A trend is the slow variation over a longer period of time, usually several years, generally associated with the structural causes affecting the phenomenon being measured. It is the variation left after time series analysis has removed accidental (irregular or random), working-day and seasonal variation from a time series.
In some cases, the trend shows a steady growth; in others, it may move either downwards or upwards. The cycle is a quasi (almost but not quite) periodic oscillation. It is characterized by alternating periods of higher and lower rates of change, possibly, but not always, involving expansion and contraction. In most cases, the trend is related to fluctuations in overall economic activity.
Generally, if the irregular part of the time series is relatively important, the trend cycle series is a better series for the analysis of longer-term past developments. However, this advantage is less clear when analyzing recent developments. This is because trend-cycle values for recent periods may undergo greater revisions (changes) than the seasonally adjusted values. Hence, the latter may be more suitable for the analysis of recent developments; this is particularly true around turning points. Trend-cycle series may, however, converge to stable results more quickly than seasonally adjusted series.