Econometric and Forecasting Models

When predicting the effect of a change in external conditions on a complex system, the most appropriate approach is usually an econometric model. Econometric models rely on data on how the system has previously responded to changes in external conditions. These models do not explicitly illustrate how the system rearranges itself to accomodate the change, but this is unnecessary because the data from which they are estimated already incorporate the internal response of the system to past changes in external conditions.

ECONorthwest has a wealth of experience developing regional and statewide econometric forecasting models. Our staff has also incorporated econometric models (with feedback) into larger integrated models that had land use and transportation components. We have refined methods for producing the best possible forecasts given limited data, including the use of Bayesian vector autoregression for time-series models and the use of a neural network for selecting the best combination of variables for each equation.

In addition to developing economic forecasting models, we have successfully applied econometric modeling techniques to markets for particular goods and services.

Representative Projects

  • Semiannual Economic and Revenue Forecasts (Excerpt, PDF, 127K)
    TriMet
    Oregon
  • Regional Economic and Tax Base Forecasts
    Puget Sound Regional Council
    Washington
  • Economic Modeling Components of the Transportation and Land Use Model Integration Program
    Oregon Department of Transportation
    Oregon
  • Toll Optimization Modeling
    ECONorthwest's proprietary Toll Optimization Model, developed by Randall Pozdena and Tom Light, has been used in many projects to determine optimal tolls and projected revenue levels.