London, 23 March 2011: Reuters

Prominent forecasters have begun to estimate the amount of political risk baked into oil prices. But even the best have only focused on assigning a dollar value. The real measure of political risk is how much demand must be destroyed to restore a comfortable risk adjusted level of spare capacity.

According to oil analysts at Goldman Sachs, one of the most respected research teams among the major banks, "Record net speculative length in WTI over 375 million barrels implies a $10/barrel risk premium in oil price from recent events."

In a thoughtful note they write: "We estimate that each million barrels of net speculative length tends to add 8-10 cents to the price of a barrel of oil. Given that net speculative length has been about 100 million barrels higher since political protests spread from Tunisia and Egypt to Libya, this suggests the oil market has been pricing a $10/barrel risk premium".

Goldman argues this is why Brent has been trading $10 above its target of $105, published before the unrest in Tunisia. Quantifying the political premium is a brave and vital step to understanding whether the market is pricing political risk properly. Most analysts and oil forecasters still talk about political threats to supply in rather vague and undefined terms.

Problems in assessing high-impact, low-probability events are well known. But unless an effort is made to quantify the impact, the concept of political risk remains meaningless and is not helpful to investors or those responsible for hedging and insurance.

Goldman's estimate for the risk premium appears reasonable -- consistent with the modest price increase and rise in volatility observed since protests spread to Egypt in late January. But there are four aspects that deserve more scrutiny, three of them rather technical and the final one more profound.

CLASSIFICATION ISSUES

Relating price impacts and political risk to net speculative length -- for example using positioning data from the U.S. Commodity Futures Trading Commission commitments of traders (COT) report -- presents inherent problems.

Even with the new disaggregated report, which divides participants into five groups rather than the old three-way classification, it remains difficult to separate speculative positions from hedges accurately. Most categories contain a mix of positions initiated for speculative and hedging purposes.

Moreover, because many positions are option-related, they vary directly with prices. Futures-equivalent positions related to upside call options appear to get bigger automatically as prices rise and the calls become more valuable.

From a regulatory perspective, such as for the enforcement of position limits, passive position increases are treated the same as active decisions to initiate new futures or option contracts. But it is not clear whether such passive position increases should be treated as the same when measuring attitudes to political risk and the price impact.

SPECULATOR IMPACT

Linking prices to changes in speculative length has to be handled with extreme care. While the recent run-up in prices has been associated with a record net long position in futures and options held by speculators, the run-up in the first half of 2008 involved only a modest speculative long position, which was liquidated well before prices peaked in June and July 2008.

Banks, exchanges and some regulators such as Britain's Financial Services Authority point to the absence of a large speculative position in 2008 to claim the price increases were driven by fundamentals, not speculation, and to explain why position limits are unnecessary and harmful. Numerous economists claim to have found no evidence of a statistical relationship between speculative positions and prices.

But it is hard to reconcile claims there was no speculative impact on prices in 2008 (because there was no net speculative length in the futures and options market) with claims there is now a speculative "political risk premium" (because there is a record speculative long position).

To sustain this apparent contradiction, speculators would need to be worried about supplies at present (when there are 3-4 million barrels per day of spare production capacity and healthy inventories) but entirely unconcerned in 2008 (when spare capacity was already down to less than 1 million barrels per day and forward demand cover was much lower). It is possible to make this argument work, but it stretches credibility.

Goldman is right to identify net speculative length as among the price drivers since the outbreak of unrest (and implicitly the build-up of speculative length in late 2010 must also have had some impact on prices in the final months of 2010).

But that must prompt a broader re-assessment of how expectations and speculative positioning contributed to previous price movements. The "no evidence of a speculative impact on prices" argument beloved of some economists and banks simply does not stand up to scrutiny.

LIQUIDITY NOT LINEAR

Goldman argues each million barrels of speculative length adds 8-10 cents to the price of a barrel. This is problematic at two levels. First, the conventional wisdom among banks, exchanges and regulators has been that liquidity in futures and options is infinite. For every buyer there is a seller.

Since liquidity is infinite, an unlimited number of futures and options can be created to meet demand at any given price level with no impact on the price.

The assumption of unlimited liquidity is obviously unrealistic. Goldman rightly acknowledges liquidity is not infinite. Prices have to move in order to discover new pockets of liquidity and allow deals to be struck. But the analysis is still problematic because it assumes liquidity is linear in price. To buy an extra million barrels of futures contracts, the buyer must push up prices by around 8-10 cents. To buy 100 million extra barrels prices must rise $8-10.

In fact, liquidity is not linear. Patches of good liquidity and liquidity holes are familiar to traders. More important, liquidity is likely to be less available for sudden, extreme price moves; at very high or very low prices; and in the face of uncertainty.

Oil markets became illiquid with violent daily price movements around the price peak in 2008. Liquidity may also have fallen recently, in which case Goldman could be underestimating the political risk premium.

BARRELS NOT DOLLARS

Apart from these technical issues, there is a more profound problem. The political risk premium is often expressed in terms of extra dollars per barrel on top of what the price would be in the absence of political concerns. While this is a good short-term way to measure political risk, in the medium and long term the appropriate measure is the number of extra barrels held in inventories or barrels per day of surplus production capacity to balance the market.

The price premium is merely a signal to build these additional buffers through some combination of extra investment, stock building, slower growth, conservation and substitution.

This points naturally to the fundamental question: How can the oil market adjust to a sustained increase in political risk? Strategic reserves could be increased and commercial stocks built higher (via a sustained contango). But stocks are already high. No realistic level of stock building could hope to insulate the market from a really major disruption to 5 million barrels per day of production capacity.

Investment in more spare capacity could relieve some pressure. But only swing-producer Saudi Arabia has an interest in building capacity to have it lie unused. Besides, more capacity in Saudi Arabia would not help. It is the risk of unrest in the kingdom and its neighbours about which markets are worried. Spare capacity would have to be built elsewhere. It is not clear how.

So a sustained increase in risk will have to build inventories and spare capacity via price increases and a reduction in demand.

If political risk rises, demand growth must slow until the market's desired level of spare capacity has been restored.

Ends --

By John Kemp, Reuters market analyst – for Commodities Now.

The views expressed here are his own.