7 OCTOBER 1938, Page 36

COMMODITY FLOW AND CAPITAL FORMATION : VOL. I By Simon

Kuznets

This formidable work (Macmillan, 2 rs.) is the first of a series of statistical-economic studies undertaken by the National Bureau of Economic Research, New York, at the request of a sub-committee of the Social Science Research Council. It is, it appears, to be followed by another embodying data not on the commodity but on the financial aspect of capital formation ; and the aim of the whole series is to provide the necessary factual matter for a constructive study of business fluctuations and of dynamic change in the economy. The present volume falls into eight parts ; a classification of manu- factured commodities according to their durability ; an estimate of output at current and 1929 prices ; three sections dealing, from various angles, with what may be described as the " middle- man's rake-off " ; an estimate of the volume of construction for the years 1929-33; a presentation of the net change in " inventories " or stocks ; and finally the resulting estimates of gross and net capital formations according to a number of alternative definitions. To the English student in particular, the interest of the book lies in its statistical procedure. There is no possible parallel in Great Britain ; the essential data are simply not there. To set beside the American Biennial Census of Manufacture we have only the decennial Census of Production, published only when completely out of date. Not that the American figures themselves are by any means perfect from the statistician's point of view ; the author makes clear, as he explains the procedure at each step, the difficulties with which he and his colleagues have had to contend and the more or less trustworthy devices to which they have resorted in order to fill in the gaps. It will be interesting to -see what use is made by American economists of the mass of material presented ; its success should certainly add to the demand, on this side of the Atlantic, for a reform of the nation's statistics.