Nonparametric estimation of the random coefficients model through regularized maximum likelihood,

Nonparametric Estimation of the Random Coefficients Model through Regularized Maximum Likelihood (RMLE) is explored. Chapter 1 establishes the underlying theory of the method and convergence rates. Simulation results are presented comparing the RMLE method with a kernel-based estimator. An application to real data is also presented. The conclusion reached is that the RMLE method is more robust with respect to tail behaviour of the design density. Chapter 2 builds on the method by expanding it to a threedimensional application. An open-source software package is also produced through Python. Chapter 3 discusses a case study of the housing market in Christchurch, New Zealand. The hedonic house price model is used to examine the changes in consumer preferences across three different time periods. The RMLE method is applied to model unobserved heterogeneity in the aforementioned consumer preferences through the inclusion of random coefficients in the hedonic model