When does momentum shift to a risky bet?

Two finance professors at the London School of Economics have introduced a new way to measure the amount of risk arbitrage in markets.

Their novel measure, dubbed “comomentum”, exploits time variation in how momentum stocks excessively “comove” together to reveal how crowded the classic price momentum trading strategy is at any point in time.

They show that comomentum predicts when momentum trading transitions from a profitable, under-reaction strategy to a risky, crash-prone bet. This new work was one of the presentations at the London School of Economics investor roundtable in June, hosted by conexust1f.flywheelstaging.com.

The professors claim the practical implications of their insights are enormous, allowing individual managers to allocate their capital more profitably as well as collectively ensuring a less risky, more efficient global financial system.

Christopher Polk, professor of finance and director of the financial markets group at the London School of Economics, is one of the professors who have coined the new word – comomentum.

And as he explains it is a measure of how momentum stocks comove together, revealing those times when momentum strategies become unprofitable and crash-prone.

“Professional investors, or arbitrageurs, play a key role in finance,” Polk says. “In our standard models, they work to ensure that markets are efficient. While this is often a realistic portrayal, a more nuanced view recognises that arbitrageurs’ positions can become crowded; with too much trading destabilising prices.”

Polk emphasises that it is essential these dynamics are understood but that it is extremely difficult to measure the amount of arbitrage activity at any given time.

“Direct measures of arbitrage capital employed are entirely inadequate. They estimate only a portion of the inputs to the investment process, for a subset of arbitrageurs,” he argues.

“Our insight is to examine the outcome of that process. In particular, we measure the correlated price impact that occurs when arbitrageurs collectively invest in similar strategies. “Our approach can then be used to test whether arbitrage activity can push prices away from fundamental value.”

Polk and his LSE coauthor Dong Lou use their novel measure to study momentum trading. Polk points out that momentum’s positive-feedback nature makes it especially susceptible to crowding.

“Prices are going up because arbitrageurs are buying, which in turn makes the positions even more attractive to them,” he says.

Polk first shows that comomentum is correlated with variables linked to the degree of momentum trading. Then he uses this new measure to forecast key properties of momentum trading strategies.

“As comomentum increases, momentum predictably transitions from being a profitable strategy exploiting under-reaction to a volatile and crash-prone overreaction bet,” he told delegates at the London School of Economics roundtable.

Indeed, Polk reveals that the comomentum measure increased dramatically prior to the recent momentum crash in March 2014.

“The problem is that momentum trading doesn’t have a fundamental anchor to tell arbitrageurs when to stop trading.” Polk states. “Contrast that with a value bet. Value investors naturally reduce their value bets when the value spread narrows.”

“We provide a new way of measuring arbitrage activity that exploits the output of the arbitrage process, namely excess return comovement. Our results suggest that ‘smart money’ can be destabilising, when positive-feedback trading becomes crowded,” he says.

“The practical implications of our insights for active investment strategies are enormous. Our new measure provides a tool that not only allows individual managers to allocate their capital more profitably but also collectively ensures a less risky, more efficient global financial system.”

The academic paper can be accessed here Comomentum – Inferring arbirtage activity from return correlations – Christopher Polk