F— ECONOMIC SCIENCE AND STATISTICS 149 



the location factor of the area) and into the urban, suburban or country 

 character of the place, for which also there are index measures. As to the 

 efficiency limit of size in manufacturing plant, the prevailing size of plant 

 in Great Britain, America or Germany is in many manufactures one 

 employing over 500 men, and for nearly all manufactures the prevailing 

 size is increasing. The typical association of this large size with mechani- 

 sation makes the policy of encouraging larger scale plants appear rational 

 since mechanisation is increasing rapidly. But mechanisation does not 

 appear to involve large plants where transport costs force plants, however 

 mechanised, to be near a consumer market. In that case plants must be 

 ubiquitous, numerous and therefore small. But even in these at present 

 ubiquitous industries such as baking, brewing and furniture, reduction in 

 transport costs ^^ may in the future permit of less ubiquity and therefore 

 larger plants. 



The extreme of ubiquity is found in the distributive trades. Here 

 tests of efficiency fail to concur. Small scale shops and firms continue to 

 prevail, yet their costs appear to be higher and their profits per given 

 amount sold, lower than in larger units. In view of the general opinion 

 about the rising costs of distribution and the widening middleman's 

 spread, further research as a basis of policy is here urgently needed. One 

 must continue to deplore the apparent unwillingness of our Government 

 to undertake a Census of Distribution such as the American, the Canadian, 

 or the Irish. Before blaming Governments, however, perhaps we ought 

 to confess that economists have not always made very clear what they 

 wanted to get out of censuses nor made much use of them once they were 

 completed ! 



4. The enlargement of the size of firms and combinations of firms so 

 often involved in rationalisation and planning has often been opposed on 

 the ground of the unwieldiness and the difficulty of any one brain managing 

 large organisations. Statistical evidence that has been offered of lower 

 profits among larger firms is, however, not easy to substantiate. The com- 

 pai^tive efficiency of larger firms probably depends on what they do. If 

 they enlarge by widening their scope and integrating either vertically or 

 laterally they may reduce their scale of operation and thus get diminishing 

 returns. If they enlarge by the increased production of given standard 

 lines, they may secure increasing return. The manager's brain is, after all, 

 just one factor determining the curve of return. It may give way in 

 attending to a thousand different varieties of integrated materials, pro- 

 cesses and markets ; but it may continue to flourish and obtain increasing 

 returns if it can specialise in a large way. Much more research is wanted 

 into the efficiency of different degrees of integration or scope within a 

 firm (or plant) in any one industry. The optimum, most efficient, pattern 

 may be specialised narrow-scope productive plants sited in selected places, 

 controlled by a wide-scope integrated firm or combine which also does 

 marketing and financing. But there is as yet little data to compare this, 

 in efficiency, with other possible patterns, though I hope to have indicated 

 by argument and examples the sort of data required for research into scope, 

 and the methods of using that data. 



*^ See H. Levy, The New Industrial System, Part III. 



