We contribute to the literature by exploring the variation in the trade-creating effect of migration networks across historical migrants and recent migrants by applying a micro-founded gravity model of trade. Both the stock of historical migrants and recent migrants variables are constructed by merging Canadian Census data with the migration data available in Statistics Canada. We address the endogeneity issue by applying IV estimators including two-step feasible GMM and PPML with IV and fixed effects by utilizing the imputed annual migration flow, the historical stock, and recent stock of migrants as instrumental variables proposed by Peri and Requena (2010). Taking Canadian geographical features into account, this study also applies carefully measured gravity variables. We control for the time-varying, exporter-year fixed effects, and importer-year fixed effects in a panel estimation. Estimated results show that annual migration, recent provincial stock of migrants, and historical stock of migrants significantly increase Canadian interprovincial trade. This impact is stronger for the stock of historical migrants relative to the stock of recent migrants and annual migration flow. Sub-sample analysis shows that both recent and historical migration consistently increase interprovincial service trade, only but not interprovincial goods trade. A greater combination of English and French-speaking provinces-pair significantly increases interprovincial trade. Regional trade agreements significantly impact the interprovincial trade of Canada. Our estimated results are robust to different estimation methods and alternative measures for the migrants’ stock and flow variables.