The Broad Money Supply is ALWAYS Endogenous

Monetary Realism starts with a simple understanding – money matters within the monetary and the dominant form of money in today’s monetary system is bank deposits.  Bank deposits are created when banks make loans.  And banks make loans when creditworthy customers have demand for loans.  So the USA has a money supply that is “endogenous” and elastic.  That is, the money supply is determined by the amount of new lending that is done and it’s elastic in that it can expand and contract (repayment of loans destroys deposits).*

The central bank exists primarily to ensure that the payments system in the monetary system is stable.  Because we have private for profit banking there’s an inherent instability in the banking system.  That is, there are times during the business cycle when banks might ease lending standards and issue more loans than their customers can actually pay back.  This can be potentially destabilizing and there are few things worse for the economy than a payments system (which is run by the banks) being unworkable.  So the Fed tries to ensure an orderly payments system before all else.

The reason I bring all this up is because I noticed Scott Sumner making a relatively basic banking error in a recent post and I think it’s important to get this stuff right if we’re going to actually understand the monetary system and how various policies might impact the monetary system.  He said:

“Individual banks are not constrained in making loans in the short run, as they can always borrow needed reserves in the fed funds market.  If they do so that will put upward pressure on interest rates, and the Fed will supply the needed reserves to maintain their fed funds target.

In the long run banks are constrained, as the Fed will adjust the monetary base to prevent economic overheating.  The endogenous money folks, who are right about the period between Fed meetings, overlook this longer run problem with their theory.  Six weeks is not a long enough period to have major macroeconomic consequences.  But in the very short run the banks are not constrained by a lack of reserves, if the Fed is targeting the short term interest rate.  The base is endogenous during that period.”

The first paragraph is correct.  The money multiplier is a myth.  The Fed will always supply reserves to the banking system if the banking system as a whole cannot meet its reserve requirements.  But the second paragraph gets things wrong.  I won’t quibble over the fact that the “long-run” is just a series of “short-runs”, but I do have an issue with the rest.  Adjusting the monetary base will not necessarily have any impact on the amount of loans the banking system can make.

Let’s say the Fed started to reduce QE tomorrow.  This would not mean banks can make fewer loans.  Banks make loans and find reserves later if they must.  But the banking system is awash in excess reserves so finding reserves to cover their requirements is not necessary in today’s environment.  The Fed could reduce the monetary base by several trillion dollars before it runs into a level where it would then NECESSARILY supply reserves to the banks if they needed to meet reserve requirements.  But even at this point there would be no “Fed choice” in the matter.  If it wants to maintain an orderly payments system the Fed MUST supply the reserves necessary for banks to settle payments and meet reserve requirements.  But that concept is largely inapplicable to a system that has a $3T+ monetary base and excess reserves through the nose….

Now, all of this might influence the spread at which banks make loans (in some cases of QE it might even impact bank capital which could impact lending), but that’s secondary and doesn’t make the broad money supply any less endogenous.  It just means the Fed can very indirectly impact the lending capabilities of the broader banking system.  But the monetary base does not directly determine the amount of lending the banks can do….

Lastly, it’s important to note that the thrust of Sumner’s post is basically correct.  That is, the Fed’s exit strategy is rather simple from here.  If it needs to raise interest rates while maintaining its current balance sheet it will simply raise the interest rate on excess reserves, which is today serving as a de-facto Fed Funds Rate.

* QE via a non-bank can also increase deposit levels, but this is not relevant to this discussion.  

** Some people might claim cash is a portion of the monetary base that is exogenous.  But this is false.  Cash is supplied to the banking system by the Treasury (not the Fed) in order to meet demand by bank customers for cash needs.  

*** Someone seems to be commenting on Scott Sumner’s site using my name.  I NEVER comment on other websites (MR or PC) so if you see my name then it’s not actually me.  Though this person appears to be quoting me via Twitter so I can’t say it’s not me at all.  

**** Endogenous money means the money supply is mostly created endogenously as credit.  This means that private banks are the primary issuers of money and do so based on the demand from creditworthy customers.  So the central bank has far less control over the money supply than one might presume if they learned the money multiplier theory.  This is the central point in understanding endogenous money.  

Cullen Roche

Cullen Roche

Mr. Roche is the Founder of Orcam Financial Group, LLC. Orcam is a financial services firm offering research, private advisory, institutional consulting and educational services. He is also the author of Pragmatic Capitalism: What Every Investor Needs to Understand About Money and Finance and Understanding the Modern Monetary System.

More Posts - Website

Follow Me:
TwitterLinkedIn

Comments

  1. Cullen, I think it might be cleaner to define endogenous money as the view that the money supply is determined independent of output and prices. In other words, the Fed cannot steer the economy by altering the monetary base or other monetary variables.

  2. LVG, you said “the money supply is determined independent of output”
    What about if you put said money into productive vs non productive action?

  3. If you don’t know how much is leaking into my offshore companies, trusts and accounts; and you don’t know how much I’m circulating around in various pools of luscious investment heavens, and various political panderings then you don’t know how much money you have, or at least how much you have left …for now.

  4. Cullen, thanks for doing a post on this. BTW, I wanted to let you know that I did get a response from Scott to my question. I already posted that in the “Ask Cullen” section, so here’s a link to the question and answer there:

    http://pragcap.com/forum?mingleforumaction=postreply&thread=122&quote=435.0

    I’ll reprint Scott’s answer here:

    “Tom, You said;

    “What exactly do you mean by “monetary base” and is it different than what you mean by “reserves?””

    The base is simple cash plus reserves. No, I would not modify my statement if the base was 100% reserves.

    BTW, prior to 2008 the base was more than 95% cash, reserves were trivial.

    I suppose you could say that the base supports a series of ffr targets, but that’s not very illuminating. It’s better to think of the ffr targets adjusting in such a way that the base adjusts in such a way as to target the price level.”

    OK, I’m glad he answered me, but now what does that mean. :(

    Specifically this bit:

    “It’s better to think of the ffr targets adjusting in such a way that the base adjusts in such a way as to target the price level.”

    I’m following him up until “…to target the price level.” Cullen, is this where you’d have a fundamental disagreement?

  5. If required reserves were in excess of system-wide reserves, the Fed could theoretically only supply the additional reserves through the discount window and at a punitive rate (although I guess they have to raise the discount rate to do so). I get that the Fed will generally always provide the required reserves so that payments clear, but they don’t technically *have* to. Could they not just sit back and let banks scramble to call in a few loans and shrink requirements or am I missing something mechanically there?

  6. I know of a lot of broads with an exogenous money supply. That was the topic right?

  7. I think the Fed could also supply reserves via “repos” or “reverse repos” (I always get confused by which) or through OMOs (specifically OMPs. Actually I’m not even sure… are the repos and reverse repos classified as OMOs? Anyway, I’m pretty sure that the discount window is not the only alternative.

  8. Excellent post. There is no question that the broad money supply is endogenous (or endogenius as I like to call it :)) It is amazing that there is any debate about it.

    But Cullen, it appears that you are arguing that even the narrower money supply (or the base) is also at least partly endogenous. Cash certainly is, but even reserves are somewhat endogenous, at least in a non-QE world. Loans create deposits, which create reserves (if needed)?

  9. I think he means that in a world with reserve requirements the Fed has to supply the level of reserves that allow the banks to meet their reserve requirements at a minimum. There’s no choice in the matter for the Fed.

  10. I asked Scott to elaborate on that “price level” bit, and here’s what he wrote:

    “Tom, Suppose prices are rising too fast. Prior to 2008 the Fed would reduce the monetary base to slow the inflation (more precisely, they’d signal that they intend to reduce the growth rate of the monetary base over time–there is no instantaneous change.) This is done by raising the ffr target. In the future they might also raise the IOR.”

    This just sounds like standard inflation targeting to me (in his “prior to 2008″ scenario).

    I guess his emphasis on statements like this (in his original post):

    “In the long run banks are constrained, as the Fed will adjust the monetary base to prevent economic overheating.”

    ..just seem odd. I don’t understand why he focuses on the “monetary base” (which seems more like a dependent rather than an independent variable here) but if all he’s really talking about is standard inflation targeting (again prior to 2008) via adjusting the FFR every six weeks, that’s not very controversial is it?

  11. sumner response: http://www.themoneyillusion.com/?p=21786#comment-254717:

    Cullen, Prior to 2008 the Fed did monetary policy by adjusting the supply of the medium of account. Now they adjust both the supply and the demand (QE and IOR).

    Colin, I’m afraid that post gets it wrong, but then almost everyone does. I’d suggest looking at this post if you want to understand money multipliers;

    http://www.themoneyillusion.com/?p=21463

    There are all sorts of factors that play into the effect of changes in the base. Is it exogenous, or responding to a change in base demand? Is it temporary or permanent? Are you looking at the effects on nominal or real bank lending? Is it accompanied by a change in IOR? Are you at full employment? Etc, etc. Most observers, including that post, just fixate on one tiny part of the big picture.

    Tom, Suppose prices are rising too fast. Prior to 2008 the Fed would reduce the monetary base to slow the inflation (more precisely, they’d signal that they intend to reduce the growth rate of the monetary base over time–there is no instantaneous change.) This is done by raising the ffr target. In the future they might also raise the IOR.

  12. Thanks. It sure sounds like reserves are endogenous, at least for the most part!

  13. Scott’s response to this post, fyi:

    I’m afraid that post gets it wrong, but then almost everyone does. I’d suggest looking at this post if you want to understand money multipliers;

    http://www.themoneyillusion.com/?p=21463

    There are all sorts of factors that play into the effect of changes in the base. Is it exogenous, or responding to a change in base demand? Is it temporary or permanent? Are you looking at the effects on nominal or real bank lending? Is it accompanied by a change in IOR? Are you at full employment? Etc, etc. Most observers, including that post, just fixate on one tiny part of the big picture.

  14. Sumner is wrong. Nobody at the Fed ascribes any importance to the base, their tool is the FFR, they adjust that to target prices. There is no point in having 2 models: FFR->base and base->output/prices if one can have one model FFR->output/prices.. The part base->output/prices won’t ever work anyway, the causality works the other way, so pushing the base up does nothing as QEs showed in Sweden, Japan, US (curiously most coincided with lowered inflation if anything) and the Fed has no option to set the base below what is demanded, which would cause a blowup of the payment system and put every political appointee at the Fed out of work.

    http://www.mnb.hu/Root/Dokumentumtar/ENMNB/Kiadvanyok/mnben_mnbszemle/mnben_szemle_cikkei/bulletin_2007june_komaromi.pdf

    Sumner responds in a typical way: glib and not engaging. When pressed he would ask that we discuss an economy without banks, I have seen it many times, he can’t discuss the actual monetary system, out of his depth. See here, after SS asked to discuss “monetary policy in a country without loans or banking” Mosler mercifully bowed out http://www.themoneyillusion.com/?p=10178

    Here is Sumner’s discussion with Fullwiler, again he can’t discuss banks: “We aren’t going to get anywhere if we keep talking about interest rates and RRs and bank reserves. ” http://neweconomicperspectives.org/2011/07/scott-sumner-agrees-that-mmt-policy.html

  15. He didn’t actually answer anything. He just says the money supply will be determined in the long run by the central bank? That doesn’t change anything at all about what Cullen said.

  16. I think the MMers are saying that the policy approach is exogenous. And I think they’re right to some degree. The Fed can add to the supply of deposits by swapping t-bonds for bank deposits. Or the Fed could actually buy alternative things like cars and directly add to the money supply. I think Cullen has noted that that’s the Fed doing fiscal though. But I see where they’re coming from when I might not have a few years ago.

  17. Does this really depend on what we define as “money”? Money seems to be different things to different schools of thought so it might just be differences in definitions. Cullen says the supply of deposits is determined by the public’s willingness to hold deposits. MM says the Fed could add deposits if it wanted. I don’t think either view is necessarily right or wrong.

  18. It seems to me that the supply of bank deposits is determined by a lot of factors, including the public’s desire to borrow money (with which to make purchases), the public’s desire to sell Tsy debt to the Fed, the interest rates, The government’s spending habits and tax policy, etc.

  19. In reading those old Sumner posts I detect a slight change on Sumner’s part on this in comparison with his current tone. Whereas before he was more like “Common guys! Can’t we just pretend banks don’t exist for a while and the Fed injects reserves straight into the hands of Tsy debt holders so I can get my bearings here??” … now he sounds like “Well the ‘endogenous guys’ ARE correct… but only on six-week intervals. Over these intervals blah blah happens [insert your favorite 'exogenous guy' explanation] but over longer time frames they’re still wrong”

    I don’t know… that actually sounds like a little progress to me, don’t you think?

  20. Money is “different things to different schools of thought”.
    Your a 20yo oil field worker who just got a piece of paper (pay check) granting you X amount of other paper (green backs) you can swap for topless lap dancer minutes for which she takes to the landlord the next day to pay her rent which the landlord pay to the mtg holder which pays the mtg servicer which pays the mtg back to the lender which blah blah ect ect………………………

  21. Can someone help clarify this for me…

    Assuming the current environment where the banking system is awah in Reserves, Let’s say there’s an individual bank. If I show up and deposit $1000

    1) Does that increase the banks Reserves by $1000?
    and
    2) Exactly what can the bank “do” with the money I just deposited? What are it’s options? Assuming no additional loans are made, does the extra $1000 just earn interest for the bank (in the current environment)? Can the bank use that money I deposited to purchase other types of financial assets?

    Thanks.

  22. Endogeneity & MMers: My conclusions based on the discussions here and other discussions elsewhere w/ Rowe, Sumner, and Glasner: Let’s look at an example situation:

    Imagine that we’re not at the ZLB (pre-2008), and that the Fed targets the federal funds rate (FFR), adjusting the target every few weeks to hit a longer term inflation target. Now lets say we find a way to deliver real-time accurate economic statistics to the Fed, and we replace the Fed chairman and board w/ an algorithm that uses that real-time data to produce a real-time continuously varying FFR and the OMOs to track that FFR, but with the overall purpose again being to track the target inflation rate. So our overall system (control law followed by the rest of the economy) prior to “closing the loop” looks like:

    Inputs:
    1. target inflation rate
    2. actual inflation rate
    3. a bunch of other inputs…

    Outputs:
    1. actual inflation rate
    2. a bunch of other outputs…

    Right after the input we’ve got our Fed computer implementing the control law, and downstream from that the rest of the economy, which in turn produces the outputs listed above.

    The first thing our control law does is subtract the actual inflation rate from the target inflation rate to produce an inflation rate error:

    inflation rate error = target inflation rate – actual inflation rate

    The control law may use as inputs other variables fed back from the economy’s output, but it uses this error signal for sure.

    Now internal dependent variables in the system include:

    1. FFR
    2. “money supply”
    3. net new borrowing
    4. etc.

    The FFR is determined directly by our control law running in the Fed computer, but some of those other dependent variables are further downstream, including “net new borrowing” which is dependent in part on the FFR but also the “desire of each private entity to borrow” which in turn is dependent on other things, perhaps including some outside independent input variables covered by item 3 above in my top list:

    3. a bunch of other inputs…

    What’s my point? I think that’s all the MMers are really saying. I also think they perhaps ignore that crucial set of unnamed independent variables which surely influence the desire to borrow, which in turn clearly helps determine the net new borrowing which in turn helps determine the “supply of money.” To the MMers, the desire to borrow is wholly determined by other variables inside the system, notably the FFR. So for them, once you close the feedback loop, this desire to borrow is wholly a dependent variable. And since the loop is closed it’s just as correct to say that the “money supply” determines the “desire to borrow” as it is to say that “the desire to borrow” determines the “money supply.” I.e. after closing the loop there’s no longer any downstream vs upstream variables, except for those inputs which remain wholly independent (e.g. the target inflation rate) and those outputs which are not fed back.

  23. BTW, in thinking about this more, it’s very difficult to draw a clean line around “the economy” to separate it from the rest of the universe. Sunspots, the tide (moon), and meteorites can have an effect, but are good candidates for independent variables — same goes for earthquakes, tsunamis, etc. But what about the weather?… humans can have some effect there. How about humans and their brains and their work ethics and powers of innovation? Some parts of humans are making purchasing and selling and saving and borrowing decisions… so should they be independent or dependent? Clearly those decisions are at least in part made after considering the state of the rest of the economy thus arguing for “dependence.” However, some parts of us humans should perhaps remain independent variables…