How long can we go on pumping oil out of the ground?

The oil industry calls the problem 'data transparency'.

As an example this week is a 'revision' to oil demand growth in the United States in 2004.

Previously the growth in oil demand was thought to be 2.4%, about 484,000 barrels per day. In fact it was 697,000 barrels per day or 3.5%.

That is in fact 46% more than was previously stated - a huge revision.

"Oil market data is generally a black art like using a set of chicken bones," says Paul Horsnell of Barclays Capital. "If Columbus had thought he'd hit India when in fact he was in the Caribbean, that's about the level of oil market data."

"The revisions to US demand growth are small in percentage terms, they are generally 99% accurate. But the change is huge in barrel terms, and this is from the USA who have the best oil data in the world."

The barrel difference was in fact 213,000 per day. Added up that is 77.75 million extra barrels per year, about one day of global production.

"Oil data is like paint thrown across a canvas, you get the broad outline of the situation. But even then it's not just a Jackson Pollack painting, the paint actually moves of its own accord after it has been applied," says Mr Horsnell.

Phantom reserves

One of the major problems surrounding oil data is in reserves.

These are the basins of crude oil that lie underground.

They are either held by governments or the 'oil majors' like BP, ExxonMobil or Shell, or a combination of both.

Many countries simply do not allow outsiders to audit the size of these fields.

This is especially true of the major Middle East oil producers of OPEC and the countries of the former Soviet Union.

Some believe that reserves stated by OPEC countries such as Kuwait and Saudi Arabia are not accurate.

"There are a lot of questions to answer over OPEC reserves," says Bruce Evers of Investec Bank. "The quality of overall oil market data is poor, but with OPEC there remains considerable debate over the reliability of their reserve estimates."

Sudden revisions

One of the main reasons is that in the 1980s OPEC decided to switch to a quota production system based on the size of reserves.

The larger the reserves a country said it had the more it could pump.

The more it could pump the more money it could make.

As a result in 1985 Kuwait revised its reserve estimates by 50% overnight.

It was soon followed by United Arab Emirates, Iran, and Iraq. In 1988 Saudi Arabia became the last to join the revised reserve estimates party, adding a whopping 88bn barrels.

Unexplained changes

"Something needs to be done," says Mr Evers. "OPEC have never fully explained the reasons behind these changes, they have never issued any guidelines. The market needs to know."

Although previous estimates may have been conservative, what troubles some analysts is that twenty years later, these reserve estimates are unchanged, in fact some have increased.

Whilst it is obviously possible to add reserves by new field discoveries it can seem a perplexing situation to market makers.

Kuwait for example still claim exactly the same reserve level as they had in 1985 despite pumping millions of barrels every day since then.

Nor are company estimates any better, with Shell forced to make four revisions downwards of its official reserves since 2002, losing around 4.8bn barrels and damaging its share price.

Unclear figures

Even current figures for OPEC production are unclear.

OPEC say they are producing exactly 28 million barrels a day (mbpd).

This includes their latest 500,000 barrels per day increase announced at their last quarterly meeting by Kuwaiti oil minister Al-Sabbah.

But OPEC have also admitted that their members break their own quotas to take advantage of high prices.

So is it really 28mbpd?

The International Energy Agency says OPEC pumped 29.3 mbpd in May 2005.

The IEA say this is actually a fall from April 2005 of 55,000bpd.

Who is correct? "There is no official OPEC output data," says Mr Horsnell. "they just kind of pass on the data they are given by their member countries. It is really not that easy for OPEC, you can't blame them, it is down to their members."

Forecasting demand

"I don't rate IEA data either," says Mr Evers. "they have horrendously underestimated demand in the past, it is one of the reasons we are where we are now. They are little more than a data collection agency, and the data they are given is already tarnished."

It is no easier to forecast the future demand for oil, and analysts are growing increasingly sceptical of oil company attempts to do so.

Energy Files director Dr Michael Smith said "it is no longer appropriate to accept glib demand forecasts from oil companies, financial institutions and governments¿suggestions that oil consumption will grow to up to 120 million barrels per day by 2020 and that automobile and airline traffic will increase at extraordinary rates are futile and damaging."

But Paul Horsnell says that gaps between data-sets can in fact show up areas of the oil market that need careful study.

"Take Russian production as an example," he says. "There are all kinds of rosy forecasts and then there are people like me who think it's all rather bad news. But there are many reasons about why it is impossible to measure oil, it's a liquid for a start.

"There are huge margins of error with oil data and it has to be treated as such. It's the nature of the product. Thinking you can measure it to the eighth decimal point, well, it's just a waste of time."

As oil prices continue to soar, the lack of accurate data could make it harder for the oil market to predict its future direction.