By MIKE MEYERS
Minneapolis-St. Paul Star Tribune
August 09, 2007
"I've done 120 short-term energy outlooks and I've probably gotten two of them right," said Mark Rodekohr, a veteran Department of Energy (DOE) economist.
"We've long been embarrassed by our mistakes," he said.
Private forecasters have done little better. Even with Monday's big drop, if oil prices don't fall a lot further, 2007 will mark the ninth year in a row that the "market consensus" guessed low on how high oil prices would go.
On average, private forecasters have undershot their target by 31 percent each year, according to a recent analysis by Deutsche Bank. In the past five years, the price of a barrel of oil has tripled. The fact is, few experts saw it coming.
"Analysts get married to a forecast or a particular view," said Michael Lewis, global head of commodities research at Deutsche Bank's London office. "It's hard to break free from history."
The results of faulty oil price forecasts can be far-reaching.
Project too low and the auto industry will build big-engine gas-guzzlers that won't sell when fuel prices surge and airlines will be punished for moving too slowly to buy fuel-efficient planes.
Project too high and millions of fuel-efficient cars suddenly will look unappealing if gasoline prices ebb and oil companies will invest billions in oil drilling and refineries that won't pay off in the long run.
Bad assumptions about oil supply and demand also can tilt government policies in the wrong direction. Governments, here and abroad, routinely make decisions about oil company taxation, subsidies and strategic reserves based on petroleum forecasting.
To cover their trails, forecasters are issuing more oil price projections than ever, responding to changes in real world oil prices that don't match what they said last month.
"The old rule of forecasting is that if you can't get it right, do it often," said Adam Sieminski, chief energy economist at Deutsche Bank's New York office.
Many problems plague forecasters trying to predict the path of oil prices. High on the list is assuming that any change in oil prices is short-lived and that, on average, oil prices will end up in the next few years near the levels of the last few years.
"You expect a reversion to the mean," said Brian Youngberg, senior area analyst for utilities and energy at Edward Jones, a St. Louis-based brokerage firm.
"Even three or four years ago, that price was $25 to $35 a barrel," he said. Today, the average price forecast is closer to $50.
"We tend to jump on a bandwagon," said Rodekohr, at the DOE. "We're almost always wrong.
"When oil prices are low, everybody says they're going to be low forever," he said. "When they're high, everybody says they're going to be high forever."
Hurricanes can disable refineries and oil rigs, pipeline trouble can cause a surprise shutdown for months of repairs, or political instability in oil-producing countries, from Venezuela to Nigeria, can disrupt supplies and force oil prices higher.
"It's very difficult for economists to know that (in advance)," said John Kingston, director of global oil at Platts, an energy news service based in New York.
Oil price forecasters are trying to learn from their mistakes.
Until recently, computer models designed to generate forecasts often weighted U.S. oil demand as the leading indicator of whether prices would go up, down or sideways.
Lately, more forecasters are tagging oil demand from rising economies -- chiefly China and India -- as a greater force in changing prices.
"Demand growth in India and China is going to make a huge difference in oil prices in the next few years," Rodekohr said. "Some private sector companies have jacked up price forecasts for the next few years, based on that assessment."
Prosperous industrial countries are showing little or no increase in demand for oil -- presumably responding to rising oil prices -- even as China and India seem almost impervious to oil price jumps.
"In those countries, higher prices don't seem to be taking a bite out of demand," Kingston, at Platt, said of China and India.
In the past, economists would have predicted that developing countries would ratchet back oil demand in the face of higher prices far faster than stronger, developed countries.
In an era of price volatility, some oil forecasters are following the lead of the Federal Reserve, which for years has issued a range of possibilities in economic forecasts -- showing growth rates if current circumstances persist and alternatives in more optimistic or pessimistic environments.
Tweaking oil price models and revising forecasts in response to changing events is showing signs of paying off. Lewis, at Deutsche Bank, said he expects the average forecasting error to remain in double digits but be improved over some whopping mistakes in recent years.
"The good news is that forecasting error is getting smaller, possibly only 13 percent this year and hence the lowest since 2001," he said. "Maybe analysts are slowly getting the message!"
Distributed to subscribers for publication by
Scripps Howard News Service, http://www.scrippsnews.com
Publish A Letter in SitNews Read Letters/Opinions