We discuss a method for estimating the convergence speed of Anderson Acceleration (AA) applied to the Alternating Direction Method of Multipliers (ADMM), for the case where ADMM by itself converges linearly. It has been observed empirically that AA may speed up asymptotic ADMM convergence by a large amount for various applications in data science and machine learning, but there are no known convergence bounds for AA with finite window size that would allow quantification of this improvement in asymptotic convergence speed. We discuss an approach to quantify the improved linear asymptotic convergence factor that is based on computing the optimal asymptotic convergence factor for a stationary version of AA applied to ADMM. Numerical tests show that this allows us to estimate the asymptotic linear convergence factor of AA with finite window size applied to ADMM. The numerical results also explain how and by how much nonlinear acceleration methods like AA can accelerate ADMM, in terms of the spectral properties of the Jacobians of ADMM and stationary AA evaluated at the fixed point.