All these factors must be taken into consideration before a full judgment of the net results of machinery upon the worker can be formed. The evidence above recorded, conclusive as it is regarding the existence of some causal connection between a high standard of living and high productivity of labour, does not necessarily justify the conclusion that a business, or a federation of employers, may go ahead increasing wages and shortening hours of labour _ad libitum_ in sure and certain expectation of a corresponding increase in the net productivity of labour. Before such a conclusion is warranted, we must grasp more clearly the nature of the causal relation between high standard of living and efficiency. How far are we entitled to regard high wages and other good conditions of employment as the cause, how far as the effect of efficiency of labour? The evidence adduced simply proves that _a_ _b_ _c_, certain phenomena relating to efficiency--as size of product, speed of workmanship, quantity of machines tended--vary directly with _d_ _e_ _f_, certain other phenomena relating to wages, hours of labour, and other conditions of employment. So far as such evidence goes, we are only able to assert that the two sets of phenomena are causally related, and cannot surely determine whether variations in _a_ _b_ _c_ are causes, or effects of concomitant variations in _d_ _e_ _f_, or whether both sets of phenomena are or are not governed by some third set, the variations of which affect simultaneously and proportionately the other two. The moral which writers like Mr. Gunton and Mr. Schoenhof have sought to extract, and which has been accepted by not a few leaders in the "labour movement," is that every rise of wages and every shortening of hours will necessarily be followed by an equivalent or a more than equivalent rise in the efficiency of labour. In seeking to establish this position, special stress is laid upon the evidence of the comparative statistics of textile industries. But, in the first place, it must be pointed out that the evidence adduced does not support any such sweeping generalisation. The statistics of Mr. Gould and Mr. Schoenhof, for instance, show many cases where higher money and real wages of American operatives are not accompanied by a correspondingly larger productivity. In such cases the "cheap" labour of England is really cheap.