Most of the existing autoregressive models presume that the observations are perfectly measured. In empirical studies, the variable of interest is unavoidably measured with various kinds of errors. Thus, misleading conclusions may be yielded due to the inconsistency of the parameter estimates caused by the measurement errors. Thus far, no theoretical result on the direction of bias of the lag order estimate is available in the literature. In this note, we will discuss the estimation an AR model in the presence of measurement errors. It is shown that the inclusion of measurement errors will drastically increase the complexity of the problem. We show that the lag lengths selected by the AIC and BIC are increasing with the sample size at a logarithmic rate.
|Number of pages||10|
|Publication status||Published - 1 May 2006|
- Akaike Information Criterion
- Autoregressive Process
- Bayesian Information Criterion
- Measurement Error