Comparing Semi-Structural Methods to Estimate Unobserved Variables : The HPMV and Kalman Filters Approaches

Laurence Boone

Organisation for Economic Co-operation and Development

关键词: unobserved component models;    Kalman filter;    standard errors;    NAIRU;   

DOI  :  https://doi.org/10.1787/112875725526

学科分类：社会科学、人文和艺术（综合）

来源: OECD iLibrary

【 摘 要 】

Economists often seek to estimate unobserved variables, representing “equilibrium” or “expected” values of economic variables, as benchmarks against which observed, realised values of these variables may be evaluated. Such comparisons are often used as economic policy indicators, for example the output gap, as measured by the ratio of actual to potential GDP, is commonly used as a measure of excess demand in assessing inflation pressures. To estimate these unobserved variables, a popular approach is the so-called semi-structural approach which includes: the Hodrick Prescott multivariate filter (developed by Laxton and Tetlow, 1992) and the Kalman filter (see, among others Harvey, 1992 and Cuthberson et al., 1992). This paper shows that the two approaches are closely linked, and specifically, it explains how to reproduce theHodrick Prescott multivariate filter using the Kalman filter. Being able to do so has at least two possible advantages. First, while the traditional HPMV filter ...