(John Kemp is a Reuters market analyst. The views expressed are his own)
By John Kemp
LONDON Commodities suffered the equivalent of a meteorite strike last Thursday and Friday.
Macro fears triggered an extraordinary plunge across a broad range of futures, with over-extended markets routed amid an acute lack of liquidity and the need to raise cash.
Commodity markets are infamous for violent price moves and lack of liquidity. But in an ironic twist, it was the largest, most liquid and normally well-behaved markets that showed the most abnormal falls -- indicating it was a synchronised macro-shock rather than deteriorating micro fundamentals which triggered the selloff.
Highly abnormal price moves were reported in silver, copper, gold, and nickel on Thursday, with lead and zinc making an appearance in the list of unusual discontinuities on Friday. In contrast, often-wild markets for U.S. natural gas, gasoline, heating oil, WTI and Brent showed only slightly elevated volatility and remained well-behaved.
The attached tables show how price moves on Thursday and Friday in the 24 liquid commodity futures included in the Standard and Poor's Goldman Sachs Commodity Index compare with typical daily volatility since 1990 (or 2005 in the case of the new gasoline blendstock contract).
Table 1 ranks the 24 contracts by their normal percentage daily price moves. U.S. natural gas, gasoline blendstock and heating oil have tended to be the most volatile, while aluminium, cattle and gold have been least prone to big daily moves.
Table 2 ranks the contracts in terms of kurtosis of their daily moves. Kurtosis measures the "peakedness" of the distribution.
The normal distribution has a kurtosis of 3.0. Distributions with a kurtosis of more than this ("excess kurtosis") have relatively more big price gains or falls and fewer small moves clustered around the average. Conversely, distributions with a kurtosis of less than 3 show lots of small price moves clustered around the average and relatively few big moves out in the tails.
As the table shows, all commodity contracts exhibit strong excess kurtosis ("heavy tails") -- which is why price moves should never be modelled with a normal distribution and big moves occur far more often than they would if the movements were normally distributed.
All commodity contracts exhibit much "wilder" moves than the normal distribution-- with lean hogs, gasoil, soy, Chicago wheat and crude (Brent and WTI) showing some of the most frequent extreme swings.
Based on Tables 1 and 2, it would have been reasonable to expect the largest price drops on Thursday and Friday to be concentrated in the petroleum complex and perhaps some grains and cattle contracts. Instead the most unusual falls were found in contracts like gold, silver and the base metals that are normally better behaved.
Table 3 ranks the percentage moves on Thursday in terms of standard deviations in each contract to correct for differences in typical volatility. Silver, copper, gold, nickel, Kansas wheat and corn all showed highly unusual declines of more than 3 standard deviations (more than 4 in the case of copper and 5 for silver). In contrast, normally volatile contracts like natgas, hogs, heating oil and gasoline barely moved.
Table 4 performs the same ranking for price moves on Friday September 23. Silver's stunning fall of almost 10 standard deviations, and gold's decline by almost 6, again stand out, while farm products showed no real movement, and the energy complex quickly settled.
No one should have been surprised by the scale of the falls last week. Even a cursory glance at historical commodity prices shows price moves are highly non-normal. The market swings from mild well-behaved volatility to a "wild" state and back again, to use terminology coined by the brilliant French mathematician Benoit Mandelbrot.
But the simultaneity of the declines and the fact the most abnormal falls were in the most heavily traded commodities (gold, copper) often treated as a proxy for global economic performance and exchange rates confirms the shock to hit commodity markets was a macro shock based on deteriorating expectations for the outlook, rather than a micro one grounded in contemporaneous declines in demand.
One interpretation is the selloff in precious and base metals has been overdone, as markets discounted a future slowdown which has not happened yet and is not certain to occur.
The data is also consistent with hedge funds and others liquidating profitable positions in gold and copper to meet losses and margin calls elsewhere.
Less optimistically, markets for oil, refined products and farm items may be concentrating too much on tight near-term supply-demand balances, failing to price in recession risks fully, as investors fail to appreciate how quickly the outlook is darkening.