Seasonal adjustment

Seasonal adjustment ( English seasonal adjustment ) is a statistical method from the field of time series analysis , with the economics is attempted seasonal fluctuations by smoothing from an economic - or curve to eliminate as if there were no fluctuations.

General

Statistics are often falsified by the fact that their data are influenced by seasonal fluctuations, i.e. movements over the course of the year that are repeated regularly in a similar form and approximately the same strength. In many countries, the unemployment rate is highest in winter due to weather conditions and lowest in summer , which is mainly due to the construction industry and other seasonal businesses. If only the original values ​​compared to the previous quarter were considered, this would lead to incorrect assessments, because no tendency for the further course can be derived from the winter decline in the construction industry . Seasonal adjustment is used to give the viewer the most unadulterated insight possible. The season is understood as an annually recurring business cycle that influences the business cycle. The point is to eliminate these influences.

Econometrics

In econometrics it is assumed that a time series is composed of different components: ${\displaystyle X_{t}}$

• Trend component ${\displaystyle T_{t}}$
• Cyclical component (also: business cycle component or business cycle component) ${\displaystyle C_{t}}$
• Seasonal component ${\displaystyle S_{t}}$

Overall, the decomposition of a time series is in the standard model: . The seasonal component represents a periodic signal, the occurrence and effect of which is normally known in the data. (For example, the unemployment rate fluctuates significantly with the seasons within a year). Since this effect is known and mostly also dominates the spectrum of the time series, it should be filtered out with the help of seasonal adjustment in order to investigate hidden periodicities, the trend or the nature of the random process in the data. ${\displaystyle X_{t}=C_{t}+T_{t}+S_{t}}$

Moving averages are often used for seasonal adjustment : If the order of a moving average corresponds exactly to the period of the seasonal effect, this is eliminated from the time series by averaging.

Many modern methods such as X-12-ARIMA or the Berlin method are based in part on convolutions of moving averages.

The problem of seasonal adjustment plays a major role, especially in official statistics. Statistics software is usually used for this. For example, the Federal Statistical Office has developed the BV4.1 program for this purpose . It is based on the current version 4.1 of the so-called Berlin method .

literature

• Literature on seasonal adjustment in the catalog of the German National Library

Individual evidence

1. ↑ Dieter Brümmerhoff / Heinrich Lützel (eds.), Lexicon of National Accounts , 2002, p. 344
2. ↑ Dieter Brümmerhoff / Heinrich Lützel (eds.), Lexicon of National Accounts , 2002, p. 344