We kick off today’s episode of the Rational Reminder by discussing when Ben will be publishing his new model portfolios and a quick look at some of our upcoming guests and resources you might want to take a look at. We have been on a roll with our guests lately, and we are certainly not slowing down anytime soon. From there, we look at some of the headlines, such as CDIC developments and the myths around inflation. Next, we move onto to listener rapid-fire questions. Some of the topics include the difference between leveraged ETFs and traditional ones as well as a small-cap investment strategy for an investor with a 30-year plus investment timeline. We then turn our attention to the core topic of the show, dollar-cost averaging versus lump-sum investing. Ben presents an overview of dollar-cost averaging along with some of the perceived benefits. We dive into his analysis of dollar-cost averaging versus lump sum investing in equity portfolios over select 10-year periods across various countries. We discuss the results based on a range of factors and variables. The crux of the argument is that dollar-cost averaging is not as compelling as it’s often sold to be. While there are psychological benefits, the empirical evidence shows that there are not real ones. We wrap the show up with a look at how the pandemic is likely to shape the annuities industry and retirement planning. Tune in today!
Key Points From This Episode:
Find out when the new model portfolios will be up. [0:03:10.0]
Some books to look at ahead of upcoming guests. [0:05:04.0]
Ben and Cameron’s takeaways from Tobi Lutke’s appearance on Invest Like the Best. [0:05:43.0]
Current affairs, including CDIC changes, Michael Kitces recent publication, and inflation. [0:09:07.0]
Rapid fire questions: Leveraged ETFs versus traditional ETFs and size as a risk factor. [0:13:47.0]
How a small cap value investment strategy could work for an investor with a long horizon. [0:23:07.0]
Why Ben and Cameron don’t talk about implementing the profitability factor with a dedicated ETF. [0:25:05.0]
A brief explanation of dollar-cost averaging and the rationale behind it. [0:29:54.0]
Find out more about Ben’s dollar-cost averaging versus lump sum investing analysis. [0:31:49.0]
The results of Ben’s analysis and some key takeaways. [0:36:44.0]
The worst 10% of lump sum outcomes versus dollar-cost averaging – the results. [0:41:26.0]
Two things people look at to try to predict positive outcomes and its influence on lump sum investing.[0:50:36.0]
How high stock prices influence lump sum versus dollar-cost averaging outcomes. [0:53:36.0]
Japan vs the US: How Ben determined if the Japanese market is expensive. [0:56:37.0]
Three key outcomes of the pandemic on retirement planning. [1:00:03.0]
How the annuity industry can encourage its products with decreasing life expectancy. [1:02:05.0]
Bad advice of the week. [1:06:16.0]
Read the Transcript:
Benjamin Felix: This is The Rational Reminder podcast, a weekly reality check on, sensible investing and financial decision making for Canadians. We're hosted by me, Benjamin Felix and Cameron Passmore.
Cameron Passmore: And we're still working from home.
Benjamin Felix: Yeah. You know, I feel funny whenever I say the four Canadians piece, because as much as it is primarily for Canadians and the data of our listenership show that, there's ever increasing contingent of non-Canadian listeners. And I've said this before, but everyone is welcome. It's not just for Canadians.
Cameron Passmore: Yeah, but it's no wonder. I mean, last week, professor Ken French was so good, picked up and retweeted by lots of people all around the world. You can see why it started to get broad appeal.
Benjamin Felix: Yeah. I mean, but the last few months might be tough to match though, between Cliff and Ken, but we'll do it. We've got some great guests coming up too, for sure. So let's talk about the discussion board. See, you've kind of retooled the rash reminder website and the discussion board is in there.
Cameron Passmore: Yeah. So I mean, it's not a discussion board makes it sound way fancier than it is. We installed an app... Installed, so that's an exaggeration too. We started using an app called Commento, which seems great. And it's really designed for commenting on individual posts. And so every podcast episode, now you can go comment as you could before, but it's just, it looks better now and you can sort by uploaded comments or the newest comments, it's all threaded discussion.
So far it has been great. We were able to port all of our old comments from our website over, but the only downside is that they came over without names. So they're all just anonymous comments, but whatever, you can still see the old discussion, some of the old posts had a whole bunch of comments on them.
Benjamin Felix: Well, we'd be able to see, we had one question from a listener asking you if you'll be able to find based on a certain topic.
Cameron Passmore: No. That's what I'm saying that a discussion board is a bit of an exaggeration. So I created a page on our website. That's just called Discussion. And it's basically just one big comment thread. You can post a top level comment, and then people can discuss underneath that. That's the benefit of it being threaded, I guess. But there's no way to sort by thread or anything like that. So over time it'll become, I don't know, a mess isn't the right word, because I think there's a lot of valuable discussion going on already, but it might become hard to organize that. So I don't know, we'll deal with that in the future. For now I think that the platform is working really well, and it's facilitating a lot of really good discussion on the website.
Benjamin Felix: And as far as discussion goes, we're also kicking around the idea of having other advisors on our team, jump in the board to start answering questions so people can get to know them as well. So I think that's a pretty cool idea.
Cameron Passmore: Yeah. So we will do that. The client facing advisors that we have in our team are going to start participating in the discussion. So that should add good elements to the discussion. But something that I've been really impressed with so far, I say this in the little intro blurb on the discussion page to feel free, to answer other people's questions and it's been happening, which is cool. People are posting questions and other people are answering them. And it's awesome. I'm participating when I have time, but I'm not anywhere close to answering every question. And I don't think that I ever will be close to that because there are a lot of questions going up every day.
Benjamin Felix: So two other questions that we had, when will the new model portfolios be up?
Cameron Passmore: Yeah, I think we're getting that question a lot for a long time. So I put quite a bit of time into that dollar cost averaging paper because I wanted to do it well, or that paper we haven't introduced yet. In this episode, we talk about a paper on dollar cost everything that I've just mostly finished probably I think. Maybe a couple more revisions, I don't know. But we talked about that paper.
So I spent a lot of time on that recently, the next one is going to be the new model portfolios, but I'm not just going to make new model portfolios. I've got to do a paper to explain why I'm doing them the way that I am, but they are going to be quite a bit different from the ones we released last time. And the paper will explain why there are different.
Benjamin Felix: Target date, roughly.
Cameron Passmore: Oh, geez. I'm going to say from now, so this is coming on Thursday. So it will be by the next episode of us. So it'll be this episode on Thursday-
Benjamin Felix: So, mid June.
Cameron Passmore: ... and then us again. I think I'll be able to at least talk about the paper, even if it's not quite ready to be published on that next episode. So two weeks.
Benjamin Felix: June 18th to 20th. Okay. And then the other thing we're working on is to get a book list. We've had a few people ask about getting a book list onto site.
Cameron Passmore: We have it. I don't actually know what we're waiting for.
Benjamin Felix: Okay.
Cameron Passmore: We work with a third person on our PWL team named Angelica on a lot of the podcast stuff behind the scenes. So we'll ask her. Yes, we have a book list and it references the episode that the book was mentioned in. And it's a pretty cool list. We have it just needs to get posted.
Benjamin Felix: Cool. So good episode again, a lot on the dollar cost averaging some planning topic, of course, the other regular stuff.
Cameron Passmore: We're not going to review all the items because someone told us not to do that softly.
Benjamin Felix: So with that, we hope you enjoy episode 101.
Cameron Passmore: Welcome to episode 101 of the Rational Reminder podcasts.
Benjamin Felix: So I have a couple of books. I suggest people take a look at ahead of upcoming guests. So one is Brian Portnoy's book, Geometry of Wealth. He'll be on next week with us. Fabulous book. I loved it. And I think we'll get a lot out of the conversation. He does go through the book, but if you have a chance to read it ahead of next week, I think that's a good idea.
Cameron Passmore: He lays out the whole concept of managing money in a way that's different than what I've seen before.
Benjamin Felix: Yeah, I agree. And the second book that I suggest you take a look at is Fred Vettese's book called The Essential Retirement Guide. He'll be on, I believe in a month, probably four, six weeks out. And that's another excellent book, a little more tactical on retirement planning, but a good book. And that was also available on your Kindle if you like. So did you listen to Toby Lucchese interview last week on Invest Like the Best?
Cameron Passmore: Yeah, I did. It's fantastic.
Benjamin Felix: Fantastic interview. And so it's Invest Like the Best with Patrick O'Shaughnessy and I thought it was a phenomenal interview and part that really made high impact on me was where he talked about quality and how growing up in Germany, quality was so high and you get so used to it. It became just part of your life. But then he linked that to the fact that US and China have lower quality standards. They can get products to market faster, which means their economies grew faster. So fascinating.
Cameron Passmore: Yeah. Did you have any more thoughts on that podcast episode?
Benjamin Felix: Oh, I have lots of thoughts on, I thought it was a really interesting episode. How about you?
Cameron Passmore: One of my big takeaways was the quality piece that you just mentioned. You can kind of relate that back to stuff like doing podcasts like this, where it would be very easy to not do it in fear of every episode not being perfect. But the fact that we're okay with not necessarily perfection in every single episode. Like there've been times where we've had to go back and maybe not necessarily correct something factual, but go back and change our minds about something that we talked about in a recent episode, because we've learned new information, same kind of concept.
Benjamin Felix: Another thing that he impacted me with was, because he is a prolific reader and he's talked about that on a number of podcasts. So folks he's recommended have been added to my own reading list. And one of them that Patrick talked about a book that he had read that I'd never heard of before, which is the Zen and the Art of Motorcycle Maintenance. Have you read that book yet?
Cameron Passmore: No.
Benjamin Felix: I never even heard of it before, but I guess it's a classic and I've had a number of friends that I've talked to about it said, "Oh yeah, I read it like 30 years ago." And it's interesting how Toby's point was just read, just read you learn about yourself. You'll make your own framework of thinking out of it doesn't mean any book is necessarily right or wrong, but it will affect how you think about the world.
The main thing is just keep learning, reading, listening to podcasts to get your own belief system. So I just started reading this book and it's such a really interesting book and it actually, it came up in a discussion around quality and that's what this book talks about is, classy motorcycles have great qualities to them, but it's a story of these people that go across the Western USA and how they view the world and quality differently.
So I'm only about a third of the way through, but it's a really interesting read and it's far more complicated. I'm sure that I'm realizing. So I think I'm either going to slow down or maybe go back and restart it because it's a very thoughtful book.
Cameron Passmore: It's one of the things Shane Parrish says in his guide to reading. I think he has a post on pharma street about that. I think he's a course about it too. Well, maybe they canceled the course anyway. One of the things he says to do is to mark places that you want to revisit in the book. And once you've read it, put it away for a bit and then go back and read it again. And it kind of speaks to what you just said. The second time you go and pick it up, you're a different person having had read the book once. And so you may appreciate the things that you highlighted in the first time more.
Benjamin Felix: And I'm actually finding that too about different podcasts. I started to curate a list of my favorite episodes, and I'm taking notes and I'm actually going back and re-listening to different episodes again, because there's such amazing quality in podcasts that are out there now, depending on what stage you're at in your life, your business.
Cameron Passmore: There's tons of content.
Benjamin Felix: Tons of content. Onto current topics.
Cameron Passmore: Yep.
Benjamin Felix: So number one, I just saw this today actually changes to the CDIC rules. The Canada Deposit Insurance Corporation don't know if you'd heard these rules yet or not, but foreign currency deposits are now protected. Same limits up to $100,000 have principal and interest per depositor. And it gets combined with the other Canadian dollar deposits. And the other thing that was added is that term deposits of longer than five years are also now covered.
Cameron Passmore: Oh, well.
Benjamin Felix: Yeah. Even if they were purchased before April 30th. So one of the questions that I saw on a chat group was whether or not the limit would be raised. So some discussion around whether the $100,000 protection limit would be raised because that limit was set 15 years ago. So apparently some committee has been given power to increase that limit. Haven't been able to get the details on that, but for now it still is a $100,000 per entity. That's kind of neat. A couple of big changes.
Cameron Passmore: Yeah, no, the foreign currency deposits that's good news. I mean, that's all good news, but that one strikes me as particularly good.
Benjamin Felix: So the next point I picked up was our friend Michael Kitces, who was a prolific commentator on our industry in the US and hopefully will be a guest on an upcoming episode. He put out a chart showing average daily trade volume by clients at TD Ameritrade. And you talk about a hockey stick chart. This thing is crazy. So you showed the average number of trades up to the date when trading became free, which was late 2019.
So on average, they were doing 800,000 to a million trades per day. Once they went free, trades quickly went up to 3 million trades per day.
Cameron Passmore: Wow.
Benjamin Felix: Isn't that crazy?
Cameron Passmore: Yeah, it is.
Benjamin Felix: Even though it was near free before, it was like a few dollars before, probably five or $10 per trade beforehand. And the last article was the article in the globe about the bank of Canada, says downward pressure on inflation likely, once this shutdown ends. And this is a question I don't know about you, but I'm getting quite a bit where many people are in the camp that all this stimulus has to lead to inflation.
This article suggests that's not necessarily the case, Bank of Canada has a target inflation rate of 2% and they've slashed their key interest rates already three times to 0.25%. And inflation rate actually turned negative in April for the first time in almost a decade. I don't know. Do you get that question as well?
Cameron Passmore: Yeah, this is one of the reasons that I wanted to get, and I still want to get proper monetary economist on the podcast to talk about the monetary and fiscal policy response to the insurance situation, because I am getting that question. On both sides, on the inflation side from the fiscal stimulus more so, and then, well, I don't know if people fully understand the difference in either case. I think there's a lot of confusion about expected inflation as it relates to quantitative easing, and also as it relates to fiscal stimulus.
Both on the inflation side and also on the asset pricing side. The comment that I see almost day online is about how money printing is propping up asset prices. In some way that kind of is, I don't know, that is something that I was doing a bunch of research on it for a while and I kind of burned myself on the topic, but I'll come back to it and we'll talk about it more detail in the podcast. At some point.
Both quantitative easing the Central Bank is trying to put downward pressure on longer-term interest rates, and they kind of do that by printing money. I mean, they do, but it's not in the way that people think. They're not making money out of thin air economists refer to that, I think is helicopter money where they're actually like dropping cash out of the helicopters into the streets.
Benjamin Felix: Mm-hmm (affirmative)
Cameron Passmore: But in reality, the federal reserve is doing it by creating bank reserves and then swapping those bank reserves with fixed income instruments, the financial institutions held. So it's really just affecting the term structure, financial assets in the private sector.
Benjamin Felix: Right.
Cameron Passmore: It's not actually dumping cash into the economy. So the idea that you're going to get inflation from that doesn't make any sense. Inflation comes from lending but lending does not come from low interest rates or cash surpluses, lending comes from credit worthy people who are willing to borrow [crosstalk inaudible 00:12:56].
Benjamin Felix: Uses for the cash.
Cameron Passmore: Yeah. We'll say same concept. Credit worthy people who have a good reason to borrow. That would drive inflation. But without that, and the other piece that a bank don't lend out, can't lend out bank reserves. Like that's not how the banking system works.
Benjamin Felix: So you're working on a zapper for this. Are you working on a YouTube video?
Cameron Passmore: I don't know what it was. I was working on just research on it, maybe for the podcast, maybe for... I think I want to do a video on it eventually, but it's a big topic and Central banks are different around the world too. Like the federal reserve operates a little bit differently from the bank of Canada. So distilling it all down into sort of concise information is not so easy, but we'll bring that back eventually.
Benjamin Felix: All right. So let's go on to rapid fire questions.
Cameron Passmore: Let's do it.
Benjamin Felix: So first question for you, how do leverage ETFs compare to borrowing to invest in traditional ETFs?
Cameron Passmore: We talked about this in a past episode. I can't remember which one, I should have written the number down. I also did a YouTube video where I mentioned this pretty extensively. Biggest difference is the decay that leverage ETFs are going to give you in volt hell markets. And I'll describe a little bit more, but what that is in a second. With leverage ETF over any single trading day, you expect to earn the multiple, whatever it is, 2X, 3X. I think I saw something about 4X, which is pretty crazy of the underlying asset.
So at the index goes up 5% on that day. You should get 10% on that day. Now it's also important to note that these instruments are specifically designed for daily replication. It's plastered all over the marketing material and websites for all of these products. Like these are designed to replicate the daily returns of the index, not necessarily the long-term returns. So that the challenge and where the decay concept comes from is that the fund has to reset its leverage at the end of each trading day.
So if you think about a 2X leverage S&P 500 ETF with a hundred million in assets, it would do like take 80 million of its assets, invest in the S&P 500 and use the remaining 20 million to enter into derivatives futures, maybe swap agreements to get the other 120 million of exposure to the index. So now it's got a hundred million of assets, 200 million net exposure to the index. So that gives you your 2X return, obviously.
So, if we say the S&P 500, got a 5% return, the funds going to gain $10 million. So obviously 10 million on a hundred million, you're getting your 2X return, 10% relative to the 5% index return. Now the problem comes in the next trading day. So the funds now got 110 million in assets, but it's a 2X leverage fund. So if it's got 110 million assets, but only 200 million exposure, or 210, I guess, it's no longer 2X leveraged.
So it has to go re-up on its leverage exposure to get back to 2X. So when you own this fund, when stocks go up, you're being forced effectively to buy more stocks. And when they go down, you're being forced to sell stocks, to keep the product on balance to its target daily exposure to the index. When markets are volatile, this is particularly detrimental. And I mentioned the word, but it's no one as decay.
And I found a paper when I did the YouTube video on this 2010 paper, in the journal of the society for industrial and applied mathematics-
Benjamin Felix: You found this?
Cameron Passmore: What's that?
Ben Felix: You found this?
Benjamin Felix: Well, when you search for... It's Google, come on, but as you may expect, it was a fairly mathematically dense paper anyway. The paper is titled Path Dependence of Leveraged ETF Returns. And they actually come up with a mathematical formula, which they verify empirically to show the theoretical relationship between volatility and decay. So how much you would expect to lose relative to the underlying asset in decay.
Cameron Passmore: So the main takeaway and the short answer to this question is that there's this time decay associated with the realized variants in the returns of the underlying asset. Now it is important to note, and I think Dave Nadig who had in the podcast, I don't know if he said this in the podcast, or it was in an article that I read. But that concept of decay can actually work in your favor too.
Cameron Passmore: Like if the market just keeps going up, it can work out positively. It's really, when you've got volatility that you start to get a real negative impact from the decay.
Benjamin Felix: Yeah.
Cameron Passmore: So I guess an easy way to think about it is that when you have a leveraged ETF, as opposed to, I guess what you'd call homemade leverage, you have leveraged exposure to the underlying asset. You have that daily leverage exposure, but you also have negative exposure to the variance in returns of the underlying asset. So if you end up with a very high variance asset, so like a triple leveraged oil type thing, relative to the underlying you'd expect your returns to be lower.
So anyway, time decay that's the big reason that you might not want to use this things. When I did the research for that video, there was a paper from people at AQR and they talked about products with built in leverage actually have a premium priced into them, which comes out as lower expected returns for the convenience of having built-in leverage. So that was another interesting point that can reduce the expected returns of products with built-in leverage, as opposed to doing leverage yourself. So time decay, potentially lower expected returns because of leverage convenience premium, I guess.
Benjamin Felix: Anything else to add?
Cameron Passmore: Nope. Can I give you the next one here. In episode 64, you mentioned that when you add the value tilt to a small cap index, you revive the factor, but haven't size and value been established as independent risk factors?
Benjamin Felix: Here's an interesting one. So no. Size hasn't really been established as an independent risk factor, which some people may be surprised to hear. There's no real good theoretical basis for a standalone size premium. I'd say it's weak at best. And in terms of statistical reliability, it hasn't been really close. I mean, somewhat close, closer than something random, I guess. But the T stats have not been nearly as high as something like value or the equity risk premium in historical data.
So theoretically, we compare not as strong, not weakness necessarily, although over some time periods, it is, there's a really good paper from AQR that digs into a whole bunch of these facts and fictions it's called Fact Fiction and the Size Effect. So they go into a lot of this. One of the really interesting takeaways from that paper is that Ralph Bonds, his 1981 paper, I can't remember the title, but that was the original plies effect paper.
And that ties into dimensional's story as well as we've probably talked about in past episodes. When you go in and repeat Bonds research, you don't actually get a statistically significant size premium over the time period that he initially looked at, which is fascinating to think about. And the reason this is a suggestion from the AQR paper, the reason they think is that crisp the center for research and securities prices. They're always making their database better.
And there's a guy named Shumway is his last name. I can't remember his first name. He did a paper in 1997 and detailed the process that he went through to correct the delisting bias that had previously existed in the crisp database. So that's 1997 and Bonds did his paper in 1981 that delisting bias, mostly affected small companies. And the deal listings were mostly related to negative events.
So Shumway went and manually dug up like what happened to this company and filled in all of those blanks. And filling in those blanks made small companies look worse on average. So when you correct for that and go back and look at the data, the original size premium paper, wouldn't be like publishable today, which is crazy to think about.
Cameron Passmore: But wouldn't one theoretical basis for small cap premium be cost of capital theoretically should be higher for a smaller company.
Benjamin Felix: Yep. I can see that argument, but I think that the counter-argument is that when you control for other things like relative price, while the other factors that we know about relative price, profitability-
Cameron Passmore: Quality.
Benjamin Felix: .... Sure. Quality when you control for all of those other things, cost of capital should be accounted for.
Cameron Passmore: In the other factors, the other factors have more power than the small-cap factor.
Benjamin Felix: Yeah. The cost of capital argument still makes a lot of sense. And the theoretical models don't include like the theoretical asset pricing models that Fama-French, which I'll mention in a sec, they don't include size. I'll just talk about analysis. So in the Fama-French 2013 paper, a five factor asset pricing model, they explained the theoretical valuation equation that helped them arrive at their five factor asset pricing model, which does not include size like I mentioned, and they address this sort of.
They say, if variables not explicitly linked to this decomposition, this decomposition being the theoretical evaluation equation, such as size and momentum, help forecast returns. And here's the important part, they must do so by implicitly improving forecasts of profitability and investment, or by capturing horizon effects in the term structure of expected returns. So I don't know, maybe that means what you said.
Cameron Passmore: I'm not going to pretend to fully understand it.
Benjamin Felix: I think there's other stuff too, like one of the big ones on the cost of capital piece and trying to remember this accurately from the AQR paper, but it was something along the lines of there's a liquidity premium that you'd expect in small stocks, which could increase their cost of capital. I think it was Cliff who wrote the paper says, and then the co-authors, say that if there's a liquidity premium, that's a liquidity premium, it's not a size premium.
I got my head into some other research recently to just looking at the value premium and its persistence. And in doing that research, there's a whole body of work on the lack of a relationship between a company's actual cost of capital and its expected returns. Fama-French, we've done work on this, or they've mentioned it repeatedly in papers. And some other people have actually done explicit research specifically on that topic.
I was talking to our director of research Raymond, about this. And that was one of the things we were talking about was this relationship between cost of capital and expected returns in the literature. That's actually not a very strong link. So maybe that answers your question, Cameron.
Cameron Passmore: Interesting.
Benjamin Felix: I think that answers that question. You think so?
Cameron Passmore: Whenever very rapid, but we're firing.
Benjamin Felix: All right. Next one. Okay. This one's for you.
Cameron Passmore: We'll see.
Benjamin Felix: Okay. The question is, I'm curious what you think about a 100% small cap value investment strategy for somebody with a long 30 plus year investment horizon. If I truly believe in the risk factors and if I'm an unemotional investor, wouldn't that be the sensible thing to do at least for the US portion?
Cameron Passmore: So this is a question that we had Larry Swedroe speak to clients a number of times, and he always talked about this. He called his barbell portfolio and I see all over red and other places at bogle heads. He calls it the Larry Swedroe portfolio.
Benjamin Felix: That actually showed up. I Googled it, it showed up on website this week called portfolio Einstein. His most recent portfolio. And it's exactly that. And he actually told us, this is what he does. He has a big slug of his portfolio in bonds and large part at that point was in bonds. But then the rest was all at the other end of the portfolio, which was largely us small cap value. So I looked on this site portfolio Einstein and it showed his current portfolio. Well, the current Larry portfolio, not his, but a current Larry Swedroe portfolio is 15% US, large cap value. 15 small cap value, 13 US small cap, 4% emerging markets, 13% international large cap value, and 40% tips.
Cameron Passmore: Given what we just talked about small caps. It's interesting that there's a big allocation to small caps.
Benjamin Felix: Well, exactly. It's somewhat conflicted, but this is something that he did talk about every time we saw him.
Cameron Passmore: Oh, I guess it's overweight for sure.
Benjamin Felix: I was thinking maybe he was bringing up the large cap value exposure just to replace market weight in small, but 13% is over. Well, maybe not actually, that's probably about market cap weight. When you look at all the different, small cap pieces, like a value neutral growth.
Cameron Passmore: But just you up to be willing to accept tracking error if you're going to have a portfolio like that. And also, deep dives in the value, I would think like serious draw downs to be able to handle that. So as long as it is growing, yeah-
Benjamin Felix: Serious products. So last question for you. Is there a specific reason Ben and Cameron are never talking about implementing the profitability factor within a portfolio with a dedicated ETF?
Cameron Passmore: I'm just going to say before I answer the question, it would be nice if we got some non-factor questions. I mean, I guess we chose the ones to answer. Like we mentioned in the introduction, we have the discussion up on the restaurant minder site and it's heavily factor oriented.
Benjamin Felix: These are the easy questions. Anyways. What say you? Feel free to ask some, I don't know, financial planning questions or something next time, how much life insurance do I need or something like that. Why don't we talk about profitability factor with ETFs? Well, if we found, and I think you did find one, Cameron, I didn't do any regression on it or anything, but if we found a profitability loaded ETF and stuck that into our portfolio, that has value in small cap value, we'd end up with something close to the market.
We kind of kill all of our other weights. And the reason is when you're overweight value relative to the market, you're naturally short profitability. Because value stocks tend to be less profitable relative to growth stocks. So by being long value, you're naturally short profitability. So if we go and stick and then on the flip side, actually profitability, your-
Cameron Passmore: Long profitability will be long growth.
Benjamin Felix: ... your long growth, right. So if we take a value ETF on one side and go and mash it together with a profitability half on the other side, you're going to negate your value with the growth and the profitability.
Cameron Passmore: Yep.
Benjamin Felix: And you're going to end up with something close to the market. So that's really the reason, you can't just go and take a profitability to have mashed together. I mean, that paper I did a while ago factor investing with ETFs. One of the things that I looked at, and this is due for a re-visitation, whatever you call it, a new addition, because there are some new ETFs in the market now, but at that time, the multi-factor taps that were supposed to look at these things together. I didn't think that they had enough exposure to the factors relative to their fees to actually add any value.
So I was like, at that point, you're just better off being market cap weighted or using the cheap, deep factor exposure, ETFs that we were able to find for us stocks. Now that doesn't mean profitability is useless. It's just like if you take a profitability, ETF and evaluate ETF and put them together, it's no good. If you take a value ETF and sort it for profitability and overweight, the most profitable value stocks now that's good. That's really good. That's better than just a regular value ETF.
And that's what this whole multi-factor investing concept is about. So there is a new, not super new anymore sort of late last year company to start launching products called Avantas. It's actually interesting because their firm largely consists of people from dimensional, who left and they formed this new firm that in a lot of ways mirrors the way that dimensional is building products.
Anyway. So we need to do more research on that firm. But one of the things that they are doing is what I just described as looking at value and also profitability, but together not separately. Separately, it doesn't work. One of their interesting point on this is IJS, which is in the old, I'm going to call it old even though we haven't released the new ones yet the old rational, minor model portfolios, IGS does have pretty strong and statistically significant exposure to the profitability factor.
And the reason I think is that the way that S&P constructs that index, it's the S&P 600 small cap value index, they have some pretty strict financial liability criteria that goes into their index construction. And so within the universe of small stocks, the ones that make it into that index, I think ended up giving it pretty significant economically significant factor exposure.
Cameron Passmore: Interesting.
Benjamin Felix: Yep.
Cameron Passmore: Not deliberate, but that's what they end up with.
Benjamin Felix: Yeah. And it's actually been, if you look at the rolling regressions over time for IJS, the profitability exposure has been remarkably consistent, considering that they're not directly targeting it.
Cameron Passmore: Fascinating. So under our main topic today.
Benjamin Felix: Yeah. Enough of the rapid fire factor of the rapid fire questions.
Cameron Passmore: Okay. So this next topic, I'm super pumped about this. When you've done a ton of work on this, this is a question that comes up all the time. All the time. If someone has a lump sum for whatever reason to invest. Should I put it in now or should I pulse it in over time? I think the work you've done on this is phenomenal. I look forward to the video on this, that paper's coming out soon, but go ahead. Set it up.
Benjamin Felix: You should used a well, but just the concept of dollar cost averaging. I mean, this is a conversation we have with clients a lot, which is one of the reasons I wanted to write something about it. Dollar cost averaging. I'm assuming people know, but I guess I won't assume because I'll explain it briefly. So you have a big chunk of cash. They have a million dollars of cash that you got from whatever and inheritance or the sale of a private business or something like that. So there's a new cash, new cash that has not been in the stock market previously, you got a million dollars cash.
You can just invest it in the market, or you can, well try and time the market I guess. That's generally not a good idea, but there's kind of a middle ground where instead of just investing the full lump sum, you can do this thing called Dollar Cost Averaging, which is systematically investing over a fixed period of time. So say you take your million dollars and split it up over whatever 10 months, 12 months, whatever it is. And I think the main idea with doing that is that you reduce the risk of investing at the worst possible time.
Cameron Passmore: It's basically regrettable avoidance.
Benjamin Felix: Yeah. Regrettable avoidance is a pretty good way to-
Cameron Passmore: If the market collapses, which is what people are looking for. Right? What if the market collapse I'd be sick over having my million dollars go down to 900,000 and a couple of months.
Benjamin Felix: Or less, it could go down to six or 500,000 like that can happen too. And I think that's the part people are really worried about.
Cameron Passmore: Anyways, so you built the framework to take a look at how rationally things have turned out.
Benjamin Felix: Yeah. I started by building a model. I just wrote some code in visual basic. The allowed me to really quickly compare dollar cost, averaging, to lump sum investing for a whole bunch of markets and a whole bunch of different time periods. I started with that. I wrote the code, and then I was like, "Oh, hey, this is kind of cool. I can compare dollar cost averaging to lesson investing for any market really quickly." So that gave me this big dataset that I could play around with.
And then I started trying to draw some insights out of it. And it all started actually just answering a client question, like someone making this decision. So I started to try and gather the data around it and then started to realize it was actually kind of neat. Now it's paper. So I ended up evaluated dollar cost averaging versus lump sum investing over 10 year periods. Rolling 10 year periods. Rolling means starting with the first month.
So let's say the first month in the dataset is January 1st, 1970, which it was for a lot of the index series that we used. That's the first data point. So we look at the 10 years following that. And then the next rolling period is February, 1970. So you keep rolling forward 10 year investment horizons and moving forward one month in the data with each step. For most of the data series, I had 485 that's from 1970 until April 2020. 485, 10 year periods.
For Canada, I had 652 periods. And for the US I had 1013 periods. I did a 12 month dollar cost averaging implementation. So taking a million dollars, although the dollar amount doesn't matter because I looked at everything in terms of presented returns or whatever. In the model, I had a million dollars and it was comparing investing that boom day one or month one, I guess, versus dollar cost averaging 12 equal monthly investments.
The cash waiting to be invested was sitting in one month, us treasury bills. And then once everything's fully invested, it was invested 100% in stocks, no bonds in the fully implemented portfolio.
Cameron Passmore: So these are all in the different countries indices, there's no balanced portfolio, correct?
Benjamin Felix: Correct. So I did not do any balanced portfolio. It was each country individually that I did look at the equal weighted data for the combination of all the countries.
Cameron Passmore: No factory loading, there's nothing like that. These are all equity portfolios in a number of different countries.
Benjamin Felix: Yeah. So as Australia, Canada, Germany, Japan, United Kingdom, and the United States.
Cameron Passmore: Right.
Benjamin Felix: And I picked those, I don't know, arbitrary is not the right word. I mean, Germany and Japan are economically large. So the UK in terms of GDP and also market cap weight, and then Australia is somewhat similar to Canada in terms of their market structure. So it wasn't a random choice anyway. It's like, what data do I have available, and then which countries do I think are relevant? So the evaluation point, like when we're talking about the relative performance of the two strategies, we're talking about the ending performance, the annualized performance difference after 10 years.
The first thing that I did, and this is how it all started, was just look at the full data series. So for all these different countries, we looked at the whole data series and just look at what are the average outcomes. And then the other thing that I did that I thought was kind of interesting was, isolated the most extreme, bad lump sum outcomes. So for getting how lump sum did relative to dollar cost averaging, just like let's rank the lump sum outcomes by best to worst and take the worst ones, and see if dollar cost averaging helped in those cases-
Cameron Passmore: So the worst cases for the ten-year period for the one year investment period.
Benjamin Felix: Tenure. So the worst ten-year outcome. So we take all the lump sum, all the months that you could have in a lump sum, rank those by best to worst, take the worst ones, and then see if dollar cost average could have helped.
Cameron Passmore: The concern isn't usually, "Am I going to get a bad outcome relative to, I guess it inherently is, am I going to get a bad outcome to dollar cost averaging." People usually just think about a bad outcome from a lump sum perspective.
Benjamin Felix: Exactly.
Cameron Passmore: For sure. So that's why I sorted by lump sum.
Benjamin Felix: And then, knowing that you can't predict when you're going to get a bad lump sum outcome. I perform similar analysis, but during bear markets. So after a 20% or greater drop in the stock market, and then again, similar analysis when stock prices have been high historically. And for that one I just did US data, because that's where I had the Shiller Cape going back to 1870, I think. Okay. So that's the kind of stuff, the analysis.
Before getting into the results, I think it's worth pointing out just the nature of stock returns in general. So the equity risk premium and for this piece, I just looked at US data because there's so much of it. The equity risk premium in the US has been super consistent over the long-term. It's been the arithmetic average of 65 basis points going back to 1926 and it's been positive.
Now there's the equity risk premium that I'm talking about. So that's equities in excess of one month T-bills. It's been positive 60% of the time, in 60% of months, it's been positive. Now there do tend to be these volatility, clustering periods, where volatility will increase a whole bunch for a period of time. And then it settles back down. I made a chart that I'll put in the paper on this. You can see it. Like you can see there's a band of returns that's pretty consistent over time, but then every now and then there's a bunch of a bunch of observations way above and below this sort of normal operating band.
Now, given that though the stock returns are positive in roughly 60%.The equity risk premium is positive, roughly 60% of months, you'd expect an investor randomly choosing a month to get a positive outcome about 60% of the time. Now for the rest of the paper, we're not referring to the equity risk premium. We're just looking at the absolute returns of dollar cost averaging relative to lump sum investing.
And when you look at that data point for US stock returns, stock returns just in absolute terms have been positive 63% of the time in the US data. Okay. So now into the actual analysis.
Cameron Passmore: The teaser's over?
Benjamin Felix: Yeah. Well, I had to set it up. Can't just start talking about the results. But for most markets, most with the exception really being Japan and sort of Australia, but we can kind of generalize and say that for most markets about two thirds of the time, and this is the same as... Like Vanguard did a paper on this a while ago in 2011, I think. And they found similar results for similar countries, although not as many countries as I looked at.
So roughly two thirds of the time lump-sum investing beats dollar cost averaging, over 10 year periods.
Cameron Passmore: Though just in general? In general?
Benjamin Felix: Yeah.
Cameron Passmore: Rolling-
Benjamin Felix: Over the full dataset. Now Japan is the worst of the bunch with 57% lump-sum beating dollar cost averaging. Well, that's one of the reasons I wanted to include Japan because we kind of know what path their stock market's taken. But since the 1990s, Japanese stocks have trailed one month US treasury bills. And keep in mind, that's the asset that we're holding our cash at a lower dollar cost averaging.
Cameron Passmore: Right?
Benjamin Felix: So Japan 57%, but you go through the numbers and I know some people are given feedback. They don't like hearing numbers in the podcast, but sorry, I have to say the numbers. I'll say them slowly. So in Australia it was 61.86% of the time, in Canada 66%, Germany 65%, Japan 57, like I said, United Kingdom, 68% of the time, and the US 70.6% of the time, lumps investing beat dollar cost averaging over 10 year periods. And if we take an equal weighted average of those, that's 65% of the time.
Cameron Passmore: So two thirds of the time you could expect lump sum or not expect, in the past lump sum, beat dollar cost averaging.
Benjamin Felix: Yeah. I mean, I think you can say expect though, because I think that's consistent with expected stock returns. Like the expected risk premium should be fairly consistent over time, especially in a diversified portfolio, globally diversified portfolio. Again, no surprise there. Lump sum investing beats dollar cost averaging. One more point to the full sample before I move on to quantify it.
So not just the percentage of the percent of the time, what does the distribution of outcomes actually look like? The annualized 10 year performance difference on average, this is an equal weighted average across all the markets was 38 basis points. So across all of the samples historically for all the markets that we looked at, you're leaving 38 basis points annualized on the table over 10 years on average. So that's the dollar cost averaging.
Cameron Passmore: It's implicit cost of dollar cost average is 0.38%.
Benjamin Felix: Correct.
Cameron Passmore: Now this also presumes you're going to behave properly and hold on for the 10 years if things, don't happen to have gone the way you had hoped?
Benjamin Felix: Oh yeah. And we're going to come back to that piece later because I think that's a really important part of this whole thing. That implicit cost was really consistent, which was fascinating across all the markets. In Australia it was 32 basis points, 37 in Canada, 38 in Germany, 39 in Japan, 40 in the UK and 41 in the US. So equal weighted average of 38 basis points. But fascinating how close they are across all the different markets.
Okay. So then the next thing that I looked at as the percentiles. So just try and get a feel for the shape of the distribution of outcomes. Now, remember that equal weighted average, that performance difference was 38 basis points. And that's important because that's a mean, and we're going to talk about the median in a second, which is important when thinking about the distribution of outcomes.
So in the 10th percentile, which is the bottom 10% of outcomes, we looked at the 50th percentile, which is a median, like I just said. And then the 90th percentile, which is the top 10%. Now the median in all of the observations is greater than the mean. Which means that this data series has a negative skew, which is characteristic of stock returns in general.
And what a negative skew means for all the people that have to go and look that up, including me, because I always get positive, negative skew confused, even though I've seen the term so many times. Negative skew means that there are more frequent and smaller gains and fewer, but more extreme losses, which is again, that is what distribution of stock returns tends to look like.
So nothing new or exciting there, the dollar cost averaging, lump sum investing outcome, or just matching the distribution of stock returns, which you'd expect. Now it is interesting when you look at the 10th and the 90th percentile, so the worst 10% and the best 10% of outcomes. The best 10% are better relative to the median than the worst 10% are worse.
Now that might sound weird because we just described this as a negatively skewed distribution. When you look further in the tails, like if we go out to the first percentile, the 99th percentile, then you start to see that negative skew show up. But that also speaks to how rare those extreme negative outcomes are. You really have to start digging in the extreme tails.
As I was preparing these notes for this discussion, I was actually thinking about this. It's almost like that cost, that implied cost of 38 basis point, it's almost like it's a really expensive insurance policy against getting that deep left tail outcome. It's a really unlikely outcome, but it can happen. The far left tail outcomes are worse than the far right tail outcomes are good.
But there are more, pretty good outcomes in the right tail, then pretty bad outcomes in the left tail. But then deep, deep, deep left tail, you've got more to lose than you have to gain in the far right tail.
Cameron Passmore: Which you go back to the behavioral part of all of this. We know that people don't like loss as much more than they like gains. So are you implying that that extreme event for a certain personality, it may warrant doing dollar-cost average versus lump sum?
Benjamin Felix: I don't know about that, because that person is going to be risk averse relative to the amount of money that they had. They're not going to be risk averse relative to how they would have done in dollar cost averaging. This is one of the other really interesting observations that I made, is that... So right now we're talking about the worst outcomes in terms of the difference between lump sum investing and dollar cost averaging.
We're not talking about the worst lump sum case, and that's an interesting distinction, because the very worst lump sum outcome relative to dollar cost averaging was not the very worst lump sum outcome.
Cameron Passmore: Fascinating.
Benjamin Felix: Right.
Cameron Passmore: This is differential. This is the differential between the two. So they still could end up positive experience though.
Benjamin Felix: And it was. It was in 1931. It wasn't a great outcome. I think it was a negative in 1931. It must've been. It was a negative annualized return, whether you were lump sum or dollar cost-
Cameron Passmore: Fascinating.
Benjamin Felix: ... average. But in that case, actually, I think you ended up with a really bad outcome from... Hold on. I got to look at it now. I was looking at this before we started talking, but then I put it away. So if we compare the very worst outcome in the whole distribution, now this is for US data only that I'm talking about. The worst difference in outcomes in terms of annualized returns was negative 6.18% in favor of dollar cost averaging. And that was for the 10 year period starting September, 1931.
Cameron Passmore: That's the biggest Delta between the two.
Benjamin Felix: The biggest Delta.
Cameron Passmore: Not worst outcome for dollar cost averaging.
Benjamin Felix: Correct. So your lump sum annualized return over that period was 3.53% annualized. Not terrible considering the time period that we're talking about. But this part's fascinating. So the dollar cost average annualized return 9.72%. So it's not like you were avoiding a horrible left tail outcome in lump sum terms. You were just kind of missing out on what just happened to be a really, really good time to be dollar cost averaging.
Cameron Passmore: Yep.
Benjamin Felix: And then you think about like the path dependence of the outcome starting in September, 1931, you probably just ended up investing at a bunch of bottoms. Like market bottoms before, because there's so much volatility over that time period. And that month just happened to be the month where you ended up with a really good outcome.
Cameron Passmore: So volatility luck at the start of that period has the impact. Fascinating.
Benjamin Felix: Here's another fascinating one. June, 2008.
Cameron Passmore: Okay.
Benjamin Felix: This is the eighth worst outcome in terms of the differential between lump sum and dollar cost averaging. So it was a negative 3.71% annualized-
Cameron Passmore: Starting when in '08?
Benjamin Felix: June.
Cameron Passmore: So just as the crisis was getting going?
Benjamin Felix: So big differential in terms of annualized return, but in lump sum terms, you made 9.47% annualized. If you dollar cost averaged, you made 12.96%.
Cameron Passmore: Wow.
Benjamin Felix: But again, we come back to the idea that you're not avoiding some horrible outcome.
Cameron Passmore: Right.
Benjamin Felix: It's like, "Yeah, dollar cost averaging would have been way better," but to speak to your point about the behavior, it's not like you're losing a bunch of money in lump sum terms. It's just, you didn't get as good of an outcome as you would have. And keep in mind also, these are the deep, deep tails. Most of the outcomes favor lump sum. We have to dig to the worst handful and to find the cases where dollar cost averaging is way better. We went onto a little bit of a digression out there.
Cameron Passmore: So now you're going to push it to the worst lump sum investment period?
Benjamin Felix: Yes. On the far left tail piece. And on the left tail being more extreme than the right tail. So that's that negatively skewed distribution. I think when people think about dollar cost averaging, that's what they're thinking about. People don't want that far left tail outcome, but I think the frequency of that outcome is so low relative to how much more frequent the good outcomes are.
Even if the good outcomes aren't quite as extreme. Hold on, I didn't expect to... I hope people don't mind that we're like talking about the data points because they can't see what we're looking at obviously.
Cameron Passmore: Yeah. I think it'd be cool to look at, it's in those extreme tails. I know I keep coming back to this, but what was going on in the markets and that might've caused someone to bail? Could you think about 2008, June of 2008? You had been able to hold on all the way through late 2008, early 2009 before things turned around in March of 2009.
Benjamin Felix: Okay. Here. So I had a histogram that I made up, I'll put this in the paper that I do to write up this whole experiment too. So the far left tail is a bucket of negative 3.66% to negative 4.08% and this is the Delta. Again, this is the difference between dollar-cost averaging and lump sum. And there are five data points in that bucket. And then you look at the far right tail and it's 3.06 to 3.48%. So again, you see that skew, right?
Like you see those worst, worst outcomes for dollar cost averaging are worse than at the other end of the distribution, then lump sum is better than dollar cost averaging. Like I said, I think what's people are worried about. People are worried about getting that deep left tail, knowing that the far right tail isn't quite as good. Like there's not quite as much to gain if you get the best possible timing by doing a lump sum, as there is to lose, if you get the worst possible outcome.
But I don't think that accounts for the frequency of good outcomes relative to bad outcomes. Like you're much more likely to get the good outcomes, even if the worst outcome is worse. It's so unlikely that you're actually going to get that. So that's why I said earlier that I think this ends up being... It's like insurance. You're probably way overpaying for insurance against that really unlikely bad outcome, if you go to the dollar cost averaging route.
Because so frequently in the historical distribution of outcomes, you've been worse off by dollar cost averaging. Okay. Anyway, I did think that that tail, that negative skew piece did give some validation to the whole concept of dollar cost averaging to avoid the worst outcome. Like sure. It can do that, but you're just way more likely to actually make yourself worse off by trying to avoid that extreme bad outcome.
Okay. So the next thing I looked at was forget about the difference in outcomes for a second. Like just now we were talking about the shape of the distribution in terms of the difference in outcomes between dollar cost averaging and lump sum investing. Let's just look at lump sum investing, sort by outcome, take the worst 10% of lump sum outcomes and see how dollar cost averaging did in those cases.
So if people thought, "Okay, I'm about to invest this lump sum of new money, and I'm really worried about getting a bad outcome." Okay, what dollar cost averaging make you better off? If it does end up being a bad time to do a lump sum.
Cameron Passmore: Wouldn't have saved even the worst periods for lump sum investing.
Benjamin Felix: Not as it randomly going to be the best time ever to do at dollar cost averaging, which is kind of what we were talking about a second ago. And I do think that's just random outcomes. Like if you happen to invest this, you end up with like naive, perfect market timing in some historical cases. But if I forget about that and just say, "Okay, pretty sure we're going to have a bad lump sum outcome, so let's dollar cost average," would that actually have helped?
You might have found that in 51.41% of historical periods, lump sum investing would have made you better off that's just in terms of number of outcomes. If we take into account the magnitude, it's kind of like the shape of the distribution that we were just talking about, how good were the good outcomes relative to how bad were the bad outcomes the annualized return difference was negative 29 basis points. So dollar cost averaging.
Cameron Passmore: These are the worst-
Benjamin Felix: Correct.
Cameron Passmore (00:49:15):
...10% of lump sum investing in your data set.
Benjamin Felix: So if we think about forward-looking, this is if we could predict the worst outcomes, would dollar cost averaging have made you better off? So in 51-
Cameron Passmore: So half the time it would have?
Benjamin Felix: Yeah. Roughly half the time it would have. And on average it was a 29 basis point advantage for dollar cost averaging. So I thought that was pretty interesting too. Now, obviously we cannot predict the thing I just said. We cannot predict what are going to be the worst future lump sum outcomes. So instead of just, "It's not compelling, oh, because we can't actually predict that." Right?
Cameron Passmore: Okay. Because that's not screamingly compelling?
Benjamin Felix: Yeah. It's a bit of a tossup.
Cameron Passmore: Well, no, it's not compelling that dollar cost average is going to save you.
Benjamin Felix: This is not a good reflection of reality because we've intentionally picked the worst lump sum outcomes. And even then, it was a tossup.
Cameron Passmore: Right.
Benjamin Felix: Or roughly a tossup. The distribution of outcomes though, and you kind of see this from that negative mean return difference, the negative 29 basis points. The distribution was much more negatively skewed this time. So the worst outcomes were worse than the best outcomes were better.
Cameron Passmore: Right.
Benjamin Felix: But again, we looked back in time and picked the worst possible lump sum outcomes. Okay. So then I thought, well obviously like what you just said Cameron this is not reality because we don't know if we're going to get the worst lump sum outcome or not. What do people think we can use to predict bad investment outcomes? So I figured there are two things. There were two things during which periods, people worry about this the most, which is bear markets. Like we just saw and-
Cameron Passmore: Bull markets.
Benjamin Felix: Yeah. And then bull markets.
Cameron Passmore: Bull markets.
Benjamin Felix: You're right. And when stock prices are high. So for bear markets, I just said, we're only going to look at starting months where the market has dropped by 20% or more in the previous month.
Cameron Passmore: So serious drop. Okay.
Benjamin Felix: 20% or more. Yeah. So it's like, "Okay, we're deciding to invest this million dollars of cash. The market just dropped 20%. Does it make sense to dollar cost average now because we're in these uncertain times, as they're often called?"
Cameron Passmore: So does dollar cost averaging protect you?
Benjamin Felix: On average, it did not protect you, which was interesting. So same thing, 12-month deployment of the cash after a 20% drop has already happened, equal weighted average across all the markets. Yeah. 53.66% of the time lump sum investing made you better off over the full period. Across all the markets it was 50% or greater except for the UK, and I had to like triple check the date on this one because it didn't make sense that it was so different, but 33% of the time only lump sum investing made you better off in that market. But the rest of the markets, Canada like-
Cameron Passmore: Was another one.
Benjamin Felix: Yeah, I double checked that data too, but Canada was 78.57% of the time lump sum made you better off. But you kind of think about practically, like what happens when the market drops 20%? Well, sometimes kind of rarely it drops more, like you get a 20% drop and then another 20% drop. More often historically anyway, you get a 20% drop and then you get a recovery. And that shows up in this data, more often you're ending up with a negative 20% and ending up being actually not the worst time to invest.
So the average annualized return difference in this case was 25 basis points in favor of lump sum investing. And it's also worth mentioning, I think that the US outcome, which was 50% of the time lump sum was better. Not that we should eliminate this part of the data, but that's heavily skewed by the 1930s where it was like every 20% drop was followed by another 20% drop for that whole time period.
So in bear markets, it's a little bit different. It's a little bit more of a tossup relative to the full dataset. And that's consistent with something that Gerard O'Reilly, Dimensional CEO said in a webcast I listened to recently. Well, it's kind of like what I mentioned earlier too, with the distribution of stock returns. The equity premium is remarkably consistent over time, except for sometimes.
And I think Gerard's data point was that I don't remember the exact numbers, but around market declines, the premium tends to be significantly negative. And then it tends to be significantly positive, like more so than usual. And then it just kind of goes back to normal. So I think this to an extent shows up in these data too.
Okay. So bear markets, still no clear advantage for dollar cost averaging. And then the last one I looked at was when stock prices are high. This is probably my favorite one to look at I think. Just in terms of going through the data. Yeah. So I mean basically when stock prices fall, everyone's panicking about investing, but when stock prices... When they rise, everyone's also panicking about investing new money, because it always is worrying about investing at a peak.
So I took Shiller CAPE Ratio data and use that to measure at a point in time, how expensive or cheap the US market was relative to history. And I only looked at observations where the Shiller CAPE was in the 95th percentile of all historical monthly observations. So I took all of the Shiller CAPE data from February, 1872 to May, 2020. And if a given month fell within the 95th percentile of historical Shiller CAPE over that full-time period, then we looked at that month and we went through the whole data set and did that.
So with that sort of expensiveness, lump sum investing beats dollar cost averaging 54.24% of the time with an average annualized return difference of only three basis points. So that makes lump sum investment look not that great, especially considering those deep left tail outcomes that we're talking about. So people might be thinking, "Well, maybe not such a good idea, maybe not worth it."
Three basis points, annualized return difference in the US data series anyway. When stock markets are expensive or when the stock market is expensive, that observation. And I took this concept from an old AQR paper, but that observation does not reflect reality because I used all of the Shiller CAPE data for the history of the US stock market to determine whether any given data point was expensive or cheap.
But if we're in 1960, we didn't know what that Shiller CAPE was going to be from 1960 through 2020. So I went back and did the correction that AQR did and the paper that made me think of this and redid the exact same analysis, but to determine whether the market was cheaper or expensive, I only looked at backward-looking Shiller CAPE data at each month.
So like saying, January 26 where that data series starts, we're comparing the Shiller CAPE on that date to the set of Shiller CAPE data from 1872 to January, 1926, not from February, 1872 to May, 2020. So it eliminates whatever you'd call that backward, looking bias. With that correction, lump sum investing beats dollar cost averaging 64% of the time.
Cameron Passmore: Wow.
Benjamin Felix: Yeah. Big deal.
Cameron Passmore: That's a little bit more compelling.
Benjamin Felix: Yeah. And this is still the 95th percentile. So you, as the investor sitting there making the decision, looking at the historical data and the market's at its 95th percentile of historical or expensiveness is still a pretty good shot of outperforming with the lump sum.
Cameron Passmore: Amazing.
Benjamin Felix: And the average annualized return difference in that case 18 basis points.
Cameron Passmore: It's crazy, eh? Even at the 95th percentile, the Shiller CAPE, is still not compelling to dollar cost average.
Benjamin Felix: Well, where till you hear this one. The next one doesn't have the same rigor in the analysis because I didn't have the data, but I looked at Japan just out of interest because I was doing the US analysis, and I'm thinking like, "Man, I wonder what it looks like in Japan. I bet it's fascinating". And it was. So, like I mentioned, I don't have Shiller CAPE data for Japan going back to, I don't know, like 1940 or something to do a similar analysis. So instead I just used the most expensive that the US market's ever been as the benchmark to determine whether or not the Japanese market was expensive.
The highest at the monthly Shiller CAPE has ever been in the US data is 44.19, which happened in December, 1999. Japan, as some people may know, got really expensive in the '80s up until 1990 in May, 1986, Japan had a Shiller CAPE of 44.31. So it exceeded the sort of tech bubble levels of the US and then it kept going up. Now the Japanese market did decline significantly. And like we mentioned at the beginning of this discussion, it hasn't really recovered even now.
So following May, 1986 with a Shiller CAPE of 44.31, there were 29 observations after that where a lump sum investing beat dollar cost averaging by a pretty significant margin too. I thought this one was fascinating. So in November, 1988, when the Shiller CAPE was 72.07, lump sum still beat dollar cost averaging in that month, in terms of 10 year annualized returns.
Cameron Passmore: Yeah I mean, the returns may not have been great at the end of it, but still lump sum beat dollar cost averaging. Because you're looking at relative, you're not looking at the returns, you're looking at relative.
Benjamin Felix: Correct. But the front end of that... I mean, I've looked at Japan on this in the past too, because people say, "Well, if you invested in Japan in 1990, you lost a ton of money and never made it back," which is true. But even if you invested in Japan and I haven't looked at the data in a while, but say you started investing in Japan in 1980 and just held until now, your annualized returns are around 6%.
And that's because the returns from 1980 to 1990 were so high that even with 30 years of flat returns afterwards, you actually still got a pretty good outcome if you just held on to Japanese stocks. That kind of speaks to this too, where prices can keep going up even if you don't expect them to. It is knowing that when prices are high, future expected returns are lower.
But based on this, you can't use that to time the dollar cost averaging versus lump sum investing decision or I mean the market timing decision in general.
Cameron Passmore: Okay. So bottom line, it's not compelling to dollar cost average? That's the punch line.
Benjamin Felix: Yeah. I mean the psychological risk piece is big. Like you mentioned that Cameron, if you're going to bail on the lump sum strategy, then that's no good. But I was thinking about that as I was writing this. And it's like, if you're investing in a portfolio that is so scary that you feel the need to dollar cost average, maybe that's not an appropriate portfolio.
I was thinking a lot about this. Like, the data are so obviously in favor of lump sum investing, no matter how you slice it up. So it's like, if you're not comfortable doing a lump sum and to make a portfolio palatable, if you have to dollar cost average with new money... I don't know. I'm not saying people should never dollar cost average. And this is like I said before, a conversation we have with clients all the time.
And we didn't have the data to this extent, but we've always had the data that lump sum tends to be better. And clients still do dollar cost average in. But I don't know like this, I wouldn't say it changed my mind, but it gave me more appreciation for how unlikely it is that dollar cost averaging is going to do anything special for you.
Cameron Passmore: Exactly. So onto the planning topic.
Benjamin Felix: Yeah.
Cameron Passmore: We'll do this quickly because I know we're running a little bit longer, but it's worth it. So I thought Alexandra Macqueen, our good friend Alexander Macqueen, she's @MoneyGal on Twitter. She wrote a great article in the April 28th episode or publication of MoneySense. Just wondering what the impact of the pandemic might be on retirement planning. And she highlighted what she thought were three potential fallouts from this pandemic, and I think it makes pretty good sense to me.
Number one, the movement towards early retirement she believes will dwindle as employment security drops. And to me, that makes sense. And you can see a scenario where housing prices fall, markets fall, potentially lower returns, less money, and fewer jobs to fall back on is her main point. I guess there's a lot of people that do retire knowing that well, if things don't go as they had hoped, they can always go back to work.
Well, her argument is more people may hang onto their jobs longer, build in a bigger cushion to make sure they're really certain that they'll be okay in retirement. So that seems sensible to me.
Benjamin Felix: Yeah.
Cameron Passmore: Another stat and this I thought was interesting. The amount of debt carried by retirees will reduce, even though she says senior debt load has increased. And can you believe this, mortgage debt for seniors has doubled since 1999 and consumer debt for seniors is up 50% since 2016? I wasn't aware of that, but the belief that weak markets and general worry about things will cause people to reduce debt, and that we're hearing from lots of people now will cause them to use current cash to pay off current debt and then also take on less debt.
So she thinks people will become more conservative generally in their spending and put more of their disposable cashflow towards debt reduction. And her last proposal, I guess you could call it is the appeal of guaranteed income will rise. And this is what we talked about here when she was on with us last summer about annuity.
So she thinks there'll be much more interested in annuities, be a lot more respect for all the security Canada pension plan. And it's also interesting she noted this said "How the Spanish flu of 1918 actually helped shape the life insurance business in North America.: So she say, well, maybe COVID-19 will actually increase the interest in the annuity business. It seems a lot pretty sensible to me.
Benjamin Felix: Yeah.
Cameron Passmore: I mean, we had chatted earlier this week about some articles that Moshe Milevsky had put onto Twitter. So he took a look at an article called COVID-19 Longevity Risk and the Economics of Annuitization. So this year during the pandemic, there has been a spike in life insurance applications, which is exactly what happened after the Spanish flu.
So the question Prof. Milevsky asked in this article was, "How should the annuity industry continue to justify and encourage longevity insurance?" When you think about it, we talked about life expectancy a few episodes ago. The result of a pandemic is that life expectancy can drop. So how does the industry increase the appeal going back to Alexandra's point of guaranteed income and protecting yourself with income, if you happen to live a long time in a period where life expectancy may be falling?
That's a pretty good question. So here I'm going to quote him. He says, "Here's my main idea. Counter-intuitively the magnitude of longevity risk which technically is a coefficient of variation of your remaining life actually increases as mortality rates spike. How's that for counter-intuitive? "Moreover, if episodes of COVID-19 are now part of the new non-normal then over the foreseeable future longevity risk will be higher rather than lower.
Ergo annuities will become even more important going forward. Longevity risk is not that you might live a long time rather, it's about the uncertainty around that expectation.
Benjamin Felix: Yeah. He called the second moment.
Cameron Passmore: How's that for a head spinner?
Benjamin Felix: Yeah. I read through his post and tried to figure out what he meant by that piece. And they actually laid it out pretty clearly, as I guess you'd expect from him. "The coefficient of variation of remaining life is the standard deviation of life expectancy divided by the life expectancy." So that number, the coefficient of variation on remaining life is the measure for longevity risk. This is what pension plan sponsors and I'm assuming annuity providers are looking at.
And then he explains it that a mortality shock reduces life expectancy and the standard deviation of life expectancy. But the standard deviation of life expectancy falls less. The ratio actually goes up. So I have the numbers that he had in his post, I think were, if I still have it open. Yeah. So in normal times, the coefficient of variation for a 65 year old is 45.3% with a 60% mortality shock.
So mortality, 60% above normal, the coefficient of variation increases to 51.1%. And that's because life expectancy has decreased from, in his example of 20.11 years in normal times to 15.56 years in the 60% mortality shock time. And the standard deviation has gone from 9.11 to 7.95. So because of that, the coefficient of variation goes up. So that of mortality or longevity of risk is actually increasing.
Cameron Passmore: So yeah, the odds of living longer have declined, but life is riskier. So that makes annuities more valuable from an economic standpoint.
Benjamin Felix: Yeah, definitely counter-intuitive but also fascinating.
Cameron Passmore: So his point is, what can the industry do to make that argument more compelling? I mean, Alexander's argument is pretty straightforward, right? More certainty, but of life expectancy is less anyways, there's the answer.
Benjamin Felix: But I guess you'd expect it to be reflected in... And he mentions pricing in the article, but you'd expect that to be reflected in annuity pricing. If life expectancy is shorter, you'd expect ... Jeez, I guess you'd expect the price to come down to reflect lower mortality. But you'd also expect it to go up a little bit because longevity risk is higher. I guess.
But net would you expect a reduction in cost? I think on net you'd expect a reduction in cost. But it also shows that I don't fully understand all the economics behind this. I think he's teaching like PhD economic, PhD level economics, which I have never taken PhD level anything. But there's one measure he shows, the annuity equivalent wealth. That number is going up.
I don't know how this affects annuity pricing, but just from an economic perspective, the value of the annuity is increasing to the annuity holder.
Cameron Passmore: Exactly.
Benjamin Felix: Which is fascinating and super counter-intuitive.
Cameron Passmore: And hopefully we can have him on to explain this to everyone one day.
Benjamin Felix: Yeah. Yeah. He explained it better than we can.
Cameron Passmore: That's a given. Anyways bad advice of the week, quickly to round this out. Article from Barron's called, It's a Weird Market. Time To Go Active. So you knew as soon as I saw that title I was going to be making to the bad advice of the week. Anyway, it was an interesting article. They interviewed a number of index advisors who are now adding active managers.
Benjamin Felix: Did you pull this from Twitter?
Cameron Passmore: Oh yeah.
Benjamin Felix: I saw the tweet. The tweet said something along the lines of like, "Can you imagine working with someone who's giving you financial advice based on evidence and data, and then all of a sudden they changed their tune and decide to start adding in active funds?"
Cameron Passmore: Some of the quotes were unbelievable. Anyway, the quote from the article says "Some financial advisors, including longtime skeptics of active management, are moving money into these strategies." Quote. They believe that the market is acting irrationally enough to give good managers the chance of beating the benchmarks.
"It makes sense to be surgical now more than ever," said one CEO of a financial investment group. "You need to find companies gaining market share, even if their prices are going down". Another advisor says, "The markets were overbought." And there is, get this Ben, "There is no cerebral cortex to a passive approach".
Benjamin Felix: Wow.
Cameron Passmore: And the article mentioned 2000 bubble as a time where 70% of active managers beat passive counterparts. "It was a Turkey shoot," says the global director of research at Morningstar.
Benjamin Felix: Wait, is that true? No, it's not.
Cameron Passmore: I'm just reporting the article, I'm not going to refute it right now, but I'm not sure if that's true.
Benjamin Felix: Maybe for a month or something when the... Because they were holding cash or something? I don't know.
Cameron Passmore: Anyways, it hasn't been the experience so far this year, 48% beating their benchmark through the recent V-shape market. Anyways get this, summer nibbling. They're choosing constitute a portfolio. So the manager has the agility and flexibility to dodge risks and snap up opportunities. Can you imagine? People moving to an active portfolio for this reason based on feeling. Like you feel they're going to dodge?
Benjamin Felix: I mean there's a lot of uncertainty and fear. It's not necessarily easy to have an appreciation for the data on active management or the theory behind why it shouldn't make a whole lot of sense. I get it. Uncertainty goes up, people start doing weird stuff. I wouldn't call it surprising.
Cameron Passmore: Anyways, the article did mention at the end that the evidence suggests that it is tough to predict which manager will perform, but that fees do matter. So it did finish on balance.
Benjamin Felix: Ken French said it perfectly in the last episode where we had him on as a guest and he talked about something along the lines of, he doesn't doubt that there are good active managers out there that can produce pretty good results before fees, but who's that benefit accruing to? Them or the clients? It's going to accrue to the manager.
Cameron Passmore: That's right.
Benjamin Felix: And that shows up in the data. And you can't really argue with the data.
Cameron Passmore: Anything else to add today?
Benjamin Felix: Nope.
Cameron Passmore: All right. Thanks for listening.
Books From Today’s Episode:
Zen and the Art of Motorcycle Maintenance: An Inquiry Into Values — https://amzn.to/3DUFB7o
The Geometry of Wealth — https://amzn.to/2AZpIkg
Links From Today’s Episode:
Rational Reminder on iTunes — https://itunes.apple.com/ca/podcast/the-rational-reminder-podcast/id1426530582.
Rational Reminder Website — https://rationalreminder.ca/
Shop Merch — https://shop.rationalreminder.ca/
Join the Community — https://community.rationalreminder.ca/
Follow us on Twitter — https://twitter.com/RationalRemind
Follow us on Instagram — @rationalreminder
Benjamin on Twitter — https://twitter.com/benjaminwfelix
Cameron on Twitter — https://twitter.com/CameronPassmore
'Path-Dependence of Leveraged ETF Returns' —https://www.math.nyu.edu/faculty/avellane/SIAMLETFS.pdf.pdf
'The Delisting Bias in CRSP Data' — https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1540-6261.1997.tb03818.x
'How the coronavirus pandemic could change the way we think about retirement in Canada' — https://www.moneysense.ca/save/retirement/planning-for-retirement-after-covid-19/
'Covid-19, Longevity Risk & the Economics of Annuitization' — https://moshemilevsky.com/covid-19-and-longevity-risk/