In concluding the discussion of the role of velocity of circulation, I think it worth while to mention Fisher's own efforts to measure them. I examine his statistics in a later chapter. I do not regard the points at issue as points which can properly be handled by inductive methods, primarily. I do not accept his conclusions with reference to the magnitudes of V, the velocity of money, partly because I do not accept his doctrine that "banks are the home of money" (p. 287).[220] He finds for V a fairly constant magnitude during the thirteen years from 1896 to 1909, the range being from 19 to 22, the figures for all the years except 1896 and 1909 being interpolations.[221] For V, however, which is much the more important magnitude, from the standpoint of his equation of exchange for the United States, since deposits do so much more exchanging than does money, he finds a wide range of variation, from 36 to 54, and he states: "We note that the velocity of circulation has increased 50% in thirteen years and that it has been subject to great variation from year to year. In 1899 and 1906 it reached maxima, immediately preceding crises" (285). I think Fisher's own statistical results show that V', at least, is a child of the "state of trade."[222] Critical analysis of these statistics show that they greatly underestimate the variability of the V's.[223] In summary: V and V' are not, as Fisher contends, independent of the quantity of money. Instead of resting on "technical conditions," and having large elements of constancy and rigidity, they are highly flexible, and vary, on the whole, with the same highly complex and divergent sets of causes which govern the volume of trade. The biggest factor affecting the variations of the V's on the one hand, and volume of trade on the other is business confidence--a factor which Fisher's normal theory is not concerned with, so far as it is considered as a variable, but which, more than anything else, does affect the concrete figures which go into the equation of exchange, either for a single year, or for an average of a good many years. The V's are not true causal entities, but merely abstract summaries of a host of heterogeneous facts. I have indicated before, and shall later demonstrate more fully, that the same is true of T. Even the "normal" causes governing the V's, however, are factors which likewise affect T, and in the same direction.