Quantification of Political Risk with Multiple Dependent Sources
Clark, Ephraim, Tunaru, Radu, Journal of Economics and Finance
In this paper, we develop a model using a conditional Poisson process for measuring the effect of a countable number of mutually dependant political risks on the outcome of foreign direct investment. We also apply a Bayesian updating process that makes it possible to re-estimate the model's parameters as new information becomes available. We then show how the model can be operationalized and provide a comparative example related to foreign direct investment. (JEL G31, D81, F21).
(Proquest Information and Learning: ... denotes formula omitted.)
Long ago when "all politics was local," political considerations figured prominently in project evaluation and capital budgeting decisions. Globalization has only increased its profile, complexity, and scope. Witness the Russian default, the Mexican peso crisis, and the Asian economic meltdown at the macro level or the constant stream of newspaper reports of bribes, strikes, taxes, decrees, political violence, asset destruction or expropriation, and the refusal to respect or enforce contracts at the micro level.
The nature of political risk is such that it is random and its sources are many and varied. These multiple sources are also dependent upon each other. For example, a tax increase can cause riots or a strike that hurts the company can cause the government to issue a decree that satisfies labor's demands but that also hurts the company. Similar relationships to one degree or another exist between most political variables. Another characteristic of political variables is that the rates at which they occur can change over time as a result of changes in the domestic and international economic, financial, and political environment. As an example of this, it is interesting to note that the number of expropriations, the most dramatic form of political risk, fell from 336 between 1970 and 1975 to 87 between 1976 and 1979 to 15 between 1980 and 1985 and, finally, to one between 1986 and 1992 (Minor 1994).
Political risk analysis has evolved along with the demands of globalization and the new "real option" capital budgeting techniques developed to evaluate managerial opportunities and responses to adversity. These option-pricing techniques used in the evaluation of financial assets have become widespread in the theory and practice of corporate capital budgeting, thereby making it possible to evaluate the effects of managerial flexibility, the so-called "real options" that are an integral part of most projects.1 They have also been applied to the evaluation of political risk that is usually associated with foreign direct investment (FDI) but which is also present in traditional domestic investments.2 Mahajan (1990), for example, models expropriation as a European style call option on an asset with a given maturity that pays no dividends, and Clark (2003) models it as an American style call option on a dividend paying asset with an indeterminate maturity and a stochastic exercise price.
Although far more sophisticated than the traditional models of Stonehill and Nathanson (1968), Stobaugh (1969), Robock (1971), and Shapiro (1978) that make a simple, ad hoc adjustment to cash flows or to the discount rate, the approach in these papers is limited in that it does not recognize the exogenous, random nature of many types of political risk. Pointon and Hooper (1995) address this issue in the particular case of expropriation, which they model as an independent Poisson process, in the presence of exchange rate risk. Clark (1997) also addresses this issue and models loss-causing political events as an independent Poisson process. Loss-causing political events take the form of legislation or decrees such as expropriations, nationalizations, devaluations, etc., or the form of direct actions such as strikes, boycotts, terrorist acts, etc. The nature of explicit events is that they arrive intermittently at discrete intervals and that they generate an actual loss or series of losses. …