F— ECONOMIC SCIENCE AND STATISTICS 141 



numerical measure of very large, 1,263,109 persons in very large com- 

 mercial and transport firms. In short, statistics collected by plants, to 

 which America and Britain are mostly confined, seriously understate the 

 degree of large-scaledness, if this is meant to refer to size of firms. 



The reason for the difference in the typical size of plant and firm is of 

 course that many firms have branch plants. The German Census of 1925 

 found that for industry and commerce as a whole, 42-4 per cent, of all 

 persons occupied were occupied in firms with branch plants. There were, 

 in fact, 45,634 firms v/ith one branch over and above headquarters plant, 

 7,318 with two branch plants and 8,051 with three or more branch plants. 

 As to the trend in the size of firm, particularly these multiplant firms, 

 no exact figures are available even for Germany ; but the view is generally 

 accepted that the combination movement is growing. Certainly, re- 

 searches now being conducted at Birmingham into the largest British 

 Joint Stock Companies, and also a random sample of all British industrial 

 and commercial companies, disclose a much greater concentration of 

 power into a few hands to-day than is usually supposed to be true for 

 Great Britain. For any company to be a Holding or a Subsidiary Com- 

 pany and to possess Interlocking Directors in common with other com- 

 panies is rather the rule than the exception. There is also the device, as 

 yet little noticed in textbooks, of the weighted vote. A few special shares 

 may be created with perhaps ten times the voting power of the ordinary 

 ' ordinary ' shares, thus helping an individual or a company that holds 

 them to ' pyramid ' control to astronomical proportions. For the moment 

 we are not concerned with policies of internal control, but merely notice 

 the wide and probably widening prevalence of large areas of control. 



When, instead of prevalence and trends, profits and costs are used as 

 measures of efficiency, it is of course the size of the firm that is being 

 tested. The view of most theoretical economists has hitherto been that 

 after a certain size is passed, in spite of marketing and financial economies 

 and plant decentralisation, the firm becomes too large to be manageable. 

 In so far as this view is at all based on statistical information, it derives 

 mainly from the experience of the big trusts formed in America between 

 1893 and 1905 as summed up by Prof. Dewing writing in the Quarterly 

 Journal of Ecottomics for November 1921. Subsequent writers have 

 not found the conclusion quite so obvious. Mead,^^ following up Dewing 's 

 thirty-seven mergers, gives a table showing that during the years 1924-31, 

 seventeen of the mergers averaged at least 6| per cent, per year in dividend 

 on the common stock, watered though it often was. Mead himself con- 

 siders thirteen of these mergers, and one other that paid lower dividends, 

 to have been conspicuously successful when accumulation of large surplus 

 and provision for depreciation of plant and depletion of mineral properties 

 is taken into account, in addition to paynlent of dividend. 



More recently Livermore ^^ has taken a wider list of the mergers formed 

 in America between 1890 and 1904. Out of 156 mergers that were 

 important enough to have the power to influence conditions in their 

 industry he found that 53 suffered early and 10 later failure, while 17 



** Corporation Finance, 7th ed., chap. XXXVIII. 

 -' Quarterly Journal of Economics, Nov. 1935. 



