Regarding statistical data on economic trends observed over time, the portion of the observed data that arises not from the economy itself, but from non-economic factors such as climatic and natural changes caused by the seasons of spring, summer, autumn and winter, or from people's social habits that are repeated at fixed times each year, is removed from the observed data, and the original data series is corrected to show trends based only on economic factors. For example, if we look at the sales figures of department stores in Japan, in every year, July and December are exceptionally high, while January, February, August and September are consistently low. Such fluctuations that are repeated every year are not caused by economic factors such as price fluctuations or production fluctuations, but rather are based on non-economic factors such as the increase in sales due to the customs of midyear and year-end gifts, which are unique to Japanese society, and the seasonal stagnation of commercial activity immediately afterwards. In such cases, unless we smooth out such large fluctuations due to external economic factors throughout the year and extract from the original statistical data the movements that are based on purely economic factors behind them, it is impossible to know accurately whether the data is really on an upward or downward trend, or even what the level itself is. The main methods of seasonal adjustment are as follows: (1) Moving average method Seasonal fluctuations, whether they occur repeatedly every four seasons or only in certain months, all have a cycle of 12 months. Therefore, to remove the fluctuations, it is sufficient to take the average value of that cycle. Based on this idea, the moving average method assigns the average value of the 12 months centered around a certain month to the value of that month, and by making successive adjustments to this value, creates a seasonally adjusted series of the original series. (2) Link-relatives method Also known as the Parsons' method, this method calculates the monthly ratio of each month in the time series to the previous month, finds the median for each month, converts this to a link ratio with January being 100, and makes some corrections to create a seasonal fluctuation index. This index is used to seasonally adjust the original time series. (3) Census Bureau method Developed primarily by the U.S. Department of Commerce Census Bureau, this method uses a computer to repeatedly calculate moving averages, but because it can perform large-scale calculations, it also allows for detailed adjustments and selection of calculation procedures. Although it received various criticisms in the 1970s, it became widely used around the world, and many improved methods have been devised depending on the purpose of use. In Japan, the Statistics Bureau of the Bank of Japan and other organizations use this method as is. [Tadashi Takashima] "Economic Statistics Reader" by Yuzo Morita (1970, Toyo Keizai Shinposha)" ▽ "Seasonal Fluctuation Adjustment Method" edited by the Economic Planning Agency's Economic Research Institute (1971, Printing Bureau of the Ministry of Finance)" ▽ "A Comparative Study of Seasonal Adjustment Methods - Application of the Census Bureau Method X-12-ARIMA to Japan's Economic Statistics" by Yoshinobu Okumoto and edited by the Economic Planning Agency's Economic Research Institute (2000, Printing Bureau of the Ministry of Finance) [Reference] | |Source: Shogakukan Encyclopedia Nipponica About Encyclopedia Nipponica Information | Legend |
経済の動きについての時間を追って観測された統計データに関して、経済自体の動きから生ずるものではなく、春夏秋冬によって生ずる気候的、自然的変動によってもたらされる動き、あるいは人々の社会的な慣行から毎年決まった時節に繰り返される行動によってもたらされる動きなど、経済外的な要素に基づいて発生した部分を観測データのなかから除去し、元のデータの系列を経済的要因だけに基づく動きに修正すること。たとえば、わが国のデパートの売上高をみると、どの年においても7月と12月については飛び抜けて大きな数字となるのに対して、1、2月と8、9月は決まって低い数字になっている。このような毎年決まって繰り返される変動は、物価の変動とか生産の変動というような経済的な要因によってもたらされるというよりも、日本社会に特有の中元、歳暮の習慣による売上げ増とその直後における商業活動の季節的停滞という経済外的要因に基づくものといえる。このような場合には、そのような経済外的要因による大幅な変動を年間を通じて平準化し、元の統計データから、その背後にある純粋な経済的要因に基づく動きを取り出してみなければ、ほんとうにそのデータが上昇傾向にあるのか、下向傾向にあるのか、さらにはその水準自体をも正確に知ることはできない。 季節調整のおもな方法には次のようなものがある。 (1)移動平均法moving average method 季節変動は、四季ごとに繰り返されるものも、特定月のみに発生する変動も、いずれも12か月を1周期とするものである。したがって、その波動を取り去るのには、その周期の平均値をとればよいことになる。この考え方に基づいて、ある月の値として、その月を中心とする12か月の値の平均値をあてることにして、順次この修正をしていくことによって原系列の季節調整系列とするのが移動平均法である。 (2)連環比率法link-relatives method パーソンズ法Person's methodともよばれる。時系列の各月の対前月比を計算して各月別の中位数を求め、これを1月を100とする連環比率に直し、補正を加えて季節変動指数を作成する。この指数を用いて原時系列の季節調整を行う。 (3)センサス局法Census Bureau method アメリカ商務省センサス局が中心になって開発したもので、方法的にはコンピュータを利用して移動平均を繰り返し行うものであるが、大規模な計算が行えるため、細部の調整や計算手続の選択なども可能となった。1970年代には種々批判を受けつつも世界的に広く利用されるようになり、利用目的に応じて多くの改良法が考えられてきている。わが国でも、日本銀行統計局などではそのまま利用している。 [高島 忠] 『森田優三著『経済統計読本』(1970・東洋経済新報社)』▽『経済企画庁経済研究所編『季節変動調整法』(1971・大蔵省印刷局)』▽『奥本佳伸著、経済企画庁経済研究所編『季節調整法の比較研究――センサス局法X-12-ARIMAの我が国の経済統計への運用』(2000・大蔵省印刷局)』 [参照項目] | |出典 小学館 日本大百科全書(ニッポニカ)日本大百科全書(ニッポニカ)について 情報 | 凡例 |
<<: Seasonally adjusted - Kisetsuchoseiichi
>>: Seasonal Nursery - Kisetsu Takujisho
The title of a Kyogen piece. A female Kyogen. A h...
In medieval Germany, a judicial assembly was held ...
...Population: 180,000 (1991). Hungarian name: Új...
A general term for several species of fish in the ...
This is an industrial synthesis method for aromati...
The deity worshipped at Kasuga Taisha Shrine in Na...
…Amateur theatrical activities were also popular ...
…There is no record of a five-story pagoda ever c...
...Decora Tricolor cv.Decora Tricolor has beautif...
…Other dipterans that create mosquito swarms incl...
...About 10 species grow wild in South Africa and...
...The hilly area around the lake has been open s...
…After examining the suitability of isoprene, but...
The face of a person who has come of age and has h...
…The first walking leg (cheliped) of crabs and cr...