A smoothing test under first-order autoregressive processes and a first-order moving-average correction

This paper focuses on two applications of time series methods. The first proposes a simple transformation of the unit root form of stationary testing to infer about the validity of smoothing by second-order running averages of a series, or of the variables in a linear model (here opposing co-integration testing). The second one advances a simple iterative algorithm to correct for MA(1) autocorrelation of the residuals of the general linear model, not requiring the estimation of the error process parameter.