Seasonal fluctuations are periodic series that are repeated on a more or less regular basis in each period, or each year. The periodic series really only exist in theory, but the seasonal variations are quite close to this model. The underlying assumption in model construction is that regular series are caused by systematic rather than accidental causes. Many examples of seasonal variations can be mentioned: sales of cars, as well as refreshments, temperatures or volume of rain registered during a year, etc. The systematic causes that produce such variations are repeated periodically, although some deviations may occur. The analysis of seasonal variations has, in our view, an obvious practical interest. The analysis allows, for example, to determine when it is time to change the seasons. This allows to explain the variations that are in some areas of production, movement of goods, etc. Note that similar to seasonal variations, that is, periodic variations can be found in some areas of economic life, for example in energy production and consumption, and for a week or during the day.
An economic time series can be affected by regular intra-temporal (seasonal) movements resulting from climatic conditions, model change, holiday practices, and similar factors. Often these effects are large enough to hide the underlying short-term movements of the series. If the effect of such repetitive intra-annual movements can be isolated and eliminated, the evaluation of a series can become more perceptive.
Seasonal movements are found in almost all economic time series. They can be regular, however, they show variation from year to year and are subject to changes in pattern over time. These changes are usually thought to evolve mainly stochastic rather than deterministic. However, seasonal adjustment practitioners have long recognized that some of the interannual variation of seasonal movements may be associated with calendar-related factors, such as the number of business days or "traded" in a month (for series whose monthly estimates are accumulations through the days of one month) or, of greater concern for some BLS series, the moment of transferring the holidays. Recently, variations in the length of the intervals between the monthly reference periods of the survey have also been found to significantly affect the seasonal patterns in some BLS series.
Because the seasonal patterns intrayearly are combined with underlying growth or slope and cyclic movements of the series (trend-cycle) and also random irregularities, it is difficult to estimate the pattern accurately. The earliest known attempts to isolate seasonal factors from time series occurred in the first half of the twentieth century. Some of the early methods depended on smoothing the curves by using personal judgment. Other formal approaches were periodic analysis, regression analysis and correlation analysis. Because these methods involved a large amount of work, relatively little implementation of seasonal factor adjustment procedures was performed.