Forecasting cyclical turning points by means of a probabilistic approach: Some South African evidence

In this paper the effectiveness of ‘transplanting’ a forecasting method based on a probabilistic approach is assessed in the South African context. The method utilises leading indicators in regression models, with a dichotomous response variable which assumes values of 0 or 1 to indicate business cycle expansion or contraction. The study closely replicates the work of Nazmi on turning point prediction. The results indicate an ability of the model to accurately forecast business cycle turning points in the 1980s. In the 1990s, the model displays a diminished capacity to forecast the turning points with acceptable accuracy. Leading indicators, in the South African experience, show reliable leading relationships with the composite coincident index over the period between 1970 and 1980 and thereafter this relationship weakens, impacting negatively on their forecasting ability.