London, 15 April 2011: Reuters

Demand destruction has taken over from geopolitical risk as the number one concern in the oil market -- as participants assess whether prices have risen enough or must go further to cut demand in line with diminished expectations about supply and spare capacity.

Preliminary data for early 2011 already show signs of slowing oil demand, according to the latest monthly "Oil Market Report" published by the International Energy Agency.

Goldman Sachs analysts claimed to see "nascent signs of oil demand destruction in the United States" in their muchcited update advising clients to close recommended long positions in crude.

Reuters reports Saudi Arabia has cut its oil production back to pre-crisis levels, citing poor demand, and lacklustre interest in the new special blend of crude offered to replace Libya's light sweet oil. Many in the industry have interpreted this as a sign high prices have started to dent consumption.

But in a note to clients Barclays Capital analysts warn "it is far too premature to signal that the first signs of demand destruction are already noticeable". They question whether Riyadh has really cut output in response to a fall in demand and conclude "price levels at which global oil demand is choked off have not been truly tested yet".

LAGGING, NOISY INDICATOR

Barclays analysts are correct to note the absence of clear signs of demand destruction, though perhaps not for the reasons the authors cite the report. Consumption data provide a notoriously backward-looking and noisy picture of the market. Most of the information is subject to a lag of at least two months -- in some cases much more. Weekly numbers published by the Energy Information Administration ( EIA) are an exception. But even they are subject to reporting and estimating errors as well as some confusion between exports and domestic consumption.

In any event the United States accounts for a diminishing share of global consumption. EIA's data provides information on a small, unrepresentative section of the global market and should not be used to extrapolate worldwide trends.

As a growing share of global consumption shifts to China and other emerging markets, a timely and accurate picture of demand is becoming even harder to obtain. Moreover, most countries and organisations still calculate "implied consumption" as a residual from other estimates for supply, inventories and cross-border trade, so any errors or uncertainties in those statistics contaminate the demand data.Demand destruction would be expected to involve a change of no more than 1-3 percentage points in consumption. Given lags in data collection, limitations of survey methods and the virtual absence of data on demand in many developing countries, hunting for demand destruction on this scale is like looking for a needle in a haystack.

Demand destruction is usually evident only in retrospect or once the process is well underway. Barcap analysts are almost certainly correct doubting whether the first signs are already noticeable. But that does not mean that demand destruction has not already begun.

FEEDBACK LOOPS

The impassioned debate about whether demand destruction is already occurring is an exact re-run of the discussion in late 2007 and the first half of 2008 with institutions taking exactly the same positions they did then. Most banks and oil bulls continued to insist through the first half of 2008 they saw little or no evidence of demand destruction, even as prices crested over $100 per barrel and went on to peak at over $140.

In retrospect, observers acknowledged rising oil prices had begun to weigh on both oil demand in the first half -- directly through incentives to economise on consumption and indirectly through their impact on the global economy. But at the time the question of demand destruction became a matter of bitter controversy and almost theological debate.

It is impossible to state precisely how much rising prices contributed to the global recession that began at the start of 2008 and accelerated following the collapse of Lehman Brothers and the global financial system in the last few months of the year. Other factors including the bursting of the U.S. housing bubble and the subprime mortgage market, spreading panic, and seizing up of trade finance all played an important role.

Likewise, it is not possible to state precisely how much the recession contributed to the subsequent decline in oil demand and prices in the latter half of 2008 and 2009. Oil prices had fallen by nearly a third before the Lehman collapse. There is no way to know what would have happened if the banking panic had not erupted.

Prices and growth are intertwined and jointly determined. Unlike a physics experiment, there is no way to re-run history holding other variables constant to isolate the price impact on growth and vice versa. Economic forecasts such as that given by Fed Vice-Chairman Janet Yellen in a speech this week, which show continued recovery and moderating inflation based on an assumption that energy and food prices will level off, are not terribly useful. The assumption dominates the forecast.

Nor are oil price forecasts based on assumptions about unchanged global growth. They are based on a basic contradiction -- recognised by the IEA when it observed "a sustained, $100 per barrel plus price environment will prove incompatible with the currently expected pace of economic recovery". Forecasters assume away or minimise complex feedbacks because it makes the job easier and gives more determinative outcomes. But real market participants understand uncertainty and indeterminate outcomes are both necessary for markets to exist and (ultimately) part of the fun and what makes them so interesting. Assuming them away misses the point.

LESSONS FROM 2008

While admitting there is no way is establish precisely what impact rising oil prices had on the global economy in 2008, it is still useful to examine the precise timeline of events back in 2007 and 2008.

The chart below shows real oil prices and the point at which activity peaked in the advanced economies measured by peak manufacturing output.

There are obvious defects with this approach. Manufacturing production is estimated. There is considerable volatility in the series (especially for some countries). In some instances production reached a sharp peak in others it hit a bumpy plateau and then declined. But the cluster of countries which started to see manufacturing activity decline between December 2007 and April 2008 is notable.

Recession had begun to sweep around the world economy in Q1 2008. It came well after the U.S. housing bubble had started to burst in 2007 but also well before oil peaked over $140. Prices were around $90-110 when the advanced economies began to tip into recession.

Oil prices impact the economy with a delay. If we assume an (arbitrary) lag of three months then recessionary forces started to work on the economy when oil prices were at $80-100. Rising prices were not the only cause of the last recession. But if they made some contribution the impact started to be felt when prices were far below $140.

The chart shows prices continued rising for some months after activity had peaked in the advanced economies -- and therefore several months after some form of demand destruction had set in. Like Roadrunner in the Warner Brothers cartoon, oil continued climbing for several months after the economy had gone over the cliff.

One final point. It is well understood the impact of any given rise in prices is magnified the longer it is sustained. Market participants come to see it as a permanent shift requiring adaptation rather than temporary spike to be endured. Demand destruction does not all occur at one moment when prices hit a new level. It is a process that intensifies over time. The absence of obvious signs of demand destruction at the moment should not cause anyone to assume it is not happening. Past evidence suggests it is more likely than not that demand destruction has already begun to occur and will intensify in the months ahead.

Ends --

The views expressed here are his own.