Applications of Frequentist Model Averaging in Economics and Operations Research
DescriptionIn Statistics, frequentist model averaging (FMA) is receiving a surge of attention. While many data-driven model weighting methods with optimal properties have evolved for FMA, all methods have been developed for models having the same structure but containing different covariates. There are instances where the concern is not about which covariates to enter the model, but rather which functional form or distribution to be specified. These features necessitate changes to the established approaches. This proposed project attempts to take the FMA literature a step further from its current level that emphasises "general" setups to one that focuses on 'local" analysis. We consider estimation problems stemming from the analysis of stochastic frontier and income inequality in Microeconometrics, and stochastic simulation in Operations Research, where the choice of an appropriate functional form or distribution for the data is the primary concern. Our objective is to develop an FMA strategy with optimal properties for estimating the parameter of interest in each case.
|Effective start/end date||1/09/17 → …|