Expert judgments of political riskiness

Professional analysts' judgments of the political riskiness of 49 focal countries for the period 1983‐1985 were studied. Data were collected on nine predictor variables; each was significantly correlated at the 0.01 level with ratings of political riskiness. The highest correlation was with infant mortality and life expectancy; either accounted for roughly 50% of the variance in ratings. Different variables were better predictors of political risk within different geographic regions. A factor analysis suggested the presence of three underlying factors. The predictor variable with the highest loading was chosen to represent each of the three factors. These were: exchange rate differential; estimated inflation rate; and infant mortality rate. Approximately 75% of the variance in ratings could be accounted for on the basis of a linear combination of the three predictor variables. These three variables were capable of good prediction even for various subsets of countries based on geographic region or other criteria. Using all nine variables as predictors resulted in only marginal improvement. A cluster analysis revealed little difference among clusters of judges. Ratings by undergraduate students closely paralled those of professional analysts. As in previous studies of expert predictions and forecasts, claims of expertise in political risk analysis were better supported by command of factual knowledge than by differentially superior predictive ability.