Forecasting economic development taking into account several turning points: Indicators, model calibration, simulation computations

Abstract

The article considers the possibility of constructing forecasts for macroeconomic indicators taking into account the turning points in their dynamics trends. The calculations are performed on the base of a discrete approximation for the constraints of a simple economic growth stochastic model using the Monte Carlo method. In the first part, the authors analyze the problems of justifying turning point indicators and show that there is no single approach to their definition. Changes in GDP, oil prices and other indicators often serve as such indicators. In the second part, the authors propose to relate turning points to a change in the value of one of the numerical parameters of the growth model under consideration — the capital depreciation rate. To determine the parameters of the model, a special calibration procedure is proposed, based on solving the optimization problem according to the criterion of the minimum discrepancy between the average calculated and actual trajectories of GDP and Consumption over the calibration period. In the third part, experimental simulations are performed taking into account turning points according to the data of the economies of Finland, Cyprus and Japan. Three turning points are allocated for Cyprus and Japan, and one for Finland. Forecasts of the GDP and Consumption dynamics for these countries at current and constant prices of 2010 are constructed. For all three countries under consideration, the results of simulations show that indirect accounting of turning points by amendment of the capital depreciation rate allows significantly improving the quality of forecasts based on the average calculated trajectory, taking into account the specified confidence interval for the selected forecast period.