Simple closed-form estimation of a binary latent variable model
提出一种闭式非参数估计量，用于估计给定未观测潜变量时二元结果变量的条件分布函数，计算简单且无需迭代优化，适用于测量误差和动态模型。
Summary This paper develops a closed-form non-parametric estimator of the conditional distribution function for a binary outcome variable given an unobserved latent variable. This type of function is commonly used in models that involve measurement error and dynamic models with agent-specific unobserved heterogeneity. This paper presents a consistent extremum sieve estimator with the following advantages: (i) it has a closed-form expression for all the sieve coefficients, (ii) it is computationally straightforward, equivalent to computing eigenvalues and eigenvectors of matrices without the use of iterative optimization algorithms. While as flexible as the sieve maximum likelihood estimator previously proposed for this model, our estimator proves computationally simpler. The finite sample properties of the estimator are investigated through a Monte Carlo study, and the developed estimator is applied to a probit model to assess the targeting performance of a social welfare program.