Eigenvalue Decomposition of Time Series with Application to the Czech Business Cycle

We follow a Beveridge-Nelson like time series decomposition method (into trend, business cycle and irregular components), and examine a stylized model of price inflation determination using the Czech data. We characterize the estimated components of CPI, IPPI and import inflations, together with the real production wage and real output, and survey their basic correlation properties; furthermore we compute structural innovations imposing restrictions on their long-run effects, draw the impulse responses, and test the results by means of bootstrap simulation. We conclude that major room for further refinement of the research is found in two areas, First, from an economist's perspective, in the construction of real marginal cost indicators, and second, from a statistiacian's perspective, in further investigation of the robustness of the method.