The Finance Paper that Changed Everything | Rational Reminder 404

0:08 This is the Rational Reminder Podcast, a weekly reality check on sensible

0:11 investing and financial decision-making from two Canadians.

0:14 We're hosted by me, Benjamin Felix, Chief Investment Officer,

0:17 and Dan Bortolotti, Portfolio Manager at PWL Capital.

0:22 Good to be back for another one.

0:23 This is a topic, I think, Ben,

0:25 that I was saying before that I feel like you've waited your whole

0:28 life to uh finally give an overview of factor investing in this way.

0:34 Yeah, we were kind of chatting about it before.

0:35 It's not It's a topic we've talked about uh lots,

0:40 but I don't think we've ever Well, we definitely have never done sort of a deep

0:42 dive into the the original paper that kind

0:46 of kicked off this whole way of thinking

0:47 about expected returns and and uh asset prices.

0:53 It was an interesting thing for me, too,

0:54 because I remember learning all of this, you know,

0:58 a decade ago when I first kind of got

1:01 into understanding index investing at a high level,

1:05 and then this kind of takes index investing to a different level.

1:10 Um but I hadn't really, you know, dove into it uh at this, you know,

1:16 depth until, you know, the last little while.

1:19 So, anyway, it'll be interesting to to kind of revisit it.

1:21 It's It's come a long way in 15 years, for sure.

1:25 Yeah.

1:26 Definitely.

1:27 Um so, we're we're we're calling

1:29 this discussion the finance paper that changed everything,

1:32 and we're we're really talking about Fama and French's

1:34 1993 paper in the Journal of Financial Economics,

1:38 which is a paper that today has nearly 15,000 citations,

1:42 which is like, you know, pretty pretty crazy for for an academic paper.

1:47 Uh and and it really was a paper that changed

1:49 financial economics and the the practice of portfolio management.

1:53 It It really changed it forever.

1:57 So, Eugene Fama and Ken French,

1:59 they found back then back in 1993 that a group of three

2:02 factors explained the vast majority of differences

2:05 in returns across diversified stock portfolios.

2:09 They did actually look at both stock and stocks and bonds in the paper,

2:12 but we're mostly going to focus on on the stock analysis.

2:16 Uh so, their their their findings in that paper,

2:18 they they really had and and continue to have sweeping implications

2:21 for the academic study of and and practice uh of uh of investment management.

2:28 So, listeners know we mentioned Fama and French quite a bit.

2:32 Their research comes up a lot.

2:34 Uh they've had such a big impact on on this space.

2:37 This paper is like one of the foundational reasons.

2:40 I mean, Fama obviously has um market efficiency, efficient markets,

2:44 but Fama and French together,

2:45 their their multi-factor asset pricing work is really foundational.

2:50 Uh I I I think it's research that every investor should understand

2:53 whether they choose to apply its findings to their portfolios or not.

2:57 Personally, this this paper is really the foundation

3:00 of how I think about investing and building portfolios.

3:02 It's actually their research is what I've probably told the story before,

3:07 but it's it's what led me to find Dimensional,

3:11 which is what led me to to get in touch

3:14 with Cameron and find PWL Capital whenever that was,

3:17 13 years ago, almost 14 years ago.

3:20 Um I was in like a corporate finance class and we

3:22 were talking about uh what is the cost of capital?

3:25 This whole big discussion in the class

3:28 and the and the the professor uh Professor Vijay

3:31 Jog talk talked about the CAPM and how

3:35 we could find the cost of capital that way.

3:37 And then he's like, "But there's some

3:38 more recent research that says you can look

3:40 at these other these other factors that that might tell you a little bit more

3:44 about the cost of capital and these other

3:46 sort of characteristics of companies that might

3:48 change the cost of capital relative to what the CAPM would say." And I was like,

3:52 oh, that's that's that's neat.

3:53 And then it ended up I think reading an article by Rob

3:57 Carrick that mentioned the Fama and French research and tied it to Dimensional.

4:01 Anyway, a bit of a tangent there, but uh yeah.

4:06 Uh now I I probably don't have to preface

4:09 this by saying that this is going to be a nerdy episode.

4:13 Um now I mean let's let's be honest with the type of listeners that we have,

4:16 that's probably why people are are listening.

4:19 But I think I think it's a discussion that's worthwhile.

4:21 So we'll cover the paper's methodology, the results, the enduring impacts.

4:26 And then we do have some comments near the end

4:28 about how investors today can pretty easily apply this information.

4:36 Any comments before you jump in here, Dan?

4:37 That's uh no, I mean it's funny.

4:39 I have a very different origin story, of course, how I found PWL.

4:42 And uh this is why it's it's an interesting um approach for me because

4:47 I kind of came at it when I first started getting interested in investing.

4:52 It was much more the sort of plain vanilla bogle type um uh indexing strategies.

4:59 And it was only later when I started to do a lot more reading,

5:02 when I, you know, came upon people like Larry Swedroe,

5:06 Rick Ferri, uh a number of others who are proponents

5:10 of taking that a step further and um you know,

5:14 exploring factor investing specifically through Dimensional.

5:17 Was no, that that for me came a few years later.

5:20 Um so but uh but definitely deepened, I think,

5:24 my understanding of how all of this comes together,

5:27 why it works, um the limitations of active management,

5:31 all of those things are tied into this.

5:33 Oh, yeah, totally.

5:34 It's all it's all so connected.

5:36 And it really is like it's all you mentioned this earlier,

5:38 it's all kind of cut from the same cloth.

5:42 Uh okay.

5:42 Yeah.

5:44 So we'll jump in here.

5:45 The the the fundamental premise of this paper, Fama and French's 1993 paper,

5:51 uh it it the fundamental premise is that multiple factors,

5:55 I don't know, I'll come back to the term factors in a minute.

5:59 Multiple factors affect asset prices and expected returns,

6:03 and that these factors can help to explain

6:06 why different types of stocks and bonds,

6:07 and therefore different investment portfolios,

6:10 have different expected returns and different realized returns.

6:15 An expected return is kind of a funny a funny term.

6:17 It's it's kind of what it sounds like.

6:19 It's the return that you expect from investing in a in a stock or bond.

6:23 Uh the study of expected returns is often referred

6:26 to as asset pricing because those two things are directly related.

6:30 An asset's price is based on its expected return,

6:33 and its expected return can be inferred from its price.

6:38 Uh expected returns are of course not guaranteed outcomes,

6:40 but as we will see, they do contain information,

6:44 and we can kind of infer that because different

6:46 types of assets have had systematically different realized returns.

6:52 It's all kind of funny stuff, right?

6:53 There's no like capital asset pricing model, CAPM, is like is like theory.

6:59 Um and the but the rest of the stuff it's often

7:03 called empirical asset pricing research because

7:06 there's some sort of light theory, but it's it's really empirically derived.

7:11 Like we you see, oh, small stocks have higher returns.

7:14 That there maybe there's a difference in expect Anyway,

7:17 um the the the idea that that different assets

7:21 have different expected returns and systematically different realized returns,

7:25 it kind of just it kind of makes sense.

7:27 You'd pay more for a safer asset than a riskier one, all else equal.

7:31 It's commonly known that stocks are riskier than bonds.

7:34 That's something that I don't think anybody really dis- disagrees on.

7:37 Uh and therefore they have higher expected returns.

7:41 But what's kind of less commonly understood is that different types

7:45 of stocks can have systematically higher expected returns than than others.

7:53 So, in Fama and French's original framing of the paper,

7:55 they talked about common undiversifiable risks

7:58 that a lot of investors are sensitive

8:00 to uh that helped to explain why some stocks have higher returns than others.

8:07 And I I I mentioned this earlier,

8:08 but I do want to just say again that this paper talks about

8:11 five different factors because they look

8:13 at both equity and fixed income factors,

8:15 but we're just going to talk mostly about

8:16 the the three equity factors that they looked at.

8:20 Uh if if a stock is exposed to more

8:23 of a certain type of risk that a lot of investors

8:25 are sensitive to, that stock must have a higher expected

8:28 return to entice investors to invest in it, all else equal.

8:34 And the idea that investors might care about multiple types

8:36 of risk wasn't a new idea when this paper came out.

8:39 This is something that Robert Merton and a few

8:42 other people had written about sort of theoretically.

8:46 But figuring out what those risks might be and how to measure them,

8:50 that was a new a new thing that Fama and French introduced in this paper.

8:55 Now, I think to understand why Fama

8:57 and French's perspective in this paper was so impactful,

9:01 it's important to understand what came before it.

9:04 So, in 1964 and 1965, researchers including Bill Sharpe developed the capital

9:10 asset pricing model or the CAPM, which I mentioned briefly earlier.

9:15 This model, the CAPM, connected a stock's expected return to how

9:18 its price moves relative to the overall market,

9:21 which is something expressed as its market beta or just its beta commonly.

9:25 A beta of one, for example,

9:27 means that a stock moves pretty much in lockstep with with the market.

9:31 A higher beta, a beta of more than one,

9:33 means the stock will tend to move up more when

9:35 the market is up and down more when the market is down,

9:37 and and vice versa for a beta less than one.

9:41 Uh according to this single factor model, the CAPM,

9:44 where exposure to market risk is the only risk that investors care about,

9:47 stocks with higher betas should on average deliver higher returns.

9:51 That's what the model predicts.

9:54 Now, the CAPM was like it was a huge deal in finance.

9:58 It was a big enough deal to win Bill Sharpe the Nobel

10:00 Memorial Prize in Economic Sciences 1990 for his work on it.

10:04 Because that model, the CAPM,

10:06 it it formalized a relationship between risk and expected return.

10:09 And it actually did a pretty good job,

10:11 a pretty good job empirically, explaining the observed returns of stocks.

10:17 You got to think like before the CAPM,

10:20 there there wasn't a model for like what should a stock's expected return be,

10:25 or what is a stock's expected return.

10:27 Um It is It is funny how we take that for granted now, right?

10:30 And like it seems so obvious, but I guess in the '60s,

10:34 no one had ever articulated it in that way,

10:37 and it demonstrated it with the math, right?

10:40 I feel that this is similar when

10:41 we talk about like Markowitz and portfolio theory,

10:44 and it's like if you combine uncorrelated assets into a portfolio,

10:49 you reduce the risk without necessarily decreasing the expected return.

10:54 We all know that today,

10:55 but I mean that was that was revolutionary when it came out.

10:59 Um so, you know, you have to go back 60 years,

11:02 but all of this stuff is just really crystallizing, I think, around this time.

11:07 Yeah.

11:08 So.

11:09 And Sharpe builds on Markowitz.

11:10 So, CAPM really builds on Markowitz portfolio theory,

11:13 and it creates uh testable predictions

11:17 that are really based on Markowitz portfolio.

11:19 But that that idea of having testable predictions,

11:21 I mean it it flows into all sorts of other stuff,

11:23 too, like how do you evaluate the the uh the performance of an active manager?

11:30 Before the CAPM, there was no model to say, "Well,

11:32 how much risk did they take and what were

11:34 their returns relative to what we would expect based

11:36 on the amount of risk that they took?" We

11:38 just didn't have a model for that before the CAPM.

11:40 And there's some interesting research that I'll I'll

11:41 mention in a bit that did that did that.

11:44 They took the CAPM and they applied it to active

11:47 mutual funds at the time and asked exactly that question.

11:49 Do they Do they earn returns higher than what

11:52 we would expect based on the risk they took?

11:54 Well, I think the to take that a step further, too.

11:56 I mean, there was a time not all that long before

11:59 this where no one really understood what market risk was, right?

12:04 Because you had to have properly constructed um

12:09 easily replicable indexes before you could do that.

12:12 And again, we take all this stuff for granted,

12:14 but if you were a stock picker in Benjamin Graham's day,

12:18 what were you measuring your performance against?

12:21 There was no benchmark.

12:22 So, um yeah, it's all again all like a lot of this stuff

12:26 during this period kind of the mid-century is when all of this stuff starts

12:30 to be formalized and now we can

12:32 actually start evaluating the performance of investments

12:36 and investment managers in a way that just wasn't pop or possible before.

12:41 Yeah.

12:41 Yeah, totally.

12:42 It really speaks to this idea that we are we are currently in this golden

12:46 age of investing where all all of this theory is is beneath us.

12:51 Like we're we're we're we're sitting on top of the foundation

12:54 of this powerful theory and not only are we aware of it,

12:57 but we have products that are just at our fingertips

13:00 to take that information and implement it in in investment portfolios,

13:04 which is pretty cool.

13:06 Uh okay, so even though the CAPM really defined and then dominated

13:11 the the study of asset pricing from its inception through the '70s and the '80s,

13:15 research research had consistently come out after

13:18 the CAPM was was published showing that certain

13:22 types of stocks had higher returns than what could be explained by the CAPM.

13:27 So, once we had this model, people started testing stuff.

13:30 I mean, I mentioned the active managers.

13:31 Do active managers generate returns in excess of what

13:34 would be expected based on the risk they took?

13:36 But, then there's also like oh, look at this certain type of stock.

13:40 Hey, it actually performs better than the CAPM would predict.

13:42 That's interesting.

13:43 People started doing all kinds of those types of tests.

13:47 Now, at the time when those observations could not be explained by the CAPM,

13:51 by the V asset pricing model at the time,

13:54 they were they would be referred to as anomalies.

13:58 Uh so, an asset pricing anomaly could mean one or two things.

14:01 This is a return that can't be explained by the CAPM.

14:05 An asset pricing anomaly could mean one or two things.

14:06 If we believe the CAPM is a perfect model for explaining expected returns,

14:13 anomalies in that case must mean that markets are not efficient.

14:17 And some of the papers at this time, uh I think one of them was

14:20 even titled titled persuasive proof of market inefficiency.

14:24 Uh so, it could mean that.

14:25 Or, if we know that markets are efficient,

14:28 the CAPM in that case must just be the wrong model.

14:33 Uh now, this this tension between those two

14:35 possibilities turns out to be impossible to resolve.

14:39 And this is something that Fama wrote about.

14:41 It's called the joint hypothesis problem.

14:43 Basically, we can't say whether markets are efficient without

14:47 having an asset pricing model to test market efficiency,

14:50 and we can't prove whether an asset pricing

14:52 model is right without knowing whether markets are efficient.

14:55 So, that's the joint hypothesis problem.

14:57 It makes all of this really non-testable, at least in a way that definitively

15:03 tells us uh whether markets are are efficient.

15:07 Uh but, even still, the original the CAPM

15:11 asset pricing research was was hugely important.

15:14 And that is where Fama and French start their 1993 paper.

15:18 So, they focus on three key problems with the CAPM.

15:21 One is that small stocks earn higher average returns than large stocks,

15:25 unexplained by differences in their CAPM betas.

15:28 The other one is that high book-to-market or value stocks

15:30 earn higher average returns than low book-to-market or growth stocks,

15:33 again unexplained by differences in their CAPM betas.

15:37 And then the relationship between beta and average returns is weaker,

15:41 was weaker than CAPM predicted.

15:43 And people may be familiar with the the anomaly there was low beta

15:46 stocks earning higher returns than the CAPM would suggest that they should earn.

15:54 Um if we just think in in basically in in in simple terms,

15:58 the CAPM suggested that two companies with the same

16:01 beta should have similar or identical expected returns

16:06 regardless of their size or value characteristics or really

16:09 any other characteristic um for for that matter.

16:13 Anything not following the model would be

16:15 considered mispricing or an asset pricing anomaly.

16:18 And so Fama and French were looking at this in this paper and saying,

16:22 you know, how how can we resolve this?

16:24 How can we resolve these these anomalies?

16:30 So they had noted uh the these these patterns

16:33 in returns that I just talked about.

16:35 Uh and they say right in the front

16:37 of their paper that the cross-section of average returns

16:40 on US common stocks shows a little relation

16:42 to the market betas of the Sharpe-Lintner asset pricing model.

16:46 Pretty pretty blunt blunt blunt language there.

16:49 Pretty like straight-up about, you know, CAPM's not working.

16:53 Um but then they follow up and this is the part that's really interesting.

16:56 They follow it up with the suggestion that maybe the standard

16:59 model was ignoring risk factors that are actually priced by the market.

17:06 And so it's on that basis that Fama

17:08 and French build their now famous three-factor asset pricing model.

17:14 So rather than just looking at the market factor like the CAPM.

17:18 Fama and French developed a a three-factor model that relates a stock's

17:22 expected returns to the market factor which is similar to the CAPM,

17:26 but they add a size factor and a relative price or or value factor.

17:33 So, the market factor again, basically same same idea as the CAPM,

17:36 how much the stock or or or portfolio moves when the overall market moves.

17:41 And and that that just shows whether the stocks

17:44 went up or down with with the market.

17:45 Uh it's still a super important factor because broad market movements do

17:50 tend to affect most stocks regardless of their size or value characteristics.

17:55 And the CAPM in Fama and French's paper,

17:57 it still explains like around 60 maybe 70

18:01 in some cases 80% of differences in return.

18:03 So, it's still a hugely important factor.

18:06 The size factor is pretty straightforward.

18:09 It's based on the size of companies.

18:11 Um more specifically, it's it's denoted as SMB or small minus big.

18:17 This factor refers to the difference in returns

18:19 between small company stocks and large company stocks.

18:22 And if small stocks outperform large stocks in a given period,

18:24 the SMB premium is positive which it had been at the time that they were

18:29 doing this research which is why it was

18:31 identified as an anomaly or a risk factor.

18:35 Uh if if large stocks do better than small stocks which have has happened over

18:40 a lot of the periods of time since this paper came out the premium is negative.

18:45 So, including this factor captures whether being a small company is

18:49 related to any systematic return variation that investors are compensated for.

18:55 And then the third one is the value factor.

18:56 And this one's denoted as HML or high minus low

19:00 which is referring to the book-to-market valuation ratios of companies.

19:04 Book-to-market is basically the company's accounting value.

19:08 It's like uh on paper value compared to its stock market value.

19:13 So like what's it worth on paper versus

19:15 what are people investors willing to pay for it?

19:18 A high book to market ratio means it's a value stock.

19:23 So they're they're cheaper in market price

19:25 relative to the book value of their assets.

19:27 And then a low book to market stock is a growth stock.

19:30 So a growth stock is where investors are willing to pay more for future

19:33 potential which pushes prices further above

19:36 the book value of the company's assets.

19:39 HML high minus low measures the difference

19:41 in returns between value stocks and growth stocks.

19:44 And again, similar to the SMB premium,

19:46 when value stocks outperform the HML premium

19:50 or the value premium is positive and then vice versa.

19:54 So that captures the effect of any return premium

19:56 that investors receive for holding value stocks rather than growth stocks.

20:01 These are all each of each of these factors

20:03 including the market factor are are called long short portfolios.

20:07 So SMB for example is long small cap stocks and short big stocks.

20:12 That part's I don't know.

20:14 That You you could have a whole

20:15 discussion about that about long short portfolios.

20:17 There's some other kind of funny things about

20:19 the way that the factor portfolios are constructed too,

20:23 um, which is even too nerdy I think for for this discussion.

20:27 But that the main idea is

20:28 that they they constructed these these factor portfolios

20:32 that are really designed to capture the return

20:34 variations related to company size and relative price,

20:37 the systematic return variations which are things

20:40 that at the time had been well documented as return anomalies.

20:45 So the interesting thing about this though is, correct me if I'm wrong,

20:48 but it doesn't seem like they addressed the third

20:52 issue which was basically what I would call

20:54 the low volatility anomaly where you had these low

20:58 beta stocks that the model suggested would underperform,

21:02 but in fact, they outperformed.

21:04 That's not really part of this, um,

21:08 there it's it's not one of the new factors added to explain volatility.

21:13 Low volatility has become a investment strategy and there

21:19 are funds in that that that try to capture it.

21:21 But I'm I'm not sure that it has the same

21:23 kind of academic rigor that that the other factors have,

21:26 but maybe we'll get into that later.

21:29 I I I think it does.

21:31 Uh it's just it's a little bit different and Fama-French's three-factor model,

21:36 you're right then it it it didn't really address the low volatility anomaly.

21:41 Their later five-factor model, which we'll touch on in a little bit,

21:44 it did go quite a bit further in addressing

21:47 or or explaining the low volatility volatility anomaly,

21:51 but yeah, the three-factor model didn't really touch it, you're right.

21:56 Uh okay, so the next thing I do in this paper, which is really neat,

22:00 is they create this model,

22:03 um but then they need to kind of take it take it for a spin.

22:06 So, to do that, to take it for a test drive,

22:09 they formed diversified portfolios by sorting

22:12 stocks based on size and value characteristics.

22:15 So, they split stocks, they these are US stocks,

22:18 into five size groups and five book-to-market groups.

22:22 Uh so, we get a total 25 test portfolios.

22:27 So, this really just allowed them to capture every combination

22:30 of of the two um of the size and value factors.

22:33 We've got like small value stocks on one end,

22:36 we've got big growth stocks in the other end,

22:38 and then all the all the possible combinations in between.

22:43 They show that these groups of stocks

22:45 do have pretty significant variation in average returns.

22:49 So, they're like, you know, there's there is something going on here for sure.

22:52 Uh and then they test whether their three-factor

22:55 asset pricing model is able to explain that variation.

22:59 Uh and this is where they did

23:00 something that at the time was pretty groundbreaking

23:04 and still shows up in tons of academic

23:07 finance papers and practical investment analysis today.

23:11 They used time series regression to test how well their asset

23:14 pricing factors explained the returns of their 25 test portfolios.

23:19 A time series regression looks at the returns

23:21 of a portfolio over time and asks how much

23:24 of the variation in returns is explained by the asset

23:28 asset pricing factors being used in the model.

23:32 It's pretty straightforward to run a time series regression.

23:34 You you can do it pretty easily in Excel.

23:36 There's also free tools online like Portfolio Visualizer where you can just drop

23:40 in a ticker and it'll run a three or a five factor regression.

23:45 I mean, on Portfolio Visualizer you can even try different asset pricing models.

23:50 Fama and French have their asset pricing models,

23:53 but there are other competing models out there.

23:55 Anyway, the tools we have I mean, talk about a golden age of investing.

23:59 Um you can just with a couple clicks run analysis that would

24:02 have taken Fama and French a lot longer to do back in 1993.

24:07 Let alone Bill Sharpe in 1965.

24:09 Yeah, yeah, right.

24:11 So, yeah, it's very easy for us to kind

24:14 of look back at a model like CAPM and say, "Oh, it's simplistic.

24:17 It doesn't explain that much." It's like, "Yeah,

24:19 but there's a reason why nobody did it until him, right?"

24:23 So, you know, it again, we we really do take a lot of this for granted.

24:28 Yep.

24:29 Uh yeah.

24:31 So, so the the the output of a time series regression is going to is

24:35 going to spit out factor loadings is what

24:37 we we call them that were were coefficients.

24:40 Uh and and it really just tells you how

24:42 the portfolio being looked at moves relative to each factor.

24:48 And it's also going to spit out an alpha,

24:49 which is the portion of returns that was

24:51 not explained by the factors in the model.

24:54 Now, that last term alpha turns out to be pretty

24:57 important to the the practice of investment management for sure,

25:01 but also to academic analysis.

25:04 So, alpha was first used to describe

25:07 excess risk-adjusted returns in a 1968 paper.

25:11 Now, this is a a Michael Jensen paper.

25:13 It's another another banger of a paper if you're into this kind of thing.

25:17 It's actually really cool paper.

25:18 It's the one I mentioned earlier where they they took the CAPM,

25:20 this new thing at the time, and applied it to active fund performance.

25:26 Uh yeah, so they they took the the the single

25:28 factor capital asset pricing model and asked whether actively managed mutual

25:32 fund managers were generating returns in excess of what should

25:35 be expected based on the amount of risk they were taking.

25:38 Uh it's kind of tangential to this discussion,

25:41 but it's worth mentioning that active managers

25:43 in that study were not able to beat the market.

25:46 Um that's a I we've talked about that paper in the past

25:50 when we were making a case for index funds, I think.

25:53 And then that model only only accounted for market risk.

25:55 It only accounted for CAPM.

25:58 Uh we'll come back to the effects of including

26:00 other factors in that type of analysis later.

26:03 In the context of Fama and French's 1993 paper,

26:07 they took each of the 25 test portfolios

26:10 and tracked their performance from 1963 to 1991.

26:15 And then they used their model to try

26:16 to explain the differences in their returns.

26:19 So, for each portfolio,

26:20 the regression estimated how its returns co-moved with the three

26:23 factors in the model and whether there were any large alphas,

26:26 the the excess returns unexplained by the model.

26:28 They also measured the R-squared values.

26:31 It's a measures the explanatory power of a regression model,

26:36 like like what percentage of a portfolio's ups

26:38 and downs over time could be attributed to the factors.

26:41 An R-squared of one would mean perfect explanatory power and and zero

26:45 would mean that the factors explained basically none of the returns.

26:49 Uh what Fama and French found in terms of alphas

26:51 and R-squareds across the 25 test portfolios was pretty incredible.

26:55 Keeping in mind this is an empirical model like

26:57 they had okay we we've got the market factor

27:00 and we've got this small companies are doing some

27:02 weird stuff value companies are doing some weird stuff.

27:04 Let's let's test them all out and see

27:06 if there's something going on here and they

27:11 find that the R-squared values ranged from across

27:16 these 25 test portfolios ranged from 0.83 to 0.97.

27:21 It was about 0.93 on average across the the test portfolios which is very high.

27:27 In 21 of the 25 portfolios tested

27:30 the R-squared value of the regression was over 0.9.

27:35 So this is where we get the common at least in in our world of factor nerdiness

27:40 the common description that the three factor model explains

27:44 around 90% of the differences in returns between diversified portfolios.

27:49 There were also there were two other key

27:51 data points that came out of these tests.

27:53 One was that each test portfolio had a beta very close to one.

28:00 So that that was interesting because it

28:01 really reinforced the fact that the variation

28:04 in returns is explained by more than just exposure to market beta.

28:07 Since the market beta exposures were basically the same

28:10 across the board but the returns varied widely.

28:13 Now that would be very surprising that finding if

28:16 market beta fully priced all assets but instead we

28:21 kind of see things following the the pattern

28:24 that Fama and French predicted with their three factor model.

28:29 Yeah so that this is where again in these tests

28:32 we see that CAPM explains 60 to 80%.

28:35 I usually hear it described as around

28:37 60% of differences in returns across diversified portfolios.

28:43 And just to clarify here like if you say for example that that the three factor

28:47 model explains 90% the 10% that is unexplained

28:52 could be skill but it could also be luck.

28:55 And the model does not differentiate between those two.

28:59 Skill, luck, or or some yet to be determined factor.

29:03 Okay.

29:05 And we will see Yeah, it means unexplained.

29:08 It doesn't mean explained by something specific other than those three factors.

29:12 It just means we don't know.

29:14 Yeah, correct.

29:15 Correct.

29:16 Something outside of the model being being tested.

29:19 And even even still, like when they add more factors later,

29:23 we get up to maybe 90 93 94 maybe 95% explanatory power,

29:28 but there's kind of always going to be

29:29 that that idiosyncratic component of of whatever of you know,

29:34 idiosyncratic risk or skill or luck or whatever you want to call it.

29:39 Um okay, so I I I do think it's worth reiterating

29:42 that we went from being able to explain somewhere between 60 70 maybe

29:48 80% of the differences in returns between diversified portfolios with the CAPM

29:52 to being able to explain over 90% with the three factor model.

29:57 Which is pretty pretty crazy.

30:00 Uh another important observation in their regressions was that the three factor

30:06 regressions resulted in near zero intercepts

30:09 or alphas on almost all test portfolios.

30:12 The one exception, the one that they really struggled with and actually

30:16 their five factor model continued to struggle with was small cap growth stocks.

30:21 They had and have much lower returns than the three factor model could explain.

30:26 So that left a lot of room for future research.

30:29 A lot of which has been done,

30:30 but there's still a a funny anomaly there where small cap growth stocks

30:33 continue to tend to have lower returns

30:36 than than asset asset pricing models predict.

30:41 Uh So the the the the minimal alphas in most cases except for small growth is

30:47 important and it might be one of the most

30:49 compelling sort of validations of the model.

30:52 Uh Fama and French demonstrate that once you

30:55 account for market size and value factor exposures,

30:59 there are basically no persistent unexplained

31:02 returns remaining in most portfolio sorts.

31:05 And again, that's the 90% like we're

31:07 explaining most of the differences in returns

31:09 between across all these portfolios that vary

31:12 across a couple of important dimensions.

31:16 The other thing they did in this paper that I I love.

31:18 Um so they had their 25 test portfolios sorted on size and value.

31:23 Cool.

31:25 Uh but they also had two other characteristics that had

31:28 been associated in academic literature as being associated with higher returns.

31:34 So they again fire up the three-factor model

31:38 to to see if they can explain those anomalies.

31:41 Uh and those were uh dividends to price and earnings to price.

31:47 And they again find that the three-factor model

31:49 explains the returns across these portfolios as well.

31:54 And so this is one of the places where Fama and French

31:56 really shattered the beliefs of dividend focused investors and showed that well,

32:00 there's nothing special about dividends even if the high

32:02 dividend to price portfolio does perform kind of well,

32:05 it's because it's a value portfolio, not because dividends are special.

32:08 Yeah, that's what I was going to say.

32:09 Is it just a a lot of overlap here, right?

32:12 Um a lot of high dividend stocks

32:15 have high dividends because they have low prices.

32:19 That's right.

32:19 It's just the inverse way of looking at it.

32:21 And I agree with this is something that I

32:23 think has not always resonated with dividend investors, right?

32:27 It's not that the high dividend is good,

32:31 it's that the dividend is high because the price

32:33 is low and the low price is what's good.

32:36 So at least I mean I think that there's

32:39 dividends high dividends can be a proxy for value.

32:43 They're just my understanding of this is that Fama

32:46 and French showed that it's not a very good proxy.

32:49 That price to book is a better measure.

32:53 Uh if you're trying to find value.

32:55 There's lots of different ways to measure value stocks.

32:58 But, they have focused, I believe,

33:01 right on price to book as being the most reliable way of measuring this factor.

33:07 Yeah.

33:08 And I I think it's it's come a long

33:09 way since this 1993 paper where once they bring

33:12 in profitability and kind of combine that with value

33:16 and and you're looking at those two metrics together,

33:18 I think that they're uh sort of precision in in looking

33:20 at what is a not just a cheap company,

33:23 but a cheap company that actually is a good investment opportunity.

33:26 Looking at both those factors is is quite good,

33:28 uh which has an inter- interesting interaction with dividends as well,

33:32 because if you look at a dividend portfolio,

33:34 it's probably going to have a value tilt.

33:36 There's a good chance it's got a profitability tilts, too.

33:39 So, dividends turned out to be a pretty good sort of naive

33:42 filter for a couple of of very robust uh academically robust factors.

33:48 I think where it gets interesting

33:48 from the dividend investing perspective is that you could

33:52 build that portfolio that loads on value

33:55 and profitability uh without needing to sort by dividends.

34:00 Right.

34:00 Cuz you're excluding a whole bunch of companies

34:02 that don't pay dividends to build a dividend focus portfolio,

34:05 when if you're really just trying to tilt toward value and profitability,

34:08 you can do that without the dividend focus.

34:10 And I think with the dividend tilt,

34:12 you'll often end up with uh with with larger companies, as well.

34:16 And you might not want that.

34:17 You might not want to tilt toward size, um or or or not.

34:20 But, when you focus on dividends,

34:21 you sort of get naive exposure to the other factors,

34:25 when you could be focusing on the factors that are

34:26 actually driving returns and get a a more diversified,

34:29 more customizable portfolio.

34:32 Mhm.

34:35 Uh okay.

34:36 Uh one thing that is worth mentioning is that is that it's still debatable.

34:44 Like Fama and French, the language they use in this paper is that hey,

34:46 these are we're modeling these as as systematic risk factors.

34:51 That's that's still debatable.

34:53 We don't really know if they're risk

34:54 factors or if they're artifacts of mispricing.

34:57 And then that's another one of those things

34:58 that's really really hard to test definitively.

35:02 Like do these things do do these systematic factors that we can see in the data,

35:06 do they exist because of mispricing or do they exist because of risk?

35:12 I don't I don't even know if it matters.

35:15 I think it matters if you believe that mispricing would be arbitraged away,

35:20 which some people do believe.

35:20 We've had some guests talk about how they think that has happened.

35:24 Um if you believe that there are limits to arbitrage

35:26 and that systematic mispricings won't go away for that reason,

35:30 then I don't I don't think it's really relevant whether we're looking at risk

35:33 or mispricing as long as we believe these things are going to persist,

35:37 which is really the big question from a a practical perspective.

35:41 Cuz if we can look at these factors and say uh cool, they exist and hey cool,

35:46 they explain differences in returns across diversified portfolios.

35:49 If we don't know or or we we can't know, but if we don't think that the premiums

35:53 for these factors are going to be positive going forward,

35:56 uh we may not want to tilt toward them.

35:57 I mean, you may even want to tilt away from them

35:59 if you believe the premiums were going to be negative.

36:02 Uh those things are kind of uh they're kind of separate, I guess,

36:05 where you can have a model that does

36:07 a good job explaining differences in returns,

36:10 but uh the premiums don't have to be positive.

36:15 But if you're going to tilt toward them,

36:16 you you kind of hope they're going to be positive.

36:20 Um but I think I think the big thing Fama and French

36:23 showed is that size and value don't need to be anomalies.

36:27 We don't have to look at the world from a CAPM perspective and say, "Hey,

36:30 these things don't make sense." Uh what Fama

36:33 and French said is maybe these things are systematic factors,

36:36 risk or or otherwise, that we need to account for when we're looking

36:40 at portfolios and assessing different types of of stocks.

36:43 Uh and when they include those in their model,

36:46 they were able to almost fully explain return

36:49 differences across this broad range of diversified portfolios.

36:55 Um Yeah, and that that really changed how we viewed markets.

37:04 Um Yeah, we we went from in a CAPM world,

37:10 that's sort of whatever you want to call it,

37:11 60% explanatory power to with a three-factor model 90% plus explanatory power.

37:19 It's a very powerful increase.

37:22 Uh And then the other area, I mentioned this earlier,

37:25 the other area that this has an impact is not just it's like, "Okay, cool.

37:29 We can explain differences in returns.

37:31 That's That's neat.

37:32 These things don't all have to be anomalies.

37:34 Okay." But it gets really interesting again when we say,

37:38 "Why have some active managers outperformed?"

37:43 If they've outperformed because they're skilled,

37:45 then maybe we should be giving them more money to invest.

37:47 If they've outperformed because they tilted toward value or size,

37:51 uh maybe that's not as interesting and and maybe

37:54 they shouldn't command a high fee for that service.

37:57 It's basically like if if you know

37:59 a manager is just tilting toward value stocks,

38:01 they're going to charge you 1.5% to do that, you

38:04 could just buy a value index for whatever, 20 basis points.

38:09 Uh and Fama and French did look that look

38:11 at that in in later research using the three-factor model.

38:14 Uh and they that's their famous paper on luck

38:18 versus skill in active uh uh fund management.

38:22 Yeah, one does wonder, right?

38:23 Like for before this was really understood,

38:26 there were managers who outperformed.

38:30 And looking back with hindsight bias, we can say,

38:34 "Well, you outperformed because you you know,

38:36 tilted more to small stocks or value stocks." But those managers identified

38:41 those opportunities presumably before the academic

38:45 research showed that they had premiums.

38:48 So, they might have been lucky.

38:50 But they might have just had very good intuition.

38:53 Um and then again, and we will probably talk a little bit about this, too,

38:57 is that once the factors get publicized,

39:01 there is some potential for them to shrink

39:03 because the market is mostly efficient, right?

39:07 And so, if some of these things were indeed mispricings,

39:11 misunderstandings of risk,

39:14 that would presumably get smaller in the future as people understood that.

39:19 But it's all This is all very difficult to tease out,

39:22 right, from real-world fund and manager performance.

39:26 Yeah.

39:27 Oh, yeah.

39:28 Very.

39:29 Very much so.

39:30 And that is true that post-publication

39:33 of factor premiums have tended to get smaller.

39:37 Uh I don't think they've gone to zero.

39:40 We we have We've had one guest, Andrew Chen, who thinks that they have.

39:43 He's looked at the US and said that maybe they've gone to zero.

39:47 Uh for for exactly that reason.

39:49 But uh you look outside the US and the factor premiums

39:52 have still been quite positive and even post-Andrew's sample in the US,

39:58 the value premium, for example, has been positive over that specific period.

40:03 Uh but it's all There's so much noise in this stuff.

40:04 It's hard to it's hard to really say what is true.

40:10 I mean, we have degrees of truthiness rather than truth.

40:15 Yeah, there's so little that we can be confident and certain about, right?

40:19 So, you know, you just try to do what what makes sense

40:23 based on the evidence without putting too much predictive power on any theory.

40:31 Altogether the findings in this paper, they they kind of they challenge but they

40:34 also built on single factor CAPM asset pricing

40:38 to create a much more comprehensive framework for how

40:41 investors should think about building and evaluating portfolios systematically.

40:47 Uh and then after this paper came out,

40:49 it it kind of caused this explosion in the a term that I mentioned earlier,

40:54 which is called empirical asset pricing.

40:58 It's really like looking at returns and figuring

41:01 out what factors might be driving them.

41:04 So, that field of study really exploded.

41:06 Uh researchers wanted to find the the best factors,

41:09 they wanted to create better asset pricing models,

41:11 and it it kind of became a bit of a problem, maybe, I don't know.

41:15 John Cochrane in his 2011 presidential

41:18 address to the American Finance Association,

41:20 he described the proliferation proliferation of factors as a zoo.

41:24 So, it's kind of become this famous term, the factor zoo.

41:27 Um you could maybe even call it factor slop, I don't know,

41:31 to use the language from my recent video on ETF slop.

41:34 That's probably not very nice to some of the researchers who

41:36 have discovered some of the the many factors, I don't know.

41:39 Uh but there was a 2016 paper from Cam Harvey,

41:43 Liu, and Zhu, and and they found at that time, this is 10 years ago now,

41:49 they found that there were 316 distinct factors

41:51 that had been published in academic journals at the time.

41:54 I think later research has kind of suggested

41:56 that maybe there were 316 published factors,

42:00 but they really all fall into a relatively small number of categories.

42:05 So, we found like 316 different flavors,

42:07 but there's when you really boil it down, there's only a handful still.

42:11 Uh but in any case, it it it did create

42:13 this whole new set of academic techniques for choosing factors.

42:18 Like, how do you how do you evaluate which factors belong in a model?

42:21 Uh, different techniques for comparing asset pricing models.

42:24 So, there's like a whole bunch of papers on this now.

42:27 Uh, but anyway, in in in the face of the of the zoo or the slop,

42:31 whatever you want to call it,

42:32 Fama and French did go on to update Uh, I don't know if they call it an update,

42:38 I guess, but they they created a new model in 2015.

42:42 And they introduced two new factors with that update.

42:45 And those were profitability and investment.

42:47 So, profitability is expressed as RMW or robust minus weak.

42:52 And that's the excess return, um,

42:54 from companies with high profitability over those with weak profitability,

42:58 or robust profitability over weak profitability.

43:01 And then investment is expressed as as CMA, conservative minus aggressive.

43:06 And that's companies that grow the book value of their assets slowly,

43:10 which is conservative,

43:11 tend to outperform those that that pour cash into rapid asset growth,

43:15 which are called aggressive.

43:17 And the five-factor model did help to solve some

43:19 of the problems that the three-factor model was not able to.

43:23 Like we talked with the low volatility anomaly,

43:25 there were a couple of other ones that were

43:26 sort of still unexplained by the three-factor model,

43:28 but Fama and French have, uh, a separate paper.

43:32 I can't remember exactly what the title of is, but it's like, uh,

43:35 something about dissecting anomalies with the five-factor model,

43:38 where they go and take a bunch of asset pricing anomalies and say,

43:41 "Well, you know, these are largely explained

43:43 by the five-factor model now." And as I mentioned earlier,

43:46 it did move the explanatory power up closer to 95%.

43:51 I think it's like, depending on which portfolio sort we're talking about,

43:54 it was sort of up to 94%

43:56 of the differences in returns across diversified portfolios.

44:01 And the five-factor model is really now, today,

44:03 I would call it the sort of workhorse.

44:05 It's like the benchmark asset pricing model in academic finance.

44:10 You think it's worth a a quick chat about profitability?

44:15 I mean, investment to me has always

44:17 been the most difficult one to understand intuitively.

44:21 Um but profitability has that issue as well.

44:23 I think it does come back to this idea of the low volatility anomaly.

44:28 Like in other words, I think we all understand intuitively small

44:33 stocks have higher expected returns in large companies.

44:37 It's usually because they're riskier.

44:40 For value, maybe it's risk.

44:42 The behavioral component I think explains a lot, too.

44:45 People generally don't like cheap boring

44:48 companies and they like big expensive companies.

44:51 They will pay more for them.

44:53 Profitability is a lot harder to understand cuz why would

44:57 investors not want to invest in companies that are clearly profitable?

45:03 And if they were attracted to those {quote} {unquote} good companies,

45:07 would they not drive up the price until the point where This is the kind

45:12 of the old idea of you can't tell

45:14 a good stock by identifying a good company, right?

45:18 Because price means everything.

45:20 So, you know, is there some explanatory reason why investors would not prefer

45:27 profitable companies and therefore impart some

45:30 kind of higher expected return on them?

45:33 I I think it's it's I've always had had

45:37 to think about it from a a multi-factor perspective.

45:39 It's really like an all else equal perspective.

45:42 If we take two companies that are otherwise identical on all characteristics,

45:46 but one has higher profitability,

45:48 but they've got the same value characteristics and all that stuff,

45:50 the higher profitability company must have a higher

45:52 discount rate applied to its expected future

45:54 cash flows in order for their valuations to be the same for these two companies.

45:59 So, if we're looking at which stock should I buy,

46:01 and we find one that's more profitable,

46:02 but it's trading at the same relative price as some other stock,

46:06 the discount rate the the inferred discount rate

46:08 or implied discount rate that the market must be

46:10 pricing in for that for those characteristics to be

46:13 the way that they are must be higher.

46:16 And I think that's the sort of story that Fama

46:18 and French talk about in their in their five-factor paper as well.

46:21 And then when you think about it from a single-factor perspective,

46:24 I think it's kind of similar for value actually.

46:26 If we think about just just value stocks uh in aggregate,

46:30 they have higher expected returns like

46:31 you said because they're boring companies,

46:33 people aren't as they're maybe riskier, people aren't as interested in them.

46:37 So, they've got lower prices, but we've also got that question of do

46:40 they have lower prices because they're crappy companies?

46:43 In which case they're not they're not good investment opportunities,

46:46 or is it because they have higher

46:48 discount rates but they're pretty good companies?

46:50 And so, I think there's there's a lot of noise

46:51 in that when you just look at the single-factor value sort.

46:55 When you introduce profitability,

46:56 it's a lot easier to find which companies have high expected returns.

47:00 They've they've got high profitability and low prices,

47:02 which is the combination that you want.

47:05 Uh but I think that they're the low prices without sorting on profitability

47:09 are enough on their own to have

47:11 higher expected returns as a single-factor portfolio.

47:14 I suspect it's something similar going on with profitability where

47:19 you really want the high profitability companies with low prices.

47:22 And that's a multi-factor sort.

47:24 If you just take the aggregate profitability portfolio,

47:28 you're probably still picking up enough of that uh for there to be a premium.

47:32 That's the way I will always thought about it.

47:34 Uh maybe there are better ways to think about it,

47:36 but I I I agree it is a less intuitive story,

47:39 especially if you try and think about it from a single-factor perspective.

47:43 Mhm.

47:44 And once you layer on the factors because again,

47:46 many maybe most companies will sort high or low across multiple factors, right?

47:54 So, in In words, a lot of value stocks are probably high profitability

48:00 as well because that's what makes them value as opposed to simply cheap, right?

48:04 Like this is kind of your argument is that a value

48:07 stock is not just one with a low price, it's a low price relative to some

48:11 positive characteristics in the company earnings, assets, whatever it is.

48:16 Um so a lot of these are going to overlap.

48:20 So maybe the companies with the highest expected returns

48:23 may score high on three or more of these factors.

48:28 Yeah, so I I So I I think in a lot of cases they're

48:31 going to be uh counter uh I don't know what the right word is there.

48:37 They're going to be negatively correlated.

48:38 So like if we just take the portfolio of value stocks,

48:41 it's going to tend to be less profitable than the market as a whole.

48:45 Uh similar things can happen with company size,

48:48 uh which is why taking that multi-factor perspective ends up

48:50 being so important because if we just take value stocks,

48:55 you end up with a basically low profitability

48:57 tilt which has a a negative expected premium.

49:01 But if you can take that value portfolio and bring

49:02 the profitability characteristics up by also doing a profitability sort,

49:06 then your expected return just goes up.

49:09 And so this is why as this research has come out,

49:11 it's been so important for firms like Dimensional and Avantis who we'll talk

49:16 more about in a second to incorporate that that research because you don't

49:20 want to just own cheap stocks when you don't really know why they're

49:23 cheap because maybe they're crappy companies

49:25 you don't actually want to own them.

49:27 And I think you do get a lot of that uh if you just do the single factor sort,

49:31 but once you add in the second factor,

49:34 and you you see that in like uh especially now that they've

49:38 changed the way that they're doing their their weighting of the factors,

49:41 uh Dimensional in in the vector portfolios which are Canadian listed well,

49:46 they have a Canadian and US funds uh that are

49:50 pretty aggressively tilted towards size and value or well,

49:53 they used to be pretty aggressively tilted towards

49:54 size and value and then took profitability into account.

49:57 But they fairly recently changed their methodology to have

49:59 a much more sort of equal emphasis on size, value, and profitability.

50:04 And when you look at their aggregate characteristics,

50:08 it used to be that the Vector portfolio would be,

50:12 you know, way cheaper than the market, but also have a little bit lower

50:15 profitability or similar profitability to the market,

50:20 which was in it itself pretty pretty impressive

50:22 because if you just did the value sort,

50:24 you'd end up with a much worse profitability characteristic.

50:28 And then now with the more equal emphasis, it's it's even more balanced.

50:33 Does raise the issue, right?

50:34 Like you you need to make some kind of active decision about how

50:39 you want to weight all of these factors because in a naive way,

50:42 like the more factors you layer on, especially if they're negatively correlated,

50:48 at some point you just end up with the market.

50:51 No, I mean like I think unless

50:53 you plan on concentrating the portfolio quite aggressively,

50:57 you just end up with a portfolio that's not

51:01 dramatically different from just a boring market cap weighted one.

51:06 And so you're going to see a lot of correlation

51:08 with just a plain vanilla index and any opportunity for outperformance,

51:13 but also risk of underperformance maybe gets narrowed unless you've got

51:18 a really compelling formula for how to combine these in the right way.

51:23 And I guess that's where that's where the skill comes into, right?

51:27 I mean, you've got the research.

51:28 Okay, that's the raw material.

51:29 Now, what do we do with it?

51:31 I think there's skill, there's also an element of of humility.

51:33 Like the way that Dimensional and I think Avantis is similar,

51:36 the way that they do this is very diversified

51:39 and it is still dominated by the market factor

51:41 and it's going to have relatively low tracking error

51:44 to the market because they are taking that super broad diversification.

51:48 There are products out there on the market that are you know,

51:52 they're they're much more concentrated.

51:54 Uh I would say they they tilt much more toward the active

51:57 portfolio construction by building concentrated

51:59 portfolios that are still quantitatively informed.

52:02 Like you could still call them factor portfolios and I think

52:05 a lot of the product names actually do call themselves factor portfolios.

52:10 But they're much more concentrated and in that case

52:12 you're making a much bigger bet on the premiums to to your point then

52:15 also going to have much more tracking error.

52:17 And you've also got much more risk of underperforming

52:20 the market if the factors don't work out as expected.

52:23 With something like Dimensional, which we haven't really talked about yet,

52:27 although listeners are probably already familiar,

52:30 you do look a whole lot like the market and from our perspective,

52:33 that's a good thing.

52:34 From someone's perspective who wants, you know,

52:37 really high conviction or or wants to really outperform the market,

52:40 they might consider that not such a good thing.

52:42 Um but we we don't have so much conviction

52:47 in these in these factors that we want to build,

52:49 you know, 25 stock portfolios that are as concentrated

52:51 as possible in uh in the the multi-factor sort.

52:57 So, yeah.

52:57 I I I I think that's right that at a certain point

53:01 you do look a lot like the market because you're broadly diversified.

53:05 And I think that's I think that's kind

53:07 of a good thing from from our perspective,

53:09 but not everybody would necessarily agree with that.

53:13 Yeah, I think that's why the the the term tilt is apt.

53:15 Right.

53:16 It's subtle, right?

53:17 It's not uh it's not an enormous fundamental

53:20 move uh away from a broadly diversified portfolio.

53:25 It's sort of starting from the premise that a diversified portfolio is

53:29 very good and now we're just trying to tweak it around the edges.

53:33 It's like Fama's famous line that you've got

53:35 to talk yourself out of the market portfolio.

53:37 Mhm.

53:39 And I think that that really comes back to, yeah,

53:41 how much conviction do you have in in these kind of tilts.

53:45 Uh it is also worth mentioning that there's

53:47 still tons of ongoing debate both in academia

53:50 and in practice about which factors make sense

53:53 and which asset pricing model we should be using.

53:56 So, Dimensional and Avantis as well are broadly using the Fama and French

54:03 five-factor model and that that that general approach and and type of thinking,

54:08 but there are tons of competing factor models,

54:09 there are tons of competing products,

54:11 there are companies that have their own factors.

54:14 So, this isn't that's not like a settled debate by any means.

54:17 Um but I'm I'm pretty comfortable saying

54:20 that when transaction costs are accounted for, which is

54:23 something that has been done in in that in that academic

54:27 literature that compares asset pricing models,

54:30 that there's one paper that shows that when you account for transaction costs,

54:35 Fama and French's five-factor model is really really strong.

54:40 Really, it's a very good foundation for thinking

54:42 about portfolio construction and evaluating portfolio performance.

54:46 And that that transaction transaction cost thing

54:48 is is important because another factor is momentum,

54:52 which is very empirically strong,

54:54 and Fama and French have always kind of it's interesting reading their papers,

54:57 they've always kind of struggled with it.

54:59 They say in one paper that they included momentum

55:02 in an asset pricing model that they were testing,

55:04 but they say that they include it reluctantly.

55:06 They they're very concerned about about data mining and about

55:10 adding factors just cuz without any sort of possible theoretical story.

55:16 Anyway, so there's one paper that compares different asset pricing models

55:20 and shows that momentum can look really good when you ignore transaction costs,

55:24 but when you account for transaction costs,

55:26 the Fama and French model without momentum starts to look a lot better.

55:32 Um Yeah, momentum is one of these things that's

55:34 always kind of famously it it makes theoretical sense.

55:40 There's a lot of data showing, like you said,

55:42 if you ignore transaction costs, it's compelling.

55:44 Except you can't ignore transaction costs.

55:47 Unlike these other tilted portfolios, which are pretty buy and hold, right?

55:52 They will need reconstitution from time to time.

55:56 But, you know, what is the the momentum factors are

55:58 typically like 1 to 3 months trading cycles and they're like,

56:02 who's implementing that?

56:04 And if you are, I mean just the the tax impacts,

56:08 the cost of the trend that you know, brokerage fees are kind of not a thing,

56:13 but certainly just the churning and the tax implications of that are

56:19 likely going to eat up any premium that was there.

56:25 So, so so hard to implement in a way that some

56:29 of these other factors are not terribly difficult to implement.

56:33 Yeah, for relatively speaking, they're they're lower turnover.

56:36 They're not super complicated.

56:38 Like all the metrics are fairly straightforward.

56:40 But even momentum, like there are a bunch of different ways you can measure it.

56:44 You don't know which measure is the best one.

56:46 We can say the same thing for for value and profitability too, though.

56:49 They're still I mean every layer of this is still hotly debated.

56:53 You could go and find five different academics who have published papers on why

56:57 their measure of value or their measure

56:59 of profitability is better than every everybody else's.

57:02 I mean they would they would all have great cases and great evidence.

57:06 It's one of the one line that I love

57:08 in this type of stuff is that for every PhD, there's an equal and opposite PhD.

57:13 You can find me someone to to tell

57:14 me why their asset pricing model or their factor

57:17 or their their product is the best and I can find somebody just as smart,

57:20 just as qualified to disagree with them.

57:23 It's one of the funny things about our our space.

57:25 And nobody will know if they're right until 30 years from now.

57:29 And then we'll look back and say, "Oh yeah, well,

57:30 that guy was right." But then we won't know if

57:33 they're going to be the right for the next 30 years.

57:35 It's so true.

57:37 Yeah, we we we work in a funny a funny space.

57:40 You can kind of just people in finance can just say stuff

57:43 and people can choose to believe them because they have a good story,

57:47 um but we won't know if what they're saying

57:49 is right or true until many years in the future.

57:53 And everyone will have probably forgotten about the whole debate

57:55 by then anyway and chased whatever new hot trend is out there.

58:00 That's right.

58:00 Uh okay, anyway.

58:02 Uh I do want to bring this back to practical relevance.

58:07 So, we're talking about these asset pricing factors.

58:09 We're talking about the idea that they

58:11 may have higher expected returns and that investors

58:14 could potentially tilt toward them to increase

58:17 the expected returns of their equity portfolio.

58:20 So, owning the market capitalization-weighted market is is

58:23 giving you exposure to the market risk premium.

58:27 But, as you said, Dan,

58:28 you can tilt toward other factors to potentially increase your expected returns,

58:33 at least if you believe what these models suggest.

58:36 Um now, how do you do that?

58:38 Uh it's, you know, for for individual to do it by selecting individual stocks

58:44 be probably kind of arduous and tricky

58:46 and probably not worth probably not worth it.

58:49 But, there are fund companies that are

58:52 using this research to build diversified portfolios.

58:54 They're they're kind of like low-cost index funds

58:56 that deliver exposure to more than just market risk.

58:58 They're usually not technically index funds because they don't track an index.

59:03 But, functionally, they're very similar, broadly diversified,

59:05 low turnover, low cost, all that stuff.

59:07 Uh Dimensional Fund Advisors has a long

59:09 history of implementing factor investing research.

59:12 They started building products before any of this research came out.

59:16 I think their first product launched around the same

59:19 time that the paper documenting the size anomaly came out.

59:23 Um so, they weren't I mean,

59:25 factor investing wasn't a thing when Dimensional launched.

59:28 They were just trying to build a small cap uh index basically,

59:32 although it wasn't actually an index.

59:34 They were trying to build a small cap product that institutions

59:36 could use to complement their often large cap tilted portfolios.

59:40 And then this academic research starts coming out

59:42 and they already had connections to the academic community,

59:44 so they continued to implement the academic research as it came out.

59:48 So, they started with the size portfolio.

59:51 Uh they eventually started using value once that research started

59:55 to come out and they've continued to implement things like profitability.

59:58 They do have a way of implementing

59:59 momentum actually that's not super high turnover.

1:00:01 They use it as part of their their trading process.

1:00:04 Uh but they're very connected to the academic roots of this idea.

1:00:10 Eugene Fama, one of the paper that we're talking about's co-authors,

1:00:15 was a founding director and remains on the board today.

1:00:17 Ken French, the other co-author,

1:00:19 has long-standing connections to Dimensional and Dimensional's track

1:00:23 record in the long run has been pretty good.

1:00:26 They've struggled a bit recently in the US in particular because large cap

1:00:30 growth has just done so well there and Dimensional tilts away from that.

1:00:34 But in international markets and even in Canada,

1:00:35 they've done they've done pretty well.

1:00:38 Uh one catch is that Dimensional used to be

1:00:41 only available through financial advisors for many many years.

1:00:45 Uh they're now available as ETFs in the US market, not in Canada though.

1:00:49 But what happened more recently is that Avantis Investors,

1:00:52 which is a Dimensional competitor basically,

1:00:55 some folks who had previously been at Dimensional

1:00:57 for many years left and started this competitor called Avantis.

1:01:01 Um they they launched similar products.

1:01:03 They launched ETFs in the US before Dimensional did.

1:01:06 And they have now just launched a few weeks ago when this episode comes out,

1:01:12 they've launched ETFs in Canada in partnership with CIBC.

1:01:16 And we did have the CIO of Avantis, Eduardo Repetto,

1:01:21 on this podcast back in March talking in lots

1:01:24 of entertaining detail about those Canadian products and and I

1:01:28 think we probably will do an an episode at some

1:01:30 point doing a deeper dive on their Canadian products.

1:01:34 I I do want to mention real quick

1:01:36 PWL we do use Dimensional funds pretty extensively.

1:01:39 We're not paid by Dimensional or Avantis and we're

1:01:42 not being paid to mention them in this episode.

1:01:45 We don't We don't do that kind of thing.

1:01:48 Okay, so what does this all mean for you the investor?

1:01:52 Evidence suggests that long-term expected returns are driven by specific

1:01:56 systematic exposures exposures that we know about today due

1:02:01 to decades of academic research implying that investors may achieve

1:02:04 higher expected returns by tilting toward certain types of of stocks.

1:02:11 Which stocks and knowing how to do it effectively and efficiently

1:02:13 at a low cost is a whole other whole other kettle of fish.

1:02:17 Uh but there are fund companies like Dimensional

1:02:19 Fund Advisors and Avantis Investors that that do that.

1:02:22 They create low-cost broadly diversified investment portfolios

1:02:24 that that are specifically built to take the academic

1:02:28 theory that we've been talking about multi-factor

1:02:30 asset pricing and apply it to investment portfolios.

1:02:34 Um so like I said at the beginning of this episode

1:02:37 this seminal paper from Fama and French really forms the foundation

1:02:40 for a big part of how I think about investing

1:02:43 and as I mentioned earlier it's part of how I arrived at PWL.

1:02:48 Uh and it's also part of the reason that PWL Capital has

1:02:51 been using Dimensional since we helped bring them to Canada back in 2003.

1:02:56 That's a whole other story.

1:02:57 That was like Cameron and a few other folks came across Larry Swedroe's writing

1:03:01 on this stuff and thought that was pretty cool and it ended up getting connected

1:03:06 with Dimensional and ended up There there is a There is a funny story in there

1:03:09 somewhere about David Booth came to Canada to talk to PWL and I think they

1:03:14 almost shut us down because we had a research department and then they assumed

1:03:19 that that meant that it was like individual

1:03:20 security research and like no no no we

1:03:22 don't want to partner with you guys but then they kind of figured out that we

1:03:25 were using index funds and it wasn't that kind of research that we were doing.

1:03:29 Anyway.

1:03:30 Yeah, it's interesting, right?

1:03:31 Because back back then I mean PWL was not that big.

1:03:34 I mean today it wouldn't surprise me if a big fund company said okay, you know,

1:03:38 we're going to listen to what you have to say

1:03:39 and maybe uh if this is a big US company

1:03:42 maybe bring the product to Canada but in 2003 it

1:03:45 was a pretty small company and uh I think you

1:03:48 know PWL sort of transitioned from being stock pickers to traditional

1:03:53 index fund investing and before that but even in 2003

1:03:57 there wasn't a whole heck of a lot of product

1:03:59 out there for people who wanted to do you know,

1:04:02 even the most basic index fund investing in Canada.

1:04:05 So that was really early days.

1:04:07 It would have been fun to have been in those rooms and Oh, yeah.

1:04:09 and heard those discussions.

1:04:11 Oh, I'm sure there are some cool stories.

1:04:13 That's something I've never really talked to Cameron in too much detail about.

1:04:16 I know the kind of the broad strokes of the story but I mean yeah,

1:04:18 I'm sure there are some some neat stories in there.

1:04:21 But you're right.

1:04:22 I I haven't seen the models that they were using back then uh but I

1:04:26 know they were using some US listed small

1:04:28 cap value ETFs to get some factor exposure.

1:04:31 They're using some uh some index index funds,

1:04:35 probably some index mutual funds I would I would imagine.

1:04:38 I think those were a little bit more more common.

1:04:41 Uh anyway.

1:04:41 I shares well, I guess it was was it I shares at that point?

1:04:45 I guess it was um Barclays, right?

1:04:47 Like uh that was the original Canadian owner.

1:04:50 And so like in the late '90s, yeah,

1:04:52 there was some ETF availability in Canada but the pickings were pretty slim uh

1:04:58 and you didn't have a whole lot of uh flexibility with what you were buying.

1:05:04 Yeah.

1:05:05 Yeah.

1:05:06 Uh okay.

1:05:06 So if if if uh listeners want to learn more about

1:05:09 how we apply this thinking to our clients' portfolios at PWL.

1:05:13 Or or not, uh which we did an episode on discussing

1:05:16 how Dan and I think about that a little bit differently.

1:05:18 You can use the the link in the episode description to book

1:05:21 a time to chat with one of our folks here at at PWL.

1:05:27 All right, that concludes our discussion

1:05:29 on the finance paper that changed everything.

1:05:34 All right.

1:05:37 Are we going to what we've been calling the aftershow?

1:05:40 Yeah, yeah, yeah.

1:05:40 Let's let's go Let's go to the aftershow.

1:05:44 Uh I think we just have one one review in there today.

1:05:48 It's a long one, though.

1:05:50 Mhm.

1:05:51 Do you want to read it, Dan?

1:05:52 I did a lot of talking today.

1:05:54 All right, do we need to read this disclaimer first?

1:05:56 true, true.

1:05:57 I'll do the disclaimer, you can do the review.

1:06:00 good.

1:06:00 We have a review from Apple Podcasts to read.

1:06:02 Under SEC regulations, we are required to disclose whether a review,

1:06:06 which may be interpreted as a testimonial, was left by a client,

1:06:09 whether any direct or indirect compensation was paid for the review,

1:06:13 or whether there are any conflict of interest related to the review.

1:06:15 As reviews are generally anonymous, including this one, yep.

1:06:20 Dave Dave down under.

1:06:23 That is an anonymous review.

1:06:24 It's It It could be.

1:06:26 Um Uh we are unable to identify if the reviewer

1:06:29 is a client or disclose any such conflict of interest.

1:06:31 They're also in Australia, so I highly doubt I think we would know.

1:06:34 Unlikely to be our clients, yes.

1:06:36 Yeah.

1:06:36 Yeah.

1:06:37 All right, here's the review.

1:06:38 It says, "Investing has been solved.

1:06:40 How this changed my money mindset." "This podcast

1:06:44 has completely changed how I think about investing,

1:06:46 money, and many other things.

1:06:48 Having been told that investing is always gambling growing up,

1:06:52 this podcast has shown me that this does

1:06:54 not need to be the case by providing evidence-based,

1:06:57 level-headed, and entertaining content.

1:07:00 While some episodes are more Canada-focused than others,

1:07:03 the principles can be applied by any retail investor around the world.

1:07:06 I especially like when Ben, his co-hosts,

1:07:09 and guests show over and over again that investing has

1:07:12 basically been solved and most people simply ask the wrong questions.

1:07:17 I do hold a globally diversified market cap

1:07:20 weighted index fund portfolio with a slight factor tilt,

1:07:23 but more importantly, I've created my own financial plan,

1:07:27 admittedly a pretty straightforward situation.

1:07:30 Starting with defining goals over over

1:07:33 creating an investment philosophy that I can

1:07:35 stick with to eventually modeling outcomes

1:07:38 to figure out the required savings rate, the risks I'm able, willing,

1:07:42 and need to take along the way to meet my future

1:07:45 spending goals has made me very confident in my decision-making.

1:07:49 Software engineering background helps, haha.

1:07:52 As a renter who has not missed a single DCA

1:07:55 contribution since I started investing 4 and 1/2 years ago,

1:07:59 and in fact has managed to increase

1:08:00 contributions consistently instead of letting lifestyle creep win,

1:08:05 I also love the episodes about the rent versus buy decision.

1:08:08 Not that it has convinced any of my friends

1:08:10 that I'm not a complete idiot by buying a property,

1:08:13 but hey, you can't win them all.

1:08:15 Thank you for all your work and all best wishes from the land

1:08:18 down under from Dave Down Under Stand from Australia on iTunes.

1:08:25 Very nice review.

1:08:26 Mhm.

1:08:26 It seems like Dave has been been paying attention.

1:08:30 It sounds like he's uh he's maybe one of these people who've gone back

1:08:33 and listened to all 400 episodes uh over the course of a few weeks.

1:08:37 We've been talking about or hearing about a few people

1:08:39 who've done that and uh that is quite a marathon.

1:08:42 Yeah, it is it is wild.

1:08:44 So.

1:08:46 Uh all right.

1:08:47 I I think that's it.

1:08:48 I I I realized I we never said the name

1:08:49 of the paper that we talked about in this episode.

1:08:51 It's common we'll we'll put it in the in the thumbnail or something,

1:08:54 but it's uh it's common risk factors in the returns on stocks and bonds.

1:08:59 That's the name of the paper.

1:09:00 the and now you know what it's called.

1:09:02 There we go.

1:09:03 All right.

1:09:03 Anything else, Ben?

1:09:04 No, we're good.

1:09:06 All right.

1:09:06 Thanks everyone for listening.

1:09:07 See you next time.

1:09:12 Hey everyone.

1:09:13 It's producer Matt.

1:09:14 Thank you so much for tuning in to this week's episode.

1:09:17 Before we sign off, here's the disclaimer you've been waiting for.

1:09:21 Portfolio management and brokerage services in Canada

1:09:24 are offered exclusively by PWL Capital,

1:09:26 which is regulated by the Canadian Investment Regulatory Organization

1:09:30 and is a member of the Canadian Investor Protection Fund.

1:09:33 Investment advisory services in the United States of America

1:09:36 are offered exclusively by One Digital Investment Advisers LLC.

1:09:40 One Digital and PWL Capital are affiliated entities

1:09:44 and they mostly get on really well with each other.

1:09:47 However, each company has financial responsibility

1:09:50 for only its own products and services.

1:09:53 Nothing herein constitutes an offer or solicitation to buy or sell any security.

1:09:58 Occasionally, we tell you not to buy crappy investments in the first place,

1:10:02 but that's not the same thing as telling you to sell them.

1:10:05 This communication is distributed for informational purposes only.

1:10:09 The information contained herein has been

1:10:11 derived from sources believed to be truthy, but not necessarily accurate.

1:10:16 We really do try, but we can't make any guarantees.

1:10:20 Even if nothing we say is fundamentally wrong, it might not be the whole story.

1:10:24 Furthermore, nothing herein should be construed as investment,

1:10:28 tax, or legal advice.

1:10:30 Even though we call the podcast Your Weekly

1:10:32 Reality Check on Sensible Investing and Financial Decision Making,

1:10:36 you shouldn't rely on us when making actual decisions, only hypothetical ones.

1:10:41 Different types of investments and investment strategies have varying

1:10:44 degrees of risk and are not suitable for all investors.

1:10:47 You should consult with a professional advisor to see how

1:10:50 the information contained herein may apply to your individual circumstances.

1:10:54 It might not apply at all.

1:10:56 Honestly, you can probably ignore most of it.

1:10:59 All market indices discussed are unmanaged,

1:11:02 do not incur management fees, and cannot be invested in directly,

1:11:06 which is a shame because it would be awesome if you could.

1:11:09 All investing involves risk of loss, including loss of money,

1:11:13 loss of sleep, loss of hair, and loss of reputation.

1:11:17 Nothing herein should be construed as a guarantee

1:11:20 of any specific outcome or profit.

1:11:23 Past performance is not indicative of or a guarantee of future results.

1:11:27 If it were, it would be much easier to be a Leafs fan.

1:11:31 All statements and opinions presented herein are those of the individual host

1:11:35 and our guests and are current only

1:11:38 as of this communication's original publication date.

1:11:41 No one should be surprised if they have all since recanted.

1:11:45 Neither One Digital nor PWL Capital has any obligation to provide

1:11:49 revised statements and or opinions in the event of changed circumstances.

1:11:53 See you next time.