Winterspeak responds to JKH's recent elaboration of a previous discussion of Paul Krugman's assertion that banks are not special based on his reading of Tobin-Brainard 1963. That reading is contested by some.It is significant in that it reveals how Krugman's idea of endogenous money compares to heterodox views that he contests. However, in the course of it a lot of information is coming out about bank operations that is of interest, especially in the discussion in the comments.
A Bank is still not a financial intermediary: redux
The question hangs on the interpretation of "intermediary." In the broad sense, an intermediary is an agent that facilitates a relationship between principles, usually for a professional fee, like a matchmaker. But this is not the only use of intermediary and banks are usually considered financial intermediaries even though they do not lend out deposit but rather fund their loans with a combination of capital and borrowing from various sources only part of which includes deposits.
The difference with banks is that their primary function is risk management rather than intermediation in the sense of savings and loan institutions and credit unions, for instance, that actually lend out deposits that they take in. The argument is not so much about this kind of intermediation, since banks are clearly special cases in that they have access to borrowing in the interbank payment system not only from other banks but also from the central bank.
Banks lend against capital and fund their assets resulting from loans by borrowing from a variety of sources, other banks, depositors and the money market. In extending credit, banks don't just receive interest as a fee for matching borrower and lender. Banks actively participate in the creation of flow, not only by arranging for existing funds to be lent, but also by adding to deposits, thereby generating a flow that increases the money stock through expansion of M1. As Neil Wilson notes in a comment at Winterspeak's, this necessitates dynamic analysis of flow rather than static analysis of stock. Other intermediaries free up existing funds to generate flow while banks create flow "out of thin air" by creating deposits and only afterward obtaining matching funding in order to balance accounts.
The question is whether this makes a difference that makes the special case of banks special economically. The fact that banks lent imprudently to borrowers who were not creditworthy based on dodgy collateral would argue that banks are indeed specially economically in generating financial instability, as Minsky hypothesized. Yet, shadow banking was also heavily involved in the crisis, which would argue against banks alone being special in this sense. However, many of the shadow banking institutions were owned and controlled by banks and were used to by banks to increase leverage beyond that which regulation allowed.