“What if I told you that everything we thought we knew about the theory of capital markets was based on a mistake made decades ago?” says Richard “Dick” Michaud. “And I can tell you exactly what that mistake was and when it was made.” As openings to interviews go, it packs a punch. But then, little about New Frontier Advisors, the company founded by Michaud and his son, Robert, is typical. Based in Federal Street, a stone’s throw from Boston’s South Street station, the firm was formed in 1999 to put the Michauds’ investment theories into action.
New Frontier is ostensibly an investment advisor that runs money—around $4 billion at last count—and technology that utilizes the Michauds’ theories. But to describe it as such is to seriously underplay the intellectual weight behind the company, and to minimize the importance of its raison d’etre.
Indeed, the work of the Michauds cuts to the very heart of the present disruption in the very foundations of asset management—the shift from active management, to passive.
“Why is it that active management doesn’t work? Because the tools and the technology that are being used don’t work,” Michaud says. The reasons behind that not only take up the majority of this interview, but explain much about the Michauds and their company, too.
To fully understand their story, it’s necessary to go back nearly 70 years, to the University of Chicago’s library, the musings of a young man named Harry Markowitz, and the creation of the efficient frontier.
Known as the father of modern investment theory, Markowitz’s landmark paper, Portfolio Selection, was published in 1952. For the first time, it effectively quantified risk as a mathematical principle rather than an intuitive one, relative to investing, by balancing the expected return of a portfolio versus its standard deviation—or, how the performance of an investment portfolio can be effectively predicted by taking into consideration the movement of risky assets.
Much of the key behind Markowitz’s theory came from an epiphany one afternoon, in that library, where he realized that parametric quadratic programming, or the process of optimizing several variables that are subject to linear constraints, could explain—or, at least, provide a mathematical framework for—modeling investor behavior.
“He realized that he could build an efficient frontier that would explain institutional investor behavior. He also explained how people could duplicate institutional investor behavior,” Michaud says. “He was inventing quantitative management at 25, and he did all of that all at once while sitting in the library of the Chicago Business School. You know, that’s pretty amazing.”
Markowitz’s paper changed the very nature of investing. Up until that point, investment strategy had largely been driven by maximizing returns, often by chasing what was considered to be the best stock. The idea of a diversified portfolio that balanced risky assets, such as stocks and bonds, with the insurance of riskless assets, like cash, in effect modeled how people make decisions in their daily lives—for instance, people don’t expect to be in a car accident, but they buy automotive insurance nonetheless. The maxim this proposed was that the risk of loss, not the prospect of gains, is what actually drives investor behavior in the long term. The efficient frontier, therefore, optimized the portfolio by considering risk and building it in rather than ignoring it.
The real fire was lit with the introduction of the Capital Asset Pricing Model (CAPM), a mean-variance preference theory that uses Von Neumann and Morgenstern game theory in its calculations. CAPM has been, essentially, the catalyst for the creation of a multi-trillion dollar asset-management industry.
The 1952 paper and later work would eventually earn Markowitz the Nobel Prize in Economics. Yet, the theory has a problem.
[Markowitz] was inventing quantitative management at 25, and he did all of that all at once while sitting in the library of the Chicago Business School. You know, that’s pretty amazing.Richard Michaud.
While an elegant mathematical model, Markowitz is information error-sensitive in practice. It is, in effect, too precise, and the outcomes it generates are sometimes bizarre in nature thanks to its tendency to treat predictions as solid truth. On the other hand, CAPM ignores the fact that rationality and order do not always manifest in human behavior. Nor does human behavior neatly fit into predictions to 16 decimal places, a measure of accuracy that is important for scientific endeavors, of course, but relatively useless for finance.
Michaud found this for himself when he was working at the Boston Company in the 1970s. His boss at the time tasked him, as a new analyst, with coming up with a quantitative model for a European capital markets mutual fund. Michaud fed his data into the Markowitz optimization program at the Massachusetts Institute of Technology (MIT), but it wasn’t until he presented the results that he realized there was a problem with how the model portfolio was weighted.
“I remember running the Markowitz optimizer program and then very proudly bringing the results to my boss,” Michaud says. “And so, we were looking at the results and he said ‘Austria? 35 percent, is that right, Dick?’ ‘Yes,’ I said. ‘That’s right.’ ‘But we could buy everything, Dick,’ he said. ‘We could buy everything in the Austrian capital market.’”
This supposedly diversified portfolio ignored, it seemed, certain inconsistent elements—and more damningly, without key data, the model sometimes drew wild conclusions. Needless to say, the problem stuck with Michaud. Indeed, it was the genesis moment that eventually led to his famous 1989 paper, which would challenge the very orthodoxy of the market to date.
That paper, The Markowitz Optimization Enigma: Is ‘Optimized’ Optimal?, laid out much of Michaud’s initial thinking around the problems with the Markowitz mean-variance efficient frontier. But it was in the 1990s, when Michaud was working at Merrill Lynch, that Harvard University approached him with the idea to write a book on portfolio optimization. His son, Robert, pitched in with what they thought was the solution to the Markowitz problem—Monte Carlo.
The problem with Markowitz, as described earlier, is its mechanical nature and its tendency to regard predicted returns as ironclad. But that doesn’t always reflect reality—sometimes, stocks beat bonds, sometimes small caps beat large caps, sometimes small Treasurys beat long-term Treasurys—the market is not simply a program running through subroutines. The solution, the two believed, was to take the essence of Markowitz and run it through thousands of Monte Carlo simulations.
Black swans, fat tails, aberrant events—you name it—were all thrown into the simulations. A spray of wildly variant outputs followed, and, after a while, a conclusion became apparent to the two mathematicians—the answer, it seemed, lay somewhere in between.
“We could find crazy, crazy, efficient frontiers,” says Michaud. “The problem is not the forecast, it’s how the optimizer behaves, how it uses statistically estimated information. The procedure is very unstable, even small differences in the inputs can make large differences in how the optimized portfolios behave. And so, with Robert’s help, we created a patented averaging process of Markowitz frontiers resulting in a new efficient frontier that did not exist before.”
The result was truly diversified portfolios with a distributed balance between risk and return, as depicted in New Frontier’s logo, which at first looks like a rainbow but is actually a graphical representation of portfolio weighting under Michaud optimization. The landmark book, Efficient Asset Management, detailing these findings, was published first by Harvard University Press in 1998, with a second edition published in 2008 by Oxford University Press. The Michauds moved quickly to patent their work, rather than simply allow it to exist in academia or start a hedge fund in secret, a move that would eventually lead to the formation of New Frontier Advisors, and the cementing of the professional relationship between father and son.
Robert, interestingly, was not all that taken with finance in his earlier career. He was a mathematician by trade, but the publication of the book changed that. After its release, he says, he started “getting calls from Nobel Prize winners who were actually interested in my work.”
After that, he says, his outlook changed, and he began to question his choice to do pure math. “I could spend my whole life, maybe getting a result that a handful of people care about. And this is something that seemed like it was actually useful to people. So, I thought I’d switch from math to a finance program.”
Robert began a doctoral program at UCLA, but the seeds of what would become New Frontier were already there. He had been writing the early software for the company, which had been geared as a consultancy, to serve as a delivery mechanism for Michaud optimization theory. It wasn’t until he realized that the technology was the key that he really began to reconsider his position, thanks to a client who asked his father whether they might be able to license it from them during the course of a consulting engagement. As such, they pivoted from pure consultancy.
“Dick, I know what you’re doing, and I’m going to beat you.”Harry Markowitz.
“Sure, you can build a nice business as a consulting firm but to really get the ideas out there, selling the technology was clearly the way to go,” the younger Michaud recalls. “And the business was doing well. We had domestic and global clients using our technology to build much more diversified portfolios, and many told us their performance was improved and their clients were benefiting. I enjoyed getting to travel around the world a little bit, but I especially liked that we were making people better off. There was even a national social security system using our technology to build retirement portfolios.”
The first version of New Frontier’s Asset Allocation System was released in 2000, around the time that Robert left his doctoral program and joined the firm full time. The Michauds’ theories had begun to work in practice, but their true importance would soon be determined by the very person whose theories they questioned—Harry Markowitz.
“Dick, I know what you’re doing, and I’m going to beat you.”
Those are the words that Richard Michaud recalls Markowitz speaking to him at a professional conference, not long after he and Robert published Efficient Asset Management. Markowitz teamed up with Nilufer Usmen, an associate professor of finance at Montclair State University, to test Michaud’s optimizer against his own.
Markowitz realized that what the Michauds were doing wasn’t something completely different from what he had formulated—rather, it worked with different inputs. And so, he pursued the same, what Richard Michaud calls “Markowitz plus.” He ran the tests dozens of times, and eventually, in 2003, the Michauds received a phone call. Markowitz was on the other end of the line.
“Dick,” he said. “We’ve finished our work. And I think you’re going to like it.” Michaud’s optimizer had beaten Markowitz’s in every single test, a result that was memorialized in a Pensions and Investments article from December 2003, titled “Markowitz Says Michaud Has Built a Better Mousetrap.” A copy of it hangs on the wall of New Frontier Advisors’ conference room.
Despite sounding adversarial, the relationship between the Michauds and Markowitz was quite cordial. Robert remembers visiting the esteemed mathematician in San Diego “every once in a while” at the elder academic’s request during his tests, in order to ensure that he had been implementing their algorithms correctly. He recalls, with a distinct fondness, coding together with Markowitz on those days, and even offering ways to improve the Nobel winner’s programming.
The Michauds remain deeply respectful of their peer. Indeed, the foreword of the second edition of Efficient Asset Management states: “The second edition is most indebted to the research, interest and ongoing support of Harry Markowitz. Our admiration of his towering body of work only increases as our understanding deepens.”
But with the technology in place, and a remarkable affirmation of their theories from one of the greatest minds in finance under their belts, there remained a further hurdle before New Frontier would become what it is today.
“People were asking us, if we’re so smart, why aren’t we managing money?” Robert Michaud says. “Occasionally we’d have people asking if we would manage money for them. These were mostly institutional mandates, but not large enough to change our business model at that time.”
Such “smaller” mandates would, of course, be substantial by any measure, but that would necessitate a radical change in how New Frontier operated. Both Robert and Richard had worked in finance, both at Acadian Asset Management, and the elder Michaud at a range of firms over his career, and they knew that investment decision-making realistically comprised a small group of people.
Although ETF portfolios are extremely popular now, we were building global multi-asset portfolios entirely out of ETFs from the very beginning.Robert Michaud
Yet there had to be hundreds supporting that small group, in terms of marketing, distribution, operations, trading, compliance and everything else. It didn’t fit with what they were trying to achieve with New Frontier. The emergence of exchange-traded funds (ETFs) changed that.
By this time, Robert was the chief technology officer of New Frontier, with Richard as the CEO. They had been watching the development of ETFs closely for some time, and the vehicle seemed eerily well-suited to what they were attempting to achieve.
“Since we were so focused on total portfolio optimization, I said, ‘Oh, ETFs are exactly what you want to optimize: low-cost, transparent, liquid building blocks of a portfolio,’” says Robert Michaud. “Although ETF portfolios are extremely popular now, we were building global multi-asset portfolios entirely out of ETFs from the very beginning.”
In many ways, it does tie in perfectly with where New Frontier sits in the new world of asset management, and the new existential struggle in the shift from active to passive management.
“Passive has to do with low cost,” Richard Michaud says. “But in spite of the fact that the index fund managers are right to think in terms of low-cost investing, they have not repealed the law of finance that says you get what you pay for. For active management to work, you need technology that works and that is what has been the problem up until now. What we are trying to do is apply state-of-the-art investment technology to very low-cost, low-risk funds. So, it’s the best of both the active and passive worlds.”
What the Michauds are trying to do, in effect, is actively managed passive investing, with a number of different strategies to achieve that, based on Michaud optimization. Assets under management passed $500 million in 2006, shortly before the second edition of the book was published. They passed $1 billion in 2008, just before the fourth patent was granted to the Michauds for their work. Currently, the firm manages multi-asset portfolios and ETFs with about $4 billion under management.
By their own admission, returns from these strategies are not spectacular. But they are consistently above the benchmark, Robert Michaud says. And their mission is succeeding—the Chartered Financial Analyst Institute, for instance, teaches Michaud optimization. Their technology and theories are used—both legally, and perhaps more dubiously so in some instances, which Richard Michaud declines to cite specifically—across finance and related professions.
And as New Frontier continues to grow, the core mission remains the same: a bedrock of science, tempered with a certain social, psychological and mathematical practicality. While the market remains dominated by Markowitz and CAPM for now, the so-called wrong turn mentioned at the beginning of this article, perhaps in the future and if this current trend toward passive continues to reshape the very nature of asset management, the name Michaud may become even more prominent.
Correction: This article has been updated to further elucidate the differences between Markowitz portfolio theory and CAPM, which is based primarily on game theory, and to correct the year of publication for Efficient Asset Management to 1998, not 1999.
Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.
To access these options, along with all other subscription benefits, please contact [email protected] or view our subscription options here: http://subscriptions.waterstechnology.com/subscribe
You are currently unable to print this content. Please contact [email protected] to find out more.
You are currently unable to copy this content. Please contact [email protected] to find out more.
Copyright Infopro Digital Limited. All rights reserved.
You may share this content using our article tools. Printing this content is for the sole use of the Authorised User (named subscriber), as outlined in our terms and conditions - https://www.infopro-insight.com/terms-conditions/insight-subscriptions/
If you would like to purchase additional rights please email [email protected]
Copyright Infopro Digital Limited. All rights reserved.
You may share this content using our article tools. Copying this content is for the sole use of the Authorised User (named subscriber), as outlined in our terms and conditions - https://www.infopro-insight.com/terms-conditions/insight-subscriptions/
If you would like to purchase additional rights please email [email protected]
IBM’s general manager of global banking and financial markets joins the podcast to talk about the importance of modernizing mainframes.Subscribe to Weekly Wrap emails
- Top venues mull offering joint EU consolidated tape for bonds
- The future of algo trading: Using deep learning to more accurately predict equity market volumes
- In pursuit of API-ness: Rise of API interfaces hints at data platforms of the future
- Waters Wavelength Podcast: IBM’s Likhit Wagle on modernizing mainframes
- Waters Wrap: Nasdaq, Quandl and the next phase of the alt data craze