December 17, 2020

EMH, Expected Returns, Factor Investing, Fundamental Topics, Historical Data, Index Funds, Portfolio Management, Portfolio Theory

Episode 129: Five Factor Investing with ETFs

Angelica Montagano

December 17, 2020

EMH, Expected Returns, Factor Investing, Fundamental Topics, Historical Data, Index Funds, Portfolio Management, Portfolio Theory

After months of research, number-crunching, and receiving listener requests on the subject, today’s episode is devoted to introducing our new model ETF portfolios — which promise to offer a smoother ride to getting reliable returns. We open our conversation with a financial news roundup and by touching on our book of the week. We then dive into the theory behind our model by first exploring how market assets are priced. We discuss historical views on asset pricing models before looking at what academia has done to overcome challenges to the idea of market efficiency. Host Benjamin Felix methodically shows how our model addresses the five systematic risk factors that are included in the Fama-French Five-Factor Model. From emerging markets to stock size, we share insights into what our model accounts for and how this should impact your portfolio distribution and premium expectations. After reflecting on how factor-loaded indexes get higher returns without extra risk, we talk about the ETFs that we use for factor exposure, as well as how you can apply our findings to your portfolio. We round-off today’s show by chatting about the latest bad financial advice. Tune in to hear more about our findings in this, our last episode of 2020.

Key Points From This Episode:

We share community and listener feedback and what you can expect from the show in 2021. [0:00:15]
Hear about The Almanack of Naval Ravikant, our book of the week. [0:04:35]
Relooking at the drive towards personalized portfolios. [0:07:28]
Insights into S&P 500 stocks having a greater dividend yield than the US Treasury. [0:08:52]
How ETFs are coming to dominate Wall Street. [0:09:56]
Dying without a will; exploring the case of Zappos CEO Tony Hsieh. [0:11:05]
Introducing today’s portfolio topic — our new model portfolios, and why index funds make sense. [0:12:29]
The risks that inform how the market prices assets and your expected returns. [0:18:00]
How academics have addressed the joint hypothesis and the brilliance of the Fama-French Five-Factor Model. [0:22:18]
The predictive power of the Fama-French Five-Factor Model. [0:30:25]
Challenges to the Fama-French Model, what it accounts for, and how links to our new model portfolios. [0:31:20]
How our model weighs the value of different markets and stock sizes. [0:36:52]
Comparing the returns of a factor-loaded index with the US total market index. [0:40:44]
Answering the question — what’s special about dividend growth investing? [0:42:15]
The role of persistence in being a factor-led investor. [0:45:50]
How our model increases the reliability of your investing outcomes compared to historical data. [0:50:10]
How to apply all the information presented in this episode. [0:54:38]
Hear about the ETFs that we use for our factor exposure. [0:58:10]
Details on when you can access our upcoming paper. [01:02:25]
This week’s bad financial advice; invest in the right region to drive performance. [01:04:05]

Read the Transcript:

Cameron Passmore: This week we'll kick it off with a book that I just finished and it's a book that's been popping up everywhere in my Twitter feed so I figured, well maybe it's some sort of message to me that I should be reading it. It's called Almanac of Naval, Ravikant. Naval was an Indian American entrepreneur and investor, and he's actually the co-founder of AngelList and he's invested in, I think over 200 early-stage companies such as Uber, Foursquare and Twitter. I did not know a lot about Naval beforehand but this book is not a book about him and this is what's pretty cool. It's actually a collection of his thoughts and philosophies that he's put out on Twitter over the past decade and in podcasts interviews but what is really cool to me is that it was actually assembled by a guy named Eric Jorgensen who's a product strategist at Zaarly. Anyways, Eric was taken by all these things that Naval has put out there that he decided to assemble it into, as he described, something more permanent because he felt that the thinking of Naval was so good that deserve to be put into a durable long-term format or a book.

On top of that, Eric went and put this out for free. You can actually download the PDF and it's in all kinds of different formats. I actually bought it on my Kindle but you can download it for free at navalmanack.com. That's N-A-V-A-L-M-A-N-A-C-K.com. It's all there for free and I really, really enjoyed it. It's an assembly so you see Eric posts all kinds of different tweets that he's given that he links and quotes from different podcasts. It answers all kinds of questions that were posted and Naval like how important is luck? Is there value in networking? What is your definition of retirement? What is happiness? And a lot of it is around happiness and wealth. It's also full of what I thought were interesting quotes and thoughts and here's some examples for you.

In any situation in life you always have three options. You can change it, you can accept it, or you can leave it. There's another one. To make an original contribution you have to be irrationally obsessed with something. That's a good one for us. Science is the study of the truth that actually changes the world. He defines the modern struggle as lone individuals summoning inhuman willpower, fasting, meditation, and exercising up against armies of scientists and statisticians weaponizing abundant food, screens and medicine into junk food, click bait news, endless games and addictive drugs. I can't say enough about it. I actually got it for my kids for Christmas. We'll see if James listens to this before it actually gets posted before Christmas but I really enjoy it. I liked how it was laid out. It's a very fast read. It's an interesting read and you can pick it up at any point and dive in. There's your book recommendation for the week.

In other news. There's a tweet last week from Nature S. He says this is more a question to you based on some of the feedback we've been getting on the community board. Nature S posted last week that table stakes from major brokerage firms within the next few years will be fractional share ownership, direct indexing, cryptocurrency-

Ben Felix: Custody of cryptocurrency.

Cameron Passmore: Custody of cryptocurrency. Easy custody of cryptocurrency exactly. I thought it's a neat link going back to our interview with Josh and Brian I think three weeks ago or so about the drive towards personal biases, and personal desires, and portfolio versus the evidence, or as Josh called it the orthodoxy. Thought it was neat how Nature S is now saying that exactly what they were arguing that the ability to do what feels right for you is going to be easier in the brokerage world.

Ben Felix: I think if we define table-stakes is what you need to be able to sell more stuff and increase margins then for sure fractional shares I think are important regardless. Direct indexing though that's... From a practical perspective and investors perspective, I don't think it's particularly useful, especially after costs because we know it's going to cost a lot more to have a direct index portfolio than it is to have a cap weighted ETF portfolio. And that's, before you start factoring in factor tilts into your direct index portfolio but if you want to make money as a brokerage, then yeah of course you need to have that. Crypto custody is the same thing. That stuff is not going away.

Cameron Passmore: No. Another data point I wouldn't mind your feedback on was the percentage of S&P 500 stocks with a dividend yield greater than the 10 year U.S. treasury yield. Currently it's 63%, which is the highest by far in this table. It goes back to 1990. The average ratio is just under 20%. Any thoughts on that?

Ben Felix: Thoughts from what perspective?

Cameron Passmore: What do you think is driving? Do you think is the low level of the ten-year treasury? Do you think it's... I guess we don't know what's really going on behind the scenes. Is it the fact that a lot of dividends paying companies might be in the smaller mid cap range, which have not had the kind of rally going-

Ben Felix: No, I mean prices are... This is yield, right?

Cameron Passmore: Yeah.

Ben Felix: Prices are high. Stock prices are high. Yields are low. U.S. companies are paying and what they're comparing this to 1990 U.S. companies have been paying less and less in dividends.

Cameron Passmore: The last story for you. Article in bloomberg.com called crushing a $5.2 trillion party. Debut ETF issuers hit a record. There's been 14 new issues of ETFs this year. This is not ETFs. This is companies issuing ETFs. There's 10 last year in 2019 so we have 14 far this year and make the argument the article that is based on new rules for active funds that made the launches easier is one of the reasons behind this. It also referenced the most dramatic new entrant being dimensional fund advisors who recently converted six of its tax managed mutual funds into ETFs. By doing that, it launched them into 11th place overall in the ETF landscape with something in the low $20 billion range of ETF assets. Now by comparison, the largest participant or the largest company in the ETF market space is BlackRock at $1.8 trillion, Vanguard at 1.1 trillion and State Street comes into number three at 700 billion. ETFs have taken into a record $90 billion in November, which they say is an unheard of $5 billion a day pace. Our year to date flows are now up to $427 billion, which is about $35 billion away from an annual record. It's been huge in the ETF market space.

Now you want to talk about Tony Hsieh, Zappos founder who passed away November 27th complications from smoke inhalation from a home fire a week before that.

Ben Felix: The only thing that jumped out to me as interesting in that story was the fact that he died and said he died without a will. I heard this in passing. I think they... One of the podcasts I listen to I can't remember which one, they aired the old episode where they'd had him on. Might've been Freakonomics or Planet Money or something. I took a quick read of the story and that jumped out at me as pretty interesting. This ultra wealthy guy who has been very successful in business and he passes away unexpectedly without a will. In most cases, and I don't know if that's true, in a lot of cases people die without expecting to die they don't have time to get everything in order. It speaks to the importance of getting all that stuff taken care of.

Cameron Passmore: He had a net worth over $840 USD-

Ben Felix: It's amazing to think about with that level of wealth, it's still not have all the proper planning in place.

Cameron Passmore: And being in business, you think of all the lawyers that would have been around his situation for years and it was never done.

Ben Felix: Maybe he didn't want to do it. Who knows what the actual full backstory is but I found it very interesting that that's how it happened for him.

Cameron Passmore: The family has asked a judge to name his father and brother as special administrators of the estate but the bottom line is that stuff happens and you should always have a will in place. Shall we jump into the feature topic this week?

Ben Felix: Yeah let's do it. This week we're going to combine not because the content spans both planning and portfolio management just because we have a bigger topic for the portfolio management piece. We're going to skip, I guess that's a better way to say it. We're not going to combine them, we're going to skip the financial planning topic but I think people will appreciate the portfolio topic that we have lined up. It's really about the new model portfolios, which people will be excited to hear. I also thought it would be good to run through the main talking points associated with the paper that we're going to put out at the same time as the model portfolios so the thinking behind all of it, which also ends up serving as a pretty good refresher of all of the main concepts behind index investing and stuff like that.

Cameron Passmore: Well that's the best part about it to me is that it's all in one spot. All the rationale, the papers, the story behind, the reasons for this portfolio construction.

Ben Felix: Yeah. So index investing, which is table stakes when you enter this world of evidence based investing. Index funds are I think pretty well understood and they're sensible as investments if markets are efficient. I think people hear that term, I don't know if everybody knows exactly what it means as Eugene Fama stated it. I'm going to read a quote from the paper where he introduced the concept of market efficiency. Fama said, and this is in 1970. "In general terms, the ideal is..." And keep in mind he said this is the ideal. He's not saying markets are efficient, which Fama has never actually argued. Back to the quote. "In general terms, the ideal is a market in which prices provide accurate signals for resource allocation. That is a market in which firms can make production investment decisions and investors can choose among securities that represent ownership of firms activities under the assumption that security prices at any time fully reflect all available information. A market in which price is always fully reflect available information is called efficient."

Now Fama also says that real markets can't be perfectly efficient in the theoretical sense. They can only approach perfect efficiency but this hypothesis and this model became as many people know the framework for much of the empirical research that's been done in finance since then. Now from the investors perspective, and Fama said this explicitly in the quote that we just read, if the market is efficient than prices; stock prices, bond prices, whatever you want, security prices, they contain information about the expected returns of a security at a point in time. Then the other thing that follows from that is that prices will change based on new information, which cannot be predicted obviously just by nature of being new information.

Given those criteria, or constraints, whatever you want to call them, in an efficient market the only way to earn extra returns is by taking more risk because security prices move randomly you can't outsmart the market, all the main talking points that people might be familiar with. Now if you can get excess returns without taking on extra risk that's known as alpha. A traditional active manager their goal is to combine picking the right stocks at the right time and timing the market as a whole in an effort to excess risk adjusted returns known as alpha. Now we also know that empirically active management is not done so well. The SPIVA reports are often cited as evidence of that. They're not my favorite source just because of the way that they construct the data. Did you know they use Series A funds, at least for the Canadian report?

Cameron Passmore: Yeah. We talked about that, didn't we?

Ben Felix: Yeah, I think so. You've got an extra 1% commission built into the fund returns then you're comparing it to an index with no fees. It no way a fair comparison.

Cameron Passmore: Did you not reach out to them to mention that?

Ben Felix: I did and they said that they would think about addressing it in a future report. I never followed up so I don't know if they did anything with that. But I don't think it would change the distribution too much. You'd still have most active funds underperforming. But anyway fortunately there are a couple of papers that were the journal of finance that address the same topic like the failure of active fund management. There is Carhart in 1997, Fama-French 2010 documented the failure of active fund managers to consistently beat the market, which supports the idea that markets are efficient because if security prices move randomly, managers should be able to consistently beat the market. Now there are some managers like Jim Simons who we've talked about and Buffet for a period of time, although he's trailed the market for about 20 years now but there are investors who are outliers and you can argue that's evidence of market inefficiency.

Cameron Passmore: It's still the same question. How do you identify them ahead of their own performance?

Ben Felix: Correct.

Cameron Passmore: What signals would you have used and can they maintain it? As we've said with Jim Simons for example, they've kept that fund forever at $10 billion of assets and every year they flush out all the gains to keep it at $10 billion and you can't buy into it-

Ben Felix: Yeah you can't buy into it anyway, there's no point of talking about it anyway. EQR will take their money but Jim Simons won't. At this point where we've talked through so far markets are efficient theoretically and the empirical failure of active management is evidence supporting that therefore index funds make sense. Nice and easy we could stop here and many people do, which is perfectly fine and there's been a lot really interesting discussion in the Rational Reminder rush reminder, the community discussion about that exact question. Should you actually factor tilt your portfolio? Should you try and get exposure to the other risks in the market? Now I'm getting ahead of myself since the introduction of the efficient market hypothesis, the field... The study of asset pricing has become this massive, massive field of study and there have been many systematic risks identified that seemed to affect the way the market prices assets.

We said earlier that the market should reflect the riskiness of an asset. The market price should reflect the riskiness of an asset but we didn't specify what risk specifically it's reflecting. Now that exact question has not necessary been answered definitively but there are some pretty strong theories and models showing which types of risk the market does include in prices. Now why is this relevant to an investor? It's relevant because the market risk factor, which is what you get by owning a market capitalization-weighted index fund. That's only one of multiple risks that currently there are really strong theories and really strong supporting evidence suggesting that there are multiple sources of expected returns. Expected returns is another loaded word, which we're going to talk more about in a second but to keep with where we've gotten far, there are other types of risks that the market systematically includes in stock prices. Now having exposure to just one of those, if you just have exposure to the market risk, that's delivering one source of expected returns. It's giving you one risk, keeping in mind in this context taking these systematic risks is actually a good thing. It's not a scary risk. It's a risk that you're taking with the expectation of a positive long-term return.

But you can access more than one risk. And this becomes quite interesting from the perspective of portfolio management for a lot of different reasons, which we're going to talk about a lot of them. There was a paper in the Journal of Portfolio Management a while ago. I don't remember the... I think it's called The Death of Diversification Has Been Greatly Exaggerated, and they argued in that paper that correlations across geographic regions in the stock market have increased over time as the world's become more globalized and all that stuff. The independent systematic risks that we're going to talk more about they have not increased in correlations. That paper that I'm talking about they actually argue that factor diversification, so getting exposure to more than just one risk is more important at this point then geographic diversification, which is a pretty interesting argument.

What we're going to talk about and what we included in the model portfolios that we're addressing here are five systematic risk factors that are included in the Fama-French 5-factor model. This is Eugene Fama and Ken French, pretty big deals in the empirical and theoretical finance space and they have an asset pricing model that includes five risk factors, the market being one of them but also four others. We're going to talk through what those are. I mentioned expected returns being a loaded word but it's an important concept. When you buy a stock... This is theory. When you buy a stock, theoretically what you're buying is a right to a company's future profits. Now this part can get confusing for a lot of people because where did the returns come from? Expected returns. Returns that you expect to receive come from how much of a discount you apply to those expected future profits. For a riskier asset you'd apply a bigger discount. You're paying less for the same amount of future profits if those future profits are riskier. Now that discount rate is your expected return.

Cameron Passmore: That greater the discount to lower the price you're paying.

Ben Felix: Right. A firm can be really profitable and not produce very good returns if you pay a lot for those profits. We see that with growth stocks. And we've talked about that quite extensively in past podcast episodes and on the Common Sense Investing YouTube channel. It's really what you pay for the asset that dictates your expected return. That discount rate concept is what ties as a prices to expected returns. I mentioned asset pricing models. I made a reference to the fact that we're talking about one of them that Fama and French have come up with. All asset pricing models are doing is trying to figure out which risks are relevant to the pricing of stocks. In this case, we're using one and there's no single correct model. I think it starts with deciding what framework you want to use to think about markets and from that will stem the type of model that you want to apply to looking at portfolio returns and stock returns.

We're using the one that Fama and French have developed, which lines up with the way that we tend to think about markets in general but it is crazy to think about too that you could just say, "Well, no. I believe markets work this way." And therefore you would use a completely different factor model, which would completely change your portfolio construction. There's no way to prove which model is the right one and that's actually one of the things Fama has written about in his past papers. This idea, the joint hypothesis problem, which basically says that a test of market efficiency using an asset pricing model is jointly a test of market efficiency and a test of the asset pricing model. That's tricky because you can prove as we're going to talk about in a moment, you can prove that the market is inefficient using that asset pricing model. You can show based on the risks in my asset pricing model, the market is not pricing securities properly. Boom! Market inefficiency. But someone else can say, "Well, no. Your model is wrong." And nobody can prove nobody what's right so it's called the joint hypothesis problem and this is a real issue.

Because of that, you can't prove market efficiency with an asset pricing model no matter how good the model is. But it does started to get pretty interesting when a model like the Fama-French 5-factor model explains the vast majority of return differences between two portfolios. It still can't be proof of market efficiency but it starts to get to a point where it's like there's not a whole lot of return differences that are left unexplained. Anyway, the original risk factor, and this is a bit of a quick history lesson on asset pricing models but the very first risk factor was the market, the equity risk premium market beta as it's known. The first time that expected returns were related to risk to market risk was through the capital asset pricing model. This is classic finance 101 capital asset pricing model. Now cap weighted index fund, total market index fund should have a market beta of one always. If we took a portfolio that was 50% market and 50% cash, it should have a beta of 0.5. If two portfolios, and this is my comment a minute ago about explaining differences in returns. If two portfolios have the same market beta but have differences in returns through a CAPM lens, that difference in returns might be portfolio managers skill, or it might be related to some as yet unidentified factor.

Cameron Passmore: At that time.

Ben Felix: Correct. Hadn't been identified yet and that ties back to the joint hypothesis problem. Now, a portfolio that... We mentioned this earlier with alpha, portfolio that delivers higher returns. Those two portfolios we were comparing a second ago, if one of them has higher returns given it's market beta, that's great for the investor. That's an excess risk adjusted return. You're getting an extra returns without taking on additional risk. Now the CAPM, that was in the early 60s I think when CAPM came out. By 1972, Fischer Black had come up with a paper showing some flaws and then Rolf Banz in 1981 came up with his paper on the relationship between return and market value of common stock. This is the first time that someone published data showing that small stocks from a CAPM perspective had persistent alpha. These stocks were beating the market without taking on extra risk as measured by the risk of the market as a whole.

Then in 1985 there was another paper that documented the same thing for value stocks. The title of their paper was actually persuasive evidence of market inefficiency. It was pretty interesting to look at how they wanted to title their paper. But then Fama and French speaking to the joint hypothesis problem, they came out in 1992 with a paper that said, okay we've got small stocks with these persistent alphas. We've got value stocks with these persistent alphas. Maybe we just need to account for the risk of those types of things in our asset pricing model. They came out with their now fairly famous 3-factor model where they said, "We think asset prices reflect the risk of the market but also the risk of company size where small stocks are riskier than big stocks and also the risk of relative price where cheap stocks are riskier than more expensive stocks. Value stocks at riskier than growth stocks. When they put those independent risks into a model together, a lot of those anomalies went away, which I guess is obvious. They built a model to make the anomalies go away. But the interesting part from a portfolio evaluation perspective, is that the ability of the model went from about 65% or so with the CAPM in terms of ability of explaining differences in returns to about 90% with the 3-factor model.

All of a sudden the explanatory power of the model becomes extremely strong. One of the really interesting things to think about with the 3-factor model is that there was a loose concept that it was related to risk but there was no real model to say why that's true at the time. Again, kind of makes sense small stocks to the riskier. Yeah it kind of makes sense value stocks be risky but there was no all encompassing evaluation framework to say how those things are related, or should relate to each other. Then later research in 2004 and in 2013, people came out with papers showing anomalies in the 3-factor model. Now we've gone from the small value being anomalies for CAPM and the 3-factor model solving it and then investment. Companies that grow their assets aggressively tended to have lower returns than companies that grew them more conservatively. Then companies with more robust profitability tended to outperform companies with weaker profitability and those alphas in the 3-factor model were persistent.

Again, we have this asset pricing problem and Fama and French more recently, in 2015 I think, came out with a new paper where they introduced their 5-factor as a pricing model. The thing that I love about the 5-factor model, if you can love a model, is that it's rooted in evaluation theory. I mentioned with the 3-factor model it was more of an empirical model where we had these empirical observations when you incorporate them into an asset pricing model that makes a lot of anomalies go away. Cool, but with the 5-factor model, they took the theoretical evaluation framework of the dividend discount model where a company's value is theoretically the present value of its expected future dividends discounted at whatever discount rate. They took that concept, applied the theory of dividend irrelevance from Miller and Modigliani and when you transform the dividend discount model to include dividend irrelevance, we talked about this on a recent episode but basically if dividends are irrelevant, the dividend discount model turns into profitability minus investment in place of dividends otherwise the equation stays the same. But that ties together all of the elements that go into the 5-factor model.

The cool thing about that is now we have there's empirical evidence that these anomalies exist. You've got low price stocks, you've got more profitable stocks and you've got stocks that they grow their assets aggressively. Those are all independently discovered empirical observations but then you've got this unifying theory that ties them all together to justify their use together in a model, which I think is pretty cool. And that's one of the tricky things with multi-factor asset pricing models is that nobody tells you. Nobody has the answer for which factors you should include in the model. I alluded to that earlier when we're talking about how do you pick a model? That's the cool thing about this one is that it starts with the rational evaluation framework and then plugs in, or spits out the factors that would make sense to include based on that. This model has a few predictions, which we've already talked about but the model predicts, and keeping in mind the empirical data are one thing but the model predicts the same thing that we observe empirically. The model predicts that cheaper stocks will outperform more expensive stocks. Again, we see that empirically. More profitable stocks will outperform less profitable stocks and then stocks that invest conservatively will beat stocks that invest aggressively.

The next challenge was measuring those things. This came before the 5-factor model. I should've said at first I guess, but measuring profitability and investment because it's not current profitability you have to measure, its future profitability because we're discounting future profits. Nobody really knew how to measure that but Novy-Marx in 2012 and there's another paper in 2013 that showed how those things could be measured with reasonable reliability to actually be used in an asset pricing model. All that culminates in the 5-factor asset pricing model that we're now talking about, how do you build a portfolio based on that? We know these five systematic risks are important to asset pricing therefore to expected returns, what can we do as investors to take advantage of that? Which is another question at hand.

Company size is a really interesting one. In all of these models you may notice that it didn't show up in the... We didn't show the actual equation but in the dividend discount model, there's no variable for company size. It doesn't directly show up in that framework, which is interesting because Fama and French still included in their 5-factor as a pricing model. We'll talk more about this as we go through but small-caps on their own, and actually people who listen to the podcast knows already because we've talked about it a bunch of times. Small-caps on their own have not had a statistically significant premium, which is pretty fascinating and you can even push that further where Banz in 1981 the data set that he used to come up with the original size premium. If you go back and run his analysis again now, because the crisp database that he used is always getting better. They're always improving the data series.

Tyler Shumway, I don't remember what year he wrote the paper, but he did a paper correcting for the delisting bias in the crisp database. I can't remember the exact details but basically it made small-caps look worse. If you re-run the Banz research with the Shumway corrections, which are now included in the crisp data you no longer get a statistically significant premium. It's never really existed on its own but that on its own is important. As we've talked on the podcast a few times, all of the other factors are significantly stronger, both economically and statistically, in small-caps. There's a paper that we talked about in a recent episode by Blitz and Hanauer, I think it was journal of portfolio management the paper was in, and they showed this by running regressions on the academic factor portfolios, which consists of 50% small-caps and 50% large-caps, running regressions on those portfolios with a market capitalization weighted factor portfolio and they found statistically significant alphas, which is another way of showing that the risk factors are much stronger in small-caps.

This is a theme that comes up a few times in these discussions. Fama and French they addressed it in their paper too. They mentioned that even though size doesn't show up in the valuation framework, empirically it's so strong that it has to be included in the model. The other big one is momentum that doesn't show up in the theoretical evaluation framework. Again, we know empirically that stocks that have been doing well tend to continue doing well for a bit. That was from Jagdish and Titman in 1993 and that's continued to be true empirically. But what do you do with that is an important question. If you buy, like in our old model portfolios we had IUSV and IGS, which are just value funds targeting value. What's the problem with targeting value from the perspective of momentum? The problem is when a stock crosses the threshold and goes from being not a value stock to being a value stock, it's going to have negative price momentum. Momentum is a well-documented empirical factor so that means you're betting against momentum by buying stocks when they become value stocks. That's a problem from an expected returns perspective, or maybe not from an expected returns perspective I don't know what you'd call that. It's a problem from an empirical finance perspective. You don't want to be short momentum.

The products that we're talking about in our new model portfolios they actually account for that by delaying trades based on momentum, delaying or accelerating, which I think is important. It's almost a 6-factor model if we're including that in a way. Why do we care about all this in the first place? The premiums have been pretty big, which makes them hard to ignore and there's lots of extra thinking and hassle that goes into this so it's important to know what the premiums are. We're going to go through that. Now we're talking about factor premiums. I think one of the other little known facts that's important to consider is that the academic long short portfolios, and we mentioned this briefly with the Blitz and Hanauer research. The academic factor portfolios are based on 50% weight to small-caps and 50% weight to large-caps. For the value premium as an example is the cheapest small-caps plus the cheapest large-caps minus the most expensive small-caps plus the most expensive large-caps.

That's the best way to construct the factor portfolios for the purpose of asset pricing to use them in a regression model to see what's driving differences in returns in portfolios, that's the best way to construct the factors but from looking at how stuff has done and saying, "Is it worthwhile to pursue this?" That's a little bit less relevant just because you're looking at a portfolio that's significantly overweight small-cap stocks, which not all portfolios are. For example, the model portfolio we're talking about here today has about 30% in small-caps. The academic factor portfolios have about 50% in small-caps and the market has about 10%. You're talking about the value premium. Market-wide value premium that's 10% in small-caps. The academic value premium is 50% in small-caps and we've talked about how factor returns are stronger in small-caps that over-weighting arguably makes the factor premiums look better than they would actually be in a portfolio that is not heavily overweigh small-caps. Sticking with me here?

Cameron Passmore: I'm with you, keep going. I don't want to break up your train of thought.

Ben Felix: Okay. For the U.S. premiums, I document in the paper devaluated and the academic premiums because Ken French has debt on both. Then for international and emerging markets I just did the academic factor returns because I didn't have the capitalization weighted. But if we go through quickly the evaluated portfolio, for size we have 1.58% premium. Small stocks beat large stocks. It's the 30% of the cheapest stocks minus the 30%, or smallest sorry, minus 30% biggest stocks from 1963 to June 2020. The small portfolio beat large portfolio by 1.58% per year on average, which is a lot.

Cameron Passmore: Just to be clear, that's the value weighted portfolio not the evaluated.

Ben Felix: Correct. Value weighted.

Cameron Passmore: Value weighted.

Ben Felix: For value... For the cheap stocks minus the expensive stocks, in the value weighted portfolio, value beat growth by 1.99%. Then for profitability it was at 2.59% premium and for investment 1.92. Now in terms of statistical significance, the small-cap premium was not statistically significant over that time period and interestingly neither was the value premium from a value weighted portfolio perspective. Now, if we go to the academic premiums again for the U.S., small minus big that the value weighted was 1.58% premium. If we go to the academic factor, weighted 50% small 50% large, premium increases to 2.04%. Then for value it went from 1.99 to 2.68. Profitability it increases from 2.59 to 2.8, and for investment the premium increases from 1.92 to 2.93 by switching to the academic factors. Now I mentioned that our model portfolio has about 30% in small-caps. It's not like one of these is right and one of them is wrong. I think in our case with the 30% weight in the small-caps it's going to be somewhere in between like what is the premium you expect?

Then if we look at the developed ex-U.S., I'm going to rip through these a little more quickly. Small stocks beat the big stocks by 0.81%. Value beat growth by 3.01. High profitability beat low profitability by 4.30% and the conservative investment beat aggressive investment by 1.34% and for emerging markets it's similar. I don't think we need to keep going through it. But the main takeaway here, the main point is that in all of these major markets that we're talking about it means basically the whole global market. The factor premiums have been economically large. For the most part they've been statistically significant. Why would you ignore them in portfolio construction? I don't know.

Cameron Passmore: That's exactly what Marlene Lee told us when she was on last January is why would you ignore information that's out there to be had?

Ben Felix: Yeah. I took a couple of examples to... How do we actually apply this to portfolios? It's important to point out too that these regression models and the factor premiums, dimensional for example is not actually using 5-factor regression models to build portfolios. They obviously know about that but it's not one of the things that they're actually using practically to build portfolios and we didn't either. We didn't rely explicitly on that. It was one of the few criteria that we looked at when we were building this model portfolio. It's also worth noting that the intention behind this model portfolio is to look somewhat similar to the portfolios that we use for our clients. We use dimensional mutual funds, which aren't available to the public except for through firms like ours. This model portfolio is designed to look somewhat similar to that in terms of regression coefficients but also in terms of characteristics, which you can argue are more important than regression analysis.

One of the things that I want to look at is... Let's take an index that we know has factor loading. I took the dimensional U.S.Corp equity index and compared its returns to a U.S. total market index. From 1975 to June 2020, the dimensional U.S.Corp index returned 13.52% with a standard deviation of 15.43 whereas the U.S. market delivered 12.12 with a standard deviation of 15.41. We see right there from a standard deviation perspective, we're getting higher returns with any extra risk. Pretty cool. If you're an active manager before multifactor asset pricing models exist, you can show this and say, "Look I'm generating alpha." But as we would expect with 5-factor regression and we can see that all of that excess return is being explained by excess exposure compared to the market to small stocks, value stocks, highly profitable stocks, and to a lesser extent in this case stocks that invest conservatively.

Cameron Passmore: That's the key takeaway. There is always reason.

Ben Felix: Right. We've just taken this hypothetical active manager that could charge 220 or whatever and we've now said actually you can do that systematically with an index, which is the same thing that the AQR paper that showed that Buffett's alpha can be explained by systematic factors. They took a similar approach and said, "Well, here's a set of systematic factors that we know drive stock returns." If you account for those and Warren Buffett's historical success, his alpha actually becomes statistically insignificant. It's the exact same thing, which I think is pretty fascinating. I also looked at dividend growth investing because that's another case where people often believe there's a superiority but the question that we can look at with the asset pricing model are dividends special in the sense that they're generating alpha, or do the dividend stocks just have extra exposure to the common risks that we're talking about? For this when I looked at VIG, the Vanguard dividend appreciation ETF from June 2006 to June 2020, and it had almost the identical return to the market. Actually I looked at this in April, people know I've been working on this paper for a while, and between April and June the VIG went from beating the U.S. market over the full time period to trailing it by three basis points annualized.

Cameron Passmore: No, no way.

Ben Felix: Just because dividend stocks got smoked in the downturn and have recovered more slowly. But even still, VIG had very similar returns off by three basis points with substantially lower standard deviation. Again, we have this question of are dividend stocks superior, or if we plug it into the asset pricing model, the Fama-French 5-factor model, the model explains 94.8% of the monthly variation returns. Basically all of it and the fund does have excess exposure to more profitable companies and companies that invest conservatively. And I mentioned earlier the relationship between the dividend growth, the dividend discount model and the 5-factor model and how they're related theoretically and we see that here. Where if a group of stocks have high dividends that are growing, we would probably expect them to be highly profitable and we would probably expect them to be investing conservatively. This is exactly what we see with this analysis. Another way of saying all of that is that the excess risk adjusted returns of dividend growth stocks over this time period are pretty well fully explained by their 5-factor risk exposures.

Cameron Passmore: Do you think these standard deviations are a lot lower than expected? That'd be the 12.5 versus the 14.7. Is there anything in that information?

Ben Felix: It should all be captured by the-

Cameron Passmore: By the factors?

Ben Felix: By the model here. So I don't think so. The annualized alpha on this fund is statistically insignificant. I guess we can say at zero. Economically it was negative 0.48% annualized but the T starts only 0.55 so it's really statistically at zero. I don't think so. I think it's... Though the model is doing quite a good job of explaining the returns here, which for a dividend investor is problematic. This is a bit of a digression but if you're a dividend investor enjoying the better risk adjusted returns of dividend stocks, if you're really getting a repackaging of systematic risk factors that exist in stocks that don't pay dividends, or that don't grow their dividends, or whatever stocks that don't meet the criteria of a dividend investor. You could be missing out on diversification, which can be detrimental.

But the bigger one I think is if we look at these companies, specifically in VIG, this ETF also has negative exposure to the size premium is larger than the market and has a negative exposure to the value premium. Because of the things targeting dividends, those risk exposures could change over time if the characteristics of dividend stock as a whole change over time. But at a point in time, at this point in time, or at least from 2006 until now you had on average negative exposure to value and to size. If someone looks at that and says, "Well, I like dividend stocks." You can also look at it and say, "Well, you're just getting a suboptimal mix, or potentially suboptimal mix of factors." You could address that by targeting the factors directly instead of-

Cameron Passmore: Bingo!

Ben Felix: Right. All that's fine, we've talked about the full period premiums which is important. I think the persistence is also important. We've looked at these long periods where factors have been positive over the full time period but if you're living through being a factor investor, as we're doing now with the value tilt, you may have periods where the factor under performs. One of the other things I want to look at on the paper is how often does that happen? I'll go through this fairly quickly but over 10 year periods... Some of these debt are fascinating over 10 year periods the U.S. market premium, that's the U.S. market minus one month U.S. treasury bills has been positive 80% of the time.

Cameron Passmore: Since 63.

Ben Felix: Since 1963.

Cameron Passmore: 57 years.

Ben Felix: Then small stocks had beaten big stocks 71.5% of the time.

Cameron Passmore: These are rolling 10 years, right?

Ben Felix: Rolling 10 year period, yes. Overlapping ten-year periods with a one month step. Value stocks beaten growth stocks 86% of the time, which is staggering. That's in more of the 10-year periods, value stocks beat growth stocks than the market beat T-bills.

Cameron Passmore: And everyone... It's a given to invest in the market for most people. You're basically saying if you believe in the market, you should believe in these other factors also.

Ben Felix: From this perspective of rolling 10 year periods they've been more reliable than the market. Statistically it's a bit more debatable. Statistically the market premium has been more reliable than value for example. Actually it's been statistically more reliable than any of the other premiums but over rolling 10-year periods will be different. It's more... If you're thinking about could you stick with a factor, or would you stick with the factor? I think that's where these data are really interesting. For the profitability premium 85.6% of the time and for the investment premium 98% of the time, which is staggering. Over 20 years, all premiums in this data from 1963 to 2020, all premiums except for small-cap were positive 100% of the time over 20 year periods. Non-overlapping periods. I don't know how much insight you can draw from that data but still interesting. For developed international excluding U.S. 10-year premiums... I don't have 20 years because it would only be one non-overlapping sample. 10-year premiums market 87.55% of the time beat T-bills. Small-caps 86% of the time value 90.87. Profitability 100% of the time and conservative minus aggressive investment 92% of the time. I'm going to skip over emerging markets. Sorry, emerging markets you can read it in the paper though. If you're listening, it's similar, it's similar across the board.

The other thing that came out of doing this paper that was really interesting was looking at how the factors behave relative to each other over time. Somebody did a post in the Rational Reminder community about this recently where they posted some really interesting charts that they generated with some code and those were neat to look at. I had a chart, I actually still have a chart in the paper showing the 10-year rolling premiums altogether on a single chart and you can see how they behave differently. But the other thing that Raymond, our director of research, suggested that I add when he was reading the paper was the correlation matrices for all the different factors. And I'm not going to try and explain them because it's just a bunch of numbers but it's in the paper and it's actually unbelievable. It's low for the most part or actually most of the time negative correlation across the factors. If we pick one example, I'll take the market. That's an easy one. The market correlation coefficients for U.S. stocks between 1960 and June 2020, the market has had a correlation of 0.29 with small-caps, with the SMB premium sorry, negative 0.22 with the value premium, negative 0.21 with the profitability premium and negative 0.38 with the investment premium.

Cameron Passmore: In plain English that means?

Ben Felix: It means the premiums perform or show up at different times than market returns as a whole. Again, we're talking about why would you want to do this? Additional expected returns are a major reason but this low to negative correlation of the other alternative, I guess whatever you want to call them, the other risk premiums with the market risk premium from a diversification perspective becomes pretty compelling and hard to ignore. We're going to talk about a couple anecdotal examples of that in a second. I have that data in here for developed and emerging markets as well. It's similar again but you can take a look at that in the paper.

Cameron Passmore: So it's reasonable to expect to smoother the right to higher return.

Ben Felix: Which is exactly what dimensional always says. They say that this increases the reliability of your outcome. And we're going to speak to a couple of examples. We talked about the rolling 10-year periods and we talked about the correlation. We've already covered this conceptually but it's really interesting to show a couple of these extreme examples. There's the last decade in the U.S., which a lot of people know about. For the 10 years ending July 2009, over that full period the U.S. market index, the crisp 110 index lost an annualized 0.19%. Pretty bad. Trailed T-bills which returned 2.95% over the period. Your risk premium was brutal. But over that same time period, U.S. small value stocks measured by the Fama-French U.S. value index return 9.51%, pretty significant difference there, and the Fama-French U.S. value research index returned 3.78%. Not as exciting but still a lot better than the losing 0.19%. Then the U.S. high profitability research index returned 2.09%.

You can see right there that's a period where the market did quite poorly over a 10-year period. Over a fairly long period of time and small-cap being the biggest out performer there delivered a substantial premium. I also looked at this in aggregate and I guess that shows up in the correlation data too but this is another way to look at it. I looked at the 111 10-year periods ending between July 1973 and June 2020, where the market premium was negative. 111 10-year periods where the U.S. market had a negative premium. Over those same periods, small minus big, high minus low value and investment, conservative minus aggressive were all positive.

Cameron Passmore: In all of those?

Ben Felix: All 111 instances of a market premium being negative, SMB, HML and CMA were all positive.

Cameron Passmore: Wow!

Ben Felix: RMW was negative in 53 of the 111 periods. Pretty crazy, right?

Cameron Passmore: Yeah.

Ben Felix: Again, we see that diversification angle for why an investor may want to do this. Here's another one. The single worst time to retire in U.S. stock market history was December 1968. From then until January 1984, the U.S. market actually gained 7.26% per year but trailed T-bills and barely kept pace with inflation. You had about a four basis point annualized return from December 1968 until January 1984 in real terms. When you run like the 4% rule analysis, that's always the time period where it breaks. Now over that same time period, small value will return to 15.8% annualized, and this isn't SMB. That's not small minus big. That is the Fama-French U.S. small value index. The long only index, not small minus big. And the value index returned 13.46%. Again, we see an instance, and I'm totally cherry picking data, but it lines up with the broader sample. I don't feel bad about it. Then the last one that we'll talk about is Japan. July 1990 to December 2019, which is a staggeringly long time for a non-existent risk premium before-

Cameron Passmore: Yeah we've done this one before, it's a great one.

Ben Felix: It's crazy. The Fama-French Japan market index returned 2.36% annualized over that full period, which trailed one month U.S. treasury bills, which returned 2.63%.

Cameron Passmore: For 29 years.

Ben Felix: It's crazy to think about. It really is. Now over that same time period, the Japan high profitability index was actually not a whole lot better, 2.27%. Actually a little worse than the market but the Dimensional Japan small-cap value index and the Fama-French Japan value index delivered annualized returns of 5.43% and 7.96% respectively. Again, we see another instance of these indexes representing exposure to the non-market factors, delivering meaningful out-performance at times when the market didn't do so well.

Cameron Passmore: So the benefit of factor diversification.

Ben Felix: 100%. It just speaks to the reliability piece where yes, you can have long times where stocks don't pay off, or the stock market doesn't pay off but there can still be other risk premiums over those periods of time that do pay off. What do we do with all this information?

Cameron Passmore: I can hear the listeners asking that now.

Ben Felix: We use Dimensional's products as we've mentioned and they're as far as we can tell from the ongoing pretty extensive due diligence that we do, they're doing a great job I guess is the easiest way to say it. The challenge particularly for people who are not our clients is that their products aren't available. You can't just go and buy them. Although that's changing a bit, they've released some ETFs. Not quite a full suite of products like you couldn't recreate our client portfolios with Dimensional EFTs yet. It seems sensible they would get there soon but I have no idea if they will. Last time we made a model portfolio, we used 3-factor model as the basis for the portfolio. We basically just ignored, not completely. We somewhat ignored profitability and investment in the construction. We didn't use products that targeted those factors directly is probably the best way to say it. IJS for example they do have a financial viability screen, which interestingly results in a profitability exposure but we didn't pick products that explicitly targeted all 5-factors because at the time they didn't exist, at least not in a format that we thought was sensible from the perspective of costs and diversification.

But now Avantis... Dimensional launched ETFs, Avantis is another company from some people that actually left Dimensional, which has also launched ETFs and Avantis has a very, very similar approach to Dimensional. You'd be splitting hairs to explain the differences in execution although they may not agree. I'm sure they'd both argue that the differences are larger, maybe they wouldn't, I don't know. But for practical purposes, they are very, very close in terms of the risks that they're delivering in their products. The nice thing about what Avantis since they launched is that they have a couple of small values ETFs. Dimensional has only launched Core ETFs, which are like a lightly tilted total market exposure. With Avantis they have a couple of products that give you pretty targeted, pretty extreme exposure to the smallest and cheapest companies that exist in the market.

That's valuable to us because one of the things... A couple of things that we're worried about in this portfolio construction process are withholding taxes, which can be problematic for Canadian investors when you're using U.S. listed ETFs, currency conversion costs, U.S. estate tax, all these additional complications that start to be introduced when you're using U.S. listed ETFs. Because they have these more extreme small-cap value tilted portfolios, we felt like it gave us an opportunity to use a mix of... Similar to what we did last time, a mix of Canadian listed ETFs, where you don't have to worry about currency conversion or the extra layer of withholding tax but you can still... With a relatively small percentage of the portfolio, you can still target a meaningful amount of factor exposure. Now, given the choice we would prefer, and I want to be clear about this. We would prefer to have a more diversified factor portfolio like Dimensional. With Dimensional's portfolios, you're tilting toward the factors across the full spectrum of market capitalization.

You've got your large value, your mid cap value, small value whereas with this model that we're introducing here, you're getting all of your factor exposure in small-caps. It's an interesting trade-off because we know the premiums are stronger in small-caps, which allows us to do this but at the same time it reduces the diversification that you're using to capture the premiums. But overall, we thought... As people know, we've been working on this for a while, we thought a lot about the different trade offs involved and this is what we decided makes the most sense at least for now. The ETFs that we're using for the factor exposure are the Avantis U.S. small-cap value ETF, and the Avantis internationals small-cap value ETF. We're getting small-cap value exposure in U.S. and international markets. Excluding emerging markets there aren't a lot of great ETFs that really target small value. Avantis does have a price but it's like the Dimensional products. More of a lightly tilted total market ETF. In the paper there's a bunch of analysis of with regression and historical characteristics and stuff like that of return characteristics of those ETFs. I won't dive into it now. You can read the paper if you want to see more about that.

The actual model, it ended up being six ETF for the equity portion. If you want fixed income it'll be another ETF or two depending on how you do that but not too complicated. The same as the previous one, six ETFs. Here it is. Everybody ready? We did a 30% an XIC. You've still got a home bias, which we think makes sense. And we've talked about that pretty extensively on the podcast in the past. 30% VUN. Now this is an important change actually. In our previous model portfolio, we had XUU. One of the things that's come up with PWS research team recently is the way that BlackRock has decided, or I-Shares has decided to execute on XUU is by using three underlying ETFs. It owns three U.S. listed ETFs of small-cap, mid-cap, and large-cap stocks. That sounds fine, right? If you can pull it off efficiently but it's actually had pretty significant tracking error relative to its benchmark. This is something that was flagged this year by our research team and basically it hasn't been super impressive execution by I-Shares, which shows up, like you can go and look at the... I didn't pull it up to talk about right now but you can go to the I-Shares website and the pretty meaningful track meter shows up. I'm going to look it up here.

That's the one year tracking error ending November 2020. XUU returned 14.12%, it's benchmark return 15.42%. It's pretty serious. Anyway, in speaking with our research team, the suggestion was even though VUN has a higher MER, XUU at seven basis points, VUN is at 16. Their suggestion was based on the continued concerns with execution of market exposure based on the way they've decided to construct that fund And what's been happening recently that it makes more sense to go with VUN. So I thought that was pretty sensible, a notable change from last year. 30% of XUU. 30% of VUN. 10% AVUV, that's U.S. small-cap value. 16% XCF, that's international core. 6% AVDV, that's in total 16% of the portfolio is in small-cap value and 8% in XCC for emerging markets.

One of the things I like about this set up is that with VUN you can flip out for VTI and an RSP to eliminate withholding tax if you want to go that route. And likewise for XEF you can switch to IEMG. In taxable accounts with XEF because it's Canadian listed and holds securities directly, you're not going to get hit with the actual layer of the holding tax that you would get from using a U.S. listed ETF and that was one of the reasons we decided to go with this overall structure. Fees end up being very close to a cap weighted benchmark. You get about 14 basis points of total management fees for the factor tilts versus 11.

Then if we look at the historical return characteristics for the tilted model versus the benchmark, which is basically the same portfolio with a small-cap and value stripped out, that's the benchmark but still using ETFs. Over 20 years, the factor tilts delivered us 5.78% and the benchmark delivered 4.96%. It's the difference between turning $10,000 into 30,800 over 20 years versus 26,300 over the same period. Lowest one year return is basically the same. Standard deviation was a little bit higher for the factor tilted model but we got some extra returns in exchange for it. That's basically it. We'll hopefully be able to post the PDF that shows the model portfolios like we've had in the past by the time this episode is released.

Cameron Passmore: The paper as well. When do you think will be posted, or is that going to come out afterwards?

Ben Felix: Probably not by Thursday. It's with our marketing team now to get cleaned up and make all the charts look nice and all that stuff. They did say by the end of this year it'll be realistic to have the full paper available but I think the model portfolio portfolios PDF I need to plug in the updated numbers and the new holdings but we already have that document created.

Cameron Passmore: I can hear the collective cheers of many listeners because this has been a very common request on all the different chat boards.

Ben Felix: Yeah. And I hope it's useful and I hope it doesn't throw... Honestly the biggest draw of the factor exposure is a big change from last time. People that were holding IJs and IUSV. Now we're looking at switching that up to be more concentrated in small-cap value but also giving international small-cap value exposure. That's a pretty meaningful change. The U.S. equity that's a bigger holding, 30% of the portfolio and changing from XUU to VUN, that's another big one that... If there was an embedded tax liability to make that change, would I do it? I don't know. That's a tough call. That's a tough call. And VUN has got a bit of a higher fee than XUU as well. It's not necessarily obvious. Given the choice to pick one now, we don't know if I-Share is going to clean up the implementation of XUU, so VUN may be a better call for now. But overall, like I said before, I think it's a relatively straightforward portfolio to implement. We didn't make it too crazy. It's obviously a lot harder than buying XEQT or something like that but if you're willing to put in the time for the extra, whatever you expect, 50 basis points or something who knows then yeah.

Cameron Passmore: Great, nice work. As you mentioned we're skipping the planning topic this week. We'll have another topic. We're planning the new year. Go quickly to the bad advice of the week. Just a shorter one this week. Who came... Joel passed this article along to us and we sent off a nice zip up Rational Reminder hoodie to Joel. That offer is still out there. If you happen to come across a good idea for bad advice then you just send it through to Ben or I, however you like; Twitter, email whatever works for you. Anyways, this is an article that came from the website Seeking Alpha that was published on November 9th 2020 called the Vanguard total international stock ETF, right idea, wrong implementation. Oops! The article goes on to say investors have multiple reasons to seek exposure outside of U.S. equities and out of control pandemic, high unemployment, high debt, and slow economic growth. So far so good. Meantime, other regions, specifically China and Asia Pacific have superior near term potential in my opinion. Oh-oh, here come the opinions.

It goes on to say, a total international stock ETF sounds like a good idea. The total part of the description means too much diversification that drags down performance. What are you talking about? As a result, the author offers investors two superior alternatives that focus more on specifically on China and the Asia Pacific region. I did not subscribe to Seeking Alpha to get the whole article but the notion of choosing regions, especially after listening to your whole talk about factors, does not make a lot of empirical or theoretical sense to me and certainly not based on near-term potential. It's absurd, right? That's not the point of this ETF. It'd be far better off to understand all the stuff that you just went through for 40 odd minutes or so about what factors to use to get higher expected returns from a theoretical and empirical standpoint, and try to pick what region. Then to go and slam an international stock ETF to me is nuts.

Ben Felix: Yeah. It is nuts. There's a little discussion in the Rational Reminder community recently about this idea of reducing exposure to the U.S., I don't know. It's tricky because... And bond ETFs have the same problem. If you're a market cap weighted equity ETF investor and this is something that I've honestly been a little bit concerned about. We haven't talked about it on the podcast much but if your total market you're getting always exposure to the most expensive stuff in the market and increasingly so as it gets more expensive. Having that constant overweight to small-cap and value it just seems to be a whole lot more sensible to me and rebalancing out of large growth into small-cap value as it increases as opposed to rebalancing into it seems sensible but-

Cameron Passmore: Keep buying cheaper assets.

Ben Felix: There's also international diversification too, right? One of the things that I didn't talk about when I went through those examples of extended periods of market under performance is an internationally diversified investor in the U.S. last decade actually did find. An internationally diversified investor from the Japan 29 year period of under performance likewise, they did fine. International diversification is worth something too and I think that's important for people to take away is that... Well, I guess this article you're talking about was about a total world fund but if you're globally diversified in market beta that's pretty good still as a starting point, we're about improving the expected outcome a bit but I don't want to scare people into thinking everybody needs factors or you're going to suffer terrible stock returns. I don't think that's... I think that's true. I think there's room for improvement, which is what we're trying to do with this model but-

Cameron Passmore: But no need to predict which areas of the world they're going to be near term winners.

Ben Felix: Yeah. To speak to your bad advice article for sure not. We did an episode a while ago on the difference between economic success and stock market success and in that episode, we talked about how economically successful or expected to be successful countries often actually have worse stock returns. It's like value investing. The least GDP growth leads to the highest stock market returns, which is super counter-intuitive but it's an empirical reality.