Episode 270: Victor Haghani and James White: The Missing Billionaires

28:01I mean, now I think they've sort of grown themselves out of a dangerous and probably a suboptimal position that they might have been in 20 years ago or 25 years ago. But, you know, again, that's all kind of speculation. You know, in the book, we talk about, you know, my experience at LTCM, which has some similarities to Rentech. It had a different outcome than Rentech. But in those early days, there were some similarities of thinking and Rentech sort of grew themselves past that dangerous period of time and LTCM didn't. I read that book too. I don't remember the specifics, but I think there were stories in there about very obviously lucky outcomes early on that contributed to the ultimate success.

28:41Yeah. And taking a lot of personal risk and the fund was running a lot of risk, but they figured it out and they got past it and they had good all of the above. I mean, really smart guys. I'm certainly not taking anything away from them, but today, I think that their main constraint is just how much of the stuff they can do rather than risk is just not a constraint for them anymore, it seems. Interesting. What does the merchant share tell us about asset allocation over time if it's possible to estimate expected equity returns? So, you know, as James was saying earlier, that the Merton share is helping us to realize that when expected returns are higher, we should allocate more to the asset.

29:21When risk is higher, we should allocate less. And so if it is the case that we can estimate the expected return of stocks at different points in time and estimate the risk of stocks at different points in time, and if those risk and return numbers are changing over time, then it's very likely that we're going to want to change our allocation to equities over time, too. When the risk and return is more favorable, we should want to have more invested in equities and vice versa. Now, it is possible that risk and return could move in exactly a particular way where our asset allocation wouldn't change.

30:01But most people believe that the risk of the equity market is probably, at least over longer horizons, a little bit more stable and that the expected return is moving around quite a bit. And when I say expected return, what I really mean is expected return in excess of whatever the safe asset is. But I think that's important to get out there. It's not just the expected return of equities, but expected return minus the safe asset expected return. And how confident do you think investors should be in estimating expected stock returns? Confident enough to act on it. So again, we're talking about estimating the expected return.

30:38We're talking about estimating the center of the distribution. Of course, the outcome is never going to be exactly the expected return. It's going to be above or below that for each period of time. But we think that using something like the cyclically adjusted earnings yields of equities is a coarse but sufficiently good measure of expected long-term returns that we think that investors should be changing their asset allocation based on that relative to the expected or offered return by the safe asset. I'd like to caveat that. We think that measures like one over CAPE are sufficiently good measures of expected return to be useful for broad markets, for things like the S &P, US equities, European equities, things like that.

31:26As you go from broad aggregates like that to individual sectors, from individual sectors to single stocks, the quality of the metrics degrade really quickly. So if you ask the question about US equities, I would say people should be confident enough that it's quite useful. If you ask the question about a single stock, I would say not useful at all. What about styles, like values cheap relative to growth right now? Yeah. So the core logic behind why measures like one over cake do a reasonable job forecasting long-term returns is because at a high enough level of aggregation, earnings look not totally bond-like, but more bond-like than stock-like.

32:16This comes back to an observation that Shiller has written extensively about, that stocks have dramatically more volatility than earnings do. It is sufficiently high level. And so when you're dealing with large aggregates where if you look at total corporate earnings in the US and the history of that over time, there's little squiggles in it, but the squiggles are really little. It looks surprisingly bond-like. And so just like you would for a bond, looking at the earnings you're getting divided by the price you're paying does a pretty good job. As you start going towards aggregates where the earnings look less and less bond-like, that logic holds less and less.

33:03So even for individual styles like value stocks or growth stocks, anything where there's really significant regime shifts over time or where if you look at the earnings and the squiggles aren't kind of de minimis, that's a pretty good suggestion that that metric is not going to help very much. Right. And by the way, when we talk about the squiggles, we're saying looking at blocks of earnings, like 10-year blocks of earnings, that in quarterly earnings or even annual earnings, they're big squiggles like you can get. But looking at cyclically adjusted earnings, they're just not bouncing around all that much.

33:41That was the Shiller paper, or maybe it was Shiller Campbell, actually. Yeah. So back to market level, dynamic asset allocation, how does that change if we decide that, hey, we actually reject the ability to estimate differences in expected returns over time? Well, if we reject the ability to estimate expected returns, then in some ways, I think you just can't invest in that asset at all. So that's the first thing, right? I mean, at a high level, like if you look at the equity market and just say, I have no idea what the expected return is of equities, then say, well, you really shouldn't be investing there.

34:15So you need to estimate it somehow. Now, if your estimate is just going to be, I always believe that equities have a 5 % long-term real expected return. Well, if that's your belief that the expected return is always the same, then your asset allocation would change because we know that the safe asset return is changing over time. So the risk premium would be changing and you'd want to change your asset allocation. If however, you said, I believe that equities always are going to give me 5 % above the risk-free asset and that their risk is constant, then your asset allocation wouldn't change over time.

34:55It's probably better than other things you could do, but it's not great. But if that leads you to just always having 65 % of your money in equities come rain, come shine, come anything, that's better, we think, than not having any money in equities over time. And it's better than being two times leveraged in equities over time, but we think that it's substantially suboptimal, not just a little bit, but substantially. I should add for listeners that whenever we talk about risk premium, what we mean by that is the expected return of an asset minus the risk-free rate. Yeah. I want to dig more into that.

35:29Can you talk about how the real return offered by long-term tips, which as John Campbell and Cochran, I think both explained to us, is the risk-free asset for long-term investors. How does that real expected return affect the optimal risky share in assets for long-term investors? It's interesting to think about a thought experiment where, let's say, your expected real return for equities was 10%, amazingly good. But in that environment, you could get 10 % risk-free from tips. Let's say you're from a 30-year tip or some very, very long-dated tip. Now you have two long-dated assets with the same expected return, equities, let's say they have 20 % annual volatility, how much equities would you want in that environment?

36:16And in contrast, imagine an environment where the equity real return is much lower. Now, let's say it's 5%. But in the second environment, the real return on tips is zero. Which environment would you want more equities in? We think it's clear that you would want dramatically more equities in the second environment. Because one interesting thing about investing that we think doesn't get talked about as much as maybe it should, is that all investing is relative. You have to invest your money in something. Even if it's in cash, you're really investing in overnight deposits. And so whenever making an investment decision, it almost never makes sense to look at one investment in isolation.

37:06You're always investing into something and out of something. And so you have to look at the relative risks and returns. And in theory, you could do that relative to any benchmark asset, but an asset that feels like a risk-free asset to you is kind of the natural baseline to use when when, you know, comparing to other investments. How do you forecast volatility for determining the optimal risky share? So, you know, there's lots of observations that we can look to in coming up with that. And in practice, you know, we think that blending some of the different sources of information makes sense. You know, I think realized historical volatility, giving more weight to long-term realized volatility than to short-term is one good input.

37:53You know, the options market is a really good input. It's not perfect. There could be risk premium that's also sort of built into the options market, but that's a good metric too. Although, you know, in general, we're making long term investment decisions, we kind of want to think about longer term volatility forecasts, and the long term options market isn't that liquid or transparent to people. But I think some combination of implied volatility from the options market, historical volatility, and also just thinking a little bit about a bottoms-up approach to estimating volatility in terms of how volatile are earnings, how volatile are discount rates.

38:30I think all of those things together are important. And what's really important also, that it's not just the volatility that's important, but it's the full distribution. And in the book, we go into this that the Merton shares for this very specific set of assumptions that are very ideal and non-realistic assumptions. We know that the stock market can gap down and we know that we can get a Black Monday in 1987 or in the 1920s. So when we're really thinking about the distribution, we really want to also build in probabilities of large tail events as well. And so, you know, it's important to get a really full distribution and not to assume that stocks are this sort of continuously trading brownie in motion sort of thing, but that, you know, there are real tails, there are fat tails, there are gaps down, all of those things.

39:23And we really need to build that into the distribution and then think about maximizing our risk adjusted return based upon that. So it's not just thinking about the center of the distribution. What we really care about more than anything almost is this small probability of these big tail events. And this expected utility framework is great for thinking about how much weight to put into those events and how much that affects your asset allocation. I do want to get into utility, but real quick, because Canada, we've decided, or the country is phasing out real return bonds. So how would dynamic asset allocation work in a country that doesn't have an equivalent to TIPS in the US?

40:07What you want to do in making your decision is try to imagine if TIPS existed in Canada or in some market, where would they be trading? And then it's like, okay, unfortunately, we don't have that to invest in. So now my choice is I can invest in these different assets that have different amount of risk. So let's say I have now equities and I have nominal bonds, but I don't have tips. So I use tips kind of as my numeraire in some ways. And now I have to make a decision between how much I want to put into equities and how much I want to put into these nominal bonds, which are not really that safe.

40:44And this framework can easily handle that. We wrote a paper about that, you know, what to do in a world where there is no safe asset. And even tips are not totally safe, right? I mean, there's default risk, there's tax risk on these things. There's the fact that the index of CPI is not going to be exactly your index. So it's quite nice that you don't need a real investable risk-free asset. You just need a risk-free asset to benchmark everything to and then decide what's the optimal risk return portfolio and size for you to have. So you don't need tips to be able to use this risk-adjusted return framework to do your asset allocation and sizing.

41:23To maybe give a slightly more concrete answer, what most people would do to kind of get that proxy tip yield is look at the return on long dated nominal bonds that do exist. And then, you know, look at inflation expectations, inflation forecasts, historical inflation, and take nominal yield, subtract out some arbitration of inflation metrics, and then you get a forecast real yield. That's not as good as having a trading inflation market, but it's probably the best you can do. All right. So we've touched on this idea of expected utility. To start this more detailed discussion on that topic, can you describe the concept of expected utility?

42:08Yeah. So I think it's useful to back up and talk about utility first before we get to expected utility. And my impression is that a lot of people have a bad experience of utility from kind of Econ 101, you know, a long time ago, where it felt like this very abstract, squirrely, unreasonable thing. But we actually think it's something that's really intuitive and that people feel really intuitively. When I talk about it with people, I usually go back to talking about gummy bears. I have this irrational, I don't even have much of a sweet tooth, but I really like gummy bears. And I think most people have had the experience, let's say, you know, you buy a pack of gummy bears and you're eating them, that the last one in the pack is still good, but it's not nearly as good as the first one was.

42:59The first one is really good. And then by the time you get to the last one, I'm still going to eat it. I'm not going to turn it down. But it's slightly grudging. And that phenomenon is true in almost everything we consume. Economists would call it diminishing marginal benefit of consumption. And I would argue that this isn't an accident. It's not something that's culturally contingent. It's not something that some people feel and others don't. It's a really deep part of normal psychological functioning. Imagine the opposite. Imagine that for anything we had, like as I eat the gummy bears, the more I eat, the more I want.

43:43You know, controlling your desires would just be impossible. And so this phenomenon that consumption isn't symmetric, and that there's this diminishing marginal value of consumption is a really innate part of human nature. And utility is really just a recognition of that. And the utility curve and the shape of the utility curve, something like a logarithm, is really just trying to mathematically capture this phenomenon. And in consumption terms, the way it tends to manifest is that more consumption is always better, but it gets less and less better as you consume more. And then dynamically, the way it tends to manifest is that from a given amount of consumption, losing 10 units of consumption and gaining 10 units of consumption are asymmetric.

44:39So losing is worse than winning in its impact on me. And as those amounts get more extreme, the asymmetry gets more extreme. Going back to the first example I talked about, no matter what your wealth is, if you compare going broke to doubling your wealth, the monetary and values involved are identical, but the impact really isn't identical. There is a massive difference, I think, for almost everybody experientially in going broke versus doubling your wealth. And the utility curve is just recognizing that. And the degree to which that asymmetry exists is different for different people. And so the exact shape and curvature of a utility function is going to be different for different people in different contexts.

45:29But the core idea is the same. Now, expected utility. Again, we think this is just a mathematical expression of something that's super intuitive for people. Let's say you have a choice to make and the outcome of the choice is uncertain. You know, there's two possible outcomes. Well, what are you going to do? In some sense, you're going to intuitively or explicitly make a judgment about what the probabilities are of the different outcomes and what the impact on you is of the different outcomes, and then you're going to weigh them. And that's just what expected utility is doing. It's taking the probability weighted average of different outcomes, except you're not probability weighting that let's say we're talking about financial decisions where there's dollars and cents involved.

46:17You're not taking the weighted average of the dollars involved because if I look at my going broke versus doubling my wealth, if that was 50-50, well, the weighted average of the dollars is zero. It looks like no big deal. So by calculating expected utility, you're taking the weighted average of the impact on you, which is your change in utility, not of the dollars involved. And that's the core insight. Everything that you were just saying matches up so well with a lot of the literature from psychology on the relationship between money and happiness, where some research finds there's a plateau in happiness.

46:52Some research finds that happiness increases with log income, but it's never a linear increase of happiness with increasing income. Yeah, and I think it can't be. For people who have that kind of profile, or even worse, a profile where it's more valuable as you get more, that really verges into what we call addiction. What is the certainty equivalent return? One challenging thing about utility, expected utility especially, is that it's a yardstick, but it doesn't have a very natural sense of scale. So it's really nice to be able to talk about something that has a scale that we're used to dealing with.

47:36And in making financial decisions, return is the obvious sense of scale, right? Right. So the certain equivalent return is the risk-free return, which delivers the same expected utility as a given risky investment. So I'm considering some risky investment. It has different outcomes with different probabilities. I can figure out its expected utility, but I can also figure out the expected utility of a risk-free investment. And the nice thing about certainty equivalent return is that it says, okay, I have this risky investment. Maybe it has a 20 % expected return, but it has a ton of risk. That's equivalent in the utility it will deliver to me of making a 3 % risk-free investment.

48:27And another word we use for certainty equivalent is risk-adjusted return. And that probably makes the sense of it even clearer, that you can think of the certainty equivalent return as an expected return on the asset with a built-in risk adjustment. We did a survey where we went to people and we said, okay, imagine that you're not allowed to invest in anything risky ever again for the rest of your life. how much would you need to be paid to forego investing in anything risky for the rest of your life? And so what we found was, how would you go about trying to answer that question? Well, if you sort of have this, if you read our book and have this expected utility framework, it would give you a way to answer it.

49:13But people really intuitively come to an answer. They realize that, well, you know, maybe equities are going to give me an extra four or 5 % return over time, but they're risky. I would be indifferent between being able to invest in equities with this 4 % or 5 % risky return. I would be indifferent between that and just getting paid an extra 2.5 % risk-free. Give me 2.5 % risk-free, and that would be about equivalent to 5 % risky, sort of scaled down for how much equities I want to own. And when we did this survey, we really found that people were answering the question. Now, these are all, you know, our client bases tends to be a lot of financing people and there's professors in their finance and all that.

49:56So it's not super surprising, but it's just interesting, you know, that when the question is framed like that, you really kind of need an expected utility framework. But at the same time, you don't because you kind of get the idea. And that risk adjusted return is a pretty cool thing. Like if you really stop people from investing in equities, you're really taking away something very valuable to them. Yeah, that's interesting. It may be obvious from the description that you just gave, but can you talk about how you'd use the certainty equivalent return to evaluate two alternative investments? Yeah.

50:27So the full credit answer is find portfolio weights, which generate the optimal certain equivalent return. Optimizing certain equivalent return or optimizing risk-adjusted return is basically the same as optimizing expected utility. And so for almost any portfolio question, the most general full credit answer is just find weights that maximize risk-adjusted return. But heuristically, the nice thing about thinking about risk-adjusted return is that it just lets you compare assets head-to-head. So if I have one asset that has a 5 % expected return with 5 % volatility and another that has a 30 % expected return with 100 % volatility, I can't just compare their returns.

51:14It's not obvious at all how those assets relate to each other in terms of quality on a standalone basis. Whereas if I had two assets, one has a higher risk adjusted return. If I'm going to do A or B, I'm going to do the one with the higher risk adjusted return. That's a higher quality investment at that size. Or another kind of interesting way to think about it is that if you look at a given investment at a given size, and its risk-adjusted return is lower than the risk-free rate, you have too much of it. Do less of it. If it has a positive expected return, you'll want to do some of it. Anything with a positive expected return in isolation, you should want some of.

51:57But if at a given size, its risk-adjusted return is lower than the risk-free rate, you have too much. How does the certainty equivalent return deal with skewed expected return distributions? Yeah, so the nice thing about the whole expected utility risk adjusted return framework is that it's not making any built in assumptions about the nature of the return distribution. So your utility function is intrinsically skewed in that the same magnitude negative outcomes hurt more than positive outcomes help. And then when you compute expected utility, you're just taking probabilities and outcomes. So the machinery of it can really deal with any kind of distribution, you'd be at normal distribution, heavily skewed, kurtotic, whatever.

52:51The outcome you see is that for normal shapes of utility, if you fix the expected return and you fix the Sharpe ratio, the framework will naturally prefer positively skewed to negatively skewed distributions, which is to say, if all the other characteristics are the same, same expected returns, the same variance, you would prefer a skewness where there's a high chance of losing a small amount of money and a small chance of making a lot to a negatively skewed investment where there's a high chance of making a little bit, but a small chance of losing a bundle. Okay. Interesting. If we take the certainty equivalent return and the Sharpe ratio, and we're using them to evaluate something like covered calls or something else with an asymmetric return profile, what's the main difference in our decision going to be?

53:49Yeah. Sharpe ratio really breaks down as you start looking at heavily asymmetric payoffs. For instance, I can construct a portfolio or an investment that has normally distributed returns and a given Sharpe ratio, or I can get the same Sharpe ratio with a heavily, heavily asymmetric payoff. And so the Sharpe ratio doesn't really tell you anything about that. It's not useful in distinguishing between those two. But what the utility framework tells us is that people with utility-like preferences should distinguish between those two. You should really prefer, let's say, normally distributed to negatively skewed, and you should prefer positively skewed to negatively skewed payoffs.

54:36And so using expected utility or risk-adjusted return will naturally deal with that, whereas Sharpe ratio won't. And one of the famous papers that kind of talks about this with respect to Sharpe ratio is the Gertzman-Ingersoll paper, Sharpening Sharpe Ratios, where they basically show that various option combinations can really be used to manufacture Sharpe ratios for people who are only sensitive to Sharpe ratio and don't care about other qualities of the return distribution. Let me just mention that we have a whole chapter in the book on the use of options within the expected utility framework.

55:15And I think we give options a bit of a hard time. And this is actually the 50th anniversary right now of the Black-Scholes model. And two of our friends and partners are, well, the Black-Scholes Merton, we could call it. Two of the three initials are friends. And I'm going to see them soon. I'm kind of hoping they haven't read that chapter, although Myron has read the chapter and objected somewhat. But in general, we kind of feel like that options are not super useful for individual investors in general. Or, you know, there's a hurdle there. And, you know, we decided to put this chapter in the book because there's a ton of books out there that are about trading options to get rich.

55:52But there's not anything that we could find out there that really addressed options within the context of individual personal finance and investing. And even great books like Jack Boggle's books or others, you know, don't even address options at all. Like in Common Sense Investing, there's no mention of options or in, you know, Charlie Ellis's book or whatever, or even William Bernstein. Authors that we like a lot, they just don't really get to talking about options. So we thought it would be useful to bring them in there and talk about them within this risk-adjusted world. The other thing I just want to add briefly here also is like all of the expected utility that we're talking about, you know, we're talking about it with one particular utility function, what we call constant relative risk aversion utility.

56:34We like that utility function. There are other utility functions that people could have, other shapes, other preferences. But for background, we're sort of making a lot of our statements based on that sort of function and shape. But again, we think it's a good one. And we also think in most other plausible ones, you get the same sorts of results out, as long as you have this diminishing marginal benefit of more wealth or more consumption that James talked about. I'll also add that this question about skewed distributions arguably is the question that ultimately resulted in our writing this book because the first note Victor and I ever wrote together back in January of 17 was called A Sharper Lens for Sizing Up Nickels and Steamrollers.

57:17and it was exactly about the differing perspective you get between a sharp ratio and a utility analysis and looking at investments that have a lot of tail risk. Incredible. I'm glad you brought that section of the book up. I'm going to ask a little bit more about it now, Victor, because actually when you sent Cameron and I a message on LinkedIn, it was the craziest thing because I knew you as a character from a book that I'd read about LTCM. And then I had your paper open on whether options make sense in the portfolios of retail investors. So I'm sitting there reading your paper and I get a message from you on LinkedIn.

57:51Anyway, it's crazy. So I can ask about that a little bit more explicitly. You touched on a little bit. Do you think that options generally make sense in the portfolios of retail investors? The answer to that is no. For a yes and no answer, I mean, could make sense sometimes, but generally speaking, no. James and I don't use options in our personal investing. We don't think it's a great fit for most individual investors under most circumstances. And we're talking here about options on financial assets, like home insurance, right, is kind of an option. Home insurance, you know, against fire or whatever, that can make sense.

58:27Sometimes it doesn't. I don't actually carry home insurance, but don't tell my wife. But for most people, home insurance is going to make sense. And that's a form of an option. But in general, we think that for individual investors, that options don't make sense in general. Not always, but in general. Yeah. That's a great paper. And I think that the section of your book on that is similar to the paper, right? Yeah. The Journal of Portfolio Management or the Journal of Derivatives let us base that chapter on that. Nice. Okay. When we started talking about utility, James, you kind of mentioned that people might not remember it so fondly, and I think it does get criticized sometimes.

59:01How do we know that expected utility as a framework is actually a good model for normative personal finance recommendations when so many people don't behave the way that the model predicts? Yeah. I think it's worth distinguishing behavior and preferences, and also between important and non-important situations. A lot of the experimental evidence violating people optimizing expected utility is really in very small dollar situations where the relevant asymmetry is just not as big. But more importantly, I think I want to distinguish between people's preferences violating expected utility and people's behavior violating it.

59:45As investment advisors, researchers working with our friends and clients and other people, our overwhelming experience is that when we really focus on people's preferences, reasonable utility functions do do a pretty good job representing people's preferences. And again, I think that's not an accident. That's really innate. That comes down to the fact that human nature would look totally different if we didn't experience some kind of diminishing marginal benefit of consumption. And so one of the first things we say in the book is this book is meant to be normative, not descriptive. What we're trying to do is help people behave better in line with their preferences, not just reinforce the behavior they have.

1:00:35And I think seeing that gap between preferences and behavior, it shouldn't be so surprising because the real world is so complicated. You know, there's taxes, there's longevity uncertainty, there's returns uncertainty, there's so many different outcomes that intuitively mapping your preferences onto a consistent set of choices, given all the choices we have to make, is really difficult. And there may be some people out there who are just intuitively amazing at that. Warren Buffett, I don't think he has an expected utility framework, but he seems really naturally good at incorporating the same kinds of information that a utility framework would, risk and return and the asymmetry of benefits and things like that.

1:01:24Most people aren't that naturally good at it. And so if I thought that there were a lot of people whose underlying preferences were really different from utility assumptions, I would be more worried. But that is not our experience at all. That's really interesting. Again, in the psychology literature, there's lots of – like I talked about how when you actually look at whether money makes people happier, there are diminishing returns. But then there's also research showing that people predict that more money will make them a lot happier, even though that's not actually the case. I think that kind of speaks to what you're talking about.

1:01:57So why hasn't this framework caught on more widely? We think there are a few reasons in different domains. If you look at Wall Street, for example, I think the big reason is that Black Scholes came along. This kind of work was really hot. People were really interested in it. And then there was the derivatives pricing revolution. and a lot of the people who had the mathematical and technical backgrounds to read Merton and Samuelson and read this kind of literature just went into derivatives pricing and risk-neutral valuation and just the whole risk-neutral world kind of took over everything. Outside of Wall Street, I would say I feel confident that the reason it hasn't caught on is not due to lack of utility, no pun intended.

1:02:49And one piece of evidence I have for that is that if you look at a group of people who is the most dedicated to making consistently good financial decisions, successful professional gamblers. Successful professional gamblers overwhelmingly use variants of the Kelly criterion, utility-based decision-making, decision-making that integrates risk and return in some consistent way. And I think the experience of professional gamblers who don't is that they don't remain professional gamblers for very long. As for why it hasn't penetrated the broader investing world more widely, I think that's a big part of why we wrote the book.

1:03:34For something to penetrate widely, you really need not just great, but really, really heavy technical papers. You need a broader conversation for a broader audience. And that's really what we hope our book to be. It really is incredible. I agree with you. We did an episode on covered calls a while ago. That's why I asked about that. And that's a topic where it's like they get evaluated with Sharpe ratios and you get people selling ETFs from mutual funds going around showing how high the Sharpe ratios are with no mention of expected utility. So I totally agree with you that it's something that people need to understand better.

1:04:14I think you guys have done a great job in the book doing that. And I think maybe another reason is that, like we talked about up front, that there's just a lot of people who have this bad experience with dealing with utility, you know, from taking Econ 101 or conflate this kind of utility-based decision-making with Benthamite utilitarianism and, you know, have this feeling or, you know, have, again, from, you know, Econ 101, try to apply utility to non-financial things, you know, where it becomes much, much squirrelier and just have this kind of gut reaction that like, oh, this stuff isn't very useful.

1:04:51But our experience that when we really take people, regardless of their priors, when we really take people through this framework and its assumptions, people actually find it really natural and intuitive. All right. So, Victor, I mentioned that I knew you prior to this as a character in the book, When Genius Failed. So I got to ask, how did your experience, because I know the story from the author's perspective, but not yours. How did your experience with LTCM affect you professionally and personally? First of all, it was very, very painful. I mean, I can sort of smile about it now, 25 years later, but it was painful.

1:05:29It was embarrassing. Yeah. I I mean, it really hurt. I think that I'm incredibly grateful that as a group of partners, we sort of had each other and maintained our friendships till today, you know, throughout. And that helped it somewhat, but it was just really, really painful. But over time, I don't know if you guys have read the book by Dan Gilbert, Stumbling on Happiness, as he and others in this sort of happiness field point out, that, you know, over time, no matter how bad a thing happens to us, we kind of drift back to our normal state of happiness. And, you know, that took a little while.

1:06:03But, you know, I think I eventually got there. professionally, it marked a change in direction for me, where, you know, I kind of wanted to take time to understand what had happened to learn from it. And I stopped in that field. And I kind of feel a little bit fortunate because in 2007 and eight, this bigger storm came through the markets. And I feel good that I didn't sort of devote those interim years to going back to the cold face of relative value trading. So yeah, I think it gave me a good feel for the range of my utility function, you could say. How did you decide how much of your wealth to invest in LTCM?

1:06:41Well, we have a chapter on this in the book about my own personal perspective and experience. It's not about trying to rewrite or set the record straight or anything on LTCM. It's just about how I experienced it and dealt with it and thought about it. And by the time we started LTCM, we had this long track record of doing relative value trading that went well beyond my time going back into the late 70s with John Merriweather and the earlier people in the group seemed like this highly profitable, high sharp ratio type of activity with a really good sense of where the money was coming from. It wasn't like a black box sort of thing at all.

1:07:18And we had a lot of market experience. We really thought about what was causing things to be out of line and what would be catalyst to bring them back. We didn't just look at valuations. We didn't just use models. We used a lot of experience looking at things and it looked very attractive. So as a partner of LTCM, my approach to begin with was to say, well, I'm going to just take a certain amount of money that I'm going to be happy with. If I lose everything else, what's an amount of money that I'll just be happy with having in safe assets on the side? And I put some money aside and then with the rest of my money, I put it into the fund.

1:07:53And I don't remember exactly, but I would say that probably 80 % or so of my liquid financial net worth. I had invested in LTCM one way or the other. And as we discuss in a chapter in the book, I think in hindsight, and without knowing what the outcome turned out to be, I mean, with perfect hindsight, I wouldn't have invested anything. I would have invested and taken it out at the end of 1997. But I think that there's a reasonable case to be made on an ex-ante basis that using an expected utility framework, I would have reached a different decision and and that other people can sort of benefit from those experiences and mistakes now.

1:08:35Yeah. Super interesting to hear you talk about it. And I'm sure that our listeners will find the same thing because that's not an everyday event to have lived through as a person. No, it wasn't. Learned a lot from it. It's actually also roughly the 25th anniversary. I think some people would say, I don't know, September 12th or something like that, 1998. So a couple more weeks and it'll be exactly 25 years. But yeah, the fact that the book is coming out now is a coincidence. The book is not about LTCM, even though the experiences of LTCM certainly have shaped. And James had his own mini LTCM sort of experiences too.

1:09:07All right. So we spent the last little while talking about sizing positions and risky assets and the expected utility framework. I want to move on to spending, which you also cover in the book. What do optimal solutions for lifetime spending from a portfolio look like? There's two characteristics that I really want to focus on. The first is that spending should be proportional to wealth. This is probably the most important one. And obviously, nobody is going to follow a spending policy that every day is literally proportional to wealth. But in some big picture, with a lag or with some averaging or whatever, for a spending policy to be close to optimal, it has to be proportional to wealth.

1:09:49And that falls out of the math, but I think it's also pretty intuitive. If you have any spending policy that's not, your wealth drops a lot, your spending doesn't change, you're going to run out of money pretty quickly. So in some big – or the other way, if your wealth increases a lot and your spending doesn't change, then you're just massively underutilizing your wealth. So the most important character is that spending should be proportional to wealth. The proportion can change over time. depending on your other preferences, it might be optimal for you to increase the proportion over time, decrease it over time.

1:10:25But in any given time, it should be proportional to wealth. This creates a really interesting relationship, which is that imagine you are, you're following it perfectly and your spending is perfectly proportional. Well, then the volatility of your spending is going to equal the volatility of your investment portfolio. And that connection is both a major constraint and also really useful in helping figure things out. That's what creates the necessity to not kind of think about your spending and your investing separately, which is in practice what we see often happens, but to set your spending and investing policies jointly because you have to be proportional to follow even a reasonable spending policy.

1:11:18And let's say you can tolerate 10 % of spending volatility, but then you go invest in a portfolio with 30 % volatility, you're going to have a huge problem. How did Merton and Samuelson solve the lifetime investing and spending problem? They solved it differently. In both cases, their solutions were really interesting and innovative. But I actually think this is a case in which the bigger part of the genius was imposing the question rather than the specific way in which they solved it. And the way they posed it was every period, say every year, I have two questions to ask and answer. How much should I spend and how much risk should I take?

1:12:04specifically, how much, I have some risky portfolio and I have a risky asset. How much should I invest in the risky portfolio versus the risky asset? And how much should I spend? I have that same question every year. What is the joint spending investing policy that maximizes my lifetime utility of consumption? That was their setup. And that setup really unlocked a huge amount of interesting results. Now, in some cases, with a lot of starting assumptions, you can get to analytic solutions for that, which Merton showed. These days, we tend to solve that system numerically rather than analytically, because we want to include a lot of real world hair, like taxes, longevity, uncertainty.

1:12:59There's a lot of real world situations that make the analytic solution unworkable. The analytic solution still gives us a lot of intuition about the character of what's going on. But I really think the number one innovation was just posing the problem like that. And one of the first things I talk about with clients now is that I still think the two most important questions clients should focus on are how much to spend as your wealth changes and how much risk to take. And a lot of people, even highly financially sophisticated people, never, almost never explicitly think about how much total risk am I taking?

1:13:43They think about individual investments, how much they have in individual stock or individual portfolios, but it's rare to encounter somebody who has explicitly framed the question for themselves as this year and next year. And how much risk should I be taking? What are the inputs? If we're talking about optimizing lifetime spending and asset allocation, what are the inputs that people should be thinking about? Yeah. So in a really highly stylized world, the major inputs are your personal risk aversion, Your personal time preference, which is basically the discount rate you apply to your utility of spending over time.

1:14:26The characteristics of the risky portfolio. So highly, highly stylized world, the expected return volatility of the risky portfolio. In a less stylized world, just the distribution of returns of the risky portfolio and the risk-free rate. And then if I add a little bit more real-world context, I would add to those your age, your longevity distribution, the priority you place on bequesting wealth versus consuming wealth. Or another way to think about that is your utility of bequest, separate from your utility of personal consumption, your subsistence spending, meaning for most people, they get to a really, really, really bad place before spending literally goes to zero.

1:15:20And for most people, there's some what economists call subsistence level of spending, at which That's basically as bad as it can get. If I'm not personally, but let's say I'm spending a million dollars a year and my spending drops to$10 ,000 a year. Well, there's really no difference between$10 ,000 and zero at that point. So identifying subsistence spending, tax rates, not just the absolute rates, but the structure of taxes. Those are all things that are inputs to our process and inputs to a less stylized process that involves a lot of real world things. So you've presumably got this in a model.

1:16:01Can you talk us through the mechanics of how you would use this framework with like a client, for example? So we start off by sending them a lot of questions. And some of the questions have factual answers. What is your net worth? Things like that. And some of the questions are less factual, like, how do you think about the value of spending on yourself versus leaving money to your family, to philanthropies, things like that? So first, we collect a lot of information from them. And that has personal factors, their age, their wealth. It has questions related to calibrating their utility functions.

1:16:41We don't treat those as kind of final answers, but I think of it more as the calibration questions give me a starting guess as to how risk averse somebody is. And by risk averse, what I really mean is how much curvature is there in their utility function? And the calibration questions also give me a sense of what is there that's consistent or inconsistent in how they're thinking about it. You know, sometimes people answer calibration questions very consistently. And sometimes they answer them quite inconsistently. And then you have to drill down into that and, you know, understand where that comes from.

1:17:20So all of the answers are kind of starting guests. We take all of those. We put them into a big optimizer we have, which simulates a risky portfolio. It has taxes. It knows about the actuarial tables. So it has a stochastic longevity built into it. And then we apply what I call the general method, which is optimize utility of consumption over time. We find the spending rule and the risk rule that maximizes their lifetime utility of consumption. We have to do that numerically by simulation rather than analytically. And then what we show people is basically what those policies look like, both in terms of, okay, spend 4 % this year, 4.5 % next year, 5 % next year, whatever, and also the heat map of if you follow these optimal policies, what is the probability at a given year of your spending falling below a given level and your wealth falling below a given level.

1:18:29And then we talk about it with people. And then there's this kind of iterative process where between the answers people gave and these heat maps that they're looking at, we see how comfortable they look. And sometimes we get it really close the first try. And sometimes people answer in a certain way, and then we look at the heat maps, and they're like, oh, no, no, This shows a 30 % chance of my spending falling below X five years from now. I can't tolerate that. That's absolutely intolerable to me. And that means that they're either more risk averse than we originally thought, or that their current spending is just too high relative to their wealth.

1:19:18And which is which will depend on the individual person and their individual situation. and we just drill into it until we feel like we understand what's going on. But we really use the starting questionnaires and these heat maps are rising from the optimal policies iteratively to ultimately get to both a set of utility and bequesting risk aversions that feel comfortable for somebody and a set of policies, both investing policies and spending policies that feel comfortable for somebody. I should say, not just comfortable, it's really important to me to have policies for clients that are comfortable, because if it's uncomfortable, you just know people won't follow it.

1:20:05But it's also important not to have policies where you only think about comfort, but ignore how suboptimal this is. There can be a policy that's comfortable for somebody, but it's just massively, massively suboptimal too. And that's not good either. One of the places where the utility framework comes in is helping people find policies which are jointly followable and comfortable. And I won't say strictly optimal, but in the vicinity of optimality. And one of the really nice things about the whole utility framework is that if you boil everything down to one dimension and you look at, let's say, expected utility or risk-adjusted return as a function of your one dimension of how much risk you're going to take, say, it's going to look more or less like a parabola.

1:20:59And in the vicinity of optimality, the slope is not going to be very great. So being really far away from the optimal is not good. But because the slope is just pretty shallow in the neighborhood of optimality, it's not important to get bang on. It's important to be in the right ballpark. So why don't we see Merton Samuelson spending built into typical financial planning softwares? We don't know what everybody does, but I think the answer is similar to why you don't see the merchant share around a lot, which is that it hasn't been written about in the mainstream very much yet. And what any investment advisor who's trying to appeal to a lot of people wants to do is take advantage of concepts that are already out there.

1:21:54So that's probably a long-winded way of saying a much more glib answer, which is people haven't read our book yet, which I say that tongue in cheek. But I think the generalized version of that is just that these concepts are only starting to move from academia to the mainstream. And as you see it enter the mainstream more, it would really surprise us if people don't move in these directions. So you mentioned a questionnaire that you use with clients to kind of work toward figuring out their utility function and their risk aversion and all that kind of stuff. How can a listener who's maybe managing their own money or maybe working with an advisor, how can they figure out their utility function and their risk aversion and their time preference?

1:22:41There's a lot of calibrating questions you can ask yourself. And I would say there's really formally three domains for thinking about utility. There's utility of wealth, utility of spending and utility of bequeathing. And in some situations, utility of wealth and utility of spending are really combined. So the Merton-Samuelson framework really combines them, where if you're following a proportional spending rule, utility of wealth and spending are really, really combined in there. Whereas let's say it's a context where you're not thinking about spending at all, like you're running a fund or you're a gambler.

1:23:21You're trying to make bet sizing decisions. You would think about utility of bankroll rather than utility of spending. But there's these calibration questions you can answer in any of those cases. And all of them have roughly the same form, which is, how would I feel about, for example, a 50-50 chance of losing 10 % of my wealth versus making 10 % of my wealth? Would I want to take that opportunity? Would I reject that? Or would I be indifferent to it? And by finding where you're indifferent, if you limit yourself to CRA utility, then you can calibrate your utility by finding that indifference point.

1:24:09And there's a lot of ways you can ask that question. My personal favorite way of asking it, let's say with respect to wealth, is let's say there's an opportunity where there's a 50 % chance of losing X percent of wealth and a 50 % chance that you'll become a multi-trillionaire. What is the highest X for which you would not take that bet? So clearly, 50-50 chance, 5 % versus a trillionaire, I would take it. 50-50 chance of losing half of your wealth versus becoming a trillionaire? Much harder question now. Your answer maps one-to-one onto your level of CRA risk aversion. I should add that we talk about this in greater detail in the book, but one thing we feel pretty good about is that for most people, institutions and circumstances, CRA risk aversion is flexible enough.

1:25:10It certainly doesn't capture all the nuances of utility in most circumstances. But in our experience, for most people in situations, it's flexible enough and it's not necessary to move beyond that. The academic literature explores tons of different classes of utility functions and can get all kinds of bizarre results depending upon how bizarre your utility function looks. But in the real world with real people, we think CRA utility really does a very good job. And you'll see that variously called CRA utility, isoelastic utility, power utility. they're all really the same thing. And maybe I'll also add that even though there are different contexts where you might want to think about utility of wealth, utility of spending, utility of bequesting, when working with clients, I usually start off with utility of spending.

1:26:07And the reason is, I think for most people at most levels of wealth, spending is really where the rubber meets the road. People have a lot of experience, real personal experience dealing with what it's like having to cut their spending by 10%, what it's like having 10 % more to spend. They really have lived experience and strong preferences about that. As you move away, the questions become somewhat more abstract. And so I like to start with spending. Now, for some people with extraordinarily wealthy people whose spending is like a tiny fraction of their potential spending, spending is less of a constraint.

1:26:55And so you tend to focus on wealth more. But for most people, I feel like the thing that constrains their risk aversion more than anything is the volatility of spending they're willing to tolerate and that people really give different answers sometimes when asked about spending versus wealth. And my experience is when you dig into that, the spending answer is almost always the one reflective of true preferences. Sometimes you encounter people who, for example, express a desire to have very, very low volatility of spending. But then when you ask the same questions about wealth look like tolerant to very, very high levels of wealth, of wealth volatility.

1:27:44And if you dig into that and kind of walk people through the consequences of that, let's say the consequences of having a very volatile investment portfolio while following a very tight spending policy, I would say in almost all circumstances, the result isn't that people say, oh, no, no, I was wrong. I could really tolerate a lot of spending volatility. That's almost never the case. People usually say, oh, I wasn't really thinking about the connection. I'm really not so tolerant to wealth volatility either. Here's a question for both of you. What advice do you have for young people as they start out their financial journeys?

1:28:23I'm going to say I have three pieces of advice, but as I say them, I might come up with some more. The first one is that you're young, your human capital is very large relative to your financial capital for most people. And you should give that a lot of focus. You should really focus on what you want to do with your life in terms of work and career and all of that. and sort of thinking about the risk adjusted value of your human capital is a good way to think about it. You know, the different careers are going to appeal to you inherently in different ways, but also you should take account of the riskiness of different kinds of careers and make some adjustments for that, you know, as well.

1:29:05So I think a big focus on human capital would be number one. Number two, I think would be make sure you're starting your financial education, you know, work some books into your routine. You know, you don't want to spend all your time reading finance books, they can be pretty boring. But make sure that you're getting a financial education by that time if you haven't had it already. And talk to people and think about it and try to really get yourself educated. And then the third thing is start to develop some good habits. Even though the academics would say, well, you want smooth consumption, so maybe you should borrow money when you're young and then save more as you get older.

1:29:42We would say develop some good habits, start doing some saving, but especially tax-deferred saving, and start thinking about investing. We would say invest in low-cost, broadly diversified index funds. That's what we would say, but do your research, do your financial education, and hopefully you'll get to that too. So I think those would be the three bits of advice. Yeah, I think that was really well stated. I mean, the only thing I would add is that that people should really read our book too. It is a really good book. For the average person off the street, it's still pretty intense in terms of the level, but it's a whole lot better than trying to read one of Merton's papers.

1:30:20You guys both have experience in academic settings and you've seen some pretty wild stuff in practice. We talked about LTCM, Victor, and James, it sounds like you've had a smaller but similar experience. What's the most important piece of wisdom in investing financial wisdom that you can leave our audience with? I mean, I think if there's just one sort of golden rule of finance or investing, it would be what I said earlier on, which is that you can't expect higher returns without taking more risk, that return and risk are bound together. This is kind of related to the theme of the book also, that we bind them together using this expected utility framework.

1:30:59But the fact that you can't expect higher returns without taking more risk doesn't mean that you can't get more risk without getting more return. It's easy to get risk without getting return. You can go to Vegas and you can get negative return for risk. So there's an important corollary there, which you shouldn't expect higher returns for risks that you can eliminate through diversification. And it's just like, I think that if you get this into your head, that you can't expect higher returns without taking more risk, and that this is like enforced by the competitiveness and efficiency of markets.

1:31:33I mean, everybody's looking for more return without risk. It's the proverbial free lunch. And that's what makes it so difficult to find. So I think if people get a respect for the efficiency and competitiveness of markets and that you can't get higher returns without more risk, that's going to be a great guide to your financial ship through your life. I agree with all that. Maybe one thing I would add is I would advise people to spend as much time thinking about the how much to invest question as they spend on the what to invest in question. And I think that goes for professional investors, whether you're at a bank or a hedge fund or individual investors, for social reasons, for a variety of reasons, almost all the focus is on what do I invest in?

1:32:24And the reality is that you're going to make a lot of investment decisions over your life. not all of them are going to be good. And if you have a good way, a robust way of thinking about how to size those investments, how to take the right amount of risk, you'll be able to survive and better than survive through all of the bad decisions you make. And it's just not plausible to think that you're not going to make a lot of bad decisions over a sufficiently long period of time in terms of what to invest in. I think that's really the key that is really neglected in books and the literature and in people's own time space.

1:33:03I love both those comments. And I think that they go together really nicely because as we talked about earlier, even somebody with a good framework might invest all of their wealth in Tesla if they believe that there's a 76 % expected return. But if they understand how risk is price and how markets work, then they might revise downward their expectations. Our final question for both of you, how do you define success in your lives? The full credit answer, for me at least, still a work in progress for sure, as it is for most people, I think. But one aspect of it for me, I was exposed early to the famous Carl Sandburg quote, time is the coin of your life.

1:33:43It is the only coin you have, and only you can determine how to spend it. Be careful lest others try to spend it for you. And, you know, certainly when at the end of every year, when I go back and look at the previous year and, you know, kind of think about the scorecard, this isn't the only thing on it. But did I use my time wisely is a really big part of it. And, you know, I say that without any judgment about what use of time is worthwhile, you know, for any person. I think there's a lot of variation across individuals and even within individuals. What I find worthwhile spending my time on is very different now than it was 20 years ago.

1:34:25But the wonky economic perspective is that time is maybe the ultimate scarce resource. And I feel really strongly when I don't use it well, that that was not successful. And when I do, that it was. I like that, James. That was great. Great to hear that. For me, I guess, I feel like I just have had so much good luck in my life in different ways. and good luck that has been delivered by different people. You know, my parents were great. Well, my mom's still alive. My parents have been great. My family and the people that have come along in my journey and have helped me and given me chances and educated me and protected me and all of that.

1:35:07Like, I just feel so, so fortunate, not only being born at a great time into the history of the earth or the history of humanity, but also I feel super fortunate in terms of what people have done for me in my life. And so, you know, I think that I'll define a successful life as one where my deficit of what I've done for other people versus what I've been given is as small as possible, maybe even positive. I think that's at the end, if there's one thing, I think that's what I would like to measure it by. And I'm still in quite a big deficit. So I'll see how things go. I'm getting older. Wow. Yeah.

1:35:45Great answers from both of you guys. This has been a fantastic conversation. We really appreciate you coming on the podcast. Thank you. Yeah. Great to meet you both. Thanks so much.