|April 22, 2011|
What would a Martian make of the oil market?
|Oil markets have remained strangely calm despite being hit by a series of large shocks since the turn of the year. Looking only at prices, what would a newly arrived Martian spaceman make of the market at the moment?
Escalating violence and the loss of oil output in Libya; intensifying unrest across the Middle East; and the earthquake in Japan. Each could have been expected to produce a significant price response and surge in volatility. But like Arthur Conan Doyle's fictional dog that did not bark in the night, price and volatility changes have been notable only for their absence.
In the last few weeks oil prices have seemed comfortable range-trading in the face of heavier than normal news flow and a broad range of positive and negative shocks.
Our Martian looking only at the time series of crude prices over the last two months and not the news wires would not guess the market had been hit by conflict spreading across the oilfields of a major producer, social tension in another, and one of the largest earthquakes recorded in a consuming country.
MONETARY POLICY AND POLITICS
Oil prices have risen strongly in the last seven months, but most of the move occurred before the outbreak of unrest in North Africa and must therefore have been driven by other factors.
Spot Brent has risen $18 per barrel (19 percent) since protests spread from Tunisia to Egypt and market participants began to worry about contagion to the rest of the region, including the major oil producing monarchies around the Gulf.
But they had already made a larger move of almost $20 (26 percent) between the time of Fed Chairman Ben Bernanke's speech at Jackson Hole in late August 2010 and the first protests in Egypt, reflecting the expected impact on liquidity, growth and sentiment stemming from the second round of quantitative easing.
Senior Fed officials continue to insist the run up in prices is mostly attributable to concerns about political unrest. The record suggests otherwise. Prices were on a strong upward trajectory before the market started to price in a greater threat of supply disruptions. Unrest explains less than half the rise in prices since August 2010 (Chart 1).
The price response to the loss of 1.3 million barrels per day of Libyan exports and a small but non-zero risk of the loss of a much larger volume of exports from the Gulf region has been exceptionally muted.
Markets become somewhat more volatile, but the response like prices has been surprisingly muted. Close-to-close volatility has risen from 20 percent per annum before the first street protests in Egypt to 36 percent at the end of last week. The protests and Japan's earthquake have resulted in a marked uptick after a period in which volatility had generally been declining (Chart 2).
But while prices have certainly become more volatile, the increase seems small given the size of the shocks to which the market has been subjected. Realised vol of 36 percent per year is not particularly high in the context of the last 20 years. Realised volatility of 36 percent lies in only the 75th percentile of the distribution for all realised volatility since 1991 (Chart 3).
In fact it could be argued that rather volatility becoming high in the wake of the crisis, volatility had become unusually low beforehand, and unrest has merely restored it to more normal levels. Prior to the crisis, volatility of 20 percent per year was in the 15th percentile of the distribution -- which was more unusual than current volatility of 36 percent in the 75th percentile (Chart 4).
The most commonly used volatility measures rely on changes from one closing price to the next, so they do not capture the full variability of within the trading day. But alternative measures such as the Parkinson volatility estimator can be calculated using daily high and low prices.
Parkinson volatility takes account of the whole intra-day range and in principle can be more efficient measures of market variability.Using the Parkinson estimator, there is some evidence of heightened intra-day volatility since the onset of the crisis as the market tries to process heavy news flow. Parkinson volatility has doubled from 20 percent to 40 percent since Jan. 24.
Parkinson volatility is now in the 93rd percentile of the distribution since 1991. Traders are not wrong to think that prices have become more volatile during the day than before the crisis erupted -- even if a significant portion of that variability looks like noise and is reversed by the time the market closes (Charts 2-4).
But again the main impact of the crisis has been to move the market from a period of unusually low intra-day volatility before January 24 (the 21st percentile of the distribution) to a period of unusually high but not extreme volatility now (93rd percentile). It is impossible to know how much of the is attributable to "fundamentals" (real changes in current and expected supply and demand) and how much is due to "noise" (over-reaction to news flow by uninformed market participants). But the sharp increase in intra-day volatility while close-to-close volatility remains more muted suggests price changes have a higher noise content than normal at the moment.
THE DOG THAT DIDN'T BARK
The lack of a bigger response in terms of either outright prices or observed volatility requires explanation. If a threat to the Middle East countries accounting for onethird of global oil production and Japan's nuclear disaster do not move prices and volatility significantly, something is clearly happening.
Several explanations suggest themselves and are not exclusive:
(1) Market participants are still struggling to comprehend and price changes which are unusually large and complex. It is well known that prices do not always adjust quickly and can take time to fully incorporate fundamental shifts in supply and demand. If that is what's happening here, the market could still see significant changes in prices and volatility in the months ahead as the full impact are more fully understood.
(2) Market participants may be assuming recent crises cancel one another out. Supply shocks from the loss of Libyan crude and Japan's nuclear generators as well as the possible threat to production in the Gulf region could be offset by a demand shock from the earthquake and reduced consumer and business confidence around the globe.
(3) Market participants may have concluded recent shocks, while frightening, do not alter the supply-demandinventory balance much. Loss of Libya's exports has been largely replaced by output from other sources. Increased demand for oil-fired power generation in Japan is not significant on a global scale (at least for the crude market rather than diesel and fuel oil).
Interruption of oil output in Saudi Arabia and other Gulf producers would be very different -- a genuine gamechanger transforming the supply-demand balance, but market participants may have concluded the risk is still remote, especially given the determined suppression of protests in Bahrain and Saudi Arabia recently.
(4) It is not clear how the market should price extreme impact but low probability events such as political unrest and disruption to oil supplies in the Gulf. It may already be pricing in a small chance of disruptions to Saudi and Gulf output.
On the other hand, participants may be seriously under-estimating and under-pricing the non-zero possibility of a really serious supply disruption. Markets often struggle pricing tail risks, sometimes wildly overpricing them and sometimes underpricing them. It is not clear which is the case at the moment.
By John Kemp, Reuters market analyst -- for Commodities Now.