A huge worldwide wave of cross-border bank acquisitions occurred in the 1990's. Europe and Latin America provided the most active game boards on which bank acquisitions played out, in terms of value and number of acquisitions, but countries in Asia, the Middle East and other regions also saw significant levels of bank sales to foreign buyers. Financial economics had traditionally analyzed bank acquisitions from an efficiency perspective, arguing that banks are driven to exploit the efficiency gains obtained through buying other banks. But I argue, with support from recent studies in banking finance, economics and international politics, that these waves of bank privatization and acquisition are driven largely by political dynamics at the national and international levels. The more controversial forms of FDI involve industries of strategic interest, of which banking is one. Banks are at the center of modern economies. Banks are monitored and governed by national and international organizations established for that purpose and they lobby these organizations. Banks are also "privately monitored" by investors and other interested parties. In short, banks and financial systems operate in a complex web of politics. Governmental actors are pulled and pushed toward banks and banks' power. They are pulled by banks' ability to deploy liquidity in productive ways, and pushed by the perceived need to regulate and monitor banking activity. I propose that cross-border bank acquisitions, driven by "closeness" between countries, could be examined through a multilevel tobit model examining the predilection for a bank in one country to buy a bank in another. I propose to modify and extend the "gravity equation" model from international economics to explain the broad variation in cross-border bank acquisitions. Through this model, I suggest that "closeness", measured in legal, political economic, cultural and spatial terms, promotes cross-border bank acquisition. Closeness is theorized to decrease transaction costs. This paper sketches a preliminary hierarchical linear tobit model utilizing longitudinal datasets of national, bilateral and international factors for OECD countries. ..PAT.-Unpublished Manuscript [ABSTRACT FROM AUTHOR]