How I Learned To Love the Bubble (Even Before it Bursts)

In seventh grade science class, we studied how things are classified – the systems used to organize various kinds of rocks, elements, species, different types of clouds. I’ve always liked the name “cumulonimbus” – the sort of big billowy heap of a cloud that has a dark grey underside – full of water, not yet transformed into rainfall. If you see those clouds and suddenly feel a cool breeze, that’s often the storm’s “gust front” racing ahead of the rain – and a good time to head for cover.

You wouldn’t look at the rainfall and think it has come from nowhere. Looking at the cloud, you know the rainfall was already there – just waiting for enough conditions to show itself. As warm, humid air rides up over the wedge of cooler air, the cloud builds – water vapor beading into droplets, gathering weight until they finally let go as rain.

We call rainfall a “conditioned” phenomenon because it depends on many other factors. When causes and conditions are sufficient, the rainfall manifests. When causes and conditions are no longer sufficient, the rainfall ceases to manifest.

We should be careful, when talking about rainfall, to consider the causes and conditions that produce rain. We might say the average amount of rainfall is this, or the average frequency is that, but if we don’t change our estimate even when there’s a cumulonimbus cloud over our head and a cool breeze in our hair, we may get soaked.

Likewise, suppose we look at historical stock market returns over any particular horizon, whether daily, weekly, or annual – regardless of valuations, market behavior, investor sentiment, monetary policy and other factors. We can collect all of those returns in a heap called an “unconditional” probability distribution. Historically, average annual market returns on the order of 10%, more or less, have been most common, so the heap is highest at that point, with progressively smaller “tails” for returns that are wildly positive or wildly negative. The overall profile looks roughly like a “bell curve.”

We might say the average market return is this, or the average frequency of a crash is that, but if we don’t change our estimate even when valuations are at the highest levels in history and market internals are ragged and divergent, we may get soaked.