Structural modeling enlightening for Crude Oil

“If you really want to revive the world economy, you’re going to have to deal with high oil prices; there’s no way around that,” said Lutz Killian of the University of Michigan in his April 22 lecture “The role of inventories and speculative trading in the global market for crude oil.”

Killian began by offering a brief background, explaining that different models for the global crude market already exist, and that they rest on two common perspectives. The first is that the price of oil is determined by shocks to flow supply of oil and flow demand and the second is that oil is an asset, and price shifts are based on speculations. Killian and his research choose instead to bridge the gap and to take a healthy mixture of these two as they created their own model.

In order to balance these perspectives, Killian’s team designed a model that provides a structural model of global market that uses a forward-looking speculation. It then allows the additional study of storage markets, in turn making it easier to solve the problem of unquantifiable expectations. Killian explains, “In principle, if you set up a structural model that jointly determines the real price of oil and the change in above ground inventories, you should be able to identify the effects of expectation shocks without ever having to parametrize what those expectations are. That is, you can write a model and tackle this problem without… the expectations, which we don’t have.”

Killian the analyzed the quantities of oil production, real activity, real oil prices, and inventories based on different shock scenarios. “So now the question is, how do you figure out which shock is which and how do you figure out what these shocks are doing to the real price of oil,” Kilian continued. To identify shocks, he established sign conventions for each scenario. In a flow supply shock, change in oil production and real activity should be negative, while change in real oil prices should be positive and inventory change should be unclear. In a flow demand shock, change in oil production, real activity, and real oil price should all be positive and inventories should be again unclear. In speculative demand shock, change in oil production, real oil price, and inventories should be positive, whereas change in real activity should be negative.
Though effective, these sign restrictions were not enough to create a complete model and Killian chose not to remain agnostic on which model variations were most accurate. He selected the fourteen most admissible models and used them to ultimately create a median to finalize his model.

With his model constructed, Killian could analyze data and offer new hypotheses for both controversial historical events and controversial recent events. He suggested that the economic events of 1990 following Saddam Hussein’s invasion of Kuwait were caused by a combination of overlapping factors made more confusing by the fact that many were cancelling each other out. Similarly, the price increase after supply shock during the 1978-1979 Iranian revolution could then be traced to the increase in speculative demand shocks. And he also suggested that the collapse of OPEC in 1986 was caused in part by overlapping shocks as well.

After illustrating that the model adequately aligned with established economic facts and even helped explain economic mysteries, he applied it to more recent events, such as the oil price increase between 2003 and 2008. He found no evidence that the increase was caused by oil trader speculation, OPEC, or any variation of peak oil, leaving only one option according to Killian. “What’s left is quite simply that we were surprised by how much the world economy was booming during this period.” This is possible because “Between ’03 and ’08, [experts] systematically underestimated economic growth… month after month they got it wrong.”

Based on the findings from his research, Killian concluded his talk by offering three key policy implications. First, the increase in action of traders is irrelevant to oil prices. Second, increased domestic oil production is “essentially irrelevant to the global pricepoint.” And third, economic recovery will require raised oil prices.