Methods of statistical diagnosis in economic and financial systems

Abstract

In 1980–1990s methods of statistical diagnosis have found increasing interest and generated new research trends in the analysis of economic and financial processes and systems. One of the most important fields is the analysis of non-stationary econometric models with ‘structural breaks’. The term ‘structural break’ refers to the situation of an abrupt change in the set of regression coefficients at some a priori unknown moment of observations. Traditional methods of regression analysis turn out to be very unstable to the influence of structural breaks and shifts caused by exogenous and endogenous ‘shocks’. The characteristic example of an exogenous ‘shock’ in economic systems is the ‘Oil crisis’ of 1970s provoked by the cartel agreement of OPEC countries. This crisis caused the sharp growth of oil prices in the West and precipitated the increase of costs and the drop of profits of the largest companies with such negative macroeconomic consequences as the sharp rise of inflation and the fall of the aggregate output and consumption in these countries.