RiskMonitor (noted RM below) is the result of a significant investment into both the conceptual aspects of country risks and new quantitative instruments able to analyse very large sets of data through non-linear relations. This version fully capitalises on the successes and performances of our previous tool, while taking on board both the new aspects of country risks that emerged during the early 2000s and the new fundamental research in quantitative data mining techniques.

RiskMonitor is based on the following principles:

• Risk is defined as a non-linear output of economic and financial circumstances, where combinations of circumstances and threshold effects play a major role. In RiskMonitor, such circumstances are described through a succession of 6 Fundamental Balances, each of them providing one aspect of the country risk by crossing two relevant economic or financial variables. The six Balances are dealing with Growth, Debt, Liquidity, Foreign Exchange, Cyclical developments and the Banking System.

• A clear distinction is made between Signals of upcoming crises (early warnings of economic and financial shocks of such a magnitude and intensity that they can derail even the best structured deal or contract), and Country Economic Ratings (comprehensive assessment of the economic and financial qualities of a developing country)

• Results are provided for three different time-horizons (less than 1 year, 1 to 3 years, 3 to 5 years)

• Results are also differentiated for three different types of crises or economic / financial difficulties (external solvency, exchange rate depreciation, economic activity reversal)

• Statistical results of RiskMonitor are impressive enough to give a very strong confidence in the Signals provided for upcoming crises. For the Country Economic Ratings, the normative approach combines the results on each of the 6 Fundamental Balances with all quantitative parameters optimised through the application of genetic algorithms. For Signals of upcoming crises, RiskMonitor applies five different non-parametric models to the 12 economic or financial variables used in the Fundamental Balances. The methodology provides a Crisis Signal, when at least 4 of the 5 quantitative models used in the methodology provide a convergent signal for a massive shock to come; for such Crisis Signals, the signal-to-noise ratio is 100% (there has never been a Signal without a Crisis over the forecasted horizon), but some crises will occur without a Signal being raised (the sensitivity, or coverage, is lower than the signal-to-noise performance). The methodology also gives an Indication of Vigilance when 3 of the 5 models are giving a convergent result: the signal-to-noise ratio is somewhat lower (some Signals will prove untrue), but the sensitivity is much higher, meaning that very few shocks will be missed. The following tables provide the statistical results of RM in the 1980-2002 backtesting period for the whole sample of 72 countries (in- and out-of-estimation sample).

• Results and outputs are provided in a fully transparent manner, and through understandable and easy-to-use products, despite the technical sophistication of the tools used in the methodology

• Finally, RiskMonitor is a service combining quantitative results obtained from such a sophisticated numerical system (using only available statistics as opposed to potentially biased forecasts or estimates) and qualitative analyses (derived from our long-standing experience of economic analyses of developing countries, and using all available information).

A comprehensive document on RiskMonitor methodology is available. An academic article on our methodology is available on Elsevier website. If you wish to have more details or more precise information, please contact us at riskmonitor@tac-financial.com.