This article explains the method used for seasonal adjustment of the monthly money series, M1 and M3.
Developments this year have made it very evident that revisions to the seasonal adjustments applied to the monetary aggregates can make a considerable difference to the apparent course of the latter month by month. Thus the new seasonal adjustments announced in May produced a seasonally adjusted growth of sterling M3 that was considerably more rapid in recent months (and particularly in February) than the adjustments previously in use had done.
The introduction of new seasonal adjustments did not mean that the previous estimates had been calculated incorrectly. The information needed for seasonal adjustment is almost always incomplete,so that the best estimates of seasonal influences tend to change over time. The adjustments are based on information about the past behaviour of the series, and they are naturally liable to revision in the light of more refIned data.
There are greater difficulties in measuring the seasonal influences affecting the money series than those for many other series: the past may not be a reliable guide to present seasonal patterns, largely because of abrupt changes in the pattern of public sector operations. As is explained more fully below, the Bank endeavour to take account of these changes, and to anticipate, on the basis of the latest available actual information and forecasts, the probable pattern in the current financial year. Not only has the public sector borrowing requirement for a financial year proved hard to forecast, but it has been even more difficult to forecast its pattern, month by month. Later information on the outturn for the fInancial year accounted for a large part of the revisions which were announced in May; the remainder reflected modifications to take account of better information about government expenditure by banking months (rather than calendar months), a soon as this became available.
Seasonal adjustment of monthly money statistics