A VIEW widely held among members of the advertising profession, as well as among academic observers of the advertising industry, is that advertising expenditures have never regained the levels of relative importance that they achieved prior to 1930. Relative importance in this case is measured in terms of the ratio of advertising expenditures to gross national product or some other index of aggregate economic activity. The commonly accepted position is that advertising first declined in relative importance during the great depression and during World War II and then rose steadily after 1945 until recent years but has never returned to its pre-1930 levels. Advertising men, on the one hand, tend to react to these data by bemoaning the fact that their industry has never regained the golden days of the 1920's. Or they anticipate with some relish the rewards that would be found if their golden age were repeated. Academicians, on the other hand, tend to react by inquiring into the factors which have led to this long-term decline.' They have usually found the explanation in greater efficiency, resulting in a decline in relative prices. This decline in relative prices, it is held, has led to larger physical quantities of advertising being consumed but to smaller relative quantities when converted to value terms. This is a somewhat surprising state of affairs in view of the apparent ubiquity of advertising today and because most of the reasons usually ascribed to the relative rise of advertising over the last two decades (see below) have been operative for prior decades as well. This anomaly suggests the possibility of some error or bias in the underlying data and has led me to an examination of the basic series from which these conclusions are drawn.2 The primary series invariably used to describe the behavior of advertising expenditures is the series presented periodically in Printers' Ink and prepared by the Central Research Department of McCann-Erickson, Inc.3 Since 1953, it has been published yearly in the Printers' Ink Advertisers' Annual, and its successor publication. In earlier years, the series as it was developed was published in sporadic issues of Printers' Ink. As we can see from Table 1, the original estimates of aggregate advertising expenditures for various years have been raised over time. But there is a general pattern to be found-the estimates for the 1930's were ultimately raised relatively little while the estimates for the late 1920's were raised quite drastically. At the time the final round of revisions were put into effect no explanation was given for this disparity in treatment. Further, the data relating to 1935 and