Knowledge that Transforms

XATITH the revisions made in recent years, a most comprehensive body of statistics relating to currency in circulation has been built up, and placed upon a homogeneous basis for an extended period.' This article presents monthly and weekly series for the volume of currency after correction for seasonal variation, and examines such figures in relation to the business and other factors affecting them. Seasonal variation, which forms a large part of the total movement of the weekly and monthly series, was discussed by Mr. Bertrand Fox in this REVIEW, Vol. XIII, pages 26-30, where the weekly indexes of seasonal variation are presented. Monthly seasonal indexes appear on a later page. Business needs for currency determine the seasonal fluctuations of money in circulation, and are the most important cause of other-thanseasonal fluctuations. But influences not of a business nature also affect the series; and explanation for certain of the movements observed in post-war years hinges upon the method used in counting the amount of currency outstanding. The figures apply to money outside (a) the Treasury and (b) the federal reserve banks. The count thus obtained does not embrace the reserve funds of the general banking system, but it does include the till money of banks other than the reserve banks.2 Such till money accounts for something less than one-fifth of the total; the rest is made up of pocket money, merchants' till money, and hoards. Probably bank holdings are to be regarded as an intermediate step through which currency Dasses from the reserve banks into business channels or back from business into the reserve banks. Nevertheless, differences in fluctuations of vault cash and other money have at times been marked. A final point to be noted is that some of the money counted is in circulation outside the United States in Cuba and Canada, for instance, as well as in European countries which absorbed United States currency during and after the war.

MONTHLY data for the building and construction industries available for the period since the War make possible a fairly satisfactory measure of the cyclical fluctuations in building activity. In an earlier issue of this REVIEW, a detailed analysis of building statistics for the United States was presented and certain important conclusions were drawn concerning the general character of fluctuations in the building industry.' The following pages are devoted to a brief description of the post-war cyclical fluctuations of building activity and to discussion of the data, methods, and statistical background of our adjusted indexes of building and real estate activity.

SINCE publication of the monthly index chart was begun in I919, two important revisions of the money curve (C) have been made. In its original form, the curve was based on rates for two grades of commercial paper, adjusted for seasonal variation on the basis of indexes computed for the interval I890-9I6, with a horizontal normal (at 4.645 per cent).' Beginning with January I922, the seasonal correction was reduced so that allowance was made for only onehalf the pre-war seasonal variation, and in May I923, the first important revision was made, the entire curve being recalculated back to January

THERE has already been published in this REVIEW an analysis of the seasonal variations in weekly figures for outside bank debits.' In presenting the results of that study a special method for determining and eliminating seasonal variations as they appear in weekly data was developed. The following paragraphs apply that method, in full or in part, to four weekly series; in addition, a revised and improved analysis pertaining to outside bank debits is in course of preparation, and will be presented in a forthcoming issue of this REVIEW.

IT is now many years since the publication of Wiebe's index of commodity prices in England for the period I50I-I702. Nineteenth-century studies of prices have latterly been carried back as far as I779, but there remains a substantial gap between the index based on Rogers' materials and the modern indexes. Recent developments of statistical method have disclosed deficiencies in the computations of Wiebe, so that we are now in position to make a number of important corrections in his figures, by interpolation of missing items. The theory of interpolation, too, suggests the possibility of connecting the earlier with the later series of index numbers by using the movements of wheat prices as the basis of interpolation. It has seemed best not to carry out the process of interpolation explicitly, but the effect is secured by the reduction of the various series to a common base and by the use of the series of wheat prices to supply the general character of movement in the period 1703-78. It thus becomes possible to establish the broader facts in regard to prices over a long historical period, and though results are necessarily tentative, they are certainly significant.