1: \begin{abstract}

2: Let $\mainset \subseteq \R^n$ be a closed set, and consider the problem of computing the minimum $\funcmin$ of a polynomial $f$ on $\mainset$. Given a measure $\mu$ supported on $\mainset$, Lasserre (SIAM J. Optim. 21(3), 2011) proposes a decreasing sequence of upper bounds on $\funcmin$, each of which may be computed by solving a semidefinite program. When~$\mainset$ is compact, these bounds converge to $\funcmin$ under minor assumptions on $\mu$. Later, Lasserre (Math. Program. 190, 2020) introduces a related, but far more economical sequence of upper bounds which rely on the push-forward measure of $\mu$ by $f$. While these new bounds are weaker \emph{a priori}, they actually achieve similar asymptotic convergence rates on compact sets. In this work, we show that no such free lunch exists in the non-compact setting. While convergence of the standard bounds to $\funcmin$ is guaranteed when $\mainset = \R^n$ and $\mu$ is a Gaussian distribution, we prove that the bounds relying on the push-forward measure fail to converge to $\funcmin$ in that setting already for polynomials of degree~$6$. 

3: \end{abstract}

4: