This paper shows how to identify nonparametrically scalar stationary d
iffusions from discrete-time data. The local evolution of the diffusio
n is characterized by a drift and diffusion coefficient along with the
specification of boundary behavior. We recover this local evolution f
rom two objects that can be inferred directly from discrete-time data:
the stationary density and a conveniently chosen eigenvalue-eigenfunc
tion pair of the conditional expectation operator over a unit interval
of time. This construction also lends itself to a spectral characteri
zation of the over-identifying restrictions implied by a scalar diffus
ion model of a discrete-time Markov process. (C) 1998 Elsevier Science
S.A. All rights reserved.