The Exchanges

Q Yeah. In addition to loss aversion, are there other specific tenets of behavioral finance that you then end up being able to map to the behavior of portfolio managers?

A We've, we've taken a practical approach and said, but let's actually analyze what goes on. And then if there is a bias that we found in the literature that relates to this, great. And if there isn't, So be it. Maybe we will end up inventing new ones based on the evidence we're gathering. We definitely do sometimes see the disposition effect as an extension of the loss aversion point, so not just holding on losers for too long, but cutting winners early. Less prevalence in professional fund managers than in retail investors, though. We see overtraining, which the literature would Call overconfidence, you know, thinking that I'm going to do something and the odds of that thing benefiting the portfolio are greater than fifty-fifty, which is overconfidence in its own way. So we see that from some people, but not others. You often see evidence of hurting around entry timing and that sort of, people are often very good at contrarian entry timing, catching it, you know, an inflection point quite well. But when something's been moving their way for a while, Take, say, a three-month run-up. They're, they're waiting and waiting, and then they start getting in, and they've come at the top, and showing that to them can be quite useful, because they can work steps into the process that get them to ask questions of themselves about, like, are we too late, you know?

AI assessment note: “We definitely do sometimes see the disposition effect as an extension of the loss aversion”

Q Yeah, let's break down those two components. So the first is there are behaviors that we know and sort of had a number of conversations the last couple of weeks about this with Michael Moveson and Annie Duke that get in our own way. And so how do you use data to get into the behaviors that are working and the behaviors that aren't?

A Well, it all comes down to what data we can get our hands on from a given portfolio manager. We start on the assumption that the very bare minimum we can get is historical daily trades and or holdings. That's the starting point, and with that, we can basically decompose every investment that you've ever made into a series of sub-decisions. No matter what you're doing, for any equity investment, you're making a picking decision, you're making an entry timing decision, you're making a decision about how big to go, and how quickly to get there, and adding and trimming decisions while you're in there, and then a set of decisions on the way out about how fast to get out, when to get out, and what to switch into. And if you look at it that way, and decompose every sort of investment story into those Constituent parts, then you can analyze each type of decision separately and say, okay, well, let's look at exit timing. Exit timing, every fund manager knows it's hard, and every book tells you that it's hard. We all know it's hard. However, you don't really appreciate it until you see your own exit timing in the mirror, and then what we're doing is saying, look, let's define what the exit is. Decision is, you know, it could be about looking at just the last day of trading, like the last shares you let go of. Are you doing it early compared with what would have been optimal, given a, a w…

AI assessment note: “we can basically decompose every investment that you've ever made into a series of sub-decisions.”

Q looks at what stock to get into, how to get in, and at what pace, trading around it when you're there, Should you, when do you get out, how do you get out, or how fast do you get out, and then the exit and potentially swap into something else. Are there other parts of the portfolio process that you're studying with the data that you gather on trading information?

A Well, so we're always looking at things by hit rate, payoff, and impact on the portfolio, and then what those things are, We might be looking at, you know, the individual skills that you just mentioned. We might be looking at just overall positions, which ones by sector, by country, by market cap, by liquidity, by a million other factors, which ones do we tend to get right most often and get most right? And then what we can do is, if you give us data about Whose ideas were these, or which research sources were you using, or what's your risk factor scores, or conviction levels, or anything like that. We can bring that into the mix, and then you can hone it that much finer. And then we also will look at, and for this we don't really need any extra data, how do you behave when you're winning, you know, when you're on a roll, versus when you're in the hole? That is a fascinating area, I think, and it's not something that anybody's really done a lot of work on in, in the industry, something that we're digging quite deep into at the moment, um, and I think that will hopefully map to things like the victory effect and snakebite effect and, and that type of stuff.

AI assessment note: “how do you behave when you're winning... versus when you're in the hole?”

Q When you go and talk to portfolio managers today about using Essentia's tools, what's the pitch?

A Depends on who we're talking to, first of all. So if we're talking to a PM directly, pitch is, are you interested in using technology to do the best possible job you can? And if the answer is, I don't care about doing the best possible job, but I can then. All right. You have bigger problems, and I will check you later. But we'll be here when you're ready to talk. Actually, what we find, though, is that there is a growing population of portfolio managers who have read a lot about behavioral finance, you know, who have taken an interest in this idea of bias and are aware that they have biases, but have never really been able to identify them or do anything about them. And so the pitch becomes about if you now have the ability to look in the mirror and see the truth about what is working, what is not working, in data terms with no judgment, knowing that it's not going to be shown to other people, this isn't about embarrassing you. It's just about helping you. Wouldn't you do that? Why would you not do that? And if technology could then be used to remind you of what you've learned from, from these insights at the moment when that, that habit starts to creep in again. So give you a sort of ex-ante prompt to use the learning in your day-to-day investment process. Wouldn't you want that?

AI assessment note: “So if we're talking to a PM directly, pitch is”

Q I'm also really curious about adoption. You mentioned that in certain instances, you shine a mirror on someone's actual activity, and they, their eyes might light up and say, whoa, wow, I've been doing that. How do you then get the behavior to change?

A That is the hardest part, right? I mean, the math, I'm not saying I can do it, but there are lots of PhDs out there, and very smart people can do the math. The hard part is actually getting somebody to use the math to change their behavior, and different people are different, you know, so we have some clients who are super low turnover, like, you know, very long term, and very, very deliberate in their process, and make very few decisions, and that means that analyzing the data is Kind of slow going. You do it once, it's not going to change by next month, right? But those people are often the sort who are very interested in codifying the investment process and using our asking nudges, as we call them, to record more information about why they're doing what they're doing. And in that way, go deeper on a smaller number of data points. We have clients who do huge amounts of turnover as well. And those guys, if they're doing, I don't know, a hundred trades a day, maybe, they're not gonna answer a nudge to ask them questions about why they're doing what they're doing. It's a question of packaging the output of the analysis with a human, you know, it's a human-machine combo. The human is an ex-fund manager, and that ex-fund manager comes and sits down with you, And get to know what your process is like, and what your attention span is like, and, and what you're really looking to get …

AI assessment note: “packaging the output of the analysis with a human, you know, it's a human-machine combo”

Q Yeah, you mentioned your nudges, and I know we, we've talked about this in the past, and it's something that I, I found really interesting, and really the second part of what we started talking about. So you now have the data, you've now figured out some things the portfolio manager maybe could do a little bit better. What are these nudges?

A So we have asking nudges and telling nudges. Um, The point of all of it is that there's too much information already out there. You can't expect anybody to actually pull anything in this. It's really like fund managers want to be served to stuff. Ideally, They want to be disciplined, but they're not going to remember to do that. It's very rare that you find somebody who remembers to journal every night and do it exactly like this and, you know, do it in some way that's actually analyzable. I mean, even the best journalers are writing in a paper notebook a lot of the time, which is not helpful beyond the act of writing it down, really. Or it is, but it's very time-consuming to go back over it. So what asking nudges do is They're triggered by a passing of time or an event in the portfolio, and they ask you questions. You just entered this name. Why did you do that? How long do you think you'll be here? What is your target price? You, you get to choose what the questions are, because it doesn't work if the questions aren't relevant to you. You drop it, like, immediately if we just asked you bog standard questions. So it has to be that these are things you've said you wanted to ask yourself because you know that down the road you want to be able to do analysis on whether we're better at this type of trade or that type, or do we ever really make it to our price target, or do we hold…

AI assessment note: “So we have asking nudges and telling nudges.”

Q So now that you've looked at all these data sets, mostly in the equity world, we know how hard it is already to just beat the market by picking stocks. We're compounding that with the individual behavioral biases that don't help. As you looked at all the data, you've worked with some teams to improve what they're doing. How optimistic or pessimistic are you about active management?

A I'm very optimistic. Charlie Ellis is a non-exec director of my company, so, you know, there are plenty of people in my life who are, who feel the opposite, except even Charlie would say, I never said there wouldn't be active managers. I said there would be fewer of them. And so I'm very optimistic for the few that will go on. And a lot of active managers will leave because it is getting so much harder to have a competitive advantage. But the ones who have behavioral alpha advantage, you know, where they actually are looking at their own behavior and saying, I want to continuously improve like an athlete, and I want to look at data to help me do that. There's low hanging fruit there. Even just pitting them against other fund managers, there's low-hanging fruit, and all it takes is the will to face the truth, the sort of Ray Dalio approach about, like, mankind needs to be honest with itself about what works and what doesn't work and what is true and what is not true, and if you're willing to do that, you do have an advantage. Some people will never be willing to do it. So I, I mean, I see managers, it's like I look at x-rays all day long of different managers, and I see a lot of skill. It's quite rare that I see somebody where you're like, oh god, you should really find a different occupation. Usually there's something to work with, and then it's all about how do I help that per…

AI assessment note: “I'm very optimistic. And so I'm very optimistic for the few that will go on.”

Q I imagine once you start You start recording data, it's trade entry data, and exit data, and now you start adding in these asking nudges, which is a whole different sort of data set. What's the most interesting data set that a portfolio manager has tried to get your help in assessing how they can improve their performance?

A The ones that are most interesting so far are ones where they're capturing a lot of data at the About the entry context. So, you know, we've, we've decided to buy this name, and we're recording. What was Joe's view? What was Fred's view? What was Jane's view? You know, was this contentious? Was it unanimous? How does it fit on, like, management quality, and valuation, and franchise? You know, all the different things that you might be rating these things on anyway as part of your process. If we can get you to do that in a structured data format, then we can start to Weave that into the analysis. And what we found so far is sometimes contentious decisions are the better decisions. For some people anyway, and, and for the first firm that we did this with, that was, that was the case, which is, and particularly around whether, so when, when they would vote unanimously on something, it was fine if they all trusted management. So there was like a link between management quality and unanimity, and that resulted in a good outcome, whereas if you didn't have that management quality agreement and it was contentious, you were more likely to end up with a good outcome. So that's just like the beginning of something about that healthy debate that everybody talks about in their marketing, and, and it's hard to do. It's way harder than, than people let on, but doing that in a way that It doe…

AI assessment note: “The ones that are most interesting so far are ones where they're capturing a lot of data”

Q It seems like all these tools start in the equity world. Probably not surprising because there is a lot of data and a lot of trade data. And you mentioned down the road being able to extend it past equities. Where are we in that time horizon and what do you think is next on the frontier?

A The most logical starting point for us will be futures because a lot of equity managers trade futures also. And that's relatively straightforward. The things that make futures complicated are more operational than anything else. There's a ton of demand from fixed income, and fixed income is very interesting because it's an area where, you know, active management, there's a strong argument for active management continuing to add value, and we'd love to address it. The hard part is figuring out how you judge a good fixed income decision. You know, and a good equity decision is typically I'm buying it because I think it's going up. I'm shorting it because I think it's going down, and it might be a relative value thing, but at the end of the day, that's how it works. Fixed income, you buying something, it's like you're throwing something into a witch's brew, and it's, you're buying it because it's going to do something to the bigger brew, and, and that could be a lot of different things. We're sort of taking a step back and looking at it more fundamentally and saying, well, Hang on a second. Is there a simpler way of doing this rather than reinventing the wheel on doing all those risk calculations and curve generation?

AI assessment note: “The most logical starting point for us will be futures”

Q What life lesson have you learned that you wish you knew a lot earlier in your life?

A Well, I've learned that failure is not only an option, but it's a necessity in life, and I think I was taught Quite young. Failure is not an option, and you do not give up, and you do not, you know, the problem with that, of course, and being an overachieving kid, you know, and those of us who are parents are now able to be conscious of this, is that when the kid finally fails, or, you know, maybe they're an adult by then, it hurts a lot. You know, it's a bit of a rude awakening, and actually, through my work at Essentia, this has all become Much more clear to me that if you start thinking about things in terms of feedback loops, then failure becomes a very valuable learning experience, and it gets easier and easier to separate the emotion of failing, you know, or getting something wrong from the learning, and recognize the emotion for what it is. Don't try to make it go away because it's going to show up somewhere else anyway, but recognize it for what it is, and then Have a rational conversation with yourself about what did I actually learn from that? What will I do differently next time? If you don't have those, then you can't do that, and, and learning becomes very difficult.

AI assessment note: “I've learned that failure is not only an option, but it's a necessity in life”

Q What's the hardest part of convincing a portfolio manager or a team to become a client?

A I'd say every team has at least one person who's into this kind of thing. You know, there's always that person. And, and often there's somebody in the head of equities or the CIO role who, Love this idea, and it's like, you know, we need data-driven feedback, so we know how to allocate our energy, and we want to prove it in our marketing material that we're skilled, and blah, blah, blah. There will sometimes be one guy who's really struggling on performance, and that's the one that they decide that we're going to do a pilot with. That's not a good plan, unfortunately. If you are going to do that, also have to do it with somebody who's doing well. Because that guy's under a lot of pressure already. He knows he's performing badly. He may or may not take well to having yet more noses in his business. Then you also will have people who They're just never gonna be up for it. You know, a lot of my old colleagues from back in the day, I love them dearly, but they're like, Claire, don't even, don't even tell me about this. I don't want it now. Like, I don't, I've been doing it the same way for 25 years, and if I lift the lid on it now, like, God help me. And it's like, you know what, fine. I'm not gonna try and sell to you, but I am gonna, Find out who the young person on your team is who's really, uh, you know, the next gen, and I'm going to sell for that person. And then also, I'm go…

AI assessment note: “you also will have people who They're just never gonna be up for it.”

Q Why don't you describe how you came to found Essentia and what your path was to get there?

A I was a long-only tech fund manager during the internet bubble in the late nineties, and it was great. You know, everything I touched turned to gold, and I won all the awards, and I was God's gift to fund management, and I actually Left and started a long short tech fund, uh, at the, uh, launched in March of oh one. By then, not everything was great, and everything I tested didn't turn to gold, and although I could short by that point, my approach in terms of fundamental analysis just wasn't working. That was fine, but my interest was in finding Way to improve. You know, what, if this isn't working, ok, but could somebody please tell me what I should be doing differently? And nobody could tell me that. All I could say was, well, one thing that would be different would be if you made more money.

AI assessment note: “I was a long-only tech fund manager during the internet bubble in the late nineties”