⛔ The Biggest Hedge Fund Lie
Models, flows, alt data, factors, "peak pod"
The biggest misconception in the hedge fund industry is that you need to be doing certain things to be successful and make money: having super detailed and eloquent models, closely tracking flows, having a best in-class alternative data process, monitoring and quantitatively decomposing your factor exposures, using AI to automate your processes, attending every single conference and management meeting…
The list goes on and on…
Let me take you inside some of the best multi-manager hedge funds in the world.
First, at the large multi-managers, there are hundreds of Portfolio Managers. If you start talking to a number of the very successful ones, you will realize there is no single set of things that all of them do. Yes, there are PMs who are extremely successful who do some or all of the things mentioned above. But there is also at least an equal number who do very little or none of those things.
Also, think for a moment: imagine there was this magic formula for doing the job, or imagine everyone is doing all of those things. If you do what everybody else is doing, by default, you won’t make money.
This is, by the way, the primary argument for “peak pod”, or the idea that the large multi-manager hedge funds are losing their competitive edge.
Since the multi-manager pod model has become widely adopted across the industry, there must be a saturation of ideas. As more funds employ the pod structure, many PMs are chasing the same alpha-generating opportunities, especially in liquid large-cap stocks or popular sectors (e.g., technology or healthcare). When too many pods pursue the same strategies (e.g., market-neutral long/short equity), the potential for differentiation decreases, and returns become more compressed. This leads to a situation where more capital is competing for fewer alpha-generating opportunities, which over time means they will make less money.
This argument implicitly assumes that most pods in a given strategy at these funds do the same things. This isn’t necessarily true.
What people are missing is that the diversity of successful investment processes at the best pod shops is second to none. It’s what makes the big platform model work. Yet among the investing community online, there is this incorrect notion that all PMs do the same things: trade on a weekly/monthly/quarterly basis, capitalize on short-term data points, run market-neutral and have “magical” financial models…
This couldn’t be further from the truth.
Time to spill some secrets…
“L/S pods have very short-term positions”
It is certainly true that their holding periods are shorter than those of the Tiger cubs or value-oriented funds. However, this doesn’t give you the full picture.
In addition to the occasional weekly/monthly/quarterly bets, some PMs have significant positions in compounder stocks in their portfolios that they hold for very long periods. These are stocks that can consistently generate strong and sustainable long-term growth through the compounding of earnings, revenue and cash flow. These companies typically have very high-quality business models with moats (strong competitive advantages) and very high ROIC (Return on Invested Capital). Historically good examples of compounders include Amazon, Apple, and Visa.
A significant number of PMs also use investment processes that don’t fit this shorter-term framework. Their portfolios consist mostly of ideas with a multi-month horizon — it could be a quarter or it could be 3-4 quarters.
“Modeling is a key part of our process”
Imagine you are a PM or an analyst covering AAPL - one of the largest companies in the world that the entire investing world follows. On every management call and meeting, there are at least 20 other people listening to the exact same things as you.
Do you really think you will form a differentiated view by having a “better” model?
For some successful investors in the multi-manager world, the model is a way to understand the business and stress-test your hypotheses and numbers, but not the other way around. You start with your thesis and then use the model to test it and quantify it. You don’t somehow magically look at your model and start forming a differentiated view. In other words, the model is often not a critical part of the process and is not a source of differentiation.