Since I started my career as a hedge fund manager, I’ve had countless investor meetings. Initially, most of the meetings were with relatively small fund of funds. As AUM grew and the investment landscape evolved, the investor mix changed to include large pension funds, sovereign wealth funds and endowments. One constant across the years and the different investor meetings was a very simple question: do your risk controls include stop losses?
My short answer to this question was – and still is – no. The long answer is more complex and nuanced. As I explained my reasoning, furious note taking ensued, I’m sure much of it negative. This guy doesn’t even embrace the simplest and most basic risk control! I’ve never been interested in simply giving the answers that people want to hear. Very few issues are black and white and the concept of unwinding a position or portion of the portfolio simple because of adverse price movements deserves a detailed analysis.
Although the concept of stop losses is very simple, their implementation is far more complex:
Is the stop loss trigger to be set on an absolute or risk relative basis?
Should the stop loss trigger be adjusted based on the security’s volatility and market volatility?
Should the stop loss be adjusted based on the latest closing price?
The answer to all three questions is yes. It’s then necessary to determine how to measure relative performance, what measures of volatility to use, and how the stop loss trigger should be adjusted based on relative outperformance. I don’t want to delve into these issues because I believe they’re moot. For most investment processes – including a robust multi-factor quant investment process – stop losses are flawed.
In broad terms, the biggest problem with stop losses – even stop losses with relatively sophisticated risk relative and volatility adjusted triggers – is they’re an overly simplistic risk control. Any robust quant factor investment process will include a diverse range of alpha drivers, including value, momentum, analyst sentiment and certainty/quality. If a broad suite of suite of quant factors are incorporated into the stock selection process, why base unwind decisions solely on momentum?
Consider a “long value” stock which is cheap and rates strongly based on certainty and quality factors. If the stock materially underperforms its risk peers then, all other things being equal, it will look even more attractive as a value opportunity. Of course, this will not necessarily be the case if something has changed which undermines the original investment thesis, such as analyst earnings downgrades.
It’s important to determine whether the share price change is due to liquidity flows or a change in fundamentals. If it’s the former, it’s not a good unwind candidate. Indeed, the adverse share price move may represent a good opportunity to increase the holding size. This is particularly the case if the share price change was on relatively low volumes.
There is one exception to this rule: share price outperformance for large short positions. It’s the exception which proves the rule! The problem with outperforming short positions is they get bigger and, especially given potential losses are unlimited, stock specific risk becomes an issue. Consider the recent GameStop and AMC short squeezes. Although the share price action was based on liquidity flows rather than fundamentals – and will undoubtedly reverse – the potential losses are too extreme to endure. Hence, several hedge funds were forced to cover their short positions and lock in large losses. Keynes famous quote – “the market can remain irrational much longer than I can remain solvent” – is probably referred to too often, but for short positions it is definitely relevant.
It is important to note that stop losses are valid for some investment processes. For momentum strategies where stock selection is based purely on technical indicators, stop losses are a valid risk control mechanism. Also, for funds with concentrated portfolios where stock specific risk is an issue, partial stop losses can be incorporated into the investment process. For a multi-factor investment process which focusses on breadth, stop losses aren’t appropriate (except for large short positions).
One of the great advantages of investors embracing a flawed investment strategy is it provides alpha generation opportunities for fund managers who take the opposite side of the trade. This was the case during the Great Quant Unwind in August of 2007. I still vividly remember buying stocks which were being decimated and selling / short selling the relative outperformers. It was the main reason why the fund I was managing generated a positive return that month. We were also able to adopt a similar strategy during the earthquake and tsunami crisis in Japan in March 2011. I wish I was managing our current hedge fund during March 2020 as similar opportunities arose, particularly in the second half of the month. Our portfolio management system (Conquest) facilitates this type of trading as it includes a range of short-term momentum and relative volume indicators. It has been a strong source of alpha in the past and we believe this will continue to be the case.
Comments