Abstract We study the asymptotic behavior of Markov operators P_ P μ defined by convolution with a probability measure μ on the unit circle T T. We prove that when μ is adapted, P_ P μ satisfies Doeblin’s condition if and only if some power ᵏ μ k is non-singular. We give an example of a symmetric probability measure μ on T T, such that the reversible stationary chain induced by P_ P μ is ρ -mixing, but P_ P μ does not satisfy Doeblin’s condition. We look at the spectra of P_ P μ in the different Lₚ L p spaces when P_ P μ is, or is not, ρ -mixing.
Cohen et al. (Thu,) studied this question.