跳到內容
描述
By: Jean-Francois Richard, Robert C. Marshall, Chaohai Shen
An operational simulation-based implementation of a fundamental research principle in the sciences—the idea that a model must account for, or encompass, findings of alternative models.
To account for the universal deficiency of all economic models, Jean-François Richard, Robert Marshall, and Chaohai Shen propose in this book an operational test of whether a model accounts for—or encompasses—key results of alternative models. This simulation-based approach is distinct from the likelihood-based approach of forty-plus years ago. It is specifically designed to be applicable to serious nonlinear applications, under the minimal requirement that the models under consideration be amenable to Monte Carlo simulations.
The method of simulated moments, increasingly applied in major empirical applications, provides the ideal framework to compute a wide range of formal simulation-based encompassing test statistics. The authors argue that simulation-based encompassing is particularly easy to implement when the model of interest incorporates novel features relative to an earlier “benchmark” model: that is, “nests” the latter.
The book should be of particular interest to practitioners, researchers, and graduate students who wish to validate their models using simulation-based encompassing as a rigorous scientific alternative to more informal model comparisons.
An operational simulation-based implementation of a fundamental research principle in the sciences—the idea that a model must account for, or encompass, findings of alternative models.
To account for the universal deficiency of all economic models, Jean-François Richard, Robert Marshall, and Chaohai Shen propose in this book an operational test of whether a model accounts for—or encompasses—key results of alternative models. This simulation-based approach is distinct from the likelihood-based approach of forty-plus years ago. It is specifically designed to be applicable to serious nonlinear applications, under the minimal requirement that the models under consideration be amenable to Monte Carlo simulations.
The method of simulated moments, increasingly applied in major empirical applications, provides the ideal framework to compute a wide range of formal simulation-based encompassing test statistics. The authors argue that simulation-based encompassing is particularly easy to implement when the model of interest incorporates novel features relative to an earlier “benchmark” model: that is, “nests” the latter.
The book should be of particular interest to practitioners, researchers, and graduate students who wish to validate their models using simulation-based encompassing as a rigorous scientific alternative to more informal model comparisons.
