Some Empirical Facts About International Trade Flows Mark N. Harris Department of Econometrics and Business Statistics, Monash University László Kónya ∗ School of Economics and Finance, La Trobe University László Mátyás Department of Economics, Central European University December 2009 Abstract The last decade has seen a proliferation of empirical models in the trade literature. Focus has ranged from the effect of particular explanatory variables to improved econometric techniques. However, there appears to be a lack of analyses on large up-to-date international trade datasets aiming at describing the key features, or “stylized facts”, of observed bilateral trade flows. Uncovering them is crucial as any empirical econometric model should reflect the basic properties of the data generating process. On the basis of a new dataset we find that bilateral trade, despite being often unbalanced, tends to be reciprocal and persistent, and that the extensive margin of trade must not be disregarded. Moreover, bilateral trade flows are probably best modeled as a mixed panel of stationary and non-stationary processes. The stationary versus non-stationary separation of these flows, although not random, does not appear to be related to some evident common characteristics of the trading partners. JEL Classification Numbers: C29, F19, Key Words: international trade, bilateral trade data, gravity model ∗ Corresponding author; Email: l.konya@latrobe.edu.au; Address: School of Economics and Finance, La Trobe University, Victoria, 3086, Australia.