Wealth and Asset Management Platform Development with MATLAB Web App Server
Dr. Martin Tarlie, GMO
Nebo Wealth is an award-winning, open architecture asset management platform that represents a leap forward in goals-based investing. Our pioneering approach to asset allocation and portfolio design—built on a big idea that risk is not volatility but “not having what you need, when you need it”—delivers the personalization that clients expect with the scale and efficiency that advisors demand. In this presentation, Dr. Martin Tarlie will motivate the problem, introduce some of the key ideas, and illustrate how GMO has used MATLAB® to create both a web-based application and a scalable API-enabled platform.
Published: 22 Oct 2024
So I appreciate the invitation to speak. So what I want to do is start with just a little background into what Nebo Wealth is. So it's really a new venture incubated within a mature asset management firm, Grantham Mayo Van Otterloo. But it really leverages a lot of the core capabilities, insights into asset allocation. And we see Nebo or the underlying concepts as sort of a new branch of asset allocation and the way that we're thinking about risk.
I think that's one of the reasons that we're winning all of these industry awards. Before we get into the details of what Nebo is, I thought it might be helpful to just share the origin story, because it really does connect, ultimately, into MATLAB. So back in 2013-- so this is more than 11 years ago-- I was actually in the equity group, Not in the asset allocation group. But I'd been doing more and more work related to asset allocation.
I got to the head of asset allocation, Ben Inker, very well. And one day, he came down to my office and he said, I just got off a call with one of our large corporate clients. They're doing more and more work in the defined contribution space and they want to if we have any thoughts on glide paths. Maybe you want to look into glide paths.
So I started looking into glide paths, and it became very clear, very quickly, that the purveyors of glide paths were not really explaining how they were being built. So that motivated Ben and I to go back to first principles and to think about the risk that the asset owner faces. And what was intuitive to us was that the risk the asset owner faces is they don't have what they need when they need it. So we said, well, why don't we build portfolios to minimize that risk?
So we put out a white paper in April of 2014, and we knew we struck a nerve, because it was one of the most downloaded white papers that year on Advisor Perspectives. It was also picked up by the Wall Street Journal, and they wrote a long article about the concepts and the ideas. And then to make a long story short, fleshed out the framework over about a five or six-year period. What we didn't realize at the time is we were really kind of offering a replacement for risk as volatility or what's called modern portfolio theory.
But early 2019, what we realized is that we had been trying to commercialize this in the institutional retirement space. And what we realized-- and it took us an embarrassingly long time-- was that the institutional retirement space were never going to be innovators or early adopters. Their primary motivation was they don't want to get sued. So we started exploring the financial advisory space, and we wanted to go out with this idea and test to see whether or not this idea about risk, the story about risk as not having what you need when you needed whether or not that resonated.
But we knew we couldn't sell them a portfolio. What we needed to do, if it was going to be successful-- we were going to have to offer them a platform where they could build their own portfolios. So I started looking around and said, OK, well, how are we going to build this platform? GMO has historically been a MATLAB shop. So I stumbled across the Designer, the App Designer, and I said, well, this reminds me of Delphi, which, for those old timers out there who remember, in the mid-1990s, Delphi was sort of a drag-and-drop but very powerful object Pascal-driven platform for creating windows-based applications very rapidly.
The Designer in MATLAB reminded me a lot of the Delphi platform. So I started building the prototypes in the App Designer. And that's what we went out with. It was a dopey prototype, but we showed it to people. We said to the advisors, what do you think about this idea on risk? Here's how we are thinking about potentially implementing it. Are you interested, and would you work with us to help us create a platform that you would use for all of your portfolio construction?
So that's how it got started. We actually are continuing that process, but now, we're at a point where we have adoption, we're signing clients-- and I'll tell you the rest of the story there. But I think it's fair to say that without MATLAB and having the Designer available to us to easily create a web-based application and being able to iterate very rapidly as we get user feedback, it's probably fair to say that Nebo Wealth would not exist. We also have additional connections
So first, in those first 3 and 1/2 years, it was really just myself writing the code, and Lawrence Barton was managing all of our networks. So he built all the pipelines. We use an Azure back end, or an Azure cloud service to actually run the web-based application. But as we've expanded, we've added additional software developers.
Paul Wang's wife actually works at The MathWorks, and he's actually an alumnus of the MathWorks. We've also added, more recently, Darshan Pungaliya, who's also a MathWorks alumnus. So we have very strong connections to MATLAB and MathWorks, and we're extremely grateful for the platform that has been available to us as we've built out Nebo Wealth. OK. So enough of that as the background.
What I want to do is I want to walk you through-- just to give you the context of the problem that we're solving for the financial advisor So this is a platform designed and built for financial advisors to help them build portfolios for their individual clients. As we were going out and talking with these financial advisors, what we learned is that one of the biggest pain points that they face is that there is a gap between what they do on the financial planning side and what they do on the asset management side. And the industry recognizes that this is a problem.
So here's a quote from Eric Clarke, who's the former CEO of Orion. Orion is a very large technology platform servicing the independent IRA market. And he says, "As an industry, we need to focus on solving investor problems and move beyond focusing so much on the investment problem." We also know that this gap between the plan and the portfolio is a problem, because we've gone out, as I said, and we've engaged with advisors. We've run a number of webinars over the years. And during those webinars, we'll ask a variety of different questions.
One of the questions is, do you observe a gap between financial planning and asset management? And a whopping 89% of advisors say they observe this gap. And anybody who's dealt with financial advisors knows that it's hard to get them really to agree on anything. So 89% is really a striking number.
The second question we ask is, do you struggle aligning your clients' financial plan with their investment portfolio? And 58% say they struggle. "Struggle" is a very strong word. The most common feedback that we get across a wide range of Advisor users is that by using our platform, they're much more confident in their entire in their entire process.
OK. So the industry knows it's a problem. We know it's a problem. What's really interesting is the industry has known about this problem for decades. If you go back into a book by Charley Ellis, Investment Policy, back to 1985, chapter 4 starts out with, "A paradox is haunting investment management. The paradox is that funds with very long-term purposes are being managed to meet short-term objectives that may be neither feasible nor important, and they are not being managed to achieve long-term objectives that are both feasible and worthwhile." This is a really nice statement of the problem that we're solving. What are you trying to achieve, when are you trying to achieve it, and is it really important?
So the industry knows it's a problem. We know it's a problem. And everybody knows it's a problem for a long time. So why is it such a hard problem to solve? In part, it's a hard problem because it's more than a portfolio problem. It's also a people problem. You have complicated clients, you have chaotic markets, and you have advisors that are caught in between these complicated clients and chaotic markets.
But it's also a theory problem. The theory problem is that modern portfolio theory is not up to the task. Why? Because modern portfolio theory generally thinks of risk as volatility. What that leads to is an approach that's generally an approach dominated by a single input. So for the retirement problem that I mentioned before, your portfolio is generally only determined by when you're going to retire. For the advisory client, it's generally only determined by what your risk score is.
What this leads to is it leads to when we talk to advisors, what they whisper to us is that 90% of our clients are in our moderate model, and we know that's not the right answer. So the key to solving this problem is to have a portfolio construction, a portfolio optimization engine, that takes the key inputs from financial planning and builds the perfect portfolio based on those goals and objectives, wants and needs, et cetera, and incorporates the capital market assumptions customized to the advisory firm.
So the fundamental problem that this ultimately solves is there is a tension between two megatrends that are operating in the industry. One is, clients expect personalization. They expect personalization in the wealth advisory services, just as they expect more and more personalization in the economy broadly. But the other side of it is that advisors want to build a scalable business. Right now, there's a trade-off. If you want a scalable business, you can't have personalization of the individual portfolios. If you want personalization of the individual portfolios, you can't have a scalable business.
Now, one of the key elements of this is that one has to preserve is it's one thing to customize or personalize a portfolio for a client, but all of these advisory firms also come with their own personalities. And so the platform has to be flexible enough to accommodate both the financial planning philosophy of the advisory firm as well as the investment philosophy of the advisory firm. And so we had to build a platform that was flexible enough and adaptable enough and open architecture enough to accommodate not only the individual needs and circumstances of the individual clients, but the idiosyncrasies of the advisory firms. So we allow each advisory firm to essentially customize an investment schema unique to their own investment framework.
So what that does is Nebo Wealth is a platform allows the advisory firm to personalize individual portfolios for the advisory firms' individual clients, but do that in a very scalable manner. So here is a quote from one of our earliest users. "Before Nebo, we customized client portfolios around five model allocations. Now, with Nebo, we can actually create 100 unique personalized client model portfolios, elevating our clients' experience and doing it in a scalable way without actually taking any more resources."
OK. So what I'd like to do now is I would like to give you a demo of the platform. I'm going to share the other screen. Oops. Let's not do that.
OK. So what I want to illustrate, just in a few minutes, is the two sides. So I said there is a gap between financial planning and asset management. So what that means is if you change the financial plan, the goals, the objectives, the individual asset owner, typically the portfolio, the asset allocation, does not change, because the asset allocation is generally only determined by a risk score.
So we take a more comprehensive approach with the time horizon, risk tolerance, cash flows, legacy goals, target return, et cetera. This application here that I have on the screen is actually built out of a whole bunch of MATLAB applications-- ML apps-- but each one of those, we had to circumvent. There's a constraint, if you're building a web-based application, that the web-based application can only have a single UI figure.
So we were able to get around that because MATLAB essentially gave us enough flexibility to create a text-based version of each of those sub-apps, if you will, and we embed those inside. And that enables us to actually do all of the development using a Git-based versioning software. And so we can see the changes over time, because we have, in addition to the ML app for each of the apps that are incorporated into here, a text-based version of that, which makes it really, really easy to see in a versioned control.
But basically, the idea is you start with the clients' basic information. We do a risk tolerance, and that drives the portfolio constraint. We get the cash flows and we have multiple layers, because we're not a planning tool. We integrate with planning tools. You can do very simple, quick and dirty financial planning. You can also do an enhanced version of that, or you can import from any one of the major financial planning tools, like MoneyGuidePro, eMoney, or RightCapital.
Then, we get to target return. So imagine we have a client-- 60 years old. They just sold their business for $10 million. They're going to retire at age 63 because they're going to stay on for three years as a consultant for the business. And they want to know-- their first objective is they don't want to run out of money. So what we set down here is a terminal wealth that creates a minimum target return that they would have to achieve, on average, over their entire remaining life, so they wouldn't run out of money.
That's, in this case, a 2.1% return, net of inflation, taxes and fees. They hit this View Allocation button. Here, what we're doing is a multi-period, wealth shortfall optimization to minimize the probability and magnitude that their wealth ends up in this red area. So this is what the pie chart would look like. You can export the portfolio, which has all of the underlying building blocks.
But let's suppose, through the course of the conversation with the prospect, let's say, they said, well, they have five kids and they'd like to leave a couple million dollars for each kid. So how would things change if, instead of just not wanting to run out of money, I wanted to leave a $10 million legacy in inflation-adjusted terms? Well, now, instead of a 2.1% compounded return and instead of having a path of wealth that ended at zero, you now have a target path of wealth that is increasing in nominal terms. You would have a different portfolio.
So this is illustrating this connection between if you change your planning goals and objectives-- and you can do that at the legacy level, you can do that at the cash flow level-- the portfolio changes in a very direct way. One last thing I will just point out is that we have extensive features for the CIO, the chief investment officer of the advisory firm, to build their own models. As I said, it's really important that the advisory firm have the capability to structure the investment building blocks that reflects their investment framework, and so we have extensive model-building capabilities that allow the advisory firms to customize according to their orientation.
They can also specify capital market assumptions. We, as an asset management firm, are known for our seven-year asset class forecasts, but we don't impose those on people. This is a fully open architecture platform. So they can view our capital market assumptions, but ultimately, they're able to use their own capital market assumptions. So I think I've got in just under the wire, and I'd be happy to entertain questions at this point.