1: \begin{abstract}

2: We study the convergence rate of a  hierarchy of  upper bounds for polynomial minimization problems,

3: proposed by Lasserre [{\em SIAM J. Optim.} $21(3)$ $(2011)$, pp. $864-885$], for the special case when the feasible

4: set is the unit (hyper)sphere. The upper bound at level $r \in \mathbb{N}$ of the hierarchy is defined as the minimal expected value

5: of the polynomial over all probability distributions on the sphere, when the probability density function is a sum-of-squares polynomial of degree at most $2r$ with respect to the

6: surface measure.

7: 

8: We show that the exact rate of convergence is $\Theta(1/r^2)$, and explore the implications for

9: the related rate of convergence for the generalized problem of moments on the sphere.

10: \end{abstract}

11: