Recent Posts

Sunday, December 15, 2019

Yes, Banks Create Money Out Of Thin Air

Unfortunately, money has not yet been abolished from economic theory, and we are stuck with pointless debates about banks and money creation. The latest salvo is "Banks do not create money out of thin air" by Pontus Rendahl, and Lukas B. Freund. As the title of this article suggests, Rendahl and Freund are incorrect in their assessment.

I assume that the Rendahl/Freund article is a followup to previous arguments, such as a Thomas Hale article I discussed recently. Since I just addressed the topic, I will keep my comments here as short as possible. (I am responding to the article since it is likely that I will do a book on fractional reserve banking, which will be an overview of incorrect theories that keep popping up. Very similar in style to Abolish Money (From Economics)!)

The key argument of Rendahl and Freund is:
Since modern money is simply credit, banks can and do create money literally out of nothing, simply by making loans”. This misconception may stem from the seemingly magical simultaneous appearance of entries on both the liability and the asset side of a bank’s balance sheet when it creates a new loan. But this is simply a reflection of double-entry bookkeeping. Economically, money creation by private banks is far from magic, nor is it out of thin air.
There are several ways in which banks’ ability to create money through lending is constrained, meaning that the idea of limitless money creation conjured up by the image of a ‘magic money tree’ is flawed. 
Very simply, the fact that the rate of money creation by banks is under a very vaguely defined constraint does not imply that it is not coming from "thin air". If one were to imagine that wizards could conjure rabbits out of thin air, but the mysterious laws of magic only allow a certain number of rabbits to be conjured in a period of time, the rabbits are still coming out of thin air. The authors very simply are attempting to impose a idiosyncratic definition of "thin air creation" and proceed to beat up up a straw man in a vigorous fashion.

There are plenty of crackpot theories about bank money creation that I have encountered over the years. Some are being pushed by random people on the internet, and are obviously flawed. But if we confine ourselves to people who have spent some time thinking about their crackpot theories, they at least implicitly accept the two following points (which are consistent with endogenous money theories).
  1. The bank "thin air" theorists argue that banks create money out of thin air, not random economic entities. In order to qualify as being a bank, an entity has to meet certain regulatory criteria. Importantly, this includes holding some form of a liquidity buffer, the details of which depend upon the national regulatory system. For example, Canadian banks do not hold "reserves," whereas their American cousins do. The implication is that is makes no sense to natter on that banks need to hold liquid assets in order to make loans (and hence, create money), since that it is part of the defining characteristic of being a bank. If one wants to be pedantic, one might note that entities with banking charters that are on the verge of failing due to lack of liquidity (or whatever) are unable to create money. So perhaps one should say that "viable banks can create money out of thin air."
  2. I have not seen any statement to the effect that banks can create an unbounded amount of money right now. There are various constraints that prevent such an outcomes. But the constraints holding for one business day do not extend very far. For example, nobody can point to an arbitrary upper limit for bank money creation on a one year horizon. (If you think there is, please explain why nominal income growth is bounded?) Given that mainstream economists appeal to models where a single time step is one generation, I fail to see how a constraint that disappears within twelve months should be taken seriously. So all we are left with is that the crackpots are vague about time scales, in much the same way as neoclassicals who invoke OLG models to "prove" something or other.
(Update: Based on a Twitter comment, I realise that there is a subsidiary point that needs to be addressed: most serious discussions of "thin air bank creation" distinguish between the money creation operation and a counterfeiting. The only group that I am aware of that do view bank money creation as a form of fraud would be Austrian economists, and even they sort-of accept that banks are not acting exactly like counterfeiters. As such, even if we accept "thin air money creation," there is no reason for a bank to create a whole boatload of money, since they are not the recipient of the money, unlike a counterfeiter.)

An alternative way of phrasing my argument is as follows: if one wants to argue that money does not "come out of thin air", where does is come from? If we look at other monetary transactions, the stock of money is generally conserved (with limited exceptions, like lighting cigars with high denomination bills). Normally, if an entity receives a monetary instrument, it comes from another entity. Bank creation of money does not obey that rule. For example, if one attempts to appeal to various bank capital constraints, one runs into the reality that bank capital does not transmogrify into deposit money.

Although I do not agree with the crackpot banking theories, the Rendahl/Freund criticisms are a dead end. 

(c) Brian Romanchuk 2019

32 comments:

  1. I like the phrase "beat up a straw man in vigorous fashion", though it makes me feel sorry for the straw man.

    ReplyDelete
  2. You really do need to separate private banks (like exist in USA and Canada) from central banks. The role of government is crucial to accurate theory.

    Money, being a human creation, can be managed. Whether money management is focused on stability or instability is decided by the managers.

    ReplyDelete
  3. Banks do not create money, nor do they lend money. As Richard Werner explains, banks are in the business of purchasing securities.

    When we sign a loan document, we create a promissory note which the bank then purchases, and in exchange gives us a matching deposit liability. That deposit liability then circulates in the economy as money.

    As Hyman Minsky said, anyone can create money, the problem is in getting it accepted. The function of banks is to backstop the money that we, the private sector, create.

    ReplyDelete
    Replies
    1. Bank deposits are part of the broader monetary aggregates, only not included in base money (M0). This means that they count as “money” within economic discourse.

      Delete
  4. "Unfortunately, money has not yet been abolished from economic theory"

    But should it be abolished?

    When you say things like this, it sounds like the usual mainstream economics [dangerous] discourse: money is something that reduces transaction costs and hence enhances economic efficiency, but other than that can be fully ignored and removed from economic models that are focused on the real (so not nominal) resources.

    ReplyDelete
    Replies
    1. It’s a plug for my book, where I explain my beef in greater depth.

      Most of my complaints about money are based on stuff pushed by Monetarists, who are undoubtedly mainstream. The real complaint with the neoclassicals is not about money, but getting rid of private credit from their benchmark models.

      Delete
  5. But when you get rid of private credit, you necessarily don't allow for a 2008 crisis, which has for sure heavy impact on real resources...

    ReplyDelete
    Replies
    1. Yes, that’s the actual problem. The stuff about money is a distraction.
      (This is Brian.)

      Delete
  6. Hmm but aren't both public and private money relevant to the real economy? You know, private money like bank deposits are almost a kind of derivative instrument where the underlying is public money (bank reserves or notes or coins)

    ReplyDelete
    Replies
    1. My argument is that the focus on “money” is misplaced, and leads to magical thinking (like Monetarism). What matters is credit more generally, and not all credit shows up in the money numbers.

      Delete
  7. One of Böhm-Bawerk's chief errors in the Austrian school was excluding money from his reasoning because it made analysis too complicated. Thus, he had to come up with theories about "roundabout production" to explain why capital should earn a higher return then the the labor embedded in it.

    However, if you include money in your analysis especially as regards liquidity preference, then there is no need to explain why capital has to earn that excess return, i.e., that you are trading yield for liquidity as Keynes explained, when you hold an investment asset in lieu of money.

    So while it's true that money is a veil, it should always be included in analysis because it does affect outcomes in the real economy.

    ReplyDelete
    Replies
    1. OK, money is cool. That does not mean monetary aggregates offer any useful information in s macro model.

      Delete
  8. I have never come across a satisfactory account explaining why banks are eager to attract funds from savers even though banks are able to create money from thin air. Can you recommend a source or would you perhaps give an explanation?

    I understand banks need deposits as a (particularly) cheap source of funding (central) bank reserves. Again, I have nowhere found details on this kind of funding or, for that matter, details corroborating that banks attract deposits only to the extent that they are needed as a source of cheap central bank funding.

    Why is there no clear account (compatible with MMT or the denial of loanable funds theory) to be found of the use of savers' funds to banks, and why and how exactly these funds are used by banks for purposes distinct from making loans?

    ReplyDelete
  9. "I have never come across a satisfactory account explaining why banks are eager to attract funds from savers even though banks are able to create money from thin air"

    That is the kind of questioning I expect someone to have when people claim that banks can issue deposits at will. Although that claim is true, it is incomplete: it just tells the beginning of the story, lacking too much.

    For example, a pizza maker has the operational capacity of issuing free pizza coupons at will. There is nothing stopping her: she just needs to print its logo and some sentences in some pieces of paper, which is usually very cheap. So the claim that a pizza maker can always issue coupons at will is true. However, this is just half of the story.

    If the pizza maker issues and distributes many coupons to many potential customers, it is inevitable that some of them will actually come to the pizza maker and order a free pizza. Then the pizza maker will have to prepare enough pizzas, or will face customer wrath, bad ratings in those food ratings apps and maybe be prosecuted for false advertising.

    So, while it is true that she has the operational capacity of issuing coupons, it is not true that issuing coupons is a good idea. The pizza maker may not have the capacity of producing enough pizzas and honoring the coupons, which is another history that was not clear in the initial claim.

    In the case of banks, yes, they can issue deposits at will. They will usually do that in two circumstances: when granting loans to customers, or when customers make currency deposits and transfers.

    Attracting deposits is usually good for banks because it earns currency in return for issuing a liability for its customer. The currency can be used to a lot of things, including paying expenses, taxes, settling obligations in general or granting loans.

    Demand deposits can be withdrawn at any time. When withdrawn, the bank needs to have enough currency to honor it. So if a bank issues deposit liabilities without having enough cash to cover potential withdrawals, it will go into trouble.

    For small banks, things are clearer. When the small bank grants a loan to a customer it also issues a liability (deposit) to the customer. It is almost impossible that the customer will keep it deposited there or will transfer the financial resources to some other customer that has an account in the same bank (because it is a small bank). Instead, the customer will probably transfer or withdraw the money in the same day. Hence, the small bank must always have enough currency before granting loans. It can do that by attracting currency from small or big investors, through demand or term deposits, or through many other funding instruments. If it receives more interest from the loan than it pays to the investors, it will may be a profitable business.

    ReplyDelete
    Replies
    1. André,

      Thank you for your most useful reply.

      I have never had a problem understanding that banks can in principle produce as much money as they want to, but do not make use of that capability in the face of constraints imposed on them largely by capital restrictions, borrower demand, borrower quality and the need to operate profitably. Brian has explained the connection between an “unlimited capacity” and “limited use of an unlimited capacity” very nicely in the post above.

      Incidentally, I do not get the analogy of “the capacity of producing enough pizzas” — presumably you have my above points (capital constraints etc.) in mind.

      What caused me difficulties (and is the reason for my questions and requests above) is that very few people write about the use of savers’ funds for fiat-money-creating banks — I have found two sources (which I can’t reference at this point) who addressed the issue, and both claimed that such funds are exclusively needed to finance bank reserves. And it is this “exclusively” that made me wonder. Is this really the only purpose of deposits?

      Certainly, banks do not need to fund their loan making, but they do have to fund the fungibility (my phrasing) of (the money created by) their loans which is based on using bank reserves as the “currency” of the payment system (of which only banks and the central bank are members) which makes loans fungible.

      The last four paragraphs of your reply are a tremendous help in better understanding deposits (other than those created by making a loan). Thank you so much.

      Still I wonder, why have I never come across anything similar to your excellent explanation.

      I think it is important to be able to differentiate the different uses/kinds/originations of deposits, if one is to have a good grasp of fiat money.

      I wish someone would write up a few pages on this issue or perhaps an article or a chapter, after all, this aspect isn’t entirely trivial, is it?

      A topic for Brian’s next book?

      Delete
    2. "Certainly, banks do not need to fund their loan making, but they do have to fund the fungibility (my phrasing) of (the money created by) their loans which is based on using bank reserves as the “currency” of the payment system (of which only banks and the central bank are members) which makes loans fungible."

      I'm not really sure I could understand the fungibility thing. For the bank, currency (bank reserves + notes + coins) is one kind of asset, and loan receivables are another, not fungible. And deposits are liabilities, so not fungible with the assets. The bank must use currency to settle withdrawals or transfers with other banks and/or financial/payments institutions.

      But yes, as soon a loan is granted to the customer, the small bank registers in its assets a loan receivable and a deposit in its liabilities. For the customer, on the other hand, the deposit is an asset. The first thing the customer will do is withdraw or transfer the amount it has in her deposits. So the small bank will zero its deposit liability, and transfer currency to the counterparty financial institution receiving the transferred amount, to settle the transfer operation. In order to do that, the small bank needs to have currency. I think you said something on those lines, and it is correct if that was the case.

      Most economists don’t know any of that (which I find very worrying). So you need to seek specialized books about banking.

      But I would recommend Modern Money Theory: A Primer on Macroeconomics for Sovereign Monetary Systems, by L Randall Wray. It covers much more than this specific issue, and I guess it will be useful to answer your question and much more.

      Delete
    3. If a bank made me a loan with the proviso that I must spend the money only on goods and services the bank is offering, the loan would be non-fungible.

      What a typical borrower needs is a loan that comes with fungible money, i.e. I can make payments to just about anyone with this money.

      Being able to make payments to virtually any participant in the banking system (fungibility) is guaranteed by the payment system which is based on a special type of currency: bank reserves.

      Bank reserves need to be acquired – at a cost. (This is how I understand MMT).

      In this sense, not loans need to be funded/financed but the fungibility of the money created by a loan needs to be funded/financed by a bank.

      The few MMT-proponents who address the question, why banks seek to attract deposits confine their explanation to the banks’ need for a cheap funding source (customer deposits) for bank reserves.

      I have not found a word in Wray (or similar texts on MMT) on the question that I have raised (different uses/kinds/forms of origination of deposits).

      Delete
    4. Just replying to the above by IGTU

      “Still I wonder, why have I never come across anything similar to your excellent explanation.

      I think it is important to be able to differentiate the different uses/kinds/originations of deposits, if one is to have a good grasp of fiat money.

      I wish someone would write up a few pages on this issue or perhaps an article or a chapter, after all, this aspect isn’t entirely trivial, is it?

      A topic for Brian’s next book?“

      Ive been tied up discussing my latest article, so I have not been able to look at this thread. My feeling is that I could do a primer with a title like “Everything You Need to Know about Fractional Reserve Banking But Were Afraid To Ask.” I still need to do Volume II of recessions, and a MMT primer, so that might get pushed off.

      But to comment on your questions, banks need to finance their balance sheet. Deposits are the cheapest form of financing. Furthermore, they can “sell” depositors various services, so they are needed beyond the narrow issue of funding. Meanwhile “depositors” and “borrowers” are typically the same; I switched to my current bank because that was where we got our mortgage.

      This servicing of a broad client base with different products is what separates a bank from a non-bank financial institution that finances positions in assets in the wholesale funding markets.

      Delete
    5. This comment has been removed by the author.

      Delete
    6. "If a bank made me a loan with the proviso that I must spend the money only on goods and services the bank is offering, the loan would be non-fungible."

      Well, but then the bank would be granting a loan where it supplies to the customer things that are not bank deposits nor currency in exchange for a promise to be repaid with interest in the future. That is very unusual: the standard is that banks supply customers with deposits or base currency (in exchange for the promise to be repaid with interest in the future). And the demand deposits are promises (made by the bank to the customer) to give base currency to her whenever she demands it. So it all revolves around base currency.

      I believe that there is some ambiguity in the word "loan". It may refer to 1) the whole process of a party lending something to another party, hence establishing a debt relation, and the later repaying of that debt somehow, with interest; or 2) to the debt relation only, which is usually registered as an asset (credit or loan receivables) to the creditor and a liability to the debtor.

      Using definition 2, the financial resources (deposits or whatever) that a bank supplies to the customer is something separated from the loan itself.

      I would avoid using the word “fungible” because it gives too much focus on the interchangeability issue, and I don’t think this is the main issue here.

      Loans are usually not fungible with other loans (maybe unless they are loans from the same customer to the same bank and with the same characteristics), deposits are not usually fungible with other deposits (maybe unless they are deposits for the same customer from the same bank with the same characteristics), and of course deposits are not fungible with loans (because one is an asset and another is a liability). And the deposit is also not fungible with its underlying asset (base currency). So loans can be thought as candies of distinct brands, deposits can be thought as cars of distinct brands, and currency can be though as rice. It is clear that candies are not fungible with cars nor rice. And the different candies are not fungible between themselves, nor the cars.

      The main issue is that the demand deposit represents a promise that the bank makes to the customer that if she demands it, the bank will convert the deposit into something else (base currency). So the bank needs to have some base currency spared in case customers want to convert the deposit into currency (via withdrawals or transfers).

      But just my opinion here. Nonetheless, I think you got the point.

      Delete
    7. Andrè,

      In my opinion fungibility is very important. I am not speaking of fungible loans or deposits (never have), but of the fungibility of the money made available by the loan. If you don't like the term fungibility, call it (general/universal) transferability / ability to be used as a general means for settling payments.

      Making it possible for the borrower to use the money made available to her by a loan (and held in a deposit created upon making the loan) to make payments to any bank customer in the banking system, is costly for the banks and hence needs to be financed/funded. For, to be able to participate in the payment system which ensures "transferability" banks need to acquire bank reserves, which are costly.

      My point is: true, banks do not have to fund loans (i.e. ask some party (e.g. savers) to give them funds so that they have money to pass on to another party in the form of a loan), but the loans typically offered by banks make available to borrowers money that can be transferred to any payee covered by the banking system, and thus requires funding bank reserves (which cannot be created out of thin air but must be acquired for a price from depositors or the capital markets (securities) or from the central bank).

      This funding requirement is a constraint on lending like capital requirements, borrower quality, borrower demand and the need to generate a profit.

      Thus, I think, one could argue that, even though banks can create money out of thin air, loans made by them still need to be funded. After all, one attribute of the sold product (the loan) - its fungibility - needs to be funded. A loan that is non-fungible will typically not be acceptable, so loan making / lending will always require funding.

      For a complete picture of why fiat money producing banks require deposits, I need to put together and unfold these aspects: (1) deposits needed to fund the fungibility of money created by making loans, (2) André's points made in his first reply to my initial question and (3) Brian's hint that banks need to finance their balance sheet.

      Delete
    8. In the real world, a viable bank doesn’t really think about the funding implications of making a loan. Yes, it is highly likely that an outflow will be triggered. But at the same time, it is getting inflows from other banks that have extended loans, plus inflows/outflows that result from capital market transactions.

      If a bank is worrying about its liquidity position, it is in the process of transitioning into a deceased bank.

      What really matters is whether the loan will default (or whether they can get it off their books into a securitisation) before said default.

      This is because banks are not allowed to make individual loans that are a huge percentage of their capital. A single loan causing a withdrawal will be a mosquito bite on a bank’s liquidity position.

      They are looking at the growth of their total loan book. If they are growing in line with their peers, inflows/outflows from lending proceeds being spent will balance out.

      However, there is another effect that most banking system primers ignore. Deposits migrate towards capital market instruments. Individual banks are forced to issue capital market instruments to compensate, and so we get an aggregate movement from deposits to bank-issued funding instruments. In the grand scheme of things, those transactions are probably more important for liquidity management than just loan book growth,

      Delete
    9. Brian,

      Is your above comment a reply to what I have written here?

      "This funding requirement is a constraint on lending like capital requirements, borrower quality, borrower demand and the need to generate a profit."

      Are you saying that in principle this statement is true, but in reality funding is rather a minor point compared to the other four.

      Why do banks compete for deposits? Not primarily to fund bank reserves, but to refinance their balance sheet?

      Why then do the few (among those in the tradition of MMT) who say something about the need to attract deposits, seem to exclusively highlight bank reserve funding? Compare here: https://www.resilience.org/stories/2014-10-28/why-do-banks-want-our-deposits-hint-it-s-not-to-make-loans/

      Delete
    10. I have been somewhat tied up, and did not have time to go through your discussion. I just published a new article, and it outlines my thinking. I think it’s related to what you are discussing, but written as a primer (that I can re-use...) that starts from scratch.

      Deposits are cheap funding, so they are attractive. My argument is that the outflow drain from lending is blown out of proportion. The drain via the shift to non-bank finance is what really shows up.

      Delete
    11. Brian,

      “In the real world, a viable bank doesn’t really think about the funding implications of making a loan”

      That is a popular view in the MMT community. I think I read many times this sort of claim from Randall Wray, Warren Mosler and Bill Mitchell.

      I have been working in financial and ALM (Asset and Liability Management) teams in banks and financial institutions for years now, and I learned that the funding is one of the core concerns of those teams. This is in contradiction with that kind of MMT message.

      The reason why a credit area or a credit officer doesn’t need to worry about cash when discussing a loan with a customer is because the finance and ALM teams (and the management) have already planned for at least the following 12 months and already could find sources of funding and/or calibrated the amount of loan origination, sometimes even slowing down or pausing some credit products or lines, if funding sources are dry.

      For small banks, 100% of the deposits created by loans are withdrawn or transferred to other banks in one or two business days. The bank necessarily needs to have currency before granting the loan. It will do that by attracting deposits from people and investors who desire to save. Usually banks rely on some regulatory or managerial KPIs (like LCR and NSFR) to control the liquidity position and guarantee that there is the required cash plus a buffer.

      And while a bank can easily invest or lend its excess liquidity into central bank repos or other public or private facilities, it is not true that a bank can easily borrow money – so there is no symmetrical relationship between lending and borrowing. It is only easy to borrow money when the bank already holds government bonds to give as collateral. Actually, most government bonds are already considered liquidity to banks, exactly because they can easily be sold and turned to cash or given as collateral. If the liquidity position (cash + gov bonds + other liquid assets) is low, it will need to attract more deposits, lower loan origination, reduce costs, etc.

      For big banks the situation is similar, but usually it has one advantage: less than 100% of customers will withdraw or transfer the deposits right away. If the bank is big enough, there is a big probability that the customer will use the deposits to pay some other customer, who will keep it deposit as savings for some time. So, in those cases, indeed the big bank doesn’t need to have currency before lending. But this sort of thing doesn’t happen to every single customer. There will still be many who will withdraw, transfer the money or pay taxes.

      Delete
    12. I.G.T.U.,

      "If you don't like the term fungibility, call it (general/universal) transferability / ability to be used as a general means for settling payments."

      Yes, in my opinion this is much better. Fungible for me means ability of a unit of something to be interchangeable with each other unit, because they are indistinguishable on from the other, like one grain of rice with another grain of rice.

      The ability of being used as general means for settling payments is, in my opinion, an entire distinct concept.

      Base currency holds this property mostly because this is what the biggest agent in the market (the government) accepts for settlement of obligations and claims (like taxes and fees) and pays in exchange for the provision of services.

      But nonetheless of this “fungibility” wording discussion, I believe your conclusion is right (like when you claim that “My point is: true, banks do not have to fund loans (i.e. ask some party (e.g. savers) to give them funds so that they have money to pass on to another party in the form of a loan), but the loans typically offered by banks make available to borrowers money that can be transferred to any payee covered by the banking system, and thus requires funding bank reserves (which cannot be created out of thin air but must be acquired for a price from depositors or the capital markets (securities) or from the central bank).”)

      Delete
    13. Andrè,

      (1) Thank you for your most instructive comments.

      (2) You are making an intriguing point in writing:

      "That is a popular view in the MMT community. I think I read many times this sort of claim from Randall Wray, Warren Mosler and Bill Mitchell.

      "I have been working in financial and ALM (Asset and Liability Management) teams in banks and financial institutions for years now, and I learned that the funding is one of the core concerns of those teams. This is in contradiction with that kind of MMT message."

      (3) Incidentally, the Swiss economist Mathias Binswanger cites in his book "Geld aus dem Nichts" ("Money from Nothing") the case of UBS bank suffering during the Great Financial Crisis a massive drain of depsosts to Raiffeisenbank, seriously threatening UBS's profitability, as more and more bank reserves had to be acquired to effect the transferrals.

      Also, I wonder, does the preference for lock step loan book growth among banks contribute to bank-driven credit cycles, which surely come at a substantial cost?

      Delete
    14. André, apologies for the wrong accent.

      And Brian, thank you for your comments, too.

      Delete
  10. I can see that you are a specialist at your field! I am propelling a site soon, and your data will be exceptionally valuable for me.. A debt of gratitude is in order for all your help and wishing all of you the accomplishment in your business. Ibotta Alternatives to save money

    ReplyDelete
  11. You’ve got some interesting points in this article. I would have never considered any of these if I didn’t come across this. Thanks!. air duct cleaning

    ReplyDelete

Note: Posts may be moderated, and there may be a considerable delay before they appear.

Although I welcome people who disagree with me, please be civil.

Please note that my spam comment filter appears to dislike long "anonymous" posts.

Note: if you want to post comments from Apple devices (iPhone, iPad), you apparently need to turn off "prevent cross-site tracking" in Safari privacy settings. (The reason presumably is that another URL handles comments, and so the user session needs to be preserved when redirected to that site. I don't like this, but this is not enough to make me switch my hosting service.)