How much do tiger hedge fund portfolios overlap? In this article we'll examine them today & historically through the Novus Alpha platform.
The Tiger Hedge Funds are without question the most recognized lineage in the industry. This isn’t the first time we’ve written about Julian Robertson’s investment tree—managers with a connection to him or Tiger Management. Three types of managers fit the bill: Tiger Cubs (those who worked at Tiger Management), Tiger Grandcubs (those who worked at a Tiger Cub), and Tiger Seeds (those who were seeded by Julian). There are ~50 managers that fall into one of these categories in our Hedge Fund Universe in any given quarter.
These funds tend to invest similarly, but how much similarity actually exists in their portfolios and across the HFU? Using the Novus Overlap engine and the recently released Q4 2016 filings we can examine aggregate trends within a lineage of managers and determine which names drive that overlap.
Q4 2016 Filings
The Novus Overlap functionality displays how much each manager’s publicly available portfolio overlaps with another’s. Here we’re analyzing the minimum overlapping position size between two managers invested in the same name. We sorted this by the managers with the highest average overlap to this “Tiger Universe” of managers. The highest average overlapping manager is Lone Pine with a 13% average overlap, but these top ten managers range from 9% to 13% average overlap across this universe.
Lone Pine has a 30% overlap with Blue Ridge, one of the highest overlaps in Q4 filings. Further scrutinizing Lone Pine and Blue Ridge reveals that the overlap stems from 11 positions. About 50% of that overlap is from their top four positions—CHTR, AMZN, FB, and NKE—each with a position size greater than 3%.
As of the Q4 filings, the most popular names in this universe were technology focused. FB, CHTR, and GOOGL were the most popular names. Joining this list is BAC, a popular name across hedge funds in Q4 during the recent financial sector rotation.
How does this information differ from previous filings—especially over 2016, a tumultuous year for hedge funds?
From December 2015 to December 2016, the highest average overlap has fluctuated from 12.8% to 14% among the whole Tiger Universe; this is a relatively tight band though not negligible. Within the top ten highest average overlapping managers, there are sixteen unique managers with seven managers always in the top ten: Coatue, Lone Pine, Blue Ridge, Viking, Foxhaven, Tiger Global, and Miura Global. Three of these appeared in the top five all five quarters of filings: Lone Pine, Blue Ridge, and Foxhaven.
Among this top-ten list, the average overlap increases, ranging from 20%–23% during 2016—if you have allocations to these managers, or are prospecting one while invested in another, this could be a concern.
Q3 2016 had peak overlap. Foxhaven and Tiger Global had the maximum overlap in the Tiger Universe at 44% in the Q3 filings. AMZN, CHTR, and PCLN made up 78% of that overlap.
Across 2016, we see a large consensus among the Tiger hedge funds for GOOGL, FB, CHTR, and AMZN. The top three names for each quarter—in terms of number of holders—consisted of these four names and had an average of 21 holders per name and an average position size of 6%. We don’t expect a lot of turnover for the Tiger Universe because of their investment styles, but the popularity of these names partly explains the average overlap. Interestingly, the largest holders of these popular positions aren’t typically in the highest average overlapping list. Of the seven unique largest holders, only three—Viking, Miura, and Light Street—are in the top ten highest average overlapping managers. Investing in a popular name (consensus) doesn’t mean that manager overlaps heavily with the rest of the Tiger Universe.
The largest average positions sizes, a proxy for conviction, aren’t necessarily also popular names. NFLX, AGN, PCLN, ADBE, JD, FLT, and APPL all are in the top three in at least one quarter, based on the average position size across the Tiger Universe (excluding positions with less than 5 managers reporting). There’s a greater variety of names and a higher average position size of 9.33% across the top three. On average, only 9.5 tiger managers on average held each of these names during the year. Even when some managers invested more heavily, it didn’t make these names popular. Of these nine Tigers that were invested in these high conviction names, only two—Tiger Global and Valiant—were in the top ten of highest average overlapping managers. Although managers may invest strongly in a name, it doesn’t mean those managers have higher average overlap across the universe.
Across both the highest conviction and consensus names, there’s still a focus on the Technology and Consumer Discretionary spaces, as expected. Of the top ten average overlapping managers to the Tiger Universe, no manager owned both a conviction and a consensus name. The Tiger Universe’s overlap exists partly from investing in similar names and with similar weights, but one factor is not the sole driver of the overlap. These factors, along with some that were not discussed like the size of their AUM and the number of their positions, all can influence overlap. Investors need to evaluate the overlap from multiple lenses to understand how the manager could diversify their portfolio. Dig further into those overlapping names to analyze how that overlap changes over time and to identify skills besides security selection like sizing decisions, trading acumen, and when the manager originally invested in a name.
These additional questions reveal that overlap is an important metric to analyze, but isn’t inherently bad. It should be avoided if you want to add diversification to your portfolio, but when continuing to evaluate a manager you must look below the surface. These questions can lead us to consider whether a manager of any lineage or universe is a leader or follower with their investment ideas. Did they initiate a trade that was followed by the herd or did they join the herd? But that is a much deeper question for a later post.