Fixing Fundamental Flaws in Probabilistic Country Risk Models

Bertrand Groslambert

This chapter will demonstrate important flaws in the standard models and approaches used for country risk assessment and management that are based on assumptions of normal distributions. Based first on theory and then on an empirical review of data on market returns and foreign exchange prices, we propose an alternate and superior probabilistic approach as a basis for country risk management.

Risk can take different forms and can range from mild to wild randomness. Mild risk is when deviations from normal are typically small. In comparison, wild risk is characterised by extreme events, abrupt changes and large fluctuations. These types of risk have very different statistical characteristics, and it is therefore essential to first understand the true nature of country risk if one wants to try to manage it relying on statistical approaches.

The chapter is structured as follows. The first section will review the foundations of probabilistic risk assessment, briefly exploring the history of probability and presenting the origin of the normal law as a natural candidate for risk assessment. The second section will highlight the flaws of the normal law and explain how misleading it