On exponential smoothing and the assumption of deterministic trend plus white noise data-generating models

The purpose of this paper is to draw attention to a segment of the literature on exponential smoothing in which a particular assumption is made about the stochastic process, generating the data. Although this assumption is both implausible and unsupported by empirical evidence, it appears to form the basis for certain methodological proposals and also for misconceptions about the properties of forecasts derived from exponential smoothing algorithms.