You are running a multi-asset portfolio and know for certain that over the next five years there is a 20% chance of an occurrence that will cause it to suffer severe losses. Fortunately, there are some assets that you could add to the portfolio that would protect it from most of the drawdown – but there is a catch. Investing in these defensive positions will cause the portfolio to meaningfully but not disastrously underperform its benchmark in the other 80% of future scenarios. What is a rational investor to do?

It depends.

If we are investing for ourselves, we would almost certainly buy the assets that provide some insurance – the risk of underperforming a benchmark is largely irrelevant. What we care about is meeting our long run objectives.

But what if we are running a portfolio where relative returns matter? What if poor relative returns might cost us our job? The incentives change quite dramatically here – what is rational for an individual investor might not be so rational for a professional portfolio manager.

What is worse: a 20% chance of severe losses, or an 80% chance of underperforming for five years? Five years is a long time. Most investors have patience for about three years of sub-par returns.

The primary goal of running a diversified, long-term portfolio should be to maximise the probability of delivering good enough outcomes, while minimising the likelihood of very bad results. It requires us to make decisions about things that could happen but don’t. This seems obvious but the structure of the industry makes it far from easy to follow. 

In the (admittedly heavily stylised) example I outlined a professional investor who makes the smart decision to protect the portfolio from a low but meaningful probability risk looks like they are doing a bad job in 80% of future worlds. Conversely, the investor focused on improving the odds of personal survival looks like they have made better choices 80% of the time.

The underlying challenge is that good portfolio management is about creating a mix of assets that is robust to a complex and chaotic world, whereas our means of measuring and incentivising success assumes the world is linear. A choice was made and it was either right or wrong.

This situation creates an agency problem where a portfolio manager is primarily assessed on an outcome that is subordinate to what should be their primary goal, and this can overwhelm their decision making (whether they admit it or not).

If the ultimate aim of a portfolio is to deliver a real return of some level over the long-run, but the portfolio manager has a separate reference point (benchmark or peer group) and time horizon (three years if exceptionally lucky, probably much shorter) their choices will almost inevitably be driven by the latter. That might be entirely acceptable, but we should not ignore that it will encourage very different behaviours.  

Of course, having these types of discussions is like howling into the void – nobody really cares. In no other industry is Goodhart’s law (when a measure becomes a target, it ceases to be a good measure) more apparent than the investment industry. You could boil the whole thing down to one dictum: “number higher good, number lower bad”.

Doing anything other than obsessing over short-run relative portfolio performance is admittedly exceptionally messy. Nobody is going to wait twenty years to find out if something has worked well, and you can excuse any investment mistake by saying: “I was just preparing for a world that didn’t occur”. Just because something is imperfect and difficult, however, doesn’t mean it is not better than the alternatives.

What has seemingly been forgotten is that there is a yawning gulf between these two statements:

“I am investing to meet my clients’ long-term outcomes, hopefully my approach will mean I can do it better than others over time.”

and

“I am investing to beat the returns of people doing similar things to me, hopefully I might also deliver good long-term outcomes”.
  
The focus of these two statements is completely distinct and the types of decisions we are likely to make similarly disparate.

Although the investment industry is not going to change, it is worth asking whether a portfolio manager’s incentives are skewed so much that decisions that are good for them might not be best for their client.   

—

My first book has been published. The Intelligent Fund Investor explores the beliefs and behaviours that lead investors astray, and shows how we can make better decisions. You can get a copy here (UK) or here (US).