I have been critical of the theoretical positions held by Krugman for a while now, even if he and DeLong, and even Summers, have been useful for policy reasons. Now a lengthy debate between Krugman and Steve Keen, a very pragmatic and reasonable post-Keynesian (and I guess part of MMT tradition) that understands endogenous money has developed [a good summary with all the links here].
First, and foremost endogenous money implies that the rate of interest is exogenous and determined by monetary authorities. That per se is not necessarily in contradiction with a neoclassical/marginalist view according to which the rate of interest equilibrates investment to full employments savings, as Krugman clearly believes. Wicksell [see here] certainly did not think so either.
For Wicksell in a giro system, in which all transactions were recorded as debit/credit relations, credit could expand indefinitely, but in the real world, bank reserves would vanish and lending would eventually collapse if the bank rate remained below the natural rate for a long period. That is fundamentally the reason why Krugman does not understand the notion that banks can create reserves, and that loans cause deposits. In other words, what regulates the bank rate is, ultimately the natural rate of interest.
Further, the natural rate of interest is NOT a banking phenomenon in marginalist analysis, and, as a result, cannot be exogenous to the system. It results from the marginal productivity of capital and the intertemporal decisions of consumers. Krugman is in fact very clear that he supports the loanable funds theory of interest.
Hence, Peter Cooper is correct to point out that ultimately the debate with Keen must revolve around a notion of a long term normal rate of interest that is institutionally determined by the central bank independent of the marginalist notion of the natural rate. That can only be obtained with the proper critique of the neoclassical notion of capital.