← Enterprising Investor
Enterprising Investor
Rob Arnott: Rethinking Risk, Fear, and the Future of Asset Pricing
00:32:02
Rob Arnott, founder and chairman of Research Affiliates, joins The Enterprising Investor to challenge long-held assumptions about the equity risk premium and asset allocation. Drawing on Fear, Not Risk, Explains Asset Pricing, the provocative paper he...
Loading summary
Transcript30 lines
- [00:03]
A
If you struggle to ramp up a refresh quickly on a long list of public companies. Listen to this Most research platforms make you dig through endless dashboards and raw data and charge a premium for it. Tenzing Memo is an AI powered market intelligence platform built for professional investors offering real time company research and insights without filters and dashboards. You just open it and get what you need in a tenth of the time at at a tenth of the cost, as one CIO put it. Tenzing Memo takes us from zero to one in understanding a candidate not in days, but in minutes. Get up to speed faster in nearly 6,000 public companies. Try memo free at Tenzing memo.com register with Code Enterprise to unlock an extended free trial.
- [00:55]
B
Hello and welcome to the Enterprising Investor, the flagship investment podcast for CFA Institute. I'm Mike Wahlberg and today's guest is Rob Arnott, Founder and Chairman of Research Affiliates and one of the most influential voices in quantitative investing. With decades of experience challenging conventional wisdom in asset pricing and portfolio construction, Rob has authored well over 100 academic papers, many of them award winning and pioneered strategies that blend theory with real world application, including the practice of tactical asset allocation. Listeners may recall we had Ed McQuarrie on the show last year to talk about his new paper that challenged the long held assumption that equities have always reliably outperformed bonds. I remember that conversation well because he told me at the time and I forget if this was offline or in the episode itself that he was headed then in January so January 25th to a CFA Institute webinar to discuss or probably defend the findings with Jeremy Siegel, Rob here, Roger Ibbotson, Elroy Dimson and Lawrence Siegel. And I admit that hearing that list of panelists made me a little afraid for Ed at the time. The good news is that today we are talking about a furtherance of that research research that Rob conducted. Together with Ed. They recently co authored a paper this past spring titled Fear Not Risk Explains Asset Pricing. This article offers a provocative rethinking of how markets actually reward investors and what we've been getting wrong for decades. I look forward to hearing more. So welcome to the show, Rob.
- [02:25]
C
Thank you very much. It's a privilege being here.
- [02:28]
B
Before we get going here, tell me a little bit about that webinar. How did that go and how did that conversation go and how did that sort of inspire you to work more with Ed on this topic?
- [02:37]
C
Well, firstly, these conversations are wonderful fun because we all have a mutual admiration society and deep respect for one another. We just have different opinions. And Jeremy Siegel's very important contribution to the world of finance has been his work on the equity risk premium and his demonstration that over long periods of time stocks have reliably beat bonds. For the investor who's patient enough, and that's the key issue. A lot of people read his writings as suggesting that an investor with a five year horizon, for instance, hardly anyone has a horizon much longer than that, even though they should. The odds of success for stocks on a rolling five year basis is actually not brilliant. And Jeremy would be the first to readily acknowledge that. Ed McQuarrie's contribution was to challenge stocks for the long run by looking at the long term history. He's an empiricist and he's a superb empiricist. He took a deep dive into the sources of long term returns going back well before the CRSP Data starts in 1926. We had the Cowles Commission taking dotted back to the 1870s. We had Schwerck taking data back to 1802 and these were based on studies that turned out to be more superficial than people realized. And Ed took the data back to 1802 and found that stocks more broadly defined than short did, more broadly defined than Cowells did, had performed a little worse than people realize and that bonds had performed a little better, which means that the risk premium had been smaller. Now the other thing that Ed and I both noticed is that the data covered by the CRSP data going back to 1926 and by the Ibbotson Sinque field data going back to 1926 is heavily dominated by a half century from 1950 to 2000. Why is that important? The stock market went from an 8% dividend yield to a 1% dividend yield in that 50 year span. That means, and we have a working paper called Revaluation Alpha which points out that if historical return includes a revaluation, and you assume that historical return is predictive of the future, tacitly you're predicting that the revaluation will continue Eightfold. Revaluation. So does that mean that we should expect 50 years hence that the dividend yield on US stocks will be 15 basis points for a dividend yield and the CAPE ratio would have to rise to about 300 to match that. Okay, do you want to assume mean reversion, that is to say when revaluation happens it reverses course? That's one possibility. That's, that's what Jeremy Grantham assumes in his seven year real return forecasts. It's a little dangerous because you can enter a new world where there's new risk tolerances. People are living longer, they don't have. If, if your life expectancy is 50, 50 chance you're going to be alive in 10 years, chances are that you're going to demand a pretty high risk premium to not spend today. And if your life expectancy is 40 years or 60 years, you're going to demand a smaller risk premium. You're going to be fine with a lower risk premium. And so if the normal risk premium changes, then revaluation can be sustained. But you don't dare assume that it's going to persist. Now back to the second half of the 20th century. That 50 year span saw an Eightfold rise in valuation of stocks measured against dividends six fold rise on a Cape ratio basis. That is huge. The revaluation alone accounts for 400 basis points of the stock market's return per annum for 50 years. 400 basis points a year for 50 years is attributable to revaluation. A, you need to take that out. So if the excess return for stocks was 8% over that span, absent revaluation, it was 4. And if you take the revaluation out of the entire century long span covered by the Ibbotson data, the Ibbotson data will have a century as of the end of this year. And that century long span would have been influenced to the tune of about 200 basis points a year. So when people look at the past and say stocks have beat bonds over the last century by four and a half percent per annum compounded, fantastic. Except two of that was revaluation. Take that out and you're looking at 2.5%. So these are important observations because the risk premium is smaller than people think. If the risk premium is smaller than people think and the variability of returns is exactly what people think it is, then that means that the amount of time you have to wait to have confidence that you're going to win with stocks goes up with the square of the decline in risk premium. So if people are assuming five and it's two and a half, then that means you have to wait four times as long to have confidence that stocks will be bonds. And sure enough, when Macquarie redid the analysis with painstakingly reconstructed stock and bond returns over the 125 year span ended in 1926, painstakingly reconstructing those returns to be more accurate and more complete picture. What you find is that there was the entire 19th century, Bonds beat stocks the entire century. Most of us don't have a century to wait for the risk premium to kick in and reward us for our choices. Which tells us that the stock versus bond decision is lightly based on a presumptive risk premium and heavily based on a tactical evaluation of which is priced to produce a higher return. Back in 2000, I wrote a paper, A Death of the Risk Premium. It got published in the Journal of portfolio management in 2001 and basically it said, look, stocks have a yield of 1. TIPS have a yield of 4. Stocks have inflation participation in the dividends. TIPS have inflation participation in the coupon. And so defining the risk premium using tips, not corporate bonds is actually a more sensible comparison. Very few people to this day, quarter century later have picked up on that. But still, it's a relevant comparison. So all stocks have to do is have real earnings and dividend growth of 3% a year, which doesn't sound like a big hurdle. 3% real growth in earnings and dividends to justify being a 3% lower yield. That's what it would take to have a risk premium of zero. All right, over the last hundred years, what's been the real growth of earnings and dividends? About 1 1/2% a year. So if real growth is 1 1/2 and you've got a 3% shortfall on the 30 year tips, we're back in approximately that circumstance today, but with a much smaller gap. You've got long tips yielding two and a half, stocks yielding one and a quarter. So real growth has to be one and a quarter in order for stocks to be long tips. The historic norm is a little over one and a half. All right, that means you might have a positive risk premium of 0.2 or 0.3%. Nobody buys stocks expecting a risk premium of a quarter percent. And so, tacitly, it's a bet. Stocks are a bet that earnings and dividend yield will be much more robust than has been the case over the last hundred years. Is that possible? Absolutely. That's what Jeremy Siegel would argue. So this all forms the basis for us recognizing. I've long thought that CAPM is built on a shaky simplification of risk aversion. And this is not a criticism. A Bill Sharpe CAPM was a brilliant innovation, an enormous leap forward, but it tacitly assumes that investors hate downside risk and upside risk equally.
- [11:57]
B
Just beta is enough to capture it. And it's volatility in all directions.
- [12:02]
C
Exactly. Now, you or anyone listening to this podcast, I would pose the question, are you more uneasy about downside volatility or upside volatility I'm guessing that somewhere around 99.9% of us say, I don't like downside volatility. I love upside volatility.
- [12:24]
B
And even within that, there's a nuance where the behavioral scientists say that we hate downside more than we appreciate upside.
- [12:32]
C
That's exactly right. So this is where there's been a clash between behavioral finance and academic neoclassical finance and finance. Neoclassical finance theory posits that you have this symmetric risk aversion, which we know is false. In our paper, fear, not risk, drives returns. In the paper points out that the fear is symmetrical. It's fear of downside risk. It's fear of missing out on upside risk. And what's beautiful about this idea, by the way, we aren't, by a long shot, the first people to think about this. Cam Harvey wrote a wonderful paper looking at upside skew. And so if you look at semivariance, which is the downside half of variance, and skew, which is the upside asymmetry, those could be the two relevant measures for a redefinition of finance theory. In our paper, we've been faulted, absolutely correctly for not advancing a new theory, for merely offering a framework for a new theory. Bluntly, I'm good at math. Ed is good at math. We're both empiricists. Neither of us is nearly as good at math as the pioneers who've shaped neoclassical finance. So the paper is really an invitation to those who are spectacular at math to flip the signs and look at aversion to downside risk, an aversion to missing out on upside risk, and it could be skew, or it could be upside semivariance. Either way, you're looking at a measure, a quantitative measure, that simply adds one more dimension. Instead of aversion to risk, it's aversion to this risk and enthusiasm for this risk. Now, if you do that, you have an even more powerful tool, because in market, in horrific bear markets, people have lost a ton of money. The fear of further losses becomes intense. The fear of missing out on upside fades to be very mild. And that creates an asymmetry where you would expect a big risk premium. And during stupendous bull markets, basically 09 to 2025 has been a stupendous bull market with a couple of interruptions. It's been astonishing. When you get a stupendous bull market, fear of missing out becomes intense and fear of downside risk fades. In that circumstance, you would expect a negligible risk premium or even a negative risk premium. So revisiting finance theory to acknowledge human nature, to acknowledge that people do not fear upside, risk, except maybe just the little hint of fear that if I just made 20%, oh goodness, I could lose 20. But there's that little twinge, and it's only a twinge because we love getting the 20%. So this is an invitation for those who are extraordinary at math, at the math of finance, to go win a Nobel Prize by reshaping capm, to embrace the asymmetric fears, which are a fear of loss and a fear of missing upside.
- [16:38]
B
It's interesting. I. If I could interrupt for a quick sec, Rob. The. One of the things I liked that you guys wrote in this, in this paper was you referenced Thomas Kuhn's reference to paradigms. You're saying theories can fail in the data forever until a new, better paradigm emerges. So I guess is that effectively what you're hoping is that, you know, the data has proven long enough that, you know that it. The former way of thinking about risk theory as the, as the driver or predictor of equity risk premium or asset pricing, really as the underpinning for that CAPM and model portfolio theory, that really the data doesn't support it as much as people have thought it has for all this time, and that perhaps fear as opposed to risk is the way that people should be thinking about it.
- [17:23]
C
That's exactly right. So we're not proposing a new paradigm. We're proposing a framework for a new paradigm that we encourage others to develop. The academic community includes a lot of people who are way better at math than Ed and I are, but who are not nearly as good as empiricists or market historians as Ed or I are. And encouraging the academic community to stop clinging to a theory that is so patently, obviously lacking. I won't say wrong outright, I'll say lacking because it lacks an acknowledgment of human nature. It lacks an acknowledgement that people like making money. They're not afraid of it, they're not averse to it. And if you simply acknowledge reality, you're going to say the beautiful, simple math that's associated with capm, where risk aversion is relative to a quadratic function, and quadratic functions are elements that most of us are introduced to in roughly eighth.
- [18:38]
B
Grade, and they source data that's readily available, right?
- [18:42]
C
And eighth grade math, supplemented a little bit with some college level whiz bang stuff, creates a beautiful, simple theory. And we're saying that theory is wrong and the correct theory will be more complicated. But just one baby step in the direction of that more complicated, simply splitting upside and downside risk and saying people hate this and love this.
- [19:09]
B
Right? And you talk in the article as well about the difficulty of quantifying that, at least historically, about getting under the skin of what people, or you know, more appropriately, entire markets full of people are feeling at any given point in time. Can you talk a bit about, you know, some potential solutions to that, to that gnarly problem?
- [19:33]
C
Well, the beauty of introducing this new structure or building out a new paradigm is that it can be multi purpose, it can help define how markets function and it can help us also identify when we're in frothy bubbles or in deep bargain circumstances. Because if you can quantify fear of downside, fear of loss, and quantify fear of missing out, and instead of having just one metric for risk aversion, have two metrics. And if you can measure the market's collective aversion to downside and aversion to missing upside and measure them relative to one another, you have a wonderful tool for those who are willing to buck their natural human emotions to buy low and sell high.
- [20:34]
B
Right. And you. And, and in order to develop this, and again this is, there's a bunch of math that has to happen. And then to develop the data around that, one idea you posit is that we could harness the power of AI to test sentiment somehow to read, you know, every, every financial journal that's published every day, all the time, social media, that sort of idea, be able to, for the first time in our history, really get a sense of sentiment, a market sentiment that's not reliant on, you know, surveys like we might have done to sort of develop our gamma, gamma signal.
- [21:10]
C
By the way, speaking of AI, AI has, is phenomenal. It is people who don't play around with it for several hours a week are, are missing an opportunity. Two and a half years ago, we have an all hands meeting every quarter. Two and a half years ago at the all hands meeting, I, I said you're going to hear a lot about AI in the months and years ahead and it's the real deal. You're going to hear a lot about millions of jobs being displaced. I want to assure everyone who works for research affiliates AI will never take your job. Somebody who knows how to use AI better than you do may take your job. So study, get used to it, play around with it and don't be that person who's displanted by somebody who knows how to use it better than you. One of the things I love doing these days is Perplexity is an AI of AIs and it's a really powerful tool. It chooses which AI is best suited to your question, which is very cool. We took the Fear paper, gave it to Perplexity, and said, critique this. Tell us what we're missing. Tell us who's done work along these lines that we might have missed in our citations. Tell us what the three most important strengths and three most important weaknesses of this paper are. And Perplexity came back with, in effect, a referee report. When I was editor of faj, I saw hundreds of referee reports. I never saw a single one that was as detailed as what you can get out of AI today. And it came back with critiques very similar to the ones we got from the FAJ referees, which were, this is an interesting paper. It plows new ground, but it doesn't posit a specific theory. It falls short. And that's exactly right, it falls short. It falls short because it, it lays out a framework and invites the academic world to please stop clinging to something you know is flawed and have known is flawed for 50 years and start exploring ways to move it towards an acknowledgment of behavioral finance. Human psyche matters.
- [23:34]
B
So I want to circle back to just the application of this framework. We talked about stock selection or asset valuation. How would you see it applying to portfolio construction given what we, we know about, you know, potentially the, the wrong conclusions that we've been making for 230 years around asset allocation between stocks and bonds?
- [24:00]
C
Well, this can be used in a number of ways. It can be used for. Early in my career, I used dividend discount models to estimate stock returns company by company for the stock market as a whole and comparing it with long bond yields and for individual companies, comparing it with their own bond yields. And came up with very simple value strategies and a very simple tactical asset allocation strategy. In fact, the first Graham and Dodd scroll that I ever won was for a 1983 paper, Systematic Asset Allocation, which was subsequently relabeled Tactical Asset Allocation. It's funny. Bill Faust was the dominant player at Wells Fargo in the tactical asset allocation arena. He had billions of assets managed to reflect the calculated risk premium for the stock market. The higher the risk premium and the higher the equity allocation. And he came up to me at a conference right after we published the paper and said, you son of a bitch, you. You told the secrets. I said, bill, you're right, but I'm not going to apologize because you're the dominant Player, watch what happens to your aum. And anyway, it was, it was a funny exchange. But risk premium can be used for stock selection, it can be used for asset allocation and redefined to recognize that the risk premium really relates to aversion to downside risk and seeking of upside risk can make that a more nuanced decision.
- [25:50]
B
So, so in, in an example, say, say that there was a huge signal for fear of missing out, which if I'm doing the math right in my head, it's basically valuations are stretched because people are popular piling into whatever the latest bubble idea is. So equities look less attractive. So the, the high FOMO signal tells you to underweight equities. Is that correct? Is that correct?
- [26:16]
C
Okay, now that also is an implication of conventional asset allocation drivers using conventional risk premium measures. But you get a more nuanced picture, you get a, a measure of how it's stretched, why it's stretched. And also very interestingly, there may be, just as an example, momentum in the appetite for upside, in the increasing fear of missing out, there may be a momentum component that can help us stay in a frothy market until the turning point is closer.
- [26:59]
B
So sort of first, first derivative on that as it's moving.
- [27:02]
C
Right? So I see so much potential in this, but mostly the paper is a call to action for academia to please stop futzing around with a framework that you know is false.
- [27:23]
B
Well, this has been an amazing conversation, Rob. Thank you so much for indulging me and to chat about this article that I enjoyed so much. We do have a two point question at the end of our, end of each one of our episodes and I'm going to ask you just like I do everyone else, Rob, what was your first job in the industry? And if you could go back and take yourself for coffee on your first day, what key piece of advice would you offer yourself?
- [27:45]
C
My first job. Firstly, I tailored my education to this career. In high school, I was kind of ambivalent about do I want to pursue astrophysics or investments. And I fairly quickly realized, certainly by my first year of college that astrophysics is pure math and I'm pretty good at math, but I'm not brilliant at math. Math and advanced math, which is central to astrophysics, I'd be good enough to be an adequate astrophysicist. I would do no path breaking work. And I also realized at that time that investing was a domain of narrative storytelling and negligible use of quantitative tools and thought. I can do some pretty powerful work here. So let me Pursue finance. Well, my college education was applied mathematics, computer science and economics. Triple major. And my first job was at the Boston Company as a computer programmer. I was offered four jobs when I came out of college. Boston Company offered me the lowest pay of the four because with each of the four, I said, I want to spend a day, a week doing pure research of my own. And you'll have credit for it and you'll be able to monetize it, but I want to do some pure research. And most of them said, do that on your own time. Boston Company said, we love the idea, we'll pay you for four days a week. I said, done deal. And if I had any advice for my 23 year old self, I actually started at 22, shortly before my 23rd birthday. The advice would be, when you learn something new that is contrary to conventional wisdom, just be aware you're going to make a lot of people very angry because what you're doing is telling them that the core foundation of their career, on which they built their career is flawed. Case in point, when I wrote the paper Death of the Risk Premium, about five years after that, I was at a Q conference and went up to a guy I knew and liked and said, how's life? And he said, oh, life is good. By the way, I no longer hate you. And I said what? He said, when you wrote Death of the Risk Premium, I hated you because you were telling me that the essential foundation of my career was wrong. And I said, okay, why do you not hate me anymore? I said, because you were right. Anyway, it was eye opening. Let's see, I was 50 when I had that conversation. And so it wasn't until my 40s that I was beginning to realize how pissed off people got when you challenge their core principles, even if you're right. So I've been endlessly curious. That's been the essence of my career. And if something is a conventionally held view, I will just automatically ask the question, has anyone tested that? And if not, I'll go test it. And often it turns out to be correct. Great. Fabulous. Publish a sharp, minor paper and often it turns out to be wrong. Publish a paper that will rattle a few cages. And so the big shock to me in the first 20 years of my career is how often I piss people off.
- [31:44]
B
Well, having that, it's courage at the root of it, right? The courage to be right and the courage to be wrong and to rattle some cages along the way. But I'm glad you continued for the next 25 years after that death of the risk premium to challenge the industry and to challenge your peers. And I'm excited to see where this fear narrative goes. And hopefully someone takes up the mantle. And your challenge to get their calculators out and figure out a way to quantify all this.
- [32:13]
C
Exactly.
- [32:14]
B
I've been speaking today with Rah Barnat, founder and CEO of Research Affiliates. The article, again is called Fear Not Risk, Explains Asset Pricing. And I'm just so thrilled that you came on the show with me today again. Thanks again, Rob.
- [32:27]
C
Thank you. This has been great fun.
- [32:29]
B
I'm Mike Wahlberg, and this is Beam, the enterprising investor.