The US yield curve dipped into inverted territory recently. But that’s not necessarily a bad omen for equities. There are several important warning signals—and lately the yield curve’s slope is the only one flashing red.

The curve plots the gap between long- and short-term US Treasury yields, and there’s a reason investors pay attention to it: the curve has inverted before each of the last seven recessions. But inversion isn’t a foolproof recession indicator, and as our colleagues have noted, it doesn’t always mean disaster for stock markets.

So what warning signs should investors be heeding? To build a dashboard, we analyzed scores of signals from our models to see how well they predicted large equity sell-offs—defined as a drawdown of 15% or more—dating back to 1950.

This period included downturns triggered by the bursting of the dot-com bubble and the Great Recession, as well as the slumps that followed the end of the Bretton Woods system of fixed exchange rates, Black Monday in 1987 and the First Gulf War.

Gauging Effectiveness from Two Perspectives

We looked for signals with good precision, measured by how often its warning signal precedes an actual sell-off. Signals also must have good recall, measured by how often there is a sell-off after a signal begins to flash.

It’s not uncommon for a signal to score highly in one category but poorly in the other. Think of the shepherd boy in Aesop’s fable who always cried wolf. He was usually wrong. But eventually, he was bound to be right. In statistical terms, he had 100% recall—when a wolf finally showed up, he had predicted it. But he had very poor precision: he warned of many other wolves that never showed up.

Market signals have similar shortcomings. This is what the late Nobel Prize–winning economist Paul Samuelson meant when he quipped that the stock market had predicted nine of the past five recessions.

Building a Risk Dashboard: Four Signals to Watch