Here is a tough question I was recently asked by an institutional investor: how do you know when to call it a day and fire your manager?
Sticking with a manager whose alpha process has appeared to have stopped working is one of the toughest decisions to make. On the one hand, you don’t want to chase recent returns. On the other hand, does that mean you should never fire a manager at all?
The consensus reason to fire your manager is when they deviate from their investment process. While appearing useful, it produces two big problems.
If a manager is sticking with their process and still continues to under-perform, the client now has to break the same principle to which they uphold their managers and “Give Up” on their process.
When a manager only sticks with their process, they sacrifice innovation - which is actually the source of all alpha (see here). Unless, constant innovation is part of the process, the system is likely to fail.
In many ways “stick with your investment process” manta makes sense. As we often say, the true risk of investing is Giving Up (see here). Sticking with a chosen investment process is the antidote to Giving Up.
But Giving Up is just the outcome of the problem not its cause. It’s the “Advil” for the headache.
If we look a layer deeper and ask: why do investors want to Give Up? The answer is usually simple: performance. Either they have lost a lot quickly (drawdown risk) or they have under-performed by a little slowly (low return risk) (see here). Either way, they lost confidence in their manager.
So how does ‘sticking with your investment process’ solve the confidence in performance problem?
If you ask the managers, the answer is simple: Mean Reversion.
Yes, there is a lot of Mean Reversion present in the markets. But it’s just the nature of normal distributions and not an indication that your process started working again. In fact, I can create a random set of numbers and if I plot their cumulative performance, it will show mean reversion. That does not mean the investment process is any good. It just means that a normal distribution generates mean reversion.
The real question behind mean reversion is if there is any Mean to revert in the first place? (see here)
Given the wide acceptance of the “stick with your process” mantra, hiring and firing managers leaves the client investor without any framework, except to react to performance.
Here are the two criteria I’d recommend looking for in a manager outside their recent performance and their adherence to their stated process: