(The holiday strength carries over to the January monthly average as well.) The seasonal change in March is minor, but in April, when individual income tax payments are due, there is a temporary seasonal bulge in deposits. Deposits tend to rise in the late spring and through the summer with a temporary downturn in August. From September to November, deposits increase month-by-month and then rise sharply in December. As can be seen from the chart, this general pattern is repeated each year. Of course, the pattern is not identical each year because of the changes in other components of the series, the trendcycle, and the irregular. It is the general repetitive pattern that seasonal adjustment procedures attempt to remove in order to measure the more basic movements in the series.

At best, seasonal adjustment procedures can only approximate the "true" seasonal component of a time series, and the approximation of the seasonal is dependent on the seasonal adjustment method used. In the past the Board staff has studied numerous alternative methods of

seasonal adjustment, and results of these methods are frequently reviewed.1/

In general, comparison of the alternative seasonal adjustment procedures indicates that a variety of seasonal methods produce roughly similar turning points and growth patterns in the seasonally adjusted money stock series. However, a wide range of growth rates is obtained for a given month using alternative measures, and growth rates computed for monthly

1/ See Fry, Edward R., "Seasonal Adjustment of M,--Currently Published and Alternative Methods," Staff Economic Studies, Board of Governors of the Federal Reserve System.

periods by any of the methods are volatile. The results underscore the

need for caution in interpreting single-month growth rates in the money stock. It is clear that seasonally adjusted movements in the money stock, month-to-month, should be viewed with caution because of the large transitory component in the series. Thus far in 1977, the money stock has been quite volatile month to month, seasonally adjusted. view of this volatility, some analysts have even suggested the discarding of seasonal adjustment altogether. Recently, such criticism has focused on the April bulge in M1.

In

The April bulge in M1 has been reviewed carefully both on a seasonally adjusted basis and not seasonally adjusted. Different methods of seasonal adjustment have been applied in order to determine if they would yield results significantly different from the Board's current method of seasonal adjustment. The alternative methods of seasonal adjustment give somewhat different readings of the seasonally adjusted growth in M1 in April--as would be expected--but all of the methods indicate that seasonally adjusted M1 expanded at a very rapid rate in April, in a range of from 14% to 22% per cent at an annual rate. The published M1 growth in April was 19.4 per cent at an annual rate, near the midpoint of the range of growth rates obtained from alternative seasonal adjustment methods.

Some analysts maintain that current seasonal factors derived from the X-11 seasonal adjustment method are improperly weighted by what has happened in the 1972-1976 period, a period of abnormal economic

changes. Frequently, it is suggested that the appropriate comparison

is year-over-year money stock growth rather than monthly seasonally

adjusted growth.

Year-over-year comparison is, however, nothing more

than a crude method of seasonal adjustment.

Practitioners of this

particular form of analysis can be as easily misled by the data as

When using

analysts depending entirely on seasonally adjusted data. year-over-year comparisons one must always be watchful that the comparisons are not distorted by unusual developments in the current year or in the base year. Such a problem occurs when the April 1976 money stock data are used as a base for year-over-year comparisons. The level of M1 in that year was relatively high because of a spurt in My growth. When April 1977 is compared to April 1976 the underlying

current strength is not measured because of the strength in the April 1976 base.

Evidence of the spurt in M1 growth in April 1977 is provided by the monthly not seasonally adjusted growth rates shown in Table 1. The first column of Table 1 shows the per cent change, not at an annual rate, in the not seasonally adjusted money stock from December to April for each year back to 1966. Column 2 shows the March to April per cent change, not seasonally adjusted and not at an annual rate, for the same period. As can be seen from the table, the December to April change in My was quite strong in 1977. In only two of the twelve years shown-1971 and 1972--was the change in M1, not seasonally adjusted, stronger than in the current year. The strength in the March to April not