The irregular component of a time series is the part of an economic series that is neither trend-cycle nor seasonal: the short, unpredictable movements that remain once those systematic parts have been stripped out. Statistical agencies decompose a series such as retail sales into a trend-cycle, a seasonal component and an irregular component, either added together or multiplied, as part of seasonal adjustment.
Irregular movements come from sampling error, strikes, unusual weather, natural disasters, one-off tax changes and similar events. Large single shocks are often treated as outliers and modelled separately so that they do not distort the estimated seasonal pattern. Programs such as X-13ARIMA-SEATS, which builds on ARIMA models, estimate the components. A seasonally adjusted series still contains the irregular component, which is why month-to-month changes can be noisy; the smoother trend estimate removes it, at the cost of being revised more as new data arrive. Unadjusted figures are described under non-seasonally adjusted data.
Ref 1100001805