1 MODELLING STOCHASTIC POLITICAL RISK FOR CAPITAL BUDGETING Ephraim Clark Radu Tunaru Middlesex University, London Metropolitan University Business School, London Economics, Finance and International Business Department, London Running title: Modeling stochastic political risk Abstract In this paper we model political risk for international capital budgeting as the value of a hypothetical insurance policy that pays the holder any and all losses arising from political events. We address three important aspects of political risk that are widely acknowledged in the literature but either missing or incomplete in existing mathematical models: 1) loss causing political events arise from a wide range of sources, which are often mutually dependent; 2) the effect of a political event in terms of actual losses can vary depending on the economic, social and political conditions when it occurs; 3) the composition and the importance of the individual sources of political risk can change over time. Thus, the multivariate nature and dependency of loss causing political events are modeled as a conditional Poisson process that allows for dependency between the increments of the counting process. To account for the random effect of a loss causing political event due to the evolution of economic, social and political conditions, we model the overall economic, social and political climate as a stochastic loss index that represents the expected size of the jump if a political event occurs. To account for changes in the composition and importance of the sources of political risk, we employ a Bayesian updating process whereby the distribution of the conditional Poisson process is updated in time as new information arrives. We then put these elements together and follow Clark (1997) to measure the total cost of political risk as the value of a hypothetical insurance policy that pays any and all losses generated by a political event. Finally, we show how the model can be implemented in practice. Key words: Capital budgeting, real options, conditional Poisson process, geometric Brownian motion, Bayesian updating JEL Classification: G31, D81, F21