispersion in public market fund performance as there is in private market fund performance. When you look at active public funds, yes, I would say so.

1:03:06That would be my expectation. There is quite a bit of dispersion. We recently were asked to do an analysis where we would look at exactly the dispersion in information ratios, right? So sort of, you know, outperformance relative to tracking error for active funds and quite a bit of dispersion just over the last five and 10 years. It's a pretty common phenomenon that when you think about active management, traditional active management, which private funds are just in the illiquid space, you tend to see a lot of dispersion between top and bottom performing managers. And the obvious next question, Mamdu, given this wide dispersion, what are the criteria to predict the successful managers?

1:03:48manager selection is always difficult the thing is that what people tend to think about is persistence so they tend to think about okay i i'm gonna go for the gps that had a stellar vintage because i expect them to have a stellar vintage next time as well the thing is that there isn't a lot of evidence of persistence in the literature lots of caveats around data here the most complete data source that academics use is the data that was formerly sort of managed by a company called Burgess, and now it's been acquired by MSCI. That's the data that we use as well. And using that data set, there really isn't a lot of evidence of dispersion.

1:04:33And I'm particularly thinking of a particular paper from 2023 by Harris, Jenkinson, Kaplan, and Stuckey. It's called Has Persistence Persisted? And they basically find that there is no evidence of persistence in buyouts. There's a little bit of evidence of persistence in VCs, but that's kind of pre-2000. Afterwards, you don't see a lot of that. We couldn't study persistence ourselves in this paper because of data issues. We don't actually have access to the individual funds or the GPs. We only have access to vintage level data, and so we couldn't study it. But using the same data set and using fairly rigorous methodology, very rigorous actually, I would say that there isn't much evidence of that.

1:05:13So given that you don't have persistence and given that there's wide dispersion, just like we talked about, you're basically back to due diligence. That should be the main criteria for selecting managers. I co-authored this paper with Wei Dai, who you also had on the podcast called Systematically Evaluating Systematic Managers, where we came up with this framework for what criteria should you look at before you hire a manager and what criteria should you look at after hiring a manager. And these things extend beyond just systematic managers. For any kind of manager, really, before hiring them, you should look at their full investment proposition.

1:05:52What kind of research do they do? How do they construct their portfolios? What is the process of going from idea generation to actual implementation? And then when you look at the track record, don't only focus on the track record maybe of the funds you're interested in or the style you're interested in. Look at their full track record. Look at their survival in particular, Look at what is the percentage of funds they have that have outperformed their benchmark. Because if you pick that one fund that works well and the other ones don't, then that's not really sort of, you know, giving you a good indication for being sustainable.

1:06:25And then, you know, after you hire them, it's of course the continuous monitoring. You have to keep assessing, are they doing what they should be doing? Are they sticking to the investment proposition or things in line with expectations? That holds in public markets just as much as it does in private markets. But in private markets, of course, you have additional things like the illiquidity and they don't report as often and things like that. And then you have to keep revisiting it. I do some volunteering on an investment committee and investment committee meetings are always about three-year performance.

1:06:57And if things aren't going well, then you're reevaluating the manager. And I think that's pretty typical. Yeah, it's tough. When you understand how much volatility there is in returns and how likely you are to see negative premiums over a three-year period. It really puts things in perspective. I always like to highlight Fama and French's volatility lessons paper from the FAJ from a few years back. That's a good read for anybody thinking that three years is enough to evaluate any kind of performance. It can be tough. And I understand manager selection is difficult, but you really have to put things in perspective.

1:07:33One last thing around manager selection in private markets, It's always difficult to do manager selection, but the illiquidity and the way that private markets are set up actually exacerbates the manager selection problems because you can't hold the market. You can't make your asset allocation sort of relatively manager agnostic. You can always do that in public markets, right? You can always go for something more systematic, something less dependent on which manager you choose. That's not possible in private markets. you have to pick a GP or some GPs. And so your opportunity cost can be a lot higher.

1:08:07Which makes going into private markets in the first place, a big decision. You're forced to make that second decision about which manager to pick, which is a very difficult decision to make. How has private fund performance relative to public benchmarks, to appropriate public benchmarks, like we talked about earlier, how has that trended over time? It's a good question. Again, because these vintages, they tend to last for anywhere between five or even seen like 12 year vintages. Sometimes it can be difficult to discern a time trend. I don't think that there is a significant time trend that has been trending sort of up or down.

1:08:39But what we have seen very clearly is higher correlations with public markets, especially after 2007. There's just a lot more alignment in terms of the returns that you get in aggregate from private funds and the returns that you get from the public market. And I think that's important to take into account when you're thinking about an allocation going forward. I love this next question. It comes up all the time. How much of the apparent diversification benefits of private assets is just an artifact of illiquid asset return smoothing? Based on what we can see around that structural break in 2007, pre-2007, a lot of the diversification you know, looked like it was overstated.

1:09:23It was because of the lack of fair value accounting across sort of a wide spectrum of these funds, the lagging and the smoothing really looked like, oh, these private asset classes are not entirely uncorrelated, but almost close to that when you compared them to public markets. This was pre-2007. After 2007, and especially when you account for lagging and smoothing, we see these really substantially higher correlations. And it allows us to, I think, more reliably quantify the diversification benefits. Yes, the sample shrinks and you also have the financial crisis as part of that sample, which tends to induce correlations by itself.

1:10:03We do some robustness in the paper where you throw out the financial crisis and get similar results, but we can kind of quantify it. This is where we find that, for instance, for buyout, you can explain up to 80 % of their returns just by their factor exposures. What's important here is that we can't explain 100 % of that. And we do some statistical testing to see if that explained variation includes 100 % as part of the set of values that you can trust. In most cases, we actually reject that you can explain 100%. And so that means that there are some diversification benefits. They're probably just not as high as you may have heard by looking at some of the marketing material.

1:10:45And especially if you look at sort of, you know, old vintages and old funds. The reason why I think we won't be able to fully explain the returns that come here, even after fair value accounting, and even after controlling for lagging and smoothing, is that private assets at a conceptual level, they should expand the opportunity set that you have as a public investor, because you're adding things that are not available in the public market. They might be highly correlated with what you see in the public market, but they're actually not there. So sort of, you know, in a conceptual framework, if you think about the total market, which includes every single asset, you are actually getting slightly closer to that.

1:11:25But again, there are lots of caveats around that. And I don't want to overstate the diversification benefits given the high correlations we see in aggregate. Yeah, it makes sense. I think when we had Ludovic Fallopo on, if I remember the example correctly, he said like tech in Europe, if you really want to get a European tech exposure, you kind of have to go into private markets. That may not be exactly the example, but conceptually, so it does make sense. If you're adding sector exposures that you can't really get in public markets, then you would expect some diversification benefit. And lots of institutional investors make that exact argument, right?

1:11:57That some exposures that they're looking for, they believe looking at the data, looking at what's available, that they can better get it through unlisted assets. And they might have an easier time accessing that than your typical retail investor or your advisor. Yeah, it's getting easier for retail investors, which may not be a good thing. And actually on that point, what effect do you think the increasing popularity of private investments and maybe also the increasing access to retail investors might have on their expected returns? This is one of the questions that I get asked. It's tough, but I can say this.

1:12:29When you bid up the price of anything, then all else equal, you should expect its returns to go down. And so we've seen the democratization, if you want to call it that, of private investing. We've seen greater demand for private funds, whether it's in retirement sort of solutions, whether it's an institutional portfolio, whether it's in retail portfolio sometimes. And so that's removed some of the exclusivity that we see around private investment. And it's also put some pressures on GPs to deliver despite their much higher AUM. We've kind of seen indications, at least in the financial press recently, of vintages prolonging their lifetimes, exits being difficult to come by.

1:13:12We've seen continuation funds where, you know, an unlisted asset gets passed on to the next vintage and maybe even the next vintage again. So there are some indications here that the asset class is trying to cope with this increased demand. Ultimately, I'm a data guy. I don't sort of forecast, but I would say just take that into account. You're not getting the same level of exclusivity that you tended to get before in private markets. There's a lot more demand. There's many more assets under management. So just be careful with that. Again, think about the market weight of that in your overall portfolio, if you're going to allocate.

1:13:49You did this paper. You obviously spent a lot of time crunching pretty cool sounding data set that you had. Sounds like you get a lot of questions about private markets too. What would you say are the implications of your findings for someone considering an allocation to private assets? I try to stay away from being prescriptive on asset allocation, right? Because I think that that depends on so many variables, your risk tolerance, your horizon, and what you actually want to achieve from this portfolio and what else you're holding. But the findings indicate that you should really have this middle of the road view of private assets.

1:14:21They're neither a fantastic unicorn, but it's not just, you know, complete bogus. All of it. There are some positives and some negatives. The negatives, I mean, we talked about the wide dispersion that you see in performance between the top and the bottom. And because you have to allocate to managers, you know, you are exposed to that dispersion. It's difficult to get away from that. It's difficult to make your allocation manager agnostic. Even sort of, you know, secondary funds and so on will only be able to capture that many GPs and that many individual funds. And there's also another negative, which is, you know, on average, you don't actually see a premium relative to some of these styles.

1:15:01The average fund performs, you know, in line with some of these styles that we know from public markets. So the positives, there is some evidence of diversification benefits. I can't get away from that. But there are lots of caveats around, you know, the types of fees that you're paying and then the GP that you select. The GP that you select might just be very highly correlated with public markets. And you don't know that extent. I would say if you're willing to stomach those caveats that I just mentioned, and you want to expand your opportunity set, fine. And you're willing to stomach the fees and the liquidity, and you might not have the liquidity exactly when you need it and all these things, then fine.

1:15:36But I would say, start with the market weight and be very clear about your expected liquidity demands and what your tolerance is for some of these investments, maybe, you know, continuing beyond what you expect, because we are seeing longer and longer vintages right now in private funds. There's a huge idiosyncratic risk component too, right? With the manager selection, unless you have a huge amount of capital, like you're not going to get private equity or private credit or whatever, you're going to get a GP. Yes, that's it. And there is talk in the literature as well around basically to get diversification in the private space, you need two types of diversification.

1:16:14You need diversification within vintages. So this is across managers, and that can be very difficult to get, especially as a retail investor or as an advisor, because they often have minimum capital requirements to be able to be part of a fund. And so spreading that out across many different GPs can be difficult. But you also need diversification across vintages because one vintage might just be a bad apple. If you need that, then you need to keep allocating year after year to vintage after vintage and to multiple managers. And so you very quickly run out of capital. Or you do a continuation fund, which might have its own set of issues.

1:16:53There are, again, idiosyncratic issues with continuation funds as well. You don't really get away from some of the liquidity constraints and some of the redemption limits and things like that. And you might get adverse selection. I don't know, maybe. Or with the evergreen funds, I mean, the perpetual funds. Yeah. Again, this is sort of the lack of exclusivity is that what is being offered right now to retail investors. Again, not to say that I know exactly what's going to happen or I have looked at the individual managers, but these are all things that you need to take into account. Manager selection is difficult, irrespective of where you go.

1:17:23Very interesting. Very cool look at the data. And like I said earlier, it's interesting how you guys looked at all the different private asset classes, not just one, which a lot of papers have done, but then taking the factor lens to private assets. Super, super interesting. This has been an incredible conversation, Mamadou. We have one more question for you. How do you define success in your life? Venturing into philosophy now. I don't want to sound too cringe, but it's probably something like doing something I love professionally, but at the same time being able to provide for myself and my family.

1:17:56I think marrying those two things together is probably some definition of success for me. I love doing research. I did it as an academic before coming to dimensional. And now I feel really privileged being able to do it at a firm that sort of appreciates that level of rigor and sort of, you know, the deep dives into the data and the theory. Compared to academia, it's great to be able to see some impact in the real world. One more aspect of success for me is to try and stay grounded and try and maintain some humility. I think in asset management in particular, there is a little bit of a tendency for successful people to exhibit some forms of overconfidence or sort of say, oh, it's kind of my way or the highway.

1:18:37If you don't agree with what I'm saying right now, then we have nothing to talk about. That's not really useful. I try to stay humble I try and sort of, you know, keep an open mind, whether it's about wonky asset pricing research or it's more general. And I think that helps when you're out actually talking to real people about their investment decisions. So staying humble would also be part of success for me, I think. Great answer. Not cringe at all. Thank you. I appreciate that. Great answer and great interviews. Thanks so much for joining us, Mamdu. Thanks, Mamdu. Pleasure. Thank you very much for having me.

1:19:10Thank you. Nothing herein constitutes an offer or solicitation to buy or sell any security. This communication is distributed for informational purposes only. The information contained herein has been derived from sources believed to be accurate, but no guarantee as to its accuracy or completeness can be made. Furthermore, nothing herein should be construed as investment, tax, or legal advice, and or used to make any investment decisions. Different types of investments and investment strategies have varying degrees of risk and are not suitable for all investors. You should consult with a professional advisor to see how the information contained herein may apply to your individual circumstances.

1:20:15All market indices discussed are unmanaged, do not incur management fees, and cannot be invested indirectly. All investing involves risk of loss, and nothing herein should be construed as a guarantee of any specific outcome or profit. Past performance is not indicative of or a guarantee of future results. All statements and opinions presented herein are those of the individual hosts and or guests, are current only as of this communication's original publication date and are subject to change without notice. Neither One Digital nor PWL Capital has any obligation to provide revised statements and or opinions in the event of changed circumstances.