Episode 23: Decoding Behavioral Finance with Merle van den Akker

Sam 1:50

Hi, everyone. I'm your host, Sam Sivarajan, and welcome to this episode of The Future Ready Advisor. Today I'm here with Merle, van den Akker, a behavioral finance expert who is here to talk about practical applications that advisors can use in building trusted relationships with their clients. Merle.

Merle 2:12

Welcome to the show.

Sam 2:13

Sam

It's a pleasure to have you and look forward to diving in to our conversation. Let me briefly introduce you to our audience. Merle has a Ph.D. in Behavioral science from the University of Warwick in England. She is a behavioral science manager at a large Australian financial institution. She's also a university lecturer in Sydney, Australia, and writes the Money on the Mind blog. And today she's joining us from Sydney, Australia. Welcome.

Merle 2:45

Thank you, Sam. And it's a very healthy time difference to be dealing with. But here we go.

Sam 2:54

You're a behavioral science manager at a large financial institution, which is interesting and fascinating in its own right. I'd love if you could walk us through a little bit about your journey to today into your current role and what you do as a behavioral science. Sure.

Merle 3:12

So I knew very early on in my life that I want to want to study psychology. I didn't understand people much growing up, but I was fascinated by them almost as like a clinical research object. You get that a lot with people who don't understand people. T end up studying them. I was no different. But what I ended up doing was not a pure psychology degree, but I ended up doing a liberal arts and science degree, which at the time was very, very popular in the Netherlands, where I grew up and worried that my aldergrove or the liberal arts degree in the Netherlands is kind of this choose your own adventure, where it is very common to have double or maybe even triple majors. So I had psychology, but I needed to figure out something else. Now my dad, for my whole life, has been an economist. I used to have this rebellious teenage phase where I was like, Oh, I'll be nothing like you, Dad. I'll never study economics. I'll never make money. Because that's something to aspire to, appar, which obviously meant as soon as I grew out of this rebellious phase, I realized that, like, economics is actually incredibly interesting, and that is one of the best applied forms of psychology. So you can call it behavioral science, behavioral economics, economics, psychology. I honestly don't mind too much, but it's it's the thing that I find most fascinating because I, like many other people, especially at the time, was just rubbish with money. And I couldn't, for the life of me, figure out why. So I had an interest in it. Throughout undergrad I did double major in economics and psychology. So thank you, Dad. An then it thanks to a great conversation I had with my my faculties. Done Dean. He told me, know, probably you're a very good fit for behavioral economics. I at the time had no idea what it was. So it's accidental behavioral economist, if you will. And he told me to go to the UK. I don't remember ever having planned on having my education in the UK, but he was like, yeah, that that's really where this field is now starting off. And you know, that education that you can do a masters degree there and that will probably be your best bet. So I was like, okay, I'll go to the UK, get my masters in the UK, and it a lot done. My PhD in the UK realize that potentially being a pure academic wasn't for me at least, definitely not at that stage. And then through connections, through friends who at the time had already moved to Australia, ended up in working in Sydney for for a bank essentially easy, doesn't.

Sam 5:49

Awesome. Yes, the accidental behavioural scientist. I think many of us can again attest to that career path, if you will. So what exactly does a behavioral science manager do? Because it's not a profession that many of us would have grown up.

Merle 6:05

No, no. It is very, very recent. And that's I get this question all the time. So obviously this looks different from unit to unit. Some people are very internal consultants, maybe some people are very data research, have your data science heavy and others are very experimental in nature. And by training it kind of depends really what sector that you're in. And of course, how open your your respective company is to to extra orientation, because that's not always the case. Now, I think, you know, within the bank we got incredibly lucky because the team has existed for, I think, close to eight years now. So a lot of these know what is behavioural science, what is experimentation, how do you do it? Well, these conversations have already been had. Yo know, as as I joined. So that was really, really convenient to a straight out of a page to be able to work in corporate and just hit the ground rolling. So what my job entails is I am an internal consultant that that is how I am classified. But like many of my colleagues,I've got a background of going to experimental backgrounds and we design both in field and in lab experimentation to study whatever questions or concerns our internal stakeholder has.

Sam 7:18

So that would be including choice architecture or how you might position particular products or services, the types of questions that you might ask to evaluate demand.

Merle 7:31

Yeah, but it might even go as far as just checking if there is even any demand for certain types of products or services or in the way that they are designed, that they make sense and they don't really lead to any unintended consequences.

Sam 7:43

Now, one of the things that I find quite interesting is that in many ways Australia is similar to Canada. I mean, in particular, the Australian financial market is quite similar to Canada. It's dominated by a few large financial institutions and I think your superannuation funds are a bit like our defined contribution pension system. Of course there are differences, but I'm mostly interested in the lessons that I think that we might be able to draw from the Australian market, particularly given how advanced. I think some of the legislation on superannuation is and I think the idea that there's the behavioral science type of units are quite well-integrated into financial institutions from the sounds of it. I think legislation is one thing encouraging people to save is another. But from your studies and from your practical work that, we have all behavioural biases and one that I find fascinating and I'm using simply as an example, is that as people retire, I think that there is a reluctance to convert that into annuities, a guaranteed life income. And funny, I've seen research that says if you ask people, do you want a guaranteed income for life? The answer is yes. But then when you ask, would you like to sign up for an annuity? Invariably the answer is no. So maybe it's a branding issue. I don't know. But, I would love for you to share any lessons that you could from the Australian market, like best practices of how you kind of help, the end client or consumer kind of, engage in the right, or the type of behaviour that they really want?

Merle 9:28

Yeah, absolutely. I think I mean you've already highlighted there is a massive difference between both Canada and Australia and with regards to, you know, a defined contribution plan, it's not the same as as a superannuation. So the way that's set up in Australia so that your listeners are completely on the same page, is that essentially super is part of your wage package? It is very unlikely that that is not discussed and I'm actually pretty sure it is illegal. So when you sign up for a job, there will be disclosure like, Hey, this is your base salary and 11% on top of that would be your super unless it is framed as. Your base salary will be 100 K, which includes the super. You always know that super at a minimum is 11% of your base. It can be higher. Some government jobs have. This percentage should be higher. Now this 11% has only been recently cranked up to 11%. I think it used to be ten and a half. And I think during just before it was ten. Now, of course, with COVID, with the cost of living crisis and the recent rent increases that we've seen, this number has gone up. And I wouldn't be super surprised if it would go up again just because obviously and this is I mean, this is from ages ago, but essentially because hypothesis being that you need to save around 10% of your gross wage to be able to retire somewhat, that number seems to be largely based on that. Obviously, you know, Bekker was doing most of his research when most of these things hadn't happened yet. So that number may be a bit outdated, but I think that is largely what it is based on. Now, the way this works is that the super gets paid as your wage gets paid. In Australia, the most common wage payout is fortnightly and it gets moved directly into your super account, Whichever super fund you may be holding that with you as a consumer may find it very, very difficult to access the super like you can see it, yo know, just in a webpage or an app or whatever, very akin to any kind of investment fund, but it is not withdraw or bolt to the same extent as in you are essentially locked out. Now, there have been cases where people can access their super early, which has been during COVID as a lot of people obviously lost their job and they were, you know, they have money sitting there and they were going into debt. So this was known as early release. Super has been a lot of research on that part. It ended up working out. It was actually was a very successful program by, you know, by any measure to what to the extent that we are aware now, it may, you know, couple of decades down the line may not look so super anymore. And then there are conditions of extreme hardship in which you can also early access your super. But there are quite a lot of processes that you would have to go through to be able to access it that way. So in that regard is quite different. Of course, again, like this 11% as a default, as you're well aware of, most people as a result will be on this default, but you can't increase your contributions on anything that is increased on top of this 11%, which is a default and there's actually a legally required percentage. What you contribute done maybe for your retirement because there is tax benefits to doing that super as super contributions are taxed very differently from your wage. But you may also use that as a way to save up for your first home. That system has also been recently introduced where there are tax benefits to locking this money away and only being able to unlock it. Once you can prove that it would be going towards your first home deposit. So there are systems there that essentially allow you to lock away this type of savings, and that is essentially what super is known as a very aggressive, locked away savings account, despite the fact that this money is invested.

Sam: 13:16

Now, just not that you would be that interested, but for the listeners, the similarities between your super and our defined contribution system is we don't get 11%. But I think again, companies will make a contribution as a part of your overall comp into the defined contribution plan. The employee can match it and they can put excess amounts up to a certain limit. T can also put it because it's again treated favourably from tax. The investments in the defined contribution plan are managed by an institution, but the employee picks which manager and which type of investment strategy, etc. and it's locked in. So until you retire you can't pull these monies out except in a certain set of circumstances, But if you do withdraw it, you pay a penalty tax to do it.

Merle: 14:21

Yeah, so for us.

Sam: 14:22

Etc., I think these superannuation I like the example of what you've done because I maybe in some ways I would say it's a further ahead of the system in Canada. I mean, especially when you're talking about 11% and, legislated, etc., is that contribution. I wanted to put that is as a parallel to kind of then go into the behavioural side of things. How does companies, your work, etc. kind of encourage people? Again, part of having this in the legislation is one thing, part of explaining to people the benefit of having this grow in a tax advantaged account, etc. is another thing. But to get people to do it, to get people to the extent that they might be able to contribute, in addition to what the company is contributing, is their best practices that, are in place in Australia, that kind of and it will tie into a little bit of some of our other conversations, questions down the road about nudges, etc.. But is there certain types of behavioural best practices I guess that's there to kind of encourage Australian consumers to maximise their use of the super funds?

Merle: 15:39

Honestly, I don't think I've seen this to the extent where there is like a nation wide or like a unified approach to this. I just happen to know what the super fund that I am with really enjoy social comparison because on your balance they, they indicate that there's an Asterix behind. Obviously the first one indicating like your balance as of this time to state you know not the most up to date thing but also just in general like your second Asterix, if you will, is essentially them saying like, hey, this balance is below or in line with or ahead of your age group. Now, I remember initially coming to Australia and I only of course just gotten super. I got my first super payments, you know, finally started my balance, very excited. And they were like, Yeah, you're like several tens of thousands behind on your age group. And I'm like, Well, I've only just thought that, please. Like, it's very demotivating. I' like, Well, I've got a long way to go, I guess, and it's a very interesting one. Now I've seen the numbers for what the averages should look like for the age groups. And this is, of course, one of the main concerns when you're using social comparison or peer effects, if you will, if those numbers aren't enough to begin with for for whatever reason, because of early release super, because of additional withdrawals, because no super is based on wages and wages haven't kept up with inflation, the realistically like this is going to again lead into problems. I think in general what the super system has done incredibly well is that because super essentially they are managed funds. So like the super funds, they are managed funds, they just happen to be for super. So what they do make very clear is that they have advice services, so they have a much more general service, you know, have a free one off call done as soon as you want more personalised advice, they will charge you for it. But here's the convenience. That money will come out of your super balance so you don't have to save up for it. You don't have to take care of it in your day to day accounting. They will sort it out for you. All happens in the background. And so, you know, the the barrier to access also the barrier to advice is very, very low because the access to advisors is relatively high. As soon as you have the money in your super account. And I think advice here, it can start. I think you can have a 90 minutes, if not even longer. Call, which I think here, you know, depending on which fund you're with will it will cost you somewhere between 75 to $100 to just get like, Hey, where should I invest my money? How aggressive should that be? Doe this align with my values? What are my timelines? You know, these kinds of conversations. So in that regard, you know, you've got the social comparison or the peer effect, the barriers to advice. You know, if you really don't know what's going on is is relatively low. And of course, you know, the biggest barrier, you know, starting to save for your retirement has been resolved for you because it's set up by default. The only thing that you do have to do yourself and of course, it's not a small barrier is to pick the the fund that is the best fit for you, b this is an issue that we do see throughout Australia where every employer has their own default super fund. So if you're with a if you work for a bank, for example, it will be your banks super fund. If you don't want to be with a super fund that's going to be with different funds, as soon as you know you get your first payment, you have to tell your your bank super fund to move it to your preferred super fund. And then you have I mean, honestly, it's one for it's not hard, but of course, no people being as inert as they are that maybe it's a tiny barrier where there are people walking around who have like seven different super faults for all the seven employers that they've had throughout their life. And that makes life a bit more difficult.

Sam: 19:41

I think it's an interesting point. can see the double edged sword of having that social comparison. As you say, it's a nice thing to have to kind of motivate behaviour, but presumably if you are in the younger cohorts, where people aren't saving or maybe they think that they can wait to save, that becomes a self-fulfilling prophecy when you're comparing it to everybody else who's also not saving. at that stage right. Let's talk a little bit about your own research and work, including your blog, where I know you've had some very interesting interviews of behavioral scientists.

I'd love if you can share some interesting findings or applications that you think might be relevant to financial advisors in general, but particularly, to the audience

Merle: 20:29

a conversation I really enjoyed recently was with Matthew Battersby, and Matt essentially describes insurance as a cabbage problem or well, he calls it a sprout. I call the cabbage because I actually like sprouts. So that analogy doesn't really working out for me. And I'm not a massive fan of cabbage. So these these sprouts or cabbage or insert unpleasant vegetable in here is essentially just for insurance on investing has a very similar problem. You are dealing with two of the most prominent biases known to mankind, where you're looking at present bias, where if it's not happening in the now, if there's no problems now, I don't need nor want to deal with it. So insurance on investing, which both have benefits later, if they fall through the cracks very quickly and at the same time just as benefits in terms of insurance but also in terms of investing, it's not a certainty. It is a risk of a kind. Now, this risk may be relatively positive in investing. You know, you obviously in the longer term, your risk becomes a lot less risky due to compounding and how the market tends to track. But there is always a chance of losing your money. It is risky. The advantages we know much ahead in the future. This is not an appealing prospect to many people because it's not now and it's not certain. And insurance suffers very similarly, but unfortunately almost the other way around where there may be a benefit of having insurance later, assuming that the worst happens and you need insurance. But again, this is not a certainty. It might be that you know, you put money towards an insurance and nothing ever happens. Does that mean that this waste of money and if you did put money towards insurance and something did happen, is that done? The good thing because at least you pay the insurance. It is very difficult to balance these things out. And I don't think the human mind really has quite figured out what to do with

it.

Sam: 22:33

Yeah, it reminds me of this. I think there's a joke, but like a religious person, they believe in an afterlife, but they don't want to get to find out what it is. Right? I mean, and I think it's a similarly, I think in this case of insurance, it's one of those things that, when it pays out is when something bad has happened. So I think there's a lot of people that don't want to think about that bad event occurring. So the best way to not think about it is just not kind of buy insurance. So it's.

Merle: 23:04

Very true.

Sam: 23:05

Right? So I think it's a really good insight that you're putting. And the present bias, I think is huge, right? I mean, the idea that, you can't see the payoffs of either investing or saving for retirement or insurance for many, many years in the future, etc., it becomes less salient, to people to sit there and say, I should make the tough decision. Now, I should forego, the new phone, the new car, the new vacation, in order to pay the premium for insurance, to pay for an investment, etc.. But, I'm not going to see any benefit of it.

Merle: 23:39

No, So I'm unaware of how the defined contribution plans work in Canada. But insurance is actually a large part of the super system here as well. Or through your super fund, you can buy life insurance, death, cover income protection, disability protections on insurance, both temporal and permanent. And so this this product, this service have two coverages combined, retirements well, actually three. So the retirement savings, the investment aspect as well as the insurance aspect, it's one hell of a product from a behavioral perspective.

Sam: 24:17

And it's a big take up. So is the findings that people through the super that they tend to then add other products like the the insurance cover.

Merle: 24:28

So I think you are defaulted into I think the minimum levels of all of them that may be fund dependent. So don't quote me on that because I have experience with one fund, that one fund only. I made my choice, ones that I'm sticking with. It's because the sunk cost fallacy is a very strong fallacy as well. But this. So this just may depend, but I and again, you can talk to a I would recommend that although it's obviously not financial advice that people go talk to their financial advisor about what your insurance options should be. Because if I'm being very honest investing, I understand quite well and I think most people can can get into investing relatively easily. There's a lot of knowledge out there. I'm not saying jump onto tech talk. That might actually be the opposite of what I'm saying, but there is quite a lot of information out there on investing. What's tracking, what's life or form? Well, you know, who did what? How does Warren Buffett say that you need to do that? So there are like there's lots of information. I personally am not aware of the same hype around insurance. I don't think insurance ever made it really onto tick tock or it's the much more boring cousin of managing your wealth, which is unfortunate for insurance.

Sam: 25:48

It is. And again, I think it you've touched on some of the reasons nobody wants to think about the circumstances under which.

Merle: 25:54

Oh, absolutely.

Sam: 25:55

Yeah.

Merle: 25:55

But it's it's actually very unfortunate because I think as as everyone in the insurance know. So I Australia obviously the climate is relatively extreme one year we've got floods, the other year we've got bushfires and some more floods, some more you know, you understand where this is going and of course most people buy insurance straight after the floods or the bushfire when unfortunately a lot of people have of course already lost their their assets and hopefully nothing more than their assets. But then it becomes very salient. Obviously the media reports on that. It's not super surprising, but realistically, Don, the event becomes so salient that maybe, ma it's time to maybe now insurance can finally become a priority or not endless to do list on this.

Sam: 26:41

Now, you wrote a very interesting paper with Cass Sunstein, another behavioural scientist who is a leader in the work on thinking about nudges and coauthored the book Nudge Together with Richard Thaler. Can you talk a little bit about the paper and explain for my audience a little bit of what exactly in layman's terms, what a nudge is and some of your key findings?

Merle: 27:05

Thank you so much. I of course, happy to. So realistically, a not just a change in choice architecture. You're not messing with incentive systems, you're not redesigning the whole lot. It is often relatively small. I hesitate to say this relatively cheap because that really depends on how you scale these things. And they can actually become very expensive very quickly depending on how you want to do them. Now, what we tested specifically in this paper is just people's general attitudes towards a plethora. I think about 36 different nudges that you can apply in the financial domain from the perspective of a financial institution. So think of system one type defaults or system to type personalized feedback, comparative feedback, things like this. And it was just across these these nudges. We just asked it like, Hey, do people prefer system one nudge or system to nudge? Does it matter if it's being ruled out by a bank or like a financial institution or something else? Like if it's not disclosed, you know, do people have a much stronger preference for it? And does it matter if we know this not just transparent, meaning that we very much explain, this is why it's ruled out, this is how it works and this is the outcome. This is going to lead to, you know, with that level of additional information, knowing that an institution may rule this out for a very specific purpose, are you still happy with it? And what we find throughout is that people do have a general preference for a system to nudges. That is not surprising because they tend to be perceived as a lot less manipulative. We do find that ourselves as well. Well done. Again.

Manipulative how? Because what what ended up happening is that we didn't actually find an effect of transparency at all. And I think the reason for that is, is that most nudges, especially system two, but the system one, a couple of examples as well, not just are really quite obvious, like I own this, I understand that, you know, the very famous cafeteria example, like I put the fruit at the till and not the lollies and the and the call me and the chocolate and whatever. I think, you know, when people see that initially they may be surprised if they get explained like, hey, this is just because people impulse buy here and we want people to be healthier. But obviously, you know, if you want chocolate, it's just like two steps behind you. No one's going to care. Call that manipulative if you like. But I don't think the average person more than our paper indicates that probably people don't go as far as saying, well, if you don't remove the choice entirely, if you don't make it impossible for me to get somewhere, why would I call it manipulative? So we see that throughout. So there was a preference for system to not just but system one, not just were not rejected and they were still rated above neutrality. So that that was fine. Now, we didn't find the fact for transparency, as I said, probably because this stuff was really quite obvious. And then the last thing that we found, which we didn't have a hypothesis for, but what we did is that most of these nudges, they were actually split. There was a savings nudge and then almost equivalence from the, gosh, of course not all of them map onto one. But what we find is throughout the people actually do have a preference for saving, not just now. We've done similar work, which unfortunately is not yet out in the public domain. But what we tend to see throughout is that a nudge is rated really to the extent that is in line with what the person themselves wants to achieve. Now, spending restraints very often can be seen as a bit too big brother. It's like all you as a financial institution are telling me you should spend less on this or you should be moving towards this or this is the way to get there. Not ideal, however it seems, was like, Hey, we're trying to help you save. Given that most people do have savings as a goal, that tends to be much more preferred. Now you may make the argument isn't saving just the reverse of spending? And to some extent you may be correct, but it clearly it isn't perceived as such. So that's that is an interesting one that we'll still need to work on.

Sam: 31:24

I think that is very interesting. And I think, yes, you're right, it's the opposite. But I think to your point, I'm conjecturing here, but to be told not to spend seems to be directive and perhaps going against the grain of what the person, might be wanting to do where if you're saying and I'm assuming and I'm giving an example but nudge, you see this with a lot of apps, if you're paying and they say you can round up the payment and the difference goes towards your savings account, etc.. to me that is seen as benign, right? To most consumers, I'm assuming that they're not seeing this as you're telling me what to do. I'm already spending this money. You're giving me a nudge to say as I'm spending this money, I I'm not even going to notice that I'm going to round up and the $0.80 or whatever is going towards, my savings on the purchase.

Merle: 32:21

And the and that seems to work quite well. I think these apps should have that implemented. I think one used to be called a I think it's called raise now in Australia, but that one is actually no. So you micro save to invest. That one's very popular. It does very well.

Sam: 32:40

And again because I think in line with your findings it's nudging them towards something that they would want to do anyways or expressed interest to do. And this is just contextually it's happening at the time when they're making that decision and they can affect it. And it's a very small ask that you're doing and a very easy way to fulfil that task. Right.

Merle: 33:02

Is it also.

Sam: 33:05

A quick question and I'm going to ask you to look into the.

Merle: 33:08

Crystal ball.

Sam: 33:10

Future of wealth management and and behavioral science. No forecasts, just thinking, we've seen the rise of robo advisors globally and I think there's more and more A.I. tools for investing. even insurance as we talked about. I'm curious as to what you think this means for human advisors. how do they compete and differentiate themselves in a world of A.I. tools when the end consumer is?

Merle: 33:37

Oh, I think that's a very, very fair question. But I also have a feeling that if as a wealth advisor, you're only asking yourself this question now wasn't how do I differentiate myself against A.I.? You know, if anyone if everyone could have access to the same materials, if everyone kind of have access to the same numbers, if you will, done, I mean, forgive the rudeness and the directness, but if you're asking yourself that question now, I think you're a bit late. I think in general when it comes to and not to underscore any one's wants talent or skill or unique ability, but I think in general, most especially the larger wealth managers, the models that they're rolling, which has which have contained machine learning on the air for a very long time already, they've already been struggling with this for for a long time. And I think what sets apart a really good wealth manager for or advisor from someone not so great is that you've always had to go beyond the numbers because you are interacting with a human being. I think telling someone what to do because the numbers back that up is one thing getting someone to actually do it or even understanding really why someone wouldn't take that advice because it's just not aligned with their values or their short to medium term future. If you and your client aren't even on the same page as to, you know, what the timeline is that we're working with, A.I. is not going to fix that for you. But then again, it was never going to. I think I will help the dearth or really already really great wealth manager set themselves apart from people who might not have been that great to begin with and those just because now the reliance will not be on what you told me, but how you tell me how you help me integrate that information into my plan, how you how you allow this to help shape the plan how you take me on that journey as as a client. So realistically, I don't think that much is going to change, but that might be a very controversial opinion.

Sam: 35:51

No, I actually don't believe it is. Certainly not. Do advisors I mean, perhaps to the people that are building? Of course that might be. But look, I couldn't agree with you more. I am an investment guy. I totally believe in the importance of asset allocation and all of the efficient frontier and all of the finance theory that's behind modern portfolio theory. But at the end of the day, anyone that has been banking on that as their differentiator over the last ten years is probably had a challenge because, more and more, asset allocation, portfolio selection, these are very important, but they're table stakes. They're not very easy to differentiate your approach from the next advisors approach or the next firm's approach. So I think your point is bang on. I truly believe that the differentiating factor is being able to understand the human being behind the portfolio or that policy that you're talking to. What are their hopes, what are their dreams, what are their preferences? Being able to kind of help them, against themselves to some extent when they have these biases or short termism And all of the present bias that we talked about. And I think that in that type of set up an advisor that can speak to the human being but then can benefit from having the AI tools to kind of streamline, like a lot. what may have been, the time demanding work to, do some of this kind of portfolio selection, etc.. I think that it's it's not a threat to those types of advisors. It's actually a benefit to kind of help them do more.

Merle: 37:32

No, absolutely. I think AI can be used as a massive time saving tool and it should, if you can use it, t make your life easier, quicker, more profitable, less annoying, you absolutely should, you know, outsources as much as you can to the extent that you are obviously comfortable, you know, you know, data, privacy and other things permitting. But it's I don't think this is going to be the the wealth management revolution that a lot of people think it is. And there is research backing this up. So I have to admit, you know, for transparency, for your listeners, robo advice is not something that really has taken off in Australia because legislation doesn't allow for it. So in Australia, most financial institutions, if if they're not a specific financial advice institution, meaning that they employ financial advisors, people who are certified and trained for this, they are not allowed to deploy these kinds of things. So for example, a bank or an insurer, if it's not a financial advisor, you're talking to, the person that you are talking to cannot give financial advice. And the Australian Government has so asked the regulator here, a consumer regulator here, has cracked down on influencers. So social media, finance influencers, they've essentially been wiped out, talked about aggressive. But you know that that area of of employment, if you will, has essentially been terminated unless these people actually got their proper certifications in order. Some of them actually have, the majority have not. But so this this is actually not allowed in Australia at all, where you do see a lot of interesting research with regards to real advice. The integration of A.I. is actually Germany. Now you might be a bit surprised why Germany? Their banking system is really quite well advanced. I'm not saying it's the most advanced in the world because that would be a bit much, but it's relatively advanced that they have a lot of open data sharing agreements in place and they're very open to academic collaborations as a result. Lots of research comes from Germany. So what they have found throughout is that even if you and of course this is all extra mental lab work, so we're talking samples of like hundreds of people. Exactly. You know, millions or thousands, as you would see with entire banks customer base. But what they find throughout these types of research is that people can learn to appreciate an algorithmic trader. So if they enter an experiment, you know, they have money at stake, they can see a human advisor perform, they can see an AI advisor or robo advisor perform throughout. If they learn that the robo advisor outperforms the human, they will slowly move more of their asset allocation to the air. So as are some people, moving everything to the AI advisor does is great. It means that at least people care somewhat about performance if you add the additional level of costs to it and then the AI becomes significantly cheaper. Excellent. Most of the money moved to the AI. If the human advisor is significantly cheaper, most of the asset allocation moves to the human advisor. People are very cost sensitive, which makes it highly ironic that Robo advisor is not as highly taken up as a lot of people expect it to be because it is significantly cheaper. At least we don't see this uptake globally as much as I think a lot of people would have expected. Not even in the states where you think this would this would go wild, but apparently it really hasn't. So that is that is a very interesting one. But I think then you're moving more into the into the domain of algorithm aversion and people's issues to to trust something that they potentially don't fully understand, but also something where they're like, well, if this does go wrong, who exactly is liable? Because these are questions that, you know, for a lot of age I haven't really been thought through, or at least they are not common knowledge throughout. So that it does make people hesitant, which is which is very understandable to come back to the to the actual German research. So throughout, you know, cost matters performance matters with cost and this was of experiment mattered most but at the very end where because the algorithm this this was a very, very simplified investing experiment. So the algorithm was based on Bayesian updating. So the algorithm would outperform in this very, very simple environment anyway, because there weren't that many states of the world

algorithm outperformed the human investor on us. The human investor turned fully Bayesian people find that's quite difficult to do. So that's all right. And then in the end, you know, algorithm outperformed. People were given the choice. So given that the algorithm outperformed by this much, again, in real life, it is very difficult to put an actual number on how much an I can outperform a human being at any step of the journey. But we may get there in the end. But when people were presented with this number significantly better human advisor do significantly worse. Would you in real life don't delegate to the AI and the answer is no. The answer is that they would like their human advisor to be AI assisted. So even if you can show. Yes.

Sam: 42:56

Really?

Merle: 42:57

So even if you can show that the AI does significantly better, people will still be like, okay, but let's give that to a human advisor so that they can see that and then we can talk.

Sam: 43:11

it's very interesting. But and as you say, it's basic and we need to see how it plays out, but it sort of makes intuitive sense that people still want and maybe I can use the analogy robotic surgical tools that in many hospital shows and doctors that are using and I think it makes a doctor be that much more precise and more effective. But I think there's very few patients that would really want the robot to be operating on them alone independently. Right. They want it to be under the control of a human physician, etc., that can understand the broader picture and the broader needs, etc.. And so I can sort of see the findings from the German research from that perspective that, you want that hybrid. And so that's good news for advisors that, take the benefit that you can from A.I. tools. But still kind of refine the, the client service, the discovery, the support that you can provide and the understanding that you can provide your clients. Fascinating We're coming to the end of our podcast, so I have a few final questions for you that I ask all of my guests. And so it's in a rapid format, so when you're ready. So number one, professional, wha is the most important lesson you've learned over the years.

Merle: 44:40

I think that might be two things. One, measure everything because you can't trust your own memory to save your life. And I, I suppose, too, is you need to convince people that they need to experiment. Doesn't matter how you do it doesn't matter what you call it, but experiments, because you're never going to be sure otherwise that what you're advising is going to work and that what they want to do is the best way of doing it. And it's better to be proven wrong really, really early on with a small sample, but that is to be proven right with a huge, huge sample and God knows how many millions invested.

Sam: 45:22

I think those two are related right you measure and you experiment and I think in our own lives we do it a little bit although we do it implicitly and I think it to your point, I think more and more companies are coming towards this experiment or pilot, type of idea. I think it's just the I think your advice is good that it just needs to be more systematized and thinking about it for pretty well all things not just big initiatives but A/B type of testing as well.

So what is one practical tip you would offer listeners keen on applying your insights?

Merle: 45:58

Read broadly, read carefully, understand the results and the context that they were gathered up and if you will not apply this to someone beyond yourself, ask yourself the question Would I want this applied to me? And like, what would I do if this were applied to me? And if the answers are unfavourable, find out why and maybe hesitate to do this to your customers.

Sam: 46:27

I love that. I think you're absolutely right. I think the golden rule should apply. And even on, behavioral nudges, etc., to think, would you like it? And I think your your point about context, I couldn't agree with you more. I think, part of where I think we talked about A.I. before, part of what I think is, unique about being human is hopefully that we can contextualize, I think sometimes it's a lost art, but I think we can contextualize I'm not entirely certain that the current generation of AI can understand context well enough to kind make recommendations.

Merle: 47:02

It is difficult. It is difficult. I think work has been done on it's called homo silica or synthetic agents. Does this work is in the public domain and there's money in financial as well as tech firms who have started this type of work where they ask you to imagine it was certain things and you think as if it were an experimental participant. But what we've essentially figured out very quickly and this makes a lot of sense for LGBT history, is that Chat doesn't understand non-American contexts because CBT is Americans, because it might be trained predominately on American data, which that is just the Internet in English, which is predominately what judges would use trained on. It may be other ways American data are much more heavily in its model, and it might just be patriotic in nature.I honestly have no idea. I've never seen the model itself, but this makes it very difficult because if you don't want to ask chop, chop, chop, chop. It's just one that one of the things I'm most familiar with. It's not it's not it's not around against open AI by any means. But this is asking these these alums like, hey, you know, imagine that you're an Australian 26 year old woman with like this much, you know, student that in the Australian context looks completely different from being a younger American woman with that level of student debt, because obviously of the systems that are in place and I can't even imagine what happens if you take us outside of the Western world. And if you ask churchy or any other large language model to pretend to be a young South African man or an older Venezuelan lady, like, I have no idea. Also, I think these alums also don't know because that data is is much less available. And then context becomes a huge problem because it can't.

Sam: 48:53

No totally agree. Let's leave it there and pick it up and a future conversation to solve that contextual problem. But Merle, this has been an immense amount of fun. Thank you for joining us from early in the morning in Sydney, Australia. If listeners want to learn more about you, your your blog or find out about your work.

Merle: 49:19

Honestly, I used to be very active on Twitter with our haven't been much lately, which would be at money mind moral, but I think your best bet honestly is to just go to the blog which is money on the mind dot org or just check out my LinkedIn or frankly, yo can follow people as well. You don't immediately have to connect. That is probably your best. But like linked in this honestly, what's Facebook used to be, which is a terrible thing to say about a career website, but here we are.

Sam: 49:47

Awesome Merle. This has been.

Merle: 49:49

A thank you so much for having me.

Sam: 49:50

Again.