Driving Regional Economic Models with a Statistical Model: Hypothesis Testing for Economic Impact Analysis

Policy models such as Input-Output (IO) or Computable General Equilibrium (CGE) are deterministic, with exogenous final demand shocks producing point estimates of local impacts. Confidence intervals around these point estimates, while desirable, are not readily available. Yet such predictions often have an underlying stochastic element, which can be derived from the underlying statistical model or sample averages used to compute the exogenous shocks themselves. Using the causal statistical model to form confidence intervals around the input/shock estimates allows for the configuration of confidence intervals around the output/impact results. The method is demonstrated on a sample policy scenario, which tests the relative significance of population versus climate change as a determinant of local economic activity in Rocky Mountain National Park’s gateway community. *Corresponding author: Stephan Weiler, CSU Economics Dept, Fort Collins, CO 80523-1771 e-mail: [email protected], Phone: 970-491-3883, Fax: 970-491-2925 Driving Regional Economic Models with a Statistical Model: Hypothesis Testing for Economic Impact Analysis ABSTRACT: Policy models such as Input-Output (IO) or Computable General Equilibrium (CGE) are deterministic, with exogenous final demand shocks producing point estimates of local impacts. Confidence intervals around these point estimates, while desirable, are not readily available. Yet such predictions often have an underlying stochastic element, which can be derived from the underlying statistical model or sample averages used to compute the exogenous shocks themselves. Using the causal statistical model to form confidence intervals around the input/shock estimates allows for the configuration of confidence intervals around the output/impact results. The method is demonstrated on a sample policy scenario, which tests the relative significance of population versus climate change as a determinant of local economic activity in Rocky Mountain National Park’s gateway community. Policy models such as Input-Output (IO) or Computable General Equilibrium (CGE) are deterministic, with exogenous final demand shocks producing point estimates of local impacts. Confidence intervals around these point estimates, while desirable, are not readily available. Yet such predictions often have an underlying stochastic element, which can be derived from the underlying statistical model or sample averages used to compute the exogenous shocks themselves. Using the causal statistical model to form confidence intervals around the input/shock estimates allows for the configuration of confidence intervals around the output/impact results. The method is demonstrated on a sample policy scenario, which tests the relative significance of population versus climate change as a determinant of local economic activity in Rocky Mountain National Park’s gateway community. The authors would like to thank Dave Theobald, Dennis Ojima, and Tom Hobbs for their valuable contributions. We gratefully acknowledge an EPA-STAR Grant and Colorado State University Ag Extension Station Project W-133 for partial funding of this research. We would also like to thank the participants in the 2000 National IMPLAN conference for valuable discussion that helped us