Theoretical and empirical research in economics suggests that bilateral migration triggers bilateral trade through a number of channels. This paper assesses the functional form of the impact of migration on trade flows in a quasi-experimental setting. We provide evidence that the relationship is not log-linear. In particular, at small levels of immigration (stocks) the elasticity of trade to migration is quite high, and it declines to zero at about 4,000 immigrants. If immigration stocks exceed such a level, the evidence suggests that trade will not increase anymore. This suggests that for cross-country network and other effects flowing from immigration to materialise at a significant level for trade, a high-enough level of immigrant stocks is necessary. But there appears to be satiation as immigrant numbers increase.