[301] The materials in this appendix are taken from an article published in the Annalist of Jan. 8, 1917, pp. 39, 53-54, and the New York Times Annual Financial Review of Dec. 31, 1916, and are reprinted by the courtesy of the New York Times Company.
[302] Vide Annalist, Feb. 7, 1916, pp. 183-184, and Feb. 21, 1916, p. 246.
[303] Wealth and Income of the People of the United States, p. 129.
[304] The justification of this procedure is argued more fully in my article in the Annalist of Feb. 7, 1916, above referred to.
[305] The figures for railway gross receipts are taken from the Commercial and Financial Chronicle, rather than from Government reports, in order to get figures for calendar rather than fiscal years, and in order to get the latest possible figures. As the absolute figures are not strictly comparable throughout, the method employed has been to calculate percentage gains or losses for the same roads for successive years. This would lead to a cumulative error, if large new roads had been built during the period, and had retained their independence. In point of fact, however, the curves for the absolute figures and for the percentage changes run pretty closely parallel down to 1909, at which time a large number of small roads, not previously counted, are brought into the figures. As the number of roads reported varies, the percentage changes on the same roads give us the more accurate measure of year by year variation. It is, at the date of writing (December, 1916), the only possible method for 1916, since the Chronicle figures which come to the end of November are based on only 37 roads, with a mileage of 84,452 out of over 240,000 miles usually reported. For these roads, a gain of 19.63%, for the first eleven months of 1916 over the same months in 1915, is reported, and our figures for 1916 rest on the assumption that the gain for the whole year over 1915 is 17.27%. (The greatest gains are for the earlier months, as the end of 1915 was a period of great activity.) Much fuller figures supplied me by Mr. Osmund Phillips, of the New York Times, for the first ten months of 1915 and 1916 serve to justify this estimate for the gain of 1916 over 1915. For the Chronicle data, see vol. 102, p. 930, vol. 103, p. 2112, and passim.
The index of prices chosen is Dun's. (See especially Dun's Review of May 11, 1907, Jan. 9, 1915, and later months, and the discussion of Dun's index number in the Bulletin of the United States Bureau of Labor Statistics, Whole Number 173, July, 1915, pp. 148 et seq.) Dun's index number is chosen partly because it is complete for 1916, and partly because it is weighted in accordance with the consumption of different classes of goods, and so particularly suited to this inquiry. I venture to express strong preference for rationally weighted index numbers, and for the use of different index numbers for different purposes. (Vide the discussion of index numbers in ch. 19.) Our price index for each year is an average of the twelve monthly figures given by Dun from 1894 to 1916. For the years 1890-94, our price index is an average of the figures for January and July. This average is lower, in most years, than the average for the whole year, and may well be lower than the average for these years, but no attempt has been made to rectify this possible source of error. The index is recalculated from Dun's figures (where it is not a percentage, but a sum of prices), and made a true percentage index, with a base in 1910.
The figures for exports and imports are for calendar years. They were obtained, for the years 1890-1909, from Statistics of the United States, 1867-1909 (National Monetary Commission Report), and, for the years since 1909 from the Commercial and Financial Chronicle. For 1916, November and December are estimated.
[306] Their indicia of variation for "trade," though failing to meet the problems for which they were designed, as shown in chs. 13 and 19, are good indicia of variation for physical production and consumption.
[307] That this should have been seriously denied during the recent Presidential campaign, on the basis of the estimate that foreign trade is minute as compared with domestic trade, gives special point to the present discussion.
[308] King's figures, for which he estimates a margin of error of 25% are used for these years. (Loc. cit., p. 129.) The export and import figures used are for fiscal years.