Title: Solving asset pricing models Via nonparametric two-stage penalized B-spline regression
Authors: Liyuan Cui - City University of Hong Kong (Hong Kong) [presenting]
Yongmiao Hong - Cornell University (United States)
yingxing li - Xiamen university (China)
Abstract: A nonparametric 2SLS Penalized B-spline regression method is presented for unknown policy functions in an exchange economy, which allows the true dynamics of state variables to determine asset prices. Unlike current numerical solution methods, this new method does not require imposing auxiliary assumptions on the conditional distributions or matching the mean and variance of state variables, which enables real state dynamics to determine equilibrium equity prices. We propose a fast generalized cross-validation method to determine the optimal penalization in the 2SLS B-splines regressions, which ensures efficient and consistent estimation of the policy function for a broad class of stationary Markov state variables. The newly proposed regression method will become a pivotal approach for obtaining a consistent estimation of the price-dividend ratio function and equity prices in the presence of misspecified state variables or those with unknown dynamics.