Market mean reversion takes longer than expected – CFM quants

Research on how long trends last could help avoid fallout from drawdowns like February’s

Jean-Philippe_BOUCHAUD.jpg
Jean-Phillippe Bouchaud: Research driving “another nail in the coffin of efficient market theory”

Quants at investment firm Capital Fund Management have identified that trending asset prices take years to self-correct – a finding that could help systematic investors navigate market turmoil such as that seen in February.

A forthcoming paper authored by six CFM staff says prices tend to trend over a two-year horizon, during which time they are on average between 50% higher and 33% lower than a notional fundamental value. But mean reversion, where prices come fully into line with their long-term “reference value”, takes place over a much longer period, the authors say. These two opposing influences happen concurrently: trends exhibit a stronger initial pull, with mean reversion becoming the dominant force subsequently.

“[Markets exhibit] a kind of tug of war between behaviour and reason,” said Jean-Philippe Bouchaud, chairman and chief scientist at CFM, speaking about the research at Risk’s Quant Summit Europe in early March.

The study’s findings challenge proponents of so-called “efficient markets”, who maintain that the return to equilibrium happens much more quickly.

CFM’s findings could help trend-following investors, in particular, avoid being caught in reversals such as that seen in February, when many such funds were blind-sided by the sharp decline in US equity prices and suffered heavy losses, having loaded up on equity exposure during the steady climb in stock markets last year.

Societe Generale’s Trend Index, which captures returns from a selection of trend-following hedge funds, fell more than 8% over the month, the third worst monthly drop since the index’s launch in 1999.

The paper – entitled Black was right: Price is within a factor 2 of value – examines data going back as far as 1800 for some time series, across markets including equities, bonds, currencies and commodities.

The aim was to test the economist Fischer Black’s intuition expressed in a 1986 paper – Noise – that markets could only be thought of as efficient in a very loose sense.

“Even highly liquid markets only equilibrate on timescales of years – and not seconds, as market efficiency enthusiasts would claim,” Bouchaud and his co-authors state.

The firm looked at how past returns on timescales from 10 days to five years predicted future returns over two days to one year, demonstrating a positive relationship over periods up to two years – higher past returns corresponded to higher future returns – but a negative relationship thereafter.

“It takes roughly six years for the price of an asset with 20% annual volatility to vary by 50%,” the authors state – a degree of drift that accords with an average level of movement as seen in the study. “Such a long timescale is another nail in the coffin of efficient market theory,” the paper adds.  

By implication trend-followers could benefit from combining momentum and mean reversion. “Trend-following strategies offer a hedge against market drawdowns; value strategies offer a hedge against overexploited trends. Mixing both strategies significantly improves the profitability of the resulting portfolios,” they write.  

Even highly liquid markets only equilibrate on timescales of years – and not seconds, as market efficiency enthusiasts would claim

CFM paper

CFM itself introduced a mean reversion element to its trend-following strategy in late February, though as part of long-term plans rather than in response to the month’s turbulence. The firm balances a trend signal and mean reversion signal, with a time lag applied to the mean reversion signal.

When markets are overpriced for longer than two years the strategy starts reducing exposure to a given trend, Bouchaud explains.

February provides a good example of when such an approach would be beneficial: the one-year trend in early February would produce a positive signal but the long-term value signal would have been negative, he said.

An additional benefit of the strategy is not having to rely on conventional metrics for value, Bouchaud observed, unlike value strategies that rely on accounting measures such as book-value-to-price which might be imprecise.

The research also has implications for high-frequency traders, he said. “On the high-frequency side there is no reason whatsoever for prices to be efficient.” 

CFM’s Institutional Systematic Trends fund was down over 7% in February, though up 3% for the 12 months to end-February. The firm’s Institutional Systematic Diversified fund, in which long-term trend following makes up about a third of the risk allocation and where the mean reversion strategy is now introduced, was down 3.29% in February and up 2% over the 12 months to end-February.

  • LinkedIn  
  • Save this article
  • Print this page  

You need to sign in to use this feature. If you don’t have a Risk.net account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an indvidual account here: