Why Calling Banks "Financial Intermediaries Between Lenders and Borrowers" Is Okay, Even if You View Banks Through the "Endogenous Money" Lens or Something Similar


There have been extensive debates on the econoblogosphere over the past couple of years regarding the degree to which banks are “special” due to the manner in which they create money and credit. Steve Randy Waldman, as is often the case, provides a helpful synthesis, discussion, and collection of links regarding these debates.

As almost always in these sorts of mainstream versus heterodox intellectual scuffles, the extent to which there is disagreement due to differences in framing and semantics, as opposed to the underlying fundamental issues, is unclear. Additionally, there are many elements to this debate, and both sides have valid things to say (disclaimer: I am admittedly amalgamating the debaters into two broad camps for simplicity, but further delineations can be made). However, I do not think that for each and every element of the debate, each side offers an equally valid viewpoint. In any case, my intention here is not to provide a comprehensive analysis of these matters. I recommend reading JKH’s commentary at the following links for more detailed thoughts along these lines.

What I’d like to focus on is the mainstream economist’s tendency to characterize banks as ‘financial intermediaries between borrowers and lenders’ with respect to their lending function. Many participants (but not all!) in this debate that lean towards the economic heterodoxy have a problem with this characterization, given their usually more nuanced understanding of banking operations and mechanics. I understand and sympathize with this perspective, but I ultimately aim to argue what’s stated in the title of this blog post. I think someone with an endogenous view of money can be okay with the mainstream characterization, even if some mainstream economists who use it don’t have a very robust understanding of how banking works operationally.* To be clear up front, none of this is to say that banks are not important to the economy or should not be modeled.

I sincerely hope I am not presenting a straw-man argument or am merely stating the obvious. I hope to offer a way of thinking about banks that I haven’t seen presented on the blogosphere, at least with the same clarity presentation and with the core concepts of endogenous money properly acknowledged. My ultimate goal is to shed light on and facilitate a unification, to the greatest extent possible, of some of the seemingly disparate views that have been expressed in this debate.* I think SRW started to go down this road, but he ultimately took a different direction than I do, his being more conceptual and complex. However, if my argument holds any water, then its value-added is likely in its relative simplicity.

Bank Lending Mechanics as “Special”

To start, some heterodox participants like to claim that banks are “special” because they can create deposits out of thin air without requiring a prior form of money (reserves or cash) - this is the element of "endogenous money" theory that I am referencing. “Special” is a term often left insufficiently defined by all sides, but in some ways excusably so given the complexity of what’s trying to be discussed and modeled. I’ll state at the outset that, putting the claim of “specialness” aside, I believe this is an accurate description of bank lending mechanics. Many mainstream economists tend not to describe banks this way or may even suggest the opposite. However, this feature of banking has been recognized for at least a century by various scholars that closely study banking operations and by practitioners that work in the business. If you want to go way back, see Charles F. Dunbar’s writing in chapter 4 of “The Check System” in Chapters on the Theory and History of Banking from 1891. (Hat tip to Perry Mehrling and his Coursera class; I highly recommend it.) I’m sure you could go even further back (if anyone has citations, please provide). Here are some supporting quotations from the more recent century from a range of experts that reside at the Fed, BoE, ECB, BIS, and IMF. There are even some eminent scholars in the mainstream tradition that have recognized this feature of banking, such as James Tobin; while Tobin has argued against the “specialness” of banks more broadly, and was referenced by Paul Krugman as having done so, that should only give you more confidence that this view of bank lending mechanics, as opposed to the “specialness” of it, is accurate. Additionally, there is a range of Post-Keynesian and other heterodox economic thinkers that recognize, if not emphasize, this element of banking. In any case, without beating this horse too much more, the following accounting should explain this simple institutional feature:

Boom. Loans create deposits without prior reserves or cash. For further explanation, please reference the plethora of blog posts and academic articles that address this issue on the internet and in print.

Of course, borrowers may ask to withdraw their deposits for currency, or banks may have to transfer deposits to another bank. In both cases, banks will need to transfer cash/reserves they already possess and/or borrow cash/reserves to complete the transaction. Furthermore, some countries, such as the U.S., impose reserve requirements on banks, which require banks to hold an amount of reserves equal to a certain percentage of their deposits.

In my mind, it’s at this point we can start to recognize that banks are like financial intermediaries between lenders and borrowers. Why? Because like anyone else, banks still have to fund themselves to do what they do, whether it is to meet reserve requirements or to settle payments. As Mehrling might say, every entity in the financial economy needs to ensure that their cash inflows will be sufficient to cover their cash outflows, including banks. Let me explicate further.

A Bank-Like Business Model as a Financial Intermediary

(As a disclaimer, I am assuming away any solvency and capital issues and am focusing solely on liquidity issues in the following example. This applies to all economic agents, including banks.)

Some of the heterodox participants in this debate may be more comfortable with describing a person who runs a lending operation as a financial intermediary between borrowers and lenders. So let’s run with this idea.

Imagine that Sally wants to start a lending operation, with a bank-like business model, out of her home. It could work like this. Sally builds a computer program that allows people to view their accounts online and maybe even attaches a machine to her house that functions like an ATM. Her neighbor John comes over looking for a loan for $100, and she excitedly accommodates his request. She opens the computer program and shows him that the loan has been approved.

John looks puzzled, so Sally offers clarification. “Just like at a bank, I’m giving you IOUs for $100 cash on demand, and I’m creating them straight out of thin air. At a commercial bank, the IOUs are called ‘bank deposits.’ At my business, I call them ‘IOUs for Cash.’ But they’re the same thing – they’re promises to pay a certain amount of cash when you demand it.”

John isn’t convinced. He’s worried you can’t actually deliver upon your promise.

You explain, “Sure I can. When you want the cash, you just come here or to the ATM and ask for cash. I’ll provide you the cash. I plan to get the cash by making cash flow projections and borrowing when and to the extent I deem necessary, just like any bank would. I also hope to source funding by convincing people to store their cash with my bank-like operation. I can also provide you with tools that perform the same function as paper checks and debit cards to draw on your account without your having to physically do so.”

Driving Home the Point of the Analogy

Let’s pause the story there. Obviously, it’s unrealistic in many regards, and I assume away many institutional features and regulations that support and shape the real commercial banking system. I am generally hesitant to create such stories and am wary of others who do so. However, here I think there is some value in it, to the extent that it can clarify some of what banks do, particularly if you’ve been mesmerized by the ‘loans create deposits without prior money’ view (although it is a really cool and very important observation for many reasons).

If the concept of a ‘financial intermediary between lenders and borrowers’ exists, then I think Sally fits the preconception pretty well. But I think we can argue banks are similar to Sally in many ways, as Sally argued. Really, Sally can create "money," or I think more aptly, "credit," without prior reserves or cash, just like banks can. The differences instead arise when we begin to take into consideration the regulatory apparatus and funding sources to which banks have access. In order to further elucidate this point, let me actually address some of these key differences, rather than the similarities, between Sally’s bank-like operation and actual commercial banks.

One really important difference is Sally’s bank-like operation doesn't have access to the fed funds market as a source of funds nor the Fed’s various credit facilities. It’s a really important difference because these are continuously flowing sources of credit for actual commercial banks and some of the cheapest credit around. If the federal funds rate is 1%, then banks can borrow in these markets for around 1% (penalty credit facilities charge higher). In contrast, Sally might have to take out a loan from a bank to meet her liquidity needs, and the bank is only going to lend to her at some spread above 1%. Maybe she’ll have to find a loan shark and borrow at a really high rate. The point is that the difference between Sally and commercial banks in this respect is that banks can fund themselves at cheaper rates. It’s a difference of degree rather than kind. (For those objecting that Sally many not even be able to find funding, please see the last section of this post.)

As another option, perhaps Sally can attract funds by convincing other people to allow her to store their cash as ‘IOUs for Cash.’ Commercial banks do this too, by attracting deposits. To convince people to do this, Sally will probably have to offer them a much higher rate of interest on their ‘IOUs for Cash’ than banks would on their deposits. Again, the point is that the difference between Sally and commercial banks in this instance is that banks can fund themselves at cheaper rates.


I think I’ve said enough at this point to support and summarize my argument. It’s as follows:
  •  Sally is a financial intermediary in that she needs funding from lenders and a pool of creditworthy borrowers to operate her lending business, and so do commercial banks
  • While the sources and cost of lending to Sally may differ from the sources and cost of lending to commercial banks, lenders exist in both cases, and they’re interconnected
    • Sally can create credit in advance of having funding just as banks can
  •  Borrowers clearly exist in both cases, but may differ in character
  •  Both Sally and commercial banks need to make sure cash inflows can meet cash outflows
  • In this way, banks are financial intermediaries between lenders and borrowers similar to Sally; many of the fundamental differences are of degree rather than kind
Why can endogenous money people be okay with this? Well, I think it’s mostly harmless. We can still describe banks as having different characteristics than other financial intermediaries, such as their sources of funding and the added flexibility this may provide them, among many other reasons. We can also still advocate for banks to be incorporated into macroeconomic models, given their systemic importance and risks. I touch on this a bit further in the next section.

Anticipated Objections

I’ve admittedly presented a simplistic story that revolves around the concept of unlimited credit at a given price. The model I am imagining is that just as the Fed always funds solvent banks at a given interest rate that the Fed determines, banks in turn always fund solvent economic agents at an interest rate they deem appropriate given the risks (here liquidity risk, since I’ve assumed away insolvency).

This is obviously too simplistic as an explanation for the real world, since sometimes quantity, independent of price, does bind. For example, during financial crises, credit sometimes isn’t supplied at any price, even if the agent in question is arguably solvent (although the lines between liquidity and solvency are often blurred in these instances). I would say that while this is an incredibly important issue and always a concern, the event itself is not a norm in modern monetary systems.

A related objection is that banks could set credit limits on a solvent Sally that could eventually bind and leave her illiquid. In contrast, the Fed is unlikely to set credit limits on solvent banks, given its lender of last resort function. I acknowledge this as being a very important difference between banks and other financial intermediaries. However, I would say two things. Firstly, it’s unclear to me that the Fed itself is willing to truly extend unlimited credit to solvent banks, since under current regulation, Fed credit creation seems to be limited by available qualifying collateral.  As such, the difference between banks and Sally in this respect may arguably remain one of degree rather than kind. Secondly, the Fed’s emergency support services are expanding through time. During the financial crisis, it took the extraordinary step of, for the first time, moving non-government financial assets (e.g., MBS, ABCP, etc.) onto its balance sheet and extended its support to the shadow banking system. As such, I could argue that Sally would fall within that safety net in my simple model.

Maybe there is a right answer here, but I think this added complexity and realism starts to move the goal posts and is beside the point I am trying to make in this essay. The point is that commercial banks are financial intermediaries in that they need to borrow in order to sustain their lending operations. The differences between banks and Sally are one of degree rather than of kind, within the scope of what’s being discussed here. That said, the differences between banks and other financial intermediaries remain good reasons to include them within macroeconomic models. These include differences in funding costs, differences in access to reliable sources of liquidity during credit crunches, and so forth.


* Some of the more "heterodox" participants, such as Ramanan and JKH, have attempted to justify this characterization through an asset allocation lens, citing James Tobin’s views for support, and in doing so, finding common ground with Paul Krugman. I am taking a different approach here.


  1. So the key idea that's special isn't necessarily banks, but "credit" in general? If we subtracted banks, we'd have personal credit systems instead, say circulating bills of exchange, or peer-to-peer credit, or store credits.

    "I’m sure you could go even further back (if anyone has citations, please provide)."

    Dunbar was inspired by Henry Dunning Macleod, an earlier British economist (and failed banker). Macleod is excellent on credit.

    1. Was hoping you'd comment! I'd knew you'd have some great econ history to contribute :).

      Maybe that's one way to put it? My thoughts here have been partly inspired by Merhling, having read his latest book, "The New Lombard Street," but actually more so by his currently ongoing Coursera class. And I get the feeling a lot of this is from Mehrling's predecessors, such as Minsky.

    2. I suppose it depends on what we define as a 'bank.' As I currently see it, Sally functions similarly to a bank. It's the absence of today's commercial banking market infrastructure and regulation that make her in large part different.

    3. I can see someone arguing the following. So banks aren't special - OK. They're mechanically just like mutual funds or broker-dealers. So what does this mean for their modeling? We abstract away from intermediaries, as I understand it, simply because of efficient markets. The fees to fund managers, or the loan spread to commercial banks, are the friction costs of obviously inefficient markets. Just as we don't need mutual funds in an EMH universe, so we don't need banks, and you seem to have shown that. If we don't like how banks are ignored in financial economic models, we should blame EMH, not the semantic predilections of elder economists. The reason they feel we can ignore banks is because they think assuming EMH is acceptable, or at the very least, the basic assumptions of Lucas critique Keynesianism - preserve cleared markets, preserve price taking - are considered acceptable. The culprit is EMH, not disagreements over the word "intermediation".

    4. Great comment. Yes, failure of EMH sounds like one good reason for modeling banks. But I'm not yet sure if it's the only one. "Loanable funds" seems like it is also related to this debate, and if loanable funds is inapplicable, is that also an argument, independent of EMH, for modeling banks? I need to look into what Steve Keen is up to, since he's modeling banks and claiming loanable funds is a central issue. Kinda was waiting for him to get those accounting / NIPA controversies sorted out though.

    5. To follow-up, what people may be objecting to when they object to the concept of calling a bank a 'financial intermediary' is really loanable funds. But I think the two concepts are different. Further thoughts TBD.

  2. I still think banks are special in a special sense: their reaction functions are completely different from real-sector lenders and borrowers. So you can't usefully model the economy as if they were the same.

    When banks lend, they aren't "saving", increasing their capacity for future spending, consumption, and contingency, like real-sector actors.

    When banks borrow, they don't spend more (on real, newly produced goods and services--the stuff of GDP).

    When loans are repaid to banks, they don't spend more.

    This is just to say that a model of the economy that consists of patient savers and impatient borrowers can't accurately or usefully represent how heterogenous economic agents' reaction functions interact. Banks (and their agents) have to be modeled differently from real-sector agents. In that sense they're "special."

    1. Steve, I think your argument survives what I was trying to say in this post. Having had time to sit back from it, I think my ultimate goal was to show that 'loans create deposits without prior reserves or cash' isn't something that only banks can do. But we're all so used to thinking of this as the province of traditional commercial banks (or at least I used to). As JP Koning said above, credit can exist with or without a traditional commercial bank.

      I think your argument takes the debate to the next level. I'm curious if you think it's one in the same, or at least related, to Zach's comment above about the efficient market hypothesis (EMH). Do you think that if we made the assumption that EMH holds, then banks would in fact be imitating real-sector actor reaction functions, as you put it?

    2. Exactly the right question. I would say that the only way we could justify modeling banks as "patient lenders" is by assuming a pretty strong form EMH, where the economy (through portfolio and incentive effects) acts *as if* the banks were 1. patient lenders, or 2. perfect, transparent, instantaneous intermediaries between patient lenders and impatient borrowers.

      IOW, when banks lend more now (so people spend more), interest rates go up because they're experiencing more demand for loans, so they lend less (hence people spend less) later.

      The world just really isn't that simple, by a long shot.

    3. Curious then that EMH has not been brought up in these debates. It's at least language Krugman et al. would understand. Maybe next time this debate rears it's head, someone should bring it up. As I stated above, though, I don't think this is the only, or even the main, issue that Post-Keynesians are disagreeing with. See below...

      Steve, could you clarify:

      "IOW, when banks lend more now (so people spend more), interest rates go up because they're experiencing more demand for loans, so they lend less (hence people spend less) later."

      First, this in the strong form EMH world? Second - I'm not sure I understand the causality. You're saying, first banks lend, which generates more demand. In turn, that increases demand for even more loans, and so the interest rate on loans goes up through a supply-demand type analysis (supply curve of loans stays the same, demand curve shifts up)? Then in the third step, why exactly do they lend less later?

      I recognize that whatever you were trying to describe there, it's something that's "too simple, by a long shot," to accurately describe the world. But is your example also assuming loanable funds? It seems to be, in that you were suggesting increased demand for loans causes a rise in the interest rate. But the interest rate should be anchored by where the Fed sets bank funding costs, with banks of course adjusting risk premia as appropriate. If anything, interest rates would go down as loans become less risky, or are perceived to be less risky, perhaps?

      This ties into what I was saying above. Both EMH and loanable funds, as independent issues, may impact whether we should and how we model banks. I think Post-Keynesians have a gripe with loanable funds at least as much as EMH.

  3. Ramanan has a post touching on similar topics:


    My point to him was that in a simple model the banks could take on the role of the NBFIs too by just issuing shares directly. Sure that glosses over some things, but for a simplified approximation I think that works out (i.e. some bank deposits get converted into shares... think of consolidating the banks and NBFIs together).