Was Renaissance’s Success Luck or Skill – And Was It Behind Trump’s Victory?
The hedge fund firm Renaissance Technologies, founded by the distinguished mathematician James Simons, has been an object of amazement, admiration, and envy for years, because of the incredibly high investment returns of its flagship Medallion fund. In a new book, The Man Who Solved the Market: How Jim Simons Launched the Quant Revolution, author Gregory Zuckerman comes as close as anyone outside Renaissance ever has to explaining how Renaissance did it. In passing, he also sheds light on the Renaissance co-CEO who, as much as any other single private person, was responsible for the earth-shaking political events that occurred recently in the Western English-speaking world.
Renaissance’s spectacular investment performance
Kurt Vonnegut’s 1965 novel, God Bless You, Mr. Rosewater, begins “A sum of money is a leading character in this tale … The sum was $87,472,033.61 on June 1, 1964, to pick a day.”
This seems a little quaint in the present time, when almost every prominent capitalist or celebrity, and her brother, can be assumed to be a billionaire or at least a multiple-hundreds millionaire. But in God Bless You, Mr. Rosewater it was meant to represent an incalculable and inexhaustible sum of money – enough to be a force of nature in itself, and the lead character in a novel.
A novel – or a non-fiction book – about Jim Simons and Renaissance Technologies could similarly begin by naming the leading character, in this case not a dollar figure but a percent, namely 63.3%, or 37.7% after fees.
That is the average annual percent by which the Medallion fund increased from 1988 through 2018. To put it another way, the fund’s first dollars grew over the 31 years by 400 million percent before fees, and by two million percent after fees. An investor who had invested $1,000 at the beginning of 1988 would have accumulated more than 20 million dollars at the end of last year; if it were to continue another 30 years that $1,000 would become $412 billion. Those figures also tell you how much the fund’s managers must have made; it would put Midas to shame.
I had an extended coffee with a friend recently and mentioned this book and the Medallion results. It turned out, surprisingly, that he wasn’t especially aware of Renaissance, Medallion, or Jim Simons. When I told him about Medallion’s investment results he didn’t believe them.
I wouldn’t have believed them either if I hadn’t been hearing about this for a long time. But rumors had spread about Madoff’s very good but much less spectacular results for a long time too. In Zuckerman’s Appendix, where he lists Medallion’s annual returns, he cites the source as “Medallion annual reports; investors” – not exactly 100 percent conclusive evidence that they are accurate. Nevertheless, I believe them. And yet, I consider it possible that I’ll discover at some time in the future that I was wrong to believe them. But henceforth, I will take them as a given.
Luck or skill?
Any report of market-beating investment performance, however jaw-dropping, raises the question: Was it luck or skill?
From 1991 through 2005, Bill Miller’s Legg Mason Capital Management Value Trust fund beat the market in every single year. Miller was celebrated for his performance – 15 years of beating the market. The chances of that happening by chance were less than one in 32,000. Surely this was skill, not luck.
But subsequently, his fund crashed and burned, essentially wiping out his past gains – and more than wiping them out for the bulk of his investors who invested with him after they saw how good his performance was.
Even before his downfall, investors should have seen the statistical possibility that it was luck and not skill. There were more than 32,000 investment funds. At least one of them could easily have experienced that kind of performance – and that is the one you would read about.
Countless other high performers over time – too many to name – have exhibited similar patterns: years or even decades of outstanding results followed suddenly by failure (often not long after the bulk of the investment money has flowed in). In each case, it could be attributed either to the fact that the high performance was not outside the range that could be attributed statistically to luck, or because the fund employed a “picking up nickels in front of steamrollers” strategy – making a lot of money until one day the steamroller wins.
Could the Medallion fund fit the pattern of the crashers and burners? Could it have been “only luck”?
What are the possibilities that it was “only luck”?
Could its high investment returns be due to leverage? Could they be considered within the range of outliers that would have to be produced at the upper end of a strongly skewed lognormal distribution of returns – the skew exaggerated further by “fat tails?”
I’ve given this some thought, and I think the answer is clearly “no.”
The first question to look at is whether the high returns could have been due to leverage.
Suppose that on January 1, 1988 you were 34 years old and had $1,000 to invest for your retirement and a house that was given to you by a rich uncle. You decide that the stock market – specifically the S&P 500 index fund – is going to do well for the next 31 years, until you retire at age 65. To leverage your measly $1,000 you take out a mortgage on your house with a balloon payment in 31 years and invest the whole proceeds plus your $1,000 in an S&P 500 index fund.
You are right, because your S&P fund returns an average of 10% (after fees) over the 31 years. You do extremely well because you were highly leveraged.
But to have done as well as you would have done if you had invested the $1,000 in the Medallion fund instead, the amount of your mortgage would have had to be one and a quarter million dollars.
Now that’s leverage: 1,250 to one; surely much more leverage than the Medallion fund ever exercised. So its performance wasn’t due to leverage.
The Medallion fund’s 31-year return of 63.3% before fees was a full 13 standard deviations above the return on the market as a whole. That’s a whale of a lot of standard deviations, fat tails or no fat tails. And in spite of the much larger annual numbers, its risk, as measured by the standard deviation of Medallion’s returns, has not been that much greater than the S&P’s.
It should be noted, however, that Medallion mostly has not invested in equities. Its domain, according to Zuckerman’s book, has mainly been currencies, commodities, and bonds. Its sister fund, the much larger Renaissance Institutional Equities Fund (RIEF), which started later, has not performed anywhere near as well. And the Medallion fund has been closed to outside investors for years. Only Renaissance employees (and in at least one case an ex-employee, see below) are allowed to invest in it. To keep it from getting too big (it’s about $10 billion now) funds were regularly returned to investors.
Let’s posit that Medallion’s success has been due to skill – how did they do it?
It turns out that there’s nothing very mysterious about it. Medallion’s managers were early practitioners of machine learning, though their techniques may not all have been those that are taught now in machine learning courses.
Renaissance’s philosophy sounds a little weird. It was that you let the machines do it, and when they find patterns that are clearly persistent over time you don’t even ask why these patterns exist or why they persist.
The principle is that if you could figure out what’s causing the patterns, others could too. Therefore, there are likely to be others trying to exploit them. Better to exploit the ones that can’t be explained.
Like artificial intelligence in recent years, their success is due not only, or not very much, to the mathematical techniques, but to the phenomenal amounts of data available and processing power to sift through them.
Zuckerman’s book does not and cannot, of course, provide details on any one strategy that Medallion employed. But we get the idea. You have reams of data of all kinds. You sift through them and apply both advanced and simple mathematical techniques to detect patterns and relationships that are not the least bit obvious to the casual – or even not so casual – observer. And sometimes a repeating pattern turns up, you have no idea why but you track it to see if it persists. It’s that simple.
One surprise in the book is that Simons himself doesn’t seem to have done much if any of the actual analytic work. He left that to the people he hired or partnered with, offering only an occasional suggestion.
But his hiring policy was the key. He would not hire or even consider anyone with a finance background. He hired only mathematicians or mathematical physicists, or similarly scientifically-minded experts.
Why was this important? I can explain from my experience. Mathematicians painfully and rigorously justify every step in their reasoning. People with finance backgrounds, by contrast, generally do not. You can tell this by reading finance journals, even the “top” ones. In them, authors of articles will write out some mathematics, then draw conclusions from it that do not follow from the mathematics. They do not painfully and rigorously justify every step in their reasoning. Rather, they take great leaps to conclusions they believe to be true regardless of the mathematics. I have written many articles about this phenomenon.
This is a plague in finance that nobody in the field recognizes. They seem not to know what painfully and rigorously justified step-by-step reasoning is, or why it is necessary. At one point, Simons assigned his employees to read hundreds of peer-reviewed finance articles to see if they could glean something useful. They found nothing of value. If I started a hedge fund, I would never hire anyone with a finance background, especially with a PhD in finance.
Simons’s hiring decisions have paid off handsomely, at least at the level of making the Medallion fund – and Renaissance itself – very successful.
Have they paid off at a larger level? That is another question.
Simons’s most controversial hire
Zuckerman reports in his book that in February 2017 a Renaissance employee named David Magerman sent an email to a Wall Street Journal reporter. In a footnote, Zuckerman disclosed that the recipient of the email was Zuckerman himself.
In a phone interview, I asked Zuckerman if that was how he began writing the book, or if it was something he was thinking about already.
He told me that he had already gathered a little information and was beginning to think about writing a book about Simons and Renaissance before he got the email.
“So,” I said, “Magerman’s email must have been a stroke of luck.”
Well yes and no, he said. It did, of course, provide a new window into Renaissance that could be valuable in writing a book. But it also shed unwanted light into a conflict within Renaissance that Simons would have preferred not to be out in the open. This may have made Simons and others at Renaissance even more reluctant than they already were to talk to or cooperate with the author.
As with virtually all hedge funds, Renaissance is a very private company. It does not seek publicity. Simons is particularly circumspect. As far as its business is concerned, it is understandable that Renaissance would want as little as possible to leak out. I once talked to a former employee of JPMorgan Chase, which was Renaissance’s broker. He told me that Renaissance insisted that the server that Renaissance’s data was on must be totally isolated from all other JPMorgan Chase systems. If any of Renaissance’s strategies leaked out, the recipient of the information could, of course, take advantage of it to compete with Renaissance’s trades.
But in this case it was not a leak of code or data that was the concern. It was the leak of an interpersonal conflict within Renaissance.
What do you do about an employee who is a political problem?
Simons was one of the biggest backers of Hillary Clinton’s election campaign, donating $27 million. But the man he hired as co-CEO of Renaissance Technologies, Bob Mercer, was not only one of the biggest backers of Donald Trump but one of the biggest backers of the Brexit campaign – and a major owner of two of the main facilitators of both: Cambridge Analytica, the company that used hacked Facebook data to target ads toward potential Brexit and Trump supporters; and Breitbart News, an alt-right Trump-backing news organization.
Zuckerman describes how almost everybody at Renaissance and at Simons’s philanthropic organization, the Simons Foundation – which backed cosmological research, autism, and other endeavors – was disappointed, if not appalled, at the Trump victory.
But Mercer was exultant. He had poured money into the Trump campaign, and had recommended to Trump that he engage Steve Bannon and Kellyanne Conway to help him win. Through Cambridge Analytica, he had helped to get the data analysis done that may have made an important contribution.
Bannon attributed much of the success to Mercer. “The Mercers laid the groundwork for the Trump revolution,” he said. “Irrefutably, when you look at the donors during the past four years, they have had the single biggest impact of anybody.”
Mercer was an extremely taciturn person, very happy saying very few words to anyone. He did support some unusual causes, scientific beliefs, and political philosophies, especially for the liberal New Yorkers among the Renaissance employees, but many of them were not that extreme and had reasonable arguments on their side. But he did believe, as have many conservatives, that government was becoming too big in the United States and there was a danger of a drift toward socialism.
Even before the Trump victory, employees were asking Simons, “Can’t you do something about him?” But Simons said, “He’s a nice guy. He’s allowed to use his money as he wishes. What can I do?”
But David Magerman, who worked for Mercer and whose career at Renaissance had been strongly supported by Mercer in Magerman’s early days, was not content. He felt, as Zuckerman describes, that, “he had personally helped provide Mercer with the resources to put Trump in office and encourage policies that Magerman found abhorrent.”
“It pisses me off,” he told his wife, Debra. “I’ve made software that makes white rich guys like Mercer even richer.”
After his email, Magerman met with Zuckerman at a restaurant in Pennsylvania. The result was a Wall Street Journal article. The morning it appeared, Magerman was suspended from Renaissance and barred from having any contact with it. A few months later, he was fired.
Mercer experienced blowback too. He found out what it’s like to be a public person accused of having aggressively pursued an unpopular agenda. Eventually, he pulled back on his sponsorship of far right-wing speakers and his giving to the Republican Party.
But he had become too much of a burden for Renaissance as its co-CEO. After too much negative feedback from clients, Simons asked Mercer to step down as co-CEO. But otherwise, he continued to work at Renaissance just as before.
Magerman sued Renaissance. They eventually came to a settlement, allowing Magerman to invest in the Medallion fund even though he was no longer an employee.
When I read Zuckerman’s book, I found Mercer surprisingly sympathetic. He was obviously something of a geek who didn’t like communicating with other people very much, and so he kept silent most of the time. This can isolate you and drive you toward unpopular, fringe beliefs – some of which may sometimes be true.
But I had the feeling that in his quiet removal from most interactions with others, he might, Billy Budd-like, not have realized his own power; the power of his money.
Yes, it is possible to have too much money.
Economist and mathematician Michael Edesess is adjunct associate professor and visiting faculty at the Hong Kong University of Science and Technology, chief investment strategist of Compendium Finance, adviser to mobile financial planning software company Plynty, and a research associate of the Edhec-Risk Institute. In 2007, he authored a book about the investment services industry titled The Big Investment Lie, published by Berrett-Koehler. His new book, The Three Simple Rules of Investing, co-authored with Kwok L. Tsui, Carol Fabbri and George Peacock, was published by Berrett-Koehler in June 2014.