This paper by Baxter and Kouparitsas is an ambitious attempt to explore which variables are robust in explaining the correlations of bilateral GDP between countries at business cycle frequencies. Most of the variables turned out to be fragile. The main contribution is to show that countries with large amounts of bilateral trade tend to have robustly higher business cycle correlations. Another interesting finding is that neither currency unions nor industrial structure are robustly related to business cycle correlations.
Published in: _Journal of Monetary Economics_ (Vol. 52, No. 1, January 2005, pp. 159-162) https://doi.org/10.1016/j.jmoneco.2004.10.001.