Forecasting Transational Terrorism and Other Series with an Unknown Number of Structural Breaks

Walter Enders
Ruxandra Prodan
Yu Liu, University of Texas at El Paso


Transnational terrorism data is difficult to forecast because it contains an unknown number of structural breaks of unknown functional form. The rise of religious fundamentalism, the demise of the Soviet Union, and the rise of al Qaeda have changed the nature of transnational terrorism. 'Old School' forecasting methods simply smooth or difference the data. 'New School' methods use estimated break dates to control for regime shifts when forecasting. We compare the various forecasting methods using a Monte Carlo study with data containing different types of breaks. The study’s results are used to forecast various types of transnational terrorist incidents.