Supplemental Appendix for “nonparametric Estimation in Random Coefficients Binary Choice Models

As is often the case when Fourier series techniques are used, we consider spaces of complex valued functions. Let Lp(S1) denote the Banach space of Lebesgue p-integrable functions and its norm by ‖ · ‖p. In the case of L2(S1), the norm is derived from the hermitian product ∫ 2π 0 f(θ)g(θ)dθ. Let Rθ and fφ denote the extension R of r according to (3.2) and fβ after the reparameterization. Our task is then to obtain fφ from the knowledge of Rθ. Rewrite (1.3) using these definitions, then divide both