(John Kemp is a Reuters market analyst. The views expressed are his own)
* Chart 1: tmsnrt.rs/2nkjmTj
* Chart 2: tmsnrt.rs/2neujUN
* Chart 3: tmsnrt.rs/2neEhWo
By John Kemp
LONDON, March 24 (Reuters) - No evidence has emerged of cheating on the OPEC/non-OPEC production agreement so far. But Goodhart’s Law is a reminder traders should expect countries to try to circumvent any target once it has been chosen.
Goodhart’s Law states that any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes (“Monetary Theory and Practice”, Goodhart, 1983).
Charles Goodhart was a Bank of England official and he was expressing the bank’s frustration that any measure of the money supply it started targeting ceased to be a good target soon after it was chosen.
Anthropologist Marilyn Strathern subsequently reformulated the law as a more general problem of measurement and accountability in any organisation or system.
“When a measure becomes a target, it ceases to be a good measure,” she wrote (“Improving ratings: audit in the British university system”, Strathern, 1997).
Central bankers have long known that once a target is chosen, banks and other institutions have a strong incentive to innovate around it to escape from control.
The question is whether OPEC and non-OPEC countries will fall victim to the same problem with their agreement to reduce oil output now they have settled on a measure to assess compliance.
The Organization of the Petroleum Exporting Countries publishes two figures for its members’ oil production in its “Monthly Oil Market Report”.
The first is based on direct communications from member countries while the second is based on a survey of six secondary sources, mostly statistical agencies and specialist oil industry publishers.
The secondary sources are Platts and Argus, the U.S. Energy Information Administration, the International Energy Agency, Cambridge Energy Research Associates, and Petroleum Intelligence Weekly.
The use of secondary sources dates from disputes in the 1980s about whether members were submitting inaccurate data to cheat on their output allocations (Goodhart’s Law in operation again).
So the OPEC agreement reached towards the end of 2016 relies on secondary sources rather than direct communications to set both the production level and assess compliance with the target.
The use of secondary sources was a key demand from Saudi Arabia to ensure the agreement was transparent, credible and verifiable.
But the selection of the secondary sources as the control target has put the system under scrutiny and pressure as never before.
Before the agreement was reached, Iraq in particular devoted a lot of effort to lobbying the secondary sources to raise their estimates of its output and bring them into line with government figures.
The apparent aim was to establish a higher production baseline from which any subsequent cuts would be calculated.
Iraq started to share detailed field-by-field production data with the secondary sources and other media organisations in an effort to bring their estimates closer to government numbers.
Now the agreement is in operation, the question is how accurately the secondary sources are capturing OPEC members’ production.
Secondary source estimates have become the focus of intense interest as OPEC officials have sought to focus oil traders’ attention on compliance.
In general, the secondary sources show aggregate compliance has been high, with Saudi Arabia shouldering more than its fair share to make up from weaker compliance by other members.
But in an ironic twist, confusion has arisen about the Saudi numbers, where have been unusually large discrepancies between government production data and secondary source estimates for January and February.
Secondary sources and government production data (notified to OPEC and the Joint Organisations Data Initiative) usually track each other closely in the case of Saudi Arabia (tmsnrt.rs/2nkjmTj).
Secondary source estimates are normally a little lower than government production data, but the gap averaged just 72,000 bpd in 2016 and was never greater than 102,000 bpd.
But in January 2017, the secondary source estimate was 117,000 bpd higher than the government data, for the first time in more than a year (tmsnrt.rs/2neujUN).
Then in February, the difference swung the other way, with the secondary source estimate 214,000 bpd lower than the government figure notified to OPEC.
The discrepancy caused significant comment in the oil market and the Saudi energy ministry responded by issuing a rare statement on March 14:
“The difference between what the market observes as production, and the actual supply levels in any given month, is due to operational factors that are influenced by storage adjustments and other month to month variables.”
Saudi sources have told Reuters that operational issues regularly require the addition of withdrawal of oil from storage (“Saudi pledges stable oil supply as market confused by data”, Reuters, March 23).
Saudi Arabia wants traders to focus on the amount of oil supplied to the market rather than the physical amount pumped from oil wells in any given month.
The difference between production and supply is caused by flows into and out from the country’s storage tanks.
Sources have told Reuters that production could fluctuate slightly from month to month, but supply to the market will remain stable at around 10 million bpd, in line with commitments to OPEC.
“What we are watching closely is the supply. Saudi Arabia will not supply the market more than 10 million bpd,” a Saudi-based industry source told Reuters.
Pumped production is the basis of the government’s direct communication to OPEC, but the supply figure is usually leaked to journalists and tends to form the basis of the secondary source estimates.
The OPEC agreement specifically referenced production, but Saudi officials insist supply gives a fuller picture of the amount of oil reaching the market and is what traders should focus on.
Saudi officials are keen to reassure the market they are committed and determined to reduce supply and bring global oil stockpiles back to more normal levels.
The discrepancy between production and supply figures in January and February may be related to an attempt to stabilise domestic crude stockpiles after an unusually large draw down over the past year.
Saudi domestic crude stocks have fallen from a peak of 329 million barrels in October 2015 to just 262 million barrels at the end of January 2017, according to government data (tmsnrt.rs/2neEhWo).
Stocks fell in 13 out of the last 14 months, with an unusually large draw down in January alone of almost 11 million barrels. Stockpile depletion cannot continue at this rate indefinitely.
Saudi Arabia remains one of the most transparent producers in OPEC but there’s much less clarity about output from some of the other parties to the agreement. For other OPEC countries, the market must rely on secondary source estimates and for non-OPEC countries, production numbers are self-reported.
Editing by David Evans