A common feature of practically all strands of post-Keynesian theory is the notion that the money supply should not be considered as fixed independently of money demand in macroeconomic models. There are, however, at least two ways to postulate money endogeneity. The first, and perhaps best known today, is Kaldor's version, where the money supply curve is assumed to be horizontal at a given interest rate level. Kaldor's approach focuses on the means-of-payment function of money, stating that money is created when firms and individuals plan to acquire goods and services and borrow from banks the necessary amount of money to do it. Kaldor's emphasis is laid on central banks' behavior, assumed to be entirely accommodating of commercial banks' demands for the reserves required to satisfy the demand for bank loans. Keynes's version, based on his Treatise on Money and other essays, focuses on money in its liquid-store-of-wealth function. To propose that money is the most liquid asset in an entrepreneurial economy rules out the possibility of accepting a horizontal money supply curve, as it is shown in the paper. In fact, the first and most important contrasting concept in Keynes's approach in comparison to Kaldor's is the notion of liquidity. Keynes proposes a hierarchical view of liquidity, while Kaldor views liquidity as a ‘flat’ concept, where different assets exhibit different degrees of liquidity but their relationship is not hierarchical. A second contrast is that Keynes's view of endogeneity is based on a theory of how banks work instead of a theory of central banking. The paper develops Keynes's approach to money endogeneity along the lines just described and evaluates Kaldor's criticisms of Keynes's views.