1: \begin{abstract}

2:   We study the convergence of iterated random functions for stochastic

3:   feasibility in the consistent case (in the sense of Butnariu and

4:   Flåm \cite{ButnariuFlam1995}) in several different settings, under

5:   decreasingly restrictive regularity assumptions of the fixed point

6:   mappings.  The iterations are Markov chains and, for the purposes of

7:   this study, convergence is understood in very restrictive

8:   terms. We show that sufficient conditions for geometric (linear)

9:   convergence in expectation of stochastic projection algorithms

10:   presented in Nedi\'c \cite{Nedic2011}, are in fact {\em necessary}

11:   for geometric (linear) convergence in expectation more generally of iterated

12:   random functions.

13: \end{abstract}

14: