Alpha Theory Featured in Net Interest Newsletter

Alpha Theory was recently highlighted in Net Interest, a newsletter offering insight and analysis from the finance world.

‍

In the article, author Marc Rubinstein discusses the significance of position sizing in investment strategies and how Alpha Theory can be beneficial.

‍

Here is the full article, "Alpha Capture: The Art of Portfolio Construction":

‍

“We can see you’re a good analyst, but are you a money maker?” — Interview question.

‍

Bets are integral to finance so it’s no surprise that business strategies in the industry should be launched by them. In 2001, Deutsche Bank’s former global head of equity trading and Mercury Asset Management’s former head of European equity investing disagreed on the worth of brokers’ stock recommendations. The pair, Ian Wace and Paul Marshall, had co-founded Marshall Wace, a London-based hedge fund, four years previously with seed capital from George Soros. Their focus was European equities and while they utilized the services of sell-side research for industry expertise, market color and so on, their views differed on how much value to place on analyst stock calls.

‍

To settle their disagreement, they tasked a summer intern to design a system to track the performance of recommendations. Initially, the goal was to reward institutions based on the profitability of their ideas, but the pair and their developer – who came on board as a full-time employee as soon as he graduated college – quickly realized that the system’s trade ideas could be optimized to try and outperform the market.

‍

In July 2002, Marshall Wace began deploying capital from its flagship Eureka Fund in a portfolio of stocks determined by the system, now known as TOPS – “trade optimized portfolio system”. Its TOPS Opportunistic portfolio returned 23.9% gross of fees in its first full year, versus a market benchmark down 21.1%. Fifteen months later, the firm initiated a parallel TOPS Fundamental portfolio to invest in stocks over longer holding periods.

By 2005, the system was drawing in 800 to 900 ideas a day in London. Rather than simply passively monitoring analyst recommendations, it asked brokers to submit them and weight them according to conviction levels, creating a more dynamic market in ideas. That year, TOPS elicited around half a million trading ideas from 2,200 individuals across 246 securities firms. Brokers were incentivised through commission payments, generously allocated according to performance. In 2004 and 2005, Marshall Wace is estimated to have paid brokerage firms over $250 million per year.

‍

For research departments, the development came at an opportune time. In April 2003, as part of a Global Analyst Research Settlement, they had been forced to sever economic ties with investment banking departments so as to avoid conflicts of interest. Stripped of their share of revenues from corporate advisory mandates, directors of research were keen to foster alternative income sources. Marshall Wace helped fill the void.

‍

By itself, the idea was not new – although TOPS took it to new heights. Alfred Winslow Jones, credited as the inventor of the original “hedged fund” and described by New York magazine as the “big daddy” of the industry, pioneered the strategy in the early 1950s. According to Sebastian Mallaby, who writes about Jones in his book, More Money than God: Hedge Funds and the Making of the New Elite:

‍

“[H]e invited brokers to run ‘model portfolios’ for his fund: Each man would select his favorite shorts and longs, and phone in changes as though he were running real money. Jones used these paper portfolios as a source of stock-picking ideas. His statistical methods, which separated the fruits of stock selection from the effect of market moves, allowed him to pinpoint each manager’s results precisely. Jones then compensated the brokers according to how well their suggestions worked. It was a marvelous technique for getting brokers to phone in hot ideas before they gave them to others.”
‍

By the early 2000s, selective dissemination of research had been outlawed, so brokers could no longer front-run ideas the way they had for Jones. Instead, they leveraged their expertise across sectors and geographies, which Marshall Wace mobilized into an investment engine. Although initially some firms were nervous the system could encourage employees to skirt regulatory guidelines, rulemakers eventually gave it their seal of approval, arguing that the electronic audit trails it introduced reduced the risk of misconduct compared with traditional communication channels.

‍

Very quickly, Marshall Wace turned the system into a flywheel. “All TOPS is a Hoover,” Wace told the Wall Street Journal. “It sucks in great quantities of ideas and tries to sieve out what is interesting.” More ideas over a longer period of time led to a better calibrated sieve. A better sieve boosted the performance of TOPS funds. Stronger performance over a larger asset base generated more fees to pay for ideas. And so it spun. The system even incentivised participation at the individual as well as the institutional level: TOPS performance became a recognised industry benchmark that brokers could use to burnish their standing in the job market.

By October 2006, Marshall Wace was managing €3.9 billion in TOPS out of its total €5.9 billion assets under management. That month, it raised a further €1.5 billion by listing a closed-end fund, MW TOPS Limited, on the Amsterdam Stock Exchange.

‍

Today, the firm runs around $30 billion across various TOPS funds. Like all profitable ideas in finance, it has also seen its innovation copied. Third-party providers of “alpha capture” software allow funds to plug into a network of idea generators without even having to recreate the Marshall Wace intern’s build-out. (The intern, by the way, was made a partner of Marshall Wace in 2004 and is now the firm’s third-largest owner. Institutional Investor magazine estimates that he personally made $330 million last year.)

‍

Funds are also increasingly scraping other sources for ideas:

• Center books embedded inside multi-manager funds pull the best ideas from across a firm’s multiple investing desks into a central portfolio. As discussed in Peak Pod, the cost of sourcing these ideas is higher than from brokers – investment teams can be paid 15-20% of their performance – but the ideas are already captive to the firm, so the incremental margin is higher.
• Newer firms, like Atom Investors and CenterBook Partners, capture ideas from other hedge fund firms in return for a share of performance fees. These funds market themselves as a solution for smaller firms to enjoy some of the advantages offered by multi-manager firms while retaining their independence. CenterBook currently curates ideas from 28 partner firms.
• SumZero, an online platform where professional investors share investment ideas, is launching a capital vehicle to invest in stock calls posted on its site. “Our fund is driven by a quant model that leverages Natural Language Processing and Machine Learning technologies,” announced founder Divya Narendra (who worked with the Winklevoss twins in Harvard on the precursor to Facebook; he is played by Max Minghella in The Social Network). “Our model reads each new idea at the end of every day and then makes its investment decision.”

‍

What each of these funds recognize is that idea selection is a different skill from portfolio construction (which is a different skill from running a firm). To see how firms reconcile the skills, read on.

When I shifted seats from the sell-side to the buy-side, one of the hardest skills to grasp was position sizing. It was no longer sufficient to like a stock, you had to have a sense of how much you liked it. And that was contingent not just on the stock’s fundamentals, but on its liquidity, its volatility and its correlation with other stocks in the portfolio. What had been a one-dimensional decision was now a multi-faceted decision.

‍

Yet it wasn’t a skill that was taught. As investment strategist Michael Mauboussin told capital allocator Ted Seides on a recent podcast, “I think most fundamental investors have some sense of why their position sizes are what they are, but in a sense, they’re not quantitative or they’re not completely structured.”

‍

And it’s not as if sizing isn’t a critical part of the investment process. While much of the glamor in the industry is in stock-picking – the ideas dinners, the charity pitches and so on – the real money is in sizing. One of the best-known hedge fund trades of all time – George Soros’s bet against Sterling in 1992 – owed its success as much to its sizing as its insight. “Druckenmiller had done the analysis, understood the politics, and seen the trigger for the trade; but Soros was the one who sensed that this was the moment to go nuclear,” writes Sebastian Mallaby. “When you knew you were right, there was no such thing as betting too much. You piled on as hard as possible.”

For some investment strategies, notably long/short equity, where portfolios are typically narrower than in macro and idea duration is longer, sizing is even more important. Yet many portfolio managers struggle with it.

‍

Consulting firm Alpha Theory, which advises many of them, estimates that its clients’ funds would have returned 17.8% last year had they optimized position sizing, rather than the 13.1% they actually delivered. Looking back over its historical data, the firm reckons optimally sized portfolios would have outperformed actual portfolios in 11 of the past 12 years, generating average outperformance of 4.3% per year.

‍

I don’t know what Alpha Theory would have made of my fund’s returns but, as with many funds, hit rate wasn’t the key determinant. Outsiders may assume that successful portfolio managers get most of their calls right but in the wild, hit rates just marginally above 50% are not unusual. “An alpha success ratio of 52-53% is already very good if it is consistent through time,” notes Paul Marshall in his book, 10½ Lessons From Experience: Perspectives on Fund Management. “A truly great manager will have a success ratio of 55%.” In fact, 55% would be quite exceptional; I’d be surprised if Marshall Wace sustained a hit rate at that level over time. Alpha Theory calculates the average success ratio over 12 years across the 100+ funds it consults for at 48%.

‍

The key, rather, is position sizing – certainly that was Stanley Druckenmiller’s takeaway from working with Soros. “What I learned was, sizing is probably 70 or 80% of the equation,” he told attendees at the 2022 Sohn Investment Conference (ironically a charity pitch event). “It’s not whether you’re right or wrong, it’s how much you make when you’re right, and how much you lose when you’re wrong.”

‍

Sadly it wasn’t available when I was managing money, but an excellent resource for understanding position sizing is Giuseppe “Gappy” Paleologo’s book, Advanced Portfolio Management: A Quant’s Guide for Fundamental Investors. Gappy has worked at multi-manager firms Citadel and Millennium and is currently on gardening leave ahead of taking on a new role at Balyasny. Over his career he has “consulted, collaborated, taught, and drank strong wine with some of the best stock-pickers in the world.”

‍

Gappy recognises the different roles at play within the investment process. He argues that estimating alpha as accurately as possible is the job of the fundamental investor; estimating beta and identifying the correct market benchmark to use is the job of the quantitative risk manager; and estimating the expected value of the market is the job of the macroeconomic investor. Sometimes the roles bleed into each other, but as in any industrial process, specialization offers competitive advantage and investment management is an increasingly industrial process.

‍

On sizing, he offers four rules. The simplest just scales positions according to expected upside or downside. This requires analysts to maintain price targets, which surprisingly many do not. According to Alpha Theory, returns on long positions on which clients carry a price target outperform those with no target by almost 5%; on shorts, the differential is 7%. Price targets aren’t an exact science, but they can help inject some order into a portfolio. Other rules factor in volatility and variance but, as a rule-of-thumb, the simplest works best.

‍

Things get more complicated when a manager wants to eliminate factor risk. Factors are the other general characteristics, alongside market return, that explain a stock’s performance. Gappy analogizes them to components of a wave. Just as a wave breaking on the shore is made up of a large wave, and then of a few smaller waves riding on it, with ripples on top, stock returns are the effect of a large shock (the market) plus a few smaller ones (sectors, styles), then a few even smaller ones. In the case of our fund – a specialist financials fund – the sector factor was clearly a major determinant of returns; the ‘value’ style contributed as well given how much of the sector exhibits a value bias. Sizing positions without simply layering up those bets requires more work. Gappy proposes a model (reproduced below).  

‍

For the typical fundamental analyst, the model may be a step too far into the dark world of quantitative finance. Which is where alpha capture comes in. By specializing in portfolio construction while outsourcing idea generation, it is able to extract some of the value left behind by stock-pickers. The latest breed of external buy-side alpha capture strategies create diversified portfolios with better calibrated portfolios than managers could necessarily muster by themselves. As better quality data emerges with longer historical time series, it’s another development that may yet signal Peak Pod.