Terrorist and insurgent groups sometimes give pre-attack warnings, informing governments of the time and place of attacks before they occur. This dissertation explains why militant groups give these warnings. It also explains why governments believe these warnings and respond to them, mobilizing emergency resources and carrying out economically disruptive evacuations. Based on interviews and other historical research on the Irish Republican Army (IRA), Euskadi Ta Askatasuna (ETA), the Tamil Tigers, Shining Path, and Túpac Amaru Revolutionary Movement (MRTA), this dissertation argues that pre-attack warnings serve a casualty-limiting function. Militant groups give warnings when civilian casualties are politically costly for the group. Civilian casualties are especially costly for groups that depend on local populations for shelter, funding and other critical resources. These conclusions are confirmed by logit analyses of a new database of more than 3,000 bombing events. A game theoretic signaling model also predicts when governments will believe and respond to warnings. Governments respond to warnings when militants are known to warn only when attacking and the frequency of prank warnings is low. The model's predictions are confirmed by interviews of police in Northern Ireland and Spain. A novel finding is that a high frequency of pranks (false warnings emanating from individuals outside the militant group) may force militants to warn truthfully. Militants may also work with governments to create clear channels for communication, using third party intermediaries, codes, and redundant messages to set militants' warnings apart from the ``noise'' of pranks. This finding substantiates a game theoretic prediction that experimental methods have so far failed to validate: that increased noise may induce separating equilibria, increasing rather than decreasing the information in a signal.
Identifer | oai:union.ndltd.org:columbia.edu/oai:academiccommons.columbia.edu:10.7916/D8JD4W3R |
Date | January 2015 |
Creators | Brown, Joseph Matthew |
Source Sets | Columbia University |
Language | English |
Detected Language | English |
Type | Theses |
Page generated in 0.0022 seconds