Taking control of economic uncertainty with science
As Congress debates the pros and cons of an automaker bailout, researchers at Illinois have developed economic models and methods that could help economists better understand the potential outcome of such stimulus packages.
The techniques use engineering principles, especially in the area of control, to explain deviations in economic systems. While many current models can’t predict a market crash, the Illinois model gives economists the chance to see potential curveballs and to better understand how to control them.
“When you authorize a $700 billion stimulus package, for example, you’d like to know it’s going to work,” said ECE Illinois Professor Sean P Meyn, a resident researcher in the Coordinated Science Laboratory on the Illinois campus. “We want to give economists the most complete preview of how their actions might play out.”
Many models used today are static and can’t account for a dynamic, complex, and uncertain world. Working with Illinois economics professor In-Koo Cho, Meyn developed a model that captures two features that are found in typical markets, but that are often not considered in economic theory: (a) constraints on increasing supply to meet demand and (b) uncertainty.
In the case of the automotive industry, these factors arise at every level of the market--from local suppliers to the global market. For an individual manufacturer, such as Ford, it is not possible to instantly decrease production of energy-efficient vehicles (constraints) when the price of oil suddenly drops (uncertainty)--even if the demand for hybrid cars goes down. The company has already established assembly lines and stocked parts to build these cars and stands to lose millions if the cars aren’t made.
More critical issues arise at a higher level, such as the overall national automotive industry and its interaction with the national economy. The researchers’ model would take into account various control factors, including loans and their conditions (even examining whether the loans should go towards investments or to bonuses and incentives for creativity). The model would display possible outcomes and degrees of uncertainty.
Meyn and Cho applied some of their research, then in its infancy, to the 2000-2001 California power crisis that followed deregulation. They were able to explicitly model “friction” (in this case, the fact that power generation was subject to ramping constraints) and discovered that their model predicted highly volatile prices similar to those seen then in California, and in many markets around the world today.
“This gave me optimism that if we take the trouble to build models that capture constraints and uncertainty, then we might learn about potential pitfalls and how to avoid them, or minimize their impact,” Meyn said.
In addition to economics, the researchers’ work has applications in finance, wireless networks, bioinformatics, and manufacturing systems. Meyn’s research has also been applied to air traffic control in an effort to reduce delays and improve safety.