The Fast Money Takes It SlowGuest Post
Just this week I heard from a CIO at an endowment who was concerned that with strategies such as ours (i.e that use Form 13F), the information was stale or old. He’s of course referring to the fact that Form 13F is filed quarterly, 45 calendar days after the end of each quarter. Therefore, a manager could have traded out of their position by the time it is disclosed. It is amazing to me how many savvy investors are still misinformed about Form 13F!
Our research as well as research from other firms has roundly debunked the myth that 13F data is stale. Maybe the perception persists because it is in the best interest of endowment/pension consultants for CIOs to remain misinformed. After all, our belief is that investors can do a much better job than most consultants picking skilled managers using little more than the Form 13F dataset.
A couple of years ago we published a research note entitled "The Fast Money Takes it Slow", which addressed this myth head on. We hope it enlightens investors on the value of the 13F dataset. If you know anyone in the endowment/pension world, please forward it to them!
The Fast Money Takes It Slow
It’s a virtual certainty. You can bet your bottom dollar on it. Whenever we discuss our methodology with prospective investors, the first question ALWAYS relates to the delayed nature of Form 13F filings. The implication is that a hedge fund manager is likely to have already exited their position by the time the manager’s quarterly filing is published 45 days after quarter end, rendering the form useless.
After nearly six years, our response is also now very predictable. The truth is hedge funds hold their positions on average for at least a year and for high conviction holdings it can be much longer than that, therefore disclosing positions quarterly can yield valuable information about a security.
In this research note we take a definitive look at hedge fund holding periods. We analyze holding periods for all Form 13F disclosed securities as well as for those held with high conviction. We also look at whether the length of a manager’s holding period is predictive in any way to the efficacy of following their holdings.
Exhibit 1 summarizes the distribution of funds in AlphaClone’s universe by holding periods. The analysis includes every holding for every fund in our universe. Holdings that appear, disappear and then appear again later are treated as separate trades. Holding periods are presented in months.
It turns out the fast money isn’t that fast after all. The average holding period across all positions is over 17 months. It makes intuitive sense then that if managers are waiting on average a year and a half to realize their investment thesis, the fact they must file quarterly means that following them based on their disclosures can make sense …. assuming the manager has skill in selecting holdings.
Let’s dig a little deeper. When it comes to active management, conviction matters. Positions that rank among the manager’s largest will drive performance, for better or worse. Therefore, high conviction positions are where the manager must have the most confidence. High conviction holdings also tend to do better when followed than the average holding overall so it’s worth taking a look at holding periods for high conviction holdings only.
Exhibit 2, summarizes the distribution of funds in AlphaClone’s universe by holding period, for high conviction positions. For the purposes of this analysis we define high conviction positions, as any position who’s size has attained a rank of 10 or lower (a rank of 1 is the highest conviction position and the largest position) in a manager’s portfolio at any time over the course of the manager’s holding period. Like in the previous analysis, holding periods are presented in months.
Surprisingly perhaps for many, the average holding period for high conviction positions is a staggering 4 years! Perhaps the most remarkable outcome from Exhibit 2 is the contrast between reality and conventional belief around how hedge funds invest. Fed by the financial media, investor perception is that hedge funds are charlatans getting rich off of high fees, playing a game that the average investor could not hope to understand let alone profit from. Hedge funds are hot money, fast money, smarter than you money when the market is up, and the money we love to hate when the market is down. The reality of course is the opposite, hedge funds don’t play by a different set of investing rules; they buy and hold, they are patient, and yes, on average they are indeed more skilled than most investors, but that’s because they are more experienced and better equipped, not be because the laws of investing physics are somehow different for them.
Too much of a good thing.
Using 13Fs to follow experienced investors can be deceptively easy. Many investors fall in love with a manager’s stellar returns, with the manager’s brand or their cult of personality. A novice 13F follower figures they can “roll their own” and be wildly successful. Again reality does not match perception. Like any investment approach, success using 13Fs takes doing your homework, it takes discipline and it takes patience.
For example, many 13F followers, including some of our competitors, believe low turnover is a key determinant for success when selecting which managers to follow. The idea is that low turnover managers hold their positions for very long periods of time and therefore following their disclosed holdings makes the most sense, especially when they also have a great historical performance record.
Exhibit 3 correlates manager holding period with the efficacy of cloning their holdings. For the purpose of this analysis we define cloning efficacy by summing the monthly excess returns (2010-2015) of a composite made up of several “follow simulations” or “clones” (e.g. follow top 5 holdings, follow top 10 holdings, etc.) over a US market factor. All simulations account for the delay inherent in 13F disclosures and include the effects of “dead” or delisted securities, thereby avoiding survivorship bias.
The scatter plot above couldn’t more clear. There is absolutely zero correlation between a manager’s holding period and the desirability of cloning their positions. Longer holding periods are necessary to make a manager’s disclosures usable but not sufficient to determine which manager to follow. Like many of the best investment disciplines, utilizing Form 13F successfully is simple but not easy.
Article by Maz Jadallah, Alpha Clone