DR. JAMES TOBIN, in his interesting analLI ysis of the shape of the liquidity preference function and the effectiveness of alternative monetary-fiscal policies,' refers to my discussion of the same problems.2 I would like to develop more fully some statements contained in my earlier discussion and to comment on Dr. Tobin's views and findings. In a passage in which he criticizes my views on interest, Dr. Tobin argues in effect that, for logical reasons, it is necessary to assume that the liquidity preference function becomes infinitely elastic at sufficiently low rates of interest (or, in any event, before the rate becomes negative). For, at the zero rate there surely exists an unqualified preference for cash as against claims;3 and if we take into account the institutional costs of credit operations, then this floor is set not at the zero rate but at some very low positive rate. An economist who maintains that, in the range of changes in interest rates, the demand for idle balances may possess little elasticity to interest, should add that the liquidity function must become perfectly elastic at the institutional floor-level (i.e., above the zero rate, at a level determined by institutional costs) . I am in agreement with Dr. Tobin. In my book in which I suggested that liquidity functions may be rather inelastic in the range of changes in interest rates I made at least one explicit statement which is the precise equivalent of the foregoing (italicized) proposition (pp. I86-87). I should have made an explicit statement also in another passage, and I will do so in the forthcoming second edition of the book (on pp. I70-7I). However, throughout my analysis, I assumed the validity of this proposition. My statements concerning the small elasticity of liquidity functions are explicitly limited to the usual range of changes in interest rates. The diagrams in my book are drawn merely for positive interest rates, and I believe that it is made clear that the lower limit is not really meant to be the zero rate proper but some very low rate including the institutional costs in question. Negative rates are not included because it is maintained that negative rates are qualitatively different phenomena in that they express subsidies.4 No private individual or institution can be made to hold securities at negative rates, if money can be held free of cost.5 Acceptance of the italicized proposition in the preceding paragraph does not imply abandonment of views expressed in my book. But Dr. Tobin is right in insisting that the implications of this proposition should be made clear. The main point in this connection is that as long as the liquidity function is inelastic above the floor-level in question, the conclusion with respect to policy is that which was presented in my book.6 The conclusion with respect to pol-