LONDON (Reuters) - Oil prices have continued to drift lower after plunging last week, highlighting the risk for traders trying to exploit mean-reversion strategies by buying futures contracts after a sharp fall in prices.
Front-month Brent futures prices tumbled by more than 7% on Thursday, a percentage change equivalent to more than three standard deviations for all daily price moves since 1990.
The one-day percentage decline was largest for more than three and a half years since February 2016, when prices were still close to their cyclical lows at just over $30 per barrel.
But if some traders were hoping prices would show a significant short-term bounce after such a severe sell off, they have been disappointed.
Front-month futures prices rose by just 2.3% on Friday, then fell again by 3.4% on Monday, and are still trading below last Thursday’s close.
Experience suggests prices do tend to bounce slightly in the days after a sharp sell-off, so there is an exploitable trading strategy, but gains tend to be small and highly uncertain.
Buying the dip is risky with relatively low expected returns and a high probability of making a loss.
Since 1990, front-month futures prices have fallen by 7% or more on a total of 44 days, including last Thursday, out of a total of more than 7,500 trading days.
By far the largest decline occurred on Jan. 17, 1991, when prices plunged by almost 35%, and was linked to the first Gulf War.
The Jan 1991 decline was an extreme outlier (the next largest percentage decline was less than 14%) so it has been excluded from the analysis that follows.
Following a decline of 7% or more, prices generally bounce in the days that follow, rising on average by a total of around 3.5% (tmsnrt.rs/2MLbDeb).
Most of the gains occur in the first 6-7 trading days after the initial slump and have largely disappeared within 10-20 days.
Rises are slightly more common than further falls in the first 10 days, but the margin is narrow, and they become almost equally likely after around 15 days.
Any post-crash bounce is small and fleeting, and it is almost as common prices will continue drifting lower.
For smaller crashes, where prices decline by 5% or 3% in a single day, the expected price bounce becomes even more marginal and transient.
Market commentators often characterise sharp one-day declines as an over-reaction and example of excessive volatility.
But if that were strictly true, prices should be strongly mean-reverting in the days after a crash, and experience suggests such effects are weak.
Many if not most sharp price falls seem to be a reaction to new information that has a significant and lasting impact on traders’ expectations.
New information could be about production and consumption, changes in the positioning of other traders, or both.
New information can therefore account for price movements following a change in fundamentals, but also seemingly unexplained flash crashes triggered by shifts in positioning.
The initial behaviour of benchmark Brent prices following the sharp decline on Thursday conforms to this pattern.
The U.S. president’s decision to impose more tariffs on imports from China, announced on Aug. 1, marked a significant escalation of the bilateral trade dispute.
Traders interpreted the escalation as making recession in the United States and around the world much more likely, explaining the fall in prices on Thursday.
Futures prices edged up on Friday as at least some tactical traders sought to benefit from any very short-term bounce, though the increase was smaller than average.
But prices eased again on Monday and Tuesday as the initial bounce faded and traders turned their attention back to the heightened risk of recession.
The sharp fall in oil is consistent with moves in government bond yields, interest rate futures contracts, and equity prices, all of which imply that recession is now much more likely because of intensifying trade tensions.
(John Kemp is a Reuters market analyst. The views expressed are his own)
Editing by David Evans