ESTIMATION OF PARAMETERS OF ECONOMIC GROWTH AUTOREGRESSIVE MODELS
The construction of time series models based on statistical data is one of the central tasks of modern econometric studies. Numerous economic models based on differential equations are used to explain the patterns of economic growth. The article is devoted to the estimation of parameters of economic growth models based on solutions of homogeneous differential equations with constant coefficients. A comparative analysis of methods for estimating the parameters of autoregression of economic dynamics series with additive interference in the output signal is carried out. The simulation results showed that the full least squares method gives the most accurate estimates. The most commonly used least squares method gives the worst estimates.
Keywords: economic growth, autoregression, parametric identification, least squares method, instrumental variables, Yule-Walker equations, trend, seasonality.
Highlights:
♦ it is proposed to use autoregression models with an additive stochastic component in the output signal to estimate the parameters of economic growth;
♦ methods for estimating autoregression parameters with an additive stochastic component in the output signal are modeled;
♦ It is shown that the application of the full least squares method gives the most accurate estimates of the parameters of economic growth models.
Dmitry V. Ivanov, Candidate of Physics and Mathematics, Associate Professor of Samara State University of Economics.