Kernel Smoothing Estimation of GDP Growth and Inflation Rate in Finland: A Nonparametric Approach
Author: Junjie Sun
Working paper, Department of Economics, Iowa State University, April 2006
Abstract:
This paper conducts a set of kernel smoothing estimations to study the behavior and relationship of the growth rate of Gross Domestic Product (GDP) per capita and the inflation rate in Finland (1950-2000). Specifically, this study first carries out a kernel density estimation (KDE) to estimate the density functions of the growth rate of GDP per capita and inflation rate in Finland. Then the kernel regression is used to estimate the potential relationships between these two variables by Nadaraya-Watson (NW) estimator and local linear estimator. Next, we estimate the conditional variance function and plot a 95% variability band. Finally, we propose parametric models for the distribution of two variables and the regression function, and construct bootstrap based goodness-of-fit (GOF) tests to compare the nonparametric and parametric model fits. The result of GOF tests suggest that the quadratic parametric fit might be adequate to capture the relationship between GDP growth rate and inflation rate in Finland. Although this study concentrates on real GDP per capita and inflation in Finland, the methods used in this paper can easily be extended to other data series in other countries.
Keywords: GDP growth, Inflation rate, Kernel density estimator, Kernel regression, Nadaraya-Watson (NW) estimator, Cross validation, Goodness-of-fit test, Bootstrap.
JEL Codes: C14, E2, E3