Heterogeneity in regional economic models: disaggregation and integration

regional economic models;input-output

Kim, Kijin

关键词: regional economic models;    input-output;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/92692/KIM-DISSERTATION-2016.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Heterogeneity in economic agents has long been one of the major discussion points in both theoretical and applied fields of economics. Particularly for regional economists focusing on macroeconomic models at the subnational level, it is of foremost importance to take into account heterogeneity of agents in regional economic models, generally used for the mid- and long-term forecasting and policy evaluation, considering the effects of time-varying socio-economic characteristics such as aging population and income inequality on simulations. Despite recent methodological advances in econometrics, modelling heterogeneity of agents for smaller regions is still a major challenge to overcome mainly due to lack of rich data. This dissertation focuses on (1) developing regional econometric models disaggregated by age and income of agents and (2) integrating the disaggregated models to existing regional macroeconomic models. Although the most typical regional models, econometric/input-output models, are illustrated as the examples of regional macro models, it is expected that the proposed integration strategies can be extended without difficulty to other modelling framework as well.In Chapter 1, I proposes an extension to the regional econometric input–output model (REIM; Conway, 1990; Israilevich et al., 1997) to which a demand system with age and income parameters is integrated. The extended model addresses concerns about household heterogeneity that has been limited to one representative in the existing REIMs. The initial testing is conducted with a model for the Chicago metropolitan area. First, using aggregate expenditure data by income and age groups, the almost ideal demand system (AIDS) with group fixed effects is constructed. Next, the estimated demand system is linked to the REIM to reflect long-term changes in the age and income distribution of households. The long-range simulation from the extended model takes into account structural changes in expenditure type stemming from changing demographic composition. The extended model further broadens the scope of impact analysis under various scenarios associated with age and income changes. In Chapter 2, drawing on a modified regional econometric input-output model (REIM) for the Chicago metropolitan region in which households were disaggregated by age, I provide an assessment of the differences generated by consumption of a representative and disaggregated households using data at the corresponding level of aggregation. The results reveal that the total effects of disaggregation that can be ascribed to population aging vary by a much smaller extent than those generated by model specification and data. The disaggregate REIM with heterogeneous households by age yields smaller RMSEs than the aggregate REIM with a representative household, but a statistical testing suggests that forecasting gains from disaggregation are modest compared to the aggregate model. In Chapter 3, I integrate an age-group-specific labor demand model into a regional input-output model to evaluate the effects of changes in age structure on a regional economy. The new integrated model suggests that ceteris paribus aging population attributes to lowering aggregate economic multipliers due to the rapidly growing number of elderly workers who earn less than younger workers. For the age-group-specific labor demand model, I find that a static labor demand model restricted with theoretical requirements yields empirically coherent wage elasticities of labor demand when the recent Census data are used. A Bayesian approach is used for more straightforward imposition of regularity conditions. The Bayesian model confirms elastic labor demand for youth workers, which is consistent with what past studies find. Comparison with other conventional methods suggests that among the regularity conditions that a cost function should satisfy, monotonicity and concavity must be checked and addressed particularly in the case where one or more factor shares are so small that monotonicity is likely to be violated.