Machine Learning in Alternative Investing

EMS:

Hello Bob, thanks for joining us today.

Bob Elliott:

Thank you so much for inviting me and I'm looking forward to being here.

EMS:

First off, tell us a bit about your company and how you got to where you are today.

Become familiar with:

Well, I've been a systematic investor for most of my career, but being on the manager side of the 2/20 universe, I increasingly realized that a lot of 2/20 business is very good for the manager, but not so good for the investor. The main reason for that is that while the manager generates a lot of alpha, they take it away with fees. So I started thinking about whether there was a way to leverage technology, specifically modern machine learning approaches, to replicate how 2/20 managers are positioned in the market. Because we're using technology instead of hiring highly paid star PMs, we can do it with much lower fees than many other 2/20 strategies. This is exactly what we set out to do with Unlimited, and that's what it's all about: the idea of ​​a technology-driven, diversified, low-cost indexing of 2/20 strategies across hedge funds, venture capital, private equity, and more.

EMS:

Bob, that leads nicely into a follow-up question I want to ask you, which is, I'd love to hear your overall perspective on using machine learning to create a product that replicates index returns on alternative investments.

Become familiar with:

I think there's been a tremendous evolution in terms of the technology commercially available to start doing this kind of replication work. Twenty or 30 years ago, the original work on replication by Andrew Lowe was to look at manager returns against asset returns using rolling regressions of returns over three years or the entire history. And the problem with that is that hedge fund managers, for example, move their positions fairly frequently and relatively agilely at any given time. So the challenge that most replication strategies had was that long back-end window. Now, fast forward to today, we have much better technology available, including Bayesian machine learning style strategies, that can probabilistically determine the portfolios that these managers were most likely holding as close to today as possible in the context of that history. So we can now get a much more timely picture of manager positions from a technology perspective than was commercially feasible three or five years ago.

EMS:

Bob, what do you see as the biggest opportunity in your industry and why?

Become familiar with:

Most investors want to allocate to alternative assets. And for good reason: active management strategies often generate alpha, meaning returns above standard index investing. The challenge is always that fees are so high that the manager gets all the profit and the investor gets very little. A typical hedge fund manager takes 80% to 100% of the alpha they generate. So by using a replication approach, you can create a portfolio that looks similar to what those hedge fund managers are positioning. But with technology, you don’t have to pay highly paid people to do the research and generate your own views, so you don’t have to pay them like you would if you were running a hedge fund. This creates an opportunity to generate what I call “fee alpha,” which is one of the most durable alphas that exists. That is, if you can get many of the benefits of a hedge fund manager’s positioning at a much lower cost than it would cost a typical manager, investors can exploit that gap and get excess returns in ways they couldn’t get before.

EMS:

Meanwhile, Bob, what is the biggest challenge you face in the investment space and why?

Become familiar with:

One of the biggest challenges in taking a systematic approach to investing is that you have to carefully build your investment technology and approach based on your depth of experience. For example, in the area of ​​replication, a lot of stuff has been developed by pure technologists who don't have a lot of investment management expertise. So I think that's a challenge because technology is often built that is disconnected from the reality of how managers actually manage their portfolios, how they think about investments, how they think about exposures. So when developing a systematic approach, it's important to marry your intuitive common sense understanding of how markets and economies and managers work with the tools available to you, and have expert-built solutions to the problem. Avoid completely discretionary solutions based on the whims of managers, or completely systematic solutions that are not rooted in the reality of how money is managed.

EMS:

Bob, we've covered a lot today. Is there anything else you'd like to share with us in closing?

Become familiar with:

Well, I think the whole industry is moving towards a world of fee compression. It started 50 years ago with Vanguard introducing diversified, low-cost index investing into stocks and bonds, which, frankly, saved investors billions of dollars. So now it's time to look at other areas of the market — advanced strategies that are still too expensive relative to what investors get — and figure out how to deliver higher returns to investors with lower fees and a better risk-return profile than we have today.

EMS:

Bob, thank you so much for sharing your perspective with our listeners.

Become familiar with:

Thank you so much for inviting me, it was a lot of fun.

EMS:

Thanks for listening to the EisnerAmper podcast series. For more information on this and other topics, please visit eisneramper.com. Join us on the next EisnerAmper podcast as we get down to business.