Position Sizing Strategies for Hedge Fund Managers

Position sizing is one of the most valuable skills a portfolio manager can demonstrate, and many owe their success to a handful of competencies. How then do the best hedge funds significantly outperform? Expressing conviction in your best ideas through sizing bets turns a good manager into a great manager and can pay dividends.

At Novus, we help our clients discover true investment acumen through innovative and intuitive analytics — that highlights core strengths and areas for improvement. We have found that there are five intrinsic sources of alpha, one of which is position sizing skill. As detailed in “How to Improve Position Sizing in 5 Charts”, thoughtful position sizing is critical for delivering alpha and managing risk.

In this article, we’ll analyze one manager who has demonstrated considerable position sizing skill, and through the lens of their publicly available portfolio, answer the following:

  1. How can we quantify the value added by position sizing?
  2. How can we think about how position sizing interacts with the portfolio?
  3. How can a manager minimize the impact of weaknesses, while leveraging strengths?

On our Data

Everything mentioned in this post is sourced exclusively from public data, including the manager’s profile, simulated performance and all other analysis and commentary. The data used here omits the short side, non-equity securities, many non-US securities and all non-public information such as actual fund performance. To simulate performance and determine portfolio attributes such as liquidity, we combine public holdings data with market and pricing data and make simple assumptions.

Coatue Management

The Tiger lineage has produced masters of position sizing, but perhaps none as great as Philippe Laffont’s Coatue Management. Though security selection skill is also evident, his major winners have been very popular names like Apple, Netflix, and Google. It’s his ability to size those names properly that drives performance. Since Jan 2001, Coatue has yielded a cumulative return of 476.4% compared to 94.3% and 71.7% of S&P 500 TR and MSCI World NR USD, respectively. More importantly, he’s generated excess returns of ~420% through position sizing. On an annualized basis, Coatue Management generated an extra 900bps by sizing up their highest conviction ideas.

In the above graph, the blue line is the actual return and the orange line is what the portfolio would have returned had it equally weighted the positions each month. Coatue Management consistently added value, as we see the gap widening over time.

He is also able to mitigate risk through intelligent sizing.

In the above graph, we highlighted three regimes of negative performance. When the fund was down, losses were relatively limited. This graph shows that position sizing fueled performance when the overall market direction was positive, while cushioning losses when the overall market was down.

To better unpack how this was achieved, let’s look at attributes of different position size buckets. The first thing we notice is that almost 50% of the exposure came from the names in 3.5% and greater buckets.

Since the beginning of our data, Coatue Management has allocated little capital to small names (0.0% — 2.0%). and concentrated more on names in the 2.0% to 7.5% range. There’s also a noticeable trend during the financial crisis where he increased his capital allocation significantly to big positions (>7.5%). Below we’ll see how he performed within each of these buckets.

As we see above, there is a clear relationship between conviction and performance — larger names generated the highest returns. Of the thousands of public portfolios we’ve analyzed, this is one of the most striking we’ve seen. Keep in mind, the chart above demonstrates position size versus return on invested capital, not profit/loss contribution. This means that Coatue Management’s allocation to larger names amplifies the P&L even more. Other the other hand, names smaller than 200 bps have significantly dragged on performance.

To analyze how efficiently Coatue sized up their positions, we look at Win/Loss ratio — which is defined as the average contribution of positions with a positive return over the analysis period versus the positions with negative returns.

As the manager moves up in position size, the Win/Loss ratio increases. The average winner in the >10% bucket yielded 1.84x more than the average loser in the same bucket. This means Coatue Management gets paid for high conviction bets, as evidenced by the asymmetric return profiles.

So far we’ve determined, using our public data, that Coatue Management is phenomenal at allocating capital to winning names, and sizing up those names to yield high returns. However, there are two potential areas of improvement Coatue Management can optimize to further enhance returns through position sizing.

  1. Reduce the number of positions in less than 3.5% buckets: there could be a number of reasons why those positions exist in the portfolio, but we’ve identified that these buckets tend to generate negative returns while representing ~50% of the portfolio.
  2. Increase exposure to high conviction names: evidently, during the financial crisis, he increased the portfolio positions significantly, at one point positions greater than 7.5% represented 74% of the portfolio. Since then, he’s gradually decreased that exposure even though the returns have been incredible for those high conviction ideas.

While there are a few common ways managers succeed, they fail in a dizzying number of ways. Through the Novus platform, we’ve discovered five specific mistakes managers make to diminish their returns, and only deliver a small fraction of the possible return to investors.

In our newest research, we:

  • List five mistakes managers are making.
  • Demonstrate how these mistakes significantly diminish returns.
  • Show managers how to search for and correct these mistakes on the Novus platform.

Download our latest research to learn more:

Originally published at www.novus.com on October 27, 2015.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.