Simulated Score Methods and Indirect Inference for Continuous-time Models

摘要: Publisher Summary This chapter describes a simulated method of moments estimator that is implemented by choosing the vector valued moment function to be expectation under structural model score an auxiliary model, where parameters are eliminated replacing them with their quasi-maximum likelihood estimates. leaves depending only on model. Structural parameter estimates those values put as closely zero possible in suitable generalized metric. methodology can also interpreted practical computational strategy for implementing indirect inference. One argues considerations from statistical science dictate should approximate true data-generating process and show using seminonparametric one means this end. When view close approximation accepted implementation, described here usually referred Efficient Method Moments literature because asymptotically efficient maximum correct specification detection error assured incorrect specification.