Details of profitable investment strategies can be found in regulatory filings, and are easily replicated. But there are tricks and pitfalls to this strategy. A recent paper analyzes the art of alpha reconstitution.
How do the best systems currently capture alpha (excess returns), and what can be improved? That is the topic of a recent paper published by François-Serge Lhabitant, professor of finance at Edhec Business School, and Denis Mirlesse, director at Steadfast Advisory Services. Both have spearheaded the hedge fund management of the Kedge Capital group.
Alpha can be identified by information gathering systems, but the process has to go beyond the sheer replication of positions. Systems aimed at capturing alpha have emerged 20 years ago and have considerably evolved. How do they work? Basically, those «alpha capture systems» that are used by institutional investors, collect, track and rank the performance of stock recommendations and trading ideas. The point is to look at the performance of highly recommended stocks provided by sell-side and buy side-sources, but also by more secretive funds and their regulatory filings.
After looking at which recommendations were the most successful, these systems select the best idea contributors and gather them into an optimized alpha-generating portfolio, tracking their profitable recommendations. A «collective brain of successful stock pickers», as Lhabitant and Mirlesse put it.
Alpha is a known rarity. Many are called, few are chosen. U.S. research shows that from 1926 to 2016, most common stocks did not outperform Treasury bills, and 4% of stocks accounted for all of the net wealth creation in the US stock market. Between 2009 and 2019, 89% of large-cap mutual funds underperformed the S&P 500, and small caps didn’t do better.
So, finding alpha can be a difficult process. One can create platforms that rely on the voluntary contributions of analysts, brokers, and other market participants, and filter the best. But these systems can be expensive and complex to set up. Fortunately, there is an alternative, less costly way: looking into regulatory documents, as the authors of the report write.
«An alternative free source of investment ideas can be found in portfolio holding disclosure. This is where the underlying securities of an investor, together with their value and portfolio weight, are mandatorily made available to the public, usually with some lag time.» In the U.S., an investor can thus access the content of any mutual fund portfolio on a quarterly basis, with a 60-day lag at worst.
«Disclosed holdings can potentially be used to free-ride on someone else’s research by running copycat portfolios», the authors explain: «In its simplest form, an investor could easily collect the list of holdings of successful fund managers from regulatory filings and replicate them in his portfolio, without paying any fee to the managers.»
Is this imitator strategy well rewarded? Yes, according to academic research: a hypothetical portfolio invested in Berkshire Hathaway’s holdings a month after they were publicly disclosed would have outperformed the S&P 500 by an annual average of 10.75% from 1976 to 2006. According to the report, copycat funds have earned returns that are statistically indistinguishable, and possibly higher after fees, compared to original funds, simply by replicating quarterly 13F filings.
But there are several pitfalls. First, discipline is needed. Regulatory information should be updated and checked as much as possible. Pair trades should be excluded. You also have to keep in mind the alpha «shelf life» to avoid entering a position when the leading source has already exited. Running after second-hand positions with expired upside potential would be a fail. An alpha capture system should prefer longer-term bets, in order to move in while the market price response to new information is slow, and when there is still alpha to be extracted once the idea is disclosed. «Talented leaders with slow alpha decay are obviously the best idea generators for alpha capture systems», suggest the authors.
Another problem is that past performance of the copied managers doesn’t guarantee that they will remain winners in the future. «Great that you can replicate Warren Buffett», says Bram Cornelisse, hedge fund manager and founding partner at Farringdon Capital Management in Amsterdam. «But he has not beaten the benchmark in the last ten years. So how would you have predicted that?»
Clearly, «mechanically riding the coattails of past winning investors and cloning their top positions with a lag does not guarantee superior returns», warn the authors of the report.
Another issue is that mutual funds don’t have much margin to outperform their benchmark since they can’t deviate from it. Therefore, the authors argue, hedge funds are a better universe to replicate, since they have more freedom and potential to add alpha.
While alpha-generating managers in long-only portfolios shouldn’t be neglected, research shows that hedge funds have a forecasting ability that is four times better than other institutional investors. Their best ideas are proven to outperform the S&P 500. «Extracting alpha from hedge fund holdings could therefore be an interesting path to explore», argue Lhabitant and Mirlesse.
They cite the case of the Goldman Sachs Hedge Industry VIP ETF, which was created in 2016 to track the firm’s hedge fund index. It would have theoretically outperformed the S&P 500 by 80% between 2001 and 2016.
Alpha capture systems shouldn’t limit their universe to hedge funds either, because the best talents are often organized into family offices, including George Soros, Leon Cooperman, Carl Icahn and many others. The best ideas should be sought everywhere, in buy-side and sell-side research, and in many different structures and sources.
Another important point is that a system wanting to capture alpha should focus only on the highest conviction ideas. The problem is that often portfolio managers end up diluting their portfolio by adding lower conviction positions. This reduces the overall risk, but also the chance of outperformance derived from specific bets, which is what defines alpha. «To be successful, systems should therefore focus on managers’ highest convictions.» Avoid crowded trades and don’t imitate the copyats, but only the originals. Strong convictions need to be correctly sized, or weighted, in the portfolio. If a manager is good, his biggest positions are supposed to generate the best performance.
One of the co-authors, Denis Mirlesse, has implemented the above advice by creating a portfolio of managers with a long-term view and a low portfolio turnover. It uses algos, or machine learning techniques, to identify behavioral characteristics, investment strengths, weaknesses, and biases of all contributors. The portfolio has outperformed the above-mentioned Goldman Sachs ETF between April 2008 and December 2020, as well as the S&P 500 (see chart). «These results clearly show that the recommended method is additive to performance and delivers much better results than just selecting managers and positions based on size», write the authors.
«One should distinguish between ‹original› alpha generation, which is what hedge funds are after, and these specific copycat strategies, which I suspect only have a fraction of the hedge fund assets», comments Bram Cornelisse. «The success of the copycat strategies is probably also partially dependent on not too much money chasing the same copied ideas».