Abstract
One of the main concerns associated with the development and use of regional CGE models is the determination of key parameter values, particularly substitution and other price elasticities. A common problem is the lack of appropriate regional data for econometric estimation. Consequently, it is important to identify key parameters that are likely to be important in determining quantitative results and then to prioritize these for estimation where appropriate data are available. In this paper, the focus is on the estimation of the regional trade (import) substitution parameters, which tend to be important in analysis for regional economies (given their openness to trade). Here, commodity import elasticities for the Illinois economy are estimated and tested in a single region CGE model of the Illinois economy. In our econometric estimation, we apply a model that takes account of market size and distance in estimating the substitutability between commodities produced in Illinois and other US states.
Acknowledgements
The econometric work reported here was initially carried out as part of Soo Jung Ha's doctoral research at the University of Illinois (under the supervision of Geoffrey Hewings, and in collaboration with Karen Turner on CGE model development). This paper has been further developed with the CGE modeling application under an ESRC Climate Change Leadership Fellowship (Grant reference RES-066-27-0029). Support from NSF grant 0818578 is also acknowledged. We are grateful to participants at the 48th Annual Meeting of the Western Regional Science Association conference, held in Napa Valley California (February 2009), and the annual conference of the Regional Science Association International: British and Irish Section, held in Glasgow, Scotland (August 2010) for comments on earlier versions of this paper.
Notes
1 More details on the derivation of market size and distance factors may be found in Erkel-Rousse and Mirza Citation(2002).
2 AMOS is an acronym for a macro-micro model of Scotland, the regional economy on which the CGE modeling software was initially calibrated (Harrigan et al., Citation1991).
3 In a fuller set of simulations (not reported here) we find that there is not a great deal of difference in CGE model results if we use the parameter estimates from Estimation Model 1 or 2.
4 Again, in simulations not reported here, this conclusion has been tested but not reported here due to the constraints of space.