Some of the earliest work in computational economics occurred a half century ago when the first efforts were made to use statistical methods to fit supply and demand curves to data from agricultural markets and fit income and expenditure data to consumption functions. Slowly, but surely, computational methods began to play a pivotal role in some fields of economics to the point that we cannot now imagine what it would be like to do analysis in these areas completely without computers, viz. macroeconomic or environmental policy analysis. At the same time computational methods have spread across the broad front of economics to the point that there is now almost no area of economic research that is not deeply affected.
This means that there is now a wide array of opportunities to push forward our understanding of economics using computational methods. These opportunities include the estimation and optimization of dynamic models with uncertainty. This kind of analysis helps us to better understand portfolio/savings decisions, patent renewal, schooling choice and retirement behavior. The opportunities also extend to models of the economics of the environment which could help to support negotiations among countries about carbon emissions. We have the opportunity to gain better understanding of the functioning of financial markets like the Chicago Board of Trade through the use of artificial intelligence methods which model high speed electronic trading systems. Models can be developed to study a single industry on a worldwide basis and thus to provide better understanding of international trade not only in final commodities, such as automobiles and computers, but also in their components such as engines, transmissions, disk drives and displays. Computable general equilibrium models can help to analyze trade agreements like the U.S.-Canada agreement which was negotiated some years ago and the Mexico-U.S.-Canada agreement which is upcoming.
We do not recommend a single large project to push forward research in computational economics. Rather we think that a number of small projects in a broad front of opportunities should be supported at a cost of about 18 million dollars per year. A part of these funds should be used to facilitate a shift to the workstation/network mode of computational research. This will permit economists to exploit the high speed computational networks which NSF is continuing to develop and to create servers for economic data and models which can be rapidly accessed from anywhere in the country.