A “True” Time Series and its Indicators: An Alternative Approach
Cheryl L. Edwards, E. Philip Howrey
Journal of the American Statistical Association · 1991 · Volume 86 · Issue 416
DOI: 10.1080/01621459.1991.10475127
Abstract
de Leeuw and McKelvey proposed a method for using two imperfect indicators to predict the values of an unobserved series. In this article we suggest an alternative method that uses the multiple indicator approach to model the true time series and its indicators. We examine the identification status of the model, estimate the parameters by maximum likelihood, and use the Kalman filter to obtain predictions of the unobserved series. The method is applied to the U.S. gross private saving rate using flow of funds and national income and product accounts data as indicators.
Enable JavaScript for citation tools and additional paper interactions.