Some economic and financial events over the recent months have pointed out the effect of uncertainty on financial market contagion. In this paper, we built on a theoretical work by Kodres and Pritsker (2002), who explain contagion among markets through a portfolio rebalancing effect, and we extend it by allowing for non-linearity driven by uncertainty. In the empirical part of the paper we estimate the TVAR model. We then calculate for each regime the spillover index of Diebold and Yilmaz (2009) extended to take into account for the switching parameters in low and high uncertainty regimes. We first show empirical evidence on non-linear effects of uncertainty on the degree of spillovers. Results also indicate that high uncertainty tends to generate more contagion among financial stock indexes of a set of advanced and emerging countries. Those results are robust to the uncertainty measures. A direct policy implication is that reducing any type of uncertainty appears to be a way to avoid rapid propagation of adverse shocks within financial systems.