A substantial literature argues that preferential trade agreements (PTAs) have an influence on trade flows. On the one hand, PTAs may increase trade flows between member countries and decrease trade flows between members and excluded countries. On the other hand, PTAs may reduce the volatility of trade flows between member countries. In this paper, we revisit this literature and argue that only looking at the presence or absence of a PTA overlooks important differences across PTAs in depth, scope and flexibility. Some agreements include far-reaching commitments to liberalize trade, while others are very shallow. And some agreements cover many policy instruments, while others only deal with tariffs on trade in goods. We hypothesize that deeper and broader agreements have a larger impact on trade flows than shallow and narrow ones. Moreover, agreements with strict dispute settlement mechanisms are more likely to reduce trade volatility between member countries than agreements with weak enforcement devices. A database on the design of more than 500 trade agreements that were signed between 1950 and 2009 allows us to test this hypothesis. The database covers substantially more PTAs in significantly more detail than any comparable database, giving us an unprecedented opportunity to revisit the PTA-trade nexus.
