Seasonally Adjusting Money: Procedures, Problems, Proposals SCOTT E. HEIN and MACK Ofl easonal variation irs economic functions is perva- sive; production, sales and leisure activities vary both substantially and systematically over the course of each week, month and year. Besides the obvious seasonal variation in agriculture, there are well-entrenched pat- terns in many other production, payment and con- sumption activities of firms and households. Auto- mobile production lines, for example, shut down in the summer and new snodels are introduced in the fall; retail consumer sales are heaviest during the Christ- mas shopping months in the late fall; income taxes are paid in April; and July is the peak month for vacation and travel. As a result, the demand for money fluctu- ates seasonally as firms and households rearrange their financial portfolios to suit these varying patterns of economic activity. For many reasons, it is useful to distinguish these seasonal variations in the data from longer-run cycles or trends. The procedures that enable these seasonal variations to be identified and, if desired, removed from the data are called seasonal adjustment tech- niques. In this article, we examirse attempts to isolate the seasonal imptilses in the money stock. WHY SEASONALL.Y ADJUST MONEY STOCK ~%4E45URES? There arc at least two difl’erent reasons for seasonally adjusting money stock measures. The first reason is fur interpretative purposes. Many analysts simply want a times series for the money stock that reveals trend and cycle impulses but excludes the effects of seasonal Scott E. Hem is an associate professor of finance a t Texas Tech University, and Mack Ott is a senior economist at the Federal Reserve Bank of St. Louis. This article was writen while Professor Rein was a senior econonust at the Bank. Thomas’ H. Gregory and Robert W, He.îs provided research assistance. variatidn. In order to exclude such variation, sosne method of identif~dng seasonal variation in the money stock is required. The second reason concerns the setting of monetary policy. The Federal Reserve states its annual and short-run objectives in terms of seasonally adjsisted monetary aggregates. These policy objectives imply that seasonal changes in money demand will be accommodated, but these changes first must be iden- tified by some method. The In.terpreta tine Reason Many economic time series are seasonally adjusted for interpretative reasons. A standard analysis of tisne series data partitions each observation into three pri- mary factors: (1) trend-cycle, C; (2) seasonal, 5; and (3) irregular or random, E. Consider, for example, tise time series for demand deposits, D. Traditional analy- sis would represent D as (1) D, = C, S~ E~. If the seasonal factor S~ is known, a “seasonally ad- justed” measure of demand deposits can he ohtained by dividing by the seasonal factor: (2) ~t= C~ E~. Since the seasonal factor is intended to remove sea- sonal variation, it will he less than 1.0 when desnand is seasonally low and greater than 1.0 when demand is seasonally high; over the year, by construction, it aver- ages 1.0. Consequently, by seasonally adjusting the data, the trend-cycle variation is revealed more clear- ly. If analysts are interested primarily in the trend- cycle elesnent in demand deposits, they will find sea- sonally adjusted demand deposit data useful in their analyses. 16