Analysis of Non-Stationary Time-Series Business Data

In time-series analysis of business and economic data (e.g. stock index data; corporate dividend payments; corporate profits; business start-ups; business survival rates) the statistical concept that has received considerable attention and gained much popularity among applied researchers is the one related to non-stationarity. As discussed in a number of econometrics textbooks (see Verbeek, 2000; Charemza & Deadman, 1997 among others), quantitative analysts are generally concerned with the concept of weak stationarity (or covariance stationarity) i.e. the mean, variances and autocovariances of the series are independent of time; that is E(yt)=c remains constant for all t; var(yt)=E(yt-c)2=?2 remains constant for all t ; and cov(yt, yt+g)=E[(yt-c)(yt+g-c)]=?g remains constant for all t and g?0. If one or more of these conditions are not fulfilled the time-series is called non-stationary (this is discussed more analytically in Seddighi et al., 2000).