We examine the ability of intergroup contact to ameliorate the effect of in-group bias on economic outcomes. Specifically, we employ randomized experiments to test whether actual and imagined contact is effective in reducing prejudice between indigenous Malawian shopkeepers (in-group), and their Chinese migrant counterparts (out-group), and test the stability of these changes over time. We find differing results with actual contact. Local Malawians’ attitude towards Chinese migrants did not improve, but their willingness to spend time with them did. In contrast, actual contact led to improvement in the Chinese migrants’ attitude toward local Malawians, but did not increase their willingness to spend time with them. These effects persisted over a time period of at least ten days. Imagined contact had no impact on Malawians’ attitude towards or willingness to spend time with Chinese migrants.