Effects of alternative seasonal adjustment procedures on monetary policy
Effects of alternative seasonal adjustment procedures on monetary policy
复制标题
替代季节性调整程序对货币政策的影响
DOI：
10.1016/0304-4076(80)90076-7
复制
发表时间：
1980
影响因子：
6.3
通讯作者：
A. Maravall
中科院分区：
文献类型：
--
作者：
A. Maravall
联委会目前用于调整M的季节性调整方法是X-11和对活期存款和货币的月序列进行的判断性修正的混合。由于按月（未经调整）数列是每日数值的平均数，另一个计算按月调整数字的程序是“调整每日数列，然后将其合计为月份”。换句话说，时间聚合可以在季节调整之前或之后。巴赫货币统计委员会建议使用每日方法。在本文中，我们比较了官方的董事会的方法与每日一次在董事会，这是一个扩展的实际使用的巴赫委员会。有人认为，有理由更喜欢一个每日的方法和理由更喜欢每月一次。似乎可以得出这样的结论：在这两种方法都不是毫无疑问的上级的情况下，它们提供了两种调整系列的替代方法。由于这两种方法得出的结果不完全相同，它们之间的差异反映了利息序列（经季节调整的月度M）计算中的模糊性。从逻辑上讲，这种模糊性为系列的测量精度设置了一个下限，因此具有测量误差的特征。当然，如果在大多数实际情况下，这两种方法之间的差异非常小，那么模糊性很容易被忽略。另一方面，有时可能是关于M的增长是否足够的判断，因此，
The seasonal adjustment method presently used at the Board for adjusting M, is a blend of X-11 and judgmental corrections applied to the monthly series of demand deposits and currency. Since the monthly (unadjusted) series are averages of daily values, an alternative procedure for obtaining the monthly adjusted figure would be’to adjust the daily series and then aggregate them into months. In other words, temporal aggregation can precede or succeed seasonal adjustment. The use of a daily method was recommended by the Bach Committee on Monetary Statistics.’In this paper we compare the official Board method with the daily one developed at the Board, which is an extension of the one actually used by the Bach Committee.It is argued that there are reasons for preferring a daily approach and reasons for preferring the monthly one. It seems safe to conclude that, to the extent that neither of the two approaches is unquestionably superior, they provide two alternative methods for adjusting the series. Insofar as the two methods do not yield identical results, the difference between them reflects an ambiguity in the computation of the series of interest (monthly seasonally adjusted M,). Logically this ambiguity sets a lower limit to the precision with which the series can be measured and therefore has the character of a measurement error. Naturally, if in most practical situations the difference between the two methods happens to be very small, the ambiguity could be easily ignored. On the other hand it may be that on occasion the judgment as to whether growth of M, has been adequate or not and therefore the