We study the determinants of new and repeated research collaborations, drawing on the co-authorship network of the International Monetary Fund (IMF)’s Working Papers series. Being an outlet where authors express their views on topics of interest, and given that IMF staff is not subject to the “publish-or-perish” conditions of the academia, the IMF Working Papers series constitutes an appropriate testing ground to examine the endogenous nature of co-authorship formation. We show that the co-authorship network is characterized by many authors with few direct co-authors, yet indirectly connected to each other through short co-authorship chains. We find that a shorter distance in the co-authorship network is key for starting research collaborations. Also, higher research productivity, being employed in the same department, and having citizenship of the same region help to start and repeat collaborations. Furthermore, authors with different co-authorship network sizes are more likely to collaborate, possibly reflecting synergies between senior and junior staff members.